The XAFS Model Compound Library

Th XAFS Model Compound Library contains XAFS data on model compounds. The XAFS data is fairly well-documented, and has been processed from raw data files from the beamlines into XAFS spectra. No smoothing, averaging, or deglitching has been done on any of the files. The files are consistently formatted, with title lines (marked with '#') at the top of the file, followed by lines with columns of Energy and XMU (measured in transmission XMU = log(I0/ IT) ), and I0 (the raw counts in the initial ion chamber). Most files have an additional column of XMU for a calibration foil.

By "model compounds", I mean compounds of homogeneous and well-known crystallographic or molecular structure. Each data files in this library has an associated atoms.inp file that can be converted to a feff.inp file using the program ATOMS. See the related Searchable Atoms.inp Archive.

This Library exists because XAFS data on model compounds is useful for several reasons, including comparing to unknown data for "fingerprinting" and testing calculations and analysis methods. The collection here is currently limited, but is growing. The focus to date has been on inorganic compounds and minerals of interest to the geochemical community. We will gladly accept data on other systems.

Table of Contents

Other Useful Links

How to use the Library

The XAFS Data Library contains data files on many model compounds. Most compounds have more than one data file, representing successive scans or scans taken under different conditions. You can get a list of all the files in the library, or (better yet) a list of files that match some search pattern, such as the central atom, or model compound name. You can then select one of these listed files, and view the data or retrieve the data file. Searching is fairly straightforward, and you can probably figure it out faster just by trying it, rather that reading on.

To do a search, type in a string in the space provided, pick one of the "search fields" to tell the program what to search for, and click on the List Matching Files button. You can perform searches on the following fields (chosen with a pop-up menu):

Some of these fields are probably not very useful to search on. You'll probably want to search the "Central Atom" , "Compound Name", or "All Title Lines" fields. You can also list all files in the library, or use a periodic table interface (which uses html forms, and so requires a sufficiently modern browser like Netscape), to select a Central Atom. After clicking on List Matching Files (or on an element in the periodic table), you will see a list of matching files. At this point you can pick one of the files (click on the little diamond-shaped button next to the file name). You can also choose the "Output Format", which changes the resulting data that's shown. To have the selected format sent to the screen, click on the Get File button. The current "Output Formats" (chosen with a pop-up menu) are: The "raw" option sends the raw data file to the screen. See the section on output formats for more details about the contents and form of this file.

The "normalized" option sends pre-edge subtracted, normalized xmu data to the screen. Again, the section on output formats discusses the contents and form of this file.
Currently not working!

The "scan information" option sends a table of information about the file to the screen. Included in this table is a link to the appropriate atoms.inp for this data.

[ top of document | table of contents | other useful links | search the Model Compound Library ]

Output formats

The data files

[ top of document | table of contents | other useful links | search the Model Compound Library ]

Format of the Top of the Data files

The data files in the library are stored in ascii text files with a fairly rigid format.

title name:     Cadmium Carbonate (otavite)
title formula:  CdCO_3
title sites:    Cd1,C1,O1
title refer1:   wyckoff, vol 1, ch VII, p 362

The title lines list "Compound Name" (this first line is the only one looked at for a Compound Name search), chemical formula, atomic sites (listing all sites in the unit cell, and used for an Atomic Sites search), literature reference, Schoenflies version of the space group symbol, and then some extra notes, if needed. An All Title Lines search will look at all these lines.

After the title lines, the space group is listed (in the Hermann-Maguin notation) on its own line. This line is used for a Space Group search. The next line contains the unit cell parameters (in Angstroms) a, b, and c. If b, and c are the same as a, they do not need to be listed. The unit cell parameters alpha, beta, and gamma (in degrees) are listed on the next line. If they have the trivial default values of 90, they do not need to be listed.

The next line contains the value of rmax (in Angstroms) used by ATOMS and written to feff.inp. Currently this value is always 6.0, which is a reasonable default -- I like it more than ATOMS default of twice the near-neighbor distance.

The central atom site is listed on the next line with the keyword core. This must be one of the Site Tags listed in the file (not atomic symbol -- see below), and by default it is the first one listed. Like rmax, this is something you may very well want to change.

The atom line denotes that the rest of the lines will contain Atomic Symbol, x, y, z, and site tag. Note that x, y, and z are the positions in the unit cell of dimension a, b, and c. This means that the order of x, y, and z need to be the same as a, b, and c.

The Site Tag gives a unique string to each of the sites in the unit cell. All the Site Tag's in the archived files consist of the Atomic symbol and an integer, usually as indicated by the literature reference.

[ top of document | table of contents | other useful links | search the Model Compound Library ]

Advanced searching

If you're familiar with the Unix pattern matching program grep, and want to try some more sophisticated searching, read on.

The patterns you type are passed as regular expressions to perl, which then uses its own pattern matching abilities. These abilities are very similar to egrep, an extension of the standard grep. For example, searching the All Title Lines with the string
" zinc[{ }+sulf] "
will return all the files with "zinc sulfide" in the title lines. So would searching for "zinc sulf", by the way (though the number of blank spaces between "zinc" and "sulf" is more restricted). What makes this particularly useful is if you search for
" zinc[^{ }+sulf] ".
This will returns files that contain "zinc" and not "zinc sulf", eliminating all the files with "zinc sulfide" in the title lines and returning a smaller set of file, including pure zinc and zincite.

If there are enough (any) complaints about the "simple" search approach, this whole business will probably be changed.

[ top of document | table of contents | other useful links | search the Model Compound Library ]

How to donate data

To donate data, send e-mail to me at The more complete the information you give me, the faster the data will make it into the Library. Note that data in this Library must have an atoms.inp available (for better or worse, this is my working definition of a "model compound"). Because there are so many possible variations in resolution and calibration in XAFS data, data which has a reference compound (preferrably a metal foil) measured at the same time is preferred.

If you think you have better data on a compound already in the Library, please send it in. If it's not completely clear which data set is better, all will be made available.

[ top of document | table of contents | other useful links | search the Model Compound Library ]

To do list

1. add more data.

2. allow simple data processing (pre-edge, normalization, spline, averaging).

[ top of document | table of contents | other useful links | search the Model Compound Library ]


This library was made in collaboration with Susan Carroll, Peggy O'Day, and Glenn Waychunas, and with support from the Earth and Environmental Sciences Directorate at Lawrence Livermore National Laboratory.

Matthew Newville
CARS, Bldg 434A, APS
Argonne National Lab, 9700 South CASS Ave
Argonne, IL 60439
(630) 252-0431
(630) 252-0443

[ CARS | Advanced Photon Source | Argonne Natl Lab | Univ of Chicago | UWXAFS Project ]