This is an XANES (Ga k-edge) data for an oxide system, in which the doped Zn flux is different. It is found that there is a significant difference in the B valley of the spectrum of the maximum doping flow. How to explain this phenomenon?
At the largest Zn concetration (magenta colored curve) at tiny peak in R-space comes up at around 2,4A. You indicated that already by an arrow. This peak is obvioulsly due to the high Zn concentration. But whether that peak is caused by Zn backscatter or from oxygen is hardly to resolve.
From the envelope function of the EXAFS contributing to that peak one could decide low ( e.g. Oxygen) or medium ( e.g. Zn) atomic number. But unfortunately this peak is overlapped by the two strong neighbor peaks, so that a backtransform make no sense.
Fitting (you mentioned that) could be an alternative, but I think the peak at 2,4A is too small for achieving reliable results in this way.
For looking closer at the geometry you should use Zn k-edge XANEX/EXAFS. Due to the quite low Zn concentration I would only expect O and Ga as neighbors atoms, but no significant Zn contribution.
Only the magenta colored line is a bit different. The others nicely coincide.
I won't draw any high sophisticated conclusions from that.
How does the XANES of your undoped sample look like?
What is your (expected) Zn doping concentration ? I guess you have got mainly oxygen and Ga as nearest neighbours and only a very, very tiny amount of Zn.
I won't expect any changes in the Ga k XANES as function of Zn doping as far as doping concentration is very small because the vicinities of only a few Ga atoms are affected of your Zn atoms (numbers of affected Ga atoms and their vicinities are about the number of Zn atoms).
Zn concentration of 16 to 29% is a lot. I expected much less than 1 % due to the term 'doped Zn flux' and 'doping flow' mentioned in your question above.
So a small back scatter contribution of Zn to the XANES could be present. The fact that only very close to the k-edge a difference in the spectra shows up is due to the decreasing factors such as 1/k and the Debye-Waller factor of the EXAFS formula.
Do you have made a Fourier analysis of your data?
Xiwen, your second (real space) plot seems to come from high-quality EXAFS data that must surely extend far further out in energy (or k-space) than the XANES data shown in your first plot!
At the largest Zn concetration (magenta colored curve) at tiny peak in R-space comes up at around 2,4A. You indicated that already by an arrow. This peak is obvioulsly due to the high Zn concentration. But whether that peak is caused by Zn backscatter or from oxygen is hardly to resolve.
From the envelope function of the EXAFS contributing to that peak one could decide low ( e.g. Oxygen) or medium ( e.g. Zn) atomic number. But unfortunately this peak is overlapped by the two strong neighbor peaks, so that a backtransform make no sense.
Fitting (you mentioned that) could be an alternative, but I think the peak at 2,4A is too small for achieving reliable results in this way.
For looking closer at the geometry you should use Zn k-edge XANEX/EXAFS. Due to the quite low Zn concentration I would only expect O and Ga as neighbors atoms, but no significant Zn contribution.
Dear Xiwen - try do apply a principal component analysis on your data - you'll at least find out if there are more than just one species hidden in your data. In a second step, add the species that you expect and repeat the PCA analysis - if the number of species does NOT increase, then you'll at least know that the species added is NOT present in your sample series. By trying several different references (and their spectra) you may get an idea if your interpretation is qualitatively correct or not. Good luck, Dirk
- Badges
- Science topic
- Similar topics
- Engineering
- Petroleum Engineering
- Gas