Path:
Periodical volume

Full text: Annual Report Issue 2008

BESSY II Annual Report 2008

Published by: Helmholtz-Zentrum Berlin für Materialien und Energie GmbH Hahn-Meitner-Platz 1 14109 Berlin Tel. +49 30 6392 2999 Fax +49 30 6392 2990 www.helmholtz-berlin.de Edited by: Dr. Kai Godehusen Dr. Markus Sauerborn

Vorsitzender des Aufsichtsrates: Prof. Dr.Dr. h.c. mult. Joachim Treusch Stellvertretende Vorsitzende: Dr. Beatrix Vierkorn-Rudolph

Geschäftsführer: Prof. Dr. Anke Rita Kaysser-Pyzalla, Prof. Dr. Dr. h.c. Wolfgang Eberhardt, Dr. Ulrich Breuer

Sitz der Gesellschaft: Berlin: Handelsregister: AG Charlottenburg, 89 HRB 5583

2

Transitions – BESSY has been in a major “phase transition” during the last 24 months. By merging with the former Hahn-Meitner-Institut Berlin, we became the new “Helmholtz Zentrum Berlin für Materialien und Energie”, which is a member of the Helmholtz Association, Germany’s largest research alliance. The new organization will be crucial for keeping the development of the Synchrotron radiation source BESSY II on track and for opening new opportunities for a next generation user facility as we are now on the roadmap to build an Energy Recovery Linac Prototype (BERLinPro). Joining the synchrotron radiation source BESSY-II and the neutron source BER-II within one research center we will be able to strengthen our user support. Additionally we have the opportunity to support the complementary use of photons and neutrons. The transition of BESSY and the re-organization of the new centre is also the reason for the delayed appearance of the last “BESSY Annual Report”. Unfortunately, there will be no BESSY Highlights 2008. Please accept our apologies, especially if you have written a contribution in anticipation of a Highlights brochure. Finally, we would like to thank our users and our staff for their ongoing enthusiasm and their patience during this transition. Sincerely

Wolfgang Eberhardt Scientific Director (Energy) Former Scientific Director BESSY

Anke Rita Kaysser-Pyzalla Scientific Director (Science with Neutrons and Photons, Large Scale Facilities) Chief Executive

3

4

Characterization of new EUV stable silicon photodiodes
F. Scholze, C. Laubis Physikalisch-Technische Bundesanstalt, Abbestraße 2-12, 10587 Berlin, Germany F. Sarubbi, L. K. Nanver, S.N. Nihtianov DIMES, TU Delft, P.O.box 5053, 2600 GB, The Netherlands Development of EUV lithography equipment has triggered a growing interest in EUV radiation detection. Several types of sensors are needed for evaluating and optimizing the imaging performance. The radiation-sensitive surface of the sensors is exposed to high photon flux doses and is affected by hydrocarbons contamination. Consequently, quite aggressive cleaning is required. Therefore, ruggedness to high photon flux and aggressive environments is a key feature of these EUV sensors alongside extreme requirements for stability, reliability, high and spatially uniform responsivity, large dynamic range, and low noise, i.e. low dark current. To meet these requirements, a new EUV photodiode technology is presently being developed and optimized for the requirements of EUV lithography systems at TU Delft in cooperation with ASML. PTB has long experience in the characterization of detectors using synchrotron radiation1,2. We characterized the EUV performance of p+ n photodiodes fabricated by using a novel doping technology3 regarding their spatial homogeneity, spectral responsivity and high-dose irradiation stability. With respect to the combination of high spectral responsivity and irradiation stability, they are already in the present state of development superior to other commercially available detectors.

Figure 1 HRTEM image of a B-layer formed after a 2.5 min B2H6 exposure at 700 ºC. The sample has been covered with PVD a-Si for TEM analysis.

Figure 2 Image and cross-section of a borondoped EUV photodiode.

For EUV diodes, the stability of responsivity under irradiation is known to be an issue4. Figure 3 and Figure 4 show the result of irradiation testing with a rather high dose 200 kJ/cm². No significant change in responsivity is observed for the boron-doped photodiode (Figure 3), while the state-of-the-art EUV photodiode suffered some responsivity loss, which might, however, partly be due to a higher susceptibility of the surface coating of this diode to carbon contamination, (Figure 4). Regarding the spectral responsivity, the boron-doped photodiode are as good as the best state-of-the-art EUV photodiodes, see Figure 5. It must be noted, that the diodes with nitrided oxide passivation shown for reference are not irradiation resistant while the more stable diodes as shown in Figure 4 have a lower responsivity.

5

Figure 3 Responsivity of a boron-doped photodiode after irradiation with 200 kJ/cm² normalized to the initial responsivity. The circles represent a diode actually exposed to the radiation while the triangles are a witness diode which was placed under the same vacuum and atmosphere conditions but was shadowed during exposure.

Figure 4 Responsivity of a state-of-the-art EUV photodiode after irradiation with 200 kJ/cm² normalized to the initial responsivity. Symbols same as in Figure 3. The solid line represents the effect of 2.5 nm carbon contamination.

Figure 5 Spectral responsivity of boron-doped silicon diodes. The highest responsivity is for a diode with no additional top layer. The lower responsivity values are for top layers of 15 nm, 30 nm, and 50 nm AlN. The dashed black line shows the theoretical limit5 for silicon responsivity of 1/3.66 A/W and the solid black line is the best state-of-the-art silicon detector with a very thin nitrided oxide passivation.

These results prove that the technology for the production of planar diffused silicon p-n diodes for EUV detection is capable of achieving the same nearly ideal responsivity as the best state-of-the-art detectors. Due to their almost ideal efficiency, the diodes also show good spatial homogeneity. A particular challenge for EUV photo sensors is the stability under irradiation. Here also the benchmark of the best state-of-the-art detectors is already met. Summarizing, the development of pure boron-doped photodiodes is proven to be a promising approach toward the improvement of EUV photo sensor performance.
References
1

R. Klein, C. Laubis, R. Müller, F. Scholze, G. Ulm, The EUV metrology program of PTB, Microelectronic Engineering 83 707 (2006) 2 F. Scholze, R. Klein, R. Müller, Characterization of detectors for extreme UV radiation, Metrologia 43 S6 (2006) 3 F. Sarubbi, L. K. Nanver, and T. L. M. Scholtes, CVD delta-doped boron surface layers for ultra-shallow junction formation, ECS Transactions 3, 3544 (2006) 4 F. Scholze, R. Klein, T. Bock, Irradiation Stability of Silicon Photodiodes for Extreme-Ultraviolet Radiation Appl. Opt. 42 5621 (2003) 5 F. Scholze, H. Rabus, and G. Ulm, Mean energy required to produce an electron-hole pair in silicon for photons of energies between 50 and 1500 eV, J. Appl. Phys. 84 2926 (1998)

6

Pilot comparison for spectral responsivity in the spectral range 11 nm to 20 nm
Frank Scholze, Gerhard Ulm Physikalisch-Technische Bundesanstalt, Abbestraße 2-12, 10587 Berlin, Germany Terubumi Saito NMIJ, 1-1-1 Umezono, Tsukuba-Shi 305-8563 Ibaraki, Japan Robert Vest NIST, 100 Bureau Drive, Stop 8411, Gaithersburg, MD 20899-8411, USA Development of EUV lithography equipment has triggered a growing interest in EUV radiation detection. To improve the metrological environment for emerging technologies using short-wavelength radiation, a pilot comparison for the spectral responsivity of photo diodes in the 11.5 nm to 20 nm spectral range was started between NMIJ, NIST and PTB. The comparison was carried out through the calibration of a group of transfer standard detectors. These detectors have been shown to have reasonable stability to be used to transfer a spectral responsivity scale maintained in a participating laboratory to that of PTB, acting as the pilot laboratory. It incorporated the comparison of different primary detector standards, ionization chamber and electrical substitution radiometer (ESR). Silicon photodiodes with different front passivation layers were used (see Figure 1); AXUV diodes having almost ideal responsivity (see Figure 2) but being sensitive to radiation damage, and SXUV diodes with an irradiation stable metal-silicide passivation. A total of six diodes, three AXUV and three SXUV, have initially been calibrated at PTB, sent to NIST for measurements, re-calibrated at PTB, sent to NMIJ for measurements and were finally re-calibrated at PTB. The re-calibrations at PTB revealed that both types of diodes were sufficiently stable. All measurements are finalized and the partners exchanged their uncertainty budgets and agreed on the final calibration results for each as collected by the pilot laboratory.
PTB Primary detector Radiation source Band width Beam divergence Beam spot size Typical radiant power /nm /mrad /mm /µW 0.025 1. 2 by 2 0.2 ESR NIST ESR NMIJ Ionization chamber

monochromatized SR 0.07 to 0.4 1.2 2.7 by 3 0.01 0.3 11 3 by 3 2.5

Table 1

Main parameters of the experimental stations used by the pilot comparison participants.

The uncertainty for the comparison (see Figure 3) is determined by the measurement uncertainty of the participating laboratories and the uncertainty attributed to the transfer detectors, due to their limited homogeneity and stability. The measurement uncertainty of the

7

participating laboratories can be separated into two major contributions, arising from the primary standard detector which is an ESR in the case of PTB1 and NIST2 and an ionization chamber for NMIJ3, and uncertainties attributed to monochromator and beamline (see Table 1). For the ionization chamber, the primary standard detector dominates the uncertainty budget, while for the ESR with its intrinsically low measurement uncertainty the total uncertainty is dominated by the contributions from monochromator and beamline. For this comparison, the uncertainty attributed to the transfer detectors is of the same order as the measurement uncertainty of PTB and NIST using an ESR. Thus further success in EUV radiometry also requires advanced detectors.

Figure 1 Spectral responsivity of the AXUV (open circles) and SXUV (closed circles) type diodes. The lines are only to guide the eye.

Figure 2 Responsivity of an AXUV diode from 0.827 nm (1500 eV) to 20 nm (red circles). The green dashed line shows the transmittance of 7.5 nm SiO2 and solid and dashed blue lines the responsivity for different levels of incomplete charge collection. Figure 3 Compilation of the relative measurement uncertainties (k=1) for the spectral responsivity; AXUV (circles) and SXUV (triangles). Data of PTB, NIST, and NMIJ are shown in blue, red, and green, respectively. The black symbols show the uncertainty resulting from the transfer detectors, mainly due to their limited homogeneity and stability.

References
1

Rabus H., Persch V., and Ulm G., “Synchrotron-Radiation Operated Cryogenic Electrical-Substitution Radiometer as HighAccuracy Primary Detector Standard in the Ultraviolet, Vacuum Ultraviolet and Soft X-ray Spectral Ranges”, Appl. Opt. 36, 5421-5440, 1997 2 Vest R. E., Barad Y., Furst M. L., Grantham S., Tarrio C., and Shaw P. S., “NIST VUV metrology programs to support space-based research”, Advances in Space Research 37, 283-296, 2006. 3 Saito T. and Onuki H., “Detector calibration in the 10-60 nm spectral range at the Electrotechnical Laboratory”, J. Optics 24, 23-30, 1993

8

UV and VUV Radiometry at PTB’s Metrology Light Source
A. Gottwald, R. Fliegauf, U. Kroth, W. Paustian, M. Richter, H. Schöppe, R. Thornagel, and G. Ulm Physikalisch-Technische Bundesanstalt, Abbestr. 2-12, 10587 Berlin In the UV and VUV spectral ranges, the main work of PTB with synchrotron radiation can be divided into (a) the calibration of radiation sources within the framework of source-based radiometry using an electron storage ring as a primary source standard of calculable synchrotron radiation, (b) the calibration of photodetectors with the aid of cryogenic radiometers as primary detector standards, and (c) reflectometry [1]. Started at the former electron storage ring BESSY I, these activities were continued from 2000 at BESSY II. However, in 2008 started the concentration of UV and VUV radiometry at PTB’s new lowenergy storage ring MLS [2], profiting there from the almost ideal measurement conditions in the UV and VUV with a characteristic wavelength which can continuously be varied from 3.4 nm to 735 nm. Moreover, the UV and VUV measurement capabilities will considerably be extended towards polarization dependent methods and photon metrology at high radiant power. The calibration of photodetectors based on cryogenic radiometry and the characterization of optical components and materials via reflectometry in the UV and VUV has been performed at BESSY I and BESSY II by using a McPherson type normal-incidence monochromator (NIM) for the wavelength range from 40 nm to 400 nm [3]. In 2008, this beamline has been transferred from BESSY II to MLS (Figure 1), together with the cryogenic radiometer SYRES II and a reflectometer system. The gracing incidence refocusing mirror was replaced by a double reflector unit for the suppression of higher spectral orders in the range from 80 nm to 120 nm. The beamline output is generally by 50 % to 100 % higher compared to BESSY II (Figure 2) when operated with the same stored electron current.

Fig. 1. Synchrotron radiation of PTB’s
Metrology Light Source passing the exit slit of the normal-incidence mochromator beamline for UV and VUV detector calibration and reflectometry.

Fig. 2. Radiant power into the respective bandpass
available at the normal-incidence mochromator beamline for UV and VUV detector calibration and reflectometry for open monochromator slits (2 mm) and a storage ring current of 100 mA at MLS (solid curves, electron energy: 600 MeV) and BESSY II (dashed curves, electron energy: 1,700 MeV).

9

For the calibration of radiation sources at MLS, a new spectrometer is under construction which combines a Seya-Namioka type NIM with a toroidal grating monochromator (TGM) under gracing incidence (Figure 3). The spectrometer covers the wavelength range from 7 nm to 400 nm and will be put into operation at MLS in 2010. In a first step, spectral radiance and spectral radiant intensity will be the radiometric quantities of a secondary source standard that can be determined by comparison with the calculable flux of MLS. In a second step, also spectral irradiance calibrations are scheduled to be realized. Within a straight section of MLS, an undulator with a periodic length of 180 mm (U180) [4] will be put into operation at the end of 2008. It provides radiation in the wavelength range from 4 nm to 20 µm. In the UV and VUV range, the radiant power available will be about two orders of magnitude higher compared to ordinary bending magnet synchrotron radiation. At the exit of a plane grating monochromtor (PGM) currently under construction, an output of 50 µW within a spectral bandpass of 0.1 % of the wavelength is expected. This PGM, again, combines a normal incidence with a gracing incidence branch and covers the spectral range from approx. 4 nm to 400 nm. It is scheduled to start operation in 2010. Since undulator radiation is also 100 % linearly polarized, the PGM undulator beamline will enable UV and VUV radiometry to be extended to high flux and polarization dependent measurements.

Fig. 3. Scheme of the
spectrometer for the calibration of radiation sources in the UV and VUV at MLS.

REFERENCES [1] M. Richter, J. Hollandt, U. Kroth, W. Paustian, H. Rabus, R. Thornagel, G. Ulm, Metrologia 40, S107-S110 (2003) [2] G. Brandt, J. Eden, R. Fliegauf, A. Gottwald, A. Hoehl, R. Klein, R. Müller, M. Richter, F. Scholze, R. Thornagel, G. Ulm, K. Bürkmann, J. Rahn, G. Wüstefeld, Nucl. Instr. and Meth. B 258, 445-452 (2007) [3] M. Richter, J. Hollandt, U. Kroth, W. Paustian, H. Rabus, R. Thornagel, G. Ulm, Nucl. Instr. and Meth. A 467-468, 605-608 (2001) [4] R. Klein, J. Bahrdt, D. Herzog, G. Ulm, J. Synchrotron Rad. 5, 451-452 (1998)

10

Absolute determination of cross sections for resonant Raman scattering on silicon carbide
M. Müller1, M. Kolbe1, B. Beckhoff1, K. Feldrapp2, S. Storm2, A.-D. Weber2 1 Physikalisch-Technische Bundesanstalt, Abbestr. 2-12, 10587 Berlin, Germany 2 SiCrystal AG, Günther-Scharowsky-Str. 1, 91058 Erlangen, Germany Reference-free total-reflection x-ray fluorescence analysis for the quantification of surface contamination requires the accurate knowledge of all experimental values as well as of the fundamental parameters involved [1]. Besides the fundamental parameters of the contaminants also the parameters of the substrate affect the result. To reduce the impact of tabulated data with unknown or estimated relative uncertainties the resonant Raman scattering of X-rays in the vicinity of the K absorption edge of silicon carbide (SiC) has been studied. The investigation was carried out at the plane grating monochromator beamline for undulator radiation of the PTB laboratory at BESSY II in Berlin. SiC is a wide-band-gap semiconductor and offers outstanding material properties for high-power electronics and optoelectronic applications. Non-destructive analytical methods like reference-free TXRF are necessary for process control. For the investigation a SiC wafer was thinned to a thickness of about 10 µm to allow for transmission measurements. We determined absolute cross sections for the energy range below the silicon K absorption edge employing calibrated instrumentation.

Theory
The resonant Raman scattering is an inelastic scattering process, which exhibits a strong resonant behavior as the energy of the incident radiation approaches from below the absorption edge of an element [2]. The KL-RRS proceeds through: • the intermediate state, where a virtual hole is created in the K-shell and the corresponding electron is transferred to an unoccupied state either in the continuum or in a bound excited state, the final state, where an electron from one of the L subshells fills the hole and a photon is emitted.

•

Experimental Set-up
For this experiment we employed monochromatized undulator radiation provided by the plane grating monochromator beamline in the PTB laboratory at BESSY II. The experimental setup (Fig.1) ensures that the scattered radiation can be measured in a well defined solid angle by means of a Si(Li) detector. The incident radiant power can be absolutely determined by using a calibrated photo diode placed behind the exit diaphragm, if the sample is out of the beam. The Si(Li) detector in use is also calibrated with respect to both its efficiency and its response behavior [3]. Additionally, in the same arrangement the thickness of the measured 2” SiC wafer and the corresponding absorption factors can be experimentally determined by employing transmission measurements.

11

monochromatized incident beam diaphragm ∅ = 2 mm

diaphragm ∅ = 1.46 mm calibrated Si(Li) detector scattered radiation 45° 45° wafer transmitted beam

Fig.1: photograph of the UHV chamber and a sketch of the experimental setup

X/Y table

diaphragm ∅ = 4 mm calibrated photo diodes

Data Analysis
Because a direct deconvolution of the measured spectrum with the response functions of the Si(Li) detector was not possible, we adopted a different approach to determine the resonant Raman scattering cross section with respect to the scattered photon energy: The theoretically calculated RRS spectral distribution was at first convoluted with the Si(Li) detector’s response functions, then the resulting spectra were fitted to the measured spectra [4]. All necessary parameters, such as radiant power, sample thickness, solid angle, absorption coefficients were experimentally deduced.

Results
We determined the cross sections of the resonant Raman scattering on SiC for three photon energies of the incident radiation and compared it with cross
104

C-Ka F-Ka

Fig.2: XRF spectrum of the thinned SiC wafer, excited below the Si-K edge (1622 eV). The spectrum was deconvoluted by detector response functions at the energies of fluorescence lines, continuous bremsstrahlung background and spectral distribution of the Resonant Raman Scattering in a least square optimization.

counts per channel

103

SiC-RRS O-Ka

102 Bremsstrahlung Rayleighscattering and Si-Ka 1500

101 500 1000 photon energy /eV

12

sections of pure silicon [4]. The RRS cross sections of Si and SiC are the same apart from a difference of 2 eV between the resulting values of the Si-K absorption edges. This difference is also found in the transmission measurements of both samples.

Conclusions
The RRS cross sections of SiC are not strongly influenced by chemical bonding, merely the chemical shift of the Si-K absorption edge has an impact on the cross sections. The results suggest that the cross sections for pure Si may be used for SiC as well, if the chemical shift is taken into account. For other Si compounds further investigations are necessary, e.g. Szlachetko et al.[5] reported that there is still a difference of about 20% between the RRS cross sections of Si and SiO2 after including the K edge shift. The relative uncertainties of the determined cross sections are only 7%[4]. Therefore the impact of the resonant Raman scattering background contribution on quantitative TXRF can be considered very accurately. This will improve the reliability of the analysis of Si and SiC wafers employing soft X-ray radiation.

References
[1] B. Beckhoff, R. Fliegauf, M. Kolbe, M. Müller, J. Weser, G. Ulm, Anal. Chem., 79 7873-7882 (2007) [2] Ch. Zarkadas, A.G. Karydas, M. Müller, B. Beckhoff, Spectrochim. Acta B 61, 189-195 (2006) [3] F. Scholze, B. Beckhoff, M. Kolbe, M. Krumrey, M. Müller, G. Ulm, Microchim. Acta 155, 275–278 (2006) [4] M. Müller, B. Kanngießer, B. Beckhoff, G. Ulm, Phys. Rev. A 74, 012702 (2006) [5] J. Szlachetko, J.-Cl. Dousse, M. Berset, K. Fennane, M. Szlachetko, Phys. Rev. A 75, 022512 (2007)

13

Characterization of self-assembled monolayers on germanium surfaces with GIXRF
M. Lommel, B. O. Kolbesen Institute for Inorganic and Analytical Chemistry, Goethe-Universität, Frankfurt/Main, Germany P. Hönicke, M. Kolbe, M. Müller, F. Reinhardt, B. Beckhoff, Physikalisch-Technische Bundesanstalt, Abbestraβe 2-12, 10587 Berlin, Germany

Introduction Germanium has gained interest in recent years as a promising material for high-performance CMOS applications. However, Ge does not form a robust passivation layer on its surface – germanium oxide is water-soluble [1]. The formation of self-assembled monolayers (SAMs) by suitable organic molecules with appropriate anchor groups on semiconductor Figure 1: The thiol (HS-) group is the so surfaces may be used to probe the chemical state and called “anchor group” with the sulphur quality of the surface or to achieve surface bonding to the germanium. The long chain is the motor of self-assembly. The “head” passivation. Molecules with thiol anchor groups are group, here –OCOCF3, makes a detection of able to bond to hydrogen-terminated germanium the monolayer easily possible. Distance: sulphur – fluorine: 1.8 nm surfaces (Ge-S bond). We have prepared SAMs of alkylthiols with different head groups on germanium. The germanium surface prior to and after SAMs formation has been characterized by AFM, XPS, SRTXRF and -NEXAFS. Since the surface preparation of germanium is neither well understood nor developed, the controlled preparation of an oxide-free completely H-terminated surface which is a prerequisite for SAM formation of alkylthiols turned out to be a major challenge. Several approaches have been studied [2]. Best results for H-termination of germanium have been obtained by HF treatment The immediate immersion of the HF treated germanium into a 1-mmol solution of thiol, e.g. dichloroethane, leads to self-assembly of thiol monolayers. Results and discussion The success of the formation of the monolayer on the substrate depends largely on several factors during its preparation e.g., the solvent used for the formation of the monolayer has to be free of water as well as traces of oxygen and has to dissolve the thiol of choice. Attempts were made to form monolayers with several 11-mercaptoundecyls [HS-(CH2)11-R] with differing head groups (-R). In this case the head groups were chosen to facilitate detection by SR-TXRF (Synchrotron Total reflection Xray Fluorescence), GIXRF (Grazing Incidence XRF) and NEXAFS (Near Edge X-ray Absorption Fine Structure) and hence determine the degree of coverage as well as to determine the properties of the monolayer, e.g. its height and the bonding angle. The preparation was performed in an argon atmosphere. The sample with the H-terminated surface was blown dry with N2 and immediately immersed into a 1mmol solution of the thiol in dichloroethane (DCE). DCE meets the needs of a

14

solvent that does not contain water, dissolves both hydrophilic and hydrophobic molecules, does not contain oxygen in its structure and has adequate vapour pressure. The thiol adsorbs onto the germanium surface about 10 seconds after immersion into the precursor solution but the process of self-assembly takes hours if not days to be finished [3]. To distinguish between physisorbed and therefore “horizontal” thiols and the chemisorbed “vertical” thiols we measured the coverage of sulphur on the substrate after different immersion times and performed GIXRF measurements. The incident angle dependant modulation of the X-ray standing wave field (XSW) above the sample surface allows for the determination of the amount as well as the distance in a vertical direction with respect to the sample surface of so-called marker atoms [4]. Here, the head group with fluorine was used as the marker (Fig.1). The GIXRF curves for different immersion times are shown in fig. 2a. The angle dependant XSW field was calculated with IMD [5] and convoluted with the model shown on fig. 2b and then fitted to each GIXRF curve [6, 7, 8].

Figure 2: a) GIXRF measurements on F-SAMS for different immersion times and the fitted curve for the 1d sample. b) Different repartitions, which best fitted the respective measurements.

It assumes that the molecules either lie on the germanium surface or that they are bound, resulting in an F-layer at a certain height. The bonding angle, the normalized fraction of vertical thiols and background parameters were used as fitting parameters. The background originates from clustering of the molecules and was assumed to be a constant fluorine distribution. The results shown in fig. 2 led to the conclusion that the fraction of self assembled molecules is optimal after a period of one day. It is lower if immersion times are shorter or significantly longer than one day. A constant background was necessary to fit the GIXRF curves. The fraction of fluorine in the background or rather in clusters is relatively high. This means that clustered molecules cannot be neglected. The curve obtained on the sample treated for five days could only be fitted by increasing the surface layer thickness. This indicates that this sample is covered with a double layer of the lying molecules. The height of the fluorine above the surface was determined to be 1.4 nm, which corresponds to a bonding angle of the Ge-S bond of 47.5°. Beyond this, we determined the coverage of the substrate by AFM [2] and NEXAFS [9]. However, time and concentration are obviously not the only factors to influence the degree of coverage, as the AFM results were not reproducible. Nevertheless, with these results we proved that under the given conditions it is possible to prepare SAMs on germanium.

15

Summary Treatments to achieve a controlled surface state of germanium, in particular with H-termination, prior to the preparation of alkylthiol SAMs have been tested and characterized. The immersion of HF treated germanium in a 1-mmol solution of thiol, e.g. dichloroethane, leads to self-assembly of thiol monolayers. The duration of exposure to the solution is a critical factor as too long treatments lead to molecule clusters. The coverage of germanium by self-assembled monolayers is a good proof of an Hterminated surface. Acknowledgement We thank Umicore AG and M. Meuris and P. Mertens from IMEC for providing the Germanium wafers, Prof. Dr. W. Jaegermann, Dr. T. Mayer and Eric Mankel for the XPS facilities and Yvonne Filbrand for helpful discussions. This work was supported by the European Commission - Research Infrastructure Action under the FP6 "European Integrated Activity of Excellence and Networking for Nano and Micro-Electronics Analysis" - Project number 026134(RI3) ANNA.

References
[1] J. He, Z.-H. Lu, S. A. Mitchell, D. D. M. Wayner; J. Am. Chem. Soc. 120 (11) 2660 (1998) [2] M. Lommel, E. Mankel, B. O. Kolbesen, ECS Transactions 11 (20) 83-90 (2008) [3] D. K. Schwartz, Annu. Rev. Phys. Chem. 52 107 (2001) [4] M. Krämer, Dissertation, Dortmund (2007) [5] D.L. Windt, Computers in Physics, 12 360-370 (1998) [6] D. K. G. de Boer, Phys. Rev. B 44(2), 498 (1991) [7] B. Pollakowski, B. Beckhoff, F. Reinhardt, S. Braun, P. Gawlitza, Physical Review B 77, 235408 (2008) [8] P. Hönicke, B. Beckhoff, M. Kolbe, S. List, T. Conard, H. Struyff, Spectrochim. Acta B 63, 1359 (2008) [9] M. Lommel, F. Reinhardt, P. Hönicke, E. Mankel, M. Müller, M. Kolbe, B. Beckhoff, B.O. Kolbesen, ECS Transactions, to be published, (2009)

16

Calibration of the NASA instrument EUNIS
W. Paustian, M. Richter, R. Thornagel Physikalisch-Technische Bundesanstalt (PTB) , Abbestraße 2-12, 10587 Berlin, Germany B. Kent Rutherford Appleton Laboratory (RAL), Chilton, Didcot, Oxfordshire OX11 0QX, UK Solar radiation in the VUV strongly influences the photochemical processes in the upper atmosphere and, thus, also the earth climate. Space-based observations of the sun in this spectral range have, therefore, attracted increasing interest during the last years. In this context, the reliable characterization of the corresponding space instruments is of high significance. In the BESSY laboratories of PTB, work on solar telescope calibration has a long tradition [1]. In the mid 1990s, the SUMER (Solar Ultraviolet Measurements of Emitted Radiation) and CDS (Coronal Diagnostic Spectrometer) instruments of the SOHO (SOlar and Heliospheric Observatory) mission were calibrated, via hollow cathode gas discharge plasma sources as the transfer standards, and in 2004 the EIS (Extreme-ultraviolet Imaging Spectrometer) instrument of the Solar-B/Hinode mission. The absolute VUV emissions of the source standards used are traceable to calculable synchrotron radiation. In the following, several underflight calibrations of these and further solar instruments were performed by means of short-term rocket missions like SERTS (Solar Extreme-ultraviolet Research Telescope and Spectrometer), MOSES (Multi Order Solar EUV Spectrograph), and EUNIS (Extreme Ultraviolet Normal Incidence Spectrometer) whose own calibrations were performed at the Rutherford Appleton Laboratory (RAL) using PTB’s CDS calibration source.

Fig. 1. Scheme of the EUNIS instrument.

17

The EUNIS instrument (Fig. 1) uses two independent optical systems in two wavelength regions: 17 nm to 21 nm with 3.5 pm resolution and 30 nm to 37 nm with 7 pm resolution. There are only two reflections in each optical channel, from a superpolished, off-axis paraboloidal pre-mirror and a toroidal grating, each coated with a high-efficiency multilayer. Hence, the throughput of EUNIS is much higher than for SERTS that have preceded it. The detector in each channel is a microchannel plate image intensifier fiber-coupled to three pixel sensors. EUNIS is supported by NASA and was launched in April 2006 and November 2007. Each EUNIS flight was accompanied by an afterflight calibration campaign within a PTBRAL cooperation, the first in October 2006 and the second in May 2008. As an example, Fig. 2 shows a part of the Ne III spectrum as emitted from the CDS calibration source and measured with the EUNIS instrument. Instrument calibration is obtained by comparing its signal output with the absolute radiant power emitted from the calibration source. For this purpose, integration along the wavelength intervals indicated by the vertical lines in Fig. 2 has to be performed which represent the ranges of integral emission of the source as calibrated at BESSY. Next flight and calibration within the very successful EUNIS program are scheduled for 2010.

Fig. 2. Part of the Ne III spectrum as emitted from the CDS calibration source and measured with the EUNIS instrument.

REFERENCES [1] M. Richter, A. Gottwald, F. Scholze, R. Thornagel, G. Ulm, Calibration of Space

Instrumentation with Synchrotron Radiation. Advances in Space Research 37, 265-272 (2006).

18

High-resolution X-ray absorption and emission spectroscopy on binary titanium compounds F. Reinhardt1,2, B. Beckhoff1, H. Eba3, B.Kanngießer2, M. Kolbe1, M. Mizusawa3, M. Müller1, B. Pollakowski1, K. Sakurai3 and G. Ulm1
2

Physikalisch-Technische Bundesanstalt, Abbestr. 2-12, 10587 Berlin, Germany Technische Universität Berlin, Institut für Optik und Atomare Physik, Hardenbergstr. 36, 10623 Berlin, Germany 3 National Institute for Materials Science, Sengen, Tsukuba, Ibaraki 305-0047, Japan

1

Introduction: While the elementary composition of samples is readily determined by use of X-ray fluorescence analysis (XRF), the chemical state of a probed atom is only accessible by techniques employing high resolution in either the excitation or the detection channel. For the chemical speciation of binary compounds of tri- and tetravalent titanium highresolution X-ray absorption and emission spectra were recorded in different energy regimes in order to evaluate and to qualify both near-edge X-ray absorption fine structure (NEXAFS or XANES) spectroscopy and wavelength-dispersive X-ray emission spectroscopy (WDXES) as spectroscopic methods for this analytical task [1]. For a comparison of the information gained from the various methods, the titanium compounds were classified according to the bonded titanium's oxidation state. Thus, it was possible to distinguish between inner atomic effects due to different oxidation states and external effects related to the respective ligand and the surrounding structure. It becomes evident, that only the combined use of the complementary methods both in the soft and the hard X-ray range allows for a reliable speciation of tri- and tetravalent titanium compounds. Instrumentation: All measurements in the soft X-ray regime, i.e. Ti-L shell absorption and emission spectroscopy, were carried out at the plane-grating monochromator (PGM) beamline for undulator radiation [2] in the PTB laboratory at BESSY II. For the Ti-K absorption spectra the measurements took place at the four-crystal monochromator (FCM) beamline [3] in the same laboratory. Soft X-ray emission spectra were obtained employing a wavelengthdispersive spherical grating spectrometer in Rowland geometry with 1200 lines/mm [4]. By aligning the CCD-detector to the focus position of the Ti-Lα fluorescence line, a resolving power of E/∆E = 430 is achieved in the Ti- Lα,β energy range, leading to an energy resolution of about 1 eV. For all samples the incident photon energy was set to 520 eV, well above the LII,III edges in a non-resonant regime. The incident photon energy was below the Ti-LI edge to create vacancies only in the Ti-2p orbitals and not in the 2s Orbital as well. With energy values for Ti-Lα and Ti-Lβ of pure titanium taken from the Elam database, (451.8 eV and 458.2 eV [5]) an energy axis was set for the recorded section of the spectrum. Complementary to the results obtained in the soft X-ray regime, additional measurements were carried out at SPring-8, beamline BL40XU with NIMS instrumentation optimized for wavelength-dispersive detection in a very broad range of hard X-ray photon energies [6, 7]. Experiment: Because of self-absorption of emitted fluorescence radiation all emission features with an energy above the Ti-LIII edge energy are reduced in intensity relative to the Ti-Lα emission and all features below. In figure 1 the X-ray emission spectra of titanium and its oxides are shown. A chemical shift is visible for all emission features consistent with the increasing oxidation state of the titanium. The asymmetry of all Ti-Lα lines is caused by selfabsorption since the LIII absorption edge energy is close by [1, 8]. On the low-energy side of the Ti-Lα emission line a satellite structure appears in the spectra of the titanium oxides.

19

Figure 1: Ti-L X-ray emission spectra of pure titanium and the titanium oxides. The CCD-rows were converted to an energy axis by use of database values for Ti-Lα and Ti-Lβ of pure titanium. a) normalized to current induced in a calibrated reference diode by the incident radiation. b) normalized to 1.

Absorption spectra were recorded for different titanium compounds, i.e. TiO, Ti2O3, TiO2, TiS2, TiN and TiC, in both the soft and the hard X-ray regime. A comparison shows the wellknown correlation between oxidation state and absorption edge energy position and an additional dependence on the ligand. Compounds with titanium in the same oxidation state and with ligands similar in Z (Ti2O3 and TiN or TiO2 and TiC) are hardly distinguishable by their K near-edge spectra. In contrast, the X-ray emission is characteristic for those compounds due to the different chemical properties of those ligands (fig. 2). Compounds where the ligand shows similar chemical properties (TiO2 and TiS2) give rise to a similar Xray emission (fig. 2) but with O and S having a considerably different atomic number and hence a very different electron backscattering ability, their XANES spectra clearly differ.

Figure 2: High-resolution X-ray emission spectra of Ti-Lα,β a) of Ti2O3 and TiN. b) of TiO2, TiS2 and TiC.

Conclusion: To evaluate the capabilities of high-resolution X-ray absorption and X-ray emission spectroscopy for chemical speciation, spectra were recorded for a set of binary titanium compounds. Other compounds, where titanium is in the same oxidation state, exhibit significant differences in their Ti-K near-edge absorption spectra only if the atomic number of the ligand itself differs significantly from the one of oxygen.

20

The Ti-Lα,β emission spectra are, in general, not a reliable indicator for different ligands. If titanium is present in a defined oxidation state, the chemical properties of the ligand play a major role for the structure of the emission spectrum. As shown for the Ti-L emission spectra of TiO2 and TiS2, X-ray emission spectroscopy hardly gives any indication to the species if the chemical bonding is similar. It is obvious that the combination of both spectroscopic methods yields a larger amount of information than each of these approaches themselves. The information from occupied and unoccupied density of states give access to complementary information and therefore allow for a reliable chemical speciation. References [1] F. Reinhardt, B. Beckhoff, H. Eba, B. Kanngießer, M. Kolbe, M. Mizusawa, M. Müller, B. Pollakowski, K. Sakurai, and G. Ulm, Anal. Chem. (2009), in print, DOI 10.1021/ac8018069 [2] F. Senf, U. Flechsig, F. Eggenstein, W. Gudat, R. Klein, H. Rabus, and G. Ulm, J. Synchrotron Rad. 5:780-782, 1998. [3] M. Krumrey, J. Synchrotron Rad. 5:6-9, 1998. [4] M. Müller, B Beckhoff, R. Fliegauf, and B. Kanngießer, submitted, 2009. [5] W. T. Elam, B. D. Ravel, and J. R. Sieber, Rad. Phys. Chem. 63:121-128, 2002. [6] K. Sakurai, H. Eba, K. Inoue, and N. Yagi, Nucl. Instr. Meth. A 467:1549, 2001. [7] K. Sakurai and M. Mizusawa, Rev. Sci. Instr. 78:066108-1-3, 2007. [8] B. Stoyanov, F. Langenhorst, and G. Steinle-Neumann, American Mineralogist 92:577586, 2007.

21

VAMAS projects on reproducibility in X-ray reflectometry
M. Krumrey
Physikalisch-Technische Bundesanstalt, Abbestr. 2-12, 10587 Berlin, Germany X-Ray Reflectometry (XRR) is a well-established technique to evaluate quantitatively electron density, thickness and roughness of thin layers. Due to the interference of X-rays reflected from different interfaces, layers and multilayers on flat substrates give rise to oscillations. The period of the oscillations depends on the layer thickness, the fringe amplitude depends on surface and interface roughness and the relative electron densities of the materials. During the last years, two worldwide XRR round-robin experiments were performed within the framework of the VAMAS project “X-ray reflectivity measurements for evaluation of thin films and multilayers”. The reproducibility of measurements obtained using different equipments has been investigated. VAMAS (Versailles Project on Advanced Materials & Standards – www.vamas.org) is an international organization that supports trade in products using advanced materials through international collaborative projects aimed at providing the technical basis for harmonized measurements, testing, specifications, and standards. For one of the round robins, a thin TaN/Ta bilayer was deposited on a 300 mm silicon wafer at the Albany NanoTech facility in the College of Nanoscale Science and Engineering (CNSE) at the University at Albany [1]. Unlike conventional Ta-based barrier metallizations, this sample was prepared with the tantalum metal being deposited first and the TaN deposited last; this was done in order to passivate the tantalum metal surface with a relatively non-reactive TaN compound layer. On the basis of the previously mapped lateral uniformity, four samples in near-center region were chosen for the round-robin analyses and delivered to the participating laboratories. The 4 samples were measured by 6 different laboratories (Table 1). While most participants used conventional X-ray sources, PTB applied synchrotron radiation at the four-crystal monochromator beamline at BESSY II [2]. The measurements were performed using the PTB UHV X-ray reflectometer [3]. The analysis of the specimens was based on the average of up to five independent XRR experiments performed, removing and mounting the sample between each measurement. Autocorrelation functions of the derivative of the density profile were obtained by Fourier transform of the ratio of reflectivity data and Fresnel reflectivity [4]. The layer thicknesses have been evaluated by a Gaussian fit of the peaks in the Fourier spectra (Fig. 1).
LAB A B C D E F X-ray source Cu Kα rot. anode BESSY II storage ring Cu Kα Cu Kα Cu Kα Cu Kα Monochromator Göbel mirror 4 Si (111) crystals Göbel mirror Göbel mirror Parabolic multilayer Si(111) channel cut crystal Slit size / μm 100 300 300 600 100 200 Detector NaI:Tl scintillator Si photodiode / Photon count. Si drift scintillator

scintillator

TABLE 1 Experimental setups of the laboratories involved in the TaN/Ta round-robin.

22

1.E+04

1.E+02

Reflectivity

1.E+00

F
1.E-02

14 12 10 8 6 4 2 0 20 15 10 5 0 20

sample 1

Number

13

13.05 13.1 13.15 13.2 13.25 13.3 13.35 13.4

E D C B A
0.1 0.2 0.3 0.4 0.5
-1

Ta layer thickness / nm

sample 2

1.E-04

1.E-06

1.E-08

0.0

0.6

0.7

0.8

0.9

Number

qz / Å

13

13.05 13.1 13.15 13.2 13.25 13.3 13.35 13.4

Number

15 10 5 0 12 10

Ta layer thickness / nm

sample 3

|FFT|2

F

13

13.05 13.1 13.15 13.2 13.25 13.3 13.35 13.4

Number

E D C B A
5 10 15 20

Ta layer thickness / nm

sample 4

8 6 4 2 0 13 13.05 13.1 13.15 13.2 13.25 13.3 13.35 13.4

0

thickness / nm

Ta layer thickness / nm

FIGURE 1. Upper left: Experimental XRR profiles acquired by different laboratories on TaN/Ta sample 1. XRR curves are scaled by one order of magnitude for clarity. Measurements in the PTB laboratory at BESSY II are labelled with B. Lower left: Relative auto-correlation function of the derivative of the electronic density profile obtained by Fourier transform. The peaks correspond to the thickness of the Ta layer and to the total layer thickness of the bilayer, respectively. Right: Distribution of thicknesses of Ta layer for different samples obtained from the peak in the auto-correlation function.

The maximum discrepancy between thickness values obtained in different labs is less than 0.2 nm. For all the samples standard deviation on the Ta layer thickness is less than 0.04 nm corresponding to about 0.3 %, despite the fact that the XRR measurements were performed using different instruments and in particular with different characteristics in terms of their angular resolution and their maximum and background intensity levels. For another round-robin with 20 participating laboratories, GaAs/AlAs multilayer samples were chosen because GaAs and AlAs readily grow epitaxially without relaxing. This kind of sample has already been considered as a possible XRR reference standard. However, a top surface oxide developed during the round robin, meaning that only the thicknesses of buried layers can be meaningfully compared. GaAs/AlAs multilayer samples were supplied from the Surface and Thin Film Standards Section of the National Metrology Institute of Japan (NMIJ), AIST Japan. Three pairs of GaAs/AlAs bilayers (Fig. 2) were fabricated on four 4-inch GaAs (100) wafers simultaneously, using a molecular beam epitaxy technique. Fabrication conditions were optimized via structural evaluation of multilayers by TEM and XRR. Uniformities of thicknesses across a wafer and from wafer to wafer were confirmed by XRR. For the analysis of the measured reflectance data for this complex structure, a modelling and fitting procedure based on the Parratt recursive formalism of the Fresnel equations was applied [6]. Each laboratory applied its own fitting procedure to obtain values for the layer thickness. In order to compare the raw data, excluding the influence of the refinement procedure and the operator choices, simulations of all datasets were also performed starting from the same model and using the same simulation routine (IMD [7]). Each XRR spectrum was fitted over the whole data range. In some cases, the scans were truncated at an angle at which the Kiessig fringes were no longer distinguishable

23

because of the low signal to noise ratio. For each experimental curve, 18 parameters (seven layer thicknesses, eight roughness values, n and k for the top layer and reflectance scaling factor) were fitted together. Logarithmic fitting was performed using the Marquardt algorithm with statistical weighting. Thickness distributions for buried layers do not differ significantly from the free-fitted ones having a sharp distribution with mean values between 9.49 nm and 9.58 nm and standard deviation values from 0.06 nm to 0.16 nm, respectively. The topmost GaAs layer thickness show a broader distribution of around 0.3 nm centred on a thickness of about 9.5 nm. In general, thickness values measured on a GaAs/AlAs multilayer proposed as a reference sample show a good intra-laboratory reproducibility of about 0.03 nm and an inter-laboratory reproducibility of about 0.1 nm for the buried layers.

FIGURE 2. Left: Schematic view of the GaAs/AlAs sample structure employed for the round-robin.

Right: Distribution of thicknesses obtained by a common fitting approach (IMD) for all data measured by all participating laboratories.

REFERENCES
[1] R.J. Matyi, L.E. Depero, E. Bontempi, P. Colombi, A. Gibaud, M. Jergel, M. Krumrey, T.A. Lafford, A. Lamperti, M. Meduna, A. Van der Lee and C. Wiemer, Thin Solid Films 516, 7962 – 7966 (2008) [2] M. Krumrey and G. Ulm, Nucl. Instr. and Meth. A 467-468, 1175 - 1178 (2001) [3] D. Fuchs, M. Krumrey, P. Müller, F. Scholze and G. Ulm, Rev. Sci. Instrum. 66, 2248 - 2250 (1995) [4] S. Banerjee, S. Ferrari, D. Chateigner and A. Gibaud, Thin Solid Films 450 (2004) 23–28 [5] P. Colombi, D. K. Agnihotri, V.E. Asadchikov, E. Bontempi, D.K. Bowen, C.-H. Chang, L. E. Depero, M. Farnworth, T. Fujimoto, A. Gibaud, M. Jergel, M. Krumrey, T. A. Lafford, A. Lamperti, T. Ma, R. J. Matyi, M. Meduna, S. Milita, K. Sakurai, L. Shabel'nikov, A. Ulyanenkov, A. Van der Lee and C. Wiemer, J. Appl. Cryst. 41, 143 – 152 (2008) [6] L. G. Parratt, Phys. Rev. 95, 359-369 (1954) [7] D. L. Windt, Computers in Physics 12, 360-370 (1998)

24

Cryogenic radiometry in the hard X-ray range
M. Gerlach, M. Krumrey, L. Cibik, P. Müller, H. Rabus and G. Ulm
Physikalisch-Technische Bundesanstalt, Abbestr. 2-12, 10587 Berlin, Germany For many applications in radiometry, spectroscopy or astrophysics, absolute measurement of radiant power with low uncertainty is essential. Cryogenic electrical substitution radiometers (ESRs) are regarded as the highest-accuracy primary standard detector in radiometry, from the infrared to the ultraviolet region; in combination with tuneable monochromatized synchrotron radiation from electron storage rings, their range of operation has been extended to the soft X-ray region [1]. ESRs are absolute thermal detectors, based on the equivalence of electrical power and radiant power that can be traced back to electrical SI units and be measured with low uncertainties. The core piece of an ESR is its cavity absorber, which is thermally linked to a fixed temperature heat sink kept at liquid helium temperature via a heat resistance. The absorber is equipped with a thermometer and a heater that allows for the supply of electrical heating power (Fig. 1). The absorber temperature is kept constant by an active control. When radiation is provided, the electrical heating power required to keep the absorber at a constant temperature undergoes a reduction equivalent to the incident radiant power so that the radiant power is obtained through the measurement of the electrical heating power. Cavity absorbers are typically made of copper which provides excellent thermal conductivity at liquid helium temperature in combination with a moderate heat capacity, which in turn ensures a short response time suitable for the measurement of monochromatized synchrotron radiation. However at photon energies above 20 keV, the use of copper absorbers, typically 100 µm in thickness, prevents the operation of the ESR due to increasing transmittance.

FIGURE 1. Operating principle of an electrical substitution radiometer. The new cavity absorber developed at

PTB consists of a gold base and a copper shell to ensure complete absorption up to a photon energy of 60 keV.

To develop a new absorber design for hard X-rays, simulations were carried out using the Monte Carlo simulation code Geant4 [2]. Extensive simulations were performed for a large variety of absorber materials including Cu and Au as well as Ag, Pt and W. Alternative absorber geometries including different thickness of base and shell were investigated as well. Also, thermal aspects had to be taken into account, such as high thermal conductivity and low heat capacity at liquid helium temperatures. Concerning these aspects, metals with low atomic number, such as Cu and Ag, show the best performances, and it became obvious that no single metal would meet all requirements. Furthermore, an appropriate way of fabrication, such as electroforming, had to be applied for the respective metal. The simulations and experiments resulted in a final design of an absorber with a gold base, 730 µm in thickness, inclined 30°, and a cylindrical shell made of copper, 90 µm in thickness, to reduce losses caused by fluorescence and scattered photons. The cavity absorber was manufactured by electroforming at PTB and was implemented into the existing ESR SYRES I [3].

25

The spectral absorptance of the new and the former cavity absorber is shown in figure 2 (left) in comparison to a pure gold plate, 500 µm in thickness, that would cause significant losses due to Au L fluorescence. These can be totally prevented by applying a copper shell of 100 µm thickness. With this cavity absorber, an absorptance close to 100 % can be achieved in the full energy range from 250 eV, which is the low energy limit for the Monte Carlo simulations using Geant4 and 60 keV, which in turn, is the maximum photon energy with fairly high photon flux at the BAMline using the double-crystal monochromator [4]. The remaining losses are depicted in figure 2 (right). For photon energies above the Au K absorption edge of 80.75 keV, Au fluorescence is the dominant contribution. At 60 keV scattering contributes with 0.4 %, whereas transmitted radiation becomes the key factor for losses in the photon energy range between 75 keV and 80.75 keV.

FIGURE 2. Left: Simulated absorption of the former copper absorber, 100 µm in thickness, compared to a pure

gold plate, 500 µm in thickness, and the new absorber with gold base and copper shell. Right: Simulated losses of the new absorber caused by transmittance, fluorescence, scattering and photoelectrons including Auger electrons.

An important application of the newly developed cryogenic radiometry in the hard X-ray range is the calibration of X-ray detectors such as semiconductor photodiodes. Monochromatized synchrotron radiation of high spectral purity was used to calibrate different silicon photodiodes against the ESR for photon energies up to 60 keV. The spectral responsivity of these photodiodes was determined with relative standard uncertainties below 0.3 %. The spectral responsivity in the photon energy range from 1.75 keV to 10 keV was measured at the PTB four-crystal monochromator beamline [5] and from 8 keV to 60 keV at the BAMline. The results for three different types of silicon photodiodes are shown in figure 3: AXUV 100 (International Radiation Detectors, IRD, USA), S3590 (Hamamatsu, Japan) and PIPS 50-500 (Canberra, Belgium). In the energy range from 4 keV to 10 keV, the photodiodes of the Hamamatsu S3590 type and the Canberra PIPS type exhibit a high responsivity close to the theoretical maximum of 0.273 A/W for silicon [6]. The responsivity declines significantly for photon energies greater than 10 keV, caused by increasing transmittance. By applying a model calculation for the energy absorption in silicon, the thickness of the active volume of the respective photodiode can be determined as 320 µm and 510 µm, respectively, and be compared to the manufacturer’s specifications, which are 300 µm and 500 µm, respectively. The thickness of the active volume of the IRD AXUV 100 photodiode is 27 µm, resulting in a much lower responsivity at photon energies greater than 4 keV. In the photon energy range just above the K absorption edge of silicon at 1.839 keV, the photodiodes of type S3590 have a reduced responsivity caused by the absorption in their thick passivation front layer, whereas the IRD AXUV 100 photodiodes exhibit high responsivity due to their SiO2 front layer thickness of only a few nm. The Canberra photodiodes of type PIPS combine both advantages, a thick active layer of 510 µm and a thin front layer, ensuring a high responsivity and making this type the most suitable for use as secondary standard in the entire photon energy range. The photodiodes of Hamamatsu type S3590 exhibit a discontinuity in their spectral responsivity at a photon energy of 25.5 keV, which is caused by the fluorescence radiation of the conductive silver with that the photodiode is attached to its mounting [7].

26

FIGURE 3. The spectral responsivity of three different semiconductor photodiodes calibrated against the ESR in the soft and hard X-ray ranges. Thicknesses of the active volume were derived from a model for photon absorption in silicon (solid lines).

In the photon energy range above 20 keV, a cryogenic radiometer was used for the first time as a primary standard detector to calibrate photodiodes. Whereas the relative uncertainty was about 2 % at PTB by calibration against a free-air ionization chamber due to the uncertainty of the mass energy absorption coefficients [8], for the present measurement with SYRES I relative standard uncertainties between 0.18 % and 0.30 % were achieved in the entire photon energy range, including hard X-rays with photon energies of up to 60 keV [9].

REFERENCES
[1] A. Gottwald, U. Kroth, M. Krumrey, M. Richter, F. Scholze and G. Ulm, Metrologia 43, S125 – S129 (2006) [2] S. Agostinelli et al., Nucl. Instr. and Meth. A 506, 250 – 303 (2003) [3] H. Rabus, V. Persch and G. Ulm, Appl. Opt. 36, 5421 – 5440 (1997) [4] W. Görner, M.P. Hentschel, B.R. Müller, H. Riesemeier, M. Krumrey, G. Ulm, W. Diete, U. Klein and R. Frahm, Nucl. Instr. and Meth. A 467-468, 703 - 706 (2001) [5] M. Krumrey and G. Ulm, Nucl. Instr. and Meth. A 467-468, 1175 - 1178 (2001) [6] F. Scholze, H. Rabus and G. Ulm, J. Appl. Phys. 84, 2926 - 2939 (1998) [7] M. Krumrey, L. Büermann, M. Hoffmann, P. Müller, F. Scholze and G. Ulm, AIP Conf. Proc. 705, 861 – 865 (2004) [8] L. Büermann, B. Grosswendt, H.-M. Kramer, H.-J. Selbach, M. Gerlach, M. Hoffmann and M. Krumrey, Phys. Med. Biol. 51, 5125 – 5150 (2006) [9] M Gerlach, M Krumrey, L Cibik, P Müller, H Rabus and G Ulm, Metrologia 45, 577-585 (2008)

27

First commissioning results in the IR/THz range at the electron storage ring Metrology Light Source
Ralph Müller1, Arne Hoehl1, Roman Klein1, Gerhard Ulm1, Jörg Feikes2, Michael von Hartrott2, Ulrich Schade2, Godehard Wüstefeld2
2

Physikalisch-Technische Bundesanstalt, Abbestraße 2-12, 10587 Berlin, Germany Helmholtz-Zentrum Berlin für Materialien und Energie GmbH Elektronen-Speicherring BESSY II, Albert-Einstein-Str. 15, 12489 Berlin, Germany

1

In April 2008, the PTB’s new electron storage ring, the Metrology Light Source (MLS), went into user operation [1-3]. Synchrotron radiation sources have major advantages in the IR range compared to conventional thermal sources, higher photon flux in the far-IR, higher brilliance, pulsed radiation in the ps range, and polarized radiation. Additionally, electron storage rings in a special operation mode with short electron bunches can deliver intense coherent synchrotron radiation (CSR) in the lower energy part of the far-IR (sub-THz to THz) with gain up to 6 to 9 orders of magnitude compared to conventional, incoherent synchrotron radiation emission. The MLS is the first electron storage ring worldwide designed and prepared for low-α operation mode based on the octupole correction scheme, for the production of CSR in the far-IR and THz region. This option strengthens the MLS as a strong THz radiation source [1 -6]. At the MLS two beamlines dedicated to the use of IR and THz synchrotron radiation are under commissioning respectively operational: (1) the MLS-IR beamline optimized for the MIR to FIR [6], and a dedicated THz beamline optimized for the FIR/THz spectral range. The commissioning of the IR beamline started in April 2008. The construction of the THz beamline was finished in the end of 2008. The IR beamline and THz beamline consist of an arrangement of mirrors which allows - in combination with a special port of the dipole chamber – the transport of the beam to the experiment (see Fig.1). After all mirror reflections the σ-polarization of the electrical wave vector of the radiation is horizontally oriented. First measurements at the IR beamline with calibrated filter radiometers and an IR camera in the visible and near infrared spectral range reveal the good adjustment of the optical path of the beamline. All the flux expected from theoretical calculations is measured at the experiment. The shape and size of the focus is also as good as expected. With this adjusted IR beamline we were able to make first measurements in the THz spectral range. The inset of Fig. 1 shows the focus of the THz radiation (all radiation with a wavelength longer than 500 mm) in the low α mode. Its FWHM size is approximately 3 mm in diameter and is located nearly at the same position as the focus of the visible and near infrared light.

28

Fig. 1. Optical design of the MLS-IR beamline. The inset shows the focus of the CSR taken with an infrared camera in the THz spectral range in the low α mode.

The MLS has a unique capability to control the higher orders of a and to achieve bunch length reductions by a factor 10 in the sub-mm range. The higher orders of α are controlled by suitably placed sextupoles and octupoles [4]. By careful tuning of the optics, stable CSR is generated at the MLS. A first proof of stable CSR has been obtained. Fig 2. shows the results for the low a mode at 630 MeV and a ring current of 19 mA. Fig. 2a shows the chopped detector signal in the THz spectral region. Without chopping, the signal amplitude is smooth and shows no signs of bursting instabilities (see Fig. 2b). So the THz radiation is stable within the time resolution of a liquid-helium-cooled InSb hot electron bolometer of few microseconds.

Fig. 2. Oscilloscope traces of stable THz signals at the MLS for 630 MeV electron energy, low α optics, 19 mA ring current, and a cavity voltage of 200 kV measured with an InSb-detector: (a) The chopped THz signal amplitude. (b) The unchopped THz signal amplitude is constant and shows no bursting instabilities.

29

Additionally we measured the absolute THz power in the focus of the IR beamline. The measured power is depending on the chosen α in the range of a few hundred micro-watts. However, the propagation of sub-terahertz and terahertz electromagnetic waves from the source point to the experiment through such a typical IR beamline is strongly affected by diffraction. This is why we decided to build a dedicated THz beamline with larger optical elements and different transmission windows compared to the IR beamline [6]. First measurements at this new beamline show up to two orders of magnitude more power in the THz range compared to the results at the IR beamline. In summary, the IR project at the MLS is well under way. The MLS-IR beamline is well adjusted and ready for measurements in the infrared and THz spectral range. The THz beamline and the undulator IR beamline are under commissioning.
[1] [2] [3] [4] [5] [6] G. Brandt et al., Nucl. Instr. and Meth. B 258, 445 (2007) G. Ulm et al., Nucl. Instr. and Meth. A 582, 26 (2007) J. Feikes et al., Proc. of EPAC08, 2010 (2008) G. Wüstefeld, Proc. of EPAC08, 26 (2008) R. Müller et al., Proc. of EPAC08, 2058 (2008) R. Müller et al., Infrared Physics & Technology 49, 161 (2006)

30

In-situ analysis of the Zn(S,O)-buffer layer preparation for chalcopyrite solar cells in the liquid phase using Zn edge X-ray absorption spectroscopy

Iver Lauermann, Timo Kropp, Ahmed Ennaoui, and Emad F. Aziz
Helmholtz-Zentrum Berlin für Materialien und Energie, Berlin, Germany

In chalcopyrite solar cells, a CdS buffer layer is needed between the p-doped absorber and the ZnO window layer. To replace CdS, Zn(S,O) has been introduced as adequate replacement buffer material. However, the chemical bath deposition (CBD) process for depositing Zn(S,O) from an aqueous solution, containing ZnSO4, ammonia, and thiourea ((NH2)2CS), is not understood in detail. The LIQUIDROM station at the BESSY II U 41 beamline allows the spectroscopic characterization of liquids and solids under ambient pressure by X-ray absorption spectroscopy (XAS). We used in-situ near edge XAS (NEXAFS) to examine the solution chemistry and to identify the species present in a Zn(S,O) CBD solution at every step of the reaction. Experimental details of the setup for soft X-ray spectroscopy of liquid solutions have been described previously [1]. In Figure 1a (red curve) the Zn L-edge spectrum of Zn2+ ions (1M ZnSO4) in pure water is presented. Three main features at 1024.3 eV, 1026.5 eV and 1034.5 eV are visible. These features are due to the transition of electrons from the p-type molecular orbitals (MOs) to the d- and s-type valence MOs localized on the Zn2+ ions according to the dipole allowed selection rules [2]. In order to elucidate the nature of the unoccupied MOs, unrestricted Hartree Fock (UHF) calculations were applied to Zn2+ with a coordination shell of 6 water molecules. In addition, Tomasi's polarized continuum model (PCM) was used to calculate the effect of the solvent [3]. Although nominally the d-MOs of Zn2+ are fully occupied (d10), upon dissolving the ions in water the d- and s-MOs hybridize with the valence MOs of the water molecules and form partially empty MOs, which can be filled by the excited pelectrons upon X-ray absorption. Comparing the XA spectrum of Zn2+ ions with the calculated MOs, the local electronic picture of the unoccupied states could be qualitatively drawn in Figure 1.a. The nature of the MOs under the first peak is s-type, the 2nd peak is dxy, xz, yz-type, and the third peak is dx22 y ,

dz2-type. Upon diluting the ZnSO4 solution to a concentration of 0.15 M (the actual concentration in

the CBD solution), the s-type MOs are decreased in intensity relative to the d-type as shown in Figure 1a, black curve. This can be due to the ion-pairing between the Zn2+ and SO4- which can take place at 1M but is reduced significantly upon dilution to 0.15 M according to our previous investigation. [4]

31

Figure 1. XAS L-edge of Zn in solution, a) Zn2+ in water, 1 M (red), and 0.15 M (black). (b) 0.15 M Zn2+ in 0.13 M aqueous NH3 solution (green). Addition of thiourea to a concentration of 0.6 M (blue), followed by heating to 80° C for 3 min, then cooling to room temperature (black curve). (c) 0.15 M Zn2+ in 0.6 M aqueous solution of thiourea (red), heated to 80° C (gray). Cooling down to 50° C and addition of NH3 to yield 0.13 M (black). The dashed vertical lines are presenting the experimental peak shift and splitting. The results of UHF calculations for the energies of the MOs are shown as vertical grey bars for the following models; (a) Zn2+ surrounded by 6 H2O, (b) Zn2+ surrounded by 5 H2O and one NH3, (c) Zn2+ surrounded by 5 H2O and one molecule of thiourea.

Replacing one water molecule with ammonia in the theoretical model causes the disappearance of the dx2-y2 and dz2 valence MOs as shown in Figure 1b. Furthermore, the s and dxy, xz, yz valence MOs are shifted to higher energies by approximately 1 eV. This qualitatively explains the change in the experimental spectrum of Zn2+ in a solution with added ammonia (0.13 M NH3), as shown in Figure 1b (green curve). This is due to electron transfer from the ammonia molecule to the empty valence dx22 y

and dz2 MOs upon complex formation; therefore they are no longer available to receive excited

electrons from oxygen p-orbitals upon XA. A second consequence of this complex formation is an increase in the intensity of the valence dxy, xz, yz relative to the s-MOs. Adding thiourea to the previous solution does not affect the local electronic structure of Zn2+ as shown in Figure 1b (blue curve). Heating this solution to 80° C and subsequent cooling to room temperature leads to a shift of the spectrum by 0.3 eV to higher energy as shown in Figure 1b (black curve) but no further change of the overall shape. Thus we conclude that the chemical environment of the Zn2+ in the CBD-N remains almost unchanged after addition of thiourea, i.e. it is still surrounded by ammonia molecules and the main interaction is between nitrogen and the zinc ion. However, adding thiourea first to the 0.15 M ZnSO4 solution (as in the CBD-T process) instead of ammonia, leads to a significantly different spectrum as shown in Figure 1c (red curve). The calculated MO energies for Zn2+ with 5 water molecules and one molecule of thiourea are in a satisfactory agreement with the experimental XAS peak positions. The optimized model shows that the main interaction between the Zn2+ and thiourea is through the C=S group rather than the NH2-groups of thiourea. This can explain the drastic difference between spectra obtained by the CBD-N method (Figure 1.b) and the CBD-T method (Figure 1.c). Heating this solution to 80° C causes a decrease in the s-valence MOs as shown in Figure 1c (grey curve). Cooling down the solution to 50° C and adding ammonia (to 0.13 M NH3 as in the actual CBD process) causes a further small decrease in the s-valence MOs intensity compared to the dx2-y2 and dz2

32

MOs as shown in Figure 1c (black curve). However, the final spectrum obtained from CBD-T is different from the one obtained by CBD-N, where the concentration of all components is identical. The XA spectrum obtained from the final CBD-T process is similar to the XA spectrum of a zinc tris(thiourea) sulphate complex (ZTS) in solution, which was synthesized chemically, crystallised, and re-dissolved in water. In this complex the zinc forms a complex with the thiourea via the sulphur atom [5]. This result confirms further that in the CBD-T process, the complex formation between the zinc and thiourea is via the sulphur atom. Comparing the CBD-N and CBD-T process, one can conclude that in the CBD-N the formation of the zinc thiourea complex is inhibited by the presence of ammonia in solution but that this complex, once formed, as is the case in the CBD-T, is not re-dissolved by ammonia. In both, the CBD-T solution and zinc tris(thiourea) complex in solution, the appearance of the peak corresponding to dx2-y2 and dz2 MOs is due to the back-donation of electrons from the zinc ion to the C=S bond. The electron back donation from metal to ligand results in a decrease of electron density in the MOs localized on the metal, which increases the transition probability of excited pelectrons. This work clearly shows the chemical background behind the observed differences in solar cell efficiency depending on deposition conditions. Furthermore, it demonstrates the value of the set-up for in-situ examinations of chemical deposition reactions and opens the door for the investigation of further preparation routes to organic and inorganic solar cells. The results described here have been published in [6].

References 1. N. Ottosson, M Henrik Bergersen, Wandared Pokapanich, Gunnar Oehrwall, Svante Svensson, Wolfgang Eberhardt, and Olle Bjoerneholm 2. J. Stöhr, NEXAFS spectroscopy (Springer-Verlag, Berlin; New York, 1992). 3. J. Tomasi, M. Persico, Chem. Rev. 1994, 94, 2027-2094. 4. E.F. Aziz, et al., J. Phys. Chem. B, 2007, 111, 4440-4445. 5. Venkataramanan, M. R. Srinivasan, H. L. Bhat, J. Raman Spect., 1994, 25, 805-811 6. I. Lauermann,T. Kropp, D. Vottier, A. Ennaoui, .E. F. Aziz, ChemPhysChem, in print

33

XAFS study of structural position of Mn impurity in SrTiO3
A.I. Lebedev1, I.A. Sluchinskaya1, A. Erko2
1 2

Physics Dept., Moscow State University, Moscow, 119992, Russia BESSY GmbH, Albert Einstein Str. 15, 12489 Berlin, Germany

It has been generally accepted for many years that Mn impurity atoms substitute for Ti sites in as-grown strontium titanate and are there in Mn4+ oxidation state [1]. Annealing in reduced atmosphere could transform Mn ions into lower oxidation states (Mn3+, Mn2+) [2]. Recently Lemanov et al. [3] revealed unusually strong dielectric relaxations in Mn-doped SrTiO3 at low temperatures. These relaxations were explained by a model, in which Mn2+ ion at the Ti site form a defect of {MnTi2+-O-} type. Later the other group of authors [4-8] found the conditions, in which the Mn impurity can be incorporated into the Sr sites and stay there in Mn2+ oxidation state. In such samples all unusual dielectric phenomena were observed; in samples with Mn atoms located at the Ti sites these phenomena were absent. To explain dielectric properties of Mn-doped samples these authors proposed that Mn2+ ions located at Sr sites are off-centre [6]. The purpose of this work was to study the local environment of Mn impurity atoms in SrTiO3 prepared in different conditions to determine directly their structural position. Two samples with a nominal composition of (Sr0.97Mn0.03)TiO3 and Sr(Ti0.97Mn0.03)O3 were prepared by solid-state reaction from SrCO3, nanocrystalline TiO2 and Mn(CH3COO)2·4H2O. Reagents were weighed in necessary proportions, grinded and annealed in air at 1100°C for 8 h. The intermediates were grinded again and annealed finally in the same conditions. One of the samples was additionally annealed in air at 1500°С for 1 h. X-ray absorption spectra were collected in fluorescent mode at the Mn K edge (6.539 keV) on the station KMC-2. The intensity of monochromated radiation was measured with an ionisation chamber; the intensity of fluorescent radiation was measured by an energydispersive RÖNTEC detector. For each sample 5–6 spectra were recorded at 300 K, they were then independently processed and the obtained spectra were finally averaged. The EXAFS data analysis was performed in the traditional way using the models that assume on-centre and off-centre positions of Mn atoms at Sr and Ti sites. EXAFS spectra for Sr(Ti0.97Mn0.03)O3 sample annealed at 1100°C and for (Sr0.97Mn0.03)TiO3 sample annealed at 1500°C are shown in Fig. 1. The conditions of preparation of these samples enable us to assume that Mn impurities enter into different sites and are in different oxidation states. The analysis of EXAFS spectra for Sr(Ti0.97Mn0.03)O3 sample annealed at 1100°C showed that the spectra are well described by a model, in which Mn atoms enter the Ti sites (see Table). To get a good agreement between experimental and calculated spectra (Fig. 1a) it was necessary to take into account two additional multiple scattering paths: Mn-O-O-Mn and Mn-Ti-O-Mn. The analysis of EXAFS spectra for (Sr0.97Mn0.03)TiO3 sample annealed at 1500°C appeared more difficult. These spectra could not be described within the model with on-centre Mn atom at Sr site. Good agreement between experimental and calculated EXAFS spectra (Fig. 1b) was obtained only for the model, in which there are two distances between Mn and Ti atoms in the second shell (3.095 and 3.467 Å). The contribution from the oxygen atoms in the first shell is characterised by a large Debye-Waller factor, thus giving evidence for a wide dispersion of Mn–O distances in this shell. The “local” lattice parameter (3.78 Å) and displacement of Mn atom from the Sr site (0.32 Å) were estimated from the two Mn-Ti distances given above. Small deviation of the “local” lattice parameter from that obtained from X-ray diffraction (3.90 Å) can be explained by the lattice relaxation around the Mn atom, the size of which is smaller than that of Ti one.

34

Fig. 1. EXAFS spectra for two Mn-doped SrTiO3 samples. Squares are experimental points, solid lines are their best fit.

Fig. 2. XANES spectra for three Mn-doped SrTiO3 samples recorded at the Mn K edge.
35

Table. Structural parameters obtained from the EXAFS data analysis for three samples. Sample Sr(Ti0.97Mn0.03)O3, 1100°C (Sr0.97Mn0.03)TiO3, 1100°C (Sr0.97Mn0.03)TiO3, 1500°C Shell 1 2 3 1 2 3 1 2 3 Ri, Å 1.916 3.313 3.67 1.914 3.38 3.70 2.32; 2.86 3.095; 3.467 3.84 σ2i, Å2 0.002 0.001 –0.006 0.002 0.002 –0.004 0.031 0.008 0.014 Atom O Sr Ti O Sr Ti O Ti Sr

XANES region of the X-ray absorption spectra for Mn-doped samples are shown in Fig. 2. Their comparison shows that absorption edges for (Sr0.97Mn0.03)TiO3 sample annealed at 1500˚C and for Sr(Ti0.97Mn0.03)O3 one are shifted by ~7 eV. This proves that Mn atoms in these two samples not only occupy two different crystallographic positions, but also are in two different oxidation states. The absorption edge energy for the sample, in which Mn is at the Sr site, is lower than for the sample, in which it is at the Ti site. This means that the Mn atom at the Sr site has lower oxidation state. Comparison of the observed shift with that obtained in [9] between Mn2+ and Mn4+ oxidation states (7.9 eV) enable to suppose that the oxidation states of Mn atoms in SrTiO3 are +2 for the Sr site and +4 for the Ti site. XANES spectrum for (Sr0.97Mn0.03)TiO3 sample annealed at 1100°C (Fig. 2) can be regarded as a superposition of XANES spectra for two samples, for which the EXAFS analysis was made. This means that Mn in this sample is present in both oxidation states, and so actually one can speak only about predominant incorporation of Mn impurity in this or that site of the lattice. From the presented results it follows that from two models proposed for an explanation of unusual dielectric phenomena in Mn-doped SrTiO3 samples, the model with off-centre MnSr2+ ion seems to be more adequate. Future systematic investigations are needed to study the effect of the preparation conditions on crystallographic positions of Mn in SrTiO3 and to develop a method of quantitative determination of Mn concentration and its oxidation state at different sites. Acknowledgement. Two authors (A.I.L. and I.A.S.) are grateful to Russian-German laboratory of BESSY for hospitality and financial support. The work was also supported by the RFBR grant No. 08-02-01436. References [1] K.A. Müller. Phys. Rev. Lett. 2, 341 (1959). [2] R.A. Serway, W. Berlinger, K.A. Müller, R.W. Collins. Phys. Rev. B 16, 4761 (1977). [3] V.V. Lemanov, E.P. Smirnova, A.V. Sotnikov, M. Weihnacht. Phys. Solid State 46, 1442 (2004). [4] A. Tkach, P.M. Vilarinho, A.L. Kholkin. Acta Mater. 53, 5061 (2005). [5] A. Tkach, P.M. Vilarinho, A.L. Kholkin. Ferroelectrics 304, 87 (2004). [6] A. Tkach, P.M. Vilarinho, A.L. Kholkin. Appl. Phys. Lett. 86, 172902 (2005). [7] A. Tkach, P.M. Vilarinho, A.L. Kholkin. Acta Mater. 54, 5385 (2006). [8] A. Tkach, P.M. Vilarinho, A.L. Kholkin, A. Pashkin, S. Veljko, J. Petzelt. Phys. Rev. B 73, 104113 (2006). [9] A. Hagen, L. Mikkelsen. In: Solid State Electrochemistry (Proceedings of the 26th Risø Int. Symp. on Materials Science), Risø National Laboratory, Roskilde, Denmark, p. 197 (2005).

36

In-situ XPS study of adsorbed oxygen species on (La,Ba,Sr)(Fe,Co)O3-d perovskites Rotraut Merkle, Joachim Maier Max-Planck-Institut für Festkörperforschung, Stuttgart Introduction Mixed conducting perovskites are used as cathode materials in Solid Oxide Fuel Cells (SOFC). In order to elucidate the mechanism of the oxygen incorporation, we performed impedance spectroscopy on well-defined dense microelectrodes [1]. These experiments are complemented by DFT calculations [2]. Knowledge about the type of adsorbed oxygen species (superoxide O2- or peroxide O22-, or atomic charged adsorbate species O-) and their concentration is important for a mechanistic understanding of the oxygen incorporation reaction into these materials. To examine such oxygen adsorbates, the samples have to be studied by XPS under a certain oxygen pressure (typically up to 1 mbar) which required the use of synchrotron radiation. The variation of sample temperature 300-600°C, pO2, p(H2O), as well as depth resolution is important for a reliable assignment of the observed peaks. The oxygen adsorbate coverages are expected to be rather small because of their negative charge (buildup of a surface dipole layer) [3], but dependent on applied DC bias. (La,Ba,Sr)(Fe,Co)O3-d dense thin films were prepared by pulsed laser deposition on YSZ oxide-ion conducting substrates and equipped with a counter electrode so that an electrical current can be drawn through the film corresponding to a steady flux of oxygen incorporated into or generated at the film surface. Thus fuel cell operation can be mimicked and adsorbate concentrations can possibly be modified by the applied electrical bias [3]. Results
0.5 mbar O2 T = 300 °C cps Ia Ib O1s 0.5 mbar O2 T = 350 °C (after 475 °C) cps Ia Ib O1s

II

II

0.5 mbar O2 binding energy / eV T = 400 °C cps

T increase in dry O2 decrease of peak II
cps

0.5 mbar O2 + H2O binding energy / eV T = 350 °C

addition of H2O:
0.5 mbar O2 binding energy / eV T = 475 °C

O2

H2O

increase of peak II strong increase and slight shift (BE +0.25 eV) of peak Ib
525

0.5 mbar O2 + H2O binding energy / eV T = 350 °C after 10 min cps 540

cps 540

535

530

535

530

525

binding energy / eV

binding energy / eV

Figure 1: In-situ XPS of dense La0.6Sr0.4CoO3-d "LSC" films at different temperatures and under applied gas pressure (pO2, p(H2O)). As exemplified in Fig.1 for LSC, the O1s region typically contains at least two peaks: II at high BE and Ia+Ib at lower BE. Peak II decreases on increasing temperature, and can at least partially be restored by H2O exposure. During hydratation, also peak Ib seems to increase, but its BE-shift indicates that a new component develops. Based on binding energy, T- and p(H2O)-dependence, peak II and the new component close to Ib are now assigned to OH-. Depth-resolved spectra (not shown here) indicate that oxygen peak II is surface-related. They also indicate that the surface region is depleted in La3+ and enriched in Sr2+. The lower

37

average cation charge in the Sr-rich surface layer can at least partially be compensated by the presence of H+ in the hydroxide groups.
0.5 mbar O2 T = 300 °C cps O1s II Ib Ia fresh fresh film: Sr excess (La0.4Sr0.6) Sr3d II I 0.5 mbar O2 energy / eV 15 h binding T = 350 °C aged at 750°C II I aged binding energy / eV 15 h film: La excess (La0.7Sr0.3) aged fresh La4d II I II I fresh

resistance change of La0.6Sr0.4CoO3-δ electrode during ageing at 750°C in air:

binding energy / eV 15 h aged

t / h vs Rs / Ohm

cps

535

530 binding energy / eV

525

135 binding energy / eV

130

110

105

100

0

5

binding energy / eV

10 time / h

15

Figure 2: In-situ XPS of freshly prepared and aged (15 h at 750 °C) dense La0.6Sr0.4CoO3-d films (Ekin = 200 eV). Right: evolution of La0.6Sr0.4CoO3-d electrode resistance during ageing. Fig. 2 demonstrates the effect of thermal ageing on the film properties. The aged film surface (Ekin = 200 eV = 0.7 nm depth) not only shows a different oxygen peak shape, but also a decrease of Sr and increase of La compared to the fresh film (quantitative analysis of cation ratios is in progress). The right panel shows the significant increase of the electrode resistance (which is determined by the oxygen incorporation rate at the LSC film surface) during the 15 h ageing time. Since the oxygen incorporation reaction is sensitive to the material's composition and defect concentrations directly at the surface, this increase must be interrelated to the observed cation redistribution. Measurements under applied electrical current (which is expected to increase adsorbate concentrations, especially under anodic conditions when oxygen is generated at the perovskite surface) were also performed, but the samle design still requires improvements. A large temperature gradient between backside counter electrode and frontside working electrode (LSC film) repeatedly lead to contact loss. Thus, so far no peak was detected that can explicitly be assigned to adsorbed oxygen species. Summary and Outlook The presence of significant amounts of OH- even at 500 °C and its accumulation close to the surface rise the question of proton involvement in the oxygen incorporation reaction (formation of HO2- etc.). This motivates further electrochemical measurements of oxygen reaction kinetics under varying p(H2O). The different surface cation compositions in fresh and aged films will help to understand the frequently observed performance decrease of these electrode materials after prolonged heating. To overcome the contact problem in the DC current measurements, a new sample design with both electrodes on the frontside will be used in the next beamtime. Then the measurements are expected either to yield evidence for molecular oxygen adsorbates, or - if not - to give an upper limit for their concentration. References: [1] F. S. Baumann, J. Fleig, H.-U. Habermeier, J. Maier, Solid State Ionics 177 (2006) 1071 [2] E. A. Kotomin, Y. A. Mastrikov, E. Heifets, R. Merkle, J. Fleig, J. Maier, A. Gordon, J. Felsteiner, ECS Transactions 13 (2008) 301 [3] J. Fleig, Phys. Chem. Chem. Phys. 7 (2005) 2027; J. Fleig, R. Merkle, Phys. Chem. Chem. Phys. 9 (2007) 2713
38

In-situ X-Ray Diffraction Study of the Influence of Sulphur on Phase Formation during the Heat-Treatment of Porphyrin-based Oxygen Reduction Catalysts U.I. Kramm1, G. Zehl1, I. Zizak2, I. Herrmann1, I. Dorbandt1, P. Bogdanoff1, S. Fiechter1 Helmholtz-Zentrum Berlin für Materialien und Energie (HZB) Lise-Meitner-Campus Wannsee, Glienicker Str. 100, D-14109 Berlin 2 Conrad-Röntgen-Campus Adlershof, Albert-Einstein-Str. 15, D-12489 Berlin
1

1. Introduction During the last decades polymer electrolyte membrane fuel cells (PEM-FC) have been optimized using platinum based catalysts. However, platinum is scarce and expensive. Hence, catalyst costs have a significant impact on the market price of fuel cell units. Therefore, the continuously growing Pt price is impeding the production of less expensive PEM fuel cells and protracting the commercial breakthrough of this technology. Thus, there is a soaring interest in the development of materials with reduced amounts of noble metals and, most promising, of Pt free oxygen reduction catalysts. It is known for decades that metal chelates can be used as catalysts for the oxygen reduction reaction (ORR) [1]. Using rotating disc electrode (RDE) measurements it has been shown recently, that heat-treated N4-metallomacrocycles reveal catalytic activities at 0.75 V even comparable to commercial Pt/C catalysts [2]. Henceforth, they are seriously considered as promising noble metal free alternatives to Pt based catalysts. However, the pyrolysis of metal-chelate compounds as a necessary preparation step leads to glassy carbon like carbonization products and only a small number of catalytic centres can participate in the ORR. Therefore, current preparation approaches are aimed to increase this side density. To prevent sintering effect, carbon blacks were impregnated with the metal chelates followed by a heat-treatment at temperatures of ≥ 700°C in an inert atmosphere. This preparation approach, however, is only suited for low amounts of metal chelate due to the limited support capacity of carbon black. Above a critical loading the excess chelate will sinter as in the case without additional carbon support, leading to similar activity losses. Therefore, using impregnation techniques the catalytic activity cannot be raised beyond a certain level because the catalysts will always be diluted by a specific amount of the auxiliary carbon [3,4]. To overcome this limitation the so called foaming agent technique was developed at the HelmholtzZentrum Berlin which works without the addition of auxiliary carbon [2,5,6]. Porphyrins are pyrolysed in the presence of iron-oxalate (and sulphur). The oxalate decomposes under release of CO and CO2. Simultaneously the porphyrin melts and carbonizes; therefore, the gaseous products cause a foaming effect to the in-situ forming carbon matrix. After the inert heating step is completed the product is allowed to cool down. Finally, a leaching process is applied to remove inorganic by-products. It was found that the addition of sulphur leads to higher catalytic activity towards the oxygen reduction and to a more complete removal of catalytically inactive inorganic metal constituents [2]. To get a better understanding of the role of sulphur during pyrolysis, the heat-treatment process of the porphyrin/Fe oxalate and the porphyrin/Fe oxalate/sulphur catalyst precursors (with optimized sulphur content) was investigated by insitu high temperature X-ray diffraction analysis (HT-XRD). 2. Experimental In-situ X-ray diffraction measurements were performed at BESSY II beamline KMC-2 using a stainless steel reaction chamber with Kapton® windows at the beam entry and exit slits and equipped with an electric graphite heater encapsulated in pyrolytic boron nitride (pBN) from Tectra GmbH. This newly developed reaction chamber allows measurements under constant Ar gas flow at reduced pressures in a temperature range from RT to 800°C. A two dimensional detector array (HiStar – Bruker AXS) was used to record the spectra. Detector, sample and synchrotron radiation beam were aligned to meet Bragg-Brentano geometry. Measurements were carried out under Ar flow of 200 ml/min at reduced pressure of p = 400 mbar with synchrotron radiation energy of 8.731 keV. Calibration of the system was done using an alumina standard for the used beam energy. A detailed description of the catalyst preparation is given elsewhere [7]. Briefly, an amount of 1.3 mmol FeTMPPCl is mixed with 28.6 mmol iron oxalate dihydrate in a mortar until a homogeneous precursor mixture is obtained. For the preparation of sulphur containing precursor mixture 1.2 mmol sulphur (S8) are

39

grounded previously before mixing with oxalate and porphyrin. The powders of the precursor mixtures were pelletized (Ø: 10 mm) and placed onto a pBN sample holder. 3. Results A detailed description of the chemical processes involved in the foaming agent technique can be found elsewhere based on the results of TG-MS measurements [2,5,6]. As described above, our Foaming Agent Technique (FAT) enables the preparation of a highly porous carbon structure with embedded catalytic centres. As already published for the CoTMPP system (plus Fe oxalate dihydrate and sulphur) [7], also for FeTMPPCl and for H2TMPP the kinetic current densities in RDE measurements were found to increase by one order of magnitude compared to the sulphur free catalysts, and hydrogen peroxide formation is decreased to less then 5 %. In RRDE measurements sulphur containing catalysts reveal a catalytic activity towards ORR in the same order of magnitude as commercial Pt/C catalysts. Beside this, the method allows us to achieve the fourfold of catalytic site density compared to similar non noble metal catalysts [8]. However, for further optimization it is essential to understand in which way the sulphur affects the pyrolysis of Foaming Agent catalysts. Therefore, the applied temperature range and the involved phase transformations have to be determined and differences compared to the sulphur free precursor should be compiled. Our newly developed in-situ HT-XRD cell allows measurement conditions comparable to the processes involved in the standard catalyst preparation. In Figure 1 the HT-XRD measurement of sulphur free (a) and sulphur containing (b) precursor mixtures are shown in a temperature range from RT to 800°C.

Figure 1: HT-XRD measurements of porphyrin/Fe oxalate precursors heat-treated without (a) and with the addition of sulphur. Figure 1 clearly demonstrates that above temperatures of approx. 350°C the processes involved during heat-treatment start to differ. To investigate this in more detail the spectra for selected heat-treatment temperatures were extracted and plotted against 2theta (as calculated for Cu Kα wave length) in Figure 2. To improve the signal to noise ratio the average of 10 consecutive measurements around the designated temperature was used. Due to the continued heating-up the error in temperature is slightly increased but still restricted to ≤ 5 °C. The room temperature spectra of both precursor mixtures are dominated by the reflexes related to iron oxalate dihydrate whereas the porphyrin (and sulphur) did not contribute to the spectra. This can be explained by the much higher amount of oxalate compared to porphyrin (and sulphur). As confirmed by HT-XRD and TG-MS earlier, the oxalate releases its crystal water within T-range from 140 °C to 190 °C (e.g. Ref. 5). Up to 350 °C, diffractograms remain similar, independently of the precursor mixture. Above this temperature structural changes are different for both precursors. Analysing first the heat treatment process of the sulphur free precursor (Fig. 2a) formation of Wuestite and Hematite can be followed in the temperature range from 435 °C to 545 °C. In parallel to a second decomposition step these oxides are reduced to a Cohenite modification, whereas above 700 °C a hightemperature phase of iron carbide is visible. This high-temperature phase, however, transforms back into Cohenite or into elemental iron and graphite during the subsequent cooling process. In contrast, the addition of sulphur (Fig. 2b) affects spectra exhibiting amorphous behaviour in a temperature range between 435 °C to 545 °C. At this temperature also Wuestite can be detected but beside this there is another phase with a main reflex at 2theta = 44.4° that might be related to an iron carbide phase. Above 545 °C this signal is decreasing and spectra of the sulphur added sample are mainly amorphous depicted in much smaller intensities of the present reflexes (Hematite and Magnetite).

40

Taking into account the findings of Grabke et al. [9] it is likely that at temperatures > 435 °C the sulphur is firmly bonded to iron, surpressing a further formation of iron carbide. Indeed, also our results of HT-XRD measurements evidence this effect. Above 600 °C the amount of iron carbide is reduced. Therefore, an amorphous carbon structure is formed that was found to enhance the catalytic properties.

Figure 2: X-ray diffractograms of porphyrin/Fe oxalate precursors heat treated without (a) and after addition of sulphur. 4. Conclusions and outlook The use of an improved reaction chamber for in-situ HT-XRD investigation under gas flow conditions allowed the study of processes involved in activation of porphyrin-type oxygen reduction catalysts under standard preparation conditions. The comparison of sulphur free and sulphur containing precursors let us conclude that the effect of “iron blocking” by sulphur preventing the formation of iron carbides starts at temperatures of 435 °C. This effect significantly enhances activity of Pt-free oxygen reduction catalysts. It was monitored for the first time by in-situ HT-XRD analysis. Further work is planned to investigate the pyrolysis processes of precursors using cheaper starting materials. 5. References

(1) Jahnke, H.; Schönborn, M.; Zimmermann, G. Topics in Current Chemistry 1976, 61, 133. (2) Herrmann, I.; Koslowski, U.; Radnik, J.; Bogdanoff, P.; Fiechter, S. ECS Trans. 2008, 13, 145. (3) Bouwkamp-Wijnoltz, A. L.; Visscher, W.; Veen, J. A. R. v. Electrochimica Acta 1998, 43, 3141. (4) Herrmann, I. Entwicklung und Optimierung neuer Präparationsverfahren für Übergangsmetallbasierte Elektrokatalysatoren für die Sauerstoffreduktion. Doktorarbeit, Freie Universität Berlin, 2005. (5) Bogdanoff, P.; Herrmann, I.; Hilgendorff, M.; Dorbandt, I.; Fiechter, S.; Tributsch, H. J. New. Mat. Electrochem. Systems 2004, 7, 85. (6) Herrmann, I.; Bogdanoff, P.; Schmithals, G.; Fiechter, S. ECS Trans. 2006, 3, 211. (7) Koslowski, U.; Herrmann, I.; Bogdanoff, P.; Barkschat, C.; Fiechter, S.; Iwata, N.; Takahashi, H.; Nishikoro, H.; Abs-Wurmbach, I. ECS Trans. 2008, 13, 125. (8) Koslowski, U.; Abs-Wurmbach, I.; Fiechter, S.; Bogdanoff, P. J. Phys. Chem. C 2008, 112, 15356 (9) Grabke, H. J.; Moszynski, D.; Müller-Lorenz, E. M.; Schneider, A. Surface and Interface Analysis 2002, 34, 369.
41

Interaction of CO2 with thin nickel oxide layers on Cu(111)
M. P. A. Lorenz, R. Streber, M.-M. Walz, A. Bayer, and H.-P. Steinrück Lehrstuhl für Physikalische Chemie II, Universität Erlangen-Nürnberg, Egerlandstraße 3,91058 Erlangen Nickel oxide is an important material for optical storage media and in heterogeneous catalysis. However, under ambient conditions, impurities such as carbonates can easily be formed by reaction with CO2. Behm and Brundle [1] for instance reported the formation of a carbonate layer on Ni(100) by simultaneously dosing molecular oxygen and CO2. As the catalytic activity of nickel oxide could be strongly influenced by the formation of carbonate, we studied the interaction of CO2 with thin NiO layers on Cu(111) by high resolution XPS applying synchrotron radiation. Cu(111) acts as inert substrate for the growth of NiO layers with defined thickness. The experiments were performed at beamline U49/2-PGM1, using a transportable apparatus described elsewhere [2]. To form NiO, first the appropriate amount of nickel was evaporated onto the Cu(111) surface at 120 K, followed by heating to 300 K to produce flat Ni films [3]. Subsequently, the Ni films were completely oxidised by dosing 120 L of molecular oxygen at 300 K, followed by flashing to 500 K. As result, no metallic component could be detected in the Ni 2p XP spectra (data not shown). Figure 1 shows a series of C 1s XP spectra, recorded while dosing CO2 onto 2 ML NiO/Cu(111) at 110 K. Two characteristic peaks arise at binding energies of 289.0 and 291.1 eV, with the former attributed to carbonate according to Behm and Brundle [1]. Its thermal stability (see Figure 2) rules out a weakly bonded physisorbed species. The other peak at 291.1 eV is assigned to physisorbed CO2; its intensity is drastically reduced, when decreasing the CO2 partial pressure (green spectrum). The corresponding O 1s binding energies are 531.6 eV (carbonate) and 534.8 eV (physisorbed CO2) (Figure 2 b).

Figure 1: Series of C 1s XP spectra while dosing CO2 onto 2 ML NiO/Cu(111) at 110 K; carbonate saturation coverage: 0.14 ML.

Subsequently, we also studied the thermal evolution of the carbonate layers during heating to elevated temperatures, by recording C 1s and O 1s XP spectra. As can be seen in Figure 2 a (C 1s) and 2 b (O 1s), the components at binding energies of 291.1 and 534.8 eV disappear completely up to 200 K, due to desorption, further confirming their assignment to physisorbed CO2. In contrast, the component attributed to carbonate is thermally much more stable, although its intensity decreases continuously with increasing temperature. But even at 500 K, a temperature range, where NiO already starts to decompose, small amounts of carbonate are still left on the surface. Therefore it is not possible to clean NiO layers, which were contaminated by carbonate (formed from e.g. CO2 from the residual gas) by simply heating to elevated temperatures.

42

Figure 2: Selected a) C 1s and b) O 1s XP spectra of CO2 and carbonate adsorbed on 2 ML NiO/Cu(111) recorded during heating to denoted temperatures.

The quantitative analysis of the C 1s data (Figure 3) clearly shows a nearly linear decay of the carbonate coverage with raising temperature. Note that the increase of the carbonate coverage for temperatures below 150 K is either caused by further formation of carbonate from CO2 and surface O and/or by reduced damping as consequence of desorbing CO2 in this temperature regime. The analysis of the O1s data (not shown) shows an overall similar behaviour. InteresFigure 3: Quantitative analysis of C 1s XP spectra of tingly, at 600 K the O 1s peak assigned to CO2 and carbonate adsorbed on 2 ML carbonate (not shown) shows still ~25 % of NiO/Cu(111) during heating. its original intensity, whereas the corresponding C 1s intensity in Fig. 3 has vanished. A possible explanation for this apparent excess of “carbonate” in the O 1s data could be the simultaneous formation of hydroxyl groups. It is known [4] that NiO surfaces are very reactive towards water, e.g., from residual gas, forming Nickel hydroxide. Dosing water onto NiO layers on Cu(111) indeed lead to the formation of NiOH, with the corresponding O 1s peak at the same binding energy (~531.6 eV) than that for the carbonate (data not shown).

Acknowledgements: The project was supported by the BMBF through grant 05 ES3XBA/5 and by the DFG through the Cluster of Excellence “Engineering of Advanced Materials”. We like to thank the BESSY staff for their support during beamtime.

[1] R. J. Behm, C. R. Brundle, Surf. Sci. 255 (1991) 327 [2] R. Denecke, M. Kinne, C. M. Whelan, H.-.P. Steinrück, Surf. Rev. Lett. 9 (2002) 797. [3] R. Domnick, G. Held, H. Koschel, Ch. Ammon, H.-P. Steinrück, Surf. Sci. 482-485 (2001) 1292 [4] N. Kitakatsu, V. Maurice, C. Hinnen, P. Marcus, Surf. Sci. 407 (1998), 36-58

43

Propylene oxidation over palladium: Operando XPS-MS study
Vasily V. Kaichev1, Andrey V. Matveev1, Vladimir V. Gorodetskii1, Valerii I. Bukhtiyarov1, Bernard E. Nieuwenhuys2
1

Boreskov Institute of Catalysis of SB RAS, Novosibirsk, Russia, E-mail: vvk@catalysis.ru 2 Leiden Institute of Chemistry, Leiden, the Netherlands Palladium is well known as the most active catalyst in a number of key industrial

processes such as the total oxidation of hydrocarbons in automotive exhausts, the natural gas combustion in gas-powered turbines, as well as selective hydrogenation of alkynes to alkenes. Its practical importance has stimulated a vast amount of works devoted to the study of adsorption, oxidation and hydrogenation of hydrocarbons over palladium. To date, most of the works was performed at high pressure under realistic catalytic conditions by monitoring the reaction products in the gas phase. Unfortunately, this approach has difficulty in unraveling the details of the reaction mechanisms that are very important for the purposeful synthesis or improvement of catalysts. Other studies were carried out under UHV conditions by surface science techniques. Certainly, these UHV studies significantly developed our understanding of the mechanisms for different catalytic reactions, however, the questions about the state of the active catalyst, as well as the reasons for catalyst activation and deactivation, remain very often under discussion. Indeed, the active state of a catalyst exists only during the catalysis and can «die» in UHV. For investigation of really «living» catalysts, different spectroscopic techniques must be applied at elevated pressure in situ, i.e., while the catalysis takes place [1,2]. Moreover, it is essential to combine the spectroscopic characterization of a catalyst surface with simultaneous monitoring of its catalytic performance. In the last several years, such an approach is usually called as operando. Here we demonstrate how the application of the operando techniques can be used to provide additional insight into the mechanism for heterogeneous catalytic reactions. The aim of our work was to elucidate the mechanism of activation and deactivation of Pd surface in the propylene oxidation. Pd is chosen as a highly active metal, used in three-way catalysts for utilization of the tailing gases. Propylene is a good model unsaturated hydrocarbon, which can be easily studied by surface science methods. We use a combination of in situ X-ray photoelectron spectroscopy and mass-spectrometry. XPS is one of the powerful tools to investigate both the surface composition and the nature of adsorbed species. When in situ XPS is coupled with mass-spectrometry, it becomes a particularly effective operando technique, which makes it possible to correlate surface properties with the catalytic performance. The experiments were performed at the ISISS beam line at BESSY in Berlin. The construction of this setup was described in detail elsewhere [2]. A differentially pumped system of electrostatic lens is the key feature of this setup, which allows investigation of the 1
44

catalytic reactions in situ in the mbar pressure range. The gas-phase analysis was carried out using a quadruple mass-spectrometer connected through a leak valve to the experimental cell. A Pd(551) single crystal was used as a catalyst. It was mounted onto a sapphire sample holder with a SiC plate heated from the back with a NIR laser. The main advantage of this heating method is the absence of any hot details, which may have a high catalytic activity. The sample temperature was monitored with a chromel-alumel thermocouple spot-welded directly to the crystal edge. The partial pressure of propylene during the experiments was 5x10-4 mbar, the propylene/oxygen ratio was 1:1, 1:10, and 1:100. The reaction was studied using the temperature-programmed-reaction (TPR) techniques based on mass-spectrometric analysis of the gas phase during heating and subsequent cooling with the constant rate of 1 K/s in the range between 100 and 500ºC. The TPR results show a very complex kinetic behavior of this reaction. In all cases, the main reaction products were CO2 and water. Fig. 1 shows the CO2 yield during the complete heating/cooling cycle as a function of the sample temperature for the different propylene/oxygen ratios. One can see at least three temperature hysteresis loops in the TPR curves. The most pronounced hysteresis loop was observed at low temperature. For example, during a heating ramp, the activity in propylene oxidation sharply increases at 285, 230 and 210ºC, for the propylene/oxygen ratio 1:1, 1:10, 1:100 correspondingly, whereas a quick decrease in the activity is observed at 150160ºC during the subsequent cooling ramp. Two other hysteresis loops appear in the oxygen excess. Certainly, such dynamic behavior is determined by strong change in the surface composition. It is well-known that depending on reaction temperature, pressure and hydrocarbon/oxygen feed, palladium surface can be covered with oxide, carbon or hydrate layers. In order to determined the reason, which causes the first hysteresis loop, we recorded under the same conditions the C1s, O1s, and Pd3d core-level spectra during stepwise heating from 100 to 250, 300, 400, and 500ºC and during the following cooling in the same manner. All the experiments showed consistent results, and therefore only the C1s spectra obtained in the equimolar propylene/oxygen mixture are presented bellow. The C1s spectrum obtained at 100ºC consists of at least 5 marked features. A small peak at 285.7 eV is due to CO, which adsorbs from the background gas or forms over palladium surface as a product of the propylene oxidation. Two weak features at the higher binding energy could be attributed to more strongly oxygenated carbon species like formate, acetate, etc. Two major spectral features at 283.9 and 284.6 eV could be attributed to carbon species located in the near-surface region. The dissolution of carbon in the Pd bulk and following formation of the PdC surface phase [1] at low temperature was also evidenced by Pd3d5/2 spectra (not shown). After heating to 250ºC, both the C1s and Pd3d5/2 spectra 2
45

essentially remained constant. In contrast, at 300ºC all the C1s features and PdC component in the Pd3d5/2 spectrum disappear that points out to full removing of carbon from the nearsurface region. During cooling to 250ºC, we again observed a similar C1s spectrum, but with very low intensity. Further cooling leads to restoration of the C1s spectrum as well as of the PdC component in the Pd3d5/2 spectrum.
0.10

C1s XPS Intensity [arb. un.]

P(O2) = 5x10-4 mbar

PdCx CO 250 C 300 C
0 0

1:1
x10

0.09

1:1

Intensity [arb. un.]

C1s / Pd3d5/2 ratio

+75

0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01

Heating

P(O2)= 5x10-3 mbar

+25

250 C

0

P(O2)= 5x10-2 mbar

Cooling

100 C
282 284 286 288

0

0.00

100

200

300

400

500

100 200 300 400 500

Temperature [C]

Binding Energy [eV]

Temperature [ C]

0

Figure 1. Temperature dependence of CO2 yield in the propylene oxidation for different propylene/oxygen ratios (left) as well as in situ C1s spectra (center) and temperature dependence of C1s intensity observed in the equimolar reaction mixture (right). Comparing the XPS and TPR data, we can postulate that the first activity hysteresis in the propylene oxidation is closely coupled with the hysteresis in the carbon content in the near-surface region. The state with the lower activity corresponded to the PdC surface phase, which is formed over the palladium surface due to carbon deposition and the following partial carbon dissolution in the Pd bulk. The PdC phase is formed even in an oxygen excess at low temperature, leading to the immediate catalyst deactivation. During heating, the carbon cleanoff reaction with oxygen proceeds, which restores the adsorption properties of the palladium surface, and, as a consequence, the propylene oxidation starts above 210-285ºC. Another interesting point of this study is an observation of strong dependence of the propylene conversion on the propylene/oxygen ratio. The higher level of propylene conversion was detected in the equimolar mixture. This phenomenon is explained by the inhibiting effect of oxygen adsorption and formation of the surface oxide. High oxygen content extinguishes the reaction. It means that the presence of oxygen-free metal surface, where propylene can adsorb and dissociate, is necessary for high catalytic activity. References [1] V.I. Bukhtiyarov, V.V. Kaichev, I.P. Prosvirin, Topics Catal. 32 (2005) 3. [2] H. Bluhm, M. Hävecker, A. Knop-Gericke, E. Kleimenov, R. Schlögl, D. Teschner, V.I. Bukhtiyarov, D.F. Ogletree, M. Salmeron, J. Phys. Chem. B 108 (2004) 14340. 3
46

An operando XPS-MS study of the oscillations in the propane oxidation over nickel foil
Vasiliy V. Kaichev1, Alexey Yu. Gladky1, Igor P. Prosvirin1, Valerii I. Bukhtiyarov1, Raoul Blume2, Michael Hävecker2, Axel Knop-Gericke2, Robert Schlögl2
1

Boreskov Institute of Catalysis of SB RAS, Novosibirsk, Russia, E-mail: vvk@catalysis.ru
2

Fritz-Haber-Institut der MPG, Berlin, Germany

One of the most interesting and unusual phenomena of catalysis is the rate oscillations [1]. To date, approximately 70 oscillating heterogeneous catalytic systems are known. A classic example is oscillations in CO oxidation over noble metals, which were intensively studied during the last thirty years. At present, a special attention is attracted to oscillations in the oxidation of light alkanes over transition metals. For example, the regular self-sustained oscillations were observed in the methane oxidation over Ni, Co and Pd supported and unsupported catalysts in oxygen-deficient conditions at ambient pressure [2]. Similar oscillations were also observed in the ethane oxidation over Ni and Co foils. In our previous work, it was found that the propane oxidation over Ni can proceed in a self-oscillation regime as well [3,4]. The characteristics of all these oscillations are sufficiently similar to suggest a common origin for the oscillatory behaviour. The stable and repeatable oscillations appear after an induction period of tens minutes, when the catalysts demonstrate very low activity. It points out that the reaction kinetics alone cannot be responsible for the oscillations. Also the oscillation mechanisms, which are in general based on UHV studies [1], cannot be simply extrapolated to the high-pressure conditions. For elucidating the oscillation mechanism in the oxidation of light alkanes over transition metals, it is necessary to apply some operando techniques. Unfortunately, self-oscillations in the methane and ethane oxidation were observed only at atmospheric pressure [2], where most of the surface-sensitive techniques, including XANES and XPS, can not be used. Therefore, as a case reaction, we chose the catalytic oxidation of propane over Ni, where regular oscillations with periods of several minutes can be observed in the mbar pressure range [3,4]. Here we present first results of an operando study of the oscillations in the propane oxidation over a Ni foil. We used time-resolved X-ray photoelectron spectroscopy in situ, i.e., while oscillations take place, simultaneously with mass-spectrometry (MS) for monitoring gas-phase components. In situ XPS is one of the most useful tools to investigate both the surface composition and the nature of adsorbed species on the catalyst surface. When the in situ XPS is coupled with mass-spectrometry, it becomes a particularly effective operando technique, which makes it possible to correlate the surface properties with the catalytic performance.
47

The experiments were performed at the ISISS beam line at BESSY in Berlin. The construction of this setup was described in detail elsewhere [5]. A differentially pumped system of electrostatic lens is the key feature of this setup, which makes it possible to investigate the catalytic reactions in situ in the mbar pressure range. During the experiments, the total pressure of the reaction mixture in the experimental cell was kept at a constant level of 0.5 mbar. The gas-phase analysis was carried out using a quadruple mass-spectrometer (Prizma, Balzers) connected through a leak valve to the experimental cell. A rectangular piece of a nickel foil (0.125×6×7 mm, purity 99.99%, obtained from Advent) was used as a catalyst. It was mounted onto a sapphire sample holder with a SiC plate heated from the back with a NIR laser. The main advantage of this heating method is the absence of any hot details, which may have a high catalytic activity. The sample temperature was monitored with a chromelalumel thermocouple spot-welded directly to the foil edge. In this operando XPS-MS study, the oscillations were observed at temperatures 550650°C and for gas mixtures with the propane/oxygen ratios from 1:1 to 20:1. The period, amplitude and waveform of the oscillations were strongly dependent on the temperature and the propane/oxygen ratio. Usually, the catalyst stayed in an inactive state for the most of time with occasional evolution of H2, CO and H2O. Such product distribution indicates that both the partial and total oxidation of propane occurred over nickel during the active half-period [4]. The periodic changes in the reactant concentration were accompanied by synchronous fluctuations of the catalyst temperature. Figure 1 shows typical oscillations of the reaction rate, which is demonstrated by the time-induced variation of MS signals at m/z = 2 (H2), m/z = 18 (H2O), m/z = 28 (CO) and m/z = 32 (O2). Changes of the catalyst temperature are also detected (Fig.1).
Temperature [ C] 630 20
∆T = 12
2 1 3

Ni2p3/2 XPS Intensity [a.u.]

853.0

610 16

MS Signal [a.u.]

600

3

O1s XPS Intensity [a.u.]

T

855.1

620

529.9

0

12

CO + 4

3 2

8

t = 140 s

H2 + 4 O2 + 1 H2O

4

2 1
848 852 856 860 864 868

t = 900 s

1
526 528 530 532 534

0

300

600

900

1200 1500

Time [s]

Binding Energy [eV]

Binding Energy [eV]

Figure 1. Oscillations of H2O, O2, H2, CO, and temperature (left) as well as Ni2p3/2 and O1s core-level spectra taken during inactive and active half-periods of oscillations (centre and right). The propane/oxygen ratio was 3:1.

48

In order to reveal the nature of the active and inactive state of the catalyst surface, time-resolved XPS spectra were measured in situ, directly during the oscillations. Fig. 1 also shows the Ni2p3/2 and O1s core-level spectra, which were taken when the system periodically passes through three characteristic points corresponding to the inactive and active states marked as 1, 2, and 3. The Ni2p3/2 spectra from the inactive surface (spectra 1 and 3) show the characteristic pattern of NiO with the main Ni2p3/2 line at 855 eV and two satellites at higher (by ~1.5 eV and ~7 eV) binding energies. According to previous XPS studies [4], the first satellite, which looks as a prominent shoulder of the main line, is assigned to NiO, while the second strong broad satellite at 862 eV is typical of Ni2+ compounds like NiO, Ni(OH)2, NiAl2O4, etc. In contrast, the Ni2p3/2 spectrum of the active nickel surface (spectrum 2) consists of a sharp single peak at 853 eV, which corresponds to nickel in the metal state. A wide low-intensive feature at 859 eV in this case is assigned to an energy loss peak due to plasmon excitation. In full agreement with these data, strong changes have been observed in the O1s spectra, when the system periodically passes from the inactive to the active state (Fig. 1). The O1s spectrum from the inactive surface (spectrum 1) exhibits an intense feature at 529.9 eV, which mainly corresponds to oxygen in NiO. Transition to the active state of nickel surface leads to a drop in the O1s intensity (spectrum 2). Again, after subsequent transition to the inactive state, the O1s spectrum is restored in the intensity to the original level (spectrum 3). Thus, in situ XPS spectra clearly indicate that during the inactive half-periods, nickel foil is covered with a nickel oxide layer, which mainly consists of NiO. The transition to the active state is accompanied by the full reduction of nickel oxide to Ni0. It means that the selfoscillations in the propane oxidation over Ni originate due to the periodic oxidation and reduction of the catalyst surface. The high-activity state is associated with metallic nickel, whereas the low-activity state is characterised by the presence of the nickel oxide layer on the catalyst surface. Considering the propane oxidation as a case reaction, we suppose that the oscillations in the oxidation of other light alkanes over transition metals proceed via a similar mechanism. References [1] R. Imbihl, G. Ertl, Chem. Rev. 95 (1995) 697. [2] X. Zhang, C.S.-M. Lee, D.O. Hayward, D.M.P. Mingos, Catal. Today 105 (2005) 283. [3] A.Yu. Gladky, V.K. Ermolaev, V.N. Parmon, Catal. Lett. 77 (2001) 103. [4] A.Yu. Gladky, V.V. Kaichev, V.K. Ermolaev, V.I. Bukhtiyarov, V.N. Parmon, Kinet. Catal. 46 (2005) 251. [5] H. Bluhm, M. Hävecker, A. Knop-Gericke, E. Kleimenov, R. Schlögl, D. Teschner, V.I. Bukhtiyarov, D.F. Ogletree, M. Salmeron, J. Phys. Chem. B 108 (2004) 14340.
49

In-situ investigations of adsorbed benzene on silvermodified Pt(322) surfaces
Sandra Wickert1, Matthias Schöppke1, Regine Streber2, Michael Lorenz2, Hans-Peter Steinrück2, Reinhard Denecke1
1

Wilhelm-Ostwald-Institut für Physikalische und Theoretische Chemie, Universität Leipzig, Linnéstraße 2, 04103 Leipzig Lehrstuhl für Physikalische Chemie II, Universität Erlangen-Nürnberg, Egerlandstr. 3, 91058 Erlangen
2

The adsorption of molecules on a surface is the first step in heterogeneous catalysis. Thus, detailed knowledge about the adsorption step is necessary for the understanding of the whole catalytic process. In a previous paper by C. Papp et al. the sitespecific adsorption of benzene on a Ni(111) surface has been investigated by in-situ X-ray spectroscopy (XPS) [1]. Among other observations, they reported about two specific adsorption sites – hollow and bridge – identified by their C 1s signature. On Pt(111), Wander et al. determined the bridge site for disordered benzene adsorption [2]. An interesting issue is the influence of lateral confinement on the adsorption properties of molecules. So in our work we study the adsorption properties of benzene on stepped Pt surfaces, as an example for larger molecules. We used a regularly stepped Pt(322) surface (with five atomic rows wide (111) terraces and (100) oriented monatomic steps) and modified it with different amounts of Ag. The thermal deposition of Ag on stepped Pt at 300 K results in monatomic rows along the step edges, at least for the first two to three Ag rows [3]. The amount of Ag was varied from 0 to 1 ML (monolayers). The adsorption of benzene was monitored by insitu XPS; in uptake experiments at 117, 190 and 300 K, the intensity and binding energy of the C 1s signature was determined. In addition, benzene was thermally desorbed, monitored again by in-situ XPS. The experiments were performed at beamline U49/2-PGM1 at BESSY-II. As a first step, benzene adsorption at different temperatures was characterized on a clean (0 ML Ag) and a completely Ag-covered Pt(322) surface (1ML Ag). At 190 K, benzene chemisorbs on the clean Pt (322) surface, with no physisorption (Fig. 1). There is only one asymmetric C 1s peak, indicating an adsorption site for benzene, where all C atoms have a similar local geometry, i.e., the hollow site is preferred [1]. The observed asymmetry is due to unresolved vibrational splitting. With increasing coverage, the peak maximum shifts to higher binding energies by about 160 meV, because of lateral interactions between the benzene molecules. After 400 s saturation of the benzene coverage was achieved. The same results were obtained for an uptake experiment at 300 K under equal conditions. For 117 K also multilayer adsorption was observed.

time or dose Fig.. 1: C 1s uptake experiment at 190 K for adsorption of benzene on clean Pt(322). Total exposure after 400 s was about 0.8 Langmuir.
50

On a Pt(322) surface completely covered with Ag one broad C 1s peak at higher binding energies (of about 285.0 eV) is observed during benzene adsorption at 190 K. By removing the benzene gas phase a decrease of the intensity of the C 1s peak identifies this species as physisorbed benzene. For 300 K there is no adsorption of benzene on the Ag layer. For the study of the lateral confinement the amount of Ag (and thereby the free Pt surface area) was varied. At 190 K, benzene simultaneous adsorbs on Ag and Pt with different relative amounts, depending on the Ag coverage. In Fig. 2 a) the contributions of chemisorbed and physisorbed benzene on Pt and Ag, respectively, are shown, as derived from a deconvolution of the C 1s spectra, The chemisorbed part decreases with increasing amount of Ag and the physisorbed part increases. At 300 K the adsorption of benzene on Ag is suppressed completely, so only adsorption on free Pt(322) areas takes place (Fig. 2 b).

4,0 3,5

a)

benzene chemisorption on Pt benzene physisorption on Ag
Coverage of benzene / a. u.

4,0 3,5 3,0 2,5 2,0 1,5 1,0 0,5 0,0

b)

Coverage of benzene / a.u.

3,0 2,5 2,0 1,5 1,0 0,5 0,0 0,0 0,2 0,4 0,6 0,8 1,0

0,0

0,2

0,4

0,6

0,8

1,0

Coverage of Ag / ML

Coverage of Ag / ML

Fig. 2: Benzene coverage (in arbitray units) in dependence of silver coverage. a) Adsorption at 190 K leads to chemisorption (on Pt) and physisorption (on Ag). Data from spectral deconvolution (fit) of C 1s data taken during C6H6 exposure. b) Adsorption at 300 K only results in chemisorption on Pt.

In both experiments there is a limiting Ag amount for benzene adsorption on free Pt. At 300 K. where no further data analysis is necessary, extrapolation yields a value of about 0.6 ML Ag for this limit. Obviously, benzene needs at least two rows of Pt to adsorb on it. This implies that benzene adsorbs in flat adsorption geometry, as expected [2]. It also agrees with results by Campbell et al. who found an ensemble of at least six Pt atoms necessary for benzene dissociation [4]. This work was supported by BMBF (05 ES3XBA/5). [1] C. Papp, T. Fuhrmann, B. Tränkenschuh, R. Denecke, H.-P. Steinrück, Phys. Rev. B 73 (2006) 235426. [2] A. Wander, G. Held, R.Q. Hwang, G.S. Blackman, M.L. Xu, P. de Andres, M.A. Van Hove, G.A. Somorjai, Surf. Sci. 249 (1991) 21. [3] P. Gambardella, M. Blanc, H. Brune, K. Kuhnke, K. Kern, Phys. Rev. B 61 (2000) 2254. [4] J.M. Campbell, S. Seimanides, C.T. Campbell, J. Phys. Chem. 93 (1989) 815.

51

Simultaneous Synchrotron X-ray Scattering and Optical Spectroscopy: The first fivefold in situ-coupling technique for on-line monitoring of catalyst synthesis
J. Radnika, U. Bentrup a, J. Leiterer b, F. Emmerling b, A. Brückner
a

a: Leibniz-Institut für Katalyse, Branch Berlin, Richard-Willstätter-Str. 12, 12489 Berlin, Germany b: Bundesanstalt für Materialforschung und –prüfung, Richard-Willstätter-Str. 12, 12489 Berlin, Germany The first fivefold coupling of in situ techniques was established at BESSY allowing simultaneous on line monitoring of precipitation processes with X-ray scattering and optical spectroscopy like Raman, UV-Vis and ATR-FTIR. The combination of these methods provides information about the species in the precipitate and the solution, the phases and the particles in the precipitate. This equipment was used for detailed investigation of the precipitation of ammonium iron molybdates used as precursor material for selective oxidation catalysts. Depending on the reaction conditions, the formation of different phases with Anderson- or Keggin-like structure could be observed during the precipitation.

