Interpreting XANES data

This talk provides an overview of the techniques of XANES interpretation. It was heavily influenced by Simon Bare’s excellent XANES talk at the 2008 APS XAFS Summer School. See http://xafs.org/Workshops/APS2008.

Transcript

1. Overview Information Measure Fingerprinting Analysis Theory Summary Interpreting XANES Data Bruce Ravel Synchrotron Methods Group, Ceramics Division Materials Measurement Laboratory National Institute of Standards and Technology & Local Contact, Beamline X23A2 National Synchrotron Light Source 1st ASEAN Workshop on X-ray Absorption Spectroscopy Synchrotron Light Research Institute Nakhon Ratchasima, Thailand 29–31 July 2010 1 / 37 Interpreting XANES Data
2. Overview Information Measure Fingerprinting Analysis Theory Summary Copyright This document is copyright c 2010-2011 Bruce Ravel. This work is licensed under the Creative Commons Attribution-ShareAlike License. To view a copy of this license, visit http://creativecommons.org/licenses/by-sa/3.0/ or send a letter to Creative Commons, 559 Nathan Abbott Way, Stanford, California 94305, USA. You are free: to Share  to copy, distribute, and transmit the work to Remix  to adapt the work to make commercial use of the work Under the following conditions: Attribution – You must attribute the work in the manner specified by the author or licensor (but not in any way that suggests that they endorse you or your use of the work). Share Alike – If you alter, transform, or build upon this work, you may distribute the resulting work only under the same, similar or a compatible license. With the understanidng that: Waiver – Any of the above conditions can be waived if you get permission from the copyright holder. Public Domain – Where the work or any of its elements is in the public domain under applicable law, that status is in no way affected by the license. Other Rights – In no way are any of the following rights affected by the license: Your fair dealing or fair use rights, or other applicable copyright exceptions and limitations; The author’s moral rights; Rights other persons may have either in the work itself or in how the work is used, such as publicity or privacy rights. Notice – For any reuse or distribution, you must make clear to others the license terms of this work. This is a human-readable summary of the Legal Code (the full license). 2 / 37 Interpreting XANES Data
3. Overview Information Measure Fingerprinting Analysis Theory Summary Acknowledgment This talk is heavily influenced by Simon Bare’s excellent XANES talk at the 2008 APS XAFS Summer School. See http://xafs.org/Workshops/APS2008. 3 / 37 Interpreting XANES Data
4. Overview Information Measure Fingerprinting Analysis Theory Summary Acronyms XANES X-ray Absorption Near-Edge Structure NEXAFS Near-Edge X-ray Absorption Fine Structure XANES and NEXAFS are exactly the same thing. Historically, the soft X-ray community says “NEXAFS” while the hard X-ray community says “XANES”. Both acronyms refer to the portion of the XAS (X-ray Absorption Spectroscopy) measurement in the vicinity of the absorption edge. The Extended X-ray Absorption Fine Structure is oscillatory data extending hundreds of volts above the edge. 4 / 37 Interpreting XANES Data
5. Overview Information Measure Fingerprinting Analysis Theory Summary Some vocabulary Words commonly used to describe specific parts of the XANES spectrum. pre-edge Small (or large, certainly meaningful!) features between the Fermi energy and the threshold edge The main rising part of XAS spectrum near-edge Characteristic features above the edge white line Large, prominent peak just above the edge, particularly in L or M edge spectra 5 / 37 Interpreting XANES Data
6. Overview Information Measure Fingerprinting Analysis Theory Summary XANES Publications Interest in XANES and its interpretation has grown steadily over the the last 3 decades as XAS has become more available and adopted by more scientific disciplines. Results of a July 2010 search at ISI Web of Knowledge for the word “XANES”: 6 / 37 Interpreting XANES Data
7. Overview Information Measure Fingerprinting Analysis Theory Summary Speciation at a glance: Coordination Here is Cr K edge data for tetragonally coordinated, hexavalent K2CrVIO7 and hexagonally coordinated, trivalent CrIII 2 O3. Trivalent Cr is insoluble and non-toxic. Hexavalent Cr is readily soluble and highly toxic. K2CrO7 Cr2O3 It is very easy to tell “good” Cr from “bad” Cr in a XANES measurement. 7 / 37 Interpreting XANES Data data from http://cars9.uchicago.edu/~newville/ModelLib/search.html
8. Overview Information Measure Fingerprinting Analysis Theory Summary Speciation at a glance: Crystallinity SiO2 is found in two forms∗ under standard conditions: crystalline (the mineral quartz) and amorphous (common glass). Again, these are readily distinguished by a XANES measurement. 8 / 37 Interpreting XANES Data ∗ Wikipedia identifies 14 other metastable, high-T, or high-P forms of SiO2.
