[molpro-user] Error for analytical RS2 gradients with a restricted active space

[Pfeifle, Mark Claudius Gerhard]

Hi,
I am trying to use RASPT2 for direct dynamics calculations, therefore analytical gradients would be very useful. I did a gradient test calculation, resulting in the following error message (see input below):

nci (cpmcscf)=                   490    nci(rs2)=                   187
? Error
? wrong nci
? The problem occurs in rs2grad_load

I should also mention that I'm using a level shift. This error happens after the RS2 energy has been calculated and the CP-CASPT2 equations have been solved. It appears to me that another call to the MULTI program is made, where the configuration restrictions are not passed correctly. 187 is the no. of CSFs in the RASSCF reference, while 490 is the corresponding no. for the full CASSCF(8e7o) wavefunction. Are analytical gradients only available for CASSCF references, not for general MCSCF?

I tried both versions 2012.1 and 2015.1 without success. Any thoughts or suggestions on this issue?
Thanks,
Mark

The following is my input file:

***,gradient_test_raspt2
file,2,record2,new
memory,800,M
angstrom
geomtyp=xyz
geometry={
           9
           m05
            C                  0.09401600    0.77190600   -0.18413000
            C                 -1.14074500   -0.45052700    0.27666600
            H                  0.11432900    0.74310700   -1.26932600
            H                 -0.20211800    1.70842000    0.27790800
            H                 -0.65384200   -1.29464400   -0.24692500
            H                 -1.04027900   -0.49614400    1.37312500
            O                  1.27204600    0.35159800    0.39429100
            O                  1.87822000   -0.61969200   -0.26472400
            O                 -2.25269500   -0.01465300   -0.20748500
           }

basis=sto-3g

{mcscf
closed,16
occ,   23
wf,40,1,0
}

put,molden,sto3g.molden

basis=cc-pVDZ

{mcscf
rotate,14.1,18.1,0.
rotate,23.1,24.1,0.
closed,16
occ,   23
wf,40,1,0
restrict,4,6,17.1,18.1,19.1
restrict,0,2,22.1,23.1
}

put,molden,ccpvdz.molden

{rs2,shift=0.20
closed,16
occ,   23
wf,40,1,0
restrict,4,6,17.1,18.1,19.1
restrict,0,2,22.1,23.1
maxiter,200,200}

force

---

----------------------------------------------------
Dr. Mark Pfeifle
Chemical Sciences & Engineering Division - Fundamental Interactions
Argonne National Laboratory
9700 S Cass Ave
Lemont, IL, 60439
mpfeifle at anl.gov<mailto:mpfeifle at anl.gov>
+1 (630) 252-3717

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.molpro.net/pipermail/molpro-user/attachments/20170602/5e9523bd/attachment.html>