TS Optimization in CCSD(T)/VTZ

[Paul Fleurat-Lessard]

Dear Molpro-users,

	A few weeks ago, I ask you for your help concerning the
optimization of a TS in CCSD(T)/VTZ. My problem was :

-------------
 At the begining of the optimization, Molpro tries to calculate a
numerical Hessian... but it stops before the end with an error message
telling me that there is no CPU time left :

 Maximum CPU time exceeded. Stopping calculation
---------

	Professor Knowles answered :

There is a trap on the cpu time for the numerical hessian calculation,
but the default value is 10^{10} seconds, ie more than 300 years. It
can be overridden with data like
cpu,max,5000
and I wonder if you have something like this in your input.
-------
Since this time, I have tried to use the (undocummented ?) 'cpu command',
but with no success. I give you my input file :

***, H2CCH-H2 doublet  CCSD(T)/VTZ, sym Cs
mem,32,m

geometry={ANGSTROM;
C1;C2, C1, dCC;
H3, C1, dCH1, C2, AHCC1;
H4, C1, dCH2, C2, AHCC2, H3, 180.;H5, C2, dCH3, C1, AHCC3, H3, 0.;
H6, C2, dCH6, C1, AHCC6, H3, 180.;H7, H6, dCH7, C2, AHCC7, C1, 0.;
}

DCC=1.32447649 ANGSTROM
[...]
DCH6=1.442

basis,VTZ
int

hf;wf,17,1,1

rccsd(T);wf,17,1,1;
optg;root,2;
numhess
cpu,max,100000
-----------------

	Unfortunately, I still have the same error message :

 Maximum CPU time exceeded. Stopping calculation

after only 10h of CPU althought I have no time limit on this calculation.

	Any suggestions ?
	
	On the other way, I have calculated the frequencies at a lower
level of calculation (CCSD(T)/VDZ) and tried to used this Hessian as a
starting point for the CCSD(T)/VTZ calculation.  Now, the error is :

 ERROR: Inconsistent FREQ-record for Hessian calculation!

 Here is my (second) input :

--------
***, TS leading from C2H5 to C2H4+H
mem,32,m

geometry={ANGSTROM;
C;C, 1, R1;H, 1, R4, 2, A9;
H, 1, R2, 2, A7, 3, DH12;H, 1, R2, 2, A7, 3, -DH12;
H, 2, R5, 1, A10, 3, DH14;H, 2, R5, 1, A10, 3, -DH14;}

R1=1.35115927 ANGSTROM
[...]
R4=1.947

basis,VDZ
hf;wf,17,1,1

rccsd(T);wf,17,1,1;
freq,forward;start,5300.2;print,force;

basis,VTZ;hf;wf,17,1,1

rccsd(T);wf,17,1,1
optg;root,2;hstart,5300.2
-------

	Once again, Any suggestion is welcome !

	Thanks in advance,
	Paul.

-------------------------------------------------------------------------
Fleurat-Lessard Paul                           pfleura at gwdg.de
Max Planck Institut Fur Stromungsforschung
Gottingen, Germany
Phone : (Int code) 49 551 5176 727
_________________________________________________________________________