Finite-size effects and optimal system sizes in simulations of surfactant micelle self-assembly

TL;DR

This study systematically investigates finite-size effects in surfactant micelle simulations, identifying optimal system sizes and validating multiscale approaches to improve accuracy in modeling micelle properties.

Contribution

The paper provides a detailed analysis of finite-size effects in micelle simulations and demonstrates the effectiveness of multiscale modeling to mitigate these effects.

Findings

Finite-size effects cause oscillations in micelle aggregation numbers.

Optimal system size supports formation of three micelles to minimize finite-size effects.

Multiscale modeling with MARTINI and CHARMM36 improves simulation accuracy.

Abstract

The spontaneous formation of micelles in aqueous solutions is governed by the amphipathic nature of surfactants and is practically interesting due to the regular use of micelles as membrane mimics, for the characterization of protein structure, and for drug design and delivery. We performed a systematic characterization of the finite-size effect observed in single-component dodecylphosphocholine (DPC) micelles with the coarse-grained MARTINI model. Of multiple coarse-grained solvent models investigated using large system sizes, the non-polarizable solvent model was found to most-accurately reproduce SANS spectra of 100 mM DPC in aqueous solution. We systematically investigated the finite-size effect at constant 100 mM concentration in 23 systems of sizes 40 to 150 DPC, confirming the finite-size effect to manifest as an oscillation in the mean micelle aggregation number about the thermodynamic aggregation number as the system size increases, mostly diminishing once the system supports formation of three micelles. The accuracy of employing a multiscale simulation approach to avoid finite-size effects in the micelle size distribution and SANS spectra using MARTINI and CHARMM36 was explored using multiple long timescale 500-DPC coarse-grained simulations which were back-mapped to CHARMM36 all-atom systems. It was found that the MARTINI model generally occupies more volume than the all-atom model, leading to the formation of micelles that are of a reasonable radius of gyration, but are smaller in aggregation number. The systematic characterization of the finite-size effect and exploration of multiscale modeling presented in this work provides guidance for the accurate modeling of micelles in simulations.