Hi, everybody! I have a problem about ICOHP. I have calculate the bcc-Fe
with 54 atoms considering spin-polarization, and the results illustrtate
that it is stable with spin-polarization.
spin:
FreeEng =    -30679.235965 eV
Ebs     =     -4376.517824 eV

no spin:
FreeEng =    -30585.048936 eV
Ebs     =     -4182.084151 eV

I calculate the ICOHP both with the following cohp.mpr
fe54
COOP
Fe-Fe
Fe
1.0 4.0
Fe

Then, I use mprop program to obtain the COHP curve data:
mprop -n 10000 -s 0.1 -M -7.2926 cohp
The fermi level is used different in the two case.

Then, I sum up the COHP data to the Fermi level. I found the sum data with
nospin is the same as the Ebs=4182eV, and the spin sum data is about
3900eV, larger than Ebs=4376eV even -4182 eV. This is error!!!!!!!

In file ../siesta-3.2/Util/COOP/Docs/UnitsAndMeaningOfCurves.txt, there is
a description about the ICOHP: "When there are two spin components, two
curves are ouput in the .cohp file, and the sum of the integrals for both
would be E_bs."
When I calculate the first nearest neighbor Fe-Fe bond energy of ICOHP, the
spin is larger (-1.52eV) than no spin (-1.72 eV)-- it's wrong!!!!!!!

I do not know what wrong it is. I do not believe that I have no mistake in
somewhere.

Best wishes to all of you!

Bill

Responder a