AMBER Archive (2008)

Subject: Re: AMBER: Problem in calculation of resp charges using RED III

From: FyD (fyd_at_q4md-forcefieldtools.org)
Date: Fri Aug 29 2008 - 10:19:23 CDT

Quoting moitrayee_at_mbu.iisc.ernet.in:

> Thanks for the reply. I have already used #P b3lyp/6-31G* Opt Test,
> the result
> being the error termination on grounds of unfulfilled convergence.

Is it a problem of SCF convergence or a problem of optimization convergence ?

If it is a problem of optimization convergence, re-run the job from
the best intermediate structure you got in your previous job (based on
the optimization criteria).

If it is a problem of SCF convergence, you might considering starting
from another initial structure (build a better initial structure) or
using another QM package to generate a better initial structure (ADF,
Jaguar or turbomole).

With Gaussian always save the .chk output file using
%chk=yourchkfile.chk at the beginning of your input file. Then,
re-read this .chk file using Guess=Read in a second job. This helps a
lot when you have problems of SCF convergence...

regards, Francois

>> Quoting moitrayee_at_mbu.iisc.ernet.in:
>>
>>> 2. Using REDIII-Tools- the input .com file was produced as follows:
>>>
>>> %Mem=128MB
>>> %NProc=1
>>>
>>> #P hf/6-31G* Opt=(VTight,CalcFC) Freq SCF(Conver=8) Test
>>
>> => See the bug fixes on the R.E.D. web site
>>
>>> Optimization output to be used by R.E.D.
>>>
>>> 1 2
>>> F 22.031 131.687 125.051
>>
>> => See the bug fixes on the R.E.D. web site
>>
>>> F 19.837 133.121 124.065
>>> F 22.106 132.579 122.776
>>> S 19.888 131.269 124.910
>>> S 21.822 130.489 123.262
>>> S 21.887 133.677 124.533
>>> S 20.081 132.614 122.029
>>>
>>> But in both the cases the following problem encountered was that the
>>> convergence criterion is not met.
>>
>> -1 See the bug fixes on the R.E.D. web site.
>> F should be Fe ? right ?
>>
>> -2 For geometry optimization of iron complexes, replace
>>
>> #P hf/6-31G* Opt=(VTight,CalcFC) Freq SCF(Conver=8) Test
>> by
>> #P b3lyp/6-31G* Opt Test
>> too see how goes the geometry optimization step...
>>
>> Once converged, you might try to re-run the job using
>> b3lyp/6-31G* Opt=(Tight) Freq SCF(Conver=8) Test

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" (in the *body* of the email)
      to majordomo_at_scripps.edu