Automated Parametrization Approach for Coarse-Graining Soil Organic Matter Molecules

Abstract

Investigating the molecular structure of soil organic matter (SOM), along with its intramolecular interactions and interactions with other soil components and xenobiotics, is essential due to its ecological importance. However, the complexity and heterogeneity of SOM present significant challenges for systematic studies. While experimental methods are commonly employed, atomistic simulations provide a complementary approach to exploring molecular-level processes. The Vienna Soil Organic Matter Modeler 2 (VSOMM2) facilitates the construction of molecular models of SOM systems with various compositions at the atomistic scale, which can then be examined through molecular dynamics (MD) simulations. This study introduces a parametrization strategy that enables the conversion of VSOMM2-generated structures into a coarse-grained representation, thus allowing larger time and length scales to be explored. By employing a conformer search technique, direct construction and analysis of coarse-grained SOM models with diverse compositions were made possible, eliminating the need for atomistic MD simulations. To demonstrate this approach, coarse-grained SOM models were created based on selected samples from the International Humic Substances Society, considering different water content levels for each model. Comprehensive analyses, including density and potential energy profile calculations, revealed a partial correlation with the SOM compositions and demonstrated that electrostatic interactions govern the structural packing. Moreover, a local phase separation process, particularly the formation of SOM voids, was observed over several microseconds, underscoring the advantages of the coarse-graining technique.