D. Steiger et R. Glaser, Lattice sum calculations for 1/r(p) interactions via multipole expansions and Euler summation, J COMPUT CH, 22(2), 2001, pp. 208-215
A method is developed here for doing multiple calculations of lattice sums
when the lattice structure is kept fixed, while the molecular orientations
or the molecules within the unit cells are altered. The approach involves a
two-step process. In the first step, a multipole expansion is factored in
such a way as to separate the geometry from the multipole moments. This fac
torization produces a formula for generating geometry constants that unique
ly define the lattice structure. A direct calculation of these geometry con
stants, for all but the very smallest of crystals, is computationally impra
ctical. In the second step, an Euler summation method is introduced that al
lo rvs for efficient calculation of the geometry constants. This method has
a worst case computational complexity of O((log N)(2) /N), where N is the
number of unit cells. If the lattice sum is rapidly converging, then the co
mputational complexity can be significantly less than N. Once the geometry
constants have been calculated, calculating a lattice sum for a given molec
ule becomes computationally very fast. Millions of different molecular orie
ntations or molecules can quickly be evaluated for the given lattice struct
ure. (C) 2000 John Wiley & Sons, Inc.