Introduction Heterogeneous Mo-based complex oxides are versatile catalysts for the selective oxidation of alkanes and olefins to the corresponding aldehydes, anhydrides or acids [1,2] and, in particular iron molybdates, for the selective oxidation of methanol to formaldehyde [3]. In general, a specific phase composition and structure in the mixed oxide catalysts is necessary for good catalyst performance. Usually, the synthesis of such catalyst material comprises different steps like the synthesis of suitable precursors by e.g. precipitation, isolation of the precipitate, drying and subsequent calcination. During the synthesis of the precursor, the preparation method as well as the nature of the used components and the reaction conditions play an important role and affect the final composition, structure and performance of the catalytic material. Systematic investigations with sophisticated methods for on-line monitoring of the synthesis process are required to understand and distinguish the influence of these different parameters. For this purpose, an experimental setup was established at the µ-spot Beamline at BESSY allowing simultaneous Small Angle and Wide Angle X-ray scattering (SAXS/WAXS), Raman, UV-VIS and ATR spectroscopy. While the structural changes of molybdate species within solution and precipitate were observed by optical spectroscopy, the scattering experiments provide information on the precipitate, such as particle and crystallite size and nature of crystalline phases formed during precipitation. Experimental An experimental setup presented formerly [4] was enhanced at the µ-spot Beamline at BESSY and is schematically shown in Fig. 1. Generally, an ammonium heptamolybdate (AHM) solution and a metal nitrate solution were prepared separately. Than, the nitrate solution was slowly added to the AHM solution by vigorous stirring. This mixture was stirred for 60 min at room temperature. In a second step, concentrated H3PO4 or a solution of diammonium hydrogenphosphate (NH4)2HPO4 were added, followed by further stirring for 30 min. Finally, the suspension was heated to 50°C and stirred for another 60 min.

Fig.1: Scheme of the experimental setup and image of the flowcell with the X-ray area detector and the Raman spectrometer

The UV-vis and ATR measurements were carried out by using respective probes directly dipped into the reaction solution. The UV-vis spectra were recorded in reflection mode using an Ava Spec 2048 fiber optic spectrometer (Avantes). Mid infrared ATR spectra were collected using a fiber optical diamond ATR probe (ifs Aachen) coupled to a FTIR spectrometer Avatar 370 (Thermo Electron). The slurry was transported by a peristaltic pump within a closed circuit of flexible tubes through a borosilicate capillary with an inner diameter

52

of 5 mm and a wall thickness of 100 µm used for Raman and X-ray scattering measurements. The Raman investigations were performed by focussing the laser beam onto the suspension flowing through this capillary using a fiber optical RXN spectrometer (Kaiser Optical Systems) equipped with a 70 mW diode laser at a wavelength of 785 nm. For the scattering experiments the capillary was irradiated with highly monochromatic X-rays (λ = 1.0336 nm). The scattered intensities were collected 20 cm behind the capillary with a twodimensional X-ray detector (MarMosaic 225). The scattering data were transformed into diagrams of scattered intensities, I, as function of the scattering vector q = 4π/λ sinθ with θ being the scattering angle. Results To obtain information about the influence of component concentration and pH-value on the formation of different molybdate phases, the system Mo/Fe/phosphate was exemplarily chosen, and the Mo/Fe ratio and the nature of the phosphate compound were varied. In Fig. 2 the X-ray scattering results of mixing AHM with the iron nitrate solution and addition of (NH4)2HPO4 aq. are presented. Bragg reflections indicating crystalline phases could be observed in all curves.

Intensity / a.u.

II III
2

10 -1 q /nm

30

Fig. 2: X-ray scattering curves and Raman spectra after mixing the AHM with the nitrate solutions (I), adding of (NH4)2HPO4 aq. (II) and heating at 50°C (III). The scattering curves were obtained every 120 sec.

In the first step one phase is formed immediately after mixing the solutions, further reflections appeared after 20 min indicating the formation of another phase. The latter phase disappears after admixture of the (NH4)2HPO4 solution. During heating at 50°C the crystallinity of the remaining solid phase lowers. The scattering results are confirmed by Raman spectroscopy. In the first step, the formation of two molybdate species is observed indicated by a band at 967 cm-1 which is due to an Anderson-type phase containing [H6FeMo6O24]3- species [5] and another one with a band at 967 cm-1 possibly correlated with [Mo8O26]4-species. The latter band disappears after addition of (NH4)2HPO4 due to the increase of pH. During heating at 50°C the band at 957 cm-1 shifts to 964 cm-1and a further band appears at 893 cm-1. The nature of the corresponding species is yet unclear. From inspection of the phosphate bands visible in the ATR spectra it can be concluded that a mixed molybdatophosphate or a mixture of Anderson-type molybdate and phosphate is possibly formed. If H3PO4 is added instead of (NH4)2HPO4, a Raman band at 980 cm-1 is observed typical for the Keggintype anion [PMo12O40]3-. The results of SAXS/WAXS measurements shown that the Mo/Fe ratio influences the crystallinity of this Keggin structure: a lower amount of Fe leads to an amorphous phase, whereas higher amounts seem to facilitate the formation of a crystalline phase.

Intensity [a.u.]

10

20 -1 q / nm

30

I

C B A

Fig. 4: X-ray scattering curves of the final precipitate obtained under different preparation conditions: A) Mo/Fe=0.12/0.01 + H3PO4 B) Mo/Fe=0.12/0.01 + (NH4)2HPO4; C) Mo/Fe=0.12/0.02 + H3PO4.

53

40

Conclusions The combination of X-ray scattering with synchrotron radiation and optical spectroscopic methods like Raman, UV-Vis and ATR-FTIR allows a comprehensive insight into the synthesis of mixed oxide catalyst precursors in the liquid phase under realistic synthesis conditions. This specific method combination proved to be a valuable tool to elucidate structural changes of the complex anions during precipitation process an on molecular scale together with changes of nanoscopic properties of respective precipitate such as agglomeration and crystallization. Acknowledgements
We thank U. Armbruster, L. Knöpke, Ch. Domke, S. Gatla and J. Kubias for technical assistance.

References
[1] [2] [3] [4] [5] R.K. Grasselli, Catal. Today 49 (1999) 141. G. Centi, F. Cavani, F. Trifirò, Selective Oxidation by Heterogeneous Catalysis, Kluwer Academic Publishers, New York 2001. A.P.V. Soares, M.F. Portela, A. Kiennemann, Catalysis Reviews 47 (2004) 125. U. Bentrup, J. Radnik, U.Armbruster, A. Martin, J. Leiterer, F. Emmerling, A. Brückner, Topics in Catalysis, in press I.L. Botto, A.C. Garcia, H.J. Thomas, J. Phys. Chem. Solids 53 (1992) 1075.

54

EXAFS study of Redox-active Metal-organic frameworks Report on application 2008_2_80132 Markus Tonigold and Dirk Volkmer Ulm University, Institute of Inorganic Chemistry 2, Materials and Catalysis, AlbertEinstein-Allee 11, D-89081 Ulm (Germany), Fax: +49 (0)731-50-23039, Tel: +49 (0)731-50-23921, E-mail: dirk.volkmer@uni-ulm.de measurements performed during 28th calendar week in 2008 on beamline KMC-2 Introduction Three metal-organic frameworks (MOFs), named MFU-1, MFU-2 and MFU-3, were investigated by EXAFS and XANES to gain deeper insights into the behaviour of these MOFs during catalysis. The structures of the untreated samples as determined by single crystal X-ray analysis are given in Fig. 1 and 2.[1]

Fig. 1: Crystal packing diagram of MFU-1 (left), MFU-2 (middle) and MFU -3. {CoON3} in MFU-1 and {CoN4} in MFU-2 and MFU-3 coordination units are represented as blue polyhedrons. Hydrogen atoms are omitted for clarity.

Fig. 2: Left: schematic drawing of MFU-1, represented by linearly connected octahedral nodes as secondary building units; Right: MFU-2 and MFU-3, represented by linearly interconnected onedimensional Co(II) as secondary building units.

All three MOFs catalyse the reaction of cyclohexene with tert-Butylhydroperoxide and show an intensive colour change during reaction. On the other hand, the XRD patterns after catalysis are unchanged for MFU-1, whereas MFU-2 and MFU-3 get amorphous during catalysis. Additionally, further experiments proofed that MFU-1 is in fact a heterogenous catalyst, whereas the Co ions of MFU-2 and MFU-3 are bleeched out from the networks, and thus homogenous catalysis is observed. Therefore, EXAFS and XANES studies were performed to gain deeper insights into the changes of the novel catalyst materials during catalysis.

55

Experimental EXAFS and XANES measurements were conducted at room temperature, using the doublecrystal monochromator (SiGe (111) graded crystals, E/ E=4200) at the beamline KMC-2 of the electron storage ring at the Berliner ElektronenspeicherringGesellschaft für Synchrotronstrahlung m.b.H. (BESSY II, Berlin, Germany). [2] The XANES and EXAFS spectra were recorded at the Co-K-edge, energy calibration was performed with cobalt metal foil. Measurements were performed in transmission mode using ion chambers, additionally the fluorescence using a fluorescence yield detector (Si-PIN photodiode) was recorded.

Results and Discussion XANES and EXAFS of MFU-1, MFU-2 and MFU-3 were recorded for the untreated samples as well as the samples after catalysis. The Co K-edge EXAFS spectra are analyzed using the standard Athena and Artemis FEFF XAFS analysis codes.[3] The raw and fitted Fourier transforms of the k2-weighted (k) spectra (calculated in the k-range 2.0–12.0 Å ) at the Co-K-edge are shown in Fig. 3-5. The Co-K edge spectra were fitted using models derived from the single crystal structure data of MFU-1, MFU-2 and MFU-3.

Fig. 3: Left: k2-weighted Fourier transformation of the EXAFS spectra of MFU-1 before catalysis (top) and after catalysis (bottom) fitted from 1.0-3.5 Å with 10 variables for 19 independent points, R = 0.004 2 = 762. The raw data and the fitting are shown in solid and dashed lines, respectively. Right: Stick representation of the cluster used for EXAFS data analysis. The absorbing Co atom is highlighted as a ball.

56

Fig. 4: Left: k2-weighted Fourier transformation of the EXAFS spectra of MFU-2 before catalysis (top) and after catalysis (bottom) fitted from 1.1-3.5 Å with 7 variables for 18 independent points, R = 0.002 2 and reduced = 213. The raw data and the fitting are shown in thin and thick solid lines, respectively. Right: Stick representation of the cluster used for EXAFS data analysis. The absorbing Co atom is highlighted as a ball.

Fig. 5: Left: k2-weighted Fourier transformation of the EXAFS spectra of MFU-3 before catalysis (top) and after catalysis (bottom) fitted from 1.0-4.0 Å with 12 variables for 19 independent points, R = 0.003 2 = 146. The raw data and the fitting are shown in thin and thick solid lines, respectively. Right: Stick representation of the cluster used for EXAFS data analysis. The absorbing Co atom is highlighted as a ball.

The excellent fits to the EXAFS data for all three compounds confirm the validity of the structure obtained from single crystal x-ray structure analysis. Particularly no further Co-containing species are observable, proofing that the catalytic activity is due to the metal sites in these three metal-organic frameworks (and not due to further unknown Co-containing impurities or side-products in the voids of the network). Whereas MFU-1 shows obvious changes between the sample before and after catalysis (especially above 2 Å in the Fourier transformed spectrum), MFU-2 and MFU-3 maintain the local structure during catalysis. Despite the obvious changes in the MFU-1 spectrum, the fit to the molecular cluster as it is before catalysis is still satisfactory (though with slightly changed bond lengths and Debye-Waller factors). Since also the XRD pattern of MFU-1 is unchanged during catalysis, we assume that due to diffusion limitation of the catalytic reaction, only a small part of the Co centers
57

close to the surface of the crystal is structurally changed. This fact is complicating the search for the structure of the catalytic center after the catalytic reaction. For MFU-1, bleeching experiments showed already that the catalysis is homogenous. The next question thus was if the catalysis is maybe only performed by defects on the external surface of MFU-1. The obvious change in the EXAFS signal for MFU-1 after catalysis now additionally proofs that the catalysis is performed on the {Co 4O} cluster centres of the framework and not only by defects on the external surface. For MFU-2 and MFU-3, on the other hand, the catalysis is due to bleeching of the Co ions into solution, and the material gets amorphous due to XRD measurements. In contrary, the EXAFS analysis for these two materials reveals that the molecular fragment is still unchanged after catalysis. Therefore, the catalyst is decomposing during catalysis into smaller, stable fragments with still the same local environment. These results are in accordance with the structural features of the three frameworks: The Co centres in MFU-1 are still accessible to potential further ligands, whereas the linker in MFU-2 and MFU-3 completely “shields” the metal centres from coordination of potential ligands. Altogether, these results indicate that only MFU-1 type catalysts are promising candidates for catalytic reactions. Since the EXAFS data for MFU-1 indicate at least a partial structural change after the catalytic process, XANES spectra were examined to obtain further information upon the valence state for MFU-1 before and after catalysis. Fig. 6 shows the XANES spectra of MFU-1 before and after catalysis together with Co3O4 and CoIII-pyrazolate as references with higher valences, showing that the edge energy is the same for MFU-1 after and before catalysis. The peak edge should be shifted to higher energy for Co of larger oxidation-state, which is often utilized to estimate the oxidation state of Co,[4] and a linear dependence of the chemical shift on the average valence can be assumed.[5] Thus the unchanged position of the peak edge indicates that the oxidation state of Co in MFU-1 is not changed throughout the catalytic cycles.

Fig. 6: Normalized Co K-edge XANES spectra.

Acknowledgements We want to thank BESSY for the financial support for this project and the granted beamtime, especially Prof. Erko at BESSY for fruitful discussions as well as scientifical and technical support.

58

1

The structure of MFU-1 is similar to the one of MOF-5, which has a CaB6 type framework topology. It encloses octahedrally-shaped {Co4O(dmpz)6} nodes reminiscent of the {Zn4O(CO2)6} secondary building units of MOF-5, and phenylene rings constituting the edges of the cubic 6-connected CaB6 net. MFU-2 and MFU-3 show PtS-type framework which contains prismatic tunnels running through the crystal, the walls consisting of layers of 1D Co(II) chains and dianionic BDPD ligands. Each Co(II) center adopts a distorted tetrahedral coordination geometry through binding to nitrogen donor atoms from four different ligands. 2 a) A. Erko, I. Packe, C. Hellwig, M. Fieber-Erdmann, O. Pawlitzki, M. Veldkamp, W. Gudat, AIP Conference Proc. 2000, 521, 415-418; b) A. Erko, I. Packe, W. Gudat, N. Abrosimov, A. Firsov, SPIE 2000, 4145, 122-128. 3 a) B. Ravel, M. Newville, J. Synchrotron Rad. 2005, 12, 537-541; b) M. Newville, J. Synchrotron Rad. 2001, 8, 322-324; c) B. Ravel, J. Synchrotron Rad. 2001, 8, 314-316; d) J. J. Rehr, R. C. Albers, Rev. Mod. Phys. 2000, 72, 621-654. 4 a) L. Barbey, N. Nguyen, V. Caignaert, F. Studer, B. Raveau, J. Solid State Chem. 1994, 112, 148. 5 a) M. Croft, D. Sills, M. Greenblatt, C. Lee, S.-W. Cheong, K. V. Ramanujachary, D. Tran, Phys. Rev. B 1997, 55, 8726; b) M. C. Sánchez, J. García, J. Blasco, G. Subías, J. Pérez-Cacho, Phys. Rev. B 2002, 65, 144409; c) M. C. Sánchez, G. Subías, J. García, J. Blasco, Phys. Rev. B 2004, 69, 184415; d) A. H. de Vries, L. Hozoi, R. Broer, Int. J. Quantum Chem. 2003, 91, 57.

59

In-situ investigation of sulfur oxidation on stepped Pt(355)
R. Streber, C. Papp, M.P.A. Lorenz, A. Bayer, R. Deneckea, and H.-P. Steinrück Lehrstuhl für Physikalische Chemie II, Universität Erlangen-Nürnberg, Egerlandstraße 3, 91058 Erlangen a Wilhelm-Ostwald-Institut für Physikalische und Theoretische Chemie, Universität Leipzig, Linnéstraße 2, 04103 Leipzig

Heterogeneously catalyzed reactions are often performed at highly dispersed particles, exhibiting a variety of defect sites, which are proposed to be very reactive. Poisoning of the active sites is a major issue in large scale applications of heterogeneously catalyzed reactions. One very efficient poison for catalysts is sulfur, which is supposed to lead to the blocking of reactive sites. We studied the reaction of sulfur with oxygen on the stepped Pt(355) surface by in-situ highresolution XPS applying synchrotron radiation. The Pt(355) surface has five atom row wide (111) oriented terraces and (111) oriented monatomic steps. Sulfur was deposited via hydrogen sulfide adsorbed at ~130 K, followed by heating to 700 K, which leads to the decomposition of H2S, with hydrogen desorbing as H2 and S remaining on the surface. The experiments were performed at beamline U49/2-PGM1, using a transportable apparatus described elsewhere [1]. Fig. 1 a shows a series of S 2p 131 K XP spectra collected during heata) 0.02 ML S 340 K after 15 min O 2 ing a mixed S and O layer pre400 K dosing at 250 K 580 K pared at 250 K by dosing molecular O2 on Pt(355), precovered with 0.02 ML sulfur. The topmost spectrum (orange) exhibits only one doublet in the S 2p region assigned to atomic S, which shows that preparation of the layer at 250 K leads almost exO2 dosing: b) 0.02 ML S clusively to the coadsorption of T = 350 K t=0s t = 40 s atomic oxygen and sulfur. t = 170 s Heating the mixed S and O layer t = 900 s to elevated temperatures leads to subsequent appearance of two new doublets in the S 2p spectra, due to the formation of SOx species during heating. The S 2p1/2 peak at 166.0 eV (green spec170 168 166 164 162 160 trum), which appears around 300 Binding Energy [eV] K, is assigned to SO2 and the one at 166.9 eV (blue spectrum), that Fig. 1: Selected S 2p spectra taken a) during heating a arises at 350 K, is assigned to mixed layer of 0.02 ML S and 0.37 ML O, prepared SO4, due to their characteristic by dosing the appropriate amounts of H2S and mo- binding energies and desorption lecular oxygen and b) during oxygen dosing on 0.02 temperatures [2,3]. SO decom4 ML S at 350 K; p(O2) = 6*10-7 mbar, hν = 260 eV poses above 500 K.
Intensity [arb. u.]

60

Fig. 1 b shows S 2p XP spectra recorded during dosing of oxygen at 350 K. Initially, the adsorption of oxygen leads to a shift in binding energy of the S 2p doublet assigned to elemental sulfur (see difference between red and orange spectrum in Fig. 1 b). This shift is caused by the displacement of sulfur by oxygen from step to terrace sites. A similar change of adsorption sites was also observed for coadsorbed CO and S on Pt(355) and Pt(322) during heating [4, 5]. After an induction period, the formation of SO2 (green spectrum) is observed, followed by a time-delayed oxidation of SO2 to SO4 (see blue spectrum).

ln S peak area

-2 => Ea = 34 kJ/mol

250 300 350 400 450

K K K K K

-4 ln k -6 -8 -10 2.0

2.5

3.0 1/T

3.5

4.0x10

-3

0

200

400 t [s]

600

800

Fig. 2: a) Quantitative analysis S 2p XP spectra recorded during oxidation of 0.02 ML S at 350 K; b) Change of S peak area (logarithmic scale) versus oxygen dosing time during oxidation of 0.02 or 0.03 ML S at different temperatures between 250 and 450 K, inset: ln k versus 1/T.

Fig. 2 a shows the quantitative analysis of the reaction of S with O at 350 K. After the short induction period the S coverage decreases exponentially with time and SO2 is formed. At ~100 s the subsequent reaction to SO4 starts and nearly all SO2 has reacted after 900 s. To determine the kinetic parameters of the reaction, the oxidation has been studied at different temperatures. For small S coverages the plot of S peak area (logarithmic scale) vs. time shows a linear decrease at all chosen temperatures, as can be seen in Fig. 2 b. The negative slopes, i.e., the rate constants k, increase with temperature, indicating pseudo first order reaction kinetics with respect to the oxidation of elemental S. Applying the Arrhenius equation, the plot of ln k vs. 1/T then reveals an apparent activation enthalpy of about 34 kJ/mol for the rate determining step in the oxidation of S to SO2 on Pt(355). Acknowledgements: The project was supported by the BMBF through grant 05 ES3XBA/5 and by the DFG through the Cluster of Excellence “Engineering of Advanced Materials”. We like to thank the BESSY staff for their support during beamtime.

[1] R. Denecke, M. Kinne, C. M. Whelan, H.-.P. Steinrück, Surf. Rev. Lett. 9 (2002) 797. [2] M. Polcik, L. Wilde, J. Haase, B. Brena, G. Comelli, G. Paolucci, Suf. Sci.181 (1997) L568. [3] St. Astegger, E. Bechtold, Surf. Sci. 122 (1982) 491. [4] R. Streber, C. Papp, M.P.A. Lorenz, A. Bayer, S. Künzel, M. Schöppke, R. Denecke, H.-P. Steinrück, J. Phys.: Condens. Matter (in press). [5] R. Streber, C. Papp, M. P. A. Lorenz, A. Bayer, R. Denecke, H.-P. Steinrück, Chem. Phys. Lett., 452 (2008) 94.

61

Structural and Electronic Properties of Mono- and Divalent Thiols Bound on Isolated Gold Nanoparticles R. Lewinski, C. Graf, B. Langer, E. Antonsson, A. Hofmann, X. Gong, B. Grundkötter, E. Rühl Institut für Chemie und Biochemie - Physikalische und Theoretische Chemie, Freie Universität Berlin, Takustr. 3, D-14195 Berlin A multivalent linker is defined as a molecule containing identical functional groups binding to acceptor sites. It has been observed in thermodynamic and kinetic studies that the stability of the complexes which are formed between multivalent linkers and substrates is significantly increased, when the binding of multivalent systems is compared to the corresponding monovalent species.[1] Moreover, the number of binding sites are not additive to the enhanced stability, implying a highly non-linear binding behavior of multivalent linkers.[1] However, these effects are not yet fully understood on a molecular level. Nanoparticles functionalized with small mono- and multivalent organic linker molecules are simple model systems for investigations of multivalent interactions. In the present study high-resolution core-level excitation is used to investigate the local electronic properties of mono- and bivalent thiols which are bound to free gold nanoparticles. For a comparison also the corresponding free ligands are studied in solid nanoparticle samples. The thiol ligands (see Figure 1) are commercially available or prepared by organic syntheses.[2] Gold nanoparticles of 25-50 nm diameter are prepared by colloidal chemistry[3] and functionalized with the mono- or bivalent ligands. The solutions of the ligands or dispersions of the functionalized nanoparticles are sprayed into the gas phase at ambient pressure by using an atomizer. Subsequently, the solvent is evaporated in a diffusion dryer. The nanoparticle beam is focused by an aerodynamic lens in the size regime r < 200 µm in the interaction region with synchrotron radiation.[4] Total electron yields are measured in the S L3,2 absorption regime in order to probe the local electronic surroundings of the absorbing sites as a function of chemical binding. This approach relies on a short interaction time of the nanoparticles with X-rays, so that radiation damage is avoided (cf. [5,6]). We investigate the binding of the three bivalent compounds lipoic acid (LA), its open form dihydrolipoic acid (DHLA), and the amino analogon of lipoic acid, 5-(1,2-dithiolan-3-yl)butan amine. 11-mercaptoundecanoic acid (11-MUDA) is used as a monovalent reference (see Figure 1). Total electron yield (TEY) spectra of the unbound molecules 11-mercaptoundecanoic acid, dihydrolipoic acid, and lipoic acid in the S L3,2-absorption regime are displayed in Figure 2.
π*(CH2) σ*(C-S) p1/2 π*(CH2)

p3/2

p1/2

p3/2

TEY [eV]

σ*(C-S)

161 162 163 164 165 166 167 168 Energy [eV]
Figure 1: Mono- and bivalent sulfur containing compounds investigated in this study. Figure 2: Total electron yield spectra of 11-MUDA (red line), DHLA (blue line) and LA (black line) in the S L3,2 absorption regime. The two vertical lines indicate the shift of the S L3,2 spectrum of LA compared to that of DHLA.

62

11-MUDA shows the typical S L3,2-spectrum of a monovalent thiol,[6] whereas the spectrum of DHLA is slightly broadened. This can be explained by the fact that in this dithiol the two sulfur atoms are in slightly different chemical environments (see Figure 1). In contrast, in the spectrum of LA the S p3/2 σ*(C-S) peak is shifted to significantly lower energies and the overall of the spectrum is significantly different, as it is expected for a disulfide.[6] The 161 162 163 164 165 166 167 168 S L3,2-spectrum of 5-(1,2-dithiolan-3-yl)butan Energy [eV] amine, also measured in this study (not shown Figure 3: Total electron yield spectra of 11-MUDA here), is practically identical with that of the bound to gold nanoparticles (red solid line) and free analogous acid because the 1,2-dithiolan ring is 11-MUDA (red dotted line) in the S L3,2-absorption identical in both molecules. regime. The vertical lines indicate the shift between The free nanoparticles consisting of the pure spectra of the free and the gold-bound 11-MUDA. ligands yield relative intense electron yield spectra. In the case of the functionalized gold nanoparticles only monolayers on the nanoparticle surface are measured. As expected, the TEY-signal of the monolayer in the in near-edge spectra is weak and superimposed to a huge electron signal from the valence continuum. Nevertheless, it turned out that the signal strength is sufficient to take near-edge spectra of such species. Even though these experiments require relatively long 161 162 163 164 165 166 167 168 acquisition times, no radiation damage was observed because continuously fresh sample Energy (eV) entered the ionization region. Figure 4: Total electron yield spectra of LA (black Figure 3 shows the S L -spectrum of 113,2 solid line) and its amino analogon (green dotted line, MUDA bound on 25 nm gold nanoparticles. see chemical formula in Figure 1) both bound on gold nanoparticles in the S L3,2 absorption regime. The spectrum is shifted by about 0.2 eV to For a comparison the spectra of free DHLA (blue higher energies compared to the spectrum of solid line) and free LA (black dotted line) are also free 11-MUDA. This result clearly indicates displayed. The vertical line indicates that the spectra that also monolayers of organic ligands bound of gold-bound LA and 5-(1,2-dithiolan-3-yl)butan to gold nanoparticles can be investigated by amine are not shifted compared to free DHLA. this approach and that these measurements are sensitive to changes in the local electronic structure between thiol-ligands bound to free gold nanoparticles and free thiol ligands in solid nanoparticle samples. The total electron yield spectrum of lipoic acid in the S L3,2-absorption regime is displayed in Figure 4. Lipoic acid is a 1,2-dithiolan (a fivemembered ring with two sulfur, see Figure 1). Hence, it can in principle bind on the gold Figure 5: Schematic view of the binding of lipoic nanoparticles as a closed ring forming a dative acid (LA) to a gold nanoparticle. bond or by ring opening (as DHLA, see Figure 1) binding as a dithiolate. A comparison with the free lipoic acid shown in Figure 4, indicates that the gold-bound lipoic acid is significantly shifted to higher energies. Besides this, the TEY spectrum of the gold-bound lipoic acid
TEY [eV] TEY [eV]

63

appears narrower and the lowest energy peak is more shifted than the other peaks. By contrast, the S L3,2-spectrum of DHLA (blue solid line) matches much better with the S L3,2spectrum of the particle-bound lipoic acid. Further, particle-bound lipoic acid has a more substantially redshifted shoulder (S L3 σ*(C-S) resonance). Therefore, we conclude that lipoic acid mainly binds by ring opening on gold. The redshifted S L3 peak is most likely due to a small fraction of lipoic acid, which is still present in the sample. It is well-known that the binding of thiols on gold nanoparticles in dispersion is an equilibrium process.[9] Therefore, we expect that this fraction is mainly due to free lipoic acid (see Figure 5). The S L3,2spectrum of 5-(1,2-dithiolan-3-yl)butan amine (the amino analogon of lipoic acid, green dotted line in Figure 4) has a similar shape as that of lipoic acid. However, the shift of the S L3-resonance is even less pronounced. Hence, in this case the fraction of the 1,2-dithiolan (closed ring) is even lower. The binding of lipoic acid on macroscopic gold surfaces and deposited gold nanoparticles has already been studied before.[7,8] Also, the results published in refs. 7 and 8 indicate that it preferably binds as an open dithiolate by ring opening on gold. This confirms that the present approach yields reliable information on the binding of divalent thiol ligands on isolated gold nanoparticles. Remarkably, neither the spectrum of lipoic acid, nor that of its amino analogon are shifted in energy compared to the near-edge spectrum of DHLA (see Figure 4), whereas the S L3,2spectrum of the monovalent gold-bound 11-MUDA is clearly blueshifted compared to the unbound 11-MUDA. This also indicates that NEXAFS is sensitive to changes between monoand multivalent bound organic molecules. A detailed analysis of the spectral shifts and the binding of the ligands to nanoparticles are currently in progress. In conclusion, the present results show that the high sensitivity of the experimental approach permits studies on the binding of mono- and multivalent ligands to free nanoparticles. It is possible to identify the particle-bound species by inner-shell excitation. Moreover, the present approach is sensitive to changes in the electronic structure between bound and free ligands as well as mono- and bivalent ones.
This project is supported by the Deutsche Forschungsgemeinschaft (SFB 765, project C5) and the Fonds der Chemischen Industrie. References 1. J. Rao, J. Lahiri, L. Isaacs, R.M. Weis, and G.M. Whitesides, Science 1998, 280, 708; P.I. Kitov, and D.R. Bundle, J. Am. Chem. Soc. 2003, 125, 16271; J.D. Badjic, A. Nelson, S.T. Cantrill, W.B. Turnbull, and S. Fraser-Stoddart, Acc. Chem. Res. 2005, 38, 723. 2. A. Hofmann, C. Graf, and E. Rühl, in preperation 2009. 3. G. Frens, Nat. Phys. Sci. 1973, 241, 20. 4. H. Bresch, B. Wassermann, B. Langer, C. Graf, R. Flesch, U. Becker, B. Österreicher, T. Leisner, and E. Rühl, Faraday Discuss. 2008, 137, 389. 5. Y. Zubavichus, M. Grunze, O. Fuchs, L. Weinhardt, C. Heske, E. Umbach, and J. D. Denlinger, Radiation Res. 2004, 161, 346. 6. J. Jalilehvand, Chem. Soc. Rev. 2006, 35, 1256. 7. T.M. Willey, A.L. Vance, C. Bostedt, T. van Buuren, R.W. Meulenberg, L.J. Terminello, C.S. Fadley Langmuir 2004, 20, 4939. 8. S. Roux, B. Garcia, J.L. Bridot, M. Salome, C. Marquette, L. Lemelle, P. Gillet, L. Blum, P. Perriat, O. Tillement, Langmuir 2005, 21, 2526. 9. M. Montalti , L. Prodi , N. Zaccheroni , R. Baxter , G. Teobaldi , F. Zerbetto, Langmuir 2003, 19, 5172.

64

Spectroscopic characterization of self-assembled monolayers of benzylmercaptan and p-cyanobenzylmercaptan on Au(111) surfaces
T. Strunskusa, L. Hallmannb, V. Staemmlerd, Ch. Wöllc, and F. Tuczekb
a

Lehrstuhl für Materialverbunde, Technische Fakultät der Christian-Albrechts-Universität zu Kiel, Kaiserstr. 2, 24143 Kiel b Institut für Anorganische Chemie der Christian-Albrechts-Universität zu Kiel, Otto Hahn Platz 6/7, 24098 Kiel c Lehrstuhl für Physikalische Chemie 1, Ruhr-Universität Bochum, Universitätsstr. 150, 44801 Bochum d Lehrstuhl für Theoretische Chemie, Ruhr-Universität Bochum, Universitätsstr. 150, 44801 Bochum

XPS Benzylmercaptan (BM) and p-cyanobenzylmercaptan (pCBM) were deposited via a selfassembly process onto gold covered silicon wafers from ethanolic solutions. After rinsing with ethanol and drying in a stream of nitrogen gas the samples were transferred into a UHV system; then XPS and NEXAFS spectra were acquired. XPS and NEXAFS experiments were performed at the HE-SGM beamline. XPS spectra were acquired in normal emission geometry at photon energies of 500 eV for the N 1s, 400 eV for the C 1s and 300 eV for the S 2p region, respectively. The XPS spectra are in accordance with the chemical formulas of the deposited molecules. The intensity of the cyano carbon in the C 1s spectrum in Fig. 1 amounts to 20%, which appears to be enhanced as compared to the stoichiometric ratio of 1:7 applying to pCBM. The deviation from stoichiometry can be explained by the high orientational order of the film, where all the cyano groups are located on the outermost layer, and by the relatively small electron mean free path of only 0.9 nm for photoelectrons at a kinetic energy of about 115 eV. In addition, our calculations indicate that the C 1s lines of the carbon atom in para position to the cyano group and of the methylene carbon should be shifted by 0.5 and 0.7 eV to higher binding energy relative to the other carbon atoms in the benzene ring. The XP spectra do not show any additional oxidized species indicating the stability of the prepared films against oxidation in air.
-1

Normalized Intensity /counts s

C 1s h = 400 eV

Normalized Intensity /counts s

-1

N 1s h =500 eV

BM

pCBM

pCBM
292 290 288 286 284 282

404

402

400

398

396

394

Binding energy / eV

Binding energy / eV

Fig.1 Synchrotron C 1s and N 1s spectra of a benzyl mercaptane (BM) and p-cyano benzyl mercaptane (pCBM) self assembled monolayer on a Au(111) surface.

NEXAFS NEXAFS spectra were acquired in the partial electron yield mode with a retarding voltage of –150 V at the C K-edge and –250 V at the N K-edge. Linear polarized synchrotron light with a polarization factor P of 82% was used. Energy resolution was 0.40 eV. The incidence angle of the light was varied from 90° (E-vector in surface plane) to 30° (E-vector near surface normal). The raw NEXAFS spectra were normalized to the incident photon flux by

65

division through a spectrum of a clean, freshly sputtered gold sample. The energy scale was referenced to the pronounced * resonance of highly oriented pyrolytic graphite at 285.38 eV. NEXAFS spectra measured at the C edge for BM and pCBM are displayed in Figures 2a and 2b, respectively. The assignments of the resonances are summarized in Table 1. BM and pCBM essentially show the same resonances at the carbon edge, but the presence of the cyano group leads to three additional * resonances in the case of pCBM. It is interesting to note that the intensity of the first * resonance of the benzene ring appears to be more pronounced for BM as compared to pCBM. This could indicate that intensity is shifted to the resonances at higher photon energies due to the presence of the cyano group. More detailed assignments of the resonances are given below. For both SAMs the dichroism of pronounced * resonance of the benzene ring has been used to determine the orientation of the phenyl ring relative to the surface normal. The angular analysis was performed as described in ref. [1]. For BM the molecular backbone shows a tilt angle of 19±5° relative to the surface normal whereas for pCBM a tilt angle of 21±5° was determined. In the NEXAFS spectrum of pCBM at the N K edge spectrum three * resonances are observed. Their positions and assignments are summarized in Table 2. The angular dependence of the * resonance of the nitrile group has been analysed in the same way as the * resonance of the benzene ring. In this case a tilt angle of 18±5° was obtained which is close to the 21±5° observed for the benzene ring at the C K edge.
BM
Normalized PEY

pCBM
Normalized PEY
1 3 4 5

1 23 4

5

90°

90°

55° 30°

55°

30°

calc.

280

290

300

310

280

290

300

310

Photon Energy /eV

Photon Energy /eV

Fig .2 NEXAFS spectra measured at the C K edge for a) BM and b) pCBM. The angles at the curves specify the angles of light incidence relative to the surface. The lines on the top indicate the positions of the resonances and are summerized in table 1. In b) also the calculated spectrum is included for comparison (the calculated transitions were broadened by Gaussian curves with 0.2 eV FWHM). Table 1. Resonance 1 2 3 4 5 Energy/eV BM 285.2 287.0 288.8 293.7 Energy/eV pCBM 285.1 286.2 286.9 288.5 293.2 Assignment 1* benzene ring 1* cyano carbon C-H*, 2* cyano carbon 2* benzene ring, 3* cyano carbon * benzene ring

The calculations of the NEXAFS spectra were performed using the Bochum suite of openshell wave function based quantum chemical ab initio programs in a similar way as in our previous studies of NEXAFS spectra [2]. The NEXAFS spectra for BM and pCBM at the carbon K edge closely resemble the spectra expected for benzene and benzonitrile, respectively. There is only a small contribution of the methylene carbon to the spectra which is also supported by the results of our theoretical calculations for pCBM. The C atom of the CH2SH group has only single bonds to its neighbours and therefore no strong π* transitions.

66

The NEXAFS spectrum of the C K edge of pCBM shows a pronounced fine structure indicating the presence of several distinct resonances. The reason is that pCBM contains six types of non-equivalent C atoms which give rise to slightly different ionization thresholds and consequently also to different NEXAFS spectra. The calculated total spectrum shown in Fig. 2b (bottom) is a superposition of the spectra of all eight C atoms. The C atoms in the benzene ring give rise to only one strong NEXAFS transition of π* character and some weaker π* and σ* transitions. The spectrum for the nitrile C atom closely resembles the one of the N K edge spectrum presented below, with three strong peaks possessing π* character. Two of the resonances overlap with resonances present also in BM (see Table 1). Therefore the most pronounced contributions of the cyano group lead to the new resonances at 286.2 eV and a pronounced intensity increase of the resonance located at 286.9 eV.
pCBM
Normalized PEY

NK
90° 55° 30°

Table 2. Resonances of pCBM at the nitrogen K edge and assignments Resonance Energy/eV Assignment 1 398.9 CN π*, out-of-plane* 2 3 399.8 401.7 CN π*, in-plane* CN π*, out-of-plane*

1

2

3

Calc.

396

400

404

408

* “out-of-plane” = perpendicular to the plane of benzene ring; “in-plane” = in the plane of benzene ring

Photon Energy /eV

Fig .7 NEXAFS spectra measured at the N K edge for pCBM. The angles at the curves specify the angles of light incidence relative to the surface. Numbers indicate the resonances as assigned in table 2. A calculated spectrum is included for comparison (the calculated transitions were broadened by Gaussian curves with 0.4 eV FWHM to mimic the experimental resolution.

At the nitrogen K edge there is a very good agreement between experiment and calculation. The N K edge NEXAFS spectrum of pCBM shows three strong bands positioned at 398.9, 399.8 and 401.7 eV. All of them are C-N π* valence excitations. The second, strongest band at 399.8 eV can be assigned to the excitation from N 1s into the “in-plane” component of the two-fold degenerate C-N π* orbital, perpendicular to the long axis of the molecule. In the other two bands the excitation leads from N 1s into two linear combinations of the “out-ofplane” π* orbital of C-N and the π* orbitals of the benzene ring. The orbital in the lower state at 398.9 eV is primarily localized at CN and has therefore a higher intensity, while the one at 401.7 eV is mainly localized at the benzene ring. It should be noted that the sum of the intensities of the out-of-plane excitations at 398.9 eV and 401.7 eV matches almost exactly the one of the in-plane excitation at 399.8 eV. These three resonances appear only when the cyano group is electronically coupled to the -system of the benzene ring. For an isolated cyano group as in a nitrile functionalized alkanethiol only one * resonance is observed at the N K edge at 400.8 eV [3].
Acknowledgments Support of the travel of L.H. and T.S. to the Berlin synchrotron source by the BMBF through project # 05 ES3XBA/5 is gratefully acknowledged.

References
1. 2. 3. Dmitriev, A.; Spillmann, H.; Stepanow, S.; Strunskus, T.; Wöll, Ch.; Seitsonen, P. A.; Lingenfelder, M.; Lin, N.; Barth, V. J; Kern, K. Chem. Phys. Chem. 2006, 7, 2197. Reiss, S.; Krumm, H.; Niklewsi, A.; Staemmler, V.; Wöll, Ch. J. Chem. Phys. 2002, 116, 7704. Frey, S.; Shaporenko, A.; Zharnikov, M.; Harder, P.; Allara, L. D. J. Phys. Chem. B, 2003, 107, 7716.

67

Characterization of a novel water-oxidizing cobalt catalyst by X-ray absorption spectroscopy
M. Risch, V. Khare, A. Grundmeier, O. Sanganas, L. Gerencser, I. Zaharieva, S. Löscher, P. Chernev, M. Haumann, H. Dau*
Freie Universität Berlin, Institut für Experimentalphysik, 14195 Berlin, Germany *holger.dau@physik.fu-berlin.de

Water oxidation occurs by the abstraction of four electrons and four protons from two water (H2O) molecules resulting in O2 formation. In nature, only one enzyme, a photosynthetic protein complex termed photosystem II (PSII), is capable of this reaction, delivering the O2 of the atmosphere [1,2]. Efficient water oxidation is also of prime interest for the production of hydrogen (H2), the fuel of the future. Therefore, in a worldwide strive, chemists search for synthetic catalysts for efficient water oxidation by low-cost catalysts. Recently several new transition-metal based synthetic compounds for water oxidation have been reported [3]. In particular, a recently introduced system using cobalt as the active metal [4,5] has attracted much interest. This catalyst is “spontaneously” assembled as a thin layer by electrodeposition on ITO electrodes from aqueous solution of cobalt and phosphate salts. Its self-assembly and self-repair mechanism bears similarities to the biological formation of the manganese complex of water oxidation in PSII [6]. The atomic structure of the cobalt catalyst so far has been unknown. For the first time, we studied the cobalt catalyst by X-ray absorption spectroscopy (XAS) and determined the Co oxidation state and coordination environment. The results enable us to postulate possible structural features, e.g. Co-Co bridging by di-µ-oxo bridges, which may be related to the function of the catalyst. A respective publication is in preparation [7]. Experimental. The cobalt catalyst was prepared as described in [2]. The catalytic film was either scratched off the ITO electrode and dried to yield XAS samples or the wet native film on the ITO support was directly studied after freezing (quasi in-situ). CoII(OH2)6(NO2)3 and CoIII(NH3)6 samples served as oxidation state standards. XAS at the Co K-edge (7709 eV) was performed at beamline KMC-1 at 20 K using a liquid-helium cryostat. A photodiode shielded by Fe foil against scattered incident X-rays served as an X-ray fluorescence detector.
Normalized Absorption (r.u.)

1.5

1.0

0.5

Co(II) Powder standard Catalyst Measurements of various films Co(III) Powder standard

0.0 7700

7725

7750 7775 Energy (eV)

7800

Figure 1: XANES spectra of the Cobalt catalyst and of Co(II) and Co(III) standards. The Co in the catalyst was in the Co(III) oxidation state, as apparent from the similar edge energy as in the Co(III) standard, no matter whether the native film on the ITO electrode or a dried powder of the scratched-off film was studied. Accordingly, the material of the catalytic film appears to be 'robust' against mechanical treatment and dehydration.

68

Figure 1 shows XANES spectra of the Co catalyst. The edge energies of the native wet Co film on the ITO electrode and the dried scratched-off powder were similar to that of a Co(III) standard. Thus, the Co film almost exclusively contains Co(III). If at all, Co(IV) was present only in very minor amounts. The unchanged K-edge of the Co catalyst in the two conditions suggests that it is robust against dehydration, air-exposure, and mechanical treatment.

10

Co-O 1.88 Å Co-Co 2.80 Å

Experiment Simulation

FT of EXAFS

Co-Co 5.6 Å 0 0 1 2 3 4 5 6 Reduced Distance / Å 7 8

Figure 2: (left) Fourier-transform of an EXAFS spectrum of the native Co catalyst. The main Co-Co distance (2.8 Å) is well determined. (right) Possible structural motif of the Co catalyst in agreement with the XAS-detected interatomic distances and Co coordination numbers (blue, Co; red, O).

Figure 2 shows the FT of an EXAFS spectrum of the native Co catalyst on ITO. The primary Co ligation by ~6 oxygen atoms at a distance of 1.89 Å was determined from a simulation and is well compatible with Co(III). A prominent Co-Co distance of 2.80 Å accounted for at least 3-4 metal-metal interactions per Co ion. There was also evidence for longer Co-Co distances. Interestingly, no positive evidence for the presence of phosphorus in the 1st and 2nd coordination spheres of Co was obtained. However, it can not be fully excluded that 1-2 Co-P interactions at ~2.9 Å may be hidden in the Co-Co FT peak (Fig. 2). This result is remarkable as phosphorus may have a key function in the self-assembly of the catalyst [4]. In summary, first XAS measurements on a novel water-splitting cobalt catalyst have been performed successfully. Our results allow construction of atomic models for the overall structure of the Co(III) catalyst, which may comprise complete or incomplete Co-oxo cubanes (Co4(µ3-O)4 or Co3(µ3-O)4 units) [7]. Further XAS experiments at KMC-1, inter alia to follow the assembly of the Co catalyst, are in preparation.
Excellent support by the scientists at KMC-1, Dr. F. Schäfers and M. Mertin is gratefully acknowledged. We thank our collaborators at the Fritz-Haber Institute Berlin, in particular S. Wasle, B. Pettinger, and G. Wagner, for complementary experiments, and SFB498, Unicat CoE Berlin, EUSolarH2, and the BMBF (BioH2 program) for financial support. [1] Dau H and Haumann M, Coord. Chem. Rev. 2522, 273-295 (2008). [2] Haumann M, Grundmeier A, Zaharieva I, Dau H, Proc. Natl. Acad. Sci. USA 105, 17384 (2008). [3] Meyer TJ, Nature 451, 778 (2008). [4] Kanan MW, Nocera DG, Science 321, 1072-1075 (2008). [5] Kanan MW, Surendranath Y, Nocera DG, Chem. Soc. Rev. 38, 109-114 (2009). [6] Barra M, Haumann M, Loja P, Krivanek R, Grundmeier A, Dau H, Biochemistry 45, 14523 (2006). [7] Risch M, Khare V, Dau H et al. Manuscript in preparation (2009).

69

X-ray study of the fluorinated double-wall carbon nanotubes produced using different fluorination methods
Yu.V. Lavskayaa, L.G. Bulushevaa, A.V.Okotruba, V.O. Koroteeva, A. Feltenb, E. Flahautc, D.V. Vyalikhd a Nikolaev Institute of Inorganic Chemistry, SB RAS, Novosibirsk, Russia b Laboratoire Interdisciplinaire de Spectroscopie Electronique, Facultes Universitaires Notre Dame de la Paix, Belgium c Centre InteruniVersitaire de Recherche et d’Ingenierie des Materiaux, UniVersite PaulSabatier, France d Institute of Solid State Physics, Dresden University of Technology, Germany Near-edge x-ray absorption fine structure (NEXAFS) spectroscopy and x-ray photoelectron spectroscopy (XPS) have been applied to investigate how synthetic conditions influence on the electronic structure of the fluorinated double-wall carbon nanotubes (DWNTs). The DWNTs were produced by catalytic chemical vapor deposition technique [1] in the result of CH4 decomposition over Mg0.9Co0.1O containing additions of molybdenum oxide. The fluorinated DWNTs were prepared using three different methods. The sample denoted with FDWNT (BrF3) was produced using the fluorination procedure described in [2]. The DWNTs were held in the vapor over a solution of Br2 and BrF3 for 7 days and thereafter, the sample was dried by a flow of N2 until the removal of Br2. The F-DWNT (СF4) sample was obtained by CF4 plasma treatment (plasma frequency is 13.56 MHz, power is 15 Watt, a chamber pressure is 0.1 Torr, exposure time is 10 min) [3]. The sample F-DWNT (F2) was synthesized using elemental fluorine flow at 200ºC. The C K- and F K- edge NEXAFS and C 1s XPS spectra were measured at the Berliner Elektronenspeicherring für Synchrotronstrahlung (BESSY) using radiation from the RussianGerman beamline. The C 1s XPS spectra were measured at the energy of monochromatized synchrotron radiation equals to 350 eV with energy resolution of 0.2 eV. The NEXAFS spectra were acquired in a total electron yield mode and normalized to the primary photon current from a gold-covered grid. The monochromatization of the incident radiation was ~100 meV in the carbon absorption region and ~410 meV in 282 284 286 288 290 292 the fluorine absorption region. Before the experiments C C 1s the samples were annealed at 70°C during 12 hours to C-CF C-F removal a residual gas. Figure 1 compares the C1s XPS spectra of the 4 initial and the fluorinated DWNTs. The spectrum of the initial DWNT has a single assymetric peak at 284.5 eV. All spectra of the fluorinated DWNTs show three main peaks corresponding to different chemical 3 states of carbon atoms. The peak C at 284.5 eV is referred to non-grafted carbon, the peak C-F in the range 288.0÷288.5 eV is attributed to the carbon atoms covalently bonded to the fluorine atoms and the 2 feature in the range 285.2÷285.7 eV corresponds to the carbon atoms positioned near the CF groups. It should be noted that the features corresponding to CF2 and CF3 groups are absent in the spectra. One can see that 1 the intensities of the selected peaks are varied 282 284 286 288 290 292 depending on the fluorination method. The C 1s XPS Binding energy (eV) spectra were fitted using a combination of three Fig. 1. C 1s XPS spectra of pristine DWNTs (1) and those fluorinated by components with a Gaussian-Lorentzian peak shape CF4 plasma (2), BrF3 (3) and F2 (4). with a Doniach-Sunjic high energy tail. The
Intensity (arb. units)

70

Table 1. The position of main components in the C 1s XPS spectra and chemical composition of the fluorinated DWNTs
EC F-DWNT (CF4) F-DWNT (BrF3) F-DWNT (F2) 284.5 284.5 284.5 EC-CF 285.2 285.4 285.7 EC-F 288.0 288.2 288.5

sample contains more fluorine that nanotubes. The C K-edge NEXAFS spectra of the initial and fluorinated DWNTs are shown in Fig. 2 (top series). The all spectra exhibit three main features and peaks at ~285.4 and ~291.7 eV correspond to π* and σ* resonances, respectively. The peak labeled with D positioned between 286.5 and 290.5 eV is attributed to 1s→σ* transition included С-F bond. The largest reduction of π* resonance is observed in the spectrum of F-DWNT (F2) sample. In case of F-DWNT (CF4), intensity of π* resonance increases compared to that of pristine DWNTs. We associate this effect with removal of defects in the DWNTs shells after CF4 plasma treatment. The F Kedge NEXAFS spectra of the fluorinated samples (Fig.2 bottom series) were aligned to the C K-edge spectra using the C 1s ∗ ∗ and F 1s core levels energies. σ∗ a σ C K-edge σ C K-edge c C K-edge b ∗ The spectra show three peaks at π π∗ π∗ the energies 690.4, 692.7 and D D D 695.8 eV labeled with A, B and C, respectively. Position of the peak C coincides with the σ* 280 290 300 310 280 290 300 310 280 290 300 310 resonance of the C K-edge C C C NEXAFS spectrum and, hence, B F K-edge F K-edge B F K-edge A can be attributed to the σ* B absorption edge. The peaks A and B, which are aligned with the A peak D of the C K-edge spectra, correspond to the C-F 690 700 710 720 690 700 710 720 690 700 710 720 interactions. Our quantumPhoton energy (eV) chemical calculations showed Fig. 2. NEXAFS near C K-edge (dark cyan line) and near F Kthat the peak B corresponds to edge (purple line) spectra of F-DWNT(CF4 ) – a, FDWNT(BrF3) – b, F-DWNT(F2) – c. C K-edge spectrum of the the σ-type antiboding between fluorine and carbon atoms from initial DWNTs showed by black line. the CF-group, the peak A is formed due to interaction of fluorine with carbon atoms situated at the CF-group. The lack of the peak A in the F K-edge NEXAFS spectrum of the F-DWNT (F2) sample is due to high fluorine coverage of nanotube surface. The work was supported by the bilateral Program “Russian-German Laboratory at BESSY”. [1] E. Flahaut, R. Bacsa, A. Peigney, Ch. Laurent. Chem. Commun. (2003) 1442-1443. [2] N.F. Yudanov, A.V. Okotrub, Yu.V. Shubin, L.I. Yudanova, L.G. Bulusheva, A.L. Chuvilin, et al. Chem. Mater. 14 (2002) 1472-1476. [3] A. Felten, C. Bittencount, J.J. Pireaux, G. Van Lier, J.C. Charlier. J. Appl. Phys. 98 (2005) 074308.
Intensity (arb. units)

component positions EC, EC-CF, EC-F and chemical compositions CFx of the fluorinated DWNTs are presented in the Table 1. The integral intensities of the CFx components, SC, SC-CF, and SC-F, were used to estimate the sample composition CF0.17 CFx by a formula х=SС-F/(SC+SC-CF+SCCF0.22 F). The largest amount of fluorine is attached to DWNT surface using F2 as a CF0.33 fluorinating agent. The energy of C-CF and C-F components increases when indicates stronger C-F bond for high fluorinated

71

XAFS study of the local structure and oxidation state of Cu impurity atoms in doped CdSe and CdSe/CdS core/shell quantum dots A.I. Lebedev1, I.A. Sluchinskaya1, S.G. Dorofeev2, P.N.Tananaev2, A. Erko3
1 2

Physics Dept., Moscow State University, Moscow, 119992, Russia Chemistry Dept., Moscow State University, Moscow, 119992, Russia 3 BESSY GmbH, Albert Einstein Str. 15, 12489 Berlin, Germany Quantum dots (QD) conjugated to biological molecules are promising luminescent markers, which can be used for molecular recognition of antigens and for labeling of specific compartments of cells in nanobiotechnology [1,2]. Unfortunately, conjugates with undoped quantum dots cannot be used in vitro because of strong absorption of luminescence by biological tissue and closeness of the luminescence decay times of QD and biological tissue. Shifting of luminescence spectra to red and infrared regions and slowing down of luminescence decay can greatly improve the sensitivity of this method. Doping of CdSe/CdS core/shell quantum dots with copper enables to achieve this purpose. However the nature of copper-containing luminescence centers is not known, and the first step in their investigation is the determination of the local structure and oxidation state of Cu in doped CdSe quantum dots. The aim of our experiment was to determine the location and oxidation state of Cu atoms in doped CdSe and CdSe/CdS core/shell quantum dots using XAFS technique. The samples studied were Cu-doped colloidal CdSe and CdSe/CdS core/shell quantum dots capped with oleic acid [3,4]. The samples were doped by three different routes: by using of copper stearate or Cu4I4(PPh3)4 as precursors during the QD synthesis, and by “etching” of undoped quantum dots in the copper stearate solution. For XAFS studies concentrated solutions of QD in dodecane were placed into silica tubes with very thin walls. X-ray absorption spectra were collected at the Cu K edge (8.979 keV) in fluorescent mode on the station KMC-2. The intensity of monochromated radiation was measured with an ionisation chamber; the intensity of fluorescent radiation was measured with a p-i-n-diode or an energy-dispersive RÖNTEC detector. For each sample 5–7 spectra were recorded at 300 K, they were then independently processed and the obtained spectra were finally averaged. A few additional EXAFS spectra were collected at the Se K edge (12658 eV) to check the structure of QD. The analysis of EXAFS spectra was performed in the traditional way. The X-ray fluorescence analysis of ten Cu-doped samples of CdSe and CdSe/CdS core/shell colloidal quantum dots was used to determine the concentration of copper in them. Six samples with highest concentration of Cu (0.3–4.5%) were selected for further EXAFS experiments. It was observed that the larger was the radius of quantum dots, the larger amount of Cu could be incorporated into them. To estimate the oxidation state of Cu in QD, XANES spectra for five Cu-containing reference samples (CuSe with klockmannite structure, Cu2Se with berzelianite structure, CaCu3Ti4O12, Cu metal and copper stearate) were measured. The spectra for two typical samples of QD are compared with XANES spectra for Cu2Se and copper stearate reference compounds in Fig. 1. It is seen that the XANES structure for QD prepared in different conditions have a different shape. In most of QD the Cu oxidation state was +1; only two samples prepared using copper stearate as a precursor demonstrated a shoulder at 8986 eV (see Fig. 1), thus indicating that 15–20% of Cu atoms in QD were in the +2 oxidation state. It is interesting that in the samples “etched” in the copper stearate solution all Cu atoms remained in +1 oxidation state. The analysis of EXAFS spectra obtained at the Cu K edge revealed two types of local structures. In samples, in which all Cu atoms were in +1 oxidation state, the neighboring Se atoms were located at a distance of 2.38±0.01 Å. For samples with both +1 and +2 oxidation states of Cu the Cu-Se distance was shorter and equal to 2.31–2.33 Å and an additional signal from oxygen atoms at a distance of 1.93–1.94 Å was detected (see Fig. 2). The number of O neighbors was 4–8 times smaller than the number of Se neighbors.
72

Fig. 1. XANES spectra obtained at the Cu K edge for two reference compounds (Cu2Se and copper stearate) and two Cu-doped CdSe colloidal quantum dots.

Fig. 2. EXAFS spectra obtained at the Cu K edge for CdSe colloidal quantum dots doped using copper stearate precursor.

We think that in the samples, where a part of Cu atoms are in +2 oxidation state, copper atoms are located on the surface of QD and are bounded with oxygen atoms of oleic acid. As a result of Cu-O interaction, the oxidation state of Cu changes from +1 to +2. In all other samples the Cu atoms are located in the inner part of quantum dots and are in +1 oxidation state. Another interesting feature of most of studied colloidal QD is a very small contributions from the second and more distant shells to EXAFS spectra obtained at the Cu K edge. DebyeWaller factor for Cu-Se bond (the first shell) was about 0.008 Å2, which is typical for thermal motion at 300 K, and so the local distortion of Cu atom in tetrahedra can be excluded. This means that some unusual kind of static disorder is present in CdSe QD. The only exclusion was

73

the sample # 923-42, the QD of which had a tetrapod shape; strong contributions from the second and more distant shells were observed in EXAFS spectra for this sample. The data analysis of EXAFS spectra obtained at the Se K edge revealed very similar Se-Cd distances of 2.61 Å in the first shell for all samples, independent of whether QD had a round or tetrapod shape. An interesting effect of increasing of coordination number in the first shell by about 20% was observed for CdSe quantum dots covered by CdS (core/shell QD). We suppose that in pure CdSe quantum dot some Se atoms are located on its surface, and so they have a decreased (2–3) number of Cd neighbors. After covering its surface with a few atomic layers of CdS, the number of Cd atoms that surround Se increases to 4, and the average coordination number also increases. Acknowledgement. The authors are grateful to Russian-German laboratory of BESSY for hospitality and financial support. References 1. M.J. Bruchez, M. Moronne, P.Gin et al. Science 281, 2013 (1998). 2. W.C.W. Chan, S. Nie. Science 281, 2016 (1998). 3. R.B. Vasiliev, S.G. Dorofeev, D.N. Dirin, D.A. Belov, T.A. Kuznetsova. Mendeleev Commun. 14, 169 (2004). 4. P.N. Tananaev, S.G. Dorofeev, R.B. Vasiliev, T.A. Kuznetsova. Inorg. Mater. 45, N 3 (2009) (in press).

74

Effective control of palladium catalytic selectivity by surface doping and structure Dmitry Zemlyanov1,2, Bernhard Klötzer3, Axel Knop-Gericke4, Robert Schlögl4
2

Materials and Surface Science Institute, University of Limerick, Limerick, Ireland Birck Nanotechnology Center, Purdue University, West Lafayette, IN 47907-2057, USA 3 Institut für Physikalische Chemie, Universität Innsbruck, A-6020, Innsbruck, Austria 4 Fritz-Haber-Institut der Max-Planck-Gesellschaft, Faradayweg 4-6, D-14195 Berlin, Germany

1

The proposal was supported through Integrated Infrastructure Initiative I3 (the beamtime project: BESSY-BM.08.2.80336). Palladium is considered as the best catalyst for the catalytic combustion/partial oxidation of methane and other small hydrocarbons, which is an environmentally benign process for power generation with low NOx emissions and for removal of residual methane from the emission gases of methane-powered vehicles [1]. Investigation of the oxidation mechanism of palladium and the mechanism of oxide decomposition is critical for developing/improving Pd-based catalysts. Palladium oxidation was found to consist of formation of several partly metastable surface oxide phases [2-4]. The methane/ethylene total oxidation mechanisms were proposed based on this study [5, 6]. Doping is also effective tool to control activity/selectivity of a catalytic reaction can be ruled. We have studied the subsurfacecarbon doped Pd [5] and the intermetallic Pd/Zn formed on Pd(111) because an ordered PdZn surface alloy forms an valence band structure similar to those of Cu(111), explaining the similar catalytic activity of PdZn and Cu in methanol steam reforming (MSR) and water gas shift reaction. The goal is to investigate how the palladium chemistry depends on the stereo-electronic influence of these different dopants. Investigation of the electronic state of palladium and the mechanism of formation/decomposition of Pd-O, Pd-C and intermetallic Pd surfaces is critical for developing/improving Pd-based catalysts. This project continues our in-situ XPS study of model palladium catalysts for selective ethylene/methanol oxidation, methanol steam reforming, water gas shift and selective hydrogenation. PdZn and PdGa surface alloys formed on the surface of Pd(111) and Pd(110) serve as novel model catalysts. First, we investigated thermal stability of the PdZn alloy in vacuum. Zinc was deposed on Pd(111) surface at room temperature and then heated in vacuum. Two Pd 3d components Tempering Pd(111) + 4 ML Zn @300 K, yielding 1 ML surface corresponding to Zn-rich and Zn-lean alloy phase were monitored as shown in Figure 1. The transformation from the Zn-rich phase to the Zn-lean phase occurred above 260°C
100.0 90.0 80.0 70.0 60.0 50.0 40.0 30.0 20.0 10.0 Relative concentration, % Pd-Zn rich 336.3eV

We found that on Zn-doped Pd, both methanol decomposition and MSR yield mainly CO2 and H2CO up to 260°C, in contrast to Zn-free Pd, mainly yielding CO and strongly deactivating with carbon. Above ~260°C the selectivity of the Zndoped Pd also changes toward CO due to Figure 1. Heating PdZn alloy in vacuum. carbon dissolution in the Pd bulk and progressive decomposition of the PdZn alloy. The surface most active toward CO2 consists of regular arrays of PdZn islands of 180 - 200 nm size. Above 300°C the regular PdZn structure decomposes and large, likely Zn-rich, islands segregate and Pd/C “alloy” forms.
0.0 0.0 100.0 200.0 300.0 400.0 500.0 600.0 Temperature, °C

Pd-Zn lean 335.4335.1eV

75

The pressure gap was localized in between 0.25 and 30 mbar. Thus, the PdZn-alloy becomes structurally strongly altered during the catalytic measurements at ~30 mbar: (i) the originally densely arranged alloy islands spatially separate at 230°C; (ii) at higher successive deactivation of the CO2 + HCHO formation and activation of steady CO-formation, due to increasing Pd metal surface was detected; (iii) at 350°C larger amounts of C become deposited and dissolved, successively deactivating the surface in analogy to clean Pd. At low MeOH/H2O pressure of 0.25 mbar the alloy is stable and the deactivation is much slower. The valence band does not change up to 300°C (Figure 2). Since gallium might have similar to zinc electronic effect on Pd, Zn doping was replaced with Ga. Figure 2. The valence band spectra First, we developed the procedure of preparation of obtained from PdZn-alloy during PdGa alloyed surface: (i) Ga2O3 was deposed on heating in 0.25 mbar MeOH/H2O. the surface of Pd foil and Pd(111) and then (ii) gallium oxide was reduced in vacuum or in H2 (Figure 3). Actually, at 250°C reduction is kinetically limited: The sharp Ga 3d peaks appear after 40min in H2 (inset of Figure 2). PdGa alloy formation on Pd(111) surface occured at 350°C. The higher temperature compare with palladium foil points to importance of the surface packing on alloy formation. Pd foil surface is supposed to have a lot of defects and therefore PdGa alloy forms at lower temperature. The valence band regions of PdZn, PdGa and pure Pd are compared in Figure 4. The spectra from the PdZn and PdGa surface alloy are very similar, therefore catalytic behavior for this surfaces is expected to be similar. The kinetic measurements are in the progress now.

Figure 3. The valence band/Ga 3d spectra demonstrating PdGa alloy formation on Pd foil. Keeping in H2 at 250°C for 40 min resulted in PdGa alloy appearance as shown in inset.

76

References [1] G. Zhu, J. Han, D.Y. Zemlyanov and F.H. Ribeiro, Journal of the American Chemical Society 126 (2004) 9896. [2] H. Gabasch, W. Unterberger, K. Hayek, B. Klötzer, G. Kresse, C. Klein, M. Schmid and P. Varga, Surface Science 600 (2006) 205. [3] D. Zemlyanov, B. Aszalos-Kiss, E. Kleimenov, D. Teschner, S. Zafeiratos, M. Haevecker, A. Knop-Gericke, R. Schloegl, H. Gabasch, W. Unterberger, K. Hayek and B. Kloetzer, Surface Science 600 (2006) 983. [4] H. Gabasch, W. Unterberger, K. Hayek, B. Kloetzer, E. Kleimenov, D. Teschner, S. Zafeiratos, M. Haevecker, A. Knop-Gericke, R. Schloegl, J. Han, F.H. Ribeiro, B. Aszalos-Kiss, T. Curtin and D. Zemlyanov, Surface Science 600 (2006) 2980. [5] H. Gabasch, E. Kleimenov, D. Teschner, S. Zafeiratos, M. Havecker, A. KnopGericke, R. Schlogl, D. Zemlyanov, B. Aszalos-Kiss, K. Hayek and B. Klotzer, Journal of Catalysis 242 (2006) 340. [6] H. Gabasch, K. Hayek, B. Kloetzer, W. Unterberger, E. Kleimenov, D. Teschner, S. Zafeiratos, M. Haevecker, A. Knop-Gericke, R. Schloegl, B. Aszalos-Kiss and D. Zemlyanov, J. Phys. Chem. C FIELD Full Journal Title:Journal of Physical Chemistry C 111 (2007) 7957. [7] H. Gabasch, W. Unterberger, K. Hayek, B. Kloetzer, E. Kleimenov, D. Teschner, S. Zafeiratos, M. Haevecker, A. Knop-Gericke, R. Schloegl, J. Han, F.H. Ribeiro, B. Aszalos-Kiss, T. Curtin and D. Zemlyanov, Surface Science 600 (2006) 2980.

Figure 4. Comparison PdZn and PdGa valence band regions.

77

Photoemission study of monodispersed Pt nanoparticles deposited onto carbon substrate for PEMFC application L.V. Yashina1, S.A. Gurevich2, M.M. Brzhezinskaya3, V.S. Neudachina1,4, Yu.A. Dobrovolsky4, E.Gerasimova4
1

Department of Chemistry, M.V. Lomonosov Moscow State University, Leninskie gory 1-3, 119991 Moscow, Russian Federation 2 Ioffe Physical Technical Institute, 26 Polytekhnicheskaya st., 194021 St Petersburg, Russian Federation 3 BESSY GmbH, Albert-Einstein-Str. 15, 12489 Berlin, Germany 4 Institute of Problems of Chemical Physics RAS, 142432 Chernogolovka, Moscow Region, Russian Federation

Recent progress in power engineering is closely connected with the low-temperature proton-exchange membrane fuel cells (PEMFC). In comparison with traditional systems, PEMFCs have many advantages, e.g. high fuel conversion efficiency and low noise level. PEMFC includes a cathode, an anode, and PEM. The catalysts deposited onto the electrode surface activate hydrogen ionization at the anode and interaction of the transferred protons with oxygen at the cathode. Nanosize particles of platinum-coated high-disperse carbonaceous substrate can be used as a catalyst for the cathode and the anode processes in the polymer electrolytes. The catalyst support material should have high specific surface area, ability to activate the catalyst, high electrical conductivity, corrosion stability and optimal hydrophobic/hydrophilic properties. In our research, we employed Toray TGP-H-060 carbon paper, which meets all of these requirements.
9 8 7 6 5 4 3 2 1 0

1,6 1,7 1,7 1,8 1,8 1,9 1,9 2,0

r,n Fig.1. Schematic view of laser electrodepositon setup (left) and size distribution of the obtained Pt clusters (right) The surface of metal nanoparticles is often charged due to numerous reasons, which are directly related to their catalytic properties; among such reasons electron transfer between the metal granules and the substrate and temperature-activated charge transfer (fluctuations) between the metal granules are worth noting. It is known from experiments that there is an optimal surface coverage, at which the surface charging effect (and hence the catalytic activity) increases drastically [1]. In order to study the nanoparticles/support system in detail, it is necessary to use a model system, in which the size distribution of the particles is very narrow (i.e. the particles are of almost the same size). In our study, we employed laser electrodispersion to obtain monodispersed spherical amorphous Pt granules 1.8 nm in diameter. The schematic representation of the method is given in fig. 1(left); fig.1 (right) represents the size distribution of the Pt particles. The method allows to deposit nanoparticles
78

uniformly over the surface of the substrate. Samples with different platinum loads were obtained (Pt load from 2 to 24 µg/cm2). The catalytic 80 1,0 13с performance of the samples 26с 70 Power, U, V was measured in a test 40с W PEMFC; the obtained data 0,8 60 indicate that the dependence 50 of catalytic performance on 0,6 the Pt load has a maximum at 40 approx. 4 µg/cm2 (see fig. 2 0,4 30 for detail). In order to explain this behavior, we have 20 0,2 performed photoemission 10 studies of the samples with different Pt load. 0,0 0 0 20 40 60 80 100 120 140 160 180 200 220 240 260 The photoemission experiments have been Ток, мА I, mA undertaken at the RussianGerman beamline (RGBL). Fig.2. The performance dependence of the PEMFCs The spectra were recorded with Pt anode catalysts of different metal load (2 using MUSTANG end station µg/cm2 -red, 3.9 - µg/cm2- green and 6 µg/cm2- blue). equipped with Phoibos 150 The cathode is E-TEK Pt/C 20%, Pt load 0.4 mg/cm2). electron energy analyzer (Specs). The Pt 4f, O 1s and C 1s spectra were recorded at a variety of photon energies (3301030 eV). For the data analysis, the spectra were fitted by the Gaussian – Lorentzian convolution functions with simultaneous optimization of the background parameters. The line asymmetry was described with Doniach-Šunjić (DS) function. Alongside with the monodispersed Pt/carbon paper samples, we also studied a metallic bulk Pt sample as a reference (Fig. 3a). It is well known that the core-level photoemission spectra of metals have several sizedependent parameters, e.g. peak position or the binding energy, asymmetry index, Auger parameters, plasmon losses etc. The spectra of Pt monodispersed nanoparticles obtained at hν=485 eV, which corresponds to high surface sensitivity, are shown in Fig. 3b-e. The sample with a minimal Pt load does not have any Pt particles at the surface of the support, since no Pt is observed in the spectra at hν=485 eV, and the spectra measured at higher energies (e.g. hν=1030 eV) include two components corresponding to Pt nanoparticles and bulk platinum which can be related to the metallic behavior of the aggregates. Hence we suppose that at low coverages Pt is aggregated in the pores of the carbon paper, thus not being detected by photoemission under high surface sensitivity conditions. The samples with higher Pt load (Fig. 3c-e) include at least two components; though the particles do not conglomerate after laser electrodispersion [1], one of these components is definitely related to the bulk Pt metal (71.2 eV, asymmetry parameter α=0.23). This component most probably appears due to a good contact between the particles, since, according to the data obtained within the framework of the present study, they are not covered with an oxide or any surface contaminant. The second broader component shifted 0.6 eV towards higher binding energies corresponds to Pt in separate clusters with the diameter of 1.8 nm, which is in line with the dependence of the Pt 4f binding energy on the particle size [2,3]. The spectral data indicate that the contact between Pt clusters (i.e. the relative intensity of the first component) is minimal for the sample with Pt load 3.95 µg/cm2 (Fig. 3b); for this sample the maximum of catalytic activity is also observed. Thus we can conclude that the catalytic activity is influenced by the size effect of the second order, i.e. not only by the size of the particles themselves, but also by the intercluster spacing, which influences the number
Мощность, мВт Напряжение, В

79

1.0

a) Pt
0.8

0.6

0.4

0.2

0.0 1.0 84 82 80 78 76 74 72

70

0.8

b) 3.95 µg/cm smaller clusters

2

0.6

0.4

0.2

of contacts between the clusters. If there are a lot of direct contacts between the particles, their behavior similar to large clusters or even to bulk metals, which annuls the nanosize effect in catalysis. It should be noted that the catalyst obtained by laser electrodispersion may help decrease platinum load in PEMFCs, since their specific catalytic performance per 1g of Pt is two orders higher than for standard electrode clusters materials. 1.8 nm The measurements were performed in the Pt bulk framework of the Russian-German bilateral program. Partial financial support of the Russian Foundation for Basic Research is acknowledged. The authors are grateful to Mike Sperling for technical assistance.

Pt bulk

intensity/a.u.

68

intensity/a.u.

0.0 1.0 84 82 80 78 76 74 72 70 68

0.8

c) 6 µg/cm

2

0.6

0.4

0.2

[1] T.N. Rostovschikova, V.V. Smirnov, V.M. Kozhevin, D.A. Yavsin, S.A. Gurevich, Nanotechnologies in Russia 1-2 (2007) 47-60. [2] J. H. Tian, F. B. Wang, Zh. Q. Shan, R. J. Wang, J. Y. Zhang, J. Appl. Electrochemistry 34 (2004) 461-67. [3] T. You, O. Niwa, M. Tomita, S. Hirono, Anal. Chem. 75 (2003) 2080-85.

intensity/a.u.

0.0 1.0 84 82 80 78 76 74 72 70 68

0.8

d) 12.9 µg/cm

2

intensity/a.u.

0.6

0.4

0.2

1.0 0.0 84 82 80 78 76 74 72 70 68

0.8

e) 24 µg/cm2

intensity/a.u.

0.6

0.4

0.2

0.0 84 82 80 78 76 74 72 70 68

Fig.3. Pt 4f photoemission spectra obtained at photon energy of 485 eV

BE/eV

80

Photoemission study of multi-wall carbon nanotubes functionalized by physiologically active substances L.V. Yashina1, M.M. Kirikova1, M.M. Brzhezinskaya2, A.S.Ivanov1, S.V. Savilov1, V.S. Neudachina1, A.Yu. Filatov1
1

Department of Chemistry, M.V. Lomonosov Moscow State University, Leninskie gory 1-3, 119991 Moscow, Russian Federation 2 BESSY GmbH, Albert-Einstein-Str. 15, 12489 Berlin, Germany

Development of drugs delivery systems is one of the ways to improve pharmacological properties of many existing medicals, as well as to develop new pharmaceuticals. Carbon nanotubes (CNTs) can potentially serve as a drug delivery system due to their unique ability to penetrate through cell membranes without any damage [1]. This is apparently due to their hydrophobic nature and affinity to bilipid layer which forms the cell walls. To decrease hazard to the living cells, recently it was proposed to modify the CNTs with different functional groups in order to provide their affinity to the inner content of cells. The present report is devoted to the experimental investigation of CNTs functionalization with izadrine (a one-substituted amine, which is a pharmaceutical analog of adrenaline) by means of SXPS and NEXAFS. The MWNTs were obtained by catalytic pyrolysis of benzene using a patented procedure [2], then purified by annealing in air atmosphere. The samples were characterized by SEM, HREM and Raman spectroscopy. The obtained MWNTs were in a shape of coaxial cylindrical or conic graphene sheets. Pristine MWNTs were treated with concentrated HCl for 6-8 h, then with 6M nitric acid for 6 h under ultrasonic or with 30% hydrogen peroxide, and finally washed by water until neutral washing solution was obtained. MWNTs were also centrifuged after each washing. The obtained carboxylated MWNTs underwent chemical modification (functionalization) with isadrine. For this purpose, a mixture of izadrine and carboxylated CNTs treated by SOCl2 (to obtain CNT-COCl) were heated up in a sealed ampoule under argon up to 500C for 6 h, then treated with ultrasound; the obtained product was washed and dried. The scheme of MWCNT functionalization is presented in Fig. 1.

Fig. 1. Scheme of MWCNT functionalization (conic nanotubes are illustrated)

The photoemission spectra have been recorded at the Russian-German beamline (RGBL) using MUSTANG end station equipped with Phoibos 150 electron energy analyzer (Specs). The O 1s, C 1s and N 1s spectra were recorded at a variety of photon energies (3301030 eV). For the data analysis, the spectra were fitted by the Gaussian–Lorentzian convolution functions with simultaneous optimization of the background parameters. Photoemission spectra of the pristine nanotubes (Fig. 2a, spectra for conic nanotubes are shown), as well as Raman and NEXAFS spectra (Fig. 3) for conic and cylindrical CNTs indicate that both CNT forms include defects (non-sp2-hybridized carbon atoms). The C 1s spectra include several components, the major of which correspond to the sp2- and sp381

hybridized carbon atoms; minor ones indicate that the surface includes some –OH and – COOH groups. The corresponding spectral parameters are given in Table 1.
1.0

a) pristine CNT
0.8

sp 2

intensity

0.6

C 1s

sp

3

0.4

0.2

C-OH
0.0 1.0 292 290 288 286 284

b) oxidized CNT
0.8

O 1s
sp
2

C=O

intensity

0.6

C 1s
C-OOH

C-OH

0.4

0.2

0.0 1.0

292

290

288

286

284

538

536

534

532

530

528

c) CNT-COCl
0.8

Cl 2p
C-OH

200.6 eV

intensity

0.6

C-OCl

C 1s
0.4 0.2

0.0 1.0 292 290 288 286 284 206 204 202 200 198 196

d) CNT-izadrine
0.8

sp 3 C-OH sp
2

N 1s
400.4 eV

C 1s
C-ONR 2 C-NR 2

intensity

0.6

0.4

0.2 406 404 402 400 398 396

292

290

288

286

284

BE, eV

BE, eV

Fig. 2. Photoemission spectra of C1s (hv=355 eV), O1s (hv=600 eV), Cl2p (hv=500 eV), and N 1s (hv=500 eV)

Spectra for the carboxylated (oxidized) CNTs are shown in Fig. 2b. Carboxylation leads to the increase of the number of –COOH and –OH groups at the surface, which is confirmed by the C 1s spectra. The spectra obtained at different photon energies show that oxidation affects not only for the upper atomic layer, but also several sublayers, as it follows from Fig. 4.

82

The spectra of the izadrine-functionalized CNTs have the most complex character. At least four different components are observed in the C 1s spectra; additional component related to C-ONR2 bond appears in the O 1s spectra, and additionally the N 1s spectrum is registered for this material. The parameters of the intensive components in C 1s spectra differ from those observed for carboxylated or pristine CNTs; we interpret them as C-NR2 and –CONR bonds. The presence of N 1s spectra (BE=400.4 eV) and the components in C 1s spectra shown in Fig. 3d clearly indicate that izadrine is chemically bonded to the surface of CNTs. The essential broadening of the N 1s peak can be attributed to different geometrical displacement of izadrine fragments at the surface. Table 1. Summary of spectral parameters for C 1s spectra obtained at photon energy 355 eV. BE, eV Peak (attribution) Relative intensities (for conic CNTs) pristine oxidized functionalized 2 sp 284.5 0.71 0.51 0.20 sp3 284.9 0.16 0.24 0.24 C-OH 286.0 0.07 0.12 0.28 C-OOH/ 288.6-288.8 0.02 0.08 0.14 C-OCl/C-NR2 C-ONR2 291.2 0 0 0.05
C1s NEXAFS spectra
1.0
0.30

CNT-COOH

0.8

0.25

sp3 -OH -COOH

relative intensity

electron yield

0.6

0.20

0.15

0.4

0.2

CNT conic pristine CNT cylindric pristine CNT-izadrin CNT conic oxidazed graphite

0.10

0.05

0.00

0.0 280 290 300 310 320

5

10

15

20

25

30

photon energy

IMFP λ, A

Fig. 3. NEXAFS-spectra C1s

Fig.4. The relative intensities of the spectral components for the C 1s oxidized CNT spectra obtained at different photon energies

By comparing the photoemission and NEXAFS spectra, it was confirmed that conic CNTs are more reactive than cylindrical CNTs due to many end-atoms of graphene layers. During carboxylation up to 3 atomic layers are modified, with partial oxidation of carbon atoms to –COOH and –OH groups. Izadrine fragments are effectively anchored to the disordered structure, and fragments apparently demonstrate different displacement at the CNT surface. Such izadrine-functionalized CNTs can be potentially applied as a drug-delivery system. The measurements were performed in the framework of the Russian-German bilateral program. Partial financial support of the Russian Foundation for Basic Research is acknowledged. The authors are grateful to Mike Sperling for technical assistance. [1] B. Кateb et al, NeuroImage 37 (2007) S9. [2] S.V. Savilov, G.A. Zosimov, V.V. Lunin, Application for Patent of the Russian Federation no. 2 005 132 267 (in Russian).

83

Al doped ZnO for methanol reforming catalyst: characterization and in-situ synchrotron study
Giulio Lolli*, Detre Teschner, Malte Behrens and Robert Schlögl Fritz Haber Institute, Max Planck Society, Department of Inorganic Chemistry, Faradayweg 4-6, 14195 Berlin (Germany) *lolli@fhi-berlin.mpg.de

Introduction The Cu/ZnO/Al2O3 catalyst for methanol production/reforming is a well industriallyestablished catalyst but the exact synergistic mechanism between the Cu metallic phase and the mixed oxide had not been understood in detail. Considering the potential importance of such catalyst in a hydrogen-based economy - methanol is a promising lightweight H2-carrier molecule [1] - it is important to understand the nature of the active site and optimize its performance to the specific application. In this contribution we want to elucidate the importance of Al doping for the properties of the ZnO substrate. It was recently shown that for extremely low Al content (in the range 2-4 mol%) the Al3+ ions are preferentially located in the zincite lattice, whereas for higher Al content a segregation of Al-rich phases (ZnAl2O4, Al2O3) is observed [2]. The typical industrial catalyst ranges between 15 and 20 mol% Al.

Materials and Methods The ZnO/Al2O3 support was prepared with Al contents ranging from 0–15 mol% by coprecipitation. The support was then characterized by synchrotron based environmental XPS and X rays absorption techniques at the Al and O K-edges and Zn L3-edge under several different working conditions: in reducing environments (H2 @ 250°C), followed by exposure to CO2 at the same temperature. This was performed at the in-situ setup at the Bessy-II beamline ISISS. Further sample characterization includes XRD, and EPR spectroscopy.
27

Al-MAS-NMR, UV-VIS

Results and Discussion The Al doped ZnO with low Al content presents a different electronic structure than the same support with no or high amount of Al. The band gap energy varies from 3.3 to 3.1 eV with the Al content. The range 2-4 mol% is characterized by a minimum of band gap, which is correlated to the amount of Al atoms in the ZnO lattice and might be due to the creation of interband states. To verify this hypothesis we studied the aluminum K-edge NEXAFS spectra (shown in figure 1). The edge transitions represents the 1s→3p transitions inside the Al atom.

84

Two features at 1565 and 1567 eV can be ascribed to aluminum in tetrahedral and octahedral coordination respectively [3]. A third broad feature at 1570 eV is a multiple scattering contribution. In our samples the two peaks at 1565 and 1567 eV were not individually resolved, but convoluted in a single peak centered between the two previous values, so a reliable fitting and quantification of the 2 components, like suggested by Shimizu [3], cannot be performed. Nevertheless many differences between the 3% and 15% Al content spectra can be seen (figure 1). The low Al content sample presents a visible shoulder at low energies (highlighted by an arrow) and the convoluted peak is very close to the value for tetrahedralAl. This suggest that the 4-fold coordinated Al component in this sample is sensibly higher than in the 15% Al that looks more similar to the NEXAFS spectra of transitional alumina with a distribution of T and O sites [3,4]. Indeed the 15 mol% Al sample presents some forms of alumina segregation. This result is in perfect agreement with the idea that at low doping level the Al ions substitute the Zn in the Tetrahedral sites of the HCP ZnO Wurtzite structure. We believe this to be responsible of the band gap energy contraction of these samples.

This different electronic structure can also lead to different properties and behavior of the material under operational conditions. By mean of environmental XPS at different excitations energies we measured the surface and near-surface composition of the material looking at the Zn3p – Al2p photoemission lines in the range 100-70 eV B.E. We can see from figure 2 that the low Al content sample shows a pronounced migration of the Al on the surface during reduction and also in contact with CO2, during the simulated reaction conditions. This evolution is extremely pronounced and is not observed in higher Al content samples. These results are consistent with a modification of the band gap energy under these conditions (not shown here). Partial surface reduction of ZnO may attract more Al to the surface and can be assumed as the driving force for formation of this newly reconstructed material, which also interacts strongly with the gas phase reactants like CO2. This suggests that a similar support can actually participate actively in the catalysis or modify the properties of the metal (Cu) that is neighboring. Again all these properties are not observed in high Al content supports, which do not modify themselves during the reaction.

Conclusions Nanocristalline Al doped ZnO (2-4 mol% Al) represents a promising material as support for methanol steam reforming catalysts. By mean of NEXAFS we were able to find that, at low concentrations, the Al occupies preferentially the tetra-coordinated sites inside the HCP ZnO framework. This however tends to migrate towards the surface if exposed to H2 environment,

85

probably due to the formation of Oxygen vacancies that favor the ion mobility. All these aspects should be considered in engineering the catalyst support.

5

Zn:Al 97:3 Zn.Al 85:15 Normalized Absorbance [a.u.]
4

3

2

1

0 1555

1560

1565

1570

1575

Energy [eV]

Figure 1. NEXAFS Al-K edge showing Te- Figure 2. Al atomic % calculated from XPS trahedral and Octahedral Al coordination spectra of the Zn3p and Al2p photoemission regions. 1s→3p transitions. References 1. 2. 3. 4. Turco M., Bagnasco G., Costantino U., Marmottini F., Montanari T., Ramis G., Busca G. J. Catal. 228(1), 43-55 (2004). Miao S., Naumann d’Alnoncourt R., Reinecke T., Kasatkin I., Behrens M., Schlögl R., Muhler M. Eur. J. Inorg. Chem., accepted. Shimizu K., Kato Y., Yoshida T., Yoshida H., Satsuma A., Hattori T. Chem. Commun. 1681-82 (1999). Bancroft G.M., Fleet M.E., Feng X.H., Pan Y. Am. Min. 80, 432-40 (1995).

86

In situ XPS study of selective oxidation reactions on Ag
T. C. R. Rocha, A. Oestereich, M. Hävecker, R. Blume, D. Teschner, A. Knop-Gericke, R Schlögl Department of Inorganic Chemistry, Fritz-HaberInstitut der MPG, Faradayweg 4-6, 14195 Berlin, Germany Introduction The remarkable catalytic activity of Ag in partial oxidation reactions has been known for decades. Particularly, the ethylene epoxidation and formaldehyde synthesis are two industrially important reactions which received most attention in the past years [1,2]. Both reactions result in important primary chemicals used to produce a wide variety of materials, which find applications in different areas. It is generally accepted that the key to comprehend the catalytic activity of Ag begins with a deep understanding of the interaction of oxygen with Ag. But the formation of different oxygen species on Ag is a complex function of temperature, gas phase composition, pressure and also the structure (defects, grain boundaries) and morphology (exposed facets, particle sizes) of the Ag catalyst [2,3]. This fact makes the correlation between the surface electronic structure of the catalyst with its activity/selectivity a very difficult task. In order to shed light on this question, we have investigated the ethylene epoxidation and methanol oxidation reactions under realistic temperatures (180 °C – 650 °C) and in the mbar pressure range by in situ X-Ray Photoelectron Spectroscopy (XPS) combined with in situ Mass Spectrometry (MS). Experimental The experiments were performed using a high pressure XPS endstation at ISISS beamline. The catalyst samples were mounted inside the reaction cell, 1.3 mm away from the first aperture to the differentially pumped stages of the lens system of the hemispherical analyser. The samples were heated from the back side using an infrared laser system. The total pressure for both reactions was kept constant at 0.50 mbar by a pressure controlled valve. The partial pressure of the gasses was regulated by calibrated mass flow controllers. The O1s, Ag3d, C1s core-levels and the valence band were recorded under working conditions by XPS together with the gas phase composition, which was monitored on-line by an electron impact Quadrupole MS (QMS) and a Proton Transfer Reaction MS (PTRMS). Results and discussion

87

The selective oxidation of methanol to formaldehyde was investigated on Ag foils with different methanol-to-oxygen mixing ratios at 450 C. Figure 1a shows the O1s core level for mixing ratios of 1:1, 2:1 and 6:1. Since the incident x-ray beam irradiates not only the catalyst surface but also the gas phase molecules, the spectra show gas phase peaks (Eb >534 eV) together with surface peaks. Three oxygen components could be distinguished by fitting the spectra, namely Oa (529.7 eV), Ob (531.2 eV) and Oc (532.8 eV). Comparing to the literature, Oa could be assigned to embedded oxygen, the so called O-gamma and Ob to dissolved O species in the Ag bulk (O-beta). Accordingly, Oc can be assigned to a combination of H2O and OH species formed during the reaction and SiO2, a contaminant deposited on the Ag surface. For the ethylene epoxidation reaction, unsupported Ag powders were used as catalysts and the investigation was done under different reaction mixtures at 180 C. In this case (figure 1b), mainly two components in the O1s XPS spectra were distinguished by fitting. They were addressed as the electrophilic (530,9eV) and nucleophilic (529,5eV) oxygen species, according to the literature [4], although the corresponding binding energies are slightly shifted.
(a)
4,0 3,8 3,6 2,3
4,4 3,8

gas phase

Oc

Ob

Oa 6:1

(b)
5,6 5,2 4,8

Oel
Onucl

2:1

CPS (10 )

2,2

4

CPS (10 )

1:2
3,6 3,4 3,2 4,2

2,1 2,0 1,9

2:1

4,2

4

1:1
4,0

4,0 3,8

1:2.5

3,8 537 536 535 534 533 532 531 530 529 528 527

3,6 534

533

532

531

530

529

528

527

Binding energy (eV)

Binding energy (eV)

Figure 1: (a) XPS O 1s core level for the methanol oxidation reaction under working conditions (0.5 mbar, 450 C) for different mixing ratios (CH3OH : O2) 6:1 (top), 2:1 (middle), 1:1 (bottom). (b) XPS O 1s core level for the ethylene epoxidation reaction under working conditions (0.5 mbar, 180 C) for different mixing ratios (C2H2 : O2) 2:1 (top), 1:2 (middle), 1:2.5 (bottom).

An important point is that the relative amounts of the different oxygen species in the silver surface formed under working conditions were found to depend on the gas phase composition. This can be qualitatively seen comparing the different spectra in figure 1a for the methanol

88

oxidation and figure 1b for the ethylene epoxidation. In order to better assign these oxygen species for both reactions, additional studies are being held to gain further information about their physical/chemical characteristics. Although the nature of the O species under methanol oxidation is still not fully addressed, they can be compared with catalytic aspects. Figure 2b shows the relative selectivity (CH2O to CO2 signal ratio) for different mixing ratios (black dots) together with the ratio of the areas of the XPS peaks Ob and Oa measured under working conditions (figure 2a). A good correlation can be observed, indicating that higher selectivities are obtained when the Ob/Oa ratio is higher.
Relative selectivity (%)
100 80 6 60 4 40 20 0 1 2 3 4 5 6 2 Ob/ Oa 8

0

Mixing ratio (CH3OH/O2)

Figure 2: Relative selectivity (CH2O/CO2) and XPS area ratios for low binding energy peaks (Ob/Oa) for different feed ratios.

Conclusions The surface of Ag catalysts under working conditions was characterized for the methanol oxidation and ethylene epoxidation. The oxygen species found in each case were found to be dependent on the gas phase composition. Additionally, in the case of methanol, the changes in the oxygen species present a good correlation with the variations in the selectivity of the reaction. These new insights on the role of different oxygen species can contribute to better understand the mechanisms of selective oxidation reactions over silver, what might lead to the development of improved catalysts. Acknowledgement BESSY staff is gratefully acknowledged for continuous support. References
[1] X. Bao, M. Muhler, B. Pettinger, R. Schloegl and G. Ertl, Catalysis Letters 22 (1993) 215-225. [2] V.I. Bukhtiyarov, A.I. Nizovskii, H. Bluhm, M. Hävecker, E. Kleimenov, A. Knop-Gericke, R. Schögl, J. Catal. 238 (2006) 260. [3] X. Bao, M. Muhler, T. Schedel-Niedrig, R. Schlögl, Phys. Rev. B 54 (1996) 2249-2262 [4] V. I. Bukhtiyarov, M. Hävecker, V. V. Kaichev, A. Knop-Gericke, R. W. Mayer, R. Schlögl, Phys. Rev. B 67 (2003) 235422.

89

XPS Area Ratio

The role of carbon and hydrogen in palladium catalyzed hydrogenation
D. Teschner1, J. Borsodi1,2, M. Hävecker1, A. Knop-Gericke1, Zs. Révay2, A.Wootsch2, S. D. Jackson3, D. Torres4, P. Sautet4, R. Schlögl1
1

Fritz-Haber-Institut der Max-Planck-Gesellschaft, Faradayweg 4-6, 14195 Berlin, Germany
2 3

Institute of Isotopes, CRC, Budapest, Hungary

WestCHEM, Department of Chemistry, University of Glasgow, Glasgow, Scotland, UK
4

Laboratoire de Chimie, Université de Lyon, Lyon, CNRS, France

Introduction Catalytic hydrogenations are one of the most important processes of the chemical industry. In heterogeneous hydrogenations the majority of catalysts include palladium that is known to be very active in hydrogenating both alkynes and alkenes. Previously we have shown that carbon dissolves in the top few layers of palladium (PdC; Pd 3d peak at 335.6 eV) in the initial stage of alkyne hydrogenation and this enables selective hydrogenation [1]. Herein, we extend our previous studies addressing the fundamental differences of carbon-carbon double or triple bond hydrogenation over Pd. Furthermore, the role of carbon and hydrogen is discussed. Experimental details In situ high-pressure XPS investigations were conducted at BESSY, while in situ Prompt Gamma Activation Analysis was performed at the cold neutron beam of the Budapest Neutron Centre, Hungary. Density Functional Theory based calculations on slab models have been carried out to investigate the accumulation of carbon on/in Pd(111). The energetics of surface and subsurface hydrogen was calculated, as well. Results We have studied the near-surface region of palladium under various alkyne and alkene hydrogenation reactions. Figure 1 summarizes the state of palladium under 1 mbar alkene or alkyne hydrogenation conditions. Clearly, all investigated alkynes induce the formation of PdC since the higher binding energy component dominates the spectra. On the other hand, hydrogenating alkenes generate much weaker signal at the higher binding energy side of bulk Pd. According to our fitting procedure this peak corresponds roughly 1/3 of the whole signal intensity, which might indicate that some carbon is indeed dissolved in the upper part of the metal lattice, but we attribute the lack of a clearly distinguishable high binding energy component to the absence of PdC. Therefore, as a general rule, the result emphasizes the

90

widely different nature of the surface under alkyne or alkene feed: alkynes are hydrogenated on PdC while no carbon is incorporated in palladium under alkene hydrogenation.

Figure 1: Comparison of in situ Pd 3d5/2 spectra of Pd foil under alkyne and alkene hydrogenation at 1 mbar (H2/CxHy: 9/1) and 343-353 K. A: 1-pentene; B: propene; C: ethylene; D: 1-pentyne; E: propyne; F: acetylene. [2]

By the help of in situ Prompt Gamma Activation Analysis, we followed the hydrogen content of palladium during hydrogenation. The results indicated that unselective hydrogenation proceeds on hydrogen saturated β-hydride, while during selective hydrogenation the activity was not a function of the hydrogen content. Furthermore, the hydrogen content of Pd during alkene hydrogenation was always high, in line with the absence of PdC that would hinder the equilibration of hydrogen between surface and bulk. In order to achieve an atomistic understanding for the formation of the PdC phase we addressed the incorporation of C into the Pd substrate. The initial stages in the formation of the PdC phase proceeds via C incorporation from the surface to the first interlayer. Independently of ΘC, structures involving C atoms adsorbed only in the subsurface are energetically favored (Figure 2 top left panel) with respect to structures involving C adatoms on the surface or on both the surface and subsurface. The most stable conformation corresponds to a √3x√3 distribution of C in the first interlayer. Regarding the C distribution between the first and second interlayer the situation changes. Placing the C atoms in the second interlayer is roughly as favorable as placing them in the first. For coverage higher than ~0.3 ML, distribution of the C atoms on both first and second interlayer is energetically more favored than the population of a single interlayer, hence creating the thermodynamic driving force for the growth of the PdC phase. We considered C incorporation into deeper interlayers and, as a general rule, we observed a significant weakening of the average C binding when an increasing number of interlayers were populated; hence this will limit for deeper extension of the PdC phase.

91

Figure 2: (Top): Contour maps showing the average binding energy (colour coded, in eV) of carbon as a function of total C coverage (X axis) and the distribution of carbon atoms between interlayers (Y axis) using a Pd(111) substrate. C atoms are distributed between surface and 1st interlayer sites (left panel) or between 1st and 2nd interlayer (right panel). (Bottom): Contour maps showing the average binding energy (colour coded, in eV) of hydrogen as a function of total H coverage (X axis) on Pd(111) (left) or on the model PdC phase (right). H atoms are distributed between surface and 1st interlayer sites. [2]

Finally, the presence of C in the subsurface affects the properties of surface hydrogen. The average binding energy of H (EHb) as a function of total H coverage (ΘH) was calculated with respect to the gas phase H2 molecule. On the clean Pd(111) surface, the binding energy does not strongly depend on the coverage (Figure 3 bottom left panel). H is most stable on the surface, but the penetration into the subsurface is more favored than desorption. In contrast, the bonding properties of H are strongly modified for PdC (Figure 3 bottom right panel): adsorption on the surface is weakened and, most importantly, the accumulation of H into the subsurface is thermodynamically disfavored. Thus one role of the PdC phase is to hinder the migration of H to the subsurface, hence decreasing the Hsub / Hon ratio in the sample. The PdC phase will, in addition, prevent the migration of bulk H toward the surface. Hence, alkynes are hydrogenated selectively by surface hydrogen, since hydrogen cannot emerge from the bulk, if present at all. On the other hand, alkene hydrogenation occurs using subsurface hydrogen, as its concentration is high and no energetic barrier is built up by a subsurface carbon population. Acknowledgement The authors thank the BESSY staff for their continual support during the measurements. References
[1] a) D. Teschner, J. Borsodi, A. Wootsch, Zs. Révay, M. Hävecker, A. Knop-Gericke, S. D. Jackson, R. Schlögl, Science 2008, 320, 86; b) D. Teschner, E. M. Vass, M. Hävecker, S. Zafeiratos, P. Schnörch, H. Sauer, A. Knop-Gericke, R. Schlögl, M. Chamam, A. Wootsch, A. S. Canning, J. J. Gamman, S. D. Jackson, J. McGregor, L. F. Gladden, J. Catal. 2006, 242, 16. [2] D. Teschner, Zs. Révay, J. Borsodi, M. Hävecker, A. Knop-Gericke, R. Schlögl, D. Milroy, S. D. Jackson, D. Torres, P. Sautet, Angew. Chem. 2008, 47/48, 9274.

92

Methodology for the structural characterisation of silica supported VxOy species by means of in situ O K-edge X-ray absorption spectroscopy and its application to VxOy/SBA-15
Michael Hävecker1, Matteo Cavalleri1, Rita Herbert1, Rolf Follath2, Axel Knop-Gericke1, Christian Hess1,3, Klaus Hermann1, and Robert Schlögl1
1

Fritz-Haber-Institut der Max-Planck-Gesellschaft, Faradayweg 4-6, 14195 Berlin, Germany
2

Helmholtz-Zentrum Berlin für Materialien und Energie GmbH, Elektronen-Speicherring BESSY II, Albert-Einstein-Str. 15, 12489 Berlin, Germany

3

Eduart-Zintl-Institut für Anorganische und Physikalische Chemie, Technische Universität Darmstadt, Petersenstr. 20, 64287 Darmstadt, Germany

mh@fhi-berlin.mpg.de (M. Hävecker)

Introduction Oxide supported vanadia particles merit special attention due to their large structural flexibility combined with chemical and physical properties that make them interesting for a wide range of applications. It has been controversially debated since years which chemical bonding configuration (terminal vs. bridging) in supported VxOy is active in various catalytic reactions. Vibrational spectroscopy is one of the most prominent probes to draw conclusions about which vanadium-oxygen bond constitutes the active site [1, 2]. However, the initially accepted view of the identification of differently coordinated oxygen species in vibrational spectra of supported vanadia catalysts has been challenged recently by a systematic experimental and theoretical study [3]. Thus, there is a strong motivation for an additional probe allowing to tackle the problem of relating structural peculiarities of silica supported vanadium oxide to their functionality. Being a functional material, supported vanadium oxide is very sensitive to the ambient conditions [4]. Thus, the probe should not only provide direct access to the molecular structure of the vanadium oxide species but it must also be applicable under reaction conditions at evaluated temperature, i.e. in situ. We introduce in situ oxygen K-near edge X-ray absorption fine structure (NEXAS) measurements based on the Auger electron yield (AEY) technique as a characterisation tool to assist vibrational spectroscopy that has been applied extensively in earlier work on this system.

93

Experimental details The silica SBA-15 supported vanadium oxide samples were prepared by a controlled grafting/ion-exchange procedure consisting of (i) surface functionalisation of silica SBA-15, (ii) ion exchange of ammonium decavanadate and (iii) a final calcination step at 550°C. In situ NEXAFS measurements have been performed at BESSY II using monochromatic radiation of the ISISS (Innovative Station for In Situ Spectroscopy) beamline as a tuneable Xray source. High pressure soft X-ray absorption spectra were obtained in the presence of oxygen at elevated temperature using the high pressure endstation designed and constructed at the FHI. Details of the set-up are described elsewhere [5]. Results a) b)

Figure 1: a) presents the Auger electron yield spectrum of VxOy(10.8wt%V)/SBA-15 before (black, “as measured”) and after (red, “transmission corrected”) the reconstruction, respectively. The O2 gas phase transmission function used for the analysis procedure is shown as well. In b) a comparison of the experimental O K-edge NEXAFS spectrum of 10.8wt%V/SBA-15 between 528eV and 535eV (black solid line, “experiment”) with theoretically obtained partial spectra of a V2Si6O14H6 dimer cluster representing the contribution of different oxygen coordinations (“theory”) is shown.

Fig. 1a shows the O K-NEXAFS of VxOy/SBA-15 in 0.5mbar O2 at 400°C. Absorption of Xrays in the gas phase modifies the photon flux impinging on the sample, visible mainly in the dip at 531eV in the “as measured” spectrum. This energy dependent variation of the photon flux is proportional to the total electron yield signal of the gas molecules for the case of thin samples i.e. low gas pressures. Thus, to obtain the non-distorted NEXAFS the “as measured” spectrum has to be corrected by the O2 gas phase transmission as depicted in Fig. 1a. The analysis of the O K-NEXAFS allows a clear distinction between separate vanadia, silica and interface contribution in contrast to vibrational spectroscopy with strongly
94

overlapping contributions due to vibrational coupling. A study of catalysts with different V loadings (0wt%, 2.7wt%, and 10.8wt% vanadium) shows that the contributions of silica to the NEXAFS appear in an energy region (above 534eV) well separated from the spectral signature of oxygen bound to vanadium (approximately 528-534eV) [6]. Differently coordinated oxygen can be identified in the O K-NEXAFS spectrum by comparison with theoretical spectra obtained by state of the art density-functional theory (DFT) calculation [7]. Fig. 1b shows a detailed region of the onset of the O K-edge where the experimental curve (solid black line, “experiment”) is compared with theoretical spectra (“theory”) obtained by DFT calculations of a V2Si6O14H6 model cluster (displayed as inset in Fig. 1b). Contributions of V-O-V and V=O to the experimental spectrum as derived from the theoretical spectra are indicated by thick blue and red arrows, respectively. The comparison of the experiment with the theoretical spectra shows clearly that a complete interpretation requires consideration of oxygen in bridging coordination (V-O-V). Therefore, O K-edge NEXAFS spectra together with theoretical support allow to conclude on the presence or absence of specific oxygen bonds in VxOy/SBA-15. Thus the comparison facilitates a detailed analysis of the molecular structure of silica supported vanadium oxide. Acknowledgement The authors thank the BESSY staff for their continual support of the synchrotron based high pressure electron spectroscopy activities of the FHI. This work was supported by Deutsche Forschungsgemeinschaft (DFG, SFB 546). C. H. thanks the DFG for providing an Emmy Noether fellowship. M. C. acknowledges financial support from the AvH foundation. References
[1] B. M. Weckhuysen, D. E. Keller, Catal. Today 78, 25 (2003). [2] I. E. Wachs, Catal. Today 27, 437 (1996). [3] N. Magg, B. Immaraporn, J. B. Giorgi, Th. Schroeder, M. Bäumer, J. Döbler, Z. Wu, E. Kondratenko, M. Cherian, M. Baerns, P. C. Stair, J. Sauer, H.-J. Freund, J. Catal. 226, 100 (2004). [4] D. E. Keller, T. Visser, F. Soulimani, D. C. Koningsberger, B. M. Weckhuysen, Vibrational Spectr. 43, 140 (2007). [5] E. M. Vass, M. Hävecker, S. Zafeiratos, D. Teschner, A. Knop-Gericke, R. Schlögl J. Phys.: Condens. Matter 20, 1 (2008). [6] M. Hävecker, M. Cavalleri, R. Herbert, R. Follath, A. Knop-Gericke, Christian Hess, Klaus Hermann, and Robert Schlögl, Phys. Stat. Solidi b, submitted. [7] M. Cavalleri, K. Hermann, A. Knop-Gericke, M. Hävecker, R. Herbert, C. Hess, A. Oestereich, J. Döbler, R. Schlögl, J. Catal. (2009), doi:10.1016/j.jcat.2008.12.013.

95

Cobalt as methanol oxidation catalyst: The effect of the reaction mixture composition to the surface oxidation state.
S. Zafeiratos1, T. Dintzer1, D. Teschner2, R. Blume2, M. Hävecker2, A. Knop-Gericke2, R. Schlögl2
1 2

LMSPC, UMR 7515 du CNRS, 25 Rue Becquerel, 67087 Strasbourg, France

Fritz-Haber-Institut der Max-Planck-Gesellschaft, Faradayweg 4-6, 14195 Berlin, Germany

Introduction Cobalt and its oxides (CoO and Co3O4) exhibit interesting electronic and magnetic properties and are used as catalysts in a range of reactions. Perhaps the major application of cobalt-based catalysts is in the Fischer-Tropsch synthesis, since cobalt has been shown to efficiently convert syn gas (CO+H2) to methane or liquid fuels [1].Recently cobalt has been proposed as a very promising catalyst to replace noble metals for H2 production by steam reforming of ethanol [2]. Pure cobalt oxide surface phases have been scarcely investigated with respect to their catalytic properties [3, 4] and always by ex-situ methods. In the present work, high pressure photoelectron and soft x-ray absorption spectroscopy are applied under working catalytic conditions, to investigate methanol oxidation reaction on cobalt. Experimental In situ x-ray photoelectron and absorption spectroscopy (XPS and XAS respectively) were performed at ISISS beamline at BESSY in Berlin. The soft X-ray absorption spectra of the Co L3,2 edges were recorded in the Total Electron Yield (TEY) mode. The Co crystal was pretreated in the XPS reaction cell by oxidation (0.2 mbar O2 at 520 K) and reduction (0.2 mbar H2 at 520 K) cycles, until all residual surface carbon disappeared. Results Figure 1 displays photoemission and absorption spectra of Co 2p core level (L3,2 edge in absorption spectroscopy nomenclature) obtained from pre-oxidized cobalt surfaces at 520 K, under various gas phase environments. The already available photoemission and absorption data of cobalt oxides in the literature provide the necessary basis for identification of the cobalt oxidation states [5,6,7,8] in figure 1. In pure O2, the Co 2p3/2 photoemission peak at 779.6 eV has a weak, broad satellite (S in figure 1a) characteristic of the Co3O4 spinel phase [5, 6]. In agreement, the L3,2 edge fine structure (figure 1b) is very similar to that previously obtained on Co3O4 reference compounds [7, 8]. Finally, the Co/O atomic ratio calculated from the Co 2p and O 1s peaks was 0.69, close to the nominal value of 0.75 for Co3O4.

96

a
S

2p3/2

XPS Co 2p3/2

b

L3

NEXAFS Co L3,2

L2

O2

Intensity / a.u.

CH3OH:O2 1: 5

CH3OH:O2 2: 1

CH3OH

790

785

780

775

775

780

785

790

795

800

Binding Energy / eV

Photon Energy / eV

Figure 1. (a) Co 2p3/2 XPS (hv= 965 eV) and (b) Co L3,2 XAS spectra of Co (0001) at 520 K under 0.2 mbar O2, 0.3 mbar CH3OH:O2 =1:5, 0.2 mbar CH3OH:O2 =2:1 and 0.1 mbar CH3OH.

All spectroscopic results are consistent with the complete transformation of cobalt surface to Co3O4 when heated in O2 at 520 K. The thickness of the oxide layer is not possible to be determined, but definitely exceeds 4 nm which is the estimated probing depth of the absorption spectra [9]. The spectroscopic characteristics undergo significant modification in oxygenmethanol mixtures (MR is defined here as the CH3OH:O2 ratio). The Co 2p3/2 photoemission peak (fig. 1a) is shifted to higher energies (780.6 eV) and the satellite structure becomes broader and more intense, especially for MR=2. These are clear indications for partial reduction of Co3O4 to CoO [6]. Additionally, in pure methanol stream, the Co 2p3/2 peak (binding energy 778.3 eV) is indicative of cobalt in metallic state, verifying that methanol is a very effective reducer for cobalt oxides. The XAS spectra presented in fig. 1b confirm the photoemission results. In particular, for MR=0.2 the Co L-edges are the sum of CoO and Co3O4 reference spectra, indicating that under these conditions both cobalt oxide phases coexist. For MR=2 and pure methanol atmosphere, the Co L-edges are very much alike to that found in previous measurements for CoO [7] and metallic cobalt [8] respectively. From the discussion above, it is evident that the composition of the gas phase significantly influences the surface chemical state of cobalt. The key in catalysis is to combine spectroscopic and catalytic data, in other words to correlate the surface chemical state to the catalytic activity and selectivity. Relative selectivities of the main products, as well as methanol conversion, were calculated based on on-line QMS data (Table 1).

97

Maximum MethanolMethanol to-oxygen Conversion Ratio % 1:5 26 2:1 2 1:0 14

S(CO) 6 25 47

S(CO2) 58 13 0

S(CH2O) 13 35 9

S(H2) 0 16 44

S(H2O) 23 11 0

Table 1. Normalized product selectivities and methanol conversion rates on cobalt, derived by on-line QMS results. Data are recorded at 520 K under CH3OH:O2 reactant gas with mixture ratios 1:5, 2:1 and 1:0.

Depending on the reaction conditions CO, CO2, CH2O, H2O and H2 were detected. Selectivities are referring to constant temperature and pressure conditions (520 K, 0.1-0.3 mbar) and are expressed as the percentage in the overall (CO, CO2, CH2O and H2) production. The maximum activity was obtained just after reaching the 520 K (500 K for pure methanol) and afterwards gradual deactivation was observed, as was evident by the decrease of methanol consumption. Comparison of spectroscopic and catalytic data presented in figure 1 and table 1 provide direct indications for the catalytic behaviour of cobalt in oxide and metallic form. In particular for MR=0.2, were Co3O4 is the dominant phase, the CH3OH consumption is high and total oxidation to CO2 is favoured. In contrast, for MR=2, partial oxidation products (CO, CH2O and H2) are detected, accompanied with significantly lower methanol conversion rates (almost ten times). As showed in figure 1, in that case the pre-oxidized cobalt surfaces were reduced to CoO. Finally, in a pure CH3OH stream, metallic cobalt favours methanol decomposition to CO and H2. It is worth mentioning that above 500 K, high rate of coke deposition was observed on metallic cobalt, which is the cause of fast deactivation of this catalyst. In summary combination of in situ spectroscopy and on-line gas phase analysis testifies for the dynamic response of the cobalt surface to the reaction mixture, indicating also the effect at the catalytic behaviour. References
[1] J. Zhang, J. Chen, J. Ren, Y. Sun, Appl. Catal. A 243 (2003) 121. [2] Meng Ni, Dennis Y.C. Leung, Michael K.H. Leung, International Journal of Hydrogen Energy 32 (2007) 3238 [3] S. Tuti, F. Pepe, Catal Lett 122 (2008) 196 [4] M. M. Natile, A. Glisenti, Chem. Mater. 14 (2002) 3090 [5] M. A. Langell, J. G. Kim, D. L. Pugmire, W. McCarroll, J. Vac. Sci. Technol. A 19 (2001) 1977 [6] H.A.E. Hagelin-Weavera, G. B. Hoflunda, D. M. Minahanb, G. N. Salaitac, Applied Surface Science 235 (2004) 420 [7] D. Bazin, I. Kovacs, L. Guczi, P. Parent, C. Laffon, F. De Groot, O. Ducreux, J. Lynch, Journal of Catalysis 189 (2000) 456. [8] T. J. Regan, H. Ohldag, C. Stamm, F. Nolting, J. Luning, and J. Stohr, R. L. White, Phys. Rev. B 64 (2001) 214422. [9] M. Abbate, J. B. Goedkoop, F. M. F. de Groot, M. Grioni, J. C. Fuggle, S. Hofmann, H. Petersen, M. Sacchi, Surf. Interface Anal. 18 (1992) 65.

98

In situ XPS study of methanol decomposition on Pd/Ga2O3 catalysts and Pd foil K. Föttingera, A. Haghofera, b, G. Rupprechtera, D. Teschnerb, A. Knop-Gerickeb, R. Schlöglb
a

Institute of Materials Chemistry, Vienna University of Technology, A-1210 Vienna, Austria
b

Fritz-Haber-Institute of the Max-Planck-Society, D-14195 Berlin, Germany

This work was financially supported by the Integrated Infrastructure Initiative I3 in FP6 Contract no. R II CT-2004-506008 Background and Objectives Methanol steam reforming is a promising reaction in terms of hydrogen production for use in PEM fuel cells. The main criterion for a catalyst in this reaction is its selectivity to CO2 and H2. The common byproduct CO, formed by side or follow-up reactions such as methanol decomposition and reverse water-gas shift (RWGS) is problematic due to its poisoning effect on the electrode of a downstream fuel cell. It was previously shown [1] that palladium supported on certain reducible oxides (ZnO, Ga2O3, In2O3) provides high selectivity combined with thermal stability. As a reason for the improved selectivity the formation of an alloy or intermetallic compound (IMC) between Pd and the reduced support has been suggested [1]. The aim of this work was the investigation of alloy formation on a 5 wt% Pd/Ga2O3 powder catalyst in different atmospheres. In-situ XRD had been used previously to determine the temperature of formation of a bulk alloy/IMC under reducing conditions. In-situ XPS allowed for a surface characterization under the conditions of methanol decomposition as well as reduction in hydrogen. In addition, we studied methanol decomposition on pure (unalloyed) Pd foil as well as on Ga2O3 (without Pd) for comparison of the carbon-containing intermediates formed on the catalyst surface.

