Coupling between thermochemical contributions of subvalence correlation and of higher-order post-CCSD(T) correlation effects -- a step toward `W5 theory'

TL;DR

This paper investigates the impact of post-CCSD(T) correlation effects on thermochemical energies of molecules, proposing a new `W5 theory' protocol that improves accuracy for second-row compounds.

Contribution

It introduces a `W5 theory' protocol that accounts for coupling of subvalence and higher-order correlation effects, improving thermochemical energy predictions.

Findings

Large (Q) contributions in molecules with neighboring second-row atoms.

Reoptimized geometries increase calculated TAE, especially for second-row compounds.

Predicted TAE0 values align well with active thermochemical tables, revising some key species.

Abstract

We consider the thermochemical impact of post-CCSD(T) contributions to the total atomization energy (TAE, the sum of all bond energies) of first- and second-row molecules, and specifically their coupling with the subvalence correlation contribution. In particular, we find large contributions from (Q) when there are several neighboring second-row atoms. Otherwise, both higher-order triples $T_3$--(T) and connected quadruples (Q) are important in systems with strong static correlation. Reoptimization of the reference geometry for core-valence correlation increases the calculated TAE across the board, most pronouncedly so for second-row compounds with neighboring second-row atoms. %just slightly increases the calculated TAE for all species, but more pronouncedly so if strong static correlation is present, as well as for second-row compounds, again especially with neighboring second-row atoms. We present a first proposal for a `W5 theory' protocol and compare computed TAEs for the W4-08 benchmark with prior reference values. For some key second-row species, the new values represent nontrivial revisions. Our predicted TAE$_0$ values (TAE at 0 K) agree well with the ATcT (active thermochemical tables) values, including for the very recent expansion of the ATcT network to boron, silicon, and sulfur compounds.