# [molpro-user] diabatisation documentation

Christian Evenhuis ceven at rsc.anu.edu.au
Tue Jun 12 13:34:59 BST 2007

Dear all,

I am trying to find out exactly how the DDR program works in molpro.
In short, my questions are:
1. What is the reference J. Chem. Phys. 102, 0000, (1999)  in section
20.5.8.
2. How does DDR calculate the unitary transformation matrix.
3. What is the first orbital correction correction to the orbitals
mentioned in section 39?  How is this calculated?

For more context see below.

Thanks,

Chris Evenhuis

Firstly, the CASSCF orbitals are rotated to maximise the overlap with
the reference orbitals using the diabatisation from section 20.5.8.
From section 20.5.8:

"Recently, an automatic procedure, as described in J. Chem. Phys.
102, 0000, (1999) has been implemented into MOLPRO. This is available
in Version 99.1 and later and is described in section 35."

In section 35 there is a reference to D. Simah, B. Hartke, and H.-J.
Werner, J. Chem. Phys. 111, 4523 (1999). Is this the mystery
reference from above?  From the description I see that the CASSCF
orbitals are diabatised by either 2x2 Jacobi rotations or block
diagonalisation.

Having a minimised the change in the CASSCF orbitals (and hence
maximising the change in CI coefficients) DDR is used to find the
unitary transformation that minimises the change in the CI
coefficients.  How is this achieved?

A.	 minimising
|< \phi_i(q')| \phi_i(q) >|^2+|< \phi_j(q')| \phi_j(q) >|^2
for each pair of active orbitals.

B. 	or from the overlap matrix S = <\phi_i(q)|\phi_j(q)> as
S(S^\dag S)^{-1/2}

C. 	or something else

The examples then mention a first order correction for the change in
the orbitals.  What is this and how is it calculated?

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.molpro.net/pipermail/molpro-user/attachments/20070612/a27dbcf1/attachment.html>