Experimental In situ X-ray photoelectron spectroscopy was performed at BESSY II at the ISIS-PGM beamline. Pure Ga2O3 and a 5 wt% Pd/Ga2O3 powder catalyst were mounted onto a temperature controlled sample holder in the form of pressed pellets. During reduction in 0.25 mbar H2, Ga3d and Pd3d regions were recorded at increasing sample temperatures and different excitation energies (160, 400 and 800 eV for Ga3d and 480, 720 and 1120 eV for

99

Pd3d) , yielding information from different depths. During methanol decomposition, the C1s region was recorded additionally. Methanol decomposition reaction was also carried out on a Pd foil at different temperatures (400, 500 and 600 K, 0.13 mbar methanol). During the reaction the Pd3d and C1s signals were followed at different excitation energies (420, 540, 660 and 1060 eV for C1s; 480, 600, 720 and 1120 eV for Pd3d). The gas phase was analyzed by online mass spectrometry.

Results Pd-Ga2O3: Previous in-situ XRD measurements under flowing H2/He had indicated that bulk alloy formation set in at 548 K. Hydrogen reduction in the in-situ XPS system was performed at temperatures of 448, 523 and 623 K and a H2 pressure of 0.25 mbar. Starting at 523 K and more pronounced at 623 K, metallic Ga appeared in the surface region, indicated by an XPS signal at binding energies 2 eV lower than that of Ga3+ in Ga2O3 (see Fig. 1). The ratio of Ga0 to Ga3+ did not strongly depend on the information depth. Methanol decomposition was investigated at temperatures of 448, 523 and 623 K as well, with similar amounts of reduced Ga species appearing as in the reduction experiment. While it can be assumed from the combination of XRD and XPS measurements that the reduced Ga is present at least partly in form of an alloy or intermetallic compound of Pd and Ga, the presence of reduced (by the atomic H supply from Pd) but unalloyed Ga cannot be excluded.

Figure 1: Ga3d region of Pd/Ga2O3 (left) and pure Ga2O3 (right) recorded at 0.25 mbar H2 and 623 K. Excitation energy: 160 eV

100

Pd-foil: For methanol decomposition reaction on Pd, two different reaction pathways can occur, with or without cleavage of the methanolic C-O bond, leading either to CO and H2 or to formation of carbon(aceous) species. Using in situ XPS, we aimed at a quantification of these two pathways at different reaction temperatures. C1s signals at 285.3 eV are attributed to adsorbed CO, CHx species give rise to XPS peaks at ~284.1 eV (see Fig. 2). At the lower reaction temperature (400 K) a significantly larger amount of adsorbed CO was detected on the Pd surface compared to 600 K (CHx:CO ratio =3 at 400 K and =6 at 600 K), but this does not result in a lower activity due to CO-blocking. At 400 K an additional Pd3d species appeared, shifted by 0.6 eV to higher binding energies, which can be attributed to an adsorbate-induced shift or to a Pd-C phase. Inclusion of such a compound was not required at 600 K reaction temperature, which rather supports the first explanation (adsorbate-induced shift). Additional C1s peaks appeared between 286 and 288 eV binding energies due to various oxygencontaining carbon species.
Pd foil
T=400 K 0.13 mbar MeOH

Pd foil

T=600 K 0.13 mbar MeOH

292.0

290.0

288.0

286.0

284.0

282.0

280.0

292.0

290.0

288.0

286.0

284.0

282.0

280.0

binding energy / eV

binding energy / eV

Figure 2: C1s region during methanol decomposition reaction on the Pd foil at 400 K (left) and 600 K (right). Excitation energy: 420 eV

Acknowledgements: We acknowledge financial support through the BESSY IA-SFS program (contract R II CT2004-506008). The support by the BESSY team is greatly appreciated.

References:
[1] Iwasa, N., Mayanagi, T., Ogawa, N., Sakata, K., Takezawa, N., Catalysis Letters, 54(3) (1998), 119-123.

101

Magnetic-field induced effects on the electric polarization in DyMnO3
R. Feyerherm, E. Dudzik, A. U. B. Wolter*, S. Valencia, O. Prokhnenko, A. Maljuk, S. Landsgesell, N. Aliouane**, and D. N. Argyriou
Helmholtz Zentrum Berlin für Materialien und Energie, Berlin, Germany * IPKM, TU Braunschweig, now at IFW, D-01069 Dresden, Germany ** now at IFE, NO-2027 Kjeller, Norway

Continuing our research program [1] on the series of magneto-electric multiferroics RMnO3, in 2008 we have performed x-ray resonant magnetic scattering (XRMS) studies of the rare earth magnetic ordering in DyMnO3, using the new superconducting magnet at the MAGS beamline for the first time. This compact magnet (Fig. 1), based on high-Tc superconductor technology, supplies a maximum field of 50 kOe and is combined with a 4 K base-T cryostat. DyMnO3 undergoes three stages of magnetic ordering of the Dy and Mn moments as function of T. At the Neel temperature TN = 39 K, a Mn sinusoidal order arises with a gradual shift of the propagation vector τMn as T is reduced. At Tlock = 18 K, τMn stabilizes simultaneously with the emergence of a spontaneous electric polarization P||c. In our previous work, we determined the commensurate Dy ordering with propagation vector τDy = ½ b* below 6.5 K. We also demonstrated that above 6.5 K the ferroelectric polarization of DyMnO3 is enhanced by a Mn-induced Dy spin order with τDy = τMn = 0.385 b* [2, 3].

Fig. 1: The new compact 50 kOe superconducting magnet at MAGS, mounted to the Euler cradle of the diffractometer together with the cryostat holder (right hand side).

DyMnO3 exhibits interesting effects induced by external magnetic field. Below 10 K, DyMnO3 shows a significant enhancement of the electric polarization P||c by a factor of up to 3.5 for magnetic fields between 10 and 50 kOe applied along a (Fig. 2). This enhancement is related to a two-step metamagnetic behavior with transitions around 20 and 50 kOe observed for T = 2 K. This suggests that a modification of the Dy magnetic ordering takes place at these transitions that, in turn, has an effect on the Fig. 2: Temperature dependence of the ferroelectric polarization. To verify this hypothesis, electric polarization Pc measured at various applied magnetic fields µ0H||a (from [4]). DyMnO3 was studied by XRMS at the Dy L3 resonance in magnetic fields H||a with scattering vectors (0 k 0). First, we have measured kscans at 4.5 K in zero field and in 20 kOe, with k values around 2+τ (Fig. 3, top). The Bragg reflection related to the individual Dy ordering τDy = ½ does vanish in an applied field of 20 kOe. Simultaneously, another reflection with τDy = 0.385 appears. This value coincides with the Mn-induced Dy ordering observed for T > 6.5 K in zero field. We conclude that the field-induced suppression of the τDy = ½ Dy ordering is accompanied by a re-emergence of the Mn-induced ordering and is directly linked to the increase of the electric polarization. From the magnetic field dependence of the (0 2+τDy 0) and (0 2+τMn 0) reflection intensities at base temperature (Fig. 3, bottom) we derive a critical field of H* = 18 kOe for the transition of the Dy ordering from τDy = ½ to τDy = τMn = 0.385. The related Bragg reflection has its maximum intensity around 20 kOe and rapidly decreases for larger fields. Thus the Mninduced ordering of Dy is gradually suppressed for magnetic fields above 20 kOe.

102

Above 40 kOe the intensity levels off, probably marking a second transition. The position of the (0 2+τMn 0) reflection does not vary significantly with the strength of the applied magnetic field (not shown), which implies that the Mn propagation vector is field-independent. These observations confirm that the first metamagnetic step in the magnetization data is indeed related to a breakdown of the independent Dy ordering. Above this transition, however, the Dy moments still carry an antiferromagnetic modulation with the same propagation vector as the Mn-induced ordering. The magnetization can only be saturated (forced ferromagnetic alignment) when this modulation is also suppressed. This point is reached above the second transition at ~50 kOe. It is also interesting to compare the magnetic field dependence of the induced ordering (Fig. 3) with the field dependence of the electric Fig. 3: (top) XRMS k-scans along (0 k 0) measured at polarization (Fig. 2). There is a direct 4.5 K in zero field and in 20 kOe applied magnetic field H||a; (bottom) Magnetic field dependence of the correspondence between the intensity of the integrated intensities of the (0 2.5 0) and (0 2.385 0) induced-ordering related Bragg reflection and reflections at 4.5 K. Open circles show the square of the magnitude of the electric polarization at 4.5 the polarization enhancement ∆Pc defined within the factor to at 20 K. We performed a quantitative analysis based text. The scaling bottomis fixedare give a match eye. kOe. Lines in the figure guides to the on the assumption that the µy and µz components of the Mn-cycloid stay unchanged (or change only slightly) under application of a magnetic field. In this case, the field-dependent electric polarization enhancement (∆Pc(H)= Pc(H) - P0) should be proportional to the size of the induced Dy moment, justified by the model presented in our previous report [3]. The related Bragg intensity is expected to be proportional to the square of the induced moment. Thus, the square of the polarization enhancement should scale with the (0 2+τMn 0) Bragg intensity. The experimental data follow this scaling well (Fig. 3, bottom) except for H ≥ 40 kOe. One possible explanation for the mismatch in this field-region is that the above assumption becomes invalid when the flop of the electric polarization at H ~ 65 kOe is approached, since this is related to changes on the Mn-sublattice. To conclude, we have shown that for H||a the intermediate field region with enhanced electric polarization for T < 6.5 K is characterized by the re-occurrence of an induced Dy ordering with τMn. Only for fields large enough to fully suppress the AFM arrangement of Dy moments with τMn, the enhancement of the electric polarization by the Dy vanishes. These observations confirm our previous conjecture [3] that this Mn-induced ordering of Dy is responsible for any enhanced electric polarization in DyMnO3. A.U.B. Wolter has been supported by the DFG under Contract No. SU229/8-1.
References [1] BESSY Highlights 2006, p. 9; BESSY Annual Report 2007, pp. 164 and 243. [2] R. Feyerherm et al., Phys. Rev. B 73, 180401(R), (2006) [3] O. Prokhnenko et al., Phys. Rev. Lett. 98, 057206 (2007) [4] T. Kimura et al., Phys. Rev. B 71, 224425 (2005).
103

Counterion distribution around charged proteins in solution studied using ASAXS Fajun Zhang1, Luca Ianaselli1, Maximilian W. A. Skoda2, Robert M. J. Jacobs3, Armin Hoell4, Dragomir Tatchev4, Frank Schreiber1
1

Institut für Angewandte Physik, Universität Tübingen, Auf der Morgenstelle 10, D-72076 Tübingen 2 ISIS, Rutherford Appleton Laboratory, Chilton, Didcot, OX11 0OX, UK 3 Chemistry Research Laboratory, University of Oxford, 12 Mansfield Road, Oxford, OX1 3TA, UK 4 Hahn-Meitner-Institut, Department of Structural Research, Glienicker Str. 100, 14109 Berlin

For the interaction of proteins, the distribution of charges is crucial. In fact, salt ions are ubiquitous in biological media, and the folding of DNA as well as protein is intimately coupled to the counterions that neutralise these macroions. Thus, a detailed understanding of the counterion distribution is essential for many biological systems. However, little is known experimentally about the counterion distribution around charged proteins. ASAXS provides the only way to study this issue by selecting the energies away and near the absorption edge of the target ions, which has been proved by recent successful studies [1-3]. The present work is devoted to the characterization of the counterion distribution around globular proteins, BSA and lysozyme, in aqueous solution using ASAXS. In the present experiment, we intend to study the ion distribution around proteins as a function of protein and salt concentrations as well as temperature and pH (as a way to control the charge) for BSA and lysozyme in solutions. Proteins with mono-, di- and trivalent salts will be studied by using ASAXS technique. Using the different valencies we will address the question to what extent the higher valency-ions are more localized and the related length scales as well as the effect of protein-ion interaction on the effective protein-protein interactions in solution. During this beamtime, we focused on BSA with trivalent salt, yttrium chloride, in solution. Our goal is to get reliable ASAXS signal for the counterions by optimizing the sample and measurement parameters, such as protein/salt concentration, measuring time and number of energies, etc. and establish the data analysis and fitting protocol for a direct description on the ion distribution around proteins in solution. Anomalous small-angle X-ray scattering (ASAXS) measurements were carried out at station 7T-MPW-SAXS of BESSY, Berlin. The detector response was calibrated using the scattering from Niobium foil at 18860 and 18800 eV. The angular scale was calibrated using the scattering peaks of Silver Behenate. We have determined the experimental absorption K-edge of yttrium in solution with and without the existence of protein as shown in Figure 1. The absorption edge is 16932 eV, with ∆E= -106 eV compared to the theoretical value (17038 eV). Five energies were selected for ASAXS measurement: with ∆E = -4, -14, -48, -152 and -538eV. Protein solutions were filled into capillaries from Hilgenberg GmbH, Malsfeld, Germany. The capillaries are made of borosilicate glass with an inner diameter of 4.0 mm and a wall thickness of 0.05 mm. The scattering of a salt solution was measured as the background, in exactly the same way as the protein solutions and was subtracted from the sample scattering. All measurements were carried out at room temperature. The raw data were corrected for transmission, fluctuation of primary beam intensity, exposure time, and the response of the detector. Bovine serum albumin (BSA) with protein concentration of 5, 10, and 20 mg/mL and salt concentration (YCl3) 30 mM, 50mM and 100 mM were measured at five energies with two sampleto-detector distance in order to cover a large q-range. Figure 2 presents the merged scattering profiles of a sample with BSA 20 mg/mL and 30 mM yttrium chloride and the deduced pure resonant term.

104

The pure resonant term was calculated from the following equation [3]: ⎡ ∆I (q, E1 , E 2 ) ∆I (q, E1 , E3 ) ⎤ 1 − S ion (q ) = ⎢ ⎥ F ( E1 , E 2 , E3 ) ⎣ f ' ( E1 ) − f ' ( E 2 ) f ' ( E1 ) − f ' ( E3 ) ⎦

(1)

f "2 ( E1 ) − f "2 ( E 2 ) f "2 ( E1 ) − f "2 ( E3 ) Where F ( E1 , E 2 , E3 ) = f ' ( E 2 ) − f ' ( E3 ) + − f ' ( E1 ) − f ' ( E 2 ) f ' ( E1 ) − f ' ( E3 ) In this study, the energies were selected that the imagery part of the scattering factor, f” remains constant for all energies. Hence the observed ASAXS signal will be determined mainly by the real part f’. The successful separation of pure resonant signal for the multivalent ion makes it possible for further understanding the binding number and thickness of counterion around charged protein molecules. Detailed data analysis and model fitting will be carried out imminently. Together with
SANS measurements on the same solution, the structure of bind-ion shell around proteins can be evaluated.

Figure 1 (left). Determination of the absorption edge for yttrium chloride in solution. Figure 2 (right). Typical ASAXS curves measured at different energies (only one was shown, E5) and the separated ASAXS curve from two energies (E1-E5). The pure resonant term deduced from the separated forms according to Eq. 1.

Nevertheless, we have demonstrated that the ASAXS measurements on the counterion distribution around charged protein molecules are feasible and that information on the ionic cloud can be obtained. The present results encourage us to continue our research on proteins with other counterions, such as Br and Rb. References [1] R. Das et al. Counterion distribution around DNA probed by solution X-ray scattering. Phys. Rev. Lett. 2003, 90, 188103 [2] K. Andresen, et al. Spatial distribution of competing ions around DNA in solution. Phys. Rev. Lett. 2004, 93, 248103 [3] G. Goerigk, et al. The distribution of Sr2+ counterions around polyacrylate chains analyzed by anomalous small-angle X-ray scattering. Europhys. Lett. 2004, 66, 33

105

Characterization of fluorinated singlewalled carbon nanotubes with X-ray absorption and photoelectron spectroscopies
M.M. Brzhezinskaya1,2, A.V. Krestinin3, G.I. Zvereva3, A.P. Kharitonov3 and A.S. Vinogradov2
2

HZB, 12489 Berlin, Germany V.A. Fock Institute of Physics, St. Petersburg State University, 198504 St. Petersburg, Russia 3 Institute of Problems of Chemical Physics RAS, 142432 Chernogolovka, Russia

1

The graphene layer forming the wall of a single-wall carbon nanotubes (SWCNTs) is a structure that is extremely stable against chemical treatment. Several problems exist, the solution to which urgently demands techniques to chemically modify the nanotube surface to be designated. First, chemical functionalization is required for the use of SWCNTs as the additive in a polymer matrix for the improvement of polymer binder in mechanical properties. Second, for the effective use of SWCNTs in many applications, it is necessary to achieve the best nanotube disintegration in the various media possible, down to the isolated tubes. Fluorination only can potentially become basis for the industrial technology of the chemically modified SWCNT production [1]. In this work, the high-resolution near edge X-ray absorption fine structure (NEXAFS) and X-ray photoelectron (XPS) spectroscopies are used to elucidate the nature of chemical bonding between carbon and fluorine atoms on the surface and inside fluorinated SWCNTs. These methods are very suitable for probing the chemical bonding in polyatomic systems due to their atomic selectivity and high C 1s absorption spectra chemical sensitivity [2,3]. A B C SWCNTs were synthesized by an electroarc D F E method using nickel-yttrium catalyst. The purified A' nanotubes had a narrow diameter distribution with an C HOPG B A D 1 E average value of ~1.5 nm. High-purity SWCNTs (~ F 98 wt.%) are obtained in the form of paper (SWCNT A' C paper). The direct fluorination was carried out at a D B E F MWCNT 2 A temperature 222 °C for 5 hours [4]. The fluorinated A' SWCNTs had ~ 35 wt.% fluorine (SWCNTs+F35%). B*C* D* SWCNT Measurements were performed at the Russian3 B* E* F* German beamline at the BESSY II. NEXAFS spectra C* of all samples were recorded at the C 1s and F 1s A B* D* A' D * SWCNT+F35% absorption edges in the total electron yield mode with 4 B* B* E* F* photon energy resolution of 75 and 150 meV, Fig. 1. C 1s NEXAFS spectra of HOPG, A respectively. The XPS spectra were recorded with the A' pristine MWCNTs and SWCNTs, exciting photon energies of 385 - 1030 eV with the MWCNT+F39% 5 SWCNTs+F35% and MWCNTs+F39%. total energy resolution of 200 meV using Phoibos 285 290 295 300 305 310 315 150 analyzer. Photon Energy (eV) The NEXAFS spectrum of the pristine SWCNTs (Fig. 1) corresponds well to the spectra of Fig. 1. C 1s NEXAFS spectra of HOPG, HOPG and MWCNTs. Small distinctions in these initial MWCNTs and SWCNTS, spectra usually appear as consistent changes in the SWCNTs+F35% and MWCNT+F39%. relative intensities of the most typical features of the A-B-C structure and inefficient broadening, resulting in “puttying” when turning from HOPG to MWCNTs and SWCNTs. First, the agreement in the spectra of HOPG and SWCNTs indicates the high structural perfection of nanotubes and a lack of any appreciable contribution of the amorphous or other phases of carbon in the samples. Next, this agreement clearly proves the key
Absorption (Total Electron Yield)
1 1 2 1 2

106

role of a single graphene layer in the formation of the C1s absorption spectra in HOPG, MWCNTs and SWCNTs [5]; it also certifies the assumption that the dependence of SWCNT conductivity band from tube bending is weak. Hereby, in both the C1s absorption spectra of nanotubes and in the HOPG spectrum, the main structures A–A' and B–F reflect transitions of 1s electron of carbon atoms into the free states of the conductivity band, which are formed from the 2p electron state of carbon atoms and have π and σ symmetry, respectively [5]. Significant differences in the structure of C1s spectra of the pristine and F-SWCNTs are observed. There is a strong decrease in A peak intensity, the emergence of new features for structures B1*, B*, C*, D* instead of A'–D structures, and isolated band E*–F* formation. We mention that, in addition to new absorption bands in F-SWCNT spectra, A band from the spectrum of the original nanotube is retained, proving the fact of incomplete nanotube fluorination. It is logical to associate the changes observed in C1s spectra of F-SWCNTs with the chemical interaction between F atoms and the nanotube wall and, subsequently, the rearrangement of the free-state spectrum in the tubes being explored by X-ray absorption. The appreciable similarity in F-SWCNT and F-MWCNT spectra allows supposing that the chemical binding of fluorine with carbon occurs in both cases, for the most part, by fluorine atoms bonding to carbon atoms on a nanotube sidewall, forming σ(C–F) bonds at the expense of the covalent mixing of F2p- and C2pz-valent electron states [5]. Obviously, such binding of fluorine atoms modifies the coordination of carbon atoms, from triangular in original nanotubes to nearly tetrahedral in fluorinated nanotubes, and it is C 1s, F 1s absorption spectra possible only when sp2 valent state B* C* D* hybridization of carbon atoms changes for E* F* sp3 hybridization in F-SWCNTs. B1* C1s In Fig. 2, the C1s and F1s absorption A SWCNT+F35% spectra for F-SWCNTs (curve 2) and FB1* C* B* 1 D* MWCNTs (curve 3) are shown. Fluorine E* F* F1s spectra were reduced to the energy scale of B1* SWCNT+F35% photons in the carbon spectrum using the C* 2 B* D* energy distance E(F1s-C1s) = 398.4 eV E* F* F1s between the ground 1s levels of F and C MWCNT+F39% atoms in F-SWCNTs, which was measured 3 by the X-ray photoemission method at 285 290 295 300 305 310 315 photon energy of 1030 eV. The F1s-spectra for F-SWCNTs and F-MWCNTs are Photon Energy (eV) characterized by very similar fine structures Fig. 2. Comparison of C 1s and F 1s absorption which, in the case of F-MWCNT spectrum spectra of SWCNT+F35% and MWCNTs+F39%. (like for carbon spectra), have more relief. F1s absorption spectrum of F-SWCNTs, energetically matched by the C1s spectrum, demonstrates a fine structure whose features perfectly correlate to new B1*–F* features, appearing in C1s spectrum as a result of SWCNT fluorination. Such correlation of F1s and C1s spectra had already been observed for F-MWCNTs [5]. This correlation means that both spectra reflect the transitions of ground F1s- and C1s-electrons into the same vacant electron states of the F-SWCNT conductivity band. New electron states are formed as a result of covalent bonding of carbon and fluorine atoms; therefore, they have a mixed (hybridized) F2p – C2pz character. C1s photoelectron spectra for F-SWCNTs are presented in Fig. 3. In accordance with the universal curve of the dependence of the electron mean free path from their kinetic energy [3], the escape depth is limited by ~0.5 to 1.5 nm, for all practical purposes restricting the sounding
B

Absorption (Total Electron Yield)

J

B

107

depth of the sample by means of the near-surface layer, which is as thick as one nanotube diameter by an order of magnitude. As it can be seen from Fig. 3, the signal belonging to carbon atoms of the graphene layer of the initial CNTs (band A) partially remains in all spectra. Moreover, in the series of spectra in Fig. 3, an explicit strengthening of this band is observed after increasing of probing depth up to ~1.5 nm from the surface of nanotube rope. That fact unambiguously indicates an appreciable decrease in the fluorination degree of CNTs located inside the rope, i.e., of all nanotubes which do not form the coating surface of the nanotube rope. This result correlates well to the previously discovered increase in the SWCNT fluorination degree upon the improvement of their dispersion, i.e., with the reduction of nanotube rope lateral size formed by them. Comparing C1s spectra of F-SWCNTs, we will mention that all of them except for the signal from the graphene layer (A band) contain three additional higher-energy B-D bands. The most interesting change in spectra that happens when the energy of exciting photons (and, therefore, the probing depth) increases from ~0.5 nm up to ~1.5 nm is the reduction of the relative intensity of the B band: it decreases more than twice in regard to C band intensity. In conclusion, the combined investigation of the F-SWCNTs by highly chemically sensitive X-ray C C 1s PES methods is presented. It was observed that fluorination process of SWCNTs is accompanied by chemical bonding between fluorine and carbon atoms C A B on the tube side walls. Fluorine atoms do not D substitute carbon atoms in graphene layers of 1030 eV C MWCNTs but they add perpendicularly to them as a B A result of covalent mixing between C 2pz and F 2p D 888 eV states. In this case the coordination of carbon atoms C B changes from sp2-triangular to sp3-tetrahedral. The A D photoelectron spectra showed that the surface of B 788 eV C SWCNT ropes is practically fully fluorinated. On the A D other hand, as deep as ~1.5 nm inside the nanotube 730 eV B ropes, the degree of nanotube fluorination significantly drops.
Intensity (Arb. Un.) A D 385 eV

282 284 286 288 290 292 294 296 298
Binding Energy (eV)

References 1. E.T. Mickelson, C.B. Huffman, A.G. Rinzler, R.E. Smalley, R.H. Hauge, J.L. Margrave, Chem. Phys. Lett. 296 (1998) 188. 2. J.G. Chen, Surf. Science Report 30 (1997) 1. 3. S. Hüfner, Photoelectron spectroscopy: Principles and Applications (Springer, Berlin 1996). 4. A.V. Krestinin, A.P. Kharitonov, Yu.M. Shul’ga, O.M. Zhigalina, E.I. Knerel’man, M. Dubois, M.M. Brzhezinskaya, A.S. Vinogradov, A.B. Preobrazhenskii, G.I. Zvereva, M.B. Kislov, B.M. Martynenko, I.I. Korobov, G.I. Davydova, V.G. Zhigalina, N.A. Kiselev, Nanotechnologies in Russia, 4 (2009) 60. 5. M.M. Brzhezinskaya, N.A. Vinogradov, V.E. Muradyan, Yu.M. Shul’ga, N.V. Polyakova, A.S. Vinogradov, Phys. Solid State 50 (2008) 587. 6. Carbon Nanotubes, Ed. by M. S. Dresselhaus, G. Dresselhaus, Ph. Avouris (Springer, Berlin, 2000).

Fig. 3. C 1s photoelectron spectra of pristine SWCNT+F35% taken at the different photon energy.

This work was supported by the bilateral program ”Russian-German Laboratory at BESSY”, the Russian Federal Special-Purpose Program (contract no. 02.513.11.3355) and the Russian Foundation for Basic Research (project no. 06-02-16998).

108

Effect of methanol on lignosulfonate macromolecules studied using small-angle x-ray scattering
¨ U. VAINIO∗ , R. A. L AUTEN† , S. H AAS , K. L EPP ANEN‡ , L. V EIGA∗1 , ‡ R. S ERIMAA , A. H OELL , AND R. G EHRKE∗
HASYLAB, DESY, Germany, e-mail: ulla.vainio@desy.de, † Borregaard Lignotech, Sarpsborg, Norway, Helmholtz-Zentrum Berlin f¨ r Materialien und Energie, Berlin, Germany, ‡ Department of Physics, University of u Helsinki, Helsinki, Finland
∗

Lignosulfonate, or sulfonated lignin, is a globular polyelectrolyte obtained from wood via chemical processing. During sulfite pulping lignin is removed from the wood and transformed into lignosulfonate. Sulfonate groups make the lignosulfonate highly soluble in water, while the rest of the macromolecule is mostly hydrophobic due to aromatic groups. The polymer has been determined to be branched, but also differing views have been presented. In aqueous solution low molecular-weight lignosulfonate has been found to exist as compact flat platelets about 2 nm in thickness. [1] Lignosulfonate has been tested as a component in low cost methanol fuel cell membranes [2]. To find out more about the inner structure of lignosulfonate and its behaviour in non-polar solvents, we added lignosulfonate into a binary mixture of alcohol and water and varied the concentration of lignosulfonate within the semi-dilute range. It is known that for some polymers the solution properties such as association into larger complexes depend on the order in which methanol and solute are added in the solution [3]. Since lignosulfonate has an overall aromatic nature a solvent with lower polarity might modulate the intra and/or intermolecular bonds. Changes in the structure of the solutions were studied using small-angle X-ray scattering (SAXS) at the 7T-MPW-SAXS beamline at BESSY. Counterions of the charged sulfonate groups are also affected by the polarity of the solvent and they were expected to attach to the sulfonate groups. With anomalous small-angle X-ray scattering (ASAXS) at the same beamline also the counterions, such as rubidium, can be seen and so the location of the charged groups can in theory then be determined. Experimental Sodium and rubidium salts of lignosulfonate were prepared at Borregaard Lignotech. The average molecular weights were 7600 g/mol and 9100 g/mol, respectively. The molecular weight of Rb-lignosulfonate is higher only due to the higher mass of Rb with respect to Na. Samples were prepared by dissolving different amount of lignosulfonate in either pure water or in water-methanol mixtures of ratio 3:1 and 1:1. (The lignosulfonate fractions were not soluble in pure methanol.) The volume fractions of lignosulfonate salts in each solution mixture were carefully adjusted to 1.0, 3.7, 7.5, 10.9, 14, and 17 %. The samples were placed in Hilgenberg glass mark tubes with about 4 mm inner diameter and 50 µm wall thickness. SAXS measurements were made with the mark tubes placed in air to avoid leakage of the methanol solution to vacuum. Due to the large background from air scattering, filters needed to be used to guard the 2D gas detector from too high intensities. Measurement time per sample at each sample-to-detector distance and photon energy was about 10 min in SAXS measurements and 30 min in ASAXS measurements. Two sample-to-detector distances were used: 1435 mm and 3804 mm. The measurements were made well below Rb K-absorption edge (15200 eV) at energies 14803, 15085 and 15166 eV to avoid fluorescence and other energy dependent
1

Present address: University of Campinas, Brazil

109

backgrounds. The magnitude of the scattering vector is defined here as q = 4π sin θ/λ, where θ is half of the scattering angle and λ is the wavelength. The data were normalized into absolute units (1/cm) using a glassy carbon of 90 µm thickness as calibration standard. Results X-ray absorption near edge structure (XANES) measurements of rubidium lignosulfonate solutions showed that no significant difference is seen in the chemical state of the rubidium in different solutions and concentrations of lignosulfonate within the measurement accuracy (Fig. 1).

1 0.9 0.8 µ d (normalized units) 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 15.19 15.2 15.21 3.7 % H2O 3.7 % H2O:MeOH 3:1 3.7 % H2O:MeOH 1:1 17 % H2O:MeOH 1:1 15.22 15.23 15.24 Energy (keV) 15.25 15.26 15.27 d µd/ dE 0.1 0.05 0 15.2 15.21 15.22 Energy (keV)

Figure 1: X-ray absorption spectra of lignosulfonate samples of different volume fraction (3.7 % and 17 %) in water, in water– methanol 3:1 and in water–methanol 1:1. µ is the linear absorption coefficient and d is the sample thickness. The inset shows the derivative of µd as a function of energy. The first inflection point of the absorption curve gives the position of the edge.

The ASAXS results for the Rb-lignosulfonate are presented in Fig. 2. A minimum visible in the separated ASAXS data of lignosulfonate in water solution is not seen for methanol– water mixtures. This feature is related to the distribution of rubidium around the lignosulfonate particles in the solution. The two lignosulfonate salts, even though prepared in the same way and having the same valency of +1, had very different behaviour in the methanol–water mixtures according to the SAXS measurements. Therefore the results obtained from ASAXS for the rubidium-lignosulfonate cannot be directly applied to other lignosulfonate salts and may not present the behaviour of the natural state of lignosulfonate.

10

0

I(q,E1)

Figure 2:

Intensity (1/cm)

10

−1

I(q,E1) − I(q,E3)
10
−2

17 % H2O 10
−3

17 % 3:1 17 % 1:1 3.7 % 3:1 3.7 % 1:1

qmin = 2.5 nm−1

10 q (1/nm)

0

SAXS intensities of some of the rubidium lignosulfonate samples at 14803 eV (I(q, E1 )) and the difference of SAXS intensities measured at 15166 and 14803 eV (I(q, E1 )−I(q, E3 )). The tail part of the separated curves differs from the total scattering and a strong minimum appears for the concentrated sample in water solution at q = 2.5 1/nm.

We gratefully acknowledge the financial support for travel costs and the excellent service at BESSY.

110

References
[1] U. Vainio, R. A. Lauten, and R. Serimaa. Small-angle x-ray scattering and rheological characterization of aqueous lignosulfonate solutions. Langmuir, 24(15):7735–7743, 2008. [2] X. Zhang, J. Benavente, and R. Garcia-Valls. Lignin-based membranes for electrolyte transference. Journal of Power Sources, 145:292–297, 2005. [3] I. Tho, A.-L. Kjøniksen, K. D. Knudsen, and B. Nystr¨ m. Effect of solvent composition on the o association behavior of pectin in methanol-water mixtures. European Polymer Journal, 42(5):1164 – 1172, 2006.

111

Sub-Terahertz excitations of Charge Density Waves in manganites
A. Nucara, P. Maselli, R. Sopracase, P. Calvani
CNR-INFM Coherentia and Dip. di Fisica, Universit` di Roma La Sapienza, Roma, Italy a

M. Ortolani, U. Schade
Berliner Elektronenspeicherring-Gesellshaft f¨r Synchrotronstrahlung m.b.H., Berlin, Germany u The optical conductivity of La-Ca and Nd-Sr manganites with commensurate charge order (CO) has been first studied in the sub-THz region, by use of the Coherent Synchrotron Radiation of BESSY. Below the ordering temperature TCO of each material, well-defined peaks are observed, associated with side bands. They have been assigned to collective excitations of the Charge Density Waves, namely, pinned phasons and combinations of phasons and amplitudons.
PACS numbers:

Several manganites, like La1−x Cax MnO3 (LCMO) with x ≥ 1/2, exhibit charge ordering (CO) phenomena below a transition temperature TCO [1]. Recent results point toward [2] a weak-coupling, or Charge Density Wave (CDW) [3, 4], approach to the CO phenomena in these materials, even if the collective excitations typical of the CDW have never been observed up to now. One of the CDW modes predicted by the theory is the phason, an infrared-active, acoustic-like, excitation with ωp (k = 0) = Ωp = 0. Either if the CDW is pinned to lattice impurities, or if it is commensurate with the lattice, one has Ωp > 0 and the phason can be observed in the very far infrared. The second collective mode is the Raman-active, optically-dispersed amplitudon at ωa (k = 0) = Ωa . CDW excitations at finite energies have been indeed detected in the subterahertz range (1 THz = 33 cm−1 ) in one-dimensional metals like the ”blue bronze” family[5] K0.3 Mo1−x Wx O3 . The manganite family offers the opportunity to study the excitation spectrum of a multi-dimensional CDW, provided that one can reach sub-THz frequencies with the desired signalto-noise ratio. We have obtained the above result by using the Coherent Synchrotron Radiation (CSR) emitted by the storage ring BESSY when working in the ”low-α” mode. We have thus studied, down to frequencies ωmin which range from 4.5 to 10 cm−1 , the optical conductivity of four manganites. One of them is a Nd1/2 Sr1/2 MnO3 (NSMO) single crystal, grown by the floating zone method, and here measured in the ab plane. Its TCO is 150 K. A second one is a single crystal of Bi0.5 Sr0.5 MnO3 , also grown by the floating zone method [6], and here measured both in the ab plane and along the c axis. The remaining three samples are polycrystalline pellets of La1−n/8 Can/8 MnO3 with n = 5,6,7, and with TCO = 270, 230, and 150 K, respectively [7]. The reflectivity R(ω) of all samples, 2-mm thick, was measured at nearly normal incidence after accurate polishing with sub-micron-thick powders. Conventional radiation sources were used between 20 and 6000 cm−1 , the CSR of the BESSY storage ring [10, 11] from ωmin to 30 cm−1 . R(ω) was extrapolated to ω = 0 by accurate Drude-Lorentz fits and σ(ω)

600
x 0.6 T = 160 K 300 140 70 10

400 Nd1/2Sr1/2MnO3 200

a
0 300 200
T = 300 K 250 200 150 100 50 10

La3/8Ca5/8MnO3

σ(ω) (Ω−1cm )
-1

100 0 400

b

200

La1/4Ca3/4MnO3

c
0 400
T = 300 K 250 200 150 100 50 10

200

La1/8Ca7/8MnO3

d
0 10 100 1000
-1

ω (cm )
FIG. 1: Optical conductivity at different temperatures of the single crystal of Nd1/2 Sr1/2 MnO3 and of the polycrystalline pellets of La1−n/8 Can/8 MnO3 .

was obtained by standard Kramers-Kronig transformations. The optical conductivity of the four samples is shown in Fig. 1. At the highest T ’s it exhibits a weak Drude term, while below TCO an optical gap 2∆ opens in the spectra of all samples. By smooth extrapolations of the lowest-T curves to σ = 0, one finds 2∆(T ≃ 0) ∼ 800 cm−1 or 0.1 eV in the polycrystalline samples of LCMO,

112

2
150
500

amplitudon combination band ωp + ωa , respectively, basing on several arguments [12].
ε1
La1/4Ca3/4MnO3 La3/8Ca5/8MnO3

σ(ω) (Ω−1cm )

-1

100
0

10

100

50

ω (cm )
-1

Both the pinned-phason mode and the amplitudon have a high density of states close to ka = 0, where their dispersion is nearly flat. If therefore ωp + ωa is peaked at Ωp +Ωa , and δp+a ≃ δ2p , one obtains for Ωa 40 (30) cm−1 for x = 5/8 (3/4) which provides and electron-phonon interaction strength λ = 0.8 (0.7) . These moderately low values show that the assumptions of the CDW model are thus justified a posteriori. By using further relations which link the frequencies and the CO gap here measured to the charge dynamics, one can obtain [12] the effective mass of the CDW. It turns out to be m∗ /mb ≃ 400 (700) for x = 5/8 (3/4) at 10 K, to be compared with m∗ /mb ≃ 800 reported for the one-dimensional CDW of K0.3 MoO3 [5]. Also the CDW contribution to the dielectric constant ǫCDW can 0 be determined. As here ωmin is lower than any CDW absorption, the experimental ǫ1 (ωmin ) at low T measures ǫCDW , after one subtracts the phonon contributions and 0 the high-frequency term ǫ∞ . From the inset of Fig. 2 one obtains ǫCDW = 100 (150) at x = 5/8 (3/4). These 0 results can be compared with the theoretical prediction of the CDW theory [4], which in the present case[12] gives ǫCDW = 90 (120) at x = 5/8 (3/4). Therefore, also under 0,calc this respect the CDW theory is in very good agreement with the experiment. In conclusion, we have first explored in the sub-THz range the optical conductivity of four manganites with TCO spanning from 130 to more than 500 K, by using a coherent, sub-Terahertz radiation source. In all samples with TCO < 300 K we have found sharp peaks which disappear above TCO and are similar to those reported for one-dimensional CDW’s. They have been assigned to pinned phasons, followed by broad combinations of phasons and amplitudons. In two La-Ca manganites, a detailed analysis based on the CDW theory has allowed us to determine the electron-phonon coupling, the CDW effective mass, and the CO-phase contribution to the dielectric constant. These parameters are consistent with a description of the charge order in La-Ca manganites in terms of charge density waves, even at commensurate doping. We acknowledge the EU support through the BESSY IA-SFS program RII 3CT-2004-506008.

La1/4Ca3/4MnO3 0 10 20 30 40
-1

50

60

ω (cm )
FIG. 2: Low-energy conductivity of La1/4 Ca3/4 MnO3 at 10 K (solid line) and 100 K (dots) with the corresponding best fits (thin-dotted lines). The side band is the sum of the overtone 2Ωp and of the combination band Ωp + Ωa , as shown at 10 K by crosses and open circles, respectively. The inset shows the CDW contribution to the real part of the dielectric function ǫ1 (ω) in two samples at 10 K.

∼ 0.2 eV in NSMO. Correspondingly, sharp conductivity peaks appear at the lowest frequencies in all panels of Fig. 1. We assign the peaks in the sub-THz range to phasons in a ”pinned state”, consistently with the presence in our samples of both impurities and commensurability. Their frequencies Ωp are all larger than in [5] K0.3 MoO3 (3.3 cm−1 ) and, in the three samples with CE-type charge order, are found to scale with TCO . In the La-Ca manganites with n = 5,6 a broad side band is also detected below 100 cm−1 . It shows the same T -dependence as the main peak and disappears into the Drude continuum above TCO . Fig. 2 shows in further detail both sub-THz bands of La1/4 Ca3/4 MnO3 , indicating that the side band is in fact the sum of two contributions. One of them broadens with T like the phason peak at 7.5 cm−1 , while the other one is nearly T -independent. A similar fit was obtained for La3/8 Ca5/8 MnO3 . We assign those two features to the overtone 2ωp and to a phason-

[1] C. H. Chen and S.-W. Cheong, Phys. Rev. Lett. 76 , 4042 (1996). [2] G. C. Milward, M. J. Calder´n, and P. B. Littlewood, o Nature (London) 433 , 607 (2005). [3] P. A. Lee, T. M. Rice, and P. W. Anderson, Solid State Commun. 14 , 703 (1974). [4] G. Gr¨ner, Rev. Mod. Phys. 60 , 1129 (1988). u [5] L. Degiorgi and G. Gr¨ner, Phys. Rev. B 44 , 7820 (1991). u [6] C. Frontera, J. L. Garc´ ıa-Munoz, A. Llobet, M. A. G.

[7] [8] [9] [10] [11] [12]

Aranda, C. Ritter, M. Respaud, and J Vanacken, J. Phys.: Cond. Matt. 13, 1071 (2001). M. Pissas and G. Kallias, Phys. Rev. B 68 ,134414 (2003). A. Douy and P. Odier, Mat. Res. Bull. 24 , 1119 (1989). K. Tobe, T. Kimura, and Y. Tokura, Phys. Rev. B 69 , 014407 (2004). M. Abo-Bakr et al., Phys. Rev. Lett. 90 , 094801 (2003). M. Ortolani et al., Phys. Rev. Lett. 97 , 097002 (2006). A. Nucara et al., Phys. Rev. Lett. 101, 066407 (2008).

113

Characterisation of Bone Mineral Density around Modified Titanium Implants in an Osteoporotic Rat Model with Synchrotron Microcomputed Tomography (SRµCT)
R. Bernhardt, J. Dudeck, D. Scharnweber, St. Rammelt , J. Goebbels , G. Vollmer , H. Worch
1Max Bergmann Center for Biomaterials, Technische Universität Dresden, Germany of Trauma and Reconstructive Surgery, University Hospital “Carl Gustav Carus” Dresden, Germany , 2Federal Institute for Materials Research and Testing (BAM), Berlin, Germany 3Institute for Zoology, Technische Universität Dresden, Germany

1

2

3

1Department

Introduction Synchrotron microcomputed tomography (SRµCT) has been applied successfully to analyse quantitatively the bone formation around biofunctionalised titanium implants resulting in a higher bone volume content in relation to uncoated implants in a goat model [1]. As the mineral content of the newly formed bone indicates the stadium of the remodelling process, a three dimensional investigation of this parameter is a promising tool to understand how biofunctionalised implant surfaces influence the local bone mineralisation in their surrounding, especially if the overall bone quality is week or influenced by bony diseases. Goal of the present work was to describe the mineral structure of bone for selected areas of explants in an osteoporotic rat model including a hormone therapy. Based on SRµCT data, the local distribution of bone mineralisation should be determined three dimensionally as a function of the conditions ‚healthy‘, ‚osteoporotic‘ and ‘hormone treated osteoporotic‘. New knowledge is expected for the osseointegration of biofunctionalised titanium implants for normal as well as for disordered bone regeneration. Materials and Methods As the animal model, ovariectomised rats with and without hormone treatments were used for the analytical investigations. Titanium wires with a diameter of 0.8 mm with biofunctional coatings of Chondroitin Sulphate (CS), Bone Morphogenetic Protein 4 (hBMP 4) and a blank control (cp-Ti) were placed in the tibia of ovariectomised rats for 4 weeks. After the animal experiments, the rat tibiae were freed from adherent soft tissue, fixed in paraformaldehyde and dehydrated in ethanol in a graded series of increasing concentrations. The embedding was performed in Polymethyl-metacrylate (PMMA). At the BAMline (BESSY II) 27 specimens, with 3 samples for each condition, were evaluated with Synchrotron microcomputed tomography (SRµCT). For each sample 720 X-ray attenuation projections with a local resolution of 9 µm were acquired using a monochromatic X-ray energy of 30 keV. Results and Discussion The SRµCT reconstruction of the explants shows a detailed visualisation of bone formation around the titanium implants and directly on the implant surface (Fig. 1). Assigning different colours to the x-ray absorption values (grey levels), local areas of higher and lower mineral content could be visualised for the bony tissue. On the implant surface a network of newly formed bone with different thickness and mineralisation could be observed. This findings indicate a specific reaction of the organism to the functionalised titanium wires. The analysis of relations between different implant surfaces, bone conditions and/or hormone therapy to the characteristics of the three-dimensional network around the implants is still under investigation. A suitable imaging procedure for a volumetric analysis related to this specific model was developed (Fig. 2).

114

Fig. 1: SRµCT visualisation of different mineralised bone on a biofunctionalised titanium wire (white) after a healing time of 4 weeks in a rat tibia.

Fig. 2: Adapted analysis procedure to obtain 3D-information of newly formed bone from SRµCT-measurements around cylindrical implants.

Acknowledgments
The authors gratefully acknowledge the support of this work by the DFG (SCHA 570/9-1), BESSY and the BMBF (05 ES3XBA/5).

References: [1] Bernhardt et al., Biomaterials 26, 3009 (2005)

115

Vacuum Ultra-Violet Spectroscopic Ellipsometry of DNA Layers Covalently Attached to Diamond and Silicon
Sylvia Wenmackers, Patrick Wagner, Hasselt University, IMO, Wetenschapspark 1, 3590 Diepenbeek, Belgium Veronique Vermeeren, Luc Michiels, Hasselt University, BIOMED, Agoralaan Building C, 3590 Diepenbeek, Belgium Simona D. Pop, Christoph Werner, Christoph Cobet, and Norbert Esser. Institute forAnalytical Sciences-ISAS, Albert-Einstein-straße 9, 12489 Berlin, Germany

Our group has developed a prototype DNA sensor on diamond. The DNA probes are covalently attached to diamond surfaces using an effective two-step protocol [1], and the location of the DNA probes can be predetermined by photopatterning the linker layer. This platform is sensitive for point mutations, as was demonstrated using fluorimetric read-out [1] as well as real-time, label-free impedance spectroscopy [2]. In the framework of this project, we are interested in the molecular organisation of DNA probe layers (brushes), both in single and double stranded form. One technique to study DNA brushes on diamond is “nano-shaving” with an atomic force microscope [3]. To study the orientation of immobilised, short DNA strands (8-36 bases) in a non-destructive way, we apply vacuum UV spectroscopic ellipsometry (SE). We demonstrated that this technique can indeed be used to detect (sub-)monolayers of short DNA strands on ultra-nanocrystalline diamond (UNCD) surfaces. Based on the 4.74 eV transition due to the π-π* transition dipole moments of the DNA bases, the orientation of the DNA strands could be calculated: we found average tilt angles ranging from 45° to 52° [4]. Since UNCD has an root-mean-square surface roughness of 17 nm, the interpretation of the obtained tilt angles is not straightforward for biological layers thinner than that. Hence, we performed additional ‘mapping SE’ experiments on UNCD samples, to qualify the lateral spread on the average DNA tilt angles. This made clear that the roughness of the DNA layer can vary from point to point: in Figure 1 and Table 1, three spots have been compared of the same UNCD sample functionalised with double stranded DNA of 29 base pairs. We also collected spectra on atomically flat surfaces to examine tilt angles caused only by the intermolecular interactions of DNA in dense layers, not by the topography of the underlying surface. At first we opted for DNA-modified single crystalline diamond (SCD). Typically, SCD samples have a surface area limited to (2 mm)². Taking spectra on such small areas is far from trivial. Therefore, we examined additional DNA-layers covalently coupled to single crystalline Si: these substrates are also atomically flat, but larger in surface area, e.g. (1 cm)². The reference spectra collected on clean SCD and Si samples correspond well with UV spectra found in literature. The additional features in the spectra of linker- and DNA-modified SCD and Si samples are similar to those previously recorded on UNCD. Further calculations based on the new data are ongoing, to reveal possible variations in the average orientation of the DNA strands between the previously analysed UNCD samples on the one hand, and the SCD and Si samples on the other, as well as between different spots on the same SCD or Si sample.

116

Figure 1: Vacuum UV ellipsometry on UNCD with double-stranded DNA of 29 base pairs. Measured (open symbols) and fitted (lines) spectra of three spots on the same sample. Table 1: Layer thickness and roughness of the DNA layer (29 base pairs) of the three spots considered in Figure 1, calculated from the fitted spectra.

Acknowledgements This work was supported by the European Community - Research Infrastructure Action under the Sixth Framework Programme (FP6) "Structuring the European Research Area" Programme (through the Integrated Infrastructure Initiative "Integrating Activity on Synchroton and Free Electron Laser Science - Contract RII 3-CT-2004-506008") and through the "NanoCharm" project under the Seventh Framework Programme (FP7). S.W. is a postoctoral research fellow of the Research Foundation - Flanders (FWOVlaanderen) and V.V. is a IWT postoctoral fellow. References [1] Vermeeren, V.; Wenmackers, S.; Daenen, M.; Haenen, K.; Williams, O. A.; Ameloot, M.; vandeVen, M.; Wagner, P.; L. Michiels, L. Langmuir 2008, 24, 9125. [2] Vermeeren, V.; Bijnens, N.; Wenmackers, S.; Daenen, M.; Haenen, K.; Williams, O. A.; Ameloot, M.; vandeVen, M.; Wagner, P.; L. Michiels, L. Langmuir 2007, 23, 13193. [3] Rezek, B.; Shin, D.; Uetsuka, H.; Nebel, C. E. physica status solidi (a) 2007, 204, 2888. [4] Wenmackers, S.; Pop, S. D.; Roodenko, K.; Vermeeren, V.; Williams, O. A.; Daenen, M.; Douhéret, O.; D’Haen, J.; Hardy, A.; Van Bael, M. K.; Hinrichs, K.; Cobet, C.; vandeVen, M.; Ameloot, M.; Haenen, K.; Michiels, L.; Esser, N.; Wagner, P. Langmuir 2008, 24, 7269.

117

BESSY II ANNUAL REPORT 2008

Near-Surface Stress Gradient in Shot Peened Ti-Alloys Measured by Energy-Dispersive Synchrotron Radiation
Proposer: Emad Maawad Institute of materials science and engineering, TU Clausthal, Germany. Co-Proposer: H.-G. Brokmeier Institute of materials science and engineering, TU Clausthal, Germany. GKSS resarch center, Geesthacht, Germany. Experimental Team: M. Klaus, C. Genzel, J. Gibmeier BESSY, 12489 Berlin, Germany.

Objective
Mechanical surface treatments such as shot peening (SP) or ball-burnishing lead to changes in the nearsurface material states. This is mainly the result of induced plastic deformation which results in workhardening and the generation of residual stresses. Residual compressive stresses are well known to enhance the fatigue performance and corrosion resistance of a number of metallic materials by retarding or even suppressing micro-crack growth from the surface into the interior. The shot peening-induced residual stresses in shot peened titanium alloys Ti-2.5Cu and Timetal LCB were evaluated by applying energy-dispersive (ED) diffraction using synchrotron radiation because of a higher penetration depth compared to conventional X-ray diffraction [1,2]. The main objective of this study is to measure the compressive residual stresses in deeper region below the surface of aforementioned alloys.

Experiment
Residual stress measurements were performed by applying hard X-rays using synchrotron radiation. Stresses are evaluated by means of the sin2ψ method. Energy dispersive X-ray diffraction technique using synchrotron radiation allows the non-destructive measurement of residual stress depth-profile. This technique uses white X-ray beam with various energy ranges, 20 ~ 120 keV in general. Table 1 shows the required parameters of hard x-ray diffractometer. Table 1. Diffractometer parameters Beam size Detector slit Beam energy 2-Theta Phi Psi Exposure time 0.50 mm x 0.50 mm 0.03 (H) x 8.00 mm (V) Up to 120 keV 8° (for Ti-2.5Cu) and 6° (for LCB) 0o 0o to 80° (~21 steps) 300 sec.

The residual stress depth profiles were measured in shot peened samples, Ti-2.5C and Timetal LCB. Peening was done to full coverage at Almen intensities of 0.11 and 0.20mmA. Blanks (20x20x5-10mm3) were cut. Some layers (50, 100, 150, 200 µm) were removed by using electro-polishing to reach the stress-free zone.
118

Achievements and Main Results
Fig. 1 shows examples of the relation between lattice spacing (d) and (sin2Ψ) for Ti-2.5Cu. No steep residual stress gradient was found.

Fig. 1 The relation between lattice spacing (d) and (sin2Ψ) for Ti-2.5Cu Fig. 2 and Fig. 3 show the residual stresses profiles σ(τ) of shot peened Ti-2.5Cu at Almen intensities of 0.11 and 0.20mmA, respectively. It is observed that the residual stress-depth profile obtained from the
synchrotron radiation is exponentially damped transform of the actual or real space stress depth. This is explained by the 'modified multi wavelength method' used in the synchrotron radiation measurements yields the residual stress depth profiles in the Laplace space, i.e. sigma(τ). In addition, The tensile residual stress is located deeper than

300 µm.
RS = -473.846-4.272T+0.007T -5.008(10 )T +1.603(10 )T -2.019(10 )T
-200 -250 -300 -300
2 -4 3 -6 4 -9 5

RS = -567.102-3.292T+0.058T -4.6388(10 )T +1.851(10 )T -2.645(10 )T
-200

2

-4

3

-6

4

-9

5

Residual Stress "RS" (MPa)

-350 -400 -450 -500 -550 -600 -650 0 30 60 90 120 150 180 210 240 270 300
0-50 micron 50-100 micron 100-150 micron 150-200 micron 200-250 micron Polynomial Fit of Concatenate

Residual Stress "RS" (MPa)

-400
0-50 micron 50-100 micron 100-150 micron 150-200 micron 200-250 micron Polynomial Fit of Concatenate

-500

-600

-700

0

30

60

90

120

150

180

210

240

270

300

Depth "T" (micron)

Depth "T" (micron)

Fig. 2 Residual stress profile of SP (0.11 mmA) Ti-2.5Cu

Fig. 3 Residual stress profile of SP (0.20 mmA) Ti-2.5Cu

119

Fig. 4 shows examples of the relation between lattice spacing (d) and (sin2Ψ) for α-phase and β-phase of TIMETAL LCB.
Alpha Beta

Fig. 4 The relation between lattice spacing (d) and (sin2Ψ) for Timetal LCB Fig. 5 and Fig. 6 show the residual stresses profiles of shot peened (0.11 mmA) α and β in Timetal LCB, respectively. It is observed that the free stress point is 270 (approx.) µm in depth for β-phase.
RS = -411-10.363T+0.202T -0.001T +2.913(10 )T -2.615(10 )T
0 -100 -200
2 3 -6 4 -9 5

RS = -510.340-37.820T+0.670T -0.005T +1.470(10 )T -1.657(10 )T
100 0 -100

2

3

-5

4

-8

5

Residual Stress "RS" (MPa)

Residual Stress (MPa)

-200 -300 -400 -500 -600 -700 -800 -900
0-50 micron 50-100 micron 100-150 micron 150-200 micron 200-250 micron Polynomial Fit of Concatenate

-300 -400 -500 -600 -700 -800 -900 0 30 60 90 120 150 180 210 240 270

0-50 micron 50-100 micron 100-150 micron 150-200 micron 200-250 micron Polynomial Fit of Concatenate

300

330

0

30

60

90

120

150

180

210

240

270

300

330

Depth (micron)

Depth "T" (micron)

Fig. 5 Residual stress profile of SP (0.11 mmA) α-phase of LCB

Fig. 6 Residual stress profile of SP (0.11 mmA) β-phase of LCB

The difference of residual stresses between α and β phases is explained by the difference in the yield strength. The β Ti-alloys have in general larger yield strength comparing to that of α alloys, so that the potential maxima of the induced residual stress by plastic deformation is larger in the β phase. Fig. 7 and Fig. 8 show the residual stresses profiles of shot peened (0.20 mmA) α and β in Timetal LCB, respectively. Also, it is observed that the residual stresses induced in α and β phases are different.

120

RS = -405.840-8.122T+0.066T -8.483(10 )T -4.168(10 )T+1.01(10 )T
-100 -200 -300

2

-5

3

-7

-9

5

RS = -400-47.420T+0.739T -0.005T +1.501(10 )T -1.707(10 )T
-100 -200 -300

2

3

-5

4

-8

5

Residual Stress (MPa)

Residual Stress (MPa)

-400 -500 -600 -700 -800 -900 -1000 -1100 0 25 50 75 100 125 150 175 200 225 250 275 300 325
0-50 micron 50-100 micron 100-150 micron 150-200 micron 200-250 micron Polynomial Fit of Concatenate

-400 -500 -600 -700 -800 -900 -1000 -1100 0 25 50 75 100 125 150 175 200 225 250 275 300 325
0-50 micron 50-100 micron 100-150 micron 150-200 micron 200-250 micron Polynomial Fit of Concatenate

Depth (micron)

Depth (micron)

Fig. 7 Residual stress profile of SP (0.20 mmA) α-phase of LCB

Fig. 8 Residual stress profile of SP (0.20 mmA) β-phase of LCB

A difficult problem concerns the neumerical inverse Laplace transformatiom (NILT), which has to be applied in order to obtain the residual stress profiles σ(z) in the real z-space from the experimentally determined Laplace space stressses σ(τ) [1].

Acknowledgements
I am indebted to the German Research Foundation (DFG) for financial support of the investigations, project number (BR961/5-1). I’d like to thank Prof. C. Genzel and M. Klaus for thier help and suggestion.

References
[1]Ch. Genzel, C. Stock, W. Reimers, Materials Science and Engineering A 372 (2004) 28-43. [2]Ch. Genzel,C. Stock, B. Wallis, W. Reimers, Nuclear Instruments and Methods in Physics Research A 467-468 (2001) 1253-1256.

121

In situ observation of liquid water evolution and transport in PEM fuel cells Ph. Krüger1, R. Kuhn1, H. Riesemeier3, J. Banhart2, I. Manke2, Ch. Hartnig1
1

Centre for Solar Energy and Hydrogen Research (ZSW) Helmholtz Centre Berlin for Materials and Energy (HZB)

2 3

Federal Institute for Materials Research and Testing (BAM)

Abstract: Liquid water transport has been visualized in situ with a so far unique spatial and temporal resolution. For the synchrotron studies two different viewing directions have been chosen in order to get a pseudo-3D insight in processes on a microscopic scale. The water content of the different components in a fuel cell and diffusion barriers for the reactant gases have been determined. The different components show a sufficiently different absorption coefficient to distinguish them. Liquid water appears at first close to the catalyst layer; the adjacent micro-porous layer is free of liquid water. In the porous gas diffusion media, water condenses and is from there transported in an eruptive manner to the channel of the flow field. Component providers and fuel cell developers will benefit from these insights for optimization and design purposes.

Detailed report: Water management is the key problem in state-of-the-art hydrogen driven fuel cells. Considerable efforts have already been reached by new designs of flow field structure which allow a well-balanced distribution of reactants and help to avoid the formation of large amounts of liquid water. Still the performance suffers from excess water in the gas diffusion layer or drying-out phenomena, e.g. at dynamic operating conditions. These problems can be addressed by, e.g., fine tuning of diffusion materials. A fundamental understanding of water distribution and transport in gas diffusion media features the key pathway to achieve the target of well distributed water content.

Figure 1: Insight in fuel cells from two different directions.

122

The observation of liquid water evolution and transport in PEM fuel cell under operating conditions is realized on a microscopic level with a resolution down to 3 µm [1, 2]. By means of synchrotron X-ray radiography the initial formation of small water clusters in the gas diffusion layer and the transport from the diffusion media to the channel of the flow field are detected. The employed fuel cell setup allows for an unperturbed and unbiased insight to the evolution and transport mechanisms of liquid water under operating conditions. In figure 1, two different viewing directions have been chosen, the through-plane view and the cross-sectional view. Through-plane imaging allows for a differentiation between areas under the land and the channel of the flow field (Fig. 2). Water evolution starts under the land of the flow field; from a certain pressure onwards the water clusters are transported in an eruptive mechanism from the GDL to the gas channel.

Figure 2: Though plane view in an operation fuel cells. On the right side liquid water clusters under the ribs can be distinguished.

The cross-sectional viewing direction allows a separate investigation of the different components: Gas diffusion layer (GDL), micro porous layer (MPL) and membrane electrode assembly (MEA). It helps to clarify transport mechanisms from the source of water evolution to the gas channel (Fig. 3).

Figure 3: Cross sectional view on liquid water formations in an operating PEM fuel cell.

The dynamic formation of water cluster at changed operating conditions is monitored and the increase of water content in the gas diffusion media can be described. Based on the experimental findings previous results from modeling and simulation approaches are confirmed, which consider the area under the ribs as the primary source of liquid water. A recently proposed eruptive mechanism for the water transport from the gas diffusion layer to the channel, which was observed in ex situ experiments, is now supported by the presented in situ results.

123

NEXAFS characterization of electronic structure for CuI@SWCNT nanocomposite M.M. Brzhezinskaya1,2, A.V.Generalov2, R. Püttner3, A.S. Vinogradov2, M.V. Chernysheva4, A.A. Eliseev4, N.A. Kiselev5, A.V. Lukashin4, and Yu.D. Tretyakov4
2

HZB (BESSY GmbH), Albert-Einstein-Str. 15, 12489 Berlin, Germany V.A. Fock Institute of Physics, St. Petersburg State University, St. Petersburg, Russian Federation 3 Institut für Experimentalphysik, Freie Universität Berlin, 14195 Berlin, Germany 4 Department of Material Science, Moscow State University, Moscow 142432, Russian Federation 5 A.V. Shubnikov Institute of Crystallography of RAS, Moscow 119333, Russian Federation

1

Encapsulated single-walled carbon nanotubes (SWCNTs) with inner channels filled by different compounds present the new class of composite materials. Such CNTs are of a big interest because of an opportunity to form 1D nanocrystals as well as quantum nanowires with new physical and chemical properties inside the tubes. Electronic properties of modified CNTs determine substantially the field of their application. As a consequence different spectroscopic methods are important for studying the electronic structure of composites. The NEXAFS [1] spectroscopy probing the partial density of empty electron states that are localized near the absorbing atom is one of the powerful methods for investigating the local atomic and electronic structure as well as the chemical bonding in various nanosystems. Therefore, the present study is aimed to characterize the possible chemical interaction between CuI and SWCNTs in CuI@SWCNTs and electronic structure of the latter. All measurements have been performed at the Russian-German beamline (RGBL) using experimental station Mustang. The CuI@SWCNT nanocomposite was produced by the filling of metallic single-walled carbon nanotubes with inner diameter of 1.1-1.4 nm by wide-gap semiconducting CuI nanocrystals using so-called capillary technique [2]. Approximate loading value of CuI equals 72.6 wt. % (i.e. one CuI molecule corresponds to about 6 C atoms.). CuI@SWCNT sample was prepared in air by rubbing powder of it into the scratched surface of stainless steel plate. Evaporated layer of CuI, powder of CuO and initial SWCNTs were used as reference samples. Thin (20-25 nm) CuI layers were prepared in situ in the preparation chamber by thermal evaporation of thoroughly dehydrated CuI powder (Alfa Aesar) from a water-cooled effusion cell onto a polished stainless-steel plate in a vacuum of ~3×10-7 mbar. Powders of CuO and SWCNTs were rubbing into the scratched substrates. NEXAFS spectra at the Cu 2p and C 1s edges were obtained in the total electron yield mode by detecting a sample current. All spectra were normalized to the incident photon flux, which was monitored by recording the photocurrent from a gold mesh placed at the outlet of the beamline. The photon energy at Cu 2p and C 1s edges was calibrated using Au 4f photoemission lines (Ebind(Au 4f7/2)=83.9 eV and Ebind(Au 4f5/2)=87.6 eV) from gold plate fastened on the same holder with sample under study. The accuracy of this procedure is evaluated to be of 0.1-0.2 eV. The total energy resolution of monochromator at C 1s and Cu 2p edges was about 100 meV and 300 meV respectively. Measurements of absorption and photoemission spectra were carried out at a pressure in the measuring chamber ~ 2·10-10 mbar. Cu 2p spectra for CuI@SWCNTs, and reference samples CuI and CuO are presented in Fig. 1. All spectra were normalized to the absorption edge jump at hν=966 eV. Structures corresponding to the electronic transitions from Cu 2p1/2 and Cu 2p3/2 core levels to the same empty electron states are denoted by the primed and unprimed letters, respectively. If the dipole rules for 2p electron transitions are taken into account, so these spectra reflect local density of empty electron states that have mainly Cu 4s- and Cu 3d-character. It can be seen from the Fig.1 that the Cu 2p3/2 NEXAFS spectra of CuI@SWCNTs and CuI show appreciable differences between their absorption structures. The main distinctions are an appearance of narrow (FWHM=1.1 eV) low-energy peak A, a broadening of all and merging (с1 and с2, e1 and e2) as well as disappearing (d, f) of some absorption structures in the CuI@SWCNT spectrum. Moreover, all structures in the spectrum of the composite have low-energy shift of order of
124

0.3 eV in comparison with their counterparts in the spectrum of CuI. It is natural to relate these distinctions to changes in electronic structure of CuI owing to its encapsulation into SWCNTs or to weak chemical interaction between the filler and the CNT in CuI@SWCNT. Clearly, these changes are accompanied by an A electron density transfer between copper, iodine and carbon atoms in the composite. With a' x10 consideration for the valence electron configuration of the Cu+ ion (3d104s0) in CuI, the appearance of narrow low-energy peak A in the a A' Cu 2p3/2 spectrum of the composite is reasonCuO able to associate with Cu 2p3/2 electron transitions to empty 3d electron states that are lackb' ing in pristine CuI and appear in b A' CuI@SWCNTs as a result of changes in elecc e tronic structure of CuI. This is accompanied A a CuI@SWCNT by the change in the valence electron configuration of the Cu+ ion from 3d104s0 for CuI to b' g b 3d10-x for CuI@SWCNTs. The direct comparic1c2 d e1e2 f son of Cu 2p3/2 spectra for the composite and a CuO confirms this interpretation. It is easy to CuI_evaporated see (Fig.1) that the spectrum for CuO with the valence electron configuration of Cu2+ ion, 3d9, has the similar narrow (FWHM=1.0 eV) low-energy peak A as well. It is well known 930 940 950 960 that this peak is due to the Cu 2p3/2 - 3d elecPhoton Energy (eV) tron transition [3]. The energy of the latter in CuI@SWCNTs (931.25 eV) is in good Fig. 1. Cu 2p NEXAFS spectra for agreement with the one in CuO (931.3 eV) CuI@SWCNTs, CuI and CuO. [3]. Comparison of relative intensities of peaks A from the spectra for CuO and CuI@SWCNTs, that were normalized equally, gives the magnitude of x= 0.06 e for the decrease of the Cu 3d charge in the composite. Right now, there is no way to unambiguously indicate a direction of the electron transfer from the copper atoms to the iodine or carbon atoms. As for the other differences between the spectra of CuI@SWCNTs and CuI, they are most likely due to the b c changes in crystal and electronic e d a f CuI@SWCNT structure of CuI owing to its encapsulation into SWCNTs. The general resemblance of absorpb tion structures for spectra of CuI c a and CuI@SWCNTs, except for d e f the peak A, is probably evidence A SWCNT for conservation of a tetrahedron coordination of absorbing Cu ion as well as relatively small magnitude of effects under consideration. In this case, the broadening and smearing of some absorption structures can be related 290 300 310 to a distortion of the tetrahedron Photon Energy (eV) environment of Cu ions by ioFig. 2. C 1s NEXAFS spectra for CuI@SWCNTs and dine atoms. SWCNTs.
Absorption, Total Electron Yield (arb. units)
125

Absorption, Total Electron Yield (arb. units)

C 1s spectra for nanocomposite CuI@SWCNTs and pristine SWCNTs normalized to the absorption edge jump at hν=317 eV are shown in Fig. 2. The main structures in the spectrum of the pristine SWCNTs are narrow (FWHM = 1.3 eV) resonance a and bands b, c that are caused by C 1s electron transitions to empty electron states of π and σ symmetry with C 2pzπ and C 2px,yσ character. These states are very similar to the corresponding states of graphite and strongly localized perpendicular and parallel to the carbon hexagon [4]. From Fig.2 it is seen that C 1s absorption spectrum for CuI@SWCNTs is practically identical to the one for pristine SWCNTs except for an additional small shoulder-like peak A at the low-energy side of the πresonance a. Besides, it is observed the decrease of total spectral weight of πstates of about 12% and the narrowing of π-resonance a of about 0.1 eV for the composite spectrum relative to the one of pristine SWCNTs. It is obvious that such a new peak reflects the interaction between valence electrons Fig. 3. C 1s and Cu 2p3/2 NEXAFS spectra for of the filler and π electron subsystem CuI@SWCNTs aligned in energy. of the nanotubes. Cu 2p3/2 and C 1s spectra for CuI@SWCNTs, which have been aligned in energy by using the binding energy separation of 648.0 eV between the Cu 2p3/2 and C 1s core levels, are represented in Fig. 3. It is easy to see that the positions of the peaks A in both spectra are nearly coincide within 1 eV. In the framework of quasimolecular approach [5], this coincidence of the absorption peaks A in both the energetically aligned spectra can be regarded as a result of the transitions of Cu 2p3/2 and C 1s electrons to the same empty state of CuI@SWCNTs which has hybridized Cu 3d-C2pz character. In this case, a slightly different position of peak A in C 1s and Cu 2p3/2 spectra can be associated with a different influence of C 1s-1 and Cu 2p3/2-1 holes on the coreexcited state under consideration. In conclusion, the present study has shown that encapsulation of CuI into SWCNTs is accompanied by the changes in electronic structure of CuI because of the ones in atomic structure of CuI in the composite and the chemical interaction between the filler and carbon nanotubes.
This work was supported by the Russian Foundation for Basic Research (projects no. 06-0216998, 09-02-01278 ) and the bilateral Program “Russian-German Laboratory at BESSY”. A.V.G. acknowledges the support from Freie Universität Berlin through the Leonhard-Euler Fellowship Program.
References J. Stöhr. NEXAFS Spectroscopy. Springer Verlag, Berlin. (1992). 403 p. M.V. Chernysheva, A.A. Eliseev, A.V. Lukashin, Yu.D. Tretyakov, et al. Physica E 37, 62 (2007). M. Grioni, J.B. Goedkoop, R. Schoorl, F. M. F. de Groot, et al. Phys. Rev. B 39, 1541 (1989). P.A. Brühwiler, A.J. Maxwell, C. Puglia, et al. Phys. Rev. Lett. 74, 614 (1995). 5. A.S. Vinogradov, S.I. Fedoseenko, S.A. Krasnikov, et al. Phys. Rev. B 71, 045127, (2005). 1. 2. 3. 4.

126

Synchrotron-based, depth-resolved analysis of elemental gradients in chalcopyrite solar cell absorbers using angle-dependent x-ray emission spectroscopy H. Mönig1, Ch.-H. Fischer1,2, R. Caballero2, C.A. Kaufmann2, A. Grimm2, B. Johnson2, T. Kropp2, M. Ch. Lux-Steiner1,2, C. Jung2, and I. Lauermann2
2

Freie Universität Berlin, Berlin, Germany, Helmholtz-Zentrum Berlin für Materialien und Energie, Berlin, Germany Synchrotron excited soft x-ray emission spectroscopy (XES) is a technique which provides bulk sensitive information1 with an information depth in the range of 0.1 to 1 μm, depending on the energy of exciting and emitted radiation and the material.. When used angle-resolved, we call this method AXES2. We quantify results from angle-resolved XES and use a simple layer model to extract depth profiles from these data. Here we examined Cu(InGa)Se2 solar cell absorbers prepared in the 3-stage process3 (Figure 1) by the technology department (SE3) of the HZB using different temperatures T2 during the deposition: T2 = 525°C (standard temperature) T2 = 425°C T2 = 330°C (suitable for polyimide substrates)

1

Figure 1: Absorber preparation by 3-stage physical vapour deposition (PVD)

We analysed these samples with AXES at the U 41 PGM beam line at BESSY II and recorded Cu L2,3 and Ga L2,3 spectra at different exit angles β to change the depth sensitivity. These spectra are shown in Figure 2. Peaks originating from Cu and Ga were then integrated and plotted vs. the exit angle. The results are shown in Figure 3 (right).

Cu L2,3

AXES hν = 1200 eV
all spectra normalized to max. of Ga L2,3

Ga L2,3

intensity (a.u.)

more surface sensitive β = 89° β = 85° β = 75° β = 65° β = 55° β = 45° 950 1000 1050 emission energy [eV] 1100 1150

Figure 2: XES spectra with Ga L2,3 and Cu L2,3 peaks, obtained at different exit angles β, normalised to the maximum of the Ga L2,3 peak. For evaluation, peaks are integrated and area ratios plotted vs. β.

127

sample depth x (nm) 6
Ga

0 A)

100

200

300

400

500

600

0

15

exit angle β (°) 30 45 60 75

90 A')

T 2= 525°C

Rel. at. conc. C /C

Cu

5 4 3 2
8 B) 6
Cu/Ga
T 2 = 425°C

1,08 T = 525°C 2 1,06
Cu/Ga

1,04 1,02 1,00 0,98 0,96
0,95
T 2 = 425°C

Ga

REI

Rel. at. conc. C /C

B')

Cu

0,90

4

REI

0,85

2 0,80 25
T 2 = 330°C

Ga

C)
5 4

0,95 T 2 = 330°C 0,90
Cu/Ga

C ')

Rel. at. conc. C /C

Cu

20 15 10 5

Zoom

2

0

100 200 x (nm )

300

REI

3

0,85 0,80 0,75

0

0

100

200

300

400

500

600

0

15

30

45

60 (°)

75

90

sam ple depth x (nm )

exit angle

Figure 3: Comparison of SNMS profiling with AXES results from 3 samples with different processing temperatures T2. Left: Experimental Cu/Ga-ratio from SNMS (dots) and two fitted depth profiles, solid line: best fit of SNMS data, dotted lines: best fit of AXES data. Right: Intensity ratios Cu/Ga from integrated AXES peaks (dots) and calculated based on literature values for x-ray absorption (solid and dotted lines as on the left). On the left of Figure 3 we show the comparison with sputtered neutral mass spectrometry SNMS data (done at the ZSW Stuttgart) from the same samples. This method yields absolute elemental concentrations4. The results can be summarised as follows: There is a clear influence of T2 on the Cu/Ga-distribution: lower T2 leads to Ga-depletion within the space charge layer. The AXES-results confirm the SNMS-profiles and allow some refinement of the elemental distribution up to a sampling depth of 500 nm. At extreme angles (high surface sensitivity), the AXES results are not consistent with a Cu/Ga gradient in the first 150 nm in samples A) and B) as suggested by SNMS. AXES results point towards a constant Cu/Ga-ratio in that region. These results prove the value of AXES for the analysis of thin films up to several hundred nm. References 1. C. Heske et. al., Appl. Phys. Lett. 74, 1451 (1999); Appl. Phys. Lett. 75, 2082 (1999) 2. H. Mönig, A. Grimm, Ch. Jung, Ch.-H. Fischer, I. Lauermann, C. Camus, C. A. Kaufmann, T. Kropp and M. C. Lux-Steiner, Applied Surface Science 255, 2474-2477 (2008) 3. C.A. Kaufmann, A. Neisser, R. Klenk, R. Scheer, Thin Solid Films 480/481, 515 (2005) 4. K. Herz, A. Eicke, F. Kessler, R. Wächter, M. Powalla, Thin Solid Films 431/432, 392 (2003)

128

Soft X-Ray Channeling in Policapillary Structures at the Condition of Anomalous Dispersion Region of Si L – Edge Absorption M. I. Mazuritskiy, P.V. Makhno
Physics Department, Southern Federal University, Sorge, 5, Rostov-on-Don, Russia, 344090

Researches of the X-ray transmission through microcapillary structures aimed at the development of a new focusing devices in particular for long-wavelength X-ray radiation is a high-priority physics problem. Its solution would lead to the appearance of a long-awaited techniques, instruments and technologies for physics, materials science, biology, and medicine. The microchannel plate (MCP) systems have been applied to focus and collimate xrays and multiply reflection properties of the microchannel surfaces are extremely important to research particularly for ultra soft x-ray radiation. Great scientific and applied interest have grazing x-ray methods based on the analysis of the secondary response of a solid to the absorption/reflection of incident radiation. If the incident photons are capable to excite the atomic levels then X-ray fluorescence is observed together with elastic scattering if the grazing angle is less than critical one. It was shown [1,2] that under certain conditions x-ray fluorescence emitted inside microcapillaries due to the excitation of atomic levels can affect the spatial distribution of the radiation intensity. We have researched channeling of the X-ray fluorescence (secondary radiation) in the anomalous dispersion region of energy range near SiL – absorption edge. In fact transportation of X-ray can be in the mode regime [3] associated in particular with the surface channeling of photons inside microcapillary structures. The MCP samples were thin ( 0.5 mm) “perforated” (with 8 m - diameter channels) plates consisting of mainly silicate glass. Samples were electrically insulated from the metal surface of the holder. The scheme in Fig. 1 shows rays (1) directed almost perpendicularly to the plate surface and corresponding to the grazing incidence onto the microchannel walls. Radiation (3) passed through the microchannel plate (2) absorbed by the plate (4) and detected in the current mode. The angle of incidence was varied by rotating the sample around the x, y axes. The XANES spectra were excited by monochromatic radiation whose photon energies were varied in the vicinity of the corresponding absorption edges. The total intensity of the X-ray radiation passed through the microchannels of the sample was measured. XANES Si L2,3 spectra (see Fig. 2,3) for the various grazing incidence of radiation onto the channel walls have been detected at the outgoing from polycapillary structures. The spectra were obtained with a resolution of 0.1 eV on spectrometer MUSTANG, RGBL-PGM at the BESSY synchrotron center.

Fig. 1 Experimental scheme

The fine structure was observed for the angles of incidence of less than 8°, i.e., under the conditions of total external reflection. The critical angle corresponds to the experimental data [4] for glass in the vicinity of the L2,3 edge of silicon absorption (100–140 eV). The L1 edge of silicon absorption is at the photon energy range of about 160 eV. The structure of spectra 2,3 in Fig. 2 corresponds to the one obtained in [5,6] for a plane SiO2 surface. However for MCP samples radiation was incident into long and narrow channels, so the reflected rays could not directly exit outside. Fig. 3 shows the spectra obtained from the MCP preliminarily subjected to thermal hydrogen reduction. The fine structure of spectra 3–5

129





µ

≈ 

differs from that seen in Fig. 2. The modified glass surface has an oxide silicon state SiOx, where 11.5 GPa incorporated nearly the maximum amount of hydrogen allowed by the staurolite structure (6 H pfu) and was subsequently overgrown and marginally replaced during the M2 stage by less hydrous, Fe-Co richer staurolite. Hydrogen zoning in staurolite is facilitated by the sensitivity of its structure to changing P-T conditions. Water in staurolite is maximized at high P and low T. Cores of staurolite from Samos represent the most hydrous staurolite compositions reported to date. References Koch-Müller, M. & Langer, K. (1998): Quantitative IR-spectroscopic determination of the component H2O in staurolite. European Journal of Mineralogy 10: 1267-1273. Wirth R. (2004) Focused Ion Beam (FIB): A novel technology for advanced application of micro-and nanoanalysis in geosciences and applied mineralogy. European Journal of Minerlaogy 16: 863 – 876.

3
149

Valence state of Co ions in nanostructured LiCoO2 obtained by high pressure torsion method
V. R. Galakhov1∗ N. A. Ovechkina1 , B. A. Gizhevskii1 , C. Taubitz2 , , M. Raekers2 , A. R. Cioroianu2 , M. Neumann2 , A. S. Semenova3 , D. G. Kellerman3 , R. Ovsyannikov4 , and S. L. Molodtsov5 Institute of Metal Physics, Russian Academy of Sciences — Ural Division, 620041 Yekaterinburg GSP-170, Russia 2 Fachbereich Physik, Universit¨t Osnabr¨ck, D-49069 Osnabr¨ck, Germany a u u 3 Institute of Solid State Chemistry, Russian Academy of Sciences — Ural Division, 620041 Yekaterinburg GSP-145, Russia 4 BESSY GmbH, Albert-Einstein-Str. 15, D-12489 Berlin, Germany Institut f¨r Festk¨rperphysik,Technische Universit¨t Dresden, D-01062 Dresden, Germany u o a
1

5

Oxides Lix CoO2 have served as cathode materials for Li batteries. Most of the electrochemical and physical properties would be affected by the electronic structure of these compounds. The electronic ground state of Co3+ ion in LiCoO2 can be written as t3 t3 e0 . We have found 2g↑ 2g↓ g that the deficiency of oxygen in LiCoO2−δ give rise of the divalent cobalt ions and this fact can be detected by Co 2p and Co 3s X-ray photoelectron spectra [1]. In lithium deintercalated cobaltites Li1−x CoO2 (x > 0) the charge compensation occurs through the formation of holes in the O 2p states, whereas the electron configuration of the cobalt ions remains unchanged [2, 3]. Here we present the results of X-ray absorption (XAS) and X-ray photoelectron (XPS) spectroscopy studies of nanostructured LiCoO2 .
ϕ=30 o ϕ=15 o ϕ=0 initial
Intensity (arb. units)
o

Co 3p LiCoO2 after deformation o ϕ=30 ϕ=15 ϕ=0
o o

LiCoO2
Intensity (arb. units)

Co 2p XPS

Li 1s

Li0.6CoO2 LiCoO2-δ LiCoO2

810

800

790

780

770

70

60

50

Binding energy (eV)

Binding energy (eV)

Fig. 1: Co 2p X-ray photoelectron spectra of nanostructured LiCoO2 obtained by high pressure torsion method. The anvil rotation angles ϕ indicate deformation degrees. For comparison, the spectrum of initial LiCoO2 is presented.

Fig. 2: Co 3p and Li 1s X-ray photoelectron spectra of LiCoO2 subjected to plastic deformation. For comparison, the spectra of initial LiCoO2 and defective cobaltites Li0.6 CoO2 and LiCoO2−δ are presented.

A single-phase, homogeneous ceramic sample of LiCoO2 was prepared by sintering a mixture of Co3 O4 and Li2 CO3 . Nanostructured LiCoO2 samples were obtained by high pressure torsion method.
∗

e-mail: galakhov@ifmlrs.uran.ru

150

High pressure torsion was realized by means of a 100-t press and Bridgman anvils. Initial powders of LiCoO2 were placed between anvils and pressed at 3–8 GPa. Shear deformation was achieved by rotation of one of the anvils. The degree of strain was specified by the anvil rotation angle ϕ. Powder X-ray diffraction was used to confirm single-phase specimens. All the nanostructured samples contain 100 % of hexagonal LiCoO2 . The X-ray photoelectron spectra of the defective and nanostructured lithium cobaltites were obtained with a PHI 5600 ci Multitechnique System XPS spectrometer using monochromatized Al Kα radiation. The energy resolution was about 0.4 eV. The Co 2p X-ray absorption spectra (XAS) were measured at the Russian-German Beam Line at the BESSY storage ring. Deformation of LiCoO2 can lead either to defects in oxygen and therefore to the formation of Co2+ ions or to defects in lithium, nominally accompanied by Co4+ ions. According to Co 2p X-ray photoelectron spectra presented in Fig. 1, there is no difference between the Co 2p X-ray photoelectron spectra (XPS) of nanostructured and initial (coarse grain) samples of LiCoO2 obtained at the anvil rotation angles ϕ ≤ 30◦ (see Fig. 1). This means that the valence state of the Co ions is not changed. Note, Co 2p XPS and XAS and Co 3s X-ray photoelectron spectra of defective cobaltites Li1−x CoO2 are nearly identical to the spectrum of the initial oxide LiCoO2 [2, 3]. On the other hand, the increase of the anvil rotation angles lead to an increase of the relative intensity of Li 1s / Co 3p XPS lines (see Fig. 2). Note that photoelectron spectra give an information about the surface of materials. This means that the relative concentration of Li atoms is increased at the sample surface region. One can see that the maximum of the Li 1s line in defective LiCoO2 is shifted to the low-binding-energy side. We suggest that it is evidence of the formation of Li2 O.

Co 2p3/2 XAS

Intensity (arb. units)

Co3O4 LiCoO2
p=8 GPa, ϕ=4x360o p=5 GPa, ϕ=30o p=3 GPa, ϕ=15o p=3 GPa, ϕ=0o

Co 2p1/2 XAS
after deformation

Li0.60CoO2 Li0.96CoO2 LiCoO2
770 775 780 initial 785 790 795 800

Photon energy (eV)
Fig. 3: Co 2p X-ray absorption spectra of LiCoO2 subjected to plastic deformation, initial LiCoO2 , defective cobaltites Li0.6 CoO2 , Li0.96 CoO2 , and Co3 O4 . The spectrum of Co3 O4 is reproduced from [4].

151

Fig. 3 shows Co 2p X-ray absorption spectra of LiCoO2 subjected to by high pressure torsion deformation. The spectra of the samples obtained at pressures of 3–5 GPa and the anvil rotation angles ϕ ≤ 30◦ present only minor changes indicating that the Co ions are not changed. The absence of changes shows that the Co ions remain in a trivalent Co3+ low-spin state. An increase of pressure to 8 GPa leads to radical changes of the Co 2p spectrum. The spectrum of nanostructured LiCoO2 obtained at a pressure of 8 GPa and ϕ = 4 × 360◦ shows the presence of Co2+ ions. XPS and XAS measurements in TEY mode are surface sensitive to a few hundred Angstroms. We suggest that after deformation, a change of the composition of lithium cobaltite occurs. The concentrations of oxygen as well of lithium decrease: Li1−x CoO2−δ (δ > x/2). The surface of the samples is enriched by Li. Such situation is described in Ref [5]. Note that extration of Li from coarse-grained LiCoO2 occurs only at temparatures higher than 1200 K. This work was supported by the Russian Foundation for Basic Research (Grants Nos 07-02-00540 and 08-03-99071) and by the bilateral Program “Russian-German Laboratory at BESSY”.

[1] V. R. Galakhov, V. V. Karelina, D. G. Kellerman, V. S. Gorshkov, N. A. Ovechkina, M. Neumann, Phys. Solid State 44, 266 (2002). [2] V. R. Galakhov, N. A. Ovechkina, A. S. Shkvarin, S. N. Shamin, E. Z.Kurmaev, K. Kuepper, A. Tak´cs, M. Raekers, S. Robin, M. Neumann, G.-N. Gavrilˇ, A. S. Semenova, D. G. Kellermann, a a T. K¨¨mbre, J. Nordgren, Phys. Rev. B 74, 045120(6) (2006). aa [3] V.R. Galakhov, M. Neumann, D.G. Kellerman, Appl. Phys. A 94, 497 (2009). [4] D. Bazin, P. Parent, C. Laffon, O. Ducreux, J. Lynch, I. Kovacs, L. Guczi, and F. De Groot, J. Synchrotron Rad. 6, 430 (1999). [5] E. Antolini, J. Eur. Ceram Soc. 18, 1406 (1998).

152

Nanostructural characterization of TiN-Cu films using EXAFS spectroscopy
F. Pinakidou1, M. Katsikini1, E.C. Paloura1, P. Patsalas2, G. Abadias3
Aristotle University of Thessaloniki, School of physics, 54124 Thessaloniki, Greece University of Ioannina, Department of Materials Science & Engineering, 45110, Ioannina, Greece 3 Laboratoire de Métallurgie Physique, UMR CNRS 6630, Université de Poitiers, SP2MI, Téléport 2, BP 30179, 86962 Futuroscope-Chasseneuil Cedex, France
2 1

Transition metal-TiNx material systems have been a subject of extensive scientific and engineering studies since they exhibit excellent mechanical properties as well as wear and corrosion resistance. The problem of toughness in superhard nanocomposites can be addressed by the formation of hard nanocrystalline phases within a metal matrix, such as TiN in Ni or Cu.1 In such coatings, one of the metals can form a hard nanocrystalline nitride phase while the other one remains unreacted. Here we apply X-ray absorption fine structure (XAFS) measurements at the Cu-K-edge in order to identify composition dependent changes in the bonding environment of Cu in TiN-Cu and Ti1-xCux films. The studied nanocomposite TiN-Cu and Ti1-xCux films were deposited at room temperature on Si substrates by reactive magnetron co-sputtering. In particular, the deposition of the former films occurred under Ar+N2 atmosphere and of the latter films under inert gas (Ar) atmosphere. Details on the growth conditions have been reported previously.2 The Cu-KEXAFS spectra of the studied samples were recorded at the (a) Ti1-xCux KMC2 beamline in the fluorescence yield (FLY) mode. A Cu 25 film, 203 nm-thick, deposited on quartz under the same conditions as the Ti1-xCux films was used as reference. 20 The χ(k) EXAFS spectra of the Ti1-xCux and TiN-Cu studied 52.7at% films were subjected to Fourier filtering in the distance range 1.615 Cu 2.6Å and the corresponding Fourier Transforms (FTs) (k range: 35.5at% 3.5-12.0Å-1) are shown in Fig. 1(a) and (b), respectively. The 10 Cu analysis of the EXAFS spectrum of the reference Cu film reveals 24.1at% Cu (quartz) that Cu is coordinated with 12.8 (±0.8) Cu atoms and the 5 respective Cu-Cu bondlength is equal to 2.55Å (±0.01), i.e. 24.1at% Cu crystalline Cu is detected. On the contrary, in the studied Ti10 0 1 2 3 4 xCux and TiN-Cu samples, the nanostructure of Cu changes as a R (Å) result of the different chemical composition. 40 (b) TiN-Cu More specifically, as shown in the upper panel of Fig. 2, in the Ti1-xCux samples the total coordination number of Cu, i.e. the 32 sum of both Cu and Ti first neighbours (NCu and NTi, respectively), increases from ≅6 to ≅11 when the Cu increases 24 from 24.1 to 52.7 at%. This change can only be attributed to the systematic increase of the number of Cu atoms bonded to Cu 16 Cu film (NCu). Indeed, as shown in the middle panel of Fig. 2, NTi is practically constant (within the error bar) and ranges between 3.5 67.7at% Cu 8 and 4.1 (±0.4-0.3), while on the contrary, as depicted in the lower 37.8at% Cu panel of Fig. 2, NCu increases linearly from 2.2 to 7.1 (±0.4-0.5). 27.3at% Cu Additionally, the EXAFS analysis reveals that the coordination 0 0 1 2 3 4 5 number of Cu does not depend on the substrate and is practically R (Å) equal between the samples with the same composition (35.5 at% Fig. 1: Fourier Transforms Cu) grown on Si and quartz: in the former case the 1st nn shell (FTs) of the filtered k3*χ(k) Cuconsists of 4.2±0.6 Cu and 3.6±0.4 Ti atoms while in the latter K-EXAFS spectra of (a) the Ti1case Cu is bonded to 3.5±0.6 Cu and 3.7±0.3 Ti atoms. In all xCux and (b) the TiN-Cu films. The raw data and the fitting are studied Ti1-xCux samples, the Cu-Ti bondlength is equal to 2.45- shown in thin and thick solid 2.47Å (±0.01) while the Cu atoms are located at 2.54-2.56Å lines, respectively. 1
153

FT (arb. units)

FT (arb. units)

(±0.02), i.e. as in the reference Cu film. Thus, it is revealed that independently of the Cu content, Cu belongs to an amorphous Cu-matrix where Ti partially substitutes Cu atoms while also forms intermetallic TiCu. As the Cu content increases, the total coordination number of Cu increases linearly. More specifically, Cu belongs to intermetallic TiCu nanocrystallites and prefers to segregate when the Cu concentration exceeds 40 at%. The EXAFS analysis of the TiN-Cu films reveals that only in the sample with the intermediate Cu concentration (37.8 at% Cu), Cu is bonded to both Cu and Ti. The 1st nn shell consists of 1.9±0.2 Ti and 2.6±0.3 Cu atoms that are 12 NTi+NCu located at 2.47Å (±0.01) and 2.55Å (±0.02) respectively, i.e. at 10 the same distance as in the Ti1-xCux samples. However, in this sample fewer Ti atoms (1.9±0.2) are bonded to Cu, compared 8 to the respective number of Ti atoms in the Ti1-xCux sample with 35.5 at% Cu (3.6±0.4). This modification in the TiN/Cu 6 sample with 37.8 at% Cu can be attributed to formation of TiN 5 NTi crystallites, which have also been detected by XRD 3 measurements. Thus it can be proposed that, when the films 4 are grown under Ar-N2 atmosphere, Ti prefers to form TiN and any excess Ti atoms bond to Cu. On the contrary, in the sample with the lowest Cu concentration (27.3 at%), only Ti 3 atoms comprise the 1st nn shell. Cu is bonded to 6.1±0.4 Ti 8 NCu atoms and the Cu-Ti bondlength is equal to 2.45Å (±0.01), i.e. the formation of intermetallic Ti-Cu bonds is identified. 6 Finally, in the sample with the highest Cu content (67.7 at%) 4 Cu is coordinated with 10.6±0.4 Cu atoms located at 2.53Å (±0.01), i.e. Cu belongs to an amorphous Cu matrix. 2 Therefore, it is concluded that in the TiN-Cu samples, the 20 25 30 35 40 45 50 55 different Cu content results in significant modifications in the Cu (at%) nanostructure of Cu: in the lowest Cu concentration limit, Cu forms intermetallic TiCu, in the sample with the intermediate Fig. 2: Modification of the Cu concentration both TiCu nanocrystallites and Cu-Cu bonds coordination number of Cu as a function of the Cu content in the are detected while finally, in the highest Cu concentration Ti1-xCux films. limit, Cu belongs to an amorphous Cu matrix. To conclude, the effect the different chemical composition on the nanostructure of Cu in Ti1-xCux and TiN-Cu nanocomposite films was studied using EXAFS measurements at the Cu-K-edge. It is disclosed that in all studied Ti1-xCux films, Cu belongs to amorphous Cu regions and to intermetallic TiCu nanocrystallites. However, the coordination number of Cu increases as a function of the Cu content in the samples, which can be attributed only to the systematic increase in the number of Cu atoms bonded to Cu. On the contrary, in the TiN-Cu films, the alteration in the chemical composition affects both the coordination number and type of atoms in the 1st nn shell. In particular, Cu is coordinated with both Cu and Ti atoms only in the sample with intermediate Cu concentration (37.8 at%). In the sample with the highest Cu content (67.7 at%), Cu belongs to pure, amorphous Cu regions, while in film with the lowest Cu content (27.3 at%), only intermetallic Cu-Ti bonds are identified.
Acknowledgements: The measurements at BESSY were supported by the IA–SFS program of the European Community (Project Nr BM.08.2.80455). References 1 J. Musil, P. Zeman, H. Hruby, P.H. Mayrhofer, Surf. Coat. Technol. 120-121, 179 (1999). 2 G. Abadias, Y.Y. Tse, J. App. Phys. 95 2414 (2004). 3 G. Abadias, private communication

2
154

Coordination Number

μ-XRF and μ-EXAFS studies of an Al matrix Fe-Ni composite
F. Pinakidou, M. Katsikini, E.C. Paloura.
Aristotle University of Thessaloniki, School of Physics, 54124, Thessaloniki, Greece

The mechanical properties of metals can be improved by either alloying with other metals or by chemical reactions which lead to the formation of intermetallic products.1 When the surface of a Fe-Ni alloy is exposed to liquid Al, dissolution occurs followed by growth of intermetallic layers at the alloy-aluminium interface.2,3 The growth of these phases is controlled by chemical reactions at the interfaces and the interdiffusion of Fe and Ni. The strength of the matrix can be further improved via a direct reaction synthesis in which reactant powders are directly added into the molten metal. The reinforcing particles are formed in situ through reactions between reactants or between the reactant and a component of the metal. For example, a mixture of K2TiF6 and KBF4 is added in to the molten Al matrix resulting in the formation of Ti-B particles which are well incorporated in the metallic matrix. Here we report on the distribution of the metals and the local coordination of Ni in a FeNi rod (97 wt% Fe and 3 wt% Ni) which was dissolved in liquid Al. A 5 wt% mixture of K2TiF6 and KBF4 was added at 1060oC and stirred for 30min. The μ-XRF and μ-Ni-KEXAFS measurements were conducted at the KMC2 beamline which is equipped with a double-crystal monochromator and capillary optics that reduce the beam diameter to 5μm. The μ-XRF maps were recorded using excitation photons of 9500eV, i.e. higher than the Fe-K and Ni-K absorption edges, using an energy dispersive (Röntec) fluorescence detector. The μNi-K-EXAFS spectra were recorded from different positions of the sample surface while an Ni-K-EXAFS spectrum was recorded from a random sample spot, after removing the capillaries. The spectra were recorded in the fluorescence yield (FLY) mode using a Röntec detector. The XRF maps of the studied sample reveal that the distribution of both Fe and Ni is inhomogeneous. More specifically, as shown in Fig.1, Ni-rich (H-regions) and Ni-poor (Lregions) regions are detected. It should be noted that in the Ni-rich islands the corresponding Fe concentration is low, while in the Ni-poor regions, Fe segregates. On the contrary, the μXRF spectra recorded from the H- and L-regions reveal that the distribution of both Mn and Cr, which are contaminants in the Fe-Ni alloys, is homogeneous (Fig. 2).
100 50

Ni
H-region L-region

0 -50 -100 -150 -150 -100 -50 0

50 100 150

100 0 0.11 50 0.22 0.33 0 0.44 0.56 -50 0.67 0.78 -100 0.89 1.0 -150 -150 -100 -50

Fe

0 0.11 0.22 0.33 0.44 0.56 0.67 0.78 0.89 1.0 50 100 150

μm

0

μm

μm

Fig. 1: 300x250μm2 μ-XRF maps of the Fe and Ni distributions. The regions denoted L and H, correspond to Ni-poor and Ni-rich regions, respectively.

In order to investigate the possible changes in the bonding environment of the Ni atoms as a function of the Ni segregation or depletion, we recorded μ-Ni-K-EXAFS spectra from the H- and L-regions, respectively. The Fourier transforms (FT) of the k3-weighted χ(k) μ-Ni-KEXAFS spectra (k-range 2.5-9.0Å-1) are shown in Fig.3. In the same figure we include the FT of the Ni-K-EXAFS spectrum (k-range 2.5-9.0Å-1) recorded after removing the capillaries. As shown in Fig.3, in the FTs of both the μ-EXAFS and EXAFS spectra, only the 1st nearest 1
155

neighbor (nn) shell is resolved, indicating that the bonding environment of Ni is amorphous.4 The μEXAFS and EXAFS spectra were fitted assuming that Ni is bonded only to Al. However, the crystalline models used in the fitting process were modified for the Ni-rich and Ni-poor regions: in the Ni-rich regions it is assumed that Al atoms substitute Ni in the structure of crystalline Ni, while in the Nipoor regions, Ni is assumed to form intermetallic NiAl nanocrystallites. The μ-EXAFS analysis reveals that the bonding environment of Ni changes as a result of the inhomogeneous distribution of Ni. More specifically, in the Ni-rich regions (H-region), the Ni atoms are bonded to 11.5±0.7 Al atoms that belong to an amorphous Ni matrix. On the contrary, in the region with low Ni content (L-region), Ni is coordinated with 8.8±0.5 Al atoms. In this region, the Ni atoms occupy sites in an amorphous Al matrix. Furthermore, the Ni-K-EXAFS analysis discloses that the coordination number of Ni is equal to 9.7±0.8, i.e. equal to the average number of Al atoms in the L- and H-regions of the sample. Additionally, the μ-EXAFS and EXAFS analysis reveal that the Ni-Al bondlength is equal to 2.472.49Å (±0.01), i.e. independent of the inhomogeneous distribution of Ni. The modification in the bonding environment of Ni can be explained on the basis of the different phases formed in the sample. In particular, in the Hregion, Ni belongs to a pure amorphous Ni phase where the Al atoms have partially substituted Ni atoms (Ni is bonded with 12 Ni atoms in metallic Ni). On the contrary, in the L-region, the number of Al atoms in the 1st nearest shell of Ni is smaller. Since in the cubic structure of intermetallic NiAl, each Al atom is bonded with 8 Al atoms, it can be proposed that in this region, Ni belongs to intermetallic NiAl nanocrystallites. Finally, the coordination number of Ni, as determined by the EXAFS analysis is, as expected, equal to the average number of Al atoms in the H- and L- regions.

9

Intensity (arb. units)

H - Region L - Region

Fe

Ni

6
Al Mn

3
1 2 3 4

Cr Fe
5 6

Ni

0

1

2

3

4

5

6

7

8

9

Energy (KeV)
Fig. 2: μ-XRF spectra recorded from the Hand L-regions, respectively. The low energy region (1 to 6.5 keV) is shown at the inset of the figure.

25

20

FT (arb. units)

15 sample 10 H-Region

5 L-Region 0 0 1 2 3 4 5

R (Å)
Fig. 3: Fourier Transforms (FT) of the μ-NiK-EXAFS spectra recorded from the Ni-rich (H-region) and Ni-poor regions (L-region). The FT shown on the top (denoted “sample”) corresponds to the Ni-K-EXAFS spectrum recorded without the capillaries. The raw data and the fitting are shown in thin and thick solid lines, respectively.

Acknowledgements: The measurements at BESSY were supported by the IA–SFS program of the European Community (Proposal Nr 2008_1_70802).

References
1

F.L. Mathews, R.D. Rawlings, “Composite Materials: Engineering and Science”, Cambridge: CRC Press, Woodhead (1999). 2 V.I. Dybkov, J. Mat. Sci., 21, 3078 (1986). 3 V.I. Dybkov, J. Mat. Sci., 25, 3615 (1990). 4 . F. Pinakidou et al., J. App. Phys., 102, 113512 (2007).

2
156

Two kinds of interfaces in SrTiO3/LaAlO3 superlattices revealed by resonant soft x-ray scattering
H. Wadati1, J. Geck1, D. G. Hawthorn1, T. Higuchi2, M. Hosoda2, C. Bell2, Y. Hikita2, H. Y. Hwang2, C. Schüßler-Langeheine3, E. Schierle4, H.-H. Wu3,5, E. Weschke4, S.-W. Huang5, D. J. Huang5, H.-J. Lin5, I. Elfimov1, G. A. Sawatzky1
1

Department of Physics and Astronomy, University of British Columbia, Vancouver, Canada 2 Department of Advanced Materials Science, University of Tokyo, Kashiwa, Japan 3 II. Physikalisches Institut, Universität zu Köln, Köln, Germany 4 Helmholtz-Zentrum Berlin für Materialien und Energie, Berlin, Germany 5 National Synchrotron Radiation Research Center, Hsinchu, Taiwan

Introduction: The interfaces of hetero-junctions composed of transition-metal oxides have recently attracted great interest. Among them, the interface between two band insulators SrTiO3 (STO) and LaAlO3 (LAO) is especially interesting due to the metallic conductivity [1] and even superconductivity [2]. Several reports suggest that n-type (LaO/TiO2/SrO) interface is metallic and the p-type (LaO/AlO2/SrO) interface is insulating. In this study we investigated the electronic structure of the STO-LAO superlattice (SL) by resonant soft x-ray scattering [3], which has recently been used to study SrMnO3-LaMnO3 SLs [4]. Experiment: The superlattice sample consisted of seven periods of 12 unit cells (uc) of STO and 6 uc of LAO. The present samples was grown on a STO (001) substrate by the pulsed laser deposition technique at an oxygen pressure of 1.0×10-5 Torr and a substrate temperature of 700 oC. A schematic view of the fabricated superlattice is shown in Fig. 1. The resonant soft x-ray scattering experiments were performed at the BESSY undulator beam line UE46-PGM. The spectra were taken at room temperature. The incident light was polarized in the vertical direction (σ polarization) or in the Figure 1: Schematic view horizontal direction (π polarization) with the of the SrTiO /LaAlO 3 3 detector integrating over both final polarizations. superlattice sample.

157

Results and discussions: Figure 2 shows the photon-energy dependence of the (002) and (003) peaks near the O 1s absorption edge measured with σ (a) and π (b) polarizations. The structure factor for this (003) peak is proportional to (fTiO2, int - fTiO2 + fAlO2, int - fAlO2), where fTiO2 and fAlO2 are the scattering factors of the TiO2 and AlO2 planes of the STO and LAO layers, respectively, and the Figure 2: Photon-energy dependence subscript “int” means the scattering of the (002) and (003) peaks near the factor for the interface. The (003) O 1s absorption edge measured with reflection is forbidden by symmetry as σ (a) and π (b) polarizations. Top and long as the interface form factor is the middle panels show intensity maps of same as the bulk. The existence of this (002) and (003) regions, respectively. peak in Fig. 2 therefore means that Here, bright parts correspond to high intensities. Bottom panels show the some kind of reconstruction occurs at (002) and (003) peak heights together the interface. From top and middle with the XAS spectra. panels, we can see strong polarization dependence. Also we notice that there are two structures in the energy region of 530 – 532 eV, which comes from hybridization with Ti 3d t2g states. The relative intensities of these two structures depend on polarizations. The value of the energy splitting is about 1 eV. From linear dichroism in Ti 2p x-ray absorption, Salluzzo et al. [5] concluded that the xy orbital is stabilized at the interface, and the value of the energy splitting between xy and the other two t2g orbitals (yz and zx) is about 50 meV, which is much smaller than the 1 eV splitting observed in this study. Now we are performing band-structure calculations to explain this 1 eV splitting. References: [1] A. Ohtomo and H. Y. Hwang, Nature 427, 423 (2004). [2] N. Reyren et al., Science 317, 1196 (2007). [3] H. Wadati et al., arXiv:0810.1926v1. [4] S. Smadici et al., Phys. Rev. Lett. 99, 196404 (2007). [5] M. Salluzzo et al., arXiv:0812.1444.v1.
158

Doping of organic semiconductors to improve electronic interface properties of heterojunction solar cells. Thomas Mayer, Corinna Hein, Eric Mankel, and Wolfram Jaegermann Darmstadt University of Technology, Institute of Materials Science Petersenstr. 23, D-64287 Darmstadt, Surface Science Division BMBF Initiative “Organic Photovoltaics” Due to low dielectric constants and strong intra and extra-molecular relaxation in organic semiconductors optically generated electron hole pairs form strongly bound excitons of 0.5 to 1eV binding energy. In organic solar cells hetero-junctions of electron donor and acceptor materials are used for dissociation of the excitons and charge separation (Fig.1). Even after injection of the electron from the donor to the acceptor electron hole pairs may be bound across the hetero-junction forming so called geminate pairs still likely to undergo recombination. Therefore the electronic structure of such organic hetero-interfaces concerning band/orbital line up and potential gradients is of basic importance for the cell efficiency. We investigate the electronic structure of the CuPc/BPEPTCDI donor/acceptor interface and the changes induced by pdoping of CuPc with WO3 using SXPS on step wise in situ prepared interfaces. CuPc and BPE-PTCDI were provided by the project partner BASF. The measurements are performed at the BESSY undulator beamline U49 using the SoLiAS endFig.1: Exciton dissociation and charge separation at the donor station. The electronic structure acceptor hetero-junction interface in organic solar cells. of the donor/acceptor system CuPc/BPE-PTCDI indicates unfavorable potential gradients that retain the separated charge carriers at the interface instead of driving them towards the external contacts (Fig.2). In order to reverse the interface band bending p-doping of CuPc and/or n-doping of BPE-PTCDI is suggested. We tested p-doping of CuPc with WO3 as an inorganic molecule of high electron affinity. Due to a high number of trap states in the energy gap of organic semiconductors much Fig.2: Band diagram of the electronic structure of the higher dopant content is donor/acceptor CuPc/BPE-PTCDI hetero-junction. The necessary in order to shift the potential gradients retain charge carriers at the interface Fermi level as compared to whereby the efficiency of the photovoltaic device is reduced. inorganic semiconductors[1].

159

. With increasing WO3 content, the Fermi level position in CuPc changes up to 0.7 eV towards the valence band as derived from 1s core level positions of C and N in CuPc. The band diagrams of the single materials and the CuPc+WO3 blend as derived with SXPS are given in Fig.3. In the Anderson model that just compares work functions, p doping of CuPc with WO3 and the reversal of the band bending at the CuPc+WO3/BPE-PTCDI interface compared to bare CuPc can be expected. The photoemission set Fig.4 taken in the course of stepwise deposition of BPE-PTCDI onto WO3 doped CuPc delivers the information for drawing the band diagram of the CuPc+WO3/BPE-PTCDI interface Fig.5.

Fig.3: Band diagram of the single materials CuPc, WO3, BPEPTCDI and a CuPc+WO3 blend. For the organic materials values for HOMO photoemission maxima have been determined using SXPS on in-situ prepared films. For HOMO LUMO energy gaps data from inverse photoemission are used. The work function is derived from the valence band spectra secondary edge.

Fig.4: Set of SXPS spectra taken in the course of stepwise deposition of BPEPTCDI onto WO3 doped CuPc. In the valence band spectra VB of the bare CuPc+WO3 film the W4f emission of the dopant is observable.

The line up of the CuPc and BPE-PTCDI HOMO bands can directly be read from the HOMO section of the valence band spectra Fig. 4 top right. The change in work function is derived from the secondary edge (SK top left). Band bending in the substrate and adsorbate film is

160

taken from induced shifts in the respective core level emissions in the course of stepwise interface formation (Fig.4 bottom).

Fig 5: Band bending in the WO3 doped CuPc substrate and in the BPE-PTCDI stepwise adsorbed film as derived from CuPc N1s and BPE-PTCDI HOMO and N1s levels. In addition the change in the energy position of the secondary edge i.e. of the work function is shown and the change in of ΔEvac which is interpreted as interface dipole is indicated.

The band diagram of the WO3 doped CuPc/BPE-PTCDI interface is displayed in Fig.6. As compared to the undoped case Fig.2, the band bending is reversed to the intended direction, supporting the transport of electrons in BPE-PTCDI and holes in CuPc towards the external contacts. Summary: Applying SXPS p-doping of CuPc using WO3 has been demonstrated to improve the potential distribution at the CuPc/BPE-PTCDI interface for solar cell applications.

Fig.6: Band diagram of the CuPc+WO3/BPE-PTCDI heterojunction. P-doping of CuPc with WO3 leads to potential gradients in the favourable direction that promotes the transport of charges away from the interface.

1.

Mayer, T., C. Hein, J. Haerter, E. Mankel, and W. Jaegermann, A doping mechanism for organic semiconductors derived from SXPS measurements on co-evaporated films of CuPc and TCNQ and on a TCNQ/CuPc interface. Proc. SPIE 2008. 7052, : p. 705204
161

On the electronic structure of the oxypnictide family
T. Kroll*,1, S. Bonhommeau2, T. Kachel2, H.A. Dürr2, F. Roth1, R. Kraus1, J. Werner1, G. Behr1, A. Koitzsch1, A. K. Ariffin3, R. Manzke3, F.M.F. de Groot4, J. Fink1, 2, H. Eschrig1, B. Büchner1, and M. Knupfer1
1 IFW Dresden P.O. Box 270016, 01771 Dresden, Germany 2 BESSY, Albert-Einstein-Strasse 15, 12489 Berlin, Germany 3 Institut für Physik, Humboldt-Universität zu Berlin, Newtonstrasse 15, 12489 Berlin, Germany 4University Utrecht, Sorbonnelaan 16, 3584 CA Utrecht, Netherlands

We investigated the recently found superconductor LaFeAsO1-xFx by X-ray absorption spectroscopy (XAS). A shift in the chemical potential is visible in both the Fe L2,3 and O K edge spectra which emphasizes the importance of band effects and moderate correlations in these compounds. From experimental Fe L2,3 edge spectra and charge transfer multiplet calculations we gain further information on important physical values such as hopping parameters, the charge transfer energy ∆, and the on-site Hubbard U. Furthermore we find the system to be very covalent with a large amount of ligand holes. Measurements of the absorption and photoemission spectra of undoped LnFeAsO (Ln=La, Ce, Sm, Gd) reveal no significant changes when exchanging the rare-earth ion or when varying the temperature below and above the magnetic/structural transition temperature. KEYWORDS: Oxypnictides, Spectroscopy, Electronic Structure Core level spectroscopic measurements such as X-ray absorption spectroscopy (XAS) are appropriate experimental methods to shed new light on the electronic structure of the recently discovered superconductor LaFeAsO1-xFx1). In XAS, a core electron is excited into an unoccupied state near the Fermi level, i.e. one probes the empty states. In this article, we present experimental data from Fe L2,3 and O K absorption edges together with theoretical descriptions such as charge transfer multiplet and local density approximation (LDA) calculations. For the presented measurements we chose undoped LaOFeAs and electron doped LaFeAsO1-xFx polycrystalline samples in a doping range between x=0.0 and 0.15. Polycrystalline samples were prepared as pellets as described in Ref. 2, the XAS signal has been taken by recording the fluorescence signal. Fe L and O K edge spectra have been normalized at 750 eV and 610 eV, respectively. Further experimental details are given in Ref. 3. According to the dipole selection rules, the Fe L2,3 absorption edges correspond to excitations of Fe 2p core level electrons into unoccupied Fe 3d electronic states. In Fig. 1(a) the experimental Fe L3 edge XAS spectra for different doping levels are shown. Two main changes appear with F doping. The energy position of the main peak around 708 eV shifts slightly with doping towards lower energies. This shift amounts to ≈150meV on going from x=0.0 to x=0.15 and can be explained by the observation that the XPS Fe 2p core level excitations do not shift relative to the chemical potential with doping within the experimental resolution4), while the chemical potential shifts by 200meV with doping from x=0.0 to x=0.25). In other words, this excitation

Fig. 1. LaO1-xFxFeAs: Doping dependence of the Fe L-edge. (a) Experimental L3-edge for various doping levels. (b) Shift of the onset of the main peak in eV relative to x=0.0 as a function of F doping x.

