Firefly and PC GAMESS-related discussion club



Learn how to ask questions correctly


Re^2: Custom hybrid functional

Miro Moman
edelmiro.moman@gmail.com


As I am in a hurry, I decided to give PBE0 a try. The calculation has been running for a couple of days now and it seems it is converging (fingers crossed ;-). So for further experiments, I think I can wait till the release of the next version. Thanks a million anyway.

In addition, following Jim's advice, I have been reading about XMCQDPT2 and I guess I should also give this a try in forthcoming projects.

Kind regards,

Miro

On Tue Jun 22 '10 8:46pm, Alex Granovsky wrote
----------------------------------------------
>Hi,

>this is impossible to do with Firefly version 7.1.G.

>However, the forthcoming version 7.1.H allows one to change
>the fraction of HF exchange used in any supported hybrid
>functional. This was done primary as a tool for stabilization
>graphs based approaches, e.g., to identify and track CT and
>Rydberg states. �

>I can provide you the latest development binaries -
>please contact me directly if you are interested.

>Another point is that P86 correlation functional is not
>implemented so you will need to try some alternatives
>e.g. PBE96 correlation.

>Regards,
>Alex Granovsky
>
>
>On Mon Jun 21 '10 6:11pm, Miro Moman wrote
>------------------------------------------
>>Hello,

>>I am modelling an iron-sulfur cluster. To the best of my knowledge, Szilagyi and Winslow (J. Comput. Chem. 2006, 27, 1385-1397) have carried out the most systematic optimisation of DFT functionals for this kind of systems to date.

>>They recommend using a hybrid functional with 5% HF exchange and 95% density functional exchange supplemented with Perdew's 1986 correlation functional. Can I implement this in Firefly? Which keywords should I use?

>>If defining custom functionals is not possible, which of the available functionals would you recommend?

>>Best regards,

>>Miro
>>


[ Previous ] [ Next ] [ Index ]           Wed Jun 23 '10 12:52pm
[ Reply ] [ Edit ] [ Delete ]           This message read 644 times