[molpro-user] CASPT2 convergence issue

[G Mahesh]

Dear Users,

I need some suggestion regarding solving the convergence problem in CASPT2
calculation for triplet (T1) state. I have optimised T1/T2 conical
intersection successfully using SA2-CASSCF method. Now I'm trying to
account for the dynamical energy with CASPT2 method (rs2c) using 4 state
(S0, S1, T1and T2) averaged CASSCF wave function as reference. Though
CASSCF has converged within 3 steps, reference CI in CASPT2 has not
converged even after 200 iterations.

--INPUT--
 ***,S1_min===>S1/S0
 memory,1000,m
 gprint,orbitals=5
 symmetry,nosym
 geometry={
---
}
{multi
 occ,41
 closed,28
 wf,72,1,0;state,2
 wf,72,1,2;state,2
 maxiter,40;
 }

 {rs2c,shift=0.3,maxiti=200
 state,1,1
 wf,72,1,2
 maxit,100}

 {rs2c,shift=0.3,maxiti=200
 state,1,2
 wf,72,1,2
 maxit,100}

 put,molden,caspt2_631g6.molden;
----

--Partial Output--
**CASCF**

ITER. MIC  NCI  NEG     ENERGY(VAR)     ENERGY(PROJ)   ENERGY CHANGE
GRAD(0)  GRAD(ORB)   GRAD(CI)     STEP       TIME

   1   60   67    0    -473.25827922    -473.26238145   -0.00410223
0.10595877 0.00000457 0.00182944  0.49D-01    384.61
   2   33   55    0    -473.26238403    -473.26238466   -0.00000063
0.00139022 0.00000156 0.00006016  0.15D-02    704.86
   3   35   29    0    -473.26238466    -473.26238466   -0.00000000
0.00000333 0.00000023 0.00000286  0.82D-05    844.69

 ** WVFN ****  CONVERGENCE REACHED, FINAL GRADIENT:  0.44D-06
.
.
.
.
Number of p-space configurations:  81

 NO CONVERGENCE IN REFERENCE CI AFTER200 ITERATIONS. DEMAX=  0.12D-07
VRMAX=  0.12D-09
 PLEASE INCREASE MAXITI AND RETRY!
---
I would be grateful for any suggestions

Thank you
Mahesh
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.molpro.net/pipermail/molpro-user/attachments/20140923/a622cc93/attachment.html>