162

Fig. 2. LaFeAsO1-xFx: (a) Doping dependence of XAS O K edge spectra. (b) Energy shift of the onset of the 1st peak as compared to x=0.0 for experimental and theoretical results.

energy as seen in Fig. 1 decreases upon doping. Moreover, the onset of the L3 edge6) shifts to higher photon energies by ≈600 meV. Note that such a shift could also cause an asymmetric peak narrowing and affect the position of the peak maxima. A shifted onset is consistent with additional electrons at the Fe sites, which diminishes the number of holes, i.e. the total intensity at the Fe L-edge. Therefore, the doped electrons reside (partially) at the Fe sites, which is supported by valence band photoemission spectroscopy (PES) 5). The observed shift of the onset of the Fe L edge spectra is especially remarkable since it emphasizes the importance of band effects and the absence of strong correlations as they have been observed for e.g. the cuprates. The onset of the spectra shifts monotonically to higher photon energies (see Fig. 1(b)) caused by a shift of the chemical potential as it has been observed by PES measurements5). Simulations of the Fe L2,3 edge require consideration of multiplet splitting, hybridization, and crystal field effects. We performed charge transfer multiplet calculations for divalent Fe2+ (d6) in tetrahedral (Td) symmetry. Note that a band effect such as the shift of the chemical potential is beyond this local approach. The parameter set that reproduces the experimental data best (not shown here, e.g. Ref. 3) is 10Dq=0.2 eV, ∆=E(d7L)-E(d6)=1.25 eV (L denotes a ligand hole), U=1.5 eV, and pdπ=0.27 eV. The core hole potential Q has been set to Q=U+1 eV. The Slater-Condon parameters have been reduced to 80% of their Hartree-Fock values as it is reasonable in solids, which leads to the two Hund's couplings Jeg=0.90 eV and Jt2g=0.78 eV
163

for the ground state. Such a parameter set leads to a highly covalent system and a high spin state of S=2. The shoulder at ≈712 eV is provoked by charge transfer effects and emphasizes the hopping values above. Since the core hole potential is rather small (Q=2.5 eV), the excited states are not shifted far out of the Fe 3d band7), and therefore band effects become visible in the experimental spectra. A second edge that is worth to investigate is the O K edge. Note that the O ions are located within the LaO1-xFx layer, and therefore do not behave like ligands at the transition metal ion contrary to cuprates or cobaltates. In Fig. 2(a), O K edge spectra for different doping levels are shown for photon energies between 529 and 539 eV. This region can be assigned to excitations from the O 1s core level into unoccupied O 2p states. In the XAS spectra, the onset of the 1st peak shifts by 350 meV towards higher photon energies with doping6), whereas the 1st peak itself (at 531 eV) shifts only by ≈100meV and the 2nd peak (at 532.7 eV) does not shift. When comparing the experimental spectra to the partial density of states (PDOS) as gained from LDA calculations, one observes that the overall agreement is good3), and LDA is able to explain all main features, and assign the 1st peak to hybridizations between O and Fe states and the 2nd peak to hybridizations between O and La states. From X-ray photoemission spectroscopy (XPS) experiments5) it has been observed that the La 4d level shifts relative to the chemical potential by about 200 meV from x=0.0 to x=0.1 while the As 3d level hardly shifts. This can be ascribed to a change of the Madelung potential between the As and La layers upon doping, in agreement with O K XAS. When focussing on the onset of the 1st peak, i.e. on the change in the chemical potential, a clear doping dependence is observed. This is further illustrated in Fig 2(b) where the shift of the onset of the 1st peak as compared to x=0.0 is shown. Such an increase is supported by the PDOS since the shift in the onset of the 1st peak between x=0.0 and x=0.125 matches well the slope found from the experimental data. This agreement between theory and experiment stresses the observation that the experimental O K edge is strongly affected by the shift of the chemical potential with doping. In LDA no Coulomb energy U is taken into account. When switching it on at the Fe site, this will have an effect on the energetic position of the Fe 3d spectral weight. As the relative position of the O K XAS peaks matches those determined by LDA calculations within 1 eV,

together with the observation that the LaFeAsO valence-band spectrum can be explained by the DOS as derived from LDA (Ref. 5) but does not match the DOS as derived from LDA+U8) or DMFT9), we conclude that U is small as compared to the conduction-band width. Note that the strong electronic correlations in cuprates such as La2-xSrxCuO2, are clearly visible in the O K absorption edges10). Furthermore, we performed X-ray absorption spectroscopy at the Fe L2,3 edge as well as core level Fe 2p and valence band photoemission spectroscopy on various undoped rare-earth oxypnictides LnFeAsO (Ln=La, Ce, Sm, Gd) at temperatures above and below the phase transition temperatures. In none of our results we find significant change of the spectral shape which leads to the conclusion that the electronic structure of the FeAs layers as seen by these experiments, does not change by exchanging the rare-earth ions or varying the temperature11). Exemplarily we show the Fe L absorption and Fe 2p excitation spectra at 20 K for different rare-earth oxypnictides (Fig. 3).

In summary, from X-ray absorption spectroscopy measurements together with LDA and charge transfer multiplet calculations, deeper insight into the electronic structure of LaO1-xFxFeAs has been proposed. The O K-edge is well described by LDA calculations. The shift in the chemical potential is clearly visible in the absorption edge. Furthermore, the band width could be assigned as an upper limit of the on-site Hubbard U. Band effects have a significant influence also on the shape of the Fe L edge absorption spectra. A shift in the chemical potential towards higher energies is observed in agreement with the results of the O K edge, which stresses the existence Hubbard U significantly smaller than the bandwidth W. Further valuable information could be extracted from Fe L2,3 absorption edge absorption together with charge transfer multiplet calculations in tetrahedral symmetry. The low Hubbard U fits to the upper bound as concluded from the comparison between O K-edge XAS spectra and DOS. Furthermore, due to small values of the charge transfer energy ∆ and the Hubbard U the system turns out to be very covalent. Measurements of the XAS and PES spectra of undoped LnFeAsO (Ln=La, Ce, Sm, Gd) reveal no significant changes when exchanging the rare-earth ion or when varying the temperature below and above the magnetic/structural transition temperature. This investigation was supported by the DFG (SFB 463 and KR 3611/1-1) and DFG priority program SPP1133.
*) E-mail: t.kroll@ifw-dresden.de 1) Y. Kamihara et al., J. Am. Chem. Soc. 130, 3296 (2008). 2) S. Drechsler et al., Phys. Rev. Lett. 101, 257004 (2008). 3) T. Kroll et al., Phys. Rev. B 78, 220502 (2008). 4) A. Koitzsch et al., to be published. 5) A. Koitzsch et al., Phys. Rev. B 78, 180506 (2008). 6) The onset of the peak has been defined as the photon energy at which the intensity reaches 10% of its value at the peak maximum 7) H. Eschrig, arXiv:0804.0186 8) C. Cao, P. J. Hirschfeld, and H.-P. Cheng, Phys. Rev. B 77, 220506(R) (2008) 9) K. Haule, J. H. Shim, and G. Kotliar, Phys. Rev. Lett. 100, 226402 (2008) 10) J. Fink et al., J. Electron Spectrosc. Relat. Phenom. 66, 395 (1994) 11) T. Kroll et a.l, accepted at New Journal of Physics

Fig. 3. LnFeAsO: (a) Fe L2,3 absorption edge and (b) Fe 2p photoemission excitation spectra at 20 K.
164

Evidence for static, site centered stripe order by photoemission on Bi2Sr1.2La0.8CuO6
Valentina Scherer, Christoph Janowitz, Beate Müller, Lenart Dudy, Alica Krapf, Helmut Dwelk and Recardo Manzke Humboldt-Universität zu Berlin, Institut für Physik, Newtonstr. 15, 12489 Berlin Taichi Okuda and Akito Kakizaki Institute of Solid State Physics (ISSP), University of Tokyo, Japan Antiferromagnetic spin correlations play an important role especially for hole underdoped cuprate superconductors. The vicinity of the Mott–insulating parent compound phase and the metallic phase in this regime requires a description beyond Fermi liquid theory. Possibly the most striking evidence is the occurrence of spin-and charge separation in form of the so called static stripe phase at a hole doping of nH = 1/8 as observed first by Tranquada [1] in a (Nd)-LaSrCuO- cuprate by neutron scattering. Photoemission measurements on cuprates of the same family revealed a dual nature of the Fermi-surface: features due to itinerant as well as due to static, localized electrons from the stripe phase were observed [2]. Since up to now most of these observations were reported on members of the LaSrCuO- family with one CuO2- layer per unit cell, it is promising to extend these studies on another single CuO2- layer cuprate from a different family. Bi2Sr1.2La0.8CuO6 single crystals with a single CuO2-layer per unit cell, a nominal hole doping around nH = 0.10 and vanishing Tc were therefore grown, characterized and studied by high resolution photoemission. Measurements at BESSY were performed at the BEST-beamline equipped with a 5mNormal Incidence Monochromator and a high-resolution photoelectron spectrometer SES-2002 and at the Photon Factory (Japan) at BL-18A at the angle-resolved photoelectron spectroscopy station for surfaces and interfaces. Spectra recorded at both locations were comparable. The spectra shown below were measured at a photon energy of 22eV and 20 K sample temperature at BESSY. A total resolution of 30meV was found sufficient to resolve the essential details. Due to the very low count rates this was a compromise to collect sufficient data without surface degradation within 48 hours. High quality single crystals of Bi2Sr1.2La0.8CuO6 were grown out of solution. The crystals were characterized using several techniques, e.g. x-ray emission (EDX), ac susceptibility, Laue diffraction and LEED, to obtain information on the chemical composition, the superconductivity, and the quality of the crystals. The five samples investigated by photoemission showed in AC susceptibility no transition to the superconducting state down to 2 Kelvin. The exact hole concentration (nH) was determined from the linear relation between La content and nH derived from X-ray absorption spectroscopy data [3,4]. This alternative method for the quantitative evaluation of the hole content uses the signal of the CuL3 edge. It consists of a so called “white line”, representing the excitation of Cu-2p-electrons into Cu-3d states, and a satellite peak, appearing as a high energy shoulder, with an intensity varying according to the hole content of the sample. To obtain a numerical value for the hole content the ratio of the intensity of the satellite peak to the intensity of white line plus satellite peak has to be evaluated.

165

Photoemission intensity maps obtained by integrating 200meV (upper picture) and 30 meV (lower picture) spectral intensity below the Fermi energy are shown in FIG. 1. The very low count rate near the Fermi energy at these doping levels posed a challenge to the experiment. It is nevertheless clearly evident, that along the nodal ΓX and ΓY lines the Fermi surface has vanished, while a buildup of spectral weight around the antinodal M-point occured. This electronic structure is decisively different from any hitherto reported one of Bi-cuprates in the underdoped regime. While no spectral weight, dispersion or Fermi surface crossings along the nodal line could be detected, a diffuse but, when compared to the missing intensity along the nodal line, clearly detectable buildup of spectral weight around the antinodal M-point occurred. Interestingly the comparison to a theoretical Fermi surface obtained by cluster perturbation theory for a static, site centered stripe model at 0.10 hole doping, i.e. at the same doping level as in our experiments, gave good qualitative correspondence [5]. The measurements will be continued at elevated temperatures, i.e. above the stripe ordering temperature and for crystals at different doping levels, especially nH = 1/8.

Figure 1. Photoemission intensity map of the ΓXΓYΓ quadrant with the Mpoint in the center. The polarization vector and the intensity scale is shown on the left. The upper picture was obtained by integrating all counts 200 meV below the Fermi energy and the lower picture by integrating 30 meV below the Fermi energy.

166

References:

1) J.M. Tranquada, B.J. Sternlieb, J.D. Axe, Y. Nakamura, S. Uchida, Nature 375, 561 (1995) 2) X.J. Zhou, T. Yoshida, S.A. Kellar, P.V. Bogdanov, E.D. Lu, A. Lanzara, M. Nakamura, T. Noda, T. Kakeshita, H. Eisaki, S. Uchichida, A. Fujimori, Z. Hussain, Z.X. Shen, Phys. Rev. Lett. 86, 5578 (2001) 3) M. Schneider, R.-S. Unger, R. Mitdank, R. Müller, A. Krapf, S. Rogaschewski, H. Dwelk, C. Janowitz, and R. Manzke, Phys. Rev. B 72, 014504 (2005) 4) A. K. Ariffin, C. Janowitz, B. Müller, L. Dudy, P. Sippel, R. Mitdank, H. Dwelk, A. Krapf and R. Manzke, Journal of Physics: Conference Series (in print) (2008) 5) M. G. Zacher, R. Eder, E. Arrigoni, and W. Hanke, Phys. Rev. B 65, 045109 (2002) Acknowledgement: We gratefully acknowledge the assistance of the staff of BESSY and Photon Factory (Japan).

167

Quantitative ASAXS studies of carbon supported RuSex/C fuel cell catalysts for oxygen reduction reaction S. Haas1, A. Hoell1, G. Zehl2, I. Dorbandt2, S. Fiechter2
1

Helmholtz-Zentrum Berlin für Materialien und Energie (HZB) Conrad-Röntgen-Campus Adlershof, Albert-Einstein-Str. 15, D-12489 Berlin 2 Lise-Meitner-Campus Wannsee, Glienicker Str. 100, D-14109 Berlin

1. Introduction Anomalous small-angle X-ray scattering (ASAXS) represents a powerful tool for the evaluation of structural parameters from element sensitive analysis of nano-materials [1, 2]. Recently, based on the use of adequate experimental setups and capable equipment, ASAXS has been proven to provide also useful data on the chemical composition even in the sub-nanometer scale. Heterogeneous precious metal catalysts consist in most instances of nanometer sized metal particles supported on chemically inert carriers. Therefore, they represent very suited model substances for demonstrating the potential of this technique and for further improvement of its skills. Currently, to meet future energy demands, there is an urgent need for higher performing heterogeneous catalysts for fuel cell applications, distinguishing by reduced amounts of nobel metals or, most promising, by replacing Pt with less expensive metals. Accordingly, selenium modified ruthenium nano-particles supported on activated carbon (RuSex/C) have been found to catalyze the oxygen reduction reaction (ORR) in acidic media with high activity and superior selectivity [3]. Therefore, this material is now in focus as methanol tolerant electro-catalyst replacing Pt at the cathode side of Direct Methanol Fuel Cells (DMFC) [4-6]. Nevertheless, a better understanding of this novel catalyst system is believed to be the prerequisite for its further optimization to pave the way towards commercial application. To reveal the structure of such highly complex materials involving constituents ranging from tens of nm down to the sub-nanometre region, comprehensive ASAXS measurements were performed. Using ASAXS one takes advantage of the so-called anomalous or resonant behaviour of the atomic scattering amplitude of an element near its absorption edge to separate its scattering contribution from other elements in the sample. The main targets were to provide a more elaborate structural investigation of novel RuSex/C fuel cell catalysts by using the possibility of separating the scattering contributions of the carbon black carrier and the different metallic constituents of the thereon supported nano-particles by means of ASAXS measurements. Based on these data a structural model of the catalytically active particles was derived. 2. Experimental Highly active RuSex/C catalysts were prepared in a multistep procedure starting from ionic metal precursors [7]. Commercially available carbon black (Black Pearls 2000 from Cabot) was used as support material. In brief, a calculated amount of RuCl3·xH2O was dissolved in water, previously purged with argon. A weighted amount of carbon black was placed inside the quartz tube, heated from room temperature to 950°C while purging with CO2 and held at this temperature for 20 min for activating the carbon support prior to impregnation. After cooling down the activated carbon sample was transferred into the round-bottomed flask with the Ru precursor solution. The solvent was rotary evaporated and the dried powder was treated under forming gas (5%H2/95%N2) at 200°C to form the Ru nano-particles. This Ru/Cintermediate was transferred into an acetonic solution of SeCl4 for selenization. After ultrasonic conditioning the solvent was removed by rotary evaporation and, finally, a reductive treatment under forming gas for 30 min at 800°C was performed. Anomalous small-angle X-ray scattering has been performed on the SAXS beamline 7T-MPW-SAXS at the synchrotron facility BESSY [8]. The contrast variation was performed at different energies near the SeK (ESe = 12658 eV) and Ru-K (ERu = 22117 eV) absorption edges (see table 1). To cover the full accessible q-range two sample-detector distances (800 mm and 3750 mm) have been selected. The scattering were recorded using an area sensitive gas-filled multi-wire proportional counter. These raw data were corrected for the dark current, dead time and sensitivity of the detector as well as the incoming photon-flux, sample transmission, scattering background and geometrical effects like the projection of the detector plane on the sphere with radius equal to the sample-detector distance. To increase the statistics of the scattering intensity the two-dimensional corrected data were circularly averaged over rings with a certain width. In order to have direct control of statistical errors and to take care of possible time dependent variations of the beam conditions the measurements were separated into several short runs and
168

then averaged [9]. The scattering of a reference sample (1 mm glassy carbon) were measured at the beginning of each run to scale the intensities obtained at different energies relative to each other. The norm of the scattering vector was calibrated using the d-spacing of silver behenate as reference. 3. Results In Fig. 1 two scattering curves for the Ru/C sample at energies E6 and E10 are shown together with the obtained scattering curve for the carbon black sample. It can be seen that the scattering curves of the Ru/C sample has a broad shoulder at q ~ 1 nm-1 compared to pure carbon black. A noticeable intensity variation with respect to the energy in this q-domain can be seen. At higher q-values a crossover of the curves occurs, being due to isotropic resonant-raman scattering (RRS) close to the absorption edge.
10
4

10 Ru-K edge: EK = 22117eV EK - 8eV (E10) 10 scattering intensity (a.u.) EK - 442eV (E6)

4

E (eV) E1 E2 E3 E4 E5 E6 E7 E8 E9 E10 12220 12524 12614 12643 12653 21675 21840 22034 22091 22109

ΔE =E - EK (eV) -438 -134 -44 -15 -5 -442 -277 -83 -26 -8
10
0

Ru-K edge: EK = 22117eV EK - 8eV (E10) EK - 442eV (E6)
3

10 scattering intensity (a.u.)

3

carbon black

carbon black

10

2

10

2

10

1

10

1

10 1 10

0

q (nm )

-1

1

q (nm )

-1

10

Table 1: Numeration of the X-ray energies used in the experiment. The EK is the energy of the corresponding absorption edge Se-K for E1-E5 and Ru-K for E6-E10.

Figure 1: Scattering curves of the Ru/C sample at two energies near the Ru-K absorption edge (E6 and E10). The SAXS curve of the carbon black is shown for comparison

Figure 2: Scattering curves of the RuSex/C sample at two energies near the Ru-K absorption edge (E6 and E10). The SAXS curve of the carbon black is shown for comparison.

In Fig. 2 scattering curves for the active catalyst sample RuSex/C at energies E6 and E10 are shown. The scattering seems to be very similar to the scattering from the sample without Se (Fig. 1) but at high qvalues above 7 nm-1 a second weaker shoulder occurs. Therefore, the corresponding structure element seems to be connected with the Se arrangement. Considering the theoretical calculated energy-dependent correction factors f'(E) and f''(E) of the atomic scattering factor, the three energy-independent partial scattering contributions I0(q), I0R(q) and IR(q) can be calculated by solving a set of linear equations [10]

I ( q, E i ) = I 0 ( q ) + 2 f ' ( E i ) I 0 R ( q ) + f ' 2 ( E i ) + f ' ' 2 ( E i ) I R ( q )

{

}

(1)

where the partial scattering contribution IR(q) contains information on the spatial arrangement of the resonant scattering element alone. The I0(q) describing the scattering behaviour if one measures at energies far from any absorption edge of the elements present in the sample. This partial scattering contribution contains information of all structure components. The I0R(q) is the cross term of I0(q) and I0R(q). The IR(q) will represent the partial scattering contribution of the Ru distribution only by solving the system of linear equation (1) for energies close to the Ru-K edge and considering the theoretical values f' and f'' listed in [11] and calculated by the method of Cromer & Liberman [12]. In order to check the reliability of the solution, the calculation was repeated several times using three, four or five different energies. The result using the five energies Ei=E6,...,E10 is shown in Fig. 3. In addition the obtained ASAXS curve at 21675 eV is also shown after subtraction of the isotropic background level which was determined at q values above 10 nm-1. The main results of this analysis were: (i) the behaviour of the non-resonant contribution I0(q) is very similar to the ASAXS curve obtained far the Ru-K absorption edge. (ii) the socalled resonant scattering contribution IR(q) behaves completely different at small and large q-values. In the limit to small q-values the IR(q) curve converges to a constant intensity contrary to the I0(q) curve which shows a strong increase. This indicates the presents of a larger non ruthenium containing structure element. In the other limit to large q-values the IR(q) curve shows a nearly linear decrease in the double logarithmic plot contrary to the I0(q) curve which shows a weak shoulder in this q-domain. Therefore, the causing structure has nearly nothing to do with the Ru present in the sample. Moreover, the shoulder deduced in

169

the I0(q) curve at large q-values has to be considered as an indication for a much smaller mainly ruthenium free structure. Discussing the measurements near the Se-K absorption edge of the Se containing sample RuSex/C, it must be noticed that the sensitivity of the ASAXS-technique to Se is much lower in contrast to Ru. Fig. 4 shows two scattering curves obtained for sample RuSex/C at energies E1 and E5 as well as the SAXS curve for the carbon black. Within in the error no variation of the intensity with respect to the different energies has been ascertained for q-values less than 4 nm-1. Only at larger q-values an increasing intensity with increasing energy has been observed. This energy-behaviour can be explained by a superposition of two effects: (i) the resonant Raman scattering effect caused by the Se-atoms. This effect gives rise to a increase in the intensity while moving the incident energy closer to the Se-K absorption edge; (ii) the anomalous scattering effect of Se-enriched small nano-structures which have a scattering contribution in the larger q-domain.
10
4

10

5

Se-K edge: EK = 12658eV

10

3

EK - 5eV 10
4

(E5)

EK - 438eV (E1) carbon black

10 intensity (a.u.)

2

scattering intensity (a.u.)

10

3

10

1

10

0

10

2

10

-1

I0(q) 10
-2

10

1

I0R(q) IR(q) I(q,E=21.675eV)
10
0

10

-3

1

q (nm )

-1

10

0.1

1 -1 q (nm )

10

Figure 3: Partial scattering contribution for sample RuSex/C calculated using the five energies close to the Ru-K absorption edge. Therefore, the IR(q) represents the scattering of the resonant scattering Ru-atoms. As comparison one measured scattering curve is shown as well.

Figure 4: Scattering curves of the RuSex/C sample at two energies near the Se-K absorption edge (E1 and E5). The SAXS curve of the carbon black is shown for comparison.

The result of the analysis of the energy dependencies of the SAXS curves is that for RuSex/C catalysts a structure model with three independent types of particles (in size and composition) must be assumed. The three contributions are assumed to be connected with the following structures: (i) the carbon black support nano-particles with a mean radius of 15.00 ± 5.70 nm; (ii) the Ru nano-particles with a mean radius of 1.25 ± 0.60 nm; (iii) the selenium aggregates with a mean radius of 0.30 ± 0.02 nm [11]. This structural model is not just supported by previous results of conventional characterisation techniques (e.g. XRD, TEM and EXAFS [13]). In fact, current ASAXS measurements revealed for the first time that Se-modified Ru nanoparticles on RuSex/C catalysts are not completely covered by Se on the particle surface as formerly assumed (core-shell model). The Se rather forms clusters on the Ru surface while the rest of the surface is covered by oxygen. Thus, ASAXS successfully demonstrated its potential as future powerful characterisation technique in the field of heterogeneous catalysis, when using adequate experimental setups. 4. References
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] G. Goerigk, H.-G. Haubold, O. Lyon, and J.-P. Simon, J. Appl. Crystallogr. 36 (2003) 425-429. A. Hoell, F. Bley, A. Wiedenmann, J.P. Simon, A. Mazuelas, and P. Boesecke, Scripta Mater. 44 (2001) 2335-2339. S. Fiechter, I. Dorbandt, P. Bogdanoff, G. Zehl, H. Schulenburg, H. Tributsch, M. Bron, J. Radnik, and M. FieberErdmann, J. Phys. Chem. C 111 (2007) 477-487. M.A. Priestnall, V.P. Kotzeva, D.J. Fish, and E.M. Nilsson, J. Power Sources 106 (2002) 21-30. D.C. Papageorgopoulos, F. Liu, and O. Conrad, Electrochim. Acta 52 (2007) 4982-4986. A.E. Comyns, Focus on Catalysts 2007 (2007) 1. G. Zehl, I. Dorbandt, G. Schmithals, J. Radnik, K. Wippermann, B. Richter, P. Bogdanoff, and S. Fiechter, ECS Trans. 3 (2006) 1261-1270. A. Hoell, I. Zizak, H. Bieder, and L. Mokrani. 2007. S. Haas. 2007. Technical University, Berlin. H.B. Stuhrmann, Advances in Polymer Science 67 (1985) 123-163. S. Haas, A. Hoell, G. Zehl, I. Dorbandt, and S. Fiechter, J. Phys. Chem. C (2009) submitted. D.T. Cromer, and D. Libermann, J. Chem. Phys. 53 (1970) 1891-1898. G. Zehl, G. Schmithals, A. Hoell, S. Haas, C. Hartnig, I. Dorbandt, P. Bogdanoff, and S. Fiechter, Angew. Chem., Int. Ed. Engl. 46 (2007) 7311-7314.

170

Magnetic moments of Fe in oxide-free FePt nanoparticles as a monitor for compositional inhomogeneities
C. Antoniak1,*, A. Trunova1, M. Spasova1, K. Fauth2, M. Farle1, and H. Wende1
Fachbereich Physik and Center for Nanointegration Duisburg-Essen (CeNIDE), Universität Duisburg-Essen, Lotharstr. 1, D–47048 Duisburg 2 Experimentelle Physik IV, Universität Würzburg, Am Hubland, D–97074 Würzburg
1

Introduction – FexPt1−x nanoparticles are currently the object of intense research activities, driven both by fundamental interest and their possible use as new ultra-high density magnetic storage media. For the latter case, nanoparticles with a high magnetocrystalline anisotropy (MCA) are needed to overcome the superparamagnetic limit, i.e. data loss caused by thermally activated magnetisation fluctuations. In addition, the magnetic moments of the nanoparticles should be low to prevent magnetisation switching caused by magnetic dipole interactions. One of the prime candidates for future application is FexPt1−x in the chemically ordered state around the equi-atomic composition due to its high MCA density of about 6×106J/m3[1]. Interestingly, the magnetic moments in FePt nanoparticles in both the chemically disordered and chemically ordered phase are rather small compared to the bulk material [2,3]. The x-ray absorption spectroscopy is a powerful tool to investigate both local structure around the absorbing atoms by the analysis of the extended x-ray fine structure (EXAFS) and magnetic properties, e.g. by measuring the x-ray magnetic circular dichroism (XMCD) as presented here.

The standard Fourier based analysis of EXAFS oscillations of oxide-free Fe0.56Pt0.44 nanoparticles give not only evidence for a lattice expansion with respect to the corresponding bulk material [4], but also a clear local deviation from the averaged composition. A visualisation of this compositional inhomogeneity is given by the Wavelet transformation (WT) method. It may be used to distinguish between different atomic species in an alloy due to their different positions of maximum backscattering amplitude in k-space (Fig. 1). This is possible since in these transformations employing wavelets both localised in real space and k-space, the resolution in k-space is not lost in contrast to a Fourier transformed signal [5]. Fig. 1 shows a contour plot of the difference between wavelet transformed EXAFS of nanoparticles and bulk material measured at the Pt L3 edge at the ID-12 undulator beamline at the ESRF indicating a reduced number of Fe nearest neighbours (nn.) and an enhanced number of Pt nn. around the Pt probe atoms in FePt nanoparticles with respect to the corresponding bulk material. Note, that the slightly different positions in r and the different inclination of the WT data are due the different EXAFS phase shifts.

Fig. 1: Contour plot of the difference between wavelet transformed EXAFS measured at the Pt L3 absorption edge of nanoparticles and bulk material (left), effective backscattering amplitude of Fe (upper right graph) and Pt (lower right graph) and EXAFS phase shift as a function of the photoelectron wave number. In both cases Pt is the absorber atom. The grey line in the coordinates plane at the bottom refers to the dependence of the phase shift on the wave number.

*

e-mail: carolin.antoniak@uni-due.de

171

K edge and the Fe L3,2 absorption edges [6]. After plasma treatment, no absorption was detected at the carbon edge and at the Fe L3,2 absorption edges, pure metallic spectra were obtained. The size distribution of the nanoparticles is lognormal around a mean diameter of 4.4nm and a standard deviation of about 0.14. In order to prevent the particles from agglomeration during the plasma treatment, less than one monolayer of particles was brought onto the substrate.

FIG. 2: Picture of the plasma chamber with ignited hydrogen plasma; Inset: schematic drawing of the FePt nanoparticles before and after plasma treatment, e.g. with and without Fe oxides and organic ligands

In this work, the influence of an inhomogeneous composition within FePt nanoparticles with a diameter around 4.4nm on the magnetic properties, especially on the element-specific magnetic is discussed. Since the magnetic moments at the Pt sites are almost independent of the local composition and distortions of the crystal lattice [2,6], we focussed on the magnetic moments at the Fe sites which are a sensitive monitor for structural changes. As reference, the magnetic moments of chemically disordered FexPt1−x alloys were probed as a function of Fe content by XMCD of different samples, i.e. bulk material and 50 nm thick films. In addition, band structure calculations were performed using the spin-polarized relativistic Korringa-Kohn-Rostoker (SPR-KKR) method [7].

Results and Discussion – The deviation of the local composition from the averaged value may strongly influence the magnetic properties of the FexPt1−x nanoparticles, e.g. the element-specific magnetic moments. In order to rate these values experimentally found in nanoparticles, the corresponding magnetic moments of 50nm thick films were measured for different compositions. In addition, SPR-KKR band structure calculations were done for different FexPt1−x bulk systems. We found both experimentally and theoretically that the magnetic moment at the Fe sites is decreasing with increasing Fe content [5]. This decrease can be essentially explained by the decrease of the lattice constant with increasing Fe content which usually yields smaller magnetic moments. Due to an enhancement in the hybridisation of Fe d-states, the decrease of the magnetic moment at the Fe sites is connected to a broadening of the density of states (DOS).

Experimental – The x-ray absorption near-edge structure (XANES) and its associated XMCD was measured in the soft x-ray regime at the Fe L3,2 absorption edges at the PM3 bending magnet beamline at BESSY II in total electron yield (TEY) mode. The measurements were performed at T = 15K in magnetic fields of μ0Hext = ±2.8T at fixed photon helicity in an energy range of 680eV  E  790eV. After each scan, either the magnetic field or the circular polarisation was reversed. Induced magnetic moments at the Pt sites were evidenced in the soft x-ray regime at the N6,7 absorption edges of Pt in the energy range 65eV  E  85eV. A portable plasma chamber [8] was attached to the experimental endstations for an in situ cleaning of the samples (Fig. 2). An exposure to a hydrogen plasma at a pressure of p = 5Pa for 20min was found to be sufficient for the complete reduction of Fe oxides and organic ligands, i.e. oleic acid and oleyl amine surrounding the wet-chemically synthesised particles [1] in the as-prepared state. The efficiency of the plasma cleaning procedure was proven by XANES measurements at the carbon

FIG 3: Spin and angular momentum resolved density of states at the Fe sites calculated for three different compositions of FexPt1-x bulk alloys using the Munich SPR-KKR package

172

XMCD shown in the lower panel of Fig. 4 that the magnetic moments at the Fe sites are reduced by 20-30% with respect to the values of the corresponding bulk material with the same averaged composition. This can be explained by the fact, that the magnetic moment strongly depends on the local environment, e.g. the number of Fe nearest neighbours which is not reflected by an averaged composition in the case of the nanoparticles examined here since the Fe atoms were found to be located in Fe-rich environments as summarised in the introduction.

Conclusion – By the analysis of absorption spectra in the EXAFS regime and XMCD measured in the XANES regime, oxide-free FePt nanoparticles were structurally and magnetically characterised. By the combination of the different x-ray absorption techniques, the reduced magnetic moments obtained at the Fe sites can be explained by an inhomogeneous composition within the particles.

Fig. 4: XANES and XMCD of FePt nanoparticles measured at the Pt N6,7 absorption edges (upper

panel) and Fe L3,2 absorption edges (lower panel) The calculated spin and angular momentum resolved DOS at the Fe sites is shown in Fig. 3 for three different compositions, Fe0.32Pt0.68, Fe0.58Pt0.42, and Fe0.68Pt0.32. The relatively narrow d-bands are strongly exchange split. A clear broadening of the DOS at the Fe sites for the Fe-rich alloy is visible and the difference between the majority and the minority band becomes smaller indicating decreasing magnetic moments in agreement to the experimental values obtained from XMCD analysis of FexPt1−x bulk-like films with different compositions [5]. One example of XANES and the corresponding XMCD of FePt nanoparticles are shown in Fig. 4. The XMCD at the Pt N6,7 absorption edges clearly indicates some induced magnetism at the Pt sites. Since there is no standard analysis method for spectra obtained at these edges, only qualitative conclusions about the ratio of orbital-and-spin magnetic moments can be made [9]. In regard to the magnetism at the Fe sites, we found from the

Acknowledgment – We would like to thank S. Sun (Brown U.) for providing nanoparticles, J.-U. Thiele (Hitachi) for thin film preparation. For help in the XMCD measurements, U. Wiedwald (U. Ulm), H.-G. Boyen (U. Hasselt), N. Friedenberger, and S. Stienen (U. Duisburg-Essen) are acknowledged as well as F. Wilhelm, A. Rogalev, P. Voisin and S. Feite (ESRF). We thank the BESSY II staff, especially T. Kachel and H. Pfau for their kind support and M. Košuth, J. Minár, and H. Ebert (LMU München) for help using the Munich SPR-KKR package. This work was financially supported by the DFG (SFB445), the BMBF (05 ES3XBA/5), the ESRF, and the EU (MRTN-CT-2004-0055667).

[1] S. Sun et al. Science 289, 1989 (2000) [2] O. Dmitrieva et al., Phys. Rev. B 76, 064414 (2007) [3] C. Antoniak et al. , Phys. Rev. Lett. 97, 117201 (2006) [4] C. Antoniak et al., Phys. Rev. B 78, 0401406 R (2008) [5] C. Antoniak et al., submitted (2009) [6] C. Antoniak and M. Farle, Mod. Phys. Lett. B 21, 1111 (2007) [7] H. Ebert et al., The Munich SPR-KKR package version 3.6 [8] H.-G. Boyen et al., Adv. Mater. 17, 574 (2005) [9] T. Shishidou et al., Phys. Rev. B 55, 3749 (1997)

173

DIELECTRIC PROPERTIES OF (NH4)2SO4 CRYSTALS IN THE RANGE OF ELECTRONIC EXCITATIONS B. Andriyevsky a), C. Cobet b), A. Patryn a), N. Esser b)
a) Koszalin University of Technology, Śniadeckich Str. 2, PL-75-453, Koszalin, Poland b) ISAS – Institute for Analytical Sciences, Department Berlin, Albert-Einstein-Str. 9, 12489 Berlin, Germany The funding source and grants numbers: EU, R II 3.CT-2004-506008, BESSY-2008_1_71033 and BESSY2008_2_80159

Ammonium sulfate, (NH4)2SO4, is a dielectric crystal with a first-order ferroelectric phase transition (PT) at the temperature TC = 223 K [1]. The crystal exhibits unusually large relative change in the dielectric constant at TC [1], a very low value (≈15 K) of the Curie-Weiss constant [2, 3] and a very large spontaneous strain [1]. The main aim of the study was the experimental investigation of the dielectric functions of a (NH4)2SO4 crystal in the photon energy range of electronic excitations, 4 – 9.5 eV, and its temperature changes in the range of 170 – 295 K. The experimental results obtained in this study could be useful for better understanding of peculiarities related to the chemical bonding and phase transition in the crystal. Measurements of the pseudo-dielectric functions <ε1>(E) and <ε2>(E) (ε = ε1 + iε2) of (NH4)2SO4 crystals were performed by spectroscopic ellipsometry [4, 5] using synchrotron radiation of the Berlin Electron Storage ring for Synchrotron radiation BESSY II in the range of 4.0 – 9.5 eV with a resolution of δE < 0.1 eV. The angle of incidence was ~68°, while the polarization of the incident beam was chosen ~20° tilted with respect to the plane of incidence during the measurements. A MgF2 polarizer and rotating analyzer ensured more than 99.998% degree of polarization. The ellipsometric measurements were conducted for mechanically polished (NH4)2SO4 crystals of three cuts perpendicular to a-, b-, and c- orthogonal axes of the unit cell and the corresponding dielectric functions <ε1(E)> and <ε2(E)> were subsequently calculated with an isotropic two layer model. The susceptibility <χ2>(E) [χ = χ1 + iχ2 = (ε1 – 1) + iε2] shows an increasing quasi monotonously behavior with a maximum at 9.37 eV (for the temperature 25 °C) and at 9.50 eV (for the temperature -98 °C) (Fig. 1). This looks very similar to the corresponding theoretical dependences of RbNH4SO4 crystals obtained from the first principles calculations using the CASTEP code [6]. An analysis of the density of electronic states of RbNH4SO4 crystals indicates that the big maximum of <χ2>(E) at 9.37 eV in case of (NH4)2SO4 mainly corresponds to the excitation of oxygen p-electrons. Distinct and big anomalies of the temperature dependences of pseudo-susceptibilities <χ1>(T) and <χ2>(T), and reflectance intensity IR(T) have been obtained at the temperature of discontinuous ferroelectric phase transition for the c-cut of (NH4)2SO4 for the superior light polarization E||a studied (Fig. 2, 3). Much smaller anomalies were seen for the samples of a- and bcuts. One can distinguish here two types of anomalies: (1) broad band with broad and small extremum in <χ1>(T) and <χ2>(T) near TC, and (2) narrow peak-like dependences of <χ1>(T) and <χ2>(T) with extremum position 5 °C lower than the corresponding position of the broad extremum. The maximum of the broad anomaly in <χ1>(T) corresponds to the minimum of the narrow and big anomaly in <χ1>(T), and vice versa, the minimum of the broad anomaly in <χ2>(T) corresponds to the maximum of the narrow and big anomaly in <χ2>(T) (Fig. 2, 3).

174

4
<χ1> 25 C
<χ1>, <χ2>, |<χ>|
o

3.0 2.5
|<χ>|

<χ1>, <χ2>

3

<χ1> -98 C <χ2> 25 C
o

o

|<χ>|

2

<χ2> -98 C

o

2.0 1.5

2.76

|<χ>| <χ1> <χ2>

2.74

1
1.0
-100 -80 -60 T/C
o

-40

0 5 6 7 E /eV 8 9

-55

-50

T/C

o

-45

Fig. 1. Spectra of the real and imaginary parts of the susceptibility <χ1>(E) and <χ2>(E) of c-cut of (NH4)2SO4 crystal for the superior light polarization E||a at the temperatures 25 °C and -98 °C.

Fig. 2. Temperature dependences of the real (<χ1>), imaginary (<χ2>) part and modulus (|<χ>|) of the susceptibility <χ> of (NH4)2SO4 crystal during heating at E = 8.5 eV. The insertion shows the temperature dependence of the modulus of the susceptibility |<χ>|(T) in a larger temperature range.

Therefore, two different temperature dependent processes can be suggested for (NH4)2SO4 crystals. The first process is a discontinuous phase transition at TC(1) taking place in the narrow temperature range of approximately 1 – 2 °C. The sharp maximum of <χ2> and the minimum of <χ1> is associated with this process (Fig. 2). The second, continuous process is associated with the slower variation in <χ1>(T) and <χ2>(T) which creates a maximum in <χ1> and a minimum in <χ2> at the temperature TC(2), approximately 5 °C higher than TC(1) (Fig. 3). The presence of these two processes agrees with results of dielectric constant measurements on (NH4)2SO4 for frequences 10 – 102 kHz [7]. In this study, a small anomaly in the dielectric constant ε was found between 6 and 12 K below the well-known ferroelectric transition at 49.5 °C. It is proposed that this new anomaly is due to the appearance of a spontaneous polarization, which is produced by the secondary order parameter and is in an anti-parallel direction to the spontaneous polarization due to the primary order parameter, which produces the well-known ferroelectric transition. According to this model, ammonium sulfate is ferroelectric just below the transition at -49.5 °C, but it gradually changes into a weak ferrielectric material below this new anomaly temperature. An interesting result has been obtained when analyzing the temperature dependence of the light intensity IR(T) reflected from the sample and the temperature dependence of the normal incidence reflection coefficient R(T) (Fig. 4). The normal incidence reflection coefficient R has been calculated according to the known relationship [8],
2 2 (n − 1) 2 + k 2 (ε12 + ε 2 )1/ 2 + ε1 (ε12 + ε 2 )1/ 2 − ε1 , , . (1) n= k= (n + 1) 2 + k 2 2 2 Taking into account the great spontaneous deformations in the crystal and the piezo(elasto)optical effect mentioned, the big and abrupt decrease of the reflectance intensity IR(T) observed for c-cut (Fig. 4) can be explained as a result of light scattering on a domain-like structure grid. This structure is generated by the non homogeneity of refractive index caused by the abrupt increase of spontaneous deformation.

R=

175

0.13

2.65 <χ1>

<χ1>

0.9

0.006
0.24 tgΨ 0.8

IR /arb. un.

0.12

0.20

cos∆

<χ2>

2.60

0.6 0.16 -50 T/C
o

-40

<χ2>

R

0.003
0.8

0.11

2.55 -100 -80 -60 T/C
o

-40

0.000 -100 -80 -60 T/C
o

-40

Fig. 3. Temperature dependences of the real <χ1>(T) and imaginary <χ2>(T) part of the susceptibility of (NH4)2SO4 crystal during heating at E = 8.5 eV in more fine vertical scale than in Fig. 2. The insertion shows the temperature dependences of the standard ellipsometric parameters tgΨ(T) and cos∆(T).

Fig. 4. Temperature dependences of the intensity of reflected light IR(T) and normal incidence reflectance R(T) calculated from the dielectric functions <ε1>(T) and <ε2>(T) of (NH4)2SO4 for the superior light polarization E||a measured at E = 8.5 eV during heating.

Taking into account that the photon energy E = 8.5 eV corresponds to the excitation of pelectrons of oxygen in (NH4)2SO4 crystal the temperature changes described above can be attributed mainly to the SO4 group. Finally, one can state that the module of complex susceptibility |<χ>| of (NH4)2SO4 crystal in the range of oxygen p-electrons excitation display maximum at the temperature T ≈ 223 K of ferroelectric phase transition. For the narrow temperature range ∆T ≈ 1 – 2 K near the main maximum of |<χ>|(T), the big relative increase ∆χ2/χ2 ≈ 1.8 of the imaginary part of the susceptibility <χ2> takes place together with simultaneous relative decrease ∆χ1/χ1 ≈ -0.25 of the real part of the susceptibility <χ1> (see Fig. 2). This result agrees qualitatively with conclusion of [9] that a considerable part of spontaneous polarization of (NH4)2SO4 crystal arises from the SO4 groups. According to [9] the main driving interaction of the phase transition has its origin in S-O bonds of SO42- ion which triggers the transition by getting more distorted structure of lower symmetry; the NH4+ ions simply follow an appropriate change. This is probably the reason that the studies related only with NH4+ ions indicate that the phase transition has the characteristics of second order [10, 11]. Taking into account the sharp spontaneous deformations at phase transition of (NH4)2SO4, which even destroy the crystal, it would be interesting to perform first principal calculations of the complex susceptibility for different unit cell dimensions and arrangements of constituent atoms and to compare results obtained with present experiment.
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] S. Hoshino, K. Vedam, Y. Okaya, and R. Pepinsky, Phys. Rev. 112 (1958) 405. H. Ohshima and E. J. Nakamura, Phys. Chem. Solids 27 (1966) 481. H.-G. Unruh, Solid State Commun. 8 (1970) 1951. R. M. A. Azzam and N. B. Bashara, Ellipsometry and polarized light, Amsterdam: North-Holland Personal Library, 1987. T. Wethkamp, K. Wilmers, N. Esser, W. Richter, O. Ambacher, H. Angerer, G. Junqk, R. L. Johnson, and M. Cardona, Thin Solid Films 313-314 (1998) 745. B. Andriyevsky, W. Ciepluch-Trojanek, V. Stadnyk, M. Tuzyak, M. Romanyuk, and V. Kurlyak, J. Phys. Chem. Solids 68 (2007) 1892. A. Yoshihara, T. Fujimura, and K. I. Kamiyoshi, Phys. stat. sol. (a) 34 (1976) 369. H. R. Philipp and H. Ehrenreich, Phys. Rev. 129 (1963) 1550. Y. S. Jain and H. D. Bist, Phys. Stat. Sol. (b) 62 (1974) 295. D. E. O’Reilly and T. J. Tsang, Chem. Phys. 46 (1967) 1291. B. H. Torrie, C. C. Lin, O. S. Binbrek, and A. Anderson, J. Phys. Chem. Solids 33 (1972) 697.

176

Residual Stress Profiles in Friction Stir Welds of 2024Al Alloy and 2124Al-25vol%SiC Composite
Ferreira-Barragáns S.[1], Gesto D.[2], Rey P.[2], Fernández R.[1], González-Doncel G.[1] [1] Dept. of Physical Metallurgy, National Centre for Metallurgical Research, CENIM-CSIC, Av. Gregorio del Amo 8, 28040 Madrid, SPAIN [2] AIMEN Technology Centre, Relva 27A-Torneiros, Porriño, 36410 Pontevedra, SPAIN

1. Introduction Metal Matrix Composites (MMCs) are known to have better mechanical properties than the corresponding unreinforced metallic alloys. However, their wide application as structural materials needs proper development of suitable joining processes. The application of the solid state welding technique known as Friction Stir Welding (FSW) to MMCs seems very attractive. On the other hand, the presence of a residual stress (RS) in welded components may have a significant influence of components performance. Thereby, knowing its magnitude and distribution is very important [1]. In this work, the sub-surface RS profile across the butt weld in a 2124Al-25vol%SiC composite plate obtained by FSW, has been studied from measurements conducted by synchrotron radiation diffraction. For comparative purposes, the RS has been also calculated on a butt weld of 2024Al alloy obtained under similar joining conditions.

2. Experimental details 2.1 Welded samples The welds were performed on 15 mm thick plates of 2024Al and 2124Al-25vol%SiCp in T6 condition. Welding process in two passes and butt joint configuration without edge preparation was carried out. A fix pin tool with threaded pin with three grind flats at 120º was used. This pin was fabricated from H13 Steel (Q&T, 49 HRc). The materials investigated were welded at AIMEN Technology Centre (Pontevedra, Spain). The joints parameters are reported in Table 1.
Material Rotation rate (rpm) 400 Counter- clockwise 300 Counter-clockwise Welding Speed (mm/min) 100 75 Force (kN)

2024Al-T6 2124Al-25vol%SiC-T6

Position Control (23 kN) Position Control (8-9 kN)

Table 1: Joints parameters.

2.2 Residual stress measurements The residual stress distribution across the weld has been measured by energy dispersive diffraction using X-ray synchrotron radiation on the beam line EDDI at BESSY, Berlin, Germany, which operate in the range 10-150 keV. The experimental set up is shown in Figure

177

1(a). The measurements were carried out in reflection mode. The use of energy dispersive allows detecting many diffraction peaks. At present only the Al-(311) and SiC-(311) reflections have been analyzed. This is a very suitable peak since it shows low plastic anisotropy and has a fairly linear relationship between stress and lattice strain. An angle 2θ = 8º was used. The incoming beam was defined by slits of 1 mm height and 1 mm width, while the diffracted beam size was adjusted by slit of 30 μm. (gauge volume: 1 mm × 1 mm × 0.03 mm). Both the longitudinal (along the weld) and transverse (across the weld) RS components were calculated. The biaxial sen2 ψ method was used [2]. The strain state was calculated using the conventional relationship

ε=

d − d0 d0

(1)

where d is the lattice spacing measured at different locations of the weld, and d0 is the unstrained lattice spacing measured at equivalent points in a stress-free “comb” sample, Figure 1(b). In the welded sample, it is very important to measure the point to point variation of d0, since this value can vary due to changes in chemical or structural composition resulting from the welding process [3]. The stress state was calculated from the two measured strain components under the assumption that there is not an out of plane component in the stress tensor (σN = 0), i.e., a biaxial stress state is assumed [2]. Hence;

σL =

E ⎛ ν ⎞ εT ⎟ ⎜ε L + 1 +ν ⎝ 1 − 2ν ⎠

(2)

where εL and εT are the values in the weld longitudinal and transverse directions, E is the Young’s Modulus and υ Poisson’s ratio. σT is obtained by exchanging the two strain parameter in Eq. (2).

(a)

(b)

Figure 1: (a) Experimental set up and welded plate, and (b) comb sample.

3. Results and Conclusions

Figure 2 shows the variation of lattice parameter across the stress-free comb samples in both alloy and composite. It can be seen that the distribution reveals structural changes as a consequence of the FSW process. The RS in the materials, Figure 3, presents the typical “M” distribution across the welds [4]. The RS state is more isotropic than in the alloy. In the latter one, the longitudinal component is clearly higher that the transverse one. On the other hand, the RS in the composite is higher than in the alloy. This should be attributed to the additional contribution of the micro-RS associated to the presence of the reinforcement in the composite material.
178

1.2170 21214Al-25vol% SiC 2024Al 1.2165

1.2160

d0(Å)

1.2155

1.2150

1.2145 -24-21-18-15-12 -9 -6 -3 0 3 6 9 12 15 18 21 24

Distance from the welding center (mm)

Figure 2: Variation of 311 lattice spacing across the welds in the comb samples in the composite material and the aluminium alloy.
100 80 60 Retreating side 40 20 0 -20 -40 -60 -80 -100 -120 weld -140 zone -160 -25 -20 -15 -10 -5 0
100 80 60 Retreating side 40 20 0 -20 -40 -60 -80 -100 -120 -140 -160 -25 -20 -15 -10

2124Al-25vol%SiC Advancing side

2024Al Alloy Advancing side

Residual Stress (MPa)

Residual Stress (MPa)

Longitudinal Transverse

weld zone

Longitudinal Transverse 5 10 15 20 25

5

10

15

20

25

-5

0

Distance from the center of the weld (mm)

Distance from the center of the weld (mm)

(a)

(b)

Figure 3: Total residual stress distribution across the welds. (a) Composite (b) Alloy.

4. Acknowledgements

Projects MAT-05-0527 and PTR95-1007-OP from MICINN, Spain and support from BESSY under contract nº RII 3CT-2004- 506008
5. References

[1] M.N. James, D.J. Huges, Z. Chen, H. Lombard, D.G. Hattingh, D. Asquith, J.R. Yates, P.J. Webster, Engineering Failure Analysis 14 (2007) 384. [2] I. C. Noyan and J. B. Cohen. “Residual Stress: Measurements by Diffraction”. Springer Verlag, 1987. [3] S. Pratihar, V. Stelmukh, M.T. Hutchings, M.E. Fitzpatrick, U. Stuhr, L. Edwards. Materials Science and Engineering A 437 (2006) 46. [4] R.S. Mishra, Z.Y. Ma, Material Science and Engineering R 50 (2005) 1-78.

179

Imaging of magnetization dynamics in Fe19Ni81/Cu/Co trilayer microstructures
J. Kurde1, J. Miguel1 , D. Bayer2, J. Sánchez-Barriga3, A. Rzhevskiy3, F. Kronast3, M. Aeschlimann2, H. A. Dürr3, W. Kuch1

1 2

Institut für Experimentalphysik, Freie Universität Berlin, Arnimallee 14, D-14195 Berlin, Germany
3

Technische Universität Kaiserslautern, Erwin Schrödinger Str. 46, D-67663 Kaiserslautern, Germany BESSY GmbH, Albert Einstein Str. 15, D-12489 Berlin, Germany (BMBF Nr. 05 KS7 KE2)

Modern recording devices contain a stack of two thin magnetic layers separated by a nonmagnetic spacer layer. Their fastest operation can be achieved by applying a magnetic pulse, which may induce the magnetization reversal of one of the layers by precession. With the aim of studying this magnetization reversal process we have imaged Fe81Ni19/Cu/Co trilayer microstructures with sizes in the range of 5x5 to 10x20 µm². We have investigated the influence of i) the thickness of the Co layer, ii) a static, external field, and iii) a magnetic field pulse in the ps time scale. In the following we will refer to Fe81Ni19 as permalloy (Py). The magnetic structures were magnetron-sputtered in Ar+ atmosphere at MAGSSY. Sizes range from 5x5 to 10x20 μm² and film thicknesses are Py (4 nm) / Cu (t Cu) / Co (15 nm), with tCu = 1.5, 2.0, 2.5 and 3.0 nm. During growth a static magnetic field of ~100 mT is applied to induce a uniaxial anisotropy. To image the magnetic domains, photoelectron emission microscopy (PEEM) is used combined with x-ray magnetic circular dichroism (XMCD). These two techniques together provide lateral resolution and element-sensitive magnetic contrast. The latter is essential for studies of systems consisting of different materials. The static images were acquired at the UE49-PGMa, a microfocus beamline with an Elmitec SPPEEM (resolution ~50 nm). A static magnetic field was applied with a special sample holder equipped with magnetic coils. The time-resolved images were acquired at the UE56/2-PGM1 beamline with a Focus PEEM (resolution ~400 nm). To apply a magnetic field pulse, the structures were deposited onto a Au stripline, which is connected to a photoconductive switch. By focusing a fs laser onto the switch, a current pulse through the stripline is generated and thereby a magnetic pulse in the vicinity of the structures [1]. Synchronization of the magnetic pulse (pump) with the x-ray

Fig. 1 a) Domain patterns in Py/Cu/Co trilayer microstructures with different Cu thicknesses. b) Line profile of a domain wall in a structure with tCu = 2.0 nm.

180

pulses of the BESSY single bunch operation mode (probe) realizes a stroboscopic pump probe experiment. To gain detailed knowledge about the coupling behavior of the two magnetic layers through the non-magnetic spacer, the domain patterns in Py/Cu/Co trilayer microstructures with different Cu thicknesses were imaged (Fig. 1a). The magnetic domains of the two ferromagnetic layers are coupled parallel, via exchange interaction and orange-peel coupling for all measured Cu thicknesses. For increasing Cu thicknesses, an antiparallel alignment in the domain walls (DW) of the two magnetic layers becomes favorable due to stray fields: At the DWs the dipolar coupling competes with exchange and orange-peel coupling. Similar behavior had been found previously in 180° and head-on walls [2]. Here this effect is also observed at 90° walls: The magnetization inside a wall in the Py layer first turns opposite to the Co layer before it turns back. Detailed statistics of the DW widths at different Cu thicknesses in the two magnetic layers as shown in Fig. 1b will reveal more information about the coupling behavior. Stray fields originating from ripples in the Co layer also lead to strong irregularities in the magnetization of the Py layer for tCu > 2.0 nm.

Fig. 2 Influence of an external field on a 5x15 µm Py/Cu/Co structure with tCu = 2.5 nm.

The strength of the antiparallel coupling inside the walls can be investigated with an external field. A trilayer structure with tCu = 2.5 nm and uniaxial anisotropy parallel to the long side (tilted by about 15°) was imaged after applying an external magnetic field along the hard axis (Fig. 2). When reversing the field, domains are switched (domain pattern changed slightly). While switching, the walls in the Py layer keep their antiparallel alignment to the Co layer, hence they are also aligned antiparallel to the applied field. In an external field of ~7.5 mT the walls start to turn partly along the field. Finally, at ~15 mT they are almost completely switched. Even higher fields are necessary to saturate the magnetization along the hard axis of the structure. Another important aspect is the dynamics of the magnetization in the ps time range. Consequently, a trilayer structure with tCu = 2.0 nm was imaged while applying a magnetic field pulse of ~3 mT amplitude and ~200 ps FWHM perpendicular to the long edge of the structure. Pump-probe delay times with respect to the magnetic field pulse from -100 to 2000 ps in steps of 50 ps where imaged (Fig. 3a). To analyze the magnetization dynamics, the XMCD contrast

Fig. 3 Time dependence of a 5x15 µm Py/Cu/Co structure with tCu = 2.0 nm. a) Magnetic domains at different delay times. b) Magnetic field pulse (line) and XMCD contrast integrated over two areas on the structure as a function of delay time (squares).

181

of two areas on the structure has been integrated and displayed as a function of delay time (Fig. 3b). The magnetization is mainly following the external magnetic field. In addition a longterm change in the Py magnetization is visible due to domain wall motion in the magnetically harder Co layer. These preliminary results are being contrasted with micromagnetic calculations (OOMMF). In summary, the following facts have been observed: An antiparallel coupling in the domain walls of the magnetic layers for spacer thicknesses tCu > 1.5 nm is found, which can be explained by a dipolar interaction. The switching behavior of a trilayer structure with tCu = 2.5 nm is studied in detail, revealing a coupling field of ~15 mT across the walls. First time-resolved measurements of a 5x15 µm² trilayer structure show a repeatable reaction of the Py layer on the magnetic field pulse. For future time-resolved measurements the setup of a new laser system at the UE49-PGMa beamline is in progress. This will combine the high lateral resolution of the Elmitec SP-PEEM with time-resolved pump probe experiments in the ps regime. We would like to thank the Nano+Bio Center Kaiserslautern and F. Radu (BESSY) for their help in the sample preparation; W. Mahler and B. Zada for their assistance during experiments. The financial support of this work by Bundesministerium für Bildung und Forschung (05 KS7 KE2) is gratefully acknowledged.

[1]

J. Miguel, M. Bernien, D. Bayer, J. Sánchez-Barriga, F. Kronast, M. Aeschlimann, H. A. Dürr, and W. Kuch, Rev. Sci. Instrum. 79, 033702 (2008) J. Vogel, S. Cherifi, S. Pizzini, F. Romanens, J. Camarero, F. Petroff, S. Heun and A. Locatelli, J. Phys.: Condens. Matter 19, 476204 (2007)

[2]

182

High-resolution XPS study of Fluorinated Carbon Nanotubes
Carla Bittencourt University of Mons-Hainaut, Parc Initialis, Av. Nicolas Copernic 1, Mons 7000, Belgium Xiaoxing Ke and Gustaaf Van Tendeloo University of Antwerp, EMAT, Groenenborgerlaan 171, B2020 Antwerpen, Belgium Alexandre Felten and Jacques Ghijsen University of Namur (FUNDP), LISE, 61 rue de Bruxelles, Namur 5000, Namur, Belgium 1. Introduction Fluorination is a promising method for expanding the use of carbon nanotubes (CNTs) to several areas where the tailoring of properties such as wettability, adhesion, chemical stability, permeation, electrical conductivity, and biocompatibility are important 1 . Due to fluorine’s high electronegativity, fluorination drastically modifies the nanotube surface chemistry, changing the electronic properties depending on the degree of fluorination and the specific addition pattern of the fluorine atoms on the nanotube surface2,3. Fluorination in fluorine gas generally results in a high reaction ratio and a deep penetration into carbon-related materials which prevents a good control of sidewall fluorination1. In this context, plasma fluorination has the potential to limit the fluorination area to the external surface of the nanotubes if performed at room temperature; in addition it requires a reaction time orders of magnitude shorter than other fluorination methods.4,5 In this work, CF4 plasma functionalization of multiwall carbon nanotubes are investigated using high-resolution x-ray photoelectron spectroscopy. 2. Method Samples were prepared using commercial MWCNT powder synthesized via catalytic chemical vapour deposition (CCVD)6. The functionalization was performed in a homemade chamber using inductive coupled plasma at the RF frequency of 13.56 MHz.4 A controlled flow of CF4 was introduced inside the chamber; the treatment was performed at gas pressure of 0.1 Torr, using 10 W. The changes induced in the CNT electronic states due to the grafting of fluorine atoms by the rf-plasma treatment were investigated by XPS. X-ray photoemission experiments were performed at UE56 beam line BESSY II (Berlin) using the Mustang end-station. 7 Photon energies were selected so that all spectra were recorded at similar kinetic energies corresponding to high surface sensitivity, namely 400 eV for recording C 1s, 800 eV for F 1s, and 110 eV for valence bands. The selected photon energy for valence band studies, together with a favourable cross-section ratio8, will lead to a higher contribution of the fluorine states on the valence band spectra. The Au 4f7/2 peak at 84.0 eV binding energy, recorded on a reference sample, was used for calibration of the binding energy scale. Results Table 1 presents the parameters used for the plasma treatment and the resulting F/C ratio evaluated by XPS analysis. As mentioned before, the photon excitation energy was chosen to generate photoelectrons with kinetic energy of the order of 100 eV, for which the inelastic electron mean free path was reported to be ~ 5 Å 9 . Therefore, considering the distance between the CNT walls (~ 3.45 Å); the outer wall contributes 50% of the intensity of the carbon peak. Thus the F/C ratio as determined by XPS includes contribution of photoelectrons belonging to the two outer walls, meaning that the actual fluorine/carbon ratio of the surface will be twice the measured F/C ratio.

183

Sample 1 2 3 4 5 6 7

Treatment time (s) 2 5 15 30 45 60 300

F/C (XPS) 0.25 0.22 0.19 0.16 0.17 0.16 0.12

F/C ratio coverage of the outer wall 0.50 0.44 0.38 0.32 0.34 0.32 0.24

Table 1 Parameters used for the plasma treatment of MWCNTs, together with the F/C ratio evaluated by XPS and the actual surface coverage on the outer wall. The F/C ratio as determined by XPS includes contribution of photoelectrons belonging to the two outer walls, so the quoted fluorine/carbon ratio of the surface is twice the measured F/C ratio.

Chemical modification by the plasma treatment is revealed by the appearance of new structures contributing to the C 1s line shape at higher binding energies. These structures are generated by photoelectrons emitted from carbon atoms participating in C-F bonding or from carbon atoms neighbour to a carbon atom bound to a fluorine atom: Due to its high electronegativity, fluorine has a very strong effect on the electronic screening of the element to which it is bound. In addition, when a C atom not bound to fluorine is first neighbour of another carbon atom bound to one or several fluorine atoms, an inductive effect arises.10 The main peak in the C 1s spectrum of the plasma functionalized samples is generated by photoelectrons emitted from the carbon atoms that do not interact directly with fluorine atoms. From the reported results1, we can suggest the following attribution for the peaks observed in Figure 1: C2F at 286.4 eV, C3F at 287.5 eV, C4F at 288.8 eV, CF at 291.0 eV, and CF2 at 293.0 eV. b) The maximum F/C ratio of 0.25 a) (equivalent to C2F surface coverage) is obtained after 2s of functionalization; for increasing functionalization time this ratio decreases (table 1). Figure 1 shows that for increasing C2F functionalization time the relative C3F intensity of the different components C4F varies. This trend suggests that certain CF bond configurations are more stable CF2 and/or are less influenced by the interaction with the CF4 plasma. We 288 292 believe that the reduction in the F/C 280 284 Binding energy (eV) 296 300 280 284 288 292 296 300 Binding Energy (eV) ratio for increasing functionalization Figure 1 : C 1s XPS spectrum recorded using hv = 400 eV on time is mainly due to the chemical CF4 plasma functionalized for a) 60 s and b) 300 s. interaction between the CF4 plasma and the less stable bond configuration — fluorine plasmas are characterized by high-rate chemical interaction.11 As the ions in the plasma interacts with carbon atoms in the graphite layer of MWCNTs, fluorine atoms can either bond to carbon atoms or cut C-C bonds. Because valence electrons are involved in bond formation, subtle differences in surface chemistry may be observed by studying the valence band region. The valence band structure of the pristine CNTs is characterized by features appearing close to 3.5 eV associated with photoelectrons emitted from the 2p-π band, extending from 5.5 to 8.0 eV associated with 2p-σ states and the mixed 2s-2p hybridized states at 13.6 eV. The σ-π hybridization resulting from the formation of the CNTs gives rise to the intensity at 11.5 eV. 12 The CNT spectrum is dominated by an intense C 2s region around 18 eV and a broad C 2p region in the range 9-15 eV.
Intensity (arb. units)
Intensity (arb. units)

184

The impact of the CF4 plasma treatment on the CNT electronic valence states can be seen in the spectrum recorded on the sample treated for 2 seconds (Figure 3). Notably the MWCNT electronic states are significantly attenuated. Photoelectrons emitted from the valence states of the plasma treated samples generate new structures in the binding energy range of 5 to 16 eV. The broad feature appearing close to 8.3 eV was reported to be generated by photoelectrons emitted from anti-bonding orbitals of the C-F bonds with contribution of F 2p states at high binding energy. 13 The main contribution of the photoelectrons emitted from the F 2p-like states appears at 10 eV. The structure close to 15 eV originates from bonding orbitals of C 2s-F bonds.

Intensity (arb. units)
30

e)
d)

c) b)
a)

25

20

15

10

5

0

-5

Bindining Energy (eV)

Figure 2: Valence band spectra recorded using hv = 110 eV on (a) pristine MWCNTs compared with those CF4-plasma treated for (b) 2, (c) 30, (d) 60 and (e) 300 s.

Summary CF4 rf-plasma treatment of CNTs effectively grafts fluorine at the CNT surface, inducing changes in the CNT valence electronic states due to the formation of C-F bonds. For increasing treatment time the relative intensity in the UPS spectra close to the Fermi energy level decreases suggesting that the functionalization of the CNT-surface can be tailored. The fluorine atomic concentration depends on the plasma exposure time. Acknowledgements This work was supported by the Belgian Program on Interuniversity Attraction Pole (IUAP 6/08), ARC-UMH, and by the European Community through the Integrated Infrastructure Initiative - Contract R II 3-CT-2004-506008 (IASFS). The support of the BESSY staff and in particular of Dr. Willy Mahler, Dr. Birgitt Zada and Mr. Mike Sperling is gratefully acknowledged. JG is a research associate of NFSR (Belgium). References (1) A. Tressaud, E. Durand, C. Labrugère, J. Fluorine Chem. 2004,125, 1639-1648. (2) H.Touhara, F. Okino, Carbon 2000, 38, 241-267. (3) K.N. Kudin, H.F. Bettinger, G.E. Scuseria, Phys. Rev. B 2001, 63, 045413 (4) A. Felten, C. Bittencourt, J.J. Pireaux, G. Van Lier, J.C. Charlier, J. Appl. Phys. 2005, 98, 074308. (5) S. Kaoru, S. Takeda, Jap. J. Appl. Phys. 2007, 46, 7977-7982. (6) www.nanocyl.com (7) http://www.bessy.de/upload/bitpdfs/mustang.pdf (8) J.J. Yeh, I. Lindau, At. Data and Nucl. Data Table 1985, 32, 1-155. (9) S. Hüfner, Photoelectron Spectroscopy, third ed., Springer-Verlag, Berlin 2003 (10) G. E. Nansé, P. Papirer, F. Fioux, A. Moguet, A. Tressaud, Carbon 1997, 35, 515-528. (11) T.E.F.M. Standaert, M. Schaepkens, N.R. Rueger, P.G.M. Sebel, G.S. Oehrlein, J.M.J. Cook, J. Vac. Sci. Technol. A 1998, 16, 239-249. (12) A. Felten, J. Ghijsen, J.J. Pireaux, R.L. Johnson, C.M. Whelan, D. Liang, G. Van Tendeloo, C. Bittencourt, J. Phys. D Appl. Phys. 2007, 40, 7379-7382. (13) E. Morikawa, J. Choi, H.M. Manohara, H. Ishii, K. Seki, K.K. Okudaira, N. Ueno, J. Appl. Phys. 2000, 87, 4010-4016.

185

Ë

Ë ×ØÙ Ý ÓÒ Ø

ÓÖÑ Ø ÓÒ Ó 
ÓÖ ¹× Ð ×× ×
× ¸

ÐÐ

»

Ù

Ò ÒÓÔ ÖØ 
Ð × Ò
º À Ù

1

¸ Àº ÃÖÙØ

1

¸ ź

Ù

Ð ¸ ˺ À

1

2

º Ì Ø

Ú ¸

2

º ÀÓ ÐÐ

2

2

Å ÖØ Ò¹ÄÙØ Ö¹ÍÒ Ú Ö× Ø Ø À ÐÐ ¹Ï ØØ Ò Ö À ÐÑ ÓÐØÞ¹ ÒØÖÙÑ ÖÐ Ò Ö Å Ø Ö Ð Ò ÙÒ Ò Ö
1

½

ÁÒØÖÓ Ù
Ø ÓÒ

Æ ÒÓ× Þ Ñ Ø Ð Ô ÖØ 
Ð × Ñ Ò ×ÙÖ 
 × Ó Ð ×× × Ö Ó Ö Ø ÒØ Ö ×Ø 
 Ù× Ó Ø Ö ÔÓØ ÒØ Ð ÔÔÐ 
 Ø ÓÒ × ÒÓÒÐ Ò Ö ÓÔØ 
 Ð Ñ Ø Ö Ð× ÓÖ Ô ÓØÓÒ 
 Ú 
 ׺ À Ö ¸ » Ù Ò ÒÓÔ ÖØ 
Ð × ´ÆÈ×µ Û Ö ÔÖ Ô Ö Ò ×Ó ¹Ð Ñ ¹× Ð 
 Ð ×× × Ý ÓÙ Ð ÓÒ Ñ¹ ÔÐ ÒØ Ø ÓÒº Ý Ù× Ò Ö ÒØ ÓÒ ÑÔÐ ÒØ Ø ÓÒ Ô Ö Ñ Ø Ö× Ø × ÔÓ×× Ð ØÓ Ù×Ø Ø ×ÙÖ 
 ÔÐ ×ÑÓÒ Ö ×ÓÒ Ò
 ´ËÈʵ Ò Ò ÜØ Ò Ö Ò Ó Û Ú Ð Ò Ø ×º Ì × × ÔÓ×× Ð 
 Ù× ÓÖ Ñ Ø ÐÐ 
 Ò ÒÓÔ ÖØ 
Ð × Ø ËÈÊ Ô Ò × ÓÒ Ø 
ÓÒØ ÒØ Ó × ÐÚ Ö Ò ÓÐ º Ø ÓÒ ÐÐݸ 
ÓÖ ¹× ÐÐ Ò ÒÓÔ ÖØ 
Ð × 
 Ò Ö ×ÙÐØ Ò × Ø Ó Ø ËÈÊ ØÓ Ö Ú Ð٠׺ Ï Ø Ë Ë¸ Ò Ð Ñ ÒØ ×Ô 
 
 ÒÚ ×Ø Ø ÓÒ ÐÐÓÛ× ØÓ Ú Ð Ø Ò ØÓ ÕÙ ÒØ Ý Ø 
ÓÖ ¹× ÐÐ ×ØÖÙ
ØÙÖ ÓÖ Ñ Ø ÐÐ 
 
ÓÑÔÓ× Ø ÓÒ Ó Ø » Ù Ò ÒÓÔ ÖØ 
Р׺ Ì Ö ÓÖ ¸ ØÓ ÙÒ Ö×Ø Ò Ø 
Ø× Û Ñ ×ÙÖ Ø ¹Ö Ý Ò Ö × ×Ð ØÐÝ ÐÓÛ Ø Ù ÄIII ¹ Ó Ö ÒØ ÑÔÐ ÒØ Ø ÓÒ × ÕÙ Ò
 × ÓÒ Ø 
ÓÒ ÙÖ Ø ÓÒ Ó Ø Ò ÒÓÔ ÖØ 
Р׺

¾

ÜÔ Ö Ñ ÒØ Ð

ËÓ ¹Ð Ñ Ð ×× × 
ÓÒØ Ò Ò ´ Ò ÑÓбµ ¾º ± Ë Ç2¸ ½ º ± Æ 2 Ǹ º ± Ǹ º¼± ŠǸ ¼º ± Ð2 Ç3 ¸ ¼º¾¼± Ã2 Ǹ ¼º¿± ËÇ3 ¸ Ò ¼º¼ ± 2 Ç3 Û Ö ÜÔÓ× ØÓ + ´¾¼¼ ε Ò »ÓÖ Ù+ ´¿¼¼ ε ÓÒ ÑÔÐ ÒØ Ø ÓÒ Ø ÖÓÓÑ Ø ÑÔ Ö ØÙÖ º ÇÒ Ð ×× × Ø× Ó ½ÑÑ Ø 
 Ò ×× Ö × Ó ¾¼Ü¾¼ ÑÑ2 Ú Ò ×Ù 
Ø ØÓ ÑÔÐ ÒØ Ø ÓÒ ÓÖ 
 ØÝÔ Ó ÓÒ× Ø 16 16 Ó× × Ö Ò Ò ÖÓÑ ¾Ü½¼ ØÓ Ü½¼ ÓÒ×»
Ñ2º Ì Ñ 
ÙÖÖ ÒØ Ò× ØÝ Û × ½ ñ »
Ñ2 º Ö Ù Ð ÙÔ Ö Ù
Ø ÓÒ Ø Ø Ð ×× ×ÙÖ 
 ÙÖ Ò ÑÔÐ ÒØ Ø ÓÒ Û × 
 Ú Ý Ð 
¹ ØÖÓÒ Ñ ÖÖ Ø ÓÒ ÓÒØÓ Ø ×ÙÖ 
 º ÌÛÓ × ÑÔÐ × Û Ø Ö ÒØ ÑÔÐ ÒØ Ø ÓÒ × ÕÙ Ò
 Û Ö ÔÖÓ Ù
 ܽ¼16 Ù+ · ܽ¼16 + ´× ÑÔÐ Ò Ñ Ù µ¸ ¾Ü½¼16 + · ܽ¼16 Ù+ ´× ÑÔÐ Ò Ñ ¾ Ùµº ÓÖ Ø Ë Ë Ñ ×ÙÖ Ñ ÒØ׸ Ø × ÑÔÐ × Û Ö Ñ 
 Ò 
 ÐÝ Ö Ò Ò ÔÓÐ × ØÓ Ø 
 Ò ×× Ó ½ ¼ñÑ ØÓ 
 Ú Ò 

 ÔØ Ð ØÖ Ò×Ñ ×× ÓÒº Ì ÜÔ Ö Ñ ÒØ Û × Ô Ö ÓÖÑ ØØ Ì ÅÈÏ Ë Ë ÑÐ Ò Ø ××ݸ ÖÐ Ò¸ Ò Ö ÐÐÓÛ Ø Ù ÄIII ¹ ´½½ ½ εº ÌÓ 
 Ö 
Ø Ö Þ Ø ËÈÊ Ù ØÓ Ø Ñ Ø Ð Ò ÒÓÔ ÖØ 
Ð × ÓÖÑ Ò Ø ÑÔÐ ÒØ Ö × Ø ÓÔØ 
 Ð Ò× ØÝ Ó Ø × ÑÔÐ × Û × Ö 
ÓÖ Ý Ñ Ò× Ó È Ö Ò¹ ÐÑ Ö ×Ô 
ØÖÓÑ Ø Ö Ò Ø Û Ú Ð Ò Ø Ö Ò Ó ¾ ¼¹ ¼¼ ÒѺ

¿
Ì

Ê ×ÙÐØ× Ò

Ó Ö ÒØ ÒØ Ò× ØÝ Ó

×
Ù×× ÓÒ
Ë
¢

Ë ÜÔ Ö Ñ ÒØ 
 Ò
i

×
Ö

Ý

dσ (Q) = dΩ

2 Ni (R)∆ηi V (R)2 |F (Q, R)|2 dR

Û Ö ∆ηi × Ø ×
 ØØ Ö Ò 
ÓÒØÖ ×ظ Ni (R) Ø × Þ ×ØÖ ÙØ ÓÒ¸ V (R) Ø ÚÓÐÙÑ Ò F (Q, R) Ø ÓÖÑ 
ØÓÖ Ó ÆÈ× Ó ØÝÔ iº ÌÓ ÒÚ ×Ø Ø Ø 
 Ñ 
 Ð 
ÓÑÔÓ× Ø ÓÒ Ó
186

½

10

4Au4Ag

10

2Ag4Au

(cm )

1

(cm ) d /d

-1

-1

1

d /d

11450eV 11867eV 0.1 11913eV 0.1

11450eV 11867eV 11913eV

0.1

q (nm )

-1

1

0.1

q (nm )

-1

1

ÒÖ

ÙÖ ½

×

Ë Ë Ñ ×ÙÖ Ñ ÒØ× Ó Ø º ÐÓÛ Ø Ù ÄIII ¹

× ÑÔÐ ×

Ù

´Ð ص Ò ¾

Ù ´Ö

ص Ø Ø Ö

ÆÈ׸ 
ÓÒØÖ ×Ø Ú Ö Ø ÓÒ Û × Ù× Ý ÙØ Ð × Ø ÒÓÑ ÐÓÙ× ×Ô Ö× ÓÒ Ø Ø Ù ÄIII ¹ º Ì Ö ÓÖ ¸ Ñ ×ÙÖ Ñ ÒØ× Ø Ø Ö Ò Ö × Ò Ö ÐÓÛ Ø Û Ö Ô Ö ÓÖÑ º ÁÒ Ø ×
 ×¸Ø ØÓÑ 
 ×
 ØØ Ö Ò 
ØÓÖ ÓÖ Ù ØÓÑ× Ú ÖÝ Û Ø Ø Ò Ö Ý E ¸ fAu (E) = ′ ′′ f0Au +fAu (E)+ifAu (E)¸ Ò Ø ×
 ØØ Ö Ò 
ÓÒØÖ ×Ø × Ò × ∆η(E) = ∆ρAu fAu (E)+ ∆ρj fj Û Ö ∆ηj × Ø Ö Ò
 Ó Ø ØÓÑ 
 Ò× ØÝ Ó Ð Ñ ÒØ j ØÛ Ò Ñ ØÖ Ü j=Au Ò ÆÈ Ò fj Ø× ×
 ØØ Ö Ò Ð Ò Ø º ÁÒ º½Ø Ë Ë Ñ ×ÙÖ Ñ ÒØ× Ø Ø Ö Ò Ö × ÐÓÛ Ø Ù ÄIII ¹ ÓÖ Ø ØÛÓ × ÑÔÐ × Ö × ÓÛÒº 
Ð Ö ÒÓÑ ÐÓÙ× 
Ø Û × Ó × ÖÚ ÓÖ ÓØ × ÑÔÐ × Ò Ø ×
 ØØ Ö Ò 
ÙÖÚ × Ò 
 Ø Ø ÔÖ × Ò
 Ó 
ÓÖ ¹× ÐÐ Ò ÒÓÔ ÖØ 
Р׺ ÖÓÑ Ì Å»ÀÊÌ Å ´× º ¾µ Ø × ÒÓÛÒ¸ Ø Ø Ø ÆÈ× 
ÓÒ× ×Ø Ó 
ÓÖ ¹× ÐÐ¹Ô ÖØ 
Ð × Ò Ø ÓÒ Ð ×Ñ ÐÐ ÓÑÓ Ò ÓÙ× Ô ÖØ 
Р׺ Ì Ö ÓÖ ¸ ÑÓ Ð Û Ø 
ÓÖ ¹× ÐÐ Ô ÖØ 
Ð × Ò ×Ô Ö 
 Ð Ô ÖØ 
Ð × Û × Ù× ÓÖ Ø × ÑÙÐØ Ò ÓÙ× Ø Ó Ø Ø Ö Ò Ö Ý Ô Ò ÒØ ×
 ØØ Ö Ò 
ÙÖÚ ×º Ý Ù× Ò Ø Ö Ø Ó Ó Ø ×
 ØØ Ö Ò 
ÓÒØÖ ×Ø ∆η(E) ØÛ Ò Ø Ø Ö Ò Ö × Ò ØØ Ò Ø Ë Ë 
ÙÖÚ × × ÑÙÐØ Ò ÓÙ×ÐÝ Û Ø Ø ÑÓ Ð¸ Ø × ÔÓ×× Ð ØÓ 
ÓÑÔ Ö Ø Ö ×ÙÐØ× Û Ø Ø 
 Ð
ÙÐ Ø 
ÓÒØÖ ×Ø Ú ÐÙ × Ò ØÓÓ Ø Ø 
ÓÑÔÓ× Ø ÓÒ Ó Ø ×Ô Ö Ò Ø × ÐÐ Ó Ø ÆÈ× ´× º ¿µº ÓÖ Ø × ÑÔÐ Ù Ø 
ÓÑÔÓ× Ø ÓÒ × ¼± Ù Ò ¼± Û Ö × ÓÖ ¾ Ù Ø × ¼± Ù Ò ½¼± º Ï Ø Ø × Ú ÐÙ × ÓÖ Ø 
ÓÑÔÓ× Ø ÓÒ¸ Ø ×
 ØØ Ö Ò 
ÓÒØÖ ×Ø ∆η ÓÖ Ø × ÐÐ Ò ×Ô Ö 
 Ò 
 Ð
ÙÐ Ø Ò Ù× Ò Ø Ò ÜØ Ø ×Ø Ôº ÆÓÛ¸ Ø × Þ ×ØÖ Ù¹ Ø ÓÒ¸ ÚÓÐÙÑ 
ÓÒ
 ÒØÖ Ø ÓÒ Ò ÒÙÑ Ö Ó Ô ÖØ 
Ð × ÓÖ ÐÐ Ò ÒÓÔ ÖØ 
Ð × Ö ×ÙÐØ× ÖÓÑ Ø ØØ Ò ÔÖÓ
 ÙÖ × Û ÐÐ × Ø ×
 ØØ Ö¹ Ò 
ÓÒØÖ ×Ø ∆η ÓÖ Ø 
ÓÖ ´× Ì º ½µº Ì Ò¸ Ø ∆η Ú ÐÙ × ÖÓÑ Ø 
ÓÖ 
 Ò Ù× ØÓ Ø ÖÑ Ò Ø 
ÓÑÔÓ× Ø ÓÒ Ó Ø 
ÓÖ º Ì × 
 Ð
ÙÐ Ø ÓÒ× × ÓÛ Ø Ø ÓÖ × Ñ¹ ÔÐ Ù Ø 
ÓÖ × 
ÓÑÔÐ Ø ÐÝ ÓÐÐÓÛº ÓÖ ¾ Ù¸ ∆η Ò 
 Ø × Ø Ø Ø 
ÓÖ ÙÖ ¾ Ì Å»ÀÊÌ Å Ñ × Ó » Ù × ÒÓØ ÓÐÐÓÛ ÙØ 
ÓÒØ Ò× Ô × Û Ø Ò ÒÓÔ ÖØ 
Ð × Û Ø 
ÓÖ ¹× ÐÐ ×ØÖÙ
ØÙÖ Ò ×Ñ ÐÐ Ò× ØÝ ÐÓÛ Ö Ø Ò Ø Ø Ó º Ø
×Ô Ö 
 Ð Ô ÖØ 
Ð × ´ Ù µº

187

¾

2Ag4Au
1.00 0.99 0.98 0.97 0.96

4Au4Ag
1.2 2x10 4x10 1.0 4x10 2x10 0.8
16

Ag Au Au 4x10 Ag 4x10
16

16

16

Ag Au

16

16

(11450eV)

0.95 0.94 0.93 0.92 0.91

optical density
11450eV 11867eV 11913eV 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

0.6

(E)/

0.90 0.89 0.88 0.87 0.86 0.85 0.84 0.0

0.4

0.2

0.0 300 350 400 450 500 550 600 650 700 750 800

Ag content

(nm)

ÙÖ ¿ Ê Ø Ó ØÛ Ò Ø ×
 ØØ Ö Ò 
ÓÒ¹ ØÖ ×Ø× Ô Ò ÒØ ÖÓÑ Ø ¹
ÓÒØ ÒØ Ò Ö ¹ ×ÙÐØ Ò 
ÓÑÔÓ× Ø ÓÒ Ó Ø × ÐÐ Ò ×Ô Ö 
 Ð Ò ÒÓÔ ÖØ 
Р׺

ÓÖ Ø ÓÐ º

ÙÖ

ÇÔØ 
 Ð ×Ô 
ØÖ ØÛÓ × ÑÔÐ × Ò

Ò ×ØÖÙ
ØÙÖ ÑÓ Ð ÓÖ ÔÙÖ × ÐÚ Ö Ò

Ù

ÓÑÔº
N fp s R ∆η

¾

Ù

ÓÑÔº
N fp s R ∆η

ËÔ Ö ¼± Ù» ¼± ½º¼¾×½¼16 ¼º¼¼ ¾± ¼º ¿ ½º½ º ×½¼10 ¼± Ù»½¼± º ×½¼15 ¼º¼½½¿± ¼º ¿ ½º¾ ¼º ¾×½¼10

Ë ÐÐ ¼± Ù» ¼± º½¼ º ×½¼10 ¼± Ù»½¼± º¼ ¼º ¾×½¼10

¾º ½×½¼13 ¼º¼¼¿ ± ¼º¾

ÓÖ ÓÐÐÓÛ ¾º ½ ¹½ º¾ ×½¼10 ÙÒ ÒÓÛÒ ¾º

¿º ¾×½¼13 ¼º¼¼ ± ¼º¾

¾º¾ ×½¼10

Ì Ð ½ Î ÐÙ × ÖÓÑ Ø × ÑÙÐØ Ò ÓÙ× Ø Ó Ø Ø Ö Ë Ë 
ÙÖÚ × 
ÓÑÔÓ× Ø ÓÒ¸ ÒÙÑ Ö Ó Ô ÖØ 
Ð × N ¸ ÚÓÐÙÑ Ö 
Ø ÓÒ f p¸ Ú Ö Ò
 s¸ Ñ Ò Ô ÖØ 
Ð Ö Ù× R¸ ×
 ØØ Ö Ò 
ÓÒØÖ ×Ø ∆η

ÓÒ
ÐÙ× ÓÒ
ÏØ Ø Ë Ë Ñ ×ÙÖ Ñ ÒØ× Ø Û × ÔÓ×× Ð ØÓ Ø Ò ÓÖÑ Ø ÓÒ ÓÙØ Ø × Þ ¸ × Þ ×ØÖ ÙØ ÓÒ¸ × Ô Ò ×ØÖÙ
ØÙÖ Ó Ø » Ù ÆÈ׺ Ð×Ó¸ Ø 
ÓÑÔÓ× Ø ÓÒ Ó Ø 
ÓÖ ¹ × ÐÐ ÆÈ× 
 Ò ÒÚ ×Ø Ø Û Ø Ø × Ñ Ø Ó º Ì ÆÈ× Ò Ø × ÑÔÐ Ù ÛØ 
ÓÑÔÓ× Ø ÓÒ Ó ¼± Ù Ò ¼± Ö ×ÙÐØ× Ò ËÈÊ ØÛ Ò Ø Ø Ó ÔÙÖ ÓÖ Ù 
 Ù× Ó Ø 
ÓÑÔÓ× Ø ÓÒ Ó Ø × ÐÐ Ò Ø ×Ñ ÐÐ ×Ô Ö 
 Ð Ô ÖØ 
Ð × ´× º µº Ì ÓÐÐÓÛ 
ÓÖ ÒÓØ 
ÓÒØÖ ÙØ ØÓ Ø ËÈʺ ÓÖ Ø ÓØ Ö × ÑÔÐ ¸ ¾ Ù¸ Ø 
ÓÑÔÓ× Ø ÓÒ ÖÓÑ Ø × ÐÐ Ò Ø ×Ñ ÐÐ ×Ô Ö 
 Ð ÆÈ× Ó ¼± Ù Ò ½¼± ÑÙ×Ø Ö ×ÙÐØ Ò × Ø Ó Ø ËÈÊ Ò Ö ØÓ Ø ËÈÊ Ó Ùº ÙØ Ò Ø × × ÑÔÐ Ø 
ÓÖ × ÒÓØ ÓÐÐÓÛ Ò Ø Ö ÓÖ Ø ËÈÊ × × Ø ØÓ Û Ú Ð Ò Ø Ú ÐÙ × Ö Ø Ò Ø Ø Ó ÔÙÖ Ùº Ì × ÒÚ ×Ø Ø ÓÒ× × ÓÛ Ø Ø Ø × ÔÓ×× Ð ØÓ Ù×Ø Ø ËÈÊ ØÛ Ò Ø Ø Ó ÔÙÖ Ò Ù ÓÖ ÓÐÐÓÛ ÆÈ× Ý Ù× Ò Ö ÒØ ÓÒ Ó× × ØÓ Ú ÖÝ Ø 
ÓÑÔÓ× Ø ÓÒº Ï Ø Ö ÒØ ÓÒ ÑÔÐ ÒØ Ø ÓÒ × ÕÙ Ò
 × Ø × ÔÓ×× Ð ØÓ ÔÖÓ Ù
 
ÓÖ ¹× ÐÐ ÆÈ× Û 
 Ö ÒÓØ ÓÐÐÓÛ Ò Ø Ö ÓÖ × Ø Ó Ø ËÈÊ ØÓ Ú ÐÙ × Ö Ø Ò Ø Ø Ó ÔÙÖ Ù × ÔÓ×× Ð º

188

¿

TXM-NEXAFS of Individual Titanate-based Nanoribbon
Carla Bittencourt LCIA, University of Mons-Hainaut, Mons, Belgium; Xiaoxing Ke and Gustaaf Van Tendeloo EMAT, University of Antwerp, Belgium, Peter Guttmann Helmholtz-Zentrum Berlin für Materialien und Energie, Elektronenspeicherring BESSY II, Albert-Einstein-Str. 15, 12489 Berlin, Germany, Jacques Ghijsen and Alexandre Felten LISE, University of Namur (FUNDP), Belgium; 1- Introduction Low dimensional TiOx- based nanostuctures, such as nanowires, nanotubes, and nanorods bring to reality the possibility to fine-tune chemical reactivity as the system structure and the occupation of the outermost energy levels can be tuned by changing preparation parameters. Several structures have been proposed for these structures, including scrolling of anatase sheets, trititanate H2Ti3O7 exfoliated sheets, (Na,H)2Ti2O4(OH)2 based layers, and lepidocrocite titanate-like sheets. Based on x-ray diffraction and transmission electron microscopy experiments, it is currently believed that the structure is closely related to the family of layered titanate H2TinO2n+1 materials, where n = 3,4,5 or even ∞ for the end-member composed of flat layers of octahedra The exact structure is still a matter of debate. In this work, electronic properties of individual TiOx-pristine nanoribbons (NR) prepared by hydrothermal treatment of anatase TiO2 micro-particles were studied using the BESSY TXMNEXAFS end-station at beamline U41. NEXAFS is ideally suited to study TiO2-based materials because both the O K-edge and Ti L-edge features are very sensitive to the local bonding environment, providing diagnostic information about the crystal structures and oxidation states of various forms of titanium oxides and sub-oxides. TXM-NEXAFS combines microscopy with spectroscopy allowing the study of the electronic structure of individual nanostuctures with spatial resolution better than 25 nm [A. Felten et al., Nano Lett., 7, (2007), 2435]. In addition, the directional electric field vector (Ē) of the x-rays can be used as a “search tool” for the direction of chemical bonds of the atom selected by its absorption edge [J. Stohr in NEXAFS Spectroscopy, Springer Series in Surface Science, vol. 25]. The electronic structure of these nanoribbons is discussed in terms of the crystal field splitting. Reference anatase and rutile samples are used as template for the interpretation of the spectroscopic signatures of the nanoribbon. 2- Results NEXAFS spectra are clear fingerprint of titanium oxides. They show significant differences depending on the crystallographic phase as well as on the Ti reduction. TiO2 single crystal spectra show well-resolved peaks in the range between 455 and 470 eV corresponding to the various Ti 2p → 3d transitions [U. Diebold, Surf. Sci. Rep. 48, 53 (2003)]. The Ti L-edge shows two groups of peaks arising from the spin-orbit splitting of Ti 2p core level into 2p3/2 and 2p1/2 levels with ~6 eV energy separation. These levels are further split due to crystalfield effects. The Ti L2,3-edges recorded on rutile TiO2 micro-particles (figure 1a), anatase TiO2 micro-particles (figure 2a), and the ribbon have similar form (figure 3a): the L3 and L2 edges are separated by the 2p core hole spin orbit coupling around 5 eV. The L3 and L2 edges are both split by the strong crystal field splitting arising from the surrounding oxygen atoms. The experimental splitting of the L2 line for anatase and the nanoribbon are of comparable magnitude (~ 2 eV), splitting for rutile being somewhat larger (~ 2.6 eV). The separation between the t2g and eg orbitals is about 1.8 eV for anatase and 2 eV for rutile in accordance with reported results [J. P. Crocombette and F. Jollet, J. Phys.: Condens. Matter 6, 10811 (1994)].

189

The O K-edges for the rutile (figure1b), anatase (figure2b), and ribbon (figure 3b) is characterized by two main features. For anatase these are split by 2.5 eV whereas for rutile it is 2.75 eV. For the ribbon the splitting is around 2.5 eV, closer to anatase. This confirms that the crystal field splitting of the titanate ribbons is close to that of the anatase phase.
Ti L - edge
Intensity (arb. units)

L2
3

TiO2-rutile
Intensity (arb. units)

L3

TiO2 -rutile
454 456 458 460 462 464 466 468 470 472

525

530

535 p h o to n e n e rg y (e V )

540

545

p h o to n e n e rg y (e V )

Figure 1: TXM-NEXAFS spectra recorded on TiO2 rutile powder: a) Ti L-edge b) O K-edge

L2
Intensity (arb. units)

TiO2-anatase

3

L3

Intensity (arb. units)

TiO2 - anatase
454 456 458 460 462 464 466 468 470 472

525

530

535 p h o to n e n e rg y (e V )

540

545

p h o to n e n e rg y (e V )

Figure 2: TXM-NEXAFS spectra recorded on TiO2 anatase powder: a) Ti L-edge b) O K-edge

Intensity (arb. units)

Intensity (arb. units)

t2g eg
454 456 458 460 462 464 466 468 470 472 photon energy (eV)

t2g eg
525 530 535 photon energy (eV) 540 545

2 µm

Figure 3 : TXM-NEXAFS spectra recorded on a single TiOx-nanoribbon: a) Ti L-edge b) O K-edge c) TXM - image at hv = 464 eV showing the two experimental geometries used to record the spectra.

Figure 3 illustrates the oxygen 1s (K-edge) and titanium 2p (L2,3-edge) absorption edges recorded on the nanoribbon in two experimental geometries: Ē parallel and perpendicular to the principal axis of the nanostructure. The Ti 2p spectra are similar to those reported for the TiO2 anatase phase [M.B. Casu et al., Surf. Sci. 602, 1599 (2008)], with the main discrepancy in the single structure at 460 eV that appears split in the TiO2. This structure results from transitions to the final state (2p3/2)-1d(3eg)1p6, the eg states are formed by d(z2) and d(x2-y2) orbitals, which are directed towards ligand anions and are sensitive to deviations from Ti Oh symmetry. Consequently, the absence of splitting of eg states into d(z2) and d(x2-y2) suggests that for TiOx-nanoribbon Ti occupies sites with high Oh symmetry in contrast to sites with distorted Oh symmetry in TiO2-anatase. The normal incidence NEXAFS spectra were measured with Ē parallel and perpendicular to the principal axis of the nanostructure. No strong evidence for anisotropic distribution of Ti
190

sites can be observed. Conversely, the O 1s spectra suggest anisotropic distribution of O sites. The O 1s transitions identified as t2g and eg in the spectra result from transitions to final states, 3d(2t2g)1(1s)-1p6 and 3d(3eg)1(1s)-1p6. The energy separation between t2g and eg (crystal field splitting) is 2.5 eV, in close agreement with the value measured in the Ti 2p spectrum. This work shows the high potential for the TXM-NEXAFS for study isolate low-dimensional nanostrutures. Acknowledgment: This work is financially supported by the Belgian Program on Interuniversity Attraction Pole (PAI 6/08), ARC-UMH and by BESSY and the European Commission under contract RII3-CT 2004-506008 (IASFS).

191

Electronic structure and bonding in tetrakis-pyridyl porphyrins S.I.Bozhko1, V.S.Bozhko1, M.M.Brzhezinskaya2,4, G.Dyker,3 A.M.Ionov1, L.V. Yashina5
1

Institute of Solid State Physics, 142432 Chernogolovka, Russia
2

BESSY GmbH, Albert-Einstein-Str. 15, 12489 Berlin, Germany;

4

Ruhr-Universitat Bochum, Universitatstrasse 150, D-44780 Bochum Germany Institute of Physics, St. Petersburg State University, 198904 St. Petersburg, Russia 5 Moscow State University, 119899 Moscow, Russia

3

Organic molecular films of porphyrins have attracted attention over the past years in view of potential applications in electronic devices. One of the most exciting properties of compounds is the ability to control electronic properties and spin states of porphyrins by optical excitation and introducing charge carriers which is important for molecular electronics. For understanding and tailoring their properties knowledge of electronic structure and bonding in these compounds is required. As shown in [1-3] in a wide class of porphyrins one can vary electronic state by ligands and metal doping. The occupied and empty states of porphyrins which are involved in chemical bonding are essentially responsible for promising properties and should be studied. In present project electronic structure and chemical bonding in tetrakis- (tetrakis-3 and -4pyridyl) porphyrins were studied by UPS and XPS photoemission spectroscopy and XAS spectroscopy. To obtain this information high resolution of N1s, C1s, valence band and absorption spectra were measured at Russian-German beamline at BESSY ( MUSTANG Phoibos 150 electron analyzer) in the photon energy range 100-1000 eV. The base pressure in the UHV chamber for measurements was about 5×10-10 mbar. The angle of the incident photons was kept fixed at 67° relative to the sample surface normal. We measured normal emission valence band and core level spectra. Total resolution (electron plus photons) was about 120 meV. Chemically (drop coating chloroform solution of the compounds and UHV in situ organic molecular beam deposited films (from oven with temperature at 590K onto Si/Au, In, layered Bi2Te3 and SnS surfaces at room temperature) of porphyrins were studied The chemically deposited samples were additionally cleaned in situ by resistive heating up to ~ 400 K. Preliminary experiments on possible interaction of porphyrins with Pt under direct UHV deposition also were performed. We have examined the electronic structure of valence band and core levels of pristine tetrapyridyl-3,4-porphyrins using synchrotron radiation PES. XPS spectra of N1s levels for TPy4P are shown in Fig.1. Similar results were obtained for TPy3P porphyrin. In TPy-porphyrins XPS spectra three peaks structure of N1s states with binding energies 400.1, 399.2 and 398 eV were assigned to pyrrol, pyridyl, and aza nitrogen
192

respectively ( similar to [3]). Different peaks observed in the C1s spectra are related to the unequivalent C atoms in the porphyrin molecules (aromatic and C-N-C groups) as well as shakeup HOMO-LUMO satellite.

Fig.1 XPS spectra of TPy4porphyrin/In: top panels – N1s (hv=600eV) and valence band (hv=120eV) spectra, bottom panel – XPS Pt4f level spectra (hν=400eV)

UPS studies of VB (performed at photon energy 120 eV) of TPy-porphyrins show that the valence band is mainly formed by peaks corresponding to π (2- 10 eV) and σ states (8-16 eV) of porphyrin macrocycles. Peak of C2s states in lied at about 18 eV in VB spectra. The C K-edge spectra of NEXAFS spectra of multilayer films (UHV sublimation in situ) of pyridyl-porphyrins exhibit resonances at 284 eV and 287 eV which could be attributed to excitation into unoccupied orbitals located on the porphyrin ring aromatic system. Similar the N
three parts K-edge spectra connected with the three different nitrogen atom types were observed in

NEXAFS spectra of these compounds. UHV deposition of a small excess of Pt atoms on this 2HTPyP monolayer results in the complex signal of Pt shown in Fig. 1 and changes in N1s spectra. The simplest possible approach to interpreting the complicated Pt signal assumes that it arises from a mixture of the product of metalation Pt in TPyP (with base and pyridyl group), and unreacted Pt. Further evidence for the existence of this species is provided by the Pt 4f XPS spectra in Figure 1, in which intermediate states of Pt are observed. The respective signal components (Figure 1, magenta line) are located at 72.1 eV), i.e., between the signal of unreacted Pt clusters 71.8 eV, (brown line) and PtTPy4P (72.7 eV, green line).

193

Preliminary photoemission studies have apparently demonstrated the interaction of monolayered film of TPy4-porphyrin with Pt on the surface after deposition in situ in UHV and changing of Pt4f states and N1s spectra. Acknowledgement
This work was supported by the bilateral program “Russian-German Laboratory at BESSY”. The authors thank the staff of BESSY and the Russian-German Laboratory for technical assistance. References.

1.G.Polzonetti et.al. Chem.Phys. 296, (2004) 87. 2.M.P. de Jong et.al. Phys.Rev.B 72 (2005) 035448. 3. F. Klappenberger, A. Weber-Bargioni, W. Auwarter, M. Marschall, A. Schiffrin, and J. V. Barth, JOURNAL OF CHEMICAL PHYSICS 129, 214702 _2008_

194

Characterization of self-assembled monolayers of terphenylthiols with amine head groups
P. Dietrich1, N. Graf1, T. Gross1, A. Lippitz1, B. Schüpbach2, A. Terfort2 and W. E. S. Unger1*
1

BAM Bundesanstalt für Materialforschung und -prüfung, VI.43 – Surface and Thin Film Analysis, D-12203 Berlin, Germany
2

Anorganische Chemie, Philipps-Universität Marburg, , D-35032 Marburg, Germany

Self-assembled monolayers (SAMs) of thiols on gold are unique systems to tailor surface structures and their chemical nature. For this reason they were investigated intensively over the past decades.[1] Especially for high-throughput diagnostic applications thiolate based SAMs on gold can be used as platform to immobilize different biological structures (e.g. peptides, proteins, DNA, carbohydrates). Therefore adequate functional head groups and the knowledge of the basic surface chemistry are a prerequisite. Within a program to improve microarray performance we studied different aminated surfaces.[2-4] Herein we report on the analysis of SAMs on gold based on ω-aminoterphenyl substituted alkanethiols. These monolayers were analyzed by XPS (X-ray Photoelectron Spectroscopy) and NEXAFS (Near Edge X-ray Absorption Fine Structure) spectroscopy at BESSY to gain deeper insight into their structural and chemical nature. The surface chemistry and reactivity of the amino groups was tested in a qualitative manner by chemical derivatization using the amine-reactive isothiocyanate reagent ITC.
CF3

F3C NCS CF3 S S NH 1-F NH

F3 C S NH 2 1 ITC

NEt 3 , DMSO, rt, 3h

Scheme 1: Derivatization of 2-(4''-amino-1,1':4',1''-terphenyl-4-yl)ethane thiol (1) with 1-isothiocyanato-3,5-bis(trifluoromethyl)-benzene (ITC).

195

In Figure 1 high-resolution C1s XP spectrum of thiolate 1-F after reaction with ITC on gold is shown exemplarily.

2 x 10 150 140 130 120 110 CPS 100 90 80 70 60 298 296 294 292 290 288 Binding Energy (eV) 286 284 282 280

Name C-C (arom) C-H, C-C C-N C=N / NHC=S shake-up shake-up CF3

Pos. %Area 284.1 68.8 285.4 14.9 3.8 286.6 2.8 287.7 1.7 289.5 1.9 290.8 6.1 292.6

Figure 1: High-resolution C 1s XP spectrum of a derivatized 1-F SAM. After chemical derivatization the most distinct peak in the C 1s region of 1-F is the signal at 292.6 eV originating from newly incorporated CF3 groups after successful reaction with ITC. This interpretation is supported by signals at 287.7 eV in the C 1s and at 399.8 eV in the N 1s (not shown) caused by carbon and nitrogen atoms of the thiourea moiety. NEXAFS analysis provides both chemical and structural information and can be used to determine a preferential orientation of the molecular moieties within the monolayer with respect to the surface normal. For this purpose NEXAFS measurements were done at BESSY’s HE-SGM beam line at different angles (20°, 55°, 90°) of the incidence synchrotron light. Figure 2 illustrates the C K-edge NEXAFS of the terphenyl SAMs 1 and 1-F. The most intense resonance of both spectra at 285 eV is associated with transitions to

π (*C =C ,C = N ) orbitals of the aromatic backbone. The spectra show further contributions
due to σ* resonances at 288 eV (C-H) and 292 eV (C-C, C-N). After chemical derivatization the C K-edge NEXAFS of the fluorinated SAM 1-F changes noticeably compared to 1. The peak at 295 eV can be assigned to excitations into the σ (*C − F ) orbitals of the CF3 groups whereas π (*C =C − N ) , π (*C = S ) , and
* π ring resonances are found at 286.4 eV, 287.8 eV and 289.0 eV.

196

4
π*(C=C, C=N) σ*(C-C, C-N)

Normalized Intensity [arb. units]

Normalized Intensity [arb. units]

3
σ*(C-H)

90° 55° 20° 90° - 20°

7 6 5 4 3 2 1 0 -1

π*(C=C, C=N)

σ*(C-H) and /or π*(C=S) σ*(C-C, C-N) σ*(C-F) π*ring σ*(C=C, C=N)

π*(C=C-N)

90° 55° 20° 90° - 20°

2
π*ring

σ*(C=C, C=N)

1

0

280

285

290

295

300

305

310

315

320

280

285

290

295

300

305

310

315

320

Photon energy [eV]

Photon energy [eV]

Figure 2: C K-edge NEXAFS at three different angles of incidence of linear polarized synchrotron light (90°, 55°, 20°) and the difference spectrum (90°-20°) of 1 (left) and its fluorinated analog 1-F (right). The intensity of the

π * resonance at 285 eV depends on the incident angle of the α = 62° for the π orbitals of the aromatic

synchrotron beam[5]. This phenomenon is known as linear dichroism effect. Using this relation an average tilt angle of

ring system in monolayer 1 was determined. And after reaction with the isothiocyanate ITC the average tilt angle was slightly increased to 66°. These results indicating that the preferred orientation within the monolayer 1-F remains unaffected by the chemical derivatization reaction. These studies will be extended to other SAMs based on aromatic thiols with amino groups in order to quantify the amount of reactive sites on the surface by chemical derivatization XPS.
References [1] [2] [3] [4] [5] J. C. Love, L. A. Estroff, J. K. Kriebel, R. G. Nuzzo, G. M. Whitesides, Chemical Reviews 2005, 105, 1103. N. Graf, E. Yegen, T. Gross, A. Lippitz, W. Weigel, S. Krakert, A. Terfort, W. E. S. Unger, submitted to Surf. Sci. 2008. N. Graf, E. Yegen, A. Lippitz, D. Treu, T. Wirth, W. E. S. Unger, Surface and Interface Analysis 2008, 40, 180. N. Graf, E. Yegen, A. Lippitz, W. E. S. Unger, in BESSY Annual Report 2007, Berlin, 2007, pp. 107. J. Stöhr, NEXAFS Spectroscopy, Springer, Heidelberg, Germany, 1992.

197

Electronic structure of LaFeAsO1-xFx from Photoemission Spectroscopy
A. Koitzsch1, D. Inosov1, J. Fink1,2, M. Knupfer1, H.Eschrig1, S. V. Borisenko1, G. Behr1, A. Köhler1, J. Werner1, B. Büchner1, R. Follath2 , and H. A. Dürr2
1

Institute for Solid State Research, IFW Dresden, P.O. Box 270116, D-01171 Dresden, Ger-

many.
2

BESSY GmbH, Albert-Einstein-Strasse 15, 12489 Berlin, Germany

The recent discovery of the superconducting oxypnictides has sparked immediate and intense scientific effort. Superconducting transition temperatures up to Tc = 43 K have been reached for LaFeAsO1-xFx [1, 2]. LaO1-xFxFeAs crystals consist of alternating LaO1-xFx and FeAs layers. It is assumed that the LaO1-xFx layers serve as ionic charge reservoirs for the covalently bound metallic FeAs layers where the superconductivity appears. According to band structure calculations the density of states in the vicinity of the Fermi energy is dominated by iron character [3]. Here we report on angle integrated photoemission measurements of LaFeAsO1-xFx (x = 0, 0.1, 0.2). We have measured the photointensity of polycrystalline material for various excitation energies ranging from hν = 15 eV to hν = 200 eV at T = 30 K. Polycrystalline samples of LaFeAsO1-xFx were prepared by using a two-step solid state reaction method, similar to that described by Zhu et al. [4]. The crystal structure and the composition was investigated by powder X-ray diffraction (XRD) and wavelength dispersive X-ray spectroscopy (WDX). Critical temperatures of Tc ≈ 23 K and Tc ≈ 10 K for x = 0.1 and x = 0.2, respectively, have been extracted from magnitization and resistivity measurements. The undoped sample shows a transition to a commensurate spin density wave below TN = 138 K [5]. The data were measured using synchrotron radiation at the “13” ARPES station with a Scienta R4000 spectrometer at BESSY. The samples have been scraped in situ before measurements at a pressure of p = 1 · 10-7 mbar. The base pressure in the measurement chamber was p = 1 · 10-10 mbar. Figure 1 presents the valence band of undoped LaFeAsO taken with different photon energies. The spectra have been normalized to the high energy shoulder of the broad peak centered at about E ≈ -5 eV (marked by the arrow). We compare the experimental data to LDA based orbital resolved density of states (DOS) calculations in panel (b). The main features of the valence band for all photon energies are a peak near the Fermi energy at E ≈ -0.25 eV and the broad peak around E ≈ -5 eV. The inset of panel (a) shows a zoom of the vicinity of EF for hν = 15 eV. The Fermi edge appears as a small slope change at the low energy tail of the spectral weight. In between the near EF peak and the broad peak a plateau-like region is observed with a small peak at E ≈ 1.7 eV. The broad peak has a complex structure and consists at least of two separate features. The low energy peak and the broad peak, have a significant dependence on the photon energy. The center of gravity of the broad peak shifts towards lower energies due to the intensity increase of the low energy shoulder, which becomes more intense than the high energy shoulder at hν = 95 eV. The presented spectra are taken at T = 30 K, but no major changes are observed when crossing the magnetic ordering temperature TN = 138 K. The reason for the hν dependent intensity variations lies in the hν dependence of the photoemission cross section. The opposite behavior of the high energy shoulder of the broad peak and the low energy peak suggests that they arise from different atomic orbitals. We show the energy dependence of the cross section of the potentially important valence orbitals, namely Fe 3d, As 4p, and O 2p as a function of photon energy in Fig 1c. As 4p is important for low
198

energies (< 25 eV) only. For energies above ~ 25 eV the spectra will be governed by Fe 3d and O 2p emission. For increasing photon energy Fe 3d dominates. Fig 1d shows the ratio of the cross sections Fe 3d/O 2p (blue line). It increases monotonically in the measured range and reaches a value of 6.5 for hν = 200 eV. Motivated by the increase of the low energy peak and the decrease of the high energy shoulder with increasing hν we assume, that the former is due to Fe 3d and the latter is mainly due to O 2p. We evaluate the intensity of the low energy peak by taking the integral from zero to E = -1 eV without any background treatment (red box). Since we normalized the spectra to the high energy shoulder this integral value corresponds automatically to the experimental intensity ratio. Those values are plotted in Fig. 1d as red points. We find satisfactory agreement and conclude a posteriori the correctness of our assumptions. To conclude we have investigate the electronic structure of the iron arsenide superconductor of LaFeAsO1-xFx and find gross agreement to band structure calculations. In particular the Fe character of the states near the chemical potential is confirmed.

Figure 1: (a) hν dependent photoemission valence band spectra of LaFeAsO. The arrow marks the point of normalization. The red rectangle is the integration window for the low energy weight shown in panel (d); the inset shows the near EF region for hν = 15 eV. Note the small change at E = 0 which indicates the Fermi edge. (b) LDA derived orbital resolved. (c) Atomic photoemission crosss section for the relevant orbitals [6]. (d) Ratio of the Fe 3d and O 2p cross section from (c) (blue line) compared to experimental values obtained by integrating the low energy peak.

[1] Kamihara et al. J. Am. Chem. Soc. 130, 3296 (2008) [2] H. Takahashi, Nature 453, 376 (2008) [3] H. Eschrig, arXiv:0804.0186 [4] X. Zhu et al., Supercond. Sci. Technol. 21 (2008) 105001 [5] H.-H. Klauss et al., Phys. Rev. Lett. 101, 077005 (2008) [6] J. J. Yeh and I. Lindau, Atomic Data And Nuclear Data Tables 32, 1-155 (1985)

199

Temperature dependence of sublattice magnetization in Heusler alloys
M. Kallmayer, P. Pörsch, T. Eichhorn, H. Schneider, C.A. Jenkins, G. Jakob, H. J.Elmers
Institut für Physik, Johannes Gutenberg-Universität Mainz, D-55128 Mainz, Germany Introduction NiMn-based Heusler-type intermetallic compounds have recently attracted much attention because Ni2MnGa has shown huge magnetic shape memory effects [1]. In Ni2MnGa based Heusler compounds a martensitic transformation from a L21 parent phase to a martensite phase of lower symmetry occurs at a specific temperature [2]. The martensitic transformation temperature (Tm) of Ni2MnGa based compounds, being below room temperature for the stoichiometric Ni2MnGa compound (202 K), is very sensitive to the composition and increases up to 410 K for slightly off-stoichiometric compounds Ni1.96Mn1.22Ga0.82. The addition of further constituents, i.e. Fe and Co [3] can remarkably change Tm and further magnetic properties, too. Theoretical investigations [4,] revealed the importance of the electron to atom ratio e/a for the martensitic transformation. A recent X-ray absorption spectroscopy (XAS) study [5] uncovered the characteristic change of the density-of-states (DOS) function in Ni2MnGa at Tm as predicted in Ref. [6]. Here, we apply circular dichroism in XAS (XMCD) to investigate changes with temperature and composition of element-specific magnetic moments and of the DOS in order to test theoretical predictions [6,7] Experimental Single crystalline Ni2MnGa based films were prepared as described in Ref. [5].The Ni2MnGa and Ni1.96Mn1.22Ga0.82 films were sputtered from targets with the denoted composition. The (Ni2MnGa)1−x(Co2FeSi)x films were prepared by sequential sputtering of thin layers with subsequent annealing. XAS revealed an intermixing of the latter samples. Results reported here were derived from transmission measurements unless stated otherwise [8]. The magnetization of the sample was switched at each energy step by applying a magnetic field of 1.6 T perpendicular to the sample. The absorption constant is calculated by µd = ln(I/Iref), where d is the thickness of the sample, µ the absorption coefficient and Iref the reference luminescence signal linearly extrapolated from energies below the respective L3 absorption edges. Results for Ni2MnGa Ni2MnGa/Al2O3(11-20) films show a martensitic phase transition at Tm = 275 K [5] as determined by Xray diffraction. A satellite peak observed at 3.8 eV above the absorption edge for the austenitic phase is nearly suppressed in the martensitic state (see bottom panel of Fig. 1).To first order absorption intensity at

Figure 1. Difference of the Ni XAS spectra (black line) measured at the indicated temperature and at low temperature (115 K) indicated by a color code (blue means negative values, yellow means positive values. Ni XAS spectra for 115 K (blue line) and 293 K (red line) and the corresponding difference spectrum (black line) is shown below.

200

the L3 absorption edge is proportional to the DOS of unoccupied states above the Fermi edge. The same holds for the L2-edge which is, however, broadened compared to the spectra at the L3 edge. Accordingly, the observed change of the spectra can be explained by a change of the electronic structure. The suppression of the satellite peak in the martenistic state is due to the lift of degeneracies in the Ni 3d related unoccupied electronic states as predicted by ab initio calculations [6] . For a detailed study of the spectral change we measured XAS for a series of temperatures covering the phase transition region around Tm with data acquisition at 3 K steps. Fig. 1 emphasizes the temperature dependence of the electronic structure in the region of Tm. The Fermi edge is located near the initial increase at the L3 edge. The difference of the high and low temperature XAS signal reveals a minimum close to EF and an additional maximum above the L3 absorption maximum in addition to the appearance of the 3.8 eV satellite discussed before. The initial minimum of the difference indicates a decrease of the local DOS at EF when the system enters the high temperature cubic phase. This is also in agreement with the theoretical prediction [6]. In order to follow the temperature dependence we plot the difference as a color map with minima indicated in blue and maxima indicated in red. The prominent feature is the rise of the 3.8 eV satellite close to Tm = 275 K. In the same temperature region the minimum at EF reveals a subtle shift of 250 meV to higher energy. The shift of the minimum indicates a shift of electronic states across the Fermi edge which might be the origin of the martensitic phase transition. Result for Ni1.96Mn1.22 Ga0.82 Due to the increased Mn content the martensitic phase transition in Ni1.96Mn1.22Ga0.82 films, increases to Tm = 410 K. The magnetic moment of both Ni and Mn is smaller for the non-stoichiometric compound. The sum rule analysis results in a total reduction of of the magnetic moment of 40 % at 300 K and 10 % at 115 K, also indicating a stronger temperature dependence for Ni1.96Mn1.22Ga0.82 (see Table 1). The ratio of the Ni and Mn spin moments of 0.12 remains the same for Ni2MnGa and Ni1.96Mn1.22Ga0.82. A

reduction of magnetic moments for non-stoichiometric Ni2MnGa compounds was predicted by theory [9]. According to this model the magnetization should decrease linearly with the deviation
Table 1. Magnetic moments (in µB per atom) for Mn and Ni in Ni2MnGa and Ni1.96Mn1.22Ga0.82 films as derived from a sum rule analysis. The summarized magnetic moment per formula unit µsum is compared to the magnetic moment per formula unit µmag measured by SQUID(VSM) magnetometry.

from the average number of valence electrons per atom, e/a, starting from a maximum value of 4 µB per atom at e/a = 7.5 for Ni2MnGa. For Ni1.96Mn1.22Ga0.82 one accordingly expects a value of 3.5 µB per atom [9]. The relative decrease, 3.5/4, is in rough agreement with our result for films, 3/3.3, however, absolute values are considerably smaller for both compounds.
Result (Ni2MnGa)1−x(Co2FeSi)x A composition of (Ni2MnGa)1-x(Co2FeSi)x represents a mixture of two Heusler compounds with an equal number of valence electrons per atom, e/a = 7.5. For a (Ni2MnGa)1-x(Co2FeSi)x film Tm = 250 K is reduced compared to the Ni2MnGa film. We attribute this reduction to the influence of the Co2FeSi content favoring the cubic structure. The temperature dependence of magnetic moments shown in Fig. 2 reveals a similar behavior as in the case of Ni2MnGa films (see Ref. [5]). The step-like increase of the Ni spin moment and the peak-like maximum of the Ni orbital moment is well reproduced although the latter appears strongly suppressed. Contrarily the spin moment shows a pronounced maximum and the orbital moment a minimum near Tm. This behavior might be induced by the lowest atomic symmetry present just at the phase transition. The Co and Fe spin moments are smaller compared to moments measured for Co2FeSi films (µspin(Fe) = 2.5 µB, µspin(Co) = 1.3 µB), while the orbital moments are increased. Smaller

201

spin moments could be intuitively expected for Co and Fe since they are interdiffused into the Ni2MnGa host with smaller average magnetization.

