In the past, corrections for the CFSE in QMC simulations have been applied using results for different simulation cell sizes and extrapolating to the infinite cell size limit [46, 33, 43, 47]. This empirical procedure is very costly as it requires several calculations at increasingly large system sizes. As mentioned in section , the CFSE is approximately inversely proportional to the number of atoms in the simulation cell. A reasonable approximation is

where *N* is the number of electrons in the system and is
defined as the radius of a sphere whose volume equals the volume per
electron so that

To obtain a value for in Eq.(), a series of calculations is performed for different system sizes. This fitting scheme is not very accurate when results for only 2 or 3 system sizes are available, especially when the estimate for the CFSE is subject to statistical noise at each system size.

Tue Nov 19 17:11:34 GMT 1996