One can not have a truly accurate description of the chemistry of the elements beyond the first row of the transition elements without taking in account the effects of relativity. The so-called lanthanide contraction and even the colour of gold can be understood only with taking relativity into account.
It is relatively easy to estimate the relativistic energy of a molecule using perturbation theory. For example, extend Bueckert et al's (1992) work done using variational Monte Carlo. More recently, Langfelder et al (1997) derived an estimator for the delta function (which enters the Darwin term) given the exact electron distribution. By employing his exact sampling algorithm, it is now possible to improve the accuracy of relativistic calculations which employ perturbation theory. Unfortunately, there exact sampling calculations are restricted to small systems.
Of course, our long term objective is heavy-atom systems. Here, for
the proper treatment of relativity one needs highly accurate cores. Exact
sampling is not feasible, but variational sampling with perturbation theory
using core-valence partitioned wavefunctions is in principle straight-forward.
To estimate the spin-orbit interaction for small atoms using Staroverov
et al's (1998) partitioned wavefunctions would be an interesting first
step in this direction.
This page is: http://chemiris.labs.brocku.ca/~chemweb/faculty/rothstein/rel.html
Revised: November 1, 1997
© copyright 1995 Brock University