Figure 2.(a) Effective spin moments per d-hole, Nh, for Mn (red squares), Co (blue circles) and Ni (cyan triangles) and (b) orbital to spin momentum ratio for Mn (red squares), Co (blue circles) and Ni (cyan triangles) for a 100 nm (Ni2MnGa)0.975(Co2FeSi)0.025 (110) film on Al2O3(11-20). (c) The increase of the peak area (after linear background subtraction) of the satellite peak A observed 3.8 eV above the Ni L3 (cyan triangles) and Co L3 (blue circles) absorption edge indicates the martensitic phase transition. The temperature range of increasing peak area is marked in gray.

In summary we find an almost temperature-independent ratio of sublattice magnetization in Ni2MnGa based compounds. An increase of the Mn concentration reduces the Mn and Ni moments. An increase of the Co2FeSi content leads to an increase in the Mn and Ni moment, and to a decrease in Tm. The authors would like to thank for financial support from the Deutsche Forschungsgemeinschaft (Ja821/3-1 within SPP1239 and EL-172/12-2) and the BMBF (ES3XBA/5) for financial aid and S. Cramm for support at BESSY. [1] K. Ullakko, J. K. Huang, C. Kantner, R. C. OHandley, and V. V. Kokorin, Appl. Phys. Lett. 69 1966 (1996). [2] P. J. Webster, K. R. A. Ziebeck, S. L. Town, and M. S. Peak, Philos. Mag. B 49, 295 (1984). [3] V. V. Khovailo et al., JSME Int. J., Ser. A 44, 2509 (2003). [4] A. T. Zayak, W. A. Adeagbo, P. Entel, and K. M. Rabe, Appl. Phys. Lett. 88, 111903 (2006). [5] G. Jakob, T. Eichhorn, M. Kallmayer, and H. J. Elmers, Phys. Rev. B 76, 174407 (2007). [6] J. Enkovaara, and R. M. Nieminen, J. Phys.: Condens. Matter 14, 5325 (2002). [7] M. Lezaic, Ph. Mavropoulos, J. Enkovaara, G. Bihlmayer, and S. BlÄugel, Phys. Rev. Lett. 97, 026404 (2006). [8] M. Kallmayer et al., Appl. Phys. Lett. 88, 072506 (2006). [9] J. Enkovaara, O. Heczko, A. Ayuela, and R.M. Nieminen, Phys. Rev. B 67, 212405 (2003).

202

Residual stress measurements on PVD-multilayers with various designs H.-A. Crostack, U. Selvadurai-Laßl, G. Fischer Lehrstuhl für Qualitätswesen; Universität Dortmund, Joseph-von-Fraunhofer Str. 20, 44227 Dortmund, Germany 1. Introduction PVD multilayer systems are used as hard coatings of cuttings tools [1]. Due to the coating process and the mismatch of thermo-elastic properties between the metallic and ceramic layers of multilayer system and the substrate (Young’s modulus, Poisson ratio, thermal expansion coefficient) high residual stresses and residual stress gradients are induced. These stresses and stress gradients can result in microcracks and the delamination of the coatings. In consequence, the life time of concerned cutting tools will be shortened. A better knowledge about the correlation between the process parameters and the residual stresses is a key factor to prevent this life time reduction. For that reason, PVD-multilayer systems were manufactured with various designs [2] and the residual stresses were analyzed in the layer materials and the substrate using Energy Dispersive Diffraction method available at beamline EDDI of BESSY II. 2. Materials and Methods The residual stress measurements were performed on the ceramic/metal multilayer system TiAlN/Ti manufactured with the four designs schematically shown in Fig. 1

Fig. 1: Designs A - D of analysed PVD-multilayer systems CrAlN/Cr and TiAlN/Cr, ceramic layers -red, metal layer - white, steel substrate - gray, the thickness ratio of layer types is drawn realistically In case of design A (5 ceramic and metal layers), the metal layer thickness was varied in the range from 10 to 60 nm at a fixed thickness of ceramic layers (500 nm). The multilayer design B (25 ceramic and metal layers) were only studied with the thicknesses of 100 and 10 nm of ceramic and metal layer, respectively. Design C is a multilayer system with graded thickness of ceramic layers and a fixed metal layer thickness (10 nm). In contrast to design C, the metal or amorphous ceramic layer was 200 nm thick. The surface of most substrates were uniform mechanical treated before depositing the multilayer system. To study the effect of mechanical

203

treatment on the residual stress, the treatment was finished after various grinding and polishing steps for selected. TiAlN/Ti samples with design A. For X-ray diffraction stress analysis on these multilayers, the materials science beamline EDDI was chosen, which provides X-rays in an energy range between about 10 and 120 keV [3]. The measurements were performed in the energy-dispersive mode which yields diffraction spectra with a multitude of reflections for fixed positions of ω (6°) and 2θ (12°). The residual stresses were analysed by means of the sin2ψ -method [4] using X-ray elastic constants calculated from the single crystal constants. Each reflection E(hkl) in the spectrum yield another average information depth and allowed for a depth-resolved analysis of the near surface residual stress fields by extending methods like the multi-wavelength method [5] to the energy-dispersive case of diffraction. By means of these methods, information on the phase-specific in-plane residual-stress state within both layer materials became available. Furthermore, the depth-depending stress states in the subsurface regions of substrate was obtained, too. The subsurface region belongs to the critical failure zones of PVD-coated tools. 3. Results The diffractograms (Intensity/E) reveal that the crystalline portion of the studied nanocrystalline ceramic and metallic layers is high. Even reflections of 10 nm thin metal layers can be detected. In all samples, the reflections (111) and (200) of TiAlN exist. In all multilayer systems high compressive residual stresses were found in the ceramic layers (-1.5 and to -5 GPa). The residual stresses in metal layers are also compressive but much lower than in the ceramic layers (-50 and -200 MPa). This effect is due to the stress relaxation by plastic deformation. A reduction of surface roughness of substrate surface by stepwise finer grinding and polishing before the deposition of the TiAlN/Ti multilayer system increases the compressive residual stress in the TiAlN layers and decreases the compressive residual stress in the substrate region near the coating (Fig. 2). But here the stresses evaluated for TiAlN (111)-reflection are lower than for the TiAlN (200) reflection. Since in the steel substrate various Fereflections could be evaluated, the depth profile of residual stress were analysed near the substrate-multilayer interface. Near the interface similar compressive stresses of about 260± 13 MPa exist in substrate after both preparations. The plot shows that the compressive stresses diminish with increasing distance to the interface (Fig.2). In the sample grinded with 1200 mesh hard particles (open symbols) the stress gradient is steeper. The compressive stress diminishes down Fig. 2: Effect of substrate preparation on the to – 60 ± 8 MPa (grinding) and 99 ± 10 residual stresses in the TiAlN layer of the MPa (grinding and polishing), multilayer system TiAlN/Ti and the steel respectively. substrate, (500 nm ceramic – 10nm metal) Correlation between stresses of coating and substrate yields to an influence of substrate preparation on the complete sample. Polishing leads to a decrease of stress values of the ceramic layers, but to an increase of stress values in deeper substrate regions. Further investigations on substrate without any preparation are needed to interpret this effect.

204

The influence of designs A, B and C was analysed on samples with various TiAlN layer thicknesses (Tab. 1). In thick coating layers with the same thickness (A) the residual stress of ceramic layers are lower than of thin layers (B). But in graded layers (C) the highest value was found. The stress values of metal layer and subsurface substrate behave contrary. Thus in metal and substrate of graded layers the smallest stress values exist. In thin layers the crystal growing process is interrupted in an early stadium and this inhomogeneous state increase residual stresses. Tab. 1: Residual stresses of TiAlN/Ti multilayers with different design A, B and C. Design Layer Thickness of ceramic / metal 500nm / 10nm-5x 100nm / 10nm-25x Gradiert / 10nm-5x Coating thickness [nm] 500 2500 1500 σ11 Fe 256±11 152±12 110±13 σ11 TiAlN (111) -1528±122 -3890±117 -5012±100 σ11 Ti (002) -191±43 -144±35 -74±27

A B C

Multilayer systems with various ceramic and partially amorphous ceramic layer thickness exhibit higher RS (Tab. 2) than systems with thick (500 nm) ceramic and thin metal layers (see design B of Tab. 1). Tab. 2: Residual stresses of TiAlN/amorph TiAlN multilayers with different design B and D. Design B D thickness TiAlN [nm] 25x100 nm Σ 1500 thickness TiAlN amorph [nm] 25x10 nm 5x200 nm σ11 [MPa] Ferrit -275±26 -300±12 σ11 TiAlN 111 -4475±64 -4022±46

4. Acknowledgement For assistance during beamtime at BAM-line we would like to thank Prof. Christoph Genzel and Manuela Klaus. We thank Dr. E. Vogli for providing the multilayer samples. 5. References
[1] [2] [3] [4] [5] T. Hanabusa, K. Kusaka, T. Matsue, M. Nishida, O. Sakata, T. Sato; JSME, Series A, Vol. 47, No. 3, (2004) 312- 317 W. Tillmann, E. Vogli: Adv. Eng. Mat., 10, No. 1-2, (2008) 79-84 M. Klaus, I.A. Denks, Ch. Genzel; Mat. Sci. Forum Vol. 524-525 (2006) 601-606 Ch. Genzel, C. Stock and W. Reimers, Mat. Sci. Eng. A 372 (2004), p. 28. Hauk, V., Macherauch, E. (Hrsg.): Eigenspannungen und Lastspannungen (HTM Beiheft), Carl Hanser Verlag München, Wien (1982)

205

Direct measurement of the electric field in semiconductor junctions using high energy, high resolution x-ray photoemission (HIKE).
Allsop, N.A.; Lauermann, I.; Mönig, H.; Gorgoi, M.; Fischer, Ch.-H.; Grimm, A.; Johnson, B.; Kropp, T.; Lux-Steiner, M. Ch.; Helmholtz Zentrum Berlin für Materialien und Energie.

One of the key parameters in semiconductor devices is the built-in electric field or band bending which results from the formation of an electronic junction. Band bending is often characterised by capacitance-based methods, however as the defects become more complicated (e.g in thin film solar cells) the interpretation becomes ambiguous. Traditional PES spectroscopy can only indirectly monitor the band bending in incomplete semiconductor junctions. We have used high kinetic energy PES (HIKE) to directly determine the band bending in completed model semiconductor junctions, including measurements under in-situ applied bias to control the band bending. High kinetic energy photoemission measurements allow material inspection down to 20 nm depth with high resolution. When measuring semiconductors the emission can change by several 100meV through this depth due to the electric field/band bending. Band bending measurements are not new to the PES community. The shift of the peaks with different doping concentrations [1] or the shift of peaks upon deposition of thin overlayer [2] is often used to follow the band bending. However, these methods involve comparing measurements taken in the presence of unknown surface charges and work function, which are usually presumed to remain constant. Due to the limited depth of conventional PES the normal measurements are rarely able to look at completed semiconductor junctions and therefore the defect occupation can also be substantially different. Sensitive measurements of the peak shifts are also complicated by changes in the work function or shifts due to realignment of the beam between samples. Therefore we also measured the broadening by fitting the shape of the photoemission peaks, a method which previous studies have argued is not realisable for conventional PES [3].
1.12 1.11 1.10 1.09 1.08 1.07 1.06 1.05 1.04 1.03 1.02 1.01 1.00 0.00 0.05 0.10 0.15 0.20 0.20 0.18 0.16

Peak shift (eV)

FWHM (eV)

0.14 0.12 0.10 0.08 0.06 0.04 0.02 0.00 0.00 0.05 0.10 0.15 0.20

Electric field (V/IMFP)

Electric field (V/IMFP)

 

Figure  1.  The  theoretical  broadening  and  peak  shift  associated  with  the  electric  field  in  Volts  per  Inelastic  Mean  Free  Path. A 1eV starting FWHM has been assumed but the field can be normalized to any FWHM.

The samples used were GaAs:Au Schottky contacts (Fig.2). Gold is known to pin the Fermi level at the surface of n-type GaAs to induce a built in voltage of approx. 0.9V. In order to measure the Ga2p and As2p in the completed junction, a very thin 5nm layer of gold was used. In conventional PES this would completely absorb the photoelectrons emitted from the GaAs, however with HIKE we can use an excitation energy of 6keV. The 5nm gold layer is also thick enough to allow a voltage to be applied in-situ.

206

Energy

5nm Au i - GaAs n - GaAs
V

GaAs
Distance

Au

Figure 2. GaAs‐Au Schottky contacts

The peak fitting procedure was performed by using the measured peak from the intrinsic GaAs, where no significant field is present, and convoluting this peak with an exponential decay function. The decay function represents Experimental Intrinsic GaAs - smoothed the shift in the peak position as the depth Experimental data 7E17 GaAs Fit increases. An example of the fitting procedure is shown in Fig. 3. In addition to the intrinsic sample, three n-type Ga2p GaAs wafers were measured with varying doping concentrations. The doping concentration was independently measured using a capacitance method (Mott-Schottky plot) and the theoretical electric field at the interface calculated using textbook semiconductor 4906 4907 4908 4909 4910 equations. The agreement between the field as Kinetic Energy (eV) measured using the peak broadening and the theoretical values based on the capacitance measurements is shown in Fig 4.
Figure 3. Example of a fitted curve 
0.25 For the highly doped sample with a strong Ga2p Fit Theory electric field the agreement is excellent but for 0.20 As2p Fit the lower doped sample the agreement 0.15 between the value extracted from the Ga2p and As2p peak is not as good. However, it should 0.10 not be forgotten that even determining the field 0.05 strength to within a factor of 2 is often very useful for electronic devices. 0.00 1E16 1E17 1E18 In order to further investigate the sensitivity of -3 GaAs doping (cm ) the technique the peaks were measured under an in-situ applied bias, in order to directly manipulate the electric field at the junction. This time both the shift in the peak position and the Figure 4. Fitted and theoretical parameters for  different GaAs doping levels.  broadening was used to estimate the electric field. The results are displayed in Fig 5.

Intensity a.u.

207

Electric field (V / IMFP)

0.20

Peak Broadending 1x10 doped GaAs Electric field (V / IMFP)
0.08

17

Peak Broadending 2x10 doped GaAs

16

Electric field (V / IMFP)

0.15

0.06

0.10

0.04

0.05

Ga2p Theory As2p 0.0 0.5 1.0 1.5 2.0 Applied Voltage
17

0.02

Ga2p Theory As2p 0.0 0.5 1.0 1.5 2.0 Applied Voltage 2.5

0.00

0.00

2.5

Peak Shift 1x10 doped GaAs
0.25 0.08

Peak Shift 2x10 doped GaAs

16

Peak Shift (V)

0.20 0.15 0.10 0.05 0.00

Peak Shift (V)

0.06 0.04

Ga2p Theory As2p 0.0 0.5 1.0 1.5 2.0 Applied Voltage 2.5

0.02

Ga2p Theory 0.0 0.5 1.0 1.5 2.0 Applied Voltage 2.5

0.00

 

Figure 5. Fitted and theoretical parameters for sample with an in‐situ applied voltage. Note that for the peak shift values  the slope is most important parameter, a y‐axis offset between samples can occur as  result of work function changes. 

The trends within the plots are clear to see and validate the method as a way of measuring the electric field. However, the number of points which deviate from the trend show that extreme care needs to be taken when making the measurements. We used the 3rd order excitation of the 111 Si crystal on the KMC-1 beamline, with a Be window to reduce the thermal load on the crystal monochromator. This gave the excellent stability which allows shifts of only 10-20meV to be measured under an excitation energy of 6030eV! Any realignment of the system must then be followed by remeasurement of a gold reference peak. In conclusion we have demonstrated that the measurement of the electric field in a completed semiconductor junction is possible using HIKE. This should provide a new tool to investigate structures such as thin film solar cells and other electronic devices. The technique places a high requirement on the stability of the beam, which the KMC-1 beamline at BESSY is able to provide. Additionally it is clear that when interpreting high kinetic energy PES data of semiconductors and insulators, the effects of band bending and electric field need to be taken into account.      References  [1] W. Eberhardt, G. Kalkoffen, C. Kunz, D. Aspnes, and M. Cardona, Phys. Stat. Sol. (b) 88, pp 135‐ 143, (1978).  [2] E.A. Kraut, R.W. Grant, J.R. Waldrop, and S.P. Kowalczyk. Phys. Rev. Lett, 44, 24, pp1620‐1622  (1980).  [3] G. Margaritondo, F. Gozzo and C. Coluzza, Phys. Rev. B, 47, 15, pp 9907‐9909 (1993)   

208

Relative Sub-shell Photoionization Cross-sections of Nickel Determined by Hard X-ray High Kinetic Energy Photoelectron Spectroscopy M. Gorgoi, F. Schaefers, W. Braun, W. Eberhardt
Helmholtz Zentrum Berlin, BESSY II, Albert-Einsteinstr. 15, 12489 Berlin, Deutschland

Recently, high kinetic energy photoelectron spectroscopy has lead to a break-through due to its non destructive way of investigating the bulk electronic properties of materials. However, due to the relatively new development of this technique there is a lack of information concerning the photoionization cross section at high energies. Currently, only calculated atomic cross section data are available [1-4] which have not yet been verified experimentally. In photoemission, the sub-shell photoionization cross-section is one of the factors that determine the intensity of particular spectral lines and their satellites. In the hard x-ray regime the core level cross-sections are decreasing rapidly, a fact which hinders the experiment considerably. This is now compensated by the photon intensity provided by third generation synchrotron sources. In most of the photoemission experiments ratios of signal intensities or energy shifts in the core levels are followed. For the analysis of these data absolute cross section values are not essential. However, whenever compound materials are investigated or when estimating signal levels and the feasibility of an electron spectroscopy experiment the knowledge of cross sections is essential. Thus the present study concentrates on describing the sub-shell photoionization cross sections of the nickel metal.

Au 4f
2 keV Intensity [arb. units] 3 keV 4 keV 5 keV 6 keV 7 keV 8 keV 9 keV
92 90 88 86 84 82 80 78 76 74 Binding energy [eV]

Ni 3p

72

70

68

66

64

62

Figure 1. The Ni 3p and Au 4f core level spectra recorded at energies ranging from 2 keV to 9 keV. The normalized spectra were shifted manually on the intensity scale for a better viewing. Thin layers of Ni were investigated using excitations energies ranging from 2 keV to 9 keV. The measurements were performed at the KMC-1 beamline employing the HIKE endstation. Spectra were recorded in normal emission geometry. The samples were prepared by evaporation in ultra high vacuum (UHV). Ni with a guaranteed purity of 99.999% was purchased from Goodfellow. The thickness of the films was controlled by means of a quartz crystal

209

microbalance. The final layer thickness was 4 nm. A thick gold film was employed as a substrate. Based on previous experimental studies of Kunz et al. [5] describing the sub-shell cross section for gold, a gold crystal Au(111) sample was also used as reference. In the Ni case the core levels characteristic to the gold substrate were still visible and thus the Au 4f core level was chosen as a reference. The 2s, 2p, 3s, 3p core levels of Ni were recorded in a stepwise like manner. A step of 1 keV in the excitation energy was employed. Moreover, for a thorough intensity calibration of the spectra the flux given by the double crystal monochromator was recorded using the available ionization chamber as well as a GaAs diode that was introduced in the beam for every measurement point. The following example shows the core level spectra of Ni 3p recorded altogether with the Au 4f core levels corresponding to the gold substrate (Figure 1). As a first evaluation step a normalization procedure to the intensity of the incoming flux was employed. Further on, the level of background on the low binding energy side was subtracted and then for a better viewing of the core level intensity evolution, the spectra were normalized to the high binding energy background side. The intensity evaluation was performed after a thorough fitting of the spectra and after the subtraction of a Shriley background. The peak areas were used in the signal evaluation. In the description of the photoemission signal [6] the angular anisotropy factors were Ni 3p 100000 neglected and electron effective attenuation lengths were considered the same for Au and 10000 Ni. A preliminary evaluation of the sub-shell photoionization cross sections taking into account 1000 the already published values of Au 4f cross sections [5] and the densities of gold and nickel was 100 performed. The result is shown in Figure 2. 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 For an exact estimation of the Excitation Energy [eV] sub-shell photoionization cross Figure 2. The evaluated cross section of Ni 3p as a function section values a very good of energy. assessment of the electron effective attenuation lengths [5] is required. In order to do that further investigation may be necessary.

Cross section [barn]

[1] J. J. Yeh and I. Lindau, Atomic Data and Nuclear Data Tables 32, 1-155 (1985) [2] M. B. Trzhaskovskaya, V. I. Nefedov, and V. G. Yarzhemsky, Atomic Data and Nuclear Data Tables 77, 97 (2001) [3] M. B. Trzhaskovskaya, V. I. Nefedov, and V. G. Yarzhemsky, Atomic Data and Nuclear Data Tables 82, 257–311 (2002) [4] J.H.Scofield, Lawrence Livermore National Laboratory Report UCRL-51326, 1973. [5] C. Kunz, S. Thiess, B.C.C. Cowie, T.-L. Lee, J. Zegenhagen, Nuclear lnstruments and Methods in Physics Research A 547, 73-86 (2005). [6] H. Bubert et al., eds. Surface and Thin Film Analysis. 2001, Wiley-VCH: Weinheim

210

NEXAFS investigation of graphite-carbon nanotube hybrid system
A.V. Okotruba, M.A. Kanygina, A.G. Kurenyaa, A.G. Kudashova, Yu.V. Lavskayaa, L.G. Bulushevaa, S.L. Molodtsovb a Nikolaev Institute of Inorganic Chemistry, SB RAS, Novosibirsk, Russia b Institute of Solid State Physics, Dresden University of Technology, Germany Angle-resolved near-edge x-ray absorption fine structure (NEXAFS) method has been applied for revealing graphitic layers distribution in the samples of graphene multi-layers and aligned multi-walled carbon nanotubes (CNTs) synthesized by aerosol-assistant catalytic chemical vapor deposition (CCVD) technique. Arrays of multi-walled CNTs have been grown on the oxidized silicon substrates using an aerosol-assistant CCVD method. A mixture of heptane and ferrocene (2 wt.%) was dispersed into reactor volume with a rate 2 g/h using an injector. The pyrolysis was performed at 800°С and atmospheric pressure in argon flow (150 cm3/min). The duration of synthesis was 5 min for the sample 1 and 40 min for the sample 2. Short time of the CCVD synthesis resulted in small thickness of the film (Fig. 1(a)). CNTs constituted the sample 1 have length ~1 μm and poor alignment. The thickness of the sample 2 is ~ 12 μm and CNTs have predominantly vertical orientation to the substrate surface (Fig. 1(b)). Bright lines on the image perpendicularly to the direction of CNT growth indicate non-uniform formation of the array. It could be suggested that thin horizontal layer consists of graphenes and iron carbide Fe3C as it was indicated in [1]. C K-edge NEXAFS spectra were measured at the Berliner Elektronenspeicherring für Synchrotronstrahlung (BESSY) using radiation from the Russian-German beamline. A vertical axis of sample rotation was oriented perpendicular to the electric field vector, which was in the horizontal plane of spectrometer. The spectra were taken for various incidence angles Θ. The NEXAFS data were acquired in a total electron yield mode and normalized to the primary photon current from a gold-covered grid, which was recorded simultaneously. The monochromatization of the incident radiation was better than 80 meV. Before the measurements the sample was heated in the vacuum chamber at ~150°C to remove the adsorbed molecules. The working pressure in the chamber was ~10-9 mbar. NEXAFS spectra measured at the C K-edge of the samples 1, 2 for different incidence angles are compared in Fig. 2 (a). The spectra have been normalized to their intensity at 330 eV. The resonances around 285 eV have the π∗-like character, while the features arisen within 291.5−293.0-eV interval correspond to the 1s→σ∗ transitions mainly. The change in the relative intensity of π∗-resonance characterizes orientation of a b graphitic layers in the samples produced. The inserts in the Fig. 2(a) show the variation in π∗resonance obtained for different angular positions of sample relative to the incidence beam. One can see that the samples 1 and 2 have opposite angular dependence of π∗-resonance. Increase of the incidence angle Θ causes decrease of the π∗-resonance intensity in the Fig. 1. SEM images of the cleavages of sample 1 grown at 800°C for 5 min (a) and sample 2 grown at 800°C for NEXAFS spectrum of the sample 1 40 min (b). Thin horizontal layers in the image (b) and enhancement of that in the correspond to graphene layers spectrum of the sample 2.

211

1,0

a
30 o 60 o 90
o

σ∗ π∗
30 o 60 o 90
o

0,8

0,6

π∗

S1

1,0

b
85° 75° 65° 55° 45° 35° 25° 15° 5°

S1

S2

Intensity (arb. units)

0,4

0,8

0,2

π∗/σ* (arb.units)

65° 60° 55° 50° 45°

0,6

0,0 1,0

σ∗
90 o 80 o 70 o 60 o 50 o 40 o 30
o

0,8

S2
90 o 80 o 70 o 60 o 50 o 40 o 30
o

0,4

0,6

π∗

π∗

0,2

40° 35° 30° 25°

0,4

0,0

Nanotube
-0,2 0 15 30 45 60 75

0,2

Graphite
90

0,0

280

290

300

310

320

330

Θ (deg.)

Photon energy (eV)

Fig. 2. a – Angular dependence of NEXAFS spectra of the sample 1 (S1 lines) and sample 2 (S2 lines). b – Theoretical dependencies of π*/σ* intensity ratio calculated for different angular distribution of carbon nanotubes and graphite layers in a sample and experimental points derived from the NEXAFS data for the sample 1 (S! line) and sample 2 (S2 line).

To determine the average misalignment of graphitic layers in the samples 1, 2 we used the approaches developed in [2, 3], which consider that graphite consists of crystallites and CNT array is a set of graphitic cylinders. Within these approaches the deviation of crystallites from the basal graphite plane and CNTs from the normal to the substrate surface is described by the Gaussian distributions. The angular dependences of the π*/σ*- intensity ratio calculated for different widths of orientation distribution of graphite crystallites and CNTs are presented in Fig 2 (b). Comparison of the theoretical curves with the experimental data allows making the numerous evaluation of misorientation of graphitic layers in the samples. The angular dependence of the relative intensity of π*-resonance derived from the NEXAFS data of the sample 1 has behavior similar to that of strongly disordered graphite. The width of the angular distribution of graphitic layers equal to ~62° corresponds to the deviation of graphene planes from the silicon surface ±31° in average. The experimental points obtained for the sample 2 follow the dependence expected for aligned CNTs with ~55°-distribution of nanotubes in array (±22.5° deviation from the vertical to the silicon substrate). This dependence is actually characteristic of array of multi-walled CNTs, having a lot of defects and intrinsic texture of the graphitic shells composing a CNT. The angular dependence of π*/σ*- intensity found for the sample 1 was surprised and it undoubtedly indicates formation of graphene layers in the result of CCVD synthesis of CNTs. In summary, using the angle-resolved NEXAFS method we detected formation of graphitic layers in process of CNT array production. The thickness and perfectness of graphenes could be controlled with changing the parameters (time, temperature, etc.) of the CCVD method. The work was supported by the bilateral Program “Russian-German Laboratory at BESSY”. [1] D. Kondo, S. Sato, Y. Awano, Appl. Phys. Express 1 (2008) 074003. [2] V.V. Belavin, A.V. Okotrub, L.G. Bulusheva, A.S. Kotosonov, D.V. Vyalykh, S.L. Molodtsov, J. Exp. Theor. Phys. 103 (2006) 604. [3] A.V. Okotrub, V.V. Belavin, L.G. Bulusheva, A.V. Gusel’nikov, A.G.Kudashov, D.V. Vyalikh, S.L. Molodtsov, . Exp. Theor. Phys. 107 (2008) 517. 2
212

Local atomic and electronic structure of ferromagnetic AlN_Cu nanorods.
authors: A.V.Soldatov, A.Guda, M.Soldatov institute: Southern Federal University, Rostov-on-Don, Russia. Nanostructured AlN – is promising material for nanoelectronics. Its unique properties are well known and can be found somewhere [1]. One can control semiconductor properties of this material and create new ferromagnetic properties by means of doping with transition metal atoms [2,3]. There are already reports about room-temperature ferromagnetic properties of AlN, doped by Fe, Cr and Mn. But there still no realistic theory that can describe ferromagnetism in all kinds of diluted magnetic semiconductors. The main problem is that it is very hard to obtain exact position of dopping atoms in a host lattice. X-ray absorption spectroscopy is unique tool for solving of such problem.

Figure 1. AlN nanorods on Si substrate (a) and Cu doped AlN nanorods on Si substrate (b)

Figure 2. HRTEM image of single AlN nanorod (a) and Cu doped AlN nanorod (b)

AlN:Cu nanorods were prepaired in tube furnace on Si (100) substrate in the presence of CuCl2 during reaction of heating as copper source [5]. The typical diameter and length of the nanorods, as follows from Figure1, is around 20-50 nm and 1 µm respectively. The concentration of the Cu atoms is about 3-5%. The spontaneous magnetization and the coercivity of the AlN:Cu are about 0.38 emu·cm-3 and 100 Oe respectively. It is important to note that AlN:Cu material is highly resistive at room temperature (>105 Ω cm). AlN:Cu nanorod is a single wurtzite AlN crystal with growth direction (0001). HRTEM images of a single nanorod (Figure2) reveal lattice distortions after incorporation of copper atoms. Cu L2,3 XANES spectra were recorded in a UHV chamber attached to Russian-German beam line (RGBL) of the third generation synchrotron BESSY in Berlin. We used total electron yield (TEY) for absorption signal measurements. Complementary to Cu L2,3 edge also Cu Kedge spectrum from the same sample was measured in NSRL, Hefei, China. X-ray absorption spectra were simulated, using self-consistent full multiple scattering theory (program package Feff 8.4). It was found that time-dependent LDA works well with Cu L2,3 edges. Magnetic properties of point Cu defects in AlN host lattice using LAPW approximation implemented in Wien2k package were also calculated. Theoretical calculations on Figure3 show that the Cu L2,3 spectrum of AlN_Cu nanorods can not be explained by only Cu point defects, that are embedded into host lattice of AlN. We supposed that small copper clusters can grow during AlN nanorods growing in the presence of CuCl2 in a tube furnace. Theoretical simulations of spectra from clusters in AlN still not finished, but meanwhile we decided to obtain complementary information about atomic structure around Cu atoms from the Cu K-edge. Spectrum for small free Cu clusters was taken from ref [6]. As can be seen from Figure 4 in AlN_Cu nanorods there is a strong probability for cluster creation, but they can have different sizes – from several atoms to several hundred atoms, which resemble bulk copper. A theory of

213

multidimensional interpolation of spectrum [7] is to be applied for this problem to answer the question about average size of copper clusters inside AlN lattice.
AlN_Cu experiment
1,2
A B C D

AlN_Cu experiment small clusters experiment metallic Cu experiment

Absorption, [rel.un.]

0,8

Abs.coeff., [rel.un.]

Theory, Cu interstitial Theory, Cu in Al site
0,0

0,8

0,4

Theory, metallic Cu Metallic Cu experiment
-0,8 932 936 940 944 948

0,0 8960 8980 9000 9020 9040 9060 9080 9100

Energy, [eV]

Figure 4. Experimental XANES spectra for Cu K-edge for AlN_Cu nanorods, small Cu clusters and metallic Cu
Energy, eV

Figure 3. Experimental XANES spectra for Cu L2,3 edges for metallic Cu and AlN_Cu nanorods vs theoretical simulations

Finally, spin-polarized band-structure calculations predict, that among all point Cu defects, that are possible in AlN lattice – Cu in N site, Cu in Al site, Cu interstitial – only Cu in Al site is responsible for magnetic moment. We continue our calculations for Cu clusters, embedded in AlN, and first results also predict that they will be magnetized, due to p-d hybridization of Cu atoms and neighboring nitrogen atoms around cluster. References. 1. A. Siegel, K. Parlinski, and U. D. Wdowik, // Phys. Rev. B – 2006, 74, 104116. 2. Y. Yang, Q. Zhao, X. Z. Zhang, Z. G. Liu, C. X. Zou, B. Shen, and D. P. Yu, // Applied physics letters 90, 092118. - 2007. 3. M.B. Kanoun, S. Goumri-Said, // Physica B 403. – 2008, 2847–2850. 4. Z G Huang, R. Wu and L J Chen // J. Phys.: Condens. Matter 19, - 2007, 056209 – Cu substitutional 5. X.H.Ji, S.P.Lau, S.F.Yu, H.Y.Yang, T.S.Herng and J.S.Chen, 2007. Nanotechnology 18, 105601 6. Sumedha Jayanetti, Robert A. Mayanovic et.al. Journal of Chemical Physics V.115 №2, 2001 7. G. Smolentsev, A.V. Soldatov, Computational Materials Science, 39, p.569, 2007 This work has been supported by the Ministry of Education and Science of Russia (project RNP 2.1.1.5932). Authors would like to thank Maria Brzhezinskaya, S.L.Molodtsov and S.I.Fedoseenko for help in organizing of XAS experiments and BESSY for partial financial support.

214

Synchrotron radiation prevents train crash.
authors: A.V.Soldatov, V.I.Kovesnikov, A.Guda, M.Soldatov institute: Southern Federal University, Rostov-on-Don, Russia. Train is the most safety transport. A big work was made to minimize the possibility of crash. But there still some problems that make a lot of harm to railwaymen. One of them is connected with brake blocks. Initially, brake blocks were made from the same material as wheels. This led to fast wear and tear of wheel and to big costs on its repair and replacement. Nowadays soft materials for brake blocks are preferable, such as composites. Last ones usually consist of graphite, synthetic rubber, aluminum oxide and sulfur compounds. Such contents significantly increase the lifetime of wheel but lead to unexpected consequences. During exploitation specific defects arise, because composite materials are aggressive for steel. These defects arise from redistribution of temperature fields on the wheel rim under friction. So, radial cracks, metal dislocations appear and wheel can derail. In order to investigate, what chemical reactions take place during the process of friction, the x-ray absorption spectra of corresponding materials were measured. This methodic was already successfully applied to exploration of complex chemical interactions between additives in engine oil and metallic surfaces [1]. We focus our attention on definite object – train wheel from crash from railway museum. We took pure steel from the wheel far from its surface – sample1 on Figure1. As can be seen from Figure1, sample3 – damaged steel during friction – has a scaled structure like mica. Millimeter-sized scales were taken from damaged wheel surface, which was in contact with brake block. We also studied a material that adhered to the material of brake block – sample2.

Figure 1. Sample1 – initial state of steel, Sample2 – scaled structure formed on the wheel surface during friction, Sample3 – part of wheel that adhered to the brake block Spectra were recorded at Russian-German beamline RGBL in high vacuum chamber. All samples were maintained on copper sample holder and a good conducting contact was provided between sample and sample holder. We used total electron yield signal from sample. Size of xray beam in experiment was about 50x100mkm2. Energy resolution was E/∆E~5000. We studied different points on a sample to find out does it sensitive to heterogeneity of the surface. We also made element analysis to see the distribution of such elements as Cr, Ni, Cu and Mn. Amount of these elements tells us how steel state is changed during its exploitation. This information is additional to the information about Fe atoms chemical state, analyzed with x-ray absorption

215

spectroscopy. Quantitative analysis of Cr, Ni, Cu and Mn distribution was made in BAM institute by means of laser induced breakdown spectroscopy [2] and results of this analysis are presented in Table1.
B

1,2 1,0

Fe L23 edges

Absorption, [rel.un.]

0,8 0,6

A C D

0,4 0,2 0,0 -0,2 -0,4 -0,6

Sample-3 Sample-2 Sample-1

705

708

711

714

717

720

723

726

729

Energy, eV

Figure 2. XANES spectra Fe L2,3 edge form initial state of wheel steel (sample1), scale from wheel surface after long time friction (sample2 and sample3). Figure2 represents the Fe L2,3 XANES spectra for initial state of steel (from a railway wheel) and steel after the exploitation. We can clearly identify the changes that took place during the reactions. Sample-2 and Sample-3 have identical spectra with small changes in relative intensities of features C and D. Intensities of peaks A and B shows chemical state of Fe atoms [3] within thin layer near surface. One can say that during a friction of the wheel with the brake blocks due to high temperatures and high pressure a chemical reaction between wheel steel and composite material takes place. Scaled structure is formed and only then follows mechanical separation of the scales. Part of them forms a layer on a wheel and other part adheres to brake block. We suppose that certain fraction of Fe atoms react with sulfur atoms of composite material, that diffuse due to high temperature from volume of brake block. Having high temperature steel surface has a contact with the air, so part of Fe atoms oxidizes, that can be seen from the features A and B in Figure1. Principle component analysis within multidimensional expansion [4] of these spectra can be applied to identify the phases that present in samples. Now we continue calculations of Fe L2,3 edges within multiplet theory [5]. Table 1. Fractions of Cr, Ni, Cu and Mn atoms in samples. Sample3 was extremely inhomogeneous, which can be seen from intervals for fractions. Cr(%) Ni(%) Cu(%) Mn(%) Sample1 0.25 0.2 0.17 1.2 Sample2 0.1 0.38 0.17 3 Sample3 0.03-0.25 0.02-0.5 0.03-0.13 0.6-2.8 Literature 1. M.Nicholls, M.N.Najman, et,al., Can.J.Chem. 85:816-830, 2007 2. http://www1.messe-berlin.de/vip8_1/website/MesseBerlin/htdocs/www.laseroptics/ pdf/1.4_Panne_-_10.20.pdf 3. S. Anders, M.Toney, et,al., J. Appl. Phys., Vol. 93, No. 10, 15 May 2003 4. G. Smolentsev, A.V. Soldatov, Computational Materials Science, 39, p.569, 2007 5. Frank de Groot, Coordination Chemistry Reviews 249 (2005) 31–63 Acknowledgements: Authors would like to thank Igor Gornushkin from BAM institute for help in organizing element analysis experiments.

216

Scanning SAXS/WAXS measurements of dentine and bone sections
P. Zaslansky, A. Maerten, C. Lange, C. Li, S. Siegel, P. Fratzl, O. Paris Max Planck Institute of Colloids and Interfaces, Research Campus Golm, 14424 Potsdam The structure of mineral crystals in human teeth and murine bone was studied using the scanning SAXS/WAXS and XRF setup at the µ-spot beamline [1]. SAXS measurements of human-tooth dentine sections Dentine in human teeth, is known to have a graded structure, that varies by depth, as one moves away from the enamel and into the tooth bulk (closer to the pulp) [2,3]. Composed of mineralized collagen fibrils, dentine exhibits a highly anisotropic structure, with micron-sized tubules running outward from the pulp, radiating in all directions through the tissue towards enamel. Dentine is tough, withstands many years of service in the mouth, yet much remains unknown about how the variation in the structure contributes to the excellent performance. We have used a micron-sized beam on the μ-Spot beamline to look closer at the structural variations. We prepared and scanned thin (100 micron thick) slices from different sides of teeth, to study how the microstructure at the lengthscale of the mineral crystals (carbonated hydroxyl-apatite crystals sized 30~50 nm) varies spatially. Variation in dentine seems to follow the outer curvature of enamel, with gradual transitions between particle orientations. The degree of particle co-alignment (Rho parameter, Fig 1, left) also varies spatially: crystals are not well aligned. The thickness T parameter, which gives a good indication about the thin dimension of the mineral platelets also varies spatially and systematically (Fig 1, right): particles become smaller at greater distances from the dentin-enamel junction. Such systematic evaluations of the particle arrangement in representative tooth samples will help to better understand the microstructure and how it is adapted to the loading forces in teeth.

Fig. 1: (Left) Rho parameter values (color coded) where 1 indicates perfect crystal co-aligning and 0 indicates complete randomness of particle orientations. The orientation is indicated by small bars. (Right) T parameter indicating the particle thickness in it's thinnest dimension. Enamel is not shown but is located on the upper right side of the sample.

Mineral deposition and growth during embryonal bone development in mice The knowledge how mineral crystals in bone nucleate, grow and organise themselves from the “birth of bone” (embryonal) is still poor. Therefore the aim of this study is to understand the development of the

217

mineral properties, the mineral deposition and organisation in growing bones (tibia, femur) from embryonal to mature stages using mouse as a model system (strain C57 BL6). This work is carried out in cooperation with the group of Prof. Stefan Mundlos (Charité and MPI for Molecular Genetics). We used a multi-method approach including optical microscopy and environmental scanning electron microscopy (ESEM) to get an overview of the mineralized sample areas. Especially, scanning smallangle X-ray scattering (sSAXS) that provides quantitative information about the crystal size, shape and orientation was used for detecting local structural changes in bone material. For very small samples like embryonal bone the SAXS facility at the µ-Spot beamline at BESSY II (Berlin) with a beam size of 25 µm is very well suited. To detect the very low amounts of mineral it is possible to measure the X-ray fluorescence (XRF) signal of calcium and to create a calcium distribution map of the sample. SAXS and WAXS maps or scans from selected regions give information on orientation and thickness of the mineral particles. Longitudinal sections of embedded bones of embryonal mice at day 16.5 as well as bone of a 6 week old mouse were measured.

Fig. 2: Light microscopy image (right) and Ca X-ray fluorescence map (left) for a typical sample. SAXS/WAXS data were collected at the spots of high Ca content, where bone mineral can be expected (dotted lines).

The embryonal bone at day 16.5 displays the first stage of mineralization revealed by the ESEM- and the X-ray images whereas the bone of the 6 week old mouse displays mature bone. The analysis of the data is in progress.

Fig. 3: Typical SAXS/WAXS patterns for an early mineralization site (left) and mature bone (right). 1. O. Paris, C. Li, S. Siegel, S., G. Weseloh, F. Emmerling, H. Riesemeier, A. Erko, P. Fratzl. A new experimental station for simultaneous X-ray microbeam scanning for small- and wide-angle scattering and fluorescence at BESSY II. J. Appl. Cryst. 40, S466-S470 (2007). 2. W. Tesch, N. Eidelman, P. Roschger, F. Goldenberg, K. Klaushofer, P. Fratzl., Graded microstructure and mechanical properties of human dentin. Calcif. Tissue Int. 69, 147-157 (2001). 3. P. Zaslansky., 'DENTIN' in Collagen: Structure and Mechanics (Ed. P Fratzl), Springer, New York (2008)

218

In – situ studies of deformation in byssus fibers
H. S. Gupta1, M. Harrington2, S. Siegel1, C. Li1 and P. Fratzl1
1Biomaterials

Department, Max Planck Institute of Colloids and Interfaces, MPIKGF Golm, D-14424 Potsdam, Germany
2

Department of Biology, University of California at Santa Barbara, Santa Barbara CA, USA

Biocomposite tissues like bone (mineralized collagen) and tendon (collagen/proteoglycan composites) consist of stiff and ductile components at the nanoscale, whose functioning is essential for understanding the mechanics of the whole system. Such systems exhibit phenomena like molecular scale self – healing due to as yet incompletely understood processes. We investigated the molecular origins of the stress healing phenomena in the byssus fiber, which attached mussel shells to the rocks on wave swept shores. They exhibit elastic behaviour followed by a long inelastic plateau region, and more remarkably have the ability to self – heal mechanically after being stretched into the inelastic zone. We applied in – situ tensile testing combined with time – resolved synchrotron small angle X – ray scattering (SAXS) to byssus fibers kept in a wet state and under applied tensile strain. Molecular strain was measured by shifts of the collagen diffraction peaks.

Figure: (A) 2D wide angle X-ray diffraction pattern of the byssus fiber: white arrow indicates collagen molecular peak (B) small angle X-ray scattering diffraction pattern (arrows) (C) integrated profile of the equatorial pattern (dashed sector in B). Plotting molecular strain against stress shows that the collagen domain is elastic with a Young’s modulus of 2.97±0.19 GPa (mean ± s.e.m.), consistent with previously measured moduli of collagen from tendon. Cyclic loading shows an elastic response at the molecular level indicating that no damage is occurring in the collagen helical domains. It is suggested that in the yield region unfolding and breakage of crosslinks in domains flanking the collagen domain in the byssal fiber occurs. A deeper understanding of selfhealing in such systems may make it possible to modify inelastic proceeses in engineered fibers.

219

Microbeam SAXS/WAXS studies on biogenic and biomimetic calcite single crystals
C. Gilow, A. Schenk, B. Aichmayer, C. Li, S. Siegel, O. Paris and P. Fratzl Department of Biomaterials, Max Planck Institute of Colloids and Interfaces, Research Campus Golm, D-14424 Potsdam

Due to their complex structure, biogenic minerals have fascinating properties. Molluscs, as an example, produce seemingly effortless high-quality minerals with an astounding degree of control. This is achieved via an organic matrix which directs and regulates the mineral deposition. We investigated calcitic prisms from the shell of Pinna nobilis in co-operation with Professor Emil Zolotoyabko (Technion, Haifa, Israel). These prisms diffract like single crystals. Additionally, a strong small angle X-ray scattering (SAXS) signal reveals a pronounced nanostructure, presumably due to intracrystalline proteins, that were previously isolated and characterized by (Marin et al. 2005). These interesting properties act as a source of inspiration for materials science and nanotechnology. In co-operation with Dr. Helmut Cölfen (MPI of Colloids and Interfaces, Potsdam, Germany) we additionally produced bioinspired calcite crystals grown in the presence of a negatively charged polyelectrolyte (PSS, Aldrich) via a gas diffusion technique (Wang et al. 2006). In this process the macromolecules are incorporated into the crystals and create inner surfaces between the mineral and the polymer phase. SAXS as a non-destructive technique probing a comparatively large sample volume is an ideal tool to explore the nano-structure of biogenic and biomimetic minerals. By recording the wide-angle X-ray scattering simultaneously with the SAXS-signal, it is possible to determine the orientation relation of the nanostructures of the crystals with respect to the crystalline lattice. The µ-Spot beamline is equipped to perform simultaneous SAXS and WAXD measurements (Paris et al. 2007). We have extended this setup, which now can be used for single crystal diffractometry. To accomplish this, the sample is placed in the axis of rotation of a goniometer with the help of a long-distance microscope. A sequence of diffraction patterns are then recorded for different rotation angles. The wavelength of the beam is chosen by a 111-Si monochromator. To calibrate the distance between the sample and the detector, a silicon-oxide powder standard (NIST-SRM 1878a) is used.
Pinholes Beamstop

Sample

Ionichamber Marmosaic, 2D-Detector

(x,y,z) ω

Fig. 1: Schematic experimental setup at the µ-Spot beamline for SAXS/WAXS on single crystals

220

Figures 2a and 2b each show typical frames from a sequence for biogenic and biomimetic calcite, respectively (100 s acquisition time). The diffraction peaks correspond to a singlecrystalline pattern of calcite. The inset shows a higher magnification of the anisotropic SAXS signal which is oriented in the direction of the indicated (104) peak. The corresponding WAXD peak is excited at a different angle of sample rotation. Taking into account geometrical corrections necessary to compare the SAXS and WAXD orientation (Lichtenegger et al. 1998), we could show that the nanostructure, which presumably arises from the organic inclusions, is preferentially oriented along both the (104) and the (006) plane. Furthermore the spherically averaged SAXS signals of the biogenic and the biomimetic calcite particles do not follow Porod’s law (decay with Q-4), but instead decay slower, which is characteristic for rough inner surfaces. In summary this setup is a versatile tool to probe the nanostructure of crystals. Especially the orientations of inner structures with respect to the crystal lattice can be probed in an effective manner. Our study of biogenic and biomimetic single-crystal calcite particles with organic inclusions serves as a proof of principle but also revealed interesting similarities between both systems, which will help to deepen the understanding of the processes involved in biomineralization. Future work will also include a three-dimensional reconstruction of the reciprocal space of both the SAXS and the WAXD signal.

(a)

(b)

(104) (104)
Fig. 2: Typical diffraction images from a calcitic prism of the pinna nobilis (a) and the biomimetic calcite particles (b). The insets show the anisotropic SAXS signals which point towards the indicated peaks.

Lichtenegger, H., Reiterer, A., Paris, O., et al. (1998). Determination of spiral angles of elementary fibrils in wood cell walls: Comparison of small-angel X-ray scattering and wide angle X-ray diffraction. Proceedings of the International Workshop on the Influence of Microfibril Angle to Wood Quality, West Port, New Zealand. Marin, F., Amons, R., Guichard, N., et al. (2005). Caspartin and calprismin, two proteins of the shell calcitic prisms of the Mediterranean fan mussel Pinna nobilis. Journal of Biological Chemistry 280 (40): 33895-33908. Paris, O., Li, C., Siegel, S., et al. (2007). A new experimental station for simultaneous X-ray microbeam scanning for small- and wide-angle scattering and fluorescence at BESSY II. Journal of Applied Crystallography 40: S466-S470. Wang, T. P., Antonietti, M., Cölfen, H. (2006). Calcite mesocrystals: "Morphing" crystals by a polyelectrolyte. Chemistry-a European Journal 12 (22): 5722-5730.

221

Microstructure and Chitin/Calcite Orientation Relationship in the Lobster Cuticle Using Microbeam Synchrotron X-ray Diffraction
Ali Al-Sawalmih, Chenghao Li, Stefan Siegel, Helge Fabritius1, Dierk Raabe1, Peter Fratzl, Oskar Paris Department of Biomaterials, Max Planck Institute of Colloids and Interfaces 14424 Potsdam, Germany 1  Max-Planck-Institut für Eisenforschung, Max-Planck-Str. 1, 40237 Düsseldorf The exoskeleton of the lobster Homarus americanus is a chitin–protein-based nano-composite, reinforced with calcite crystallites[1] and amorphous CaCO3 (ACC)[2, 3]. Earlier studies identified that two components of chitin/protein fibers networks exist in the cuticle: in-plane fibers (i) which exhibit the well known twisted plywood structure[4], and out-of-plane fibers (o) which run vertically toward the surface through the pores of the former.[5] We employed microbeam X-ray diffraction to investigate the local microstructure, the orientation relationship between the organic (α-chitin) and the inorganic (calcite) components, and their corresponding spatial distribution.[6] Our synchrotron measurements were performed at the μ-Spot beamline[7] of the BESSY synchrotron facility in Berlin. We found that calcite exhibits a simple fiber texture (preferred orientation) with the c-axis perpendicular to the cuticle surface, as well as parallel to the chitin c-axis. Fig. 1a shows a microbeam diffraction pattern from the lobster exocuticle recorded with the X-ray beam parallel to the cuticle surface. In Fig. 1b, the azimuthal coincidence of the calcite 00.6 reflection and the chitin 013 pair indicates that the calcite c-axis and the chitin c-axis of the vertical fiber component are collinear. Detailed analysis of the microdiffraction patterns reveals that calcite is associated with the out-of-plane chitin/protein fibers, rather than the in-plane fibers.[6] The spatial distributions of the two major mineral phases, calcite and amorphous calcium carbonate (ACC), as well as chitin were determined by performing line scans across the cuticle cross-section covering exocuticle and endocuticle. Calcite was only found in the distal layers of the exocuticle in a region extending about 20–50 μm from  the cuticle surface. However, the whole cuticle is fully mineralized with ACC in the non calcified regions, which can be uniquely concluded from the occurrence of an amorphous halo in the recorded X-ray profiles of the area of interest, Fig. 3.[6] For a better understanding of the functional role of amorphous calcium carbonate and its stability in the cuticle, the temperature induced transformation of ACC into calcite was studied. We employed in-situ heat treatment of the cuticle cross section using a custom made lamp-furnace [8] . Figure 3 shows diffraction profiles for three different temperatures from the same representative region within the endocuticle. The crystalline chitin decomposed completely at temperatures between 250–300 oC, and only a broad hump from ACC remained in the diffraction profile at 325 oC. At about 425 oC, the amorphous halo from ACC started to vanish, and calcite reflections appeared. The whole cuticle was fully calcified at 450 oC indicating that ACC has fully transformed into calcite (see Fig. 3). The details of the transformation behavior, in particular the changes of the cuticle nanostructure and the formed crystalline phases at even higher temperatures allow drawing conclusions about the mechanisms of ACC stabilization in the lobster cuticle[9].

222

Fig. 2: Amount of chitin, calcite and ACC, based on the integrated intensity from the corresponding diffraction peaks.

Fig. 1: a) Diffraction pattern from the mineralized lobster exocuticle recorded with the X-ray microbeam, b) Azimuthal integration of selected X-ray diffraction rings of chitin and calcite.

Fig. 3:  Microbeam X-ray diffraction profiles from the endocuticle as a function of temperature
[1]  [2]  [3]  [4]  [5]  [6]  [7]  [8]  [9]  A. Neville, Biology of fibrous composites, Cambridge University Press, 1993.  F. Boßelmann, P. Romano, H. Fabritius, D. Raabe, M. Epple, Thermochim Acta 2007, 463, 65.  A. Becker, U. Bismayer, M. Epple, H. Fabritius, B. Hasse, J. M. Shi, A. Ziegler, Dalton Transactions  2003, 551.  Bouligan.Y, Tissue & Cell 1972, 4, 189.  D. Raabe, A. Al‐Sawalmih, S. B. Yi, H. Fabritius, Acta Biomaterialia 2007, 3, 882.  A.  Al‐Sawalmih,  C.  H.  Li,  S.  Siegel,  H.  Fabritius,  S.  Yi,  D.  Raabe,  P.  Fratzl,  O.  Paris,  Advanced  Functional Materials 2008, 18.  O. Paris, C. H. Li, S. Siegel, G. Weseloh, F. Emmerling, H. Riesemeier, A. Erko, P. Fratzl, Journal of  Applied Crystallography 2007, 40, S466.  G. A. Zickler, W. Wagermaier, S. S. Funari, M. Burghammer,  O.  Paris, Journal of Analytical and  Applied Pyrolysis 2007, 80, 134.  A. Al‐Sawalmih, C. Li, S. Siegel, P. Fratzl, O. Paris, submitted.

223

Widening the fullerene band gap in fluorination
V.M. Mikoushkin1, V.V.Bryzgalov 1, S.Yu. Nikonov1, V.V. Shnitov1, Yu.S. Gordeev1, Boltalina2,3), I.V. Gol’dt2, M.M. Brzhezinskaya4
) Ioffe Physical-Technical Institute RAS, St.-Petersburg, 194021, Russia ) Chemistry Department, Moscow State University, Moscow, 119899, Russia 3 ) Chemistry Department, Colorado State University, Fort Collins, 80523, USA 4 BESSY, German-Russian laboratory
2 1

Abstract. Study of valence band structure of highly fluorinated fullerenes C60F48 in comparison with that of ordinary fullerenes C60 was performed at the conditions of negligible static charging of samples. It was shown that the band gap of carbon clusters becomes wider in fluorination due to disappearance of molecular levels starting from the highest ones which takes place due to participation of π-electrons in creation of C-F bonds. The project was supported by the bilateral program “Russian-German Laboratory at BESSY” and by the by the Russian Academy of Sciences (Program “Quantum physics of condensed matter”).

Introduction Fluorinated fullerenes C60Fx have been extensively studied last fifteen years in many laboratories by different techniques, as these objects make possible investigating the influence of fluorination on the electronic structure and basic properties of well defined carbon clusters [1-9]. In the comprehensive research of several groups [5] the following important conclusion was made that the highest occupied electron levels (HOMO) on highly fluorinated fullerenes, such as C60F46, are shifted compared to the levels of ordinary fullerene C60 to higher binding energies due to inductive effect of electronegative fluorine atoms, which makes the band gap wider. Shifts of HOMO levels of less fluorinated fullerenes were observed also in our former research [8-10]. But these shifts were of different origins. One of them was directed to lower binding energies and was shown to be connected with fragmentation of fluorofullerenes induced by radiation. Another one was directed to higher binding energies analogous to the result of Ref. [5], but contrary to it, the shift was explained by systematic fault, namely, by static charging of films under irradiation due to photoemission. The main objective of this research was to study the HOMO and valence band electronic structure of highly fluorinated fullerenes C60F48 in conditions of negligible radiation degradation and charging, to compare it with the electronic structure of ordinary fullerenes C60 and to attempt answering the question whether HOMO levels shift or not. Conductive organic thing films were used as a solid solvent of fluorinated fullerenes C60F48 to prevent charging of these wide band gap dielectric molecules. We have found signs that electron levels do not shift, and that the band gap becomes broader in fluorination due to depletion of π-electron system and disappearance of the upper π-electron derived HOMO levels. Experimental details Films of C60F48 were grown on the surfaces of conductive n-Si(100) wafer and Au. Silicon substrate was annealed in high vacuum before the deposition of the film to remove thick layer of native oxide and to create ultra-thin SiC layer which was expected to prevent chemical interaction of C60F48 molecules with Si substrate. Powder consisting of molecules with certain and known number of fluorine atoms was used in the experiment. This material was synthesized at Chemistry department of Moscow State University by the technology described in [1]. Thin composite films were grown by co-evaporation of C60F48 molecules together with molecules of phthalocyanine PcCu in ratio 1:2 which was controlled by C1s photoemission lines. PcCu semiconductor matrix was expected to provide charge outflow and to prevent static charging of the sample surface under irradiation. Thickness of the films was estimated by the weakening of SiLVV Auger- and Au4f photoemission lines and did not

224

2200

F1s
2000

3 min

15 min
1800

60 min
1600

exceed 2-3 monolayers. Photoemission spectra of the films were measured using monochromatic synchrotron radiation of the German-Russian beamline equipped with the plane-grating monochromator (PGM) and the photoelectron spectrometer with hemispherical analyzer VG CLAM-4 whose total energy resolution was better than 300 meV [11].
Fig.1. F1s photoelectron spectra of C60F48 measured after 3, 15 and 60 min of irradiation by diagnostic x-ray beam (hν ~ 850 eV).

I , a.u.
1400

160

162

164

EK , eV

166

Experiment duration was short enough to prevent essential degradation of studied films resulting in shift of photoemission lines towards to the Fermi-level. The photon energy scale of the monochromator was calibrated using the photoemission Au 4f7/2 line of gold (Eb = 84.0 eV). Results and discussion X-ray photoemission spectra of core-electrons C1s and F1s were measured and the corresponding binding energies of C60F48 molecules in the organic matrix were obtained relative to the Fermi level: Eb (C1s, C-F) = 288.5 eV and Eb (F1s) = 687.1 eV. These values proved to be very close to the values obtained by us earlier for pure thing film of C60F36 [10]. The value of the binding energy Eb (C1s, C-F) = 288.5 eV corresponding to carbon atoms bound with fluorine practically coincides with that (288.4 eV) obtained in Ref. [5]. This fact evidences absence of static charging in experiments with thing (~10 nm) pure films of highly fluorinated fullerenes in photoemission experiments [5,10]. Though, the compensation of static charging effect by the effect of degradation of fluorofullerenes under irradiation also can result to the agreement of the experiments. The scale of the degradation effect is illustrated by Fig.1 presenting the radiation induced shift of F1s line of C60F48 observed in this work. Fig.1 demonstrates low radiation stability of highly fluorinated fullerenes C60F48 as compared to C60F18 [8]. Visible shift of about 0.1-0.2 eV is observed after irradiation of the sample by diagnostic x-ray beam during several minutes. After one hour the shift exceeds 0.4 eV. The shift must be large by the value of static charging the sample surface, if there is a case, because of increasing conductivity of the material in the course of its degradation [10]. Therefore measurements of photoemission spectra of highly fluorinated fullerenes need careful analysis of static charging and degradation effects.
0.3

I, a.u.

C60F48 + Substr.

0.2

Au

Fig.2. Valence band photoelectron spectra of gold, of Pc on Si substrate and composite PcCu+C60F48 thing film grown on Si substrate. Photon energy is hν = 120.6 eV.

EF
0.1

Substrate 0.0

100

105

110

115

E, eV

Fig.2 shows the photoemission spectra of valence band of the studied composite film and the film of organic matrix of smaller but comparable thickness. Contribution of the spectrum of the Si substrate is seen because of low thickness of the studied films. The valence band photoemission spectrum of

225

I, a.u.

0.2

"F2p+C2p"

C60F48 was obtained as a difference between these two spectra. Fig.3 demonstrates the obtained spectrum of C60F48 .
Fig.3. Valence band photoelectron spectrum of C60F48 without contribution of substrate. Photon energy is hν = 120.6 eV.
EF

HOMO-4 0.1 HOMO-3

C60F48

0.0

100

105

110

115

E, eV
32000 30000

C60

HOMO-4

HOMO-3

Position of the Fermi-level corresponds to the valence band edge of Au. Unfortunately the fine structure of the spectrum is not seen because of the residual contribution of the matrix. Comparison of this spectrum with the corresponding spectrum of ordinary fullerite C60 (Fig.4) shows the grate difference of the electronic structure of these molecules.
Fig.4. Valence band photoelectron spectra of C60 film. Photon energy is hν = 120.6 eV.

I, a.u.

28000 26000 24000

HOMO-2

One can see that lines of the upper molecular states of fullerene HOMO-1, 20000 HOMO-2 and HOMO-3 completely disappear and the line HOMO-4 radically de18000 creases when 48 fluorine atoms attach to -16 -14 -12 -10 -8 -6 -4 -2 0 2 fullerene C60 and form C60F48 . As a result, EB, eV the band gap of C60F48 becomes wider than that of ordinary fullerite C60 by ~5 eV. Fig. 3 does not show any shifts of molecular lines: the line HOMO-4 saves its position. Therefore we came to the conclusion that the reason of widening of the band gap of carbon clusters in fluorination is disappearance of the upper π-electron states due to participation of π-electrons in creation of C-F bonds. These former π-electrons contribute to the “F2p+C2p” peak in the photoemission spectrum of C60F48 .
22000

HOMO-1

References:

[1] O.V.Boltalina, S.H.Strauss, “Fluorinated Fullerenes”, in Dekker Encyclopedia of Nanoscience and Nanotechnology: New York, Marcel Dekker, 2004, eds. J. A. Schwarz, C. Contescu, K. Putyera, p. 1175-1190. [2] P.J.Benning, T.R.Ohno, J.H.Weaver, et al., Phys. Rev. B 47, 1589 (1993). [3] D.M.Cox, S.D.Cameron, A.Tuinman, et al., J.Am. Chem. Soc. 116, 1115 (1994). [4] Y. Matsuo, T.Nakajima, S.Kasamatsu, et al., J. Fluor. Chem. 78, 7 (1996). [5] R. Mitsumoto, T. Araki, E. Ito, et al., J. Phys. Chem. A 102, 552 (1998).
[6] A.P.Dement'ev, O.V.Boltalina et al., Proceeding Electrochemical Society 2001-11, 559 (2001).

[7] S.V.Amarantova, V.N.Bezmelnitsyn et al., Nucl.Instrum Meth.A 470, 318 (2001). [8] V.M. Mikoushkin, et al., Bulletin of the BESSY Annual Report (2006), Berlin, 202 (2007). [9] V.V. Shnitov, et al., Fullerenes, Nanotubes and Carbon nanostrucrutes 14, 297 (2006). [10] V.M.Mikoushkin, V.V.Shnitov, V.V.Bryzgalov, Yu.S. Gordeev, O.V. Boltalina et al., “Core-level structure of С60F18 and С60F36 studied by photoelectron spectroscopy”, this book. [11] S.I.Fedoseenko, D.V.Vyalikh, et al., Nucl. Instr. and Meth. A 505, 718 (2003).

226

The oscillator strength distributions in the NEXAFS spectra of bacterial surface protein layers: X-ray damage. V.N. Sivkova, S.V. Nekipelova , D.V. Vyalikhb, K. Kummerb, V.V. Maslyukc, A. Bluherd, I. Mertigc, M. Mertigd, S.L. Molodtsovb
b d

Komi Science Center, Russian Academy of Science, Ural Division, Syktyvkar 167982, Russia Institute of Solid State Physics, Dresden University of Technology, D-01062 Dresden, Germany c Martin-Luther-Universit¨at Halle-Wittenberg, Fachbereich Physik, D-06099 Halle, Germany

a

BioNanotechnology and Structure Formation Group, Max Bergmann Center of Biomaterials and Institute of MaterialsScience, Dresden University of Technology, D-01062 Dresden, Germany
backbone and different functional groups of the amino acids. A series of characteristic NEXAFS peaks was assigned to particular molecular orbitals of the amino acids by applying a phenomenological building-block model.

The possible applications of the regular bacterial surface layers (S layers), which are the self-assembly biomolecular structures with welldefined structural and physico-chemical surface properties [1], for the organization of inorganic material at the nanometer scale have stimulated the spectroscopic researches of the surface protein layer of the bacteria B. sphaericus NCTC 9602 in previous years. The S layer protein is known to be absorbed in a regular, twodimensional structure onto solid surfaces which is formed via self-assembly [2, 3]. At that, surface coverage of up to 90% is routinely achieved as had been demonstrated by scanning electron microscopy. The S layer shows a p4 symmetry and reveals a lattice constant of 12.5 nm, and a complex pattern of pores and gaps that are 2–3 nm wide. The thickness of such protein layer amounts to approximately 5 nm. According to its presently known primary structure a single S layer protein molecule is composed of 1228 amino acids, including 1517 nitrogen atoms and has a molecular weight of ~130 kDa [4]. With the exception of Cysteine all 20 standard amino acids are present. The protein contains 3.91 carbon, 1.22 oxygen, 7.7 hydrogen and the neglectable amount of 0.0027 sulfur atoms per one atom of nitrogen. The 83.1% of the nitrogen, 21.4% of the carbon and 64.8% of the oxygen atoms are bounded in the peptide backbone of the protein. Details of the sample preparation are described elsewhere [5]. Recently, we have performed photoemission (PE) and near edge x-ray absorption fine structure (NEXAFS) experiments in order to characterize electronic properties of the regular two dimensional S layer of Bacillus sphaericus NCTC 9602 [5-7]. Both PE and NEXAFS at the C 1s, N 1s, and O 1s core levels showed similar chemical states for each oxygen atom and also for each nitrogen atom, while carbon atoms exhibited a range of chemical environments in

14 C1s

Absorption cross section, Mb

12 10 8 6 4 2 0 300 350 400 450 500 550 600 NC0.1O0.33Si1.24

NC4.0O1.53Si1.24 N1s O1s

E, eV

Fig. 1. Spectral dependence of the TEY signal of S layer after normalization to the incident photon flux (black). Comparing with tabulated atomic cross sections [12] we found the best agreement for a chemical composition NC4.0O1.53Si1.24 (dotted line). The additional signal C0.1 O0.33 Si1.24 is easily assigned to contributions by the SiOx/Si substrate (blue). In the issue of the researches had shown that the radiation induced damage of the protein samples can substantially change the line shape of the data also resulting in appearance of new intense spectral feature [8]. Therefore the valid irradiation doses data are of severe importance for researches allowing feasible adjustment of experimental conditions and provide perspectives of measurements of biological samples with spectroscopic methods applying soft x-ray radiation. The goal of the project is to carry out the study of the radiation-induced decomposition of the S-layer by the analysis of the oscillator

227

12 C1s Absorption cross section, Mb N, fotons / s / m / mA*10
15

4,0
N1s

Ti2p

10 8 6 N1s 4 2 0 300 350 E, eV 400 450

3,5 3,0 2,5 2,0 1,5 1,0 0,5 0,0 200 300 400 E, eV 500 600
O1s
Ti2p two order C1s

Fig.2. The absorption cross section spectral dependences of S layer. The dash lines are extrapolated curves. strength distributions in the C1s, N1s and O1s NEXAFS-spectra as a function of the gained skin dose. An irradiation doses gained by the surface area of the samples that is much thinner than the light penetration depth (skin doses, D) were calculated:
D=

Fig.3. The photon-flux density of monochromatic synchrotron radiation normalized at the beam current.

2

the had

s
M

∫ σ ⋅ N ⋅ Edt =

s
M ⋅ VE

∫ σ ⋅ N ⋅ EdE

(1)

where σ is absorption cross section of the single S layer protein molecule, N denotes the photonflux density of the monochromatic synchrotron radiation, S is the light spot size at the sample position, E is the photon energy, t is the exposure time, VE is the energy scanning velocity and M is weight of a single S layer protein molecule (~130 kDa). The absolute absorption cross sections were measured by Total electron Yield (TEY) method with using Ti-filter for suppressed nonmonochromatic background [9]. All presented data were obtained at the BESSYII using radiation from the Russian–German beamline (RGBL) [5, 6]. The light spot size at the sample position was 200×87 μm2 [10] which allowed to study homogeneous parts of the S layer. All spectra were acquired in TEY mode by recording the sample drain current. The spectral dependence of the photon flux was determined using a clean Au photocathode. The flux curve exhibits huge dips in the region of the O 1s absorption threshold (530-570 eV) due to extensive oxygen contamination of the optical elements. Hence, we decided to repeat measurements in the O 1s range at the beamline D1011 of the MAX II storage ring at MAXlab (Lund University, Sweden). Fig. 1 shows the TEY signal obtained from the surface-adsorbed protein layer after normalization to the incident photon flux. The

curve features distinct jumps when passing the C, N and O 1s edge. More precisely, at each absorption edge it exhibits a near-edge fine structure followed by a monotonically descending tail. While the NEXAFS reflects transitions into free molecular states, the structureless tail is due to transitions into continuum states far beyond the vacuum level [11]. However, beside the strong evidence in the already mentioned previous studies, our data itself indicates direct proportionality between the curve in Fig. 1 and the total X-ray absorption cross section of the sample. In fact, using atomic calculations of absorption cross-sections, we were able to reproduce the full measured spectrum. The best agreement was found assuming an average atomic composition of NC4.0O1.53Si1.24. The corresponding curve is included in Fig. 1 as a blue line. It should be noted that the close similarity in the regions of continuum transition where atomic calculations hold validity from the energies of the synchrotron radiation. The values of the atomic absorption cross section were selected from tabulated data [12]. Fig. 2 shows the absorption cross section of S layer which had measured as difference between the curves (1) and (2) in fig.1. The photon-flux density of the monochromatic synchrotron radiation was measured by TEY clean Au-photocathode with using the front-surface total quantum yield data of Au [13] which show in fig.3. The exposure doses gained by the S-layer samples were calculated by formula (1) for the any certain time and energy position. The dose was obtained for monochromatic radiation only, because Tifilter full suppressed higher order radiation. The researches had shown that higher order radiation induced strong damage of the protein samples. X-ray damage effects were studied at C 1s and N

228

1s thresholds of excitation upon scan-by-scan irradiation of the S-layer proteins and in the extreme case of irradiation by outmost high fluxdensity nonmonochromatized (“zero-order”) synchrotron radiation. Only the S-layer signal was used to describe the radiation damage. In the fingers 4 and 5 showed the irradiation-induced changes of NEXAFS C1s and N1s spectra of S layer. Fig.4-5 are shown that the oscillator strength of the C1s - πCONH and N1s - πCONH transitions decrease with increased the exposure doses gained by the S-layer sample. At that sum oscillator strength for C1s spectrum is unvaried (unshow), but for N1s spectrum is obvious decrease. The two structures (a, b) developed in the N1s edge with increasing irradiation doses. The analysis of the NEXAFS 1s spectra and the chemical structure of S layer have shown that 90% from area A resonance in N1s spectra fit from the nitrogen atoms of peptic group and 82% in from area A resonance C1s spectra fit from the carbon atoms of peptic group. This observation indicates that, while the peptic bonds get damaged by the applied radiation, N atoms partially leave from the sample surface,
0.199 2.74 9.68 16.0 21.53 26.60 31.25 35.53 39.61 43.18 46.58 50.04 81.33 zero-order

C atoms - in contrast to nitrogen – remain at the sample surface, and interact presumably with N atoms likewise released from the peptide units or resided by cracking at the surface of the Si wafer. We suspect that also in the case of radiation-damaged proteins some of the dispensed N atoms may interact with C atoms forming carbon-nitride compounds. Acknowledgements: This work was supported by grant RFBR 09-0200049a and by the Bilateral Program of the Russian--German Laboratory at BESSY II.V.N. Sivkov gratefully acknowledge the financial support by BESSY-II and the Technische Universitat Dresden. References
1. 2. Mann, S. Nature 1993, 365, 499. U.B. Sleytr, P. Messner, D. Pum, M. Sara, Crystalline Bacterial Cell Surface Proteins (Academic Press, San Diego, 1996). [3] Sleytr, U.B.; Sara, M. Trends Biotechnol., 15, 20 (1997). UniProt database, www.uniprot.org Vyalikh D.V.; Kirchner, A.; Kade, A.; Danzenbacher, S.; Dedkov, Yu.S.; Mertig, M.; Molodtsov, S.L. J. Phys. Condens. Matter, 18,131 (2006). Vyalikh D.V.; Danzenbächer S.; Mertig, M.; Kirchner, A.; Pompe, W.; Dedkov, Yu.S.; Molodtsov, S.L. Phys. Rev. Lett., 93, 238103 (2004). Vyalikh D.V.; Kirchner, A.; Danzenbächer, S.; Dedkov, Yu.S.; Kade, A.; Mertig, M.; Molodtsov, S.L. J. Phys. Chem. B, 109, 18620 (2005). Kade, A.; Vyalikh D.V.; Danzenbächer S.; Kummer K.; Bluher A., Mertig, M., Scholl A,, Doran A., Molodtsov, S.L., J. Phys. Chem., B, 111, 13491 (2007). Sivkov V. N., Vinogradov A. S., Nekipelov S. V., Sivkov D. V., Vyalykh D. V., Molodtsov S. L., Optika i Spektroskopiya 101, N5, 782 (2006) Fedoseenko S.I., Vyalikh D.V., Iossifov I.F., Follath R., Gorovikov.S.A., Püttner R., Schmidt, J.-S., Molodtsov S.L., Adamchuk V.K.; Gudat W., Kaindl G., Nucl. Instr. and Meth. in Phys. Res. A, 505,718 (2003). Stohr, J. NEXAFS Spectroscopy (Springer, Berlin, 1992). J.J. Yeh, Atomic Calculation of Photoionization Cross-Sections and Asymmetry Parameters (Gordon and Breach Science Publishers, Langhorne, PE (USA), 1993). J.J. Yeh, I.Lindau, Atomic Data and Nuclear Data Tables, 32, 1-155 (1985) Henke B.L., Knauer J.P., Premaratne K.,J. Appl. Phys., 52,1509 (1981)

3. 4. 5.

S layer, C1s spectrum A(πCONH)

3,0

Prtial C1s cross section, Mb

2,5 2,0 1,5 1,0 0,5 0,0 283

6.

A(πAr.ring)

7.

284

285

286

287

288

289

290

E, eV

8.

Fig.4. Irradiation-induced changes of the Partial C1s absorption cross section for S layer.

9.

1,2 Partial N1s cross section, Mb 1,0 0,8 0,6 0,4 0,2 0,0 390
a

A (πCONH)

S layer, N1s spectrum
4.85 39.17 62.57 104.5 zero-order

10.

11.
b

12.
410 E, eV 420 430

400

Fig.5. Irradiation-induced changes of the Partial N1s absorption cross section for S layer.

13.

229

Analysis of 3D Microstructure and Deformation of Multiphased Materials by Micro Computertomography
H.-A. Crostack, J. Nellesen1, G. Fischer1, K. Ehrig2, H. Riesemeier2 and J. Goebbels2 Lehrstuhl für Qualitätswesen, TU Dortmund, D-44221 Dortmund, Germany 1 RIF e.V., Joseph-von-Fraunhofer-Str. 20, D-44227 Dortmund, Germany
2

BAM Bundesanstalt für Materialforschung und -prüfung, Unter den Eichen 87, D-12205 Berlin, Germany

1. Introduction
Comprehension of micro deformation and micro damaging processes running before the macroscopic failure of metal materials is essential in order to predict the material behavior and to tailor materials with respect to desired mechanical properties. To gain this understanding one has to analyze these processes in the course of loading and to study the impact of microstructure. In case of multiphased materials for example the effect of volume fraction, shape, orientation and arrangement of the phase dispersed in the matrix on these processes has to be considered. Micro deformation which precedes the micro damaging is often accompanied by strain localisation. This phenomenon affects tremendously the locus and the type of the successive failure, the behavior during machining the material, the roughness of the surface and the fracture strain.

2. Materials and Methods
For this reason, in this work strain localisation is analysed non-destructively by mapping 3D microstructure of multiphased materials in different stages of deformation with synchrotron radiation based micro-tomography. The 3D tomograms are evaluated by digital image correlation to get strain tensor and displacement vector fields: Exploiting gray value gradients (i.e. contrast between the phases), an affine and radiometric transform between a less deformed and a more deformed state are iteratively optimized for a set of corresponding tomogram subvolumes. From the final affine transform of each subvolume pair the Lagrangian strain tensor γ is calculated. After a principal axis transformation of this tensor the scalar equivalent strain εequ can be gained. In order to guarantee a sufficient number of sites with a high gray value gradient for the digital image correlation algorithm the size of the subvolumes has to be adapted to the characteristic microstructural length which is given by the mean distance between vicinal microstructural objects. A detailed description of this approach can be found in [1].
´ µ ´ µ

¼¼

µÑ

Figure 1: (a) 2D-image of microstructure from the bulk of a plain tensile specimen made of the multiphased material Al98Sn2, tin segregations (white) at the grain boundaries; (b) 3Drepresentation of the segregated tin.

230

In the scope of the experiments tiny plain dog-bone shaped tensile specimens with a crosssectional area A = 2 x 1 mm2 made of the alloy Al98Sn2 were investigated which were manufactured by spark cutting. The experiments were performed at BAMline at BESSY II [2, 3]. Monochromatic photons of E = 35 keV were selected by a double multilayer monochromator. After each deformation step the gauge length of the tensile specimen was imaged by microtomography in the unloaded state. During image acquisition the specimen was turned on its longitudinal axis 2000 times through an angle increment of 0.09 ° yielding the semi circle. Due to parallel beam geometry the reconstruction was done slice-by-slice. The 3D tomogram was formed by the stack of 2D slices.

3. Results
From the 2D-image (xy-slice) in fig. 1(a) extracted from the 3D-tomogram of the undeformed specimen it is obvious that tin segregates at the grain boundaries. For this reason, tin serves as a marker for the 3D image correlation algorithm. Thus, in this case the characteristic microstructural length is given by the mean grain size. The spatial distribution of the segregated tin can be grasped in fig. 1(b). Since the characteristic microstructural length in the investigated material approximately amounts to 100 µm a cubic tomogram subvolume with an edge length of about 250 µm was chosen for the correlation analysis.
´ µ

Ö×Ø

ÓÖÑ Ø ÓÒ ×Ø Ô

´ µ

× 
ÓÒ

ÓÖÑ Ø ÓÒ ×Ø Ô

½º ±
εequ εequ

º½ ±

¼
x y
ÓÖÑ Ø ÓÒ ×Ø Ô

z

¼
Ø Ö

´
µ

Ø Ö

´ µ

ÓÖÑ Ø ÓÒ ×Ø Ô ´εequ

Ò

γij µ

½¼ ±
εequ εequ

½¼ ±

¼

¼

Figure 2: (a)-(c) 3D representation of the color-coded equivalent strain in the course of loading, (d) 3D representation of the field of Lagrangian strain tensors γ with cuboids; cuboids are color-coded according to the scalar equivalent strain εequ ; only the tensor cuboids of the planes visible in subfigure (c) are shown.

In the subfigures 2(a)-(c) the development of equivalent strain εequ in the first, second and third deformation step are visualized. In order to achieve the smooth representation of equivalent strain the distribution of the discretely sampled scalar field was tesselated. From the comparison of these subfigures it can be deduced that bands of elevated equivalent strain which emerge already in the first deformation step retain their position but the strain values increase. In subfigure 2(d) the 3D field of Lagrangian strain tensors γ in the third

231

deformation step is depicted by cuboids. For the sake of clearness only the tensor cuboids of the planes visible in subfigure 2(c) are shown. The bricks which are color-coded according to the scalar equivalent strain are oriented and scaled according to the Lagrangian strain tensor. The comparison of subfigures 2(c) and 2(d) reveals that in the band of elevated strain which can be seen in the facing xz-plane the largest edge of the cuboids (maximum principal strain) is preferentially aligned with the direction of global load indicated by the red line with arrows at both ends. Up to now the effect of microstructure on the initiation and growth of bands has not been studied but will be analyzed in future work.

4. Acknowledgment
For assistance during beamtime at BAMline we would like to thank Ralph Britzke, Thomas Wolk and Jörg Nötel. This work was financially supported by the BMBF (support code: 05ES3XBA/5) which is gratefully acknowledged.

5. References
[1] H.-A. Crostack, J. Nellesen, G. Fischer, S. Schmauder, U. Weber, F. Beckmann, 3D analysis of MMC microstructure and deformation by µCT and FE simulations, Developments in X-ray Tomography VI, edited by Stuart R. Stock, Proceedings of SPIE 7078, 70781I-1-12, 2008 (SPIE, Bellingham, WA). [2] H. Riesemeier, K. Ecker, W. Görner, B.R. Müller, M. Radtke, M. Krumrey, Layout and first XRF Applications of the BAMline at BESSY II, X-ray Spectr. 34 (2005) 160-163 [3] A. Rack, S. Zabler, B. R. Müller, H. Riesemeier, G. Weidemann, A. Lange, J. Goebbels, M. Hentschel, W. Görner, High resolution synchrotron-based radiography and tomography using hard X-rays at the BAMline (BESSY II), Nucl. Instr. Meth. A 586 (2008) 327-344

232

Valence Electronic Structure of Ruthenium Based Complexes Probed by Photoelectron spectroscopy at High Kinetic Energy (HIKE) E. M. J. Johansson1, M. Odelius2, M. Gorgoi3, O. Karis1, R. Ovsyannikov3, F. Schäfers3, S. Svensson1, H. Siegbahn1, H. Rensmo1
1

Dept. of Physics, Uppsala University

2

Fysikum, Stockholm University
3

BESSY GmbH, Berlin

Different ruthenium-polypyridyl complexes are extensively investigated today due to their potential use in photoconversion, e.g. in dye-sensitized solar cells. Some rutheniumpolypyridine complexes such as Ru(dcbpy)2(NCS)2 have resulted in solar cell systems having very high photon to current conversion efficiency in the visible region. In most of the ruthenium based dyes, the dominant absorption process is traditionally considered as a metal to ligand electron transfer, induced by absorption of the incoming photon. In for example the dye-sensitized solar cell, the electron is thereafter transferred from the ligand to the oxide nanoparticle. The efficiency of the photoconversion process is largely dependent on the properties of the interfacial region and specifically the molecular properties of the dye molecule. Insight into the detailed electronic structure of the dye molecule will thus facilitate the understanding of the mechanisms determining observed functions and efficiencies. Photoelectron spectroscopy (PES) is a technique that can be used to obtain element specific information on the electronic and molecular surface structure at the interfaces in the dyesensitized solar cell. In the present report we use photoelectron spectroscopy at high kinetic energy (HIKE) using hard X-ray as an experimental tool to understand the bulk valence electronic structure of such metal-centered complexes. We report a series of ruthenium polypyridine complexes with CN-, or Cl- or NCS- ligands. The different ligands are exchanged stepwise to follow how the metal-ligand interactions develop. Such experimental information is of general interest in the design of such inorganic coordination complexes. The experimental results are modeled well by DFT calculations in a crystalline environment whereas the gas phase models show distinct differences in the mixing of ruthenium and ligand orbitals. The bulk sensitivity using HIKE is particularly useful for measurements of valence structures on molecular films that are difficult to prepare in-situ (in a UHV environment) due to difficulties in evaporation of the material without decomposition. For such ex-situ samples the valence electron spectra may contain contributions from a contamination layer making, for example, the interpretation of traditional ultraviolet photoelectron spectroscopy (UPS) measurements difficult. However, at a kinetic energy of 2800 eV, as used in this study, the mean free path is in the range of 5 nm, that is, about 5 molecular layers. This means that about 40 percent of the signal originates from layers buried deeper than 5 nm and that the contribution from the surface contamination in the interpretations therefore is substantially diminished. Another important feature when using hard X-rays compared to UPS or PES based on soft X-rays, is the general influence of the cross section of the different energy levels. Specifically, in the ruthenium based dyes, the cross section of Ru 4d relative C 2p, N 2p, O 2p, Cl 3p or S 3p will increase considerably when using hard X-rays. Photoelectron spectroscopy with HIKE in the valence region will therefore mainly probe the Ru 4d partial density of states (PDOS) and can be used as an experimental tool to delineate the molecular orbital character for the highest occupied orbitals in the complexes. The figure shows the experimental valence spectra of the different Ru-polypyridyl complexes. The frontier orbital structures of these kinds of ruthenium complexes are often interpreted within an octahedral symmetry in which the highest occupied molecular orbital contains three degenerate orbitals mainly localized at the Ru center (The t2g set).

233

However, experiments on Ru(dcbpy)2(NCS)2 (N719 in the figure), Ru(tcterpy)(NCS)3 (BD in the figure) indicate that the t2g Ru 4d levels split substantially due to the symmetry and the interaction with the different ligands and accordingly that other orbitals than those from the metal center have to be mixed into the density of states (DOS) between 0-5 eV. Comparing the experimental spectra of the different dye molecules, we observe that the highest occupied electronic structure differs substantially. At high photon energies, the ruthenium contribution is high as explained above and the interaction between the Ru 4d and the X-ligands can thus be followed directly for the different dyes. The splitting between the two peaks with lowest binding energy are very similar for dye 1 (BD) and dye 2 (N719), about 1.5 eV, and the intensity ratio in both cases is also similar and slightly below 2:1. For the dye with CN ligands and the dye with Cl ligands, there is no clear large peak close to the peak at lowest binding energy, instead a broad feature at 3-4 eV from the first peak is observed, which suggests that the mixing of the ligand molecular orbitals with the Ru 4d levels is different for these dyes. Calculations of the electronic structure of single molecules showed that the Ru 4d level splits up. However, the intensity relations for the ruthenium contribution in the highest occupied electronic structure do not agree with experiment. Calculations of the dye molecules in crystal structures were also performed and the agreement with experimental data for these calculations is found to be much better. This showed that the intermolecular interactions in the solid state are important and largely influence the orbital composition in the frontier electronic structure. The precise electronic structure is a key parameter for photoconversion involving charge transfer processes and the result emphasises the importance of intermolecular interaction for modeling the initial conversion process in the dye-sensitized solar cell.

Figure 1: Valence spectra of the ruthenium complexes together with the calculated ruthenium contribution in a gas phase calculation CPMD(g) and in the calculation on a crystal of molecules CPMD(s)

234

High energy X-Ray photoelectron spectroscopy of HBC-C14
Vajiheh Alijani Zamani, Shahabeddin Naghavi, Andrei Gloskovskii, Gerhard Fecher and Claudia Felser
Institut f¨r Anorganische Chemie und Analytische Chemie, u Johannes Gutenberg - Universit¨t, Mainz a

A charge transfer (CT) complex is defined as an electron donor – electron acceptor complex, characterized by electronic transitions to an excited state. In this excited state, there is a partial transfer of charge from the donor to the acceptor. Almost all CT complexes have unique and intense absorption bands in the UV and visible range. They are good candidates as photovoltaic cells, transistors and superconductors. Hard X-ray photoelectron spectroscopy (HAXPES) is a powerfull tool to investigate the bulk electronic structure of solids. It allows to study the density of occupied states and to estimate the position of HOMO. We started our theoretical and experimental investigations of CTs from HBC-C14 which is a strong polyaromatic donor. The structure of HBC-C14 is shown in Figure 1. Thin films of HBC-C14 were prepared by spin coating. The concentration of the solution was 3 mg/ml in chloroform as solvent. To reach this concentration, the solvent was heated to 55◦ C. The optimal rotation speed for the films preparation was 1500 rpm. Lower rotation speeds leads to thicker films but we can see precipitates on the surface of the films, whereas higher rotation speeds results in thinner noncontinuous films. The process for cleaning of the Si/SiOx substrate was acetone ultrasonication for 30 minutes at 35◦ C followed by rinsing with isopropanol and finally drying with N2 . The thickness of the film measured by using a tencor profiler was estimated to be 30 nm. AFM images and thickness profile (TP) line of the film chosen for the photoemission studies are shown in Figure 2a, b, respectively. The TP was taken along the line AB drawn through a defect on in the film which most likely is a hole to the substrate. Its depth is 25 nm and can be used as a measure for the film thickness. Generally, the TP line excibits the oscillations with an amplitude of 10 nm. This means that the film consist of two layers: 15 nm thick 2D continious layer and 10 nm thick 3D layer on the top. Thin film was investigated at KMC-1 beamline of BESSY with different excitation energies between 2.2 keV and 8 keV. Prior to synchrotron measurements, the film was investigated using laboratory XPS (Al Kα with an excitation energy of 1.486 keV). Figure 3 shows survey XPS spectrum. No trace of Si from the substrate and negligible oxygen signal was observed in this measurement. The resolution of the non-monochromatised laboratory X-Ray sources is not enough for valence band studies, therefore HAXPES measurements using tunable synchrotron radiation were done.

235

A
C14H29

B

C14H29

C14H29

25 20

Film height (nm)

15 10 5 0
15 nm 10 nm

C14H29

C14H29

0

500

1000

1500

C14H29

Distance (nm)

FIG. 1: Structure of HBC-C14.

FIG. 2: (a) AFM images of the HBC-C14 film; (b) thickness profile line taken along line AB shown in (a) and a schematic drawing of the structure of the film.

C 1s

0.25 0.20

Intensity (arb. units)

(arb. units)

0.15 0.10 0.05 0.00 -5

O 1s

Intensity

-4

-3

-2

-1

0

Si 2s Si 2p

-600

-500

-400

-300

-200

-100

0

Binding energy (eV)

-25

-20

-15

-10

-5

0

Binding energy

(eV)

FIG. 3: XPS spectrum of thin film of HBC-C14 on Si/SiOx substrate. The energy position of Si and O core levels are marked with arrows.

FIG. 4: Valence band spectra of HBC-C14 at Eexc =3 KeV.

Still at excitation energy of 2.2 keV, the survey spectrum shows mainly C 1s core level emission and Si is not visible. With increasing photon energy the inelastic electron mean free path increases from 5 nm at 2.2 keV to 15 nm at 7 keV and photoemission signal from the Si/SiOx substrate becomes visible at 7 keV (not shown here). We can conclude that HAXPES measurements at intermediate energies (about 4 keV) are better suited than at high energies (7 keV), in

236

particular for uncapped thin films with a thickness of several tens of nanometers. Intermediate excitation energies already ensure good bulk sensitivity with a probing depth slightly less than the film thickness. In the case of very high excitation energies the inelastic electron mean free path becomes comparable with the film thickness. This results in the experimental probing depth much larger than the film thickness. The signal from the substrate superimpose with the spectrum from the thin film what makes the data processing much more complicated. Moreover, at very high excitation energies the intensity from p states is moderately low, as their absorption cross section drops down at high kinetic energies much faster than the one of the s states. The best resolution at high signal to noise ratio was achieved at Eexc =3 keV (Figure 4). The valence band onset is at about -2 eV. This result is in a good agreement with the calculations predicting the HOMO-LUMO gap of 2.6 eV. Additional inversed photoemission studies are needed to investigate the density of states with positive energies where LUMO is situated.

E

exc

= 3KeV 4KeV

Intensity (arb. units)

5KeV 6KeV 7KeV

-25

-20

-15

-10

-5

0

Binding energy (eV)

FIG. 5: Valence band spectra of HBC-C14 measured at different excitation energies. Pronounced excitations are visible from -14 eV to -20 eV that are related to aliphatic chains that exist in the HBC center [1]. The intensity of these excitations decreases with increasing the photon energy. This is, indeed, a prove that those excitations belong clearly to the compound and not to the substrate. At -6 eV to -10 eV excitations with lower intensity and an unresolved structure are visible. The intensity of the excitations between -6 eV and the threshold are even lower. The intensity in the latter two ranges does not change with increasing photon energy (Figure 5). The support of KMC-1 crew at BESSY is gratefully acknowledged. This work was funded by SFB (Project B8 in TRR49) and BMBF 05KS7UMI.

[1] J. J. Pireaux and R. Caudano, Phys. Rev. B 15, 2242 (1977).

237

Characterisation of Photonic Band Gap Phosphors Alan Michette and Matthew Shand King’s College London, Department of Physics, Strand, London, WC2R 2LS, UK George R Fern, Jack Silver, Robert Withnall and Terry Ireland Wolfson Centre for Materials Procesisng incorporating the Centre for Phosphors and Display Materials, Brunel Univesity, Kingston Lane, Uxbridge, Middlesex, UB8 3PH, UK. Several phosphors were characterized using the KMC-2 beamline and a user-provided Jeti Spectroradiometer 1201 linked to a standard PC via a USB cable. This instrument records the luminance of the material over the wavelength range 380– 780 nm. The measurements were taken at 1 keV intervals in the range 6–12 keV and in some cases measurements were also taken at smaller energy intervals around the xray absorption edges of the activator ions and the host lattice ions. Many samples were investigated to ascertain which materials show the brightest luminance using these excitation conditions. The samples that were studied were powders of: Y2O3:Eu 1%, Y2O3:Ce 1%, Y2O3:Dy 1%, Y2O3:Er 1%, Y2O3:Gd 1%, Y2O3:Ho 1%, Y2O3:Nd 1%, Y2O3:Pr 1%, Y2O3:Sm 1%, Y2O3:Tb 1%, Y2O3:Tm 1%, Y2O3:Yb 1%, Gd2O2S:Pr 1%, 2%, 3% and 5%, and the 1% dopant concentration with several deposition thicknesses, Y2O3:Eu 2% Tb0.1% Gd0.1%, Y1.2Gd1.8Al5O12:Ce, Y2O3:Eu 2% Tb 0.1%, Gd 0.1%, ZnO:Zn, several types of ZnS:Cu, ZnS:Cu,Pb, ZnO:Ga, Tb3Al5O12:Ce LaPO4:Tb, Y2O2S:Tb and of some of these materials deposited as photonic band gap materials comprising of face centred cubic lattices with a cell parameter in the region of 300 nm. Hence around 55 sets of measurements were made. A significant amount of time was spent optimising the measurements with the Jeti spectroradiometer since this is a non-standard detector and has non-ideal resolution and sensitivity. The spectrometer should have ideally been placed 6–20 cm from the sample but in this range of distances with the first samples measured it was found that no visible light emission could be detected. After some experimentation it was found that the optimised position for the Jeti was as close to the sample as possible without interfering with the x-ray beam, which meant that it had to be around 30o to the incident x-ray beam. A monochromator system would give better resolution, and sensitivity would be greatly enhanced if the detection method was via a photomultiplier tube. It was also found that the mounting of the samples for measurement could be improved. The phosphor powders were mounted onto adhesive tape which resulted in a thin uneven covering of powder and may have been one factor that produced some of the poor emission spectra. By modifying the sample holder it was possible to produce 2 mm thick samples that gave improved emission spectra. It was surprising that most of the phosphors studied gave poor visible emission spectra, < 1 Cd m–2. Most of the materials analysed are well known scintillator materials and have already shown good photo and cathodo-luminescence; the Gd2O2S:Pr phosphor had also previously been tested on the King’s laser plasma x-ray source where it showed excellent emissive properties.(1) This is likely to be due to the inability of the detector to measure small sample areas, since the x-ray beam spot was smaller than the sampling area of the detector. The most intensely emitting samples measured all contained Tb doping; LAP:Tb, (YGd)2O3:Tb, Y2O2S:Tb and Gd2O2S:Tb

238

gave very strong, clean emission spectra with luminance values of 10–60 Cd m–2 (see figure 1).

Figure 1. Typical emission spectrum of 1% terbium doped yttria obtained at BESSY in the energy range 6–12 keV.

The use of Tb3+ is advantageous since it has many sharp emission bands spanning the visible spectrum (380–780 nm) which will enable the analysis of the effect of photonic band gaps in the structured materials. Attenuation of a Tb3+ emission band was observed at 427 nm in the lattice of a small PBG crystal. This coincides with the calculated stop band for this inverse opal phosphor lattice, and requires further investigation since the observed emission was weak. Hence, future investigations on photonic band gap materials will need to consider the use of a better detector, improved sample mounting and the use of Tb doped scintillator materials. Acknowledgements This work was supported, in part, by the UK Smart X-Ray Optics consortium (EPSRC Basic Technology grant D04880X), by the ESF COST Action MP0601 “Short Wavelength Laboratory Sources” and by the European Union through the BESSY-EC-IA-SFS. Reference (1) George Fern, Terry Ireland, Jack Silver, Robert Withnall, Alan Michette, Chris McFaul, Slawka Pfauntsch, Characterisation of Gd2O2S:Pr phosphor screens for water window X-ray detection, Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, In Press, Available online 6 December 2008, ISSN 0168-9002, DOI: 10.1016/j.nima.2008.11.116.

239

Synthesis and material analysis of carbon-nitride nanoparticles
E. Kovacevic1, J. Berndt1, Th. Strunskus2, J. Winter3, L. Boufendi1 and Ch. Wöll4
2

GREMI Université d'Orléans, Polytech’Orleans, Orleans Cedex 2, France Christian-Albrechts-University of Kiel, Institute of Material Science, Kiel, Germany 3 Ruhr University Bochum, Institute for Experimental Physics II, Bochum, Germany 4 Chair of Physical Chemistry I, Ruhr University Bochum, Bochum, Germany

1

The aim of the investigations is the production and analysis of nitrogen containing hydrogenated carbonaceous nanoparticles. The incorporation of nitrogen into carbonaceous materials is interesting in a wide area of applications - from technology to astrochemistry. A large variety of nitrogen containing carbonaceous materials have attracted a great interest due to their mechanical and electrooptical properties. Nowadays many of these materials already found applications in industry, e.g. as coatings for hard disks, microelectromechanical elements (MEMs) or emitters. Furthermore nitrogen containing carbon compounds can also be found in astrophysical environments. The existence of nitrogen both in gas and solid phase is clearly seen in observational IR spectra of the interstellar medium and it is known that the prominent feature of Titan's atmosphere is a thick haze region that acts as the end product of hydrocarbon and nitrile chemistry. Carbon-nitride materials miming the Titans aerosols are named tholins and low temperature plasmas can be considered as one of the best sources for their production. One of the crucial questions for the application and production of plasma produced polymers concerns the stability of these materials. An important point for the production of stable polymers is of course the material temperature - during the production process and afterwards. It was found (see [1] and the references therein) that plasma polymers can be stabilized by thermal annealing, since the thermal treatment enhances the mobility of cross-linked polymer chains and therefore favors the recombination of trapped free radicals. Vice versa, annealing experiments can deliver important information about the stability of the polymer, about the stability of their bonding structure, and thus, indirectly about the conditions during the production process. The nanoparticles we present in this contribution are produced in a low temperature, low pressure and low power radio-frequency discharge, from nitrogen/methane gas mixture (10% methane). The capacitively coupled discharge was driven by 13.56 MHz , input power 40W at 1mbar total gas pressure. The particle formation is monitored by the analysis of electrical characteristics of the discharge (e.g. self-bias ) as well as by emission spectroscopy and laser light scattering. This observation of the plasma allowed the in-situ identification of different growth phases - that are related to specific particle sizes, and thus a controlled production of the particles. Our further analysis was obtained on particles with diameter of about 100nm. The annealing experiments have been obtained in high vacuum chamber at the HEGSM beam-line The effect of the annealing on the polymer material was investigated (in situ i.e. during the annealing) by means of Near Edge X Ray Absorption Fine Structure spectroscopy (NEXAFS). The NEXAFS measurements are obtained in the soft X-ray spectral region where sharp core levels are found for C, N, and O elements contained in our samples. Figure 1 shows the ex-situ IR spectrum of the particles produced in nitrogen/methane gas mixture. The broad strong features with maximum at about 3300 cm-1 can be identified as NH stretching vibrations (NH2 and NH3, overlapping with certain small amount of OH stretching bonds due to impurities present in the chamber). The structured shoulder at 2930 cm-1 is identified as CH stretching feature (CH2 and CH3, symmetric and asymmetric). The dominant feature in the spectrum is a strong, broad absorption peak centered at about 1625 cm1 (6.15 µm) which can be interpreted mostly by the 0.010 formation of CN double bonds, with possible C=C 0.008 presence as well as a superposition of vibrations with different origins (e.g. C=O, CH deformation bands, see 0.006 [2]). The presence of triple bonds is observable from the feature with maximum at about 2200cm-1. The origin of 0.004 this feature is C≡N (nitrile, isonitrile) and possibly C≡C stretching vibrations.
Absorbance 0.002 0.000 4000 3500 3000 2500
ν(cm )
-1

2000

1500

1000

Figure 1 Ex-situ FTIR spectrum of CN dust particles (diameter about 100nm).

NEXAFS measurements on the nanoparticles gave us better insight in the charachteristics of nanoparticles. From normalized C K edge spectra, as in Figure 2, we can see almost negligable presence of C=C bonds at 285 eV (signed within the Figure 2 as C1)). Therefore, and according to explanations proposed in the literature, we can suppose that nitrogen intensified the C=C vibrations in the IR spectrum. This can partially explain the dominant feature in the fingerprint area (1000-1800 cm-1). Further resonance line can be observed at 286.6 eV (C2), in the

240

literature correlated with C1s→π* (C≡N), and with some π* resonances from C=O [3]. Some authors [4] associate this peak also with some C=N bonds. This possibility can be in our case not excluded, for one reason that we observe C=N in the FTIR spectrum. The other important reason why not to exclude several possibilities is that nanoparticle formation in plasma processes results in complex polymers with large variety of bonding situations. The resonances at 288eV (C3) can be correlated with C-H σ*; although there is also a possible correlation with C=O originating from amide groups.. This peak has a shoulder at about 287.7 eV, correlated with aliphatic and ketonic C-H groups. Around and above 292 eV (edge, C4) we can observe different C1s→σ* resonances, with shape and position, together with previously mentioned σ*(C-H) states, typical for plasma polymerized samples [5]. We observe also a shoulder around 301eV (C5), probably coming from σ* C=C and C=N resonances [5].
3
2.0 Normalized total electron yield (TEY)
C4 C5 C3 C2

N1 Normalized total electron yield (TEY)

1.5

2

1.0

1 N2

0.5
C1

0

0.0 280 285 290 295 300 305 310 315 Photon energy (eV)

390 395 400 405 410 415 420 425 Photon energy (eV)

Figure 2 NEXAFS spectrum: C K edge

Figure 3 NEXAFS spectrum: N K edge

The dominant pre-edge peak is positioned at 399.7 eV(N2), with a shoulder at about 399 eV. The main peak can present a variety of π nitrogen bonds; hence the unique identification is obviously very complicated. On the basis of calculation data for carbon-nitride films some authors [6] show that such kind of low energy peak can be related to constrained CN bonds, the presence of pyridine-like double bonds (–C=N–), or imines. Some authors ascribe this peak to N≡C bonds (nitriles, isonitriles)[3]. This kind of bonds is, actually, for most of polymer materials in IR spectra (around 2100 cm-1) either not observable or present as a weak feature. One explanation in favor of nitrile presence in the material is: aliphatic nitriles absorb weakly in IR, so it is possible that there is a more significant presence of these groups as assumed. On the other hand, NEXAFS is very sensitive to the presence of these groups and could overemphasize the real picture. Furthermore we can observe a change of the slope around 402 eV and the strong broad σ* resonance around 407 eV. The presence of σ bonds is observable through the strong broad resonance above 407 eV (N-C, N-H). The annealing of the above analyzed nanoparticles was performed in several steps, with temperature increase of about 20 K/min. The NEXAFS spectra from annealing experiment are presented without normalization in order to stress losses we could observe from nanoparticle surface and bulk material. Although measurements have been obtained after every 50K step, we present only certain steps, in order to present important changes more clearly. Figure 4 shows the C K edge at different temperatures. The changes in spectra start first above 550K. Above this temperature we can observe the increase of C=C sites (peaks C1+C5), and a gradual loss of C=O and C-H correlated peaks (C3). After 700K spectra reveal changes in the C2 peak, correlated with nitrogen presence. The changes observed above 550K can be correlated with the behavior of oxygen, i.e. oxygen removal. From the NEXAFS spectra on the O K edge (Figure 5) we can observe the start of this oxygen removal at 550 K. The oxygen presence is negligible above 650 K. Fig. 6 shows the changes of the NEXAFS spectra at certain temperatures for the N K edge. Spectra were unperturbed until 650K (oxygen loss). Above 700K (Figure 7) spectra reveal drastic changes in the peaks N1 and N2 (for identification see Fig 3). N1 increases, while N2 decreases, and new peaks arise, like the feature around 401-402eV. The whole preedge region becomes weaker compared to the 407 eV region (σ* resonances). In comparison with the previous reports on CN film annealing (e. g [7]), where the nitrogen loss starts already around 450K, we observed losses around 800K. After 850 K the material is rather stable.

241

However, due to the complexity of the material and the complicated identification of spectral features, this work requires further analysis like elemental analysis and especially theoretical calculations of NEXAFS spectra, which are beyond the scope of this report.
4.0

30 800 K 25 Total electron yield (TEY) 700 K 650 K 20

900 K
3.5 Ttotal electron yield (TEY)

800 K 750 700 3.0 650 600 2.5 550 500 400 1.5 300 1.0
520 530 540 550 560 570 580

15

600 K 550 K

2.0

10

500 K 300K

5 280 285 290 295 300 305 310 315 Photon energy (eV)

Photon energy (eV)

Figure 4 C K edge NEXAFS spectra: annealing

Figure 5 O K edge NEXAFS spectra: annealing

650 K 6 Total electron yield (TEY) 600
Total electron yield (TEY)

3.0

1020 K 950 K

2.5

550 4 500 400 2 300

900 K 850 K

2.0

800 K

1.5

750 K 700 K

1.0

0 390 395 400 405 410 415 420 425 Photon energy (eV)

390 395 400 405 410 415 420 425 Photon energy (eV)

Figure 6 N K edge NEXAFS spectra: up to 650K

Figure 7 N K edge NEXAFS spectra: above 650K

Acknowledgments: The authors also acknowledge support by BMBF 05 ES3XBA/5 and EC under IA-SFS Contract RII3-CT2004-506008. References [1] H. Biederman, Plasma polymer films, Imp. Colledge Press, London, UK (2004). [2] A. C. Ferrari, S.E. Rodil, and J. Robertson, Physical Review B 67, 155306 (2003). [3] A.G. Shard, J. D. Whittle, A. J. Beck, P. N. Brookes, N. A. Bullett, R. A. Talib, A. Mistry, D. Barton, S. L. McArthur, J. Phys. Chem. B 108, 12472 (2004). [4] C. Lenardi, M.A. Baker, V. Briois, G. Coccia Lecis, P. Piseri, W.Gissler, Surf. Coat.Technol. 125, 317 (2000). [5] J. Stöhr, “NEXAFS Spectroscopy”, Springer-Verlag, Heidelberg (1992). [6] F. Alvarez and M.C. dos Santos, Journal of Non-Crystalline Solids 266, 808 (2000) [7] G-Q. Yu, S-H. Lee, J-J. Lee, Diamond and Related Materials 11, 1633 (2002).

242

Features of resonant F KLL Auger spectra from titanium trifluoride TiF3 and fluorinated multi-walled carbon nanotubes F-MWCNTs.
А.S. Vinogradov1, M.M. Brzhezinskaya1,2, A.V. Generalov1, A.Yu. Klyushin1, and R. Püttner3
1

V.A. Fock Institute of Physics, St. Petersburg State University, St. Petersburg, 198504 Russia 2 HZB (BESSY GmbH), 12489 Berlin, Germany 3 Institut für Experimentalphysik, Freie Universität Berlin, 14195 Berlin, Germany

The electronic structure of solid metal fluorides, that present the most ionic crystals, is usually considered in the framework of the simple ionic model. At the same time, the covalent mixing (hybridization) between the fluorine 2p electrons and valence electrons of metal atoms has a dramatic effect on the electronic structure of their compounds. Using x-ray absorption spectroscopy at the F 1s edge, we have shown previously that even in the case of such strongly ionic compounds of the 3d transition-metal (TM) fluorides, covalent bonding plays an important role, causing a mixed TM 3d – F 2p character of the empty electronic states near the bottom of the conduction band [1]. Recently we also have observed the strong effect of the covalent bonding on electronic structure of multi-walled carbon nanotubes (MWCNTs) upon their fluorination [2]. The main aim of the present study was to gain a deeper insight into the nature of chemical bonding and electronic structure features for the titanium trifluoride, TiF3, and the fluorinated multi-walled carbon nanotubes, F-MWCNTs, using highresolution measurements of resonant F KLL Auger spectra. Resonant Auger spectroscopy is well known to provide a new and powerful method to study the core-electron excitations in atoms and ionic crystals [3-5]. In principle, there is a close analogy between the core-excited states of atoms and those in solids: in both cases the excited electron is bound to the inner hole of the parent atom by electrostatic forces. At the same time, this electron can be localized in solids only in the volume that is confined to neighboring atoms of the parent one for the excited electron. These atoms are chemically bonded to the parent atom and therefore are involved in a formation process of possible excited states for the core electron under consideration. From this it is clear that the covalent bonding between the parent atom and its neighbors can strongly influence the localization character of the excited electron and its localization extent on the parent atom which define resonant spectator and participator Augerdecay processes for the core-excited electron and spectral shape of resonant Auger spectra. All measurements have been performed at the Russian-German beamline (RGBL) using experimental station Mustang. Thin (20-25 nm) TiF3 layers were prepared in situ in the preparation chamber by thermal evaporation of thoroughly dehydrated TiF3 powder from a effusion cell heated by an electron beam onto a polished stainless-steel plate in a vacuum of ~3×10-7 mbar. A sample of F-MWCNTs (10 wt. % of fluorine) was prepared in air by rubbing powder of it into the scratched surface of a clean stainless-steel plate. NEXAFS spectra at the F 1s edges for these samples were obtained in the total electron yield mode by detecting a sample current. The photon energy was calibrated using the known energy position of the first narrow peak in the F 1s absorption spectrum of solid K2TiF6 (F 1s → t2g; 683.9 eV [6]) and the photon-energy resolution was set to 200 meV at the F 1s edge (~685 eV). Photoelectron and F KLL Auger spectra for TiF3 and F-MWCNTs were collected in the angle-integrated mode with the total energy resolution of 400 meV using a Phoibos 150 electron analyzer. No sample charging effects were observed during the absorption and photoelectron measurements that were carried out at a pressure in the measuring chamber about 2·10-10 mbar. The F 1s absorption spectrum recorded for TiF3 is compared in Fig. 1 with the spectra for other titanium fluorides, K2TiF6 and TiF4, that were measured previously [1]. As is well seen, the spectra for ionic compounds TiF4 and TiF3 as well as for a molecular anion TiF62- (in solid K2TiF6) are strongly similar in their spectral shape. Taking into account the similar octahedral fluorine environment of the titanium atom in these compounds and the quasimolecular analysis of the spectra for TiF4 and TiF62- [1], the low-energy absorption band a –

243

b΄ in F 1s absorption spectrum of TiF3 can be also associated with core-electron transitions to the empty weakly antibonding MOs t2g and eg of the octahedral TiF6 quasi-molecule formed by the Ti atom and the surrounding fluorine atoms. These MOs result from the covalent mixing between the Ti 3d electron states, which are split into 3dt2g and 3deg components in the field of the fluorine octahedron, and the a c F1s absorption spectra 2p states of the fluorine atoms. It should be b noted that in the context of a simple ionic d e e' h f f' g model these MOs are essentially localized at the central (Ti) atom and are primarily K2TiF6 a' b d determined by the 3dt2g and 3deg states, b' c h e which can be observed in the F 1s f a 8 6 a'b b' absorption spectrum of the TiF62- anion due d 10 5 TiF4 9 h 7c to covalent bonding between the Ti and F g e a 4 f atoms. Finally, the variations of the low3 TiF3 2 1 energy absorption structures, when going from TiF62- to TiF4 and TiF3,.result from a tetragonal and trigonal 680 690 700 710 720 730 considerable distortions of the TiF6 octahedra in these Photon Energy, eV titanium fluorides. Fig.1. F 1s absorption spectra of TiF3, TiF4, and K2TiF6. The electron (valence-band, corelevels and Auger) spectra for TiF3 excited by photons of various energies in vicinity of the F 1s absorption edge (as marked by numbered arrows in the absorption spectrum, Fig.1) are presented in Fig. 2. It is seen from Figure that contributions from the valence-band and corelevels signals to Auger spectra are small. The nonresonant F KLL Auger spectrum excited by photons with energy far above the F 1s edge (curve 10; hν = 719 eV) exhibits a spectral shape that is typical of the normal Auger spectrum of the neon-like F- anion in alkali fluorides [4]. It comprises three groups of lines (a a F KLL hν, eV and b, cd and e, f) associated with KL2,3L2,3 b e d c 719.0 f 10 (1D and 1S), KL1L2,3 (3P and 1P), and KL1L1 713.0 Ti 3p 9 Auger transitions. Examining the (1S) series of Auger spectra (curves 2 – 7), it is 695.6 F 2s 8 clear that all the Auger signals change their 689.0 intensities and energy positions, when the 7 photon energy is scanned across of the low-energy absorption band (resonance) a 687.3 6 – b΄. All resonant spectra are characterized a by a high-energy shift with respect to the 685.5 5 normal one that decreases nonlinearly in its b 684.8 magnitude from 5.5 eV to 0.3 eV with 4 a b 684.2 e d c increasing the photon energy in the range f 3 a 682.5 between 682.5 eV (curve 2) and 689.0 eV e 2 ×3 675.0 (curve 7). Note that the valence-band ×3 1 Ti 3p F 2s Ti 3s VB photoelectron signal VB therewith does not 600 620 640 660 680 substantially change. Evidently these Kinetic Energy, eV findings are caused by characteristic Fig.2. Resonant F KLL Auger spectra of TiF3. features of decay processes for the coreexcited F atoms arising from transitions of the F 1s electrons to low-energy unoccupied electron states of TiF3 (MOs t2g and eg; absorption band a - b΄). Owing to the appreciable localization of the excited electron on the fluorine atoms, it has effect on the Auger decay processes for 1s hole in the fluorine atoms. Only slight modifications on the line shape and intensity of the VB signal upon the resonance excitation are indicative of an absence of participator Auger transitions. In such a situation it
Photoelectron Intensity
Total Electron Yield

244

is clear that the low-energy F 1s excitations in TiF3 decay predominantly via spectator Auger transitions. In this deexcitation process the electron excited to the Ti 3d – F 2p hybridized electron state (MOs t2g and eg) will remain as a screening electron both in the initial and final states, thus causing an increase in the energy of the outgoing electron compared to the normal Auger decay (with a fully delocalized excited electron). The observed decrease of this high-energy shift for the resonance Auger transitions, 6 F-MWCNT C* * * when the photon energy is scanned across of D1 7 D* B1 B * 2 2 E* F* the resonance a – b΄, points to the fact that 5 8 the localization extent of the low-energy 4 core-excited states is rapidly reduced with F 1s absorption increasing the photon energy over the 12 3 spectrum absorption resonance. Further inspection of the resonant F 680 690 700 710 720 KLL Auger spectra for F-MWCNTs Photon energy, eV containing 10 wt% of fluorine atoms (Fig. 3) a hν, eV F KLL discloses similar regularities that also can be 713.0 e d explained by the spectator Auger f F 2s deexcitation processes of the low-energy 8 core-excited states which have the 697.0 hybridized C 2p – F 2p character and are responsible for the band B1* in the F 1s 688.5 7 absorption spectrum. Differences which are 687.8 6 observed in the resonant Auger spectra of FMWCNTs as compared to those of TiF3 (a 5 lower resolved spectral shape and a lower 687.3 energy shift of 4.2 eV) can be understood taking into account the presence of two C-F 685.6 4 phases [7] and a weaker covalent bonding a d 684.4 f between the fluorine and carbon atoms in F2 683.4 MWCNTs. 1 F 2s Val-band In conclusion, the analysis of the 600 620 640 660 680 resonant F KLL Auger spectra measured with Kinetic Energy, eV a high energy resolution for TiF3 and FMWCNTs has shown that these spectra Fig. 3. Resonant F KLL Auger spectra contain direct and otherwise inaccessible of F-MWCNTs. information concerning a localization and hybridization character for the low-energy unoccupied electronic states.
This work was supported by the Russian Foundation for Basic Research (projects no. 06-02-16998, 09-0201278) and the bilateral Program “Russian-German Laboratory at BESSY”. A.Yu.K. acknowledges the support from Freie Universität Berlin through the Leonhard-Euler Fellowship Program. References 1. A.S. Vinogradov, S.I. Fedoseenko, S.A. Krasnikov, et al. Phys. Rev. B 71, 045127 (2005). 2. M.M. Brzhezinskaya, N.A. Vinoradov, V.E. Muradyan, Yu.M. Shul’ga, N.V. Polyakova, A.S. Vinogradov. Fizika Tverdogo Tela 50, 565 (2008). [Phys. Sol. State 50,587 (2008)]. 3. H. Aksela, S. Aksela, J. Tulkki, et al. Phys. Rev. A 39, 3401 (1989). 4. H. Aksela, E. Kukk, S. Aksela, et al. Phys. Rev. B 49, 3116 (1994). 5. E. Kukk, S. Aksela, H. Aksela, et al. Phys. Rev. B 50, 9079 (1994). 6. A.S. Vinogradov, A.Yu. Dukhnyakov, V.M. Ipatov, D.E. Onopko, A.A. Pavlychev, and S.A. Titov. Fizika Tverdogo Tela 24,1417 (1982) [Sov. Phys. Solid State 24, 803 (1982)]. 7. M.M. Brzhezinskaya, N.A. Vinoradov, V.E. Muradyan, Yu.M. Shul’ga, R. Püttner, A.S. Vinogradov, and W. Gudat. Fizika Tverdogo Tela in print (2009). [Phys. Sol. State in print (2009)].

