Pragmatic Many-Body Approach for Economic MP2 Energy Estimation of Molecular Clusters

We propose a procedure, within the many-body analysis (MBA) framework, for an economic yet accurate estimation of the correlated method-based energies of large molecular clusters employing Dunning's augmented basis sets. The basis of the procedure is to segregate the Hartree-Fock ( E_HF) and correlation energy ( E_C) estimations. E_HF is found to differ by tens of millihartrees (mH) from its full-calculation (FC) counterpart on truncating the MBA expansion at the two-body (MBA-2) level. On the contrary, E_C is estimated with smaller error on modest hardware with limited computation time at the (MBA-2) level. In view of this, we adopt a pragmatic method wherein the E_HF (accurate to five decimal places) is taken from the FC, whereas E_C is estimated at the MBA-2 level. This method is applied to a variety of medium to large molecular clusters at the MP2 level. Preliminary results at the CCSD(T) level for (H₂O)₁₆ and (H₂O)₁₇ are also reported with tremendous savings in wall-clock time and resources. The typical errors in MP2 and CCSD(T) energies per monomer are up to 0.1 and 0.2 mH, respectively. Thus the present method, balancing accuracy and computational economy, opens a way for estimating energies of large molecular clusters using correlated theories with large basis sets.