[WIEN] DOS question

[Jorissen Kevin]

Hi Yam Chi Yung,
I have seen similar things, but forgot to put them to the mailing  list or to investigate them further.  I saw discrepancies between the number of electrons inside the MT (as reported in case.scf2) and the integral of the DOS for the same atom.
One point may be that the charges in the MT are (I think) derived from the LAPW-representation in A,B and C coefficients, while the DOS is definitely calculated in APW-representation (using 'D' - coefficients).  The change of representation is not 'exact'.
 
Here's an example for MnO, just a normal calculation, using good RKmax and K-mesh, and gmax of 20 and l,m up to l=8 ; so the charge should definitely be well represented, I guess : 
* output of grep :QTL001 case.scf
s : 2.1134
p : 6.0771
d : 4.774
f : 0.0282
total : 12.999 (sum of the above values actually yields 12.993 ; probably a small rounding effect)
* integrating the local density of states up to the fermi level on a very fine mesh
s : 2.0031
p : 6.0835
d : 4.77 à 4.78 (strong contribution at the fermi level)
f : 0.958
total : 12.90 (sum of the above values actually yields 13.819)
 
 
Regarding the weird number of f valence electrons in the second approach,  remark that in lapw2 when the charge (or rather the wave functions) is converted to APW and is then decomposed into different contributions, f charge is not really treated correctly; so probably we should just forget about that value?
 
Kevin.
PS Your warning about the energy can usually be ignored.
 

	-----Oorspronkelijk bericht----- 
	Van: Yam Chi Yung [mailto:yamcy at yangtze.hku.hk] 
	Verzonden: wo 11/5/2003 3:14 
	Aan: wien at zeus.theochem.tuwien.ac.at 
	CC: 
	Onderwerp: [WIEN] DOS question
	
	

	Dear all wien users,
	
	I tried to calculate the electronic structure of KI. However, when I
	integrate the density of states up to the Fermi level, I found that it
	does not equal to the number of electrons. What was the problem? I also
	got warnings in the SCF cycles:
	
	:ENE  : *WARNING** TOTAL ENERGY IN Ry =       -60104.219684
	
	The SCF converge successfully, is that normal?
	
	
	Regards,
	ChiYung, Yam
	
	
	
	_______________________________________________
	Wien mailing list
	Wien at zeus.theochem.tuwien.ac.at
	http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
	
	

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/ms-tnef
Size: 5670 bytes
Desc: not available
Url : http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20031105/41d1c054/attachment.bin