Several short-time approximations of the imaginary-time propagator of the Schrodinger equation are compared working on small helium and neon clusters. A recently discussed fourth order short time approximation of the propagator [Phys. Rev. E 61, 2050 (2000)] allows us to compute several properties practically unaffected by the time step bias. The comparison among simulations of the same length shows that this algorithm permits the use of larger time steps, leading to more accurate statistics than the ones obtained by employing commonly used schemes. Results of the mixed estimator of the potential energy, of the first two momenta of the interparticle distribution, and of the particle-center-of-mass distribution seem to indicate that the new propagator is able to perform unbiased sampling even when very large time steps are used. Also, the relative population of the four Ne-7 isomers sampled using the fourth order propagator does not show any time step bias in the 200-1000 hartree(-1) time step range. This fact indicates that using the fourth order propagator with large time steps is a viable approach to tackle ergodicity problems in semirigid clusters. (C) 2003 American Institute of Physics.
Comparison of different propagators in diffusion Monte Carlo simulations of noble gas clusters
MELLA, MASSIMO;MOROSI, GABRIELE;BRESSANINI, DARIO
2003-01-01
Abstract
Several short-time approximations of the imaginary-time propagator of the Schrodinger equation are compared working on small helium and neon clusters. A recently discussed fourth order short time approximation of the propagator [Phys. Rev. E 61, 2050 (2000)] allows us to compute several properties practically unaffected by the time step bias. The comparison among simulations of the same length shows that this algorithm permits the use of larger time steps, leading to more accurate statistics than the ones obtained by employing commonly used schemes. Results of the mixed estimator of the potential energy, of the first two momenta of the interparticle distribution, and of the particle-center-of-mass distribution seem to indicate that the new propagator is able to perform unbiased sampling even when very large time steps are used. Also, the relative population of the four Ne-7 isomers sampled using the fourth order propagator does not show any time step bias in the 200-1000 hartree(-1) time step range. This fact indicates that using the fourth order propagator with large time steps is a viable approach to tackle ergodicity problems in semirigid clusters. (C) 2003 American Institute of Physics.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.