9. Overview Information Measure Fingerprinting Analysis Theory Summary Speciation at a glance: Oxidation There is an 11 eV shift from S2− to S6+ with lots of variation among species. S speciation is of importance across a broad range of disciplines, including life science, catalysis, petroleum science, photovoltaics, environmental science and more. P and Cl are similarly rich in their XAS. XAS of S, P, and Cl are a particular strength of BL8 ¨ 9 / 37 Interpreting XANES Data Sulfur K-edge x-ray absorption spectroscopy of petroleum asphaltenes and model compounds, G.N. George, M.L. Gorbaty, J. Am. Chem. Soc. (1989) 111:9, 3182 DOI: 10.1021/ja00191a012
10. Overview Information Measure Fingerprinting Analysis Theory Summary Oxidation and edge position There is a relationship between formal valence of a metal and the position of the edge in the XANES spectrum. Here is Re metal along with 4+, 6+, and 7+ oxides of Re. The shift to higher energy is, to first order, a Coulomb effect. Less charge on the atom means less screening of the core. 0 0.5 1 1.5 2 2.5 3 3.5 10510 10520 10530 10540 10550 10560 10570 10580 10590 10600 Normalized Absorption Energy (eV) Re0 metal Re4+O2 Re6+O3 NH4 Re7+O4 0 1 2 3 4 5 6 7 Edge position (eV) Formal valence Some more examples: Mo S.P. Cramer et al. J. Am. Chem. Soc., (1976) 98:5, pp 1287 V J. Wong et al. Phys. Rev. B30, 5596–5610 (1984) 10 / 37 Interpreting XANES Data Simultaneous XAFS measurements of multiple samples, B. Ravel, C. Scorzato, D.P. Siddons, S.D. Kelly and S.R. Bare, J. Synchrotron Rad. (2010) 17, 380-385 doi:10.1107/S0909049510006230
11. Overview Information Measure Fingerprinting Analysis Theory Summary Mixed phases Here we see trivalent V2O3, pentavalent V2O5 and an unknown Vanadium compound plotted together. Like in the Cr example, we see a distinct difference between 6-coordinated and 4-coordinated V. Our unknown is partially reduced, as can be seen by the reduction in pre-edge peak and the left-ward shift of the main edge. Later we will discuss ways of determining the content of the unknown. 11 / 37 Interpreting XANES Data
12. Overview Information Measure Fingerprinting Analysis Theory Summary Evolution of redox state The edge features are often large enough that their evolution can be measured in an in situ experiment. Here we see the kinetics of CrIII → CrVI oxidation by Mn oxide over the course of four minutes of reaction time. Each scan was measured in 3 second. The in situ experiment could involve a chemical reaction, a change in temperature, electrochemical cycling, and so on. 12 / 37 Interpreting XANES Data Kinetics of Chromium(III) Oxidation by Manganese(IV) Oxides Using Quick Scanning X-ray Absorption Fine Structure Spectroscopy (Q-XAFS), G. Landrot, M. Ginder-Vogel, and D.L. Sparks, Environ. Sci. Technol., (2010) 44:1, pp 143-149 DOI: 10.1021/es901759w
13. Overview Information Measure Fingerprinting Analysis Theory Summary Ligands We see a significant edge shift between aqueous As3+ and aqueous As5+ , as we expect. Note that the As3+ and As5+ methionine solutions are similar, but shifted to lower energy. The same shift is seen between divalent hematite (Fe2O3) and the nominally divalent pyrite (FeS2). 13 / 37 Interpreting XANES Data
14. Overview Information Measure Fingerprinting Analysis Theory Summary Sample preparation Sample preparation for XANES is fairly forgiving and usually pretty easy. Here are three examples: 1 A Spex XRF vessel is great for solution samples. 2 A simple frame and tape is fine for powders. 3 A carbon tape with powder sprinkled on is useful for lower energy edges. Ideally, the sample is homogeneous, but in practice almost anything can be used. For example, cultural heritage samples are often placed as is in the beam. 14 / 37 Interpreting XANES Data
15. Overview Information Measure Fingerprinting Analysis Theory Summary Sensible scan parameters Through the edge, the measurement grid must be very fine to adequately measure the quickly changing data. If using a third derivative analysis (see the reference below), your data grid through the edge must be much finer than the experimental resolution. Enough of the pre-edge and post-edge must be measured to do a good job normalizing the data. region step size pre-edge 5 eV edge 0.25 eV EXAFS 0.05 ˚ A −1 0.25 eV is reasonable for ∼5 keV. A larger step is OK at higher energy, a smaller step might be needed for lower energies. 15 / 37 Interpreting XANES Data Sulfur K-edge x-ray absorption spectroscopy of petroleum asphaltenes and model compounds, G.N. George, M.L. Gorbaty, J. Am. Chem. Soc. (1989) 111:9, 3182 DOI: 10.1021/ja00191a012
