21.5 Level shifts

Level shifts are often useful to avoid intruder state problems in excited state calculations. MOLPRO allows the use of shifts as described by Roos and Andersson, [Chem. Phys. Lett. 245, 215 (1995)]. The shift can be specified on the RS2 or RS2C card


Typical choices for the shift is are $0.1 - 0.3$. Only two figures after the decimal point are considered. The shift affects the results, the printed energies as well as the ENERGY variable include the energy correction for the shift as proposed by Roos and Andersson. At convergence, also the uncorrected energies are printed for comparison.

Alternatively (or in addition), the IPEA shift of G. Ghigo, B. O. Roos, and P.A. Malmqvist, Chem. Phys. Lett. 396, 142 (2004) can be used. The implementation is not exactly identical to the one in MOLCAS, since in our program the singly external configurations are not (RS2) or only partially (RS2C) contracted. In Molpro, the shift is implemented as follows:

$\frac{1}{2} D_{pp} \epsilon$ is added to the occupied part of the Fock matrix; in addition, $2 \epsilon$ is added as a general shift (not corrected). $\epsilon$ is the value specified with the IPEA option (default 0). A value of 0.20-0.25 is recommended. This removes intruder state problems to a large extent and usually improves the results. Note that the method is not exactly orbital invariant, and pseudo-canonical orbitals should be used (see CANONICAL option in MULTI).

It is possible to use SHIFT and IPEA simultaneously, but it does not make sense to use one of the G-options together with IPEA.

molpro@molpro.net 2019-06-17