Hi all,

I’m interested in computing dipole moment gradients w.r.t. to nuclear coordinates. I’ve come across the `MintsHelper::dipole_grad(SharedMatrix D)`

method. I assume it takes a density matrix as argument, thus it returns the dipole moment gradient in the MO basis. Is this assumption correct? Also, does this method return the total dipole moment gradient or just the term that depends on the nuclear coordinates explicitly, i.e., I would still need to compute the response part?

Thanks!

I don’t understand what a “dipole moment gradient **in the MO basis**” means. If you are talking about the derivative of the molecular dipole *with respect to nuclear displacements*, the basis enters in the intermediates used to compute it, but you don’t need a particular choice of orbital basis to define the quantity.

Yes, you still need to compute the response part by CPHF. A dipole moment gradient is a second derivative of the energy, so the density matrix doesn’t contain enough information to compute it, even at the Hartree-Fock level.