Abstract
Reliable prediction of molecular crystal energetics is a vital goal for computational chemistry. Here we show that accurate results can be obtained from a monomer-based many-body expansion truncated at the two-body level, with the monomer and dimer calculations suitably embedded in a model of the crystalline environment. By including the two dominant effects-electrostatics and exchange-repulsion-we are able to capture the important nonadditive terms in the energy, and approach very closely results from full periodic second-order Moller-Plesset calculations. The advantage of the current scheme is that extension to coupled-cluster and explicitly correlated F12 methods is completely straightforward. We demonstrate the approach through calculations on carbon dioxide, hydrogen fluoride, and ice XIh and XIc. In accord with previous studies, we find these two ice polymorphs to be very close in energy, with our periodic coupled-cluster single double triple-F12 calculation giving the hexagonal structure more stable by around 0.3 kJ mol(-1). (C) 2012 American Institute of Physics. [http://dx.doi.org/10.1063/1.4759079]
Original language | English |
---|---|
Article number | 164102 |
Number of pages | 9 |
Journal | Journal of Chemical Physics |
Volume | 137 |
Issue number | 16 |
DOIs | |
Publication status | Published - 28 Oct 2012 |
Keywords
- PERIODIC-SYSTEMS
- FUNCTIONAL-THEORY
- DENSITY FITTING APPROXIMATIONS
- SOLID CARBON-DIOXIDE
- HARTREE-FOCK
- LITHIUM HYDRIDE
- MOLLER-PLESSET THEORY
- HYDROGEN-FLUORIDE
- STRUCTURE PREDICTION
- CORRELATION-ENERGY