Photoemission Intensity

Total Electron Yield

245

Magnetic domain structure of Heusler/MgO/Heusler trilayer systems
A. Kaiser, C. Wiemann, S. Cramm, C.M. Schneider
Forschungszentrum J¨lich, Institut f¨r Festk¨rperforschung IFF-9, 52425 J¨lich, Germany u u o u

Heusler alloys [1] are considered as interesting ferromagnetic electrode materials for magnetic tunnel junctions (MTJ). Due to their high spin polarization at the Fermi level they are expected to show extremely high tunnelling magnetoresistance (TMR) values. MgO as a tunneling barrier material has a comparable lattice constant and thus provides the possibility of epitaxial growth of trilayer systems [2]. Due to the reduction of defects and the onset of resonant tunnelling mechanisms an increase of the TMR effect can be expected [3].

5 µm
(a) CMS 10x10 µm2 (b) Sketch of domain pattern in CMS 10x10 µm2

Many macroscopic studies of the magnetic, electronic and structural properties of Heusler-based MTJs have been carried out. From a fundamental point of view, but also with respect to technical aspects, the micromagnetic behaviour cannot be neglected. Co2 FeSi (CFS) and Co2 MnSi (CMS) are two protagonists of the class of half-metallic Heusler compounds. They have similar lat5 µm 5 µm tice constants providing structural compatibility to MgO. Both materials have very high Curie temperatures around (c) CMS 20x20 µm2 (d) CFS 20x20 µm2 1000 K [4, 5] and magnetic moments per formula unit of 5.07 µB (CMS) and 6 µB , respectively. Hysteresis measurements reveal clearly distinguishable coercive fields of Figure 1: Magnetic domain structures in patterned CMS and CFS elements. The magnetic contrast has been obtained at 2.8 mT (CMS) and 6.5 mT (CFS).
the Co L3 edge.

Single films and trilayer structures with asymmetric electrode configurations have been prepared by magnetron sputtering. The films have been subsequently microstructured by optical lithography and argon ion beam milling into squares with areas ranging from 2×2 to 100×100 µm2 . A more detailed description of the growth conditions and the experimental results can be found elsewhere [6]. The micromagnetic structure of the films has been studied by photoemission electron microscopy (PEEM). PEEM exploiting the XMCD effect is a powerful technique for the element-selective study of magnetic domain configurations [7, 8]. The measurements have been conducted using an Elmitec PEEM III at the beamline UE56/1-SGM at BESSY-II. All measurements shown in this report have been generated by tuning the photon energy to the appropriate L3 absorption edge and calculating the XMCD asymmetry value A = (Iσ+ − Iσ− )/(Iσ+ + Iσ− ) for each pixel with Iσ± being the intensity values for right and left circularly polarized light. Fig. 1 shows the magnetic domain patterns from single CMS and CFS films. Under the influence of the shapeinduced demagnetizing field, the magnetization configuration of elements of comparable size is distinctly different. The CMS film develops a so-called concertina or buckling pattern (shown in fig. 1(c)) also known from Permalloy elements in a similar thickness range [9]. It is formed by alternating low-angle walls with the local magnetization direction varying around the average magnetization. With decreasing element size the effect of the demagnetizing field becomes stronger and successively simpler flux-

closure patterns reminiscent of Landau states start to form (Fig. 1(a) and 1(b)), which are still accompanied by buckling structures. The latter disappear for elements in the micrometer regime. However, the occurrence of the buckling state is not necessarily the magnetic ground state configuration, but may arise due to a local energetic minimum caused by neighbouring domains blocking each other. A completely different response is observed in the CFS films. Even under the influence of the demagnetizing field in small 10 × 10 µm2 elements (Fig. 1(d)) the polycrystalline nature of the film is dominating the magnetization pattern and the fine-grained domain structure remains essentially unchanged from that observed in the extended film (not shown). This result shows that the intrinsic anisotropy of the CFS-film is much stronger than the demagnetizing field of the square element. In a second step the single Heusler films have been combined into trilayer structures with a MgO interlayer of 3 nm thickness. In order to separate the magnetic response of the individual layers in this stack, the full versatility of soft x-ray PEEM is needed. By tuning the photon energy of the incident beam to the L3 -absorption edges of Fe and Mn the magnetization of both ferromagnetic layers can be investigated independently. Resulting domain images for a square element of CMS(20 nm)/MgO(3 nm)/CFS(2 nm) with an edge length of 10 µm are compiled in fig. 2. Due to the limited escape depth of the photoelectrons, the Mn signal is rather weak and had to be upscaled by a factor

246

5 µm
(a) Fe-L3 (b) Mn-L3

5 µm

Figure 2: Element-selective domain imaging in the layer system CMS/MgO/CFS, revealing a parallel magnetic coupling of the CFS and CMS films.

(a) CMS (20 nm)/ MgO (b) CFS (20 nm)/ MgO (3 nm)/ CFS (2 nm) (3 nm)/ CMS(2 nm)

of five. Comparing the domain patterns of the Fe and Mn data reveals identical structures consisting of Landau fluxclosure pattern superposed by concertina features in both films. The reasons for this coupling can be a roughnessinduced N´el/orange-peel mechanism [10] or pinholes in e the MgO layers, which favour a ferromagnetic contact between the CFS and CMS layer through a direct exchange interaction. The domain patterns of the trilayer film resemble the situation of the single CMS film (fig. 1a). Due to the difference in thickness in both films the micromagnetic structure is strongly dominated by the CMS bottom layer. For larger 20 × 20 µm2 elements (fig. 3a) the magnetic structure is no longer determined by the flux-closure. Instead the competition between local anisotropy fluctuations and demagnetizing field is won by the anisotropy and a magnetization ripple due to the polycrystalline structure of the films is formed. In the inverse trilayer system the magnetic microstructures changes drastically. Instead of the anisotropy-dominated ripple pattern we find a different behaviour with higher average domain size and the formation of a low-remanence magnetization pattern consisting of two antiparallel Landau domains, as shown in fig. 3b. Some of the 90◦ -walls have been replaced by an additional domain with two lowangle walls. This configuration is known as “Tulip” state [11]. The 180◦ -walls between neighbouring antiparallel domains are modified by a high density of cross-tie structures. By the formation of cross-ties the magnetostatic energy is reduced, since 180◦ -walls are replaced by energetically more favorable 90◦ -walls [12]. In this trilayer structure we do not find a magnetic contrast at the Mn edge. This fact is surprising since the CMS film is the top layer and is expected to yield a higher intensity than in the reversed stack. Thus we must conclude that the CMS film is nonmagnetic at room temperature. This behaviour may be attributed to a strong thickness dependence of the CMS magnetic moment in thin films below 8 nm thickness supporting the results of previous XAS and FMR measurements [13, 14, 15]. The strongly reduced Curie temperature in the 2 nm CMS film may be explained by interdiffusion at the interface leading to a higher atomic disorder. Furthermore, this result seems to indicate that the MgO barrier in this layer has only a negligible density of pinholes, because a direct exchange coupling to the bottom CFS layer should also result in a common Curie temperature for both layers.

Figure 3: Comparison of magnetic domain patterns acquired at the Co L3 edge of 20 × 20 µm2 square elements of both trilayers.

In conclusion our element-selective domain imaging experiments reveal the complexity of the magnetic microstructure in Heusler-based thin film systems. The results also show that the micromagnetic structure depends on fine details of the formation process of the Heusler phases. Analysis of the domain configurations shows that the ferromagnetic coupling observed in the dual-Heusler trilayers can be attributed to roughness-induced N´el coupling. This can e be overcome by an improvement of the preparation conditions. The surprising difference of the magnetic behaviour between the CMS/MgO/CFS and CFS/MgO/CMS trilayer structures is due to a strong thickness dependence of the magnetic ordering in CMS and must be taken into account for the construction of magnetic tunnelling junctions. Acknowledgements We thank D. Rata and D. Banerjee for the deposition of the samples and hysteresis measurements. This work was financially supported by the DFG (SFB 491). References
[1] Heusler, F. Verh. Dtsch. Phys. Ges. 12, 219 (1903). [2] Yamamoto, M., et al., J. Phys. D: Appl. Phys. 39, 824 (2006). [3] Yuasa, S., et al., Nature Materials 3, 868 (2004). [4] Webster, P. Contemp. Phys. 10, 559 (1969). [5] Wurmehl, S., et al., Phys. Rev. B 72, 184434 (2005). [6] Kaiser, A., et al., J. Magn. Magn. Mater. (2008). doi:10.1016/j.jmmm.2008.10.037. [7] Sch¨nhense, G. J. Phys.: Cond. Matt. 11, 9517 (1999). o [8] St¨hr, J. and Anders, S. IBM J. Res. Develop. 44, 535 o (2000). [9] Chumakov, D., et al., Phys. Rev. B 71, 014410 (2005). [10] Neel, L. C. R. Acad. Sci. 255, 1676 (1962). [11] Hubert, A. and R¨hrig, M. J. Appl. Phys. 69, 6072 (1991). u [12] Huber Jr, E., et al., J. Appl. Phys. 29, 294 (1958). [13] Schmalhorst, J., et al., Phys. Rev. B 70, 24426 (2004). [14] Wang, W., et al., Phys. Rev. B 71, 144416 (2005). [15] Rameev, B., et al., phys. stat. sol.(a) 1, 10 (2006).

247

Effect of the partly filled HOMO in CoPc on formation and decay processes of Co 2p3/2 excitations from a comparison with NiPc.
А.S. Vinogradov1, M.M. Brzhezinskaya1,2, A.V. Generalov1, A.Yu. Klyushin1, and K.A. Simonov1
1

V.A. Fock Institute of Physics, St. Petersburg State University, St. Petersburg, 198504 Russia 2 HZB (BESSY GmbH), 12489 Berlin, Germany

The 3d-metal phthalocyanines (3d-MPc’s) are very stable planar complexes that show various interesting properties and have extensive applications in the areas of catalysis, pigments, semiconductors and sensors [1]. The central part of the 3d-MPc’s (including the 3datom and its nearest surroundings) is believed to be their most reactive part and to determine the most important applications of these compounds. The occupied and empty 3d electron states of the metal atom, which are located near the Fermi level, are essentially responsible for the unique properties of the 3d-MPc’s. Since the practical application of these phthalocyanines is based on the understanding their electronic structure, its investigation remains a subject of intense research over several decades. In particular, the occupied electronic states of 3d-MPc’s were investigated by photoemission (PE) spectroscopy [2,3], while x-ray absorption (XA) and inverse photoemission spectroscopy were used for studying the unoccupied states [4-6]. The similar studies are usually performed for the whole series of MPc’s (M = Fe, Co, Ni, Cu and Zn) since their electronic structure is mainly characterized by a gradual increase of the 3d electron number for the metal atoms along this series. Further, the metal 3d derived electronic states of MPc’s are usually regarded, up to now, as the nearly pure atomic 3d components split by a square-planar (D4h) molecular field of the complex into the eg (dxz,yz), b2g (dxy), a1g (dz^2), and b1g (dx^2-y^2) components [7]. As a result, in going from FePc to ZnPc the metal atom 3d electron configuration varies from eg4a1g2b2g0b1g0 to eg4 a1g2b2g2b1g2. The similar electron configurations are known to be best probed in an x-ray absorption experiment by excitation of metal 2p core electrons to the unfilled 3d derived electronic states [8]. In such the probing, a gradual filling of emty b2g and b1g states must be observed in the metal atom 2p absorption spectra of MPc’s. A most interesting situation is expected in the case of Co 2p excitations of CoPc since the 2p 6 Ni 2p3/2 spectrum NiPc electron excited to the unfilled 3db1g will interact 5 A additionally with the 3d electron in the partly 4 filled b2g HOMO. At the same time this HOMO 7 3 is fully occupied for the NiPc complex and the LUMO is the empty 3db1g state. The main aim of 89 10 11 the present study was to gain a detailed 2 12 13 1 information on formation and decay processes for Co 2p excitations from the high-resolution 848 852 856 860 864 absorption and resonant PE (Res PE) spectra 9 5 11 Co 2p CoPc measured at the Co 2p edges and to analyze 3/2 spectrum A* A1 10 A2 effects of the partly filled HOMO and 3d-3d 4 12 electron interaction on decay processes of the Co 6 2p excitations on the basis of comparison 8 13 between corresponding spectra for CoPc and 7 3 NiPc. Below are presented the first results of our 14 2 1 measurements performed last December. All measurements were performed at the 772 776 780 784 788 Russian-German beamline (RGBL) using Photon Energy, eV experimental station Mustang. The NiPc and Fig. 1. Metal atom 2p3/2 absorption spectra CoPc chemicals were purchased from Aldrich for NiPc and CoPc. Inc.. The samples were thin (30-50 nm) polycrystalline NiPc and CoPc films prepared in situ in the preparation chamber by thermal
Photoabsorption (Total Electron Yield)
248

evaporation of powders from a Knudsen cell onto a polished stainless-steel plate in a vacuum of ~3×10-7 mbar The NEXAFS spectra of these samples were obtained in the total electron yield mode by detecting a sample current. The photon-energy resolution was set to 550 meV at the Ni 2p3/2 edge (~850 eV) and 480 meV at the Co 2p3/2 edge (~780 eV). The photon energy was calibrated using the known energy position of the first narrow peak in the Ne 1s absorption spectrum (Ne 1s → 3p; 867.13 eV [5]) and in the F 1s absorption spectrum of solid K2TiF6 (F 1s → t2g; 683.9 eV [6]). Valence-band and core-level photoelectron (PE) spectra for NiPc and CoPc were collected in the angle-integrated mode with a Phoibos 150 electron analyzer. The total energy resolution was about 580 meV and 520 meV for the Res PE spectra at the Ni 2p3/2 and Co 2p3/2 edges, respectively. The energy scale was aligned by measuring the 4f7/2 PE lines of the reference gold or platinum foils. All spectra were normalized to the incident photon flux, which was monitored by recording the photocurrent from a gold mesh placed at the outlet of the beamline. No sample charging effects were observed during the absorption and photoelectron measurements that were carried out at a pressure in the measuring chamber about 2·10-10 mbar. The long x-ray irradiation of the CoPc sample was found to result in its minor deterioration which was only manifested in small changes of relative intensities of absorption structures in Co 2p3/2 spectrum. Nevertheless, the position of the light spot on the sample was varied every 2 h in order to ensure insignificance of this photochemical effect. Ni 2p3/23d3d The metal NiPc hv,eV atom 2p3/2 absorption ×3 859.7 d c ba spectra recorded for 13 857.7 ×3 CoPc and NiPc are 12 856.5 ×2 compared in Fig. 1. 11 855.6 ×2 As is well seen, the 10 Co spectrum is 855.3 significantly different 855.0 9 from the Ni one and 8 shows two additional low-energy 854.8 absorption structures A* and A1. The more 7 complicated structure of Co 2p3/2 absorption 854.3 spectrum can be 854.1 understood taking into account a 6 854.0 presence of the partly 5 filled HOMO 3db2g in 853.8 4 the cobalt complex. 852.8 3 f Compared to the e d 850.7 c b a 2 spectrum of NiPc, the 1 additional low-energy 30 25 20 15 10 5 0 band A* is attributed to a Co 2p3/2 -> 3db2g Binding Energy, eV transition while the Fig. 2. Res PE spectra of NiPc at the Ni 2p3/2 edge. main absorption transition of Co 2p3/2 electrons to a 3db1g state (band A in the Ni2p3/2 spectrum) should be regarded to be split into two triplet and singlet components (A1 and A2) due to an 3d-3d exchange interaction between the electron excited to the 3db1g state and the 3db2g electron.

Photoelectron Intensity

249

PE spectra of NiPc and CoPc excited by photons of various energies in vicinity of the Ni 2p3/2 and Co 2p3/2 absorption edges (as marked by numbered arrows in the absorption spectra, Fig.1) are shown in Figs. 2 and 3, respectively. The bottom curves in both figures correspond to photon energies below the Ni 2p3/2 and Co 2p3/2 excitations. Structures a΄ - f in these spectra are superpositions of the valence MO PE signals originating from the nitrogen and carbon 2p and 2s atomic states as well as from Ni or Co 3d states. Examining the series of PE spectra in Fig.2 and 3, it is clear that CoPc e ×1.5 788.6 some of valencec b a d a' 14 Co 2p3/23d3d band PE signals are significantly 782.2 enhanced when the 13 photon energy is 781.5 scanned across the 2p – 3d absorption 780.9 12 resonances (band A 780.4 in Ni spectrum and 11 A*, A1 and A2 in Co spectrum). It is of 10 779.6 interest that the 779.0 enhancement effects 778.4 ×1,3 are different for Res 9 PE spectra of NiPc 778.0 7 and CoPc. Evidently 6 these findings are caused by the 777.7 presence of the partly filled b2g HOMO in 777.4 ×1.3 5 ×2 CoPc. More detailed 777.0 4 x2 analysis of these 776.6 3 f ×2 experimental data is e 2 760.3 d c b a a' 1 in progress.

Photoelectron intensity

30

25

20

15

10

5

0

Binding Energy, eV Fig. 3. Res PE spectra of CoPc at the Co 2p3/2 edge.
This work was supported by the Russian Foundation for Basic Research (projects no. 06-0216998, 09-02-01278) and the bilateral Program “Russian-German Laboratory at BESSY”.
References 1. 2. C.C. Leznoff, A.B.P. Lever, Phthalocyanines, Properties and Applications. Vol. 3, VCH Publishers, Inc., NY, 1993. W.D. Grobman, E.E. Koch, in: L.Ley, M. Cardona (Eds.), Photoemission in Solids II, Springer-Verlag, Berlin 1979, Chapter 5.; E.E. Koch, in: P. Reineker, H. Haken, H.C. Wolf (Eds.), Organic Molecular Aggregates, Springer-Verlag, Berlin 1983, p.35. M. Iwan and E.E. Koch, Solid State Commun. 31, 261 (1979). E.E. Koch, Y. Jugnet, and F.J. Himpsel, Chem. Phys. Lett. 116, 7 (1985). M.L.M. Rocco, K.-H. Frank, P. Yannoulis, and E.E. Koch, J. Chem. Phys. 93, 6859 (1990). H. Yoshida, K. Tsutsumi, N. Sato, J. Electron Spectrosc. Relat. Phenom. 121, 83 (2001). H. Höchst, A. Goldmann, S. Hüfner, and H. Malter, Phys. Stat. Sol. (b) 76, 559 (1976). H. Ebert, J. Stöhr, S.S.P. Parkin, M. Samant, A. Nilsson. Phys. Rev. B 53, 16067 (1996).

3. 4. 5. 6. 7. 8.

250

Charge order in La1.8-xEu0.2SrxCuO4 studied by resonant soft X-Ray diffraction
V. Soltwisch , E. Schierle , E. Weschke , J. Geck , D. Hawthorn , H. Wadati , H.-H. 3 1 4 4 5 2 1,4 Wu , H. A. Dürr , N. Wizent , B. Büchner , P. Ribeiro ,G. A. Sawatzky , J. Fink 1 2 Helmholtz Zentrum Berlin für Materialien und Energie, University of BC Vancouver 3 4 Canada, II.Physikalisches Institut, Universität Köln, Leibniz Institute for Solid State and 5 Materials Research Dresden, IFW Dresden In high-Tc superconductors at low doping concentrations x the doped holes tend to order within the CuO2 planes. This ordering is determined by a complex interplay between charge, spin and lattice degrees of freedom and results in stripe-like or checkerboard phases. The charge stripes constitute antiphase domain walls between antiferromagnetic domains and their role for the occurrence of O-K 6K superconductivity is not yet understood. Stripe order, 20 K 40 K observed near x = 1/8, is fluctuating in many hole50 K doped cuprates, but can be stabilized by corrugation of 60 K 70 K the CuO2 planes as in La2-xBaxCuO4 (LBCO), 100 K La1.6-xNd0.4SrxCuO4 (LNSCO) as well as in the compound La1.8-xEu0.2SrxCuO4 (LESCO). LESCO 0.20 0.25 0.30 in particular has a strong corrugation due to the small h (r.l.u.) Eu ions and exhibits static stripes in a large range of x, Fig. 1. (ε 0 l) superstructure peaks the superconducting range between 0.080.2 [6]. A large doping concentration total fluorescence yield (black) and calculated per Cu site in the charge stripes was also intensity obtained from a Kramers-Kronig concluded in [4]. transform of absorption data (blue) are shown. The data of Fig. 1 can be fitted by Lorentzian line shapes (solid lines), the resulting integrated intensities as a function of temperature for doping concentrations ranging from x = 0.1 to x = 0.15 are shown in Fig. 3. Evidently, the intensities tend to vanish at temperatures between 40 K (x = 0.1) and 80 K (x = 1/8). These temperatures are significantly smaller than the transition temperature of the low-temperature tetragonal phase of approximately 120 K [1], indicating that charge order is not driven by this structural transition. More significantly, the charge ordering temperature is substantially larger than the spin ordering temperature [1]. This means that while the charge stripes occur as a periodic arrangement of domain walls between antiferromagnetic antiphase domains, they do no depend on long-range magnetic order. Financial support by the DFG is appreciated by the IFW based researchers (Forschergruppe FOR 538) and by J.G.. The UBC based researchers acknowledge financial support from the Canadian granting organizations NSERC, CIFAR and CFI.
1.0 0.5 0.0 1.0 0.5 0.0 0 Integrated intensity (arb. units) x=0.15 O-K x=0.125 O-K

References:
x=0.11 O-K x=0.1 O-K

20 40 60 80 0 Temperature (K)

20 40 60 80 Temperature (K)

Fig. 3: Integrated intensity of the (ε 0 0.75) superstructure reflection as a function of temperature for various doping levels x as determined from data obtained at the O-K resonance.

[1] H. H. Klauss et al., Phys. Rev. Lett. 85, 4590 (2000). [2] J. M. Tranquada et al., Nature 375, 561 (1995). [3] M. Von Zimmermann et al., Europhysics Letters 41, 629 (1998). [4] P. Abbamonte et al., Nature Physics 1, 155 (2005). [5] K. Yamada et al., Phys. Rev. B 57, 6165 (1998). [6] J. Fink et al., Phys. Rev. B, accepted (2009).

252

Charge order in doped RTiO3
A. C. Komarek,1 H. H. Wu,1,2 C. Trabant,1 R. Feyerherm,3 E. Duzik,3 L. H. Tjeng,1 M. Braden,1 C. Schüßler-Langeheine1
1

II. Physikalisches Institut, Universität zu Köln, Zülpicher Str. 77, 50937 Köln, Germany 2 National Synchrotron Radiation Research Center, Hsinchu 30076, Taiwan 3 Helmholtz-Zentrum Berlin, Albert-Einstein-Str. 15, 12489 Berlin, Germany

Funded by the DFG through SFB 608 and by the BMBF through project 05 ES3XBA/5. Inducing a metal insulator transition by doping into a Mott insulator is intensively studied at the moment due to its relevance for the phase-diagram of the cuprate high temperature superconductors [1]. The titanates RTiO3 are particularly interesting since they posses a single electron in the 3d-shell compared to the cuprates with a single hole. But in contrast to the cuprates, the orbital degree of freedom plays an important role in the titanates [2] and they exhibit a 3-dimensional crystal structure. Doping Sr and Ca into RTiO3 allows to render the titanates metallic; concomitantly the magnetic order is suppressed. However, while a small amount of Sr is sufficient to render RTiO3 with larger rare earth ionic radius (like R=La) metallic, R1-xCaxTiO3 samples with small­ er rare earth ionic radius stay insulating up to much higher doping levels of about 40% of Ca and beyond. This surprising difference has remained unexplained so far. There are several phase diagrams known, where doping into a Mott-insulator does not yield metallic behavior even at high amounts of doping. In such systems charges sometimes order, forming checker-

Fig. 1 (a) Temperature dependence of the (011) superstructure reflection from a sample of Y1-xCaxTiO3 with x = 0.36. (b) Energy scans of the (011) superstructure and the (022) fundamental reflections together with the fluorescence signal (x = 0.36). (c) polarization analysis: ω-scans of the (011) reflection for the rotated and unrotated channels (x = 0.36).

253

Fig. 2 Energy scans of the (011) superstructure reflection for (a) Y0.67Ca0.33TiO3 and (b) YTiO3 at different azimuthal angles. board or stripe arrangements, like for example in Pr1-xCaxMnO3 or in La2-xSrxNiO4 [3,4]. In the titanates, however, so far no evidence for charge ordering has been reported. We have recently found evidence for charge ordering in R1-xCaxTiO3 samples with R=Y, Er, Lu in different single crystal neutron diffraction measurements as well as in single crystal Xray diffraction measurements using synchrotron radiation. Charge ordering in these systems leads to a symmetry reduction from an orthorhombic to a monoclinic symmetry: Pbnm → P21/n. In the monoclinic phase, two distinct Ti-sites appear. This symmetry reduction was al­ ready observed in synchrotron radiation powder X-ray diffraction measurements by K. Kato et al. [5] and has been verified in similar synchrotron measurements on our own samples. The structural refinement of various single crystal neutron and X-ray diffraction measurements re­ veals two different oxygen environments of the two distinct Ti-sites with one TiO6 octahedron being enlarged. Using these structural data the nominal valence of the two Ti-sites can be cal­ culated by the bond valence sum (BVS) formalism. The difference of the oxidation state of both Ti-ion amounts to 0.3 electrons. In order to verify this finding spectroscopically, we performed resonant X-ray diffraction measurement at the MAGS beamline, where we studied the signatures of charge ordering for several R1-xCaxTiO3 samples with R=Y, Er, Lu for the hole-doping regime 0.33 < x < 0.5. In Fig. 1(a) the temperature dependence of the (011) charge ordering superstructure reflec­ tion of a Y1-xCaxTiO3 sample with x = 0.36 is shown for two different incident X-ray energies. Below the charge ordering temperature this superstructure reflection intensity increases strong­ ly. In Fig. 1(b) an energy scan of this superstructure reflection is shown together with the fluo­ rescence spectrum and the energy dependence of the fundamental (022) peak. In contrast to the (022) behavior, there is a strong resonant enhancement of (011) at the absorption threshold

254

Fig. 3(a,c) Temperature dependence of the (011) superstructure reflection and (b) tempera­ ture dependence of the peak width (HWHM) of the (011) reflections for different samples. with a peak at the maximum of the derivative of the fluorescence signal. This is a clear signa­ ture of charge ordering. To corroborated this interpretation, we present in Fig. 2(a) a direct comparison of energy scans of the (011) reflections for Y1-xCaxTiO3 (x = 0.33) and YTiO3. In the undoped YTiO3 system this reflection is indicative for orbital ordering since the antiferro-orbital ordering scheme is alternating between neighboring Ti-sites. As can be seen in these figures, the maximum intensities can be found at the peak of the absorption in the undoped system, while in stark contrast the maximum (011) intensity for 33% hole-doped system is found at clearly lower photon energies. Furthermore, the polarization analysis reveals the bulk of the (011) signal from the doped samples is found in the unrotated σ-σ channel and not in the σ-π channel where orbital ordering contributions would be expected [Fig. 1(c)]. All these results support the charge ordering picture in the R1-xCaxTiO3 system with R=Y, Er, Lu. In Fig. 3(a) the temperature dependence of the superstructure reflection intensity is com­ pared for different samples with Ca-doping levels 0.33, 0.36, 0.37 and 0.40. Upon cooling the (011) charge ordering superstructure reflection intensity starts to increase at higher tempera­ tures for samples with larger hole-doping level. Hence, the charge ordering seems to be stabi­ lized for samples closer to the optimum/half-doped composition. In Fig. 3(b) the peak widths are plotted as a function of temperature indicating a loss of spatial coherence at the lower tran­ sition temperature. Finally, in Fig. 3(c) also the temperature dependence of the (011) super­ structure reflection intensities of the higher doped samples are shown. These exhibit a much smoother transition and show signatures of charge order already at room-temperature. [1] M. Imada et al., Rev. Mod. Phys. 70, 1039 (1998). [2] Y. Tokura et al., PRB 48, 14063 (1993). [3] Y. Tokura et al., Science 288, 462 (2000). [4] J. M. Tranquada et al., Nature 375, 561 (1995). [5] K. Kato et al., JPSJ 71, 2082(2002).

255

Temperature dependent quasiparticle renormalization in nickel metal
R. Ovsyannikov, J. S´nchez-Barriga, J. Fink, A. Varykhalov, H. A. D¨rr a u Helmholtz Zentrum Berlin, BESSY II, Albert-Einstein-Strasse 15, D-12489 Berlin, Germany
Spintronics, also known as magnetoelectronics, is an emerging technology involving detection and manipulation of electron spin for data storage and processing. In order to perform such an operation two primary types of devices are required – a device that could generate a spin polarized current, the so called spin-injector and a device that is sensitive to spin polarization, i.e., a spin detector. One of the known spintronics systems are the giant magnetoresistance (GMR) devices[1]. A typical GMR devices consist of two ferromagnetic layers separated by a metallic nonmagnetic spacer layer. When the magnetization direction of two ferromagnetic layers is aligned the resistance will be smaller than that for antialigned layers. Tunnel magnetoresistance (TMR) devices[2] use a similar idea. In this case the spacer layer is made out of insulating material and the magnetization vector alignment/antialignment will change the tunneling resistance. Such devices find common use as read heads of modern hard drives. Ultimate performance of such devices is determined by spin-dependent lifetimes of spin-polarized charge carriers, thus focusing attention on the electronic structure and the dynamic properties of strongly correlated systems like 3d-ferromagnets, transition metal oxides and 4f- rare-earths systems. One of the fundamental consequences of electron correlation effects is that the bare particles in solids become ’dressed’ with the excitation cloud resulting in quasiparticles. Such a quasiparticle will carry the same spin and charge as the original particle, but will have a renormalized mass and a finite lifetime. The properties of many-body interactions are described with a complex function, the called self energy Σ = Re Σ+Im Σ. The real part of the self energy Re Σ depicts the change of the dispersion or the increase of the effective mass, while the imaginary part of the self energy Im Σ contains the information about the lifetime of the quasiparticle. In case of electron scattering processes due to electron-electron, electronphonon or electron-impurity interactions their contribution to the lifetime broadening, Γ, is independent of each other and can be expressed as a sum of each contribution: Γ = Γel−el +Γel−ph +Γel−imp [4]. Γel−imp represents an energy-independent term arising from the electron-impurity scattering, while Γel−el and Γel−ph , which are contributions from electron-electron and electron-phonon scattering respectively, are energy dependent. For 3d -metals Γel−el is well described at low energies by a Fermi liquid behavior and depends quadratically on the binding energy. In case of electron-phonon interaction the energy dependence of lifetime broadening is more complicated and will lead to the appearance of a kink – a sudden change in the dispersion below a finite binding energy. We studied the spin dependent quasiparticle band structure of Ni (111) with high resolution angleresolved photoemission spectroscopy (ARPES) to obtain detailed insight into self energy. The experiment was performed at the UE-112 beamline in the BESSY-II storage ring using ”13 ” spectrometer. We grew a thick (≈100˚) Ni(111) layer on a W(110) surface. ARPES spectra were measured at a photon energy of A 136 eV. The accessible plane in the Ni bulk Brillouin zone is shown in Fig. 1. The measured Fermi surface is shown in Fig. 2. The solid yellow line indicates the line along which ARPES data were taken. Dashed lines indicates positions of d and sp-like bands. In order to quantitatively analyze the spectral line shape we used Lorentzians on a linear background to fit the momentum distribution curves (MDC’s) – the intensity distribution curves as a function of momentum for a given binding energy. Data measured at low temperature (≈50 K) for sp-like and d bands are presented in Fig. 3. The plots (c) and (d) of Fig. 3 show the intensity distribution of a band while the (a) and (b) plots show the MDC data (symbols) and fits (solid lines) at Fermi level. In the present analysis we used the MDC widths δk to estimate the imaginary part of self energy Im Σ = vf δk, where vf is the Fermi velocity [4]. The band dispersion in the absence of the kink was assumed to be linear and the Fermi velocity was estimated as

256

Figure 1: Sketch of the Ni Brillouin zone

Figure 2: Experimentally obtained Fermi surface of Ni with the assignment of minority d (d ↓ ), majority sp-like (sp ↑ ) and minority sp-like (sp ↓ ) bands.

Figure 3: ARPES images measured at low temperature (≈ 50K) of the Ni (111) surface for sp-like bands (c) and d band (d). Corresponding MDC data (symbols) and fits (lines) at Fermi are shown in (a) and (b), respectively a slope of the linear fit of the first ≈ 50 − 70 meV of the fitted dispersion. The resulting imaginary and real parts of the self energy of the sp-like minority, sp-like majority and d bands are shown in Fig. 4(a) and Fig. 4(b), respectively. Symbols represent the observed imaginary and real parts of the self energy. Solid and dashed lines display modeled electron-phonon and electron-electron lifetimes, respectively. For binding energies E 35meV the lifetimes of minority sp-like and d bands decrease while the majority sp-band does not show any significant change. Such behavior is in a good agreement with literature [3] and has been explained by electron-phonon interaction. The electron-phonon coupling to a phonon of energy Ω can be expressed as:
Ωmax el−ph

Im Σ

= 2π
0

α2 Fk (E ) [2n (E ) + f (E + E) + f (E − E)] ,

where n(E) is Bose-Einstein distribution, f (E) is Fermi-Dirac distribution and α2 Fk is the Eliashberg function[4]. In the present analysis we used a Debye model to describe the phonon Eliashberg function:
E α2 Fk = λ ΩD for E ≤ ΩD and α2 Fk = 0 for E > ΩD [4], where λ is electron-phonon coupling constant. We obtained λsp↓ ≈ 0.21 for minority sp-like states and λd↓ ≈ 0.44 for d states. In case of majority sp-like states the coupling constant is much smaller λsp↓ ≈ 0. This difference in the λ-values is especially prominent for the dispersion close to the Fermi level in Figs. 3(c) and 3(d). Measured room temperature data (≈320 K) for sp-like and d bands are presented in Fig. 5. At high temperatures kink disappear in line broadening but we still could compare MDCs at the Fermi

257

Figure 4: Experimental imaginary (a) and real (b) parts of the self energy of sp-like minority, sp-like majority and d bands. Symbols represent the observed imaginary (real) parts of the self energy. Solid and dashed lines displays modeled electron-phonon and electron-electron lifetimes respectively.

Figure 5: ARPES images measured at room temperature (≈ 320K) of the Ni (111) surface for sp-like bands (c) and d band (d). Corresponding MDC data (symbols) and fits (lines) at Fermi are shown in (a) and (b), respectively level. With increasing temperature we observe a decreasing quasiparticle lifetime at the Fermi level for all probed minority spin bands as expected for electron-phonon coupling. The sp-like minority spin band displays an increase of lifetime broadening from 0.014˚−1 to 0.029˚−1 , while d minority band lifetime A A broadening increases from 0.038˚−1 to 0.096˚−1 . Surprisingly the majority spin states behave differently. A A We actually observe a slightly increased lifetime at room temperature. The width of the majority band decreases with increasing temperature from 0.029˚−1 to 0.023˚−1 corresponding to an actual increase A A of the elastic mean free path from 33˚ to 43.5˚. The corresponding increase in Fermi velocity from A A 1.73˚/fs to 2.06˚/fs points to a temperature dependent reduction of the majority spin quasiparticle A A renormalization. Although the origin of this novel effect is still unclear it could possibly arise from spin excitations coupling to phonons. [1] P. Gr¨nberg et al., Phys. Rev. Lett. 57, 2442–2445 (1986) u [2] M. Julliere, Phys. Lett. 54A, 225226 (1975) [3] M. Higashiguchi et al., Phys. Rev. B 72, 214438 (2005) [4] S. H¨fner, Very High Resolution Photoelectron Spectroscopy,1st ed. (Springer-Verlag Berlin Heidelu berg 2007)

258

Static and fluctuating stripe order in 1/8-doped LNSCO and LSCO
H. H. Wu,1,2 M. Buchholz,1 C. Trabant,1 F. Heigl,3 E. Schierle,4 M. Cwik,1 M. Braden,1 L. H. Tjeng1 and C. Schüßler-Langeheine1
1

II. Physikalisches Institut, Universität zu Köln, Zülpicher Str. 77, 50937 Köln, Germany 2 National Synchrotron Radiation Research Center, Hsinchu, Taiwan 3 CELLS-ALBA, 08193 Bellaterra, Barcelona, Spain 4 Helmholtz-Zentrum Berlin, Albert-Einstein-Str. 15, 12489 Berlin, Germany

Funded by the DFG through SFB 608 and by the BMBF through project 05 ES3XBA/5 In hole-doped layered copper-oxides static and fluctuating stripe order has been considered to be a crucial issue for the understanding of the basic physics of the high-Tc superconductor. Stripe order is a complex collective order involving charge and spin degrees of freedom. While in La2-xSrxCuO4 (LSCO) fluctuating stripe order gives rise to peaks in inelastic neutron scattering, partial substitution of La by Nd leads to a pinning of spin and charge order and the formation of static stripes as revealed by peaks in the elastic neutron diffraction signal [1]. Earlier experiments from the isostructural nickelate system gave us strong indications that resonant soft x-ray diffraction RSXD is not only sensitive to static but also to fluctuating stripe order. This finding suggests that RSXD may be an attractive alternative method to study in particular fluctuating charge order, which is difficult to probe by neutron scattering. For this experiment we chose samples with doping level near 1/8, namely La1.48Nd0.4Sr0.12CuO4 (LNSCO) and La1.88Sr0.12CuO4 (LSCO) single crystals. The samples were cut and polished ex situ. The experiments were performed at the beam line UE46-PGM using the UHV diffractometer built at the Freie Universität Berlin. As seen in Figs. 1, 2 and 3, both systems exhibit a pronounced incommensurate charge order peak at the oxygen K and copper L2,3 edges, thus confirming our basic assumption that fluctuating stripe order can in fact be probed by RSXD. In fact for the LSCO sample this is the first direct observation of charge order. Earlier neutron diffraction work found magnetic order, but no direct evidence for charge order [3].

FIG. 1. (a) Photon energy dependence of charge order at oxygen K edge in LNSCO and LSCO. (b) X-ray absorption spectra (XAS) at oxygen K edge (red line) and the resonant x ray scattering intensity of charge order near O-K edge in LSCO (black line).

259

In Figure 1(a) and 1(b), the charge order in LNSCO and LSCO shows the strong photon energy dependence near oxygen K edge. The resonance energy for the charge order in both systems is located at photon energy of 529 eV, which is mainly from a 3d9L to 1s3d9 transition. The resonance behavior at oxygen K edge has similar feature for LNSCO and LSCO supporting a common electronic origin of both diffraction features. As can be seen in Figure 2(c), and 2(d), the two systems have slight difference in the RSXD spectra near Cu L2,3 edges, which require further analysis. We studied the temperature dependence of both signals (Fig. 3). The Nd-stabilized system (LNSCO) shows a structural phase transition from low-temperature-orthorhombic (LTO) to low-temperature-tetragonal (LTT) at about 70 K [1,2] while the Nd-free La1.88Sr0.12CuO4 always remains in the LTO phase. As for LNSCO, static charge order is known to exist in the LTT phase [4], and in fact we find no

FIG. 2 (c) and (d) Photon energy dependence of charge order near Cu L2,3 edge (blue line) and XAS at Cu L2,3 edge in LNSCO and LSCO.

superstructure peak above the LTT-LTO phase transition temperature. In LSCO, the intensity of charge order vanishes slightly above the critical temperature similar to what has been observed for the magnetic order [5].

[1] J. M. Tranquada et al., Phys. Rev. Lett. 78, 338 (1997). [2] J. M. Tranquada et al., Phys. Rev. B 54, 7489 (1996). [3] K. Yamada et al., Phys. Rev. B 57, 6165 (1998). [4] J. M. Tranquada et al., Nature 375, 561 (1995). [5] H. Kirmura et al., Phys. Rev. B 59, 6517 (1999). FIG. 3: Temperature dependence of charge order at O-K and Cu L3 edges in LNSCO and LSCO.

260

X-ray photoelectron spectroscopic analysis of HfO2/Si structure synthesized by ALD and MOCVD methods E O Filatova1, A A Sokolov1, V V Afanas’ev2, E Yu Taracheva1, M M Brzhezinskaya3 and A A Ovchinnikov1
2

Institute of Physics, St-Petersburg State University, St-Petersburg, 198504, Russia Department of Physics, University of Leuven, Celestijnenlaan 200D, B-3001 Leuven, Belgium 3 BESSY, Albert Einstein Str. 15, D-12489 Berlin, Germany

1

HfO2 is one of the most promising contenders to become the insulating material for the future MOS technology nodes, it combines high dielectric constant (κ=16-45), wide bandgap, and sufficient thermodynamic stability with respect to interaction with silicon to withstand the thermal budget of the MOS device processing [1,2]. Successful implementation of HfO2-based insulating stacks requires basic understanding of the HfO2/Si interface chemistry and its impact on the electrical properties [3,4]. In particular, the properties of the interlayer (IL) between Si and HfO2 are of importance as they have direct effect on the electron transport in the Si surface channel. The main goal of this project consisted in layer-by-layer X-ray photoelectron spectroscopic (XPS) investigation (Hf 4f, Si 2p, and O 1s electron states) of thin layers of HfO2 grown on (100)Si crystal surface using atomic layer deposition (ALD) or metallo-organic chemical vapour deposition (MOCVD) with purpose to receive the information about chemical phase composition of the IL depending on preparation method. In framework of present work were investigated next samples: 1. The HfO2 (ALD, 5 nm)/n-Si and HfO2 (ALD, 5 nm)/p-Si structures were fabricated at 300οC using ALD from HfCl4 and H2O precursors on the IMEC-cleaned, i.e., covered with 0.8 nm-thick chemical Si oxide, surface of n- and p-type (100)Si crystals, respectively. 2. The HfO2 (MOCVD, 5 nm)/n-Si sample was prepared on the identically prepared n-type (100)Si surface using MOCVD of HfO2 at 485οC from (Hf((CH3)2N)4) and O2 precursors. 3. The HfO2 (ALD, 1.5nm)/p-Si sample was synthesized by ALD at 280οC from (Hf((CH3)2N)4) and H2O precursors on Si-p-type (100) substrate etched prior to deposition in 1% aqueous HF solution for 30 s without final water rinse. To obtain the depth profiles, Figure 1. Hf 4f photoelectron spectra from the HfO2/Si the Ar+ ion sputtering was applied samples before and after different Ar+ ion sputtering steps: a) using 2.5 keV in situ ion beam with HfO2(ALD, 5 nm)/n-Si; b) HfO2(MOCVD, 5 nm)/n-Si; c) the size of the sputtered area being 15 HfO2(ALD, 5 nm)/p-Si; d) HfO2(ALD, 1.5 nm)/p-Si. mm x 15 mm. All measurements were performed at the Russian-German beamline. The XPS spectra were taken at the exciting photon energies of 200 eV and 700 eV using hemispherical electron energy analyzer VG Clam 4 with the energy resolution better than 200 meV. All spectra were collected at the analyser pass energy of 10 eV and the binding energies (BE) were referenced to the bulk Si0 2p3/2 core level at 99.2 eV [5]. Fig.1 shows the Hf 4f XPS spectra obtained at the excitation energy of 200 eV from the pristine HfO2/Si samples and those observed after different sputtering steps. Before sputtering, analysis of the Hf 4f spectra reveals the presence of the Hf 4f doublet peak with spin-orbit splitting of l.7 eV and intensity ratio between components of 0.8. The BE position of these peaks is varied from

261

18.6 eV (4f7/2) (5-nm thick films) to 17.6 eV (4f7/2) (1.5-nm thick layers). The most plausible explanation of these differences is HfO2 thickness-dependent irradiation-induced oxide charging. During sputtering the Hf 4f doublet peak becomes weaker and ultimately disappears. At the same time the new peaks centred at 14.7 eV and 16.4 eV become detectable in the spectra after 40 min sputtering. As these peaks are associated with Hf0 4f doublet in metallic hafnium [6, 7], this observation suggests that oxygen is preferentially removed by Ar+ ion sputtering and while the remaining hafnium is reduced to Hf0. Fig.2 shows the Si 2p XPS spectra obtained at the excitation energy of 200 eV. Fig. 2a),b),c) shows the spectra obtained from the HfO2(5nm)/Si samples for sputtering steps after 40 minutes of the sputtering. Analysis of the Hf 4f and Si 2p XPS spectra of HfO2(ALD, 5 nm)/Si samples after the 70 minutes of ion sputtering reveals that only Hf0 4f and Si0 2p lines remain in the spectra. Further sputtering is found Figure 2. Si 2p XPS spectra from the HfO2/Si samples after to have no measurable effect on the ion sputtering for different time (given in min): a) HfO2(ALD, shape and the intensity of these 5 nm)/n-Si; b) HfO2(MOCVD, 5 nm)/n-Si; c) HfO2(ALD, 5 nm)/p-Si; d) HfO2(ALD, 1.5 nm)/p-Si. peaks. 0 The intensities of the peaks Hf 4f and Si0 (2p) were compared using normalization on the photoionization cross-section yielding the ratio close to 5:1. Apparently, a thin layer of metallic Hf, was formed on the Si-surface during the Ar+ ion bombardment impeding further sample sputtering. Interestingly, as can be seen from figure 2, in the sample with MOCVD HfO2 the Si0 (2p) peak appears after longer sputtering time as compared to the samples with ALD HfO2 suggesting difference in the Ar+ sputtering rate. All peaks in the Si 2p XPS spectra were deposited into peaks which corresponds to different oxidation states Six+ (x=0,1,2,3,4). Figure 3 shows the fit of the Si 2p XPS spectra of the HfO2(ALD, 5 nm)/n-Si samples after 45 and 55 minutes of the sputtering. In its turn, figure 4 presents curve fitting of the Si 2p emission band in the HfO2(ALD, 1.5nm)/p-Si sample Figure 3. Curve fitting of the Si 2p XPS spectra taken after 45 before the sputtering, and after 17 and 55 minutes of ion sputtering of the HfO2 (ALD, 5 nm)/nand 32 minutes of the sputtering. Si sample. Such curve fitting was carried out for all etching steps. Analysis of all curve obtained indicates the presence of small amount of Hf-silicate and SiO2 at the interface for HfO2 (5 nm)/Si samples and substantial quantity of SiO2 and SiOx for HfO2 (ALD, 1.5 nm)/p-Si sample.

262

Figure 4. Curve fitting of the Si 2p XPS spectra from the HfO2 (ALD, 1.5 nm)/p-Si sample before the sputtering (a) and after sputtering for 17 ( b) and 32 (c) minutes. Fig.5 shows O1s XPS spectra obtained at the excitation energy of 700 eV. There is observed a shift of the O 1s peak towards lower BEs as illustrated in figure 3. Importantly, variation of the samples surface charging would result in a correlated shift of the both Hf 4f and O 1s peaks energy positions. Analysis of Hf 4f, Si 2p and O 1s photoelectron spectra in the HfO2/Si samples prepared using Figure 5. O1s XPS spectra of the HfO2(ALD, 5 nm)/n-Si (a) different oxide growth techniques and HfO2(MOCVD, 5 nm)/n-Si (b) samples before and after indicates the presence of small different Ar+ ion sputtering steps. amount of Hf-silicate and SiO2 at the interface between (100)Si crystal and HfO2 grown on the 0.8-nm thick native oxide under-layer. On the contrary, the presence of considerable amounts of SiO2 and SiOx is revealed at the interface formed by depositing hafnia on HF-dipped, i.e., hydrogen-passivated, silicon. Further, it is found that Ar+ ion sputtering leads to the formation of a metallic Hf on the Si-substrate with no signs of silicate phase. The sputtering induced formation of metallic Hf may be important not only from the point of view of optimizing dry etching of the HfO2 films, but, at the same time, may also be used to incorporate metallic particles or conducting layers to the Si/HfO2–based heterostructures. Acknowledgments This work was supported by the ISTC (Project № 3401) and bilateral Program “Russian-German Laboratory at BESSY”. References [1] Kukli K, Aarik J, Aidla A, Siimon H, Ritala M and Leskela M 1997 Appl. Surf. Sci. 112 236 [2] Hsu C T, Su Y K and Yokoyama M 1992 Jpn. J. Appl. Phys., Part 1 31 2501 [3] Sayan S, Garfunkel E, Nishimura T, Schulte W H, Gustafsson T and Wilk G D 2003 J. Appl. Phys. 94 928 [4] Misra V, Lucovsky G and Parsons G 2002 MRS Bull. 27 212 [5] Cardona M and Ley L 1978 Eds., Photoemission in Solids I: General Principles (Springer-Verlag, Berlin) with additional corrections [6] Tan R et al 2005, Applied Surface Science 241 135–140 [7] Suzer S et al 2003., J. Vac. Sci. Technol. A21 106

263

X-ray reflection spectroscopic characterization of HfO2/Si structure synthesized by ALD and MOCVD methods
E O Filatova1, A A Sokolov1, I V Kozhevnikov2 E Yu Taracheva1, F Schaefers3and W Braun3
1 2

St-Petersburg State University, Institute of Physics, St-Petersburg, 198504, Russia Institute of Crystallography, Moscow, 119333, Russia 3 BESSY, Albert Einstein Str. 15, D-12489 Berlin, Germany HfO2 is one of the most perspective materials for the nanoelectronics industry to replace SiO2 because it has a high dielectric constant and is expected to be stable in contact with Si. So far one stable monoclinic phase and four metastable phases: cubic, tetragonal, orthorhombic I and orthorhombic II have been identified for HfO2. An amorphous modification can also be fabricated. Although the tetragonal HfO2 structure has the highest dielectrical permittivity an amorphous structure is to be preferred for several reasons. First of all amorphous materials do not contain grain boundaries or dislocations that can trap charge and offer fast diffusion pathways for leakage current. In addition, stresses in amorphous materials can be taken up by small variations in the random network, where they may likewise be taken up by misfit dislocations in a polycrystalline material. Amorphous structures also tend to minimize electronically active defects, but may give rise to shallow traps. Epitaxial films can be produced free of grain boundaries and defects and, therefore, are usually preferred for gate dielectrics. Therefore, most of the work on HfO2 has been focused on manufacturing amorphous films to replace SiO2, but this process is unpredictable, because the microstructure of films is largely depending on thickness, technology of synthesis and utilizable precursors. Therefore the technology of preparing films, i.e. method and condition of synthesis (rate of deposition, temperature of substrate, etc.), also as utilizable precursors are an important issue to fabricate films with the required properties. A further study of the effect of different factors is required. In this work we were focused on the discussion of the crystalline and electronic structure of thin films of HfO2 of different thicknesses and prepared by two different methods (Atomic Layer Deposition (ALD) and Metal Organic Chemical Vapour Deposition (MOCVD)). The HfO2 films were deposited on (100) n-type silicon wafers by ALD, using hafnium tetrachloride (HfCl4) and water at 300 °C substrate temperature. Different film thicknesses (5, 20 and 100 nm) were obtained by varying the number of ALD cycles only. When the MOCVD method was used, the HfO2 films were deposited on (100) n-type silicon wafers using tetrakisdiethylaminohafnium (Hf(N(C2H5)2)4) and O2 at 485 °C substrate temperature. The angular and spectral dependences of the reflectivity in the vicinity of O K(1s) - absorption edge were measured using s-polarized synchrotron radiation in the reflectometer set-up on the optics beamline (D-08-1B2). The O1s reflection spectra of HfO2 films of thicknesses 5, 20 and 100 nm synthesized by ALD and measured at grazing incidence angle of 4o are presented in figure 1. The absorption spectra calculated on the basis of the measured reflection spectra by means of the Kramers–Kronig transform using the method described in details in [1] are presented in figure 2. The molecular orbitals of HfO2 derived from a linear combination of atomic orbitals (LCAO) are characterized by four unoccupied orbitals: eg(Hf5d+O2pπ), t2g(Hf5d+O2pσ), a1g(Hf6s+O2p) and t1u(Hf6p+O2p). Peaks a and b (fig.2) reflect core –electron transitions in the oxygen atoms to the lowest unoccupied Hfdeg and Hfdt2g electronic states that are hybridized with the 2p states of the oxygen atoms. Obviously, there is a broadening of the feature a with the growth of the film thickness. Moreover the feature a demonstrates a double structure a –a´ in the absorption spectrum of the film of 100 nm thickness. This splitting appears because of Jahn-Teller d-state degeneracy and specifies on presence of monoclinic phase in the film. At that the grains size must be ~/>2 nm. Otherwise it takes place suppression of the Jahn-Teller effect. The manifestation of details c and d is also signature of ordering of structure. The probability of the transitions to the (Hf 6s,6p+O 2p) states is very low and the double structure c-d cannot be observed in the amorphous state of the film. Above mentioned arguments have allowed concluding that the 100 nm thick film has the polycrystalline structure with a size of the grains ~/>2 nm and its microstructure involves generally the monoclinic phase. At that films of thickness 5 and 20 nm have amorphous structure. Figure 4 shows the O1s-reflection spectra of films of equal thickness of 5 nm, but fabricated by two different methods (ALD and MOCVD). The O1s - absorption spectra calculated from reflection spectra (figure 4) using a Kramers – Kronig analysis are presented in figure 5.

264

Figure 1. O1s reflection spectra of HfO2 5, 20 Figure 2. Absorption spectra calculated on the and 100 nm thick films synthesized by ALD and bases of the measured reflection spectra measured at a grazing incidence angle of 4o. (figure 1) using a Kramers - Kronig analysis. Analysis of the shape of the reflection and absorption spectra (figure 4, 5) allows concluding, that the film manufactured by ALD is amorphous, whereas the film prepared by MOCVD method has signatures of ordering of the structure because of the pronounced details c and d. In contrast to the 100 nm thick film manufactured by ALD (figures 1 and 2) there is no sign of a Jahn-Teller splitting in the spectrum of the film prepared by MOCVD that allows assuming that grain size in the film under consideration is < 2 nm. It can be argued that the film prepared by MOCVD shows evidence for crystallization.

Figure 4. O1s- reflection spectra of films 5 nm Figure 5. O1s- absorption spectra of 5 nm thick fabricated by ALD and MOCVD measured thick films calculated from reflection spectra at a grazing incidence angle 4o. (figure 4). In addition to reflection spectra the reflectivity curves versus grazing angle were measured at different wavelengths of the incident beam for HfO2 films under consideration. The reflectivity curves versus grazing angle measured at energy 520 eV are shown in figure 6, circles, respectively. The depth-distribution of chemical elements obtained within the framework of the approach developed in [2] is shown in figure 7. The corresponding reflectivity curves are shown in figure 6, curves 3, for the ALD sample, which are in a perfect agreement with the experimental data. The calculated reflectivities of the MOCVD samples are not shown in figure 6, because they cannot be distinguished from curves 2, by applying the fitting procedure. Such situation can takes place when the film is structurally inhomogeneous.

265

Figure 6 The measured reflectivity curves (circles) versus grazing angle at energy 520 eV of the 5 nm thick HfO2 film deposited onto a Si substrate by the ALD (left) and MOCVD (right) technique. Curves 1 and 2 are the result of fitting with the use of the single-layer (1) or three-layer (2) models. Curves 3 are the result of a numerical refinement of the three-layer model.

Figure 7 Depth-distribution of chemical elements for the 5 nm thick films deposited onto a Si substrate by the ALD (left) and MOCVD (right) technique: three-layer model. The microstructure of films of different thickness synthesized by ALD and MOCVD was characterized by reflection spectroscopy and soft x-ray reflectometry techniques. It was shown that near – edge reflection spectra and absorption spectra calculated on the basis of reflection spectra are very sensitive even to signatures of ordering on a subnanometer scale in the film. It was established that the microstructure of the HfO2 film is strongly depends on the film thickness. First results on the reconstruction of the depth-distribution of chemical elements based on the analysis of reflectivity curves were discussed. The element depth profiling shows that there are no measurable differences between 5 nm ALD and MOCVD films would provide a direct support to the importance of structural factor. The calculations carried out allow assuming that MOCVD film is inhomogeneous in depth. Acknowledgments This work was supported by the ISTC (Project № 3401). E.Yu.Taracheva gratefully acknowledges financial support from BESSY. Authors thank Dr. Stefan DeGendt (IMEC, Belgium) for providing with ALD and MOCVD HfO2 layers, Dr. Fred Senf from Optics Beamline, Centre of Synchrotron Radiation BESSY for the help in experiment carrying out. References [1] Filatova E, Lukyanov V, Barchewitz R, Andr´e J-M, Idirk M and Stemmler Ph 1999 J. Phys.: Condens. Matter 11 3355 [2] Kozhevnikov I V 2003 Nucl. Instrum. Methods A, 508 519-541.

266

Investigations of a structure of thin TiO2 films by soft x-ray reflectometry techniques
E O Filatova1, E Yu Taracheva1, G S Shevchenko1, A A Sokolov1, I V Kozhevnikov2, S Yulin4, F Schaefers3, W Braun3
1 2

Institute of Physics, St-Petersburg State University, St-Petersburg, 198504, Russia Institute of Crystallography, Moscow, 119333, Russia 3 BESSY, Albert Einstein Str. 15, D-12489 Berlin, Germany 4 Fraunhofer Institut Angewandte Optik und Feinmechanik, Albert Einstein Strasse 7,D-07745 Jena

TiO2 films are extensively studied because of their interesting chemical, electrical and optical properties. TiO2 is a high bandgap semiconductor that is transparent to visible light and has excellent optical transmittance. TiO2 has high refractive index, good insulating and catalytic properties as well as it possesses unique wettability and biological compatibility. As a result TiO2 thin films are widely used in many practical applications (catalysis, photocatalysis, dye-sensitized photovoltaic cell, gas sensors, and etc.) Besides TiO2 nanofilms are widely used as protective layer for very large scale integrated (VLSI) circuits, for manufacture of optical elements, as antireflective (AR) coatings, for electrochromic displays and planar waveguides. The high dielectric constant of TiO2 allows its consideration as an alternative to silicon dioxide for ultrathin gate oxide dielectrics used in memory and logic devices. The mechanical, physical, chemical and electrical properties of TiO2 films in many respects depend strongly on their crystallinity. As it is known, TiO2 films can be synthesized in different crystal modifications (anatase, rutile and brookite) or in amorphous phase. Rutile TiO2 has higher opacity, greater density and greater inertness than the anatase one and it is used as a convertible pigment in paints and dyes industry. Besides because of its highest refractive index TiO2 rutile type is required for photonic crystals application. While the anatase type of TiO2 can enhance the photocatalytic and hydrophilic property of TiO2 due to its higher photoactivity and more crystal defects than other types of TiO2. Fabrication of films with the necessary microstructure depends on different deposition parameters and is the important technological problem. To adjust a technological process it is necessary to use precise technique for thin films characterization. Traditionally XRD technique is applied to identify crystallinity of films. At the same time the sensitivity of this method is limited by the grain size. The TiO2 films 10nm and 70 nm thick synthesized by magnetron sputtering method on Si(100) wafers were investigated using X-ray reflection spectroscopy. The angular and spectral dependences of the reflectivity in the vicinity of O K(1s) - absorption edge and Ti L2,3 (2p) - absorption edge were measured using s-polarized synchrotron radiation in the reflectometer set-up on the optics beamline. Reflection spectra of TiO2 films 10 and 70 nm thick measured in the vicinity of Ti L2,3 (2p) – and O K(1s)- absorption edges at grazing incident angles 4o and 2o, correspondingly, are presented in Fig.1. One can see that both the Ti 2p and the O (1s) spectra for different films thickness correlate well in energy position of main details of the structure and in absolute value of the reflectivity. At the same time, the reflection spectra of films are characterized by different contrast, especially in the case of OK-spectra. Moreover the reflection spectrum for film 70nm thick shows additional details in the structure. The absorption spectra calculated on the bases of the measured reflection spectra with help of Kramers - Kronig relationship described in details in [1] are presented in the Fig.2. The peaks a and b (fig.2a) in Ti 2p3/2 spectrum stem from dipole allowed transitions of Ti 2p3/2 electrons to unoccupied 3d states. These states are split into 3dt2g (peak a) and 3deg (peak b) components by octahedral crystal field created by the O ions. What is the most interesting that there is a further splitting of the peak b into two bands b1-b in the absorption spectrum of thick film (70nm). The manifestation of doubling of the Ti 2p-3deg- transitions depends on the crystalline structure of TiO2. The main difference in Ti 2p spectra of different polymorphs is a change in intensities of components of peaks b and b1. It is important for our investigations that the intensities of components of peaks b1 and b in the spectrum of thick film (70nm) correlate well with anatase structure. The essential difference between rutile and anatase lies in the secondary coordination. It is shown in [2] that the experimental XAS spectra of anatase and rutile cannot be solely viewed in terms of a localized atomic cluster. According to [2] clearly the effect of the p-d and d-d interactions is important but this must be modified by the crystal field and further longer-range effects.

267

  Figure 1. Reflection spectra of TiO2 films 10 and 70 nm thick measured in the vicinity of (a) Ti L2,3 (2p) - and (b) O K(1s)- absorption edges

Figure 2. Absorption spectra of TiO2 films 10 and 70 nm thick in the vicinity of (a) Ti L2,3 (2p) - and (b) O K(1s)- absorption edges calculated from reflection spectra (Fig.1). The Ti 2p1/2 structures are marked by asterisks. The OK-edge features (fig.2b) at higher energies are very different for investigated films, too. The structure in the spectrum of TiO2 film 70 nm thick is characterized by d, c and f peaks and is related with transitions into the empty electronics states with mixed Ti 4s, 4p+O2p character. Analysis of the energy positions of c, d and f peaks in the spectrum of TiO2 film 70 nm thick shows a good correlation of this structure with the structure of anatase. Analysis of the energy positions and intensity of main peaks in the spectra of investigated TiO2 films shows that the TiO2 film 70nm thick is characterized by anatase structure and the TiO2 film 10nm thick is amorphous. The TiO2 film 70nm thick was investigated in the vicinity of OK-absorption edge at different grazing incident angles, which means at different depth formation of the reflected beam. The measured OK-reflection spectra are plotted in the fig. 3. One can see that there is no change in the shape of reflection spectra with growth of the depth formation of the reflected beam that points on the homogeneous of the investigated film. In addition to reflection spectra the reflectivity versus grazing angle was measured at Figure 3. Reflection spectra of TiO2 film 70 nm λ = 0.154 nm for TiO2 film 70nm thick. thick measured in the vicinity of O K(1s)Measured reflectivity versus the grazing angle of absorption edge at different grazing incident a probe beam is shown in Fig.4, black curve. angles.

268

Approach developed in [3] was used to reconstruct the depth-distribution of the dielectric constant basing on the experimental reflectivity curve. The found dielectric constant profile δ(z) = 1 – Re[ε(z)] is shown in Fig.5, where the Z axis is directed into the depth of a sample. The film density is constant with depth and is equal to 3.6 g/cm3. An accuracy of fitting process is demonstrated by red curve in Fig.5. So that one can conclude that TiO2 film 70nm thick is homogeneous with depth.

Figure 4. Reflectivitie versus grazing angle, measured at λ = 0.154 nm for TiO2 film 70nm thick. Measured reflectivity is red curve. Accuracy of fitting process is blue curve.

Figure 5 Dielectric constant profile δ(z) = 1 – Re[ε(z)] reconstructed from measured reflectivity (fig.4).

The x-ray reflection spectroscopy data indicate that TiO2 film 70nm thick is characterized by anatase structure. Following to dielectric constant profile reconstructed from measured reflectivity and angular dependence of the OK-reflection spectra one can conclude that TiO2 film 70nm thick is homogeneous. The TiO2 film 10 nm thick is amorphous. The investigation carried out shows the high sensitivity of the reflection spectroscopy to the middle atomic ordering. Acknowledgments This work was supported by the ISTC (Project № 3401). E.Yu.Taracheva gratefully acknowledges financial support from BESSY. Authors thank Dr. Fred Senf from Optics Beamline, Centre of Synchrotron Radiation BESSY for the help in experiment carrying out. References [1] E. Filatova, V. Lukyanov, R. Barchewitz, J.-M. Andre, M. Idir, Ph.Stemmler. J. Phys.: Condens. Matter 11 (1999) 3355 [2] R Brydson, H Sauer, W Engel, J M Thomas, E Zeitler, N Kosugilland H Kurodall. J. Phys.: Condens. Matter l(1989) 797-812. [3] I.V.Kozhevnikov, "Physical analysis of the inverse problem of X-ray reflectometry", Nucl. Instrum. Methods A, 508 (2003) 519-541.

269

Chemical and magnetic spectromicroscopy of individual nanoparticles

Florian Kronast1, Nina Friedenberger2, Katharina Ollefs2, Michael Farle2, Hermann A. Dürr1 1 Helmholtz Berlin, Albert-Einstein-Str. 15, 12489 Berlin 2 Universität Duisburg-Essen, Lotharstr.1, 47048 Duisburg
Overcoming the superparamagnetic limit represents a mayor challenge in high-density magnetic data storage technology. The bottom-up self-assembly technique for colloidal monodisperse hardmagnetic nanoparticles may provide a new cost effective approach. For instance the magnetocrystalline anisotropy of FePt alloys and nanoparticles can be tuned from zero to about 10^7 J/m^3 depending on the chemical composition and order [1,2]. In self-assembled nanoparticle arrays the “blocking temperature”, i.e. the critical temperature below which the magnetization is considered to be stable over the timescale of the measurement [3], depends not only on the magnetic anisotropy of the individual particle but also on the dipolar interaction which in turn depends on the magnetization of the particle and the distance between them. Up to now there have been no experimental results that clearly identify why some results can be interpreted in a way that the dipolar interactions between the nanoparticles result in an enlarged or reduced blocking temperature. While structural properties of such self-assembled particles can be routinely studied using electron microscopy, a magnetic characterization is presently mainly based on ensemble averaging. Here we use photoelectron emission microscopy (PEEM) to overcome this limitation. 30nm spatial resolution at the spin-resolved photoelectron emission microscope (SPEEM / UE49 PGM1) allows us to measure x-ray absorption spectra and x-ray magnetic circular dichroism (XMCD) of individual nanoparticles. To correlate magnetic properties of the nanoparticle with its shape and configuration, which cannot be resolved by the SPEEM, we match our PEEM data with previously taken high-resolution electron microscopy images (SEM). A grid of Au markers on the sample substrate serves a map to identify identical places. A specially designed sample holder with integrated magnetic yoke allows us to apply magnetic fields up to 33mT during imaging without significant reduction of the spatial resolution of the SPEEM instrument. Therefore we could study the magnetization reversal of individual nanoparticle as a function of applied magnetic field. During our first study we investigated Fe nanoparticles with a cubic shape, synthesized by the group of Luis M. Liz-Marzan in Vigo [4]. As a result of the colloidal preparation technique the metallic core of the as grown nanoparticles is surrounded by an oxidation shell and organic ligands. The ligand shell also acts as a spacer keeping neighbouring particles at a minimum distance of ~2nm. We performed spectromicroscopy to obtain oxidation state of single nanoparticles in the as grown state. In order to extract local XAS spectra we recorded a stack of PEEM images at photon energies ranging from 700 to 740eV. Fig. 1a) shows a PEEM image taken at an energy close the Fe L3 resonance. Bright spots in the PEEM image correlate perfectly with the position of single Fe nanoparticles visible in the SEM image Fig.1b). After drift correcting the PEEM image stack we extracted XAS spectra of individual nanoparticles by reading out the intensity from a particular region of interest. The XAS spectrum of the indicated nanoparticle is presented in Fig.1c). In the as grown state we find a spectral line-shape that can be explained by a combination of Fe and Fe2O3 contributions. Prior to the magnetic measurements we removed the organic ligands and reduced the iron oxide by a plasma etching technique. Finally we protected the samples by a thin Al layer against re-oxidation.

270

Fig. 1: The PEEM image in panel a) shows the chemical contrast of the Fe nanoparticles recorded at a photon energy close to the Fe L3 resonance. The field of view is about 3µm. Panel b) shows the corresponding image taken with a scanning electron microscope. Small spots around the centre are nanoparticles; the big structure at the top is part of the Au markers on our substrate. Panel c) shows the XAS spectrum obtained from a single nanoparticle located at the position indicated in a) and b).

We started to investigate the magnetic properties of these Fe nanoparticles at the lowest possible temperature in the SPEEM (~115K). At low temperature we expect less thermal fluctuations and therefore a stronger magnetic contrast. For magnetic imaging we exploited the x-ray magnetic circular dichroism at the Fe L3 resonance. In combination with applied magnetic fields during imaging we could measure hysteresis loops of single nanoparticles. Fig. 2a) shows a PEEM image recorded at a different place of that sample where the nanoparticle coverage was significantly higher than at the position shown in Fig.1a). Due to the preparation we find the particles in different configurations that range from single particle sites to small clusters formed by a few nanoparticles. The red marker indicates a position with single particles. The image with the corresponding magnetic contrast is shown in Fig. 2b). It displays the component of the magnetization vector M parallel (red) or antiparallel (blue) to the propagation vector k of the incident x-ray beam. The latter differed from the direction of the applied magnetic field B only by the grazing incidence angle of 16°. At 12mT external field we observe a few Fe nanoparticles with their magnetization antiparallel to the magnetic field possibly due to a strong dipolar coupling with their neighbours. Variations in the magnetic contrast with the cluster size already indicate the influence of the particle configuration on its magnetic properties. One of the main goals of our study was to measure the hysteresis loops of individual nanoparticle configurations, extracting them from a stack of magnetic images recorded at different magnetic fields. The hysteresis of the single Fe nanoparticle indicated in Fig.2a) is shown in Fig. 2c). For that isolated particle we find a closed hysteresis loop indicating a superparamagnetic state. Comparing hysteresis loops of different nanoparticle configurations we find that the magnetic switching behaviour of the nanoparticles is sensitively influenced by their configuration.

271

In summary we have shown that chemical and magnetic properties of individual nanoparticles can be measured using X-PEEM in combination with applied magnetic field during imaging. In combination with SEM data also the influence of different particle configurations on the magnetic switching behaviour can be resolved. This work was supported by EU grant Syntorbmag and the DFG (SFB445).

Fig. 2: The PEEM image in panel a) shows different nanoparticle configurations which are mainly clusters consisting of a few particles. The arrow indicates the position of a single nanoparticle. Panel b) shows the corresponding XMCD contrast recorded with an applied field of 12mT at a temperature of 115K. As XMCD contrast we display the difference of two images taken with circular polarization and opposite helicities divided by their sum. In panel c) we present the hysteresis loop of the indicated single nanoparticle extracted from a stack of PEEM images taken at magnetic fields varying from –23 to 23mT.

References:
[1] O. Margeat, M. Tran, M. Spasova, M. Farle, Magnetism and structure of chemically disordered FePt3 nanocubes, Phys. Rev. B 75, 134410 (2007). [2] C. Antoniak, J. Lindner, M. Spasova, D. Sudfeld, M. Acet, M. Farle, K. Fauth, U. Wiedwald, H.-G. Boyen, P. Ziemann, F. Wilhelm, A. Rogalev, Shouheng Sun, Enhanced Orbital Magnetism in Fe50Pt50 Nanoparticles, Phys. rev. Lett. 97, 117201 (2006). [3] C. Antoniak, J. Multifrequency magnetic resonance phys. stat. sol. (a) 203, 2968 (2006) Lindner, V. Salgueiriño-Maceira, M.Farle, and blocking behavior of Fex Pt1-x nanoparticles,

[4] Alexey Shavel, Benito Rodríguez-González, Marina Spasova, Michael Farle and Luis M. LizMarzán; Adv. Funct. Mat. 17 (2007) 3870-3876

272

Resonant photoemission at 2p edges of magnetic transition-metal oxides exhibiting half-metallic properties
M. C. Richter1, D. R. Batchelor2, J.-M. Mariot3, P. De Padova4, M. Sabra1, O. Heckmann1, A. Taleb-Ibrahimi5, and K. Hricovini1
1

LPMS, Université de Cergy Pontoise, 5 Mail Gay-Lussac, 95031Cergy Pontoise, France
2

BESSY, Albert-Einstein-Str. 15, 12489 Berlin, Germany
3

3

LCP-MR (UMR 7614), Université P. et M. Curie, 11 rue P. et M. Curie, 75005 Paris, France CNR-ISM, via Fosso del Cavaliere, 00133 Roma, Italy

5

Synchrotron SOLEIL, L'Orme des Merisiers, Saint Aubin, BP 48, 91192 Gif-sur-Yvette, France

1. Introduction Transition metal (TM) oxides, for instance Fe3O4, CrO2, and La0.7Sr0.3MnO4, have become the most actively studied half metals. The electronic structure of TM oxides is not well understood yet because d electrons are strongly correlated and cannot be adequately described within a standard band theory framework [1]. Moreover, there is a lack of experimental results proving the true half-metallic behaviour. From the experimental point of view, half metallicity does not show any clear electrical signature and therefore it is not easy to determine. The most direct method to measure the spin polarization is based on photoemission because this technique directly probes the occupied electronic states [2]. But the low efficiency of the spin detection severely limits its use. In the case of nickel (not a halfmetal) it has already been demonstrated that intensities of the resonant photoemission spectroscopy (RPES) spectra induced by circularly polarized light are sensitive to the sample magnetism [3]. In our experiments, aside from the intensities, we expect to see also a difference in kinetic energies of photoelectrons when switching the circular polarization of the light on a magnetized sample. For a given photon energy, the difference in the kinetic energy for left and right light polarization is a consequence of the energy gap in one of the spin channels. This will be a qualitative proof of half metallicity. Hopefully, we will also be able to obtain a quantitative estimate of the gap width. Therefore RPES at TM 2p thresholds may be used as an alternative technique to direct spin detection. Using left and right hand circular polarisation of the light on magnetized samples, a signature of half metallicity can be extracted. 2. Experiment With this idea in mind, we have used the dispersive mode available at BESSY on the photoemission endstation of beamline UE52-PGM to obtain image shots [4]. Such measurements in other synchrotron radiation centres require at present a time-consuming recording of individual spectra for different excitation energies. The set-up at BESSY gives opportunity to take, in one shot, a complete electron spectrum over an energy range large enough to follow the Raman-Auger behaviour with an unsurpassed accuracy. This opens up new possibilities for the study of electron correlation and of subtle dichroic effects in magnetised samples which are expected to be significant in the Raman region. In the present work, we have been able observe the evolution of three Auger channels after excitation in the vicinity of the 2p edges of the TM. This is an important step forward in opening the possibility to use circularly polarised light to investigate such systems.
273

3. Results In this test experiment, we have been able to study in great detail the Raman-Auger behaviour in the 2p-3p3p, 2p-3p3d, and 2p-3d3d non-radiative transitions for Fe and Cr in the halfmetallic systems Fe3O4 and CrO2, respectively. In this report, we present only a preliminary analysis of the data obtained on Fe3O4. Figure 1 shows the 2p absorption spectra for both light helicities together with the x-ray magnetic circular dichroism (XMCD) which indicates that we successfully magnetised remanently the sample.

Fig. 1: Fe 2p x-ray absorption and XMCD spectra of Fe3O4

Fig. 2: Intensity of non-radiative Fe 2p decays as a function of the excitation energy in Fe3O4

In Fig. 2 we show an image of the intensity of the above-mentioned non-radiative processes as a function of the incident photon energy, obtained after normalising the spectra and putting together eight different photon energy intervals. The three Raman-Auger transitions appear as bright spots in the figure. Profiles taken at different photon energies over an energy interval of 0.2 eV fully confirm the presence of the « over-shoot » behaviour for the Fe 2p-3p3p feature which we have observed and discussed recently [5]. The quality of the image spectra shows that it will be possible to do a subtraction between images taken with different light polarisations in order to extract even only subtle differences. We hope that these studies will establish resonant photoemission as an efficient tool for the determination of magnetic structure in half-metallic systems.

274

References: [1] Y. Tokura (ed.), Colossal Magnetoresistive Oxides (Gordon & Breach, London, 2000). [2] See, e.g., J.-H. Park et al., Phys. Rev. Lett. 81 (1998) 1953. [3] L. H. Tjeng et al., Phys. Rev. B 48 (1993) 13378. [4] D. R. Batchelor et al., Nuclear Instrum. Methods Physics Research A 575 (2007) 470. [5] M. C. Richter et al., European J. Phys. ( accepted for publication).

275

NIR-VUV-ellipsometric study of bulk c-GaSb and nano-patterned GaSb surfaces, from 0.7 to 23.6 eV* M. Kildemo1, I. S. Nerbø1, S. Leroy2, C. Cobet3, S. Wang3 and E. Søndergård2 Department of Physics, Norwegian University of Science and Technology (NTNU), NO7491 Trondheim, Norway 2 Surface du Verre et Interfaces, Unité Mixte de Recherche CNRS/Saint-Gobain Laboratoire, UMR 125, 39 Quai Lucien Lefranc, F-93303 Aubervilliers Cedex, France 3 ISAS-Institute for Analytical Sciences, Department Berlin, Albert-Einstein-Strasse 9, D12489 Berlin, Germany Spectroscopic Ellipsometry (SE) of nano-patterned surfaces has been shown to result in accurate estimates of particularly the height of a uniform nano-pattern, and a reasonable estimate of the average shape of the nano-feature [Nerbø et al. Appl. Opt. 47 (2008) 5130]. In particular, nanostructured GaSb produced by low ion-beam-energy sputtering with high flux, is an interesting model system [Facsko et al. Science 285 (1999) 1551]. The primary idea behind the beam-time at BESSY using the Vacuum Ultra Violet (VUV) Rotating Analyzer Ellipsometer (RAE), was to investigate the fundamental question with regards to the polarimetric optical response of nano-patterned surfaces at shorter wavelengths. Furthermore, through the measurement of the dielectric function of bulk-GaSb, and the estimation of the dielectric function of GaSb-oxide and amorphous GaSb (a-GaSb), it is possible to compare the fitted uv-visible graded anisotropic effective medium model to the results from the VUV region, and possibly deduce empirically at what point effective medium theory breaks down. The measurement of not previously reported dielectric functions is a continuous matter of research as instrumentation allows the spectral range to be extended. In this study, the dielectric function of c-GaSb has been accurately measured and extended from 6 eV to 23.6 eV using VUV-RAE. Furthermore, in order to compare the VUV results from the effective medium models fitted in the UV-visible range, the dielectric function of thin GaSb-oxide has also been estimated. Figure 1 shows an overview of the dielectric function of c-GaSb in the range 0.7 to 23.6 eV. In particular, the dielectric function from the reference data from [Aspnes et al. J. Appl. Phys. 48 (1997) 3510] is used in the range 1.5 to 6 eV. Furthermore, the dielectric function was extended from 0.7-1.5 eV using PMSE measurements at 55 and 70 degrees angle of incidence, together with an appropriate overlayer removal. The pseudo-dielectric function measured in this work from Te-doped GaSb in the range 4 to 9.8 eV using a MgF2 rotating analyzer at 68 degrees angle of incidence, and using a triple gold rotating analyzer from 10-23.6 eV at 45 degrees angle of incidence, is also shown in Figure 1. The overlayer from the “clean” HCL etched sample heated in-situ to 300 degrees Celsius was found to be best approximated by a surface roughness, and not by a GaSb-oxide. A 50 % void and 50 % c-GaSb in conjunction with the Bruggeman isotropic effective medium approximation was used to determine an overlayer thickness. The dielectric function of bulk c-GaSb was calculated by inverting (Ψ,Δ) data, and is represented by the full lines overlapping the spectral data from ref. Aspnes from 4 to 6 eV, and covering the new range 6 to 9.8 eV and 10 to 23.6 eV. New features in the dielectric function is observed at 8 eV, 10-11 eV, 15 eV and around 20-23 eV. The most surprising result is a strong feature at 10-11 eV, which experimentally unfortunately falls just in the overlapping region between measurements with the MgF2 rotating analyzer and the triple gold rotating analyzer. By focusing onto the details of the high energy end of the dielectric function of GaSb in Figure 1, the Ga 3d to the conduction band
1

276

excitations are clearly observed (not shown here), as recently described for GaN [Rakel et al. Phys. Rev. B 77 (2008) 115120]. The optical response of random nano-patterned GaSb surfaces consisting of 32 nm high nanocones laterally distributed with an average spacing of 49 nm, was thus investigated by VUV-RAE using synchrotron radiation from 4 to 23.6 eV. The VUV results have been combined with ex-situ ultraviolet-visible spectroscopic ellipsometry. The polarimetric response in terms of the more general Mueller matrix components m12 and m33 has therefore been fitted and compared to the graded uniaxial effective medium model recently reported for nano-patterned GaSb [Nerbø et al. Appl. Opt. 47 (2008) 5130]. The ex-situ SE measurements of m12, m33 and m43 are shown in the left Figure 2, together with the simulated intensities using the graded uniaxial effective medium model. The same nano-patterned surface was also measured by RAE using synchrotron radiation in the range 4-9.6 eV at 68 degrees incidence. The latter RAE measurements of m12 and m33 are shown in the middle Figure 2, together with the simulations using the uv-visible model. The middle Figure 2 additionally shows m43 calculated from m12 and m33 assuming no depolarisation. A future beam-time would preferably include a retarder in order to measure m43 in order to determine depolarisation. The RAE measurements of m12 and m33 in the range 10-20 eV at 45 degrees incidence, are shown in the right Figure 2 together with the simulations using the uv-visible model. The right Figure 2 shows a strong spectral feature from the nano-patterned surface near the region of Bragg condition for ordered surfaces and a strong deviation from the effective medium model.
Aspnes data
24 20 16 12

New Bessy Data

26 24 22 20 18

ε1

16 14

8 4 0 -4 -8 -12 2 4 6 8 10 12 14 16 18 20 22 24

overlayer removal

12 10 8 6 4 2 0

ε2

Eph (eV)

Figure 1. The compiled dielectric function of GaSb from 0.6 to 23.6 eV used in this work. The data from 1.5 to 6 eV is the data from Aspnes et al.
PMSE Ellipsonmetric Intensities (m12,m33 and -m43)

55 degrees incidence UVISEL Measurements
Mueller matrix elements (m12,m33,-m43)
1.0 0.8 0.6 0.4 0.2 0.0 -0.2 -0.4 -0.6 -0.8 -1.0 2 4 6 8 10 12 14 16 18 20 22 24

1.2 1.0 0.8 0.6 0.4 0.2 0.0 -0.2 -0.4 -0.6 -0.8 -1.0

67.85 degrees incidence BESSY Measurements
Mueller matrix elements m12, m33
Is/m43 Ic/m33 Ic2/m12 m12 m33 m43P

1.0 0.8 0.6 0.4 0.2 0.0 -0.2 -0.4 -0.6 -0.8 -1.0

45.5 degrees incidence RAE-BESSY measurements
m43-sim m33-sim m12-exp m33-exp

Is/m43 Is Ic

Ic/m33 Ic

Ic2/m12

2

4

6

8

10

12

14

16

18

20

22

24

2

4

6

8

10

12

14

16

18

20

22

24

Eph (eV)

Eph (eV)

Eph (eV)

Figure 2. Uv-visible spectroscopic ellipsometric measurements (symbols, left Figure) and VUV Rotating Analyzer spectroscopic ellipsometry at Bessy (symbols, middle and right Figure) of nanostructured GaSb. The full lines are the simulations using the uv-visible fitted graded unaixal effective medium model.

* This short report is an extract from a manuscript in preparation and will be submitted during 2009.

277

Pt and Pd-Carbon Nanotube Interaction
Carla Bittencourt University of Mons-Hainaut, Parc Initialis, Av. Nicolas Copernic 1, Mons 7000 Belgium Xiaoxing Ke and Gustaaf Van Tendeloo University of Antwerp, EMAT, Groenenborgerlaan 171, B2020 Antwerp, Belgium Alexandre Felten and Jacques Ghijsen
University of Namur, LISE, 61 rue de Bruxelles, Namur 5000, Namur, Belgium

Among the metals used to decorate CNTs, Pd and Pt appear to be particularly important since besides the catalytic behavior found for Pt decorated CNTs, Pd is considered the most promising metal to achieve transparent contacts; also, ballistic transmission of electrons was reported for the Pd-CNT contact [1]. Thus, the understanding of the interaction between Pd atoms and CNT-surface is a key issue in the design and optimization of practical applications. The nature of the metal coating CNTs was reported [2] to depend on the interaction energy of a single metal atom on a CNT as well as on the metal cohesive energy (Ecoh). The total binding energy of a cluster of size N was estimated to be equal to Ecluster = - Ecoh + α Esurface N2/3, (α is a geometrical factor and Esurface is the average surface energy per atom). Suggesting the existence of a critical cluster size such that metal nanoparticles smaller than such size will efficiently wet the surface, while larger particles will coalesce into still larger clusters. In the present study of the Pt-CNT and Pd-interaction, pristine and oxygen-plasma, treated MWCNTs with different amounts of metal evaporated onto their surface were analyzed. The morphology of the “overlayer” is observed by transmission electron microscopy; its electronic structure and its interaction with the CNT surface is investigated by photoelectron spectroscopy (PES); the latter measurements were performed at UE56 beam line BESSY II (Berlin) using the Mustang end-station. Figure 1, shows TEM images of pristine CNT decorated with Pt (figure 1a) and Pd (figure 1b) thermally evaporated. It can be seen that discrete particles form at the CNT-surface.

Figure 1, show TEM images of a pristine CNT decorated with 5 Å (a) of Pt and (b) of Pd thermally evaporated. It can be seen that discrete particles form at the CNT-surface. Figure 2a shows the comparison of the C 1s XPS spectrum before and after the plasma treatment, the chemical modification of the CNT surface produced by the plasma treatment is revealed by the appearance of a broad structure at higher binding energy. This structure was reported to be generated by photoelectrons emitted from C 1s atoms belonging to hydroxyl, carbonyl and carboxyl (or ester) groups that were grafted at the CNT surface by the oxygen plasma treatment. The inspection of the evolution of the C 1s peak for a sequence of Pt (or Pd) evaporations (not shown) revealed that no additional feature can be observed.

278

a)

b)

c)

Intensity (arb. units)

294

292

290

288

286

284

282

280

Intensity (arb. units)

294

292

290

288

286

284

282

280

Intensity (arb. units)
294

292

290

288

286

284

282

280

Biding Energy (eV)

Biding Energy (eV)

Biding Energy (eV)

Figure 2: C 1s core-electron spectra recorded on a) CNTs, b) Pt/CNTs and Pd/CNTs. Black line – pristine CNTs and red line oxygen plasma functionalized CNTs Metal-CNT interaction can be studied by X-ray photoelectron spectroscopy. If there is a chemical reaction at the interface, then the new chemical environment of the atoms at the interface will lead to the appearance of new features at the XPS spectra. The inspection of the evolution of the C 1s peak for a sequence of Pt (or Pd) evaporations (not shown) revealed that no additional feature can be observed. Figure 2 shows the comparison between the C 1s core-level before and after evaporation of 10 Å of Pt and Pd on pristine and plasma functionalized CNTs. The increase in the asymmetry of the C 1s peak is associated with the many-electron response to the sudden creation of a photohole [2]. Within the Born approximation, it can be shown that the magnitude of this asymmetry is proportional to the square of σ(Ef), the density of states (DOS) near the Fermi energy level, as well as to the effective charge of the photohole seen by the conduction electron, χq [2]. Hence, changes in the asymmetry of the C 1s peak following the Pt deposition, suggest changes in χq and σ(Ef) due to Pt-CNT interaction, it affecting the screening process of the C 1s core hole by perturbing the DOS near Ef. Theoretical results have shown that when a metal atom such as Ni and Pd (which belong to the same column of the periodic table as Pt) replaces one C atom of a graphene sheet, a few electronic levels resulting from the metal atom interaction with the sheet appear close to the Fermi energy level. This effect suggests an increase in the metallicity of this system (when compared to the graphene sheet) what explains the increase of the asymmetry in the C 1s peak. In addition, based on these results it can be suggested that a small shift to higher binding energy of C 1s peak following the Pt evaporation corresponds to an upward displacement of the Fermi energy level and consequently a rigid shift of the electronic states [3]. The decrease in the intensity of the structure associated to oxygen groups is less conspicuous for the Pt/CNT samples (figure 2b), this can be associated to the formation of C-O-Pd bonds in the Pd/CNT samples. Acknowledgements This work was supported by the Belgian Program on Interuniversity Attraction Pole (PAI 6/08), ARC-UMH, by DESY-HASYLAB and the EC under contract RII3-CT 2004-506008 (IASFS). JG is research associate of NFSR (Belgium). The support of the BESSY staff and in particular of Dr. Willy Mahler, Dr. Birgitt Zada and Mr. Mike Sperling is gratefully acknowledged.

279

References
[1] [2] [3] [4] [5] A. Javey, J. Guo, Q. Wang, M. Lundstrom and H. Dai, Nature, 424, 654 (2003). A. Maiti, and A. Ricca. Chem. Phys. Lett. 395, 7 (2004) S. Hüfner, Photoelectron Spectrsocopy, third ed., Springer-Verlag, 173-187, (2003) F. Banhart, J.C. Charlier, P. M. Ajayan, Phys. Rev. Lett. 84, 686 (2000) G.K. Wertheim, S.B. DiCenzo and S.E. Youngquist, Phys. Rev. Lett. 51, 2310, (1983)

280

Internal load transfer in a novel metal/ceramic composite exhibiting lamellar microstructure
Siddhartha Roy, Jens Gibmeier and Alexander Wanner Universität Karlsruhe (TH), Institut für Werkstoffkunde I, Kaiserstr. 12, D-76131 Karlsruhe Funded by the DFG, grant no. WA1122/4-1 Metal/ceramic composites (MMCs) are technologically important because of their high specific stiffness and strength, high wear resistance, better properties at elevated temperatures etc [1, 2, 3]. The aim of our study is to analyze the mechanics of a new class of metal/ceramic composites on a mesoscopic length scale. These composites are produced by melt-infiltration of freeze-cast and sintered alumina preforms by the eutectic aluminum-base alloy Al-12Si. The as-produced material exhibits a hierarchical domain structure with each domain composed of alternating and interpenetrating layers of metallic and ceramic lamellae. Figure 1 shows a typical microstructure of the face perpendicular to the freezing direction for this composite. The ceramic preform for this composite was freeze-cast at -10°C. After sintering and prior to melt-infiltration the porosity of the preform was 56 vol%. After melt-infiltration the porosity was negligible. Diffraction-based techniques are most suitable for the in-situ study of internal load transfer in composites because they provide phase-selective information [4]. Such measurements have e.g. successfully been carried out in particle reinforced [5], whisker reinforced [4] and short fiber reinforced [6, 7] MMCs. Our experiments at beamline EDDI@BESSY aim at studying the internal load transfer in composite samples having one single domain under external compressive load using energy dispersive X-ray diffraction. High-energy synchrotron X-ray diffraction is favoured over laboratory X-ray diffraction or neutron diffraction because this experimental approach combines a large penetration depth with a sufficiently small gauge volume at acceptable measuring times [8, 9]. The energy-dispersive set-up at the EDDI beamline provides a white beam which allows recording the complete diffraction spectra over a wide energy range up to photon energies of 150 keV. Thus several lattice planes of each phase can be taken into account for residual stress analysis and moreover the crystallographic texture can be determined simultaneously [10, 11]. Schematic layout of the beamline components as well as the technical specifications can be found in ref. [12]. For the present study, the energy range 20 – 90 keV was selected for analysis and a scattering angle 2θ = 7° was chosen as it gave good energy separation as well as sufficient peak intensities. A gauge volume having a nominal volume of 0.12 mm³ was defined by the primary and the secondary slits. The slit system as well as the dimension of the gauge volume within the sample is shown in Figure 1. The slit size for the incoming beam (S2 in Figure 1) was maintained at 1 mm × 1 mm while each of the two slits in the diffracted beam (S3 and S4) had dimensions of 60 µm × 5 mm. For in-situ analysis of internal load transfer under external compressive loading, the sample was placed between two hardened steel punches in a miniature mechanical testing rig manufactured by Kammrath & Weiss GmbH, Dortmund, Germany. The testing rig was mounted on the 5 axes sample positioning table of the diffractometer unit. The sample was aligned in a manner that the centre of mass of the gauge volume coincincides with the centre of the sample. Crosshead velocity during compression was maintained at 2 µm s-1, corresponding to a nominal strain rate of 10 -3 s-1. The compressive load on the sample was increased stepwise and at every loading step measurements were carried out according to the sin²ψ method of X-ray stress analysis [13] by tilting the test rig along with the sample in the inclination angle range 0° ψ 90°. After each load application and before the corresponding diffraction measurement, sufficient waiting time (in the range of 5-10 minutes) was maintained to minimize the effects of stress relaxation during the single diffraction measurements. At any applied load for each ψ tilt the acquisition time was 1 minute. In diffraction measurements the lattice strain is always measured parallel to the scattering vector. For ψ = 0°, the scattering vector is approximately transverse to the loading direction while for ψ = 90° it is almost parallel to the loading direction. A pre-load in the range of 20–30 N (corresponding to a stress of 3–5 MPa) was first applied

281

to ensure that the sample did not fall down during tilting of the rig and the strain at this initial state was used as the reference value for further calculations. Hence, no effect of processing induced thermal residual stresses (refer to [14]) is considered and only the extent of internal load transfer under an applied external compressive load is measured, irrespective of the process history the sample experienced. Here onwards, lattice misrostrains and stresses would correspond to changes in lattice microstrains and stresses with respect to this reference state. Volume average lattice strain analyses are carried out in all three phases of the composite in transmission mode. Individual diffraction lines were fitted by a “Pseudo-Voigt” function to determine the line positions.

Figure 1: Schematic diagram showing the measurement geometry for energy dispersive synchrotron X-ray diffraction. The zoomed image shows the microstructure of the face perpendicular to the freezing direction for the actual single domain sample used in the study. Actual dimensions and shape of the gauge volume are marked within the sample and ”q” shows the orientation of the scattering vector.

Composite samples with varying orientations of the loading axis with the lamellae along with samples with preforms coated prior to melt infiltration were studied (details about the effect of domain orientation on the elastic and elastic plastic flow behavior have been studied by Roy and Wanner [5], Ziegler et al. [15] and Roy et al. [16]). Simple monotonic compression as well as loading – unloading – reloading experiments were performed. Only a brief description of the internal load transfer taking place in the single domain sample with uncoated preform compressed along the freezing direction (actual microstructure shown in Figure 1 with direction 1 being the freezing direction) will be given here. Details of the internal load transfer mechanism operative in this sample are given in ref. [17].

Figure 2: (a) combined plot showing the evolution of continuum mechanics average microstrains for all three phases at different applied stresses and (b) Load fraction vs. applied stress plot for alumina and aluminum phases; minor effects of Si has been neglected (ref. [17])

A typical diffraction spectrum for the composite under study can be found in ref. [14]. 8 diffraction peaks of alumina, 4 of aluminum and 3 of silicon were indexed for analysis. Figure 2a shows the change in continuum mechanics equivalent lattice microstrain measured along the loading direction in all three phases of the composite at different externally applied compressive stresses.

282

These continuum mechanics equivalent strains for each phase were determined from the analysis of the multiple diffraction peaks according to the method described in [18]. For this calculation the phases were assumed to be texture-free. Two distinct regions can be identified in Figure 2a. Until about 90 MPa (region I), the lattice microstrains in all three phases increase almost linearly. Afterwards, at higher applied stresses, the microstrain vs. applied stress plots for the aluminum and silicon phases reach a plateau while the microstrain in the alumina phase keeps on increasing almost linearly. Figure 2b shows the calculated load fraction in the alumina and aluminum phases (minor effects from the presents of the Si phase have been neglected) at different applied stresses. The plot clearly shows that at all applied stresses, most of the load is being carried by the alumina phase. At low values of applied stress, the load fraction in alumina first increases and correspondingly, the load fraction in aluminum decreases. At higher stresses, the load fraction in alumina reaches a plateau while the aluminum load fraction slowly tends towards zero. Conclusions Energy dispersive synchrotron X-ray diffraction has been shown to be a powerful tool for insitu study of internal load transfer in a lamellar metal/ceramic composite under external compressive load. Load fractions calculated at various applied stresses in alumina and aluminum phases show that as the applied stress is increased, the load fraction in the aluminum phase decreases, while the load fraction in the alumina phase increases to about 80%. The load transfer from the metallic alloy to the ceramic phase is significant but not complete. A plateau well below unity is observed, which may be attributed to the internal damage processes within the ceramic phase. This work gives a first insight into the mechanism of internal load transfer in this novel composite. Further work will concentrate on the effect of domain orientation and interface modification on load transfer behavior.

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.

Chawla N, Chawla KK. Metal matrix composites. Springer (2006) Clyne TW, Withers PJ. An introduction to metal matrix composites.Cambridge: Cambridge University Press (1993) Berthelot J-M. Composite materials: mechanical behavior and structural analysis. Springer (1999) Bourke MAM, Goldstone JA, Shi N, Allison JE, Stout MG, Lawson AC, Scripta Met. Mater. 29 (1993) 771 - 776 Wanner A, Dunand DC, Met. Mater. Trans. A. 31 (2000) 2949 - 2962 Garces G, Bruno G, Wanner A, Scripta Mater. 55 (2006) 163 - 166 Garces G, Bruno G, Wanner A, Acta Mater. 55 (2007) 5389 - 5400 Wanner A, Dunand DC. Metallurgical and Materials Transactions A. 31A (2000) 2949 - 2962 Fitzpatrick ME, Lodini A. Analysis of residual stress by diffraction using neutron and synchrotron radiation. Taylor & Francis (2003) Reimers W, Pyzalla AM, Broda M, Brusch G, Dantz D, Schmackers T, Liss KD. Journal of Materials Science Letters. 18 (1999) 581 - 583 Pyzalla A. Journal of Nondestructive Evaluation. 19 (2000) 21 - 31 Genzel Ch., Denks IA, Gibmeier J, Klaus M, Wagener M, Nuclear Instruments & Methods in Physics Research: Section A. 578 (2007) 23 - 33 Macherauch E, Müller P., Z. f. angewandte Physik. 13 (1961) 305-312 Roy S, Gibmeier J, Wanner A, Advances in X-ray Analysis. 2008, submitted Ziegler T, Neubrand A, Roy S, Wanner A, Piat R. Compos. Sci. Technol. doi:10.1016/j.compscitech.2008.12.009 Roy S, Butz B, Wanner A. 13th European Conference on Composite Materials (ECCM13), Stockholm, Sweden, June 2 – 5, 2008, paper no. 0303 Roy. S, Gibmeier. J, Wanner. A. Adv. Engg. Mater., accepted Daymond MR. Journal of Applied Physics. 96 (2004) 4263 - 4272

283

Spontaneous capillary filling of silica nanochannels arrays studied by energy-dispersive small-angle x-ray diffraction measurements Daniel Rau1, Stefan Mörz1, Wolfram Leitenberger2, and Patrick Huber1 1 Faculty of Physics and Mechatronics Engineering, Saarland University, D-66041 Saarbruecken 2 Institute for Physics and Astronomy, Potsdam University, D-14476 Potsdam-Golm We studied the capillary filling (spontaneous imbibition) of silica grains permeated by arrays of hexagonally arranged nanocapillaries (SBA-15, mean channel diameter 7 nm) by time dependent energy-dispersive SAXS measurements. The study was aimed at a quantitative understanding of this filling process, in particular with regard to the filling dynamics. The Bragg peaks characteristic of the hexagonal arrangement of the empty nanochannels could be resolved by energy-dispersive SAXS measurements (50s exposition time) – see Fig. 1. Upon invasion of the liquid hydrocarbon (n-C25H52) the intensity of the hexagonal reflections reduced according to a square root of time power law – see Fig. 2. Such LucasWashburn dynamics are expected for a capillarity-driven filling process of the channels, which leads to a reduction in electron density contrast between empty and filled nanochannel. The overall dynamics are significantly slower than expected from the fluid parameters of the n-alkane investigated (viscosity, surface tension, and contact angle), a finding under current further investigation.

Fig. 1: SAXS pattern of the empty SBA-15 powder. The reflections are indexed based on two-dimensional, hexagonal mesh with lattice parameter a=10.8 nm.

Fig. 2: Relative decrease of the (10) reflection upon of invasion of liquid n-C25H52 into the nanochannels of SBA-15 (symbols) in comparison with a square-root-of-time power-law fit.

284

PEEM of Rb/TaS2 : Probing a Metal-to-Insulator Transition with Nanoscale Precision
D. Rahn1 , E. Ludwig1 , J. Buck1 , K. Rossnagel1 , F. Kronast2 , H. Durr2 , and L. Kipp1 ¨
1

Institute for Experimental and Applied Physics, University Kiel 2 Helmholz Zentrum Berlin fur Materialien und Energie ¨

The effects of alkali metal deposition on layered transition-metal dichalcogenides (TMDCs) are of great interest in basic research and in applications, for example battery development. However, the rather simple question of how the alkali metals adsorb on TMDCs has not been satisfactorily answered yet. On the one hand, a certain amount of alkali atoms intercalates into the van der Waals gaps of the TMDCs, while a small amount also remains on the surface [1]. On the other hand, the formation of nano structures in the uppermost layer is observed [2, 3]. Beside these considerations on the kinetics of the intercalation process, the TMDC 1T-TaS2 is of particular interest. It is a correlated material showing an extraordinarily rich phase diagram including various charge-density-wave (CDW) phases and a first-order metal-insulator transition (MIT) at about 180 K with an increase of in-plane resistivity of about one order of magnitude. This transition is widely understood as a Mott-Hubbard localization caused by the favorable valence band structure created by the CDW. It is also accompanied by a modification of √ √ the periodic lattice distortion from a nearly commensurate c( 13 × 13)R13.9◦ reconstruction to a commensurate one. This results in an increased splitting of the Ta 4f core levels which provide a very sensitive tool for investigating changes in the atomic distortion pattern associated with the CDW. Using angle-resolved photoemission spectroscopy, it has recently been shown that a similar metal-insulator transition at the surface of 1T-TaS2 can be induced already at room temperature by adsorption of Rb [4].
a)
µm 230
“sha rp b

b)
oun dary ”

UV lamp

d)
60

Ta 4f

Rb s

tripe

35

y/µm

µm

40

“sha

rp b

oun

dary

E-EF = -23.5 eV 20 hn = 170 eV 20 40 60 80
”

c)
60

e)
60

x/µm
Valence band

Rb 3d

40

40

20 hn = 170 eV 20

E-EF = -109.75 eV

40

60

80

E-EF = 0 eV 20 hn = 120 eV 20 40 60

80

x/µm

x/µm

Figure 1: PEEM images of 1T-TaS2 after spatially Rb evaporation. a) Overview image taken with a Hg discharge lamp. b) Enlarged image showing smaller islands. The contrast is due to changes in the work function. c-e) Same sample area as in b) but using the Rb 3d-, and Ta 4f core levels and the Ta 5d valance states for imaging, respectively.

To further investigate the adsorption of Rb and the accompanying structural and electronic transitions, we prepared a spatially sharply defined Rb stripe on the crystal surface. To this end, we developed a new dispenser device which allows a spatially well-defined deposition. It consists of a SAES-Getters dispenser in a liquid nitrogen cooled copper housing. Two apertures are embedded in the housing to collimate the Rb beam to a specific divergence. The first one, directly positioned behind the SAES-Getters dispenser, has a circular shape and a diameter of 2 mm, and the second one, positioned at a distance of 50 mm behind the first one, consists of a slit aperture with the dimensions of roughly 5 mm × 0.3 mm. During the deposition, the exit slit aperture was very close to the sample (≈ 500 µm) to minimize the divergence of the Rb beam, and the housing was cooled to a temperature of around 150 K. After a spatially well-defined deposition for 390 s at 6 A, photoemission images and spectra were taken with the Elmitec PEEM at beamline UE49 PGMa with an energy resolution of about 350 meV. Photoemission intensity images of the sample surface taken with a Hg discharge lamp (fig. 1a) show a well defined stripe composed of many differently sized islands. This stripe exhibits a width of approximately 230 µm, which roughly agrees with the 300 µm wide exit slit of the evaporator. The contrast in this image and also in fig. 1b) is due to local changes in the work function induced by the Rb adsorption. High intensity regions reflect a local lowering of the work function. The three images (fig. 1c,d,e) show the same sample area as fig. 1b) using adsorbate and substrate core levels and substrate valance states for the image contrast. The first image (fig. 1c) measured at the Rb 3d core level (E-EF = -109.75 eV) shows the same contrast as fig. 1b) and therefore reveals a correlation between the work function lowering and the Rb adsorption. The photoelectron spectra in fig. 2b) (along x-direction
285

a)
Ta 4f
50

b)
-108 intercalated Rb

S

I

Rb 3d
intensity/arb.units
-110

E-EF/eV

surface Rb island substrate

y/µm

-112

40

-114

E-E F= -109.75 eV
-116 -114 -112 -110 -108

hn = 170 eV
20 30 40

E-EF= -23.5 eV
50

-116 28 30 32 34 36

E-EF/eV

x/µm

x/µm

c)

-22

S

I
intensity/arb.units

d)
1 0

-26

island substrate

intensity/arb.units

-24

Ta 4f
E-EF/eV
-1

I

S

E-EF/eV

valence band
island substrate -6 -4

-28

E-EF = -23.5 eV
-30 -30 28 30 -28 -26 -24 -22

-2

E-EF = 0 eV
-2 0

-3 28 30

x/µm

32

34

36

E-EF/eV

x/µm

32

34

36

E-EF/eV

Figure 2: Spatially resolved photoemission spectra. a) Same sample area as in fig. 1b), but slightly enlarged . b) Rb 3d spectra (taken in the box in a) from the islands show an increased Rb signal and an asymmetry orginating from intercalated Rb. Corresponding Ta 4f (c) and Ta 5d (d) spectra reveal an enhanced splitting of the Ta 4f CDW components and a shift of spectral weight away from EF respectively.

in marked box in fig. 2a) show that this contrast is not only due to an overall higher intensity in the island regions. The Rb 3d core level spectra on the islands also exhibit an asymmetry on the low binding energy side which can be attributed to Rb atoms intercalated into the van der Waals gap between the TaS2 triple layers. The second image (fig. 1d) displays the same sample region using the Ta 4f core level (E-EF = -23.5 eV) for imaging. This image reveals an inverted contrast in comparison to the Rb 3d and the UV lamp images. The Ta 4f core level spectra on the islands in fig. 2c) show an overall loss in intensity but also an enhanced CDW induced splitting of the two different Ta 4f spin orbit components. To get a comprehensive picture of the electronic changes, the third image (fig. 1e) shows the same sample region again but now using the Fermi edge (E-EF = 0 eV) for imaging. Similar to the Ta 4f image, a lower intensity on the islands is observed. The spectra in fig. 2d) illustrate that this intensity loss is due to a shift of spectral weight away from the Fermi level to higher binding energies. In conclusion, results of our measurements suggest the following picture. In the sample region directly hit by Rb atoms, a sharp image of the evaporator exit slit could be found. This stripe consists of many smaller concentric islands of intercalated Rb. The different diameters of these islands, especially in the boundary regions of the stripe, and the clustering of intercalation seeds in the directly hit region suggest that these islands orginate from small dot-like intercalation channels generated by one or a few of the fastest Rb atoms that leave the evaporator. After the creation of such an intercalation channel, a concentric expansion probably takes place as long as additional Rb is provided. Thus, the local dimensions of the islands are not correlated with the diffussion length of the Rb atoms inbetween the layers but rather with a spatially varying Rb amount deposited onto the directly hit sample surface. In addition to the elucidation of the intercalation kinetics, the results of our measurements also suggest a spatial correlation between intercalated Rb, a change in the periodicity of the lattice distortion as evidenced by an enhanced splitting of the Ta 4f lines, and the metal-insulator transiton. Moreover, the Ta 5d valance states are more intense and narrower on the islands, which supports the picture of a transition driven by a Mott-Hubbard-type localization.

References
[1] S. E. Stolz, L. J. Holleboom, and H. I. Starnberg, Phys. Rev. B 71, 125403 (2005) [2] R. Adelung, L. Kipp, J. Brandt, L. Tarcak, M. Traving, C. Kreis, and M. Skibowski, Appl. Phys. Lett. 74, 3053 (1999). [3] E. Spiecker, A. K. Schmid, A. M. Minor, U. Dahmen, S. Hollensteiner, and W. J¨ ger, Phys. Rev. Lett. 96, a 086401 (2006). [4] K. Rossnagel, E. Rotenberg, H. Koh, N. V. Smith, and L. Kipp, Phys. Rev. Lett. 95, 126403 (2005).

286

Photoelectron microscopy of Rb/TaS2 : Evidence for a spatially confined 2H to 1T structural transition
D. Rahn1 , E. Ludwig1 , K. Rossnagel1 , C. Enderlein2 , K. Horn2 , and L. Kipp1
1

Institute for Experimental and Applied Physics, University Kiel 2 Fritz-Haber-Institut der Max-Planck-Gesellschaft Berlin

The layered transition metal dichalcogenide TaS2 is of particular interest because it grows in a variety of different polytypes which display effects like superconductivity and charge-density-wave (CDW) formation. The 1T polytype, in particular is a correlated material with an extraordinarily rich phase diagram including various CDW phases and a first-order metal-insulator transition (MIT) at about 180 K (TCDW 1T ). This transition is widely understood as a Mott-Hubbard-type localization caused by the favorable valence band structure created by the CDW. It is also accompanied by a modification of the periodic lattice distortion from a nearly commensurate √ √ c( 13 × 13)R13.9◦ reconstruction to a commensurate one. This ‘Star of David’ like reconstruction exhibits three non-equivalent tantalum positions (CDW sites a, b, c) 1:6:6. In Ta 4f core level spectra emissions from sites b and c are clearly distinguishable by relatively large splitting of the two Ta 4f spin orbit components. This CDW induced splitting, which is increased after the MIT, provides a very sensitive tool for investigating changes in the chemical environments of the tantalum atoms, and in particular for changes in the periodicity of the reconstruction.

a)
Ta 4f core level map (E-EF = -22.31 eV)

b)
-22
DE= 0.31 eV DE= 0.73 eV

c)
intensity/arb. units

DE = 650meV 4f 5/2

4f 7/2

-23 0.8

DE = 300meV

2H bulk crystal
x. pro ap 0 µm 35

E-EF /eV

-24

y/mm

-25 -26 0.0 0.1

2H polytype
DE= 0.73 eV

0.4
1T domain

1T

H /2

ixe m

d

1T polytype
0.2 0.3 0.4 0.5

0.0 0.0 0.4 x/mm

ary ound ple b sam

x/mm intensity/arb. units

-26

-25

E-EF /eV

-24

-23

-22

0.8

2H Ta4f 1T Ta 4f CDW site b 1T Ta 4f CDW site c

Figure 1: a) Constant binding energy map of the Ta 4f photoemission intensity at room temperature (E-EF = -22.55 eV) after Rb deposition on a liquid nitrogen cooled sample. Red dashed lines indicate the Rb deposited region. b) Ta 4f core level spectra (along the solid red line in image a). The spectra show a mixture of 2H and 1T polytype phases. c) Ta 4f spectrum at black arrow in b) fitted with Doniach Sunjic profiles. The spectrum is composed of two 1T related spin orbit doublets (CDW sites b and c) and one 2H related feature. The core hole screening induced asymmetry, typical for the 2H polytype, is fitted by an additional peak.

