[Pw_forum] lambda calculation from the a2F(w) files

Alexander Hernandez alexande at cab.cnea.gov.ar
Thu Feb 1 00:02:17 CET 2007


Dear pw_users,
        does someone know which is the normalization of the a2F function
generated by the espresso 3.2 during an electron-phonon calculation?

We tried to reproduce the lambda values generated by matdyn.f90 using
the a2F(w) files. At first, we used the formula that appears in the
majority of the papers, including the McMillan and the Allen-Dynes
papers that is  (lambda=2*\int a2F/w dw), and we could not obtain the
same results. A quick inspection of the source code suggests that in
order to used the above formula, the a2F values should first be divided
by  2*pi. Or what is the same, one should instead use the formula
(lambda=(1/\pi)*\int a2F/w dw) that seems to be implemented in matdyn.
Is this correct?

Thanks in advance,
Alexander








More information about the users mailing list