16. Overview Information Measure Fingerprinting Analysis Theory Summary Self-absorption As µ(E) changes through the edge and fine-structure, the penetration depth of the sample changes, thus changing the sample volume contributing to the fluorescence measurement. This has the effect of attenuating the XAS. (See A. Manceau, M.A. Marcus, N. Tamura (2002) Reviews in Mineralogy and Geochemistry, 49, 341-428.) Zirconolite CaZrTi2O7 powder in transmission and a sintered pellet in fluorescence. The data can be approximately corrected for the effect of self-absorption using . 16 / 37 Interpreting XANES Data D.P. Reid, et al., Nuclear Instruments and Methods in Physics Research B, 268:11-12, (2010) 1847 DOI: 10.1016/j.nimb.2010.02.026
17. Overview Information Measure Fingerprinting Analysis Theory Summary Avoiding self-absorption It is better to avoid than to correct! The best approach is to prepare your sample in a way that will see a negligible self-absorption effect. Measure in transmission, if possible. Make your sample thin compared an absorption length. Make your sample dilute in absorber concentration. If possible, measure a standard that can be used to guide the self-absorption correction, as shown on the previous slide. Your sample is your sample Sometimes self-absorption cannot be avoided. ¨ 17 / 37 Interpreting XANES Data
18. Overview Information Measure Fingerprinting Analysis Theory Summary Dead time When using an energy discriminating detector, data can be distorted due to “pile-up”, which is due to photons arriving faster than the discriminating electronics can process them. SrMnO3 measured at a very high count rate with a Si-drift detector, compared to data measured with a Stern-Heald detector. Data can be accurately corrected to restore the true measurement. 18 / 37 Interpreting XANES Data Performance of a four-element Si drift detector for X-ray absorption fine-structure spectroscopy, J. C. Woicik, B. Ravel, D. A. Fischer and W. J. Newburgh, J. Synchrotron Rad. (2010) 17, 409-413, DOI: 10.1107/S0909049510009064
19. Overview Information Measure Fingerprinting Analysis Theory Summary Fingerprinting Fingerprint, tr.v. To identify by means of a distinctive mark or characteristic. One of the most powerful uses of XANES data is to simply identify what is in front of the beam. Looking back at the CrIII /CrVI example, what might you say about the valence of the chromium contained in coal combustion residue? K2CrO7 Cr2O3 19 / 37 Interpreting XANES Data Quantifying Hazardous Species in Particulate Matter Derived from Fossil-Fuel Combustion, F.E. Huggins, et al., Environ. Sci. Technol. (2004) 38:6, 1836 DOI: 10.1021/es0348748
20. Overview Information Measure Fingerprinting Analysis Theory Summary Categorizing spectra In an study of Ti-containing standard materials, the different coordination environments were found to aggregate when plotting pre-edge peak height v. peak position. Here we see the data from the reference below along with Ti K-edge data from various Zirconolite (CaZrTi2O7) samples, including the one from the self-absorption slide. 20 / 37 Interpreting XANES Data Coordination chemistry of Ti (IV) in silicate glasses and melts: I. XAFS study of titanium coordination in oxide model compounds , Geochimica et Cosmochimica Acta, 60:16, 3023, 1996, DOI: 10.1016/0016-7037(96)00144-5
21. Overview Information Measure Fingerprinting Analysis Theory Summary XANES and disorder The details of the XANES can often give information about structural disorder about the absorbing atom. EuTiO3 is a true cubic perovskite. PbTiO3 is a tetragonally distorted perovskite with substantial disorder in the oxygen octahedron. Consequently, the pre-edge peak is much larger for PbTiO3. 21 / 37 Interpreting XANES Data Local structure and the phase transitions of BaTiO3 B. Ravel, E. A. Stern, R. I. Vedrinskii, V. Kraizman Ferroelectrics, 206:1 (1998) 407, DOI: 10.1080/00150199808009173
22. Overview Information Measure Fingerprinting Analysis Theory Summary Why are disorder and the pre-edge peak related? XAS is a dipole transition. The photoelectron changes angular momentum by 1: ± 1. For a K-edge spectrum, the initial state is s: = 0. Thus the final state must be = 1. Ti has a filled p shell but a completely empty d shell. With centro-symmetry, as in a true perovskite, the p and d states cannot hybridize. Broken symmetry leads to mixing of p and d states around the Fermi level. Disorder-driven admixture of d character results in an enhanced pre-edge peak. 22 / 37 Interpreting XANES Data
