You may want to know what the distribution of different chemical species is in your samples. For instance, imagine you have some sort of sediment sample in which mackinawite (an Fe(II) mineral) is being oxidized to an Fe(III) oxide, like ferrihydrite. In this case, you can collect multiple energy maps at energies where the XANES spectra are very different. Typically, n+1 energies are collected, where n is the number of species present in your samples. So, if you have a combination of mackinawite and ferrihydrite, you should collect maps at three energies. XANES fitting can be applied to multi energy maps to provide the proportion of each chemical species that is present in every single pixel in your image.
XANES Fitting Step-by-step procedure
1. Select reference XANES spectra with which to do the fitting. For instance, imagine some samples contains a mixture of different Fe minerals, referred to as “Ref1”, “Ref2” and “Ref3”. The yellow circles denote energies at which multi energy maps were collected:
2. Load multi energy maps (follow instructions provided above for rapid data importing).
3. Under the Analyze menu, select XANES Fitting. There are three options (under “XANES Fit Options”: LS, NNLS-Fit A, and NNLS-Fit B) for the type of fitting algorithm you can use. LS is the default option. We suggest you try all three.
4. When you select XANES Fitting, a new window will pop up:
5. Select the number of reference XANES spectra you would like to use in the fitting under “Number Standards”. In this instance, we select 3.
6. Select the channels you would like to fit by highlighting (control+left click) all the channels for your element that were collected at different energies. In this example, the following channels are selected: “Fe” (which corresponds to Fe collected at 7122 eV), “Fe7124.0”, “Fe7126.0”, “Fe7130.0”, “Fe7150.0”.
7. Finally, rename the labels “Fit1”, “Fit2” etc to the name of your reference materials, e.g. ferrihydrite, pyrite, siderite, etc. In this case, we have simply called each of the reference compounds “Ref1”, “Ref2” and “Ref3”.
8. The window should now look like this:
9. The next step is to “Get mu Parameters”, under the file menu. The XANES fitting works by linearly combining the intensity of the references XANES spectra at each of the energies from the multi energy maps (i.e. mu(E) at each energy). When you select “Get mu Parameters” SMAK makes a table of the mu intensities from each reference spectrum at each of the energies used in the map.
10. The following window will pop up:
11. Use “Open” to select your data XANES reference file from your computer (Note, this must be a normalized XANES file, e.g. *.mu file from SixPack).
12. Click “Yes” next to “Ref1” to tell SMAK to use that reference in its fitting procedure. Then click ok to return to the main fitting window.
13. You can now see that the table has been filled in with all of the mu parameters. At this point, you can select “Save Parameters” from the file menu to save this table. In future fits, you can then reload this table simply by selecting “Load Parameters” from the file menu.
14. At this point, you can fit the data. This may take a few minutes, so be patient. When the fit is complete, new channels with the names of your references (in this case, “Ref1”, “Ref2”, and “Ref3” will appear in your SMAK channel list in the main SMAK window. You will also find a channel labeled “fiterror” which provides the fit error for each pixel. We suggest you examine this closely to find whether spots have been poorly fit (the z-value in the bottom left hand of the display window will read the fit error).
15.You can now look at a Tri-Color Image of the speciation of Fe in your sample. In this instance, the majority of the sample is described by Ref3, with Ref1 and Ref2 being found as small, discrete particles embedded in the larger matrix: