[molpro-user] problems with HSTART and CCSD(T)

[H. -J. Werner]

Dear Juan,
this problem is known and will be fixed in the new Molpro version.
Currently you can only use a hessian if it was computed
numerically. But your job also works without the hstart directive.
Best regards
Joachim Werner
 

On Mo, 04 Apr 2005, Senosiain, Juan Pablo wrote:

>Dear colleagues,
> 
>I'm trying to optimise a series of transition states using CCSD(T) and
>reading the Hessian from a previous HF frequency job.   An example for
>this is illustrated in section 32.2.16 of the manual and works fine for
>MP2  but fails for CCSD or CCSD(T) optimisations.  In the case of
>CCSD(T), the optimisation job terminates with the message:
> 
> 
> Numerically approximating hessian for CCSD(T) by central energy
>differences
> 
> Total number of displacements:      7
> 
> Incrementing variable L1     by +/- 0.005 ANG
> Incrementing variable L2     by +/- 0.005 ANG
> 
> ERROR IN GETHIS: NO DATA FOUND FOR CCSD(T)
> CURRENT STACK:      MAIN
> 
> 
> 
>Any insights would be greatly appreciated.
> 
>Regards,
> 
> Juan Senosiain
> 
>
>PS. a sample input file follows:
> 
> ***,HCN  <--> NCH TS opt
> memory,1,m
> 
> l1=1.1826 ang
> l2=1.4074 ang
> a1=55.05 degree
> 
> basis=3-21G
> 
> geometry={ nosymm;
>  C,
>  N,1,l1
>  H,2,l2,1,a1 }
> 
> int
> hf
> frequencies,analytical
>  save,5300.2
>  print,low
>  print,imag
> 
> ccsd(t)
>  optg
>  root,2
>  method,rf
>  step,1,0,3,10
>  hstart,5300.2
> 
> frequencies
>  print,low
>  print,imag
> ---
>
> 
>~~-~~-~-~-~~-~-~-~~-~--~-~~-~-~-~~-~-~
>Juan Pablo Senosiain
>Sandia National Laboratories
>P.O. Box 969, MS 9055
>Livermore, CA 94551-0969
>phone:    925-2942305
>fax:           925-2942276
> 

-- 
Prof. Hans-Joachim Werner
Institute for Theoretical Chemistry
University of Stuttgart
Pfaffenwaldring 55
D-70569 Stuttgart, Germany
Tel.: (0049) 711 / 685 4400
Fax.: (0049) 711 / 685 4442
e-mail: werner at theochem.uni-stuttgart.de