Cluster-in-Cluster Approach for Computing MP2-Level Vibrational Infrared Spectra of Large Molecular Clusters

Constructing the Hessian matrix (HM)
for large molecules
demands
huge computational resources. Here, we report a cluster-in-cluster
(CIC) procedure for efficiently evaluating HM and dipole derivatives
for large molecular clusters by employing the second-order Møller–Plesset
perturbation (MP2) theory. The highlight of the proposal is the separation
of the estimations of Hartree–Fock (HF) and post-HF components.
The parent cluster with n molecules is divided (virtually)
into n subclusters centering each monomer and accommodating
its near neighbors decided by a distance cutoff. The HF-level HM is
obtained by doing full calculation (FC), while the correlation part
is approximated by the respective subclusters. A software automating
the procedure [followed by calculating infrared (IR) frequencies and
intensities] is applied to deduce the IR spectrum for a variety of
molecular clusters, particularly water clusters of various sizes,
containing up to ∼2000 basis functions. The accuracy of the
IR spectrum constructed using CIC is remarkable, with a substantial
time advantage (with respect to its FC counterpart). The reduced computational
resources and the tractability of the computations are other major
benefits of the procedure.