23. Overview Information Measure Fingerprinting Analysis Theory Summary Analysis There are a number of ways to get quantitative results from XANES spectra: Linear Combination Fitting Interpret data by comparison with standards Peak Fitting Fit Gaussian and Lorentzian line-shapes to the peaks in XANES data Principle Components Analysis Decompose an ensemble of data into a mathematical basis Difference Spectra Subtract one normalized spectrum from another Software has mature implementations of LCF and difference spectra. It recently gained a PCA implementation. Peak fitting has been implemented in but is not yet in the GUI. 23 / 37 Interpreting XANES Data
24. Overview Information Measure Fingerprinting Analysis Theory Summary LCF The working assumption of LCF The spectrum from an unknown sample can be understood as a linear superposition of the spectra of two or more known samples. That is: + = + LCF requires: 1 A complete set of standards (i.e. every possible constituent) 2 High quality measurements of each standard 3 Data and standard are aligned to a common absolute energy grid and processed (i.e. normalization and background removal) consistently 24 / 37 Interpreting XANES Data
25. Overview Information Measure Fingerprinting Analysis Theory Summary Economic geology (I) One way that gold deposits form is by having Au chloride fluids rise from the deep earth, wash over cyanobacteria colonies, and eventually reduce to metallic gold. Au3+Cl Before Exposed After We simulated this process at the beamline by exposing cyanobacteria to an Au3+ solution and “watching” the evolution of the Au XAS from Au3+ to Au0 . Questions What is the rate constant? Is there an intermediate species? 25 / 37 Interpreting XANES Data M. Lengke et el., Mechanisms of Gold Bioaccumulation by Filamentous Cyanobacteria from Gold(III)-Chloride Complex, Environ. Sci. Technol. 40(20) p. 6304-6309. (2006)
26. Overview Information Measure Fingerprinting Analysis Theory Summary Economic geology (II) We see that 7 minutes after injection, the data strongly resemble the Au3+ Cl. After one week, the data resemble Au metal. Over the course of the time series, the white line ∼ 11921 shrinks while the bump ∼ 11945 grows, suggesting the reduction to Au metal. 26 / 37 Interpreting XANES Data M. Lengke et el., Mechanisms of Gold Bioaccumulation by Filamentous Cyanobacteria from Gold(III)-Chloride Complex, Environ. Sci. Technol. 40(20) p. 6304-6309. (2006)
27. Overview Information Measure Fingerprinting Analysis Theory Summary Economic geology (III) We can analyze these data as a linear combination of species, including Au3+ Cl, Au metal, and Au1+ sulfide. We plot the contributions from these species as a function of time to see reaction rates. 27 / 37 Interpreting XANES Data M. Lengke et el., Mechanisms of Gold Bioaccumulation by Filamentous Cyanobacteria from Gold(III)-Chloride Complex, Environ. Sci. Technol. 40(20) p. 6304-6309. (2006)
28. Overview Information Measure Fingerprinting Analysis Theory Summary Peak fitting The working assumption of peak fitting A spectrum can be meaningfully deconstructed into a set of step-like (atan or erfc) and peak (Gaussian, Lorentzian, Voight) functions. In this case, various Gaussians are interpreted as being the contributions from differently S valence states, with thiol and sulfonate dominating in a humic material from an aquifer in northern Germany. This sort of analysis is most meaningful when performed across an ensemble of related data. The drawback is that the physical significance of the line-shapes is sketchy, at best. 28 / 37 Interpreting XANES Data Origin and mobility of fulvic acids in the Gorleben aquifer system, T. Sch¨ afer, et al., Organic Geochemistry, 36:4, (2005) 567, DOI: 10.1016/j.orggeochem.2004.10.011
29. Overview Information Measure Fingerprinting Analysis Theory Summary Principle Components (Factor) Analysis The working assumption of PCA An ensemble of data represents a number of physical components that is equal to or smaller than the size of the ensemble. 1 Decompose the ensemble of spectra into an orthogonal set of eigenvectors. These are purely mathematical vectors that do not individually represent physical components of the data. 2 When done correctly, the number of significant eigenvalues equals the number of physical components of the system. 3 Target transforms can be done to test potential physical components against the eigenvectors. 4 Great care must be taken to align and normalize the data properly. Any errors in data processing show up as additional non-negligible eigenvectors. 5 has a brand new implementation. 29 / 37 Interpreting XANES Data S.R. Wasserman, J. Phys. IV France (1997) C2-203-C2-205; S.R. Wasserman et al., J. Synchrotron Rad. (1999) 6, 284-286; references therein
