Firefly and PC GAMESS-related discussion club


 
Learn how to ask questions correctly  
 
 
We are NATO-free zone
 



Re^4: Strange B3lYP geometry opt for EtSnCL3 with LANL2DZdp ecp and basis for Sn and Cl

olgerdovich
olgerdovich@yandex.ru


In this case I'd like to point out another feature which (I hope) is worth your attention.
It appears that once ECP is used, there is no checking for consistensy in ICHARG and MULT variables for the calculated species and the type of wavefunction (RHF or UHF). That is, running (accidentelly) a run with SnCl6 dianion with scftyp=rhf, icharg=-1 and mult=1 I find no alarms and the run is stopped just by exhausting of allowed number of SCF-iterations (actually, it is not surprising that for radical species SCF does not converges with RHF). It appears to be a problem of the same sort as one discussed above.

Thank you for your kind attention.


On Thu Jul 28 '16 3:43pm, Alex Granovsky wrote
----------------------------------------------
>Dear olgerdovich,

>>I'd like to offer to add such an option in later version of firefly.

>I'm sorry for some delay with my reply. Thanks for your suggestion.
>We'll take a look at this problem.

>All the best,
>Alex Granovsky
>
>
>
>
>
>
>On Sun Jul 24 '16 10:23pm, olgerdovich wrote
>--------------------------------------------
>>Wonderful!
>>Thank you a lot. I was sure that the problem was in my input and lack of experience in composing such inputs, but I couldn't realise that it was so silly mistake. The forth chlorine came from test run win tin tetrachloride - I just forgot to remove it in next task.

>>Eventually, I overcome this problem by using Chemcraft feature of composing basis and ecp section, but your kind reply at least explains me that doing it manually is simpler (and is conceivable, at least) than I was afraid it was. Just add a little bit more attention.

>>I'm somewhat confused that my mistake in ECP-group passes input processing: it appears that mismatch of the number of atoms and the number of lines could be easily found, let alone an attempt to remove 10 electrons from carbon with its total six electrons. It is very strange for me that there is no check for consistency between $ecp and $data groups.

>>I'd like to offer to add such an option in later version of firefly.
>>
>>
>>
>>On Sun Jul 24 '16 9:42pm, Pedro Silva wrote
>>-------------------------------------------
>>>The ECP group in your input contains 4 chlorine atoms, whereas the $data deck only contains 3.  FF tries to remove 10 electrons from the 4th chlorine (which happens to be a carbon instead of a Cl), and all hell breaks loose :-) After removing the 4th Cl from the ECP deck, partial optimization yields the following very reasonable geometry after ca. 45 steps. RMS is still around 5e-4, but you may proceed from here :-)
>>>
>>>
>>> $DATA
>>> SnC2H5Cl3
>>>C1
>>>TIN        50.0      0.084818635         0.569199412        -2.137983974
>>>CHLORINE   17.0     -1.561661805        -1.078523340        -2.313366895
>>>CHLORINE   17.0      0.636333799         1.190641688        -4.320186432
>>>CHLORINE   17.0      1.972179867        -0.453258648        -1.225126104
>>>CARBON      6.0     -0.627727268         2.222290346        -1.001772764
>>>HYDROGEN    1.0     -0.602346555         1.902787799         0.044500035
>>>HYDROGEN    1.0      0.110035103         3.021291907        -1.124176052
>>>CARBON      6.0     -2.031519706         2.659585996        -1.440750083
>>>HYDROGEN    1.0     -2.035172355         3.030199727        -2.469788674
>>>HYDROGEN    1.0     -2.753831215         1.840882094        -1.376428877
>>>HYDROGEN    1.0     -2.387923870         3.470002175        -0.796394787
>>> $END
>>>
>>>
>>>
>>>
>>>


[ Previous ] [ Next ] [ Index ]           Sun Aug 14 '16 3:28am
[ Reply ] [ Edit ] [ Delete ]           This message read 481 times