[molpro-user] computing density differences

[Francesco Tarantelli]

I would like to plot/visualize density differences between an adduct AB
and its two (say) fragments A and B (in a given basis set). If you know
how to do this please help. I have tried the following:

1. Compute AB in a fixed symmetry and orientation, compute
    density and store. Then declare all of A dummy and repeat. Then dummy
    B. Then use matrop to subtract A and B densities from the AB one.
    This works fine, but I would like to include basis superposition, so
    I would like to subtract from AB the isolated A and B densities, each
    calculated without the basis functions of the other.

2. Compute AB as above and output (gopenmol) its density. Then compute A
    and B in the same space position they have in the adduct and again
    output the plt. Now gopenmol should be able to read the separate
    densities and subtract them, but it fails to do so because the plt
    files have different contour limits. As there's apparently no
    documented way to completely define the grid molpro uses, I'm stuck.
    The same problem occurs with cube.

3. I guess one could merge natural orbitals and outer-multiply them via
    matrop to get the superimposed non-interacting density matrix, but
    how to access occupation numbers?

BTW: the manual says molpro should be able to compute densities for
closed shell ccsd (p. 163). The dm command is indeed accepted but
it does nothing.  Am I doing something wrong?

Thanks,
Francesco