[molpro-user] Non-convergence error in OPTG program
Cayo Emilio
cayoquimica at gmail.com
Wed Aug 5 15:27:19 BST 2015
---------- Forwarded message ---------
From: Cayo Emilio <cayoquimica at gmail.com>
Date: ter, 16 de jun de 2015 às 15:02
Subject: [molpro-user] Non-convergence error in OPTG program
To: <molpro-user at molpro.net>
Dear MOLPRO community.
I am trying to compute the energy of a 3-atom system using the program
MULTI in different multiplicities. The singlet works just fine, but triplet
is a headache
My input is:
memory,200,m
thresh,pspace=20
basis=AVQZ
SYMMETRY,X,Y
R1=8.20 bohr
R2=[0.5,1.0,1.5,2.0,2.25,2.5,2.75,3.0,3.25,3.5,3.75,4.0,4.5,4.75,5.0,5.5,6.0]
theta1=90 deg
theta2=0 deg
theta3=-90 deg
R3=1.0 bohr
geometry={Si1;
q1,Si1,R1/2;
q2,q1,R3,Si1,theta1;
Si2,q1,R1/2,q2,theta3,si1,theta2;
Si3,q1,RR,Si1,theta1,q2,theta2}
DO i=1,#R2
RR=R2(i)
R(i)=RR
{multi;maxit,40;wf,42,1,2;state,2}
{optg,maxiter=60,VARIABLE=energy(1)
active,R1}
RR1(i)=R1
theta(i)=theta1
ecas1(i)=energy(1)
ecas2(i)=energy(2)
table,R,ecas1,ecas2,RR1
ENDDO
I am trying to approximate one of the atoms point-by-point to the dimer,
and usually after two or three points of the loop the MOLPRO crash with the
error: "No convergence. The problem occurs in Multi.". The log says
" WARNING: exit with wrong ci vector!!" for many interactions, and i don't
understand what this means. Sometimes I try to reedit the input to start
from the point (of the R2 vector) it had just crashed and it goes fine, but
after several tries and small changes on my parameters values.
The problem occur only in triplet calculations, with all symmetries of C2v
group. The calculations for all four symmetries of the singlet state, for
the same parameters of the input above, end correctly.
If I do not give a higher thresh for p-space the program crashes with the
"Non-zero p-space gradient" error, precisely in the same point.
I appreciate any help.
Thank you,
Cayo E. M. Gonçalves.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.molpro.net/pipermail/molpro-user/attachments/20150805/385d0895/attachment-0001.html>
More information about the Molpro-user
mailing list