The 2H polytype of TaS2 shows a less complex phase diagram with only one CDW transition from the undistorted state to a 3×3 superlattice below 70 K (TCDW 2H ) and a superconducting phase below 0.8 K (TSC ). Recent studies on both show that by intercalation of foreign atoms into the van der Waals gap of these layered compounds one can effectively tune the different transition temperatures [2]. For in situ Rb intercalated 1T-TaS2 , an increased transition temperature of the MIT could be observed, leading to an insulating phase already at room temperature [3]. For the 2H polytype, a recent study [4] has shown that Na doping reduces TCDW 2H and simultaneously increases TSC . To further investigate the interplay between the different phases, a spatially well-confined stripe shaped Rb deposition on a 2H-TaS2 crystal was carried out. The spatially well-defined Rb deposition was achieved with a newly developed evaporator device consisting of a SAES-Getters dispenser inside a liquid nitrogen cooled copper housing. Two apertures are embedded in the housing. The first one, directly positioned behind the SAES-Getters dispenser, has a circular shape and a diameter of 2 mm, and the second one, positioned at a distance of 50 mm behind the first one, is slit-shaped with the dimensions of approx. 5 mm × 0.3 mm. During the deposition the dispenser was running at 6 A for 390 s and the sample was cooled with liquid nitrogen. The exit slit aperture was very close to the sample surface (≈ 500 µm) and the housing was cooled to a temperature of around 150 K. The pressure during deposition increased from 9.7×10−10 mbar to 5×10−8 mbar. The photoelectron microscopy measurements were carried out at beamline UE112 PGM-1 at BESSY (spot size: 8 × 11 µm2 ) using a hemispherical electron analyzer (Specs Phoibos 100) and a sample on a stepper motor controlled scanning stage. Photon energies of 80 eV and 180 eV were used. The total energy resolution was about 0.1 eV.
287

a)

Rb 3d core level map (E-EF = -110.28 eV)

b)
-109

c)
DE = 560meV DE = 720meV

E-E /eV F

0.8

I
-111

intensity/arb. units

S

I

I+S

x. pro ap µm 0 35

y/mm

0.4

Rb
intercalated Rb

i str

pe

-113

S
0.0 0.1 0.2 0.3 0.4 0.5
-113 -111 -109

x/mm
intensity/arb. units

E-E /eV F
intercalation peak surface peak 2H surface peak 1T

0.0 0.0 0.4 x/mm 0.8

Figure 2: a) Constant binding energy map of the Rb 3d photoelectron intensity at room temperature (E-EF = -110.28 eV) of the same sample region as in fig. 1 a). b) Rb 3d core level spectra (along the solid red line in a). The Rb signal indicates a sharp Rb stripe c) Rb 3d core level spectrum (at arrow in fig. 1b) fitted with Doniach Sunjic profiles. The spectrum is composed of one surface 1T, one surface 2H and one Rb intercalation related spin orbit doublet.

Figure 1a) shows a Ta 4f photoemission intensity map taken after Rb deposition. A significant reduction of intensity in a stripe shaped sample region is observed, as expected for the deposition with a slit aperture. Taking into account the divergence of the Rb beam, the width of this stripe (350 µm) agrees very well with the one from the slit aperture (300 µm). The changes of the Ta 4f core level in this low intensity domain (fig. 1b) is visualized by core level spectra taken along the solid red line in fig. 1a). The most striking changes are located between x = 0.15 mm and x = 0.35 mm. Here the Ta 4f core level shows an additional splitting of the spin orbit components typical for the 1T polytype. Between the 2H- and the 1T- like sample regions, a continuous transition is observed. The transition region exhibits three peaks in each spin orbit component (see fig. 1c), the intensity decreases and the emissions shift 310 meV to higher binding energies. This shape of the spin orbit components orginates from a mixture of three Ta 4f spin orbit doublets: one from the 2H and two from the 1T domain. The intensity decrease can be explained by remaining surface Rb. The shift results from a charge transfer between the topmost 1T- like and the underlying 2H- like layer and is an evidence that mainly the topmost layer is of 1T configuration. This core level shift has also been observed for TaS2 crystals of the 4H b polytype [5, 6], whose unit cell consists of two 1T-like and two 2H-like layers in alternating order. In fig. 2a) a Rb 3d intensity map of the same region as in fig. 1a) is shown. The high intensity region in this map corresponds to the low intensity region in fig. 1a), which indicates a connection between the Rb distribution and the changes of the Ta 4f core level, especially the polytype transition marked in fig. 1a). Furthermore, the Rb 3d core levels in fig. 2b) show an intercalation related asymmetry only in the 1T region. This suggests that the 2H to 1T transiton may be initiated by Rb atoms intercalated into the van der Waals gap between the topmost and the first underlying layer.

a)
0.5

b)

intensity/arb. units

E-EF /eV

0.0

-0.5

-1.0

0.0

0.1

0.2

0.3

0.4

0.5

x/mm
intensity/arb. units
-1.0 -0.5 0.0 0.5

x-direction

E-EF /eV

Figure 3: a) Ta 5d spectra (taken along the solid red line in fig. 1 a) near normal emission. In the 1T domain, the Ta 5d peak is shifted to higher binding energies indicating a metal-insulator transition, as expected for Rb intercalation into this polytype. b) The spectra taken at the marked positions in a) show a shift of 285 meV on the 1T domain. The altered peak shape supports the picture of a Mott-Hubbard like transition. Both graphs are slightly filtered with a Gaussian.

288

DE=-0.285eV

The surface feature of the Rb 3d core level exhibits a shift (560 meV) between the two polytypes, similar to that of the Ta 4f. Therefore, the Rb 3d core level spectrum consists of two surface and one intercalation related Rb 3d spin orbit doublet (fig. 2c). This explains the reduced separation between surface and intercalation related features. In addition, the spatial confinement of the Rb atoms helps to understand the intercalation process. Since no surface Rb is observed outside the stripe boundaries, already during the adsorption of Rb atoms on the cooled sample some process has to be at work in the directly hit region inhibiting nearly all surface diffusion on warming up. Otherwise surface diffusion would lead to a surface Rb related feature over the entire sample surface. One possible explanation could be that the Rb atoms generate intercalation channels in the directly hit sample area. This would lead to an increased intercalation probability in this local area and thus to a lowering of Rb atoms that leave this region by surface diffusion. Whether the 1T stripe is Rb induced or not, it is expected that this sample area should respond with a MIT when the Rb atoms reach the intercalation sites between the layers [7]. Indeed, a Ta 5d core level map (not shown) exhibits a slight intensity decrease in the whole stripe area and an even stronger intensity loss in the 1T part of the stripe. Fig. 3a) shows Ta 5d spectra (taken along the solid red line in fig. 2a), revealing that this intensity loss is caused by a shift of spectral weight from the Fermi level to higher binding energies (∆E ≈ 285 meV). This may be attributed to a local MIT on the 1T- domain. In the depleted region the Ta 5d peak becomes more intense and slightly narrower, indicating an increased localization of the Ta 5d electrons, as expected for the MIT. Furthermore, the Ta 4f core levels show an increased CDW splitting (ECDW = 730 meV) of the two spin orbit components in this sample region compared to pristine 1T-TaS2 (ECDW = 500 - 550 meV). This serves as evidence for a CDW in the 1T domain and a changed CDW periodicity induced by intercalated Rb. We can therefore conclude that there is a close connection between intercalated Rb and the MIT.

References
[1] E. Morosan, H. W. Zandbergen, B. S. Dennis, J. W. G. Bos, Y.Onose, T. Klimczuk, A. P. Ramirez, N. P. Ong and R. J. Cava, Nature Phys. 2, 544 (2006). [2] L. Fang, P. Y. Zou, Y. Wang, L. Tang, Z. Xu, H. Chen, C. Dong, L. Shan, and H. H. Wen, Science and Technology of Advanced Materials 6, 736-739 (2005). [3] K. Rossnagel, E. Rotenberg, H. Koh, N. V. Smith, and L. Kipp, Phys. Rev. Lett. 95, 126403 (2005). [4] L. Fang, Y. Wang, P. Y. Zou, L. Tang, Z. Xu, H. Chen, C. Dong, L. Shan, and H. H. Wen, Phys. Rev. B 72, 014534 (2005). [5] S.E. Stolz, and H. I. Starnberg, J. Phys.: Condens. Matter 15, 52331 (2003). [6] H. P. Hughes, and J. A. Scarfe, J. Phys.: Condens. Matter 8, 1457 (1996). [7] D. Rahn, H. I. Starnberg, M. Marczynski-Buehlow, T. Riedel, J. Buck, K. Rossnagel, and L. Kipp, DPG spring meeting, O27.2 (2008).

289

Band structure of hard carbon films assessed from the optical measurement
L. Zaj´ckov´1 , D. Franta1 , D. Neˇas1 , V. Burˇ´ a1 , C. Cobet2 ıˇ a c sıkov´
1

Department of Physical Electronics, Masaryk University, Kotl´ˇsk´ 2, 61137 Brno, Czech Republic ar a 2 Institute for Analytical Sciences, Department Berlin, Albert-Einstein-Str. 9, D-12489 Berlin

Introduction
Hard carbon materials find many applications in tribology, MEMS and as protective coatings due to their unique properties such as high hardness, chemical inertness and biocompatibily. In case of microcrystalline and nanocrystalline diamond, diamond-like carbon (DLC) and their modifications (doping, nanocomposites) it is important to control the structure of these materials, namely the content of sp2 and sp3 bonded carbon. The relative content of sp2 and sp3 bonded carbon can be determined from the ratio of pi and sigma electrons. Recently, we published the Kramers-Kronig consistent model of dielectric response based on parameterization of density of states (PDOS) that was employed for various disordered materials (DLC, SiOx , a-Si) [1, 2]. Using this model the ratio of π-to-σ electrons in the carbon materials can be determined from optical measurements provided that the set of thermally annealed DLC films is studied. While the absorption peak corresponding to π electrons (responsible for the weak graphitic bonds) lies in the visible region, the absorption corresponding to σ electrons (responsible for the strong diamond bonds) lies in the VUV region, with the maximum of absorption at about 13 eV. Therefore, it is necessary to carry out the ellipsometric measurements of the films in the VUV region.

Experimental
The DLC films were prepared by plasma enhanced chemical vapor deposition (PECVD) in radio frequency capacitively coupled discharges. The gas mixture consisted of methane and hydrogen. The substrate was double side polished silicon single crystal (c-Si) wafer that can be used also for transmittance measurements in the IR range. The optical measurement of DLC films was performed by following methods. • Transmittance data in MIR (0.046–0.62eV) measured with Bruker Vertex80v FTIR spectrophotometer. • Ellipsometry in the NIR, visible and UV range (1.25–6.5eV) measured with Jobin Yvon UVISEL phase-modulated variable-angle spectroscopic ellipsometer. • Transmittance in NIR (1.13–1.35eV) measured with PerkinElmer Lambda 45 spectrophotometer. • Reflectance in NIR, visible and UV region (1.24–6.5eV) measured with PerkinElmer Lambda 45 spectrophotometer. • Ellipsometry in the UV and VUV range (5.5–9eV) measured with BESSY II synchrotron rotatinganalyser ellipsometer.

Results and Discussion
For the optical characterization of films prepared on c-Si it is essential to use reliable optical constants of silicon. Since such data are not available in wide spectral range applied in present investigation it was necessary, as the first step, to characterize the particular silicon wafer used as the substrate in PECVD. Except for the reflectance, the same methods listed above were employed. The overview of obtained data is shown in Fig. 1.

290

1.0 0.8 0.6 0.4 optical quantities 0.2 0.0 -0.2 -0.4 -0.6 -0.8 -1.0 0 1 2 3 4 5 6 7 8 9 10 photon energy [eV] Is IcII IcIII C2 S2 T Fit

1.0 0.8 0.6 optical quantities 0.4 0.2 0.0 -0.2 -0.4 -0.6 -0.8 0 1 2 3 4 5 6 7 8 9 10 11 photon energy [eV] Is IcII IcIII C2 S2 R T Fit

Figure 1: Overview of experimental data for (left) the silicon substrate and (right) selected DLC film obtained by several methods in wide spectral range and their comparison with the proposed model. The abbreviation of the methods used in caption are following: IS , ICII and ICIII are associated ellipsometric parameters measured with UVISEL ellipsometer, C2 and S2 are second harmonics coefficients measured with BESSY ellipsometer, R and T are reflectance and transmittance, respectively. All experimental data on c-Si were fitted simultaneously by a single Kramers-Kronig (KK) consistent dispersion model covering both interband transitions and phonon absorption. The interband transitions were modelled using true Gaussian-broadened functions describing individual contributions of direct transitions in the one-electron approximation. The indirect transitions forming the indirect band gap were modelled using the parameterized density of states (PDOS) model. The sharp structures at critical points were modelled using hyperbolic excitons. Efficient numerical methods were developed for calculation of true Gaussian broadening of arbitrary functions and for calculation of KK images of such broadened functions. The latter is based on expression of the KK integral using convolutions of the unbroadened function with Dawson integrals. These methods can also be used for modeling of other crystalline materials. Finally, the absorption in MIR region was modelled using a set of Gaussian peaks. The sound agreement between the measurement and the model is apparent from Fig. 1. The determined optical constants of silicon substrate can be seen in Fig. 2. All experimental data (Fig. 1) on DLC films were fitted simultaneously by a single KK consistent dispersion model covering both interband transitions and phonon absorption. The π → π ∗ and σ → σ ∗ interband transitions were modelled using the parameterized density of states (PDOS) model while the absorption in MIR region was modelled using a set of Gaussian peaks. This model had been already applied to characterization of these samples, however, without the synchrotron ellipsometry data [2]. The extension of the spectral range to 9eV resulted in substantial changes in determined band gaps and transition high energy limits found by this model. The most important change is the increase of the σ → σ ∗ band gap from 1.6 to 6.7eV – the new value is much closer to the crystalline and polycrystalline diamond band gap. Consequently, the π-to-σ ratio predicted by the model increased considerably. This change also strongly influenced the dielectric function above the measured range, i.e. above 9eV (see Fig. 3): the sharp structure around 13eV resembles the similar structure observed for diamond.

291

10 10 optical constants 10 10 10 10 10 10 10

1 0

-1 -2 -3 -4 -5 -6 -7

n k

10

-7

10 wavelength [m]

-6

10

-5

Figure 2: Optical constants on the silicon substrate determined by fitting the measurement with proposed model as shown in Fig. 1.
3.0 imaginary part of dielectric function 2.5 2.0 1.5 1.0 0.5 0.0 0 10 20 30 40 50 photon energy [eV] 60 70 80 without BESSY data with BESSY data

Figure 3: Imaginary parts of dielectric function of the DLC film determined from data including/excluding the VUV synchrotron ellipsometric measurements.

Conclusion
The measurement of DLC films at the synchrotron ellipsometer at BESSY II performed in the range 5–9 eV brought additional information about the band structure of DLC material. This is obvious from the fitting results obtained including these VUV data to the data sets obtained by other methods in the range restricted by the highest energy of 6.5 eV. Therefore, it is important to extend further the measurement range to the XUV region.

Acknowledgment
This work was supported by EC under IA-SFS Contract RII 3-CT-2004-506008, project num. BESSYBM.08.2.80317, Czech Science Foundation under contract 202/07/1669 and by MSM 002162241.

References
[1] D. Franta, D. Neˇas, L. Zaj´ckov´, Opt. Express 15 (2007) 16230–16244. c ıˇ a [2] D. Franta, D. Neˇas, L. Zaj´ckov´, c ıˇ a doi:10.1016/j.diamond.2008.10.037 V. Burˇ´ a, sıkov´ Diamond Relat. Mater. (2008),

292

Electronic structure studies of BaFe2As2 by angle-resolved photoemission spectroscopy
S. Thirupathaiah, J. Fink, R. Ovsyannikov, H.A. Dürr, R. Follath, Y. Huang, S. de Jong, M.S.Golden, 4 4 5 5 6 6 Yu-Zhong Zhang, R. Valenti, C. Felser, S. Dastjani Farahani, M. Rotter, D. Johrendt,
1 2

1

1,2

1

1

1

3

3

3

Helmholtz-Zentrum Berlin, Albert-Einstein-Strasse 15,12489 Berlin, Germany

Leibniz-Institute for Solid State and Materials Research Dresden, P.O.Box 270116, D-01171 Dresden, Germany
3

Van der Waals-Zeeman Institute, University of Amsterdam, NL-1018XE Amsterdam, The Netherlands
4

Inst. für Theor. Physik, Goethe-Universität, Max-von-Laue-Straße 1, 60438 Frankfurt, Germany

5 6

Inst. für Anorg. Chemie und Anal. Chemie, Johannes Gutenberg-Universität, 55099 Mainz, Germany

Department Chemie und Biochemie, Ludwig-Maximilians-Universität München, 81377 München, Germany .

Introduction
The discovery of high-Tc superconductivity in iron pnictides [1] is a new era which has acquired much attention from the superconductor’s community. In the case of cuprates the parent compounds are antiferromagnetic Mott-Hubbard insulators and when doped undergo a metal-insulator transition which leads to the superconducting state. On the other hand the parent compounds of the iron pnictide superconductors are metallic in nature and at low temperature a structural transition accompanied by a spin density wave (SDW) order occurs. When iron pnictides are doped the SDW order is suppressed and superconductivity appears [2]. There are several reports on both electron doping and hole doping into pnictides for the achievement of superconductivity [3-4]. The highest transition temperature, Tc that was achieved so far is 56K in SmO1-xFxFeAs [5]. There are other iron pnictides containing no oxygen ions. The prototype of one of them is AFe2As2 (A = Ba, Sr, etc.). The highest Tc obtained in these compounds is 38 K in Ba1-xKxFe2As2 [3]. These materials undergo a structural transition from a high temperature tetragonal paramagnetic phase to an orthorhombic antiferromagnetic phase at low temperatures.

Fig.1. Crystal structure of BaFe2As2. FeAs tetrahedra form two-dimensional layers, surrounded by the layers of Ba. Fe ions inside the tetrahedra form a square lattice.

Along with these prototypes, two more compounds of simple structure were developed. They are namely LiFeAs with highest Tc = 18K in Li1-xFeAs [6] and FeSe (Te) whose highest Tc is 27 K under pressure of 1.48 GPa [7]. In iron pnictides the existence of SDW ordering is supposed to occur because of a Fermi surface nesting scenario between hole pockets and electron pockets. Interestingly the electronic structure of all these compounds is found to be similar. The clarification of electronic structure of iron pnictides is an important step for the understanding of the physical properties and the mechanism for high-Tc superconductivity.

293

We report here our results on the electronic structure of BaFe2As2 using angle-resolved photoemission spectroscopy (ARPES). BaFe2As2 is the parent compound of the superconductors Ba1-xKxFe2As2 and BaFe2-xCoxAs2. Fig. 1 represents the crystal structure of BaFe2As2 in its tetragonal form [8].

Experimental Details
Single crystals of BaFe2As2 were grown out of a Sn flux in Amsterdam and München, using conventional high temperature solution growth techniques. Elemental analysis of the Amsterdam crystals was performed using wavelength dispersive X-ray spectroscopy (WDS). Further elemental analysis was obtained from X-ray induced photoemission spectroscopy on the core levels. Both methods yielded a Sn contamination of approximately 1.6 atomic %. The ARPES experiments were carried out at the BESSY synchrotron radiation facility using the U125/1-PGM beam line and the ”13-ARPES” end station provided with a SCIENTA R4000 analyzer. Spectra were taken with various photon energies ranging from hν = 30 to 175 eV. The total energy resolution ranged from 10 meV (FWHM) at photon energies hν = 30 eV to 20 meV at hν = 175 eV. The angular resolution was 0.2◦ along the slit of the analyzer and 0.3◦ perpendicular to it. The linear polarization of the radiation could be changed from the horizontal direction to the vertical direction. Due to matrix element effects the experimental results are strongly affected by the orientation of the polarization of the photons relative to the scattering plane (defined by the direction of the incoming photons and that of the outgoing photoelectrons).

Results and Discussion
The ARPES measurements on BaFe2As2 were focused to two points of the Brillouin zone, the Γ and the X-point. The electronic properties of the material are determined by Fe states located at these points. To obtain information on the character of the antiferromagnetic order we performed measurements in the paramagnetic tetragonal state and the antiferromagnetic orthorhombic state. Typical ARPES data on BaFe2As2 are depicted in Fig. 2, where we show the measurements near the Γ-point at T = 20 K using 75 eV photons with linear vertical polarization. In the panel (a) of Fig. 2 we have plotted momentum distribution maps for different energies relative to the Fermi level. We see an almost circular Fermi surface which increases with increasing binding energy. This clearly demonstrates that at the Γ-point there is a hole pocket. Comparing our data with band structure calculations indicates that the hole pocket is formed by 3 Fe bands having Fe 3dxy, Fe 3dxz and Fe 3dyz character. This assignment is supported by our polarization dependent measurements (not shown). More resolved data show a small splitting between the Fe 3dxy and the other two Fe 3d bands. In Fig. 3 we show the measurements near the X-point at T = 20 K using 75 eV photons with linear horizontal polarization. In the panel (a) of Fig. 3 we have plotted momentum distribution maps for different energies relative to the Fermi level. The size of the hole pocket is smaller than the electron pocket near the Fermi level. This is in good agreement with previous reports and with bandstructure calculations. According to bandstructure calculations the in-plane Fermi surfaces around this high symmetry point (X-point) should be caused by two concentric electron pockets. The character of these pockets is supported by the experimental data shown in Fig. 3 near the Fermi level. With increasing binding energy the diameter of the Fermi surfaces at the X-point increases. At higher binding energies the Fermi surfaces transform into a blade-like feature, in good agreement with bandstructure calculations. Performing similar measurements at T = 300 K, no differences relative to the low temperature data could be resolved.

294

Fig.2. ARPES data around the centre of Brillouin zone (Γ-point) measured with hν = 75 eV at a temperature of 20 K. The photon polarization was vertical. (a) Momentum distribution maps, (b) spectral intensities as a function of ky at kx = 0, (c) momentum dispersive curves and (d) energy dispersive curves.

Fig.3. ARPES data around X-point measured with hν = 75 eV at a temperature of 20 K. The photon polarization was horizontal. (a) Momentum distribution maps, (b) spectral intensities as a function of ky at kx = 0, (c) momentum dispersive curves and (d) energy dispersive curves.

Summary
In the present ARPES study we have determined the electronic structure of BaFe2As2, a parent compound of doped FeAs-based high Tc superconductors. Performing temperature dependent measurements, important information has been obtained both for the paramagnetic tetragonal state and the antiferromagnetic orthorhombic state. No significant changes could be resolved between the two phases indicating that the real differences for the electronic structure between the two phases are probably much smaller than that predicted by spin-dependent bandstructure calculations. This supports an itinerant character of the Fe 3d electrons in these systems.

References
[1]. Y. Kamihara, T. Watanabe, M. Hirano, and H. Hosono, J. Am. Chem. Soc. 130, 3296 (2008). [2]. H. Chen, Y. Ren, Y. Qiu, Wei Bao, R.H. Liu, G. Wu, T. Wu, Y.L. Xie, X.F. Wang, Q. Huang, X.H. Xhen, arXiv: 0807.3950. [3]. M. Rotter, M. Tegel, and D. Johrendt, Phys. Rev. Lett. 101, 107006 (2008). [4]. A. S. Sefat, R. Jin, M. A. McGuire, B. C. Sales, D. J. Singh, D. Mandrus, arXiv:0807.2237. [5]. X. H. Chen, T. Wu, G. Wu, R. H. Liu, H. Chen, and D. F. Fang, Nature (London) 453, 761 (2008). [6]. X.C. Wang, Q.Q. Liu, Y.X. Lv, W.B. Gao, L.X. Yang, R.C. Yu, F.Y. Li, C.Q. Jin, arXiv: 0806.4688. [7]. Y. Mizuguchi, F. Tomioka, S. Tsuda, T. Yamaguchi, Y. Takano, arXiv: 0807.4315. [8]. F. Ma, Z. Y. Lu and T. Xiang , arXiv:0806.3526v1.

295

Characterization of Carbon Nanotubes Using X-ray Microscopy J. Sedlmair1, S.-C. Gleber1, H. Zänker2, P. Guttmann3, Heim3, J. Thieme1 Institut für Röntgenphysik, Georg-August-Universität Göttingen, Friedrich-Hund-Platz 1, D-37077 Göttingen 2 Forschungszentrum Dresden-Rossendorf, Bautzner Landstraße 400, 01328 Dresden 3 BESSY GmbH, Albert-Einstein. 15, 12489 Berlin
1

Introduction Carbon nanotubes (CNTs) of different origin have been studied with the Scanning Transmission X-Ray Microscope (STXM) at U41 at BESSY II. The combination of NEXAFS absorption spectra with spatial resolution is ideally suited for a detailed analysis of these samples. In recent time, CNTs have drawn a lot of attention due to their unique properties. However, to fully understands and make use of these properties, the characterization of CNTs needs to be optimized. In our study, the differences between the species are observable both in the microscopic images and the spectral data. To be more specific, we have investigated two CNT-samples with energies around the C1s K-shell absorption edge (∼ 284 eV).

Materials and Methods We have investigated CNTs from two different productions: the first ones were from Bayer GmbH (in the following called BayCNT), while the other specimen came from the Forschungszentrum Dresden-Rossendorf (in the following referred to as FZD-CNT) [1]. The samples have been treated with ultrasound for several hours. However, they did not do into solution with the water, but rather precipitated at the bottom of the glass vessels we used. Nevertheless, all CNT-samples have been suspended in water (with concentration of 10 – 40 µg per 300 - 1000 ml) for better handling. The dispersion was dropped onto Si3N4-foils and left there to dry, so that the CNTs would stick to the foils. The measurements were done at the K-shell absorption edge of carbon, i.e. around 285 eV. This energy range lies in the so-called ’water window’, which extends between 280 eV and 523 eV (4.4 nm and 2.1 nm), meaning that here the absorption from water is much lower than from other elements. This is important, as the goal of these studies is to investigate the CNTs in aqueous media. If a sample is imaged at different energies ascending over an absorption edge, each pixel of this so-called image stack contains its own NEXAFS-spectrum. For the stacks here, we recorded images from 280 eV to 300 eV with a stepsize of 0.2 eV. These data were evaluated with "Stack_analyze”, which runs on IDL. The scanning transmission x-ray microscope used for the studies has been described elsewhere [2] [3], as well as the Stack_analyze [4] software used for further evaluation of the stacks.

296

Results and Discussion The first image shows the FZR-CNTs at high resolution (11x11µm2, 220x220px2). The flexible structure of the CNTs is obvious and resembles to the SEM-images we got from the FZD. We also performed a stack of an interesting area of the CNTs on that foil, with a less good resolution (7x7µm2, 30x30px2). The resulting spectrum can be seen in Figure 1b. The program SpecFit, developed in our group, has been used to derive information about single binding states. The second figure shows the results of the same measurements performed with the BayCNTs. Interestingly, both the visible structure as well as the spectral signature differ immensely, although both samples were expected to be similar. 1µm

1a) FZD-CNT FZ 1µm 1b)

2a) Bay-CNT 2b)

Figure 1 and 2: The images show the x-ray microscopic pictures of the two CNTsamples. The NEXAFS-spectra show the absorption of the different CNT-sample depending on the energy of the incident x-radiation, from which information about the binding states in the samples can be achieved.

297

Comparing the spectra it becomes visible, that these CNTs have been fabricated in a way, that functional groups appear differently, thus causing different properties. This makes clear how important it is to have a suitable method to characterize these new materials [5]. The STXM is ideal for this purpose since it combines high spectral resolution with spatial information. Outlook In the future, we want to investigate the role of CNTs within material sciences, such as the interaction with concrete. The incorporation of CNTs in this building material makes it a lot more stabile and lighter at the same time. Since CNTs are applied not only in material sciences, but as well in other fields of research it is also important to know what happens, when they are disposed and get in contact with the environment. This interaction between CNTs with living and non-living organic matter is also subject of our future work. Acknowledgwments This work has been funded by the DFG within the Collabborative Research Center SFB755. We want to thank the staff of BESSY for excellent working conditions.

References [1] Schierz A, Zänker H.: Aqueous suspensions of carbon nanotubes: Surface oxidation, colloidal stability and uranium sorption. In: Environ Pollut. 2008 Nov 14 [2] U. Wiesemann, J. Thieme, P. Guttmann, R. Früke, S. Rehbein, B. Niemann, D. Rudolph, and G. Schmahl: First results of the new scanning transmission X-ray microscope at BESSY-II. In: J. Susini, D. Joyeux, F. Polack, eds., X-Ray Microscopy 2002, Journal de Physique IV 104 (2003) 95–98. [3] P. Guttmann, S. Heim, S. Rehbein, S. Werner, B. Niemanns, R. Follath, G. Schneider: X-ray microscopy at the new U41-FSGM beam line. In: BESSY Annual Report 2007, p 297-300 [4] Jacobsen, C., Medenwaldt, R., Williams, S., 1998. A perspective on biological Xray and electron microscopy. In: Thieme, J., Schmahl, G., Rudolph, D., Umbach, E. (Eds.), X-ray Microscopy and Spectromicroscopy, Springer, Heidelberg, pp. II-93– II-102. [5] Felten A, Bittencourt C, Pireaux JJ, Reichelt M, Mayer J, Hernandez-Cruz D, Hitchcock AP. Individual multiwall carbon nanotubes spectroscopy by scanning transmission X-ray microscopy. In: Nano Lett. 2007 Aug;7(8):2435-40

298

Synchrotron-based infrared reflectance microspectroscopy of metamorphosed CM2 carbonaceous chondrite NWA 4757
Lyuba Moroz1,2, Martin Schmidt3, Ulrich Schade4, Marina Ivanova5
Westfälische Wilhelms-Universität Münster, Institut für Planetologie, Wilhelm-Klemm-Str. 10, D48149 Münster, Germany; 2Deutsches Zentrum für Luft und Raumfahrt e.V. (DLR), Institut für Planetenforschung, Rutherfordstr. 2, D-12489, Berlin, Germany; 3University of California Berkeley, Berkeley CA 94720, USA; 4Helmholtz-Zentrum Berlin für Materialien und Energie, BESSY II, Albert-Einstein Str. 15, D-12489, Berlin, Germany; 5 Vernadsky Institute of Geochemistry and Analytical Chemistry, Russian Academy of Sciences, Kosygin St. 19, 119991 Moscow, Russia.
1

Introduction. Chondritic meteorites (chondrites) are the most ancient rocks formed in our solar system. They provide unique opportunities to constrain the physical and chemical processes that were active in the accretionary disk (solar nebula) surrounding the ancient Sun ~ 4.57 Ga ago. The most chemically primitive carbonaceous chondrite (CC) class consists of 9 groups with distinct bulk compositional and oxygen isotopic characteristics: CI, CM, CV, CO, CK CR CH, and CB. Several meteorites are ungrouped. Several meteorites found in Antarctica have been thermally metamorphosed in their parent bodies (asteroids), but their mineralogical and chemical characteristics would appear to have classified them as CI or CM chondrites before metamorphism [1]. More recently, two non-Antarctic metamorphosed CM2 carbonaceous chondrites (MCCs) - Dhofar 225 and Dhofar 735 - were found in a hot desert: [2, 3, 4]. They differ from typical CM2 chondrites in mineralogy, oxygen isotopic compositions, H2O content, bulk chemistry, and infrared spectra of their matrices [3, 4, 5]. The carbonaceous chondrite NWA 4757, recently found in Morocco, has similarities to both groups of carbonaceous chondrites, metamorphosed and normal CM2s [6]. The meteorite contains 1.9 wt.% H20, 0.68 wt.% C and consists of fine-grained matrix material, round objects sometimes with halos of phyllosilicate and carbonates, and relict aggregates embedded into the altered matrix [6, 7]. Silicates of these objects and the matrix correspond to serpentine in chemical composition [6]. The minor phases are ilmenite, chromite, sulfides, kamacite, taenite, tetrataenite, phosphates, Ca,Mg-carbonates. Average bulk oxygen isotopic compositions are δ17O = 12.84; δ18O = 23.83; Δ17O = 0.45 (all ‰) [7]. The oxygen isotopic composition of NWA 4757 is out of the range of typical CM2 meteorites, and is similar to that of MCCs. The matrix is more homogeneous in chemical composition compared to matrices of metamorphosed and normal CM2 chondrites [6]. In this study we used synchrotron-based IR microspectroscopy at the BESSY IR beamline IRIS [8] to characterize hydration states of NWA 4757 matrix minerals and altered objects. Experimental procedure A polished section of NWA 4757 provided by the Meteorite Committee RAS was used for IR reflectance measurements. The IRIS infrared beamline at the electron storage ring facility BESSY II [8] is equipped with a Thermo Nicolet Continuμm IR microscope coupled to a Nexus 870 FTIR spectrometer. We used an LN2-cooled HgCdTe (MCT) detector, a KBr beam splitter and a 32x Cassegrain objective with a numerical aperture of 0.65. We acquired reflectance spectra in the range between 1.4 and 14 μm. The spectral resolution was 8 cm-1. We probed the meteorite matrix and hydrated objects with a spot size of 15 x 15 µm2 and 20 x 20 µm2. To provide gold-plated surfaces for standard measurements, we deposited gold layers 1mm wide and 100 nm thick directly onto the samples. It is also possible to deposit gold onto a polished epoxy next to the polished meteorite surfaces.

299

1.0 0.9 0.8 0.7
Dhofar 225 (MCC)

AVERAGE MATRIX

Reflectance

0.6 0.5 0.4
NWA 4757 (MCC) Dhofar 735 (MCC)

0.3 0.2 0.1 0.0 8 9 10 11 12 13 14 15 16 17 18 19 20
Kainsaz CO3 Average normal CM2

Wavelength (microns)

Fig. 1. The average matrix IR reflectance spectra of NWA 4757 compared to the matrix spectra of other nonAntarctic metamorphosed CM chondrites (MCCs), “normal” non-metamorphosed CM2 chondrites and CO3 chondrite Kaisaz. Each spectrum is offset for clarity from the previous one.

Results and discussion: The spectra of normal (hydrated) CM2 matrices are dominated by smooth broad Si-O peaks of hydrated silicates, observed near 10 µm (Fig. 1). Anhydrous (or nearly anhydrous) matrices of thermally metamorphosed CMs chondrites Dhofar 225 and Dhofar 735 are spectrally dominated by fine-grained Fe-rich olivine, whose spectra are characterized by three Si-O reflectance peaks between 10.5 and 12.5 µm (Fig. 1). The average IR spectrum of anhydrous matrix of CO3 chondrite Kainsaz shown in Fig. 1 also exhibits similar reflectance peaks. Unlike matrices of typical CM2 chondrites, studied previously [5], the matrix of NWA 4757 is depleted in hydrated silicates and is dominated by Fe-rich fine-grained olivines. In this respect, the matrix of NWA 4757 strongly resembles previously studied matrices of metamorphosed CM2 meteorites Dhofar 225 and Dhofar 735 [5], but is somewhat enriched in hydrated silicates compared to the latter two meteorites. Furthermore, our IR study shows that NWA 4747 contains strongly hydrated objects and halos, absent in Dhofar 225 and Dhofar 735. These objects and halos are significantly enriched in hydrated silicates and carbonates compared to the matrix (Fig. 2). Some of hydrated objects and halos do not show detectable IR signatures of anhydrous silicates, while others contain fine-grained olivines, resembling the matrix olivines in composition. Our study indicates that phyllosilicates in NWA 4757 are more homogeneous in composition than hydrated silicates of typical CM2 chondrite matrices. It is possible that these hydrated objects and halos in NWA 4757 result from terrestrial weathering of the meteorite. A study of trace element distribution will help to understand whether this is the case [6]. Based on our IR study and preliminary mineralogical, chemical and oxygen isotopic analyses [6, 7], NWA 4757 belongs to the metamorphosed CM2 chondrites, and appears to be a mixture of dehydrated matrix material and strongly hydrated objects. If these strongly hydrated objects and halos formed in deep space, they could survive only if the matrix had been dehydrated before their incorporation into the parent asteroid of NWA 4757.

300

olivine (former phyllosilicate)

0.20
Phyllosilicate

Reflectance

0.15

carbonate

0.10
carbonate

0.05
NWA 4757 average matrix NWA 4757 part. hydrated object with carbonate NWA 4757 hydrated rim with carbonate

0.00 6 7 8 9 10 11 12 Wavelength (microns) 13 14

Fig. 2. IR reflectance spectra of (blue) a partly hydrated object containing phyllosilicate, fine-grained olivine and carbonate, and (magenta) a strongly hydrated halo composed of phyllosilicates and carbonates. The average matrix IR spectrum of NWA 4757 is shown for comparison. Characteristic reflectance peaks of phyllosilicates, carbonates and olivine are indicated by arrows.

Acknowledgments We thank Dr. A. Firsov (Helmholtz-Zentrum Berlin, BESSY II) for assistance with the sample preparation. This work was supported by the DLR MERTIS project, Germany, grants RFBR-BSTS (projects N14/04 and 03-05-20008), Austrian Academy of Sciences (FWF, Austria) and PPARC, UK. References [1] Ikeda Y. 1992. Proceedings of NIPR Symposium on Antarctic Meteorite:5, 49-73. [2] Ivanova M.A. et al. 2004. Meteoritics & Planetary Science 39. Abstract #5113. [3] Ivanova M.A. et al. 2005. 36th Lunar & Planetary Science Conference, Abstract #1054. [4] Ivanova M.A. et al. Submitted to Meteoritics & Planetary Science. [5] Moroz L.V. et al. 2006. Meteoritics & Planetary Science 41, 1219-1230. [6] Ivanova M.A. et al. 2004. Meteoritics & Planetary Science 43. Abstract #5103. [7] Ivanova M.A. et al. 2008. In Meteoritical Bulletin 93, 2008 March, 578-579. [8] Schade, U. et al. 2002 Rev. Sci. Instr., 73, 1568-1570.

301

Dynamics within nanoparticle dispersions from the environment studied by soft X-ray spectromicroscopy
1

S.-C. Gleber1 , J. Sedlmair1 , S. Heim2 , P. Guttmann2 , J. Thieme1 Institut f¨r R¨ntgenphysik, Georg-August-Universit¨t G¨ttingen, Friedrich-Hund-Platz 1, u o a o 37077 G¨ttingen o 2 BESSY GmbH, Albert-Einstein-Str. 15, 12489 Berlin

Introduction It has been pointed out that soft X-ray spectromicroscopy is an important tool for investigations of aqueous colloidal samples from the environment (Thieme et al., 2007 (1), Thieme et al, 2007 b (2)). The BESSY scanning X-ray microscope (STXM) provides a combination of high spatial resolution in the range of 30 µm to 50 µm and high spectral resolution of about 3000. This has been used for elemental mapping and chemical analysis of soil samples (Mitrea et al., 2007 (3), Mitrea et al., 2008 (4)). Taking advantage of the natural contrast mechanism, sample preparation is not necessary, it can be imaged in transmission within its original aqueous media up to 10 µm thickness In soil science, the element distribution within soil colloid clusters is of great interest. The here presented application of stereo imaging to the STXM provides a tool for that and can be used to reveal the spatial arrangement of e. g. iron oxides in soil colloid clusters. Changing the chemical conditions of aqueous medium leads to changes in the spatial arrangement, which can be done directly in the X-ray microscope. Thus, in combination with stereo imaging, dynamical behaviour of spatial arrangements can be investigated. The analysis of stereo images is done by the self-written programme xstereo based on IDL (Interactive Data Language). It provides the reconstruction of spatial distances, lengths and edge structures and spatial plots of the marked structures (Gleber et al., 2003 (5), Gleber et al., submitted (6)). Stereo experiments at STXM Stereo experiments at the STXM required the modification of the STXM object plate (Gleber et al, 2007 (7)). Now, the detachable tilt stage, introduced by Weiß, 2000 (8), is implemented as shown in figure 1 right. To mount it to the STXM object plate, an aluminium adapter is applied. A tapered borocilicate glass capillary is used as tiltable sample holder for aqueous samples. It is tilted via an axle by a stepper motor. The axle is inserted into the STXM object plate due to the small distance of both STXM vacuum vessels for X-ray operation. The insertion causes a difference in z-position between the stepper motor and the axle, resolved by a spring. Since 2008, a compact digital camera with USB hub is implemented in the STXM setup, allowing for displaying the sight by the visible light microscope (VLM) with a computer and storing the image. The computer display provides a resolution of about 1 µm when combined with the 40 x objective of the VLM. The resulting improve in prealignment reduces focusing efforts with X-radiation, and thereby minimises dosage and expenditure of time. Both effects are important for stereo experiments. Furthermore, the possibility to store the prealignment images is helpfull to recognise the initially imaged sample region after tilting the object holder. The minimum distance of the capillary from the STXM zone plate is limited by the minimum distance necessary between the capillary and the vacuum window of the zone plate vacuum vessel. The tapered part of capillary has to be quite short for stability reasons. However, the vacuum window of the STXM is not on top of a cone, but of a cylindrical device of about 1.5 cm diameter (figure 1 left). Thus, attaching of the part of the capillary increasing in diameter to the cylindrical mount of the vacuum window limits the achievable distance of the tapered capillary tip from the vacuum window. On the other hand, the minimum distance between vacuum window and zone plate is limited due to the limited possible drive of the zone plate in
302

Fig. 1: Photographs of the STXM object holder plate modified for stereo imaging and the detachable object holder. Left: Photograph of the STXM object holder plate modified for stereo imaging. The detachable tilt stage is mounted via an adapter (shown right besides) in front of the vacuum window of the zone plate vacuum vessel. The axle is inserted in STXM object plate to connect tiltable sample holder with a stepper motor, divided in two parts for easy object holder demounting. Right: Photograph of the detachable object holder for stereo imaging with the STXM with a capillary inserted. Detachable object holder on an adapter for mounting to the object plate. Object-side part of the divided axle visible.

z-direction. Therefore, the capillary tip could not be focused at energies below 300 eV yet. This has to be improved in order to realise carbon mapping in combination with stereo microscopy. Experimental results With the actual setup, it is possible to investigate spatial arrangements within aqueous samples with in-situ manipulation. This is demonstrated with an aqueous soil sample (calcaric phaeozem) and an iron oxide abundand in the environment (haematite) (Gleber et al., submitted (6)). To show the influence of haematite added to an aqueous calcaric phaeozem sample as a progression in time, three pairs of stereo micrographs taken at the STXM are presented in figure 2. The tilt angle was 12◦ for all three stereo pairs, where the upper images and the lower images were taken under equal tilt position, respectively. The first pair on the left side shows a cluster of a dispersion of pure phaeozem filled into a capillary. A cluster of phaeozem particles of different sizes is shown. The image pair presented in the middle of figure 2 shows the same phaeozem cluster, but after the addition of haematite. The lower image is taken 12 min and the upper image 38 min later. Structural changes are visible compared to the images taken before the injection of haematite. The image pair shown on the right of figure 2 is taken from the same phaeozem cluster after third addition of haematite dispersion to the capillary, 2 hrs 40 min after first haematite addition. The image recording started 44 min for the upper image and 55 min for the lower image after third haematite addition. Further changes in the cluster arrangement are visible. To determine the changes induced by the addition of haematite, the stereo images shown in figure 2 were processed with xstereo. The markers applied to the first stereo pair are shown in figure 3. The same structures were marked and the revealed spatial distances are included in the following discussion. The spatial distance between the structures marked as 1 and 2 stays stable at 1.1 µm within the estimated error range of 0.1 µm due to the image quality. The distance between the particle edges marked as 3 and 4 increases after the addition of haematite from 0.6 µm in the first stereo pair (figure 3 left) to 0.8 µm in the second stereo pair and 0.9 µm in the third stereo pair. The plots of the spatial distribution of the marked structures are revealed by xstereo , but not presented here.

303

Fig. 2: Three pairs of stereo micrographs of aqueous calcaric phaeozem in a capillary. Upper images were taken at equal capillary position, and lower ones under a tilt angle of 12◦ . Left: Pure aqueous phaeozem sample. Middle: Same phaeozem cluster after first addition of a haematite dispersion. Imaging time 12 min (lower image) and 38 min (upper image) after addition of haematite. Right: Same phaeozem cluster after third addition of haematite. Imaging time 44 min (upper image) and 55 min (lower image) after third haematite addition. All images taken at E = 400 eV with 50 nm pxl size and an exposure time of 18 ms per pixel. The scale bars indicate 1 µm

Fig. 3: Sample structures marked for xstereo processing of sample shown on the left in figure 2. Only first stereo pair presented.

Outlook To be able to identify particles within the distribution and relate them with the morphological changes of the colloidal soil structures, it is necessary to combine the stereo experiment with elemental mapping. With further adaption of the STXM, it is possible to perform stereo experiments as presented here at the carbon absorption edge. Then, for example the dynamical behaviour of added carbon nanotubes within aqueous soil samples (Sedlmair et al., this volume (9)) can be spatially investigated. Acknowledgement This work has been funded by DFG within the Collaborative Research Center SFB 755 “Nanoscale Photonic Imaging” and under contract number Th 445/8-1 and Th 445/8-2. We would like to thank the staffs of BESSY for providing excellent working conditions, as well as the staffs of our institute for important technical and methodical support. Literature [1] Thieme J, McNulty I, Vogt S and Paterson D 2007 Environmental Science and Tecnolology 41 1027 [2] Thieme J, Gleber S, Mitrea G and Guttmann P 2007 Optics and Precision Engineering 12 1887–1885 [3] Mitrea G, Thieme J, Novakova E, Gleber S C, Guttmann P and Heim S 2007 BESSY Annual Report 2006 190–192 [4] Mitrea G, Thieme J, Guttmann P, Heim S and Gleber S C 2008 Journal of Synchrotron Radiation 15 26–35 [5] Gleber S C, Kn¨chel C, Thieme J, Rudolph D and Schmahl G 2003 X-Ray Microscopy 2002 o (Journal de Physique IV Proceedings vol 104) ed Susini J, Joyeux D and Polack F (EDP Sciences, Les Ulis) p 639 [6] Gleber S C, Thieme J, Chao W and Fischer P Journal of Microscopy submitted [7] Gleber S C, Thieme J, Guttmann P, Heim S and Mitrea G 2007 BESSY Annual Report 2006 234–236 [8] Weiß D 2000 Computed tomography based on cryo X-ray microscopic images of unsectioned biological specimens Ph.D. thesis Georg-August-Universit¨t G¨ttingen a o [9] Sedlmair J, Gleber S C, Z¨nker H, Guttmann P, Heim S and Thieme J BESSY Annual a Report 2008
304

DEAD SEA SCROLLS: CHARACTERIZATION OF PARCHMENT SURFACE AND INKS Admir Masic1, Emanuel Kindzorra2, Ira Rabin2, Ulrich Schade3, Oliver Hahn2, Gisela Weinberg4 1Max Planck Institute of Colloids and Interfaces, Potsdam, Germany 2BAM Federal Institute of Materials Research and Testing, Berlin, Germany 3BESSY, Berlin, Germany 4Fritz Haber Institut der MPG, Berlin, Germany

The Dead Sea scrolls (DSS) were discovered between 1947 and 1956 in eleven caves in and around the settlement Khirbet Qumran at the west shore of the Dead Sea. The texts, of great religious and historical significance, pose a number of historical questions that might be solved with the help of material study. In addition, the scrolls and scroll fragments have experienced complicated and seldom accurately documented post discovery treatments. For longterm preservation and historical study of the scrolls the recognition of the treatment materials is of primary importance. Our previous study was focused on the physical characterisation of parchment surface by means of infrared external reflection spectroscopy (IR-ERS) [1]. The present study is devoted to the recognition of the various deposits on the parchment surface and characterization of the inks. The measurements were performed using infrared synchrotron radiation of the BESSY storage ring (IRIS line) and a FT-IR microscope (Nicolet) equipped with a liquid nitrogen cooled MCT detector. The microscope has a computercontrolled XY-mapping stage with a precision of 0.5 microns and a range of 10 cm, hence allowing for reproducible measurements over several selected microscopic areas of quite large fragments. The full mapping of a 3x3 cm fragment takes about 12 hours, with 128 scans acquired in each point with a spectral resolution of 4 cm-1 and a spatial resolution of 5 microns. For acquiring transmission FT-IR spectra, micro samples were prepared in a Diamond micro compression cell and measured at a room temperature. A 32x Cassegrain objective with an aperture of 20 μm × 20 μm was used to focus the beam on the sample. A total of 256 scans were co-added per sample spectrum (wavenumber range: 4,000– 700 cm−1). Surface characterization.

Fig. 1: EDX spectrum (upper left) and Mid μ-FT-IR absorbance spectrum (lower left) of the DSS sample (IQ10_v) obtained from the reflectance spectra by applying Kramers-Kronig transformation measured at the spot indicated by the black arrow in the scanning electron micrograph (right).

In Fig. 1 an example of absorbance spectra of IQ10 sample (verso) is shown. Beside a Reststrahlen band observed at about 1400 cm-1 associated to calcite (CaCO3) the spectra is characterised by bands which can be assigned to C=O stretching (1741 cm-1), Amide I (1661 cm-1), Amide II (~1550 cm-1), oxalates (1328 cm-1), silicates (1184 cm-1, 1141 cm-1, 1095 cm-1) and calcite (875 cm-1). Since it is known that this sample has not been treated in the post-discovery period we believe the band due to C=O stretching to be a manifestation of the tannins on the surface of parchment, as suggested by John Poole and Ronald Reed. According to their hypothesis the finishing stage of the production of Jewish parchments in antiquity included surface treatment with vegetable tannins. [2] In many cases, however, appearance of the C=O stretch is due to the treatments of the post-discovery period. The summary of results (Table 1) for a number of the Dead Sea Scroll parchments have shown that the spectral features obtained with the IR microscope are directly related to the specific areas analyzed additionally by means of optical and scanning electron microscopy accompanied by EDX and micro-XRF scans. Fibres are always characterized by enhanced Amide I and

305

Amide II bands whereas the dusty surfaces show the dominant presence of calcite and silicates. In the same way, the elemental distribution measured by EDX and micro-XRF that reflects the composition of the analyzed area corroborated the results obtained by IR-microscopy. Table I: Summary of the results obtained by IR mapping of the fragments. Sample r=recto v=verso Specific area EDX,XRF,SEM identified I_Q10_r I_Q10_v GA_v GA_v GA_v Fibers Fibers Border Fibers Gelatinized surface Fibers Gray-yellow deposit Fibers White deposit Glossy surface + + + + + + + + + + + + + + + + + + + + Amide I Amide II Calcit e Silicat es C2O42C=O stretch

+ + + + -

4Q12_r 4Q12_v

+ +

+ +

+ +

+ +

-

4Q12_v 4Q12_v 4Q28_r

+ +

+ +

+ + +

+ +

+ +

4Q28_v

Single fiber

+

-

+

+

-

4Q28_v 4Q28_v

Fibers and dust Fibers

+

+

+ +

+ -

-

Ink characterization by FTIR transmission spectroscopy. Mixing carbon black with the gums and distilled water produces a black ink similar to those used in the ancient times. We have also investigated a sample of commercial Chinese inks. Parchment has strong amide absorptions in the finger print region, therefore sample preparation is of paramount importance: presence of parchment in the sampls can easily mask the characteristic features of the binding agents. Since the samples for the study must be taken from the inscriptions on parchment the extraction of of inks only, i.e. without parchment becomes one of the greatest limitations of this type of the measurement. In Fig. 2 we show the spectra of the inks containing rosins of Acacia trees, oak galls. Note that nature of the binding agent in the commercial Chinese ink was unknown to us.

306

Fig. 2. FTIR spectra of the inks. For comparison two curves in the upper part correspond to the pure binding agents.

The prominent broad peak at 1043 cm-1 clearly indicates the presence of Gum Arabic [3]. Hydrolyzable tannins from the oak gall in the same ink manifest themselves through the series of the peaks between C—O stretch 1190 cm-1 and C==O stretch at 1715 cm-1 (red curve). Bands corresponding to the condensed tannins of acacia Raddiana (around 1000 cm-1) couldn’t be resolved in this sample (blue curve) [4,5]. The binding agent of the Chinese ink (green curve) could be clearly identified as animal glue due to the characteristic pattern of the amide bands of the collagen. Conclusion FTIR reflection microscopy allows to distinguish between the collagenous portions and inorganic deposits on the surface. The applicability of this method is, however, severely affected by the texture and variable thickness of parchments. Extraction of minute samples is advantageous in the cases of well preserved fragments or dummy studies. It is extremely useful for the construction of the data base of mixed media. However, with historical fragments in advanced gelatinization stage, and thus highly inhomogeneous, extraction of a minute sample might not lead to representative results. Acknowledgement The project is sponsored by the Stiftung Preußischer Kulturbesitz. Bibliography [1] A. Masic, M. Ortolani, I. Rabin, U. Schade, G. Martra, O. Hahn, S. Coluccia, Dead sea scrolls: From mid to far IR non-destructive characterization of parchments and their conservation treatment materials, in BESSY – Annual report (2007), 257-259 [2] J.B. Poole, R. Reed, The preparation of leather and parchment by the Dead Sea Scrolls community, Technology & Culture 3 (1962), 1-36 [3] P. A. Williams, G. O. Phillips, Gum Arabic, in Handbook of hydrocolloids, eds. G. O. Phillips and P. A. Williams, CRS press, Cambridge 2000 [4] M. Arshad, A. Beg, Z.A. Siddiqu, Infrared Spectroscopic Investigation of Tannins. Die Angewandte Makromolekulare Chemie 7 (1969), 67-78. [5] M. Giurginca, N. Badea, L. Miu, A. Meghea, Spectral technics for identifying tanning agents in the heritage leather items. Revista De Chimie 58 (2007), 923-928

307

ARCHAEOLOGICAL GOLD PROVENANCE STUDIES – THE CASE OF CARPATHIAN MOUNTAINS NATIVE GOLD
1 1 1 2 2

Bogdan Constantinescu , Angela Vasilescu , Martin Radtke , Uwe Reinholz Department of Applied Nuclear Physics, “Horia Hulubei” National Institute of Nuclear Physics and Engineering, PO BOX MG-6, Bucharest 077125, Romania
2

Bundesanstalt für Materialforschung und – prüfung (BAM), Richard-Willstätter Strasse 11, 12489, Berlin, Germany Work supported by EU grant R II 3 CT-2004-506008, (I3) IA-SFS project “Integrating Activity on Synchrotron and Free Electron Laser Science” The study of trace elements in archaeological metallic objects can provide important clues about the metal provenance and the involved manufacturing procedures, leading to conclusions regarding the commercial, cultural and religious exchanges between the old populations [1]. Ancient metallic objects are inhomogeneous on a micrometric scale, containing remains of imperfect smelting and inclusions (small areas with composition different from the surroundings). The goal of the study is to verify if Transylvanian gold was used to manufacture Romanian archaeological objects. This is realized by using information related to trace elements: Sb, Te, Pb - recognized fingerprints for Carpathian Mountains mines and Sn characteristic for the panned river-bed (alluvial) gold [2]. To solve these issues, samples (grains, nuggets, fine gold "sand") from various Transylvanian mines and rivers and some very small (few milligrams) fragments of archaeological objects are measured. Another outcome of this Synchrotron Radiation X-Ray Fluorescence (SR-XRF) experiment is to obtain the elemental characterization (Au, Ag and Cu content) of representative Transylvanian gold mines, subject of interest for the assignment of any other archaeological artifacts to one of the Central European gold sources. During the 2008 (March) experiment, point spectra for 15 natural gold samples from Transylvania and 12 "micronic" samples from archaeological objects were acquired at 32.5 keV excitation SR energy, using a spatially resolved SR-XRF set-up mounted for analyses at the hard X-ray beam line – BAMline [3] . The beam was focused to a beam size of 100×100 μm . The gold samples were mounted in air in a special frame for passe-partouts on a motorized xyz stage at an angle of 45◦ to the X-ray beam. Fluorescence signals were collected for 300 s each by a Silicon Drift Detector (SDD) detector covered with a with respect to the incident beam. Data are performed by means of AXIL software [4]. Relative concentrations of minor components are determined using a procedure based on different metallic standards and fundamental parameter calculations with SNRLXRF. A summary for the characterization of Transylvanian native gold is the following: • high (7 - 35%) Ag amounts and low (0.15 - 1%) Cu amounts; • placer deposits (Valea Oltului, Stanija, Valea Ariesului) contain as fingerprint Sn (150-300 ppm) – most probably from river bed cassiterite, but also palladium (Lipova – Arad); • primary deposits present as main fingerprints Te (200-2500 ppm), Sb (150-500 ppm) - however, the samples are very inhomogeneous; • primary deposit of Rosia Montana contains te (500 – 5000 ppm) and a mixture of metallic copper and copper from chalcopyrite; • primary deposit Fizesti presents a big amount of Pb = 1%, Sb (350 ppm), traces of Te and also Sn. Practically the native gold samples are inhomogeneous - mixtures of gold-silver-copper metallic alloy with areas of tellurides and antimoniates with galena (lead), pyrite (iron), chalcopyrite (irin, copper) and sphalerite (zinc) inclusions.
2

308

Many isolated pieces or large treasures of ancient gold coins (staters) were discovered in Romania, especially around Sarmizegetusa – the capital of Dacia [5]. The coins analyzed in this work (6 pieces) come from the recently discovered treasure (around 1000 kosons) at Târsa-Luncani [6]. This quaint type of coins is usually considered the only kind of gold coins issued by the Dacians. The strangeness of these coins consists in their Roman iconography. The obverse - an eagle standing left on a scepter holding a wreath in one claw - is inspired by the silver denarii issued by O. Pomponius Rufus; while the reverse - three togate male figures advancing left, the first and third of which carry an axe on their left shoulder - seems to be inspired by a silver denarius issued in 54 BC by M. Junius Brutus. The controversies around these coins are connected with the significance of the inscription, the place of mint and the issuer. There are 2 main types of koson coins: with and without monogram, undeciphered up to date. The study intended to determine whether the gold used for the koson coins is native or refined. This study shows that the type "with monogram" is made from refined (more than 97%) gold with no Sb, Te or Sn traces (re-melted gold), and the type "without monogram" is manufactured from native alluvial gold, partially combined with primary Transylvanian gold (Sn and Sb traces detected). For a comparison, the Greek "pseudolysimachus" type staters (contemporary with "kosons") are made of refined re-melted gold (no Sn, Sb, Te presence). Concerning the Early Bronze Age hair ring from the Tauteu hoard and the Late Bronze Age Vulchitrun disk [7] we found that both are made of alluvial gold.

The results obtained in 2007 and 2008 have been presented at the 3rd International Symposium on Synchrotron Radiation in Art and Archaeology - SR2A - Barcelona, Spain, October 2008, and, consequently, submitted to Applied Physics A. 1. E. Pernicka, Nucl. Instr. Meth. B47, 24 (1986) 2. B. Constantinescu, R. Bugoi, V. Cojocaru, D. Voiculescu, D. Grambole, F. Herrmann, D. Ceccato, Nucl. Instr. Meth. B 231, 541 (2005) 3. I. Reiche, M. Radtke, A. Berger, W. Goerner, S. Merchel, H. Riesemeier, H. Bevers, Appl. Phys. A 83, 169 (2006) 4. B. Vekemans, K. Janssens, L. Vincze, F. Adams, P. Van Espen, X-Ray Spectrometry 23, 278 (1994) 5. O. Iliescu, Quaderni ticinesi di numismatica e antichità classiche, 19, 185 (1990) 6. C. M. Petolescu, The treasure of King Koson (Booklet), Romanian National History Museum, Bucharest (2000) (in Rom.) 7. Al. Bonev, The gold treasure of the Vulchitrun village, in D.W. Bailey, I. Panayotov (eds), Prehistoric Bulgaria, Monographs in World Archaelogy 22, Madison, 277 (1995).

309

Toward a better understanding of the use of bone materials in the prehistoric times: microXRF/SAXS/WAXS investigations of artificially heated bone samples at low temperatures.
Céline Chadefaux¹, Aurélien Gourrier², Oskar Paris 3 and Ina Reiche 1 ¹ Laboratoire du Centre de Recherche et de Restauration des Musées de France, UMR 171 CNRS, Palais du Louvre - 14, quai François Mitterrand, 75001 Paris, France, supported by EU project R II 3.CT-2004-506008, BESSY-ID.08.1.70810 and ID.08.2.80355 as well as ANR project n°: 07-JCJC-0149-01 2 Laboratoire de Physique des Solides, UMR 8502 CNRS, Université Paris Sud, Bât 510 Orsay, France. 3 Max-Planck-Institut für Kolloid- und Grenzflächenforschung, Potsdam, Germany. Ancient bone materials record information in their structure and in their chemical and isotopic composition about the ways of life in the past. This is particularly true for burnt or boiled archaeological samples as they can represent evidences of the mastering of fire by man in prehistoric times. Therefore, it is very important to get access to this information by chemical and structural analysis. Bone materials exhibit a high degree of structural hierarchy. They essentially consist of collagen molecules and hydroxyapatite crystallites. Both phases are closely linked at the nanoscale and the complexity of bone structure and composition is enhanced in archaeological bone materials because of post-mortem alteration processes in the burial environment. These chemical, physical or biological processes may alter the material and so, the archaeological information may be lost [1]. This is the reason why an exact characterization of the state of conservation of archaeological bone material is required. This investigation of the archaeological bone structure and composition will help us to preserve the material itself, to find characteristic markers of ancient human activities and to evaluate if the information contained in its structure can have a real archaeological signification. For this reason, an analytical strategy adapted to the complexity of this nanocomposite biomaterial was developed at the LC2RMF in Paris to characterize the state of conservation of the archaeological bone materials at different scales [2]. Micro-XRF/SAXS/WAXS experiments were conducted at the MySpot beamline at BESSY and were complemented by other technique in order to obtain a complete picture of structural and chemical changes at all levels [3]. In particular, micro-SAXS/WAXS experiments provide average values of the thickness and the orientation of mineral particles across a transverse bone section [4,5]. These measurements will be correlated to the morphological and structural analyses conducted by TEM on ultrathin sections at nanoscale [6]. This will allow us to obtain both a precise structural description using the TEM images, although limited to a highly localized part of the samples, and the corresponding average parameters over much larger sampling regions using scanning SAXS/WAXS. One of the purposes of our project deals with the characterization of structural modifications in bone material induced by heating at low temperatures (100-300 °C). Most of the previous research concerns studies of the heat-induced modifications at high temperature (500-940 °C) and its effect on the mineral part of the bone material [7,8].

310

Material In order to investigate the effect of low temperature treatments on the particle size and orientation, three modern sheep bone samples artificially heated have been analysed by micro-SAXS/WAXS. These samples were heated during 60 minutes at 150, 170 and 200°C. A modern sheep sample not submited to any heat treatment was also measured for reference and archaeological samples were also analysed to distinguish alteration features linked to the heating from those linked to the diagenesis. Experimental Scanning SAXS/WAXS experiments were conducted at the MySpot beamline [9] by working at an energy of 15 keV using a Si/W multilayer on Si(311) as monochromator. The beam size was reduced by collimation to ~ 30 µm in dia meter. Simultaneously, X -Ray fluorescence (XRF) measurements were acquired to correlate the structural crystalline modifications with the elemental distribution across the sample. Two dimensional mappings were performed from the outer part of the bone material section to the inner part to evaluate the structural and chemical modifications. Indeed, during burial, the outer part of the bone section, directly in contact with the soil, is more sensitive to interactions with the environment and will be more intensively modified [10]. Thus, scans were carried out in steps of 30(h) x 30(v) µm2 across the transverse sections to use the full resolution of the instrument in direct space corresponding, in first approximation, to the beam diameter. The sample -todetector distance and the detector tilt and centre were calibrated using a Ag behenate standard [11]. The data were reduced using the FIT2D software [12] and analysed with software developed in Python by Aurélien Gourrier using a procedure described in detail elsewhere [5]. The analysis of the radial profiles provides parameters related to the shape and the thickness (so called T-parameter) of the bone crystals while the azimuthal profiles allow deriving orientation parameters [13]. Preliminary results and discussion The figure 1 presents the results obtained on the bone sample artificially heated at 170°C. A region of the sample indicated in fig. 1.A by a rectangle was scanned and analysed to derive a map of the particle thickness shown in fig.1.B. About 80% of the T-parameter values fall in a range of 2.5-3.2 nm as observed in the histogram of the parameter distribution (fig. 1.C). The lower values seem to correspond to a network of porosity in the sample , which can be related to the channels formed during the remodelling process of the living tissue as observed from fig.1.B. It is not clear, at present, whether this is an effect of the heat treatment. In fact, lower T-parameter values have also been reported to be linked with a late stage of the mineralization process during the remodeling phase [5]. We will process the measurements of the other heated samples to compare them to the modern reference in order to find out characterisic heat induced modifications at low temperature. Last but not least, it is also important to be able to differentiate heating effects from those of diagenesis in ancient bones. Work is in progress to clarify these issues.

311

Figure 1: Evaluation of the thickness of the bone crystals in an artificially heated bone section: Alight microscopy image of the transverse bone section analysed B- map of the T-parameter and Ccorresponding % distribution over the scanned area.
References [1] R.E.M. Hedges, Archaeometry 44, 3 (2002) 319 [2] I. Reiche, L. Favre-Quattropani, C. Vignaud, H. Bocherens, L. Charlet, M. Menu, Meas. Sci. Technology 14 (2003) 1608. [3] C. Chadefaux, A. Staude, J. Nötel, H. Riesemeier & I. Reiche, Appl. Phys. A (in prep.) [4] S. Rinnerthaler, P. Roschger, H.F. Jakob, A. Nader, K. Klaushofer, P. Fratzl, Calcif. Tissue Int. 64 (1999) 422 [5] A. Gourrier, W. Wagermaier, M. Burghammer, D. Lammie, H.S. Gupta, P. Fratzl, C. Riekel, T.J. Wess & O. Paris, J. Appl. Cryst 40 (2007) 78 [6] C. Chadefaux & I. Reiche, J. nano Res. (accepted). [7] I. Reiche, C. Vignaud, M. Menu, Archaeometry 44, 3 (2002) 447 [8] J. Hiller, T.J.U. Thompson, M.P. Evison, A.T. Chamberlain, T.J. Wess, Biomaterials 24 (2003) 5091 [9] O. Paris, C. Li, S. Siegel, G. Weseloh, F. Emmerling, H. Riesemeier, A. Erko, P. Fratzl, J. Appl. Cryst. 40 (2007) 466 [10] I. Reiche, L. Favre -Quattropani, T. Calligaro, J. Salomon, H. Bocherens, L. Charlet, M. Menu, Nucl. Instr. Meth. Phys. B 150 (1999) 656 [11] T. N. Blanton, T. C. Huang, H. Toraya, C. R. Hubbard, S. B.Robie, D. Louër, H. E. Göbel, G. Will, R. Gilles, & T. Raftery, Powder Diffr 10 (1995) 91 [12] A.P. Hammersley, ESRF Internal report HA02T (1997) [13] P. Fratzl, S. Schreiber, P. Roschger, M.H. Lafage, G. Rodan & K. J. Klaushofer, Bone Min. Res. 11 (1996) 248

312

SR-XRF analysis of Pt trace contents in ancient gold alloys at the BAM line: a new approach Maria F. Guerra¹, Martin Radtke², Ina Reiche¹, Heinrich Riesemeier², Erik Strub² ¹ Centre de Recherche et de Restauration des Musées de France, UMR 171 CNRS, Palais du Louvre 14, quai François Mitterrand, 75001 Paris, France, supported by EU project R II 3.CT-2004-506008, BESSY-ID.07.1.693 ² Bundesanstalt für Materialforschung und –prüfung (BAM), Richard-Willstätter-Straße 11, 12489 Berlin, Germany. Queries on the origin and provenance of gold used in the past require for precious and rare items the use of non-destructive analytical techniques providing information on the characteristic elements of the metal. Gold sources exploited in the antiquity were mainly of secondary type (especially alluvial) and in a few cases of primary type (especially auriferous quartz). The different metallurgical steps from the gold ore to the gold alloy – in the past a gold alloy consisting of a combination of gold with different amounts of silver and copper – results in the loss of information on the original gold both by evaporation or absorption in the cupels of a large number of elements present in the ore and by addition of other different metals. For this reason, together with the fact that gold has always been remelted and reused to produce other objects, only very few elements are characteristic of the gold source. Among the characteristic elements of gold, Pt together with elements such as Pd, Sn, Sb,... provides important information on the geological source [1-3]. For example, the presence of Pt in gold, together with Sn, is essential to distinguish secondary from primary gold deposits [4]. In the case of available geo-chemical information on the ancient gold sources, Pt can b used to provenance gold and follow e its circulation in the past [5]. Few analytical techniques are at present available to measure Pt directly on the gold object (noninvasive) or, in a few cases, on tiny samplings. Nowadays, the most accepted techniques are ICP -MS, PIXE and PIXE-XRF. However, ICP-MS requires sampling and IBA techniques, in spite of being totally non-invasive, have inadequate detection limits for certain elements such as Pt. We could show that with PIXE-XRF we could only attain a minimum detection limit (MDL) of 80 ppm. For these reasons, SR-XRF appears to be a good solution for the analysis of goldwork. In fact, SRXRF fulfils all the necessary requirements, which means MDLs that can be optimised by adapting the excitation energy to the element to be measured, and micro-beams for a non-invasive analysis with good spatial resolution. Our first attempt to develop a protocol in order to measure low contents of Pt in gold alloys [6] was carried out with excitation energy at the Pt L3 -edge (11.564keV). The evaluation of the Pt contents in gold (Z-1 compared to Au) required a specific procedure taking into account the resonant Raman scattering [7,8]. Using a simple model, quantification of the Pt content was carried out by Monte Carlo simulation using the program MSIM7 [9] and by subtraction of the Au and Pt spectra obtained on pure standards. The MDL could be estimated to 20ppm. In order to solve the problems connected to both the rather poor MDL obtained at low energy and the difficult quantification process of Pt contents, we made new developments of both the analytical SRXRF set-up and the quantitative calculation procedure at high energy [10]. In the case of the BAMline, the continuous radiation ranges from 5keV up to 100keV allowing the use of the K-lines of Pt. An intense monochromatic photon beam of 0.2x0.2mm² was obtained with a double multilayer monochromator (DMM) leading to a 100 times higher photon flux attaining the sample compared to often used Si 111-Monochromators. Energies for the K-edges of high Z elements can only be reached with reasonable intensities with the DMM excitation. The size of the beam was defined by a slit. We used a 20mm² HPGe detector with nominal resolution of 130eV at photon energy of 5.9keV, placed at a distance of about 2.5cm from the sample, perpendicular to the photon beam in the plane of the electron storage ring, in order to minimize the background contribution from the scattered radiation; the signal delivered by the detector was processed with a Saturn digital signal-processing unit from XIA. An Al foil insulates the detector from the low energy fluorescence spectrum. The sample was inclined by 45° with respect to the photon beam and the detector. All spectra were normalized to the integrated excitation intensity. A Pt free Au standard was measured under the same experimental conditions as the sample for the estimation of the Pt content. Calculation is carried out by subtraction of the pure Au spectrum from the

313

spectrum of the sample. The peaks positions and widths of the difference signal and of the Pt sample were compared. Excitation energy of 79.5keV, lying between the Pt and the Au K-absorption edge, respectively at 78.395keV and 80.725keV, was chosen. Because of the limited energy resolution of the DMM, around 1.2keV FWHM at this energy, an Au fluorescence signal is still present in the spectra. The Au Ka2 peak at 66.989keV interferes with the Pt Ka1 peak at 66.832keV. The Raman scattering from the gold is also located in this energy region. Therefore we used the Kß1 line to carry out the calculation. Table 1: Composition by ICP-MS and MDL Kβ 1 activation analysis of different ancient gold Au % Ag % Cu % Pt ppm Pt ppm samples of infinite thickness used to evaluate ancient alloys S1 98,2 1,7 0,05 61 89 the MDL of Pt using the Kß1 line. S4 92,9 3,0 3,9 634 40
S5 S6 S2 S3 68,3 28,5 90,2 6,3 87,8 9,8 90,3 0,3 3,1 2,9 2,3 9,4 291 2233 78 232 67 73 71 74

ancient coins

The MDL for Pt was calculated with the fit of the un-normalised spectra based on the formula suggested by IUPAC [11]. Table1 shows that the Pt MDL calculated for a set of gold coins and plaques with different Pt contents range from 40 to 90ppm. In this table all the samples are of infinite thickness. Figure 1: Comparison of the Pt MDL (ppb) 10000,00 MDL (in ppb) obtained for Pt by 1000,00 SR-XRF at high and low energy at the BAMline with other 100,00 analytical techniques usually 10,00 used for gold provenancing: 1,00 classical PIXE and with a 75µm Zn selective filter, PIXE-XRF 0,10 with As primary target and 12 0,01 MeV PAA [1] together with ICP0,00 MS and LA-ICP-MS [5,12].
0,00 ICP-MS 1 LA-ICP-MS 2 PAA 3 PIXE 4 PIXE filter 5 PIXE-XRF 6 SR-XRF[6] 7 SR-XRF 8

Figure 1 compares these MDLs with those obtained for other techniques: high energy SR-XRF shows a MDL equivalent to PIXE-XRF whilst at low energy the MDL enhanced by a factor of about 3. We tested our new protocol with a 79.5keV incident beam of 0.2x0.2mm², based on the fact that gold samplings are usually very tiny and included in resin for non-destructive analysis and that goldwork is of very different type and size: from large statues to mini-beads, from plain objects to thin foils. However, with our set-up only objects or samples with a thickness superior to 300µm (the analytical depth for PtKβ1 in gold [13]) can be considered as infinite thick during quantification and only in certain cases the beam is smaller than the sample. The analysis of 3 thick ancient gold coins showed that the presence of additional elements in the gold alloys, such as Os and Ir, increases the Pt MDL. The analysis of four archaeological gold samples illustrated the difficulties connected to the study of such small or thin items: for two samples included in resin we could calculate 120ppm of Pt for one and 420ppm for the other with a MDL of 80 ppm; for two thin foils of about 50µm thickness Pt concentrations were under the MDL. Confirming the significant role of SR-XRF in gold provenancing, these results showed that the use of L-edge measurements and of a Si SDD detector accepting higher count rate should simplify the analytical task as samples can be assumed as infinite thick. This assumption does not held for the highenergy measurements and can introduce a significant source of error for quantitative analysis. The first results at the Pt L-edge are in progress. They were obtained with excitation energy of 11.58 keV. This energy was chosen after a series of test measurements with different energies to optimize the detection limits. DMM and DCM have been used in series to suppress harmonics. An SDD detector was used to get maximum count rate with acceptable energy resolution of 380 eV for Pt with a count rate of typically 70000 cps and a dead time of 50%.

314

The calculation of the MDL has been done with respect to pure element standards. In Figure 2, the fit for the Standard RM8058 Pt is shown exemplarily. For this case a MDL of 3.1ppm for Pt has been reached, using the standard IUPAC definition. Anyhow, the real detection limit must be assumed a little worse due to the uncertainty of the non standard fit procedure. Quantification of unknown samples has been done with a rule of proportion, using RM8058 as standard. Figure 2: Fit of the measured signal with a sum of pure Pt, Au, Ag, Zn and Cu measurements. The peak of Pt is located in Channel 19. The yellow line shows a pure Pt signal. The black line shows the fraction of the signal of the sample whish is attributed to Pt. These signals are correlated with a correlation factor of 0.93. This is strengthening the position, that the signal can really be attributed to Pt.

Standard RM 8058 containing 40.8ppm of Pt was used to calculate the Pt MDL with the new protocol together with two coins analysed by 12MeV proton activation analysis (PAA) and LA-ICP-MS. We could show in table 2 that the Pt MDL ranges from 3.1 to 4.7, attaining 3ppm for standard RM 8058. This value is equivalent to the MDL obtained by PAA at optimised conditions. In this same table we compare the Pt concentrations obtained by SR-XRF at low energy with the expected values of the two coins measured by PAA and ICP-MS. The good agreement between the expected and obtained concentrations is evident. Table 2: Expected Pt concentrations for SR-XRF SR-XRF expected Pt Pt / ppm MDL / ppm ppm standard RM 8058 and a set of ancient gold coins measured by ICP-MS and activation RM 8058 40,6 3,1 40,8 analysis compared with the concentrations coin NAP 624,8 4,7 630 obtained by SR-XRF at low energy. The Pt coin FRA 2665 3,6 2924 MDL is shown for the three samples. The application of this new protocol to a few thin ancient gold foils of about 50µm thickness showed that Pt concentrations could be calculated at low energies. A few archaeological samplings included in resin were also analysed and their Pt concentrations could as well be obtained.
References [1] M.F. Guerra, Nucl. Instr. and Meth. B 226 (2004) 185. [2] M.F. Guerra, J. Archaeological Sciences 31 (2004) 1225. [3] A. Gondonneau, M.F.Guerra, Archaeometry 44 (2002) 473-599. [4] R.K Dubé, Gold Bull. 39 (2006) 103. [5] M.F. Guerra, in: R. Van Griecken, K. Janssen (Eds.), Cultural heritage conservation and environmental impact assessment by non-destructive testing and micro-analysis, Balkema, London, 2005, 223. [6] M. F Guerra, T. Calligaro, M. Radtke, I. Reiche, H. Riesemeier, Nucl. Instr. and Meth. B 240 (2005) 505. [7] T. Fujikawa, T. Konishi, T.Fukamachi, J. Electron Spectrosc. Relat. Phenom. 134 (2004) 195. [8] P.P. Kane, Physics Reports-Review Section of Physics Letters 218 (1992) 67. [9] L. Vincze, K. Janssens, F. Adams F, M.L. Rivers, K.W. Jones, Spectrochim. Acta Part B 50 (1995) 127. [10] M. F Guerra, M. Radtke, I. Reiche, H. Riesemeier, E. Strub, Nucl. Instr. and Meth. B 266 (2008) 2334. [11] P. Kump, Spectrochim. Acta Part B – Atomic Spectroscopy 52: (1997) 405 [12] A. Gondonneau, M.F. Guerra British Archaeological Reports IS 792 (1999) 262 [13] R.E. Van Grieken, A.A. Markowicz (Eds.). Handbook of X-Ray Spectrometry: Methods and Techniques. Marcel Dekker, Inc., New York, 1993.

315

Photon-energy calibration for the 10m-NIM beamline by establishing an atlas of H2 absorption lines between 85 and 72 nm (14.5 – 17.3 eV).
M. Glass-Maujean, , A. Knie*, I. Haar*, R. Hentges*, W. Kielich*, K. Jänkälä*, A. Ehresmann*, and H. Schmoranzer** Laboratoire de Physique Moléculaire pour l’Atmosphère et l’Astrophysique, Université P. et M. Curie /CNRS, 4, pl Jussieu, F-75252 Paris Cedex 05, France *Institut für Physik, Universität Kassel, D-34132 Kassel, Germany ** Fachbereich Physik, Technische Universität Kaiserslautern, D-67653 Kaiserslautern, Germany The U125/2-10m-NIM beamline at BESSY is probably the beamline with the highest resolution world wide in the VUV spectral range. In any case, there is no comparable instrument at the other synchrotron facilities in terms of spectral resolution and light intensity. However, in order to fully use its capabilities, it is necessary to characterize precisely its apparatus function and ultimate resolution and, particularly, to calibrate the energy of the delivered photons in the relevant energy range. The molecular hydrogen absorption spectrum was used for that purpose as it presents hundreds of lines in the 14.5 – 17.3 eV energy or 85 – 72 nm wavelength range, many of them being still very narrow. At these energies, however, the hydrogen molecule may dissociate and, above 15eV, also ionize; nevertheless, many excited levels still decay via molecular fluorescence connected with a sufficiently long lifetime to lead to very narrow absorption lines. The absorption spectrum was recorded simultaneously with the ionization, dissociation and fluorescence excitation spectra using a new redesigned target cell (fig. 1). The target
Figure1: redesigned target cell

gas pressure used amounted to 5 or 20 mTorr, at room absorption 5x10 temperature. The 10m-normal-incidence 0 monochromator was equipped 76,55 76,60 76,65 76,70 76,75 76,80 5x10 with a 1200-lines/mm grating dissociation and was used at third order with 45µm-wide slits. The translation 0 76,55 76,60 76,65 76,70 76,75 76,80 of the grating had been adjusted 5x10 ions to get the best resolution at the third order. 0 The apparatus function 76,55 76,60 76,65 76,70 76,75 76,80 5x10 and the resolution were fluorescence measured on lines known to 0 present a width equal to the 76,55 76,60 76,65 76,70 76,75 76,80 Doppler width, i.e. 0.00035nm incident wavelength (nm) (HWHM) at 80 nm. In this case, Figure 2: Absorption, dissociation, ionization and fluorescence cross the measured profile was found sections, (for the fluorescence, the values are only qualitative). nearly symmetrical with nearly an expected triangle-shape with a width of 0.0016±0.0002 nm (HWHM) which is by 50% larger than last year. The flux transmitted through the monochromator, as measured on the on the last refocusing mirror before the experiment, was around 3 times lower than last year.
-17 -17 -17 -17

cross sections (cm )

2

316

2

1

0

-1

-2 78 80 82

incident wave length (nm)

Figure 3: Energy differences between the measured positions and the reference data

We compared the positions of the absorption lines in the 83-81 nm and 79-78 nm spectral ranges with reference data obtained thirty years ago, on photographic plates, at low temperature, with a resolution of 300 000, simultaneously with Ritz standards1, precise within 1.0 or 0.3 cm-1 (for the data corrected for pressure shift) and confirmed for several lines by recent laser measurements2. Obviously, the recent corrections in the 10m-NIM grating drive control loop (Heidemann correction) improved greatly the precision of the measurements. The measured values are now precise within one cm-1 at 80 nm (125000 cm-1), instead of six in previous years. As can be seen from figure 3 the

absorption and dissociation cross sections (cm )

2

σDATA - σBESSY (cm )

-1

measured positions are really much improved and consistent. The complete scan of the lines still contains a few steps much larger than demanded; but these spikes can be easily identified and removed.

As the hydrogen molecule is a simple molecule; the potential curves of the lower states have been calculated at high precision with the adiabatic corrections. It is quite straightforward to solve the Schrödinger equation and calculate the positions of the rovibronic levels in the adiabatic approximation. It is also easy to calculate the intensities of the absorption lines in this approximation. This approach is 8x10 − quite good for the 1 Π u levels
-17

6x10

-17

excited from the ground state with 1 + Σ g symmetry through Q lines; most of the levels with a long radiative − lifetime belong to the 1 Π u symmetry.

4x10

-17

2x10

-17

However, this approach is only fair + or even poor for the 1 Σ u levels

which are very sensitive to the nonadiabatic perturbations. 0 The positions of the Q lines 75,8 76,0 76,2 76,4 were reproduced by the calculations -1 incident wavelength (nm) within a few cm as tested on the existing reference data. At higher Figure 4: Part of the spectrum: in grey the absorption cross energies, where no precise section, in blue the dissociation one, with the calculated Q(1) lines experimental reference data are (green stars) available, we compared the computed positions of the Q lines with our measured positions. The calculated positions were found in very good agreement with the observed values of our spectrum3 and, in most cases, the intensities were also well reproduced; the non-adiabatic couplings affect the intensities of the Q lines. (see figure 4) The planed energy range could not be fully covered during this beam time. Therefore, the project should be completed during the next beam time in spring 2009.
G. Herzberg and Ch. Jungen, J. Mol. Spectrosc. 41, 425 (1972) and S. Takezawa, J. Chem. Phys., 52, 2575 (1970) 2 M. Sommavilla, Diss. ETH Zürich nr 15688 (2004) 3 M. Glass-Maujean, S. Klumpp, L. Werner, A. Ehresmann, & H. Schmoranzer, Mol. Phys. 105,1535 (2007)
1

317

Initial Alignment and Commissioning of an Autocollimator Based Slope Measuring Profiler at the Advanced Light Source
Frank Siewerta, Valeriy V. Yashchukb, Jonathan L. Kirschmanb, Gregory Y. Morrisonb, Brian V. Smithb Helmholtz Zentrum Berlin für Materialien und Energie, Elektronenspeicherring BESSY-II, Albert-Einstein-Str. 15, 12489 Berlin, Germany b Lawrence Berkeley National Laboratory, One Cyclotron Road, Berkeley, California 94720, USA
a

Abstract
We describe the initial alignment of an autocollimator and beam guiding optical components on a carriage–traverse setup to be used as a slope measuring instrument for the inspection of highly precise reflective synchrotron optics at the LBNL/ALS Optical Metrology Laboratory. After the alignment, first commissioning measurements have been performed to characterize the new instrument. A discussion of the results will show the achieved performance and define the direction of future upgrade work to improve the instrument.

Introduction
The use of an autocollimator based slope measuring profiler to measure reflective optical surfaces was first proposed by Debler and Zander in 1979 [1]. Recent developments of the Nanometer Optical Component Measuring Machine (NOM) at BESSY [2, 3, 4] and the Extended Shear Angle Difference (ESAD) Instrument at the Physikalisch Technische Bundesanstalt (PTB) [5, 6, 7] demonstrated a new level of accuracy for slope measuring instruments. Both instruments achieved up to 10-fold superior measurement accuracy, compared to previous instruments, such as the Long Trace Profiler (LTP-II) [8, 9]. 2nd generation slope measuring profilers like the NOM or the ESAD-device are sophisticated engineering solutions combining highly accurate mechanical and motional parts, precise optical components and calibrated sensors. These instruments are operated under defined environmental conditions to guarantee excellent physical stability during the measurements [7, 10]. The Autocollimator- based Slope Measuring Profiler (ASMP)discussed here is a low budget realization of a NOM-like profiler using the Developmental Long Trace Profiler (DLTP) setup at the ALS-OML [11] and an autocollimator “Elcomat 3000 special” (Moeller Wedel Optical [12]) calibrated at the PTB.

Initial Alignment of the DLTP Optical Components
The inspection of reflective optical elements using a scanning pentaprism and an angle measuring sensor is a well known setup [2, 3, 13, 14, 15]. It prevents the measurement from being influenced by the pitch of the moving carriage as a first order error source. An optimal utilization of this setup requires the precise alignment of the autocollimator with respect to the beam guiding optics on the air bearing based carriage. Nevertheless, the accuracy of the carriage movement is an essential point for the achievable measurement accuracy. The test beam and moving direction of the carriage have to be very accurately aligned [14]. Otherwise the yaw movement of the carriage could cause cross-talk effects in the autocollimator. The DLTP setup model is shown in Fig. 1. The main mechanical parts like the carriage-traverse system and the alignment table for the autocollimator are mounted on an optical table. The complete setup is covered by a hutch to enable stable environmental conditions during a measurement. A home-made adjustable laser was used to align the autocollimator parallel to the moving direction of the carriage. A diaphragm at the carriage, mounted telecentric to the measuring axis of the autocollimator, was used for fine alignment. The projection of the diffraction

318

pattern from the diaphragm onto an image plane showed a stable image while moving the carriage over the entire range of motion of 900 mm.

Figure 1:

Mechanical setup of the Autocollimator based Slope Measuring Profiler (ASMP) at the ALS Optical and Metrology Laboratory

a

b

c

Figure 2:

Setup for the alignment of the autocollimator and pentaprism: a) view from autocollimator to the pentaprism, the image plane for detecting the diffraction pattern of the diaphragm is in the background, b) ghost reflections of the pentaprism at an image plane between pp and autocollimator, c) alignment laser, alignment plumb line and further image plane in the back.

319

Finally the beam guiding pentaprism was assembled on an alignment stage at the carriage (see Fig. 1). The ghost reflections of the front and reverse side of the pentaprism were used for the alignment, see Figure 2. When alignment is achieved both spots can be detected at the alignment plumb line independently from the current position of the carriage along the complete range of motion. A more sophisticated alignment procedure for a pentaprism, but not used for this first approach, was recently given by R. Geckeler [14].

Inspection of the DLTP Pentaprism
The quality of the beam guiding pentaprism is an accuracy limiting parameter for the performance of the instrument. The inhomogeneity of the glass, the flatness of the pentaprism faces and the angle errors of the faces will have a direct influence on the measurement. Thus, a range of five pentaprisms has been inspected by use of the ZYGO-GPI interferometer in the OML to identify the piece of best quality (see Figure 3). First a plane reference mirror of λ/40 (15 nm pv) shape accuracy was measured. After alignment of the pentaprism in the beam path the reference mirror was measured again.

Figure 3:

Set up at the ZYGO-GPI for inspection of penta prisms (left) and for measurement with the penta prism in orientation similar to the mounting at the instrument (right)

Figure 4:

ZYGO-GPI interferometer measurements, left: result for the plane reference mirror, height deviation = 15 nm pv, radii = -8.85 km and -14.58 km, right: result with pentaprism in the test beam height deviation 311 nm pv, radii = 0.35 km and -2.43 km

Then the pentaprism was mounted and inspected at an orientation similar to its position in the instrument (see Figure 3, right). Figure 4 shows the strong influence of the pentaprism quality on the measurement. Instead of 15 nm pv shape accuracy, measured for the reference mirror, a shape deviation of about 311 nm pv, more than one order of magnitude, was measured for the aligned pentaprism in the test beam. The values for the radii show a comparable characteristic.

320

Initial Commissioning
A long term stability scan over 10 hours was chosen for a first step evaluation of the instrumental setup. A plane mirror was placed at a distance of about 200 mm from the autocollimator. Directly after alignment of the mirror and closing the hutch at 6 pm a scan was run overnight. Figure 5 shows the signal of the X- and Y-angles (a strong drift, typical for the system immediately after alignment and a long term transient over the complete measuring time can be seen). In addition the Y-angle for the first 2-3 hours shows a higher noise level, probably caused by other equipment under operation close to the lab. Especially for the Y-angle an oscillation over a time period of about 12 minutes is clearly detected for the entire scan time. It shows a strong correlation to the switching-on/switching-off cycle of the lab air conditioner. Related to the required time for a line scan of about 5 – 30 min this oscillation would be a first order error source and needs to be eliminated by further improvement of the laboratory environment. In general, the noise level of the Y-angle is lower compared to the X-angle. It can be assumed that the pentaprism lessens the “pitch”-part of the random error influencing the measurement.

Figure 5:

Stability scan of the autocollimator showing the read out of Y- and X-angles. Variations due to sample alignment and temperature variations in the OML can be seen. The higher frequency oscillations with a period of about 12 min are correlated with the switching on/switching-off cycle of the laboratory air conditioner

The option of cross measurements of well known reference surfaces is a sure method to characterize a new measuring device [16, 17]. A spherical mirror R = 10m with a residual slope error of 1 µrad rms and 60mm in length was measured along its tangential center line by use of the BESSY-NOM. The substrate material is quartz glass with a gold coating of 30 nm thickness. The same line of inspection was scanned at the DLTP. The scan-line of interest was measured under different conditions, from mirror edge A to edge B and after a 180° rotation from edge B to edge A. Each measurement consisted of a group of 30 up to 34 line scans traced in the forward and reverse directions. A measuring point spacing of 0.2 mm was chosen to enable a higher angular resolution for an optimal line-array characterization of the autocollimator. All results were obtained by application of a calibration curve, the result of a highly accurate calibration procedure of the autocollimator at the PTB [18]. The air conditioner was running during two measurements (Fig. 6) and was switched off during two further measurements (Fig. 6). Like the stability scan, these measurements were performed overnight, starting at about 5 pm, and ending next day between 9 – 10 am. The results of these tests, shown in Figures 6 clearly demonstrate the need of a climate control system to guarantee thermal stability of the sample and instrument during a measurement series. It can be seen in Figure 6 that the system starts drifting after switching off the air conditioner. It takes about 6 hours to achieve a more or less stable state. In the morning around 8 – 9 am the overnight-achieved state of stability has disappeared by initial staff activity in the laboratory and heating up of the building. As in the stability scan, a higher level of

random error is identified for the measurement time until about 8-9 pm, see Fig. 6.

321

ba ccu on

ba ccu off

6 5 4

6 5 4

Slope [arcsec]

2 1 0 -1 -2 -3 -4 825 835 845 855 865 875

Slope [arcsec]

3

3 2 1 0 -1 -2 -3 -4 825 835 845 855 865 875

x-Position [mm]

x-Position [mm]

Figure 6: Left: Measurement from edge B (xB=825mm) to edge A(xA=878mm), air conditioner switched on, profile of residual slope for 30 single line scans forward and reverse trace. Right: Measurement from edge B (xB=825mm) to edge A xA=878mm), air conditioner switched off, profile of residual slope for 33 single line scans forward and reverse trace.

In general, the noise level at night is clearly lower, see again Fig. 6. When the air conditioner is off the 12-minute periodic error, identified in the stability scan, will not influence the measurement. But this advantage is canceled out by a strong drift in the temperature of the test mirror and the different components of the instrument. The level of random error of the DLTP is about 2-3 times higher than at the BESSY-NOM. Figure 7 shows 33/34 single line scans traced forward and reverse of the reference sphere by use of the DLTP and the NOM (with running climate control unit). The measurements were started 24 hours after alignment of the sample. Expressed as a 95% standard deviation (k=2) the reproducibility of the 33 DLTP measurements shown in Figure 10 is 0.143 µrad rms. The level achieved with the NOM is 0.053 µrad rms for a comparable set of 34 measurements.
2

Slope [arcsec]

1

0

-1

-2 825

835

845

855

865

875

x-Position [mm]

Figure 7: Comparison of BESSY-NOM (read) and ALS-DLTP (black) measurements (reference sphere R=10m, measurement from mirror edge B to edge A, climate control unit switched on), profile of residual slope for 33 single line scans forward and reverse traced, the waiting time before scan start was 24 hours after sample alignment. For a better visualization a shift of 0.8 arcsec is subtracted from the NOM-measurements.

Discussion of Results of First Measurement with the DLTP
Figure 8 shows the results of measurements obtained under different environmental conditions by use of the DLTP and the NOM. Both instruments yield comparable results for the radius of curvature and in part for the residual slope deviation, see also Table 1. The deviation for the measured radius of curvature values is between 0.006 and 0.04 %. Note that unlike the DLTP measurements presented here, the BESSY-NOM reference measurement was achieved by a stitching procedure of 4 times 4 sub-scans. For the reference measurement at the NOM, the well-characterized, linear view field of the autocollimator of ± 200 arcsec was applied. The reference curve is an average of 2 measurements (edge A to edge B and edge B to edge A). Compared to the NOM measurements, a higher random error is observed for the DLTP measurements. The excellent agreement for the measured radius of curvature is a strong indication of the effective application of the calibration curve generated by the PTB. The subtraction of a measurement, traced
322

from mirror edge A to edge B, by a measurement traced from edge B to edge A, enables an estimation of the achieved measurement uncertainty and partial identification of the systematic error. Note this method is applicable only if the same x-travel range of the scanning carriage is used and for both scans the slope is detected by the same field of view of the autocollimator. Figure 9 shows the difference curves for the measurements with the DLTP for the air conditioner switched on and off, (shown in Fig 8) and for measurements at the NOM (air conditioner switched on). The differential curves of the DLTP measurements clearly displays systematic error structures in the profile (see Figure 9). Note that a significant part of the systematic error has a periodical (oscillation-like) structure. That is probably a result of a limited performance of the calibration of the autocollimator at a stationary baseline distance. While the standard working state (air conditioner switched on) shows an error level of about 0.5 µrad rms, a slightly lower level of 0.4 µrad rms is achieved for the switched off state. For additional observations see also Figure 11. Subtracting the reference curve, the residual slope profile of the single measurements shows a slightly lower noise level for the switched off state. But the gained improvement for the switched off state is marginal at best. Systematic error sections at both curves are specially labeled. Subtracting the curves for the switched on and off states, the rms value for the achieved deviation profile can be interpreted (very carefully) as the level of random error of the instrument (see Figure 10).
Test Sphere R=10m - profile of residual Slope
6 5 4 3 Slope [µrad] 2 1 0 -1 -2 -3 -4 5 15 25 35 x-Position [mm] 45 55

0

BESSY - Reference Measurement BESSY A-B BESSY B-A ALS A-B, Climate Control Unite on ALS A-B, Climate Control Unite off ALS B-A, Climate Control Unite on ALS B-A, Climate Control Unite off

Figure 8: Reference sphere R=10m, profile of residual slope, for different measurements with the BESSYNOM and the ALS-DLTP. With both instruments the line of interest was measured from edge A to edge B and after realignment from edge B to edge A. The measurements at the ALS were taken with the Climate Control Unit switched both on and off

Figure 12 shows the PSD spectra of the measurements on the 10m test mirror performed with the ALS-DLTP and the BESSY-NOM. Both spectra have very similar high spatial frequency behavior that can be expressed with an inverse power law of the second order. Such behavior is characteristic for a random noise related to air turbulence. Moreover, the same inverse power law describes the spectra of slow drifts.

Figure 9: Residual error budget for different measurements with the ALS-DLTP and the BESSY-NOM. The measurements at the ALS were taken with Climate Control Unit switched both on and off.

323

Figure 10: Difference of two measurements with the ALS-DLTP. The difference can be interpreted very carefully as the achieved level of random error for the ASMP of 0.34 µrad rms. Residual Slope [µrad rms] BESSY NOM Reference Curve BESSY NOM A-B BESSY NOM B-A ALS DLTP A-B (CCU on) ALS DLTP B-A (CCU on) ALS DLTP A-B (CCU off) ALS DLTP B-A (CCU off) Table 1: 1.02 1.05 0.97 1.25 1.27 1.22 1.22 Radius of Curvature [m] 9.9609 9.9614 9.9603 9.9652 9.9645 9.9626 9.9642

Measurement results obtained by use of the ASMP and the NOM, for a reference sphere, R=10m, scan length: 50 mm, applied step size: 0.2 mm

The inspection of highly accurate plane mirrors characterized by residual slope errors below the 0.25 µrad rms limit is a specific challenge for metrology. In contrast to the measurement of curved surfaces, a very small field of view of the CCD-array of ± 2 µrad or less is used for the measurement. Thus, the error budget strongly depends on pixel errors and the local homogeneity of the CCD-array in general. Furthermore, vibrations and air-turbulence will limit the achievable accuracy. Figure 13 shows eight scans (blue curves) obtained by DLTP measurements on a super polished plane mirror for the MERLIN-beamline at the ALS [19]. The NOM measurements (red curves) were performed on a 310 mm long super polished mirror for beamline UE46 at BESSY [20]. In both cases the climate control unit was switched on and the measurements were started more than 24 h after alignment. The DLTP measurement shows a 3 times higher noise level than the NOM result. It can be assumed that this is mainly caused by the high level of vibration in the lab and the thermal cycling of the air conditioner. The achieved measurement result for the ASMP of 0.04 arcsec rms (0.2 µrad rms) is dominated by random noise. In this case pixel errors and the local homogeneity of the CCD-array are not relevant to the accuracy achieved. The inspected mirror under test is more accurate than the instrument used. The reproducibility of the 8 DLTP measurements is 0.06 arcsec rms (0.3 µrad rms). This result is obtained

Figure 11: Profile of residual error and rms values for different measurements with the DLTP and the NOM after subtraction of the NOM reference measurement. For the BESSY-NOM measurements the Climate Control Unit was on.

324

when a 0-order fit is used to correct, in part the drift on the measurement. The 8 NOM measurements have a reproducibility of 0.02 arcsec rms (0.1µrad rms) for a 70 mm scan length.

Figure 12:
0,5 0,4 0,3 0,2 0,1 0 -0,1 -0,2 -0,3 -0,4 -0,5 0

PSD of residual slope data for different measurements on the R=10m reference sphere

a

b

- BESSY-NOM, 8 scans - Average of 8 scans

Slope [arcsec]

- ALS-ASMP, 8 scans - Average of 8 scans
20 40 60 80 100 120 140

x-Position [mm]

Figure 13: Measurements on plane substrates. Profiles of residual slope, a) 8 scans traced with the DLTP blue curves left and b) 8 scans traced with the NOM red curves right. The reproducibility is 0.06 arcsec rms, after drift correction, for the DLTP and 0.02 arcsec rms for the NOM. Both test substrates are chracterized by a residual slope of better than 0.04 arcsec rms (0.2 µrad rms).

Conclusions and Outlook
The instrumental set up of a Developmental Long Trace Profiler1 assembled, preliminary aligned, and inspected in June 2008 at the ALS enables the slope metrology of synchrotron optics with an estimated measurement uncertainty of about 0.5 µrad rms for significantly curved optics. A lower measurement uncertainty in the range of 0.25-0.3 µrad rms can be achieved in the case of plane optical elements. It was shown that the use of calibration curves gained by calibration of the autocollimator at the PTB provides an excellent agreement with measurement results of the BESSY-NOM. Comparing the results for the measured radius of curvature, of a R = 10 m sphere, a deviation of less than 0.04 % is observed. It was shown that the method of curve correction by calibration is applicable in general. Some fitting errors in the software used have been observed during the commissioning (see Figure 6), and will be debugged in the near future. The achieved performance of the instrument can be improved by further upgrade of the laboratory environment. It is to be noted that the measuring device, data acquisition method, and the surrounding environment have to be understood as one unit of operation. The air conditioning controller located close to the DLTP should be replaced by a more advanced system. A new climate control unit should be placed at a larger distance from the metrological devices in the laboratory, to avoid interfering parasitic effects. The limited quality of the beam guiding optics (pentaprism) was shown. It can be

325

assumed that a part of the shown systematic error (see Figure 9 and 11) is caused by the inhomogeneity of the pentaprism. It can be easily replaced by a double mirror as done at the NOM. A more accurate traverse and stone pillars instead of metal parts, as currently in use, will improve the physical stability of the instrument. An upgraded hutch to cover the measuring set up will provide better shielding against environmental instability. However, better shielding would significantly increase the characteristic time of the system that describes the temperature transition process after opening/closing the hutch. For an easier alignment of optical components a more robust/stable alignment laser would be helpful. A topic of the future work is the high precision characterization and calibration of the instrument. Highly accurate metrology devices have to be calibrated for the specific conditions during a measurement. Thus, the development of dedicated calibration tools is mandatory. A first set up of a vertical angle comparator has been realized at BESSY [21] and shown to be reasonable in its use for additional calibration of the NOM. A more sophisticated design of Universal Test Mirror (UTM) was proposed recently [22] and is under development by a cooperation of the ALS, BESSY and the PTB. The presented results can not provide a complete characterization of the instrument. Further investigation is required. Additional measurements of well known reference optics, provided by the Global high accuracy Round Robin (GRR) [17] cooperation are recommended. Stability scans for different pentaprism-to-autocollimator distances as well as variations of the diaphragm size and position are recommended. Using the autocollimator the straightforward movement of the carriage system can be characterized.

Recent Developments
After the first DLTP setup was assembled the movable pentaprism was precisely re-aligned by R. Geckeler (PTB) [23] using his original procedure for alignment [14]. The positive effect of the realignment [24, 25], was demonstrated by measuring a spherical mirror (R=15 m) supplied by InSync. Inc. In addition, the application of optimal scanning strategies enables a highly accurate metrology by use of the DLTP [25], (related to scanning strategies see also [26]). Finally, the high performance of the ALS-DLTP after the last upgrade [27] was shown by measuring the GRR reference sphere S3 (R=1280 m) [28]. The result of the measurement is shown in Fig. 14. The resulting residual slope traces obtained with the DLTP and the NOM are of excellent coincidence with an absolute deviation of less than 0.1 µrad. A further cross check, of DLTP and NOM by use of the 15 sphere is in preparation. Note: this test will cover a much larger view filed of the sensors than applied for the S3 sphere.

Figure 14: Profile of residual error and rms values for different measurements with the DLTP and the NOM after subtraction of the NOM reference measurement. For the BESSY-NOM measurements the Climate Control Unit was on.
Note the DLTP does not use a LTP-head for performing slope measurements. The used sensor is an autocollimator made by Moeller Wedel GmbH
1

326

Acknowledgements
The authors are grateful to Thomas Zeschke and Gerd Reichardt (BESSY), Ralf D. Geckeler and Andreas Just (PTB), Samuel Barber, Tony Warwick, Wayne McKinney and Howard Padmore (ALS), for useful discussions. F.S. wishes to particularly acknowledge the Advanced Light Source, and its staff, for their collaboration, kind hospitality, and encouragement during his visit of May-June 2008, when most of this work was accomplished, and measurements where taken with the DLTP setup. The Advanced Light Source is supported by the Director, Office of Science, Office of Basic Energy Sciences, Material Devision, of the U.S. Departement of Energy under Contract No. DE-AC02-05CH11231 at Lawerence Berkeley National Laboratory.

References:
1 E. Debler, K. Zander, Ebenheitsmessung an optischen Planflächen mit Autokollimationsfernrohr und Pentagonprisma, PTB Mitteilungen Forschen + Prüfen, Amts und Mitteilungsblatt der Physikalisch Technischen Bundesanstalt, Braunschweig und Berlin, 1979, pp. 339-349 Frank Siewert, Tino Noll, Thomas Schlegel, Thomas Zeschke, and Heiner Lammert, The Nanometer Optical Component Measuring machine: a new Sub-nm Topography Measuring Device for X-ray Optics at BESSY, AIP Conference Proceedings 705, American Institute of Physics, Mellville, NY, 2004, pp. 847-850 H. Lammert, T. Noll, T. Schlegel, F. Siewert, T. Zeschke, Optisches Messverfahren und Präzisionsmessmaschine zur Ermittlung von Idealformabweichungen technisch polierter Oberflächen, Patent No.: DE 103 03 659 (28 July 2005) F. Siewert, H. Lammert, T. Zeschke, The Nanometer Optical Component Measuring Machine; in: Modern Developments in X-ray and Neutron Optics, Springer 2008 R.D. Geckeler, I. Weingärtner, Sub-nm topography measurement by deflectometry: flatness standard and wafer nanotopography, Proc. of SPIE, 4779, Bellingham, WA, 2002, pp. 1-12 Jens Illemann and Matthias Wurm, Deflectometric Measurements of Synchrotron-Optics for Postprocessing, AIP Conference Proceedings 705, American Institute of Physics, Mellville, NY, 2004, pp. 843-846 Ralf D. Geckeler, ESAD Shearing Deflectometry: Potentials for Synchrotron Beamline Metrology, Proc. of SPIE, 6317, Bellingham, WA, 2006, P. Takacs, S. N. Qian and J. Colbert, Design of a long trace surface profiler, Proc. of SPIE, 749, Bellingham, WA, 1987, pp. 59-64 P. Takacs, S. N. Qian, Surface Profiling interferometer, US patent No.U4884697, Dec. 5, 1989 F. Siewert, H. Lammert, T. Noll, T. Schlegel, T. Zeschke, T. Hänsel, A. Nickel, A. Schindler, B. Grubert, C. Schlewitt, Advanced metrology: an essential support for the surface finishing of high performance x-ray optics, in Advances in Metrology for X-Ray and EUV Optics edited by Lahsen Assoufid, Peter Z. Takacs, John S. Taylor, Proc. of SPIE, Vol. 5921-01, Bellingham, WA, 2005, J. Kirschmann, E. E. Domning, G. Y. Morrison, B. V. Smith, V.V. Yashchuk, Precision Tiltmeter as a Reference for Slope Measuring Instruments, Scholarship Repository, Univ. of California , Berkeley 2008 http://www.moeller-wedel-optical.com/El-Autocolimators/G_Elcomat3000.htm Shinan Qian, Werner Jark, Peter Z. Takacs, The penta-prism LTP: A long-trace-profiler with stationary optical head and moving penta prism, Rev. Sci. Instrum. 66 (3), March 1995 R.D. Geckeler, Optimal use of pentaprisms in highly accurate deflectometric scanning, Meas.Sci.Technol., 18(2007), pp. 115-125 J. Susini, R. Baker and A. Vivo, Optical metrology facility at the ESRF, Rev. Sci. Instrum. 66 (2), February 1995 A. Rommeveaux, M. Thomasset, D. Cocco, F. Siewert, First report on a European round robin for slope measuring profilers, Proc. of SPIE, 5921-01, Bellingham, WA, 2005 F. Siewert et al., Global High Accuracy intercomparison of Slope Measuring Instruments, AIP Conference Proceedings 879, American Institute of Physics, Mellville, NY, 2007, pp. 706-709 R.D. Geckeler, I. Weingärtner, A. Just, R. Probst, Use and traceable calibration of autocollimators for ultraprecise measurement of slope and topography, Proc. of SPIE, Vol. 4401, Bellingham, WA, 2001, pp. 184-195 MERLIN (Mili-Electron-volt Resolution beamLINe) project developed by the ALS Scientific Support Group, Beamline 4.0.1: www-als.lbl.gov/als/techspecs/ bl4.0.1.html F. Siewert, BESSY-Meßprotokoll für planen Spiegel M2 – UIE46XM-Beamline, Berlin, 2007 F. Siewert, Calibration: Calibration and Autocalibration, 3rd International Workshop on Metrology for X-ray Optics, Daegu, Korea, June 2006 V. Yashchuk, W. McKinney, T. Warwick, T. Noll, F. Siewert, T. Zeschke, R. Geckeler, Proposal for a Universal Test Mirror for Characterization of Slope Measuring Instruments, Proc. of SPIE, Vol. 5856-121, 2007 S. Barber, R.D. Geckeler, Tony Warwick and V. Yashchuk, Optimal alignment of Pentaprism in DLTP, LSBL Note LSBL-924 (September 12, 2008) V. Yashchuk, J. Kirschmann, G. Y. Morrison, B. V. Smith, E. E. Domning, F. Siewert, T. Zeschke, R.D. Geckeler, A. Just, Autocollimator based slope measuring profiler (a la BESSY NOM) under development at the ALS OML , LSBL-920 (June 26, 2008) V. Yashchuk, Optimal Scanning Strategies for Slope Measuring Profilers, (in preparation) H. Lammert, Subnanometer Accuracy Measurements with the LTP at BESSY, Int. Workshop on Metrology for Xray and Neutron Optics, Chicago 2000 V. Yashchuk, DLTP Performance Test with a 1280-m Spherical Reference Mirror, LSBL Note LSBL-931 (January 12, 2009)

2

3

4 5 6 7 8 9 10

11 12 13 14 15 16 17 18 19 20 21 22 23 24

25 26 27

327

Time-of-flight electron spectroscopy in multi-bunch mode
Johannes Floß,1 Marko Förstel,2,3 Toralf Lischke,2 Tiberiu Arion,2 Melanie Mucke,2 Brandon Jordon-Thaden,3 Andreas Wolf 3 and Uwe Hergenhahn2,a
1: Freie Universität Berlin, Institut für Chemie und Biochemie, Takustr. 3, 14195 Berlin 2: Max-Planck-Institut für Plasmaphysik, EURATOM association, Boltzmannstr. 2, 85748 Garching 3: Max-Planck-Institut für Kernphysik, Saupfercheckweg 1, 69117 Heidelberg

Electron spectroscopy carried out via a time-of-flight (TOF) measurement offers some advantages compared to the use of a conventional dispersive spectrometer. Most notably, the parallel acquisition of a band of electron energies can be accomplished. Additionally, by a suitably shaped magnetic field, a large solid angle acceptance of almost 4π sR is achievable ('magnetic bottle spectrometer', [1,2]). The times-of-flight of the electrons for typical instrument dimensions range from some ns to some µs. The use of a pulsed light source with a low repetition rate, such as a laser or a synchrotron radiation source running in single-bunch mode, therefore seemed a necessity for this type of spectroscopy. In this report we give first results from an alternative approach, in which the time structure of the fill pattern of BESSY is removed from the TOF spectra by a deconvolution procedure. A typical BESSY fill pattern, as measured by time-correlated single photon counting [3], is shown in Fig. 1. Mathematically the deconvolution of such functions from a signal recorded as a function of time is an exact operation. In practice however the inevitable noise on any experimental data renders a straight-forward inversion of the convolution impossible. Some strategies used to overcome these problems are frequency filtering in Fourier space, deconvolution guided by the maximum entropy principle and least squares analysis of the measured data using a parametrized form of the underlying deconvoluted signal ('peak fitting'). In our experiment, we have measured the TOF photoelectron spectra of He and Ne using a magnetic bottle spectrometer (Fig. 2) at the TGM4 beamline during multi-bunch operation. For analysis, spectra were modelled by 1-4 Gaussian peaks of variable width and position, which were convoluted by the fill pattern recorded simultaneously by the BESSY control system used to monitor the hybrid bunch position. The parameters of the Gaussian curves were adapted to optimally reproduce the measured TOF spectra (Fig. 3). Results for the photoelectron spectrum of Ne excited with 110 eV photons are shown in Fig. 3. At this photon energy, the spectrum is dominated by photoionization of the 2p main line. Excitation of this feature by the hybrid bunch can be visually distinguished. It is therefore most interesting whether the deconvolution algorithm reveals any further photoelectron lines, such as the 2s inner valence line. The deconvoluted spectrum (Fig. 3 right hand side panel) does show a second, intense photoelectron line, which after time-to-energy conversion (not shown) appears however at a binding energy of 73 eV. This somewhat unexpected result can be explained by a Xe contaminant in the gas inlet system from a preceding experiment – this gas gives rises to an intense photoelectron line from ionization of the 4d orbitals, with reference binding energies of 67.5 and 69.5 eV. These first results show that deconvolution TOF spectroscopy in multi-bunch mode has some potential, which we will explore further by comparison of identical spectra measured in multi and single bunch.
[1] [2] [3] P. Kruit and F. H. Read, J. Phys. E 16, 313 (1983). P. Lablanquie, L. Andric, J. Palaudoux, U. Becker, M. Braune, J. Viefhaus, J. H. D. Eland, and F. Penent, J. Electron Spectrosc. Relat. Phenom. 156-158, 51 (2007). K. Holldack, M. v. Hartrott, F. Hoeft, O. Neitzke, E. Bauch, and M. Wahl, in Advanced Photon Counting Techniques II, edited by W. Becker (SPIE, 2007), Vol. 6771, p. 677118.

a

Mail address: IPP, c/o BESSY, Albert-Einstein-Str. 15, 12489 Berlin; email: uwe.hergenhahn@ipp.mpg.de

328

Intensity (arb. units)

200

210

220 x10

-9

0

200

400 Time (s)

600

800x10

-9

Figure 1: Fill pattern of BESSY in multibunch operation, recorded with an avalanche photodiode (APD). The inset of the figure shows the 2 ns spacing of the individual photon bunches. Here, the true pattern is broadened by the APD risetime of approx. 100 ps.

Figure 2: Sketch of the magnetic bottle time-of-flight spectrometer. The synchrotron radiation direction, gas jet and central axis of the spectrometer are all perpendicular to each other. In the set-up actually employed, the spectrometer was mounted vertically pointing upward.

Figure 3: Time-of-flight spectrum of Ne photoelectrons at 110 eV kinetic energy (left panel) and deconvoluted structure consisting of four Gaussians (right panel).
Partial funding by the Deutsche Forschungsgemeinschaft, the Advanced Study Group of the Max-PlanckGesellschaft, the BMBF and the Fonds der Chemischen Industrie is gratefully acknowledged.

329

The new BESSY TXM for cryo-tomography and nano-spectroscopy
P. Guttmann, S. Heim, S. Rehbein, S. Werner, G. Schneider Helmholtz-Zentrum Berlin für Materialien und Energie, Elektronenspeicherring BESSY II, Albert-Einstein-Str. 15, 12489 Berlin

1. Introduction The BESSY x-ray microscopy group has developed a new full-field transmission x-ray microscope (TXM) which employs an advanced x-ray optical concept [1]. In our new BESSY microscope, a standard monochromator beam line provides a high energy resolution of up to 10,000 which permits NEXAFS studies. An elliptically shaped mono-capillary is used to form the hollow cone illumination necessary for sample illumination and to match the aperture of the objective. The new TXM also significantly improves the possibilities to study cryogenic or heated samples for life and materials sciences. Taking advantage of the developments over the last decades in cryo electron tomography, the sample stage of the TXM is an adaptation of a stateof-the-art TEM stage. Therefore, the microscope provides the same excellent mechanical accuracy and temperature stability known from cryo TEM, a tilt range of ±80° and lateral travel ranges of ±1mm with a bidirectional reproducibility <15 nm. The translation axes are mounted with respect to the rotation axis in order to make it possible to establish eucentricity for any part of the specimen, i.e. bring any point of the sample onto the spatially fixed tilt axis. Together with the high photon flux and the fully automated operation of the microscope's control software, the acquisition time for a full tomography tilt series is reduced to less than 20 minutes.

