Abstracts

Dr. Edward A. Stern
University of Washington, Seattle
Avoiding Some Pitfalls in XAFS Analysis
Because analysis of XAFS requires fitting multiple parameters simultaneously in a non-linear way local minima can occur which do not give the correct model. It is important to be able to eliminate the false models and to assess correctly the errors in the correct model. A discussion will be given of physical conditions the correct model should satisfy such as limitations on spatial resolution, positive sigma squared, and minimum bond distances. The importance of including systematic uncertainties in estimating the errors in the final model will be emphasized.

View the HTML file.

---

Dr. Alain Michalowicz
University of Paris XII
Two-Pi Shift and Ligand-Swapping EXAFS Modeling Tricks in Applications to Catalysis and Biology
Since it's emerging in the 70th as major local structural method using synchrotron radiation, EXAFS provided many important results, especially in catalysts and biological chemistry. Unfortunately, among other published results, some reports of EXAFS-determined bond lengths were criticized as chemically unrealistic. The aim of this talk is to show how EXAFS modeling can lead to the systematic artifacts and erroneous nearest neighbor distances. Some of the examples to be discussed are:
- If the energy threshold is free to vary in a fitting procedure, it is not impossible to obtain wrong Delta Eo values, and associated with them false distances. An example of such an error published in the framework of catalysis research will be discussed.
- Structures of the metallic site of metalloenzymes frequently contains inequivalent axial and equatorial ligands. EXAFS modeling can easily confuse these ligands, leading to systematic errors. Modeling tricks allowing to refine the structure will be examined. In both cases we shall discuss the proposed method to avoid the trick and find the most likely solution among all the posssible ones.

Download the PDF file.

---

Dr. Stephen R. Wasserman
MediChem Life Sciences
Principal Component Analysis: Getting an Edge on EXAFS
XAFS, particularly in environmental, biological, and catalytic applications, is often measured for related samples that differ in some experimental variable, such as pH, time of reaction, concentration of ligands, or geological location. Traditionally the analysis of EXAFS from these samples has focused on the individual coordination shells for each sample. Principal component analysis (PCA) supplements the traditional approach through a global view of speciation within the entire series of spectra. The results from PCA offer constraints that can then be applied to the traditional non-linear least-squares analysis. Recent extensions of the method in EXAFS and a comparison to standard fitting scheme will be presented.

Download the PDF file.

---

Dr. Matthew Newville
University of Chicago
The Use of Bond Valence Sums in XAFS Analysis
For many inorganic crystal structures, the bond valence model gives a simple and accurate prediction of bond distances given cation and anion type, valence, and coordination number. Such information is complementary to the information available from XAFS and can be used as a reality check on the results of fitting XAFS data for a wide class of systems. The bond valence model and its use in XAFS analysis will be reviewed and discussed. While this powerful model can be used as an independent check on the results of XAFS analysis, it may also prove valuable as 'prior knowledge' to place constraints on the fitting done in the XAFS analysis itself. Since the bond valence sum may not be perfectly satisfied for all atoms, its use as both an exact constraint and an inexact restraint in XAFS analysis will be compared.

Download the PDF file.

---

Dr. James E. Penner-Hahn
University of Michigan, Ann-Arbor
X-ray Absorption Spectroscopy of Zinc Sites in Proteins. When is a Tetrathiolate not a Tetrathiolate?
Numerous reviews describe the accuracy of EXAFS, and most state that scatterer type can be determined to the nearest row of the periodic table (some claim even better accuracy). Given this, it would seem a simple matter to completely characterize Zn sites by EXAFS since the structural questions of interest center on determining the number of cysteine ligands and the number of low-Z (histidine, carboxylate, water) ligands. Since S and N differ by 8 in atomic number, they should be readily distinguished. The reality, however, is that such distinctions can be extremely difficult, and there are now several examples in which EXAFS-predicted structures were not consistent with subsequent high-resolution protein crystal structures. In order to better understand the precision of Zn EXAFS, we undertook a study of a series of well-defined model compounds with variable sulfur/nitrogen ligation (J. Am. Chem. Soc., 120 (1998) 8401-8409). Through this work, we developed a protocol that can be used to obtain an accurate determination of Zn ligation. The basis of the protocol is careful control of the num-ber of variable parameters used in data analysis. This approach, and its limitations, will be described. Through this work, we have gained insight into possible sources of error in interpretation of EXAFS data.

Download the PDF file.

---

Dr. Bruce Ravel
Naval Research Laboratory
Robust Structural Modeling Using Multi-Edge Refinements
Sometimes the interesting physics in an EXAFS analysis problem is only revealed when the analysis is extended beyond the first shell. This probably precludes the use of simple Fourier filtering techniques, requiring instead the refinement of sophisticated structural models. Among the many tools from the EXAFS bag of tricks that can be applied to a challenging problem is multi-edge refinement. Many materials have two or more edges accessible to an EXAFS measurement. Data can be collected from all accessible edges and refined simultaneously to yield more and better results than are available from the analysis of individual edges. In this talk, I will explain how I applied the tool of multi-edge refinement to Co2MnSi and Co2MnGe. These are two of a class of materials known as half-metallic conductors and can, in principle, support 100% spin-polarized conduction. It is thought that anti-site disorder among the Mn and Si or Mn and Ge ions degrades the magnetic and transport properties of these materials. This anti-site disorder is a subtle effect in the EXAFS data and is difficult to measure with precision using only data collected at a single edge. By co-refining data from two or three edges, we are able to place much tighter error bars on the anti-site disorder parameter in these materials. While FEFF is used to compute fitting standards and FEFFIT for performing the fits, the techniques discussed in this talk can be applied to analysis using any adequate theory and fitting codes.

Download the PDF file.

---

Dr. Ralph G. Nuzzo
University of Illinois, Urbana-Champaign
Emergence of Materials Properties at the Nanoscale
The length scales of interest in nanoscale materials are ideally matched to the characteristic probe sensitivities of EXAFS. This correspondence, along with the development of powerful analytical formalisms and models, empowers studies of these systems at remarkable levels of sophistication. In this talk I will describe the application of EXAFS to the study of systems based on well-defined metal nanoparticles derived from directed chemical synthesis, highlighting modeling methods useful for their characterization. The types of questions we are addressing regarding the very interesting structural and dynamical properties of these materials will be examined. The powerful synergies that develop from the comparison of single particle structural data (as obtained from real space probes based on high resolution electron microscopy) and the information on particle ensembles derived from EXAFS will be discussed.

Download the PDF file.

---

Dr. Daniel Haskel
Argonne National Laboratory
XAFS in Anisotropic Structures: Exploiting Angular Dependence for Better Modeling
"Angular resolved" XAFS refers to the dependence of the absorption coefficient on the relative orientations of the incident electric field and crystallographic axes. We will review the origin of such angular dependence and will discuss sample requirements needed to observe it. The cases of powders, partially aligned powders and single crystals will be used to exemplify the different types of angular averages performed in an XAFS measurement. We will then discuss how to exploit this angular dependence in order to reduce the number of fitting parameters needed to solve the local structure of anisotropic materials. Examples will include high Tc superconductors, magneto-resistance materials and others.

Download the presentation.

---

NSLS home page
BNL home page