On 9/6/06, Ameya Joshi <ajoshi at mines.edu> wrote: > Axel and Kostya, > > Thanks for the suggestions. My first concern was convergence, and as pointed > out by Axel, I did try increasing the cutoff, but even after increasing the > cutoff from 30 to 50 Ryd, the BDE changes only by 0.5 kcal/mol (and makes it > worse), so its safe to say its converged with respect to that. I've also
i agree for the ultra-soft potentials (and even there are differences, as one can kindof define during psp generation, how soft the potential should be) but for the norm-conserving ones used with BLYP, you'll need 70-90ry. and in general you always have to find the optimal cutoff for each pair as the 'hardest' pseudopotential determines the required cutoff. cheers, axel. > confirmed that increasing the box size has no further effect. > > Ameya > > > > > ----- Original Message ----- > From: "Konstantin Kudin" <konstantin_kudin at yahoo.com> > To: <pw_forum at pwscf.org> > Sent: Wednesday, September 06, 2006 4:20 PM > Subject: Re: [Pw_forum] Bond dissociation energy of CO (gas-phase) > > > > --- Ameya Joshi <ajoshi at mines.edu> wrote: > > > > Hi Ameya, > > > >> I'm a recent convert to Espresso. > > > > Yes, it is about time that this crusade has started to bring in some > > fresh converts :-) Now, if only people could get converted away from > > using Microsoft formats such as *.xls ... > > > >>... > >>... > >>... > > > >> Finally, I'd like to share my ultimate motive behind these > >> calculations - I'm looking to calculate the binding energies of small > >> molecules on nickel surfaces. I find that in case of CO, the binding > >> energy is lower by about 15 kcal/mol - the same difference as the BDE > >> - as compared to CASTEP. So my guess is that the absolute energy of > >> CO molecule is not being calculated correctly with my current > >> settings. > > > > It seems that you could always compute the energy of CO on the nickel > > surface, as well as a few Angstroms away from the surface. Then you get > > the energy difference from an identical setup, just different > > geometries. See if you still get the 15 kcal/mol difference. Now, if > > you computing the energy of Ni-CO with respect to (Ni+C+O), that one > > can easily find this 15 kcal error coming from the improperly handled > > reference substances. > > > > Anyway, BDEs are often rather challenging to compute correctly (one > > has to deal with reference atomic states and such), so try to avoid > > doing that if possible, and stick to the smallest possible changes. > > > > And, if you really want to learn about minimizing errors for chemical > > reactions, check out the definition of "isodesmic" : > > http://en.wikipedia.org/wiki/Isodesmic_reaction > > > > Kostya > > > > __________________________________________________ > > Do You Yahoo!? > > Tired of spam? Yahoo! Mail has the best spam protection around > > http://mail.yahoo.com > > _______________________________________________ > > Pw_forum mailing list > > Pw_forum at pwscf.org > > http://www.democritos.it/mailman/listinfo/pw_forum > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > > -- ======================================================================= Axel Kohlmeyer akohlmey at cmm.chem.upenn.edu http://www.cmm.upenn.edu Center for Molecular Modeling -- University of Pennsylvania Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323 tel: 1-215-898-1582, fax: 1-215-573-6233, office-tel: 1-215-898-5425 ======================================================================= If you make something idiot-proof, the universe creates a better idiot.
