[Pw_forum] Electron-phonon coupling

ma Yanming yanming_ma at hotmail.com
Wed Mar 19 16:43:10 CET 2003


Dear PWSCF Users,

I have been puzzled by the calculation of the electron-phonon (EP) coupling 
coefficient lambda using PWSCF for a long time. I have two questions in 
this issue. 
(1)Usually, using PWSCF to calculate EP coupling lambda, we should choose 
many (say one hundred or more, I am not sure. It may depend on system) q 
vectors in first BZ. After we finish the EP coupling calculation at all q 
vectors, we make an average over these q points. Then we can get the lambda 
value. To choose the desired q vectors in first BZ, as I know the q points 
can not be the zone boundary points, because it shares by other BZ. But I 
am not sure If the q points for high symmetry points or along high symmetry 
lines in the first BZ can be chosen.  
My question is how to choose the q points properly in the first BZ? What is 
the rule for this choice? 

(2)I tried to calculate the lambda of BCC Nb. I got almost perfect 
agreement with experimental data for phonon calcualtion at several high 
symmetry q points (say H, P, N) using 16 16 16 k mesh. Then I use 32 32 32 
kmesh to calculate the EP coulping, the calculated linewidth Gamma is quite 
different with experimental results. I was puzzled, It should not be this 
case. because phonon contribute mainly to the electron phonon coupling. 
So I doubt the convergency of this EP coulping calculation. By default, the 
pwscf uses ten step Gaussian broadening to calculate the lambda and 
linewidth to test the convergency. So I tried to increase the Gaussian 
broadening steps to see if the EP coupling calculation converges.In some 
steps it seems that the calculation converges.But for further steps the 
results go away (say several GHZ changes in Linewidth). It doesn't help 
even I increase the steps to 40. So My question is what is the criterion 
for the EP coulping convergency. Is it correct that "more the Gaussian 
broadening steps there are , more reliable the results are? 

Can someone who has any successful experiences on EP coulping lambda 
calculation help me? I will high appericate your help. Note that to 
calculate the EP coulping is one of the most powerful function in PWSCF. 


Best Wishes!


Yanming Ma PhD
Steacie Institute for Molecular Sciences,
National Research Councils of Canada.
100 Sussex
K1A 0R6




_________________________________________________________________
与联机的朋友进行交流,请使用 MSN Messenger:  http://messenger.msn.com/cn  




More information about the users mailing list