[molpro-user] Problems computing properties with CASSCF

[Lorenzo Lodi]

I would like to compute some expectation-value properties for NiH using 
MCSCF wavefunctions, as described in section 6.13 of the manual. I am 
using molpro 2009.1

First of all, there seems to be differences in behaviour between these 
two cases:
1) a GEXPEC command is put at the beginning of the input
2) a EXPEC command is put within the MCSCF program block.

For example, consider the following input:

***, Calculation for NiH --- Input for MOLPRO 2009
GEXPEC, DM, DARW, MASSV, REL, LOP, DELTA

basis=cc-pVDZ
r=3.0
geometry={Ni; H, Ni, r}

{MCSCF; OCC,  9,3,3,1; CLOSED, 5,2,2,0; FROZEN,0,0,0,0; MAXITER,20; 
ACCURACY, ENERGY=1.e-12;
    WF, NELEC=29, SYM=1, SPIN=1; STATE,1;
    WF, NELEC=29, SYM=2, SPIN=1; STATE,2;
    WF, NELEC=29, SYM=3, SPIN=1; STATE,2;
    WF, NELEC=29, SYM=4, SPIN=1; STATE,2;
}

[The reference space above is not appropriate, but never mind that for now.]

Molpro with the input above stops after computing the integrals with the 
message
  ? Error
  ? Unknown operator:DELTA
  ? The problem occurs in Prpdrv

On the other hand if the GEXPEC line is eliminated and in its stead a line
EXPEC, DM, DARW, MASSV, REL, LOP, DELTA
is inserted in the MCSCF block no errors are reported and the 
expectation values of DELTA are printed.
However, with respect to the GEXPEC case, the expectation values for DMX 
and DMY are not reported in this case...

In any case, what matters to me is that I didn't find a way to compute 
are the various LOP2 operators such as LXLX, etc. If I add LOP2 either 
as GEXPEC or EXPEC I get an error message:
Invalid operator on EXPEC directive: LOP2 (2-electron operator)

The operator LOP is accepted when put in GEXPEC (expectation values are 
printed out) but ignored completely when put in EXPEC.

What am I doing wrong? How can I compute those LXLX, LYLY and LZLZ 
operators?
Any suggestion would be appreciated :-)

Lorenzo