2. The way towards sub-10 nm real space x-ray imaging 2.1 Improvements of the condenser The parameters of the ellipsoidal capillary are determined by the beam divergence of the monochromator and by the desired aperture matching to the high resolution objective of the microscope. Ray tracing calculations show that the slope error of the capillary should be below 100 µrad and the alignment accuracy in the tilt angle of the capillary with respect to the incoming beam should be better than 100 µrad [2]. A single-bounce ellipsoidal glass capillary fabricated and evaluated by optical measurements during and after fabrication by Xradia was applied [3]. The resulting slope error of 80 µrad is well below the limit calculated by ray tracing. The x-ray performance of the capillary was tested with the BESSY x-ray microscope at the U41-FSGM beam line at a photon energy of 510 eV. Firstly, the capillary was adjusted to get the smallest possible focus. Figure 1a shows an x-ray image of the focus which is in size 620 nm x 990 nm (FWHM-values). Secondly, a Siemens star test pattern with structure sizes down to 25 nm was used to demonstrate the imaging capabilities of the x-ray microscope using a capillary as condenser. To obtain a large homogeneously illuminated object field, the condenser is helically scanned. Figure 1b shows the x-ray image of this test pattern using a micro zone plate objective with 30 nm outermost zone width.

330

a)

b)

Fig. 1: a) Focus of the capillary imaged by a micro zone plate at E/E 9000, 510 eV, 100 ms exposure time. b) Test pattern imaged by a micro zone plate (both fabricated at BESSY by S. Rehbein, S. Werner) at E/E 9000, 510 eV, 2 s exposure time with scanned condenser. To analyze the performance of the new condenser, quantitative measurements were performed to measure the efficiency of the capillary. The focusing efficiency of the capillary was measured to be 80% for a photon energy of 510 eV. Note, this value exceeds the efficiency of zone plate condensers by nearly an order of magnitude. 2.2 Developments of the x-ray objective Fresnel zone plates are the key optical elements for soft and hard x-ray microscopy. The resolving power of zone plates scales with the order of diffraction. By employing high orders of diffraction, it is possible to increase the resolution without the need of manufacturing smaller outermost zone widths. In order to demonstrate the resolution achievable with high order imaging, we used a multi-layer test sample with B4C/Cr lines and spaces of 5 different widths down to 7.8 nm. Figure 2 shows the result of a resolution test experiment performed with the BESSY TXM at 700 eV photon energy with a 25nm zone plate operating in the 3rd order. Note that lines and spaces of 14.3nm are clearly resolved while the 11.0nm structures still show some modulation.

Fig. 2. B4C/Cr multilayer test structure imaged in the 3rd diffraction order using a drN = 25nm zone plate, E = 700 eV, E/ E =1732, texp = 15.7 s

331

Until now no sub-20 nm (half-pitch) resolution x-ray images were presented in the literature from extended relevant samples. To demonstrate the improved resolution in the 3rd order of diffraction for relevant samples, an advanced copper interconnect stack which was thinned by FIB milling to a thin lamella of about 1.5 µm thickness was imaged at E = 900 eV. In this image the cut-off frequency corresponds to a feature size of about 12 nm which is in good agreement with the obtained resolution demonstrated with the multilayer structures [4]. For short exposure times and minimum radiation load of the specimen the diffraction efficiency of the zone plate objectives has to be maximized. As the efficiency strongly depends on the height of the diffracting zone structures the achievable aspect ratio of the nanostructures determines these limits. To reach aspect ratios 20:1 for high efficient optics we superimposed zone plates on top of each other [5]. With this approach the final aspect ratio is only limited by the number of stacked zone plate layers. For the stack process several nano-structuring process steps have to be developed and/or improved [6]. In first experiments two layers of nickel zone plates were stacked on top of each other. The height of the single layers was about 160 nm corresponding to a total height of 320 nm for the stacked structures. After electroplating of the second layer, a trench was cut into the nanostructures by FIB to allow a look at the cross section (thanks to Y. Ritz and D. Chumakov, AMD Dresden). The stacked zone plate with an outermost zone width of 30 nm is shown in Fig. 3. A mismatch of about 30 nm between the first and the second zone plate layer is observed. The second layer has not been polished yet and is still slightly overplated. The required overlay accuracy is 2drn/7 to avoid a degradation of the zone plate efficiency and resolution [7]. This means our achieved overlay accuracy has to be improved in order to fabricate efficient high resolution zone plates. Recently, an overlay error of about 2 nm was reported [8] showing that a sufficient accuracy is possible with advanced e-beam lithography. For this reason a new stateof-the-art e-beam writer is ordered and will be setup in the second semester of 2009. The polished zone plates are fabricated on a thick substrate due to the mechanical stability for the Chemical Mechanical Polishing (CMP) process step. For using this zone plates as x-ray optics the substrate has to be thinned afterwards by wet chemical etching. A process to protect the structures during the backside etching is under development.

20 µm

(a)

500 nm

(b)

Fig. 3: SEM micrograph of a stacked nickel zone plate with 30 nm outermost zone width. (a) Overview with cross marks and FIB produced trench. (b) Stacked zones are slightly overplated.

332

3. Cryo-tomography of frozen-hydrated mammalian cells Full-field x-ray microscopy has proven to be an excellent tool for 3-D studies of whole frozen hydrated biological samples [9, 10, 11]. However, tomography of flat grown mammalian cells was not possible as they do not fit into glass capillary sample holders. Our new TXM overcomes this severe limitation, because its optical concept requires no monochromator pinhole close to the object plane. In the new TXM, it is possible to study adherent cells grown on TEM grids. As an example, Fig. 4 shows one slice of a tomography reconstruction of a frozen-hydrated mouse cell imaged at 510 eV. The voxel size is 9.8 nm, which is also the thickness of the slice. The nuclear membrane with its double structure and internal structures of the nucleus are visualized.

Fig. 4: Slice through the reconstructed local linear absorption coefficient of the nucleus of a mouse adenocarcinoma cell; 131 projections recorded at E = 510 eV.

4. Stress-migration in advanced copper interconnect structures Besides electromigration in semiconductor devices, stress migration is one of the most important failure mechanisms in advanced copper interconnects. As a result of lattice vacancy migration due to local mechanical stress at the via/plate interface small holes called voids form. However, the void formation mechanism is not fully understood. X-ray microscopy permits to study dynamic void evolution during stress migration. An exemplary micrograph from the stress-migration studies is shown in Fig. 5. This slice through the reconstructed volume of a 1 µm thick IC lamella shows that the void is located in the copper interconnect line directly below the via. Depending on their size, larger voids lead to unacceptable resistance increase and ultimately cause open circuit or other circuit malfunction.

333

Fig. 5: Slice from a tomographic reconstruction of a stress migration sample imaged at E = 700 eV, visualizing the location of the void close to the via.

5. Nano-spectroscopy of TiO2 particles The spectromicroscopy capabilities of the new TXM are demonstrated with TiO2 particles. An image stack was taken in the energy range from the Ti L-edge up to the oxygen K-edge with an energy resolution of E/ E 7214 (Fig. 6). Note the combined excellent spectral and spatial resolution in a large field of view.

d)

Ti 2p3/2

Ti 2p1/2

O 1s

Fig. 6: a) - c) Selected micrographs of TiO2 particles imaged at different photon energies. d) Spectrum at the position of the green dot in a)-c) for the whole energy stack (440 - 550 eV, E/ E = 7214, total 380 images).

334

Acknowledgements The authors are greatly indebted to S. Braun (IWS, Dresden) and Y. Ritz (AMD, Dresden) for the fabrication of the multi-layer test sample and W. Müller (NIH, Bethesda) for the cell preparation. This work was supported by the EU in the 6th framework program under contract number RII3-CT-2004-506008, the Human Frontier Science Program (HFSP) Research Grant Ref. RGP0053/2005-C and the German Federal Ministry of Education and Research (BMBF) under contract number 05KS4BY1/7.

References 1. P. Guttmann, S. Heim, S. Rehbein, S. Werner, B. Niemann, R. Follath, G. Schneider: X-ray microscopy at the new U41-FSGM beam line, BESSY Annual Report 2007 (2008), 297 - 300 2. P. Guttmann, X. Zeng, M. Feser, S. Heim, W. Yun, G. Schneider: Ellipsoidal capillary as condenser for the BESSY full-field x-ray microscope, submitted to XRM 2008 proceedings 3. Zeng X, Duewer F, Feser M, Huang C, Lyon A, Tkachuk A, Yun W., Ellipsoidal and parabolic glass capillaries as condensers for x-ray microscopes, Appl Opt. 2008 May 1; 47(13):2376-81 4. S. Rehbein, S. Heim, P. Guttmann, S. Werner, and G. Schneider: Ultrahigh-Resolution XRay Microscopy with Zone Plates in High Orders of Diffraction, submitted 5. S. Werner, S. Rehbein, P. Guttman, S. Heim, G. Schneider: Towards stacked zone plates, submitted to XRM 2008 proceedings 6. S. Rehbein, P. Guttmann, S. Werner and G. Schneider, J. Vac. Sci. Technol. B 25 (1789), 2007 7. G. Schneider, Appl. Phys. Lett. 73, 599 (1998) 8. W.Chao, B. D. Harteneck, J. A. Liddle, E. H. Andersson and D. T. Attwood, Nature (London) 435, 1210 (2005) 9. Schneider G, Ultramicroscopy 75 (1998) 85 - 104 10. Weiss D, Schneider G, Niemann B, Guttmann P, Rudolph D, Schmahl G, Ultramicroscopy 84 (2000) 185 - 197 11. Schneider G, Anderson E, Vogt S, Knöchel C, Weiss D, LeGros M and Larabell C, Surf. Rev. Lett. 9 (2002) 177 - 183

335

Project title:  X­ray microCT study of ex vivo engineered heart tissue  Objectives (max. 200 words)   The aim of this study is to obtain 3D visualization of spatial distribution of injected neonatal rat  cardiomyocytes (labelled cells) inside the heart tissue of infracted mice.  Achievements (max. 500 words)  The microCT was used to image and characterize 3D distribution of injected cells plated on  collagen scaffolds inside the heart tissue of infracted mice (Fig. 1a).  3D visualization of the  spatial distribution of the grafted cells in respect with the host myocardium, veins and capillary  system was obtained (Fig. 1b). In particular, the X‐ray absorption of the labeled cells by  magnetic iron oxide nanoparticles was higher than the other heart tissues, allowing their  visualization as bright spots in the 2D images (Fig.  1c). These slice images were compiled and  analyzed to render 3D images and to obtain a better visualization of cell distribution within the  samples. 3D visualization of the spatial distribution of the grafted cells is shown on Fig. 1d.  

  Fig. 1. MicroCT image of the infracted heart (a). 3D image of the spatial distribution of the  grafted cells in respect with the host myocardium, veins and capillary system (b). 2D original  slice (c). 3D visualization of the spatial distribution of the grafted cells (d).   
1   
336

This result is of special importance that after injection cell population is manly heterogeneous  distributed at earlier stage as well as was located in one area. The 3D images so obtained  constituted a very innovative progress, as compared to the usual 2D histological images, which  do not provide the correct position of the cardiomyocytes within the heart. Moreover, this is  potentially interesting for future research on determining time variable (early and late  differentiation stage), because through the microCT we hope to be able to observe in 3D the  migration of cells with respect to the cardiac vessel, with important structural details not  observable by the conventional bidimensional imaging techniques. However, we should note  that due to a complicate nature of the investigated object the data treatment is still in progress.   
   

2   
337

Morphological Characterization of Articular Cartilage on the Cellular Level

Rolf Zehbe1, Christoph Brochhausen2, Astrid Haibel3, Franziska Schmidt1, Ulrich Gross4, C. James Kirkpatrick2, Heinrich Riesemeier3, Helmut Schubert1
1

Technische Universität Berlin, Institute of Materials Science & Technologies, Englische Strasse 20, 10587 Berlin, Germany 2 GKSS Research Centre at DESY, Petra III, Notkestrasse 85, 22607 Hamburg, Germany 3 BAM, Federal Inst. for Materials Research & Testing, Division I.3, Richard-Willstätter-Strasse 11, 12489 Berlin, Germany 4 Charité University Medicine Berlin, Biomaterials Laboratory, Aßmannshauser Strasse 4-6, 14197 Berlin, Germany 5 Johannes Gutenberg University, Institute of Pathology - REPAIRlab, Langenbeckstrasse 1, 55101 Mainz, Germany

1

Corresponding author: Rolf Zehbe (rolf.zehbe@tu-berlin.de)

Introduction: Articular cartilage covers the end of bones allowing their relative movement. It forms a smooth, glassy tissue, also known as hyaline cartilage, showing a characteristic zonal structure of the extracellular matrix components and specialized cells, the chondrocytes. Depending on the localization, the chondrocytes are either stretched along the surface (superficial zone) or are shaped elliptical or rounded in the middle zone, where some cells form isogeneous cell groups. Chondrocytes are generally surrounded by a low density fluidic region called lacuna. In deeper regions the extracellular matrix proteins become more perpendicularly oriented to the surface. Further below, calcification increases, which is indicated by the so called tide mark. Ultimately, this calcification results in the formation of the subchondral bone. Unlike the cartilagineous tissue region, the subchondral bone is vascularised and varies in cell types (osteoblasts, osteoclasts, osteocytes, fatty cells and cells attributing to the nervous system). The tissue morphology is well investigated via conventional means, including light microscopy and electron microscopy. Unfortunately, due to the opacity of the tissue for both electrons and optical wavelength, it is only possible to investigate thin tissue slices. To reveal the true 3D morphology, it is therefore
338

required to section and register multiple slices, which is prone to errors due to sectioning losses and sectioning artefacts (1).

Figure 1: Experimental measurements

setup

for

SR-µCT

Contrary to electrons and optical wavelength, X-rays show only marginal refraction, resulting in poor absorption but high penetration in biological tissues. It is therefore possible to image the true 3D morphology by recording projection images and mathematically transform this data into subsequent slice images (2, 3, 4 & 5). Although the resolution offered by X-ray experimental setups and detector systems is sufficient to image structural details down to several nanometers in theory, conventional X-ray tube setups have beam properties not fit well to image biological tissues (e.g. non-monochromaticity, non-

coherence and low photon flux). Even with synchrotron experimental setups, biological tissues are difficult to image due to the very low differences of structures in absorption contrast. Therefore, two different approaches exist; first, elevating the absorption contrast by labelling tissue structures with metals as is common in electron microscopy and second, attenuating the phase contrast which can be achieved with a coherent X-ray beam (6, 7). While staining might introduce artefacts due to non homogeneous diffusion of the stain or later during data reconstruction (8), the use of phase contrast imaging requires an experimental setup with a highly coherent beam. These experimental requirements are met by the BESSY BAMline setup, which is displayed in Figure 1 and consists of a high precision sample stage (rotation table combined with translation tables), with a sample holder, a scintillator and subsequent light optics with a CCD-camera.

sinogram correction. Reconstruction of the obtained projected images was performed using filtered back-projection as described by Basu and Bresler (9). The data was tone-mapped to 8-bits while preserving a high contrast between all relevant tissue structures. These data were saved as multiimage TIFF-stack (1). 2D-sliced and 3D-rendered data were obtained using the following software OsiriX, VGStudio MAX and ImageJ64. To further enhance the cellular distribution and to more easily distinguish between soft and calcified tissue, the histogram grey values were remapped corresponding to Figure 2.

Methods: A cylindrical cartilage-bone plug was harvested from the joint of a 24 month old cow and was immediately fixed in phosphate-buffered formaldehyde. Afterwards, it was rinsed in distilled water, decalcified using EDTA, followed by dehydration using a series of graded ethanol, exchanging the ethanol with methyl benzoate and finally exchange with xylene as described elsewhere (1). Using the BESSY BAMline setup, the sample was positioned 15 cm away from the scintillator screen to permit both phase contrast and absorption contrast. The region of interest recorded by the CCD camera was set to 2048 x 1500 pixels allowing for a maximum sample width of 3.3 mm. The sample was rotated in steps of 180°/ 1200 and was exposed to the beam at 14 keV for an exposure time of 2.0 s. Both, darkfield and flatfield corrections were applied to limit formation of ring artefacts during reconstruction. Persistent ring artefacts were compensated later via

Figure 2: Rendered SR-µCT volume data of bovine articular cartilage

The representation of a single chondrocyte inside its lacunae was achieved by digitally magnifying and median filtering of the cropped TIFF-stack in each axial direction using ImageJ. The filtered data were rendered using OsiriX, while estimating the volume of a single lacuna with its chondrocyte using ImageJ (Figure 3). To compare the SR-µCT data to a higher resolving methodology, SEM imaging combined with a focussed ion beam for milling was carried out using a Zeiss Cross Beam EsB 1540 with a gallium source. The sample was previously gold sputtercoated, FIB milled to achieve a planar surface and finally imaged in SE-detector mode at a region showing a lacuna doublet.

339

Results: The µCT data (voxel size of 1.6 µm), shows both phase contrast and absorption contrast, due to both coherence and monochromacity of the beam. The used beam energy of 14 keV proved sufficiently high to penetrate both soft tissue and hard tissue zones, while achieving good absorption contrast in both. Sample positioning 15 cm away from the scintillator screen favoured the phase contrast mode. Consequently, all cartilage structures can be revealed. Figure 2 shows the distribution of the slightly arched chondrocytes in the soft tissue region, a region of decalcified bone and the calcified subchondral bone, with its distinct microand macro-porosity.

further metal staining using a combination of phase contrast imaging mode and absorption contrast imaging.

Acknowledgements: This study was supported by the Deutsche forschungsgemeinschaft (SCHU 679/27-1 and SCHU 679/27-2). This work was submitted to the “Journal of the Royal Society – Interface” (1).

References:
(1) Zehbe et al., Going beyond Histology: Synchrotron µCT as a Complementary Methodology for Biological Tissue Characterization, Journal of the Royal Society Interface, submitted (2) Hounsfield, G.N., Computerized transverse axial scanning (tomography). Part I: Description of system. Part II: Clinical applications. British Journal of Radiology 46, 1016-1022 (1973). (3) Cormack, A.M., Reconstruction of densities from their projections, with applications in radiological physics. Physics in Medicine and Biology 18, 195-207 (1973). (4) Radon, J. Über die Bestimmung von Funktionen durch ihre Integralwerte längs gewisser Mannigfaltigkeiten, Berichte Sächsische Akademie der Wissenschaften 69. 262-267, (1917). (5) Kak, A.C. & Slaney, M. Principles of Computerized Tomographic Imaging, IEEE Press (1988).

Figure 3: Comparison of a FIB-milled crosssection of a lacuna doublet in SEM with the SRµCT data showing cells, lacunae and the volume estimation

(6) Germann, M. et al., Strain fields in histological slices of brain tissue determined by synchrotron radiation-based micro computed tomography, Journal of Neuroscience Methods 170, 149-155 (2008). (7) Cloetens, P. et al., Quantitative phase tomography of Arabidopsis seeds reveals intercellular void network, PNAS 103 (39), 1462614630 (2006). (8) Zehbe, R. et al., Characterization of oriented protein-ceramic and protein-polymer-composites for cartilage tissue engineering using synchrotron µ-CT. Int J Mat Res 98, 562-568 (2007). (9) Basu, S. & Bresler, Y. O(N2 log2N) Filtered Backprojection Reconstruction Algorithm for Tomography, IEEE Transactions on image Processing, 9 (10), 1760-1773 (2000).

The magnified spatial data of three independent chondrocytes embedded in their lacuna is shown in Figure 3. One of the lacunae with its cell was analyzed using ImageJ to have a combined volume of 1805 µm³. The SEM data of a similar lacuna doublet corresponds directly to the presented SR-µCT data. Discussion: In this study we have demonstrated that SR-µCT reaches a spatial resolution which enables the tomographic representation of single cells inside tissues without any

340

The ββα-Me type II restriction endonuclease Hpy99I
Monika Sokolowska1,2, Honorata Czapinska1,2, Matthias Bochtler1,2,3,§
1

International Institute of Molecular and Cell Biology, Trojdena 4, 02-109 Warsaw,

Poland
2

Max-Planck-Institute of Molecular Cell Biology and Genetics, Pfotenhauerstr. 108,

01309 Dresden, Germany
3

Schools of Chemistry and Biosciences, Main Building, Park Place, Cardiff University,

Cardiff CF10 3AT, UK

ββα-Me nucleases have been named for their conserved active site region, which consists of a β-hairpin (ββ) and α-helix (α) that together anchor a catalytic metal ion (Me). They are present in all three kingdoms of life and fulfill diverse biological functions. Structurally characterized members include unspecific nucleases, homing endonucleases and also Holliday junction resolvases. Structures of ββα-Me restriction endonucleases have not yet been reported, even though these enzymes form the second largest group of type II restriction endonucleases after PD-(D/E)XK enzymes and before GIY-YIG, phospholipase-derived and half-pipe endonucleases. The type II REase Hpy99I from the gastric pathogen Helicobacter pylori can be classified as a ββα-Me endonuclease on the basis of statistically significant sequence similarity to the Holliday junction resolvase T4 endonuclease VII, a bona fide ββα-Me endonuclease. The enzyme recognizes the nearly symmetric (pseudopalindromic) recognition sequence CGWCG. At the center of its target sequence, Hpy99I distinguishes A-T from G-C pairs, but not individual nucleotides. Interestingly, the enzyme cleaves its target just downstream of the recognition sequence, i.e. with a unique five base pair stagger not observed previously with this polarity in any structurally characterized restrictase DNA complex. Starting from synthetic genes, we have overproduced N-terminally histidinetagged Hpy99I in the presence of its cognate methyltransferase to protect host DNA from the endonuclease activity of Hpy99I. The recombinant restrictase was purified by metal affinity chromatography and confirmed to be active in the presence of Mg2+ or Mn2+ ions,

341

but not in the presence of Ca2+ ions. Crystals in space group H32 were obtained in the presence of EDTA or Ca2+ ions and diffracted to 1.75 and 1.50 Å resolution, respectively. Phases were obtained by the SAD method, exploiting the anomalous signal of structural Zn2+ and soaked in Br- ions. Crystals contained one Hpy99I dimer with bound DNA in the asymmetric unit. Instead of the co-factors Mg2+ or Mn2+, the active sites contained a surrogate Na+ ions of similar size and coordination preference, which do not support DNA hydrolysis.

Figure 1: Schematic representation of the Hpy99I-DNA complex. The DNA and protein are shown in all atom and ribbon representations, respectively. The two protein subunits are colored in cyan and magenta, respectively. The metal ions in the active sites are shown in orange, and structural Zn2+ ions are presented as yellow balls.

According to the crystal structure, each Hpy99I subunit recognizes one half-site of the DNA and consists of an antiparallel β-barrel and two β4α2 repeats. The β-barrel makes no contacts with DNA and lacks an obvious function. The two β4α2 repeats bind structural zinc ions. The first and second repeats make sequence specific major and minor groove contacts, respectively. The second repeat comprises the ββα-Me active site region and binds the catalytic metal ion (or its surrogate). The crystal structure illustrates the recognition of the CGWCG target sequence and attributes the partial specificity for the central base pair to exclusive minor groove readout by two arginine residues. It provides

342

the first detailed illustration of the ββα-Me site in REases and complements structural information on the use of this active site motif in other groups of endonucleases, such as homing endonucleases (e.g. I-PpoI) and Holliday junction resolvases (e.g. T4 endonuclease VII). A manuscript describing the above results is currently under consideration by the journal Nucleic Acid Research.

Acknowledgment:
We thank Dr. Uwe Mueller for generous allocation of beamtime on MXbeamline-14.1 (BESSY, Berlin) and Dr. Georg Zocher for excellent assistance at the beamline. This work was supported by the European Community - Research Infrastructure Action under the FP6 "Structuring the European Research Area" Programme (through the Integrated Infrastructure Initiative Integrating Activity on Synchrotron and Free Electron Laser Science - Contract R II 3-CT-2004-506008). M.B., M.S. and H.C. acknowledge the support from the Polish Ministry of Science and Higher Education (grant numbers 0295/B/PO1/2008/34, N N301 028934 and

PBZ/MEiN/O1/2006/24, respectively). M.B. thanks the European Molecular Biology Organization (EMBO) and HHMI for a Young Investigator Award.

343

Human Cystathionine Gamma-Lyase in complex with DLpropargylglycine (PDB entry 3CO6)
Collins, R., Karlberg, T., Lehtiö, L., Berglund, H., Dahlgren, L.G., Flodin, S., Flores, A., Gräslund, S., Hammarstrom, M., Johansson, I., Kallas, Å., Kotenyova, T., Moche, M., Nilsson, M.E., Nordlund, P., Nyman, T., Olesen, K., Persson, C., Schuler, H., Svensson, L., Thorsell, A.G., Tresaugues, L., Van den Berg, S., Sagermark, J., Busam, R.D., Welin, M., Weigelt, J., Wikström, M. Structural Genomics Consortium#, Department of Medical Biochemistry and Biophysics, Karolinska Institutet, Scheeles väg 2, 171 77 Stockholm, SWEDEN

Cystathionine- -lyase (EC 4.4.1.1, CTH, CSE, cystathionase) is a pyridoxal-phosphate (PLP) dependent enzyme that catalyses the conversion of L-cystathionine to L-cysteine in the transulfuration pathway. Deficiency of CTH causes the autosomal recessive disease cystathioninemia. CTH can also convert L-cysteine to H2S, a gas transmitter highly interesting from a medical perspective since it functions as a neuromodulator in the central nervous system and as a smooth muscle relaxant in the vascular system. It has also been suggested to be linked to several cardiovascular diseases, to (anti-) inflammatory responses and to gastric injury caused by non-steroidal anti-inflammatory drugs (1). H2S is produced in the liver, kidney, vascular system and gut by CTH and in the brain by cystathionine beta-synthetase (CBS). Yang et al. (2) showed that H2S is physiologically generated by CTH and that genetic deletion of CTH in mice markedly reduces H2S levels in the serum, heart, aorta, and other tissues. DL-Propargylglycine (PAG) is one of few known inhibitors acting on CTH (3). Here we present the crystal-structure of CTH in complex with PAG at 2.0Å resolution together with kinetics of H2S formation and inhibition studies using PAG (4). The enzyme is crystallized as a tetramer with PLP observed covalently bound to Lys212 forming a Schiff base. Both monomers in the subunit interfaces contribute to the active site pocket. The inhibitor is observed bound to Tyr114, the interaction of was not predicted based on existing information from known PAG complexes. The structure of the PAG complex highlights the particular importance of Tyr114 in and the mechanism of PAG-dependent inhibition of human CTH. These results provide significant insights, which will facilitate the structure-based design of novel inhibitors to aid in the development of therapies for diseases involving disorders of sulfur metabolism.

References
1. 2. 3. 4. S. Fiorucci, E. Distrutti, G. Cirino, J. L. Wallace, Gastroenterology 131, 259 (Jul, 2006). G. Yang et al., Science 322, 587 (Oct 24, 2008). C. Steegborn et al., J Biol Chem 274, 12675 (Apr 30, 1999). Q. Sun et al., J Biol Chem (Nov 19, 2008).

# Structural Genomics Consortium is a registered charity (number 1097737) that receives funds from the Canadian Institutes for Health Research, the Canadian Foundation for Innovation, Genome Canada through the Ontario Genomics Institute, GlaxoSmithKline, Karolinska Institutet, the Knut and Alice Wallenberg Foundation, the Ontario Innovation Trust, the Ontario Ministry for Research and Innovation, Merck & Co.,Inc., the Novartis Research Foundation, the Swedish Agency for Innovation Systems, the Swedish Foundation for Strategic Research and the Wellcome Trust.

344

The TIR domain of human Toll-Like Receptor 10 (PDB entry 2J67)
Stenmark, P., Ogg, D., Berglund, H., Busam, R., Collins, R., Ericsson, U.B., Flodin, S., Flores, A., Graslund, S., Hammarstrom, M., Hallberg, B.M., Holmberg Schiavone, L., Hogbom, M., Johansson, I., Karlberg, T., Kotenyova, T., Magnusdottir, A., Nilsson, M.E., Nilsson-Ehle, P., Nyman, T., Persson, C., Sagemark, J., Sundstrom, M., Uppenberg, J., Thorsell, A.G., Van Den Berg, S., Wallden, K., Weigelt, J., Welin, M., Nordlund, P. Structural Genomics Consortium#, Department of Medical Biochemistry and Biophysics, Karolinska Institutet, Scheeles väg 2, 171 77 Stockholm, SWEDEN

The Toll-like receptors (TLRs) play an essential role in initiating the immune response against pathogens. The 10 identified human TLRs recognize a wide variety of pathogen-associated molecular patterns from bacteria, viruses, and fungi, as well as certain host-derived molecules (1). TLRs are type I transmembrane glycoproteins with an extracellular domain composed of numerous leucine-rich repeats and an intracellular region containing a Toll/IL-1 receptor homology (TIR) domain (2). The intracellular domains interact with several TIR domaincontaining adaptor molecules that conveys the downstream signaling, resulting in transcription factor induction. Dysfunction in the TLRs results in desensitization of the recognition of certain (3) pathogenic ligands, and might render the host susceptible for microbial invasion. To date, TLR10 remains the only orphan member among the human TLRs. TLR10 has been reported to homodimerize and also heterodimerize with TLRs 1 and 2, and to directly associate with the common TLR adaptor MyD88 (4). The 2.2 Šcrystal structure of the TLR10 TIR domain represents the third TLR TIRdomain structure alongside with those of TLRs 1 and 2 (5). The asymmetric unit is comprised of a dimer with a two-fold symmetry axis. The buried surface area of 974 Ų shows a high degree of compatibility between the two monomers, and the structure is in accordance with functional mapping done with TLR10 and other TLR family members. Analysis of the amino acid sequence conservation between members of the TLR family has revealed a sequence that might constitute an adaptor protein recruitment site (Hasan, 2005). In the TLR10 dimer, these residues form an extended patch. It is concluded that this structure might represent the physiological dimer. References 1. 2. 3. 4. 5. H. Heine, E. Lien, Int Arch Allergy Immunol 130, 180 (Mar, 2003). R. J. Ulevitch, Nat Rev Immunol 4, 512 (Jul, 2004). K. A. Fitzgerald et al., J Exp Med 198, 1043 (Oct 6, 2003). U. Hasan et al., J Immunol 174, 2942 (Mar 1, 2005). T. Nyman et al., J Biol Chem 283, 11861 (May 2, 2008).

# Structural Genomics Consortium is a registered charity (number 1097737) that receives funds from the Canadian Institutes for Health Research, the Canadian Foundation for Innovation, Genome Canada through the Ontario Genomics Institute, GlaxoSmithKline, Karolinska Institutet, the Knut and Alice Wallenberg Foundation, the Ontario Innovation Trust, the Ontario Ministry for Research and Innovation, Merck & Co.,Inc., the Novartis Research Foundation, the Swedish Agency for Innovation Systems, the Swedish Foundation for Strategic Research and the Wellcome Trust

345

Human Cleavage and Polyadenylation Specificity Factor 5 (PDB entry 2CL3/2J8Q)
Moche, M., Stenmark, P., Ogg, D., Berglund, H., Busam, R., Collins, R., Ericsson, U.B., Flodin, S., Flores, A., Graslund, S., Hammarstrom, M., Hallberg, B.M., Holmberg, S.L., Hogbom, M., Johansson, I., Karlberg, T., Kosinska, U., Kotenyova, T., Magnusdottir, A., Nilsson, M.E., Nilsson-Ehle, P., Nyman, T., Persson, C., Sagemark, J., Sundstrom, M., Uppenberg, J., Upsten, M., Thorsell, A.G., Van Den Berg, S., Wallden, K., Weigelt, J., Welin, M., Nordlund, P. # Structural Genomics Consortium , Department of Medical Biochemistry and Biophysics, Karolinska Institutet, Scheeles väg 2, 171 77 Stockholm, SWEDEN

Modifications in the 3´ end of mRNA are carried out in two steps: ii) endonucleolytically cleavage at the polyadenylation site and ii) polyadenylation. One of the factors involved in this process is the pre-mRNA cleavage factor Im which is a heterodimeric protein consisting of one 25 kDa subunit and a subunit of either 68, 59 or 72 kDa (1). We have determined the structure of the 25 kDa subunit called “Cleavage and Polyadenylation Specificity Factor 5” (CPSF5) to a resolution of 1.9 Å using SAD phasing (2). It is a member of the NudiX protein family but harbours large structural differences with other members of the family (e.g. several extra beta strands and one additional long helix). CPSF5 appeared to be a dimer in the crystallographical structure, which opens three possibilities: i) the complex is actually a heterotetramer, ii) one dimer of CPSF5 interacts with one monomer of the large subunit or iii) the CPSF5 dimer dissociates when interacting with the large subunit forming a heterodimer. A crevice that is very likely to be the RNA binding site of the monomer is prolonged in the dimer making a long continuous crevice over the dimer. Contacts between symmetry-related mates also reveal the existence of a putative protein-binding site. References
1. 2. U. Ruegsegger, D. Blank, W. Keller, Mol Cell 1, 243 (Jan, 1998). L. Tresaugues et al., Proteins 73, 1047 (Dec, 2008).

# Structural Genomics Consortium is a registered charity (number 1097737) that receives funds from the Canadian Institutes for Health Research, the Canadian Foundation for Innovation, Genome Canada through the Ontario Genomics Institute, GlaxoSmithKline, Karolinska Institutet, the Knut and Alice Wallenberg Foundation, the Ontario Innovation Trust, the Ontario Ministry for Research and Innovation, Merck & Co.,Inc., the Novartis Research Foundation, the Swedish Agency for Innovation Systems, the Swedish Foundation for Strategic Research and the Wellcome Trust

346

Human Argininosuccinate synthase (PDB entry 2NZ2)
Karlberg, T., Uppenberg, J., Berglund, H., Busam, R.D., Collins, R., Ericsson, U.B., Flodin, S., Flores, A., Graslund, S., Hallberg, B.M., Hammarstrom, M., Hogbom, M., Johansson, I., Kotenyova, T., Magnusdottir, A., Moche, M., Nilsson, M.E., Nordlund, P., Nyman, T., Ogg, D., Persson, C., Sagemark, J., Stenmark, P., Sundstrom, M., Thorsell, A.G., Van Den Berg, S., Wallden, K., Weigelt, J., Holmberg-Schiavone, L. # Structural Genomics Consortium , Department of Medical Biochemistry and Biophysics, Karolinska Institutet, Scheeles väg 2, 171 77 Stockholm, SWEDEN

Argininosuccinate synthase (EC 6.3.4.5, ASS, Citrullineaspartate ligase) is responsible for the condensation reaction between citrulline and aspartate in an ATPdependent manner to form argininosuccinate, AMP and pyrophosphate. It is the rate-limiting enzyme in the synthesis of arginine and part of the Urea-cycle and also a potential limiting step for the NO-production via the Arginine-citrulline cycle (1). The gaseous messenger NO has been widely studied and is of significant importance to human cell physiology (2). Deficiency of ASS can lead to the rare autosomal disorders citrullinemia or hyperammonemia where increased levels of citrulline and ammonia could affect the brain or manifest as intermittent ataxia (3). We have solved the structure of the human tetrameric argininosuccinate synthase in complex with two of its substrates, aspartate and citrulline at 2.4Å resolution (4). The enzyme comprises two domains, one nucleotide-binding domain and one catalytic domain. The nucleotide domain resembles the Adenine nucleotide alpha hydrolase-like fold, whereas the catalytic domain has a unique fold for this type of enzyme. An additional C-terminal extension, a long –helix is stretching out from the catalytic domain and is making interactions with neighboring monomers and thus is important for the oligomerization. The overall structure is a homo-tetramer with tight subunit-subunit interactions. Two of the substrates were found bound to the active site pocket, in a cleft between the two domains. The binding site for the aspartate lies between citrulline and a loop from the nucleotide-binding domain whereas the citrulline is bound to the catalytic domain. The residues in the active site cleft are highly conserved. The structure gives new insights into the function of the numerous clinical mutations identified in patients with type I citrullinemia. References 1. 2. 3. 4. A. Husson, C. Brasse-Lagnel, A. Fairand, S. Renouf, A. Lavoinne, Eur J Biochem 270, 1887 (May, 2003). M. Mori, T. Gotoh, Biochem Biophys Res Commun 275, 715 (Sep 7, 2000). L. C. Pendleton, B. L. Goodwin, L. P. Solomonson, D. C. Eichler, J Biol Chem 280, 24252 (Jun 24, 2005). T. Karlberg et al., Acta Crystallogr D Biol Crystallogr 64, 279 (Mar, 2008).

# Structural Genomics Consortium is a registered charity (number 1097737) that receives funds from the Canadian Institutes for Health Research, the Canadian Foundation for Innovation, Genome Canada through the Ontario Genomics Institute, GlaxoSmithKline, Karolinska Institutet, the Knut and Alice Wallenberg Foundation, the Ontario Innovation Trust, the Ontario Ministry for Research and Innovation, Merck & Co.,Inc., the Novartis Research Foundation, the Swedish Agency for Innovation Systems, the Swedish Foundation for Strategic Research and the Wellcome Trust

347

Human DEAD-box RNA-helicase DDX19 ADP complex (PDB entry 3EWS)
Lehtio,L., Karlberg,T., Andersson,J., Berglund,H., Collins,R., Dahlgren,L.G., Flodin,S., Flores,A., Graslund,S., Hammarstrom,M., Johansson,A., Johansson,I., Kotenyova, T., Moche,M., Nilsson,M.E., Nordlund,P., Nyman,T., Olesen,K., Persson,C., Sagemark,J., T horsell,A.C., Tresaugues,l., Van den berg,S., Weigelt, J., Welin,M., Wikstrom,M., Wisniewska,M., Schueler, H Structural Genomics Consortium#, Department of Medical Biochemistry and Biophysics, Karolinska Institutet, Scheeles väg 2, 171 77 Stockholm, SWEDEN

The DExD/H family of RNA-binding helicases consists of a large group of proteins involved in general cellular RNAmetabolism such as transcription, splicing, RNA nucleocytoplasmatic transport, translation and ribosome biogenesis (1). All DExD/H helicases bind and hydrolyze ATP, and are believed to unwind RNA-secondary structure or assist in the folding of RNA/RNP complexes thus acting as RNA chaperones. Proteins in this family normally consist of two domains: an N-terminal domain with the conserved DExD/H motif and a C-terminal helicase domain. There are at least eight conserved motifs, based on primary sequence alignments that are involved in coordination and hydrolysis of ATP and binding of RNA (2, 3). DDX19 (or Dbp5) is a human DEAD-box helicase that is required for mRNA export from the nucleus (4). DDX19 is located primarily in the cytosol, but is recruited to the nuclear pore complex, where it assists in the mRNA export. Futher, the helicase is involved in translation termination and interacts with release factors eRF1 and eRF3 (5). We have solved the structure of the conserved core of DDX19, containing the DEADdomain and the helicase domain in complex with ADP, at 2.7 Å resolution. DDX19 has a fold that is typical of a DEAD-box helicase with two - RecA-like domains connected via a flexible linker. The nucleotide binding site is located in a pocket between the two domains. ATP hydrolysis at this site enables DDX19 to translocate along ssRNA and remove paired strands or proteins (3).

References
1. 2. 3. 4. 5. P. Linder, Curr Biol 18, R297 (Apr 8, 2008). O. Cordin, J. Banroques, N. K. Tanner, P. Linder, Gene 367, 17 (Feb 15, 2006). A. M. Pyle, Annu Rev Biophys 37, 317 (2008). C. Schmitt et al., Embo J 18, 4332 (Aug 2, 1999). T. Gross et al., Science 315, 646 (Feb 2, 2007).

# Structural Genomics Consortium is a registered charity (number 1097737) that receives funds from the Canadian Institutes for Health Research, the Canadian Foundation for Innovation, Genome Canada through the Ontario Genomics Institute, GlaxoSmithKline, Karolinska Institutet, the Knut and Alice Wallenberg Foundation, the Ontario Innovation Trust, the Ontario Ministry for Research and Innovation, Merck & Co.,Inc., the Novartis Research Foundation, the Swedish Agency for Innovation Systems, the Swedish Foundation for Strategic Research and the Wellcome Trust.

348

Human FLJ10324 (RADIL), RA domain (PDB entry 3EC8)
Wisniewska, M., Lehtio, L., Andersson, J., Collins, R., Dahlgren, L.G., Flodin, S., Flores, A., Graslund, S., Hammarstrom, M., Johansson, A., Johansson, I., Karlberg, T., Kotenyova, T., Moche, M., Nilsson, M.E., Nordlund, P., Nyman, T., Olesen, K., Persson, C., Sagemark, J., Schueler, H., Thorsell, A.G., Tresaugues, L., van den Berg, S., Weigelt, J., Welin, M., Wikstrom, M., Berglund, H # Structural Genomics Consortium , Department of Medical Biochemistry and Biophysics, Karolinska Institutet, Scheeles väg 2, 171 77 Stockholm, SWEDEN

FLJ10324 (also called RADIL, KIAA1849, AF6L) is a multi-domain protein composed of a Ras association (RA) domain, a forkhead-associated (FHA) domain mediating phospho-peptide interactions, a dilute (DIL domain) of unknown function, and a protein-protein interacting PDZ domain. The RA domain is structurally similar to ubiquitin and is present in one or two copies in a number of signalling molecules that bind and regulate a small GTPase called Ras or the Rasrelated GTPases, Ral and Rap. RA-containing proteins include RalGDS, AF6, RIN1, RASSF1, SNX27, CYR1, STE50, and phospholipase C epsilon. At date there is limited knowledge about the specific biological role played by FLJ10342. This protein has a similar domain composition as afadin (AF6), a protein involved in linking membrane-associated proteins to the actin cytoskeleton and in Rap-induced cell adhesion that suggests a related role for FLJ10324. The RA domain of FLJ10324 has been shown to bind to the small GTPase Rap1 and reduce the exchange rate of the nucleotide but seems not to effect Rap-induced cell adhesion in a similar way as afadin (1). Later FLJ10324 has been shown to be required for cell adhesion and migration of neural crest precursors, a highly motile embryonic cell population that give rise to different cell types (2). Here we present the crystal structure of RA domain of FLJ10324 at 2.6Å resolution. The structure was solved by MIRAS phasing using heavy atoms derivatives (lead and mercury). The overall structure reveals the ubiquitin-like fold consisting of five-stranded -sheet and two -helices. Additionally, at N- and C- terminus we can observe two helices that do not belong to the RA domain.

References
1. 2. Z. Zhang, H. Rehmann, L. S. Price, J. Riedl, J. L. Bos, J Biol Chem 280, 33200 (Sep 30, 2005). G. A. Smolen et al., Genes Dev 21, 2131 (Sep 1, 2007).

# Structural Genomics Consortium is a registered charity (number 1097737) that receives funds from the Canadian Institutes for Health Research, the Canadian Foundation for Innovation, Genome Canada through the Ontario Genomics Institute, GlaxoSmithKline, Karolinska Institutet, the Knut and Alice Wallenberg Foundation, the Ontario Innovation Trust, the Ontario Ministry for Research and Innovation, Merck & Co.,Inc., the Novartis Research Foundation, the Swedish Agency for Innovation Systems, the Swedish Foundation for Strategic Research and the Wellcome Trust.

349

Human guanine monophosphate synthetase (PDB entry 2VXO)
Welin, M., Lehtio, L., Andersson, J., Arrowsmith, C.H., Berglund, H., Collins, R., Dahlgren, L.G., Edwards, A.M., Flodin, S., Flores, A., Graslund, S., Hammarstrom, M., Johansson, A., Johansson, I., Karlberg, T., Kotenyova, T., Moche, M., Nilsson, M.E., Nyman, T., Olesen, K., Persson, C., Sagemark, J., Schueler, H., Thorsell, A.G., Tresaugues, L., Van Den Berg, S., Wisniewska, M., Wikstrom, M., Nordlund, P. # Structural Genomics Consortium , Department of Medical Biochemistry and Biophysics, Karolinska Institutet, Scheeles väg 2, 171 77 Stockholm, SWEDEN

GMP synthetase (GMPS, E.C. 6.3.5.2) is a glutamine amidotransferase involved in the de novo synthesis of purines. It catalyzes the conversion of XMP to GMP in the presence of glutamine and ATP. GMPS is a bifunctional enzyme with an N-terminal glutaminase domain that generates ammonia from glutamine, and a C-terminal synthetase domain that aminates XMP to GMP (1). XMP + ATP + GLN + H2O -> GMP + AMP + GLU + PPi GMPS is a potential target for anticancer therapies and acivicin is known to inhibit GMPS (2-4). The E. coli GMPS structure suggested large movements between the two domains, separated by 30 Å, during reaction since no obvious route for ammonia channeling was visible (1, 5). Human GMP synthetase structure was solved to 2.5 Å resolution by molecular replacement using the glutaminase domain of human GMP synthetase (2VPI) and the synthetase domain of Thermus thermophilus (2YWC). GMPS belongs to the class 1 glutamine dependent amidotranferases which has a conserved catalytic triad consisting of a Cys-His-Glu [5]. The asymmetric unit contains a homodimer with XMP bound to each active site. The human GMPS is slightly larger than the T. thermophilus and E.coli GMPS structures, having an additional domain in the dimer interface stretching from residue 449 to 579. This domain is built up by a three stranded β-sheet flanked by five αhelices. A β-hairpin in this domain is also interacting with the same β-hairpin in the other subunit and involved in binding of the XMP phosphate. The human GMPS dimer is more tightly packed than the bacterial GMPS structures. This is the first GMPS structure of eukaryotic origin and the information regarding substrate binding will aid potential drug design.

References
1. 2. 3. 4. 5. X. Huang, H. M. Holden, F. M. Raushel, Annu Rev Biochem 70, 149 (2001). S. V. Chittur, T. J. Klem, C. M. Shafer, V. J. Davisson, Biochemistry 40, 876 (Jan 30, 2001). M. Hirst, E. Haliday, J. Nakamura, L. Lou, J Biol Chem 269, 23830 (Sep 23, 1994). J. Nakamura, K. Straub, J. Wu, L. Lou, J Biol Chem 270, 23450 (Oct 6, 1995). J. J. Tesmer, T. J. Klem, M. L. Deras, V. J. Davisson, J. L. Smith, Nat Struct Biol 3, 74 (Jan, 1996).

# Structural Genomics Consortium is a registered charity (number 1097737) that receives funds from the Canadian Institutes for Health Research, the Canadian Foundation for Innovation, Genome Canada through the Ontario Genomics Institute, GlaxoSmithKline, Karolinska Institutet, the Knut and Alice Wallenberg Foundation, the Ontario Innovation Trust, the Ontario Ministry for Research and Innovation, Merck & Co.,Inc., the Novartis Research Foundation, the Swedish Agency for Innovation Systems, the Swedish Foundation for Strategic Research and the Wellcome Trust

350

Human heat shock 70kDa protein 6 (HSP70B), ATPase domain in complex with ADP (PDB entry 3FE1)
Wisniewska, M., Lehtio, L., Berglund, H., Collins, R., Dahlgren, L.G., Flodin, S., Flores, A., Graslund, S., Hammarstrom, M., Johansson, A., Johansson, I., Karlberg, T., Kotenyova, T., Moche, M., Nilsson, M.E., Nordlund, P., Nyman, T., Persson, C., Sagemark, J., Siponen, M.I., Thorsell, A.G., Tresaugues, L., Van Den Berg, S., Weigelt, J., Welin, M., Wikstrom, M., Schueler, H. 1 Structural Genomics Consortium , Department of Medical Biochemistry and Biophysics, Karolinska Institutet, Scheeles väg 2, 171 77 Stockholm, SWEDEN

Heat shock-70 proteins (Hsp70) are a family of chaperones involved in the heat shock and unfolded protein response as well as housekeeping functions (1). The human genome contains at least eight Hsp70 isoforms – some stress-induced, some constitutively expressed; some cytosolic and some organelle-specific. Hsp70 proteins consist of an N-terminal ATPase domain and a C-terminal peptide binding domain joined by a flexible linker. They bind extended peptide segments with a net hydrophobic character exposed during translation, membrane translocation, or following stress-induced damage, and allow refolding of such peptides. Refolding activity is coupled to ATPase activity of the N-terminal domain, which is also regulated by co-chaperones (2). The human HSPA6 protein (also called Hsp70B’) is a stress-induced member of this family that localizes to the nucleus and cytosol (3-5). We have determined the crystal structure of the HSPA6 ATPase domain (residues E6 – D385), which shares 85% identity with the corresponding part of the HSPA1AA (Hsp72) protein. Our structure shows the protein with one magnesium ion and the products of ATP hydrolysis (one molecule of ADP and one phosphate group) bound in the cleft between the two major lobes of the ATPase domain. Coordinates and structure factors have been deposited in the pdb (accession code 3FE1).

References
1. 2. 3. 4. 5. M. Daugaard, M. Rohde, M. Jaattela, FEBS Lett 581, 3702 (Jul 31, 2007). T. K. Leung, C. Hall, M. Rajendran, N. K. Spurr, L. Lim, Genomics 12, 74 (Jan, 1992). T. K. Leung, M. Y. Rajendran, C. Monfries, C. Hall, L. Lim, Biochem J 267, 125 (Apr 1, 1990). M. P. Mayer, B. Bukau, Cell Mol Life Sci 62, 670 (Mar, 2005). E. J. Noonan, R. F. Place, C. Giardina, L. E. Hightower, Cell Stress Chaperones 12, 219 (Autumn, 2007).

1 Structural Genomics Consortium is a registered charity (number 1097737) that receives funds from the Canadian Institutes for Health Research, the Canadian Foundation for Innovation, Genome Canada through the Ontario Genomics Institute, GlaxoSmithKline, Karolinska Institutet, the Knut and Alice Wallenberg Foundation, the Ontario Innovation Trust, the Ontario Ministry for Research and Innovation, Merck & Co.,Inc., the Novartis Research Foundation, the Swedish Agency for Innovation Systems, the Swedish Foundation for Strategic Research and the Wellcome Trust.

351

Human Nudix motif 16 (PDB entry 3COU)
Tresaugues, L., Moche, M., Berglund, H., Busam, R.D., Collins, R., Dahlgren, L.G., Flodin, S., Flores, A., Graslund, S., Hammarstrom, M., Herman, M.D., Johansson, A., Johansson, I., Kallas, A., Karlberg, T., Kotenyova, T.,Lehtio, L., Nilsson, M.E., Nyman, T., Persson, C., Sagemark, J., Schueler, H., Svensson, L., Thorsell, A.G., Van Den Berg, S., Welin, M., Weigelt, J., Wikstrom, M., Nordlund, P. # Structural Genomics Consortium , Department of Medical Biochemistry and Biophysics, Karolinska Institutet, Scheeles väg 2, 171 77 Stockholm, SWEDEN

NUDT16 is a 195 aminoacid long protein which belongs to the Nudix hydrolase superfamily (1). The activities of these enzymes are targeted against a broad range of substrate containing a nucleoside diphosphate linked to another moiety (NDP-X) leading to NMP + P-X. NUDT16 has a “decapping” activity consisting in removing the 5´cap of the mRNA rendering it accessible for being degradated by a 5´-3´ exonuclease (2). The specificity of NUDT16 and its homologuous from Xenopus laevis X29 has been shown to be dependent of divalent cations required for hydrolysis. In vitro, when Mg2+ is added, the protein is only able to hydrolyze U8snoRNA whereas in presence of Mn2+ or Co2+ the activity is higher and the specificity is extended to other RNAs (3). Structures of the both apo and holo- X29 protein of Xenopus laevis in complex with Mn2+ and m7GpppA were available (4). We have solved the structure of human NUDT16 to a resolution of 1.8Å (PDB code : 3COU). Applying a crystallographic symmetry operator on the monomer present in the asymmetrical unit allows retrieving the dimerical quaternary structure previously determined in the Xenopus homologous X29. If the sequence identity between this X29 and NUDT16 is only 54%, residues involved in Mn2+ chelating and CAP binding are strictly conserved. One difference between these structures is that the side-chain of Glu158, involved in Mn2+ chelating, is in the same orientation as the corresponding residues (Glu150) in the holo-X29 despite the lack of a cation in this area. References
1. 2. 3. 4. A. G. McLennan, Cell Mol Life Sci 63, 123 (Jan, 2006). T. Ghosh, B. Peterson, N. Tomasevic, B. A. Peculis, Mol Cell 13, 817 (Mar 26, 2004). B. A. Peculis, K. Reynolds, M. Cleland, J Biol Chem 282, 24792 (Aug 24, 2007). J. N. Scarsdale, B. A. Peculis, H. T. Wright, Structure 14, 331 (Feb, 2006).

# Structural Genomics Consortium is a registered charity (number 1097737) that receives funds from the Canadian Institutes for Health Research, the Canadian Foundation for Innovation, Genome Canada through the Ontario Genomics Institute, GlaxoSmithKline, Karolinska Institutet, the Knut and Alice Wallenberg Foundation, the Ontario Innovation Trust, the Ontario Ministry for Research and Innovation, Merck & Co.,Inc., the Novartis Research Foundation, the Swedish Agency for Innovation Systems, the Swedish Foundation for Strategic Research and the Wellcome Trust

352

Human Phosphate Cytidylyltransferase 2 (PDB entry 3ELB)
Karlberg, T., Welin, M., Andersson, J., Arrowsmith, C.H., Berglund, H., Bountra, C., Collins, R., Dahlgren, L.G., Edwards, A.M., Flodin, S., Flores, A., Graslund, S., Hammarstrom, M., Johansson, A., Johansson, I., Kotenyova, T., Lehtio, L., Moche, M., Nilsson, M.E., Nordlund, P., Nyman, T., Persson, C., Sagemark, J., Thorsell, A.G., Tresaugues, L., Van Den Berg, S., Weigelt, J., Wikstrom, M., Wisniewska, M., Schuler, H. # Structural Genomics Consortium , Department of Medical Biochemistry and Biophysics, Karolinska Institutet, Scheeles väg 2, 171 77 Stockholm, SWEDEN

Ethanolamine phospholipids (PE) are a major component of cell membranes in eukaryotes and the most abundant in prokaryotic membranes. Three biosynthetic pathways of PE formation are known in eukaryotes: the major route CDPethanolamine pathway (Kennedy pathway), decarboxylation of phosphatidylserine and base exchange with phosphatidylserine (1). Phosphate cytidylyltransferase 2 (CTP:Phosphoethanolamine cytidylyltransferase, PCYT2; ECT) catalyzes the transformation of CTP and phosphoethanolamine into CDP-Ethanolamine and pyrophosphate using the hydrolysis of CTP to CMP, the penultimate step in the Kennedy pathway of de novo synthesis of phospholipids (1-3). After which CDPEthanolamine:1,2-diacylglycerol ethanolaminephosphotransferase catalyses the final step producing PE. Recent studies suggest that increase in phosphoethanolamine in some breast cancer cells is caused by the down-regulation of PCYT2 (4). Here we present the crystal structure of PCYT2 in complex with CMP at a resolution of 2.0Å. The structure was solved by SAD using selenomethionine-labelled protein. PCYT2 is a monomer which consists of two cytidylyltransferase domains and each domain contains a nucleotide-binding motif HxGH. The N-terminal domain (residues Gly19Leu155) consists of a five-stranded parallel -sheet with topology 3-2-1-4-5 flanked by five -helices and a sixth helix making considerable domain-domain interactions. The Cterminal domain (Trp187-Thr351) has a similar fold. Interestingly, only the C-terminal domain has CMP bound in its nucleotide binding pocket. CTP and MgCl2 were added to the protein in the crystallization trials.

References
1. 2. 3. 4. M. Bakovic, M. D. Fullerton, V. Michel, Biochem Cell Biol 85, 283 (Jun, 2007). O. B. Bleijerveld, J. F. Brouwers, A. B. Vaandrager, J. B. Helms, M. Houweling, J Biol Chem 282, 28362 (Sep 28, 2007). A. Signorell, M. Rauch, J. Jelk, M. A. Ferguson, P. Butikofer, J Biol Chem 283, 23636 (Aug 29, 2008). L. Zhu, C. Johnson, M. Bakovic, J Lipid Res 49, 2197 (Oct, 2008).

# Structural Genomics Consortium is a registered charity (number 1097737) that receives funds from the Canadian Institutes for Health Research, the Canadian Foundation for Innovation, Genome Canada through the Ontario Genomics Institute, GlaxoSmithKline, Karolinska Institutet, the Knut and Alice Wallenberg Foundation, the Ontario Innovation Trust, the Ontario Ministry for Research and Innovation, Merck & Co.,Inc., the Novartis Research Foundation, the Swedish Agency for Innovation Systems, the Swedish Foundation for Strategic Research and the Wellcome Trust

353

Human sorting nexin-17, PX domain (PDB entry 3FOG)
Wisniewska, M., Tresaugues, L., Berglund, H., Collins, R., Dahlgren, L.G., Flodin, S., Flores, A., Graslund, S., Hammarstrom, M., Johansson, A., Johansson, I., Karlberg, T., Kotenyova, T., Lehtio, L., Moche, M., Nilsson, M.E., Nordlund, P., Nyman, T., Persson, C., Sagemark, J., Siponen, M.I., Thorsell, A.G., Van Den Berg, S., Weigelt, J., Welin, M., Wikstrom, M., Schueler, H. 1 Structural Genomics Consortium , Department of Medical Biochemistry and Biophysics, Karolinska Institutet, Scheeles väg 2, 171 77 Stockholm, SWEDEN

Sorting nexins (SNXs) are a family of membrane associated proteins characterized by the presence of a phox (PX) homology domain. This domain confers specificity towards phosphoinositides, and targets the SNX proteins to membrane domains enriched in specific phospholipids (1, 2). Sorting nexin-17 was initially identified as a binding partner for P-selectin (3). This interaction accelerates P-selectin internalization and inhibits its lysosomal degradation (4). SNX17 is also a binding partner for several members of the low-density lipoprotein (LDL) receptor family such as LDLR, VLDLR, ApoER2 and LDLR-related protein (LRP) (5). Recently it was shown that SNX17 interacts with the NPVY motif in the LDL receptor tail (6), and is a part of the cellular sorting machinery that regulates cell surface levels of LRP by promoting its recycling (7). SNX17 consists of an N-terminal PX domain, followed by a B41 (band 4.1 or FERM) domain. Here we present the crystal structure of the PX domain at 2.8 Å resolution. The structure was solved by molecular replacement using MOLREP with the cytokine-independent survival kinase CISK-PX (1XTE) as a search model. The asymmetric unit consisted of one polypeptide chain. The overall fold of the SNX17-PX domain is composed of a sheet with three antiparallel -strands and a helical subdomain consisting of three helices. Coordinates and structure factors have been deposited in the PDB with accession code 3FOG.

References
1. 2. 3. 4. 5. 6. 7. M. L. Cheever et al., Nat Cell Biol 3, 613 (Jul, 2001). F. Kanai et al., Nat Cell Biol 3, 675 (Jul, 2001). V. Florian, T. Schluter, R. Bohnensack, Biochem Biophys Res Commun 281, 1045 (Mar 9, 2001). R. Williams et al., Mol Biol Cell 15, 3095 (Jul, 2004). W. Stockinger et al., Embo J 21, 4259 (Aug 15, 2002). J. J. Burden, X. M. Sun, A. B. Garcia, A. K. Soutar, J Biol Chem 279, 16237 (Apr 16, 2004). P. van Kerkhof et al., Embo J 24, 2851 (Aug 17, 2005).

1 Structural Genomics Consortium is a registered charity (number 1097737) that receives funds from the Canadian Institutes for Health Research, the Canadian Foundation for Innovation, Genome Canada through the Ontario Genomics Institute, GlaxoSmithKline, Karolinska Institutet, the Knut and Alice Wallenberg Foundation, the Ontario Innovation Trust, the Ontario Ministry for Research and Innovation, Merck & Co.,Inc., the Novartis Research Foundation, the Swedish Agency for Innovation Systems, the Swedish Foundation for Strategic Research and the Wellcome Trust.

354

Human TULP1 in complex with IP3 (PDB entry 3C5N)
Busam, R.D., Lehtio, L., Collins, R., Dahlgren, L.G., Flodin, S., Flores, A., Graslund, S., Hammarstrom, M., Hallberg, B.M., Herman, M.D., Johansson, A., Johansson, I., Kallas, A., Karlberg, T., Kotenyova, T., Moche, M., Nilsson, M.E., Nordlund, P., Nyman, T., Persson, C., Sagemark, J., Svensson, L., Thorsell, A.G., Tresaugues, L., Van den Berg, S., Weigelt, J., Welin, M., Berglund, H. # Structural Genomics Consortium , Department of Medical Biochemistry and Biophysics, Karolinska Institutet, Scheeles väg 2, 171 77 Stockholm, SWEDEN

Monogenic murine obesity models are of value for pinpointing genetic causes for obesity (1). One murine obesity model is ‘tubby’ where the phenotype is characterized by increased body-weight at three to six months of age. Once the weight gain is induced it rapidly progresses and results in doubling of weight compared to controls. The phenotype is caused by an autosomal recessive mutation (2, 3). Genetic approaches aimed at understanding the molecular basis for the tubby mouse phenotype revealed a family of four genes: the founding member TUB and three Tubby-like proteins (TULP1-3) that share significant sequence homology in a 260 residues long C-terminal domain (4, 5). The C-terminal domain of TUB has been found to anchor in the membrane by binding to phosphatidylinositol-(4,5)-bisphosphate (6). Upon cleavage of PIP2 by phospholipase C, a nuclear localization signal in the N-terminal part of TUB mediates transport into the nucleus. In the nucleus, TUB may regulate the transcription of one or several so far unknown genes (7). Here we solved the structure of TULP1 in complex with IP3 to 1.8Å resolution that is a follow on structure of 2FIM. References
1. 2. 3. 4. 5. 6. 7. J. Naggert, T. Harris, M. North, Curr Opin Genet Dev 7, 398 (Jun, 1997). P. W. Kleyn et al., Cell 85, 281 (Apr 19, 1996). D. L. Coleman, E. M. Eicher, J Hered 81, 424 (Nov-Dec, 1990). P. M. Nishina, M. A. North, A. Ikeda, Y. Yan, J. K. Naggert, Genomics 54, 215 (Dec 1, 1998). M. A. North, J. K. Naggert, Y. Yan, K. Noben-Trauth, P. M. Nishina, Proc Natl Acad Sci U S A 94, 3128 (Apr 1, 1997). S. Santagata et al., Science 292, 2041 (Jun 15, 2001). T. J. Boggon, W. S. Shan, S. Santagata, S. C. Myers, L. Shapiro, Science 286, 2119 (Dec 10, 1999).

# Structural Genomics Consortium is a registered charity (number 1097737) that receives funds from the Canadian Institutes for Health Research, the Canadian Foundation for Innovation, Genome Canada through the Ontario Genomics Institute, GlaxoSmithKline, Karolinska Institutet, the Knut and Alice Wallenberg Foundation, the Ontario Innovation Trust, the Ontario Ministry for Research and Innovation, Merck & Co.,Inc., the Novartis Research Foundation, the Swedish Agency for Innovation Systems, the Swedish Foundation for Strategic Research and the Wellcome Trust

355

Human dCMP deaminase (PDB entry 2W4L)
Siponen, M.I., Moche, M., Berglund, H., Collins, R., Dahlgren, L.G., Flodin, S., Flores, A., Graslund, S., Hammarstrom, M., Johansson, A., Johansson, I., Karlberg, T., Kotenyova, T., Lehtio, L., Nilsson, M.E., Nyman, T., Persson, C., Sagemark, J., Schuler, H., Thorsell, A.G., Tresaugues, L., Van Den Berg, S., Weigelt, J., Welin, M., Wikstrom, M., Wisniewska, M., Nordlund, P. 1 Structural Genomics Consortium , Department of Medical Biochemistry and Biophysics, Karolinska Institutet, Scheeles väg 2, 171 77 Stockholm, SWEDEN

2´-Deoxycytidylate deaminase (dCMP deaminase or dCD) is a key enzyme in the pyrimidine metabolism catalysing the deamination of deoxycytidine monophosphate (dCMP) into deoxyuridine5´-monophosphate (dUMP). In humans, dCMP deaminase is hexameric and allosterically activated by magnesium associated dCTP (Mg*dCTP) while inhibited by deoxythymidine-5' -triphosphated (dTTP)(1-3). In Humans, dCD modifies several anticancer(4) and antiviral(5) drugs by deamination thereby reducing drug efficiency giving dCD inhibitors therapeutic potential. A crystal structure of the dCD enzyme from Streptococcus mutans was recently determined with the active site zinc to 3.0Å resolution (2HVV) and in complex with zinc, Mg*dCTP and the substrate analogue DHOMP to 1.7Å resolution (2HVW) (6). We determined the crystal structure of human dCMP deaminase to 2.1Å resolution with a zinc ion in the active site (2W4L). The enzyme is found to be hexameric and current experiments aim to determine the structural basis for its allosteric mechanism. Comparing this human and S. mutans dCD enzymes reveal that the protein part, bridging the allosteric effector and the zinc active sites, is quite different between the two enzymes. References
1. 2. 3. 4. 5. 6. G. F. Maley, F. Maley, Science 141, 1278 (Sep 27, 1963). G. F. Maley, F. Maley, Biochemistry 21, 3780 (Aug 3, 1982). W. R. Mancini, Y. C. Cheng, Mol Pharmacol 23, 159 (Jan, 1983). B. Hernandez-Santiago et al., Antimicrob Agents Chemother 46, 1728 (Jun, 2002). J. Y. Liou, P. Krishnan, C. C. Hsieh, G. E. Dutschman, Y. C. Cheng, Mol Pharmacol 63, 105 (Jan, 2003). H. F. Hou, Y. H. Liang, L. F. Li, X. D. Su, Y. H. Dong, J Mol Biol 377, 220 (Mar 14, 2008).

1 Structural Genomics Consortium is a registered charity (number 1097737) that receives funds from the Canadian Institutes for Health Research, the Canadian Foundation for Innovation, Genome Canada through the Ontario Genomics Institute, GlaxoSmithKline, Karolinska Institutet, the Knut and Alice Wallenberg Foundation, the Ontario Innovation Trust, the Ontario Ministry for Research and Innovation, Merck & Co.,Inc., the Novartis Research Foundation, the Swedish Agency for Innovation Systems, the Swedish Foundation for Strategic Research and the Wellcome Trust.

356

Human DEAD box RNA helicase DDX5 (p68) (PDB entry 3FE2)
Karlberg, T., Siponen, M., Berglund, H., Collins, R., Dahlgren, L.G., Flodin, S., Flores, A., Graslund, S., Hammarstrom, M., Johansson, A., Johansson, I., Kotenyova, T., Lehtio, L., Moche, M., Nilsson, M.E., Nordlund, P., Nyman, T., Persson, C., Sagemark, J., Thorsell, A.G., Tresaugues, L., Van Den Berg, S., Weigelt, J., Welin, M., Wikstrom, M., Wisniewska, M., Schuler, H. 1 Structural Genomics Consortium , Department of Medical Biochemistry and Biophysics, Karolinska Institutet, Scheeles väg 2, 171 77 Stockholm, SWEDEN

DDX5 (EC=3.6.1.- DEAD box protein 5, RNA helicase p68) belongs to the ATPdependent DEAD box subfamily of RNA helicases that is characterized by the presence of several conserved motifs, including the signature DEAD sequence. The DEAD box RNA helicases play important roles in all aspects of cellular RNA metabolism e.g. pre-mRNA splicing, transcription, ribosome biogenesis, export, translation (1). DDX5 shares 90% protein sequence similarity with DDX17 (p72) in the central core, but the N- and C-terminal extensions are significantly different. DDX5 and DDX17 have been suggested to exist as heterodimers in a variety of complexes in the cell (2). Both proteins act as transcriptional co-activators; this has been shown for the activation of transcription of ER (3), MyoD (4) and p53 tumor suppressor (5). DDX5 and DDX17 may be important for the recruitment of specific components of the transcription machinery, including chromatin remodeling factors, and they may facilitate formation and stabilization of the initiation complex (6). The structure of the DEAD domain of DDX5 was solved and refined to 2.6 Å resolution. It shows the typical DEAD domain structure, with a central 8stranded -sheet sandwiched between 5 -helices on each face, and one ADP molecule in the nucleotide binding pocket. References
1. 2. 3. 4. 5. 6. O. Cordin, J. Banroques, N. K. Tanner, P. Linder, Gene 367, 17 (Feb 15, 2006). V. C. Ogilvie et al., Nucleic Acids Res 31, 1470 (Mar 1, 2003). H. Endoh et al., Mol Cell Biol 19, 5363 (Aug, 1999). G. Caretti et al., Dev Cell 11, 547 (Oct, 2006). G. J. Bates et al., Embo J 24, 543 (Feb 9, 2005). G. Caretti, E. P. Lei, V. Sartorelli, Cell Cycle 6, 1172 (May 15, 2007).

1 Structural Genomics Consortium is a registered charity (number 1097737) that receives funds from the Canadian Institutes for Health Research, the Canadian Foundation for Innovation, Genome Canada through the Ontario Genomics Institute, GlaxoSmithKline, Karolinska Institutet, the Knut and Alice Wallenberg Foundation, the Ontario Innovation Trust, the Ontario Ministry for Research and Innovation, Merck & Co.,Inc., the Novartis Research Foundation, the Swedish Agency for Innovation Systems, the Swedish Foundation for Strategic Research and the Wellcome Trust.

357

An Intersubunit Active Site between Supercoiled Parallel β Helices in the Trimeric Tailspike Endorhamnosidase of Shigella flexneri Phage Sf6
Jürgen J. Müller,1,4 Stefanie Barbirz,2,4 Karolin Heinle,2 Alexander Freiberg,2,5 Robert Seckler,2 and Udo Heinemann1,3,*
1 2

Max-Delbrück-Centrum für Molekulare Medizin, Robert-Rössle-Str. 10, 13125 Berlin Physikalische Biochemie, Universität Potsdam, Karl-Liebknecht-Str. 24-25, 14476 Golm 3 Institut für Chemie und Biochemie, Freie Universität, Takustr. 6, 14195 Berlin 4 These authors contributed equally to this work. 5 Present address: The University of Texas Medical Branch, Department of Pathology, Keiller Building, 3.144, 301 University Boulevard, Galveston, TX 77555-0609, USA.

Sf6 belongs to the Podoviridae family of temperate bacteriophages that infect gram negative bacteria by insertion of their double-stranded DNA. They attach to their hosts specifically via their tailspike proteins. The structure of Shigella phage Sf6 tailspike protein lacking the head binding domain (Sf6 TSP∆N) was determined to a resolution of 1.25 Å at the EMBL BW7B beamline of DESY (Hamburg), whereas the diffraction data of the Sf6 TSP∆N in complex with the oligosaccharide were measured at the Protein Structure Factory Beamline BL 14.2 at BESSY (Table 1). The final model shows one monomer in the asymmetric unit. The biologically active trimer is built by crystallographic three-fold symmetry. The monomer structure reveals a conserved architecture with a central, right-handed β helix. In the trimer of Sf6 TSP∆N, the parallel β helices form a left-handed, coiled β coil with a pitch of 340 Å. The C-terminal domain consists of a β sandwich reminiscent of viral capsid proteins. Further crystallographic and biochemical analyses show a Shigella cell wall O-antigen fragment to bind to an endorhamnosidase active site located between two β helix subunits each anchoring one catalytic carboxylate. The functionally and structurally related bacteriophage, P22TSP∆N, lacks sequence identity with Sf6 TSP∆N and has its active sites on single subunits. Sf6 TSP may serve as an example for the evolution of different host specificities on a similar general architecture.

Reference
Mueller, J.J., Barbirz, S., Heinle, K., Freiberg, A., Seckler, R., Heinemann, U. (2008) Structure 16, 766-775.

358

Figure 1. Crystal structures of the monomer (left) and biological active crystallographic trimer (right) of Sf6 TSP∆N. (PDB code 2VBK).

Figure 2. Localization of the Endorhamnosidase Active Site of Sf6 TSP∆N (PDB code 2VBM). (A) Difference electron density (contoured at 3σ) observed in a complex of the protein with one repeating unit (RU) of an O-antigen hydrolysis product (a-L-Rhap-(1-3)-b-L-GlcpNAc-(12)-a-L-Rhap-(1-2)-a-L-Rhap). Glu293 (chain C) and Asp247 (chain A) belong to the binding site. (C) An octasaccharide (2 RU) modeled into the binding site with its reducing end reaching the catalytic residues Asp399 and Glu366, which lie on different chains, as indicated (right). Bridging water molecules are colored purple.

359

Table 1. Summary of data collection and refinement statistics of Sf6 TSP∆N, and of the Sf6 TSP∆N-tetrasaccharide complex. Sample Sf6 TSP∆N A. Data collection Resolution (Å) Observed / unique reflections Completeness (%)  Rsym (%) Redundancy B (Å ), Wilson statistics / mean value Mosaicity (°) B. Refinement Unit cell dimensions a / c (Å) R / Rwork / Rfree (%) Overall coordinate error (from Rwork and Rfree) (Å) Rmsd bond lengths (Å) Rmsd bond angles (°) / torsion angles (°) , protein mainchain (Å ) / sidechain (Å ) , protein all atoms (Å ) , solvent and hetero atoms (Å ) Ramachandran: Favored / allowed / outlier(%) C. Structure model Protein: Residue range/atoms Number of non-hydrogen atoms Water molecules / phosphate ions PEG/ethylene glycol/glycerol/tetrasaccharide Mg / Mn
2+ 2+ 2 2 2 2 2

native

complex

15-1.25 (1.27-1.25) 355,196 / 169,746 97.4 (79.8) 11.6 (2.0) 5.1 (25.4) 2.2 (1.9) 9.2 / 11.2 0.16

47-2.0 (2.07-2.0) 205,525 / 41,871 99.7 (97.2) 17.4 (7) 5.5 (21.5) 4.9 (4.0) 17.6 / 11.5 0.2

96.23 / 182.43 12.0 / 11.9 / 14.2 0.031/0.032 0.016 1.864 / 7.69 9.7 / 12.0 10.8 24.5 95.7 / 4.1 / 0.2 112-622 / 3857 4874 665 / 3 2 / 12 / 6 / 0 2/0 44 / 23

95.36 / 182.76 19.2 / 19.0 / 23.2 0.164/0.153 0.012 1.362 / 6.77 19.4 / 21.0 20.2 26.3 95.9 / 3.9 / 0.2 114-622 / 3833 4139 248 / 2 -/-/-/1 1/2 2/-

Alternative conf. / sites of radiation damage

Values for the highest resolution shell are shown in parenthesis. Rsym= Σ|I-| / Σ I. R = Σ||Fo|-|Fc|| / Σ |Fo|, where Fo is the observed and Fc is the structure factor amplitude calculated from the model.

360

Crystal structure of E. coli phage HK620 tailspike reveals that Podoviral tailspike endoglycosidase modules are evolutionarily related
Stefanie Barbirz1*, Jürgen J. Müller2*, Charlotte Utrecht1, Alvin J. Clark3, Udo Heinemann2,4, and Robert Seckler1
Physikalische Biochemie, Universität Potsdam, Karl-Liebknecht-Str. 24-25, 14476, Golm 2 Max-Delbrück-Centrum für Molekulare Medizin, Robert-Rössle-Str. 10, 13125, Berlin 3 University of Arizona, Department of Molecular and Cellular Biology, Life Sciences South, P.O. Box 210106, Tucson, AZ 85721-0106, USA. 4 Institut für Chemie und Biochemie, Freie Universität, Takustr. 6, 14195 Berlin * Authors contributed equally to the work Temperate phages of the λ family possess mosaic genomes as a result of horizontal exchange of genetic material. Their genomes mutate fast, so that structurally and functionally homologous proteins frequently lack detectable sequence homology. The Podoviridae HK620, Sf6 and P22 are evolutionarily related. With their tailspike proteins (TSP) they recognize and cleave their host cell receptor, i.e. the O-antigen of gram-negative bacteria. The trimeric TSP have highly homologous N-terminal particle binding domains but no homology in their receptor binding parts, designated TSP∆N. Single-wavelength anomalous diffraction (SAD) data from HK620 TSP∆N (SeMet) crystals were collected at the selenium peak wavelength of 0.9797 Å under cryo-cooling (100 K) using a MarResearch imaging plate MAR345 at the PSF beamline BL14.2 of the Free University Berlin at BESSY, Berlin. The high resolution native dataset was collected with an ADSC Q4 CCD detector at the beamline ID14-2 at the ESRF (Grenoble, France) (Table1). Crystals of oligosaccharide-HK620 TSP∆N complexes diffracted to 1.6 Å resolution at 100 K at the BESSY BL14.2 beamline using a MAR345 imaging plate (Table 1). We have determined the crystal structure of the TSP∆N of Escherichia coli phage HK620 to 1.38 Å resolution (Fig. 1). The protein has endo-N-acetylglucosaminidase activity and produces hexasaccharides of an O18A1 type O-antigen. Its active site is revealed in the structure of a hexasaccharide bound to HK620 TSP∆N determined at 1.6 Å resolution (Fig.2) (1). As in Sf6 TSP (2) and P22 TSP (3), the main central part of HK620 TSP is a right-handed parallel β-helix. The C-terminal domain forms a β-sandwich, as in Sf6 TSP. Although P22 TSP has a different fold in its C-terminus, sequence alignments with the other TSP C-terminal
1

361

domains reveal conserved motifs (1). We therefore propose that the three TSP share conserved functions in the infection process.

References
1 Barbirz, S., Mueller, J.J., Uetrecht, C., Clark, A.J., Heinemann, U., Seckler, R. (2008) Molecular Microbiology 69, 303-316 Mueller, J.J., Barbirz, S., Heinle, K., Freiberg, A., Seckler, R., Heinemann, U. (2008) Structure 16, 766-775 Steinbacher, S., Seckler, R., Miller, S., Steipe, B., Huber, R. Reinemer, P. (1994) Science 265, 383-386

2

3

A

B

Figure 1. (A) Crystal structure of the biological active crystallographic trimer of HK620 TSP∆N. (PDB code 2VJI). The monomers are coloured red, yellow, orange (B) Crystal structure of a hexasaccharide repeat released from the O-antigen of the host cell E. coli H TD2158 bound to HK620 TSP∆N in the solvent exposed groove of one monomer. The electron density for the bound hexasaccharide was contoured at 3σ. Putative active-site residues and corresponding hexasaccharide model are shown in ball-and-stick representation (PDB code 2VJJ).

Table 1. Summary of data collection and refinement statistics of HK620 TSP∆ and of the HK620 TSP∆N-hexasaccharide complex at pH 6.0.

362

Sample HK620 TSP∆N A. Data collection X-ray source Detector Wavelength (Å) Resolution range (outer shell) (Å) Wavelength (Å) Unique reflections Completeness (%) I / σ(I) Rsym (%) Redundancy Wilson B-factor (Å ) Mosaicity (°) B. Refinement and model statistics Unit cell dimensions a / c (Å) Rwork / Rfree
2

native

complex

ESRF ID14-2 ADSC Q4 CCD 0.9330 20 -1.38 (1.38-1.428) 0.9330 115,675 97.5 (90.2) 12.7 (5.1) 6.4 (26.3) 4.6 (4.0) 14.6 0.36

BESSY BL 14.2 MAR345 0.9500 43.6-1.59 (1.585-1.64) 0.9500 73,235 95.6 (83.5) 26.7 (10.9) 4.1 (12.8) 5.0 (4.1) 10.9 0.30

74.22 / 175.16 0.152 / 0.182 (0.209 / 0.238)

73.91 / 174.59 0.133 / 0.165 (0.260 / 0.334) 0.009 1.30 7.1 9.2 87.6 / 12.2 / 0.2 / 0

Rmsd bond lengths (Å) Rmsd bond angles (°) Rmsd torsion angles (°) Averaged main-chain B-factor (Å2) Ramachandran statistics C. Structure model Protein: Residue range/atoms Hetero compounds: H2O Ca , K , Cl , HPO4
2+ + 2-

0.011 1.55 5.1 17.0 87.8 / 12.0 / 0.2 / 0

112-709 / 4535 935 0, 0, 1, 1 2, 0

112-709 / 4604 731 1, 1, 1, 0 7, 1

cryo components, hexasaccharide

Values for the highest resolution shell are shown in parenthesis. Rsym= Σ|I-| / Σ I. R = Σ||Fo|-|Fc|| / Σ |Fo|, where Fo is the observed and Fc is the structure factor amplitude calculated from the model.

363

Preliminary

X-ray

analysis

of

a

novel

haloalkane

dehalogenase DbeA from Bradyrhizobium elkani USDA94
Tatyana Prudnikovaa, Tomas Mozgac, Pavlina Rezacovad,e, Radka Chaloupkovac, Yukari Satof, Yuji Nagataf, Jiri Bryndad,e, Michal Kutya,b, Jiri Damborskyc, and Ivana Kuta Smatanovaa,b*
a

Institute of Physical Biology University of South Bohemia Ceske Budejovice, Zamek 136, 373 33 Nove

Hrady, Czech Republic
b

Institute of Systems Biology and Ecology, Academy of Science of Czech Republic, Zamek 136, 373 33

Nove Hrady, Czech Republic
c

Loschmidt Laboratories, Institute of Experimental Biology and National Centre for Biomolecular

Research, Faculty of Science, Masaryk University, Kamenice 5/A4, 625 00 Brno, Czech Republic
d

Institute of Molecular Genetics, Academy of Sciences of the Czech Republic, Flemingovo nam. 2, 166

37 Prague, Czech Republic
e

Institute of Organic Chemistry and Biochemistry, Academy of Sciences of the Czech Republic,

Flemingovo nam. 2, 166 37 Prague, Czech Republic
f

Department of Environmental Life Sciences, Graduate School of Life Sciences, Tohoku University, 2-

1-1 Katahira, Sendai 980-8577, Japan
*

Correspondence e-mail: ivas@greentech.cz

A novel enzyme, DbeA, belonging to the family of haloalkane dehalogenases (EC 3.8.1.5) was isolated from Bradyrhizobium elkani USDA94. This haloalkane dehalogenase is closely related to DbjA enzyme from Bradyrhizobium japonicum USDA110 (71% sequence identity), but has different biochemical properties. DbeA is generally less active and has a higher specificity towards brominated and iodinated compounds compared to DbjA. To understand the altered activity and specificity of DbeA enzyme, its mutant variant DbeA1, carrying the unique fragment of DbjA, was constructed. Both the wild type DbeA and the DbeA1 were crystallised using the sitting-drop vapour-diffusion method. The crystals of DbeA belong to the primitive orthorhombic space group P212121, while the crystals of the mutant DbeA1 belong to

364

the monoclinic space group C2. Diffraction data were collected to the resolution of 2.2 Å for DbeA as well as for DbeA1. A mutant of DbeA, designated DbeA1, was constructed using inverse PCR to study the importance of the insertion in the N-terminus of the cap domain for activity and specificity of these enzymes. The fragment
143VAEEQDHAE151

equivalent to the

unique sequence of DbjA that is not present in DbeA and other HLDs (Ikeda-Ohtsubo et al., in preparation) was inserted between D142 and A143 of DbeA. Crystallographic analysis of DbeA and DbeA1 was initiated to understand the structure-function relationships of the wild type and the insertion mutant. Here we report the crystallisation and diffraction data analysis of DbeA and DbeA1.

The purified sample of DbeA protein (Fig. 1) was used for crystallisation experiments at concentration of 4-6 mg.ml-1 in 100 mM Tris–HCl buffer pH 7.5. Sitting-drop procedure (Ducruix & Giege, 1999) for DbeA protein crystallisation was performed in Cryschem 24-well plates (Hampton Research, Aliso Viejo, USA). Reservoir contained 300 – 1000 µl of the precipitant reagent. Each droplet contained 3 µl of DbeA protein solution mixed with 1 µl of reservoir solution (100 mM Tris – HCl pH 7.5, 20% (w/v) PEG 3350 or 4000 and 150 mM calcium acetate). Experiments were carried out at 277 and 292 K temperature. Colourless single needle-shaped crystals with dimensions of approximately 0.05 x 0.05 x 0.3 mm were obtained within two weeks. Crystallisation of DbeA1 protein was performed using the same crystallisation method and conditions with protein concentrations of 6.5–9 mg.ml-1 in 100 mM Tris– HCl buffer pH 7.5. Colourless, well shaped crystals of DbeA1 mutant grew within 5-8 days in drops containing 3 µl of protein solution and 1 µl of reservoir solution composed of 22-25% (w/v) PEG 3350 or 4000, 130 - 150 mM calcium acetate and 100 mM Tris–HCl buffer pH 7.5. Drops were equilibrated over 800 ml reservoir solution.

Freshly prepared single crystals were used to collect diffraction data using synchrotron radiation sources. Diffraction data sets were collected to the 2.2 Å resolution for both the DbeA and DbeA1 (Fig. 3).

365

DbeA protein crystallised in the primitive orthorhombic space group P212121 with unit-cell parameters a = 62.7 Å, b = 121.9 Å, c = 161.9 Å. Scaling and merging of data in resolution range 50 – 2.2 Ǻ resulted in completeness of 92 % and an overall Rmerge of 6.6 %. Diffraction data for the DbeA1 were processed in the centred monoclinic space group C2 with unit-cell parameters a = 133.8 Å, b = 75.1 Å, c = 77.6 Å, α = γ = 90°, β = 92°. Scaling and merging of data in resolution range 50 – 2.2 Ǻ resulted in completeness of 98.9 % and an overall Rmerge of 12.9 %. Evaluation of the crystal-packing parameters indicated that the lattice could accommodate four DbeA proteins in one asymmetric unit with a solvent content of approximately 46 % and two molecules in the asymmetric unit with about 45 % solvent content.

This research was supported by the Ministry of Education of the Czech Republic (LC06010 and MSM6007665808) and by the Academy of Sciences of the Czech Republic (AV0Z60870520 and AV0Z50520514). We thank Uwe Müller for his assistance with data collection at MX 14.1 BESSY beamline in Berlin and the EMBL for access to the X11 beamline at the DORIS storage ring of DESY in Hamburg. We thank Matthew Groves for his help during DbeA1 data processing procedure in XDS program.

References
Chaloupkova, R., Sykorova, J., Prokop, Z., Jesenska, A., Monincova, M., Pavlova, M., Tsuda, Nagata M. Y. & Damborsky, J. (2003). J. Biol. Chem. 278, 52622–52628. Ducruix, A. & Giege, R. (1999). Crystallization of Nucleic Acids and Proteins. A Practical Approach, 2nd ed. Oxford University Press. Ikeda-Ohtsubo, W., Goto, Y., Sato, Y., Ohtsubo, Y., Minamisawa, K., Tsuda, M., Damborsky, J. & Nagata, Y. (2009) (in preparation) Kabsch, W. (1993). J. Appl. Cryst. 26, 795–800. Kulakova, A. N., Larkin, M. J. & Kulakov, L. A. (1997). Microbiology 143, 109–115. Minor, W., Cymborowski, M., Otwinowski, Z., Chruszcz, M. (2006). Acta Cryst. D62, 859866. Murshudov, G. N., Vagin, A. A. & Dodson, E. J. (1997). Acta Cryst. D53, 240–255. Nagata, Y., Hynkova, K., Damborsky, J. & Takagi, M. (1999). Protein Expr. Purif. 17, 299304.

366

Newman, J., Peat, T. S., Richard, R., Kan, L., Swanson, P. E., Affholter, J. A., Holmes, I. H., Prokop, Z., Damborsky, J., Nagata, Y. & Janssen, D. B. (2004). WO 2006/079295 A2. Prokop, Z., Damborsky, J., Oplustil, F., Jesenska, A. & Nagata, Y. (2005). WO 2006/128390 A1. Sato, Y., Natsume, R., Tsuda, M., Damborsky, J., Nagata, Y. & Senda T. (2007). Acta Cryst. F63, 294-296. Vagin, A. & Teplyakov, A. (1997). J. Appl. Cryst. 30, 1022–1025.

367

Figure 1
12 % SDS-PAGE stained by Coomassie Brilliant Blue R-250 monitoring the purity of the protein samples and dissolved crystals. Lanes 1 and 4: SDS-PAGE Standards, high range molecular weight marker (Bio-Rad Laboratories), lanes 2 and 5: 3.0 ng of DbeA and 4.0 ng of DbeA1 samples used for crystallisation experiment, respectively; lanes 3 and 6: dissolved crystals of DbeA and DbeA1, respectively.

Figure 2
Optimised crystals of (a) DbeA and (b) DbeA1derived from B. elkani used for diffraction analysis.

368

Figure 3
Diffraction images of (a) DbeA and (b) DbeA1 crystals.

369

Haloalkane dehalogenase DhaA from Rhodococcus rhodochrous NCIMB 13064
A. Stsiapanava1, M. Strakova 3, J. Damborsky3 and I. Kuta Smatanova1,2
of Physical Biology, University of South Bohemia Ceske Budejovice, Zamek 136, CZ-373 33 Nove Hrady, Czech Republic 2Institute of Systems Biology and Ecology, v.v.i., Academy of Science of the Czech Republic, Zamek 136,CZ-373 33 Nove Hrady, Czech Republic 3Loschmidt Laboratories, Faculty of Science, Masaryk University, Kamenice 5/A4, 62500 Brno, Czech Republic
1Institute

Haloalkane dehalogenases (EC 3.8.1.5), which are members of the α/β-hydrolase fold family, catalyze hydrolytic conversion of a broad spectrum of hydrocarbons to the corresponding alcohols [1]. Dehalogenation is a key step in the aerobic mineralization pathways of many halogenated compounds that represent environmental pollutants. So haloalkane dehalogenases are potentially important biocatalysts with both industrial and bioremediation applications. Moreover, they can be applied as active components of biosensors or in decontamination mixtures for warfare agents [2]. The haloalkane dehalogenase DhaA was isolated from Gram-positive bacterium R. rhodochrous NCIMB 13064 [3]. Besides a wide range of haloalkanes, DhaA can slowly convert serious industrial pollutant 1,2,3trichloropropane (TCP) [4]. The aim of our project is solve structure of DhaA wild type in the complex with TCP. Diffraction data for DhaA were collected to 1.04 Ǻ at the BESSY in Berlin. Crystals belong to the triclinic space group P1. The known structure of the haloalkane dehalogenase from Rhodococcus species (PDB code 1bn6) [5] was used as a template for the molecular replacement. Currently, structure of the DhaA protein is in the process of being further refined and interpreted.

References
[1] Janssen, D.B. (2004). Cur. Opin. Chem. Biol. 8, 150-159. [2] Prokop, Z., Oplustil, F., DeFrank, J. & Damborsky, J. (2006). Biotech. J. 1, 1370-1380. [3] Kulakova, A. N., Larkin, M. J. & Kulakov, L. A. (1997). Microbiology 143, 109–115. [4] Schindler, J. F., P. A. Naranjo, D. A. Honaberger, C.-H. Chang, J. R. Brainard, L. A. Vanderberg, & C. J. Unkefer. (1999). Biochemistry 38, 5772–5778. [5] Newman, J., Peat, T.S., Richard, R., L., Swanson, P.E., Affholter, J.A., Holmes, I.H., Schindler, J.F, Unkefer, C.J., & Terwilliger, T.C. (1999). Biochemistry 28, 16105-16114.

370

A CK2α point mutant without backbone tension at the P+1 loop
J. Raaf1, K. Klopffleisch1, B. B. Olsen2, O.-G. Issinger2, K. Niefind1 Institute of Biochemistry, University of Cologne, Zülpicher Straße 47, D-50674, Köln, Germany 2 Institute of Biochemistry and Molekylær Biology, University of Southern Denmark, Campusvej 55, 5230 Odense, Denmark The Ser/Thr protein kinase CK2 (former name “casein kinase”) is composed of two catalytic subunits (CK2α) attached to dimer of non-catalytic chains (CK2β)1. Together with the cyclin dependent kinases and the mitogen-activated kinases, CK2α is a member of the CMGC family of eukaryotic protein kinases (EPKs)2. CK2 is involved in several important cellular processes like proliferation and apoptosis and its overexpression is associated with a variety of tumors3. In contrast to other EPKs, the regulatory mechanism and the detailed function of CK2 in signaling pathways are not illucidated yet. One remarkable feature of CK2α is that all known structures have the characteristic active kinase conformation including an open activation loop. In CK2α the activation loop is constrained in an open position by the Nterminal segment together with a structural chloride ion and a conserved water cluster. Another structural speciality of CK2α is an unfavorable backbone conformation of Ala193 within the P+1 loop of the substrate binding site, which is found in all CK2α structures published so far4 (see Figure 1). This tension is stabilized by hydrogen bonds between the proximate carbonyl groups of Ala193 and Lys198. In closely related kinases of the CMGC group the P+1 loop tension is released in the inactive state and the equivalent position is responsible for their proline preference at the P+1 position of substrates. As this preference and regulatory mechanism is not valid for CK2α, the backbone strain has no evident function and is regarded as an evolutionary remnant4. a) b)
1

Figure 1: Ramachandran plots of a) hsCK2α1-335 (2pvr4) and b) hsCK2α1-335Ala193Gly. The outlier (Ala193) in the case of hsCK2α1-335 is representative for the backbone strain of position 193 in all published CK2α structures. In the point mutant hsCK2α1-335Ala193Gly this tension is released.

To study the case of a relaxed P+1 loop of a human CK2α C-terminal deletion mutant (hsCK2α1-335)5, we replaced Ala193 by a glycine residue. Crystals (Figure 2) were obtained with 1 M Na K phosphate, pH 5, as precipitant. For cryo protection we used a 1 M Na K phosphate, pH 5; 25% glycerol solution. The crystals with the space group P43212 diffracted to 2.29 Å. An X-ray diffraction data set was collected at beamline 14.1 of BESSY. The structure was refined to a final Rfree-value of 24.04 % and al Rwork-value of 17, 61 %.

371

Figure 2: Protein crystal of hsCK2α1-335Ala193Gly, grown in 1 M Na K phosphate, pH 5.

In the point mutant structure we observed a relaxed backbone in position 193. The hydrogen bonds mentioned above are less strong, so that Lys198 can turn away and releases the backbone (Figure 3). The relaxation is also visible in the corresponding Ramachandran plot (Figure 1). According to this decrease of internal tension we also found an increased thermostability of ∆=+3.1°C of hsCK2α1-335Ala193Gly in comparison to hsCK2α1-335 in differential scanning calorimetry (data not shown).

Figure 3: Detailed view on position 193 where alanine was replaced by glycine. The backbone tension which can be observed in all CK2α structures is stabilized by hydrogen bonds between the flanking carbonyl groups of Ala193 and Lys198. In hsCK2α1-335Ala193Gly Lys198 turns away and releases the backbone. Yellow carbon atoms belong to hsCK2α1-335Ala193Gly, black atoms belong to hsCK2α1-335(2pvr4). Electron densities are drawn with a contour level of 1σ. Dashed lines indicate hydrogen bonds (blue: hsCK2α1-335Ala193Gly, purple hsCK2α1−335), distances are given in Å.

References
1 2

Niefind, K., Guerra, B., Ermakowa, I. & Issinger, O.-G. , EMBO J, 20, 5320–5331, 2001 Hanks, S.K. & Hunter, T. , The FASEB Journal, 9, 576-596, 1995 3 Guerra, B. & Issinger, O.-G , Electrophoresis, 20, 391-408, 1999 4 Niefind, K., Yde, C. W., Ermakova, I. & Issinger, O.-G. , J Mol Biol, 370, 427-438, 2007 5 Ermakova, I., Boldyreff, B., Issinger, O.-G. & Niefind, K., J Mol Biol, 330, 925-934, 2003 This work was supported by the Deutsche Forschungsgemeinschaft (DFG) (grant no. NI 643/1-3).

372

Tomographic insights into evolution. Microtomographic investigations of the external and internal morphology of male and female insect genitalia Seidel, Sophia, Museum für Naturkunde Kühbacher, Markus, Helmholtz Zentrum für Material und Energie Wessel, Andreas, Museum für Naturkunde Hartung, Viktor, Museum für Naturkunde Hilger, Andre, Helmholtz Zentrum für Material und Energie Choinka, G., Helmholtz Zentrum für Material und Energie Manke, Ingo, Helmholtz Zentrum für Material und Energie Riesemeier, Heinrich, BAM Kyriakopoulos, A., Helmholtz Zentrum für Material und Energie Hoch, Hannelore, Museum für Naturkunde

Comparative morphology of genital structures in Butterfly Bugs (Insecta: Hemiptera: Fulgoridae) The Fulgoridae, or butterfly bugs, are conspicuous, fairly large (up to 10 cm) insects with wings that are often brightly colored. They feed on trees and wooden shrubs by sucking plant juices from the phloem. At present ca. 600 species are described worldwide (Nagai & Porion, 1996). The majority of species are tropical (O'Brien & Wilson 1985). Interestingly, neither the degree of their diversity (i.e. the exact number of species) or the boundaries between many of the species already described are sufficiently clear. In many insects, the male genitalia show a complicated structure and differ in closely related taxa. Therefore, they are widely used in taxonomy. A combination of established and new methods like Scanning Electron Microscopy (SEM) and Micro-Computer Tomography (Micro-CT) provides a more detailed view of the genitalia than was possible previously. Synchrotron radiation-based X-ray microtomography allows the acquisition of high resolution images while maintaining the physical integrity of the specimens (which is of immense importance if unique museum specimens, such as holotypes, are to be studied). In the present study we apply these methods to species of the butterfly bug Penthicodes, which is widespread in South-East-Asia (10 species described: Nagai & Porion, 1996).

In the bilaterally symmetrical male genitalia, the parameres and the anal tube enclose and conceal the copulatory organ, the aedeagus.The membranous expandable lobes of the aedeagus are wrinkled in repose. During copulation they are inflated apparently due to pumping haemolymph into the aedeagus. As the female gonoporus is very small, inflation probably happens after insertion of the aedeagus into the female reproductive duct.

In fulgorids, the structure of the aedeagus has never been studied in such detail. It was only occasionally mentioned or drawn in species descriptions, for example from Lallemand (1963). Homologising morphological structures and consequently taxonomy above the species level suffered from the lack of knowledge about the functional morphology.

373

Preparation of specimens. Inflating the genital structures with potassium hydroxide (KOH). For examination the male abdomen is placed in 50 °C KOH solution (10 %) for 45 min (histolysis) and then in distilled water: the difference in osmotic pressure leads to inflation of the membranous parts. After inflation the sample is exposed to ascending ethanol series and then critical-point-dried. Perspectives: High resolution Micro-CT enables us to create a virtual 3D-model without destroying the original specimen, and thus allows us to study unique specimens such as holotypes. Micro-CT will for the first time help us visualize the internal structures of male and female genitalia, contributing to a deeper knowledge of their functional morphology, and eventually, the copulatory mechanism. This in turn, is subject to sexual selection, and hence, at the root of evolutionary change. Preliminary results of this study were presented at - Annual Meeting of the German Zoological Society (DZG, Deutsche Zoologische Gesellschaft), Jena, September 2008 (poster) - Annual BESSY Users´ Meeting, Berlin-Adlershof, December 2008 (poster; awarded 1st prize)

374

Microfluidic Mixers for the Investigation of Rapid Protein Folding Kinetics Using Synchrotron Radiation Circular Dichroism Spectroscopy P. Baumgärtel1, A.S. Kane2,3, A. Hoffmann4, G. Reichardt54, R.Seckler1, O. Bakajin3, B. Schuler4 1 Department of Physical Biochemistry, University of Potsdam, Germany 2 Department of Electrical & Computer Engineering, University of California, Davis 3 BioSecurity and Nanosciences Lab., Lawrence Livermore National Lab., Livermore, CA 4 Department of Biochemistry, University of Zurich, Switzerland 5 Helmholtz-Zentrum Berlin, 12489 Berlin, Germany Funding: BMBF 05KS7IP1 In the last year we worked on optimizing our microfluidic mixer which we have developed for monitoring rapid protein folding reactions using synchrotron radiation circular dichroism (SRCD) spectroscopy. The detailed design and the fabrication process as well as the quantification of the mixing efficiency and performance SRCD spectroscopy measurements can be found in our newest publication1. The SRCD spectroscopy measurements were performed on the small, fast folding, protein cytochrome c (cyt c) at the undulator beamline U125/2-10m NIM at BESSY. Our results show that the combination of SRCD with microfluidic mixing opens new possibilities for investigating rapid conformational changes in biological macromolecules that have previously been inaccessible. Cytochrome c folding kinetics The refolding reaction was initiated by a 4-fold dilution of cyt c in 4 M GdmCl (guanidinium hydrochloride) with refolding buffer to a final GdmCl concentration of 0.8 M. The measurement was performed at a total flow rate of 250 μL/min. At the first measurement position (Figure 1C) in the channel, corresponding to 180 μs after mixing, spectra were taken both for the refolding reaction and under equilibrium start (4 M GdmCl) and end (0.8 M GdmCl) conditions (Figure 2B). The reference spectra at 4 and 0.8 M are typical of an unfolded protein and a folded a-helical protein, respectively. A linear combination of both reference spectra shows that the spectrum acquired 180 μs after the start of refolding corresponds to about 28 ± 6% folded signal. Note that the reference with 4 M GdmCl can be measured down to 205 nm, significantly lower than accessible in conventional stopped-flow CD instrumentation. We also measured the kinetic progress curve of the refolding reaction at 220 nm along the entire observation channel (up to 30 ms, Figure 2A). The refolding trace was fit with a single exponential with a refolding rate of 190 ± 25 s-1. The accessible time window covers 44 ± 2% of the total signal change. A change of 22 ± 2% already occurred in the dead time of our instrument, and 33 ± 2% of the signal is still missing Figure 1: (A) drawing of the optimized SRCD mixer chip. (B) Filter posts are designed to minimize clogging of the mixer. (C) Design details of the optimized serpentine mixer. The smooth curvature of the diffuser minimizes the creation of any recirculation vortices in the observation channel where kinetics are measured.

side channel inlet

center channel inlet

observation channel exit

A
filter post mixer diffuser first measuring position

B

C

375

Figure 2: Synchrotron radiation circular dichroism measurements of cyt c in the mixing device. Kinetic measurements under refolding conditions are show in blue, measurements under equilibrium conditions representing the start (4 M GdmCl) and end conditions (0.8 M GdmCl) are shown in red and green, respectively. (A) Refolding kinetics measured at 220 nm and single exponential fit to the data (black). (B) CD spectra measured at 0.18 ms after mixing and linear combination of 28% native and 72% unfolded spectra (black dashed line). after 30 ms. As expected, the reference spectra under equilibrium conditions are invariant along the channel and show the CD signal characteristic of folded and unfolded protein, respectively. The fast refolding kinetics of cyt c have been studied in great detail,1 which makes cyt c an ideal reference for rapid mixing experiments. The novel combination of SRCD spectroscopy with microfluidic mixing devices opens new opportunities to probe structural changes in biomolecules on previously inaccessible time scales and wavelength ranges. Our device provides an unprecedented dead time for CD experiments in the presence of buffers with high absorption (GdmCl) and clearly demonstrates the potential of the method. Future improvements in the approach described here are to be expected, especially with the advent of a dedicated SRCD beam lines, which is currently under construction at BESSY and will be available for users in the second half of 2010. Higher photon flux will improve the signalto-noise ratio, accelerate data acquisition, thus reduce sample consumption, enable the use of lower protein concentrations, and possibly allow a further extension of the accessible wavelength range and the use of deeper channels. Of particular importance for the combination with microfluidic mixing is the brilliance of synchrotron sources, allowing the efficient focusing of light into microstructures. Also future developments in microfabrication, here we plan to work in cooperation with the “Anwendungszentrum für Mikrotechnik” at the HZB, are thus expected to enable further improvements in dead time and time resolution. [1] Avinash S. Kane, Armin Hoffmann, Peter Baumgärtel, Robert Seckler, Gerd Reichardt, David A. Horsley, Benjamin Schuler, und Olgica Bakajin. “Microfluidic Mixers for the Investigation of Rapid Protein Folding Kinetics Using Synchrotron Radiation Circular Dichroism Spectroscopy.” Analytical Chemistry 80, 9534-9541 (2008)

376

Structural and functional characterization of human Iba proteins
Jörg O. Schulze 1 , Claudia Quedenau 2 , Yvette Roske 1 , Thomas Adam 3 , Herwig Schüler 1 , Joachim Behlke 1 , Andrew P. Turnbull 1 , Volker Sievert 2 , Christoph Scheich 2 , Uwe Mueller 4 , Udo Heinemann 1,5 and Konrad Büssow 2,6
1 Max Delbrück Center for Molecular Medicine, Berlin, Germany 2 Max Planck Institute for Molecular Genetics, Berlin, Germany 3 Institute of Microbiology and Hygiene, Charité Medical School, Berlin, Germany 4 Macromolecular Crystallography, BESSY GmbH, Berlin, Germany 5 Institute of Chemistry and Biochemistry – Crystallography, Free University, Berlin, Germany 6 Department of Structural Biology, Helmholtz Centre for Infection Research, Braunschweig, Germany

Iba2 is a homolog of ionized calcium-binding adapter molecule 1 (Iba1), a 17-kDa protein that binds and cross-links filamentous actin (F-actin) and localizes to membrane ruffles and phagocytic cups. Human Iba1 was found to bind calcium ions in overlay assays, but structures solved by NMR and X-ray crystallography revealed a monomeric, Ca(2+)-free protein. In contrast mouse IBa1 showed a homodimeric protein with Ca(2+). Here, we present the crystal structure of human Iba2 and its homodimerization properties, F-actin cross-linking activity, cellular localization and recruitment upon bacterial invasion in comparison with Iba1.

Figure 1. Cartoon representation of the homodimer Iba2. One subunit of the dimer is rendered in gray, the other subunit is shown in blue with a transparent surface. Important residues in the dimerization interface are depicted in stick representation with oxygen atoms in red and sulfur atoms in yellow.

The Iba2 structure comprises two central EF-hand motifs lacking bound Ca(2+). Iba2 crystallized as a homodimer stabilized by a disulfide bridge and zinc ions. These ions are not located in potential dimerization or oligomerization interfaces. Structurally, the Iba2 monomer is very similar to monomeric human Iba1. Both structures differ significantly only in the conformation of EF-hand2. The homodimer of Iba2 has a unusual small dimerization interface of 360A² corresponding to 5% of the total protein surface. Equilibrium dialysis showed that neither Iba1 nor Iba2 bind Ca(2+) in presence of 100µM Ca(2+). Analytical ultracentrifugation showed that human Iba1 and Iba2 form homodimers
377

under reducing conditions. The presence of Ca(2+) had only marginal effect on dimerization of both Iba proteins. Sedimentation experiments and microscopy detected pronounced, indistinguishable F-actin binding and cross-linking activity of Iba1 and Iba2 with induction of F-actin bundles. Fluorescent Iba fusion proteins were expressed in HeLa cells and colocalized with F-actin. Iba1 was recruited into cellular projections to a larger extent than Iba2. Additionally, we studied Iba recruitment in a Shigella invasion model that induces cytoskeletal rearrangements. Both proteins were recruited into the bacterial invasion zone and Iba1 was again concentrated slightly higher in the cellular extensions. According to our studies, the most outstanding difference between both Iba proteins seems to be their distinct expression patterns in various tissues of the body.

References: Imai Y, Ibata I, Ito D, Ohsawa K & Kohsaka S (1996) A novel gene iba1 in the major histocompatibility complex class III region encoding an EF hand protein expressed in a monocytic lineage. Biochem Biophys Res Commun 224, 855–862. Kawasaki H, Nakayama S & Kretsinger RH (1998) Classification and evolution of EF-hand proteins. Biometals 11, 277–295. Ohsawa K, Imai Y, Kanazawa H, Sasaki Y & Kohsaka S (2000) Involvement of Iba1 in membrane ruffling and phagocytosis of macrophages/microglia. J Cell Sci 113, 3073–3084. Yamada M, Ohsawa K, Imai Y, Kohsaka S & Kamitori S (2006) X-Ray structures of the microglia/macrophage-specific protein Iba1 from human and mouse demonstrate novel molecular conformation change induced by calcium binding. J Mol Biol 364, 449–457

This study was funded by the German Federal Ministry for Education and Research (BMBF) through the Leitprojektverbund Proteinstrukturfabrik’, the National Genome Network (NGFN; FZK 01GR0471, 1GR0472) and with support by the Fonds der Chemischen Industrie to U.H.

378

Crystal Structure of KorA Bound to Operator DNA: Insight into Repressor Cooperation in RP4 Gene Regulation

Bettina Königa, Jürgen J. Müllera, Erich Lankab und Udo Heinemanna,c
a b

Max-Delbrück-Center for Molecular Medicine, Robert-Rössle-Straße 10, 13125 Berlin, Germany Max-Planck-Institute for Molecular Genetics, Ihnestraße 73, 14195 Berlin, Germany c Institute for Chemistry and Biochemistry, Free University, Takustraße 6, 14195 Berlin, Germany

heinemann@mdc-berlin.de

KorA is a global repressor in RP4 which regulates cooperatively the expression of plasmid genes whose products are involved in replication, conjugative transfer, and stable inheritance. The structure of KorA bound to an 18-bp DNA duplex that contains the symmetric operator sequence and incorporates 5-bromo-deoxyuridine nucleosides has been determined by MAD phasing at 1.96-Å resolution (Figure 1, Table I). KorA is present as a symmetric dimer and contacts DNA via a helix-turn-helix motif. Each half-site of the symmetric operator DNA binds one copy of the protein in the major groove. As confirmed by mutagenesis, recognition specificity is based on two KorA side chains forming hydrogen bonds to four bases within each operator half-site. KorA has a unique dimerization module shared by the RP4 proteins TrbA and KlcB. We propose that these proteins cooperate with the global RP4 repressor KorB (1, 2) in a similar manner via this dimerization module and thus regulate RP4 inheritance.

Figure 1. Overall structure of the KorA-OA* complex. (A) Image of the KorA dimer (yellow and red) on a semitransparent molecular surface bound to its operator site, looking down the DNA axis. For clarity, the DNA (gray) is shown in only one of the two orientations present in the complex.

379

Table I. Data collection and refinement statistics Data collection Data set Wavelength (Å) Resolution range (Å)a Total reflectionsa,b Unique reflectionsa Completeness (%)a a Rsym (%)a Rmeas (%)a,c

Peak 0.91991 50-1.96 (2.08-1.96) 229,922 (19,830) 60,942 (7,376)b 92.6 (68.4) 32.2 (12.7) 2.5 (6.5) 2.9 (8.3)

Inflection point 0.92004 50-1.96 (2.08-1.96) 216,313 (18,556) 60,954 (7,347)b 92.5 (68.0) 37.1 (11.4) 2.2 (7.5) 2.5 (9.7)

Native-like merged peak 0.91991 50-1.96 (2.08-1.96) 230,648 (20,085) 31,085 (3,817) 93.3 (70.1) 42.5 (17.4) 2.7 (7.4) 2.9 (8.2)

Refinement Resolution range (Å) 50-1.85 Rwork (%) 17.0 Rfree (%) 20.1 32,666 Reflections in Rwork 1,753 Reflections in Rfree R.m.s. deviations Bond lengths (Å) 0.006 Bond angles (° ) 1.522 No. of atoms Protein and DNA 2961 Water oxygens 317 Ramachandran (%) Most favored 98.8 Additional allowed 1.2 a Values in parentheses are for the highest resolution shell. b Friedel pairs not merged. c Multiplicity-corrected Rsym as defined by Diederichs and Karplus (41).

References 1. Delbrück, H., Ziegelin, G., Lanka, E. and Heinemann, U. (2002) An Src homology 3-like domain is responsible for dimerization of the repressor protein KorB encoded by the promiscuous IncP plasmid RP4. The Journal of Biological Chemistry, 277, 4191-4198. 2. Khare, D., Ziegelin, G., Lanka, E. and Heinemann, U. (2004) Sequence-specific DNA binding determined by contacts outside the helix-turn-helix motif of the ParB homolog KorB. Nature Structural & Molecular Biology, 11, 656-663. Funding This work was supported by the Fonds der Chemischen Industrie.

380

Structural basis of S-norcoclaurine synthase enzymatic activity.
Andrea Ilari 1, Stefano Franceschini 1, Alessandra Bonamore 1, Fabio Arenghi 2, Bruno Botta 3, Alessandra Pasquo 4, & Alberto Boffi 1
1. Dipartimento di Scienze Biochimiche and Istituto di Biologia e Patologia Molecolari (IBPM), CNR, Università “Sapienza” P. Aldo Moro 5, 00185, Rome, Italy. 2. Dipartimento di Studi di Chimica e Tecnologia delle Sostanze Biologicamente Attive, Università “Sapienza” P. Aldo Moro 5, 00185, Rome, Italy. 3. CPC Biotech s.r.l., via dei Mille 111, 80121, Napoli, Italy. 4. ENEA Casaccia Research Centre, Dipartimento BIOTEC, Sezione Genetica e Genomica Vegetale, P.O. Box 2400 I-00100 Rome, Italy
Project partially supported by the MIUR, Italy (FIRB 2003).The European Community - Research Infrastructure Action under the FP6 ‘‘Structuring the European Research Area Programme’’(R II 3-CT-2004-506008) is acknowledged for travel and accommodation support.

Benzylisoquinoline alkaloids are among the most important plant secondary metabolites and include a number of biologically active substances that are widely employed as pharmaceuticals such as morphine, codeine, berberine and papaverine. The enzymatic pathways leading to the amazing diversity of benzylisoquinoline derivatives have been largely unveiled in several plant species. These pathways have been shown to originate from a common route in which the first committed step consists in the stereospecific Pictet-Spengler condensation of dopamine with 4hydroxyphenylacethaldehyde (4-HPAA) to yield the benzylisoquinoline central precursor, Snorcoclaurine. The S-norcoclaurine biosynthesys is catalyzed by the Norcoclaurine synthase (NCS). We solved he first crystallographic structure of NCS (from T. flavum) without substrates and in complex with dopamine and the non-reactive substrate analogue 4-hydroxybenzaldehyde (PHB). A three-wavelength MAD data set was collected from Se-Met -NCS on the BL14-2 beamline at the synchrotron radiation source BESSY, Berlin, Germany, using a CCD detector. Complete data sets (120° of φ rotation each) were collected at the peak (λ= 0.97966 Å), inflection (λ= 0.97984 Å) and remote (λ= 0.97800 Å) wavelengths, at a temperature of 100 K. Each frame was collected with an exposure time of 2 s and a 1.0° oscillation range. The data scaling indicated that the crystal is trigonal (P3121) with the following unit-cell dimensions: a=b=86.31 Ǻ, c=118.36 Å, α=β=90°, λ=120° (1). The structure of NCS without substrates was solved by MAD technique using the program SOLVE (PDB code 2VNE) and the structure of NCS in complex with dopamine and the nonreactive substrate analogue PHB was solved at 2.1 Å resolution by molecular replacement using the monomer B of substrate free NCS as search probe (PDB code 2VQ5) (2). Each monomer shows an accessible cleft, located between the seven stranded antiparallel β-sheets and the three α-helices, that extends through the protein matrix forming a 23.4 Å long tunnel. The wider opening (4.2 Å diameter), is formed by an array of hydrophobic residues and a polar patch composed by Tyr108,131,139 and Glu103 side chains located at the entrance of the cavity. Deeper in the cavity, the side chain of Lys122 protrudes towards the interior of the tunnel forming a “hook” capable of intercepting the carbonyl group of the aldehyde substrate. As shown in Fig.1, the geometry of NCS active site is dominated in addition to Lys122 by the presence of other two strong proton exchanger, Asp141 and Glu110, and of a hydrogen bonding donor, Tyr108. These residues shape the binding site of the two aromatic substrates and dictate the mechanism proposed in Fig.2. The reciprocal orientations of dopamine and PHB and their relationships with neighbouring aminoacid side chains immediately suggest a general acid-base reaction mechanism that matches closely the classical two-steps Pictet-Spengler scheme and

381

eventually leads to the stereospecific ring closure to yield S-no