[adf-list] Adflist Digest,第23卷,问题1

Mitoraj. Mitoraj在Chemia.uj.edu.pl
Fueb 9 12:34:05 CET 2016

Hi Mark and Tobias,

Indeed, Alexei is right, the remaining 4.3 kcal/mol originates from the 
remaining NOCV pairs;
they are summed at the end of SFO section; in order to
get the corresponding NOCV deformation density one can
substract the most important channels from the overall deformation
density (density-densties of orthogonalized fragments);

very best greetings,

Mariusz







W dniu 2016-02-09 10:18, Alexei Yakovlev napisał(a):
> Hi Mark,
>
>  Tobias is absolutely correct about the source of the -33 vs. -37
> kcal/mol discrepancy. It has nothing to do with the calculation
> accuracy but is caused by the fact that only the top 8 pairs are used
> to obtain the -33.41 number. The other 480 pairs are most likely
> responsible for the remaining 4.3 kcal/mol.
>
>  Alexei
>
> On 08/02/2016 17:23, Marc Raupach wrote:
>
>> Dear Tobias,
>>
>> I am happy that my answer was useful. Nevertheless, I would like to
>> point something out which is maybe still unclear.
>>
>>> Orbital Interactions:
>>> A: -37.77
>>> --------------
>>> Total -37.77 (kcal / mol)
>> This energy is calculated solely by the ETS method in the basis of
>> the orthogonalized fragment orbitals.
>>
>>> ETS-NOCV-decomposition: taken restricted fragments
>>> ===================================================
>>>
>>> 1. Eigenvalue pairs from diagonalization of DeltaP expressed in
>>> Lowdin:
>>> 1 -0.45650 976 0.45650
>>> 2 -0.26709 975 0.26709
>>> 3 -0.26685 974 0.26685
>>> 4 -0.18012 973 0.18012
>>> 5 -0.05494 972 0.05494
>>> 6 -0.05493 971 0.05493
>>> 7 -0.05253 970 0.05253
>>> 8 -0.05240 969 0.05240
>>> 2. Orbital Interaction Energy Contributions from each NOCV pair
>>> (in a.u.)
>>> 1 -0.02352
>>> 2 -0.00958
>>> 3 -0.00957
>>> 4 -0.00848
>>> 5 -0.00078
>>> 6 -0.00078
>>> 7 -0.00026
>>> 8 -0.00026
>>> 3. Orbital Interaction Energy Contributions from each NOCV pair
>>> (in kcal/mol)
>>> 1 -14.76153
>>> 2 -6.01314
>>> 3 -6.00468
>>> 4 -5.32319
>>> 5 -0.49070
>>> 6 -0.48816
>>> 7 -0.16481
>>> 8 -0.16229
>> These energy terms are calculated by the ETS method in the basis of
>> the NOCV orbitals. So an additional basis transformation is carried
>> out, which by itself can introduce some (small, < 1 kcal/mol)
>> errors. Since the sum of the NOCV contributions 1 to 8 does give
>> only -33.41 kcal/mol, there might be some other problems.
>> Possibility 1: the integration grid is too coarse.
>> Possibility 2: the basis set itself is not flexible enough. (Maybe
>> the frozen core is too large)
>> Possibility 3: sometimes the NOCV orbitals with "large" NOCV
>> eigenvalues do not catch (very) diffuse charge flow.
>>
>> So, if you want to try to fill the gap, you could try to make the
>> calculation a bit more accurate. (And hope it is not possiblity
>> 3...there is no solution for that one)
>>
>> Greetings,
>> Marc
>>
>> PS: Symmetry and NOCV might sound interesting but imho it is not.
>> The NOCV combined with symmetry would then be applied to each sub
>> space (for each irreducible representation) of the density matrix
>> and one would not allow for charge flow between the irreducible
>> representations.(which sometimes might be the case - one could argue
>> about that) Without symmetry there is only one "sub" space which
>> contains all the information.
>>
>> _______________________________________________
>> ADFlist mailing list
>> ADFlist at scm.com
>> http://lists.tofoba.com/mailman/listinfo/adflist

-- 
dr Mariusz Mitoraj
Department of Theoretical Chemistry
Faculty of Chemistry, Jagiellonian University
30-060 Krakow, Ingardena 3
Poland


有关Adflist邮件列表的更多信息