Firefly and PC GAMESS-related discussion club



Learn how to ask questions correctly


Re^2: Perfluorinated carboxylic acid geo opt fails with no indication why.

Gordon
gordon.driver@abo.fi


Hi Alex,

now that I have time to look into this, it seems like the problem is solved, as you directed.  Thanks very much.  For anyone needing to use differing basis sets for different atoms in the same structure, see the following link in this forum: http://classic.chem.msu.su/cgi-bin/ceilidh.exe/gran/gamess/forum/?C34df668afbHW-7080-1259+00.htm.

Here is the form of my input file, that alleviated the warning messages:

$p2p p2p=.t. dlb=.t. xdlb=.t. mxbuf=2048 $end
$SYSTEM MWORDS=125 TIMLIM=8000 $END
$CONTRL RUNTYP=Optimize $END
$STATPT  HSSEND=.T. OptTol=1e-5 NStep=500 $END
$CONTRL SCFTYP=RHF COORD=UNIQUE $END
$CONTRL MPLEVL=2 d5=.TRUE. $END
$CONTRL ICHARG=0  MULT=1 $END
$BASIS GBASIS=N311 NGAUSS=6 $END
$BASIS NDFUNC=1 NPFUNC=1 $END
$BASIS DIFFSP=.TRUE. elneg(1)=8,9 $END
$FORCE PURIFY=.T. NVIB=2 $END                                                
$ZMAT DLC=.T. AUTO=.T. $END  
$SCF DIRSCF=.TRUE. NCONV=7 $END
$DATA

I had previously seen the warning but was trying to deal with it with adjustments to NCONV, ICUT etc (across several different runs).

The use of "mixed" basis sets leads to one further question for someone to help me with...if I apply DIFFSP=.T. only to O and F (and hence no diffuse functions on the C and H atoms present), how would the nomenclature apply?  For CF3(CF2)7CH2CO2H, would I write "O,F/MP2 6-311+G(d,p)-C,H/MP2 6-311G(d,p)?

Cheers,
Gordon

On Tue Sep 7 '10 7:40pm, Alex Granovsky wrote
---------------------------------------------
>Hi Gordon,

>Actually, you have the error message at the end of output:

> SCF DID NOT CONVERGE...NO MPLEVL=2 CALCULATION

>There is a partial linear dependence in your basis set that negatively
>affects convergence. There is also the corresponding warning in the output:

> THE OVERLAP MATRIX HAS    4 EIGENVALUES BELOW 1.0E-05.
> THE SMALLEST OF THESE IS 2.63936E-06.
> THIS INDICATES A PARTIAL LINEAR DEPENDENCE IN YOUR ATOMIC BASIS.

> TO OBTAIN SCF CONVERGENCE MAY REQUIRE
> MORE ACCURATE INTEGRAL EVALUATION (INTTYP=HONDO, ICUT=11, ITOL=30 IN $CONTRL),
> MORE ACCURATE DIRECT SCF FOCK MATRIX FORMATION (FDIFF=.FALSE. IN $SCF), OR
> CHANGING CONVERGERS (DIIS=.T. SOSCF=.F. IN $SCF).

>Note, 6-311 is a spherical basis set, so some part of this dependence
>may be because of missed d5=.t. keyword. Next, you do not actually
>need diffuse functions on carbons, so you may add proper elneg
>directive to improve convergence. Finally, to get reasonable
>precision with MP2, you actually need to tighten nconv for all
>calculations with partial linear dependence; see the documentation
>on the MP2 code.

>regards,
>Alex Granovsky
>
>
>On Tue Sep 7 '10 2:56pm, Gordon wrote
>-------------------------------------
>>Dear forum followers,

>>I have a compound CF3(CF2)7CH2COOH which I am having problems with in terms of geometry optimisation.  I am using MP2 6-311+G(d,p) level of theory, and am correcting with vibrational frequencies.  For some reason, this run always stops, without any other indication than "EXECUTION OF FIREFLY TERMINATED ABNORMALLY...".

>>I initially optimse the "hand drawn" structure using a forcefield (MMFF94, 100 iterations, steepest decent etc) in Avogadro, export as an .xyz file, and import to GabEdit to construct the input file.  I have tried changing the optTol = 1 E-5 to optTol = 5 E-4 without any improvement.  This is a typical method for me, and in general I do not encounter problems, and definitely never like this.

>>I've attached the .inp and .out files here, in case anyone might have some insight as to why this problem is occurring.

>>Cheers,
>>Gordon


[ Previous ] [ Next ] [ Index ]           Sat Sep 18 '10 9:35am
[ Reply ] [ Edit ] [ Delete ]           This message read 2456 times