UT theoretical chemistry code forum

Dear all,

I have some questions about the two scripts of "dosanalyze.pl and doslplot.pl".

First, when we use dosanalyze.pl to calculate the d band center, how can we choose the energy range? Is it right to choose the energy around the fermi energy, with the beginning and ending point located at the zero DOS? If true, then I have another question. Sometimes deep level state forms, how can we deal with that when trying to include the deep level states?

With respect to the script doslplot.pl, the first question is whether the script adds all the DOS of spin-up d orbits and plot it? Another question is that how can I change the doslplot.pl script to remove the DOS plot for the whole system (the red line).

Any suggestions are greatly appreciated.
Regards
Chris

For calculating a d-band center, you want to choose a window that contains the full d-band, as best you can. If you are looking to correlate the d-band center with some adsorbate binding, it would not be appropriate to include core states. I'm not sure what you mean by deep level states, but again, you probably just want to find the center of the metallic d-band, rather than some specific localized state.

A second point is that you typically correlate the d-band center with some binding energy, but do not attempt to determine absolute values. For that correlation, some of the details of the integration range are less important, as long as you are consistent across whatever set of metals you are looking at.

For the doslplot.pl, it is easy to modify the script or the output. First, since the output is eps, you can delete any line in a program that edits vector graphics files. Then, you can also just modify the gnuplot file ldosplot.gnu, to format the plot as you want and/or add delete items. Finally, you could modify the script and remove the line:

print GNUPLOT " w l lt 3 lw 2.0,\\\n";

in the appropriate block of code.

Dear Graeme,

Thanks a lot for the advice! Another question is that to the best of my knowledge, we can use the d band center value to evaluate the adsorption reactivity only if the total filling electron numbers are similar to each other, but if not, which one is more important, the filling electron numbers or the d band center? are there any other parameters to do better?

If you are moving from left to right across the periodic table, the filling of the d-band should be a very significant factor. It is a little hard to separate this from the d-band center, however, because as you move to a filled d-band, the band center will nessecarily become lower with respect to the fermi level. If you are moving down the periodic table, of course the band with and center will dominate.

Many thanks for the response, I think I get a better understanding of this now!