[Pw_forum] Clarification on Calculating U based on linear response

[Joshua Davis]

Dear QE Users

I want to know how to calculate U based on linear response
<http://journals.aps.org/prb/abstract/10.1103/PhysRevB.71.035105>, but I
am having trouble understanding what occupation values to use from a
calculation for a response matrix.

I have fallowed the tutorials

http://media.quantum-espresso.org/santa_barbara_2009_07/slides-exercices/Handson_ldau.pdf

and

http://hjklol.mit.edu/content/calculating-hubbard-u


Do I use the Trace of the d-electron occupation matrix (like in the first
link) or am I to use sum of all the elements of the matrix (like in the
second link)?

Secondly when calculating U, if the atoms that require a U are not very
close to each other, can you get away with not calculating the off-diagonal
response matrix elements and just invert the diagonal response matrix
elements directly?

I am using QE 5.1.
Thanks for any help
----------------------------------------------------------------------------------------------------------------
Joshua D. Davis

Graduate Assistant
Department of Chemistry
Michigan State University

East Lansing, MI 48824
-----------------------------------------------------------------------------------------------------------------
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20150302/2d136645/attachment.html>