30. Overview Information Measure Fingerprinting Analysis Theory Summary Difference Spectra Difference spectra Subtract one spectrum from another. The most common use is for X-ray Magnetic Circular Dichroism (XMCD) The areas under the difference spectra tell you about moment and magnetic ordering. Difference spectra can also be used to highlight a subtle change in a data sequence. Here, Pt nanoparticles are being covered with H. 30 / 37 Interpreting XANES Data X-ray magnetic circular dichroism study on CeFe2, A. Delobbe, et al., Europhys. Lett. 43 320 (1998), DOI: 10.1209/epl/i1998-00359-2; Pt data courtesy of Simon Bare
31. Overview Information Measure Fingerprinting Analysis Theory Summary Feff is a widely popular, real-space, multiple scattering code used for XANES, EXAFS, and other spectroscopies. It is fully integrated into and not hard to use on its own. A GUI exists. Simulating XANES data is not hard, although much thought needs to go into the cluster and the parameters to make the calculation useful for comparison with real data. is a single electron, dipole approximation code  multi-electron excitations, quadrupole contributions to the XANES, and multiplet effects are not calculated. Works on all computer platforms. homepage: http://leonardo.phys.washington.edu/feff/ 31 / 37 Interpreting XANES Data
32. Overview Information Measure Fingerprinting Analysis Theory Summary Multiplets One electron codes can fail for systems with strong overlap between the core and valence states. Transition metal L edges are a good example: strong overlap between the 2p and 3d bands. Here is an Fe L2,3 calculation for a low-spin ferric complex. The quantum chemical calculation is made to find the many discrete states of the system and an instrumental/lifetime broadening is applied. One popular multiplet code is by Frank de Groot. Windows only. 32 / 37 Interpreting XANES Data Multiplet effects in X-ray spectroscopy, F. de Groot, Coord. Chem. Rev. 249:1-2 31 (2005), DOI: 10.1016/j.ccr.2004.03.018
33. Overview Information Measure Fingerprinting Analysis Theory Summary FDMNES The Schr¨ odinger equation is solved in a discrete form on a three-dimensional grid using a finite difference method. From this, many calculations are possible, including XAS, XES, RIXS, anomalous scattering, and others. This approach avoids certain shortcomings of 8 in highly non-centro-symmetric systems. Windows and linux. FDMNES homepage: http://www.neel.cnrs.fr/spip.php?article876 33 / 37 Interpreting XANES Data X-ray absorption near-edge structure calculations beyond the muffin-tin approximation, Y. Joly, Phys. Rev. B 63 125120 (2002), DOI: 10.1103/PhysRevB.63.125120
34. Overview Information Measure Fingerprinting Analysis Theory Summary FitIt Fit XANES spectra by multidimensional interpolation using or calculations. This software helps you set up the calculations and parametrize your fit. Windows only. FitIt homepage: http://www.nano.sfedu.ru/fitit.html 34 / 37 Interpreting XANES Data
35. Overview Information Measure Fingerprinting Analysis Theory Summary Using theory effectively Convergence in cluster size. Near the edge, the mean free path of the photoelectron is quite large. The only way to know the proper cluster size is the do the calculations. Often the theory is an imperfect match to the data, but a sequence of calculations shows a same trend that can be related to the data. In this case, a sequence of calculations on PbTiO3 shows the relation between pre-edge peak and amount of distortion. 35 / 37 Interpreting XANES Data
36. Overview Information Measure Fingerprinting Analysis Theory Summary Summary XANES is a much larger signal than EXAFS Good XANES spectra can be collected at lower concentrations and with less-than-perfect samples. XANES is easier to crudely interpret than EXAFS In many cases, XANES can be a fingerprinting tool. For many systems, the XANES analysis using known spectra from model compounds is sufficient. XANES is harder to fully interpret than EXAFS The exact physical and chemical interpretation of all spectral features is still difficult to do accurately, precisely, and reliably. This situation is improving... 36 / 37 Interpreting XANES Data
37. Overview Information Measure Fingerprinting Analysis Theory Summary Notes 37 / 37 Interpreting XANES Data