[molpro-user] two questions: freq for CASSCF and gradients for QCISD(T)

[eurisco1 at pochta.ru]

Dear Molpro Users and Developers,


Would you allow me to address to you
necessary questions:



I have calculated frequencies
	(numerically) as well as analytically by using routes on Version
	2008.1 linked 05 Sep 2008 (trial) at the level CAS(10,10)/cc-pVDZ



First input: 


geomtyp=xyz                


geometry={

 6





 my molecule _xyz

}      


basis=VdZ                 


{hf;wf,46,1,0,0}

{multi;occ,28;closed,18;wf,46,1,0,0;cpmcscf,hess}

{frequencies}                   !do 


---


Second input:

geomtyp=xyz                


geometry={

 6





my molecule _xyz

}      


basis=VdZ                 


{hf;wf,46,1,0,0}

{multi;occ,28;closed,18;wf,46,1,0,0}

{frequencies;


print,low,img;}                   !do

and I have always observed only real
frequencies (which were slightly  different). But if I used Gaussian
(both 03 and 09) to calculate (numerically) frequencies I observed
only one imag. (~100i) frequency (I performed IRC and achieved both
valleys).




Why I got two different results? How
can I treat this situation?



In the Molpro manual states that
	analytical gradients are available
	for qcisd(t). I have tested simple input 
	


! $Revision: 2008.1 


***,h2o                   !A title

memory,150,m

r=2.5,theta=120

geometry={o;

          h1,O,r;

          h2,O,r,H1,theta

}      


basis=vdz               


rhf

rqcisd(t)                  


optg





but I have observed only messages about
numerical calculations of gradients

Numerical gradient completed.
CPU-time:     0.75 sec, Elapsed:     2.50 sec


What is wrong in my input? Are
analytical gradients available for geometry optimizations if qcisd(t)
method is used for calculation of energy?

Sincerely,


PhD student Oleg B. Gadzhiev


ps: I have downloaded molpro 5 days
ago.













-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.molpro.net/pipermail/molpro-user/attachments/20091122/8ee019f7/attachment.html>