PRODUCT REPRESENTATION OF POTENTIAL-ENERGY SURFACES - II

An efficient method was recently introduced [J. Chem. Phys. 102, 5605 (1995); 104, 7974 (1996)] to represent multidimensional potential ener gy surfaces as a linear combination of products of one-dimensional fun ctions, so-called natural potentials. Weight functions were shown to b e easily implemented in the product representation scheme as long as t hey are separable, i.e., defined as a product of one-dimensional weigh t functions. Here the constraint imposed by the special product form o f the separable weights is removed. Nonseparable weights are emulated by dividing the potential energy surface in arbitrary regions of minor and major physical relevance and by utilizing a so-called relevant re gion iteration procedure. Maintaining the advantageous computational s caling properties of the product representation scheme, this relevant region iteration procedure allows the stepwise improvement of the surf ace representation in the regions of major relevance. The quality of t he product representation in the regions of minor relevance remains ne vertheless acceptable. As a consequence, the number of potential expan sion coefficients can be reduced substantially. The product representa tion of potential energy surfaces is especially well suited to be empl oyed within the framework of the multiconfiguration time-dependent Har tree (MCTDH) approximation. To check the performance of the proposed m ethod the Liu-Siegbahn-Truhlar-Horowitz (LSTH) surface is represented in Jacobian coordinates, and initial-state selected reaction probabili ties for the H+H-2(nu=j=0)-->H-2+H exchange reaction are computed. (C) 1998 American Institute of Physics. [S0021-9606(98)01034-4].