Schmitt-Monreal Daniel, Jacob Christoph R
Institute of Physical and Theoretical Chemistry, Technische Universität Braunschweig, Gaußstr. 17, 38106 Braunschweig, Germany.
J Chem Theory Comput. 2021 Jul 13;17(7):4144-4156. doi: 10.1021/acs.jctc.1c00340. Epub 2021 Jul 1.
Fragmentation methods based on the many-body expansion offer an attractive approach for the quantum-chemical treatment of large molecular systems, such as molecular clusters and crystals. Conventionally, the many-body expansion is performed for the total energy, but such an energy-based many-body expansion often suffers from a slow convergence with respect to the expansion order. For systems that show strong polarization effects such as water clusters, this can render the energy-based many-body expansion infeasible. Here, we establish a density-based many-body expansion as a promising alternative approach. By performing the many-body expansion for the electron density instead of the total energy and inserting the resulting total electron density into the total energy functional of density functional theory, one can derive a density-based energy correction, which in principle accounts for all higher-order polarization effects. Here, we systematically assess the accuracy of such a density-based many-body expansion for test sets of water clusters. We show that already a density-based two-body expansion is able to reproduce interaction energies per fragment within chemical accuracy and is able to accurately predict the energetic ordering as well as the relative interaction energies of different isomers of water clusters.
基于多体展开的碎片化方法为大分子体系(如分子簇和晶体)的量子化学处理提供了一种有吸引力的途径。传统上,多体展开是针对总能量进行的,但这种基于能量的多体展开在展开阶数方面往往收敛缓慢。对于表现出强极化效应的体系,如水簇,这可能使基于能量的多体展开变得不可行。在此,我们建立了一种基于密度的多体展开作为一种有前景的替代方法。通过对电子密度而非总能量进行多体展开,并将得到的总电子密度代入密度泛函理论的总能量泛函中,可以导出一种基于密度的能量修正,原则上它考虑了所有高阶极化效应。在此,我们系统地评估了这种基于密度的多体展开对于水簇测试集的准确性。我们表明,仅基于密度的二体展开就能够在化学精度内重现每个片段的相互作用能,并且能够准确预测水簇不同异构体的能量排序以及相对相互作用能。