abstract
Glucose is an important carbohydrate, relevant both for its biological functions and as a raw material for industrial processes. As a monomer of cellulose, the most abundant biopolymer, it is an alternative feedstock for fuels and chemicals in the biorefinery framework. Since glucose is often used and processed in aqueous solutions, it is important to understand the structural, volumetric, and dynamic properties of aqueous glucose solutions at varying concentrations. Molecular dynamics (MD) simulations are an effective means for computing the properties of liquid solutions, but the technique relies upon accurate intermolecular potential functions (i.e., "force fields"). Here we evaluate the accuracy of the recently developed GROMOS 56A(CARBO) glucose force field for its ability to model the density, viscosity, and self-diffusivity of aqueous glucose solutions as a function of concentration. We also compute different structural properties, including hydrogen bonds, radial and spatial distribution functions, and coordination numbers. The results show that the force field accurately models the density and viscosity of dilute solutions up to a glucose mole fraction of 0.1. At higher glucose concentrations, the force field overestimates the experimental density and viscosity. By analyzing the liquid structure, it is found that the glucose molecules tend to associate at higher concentrations, which leads to deviation from the experimental results. This suggests that, while the GROMOS 56A(CARBO) force field performs well for highly dilute glucose solutions (conditions under which it was developed), it is not appropriate for carrying out simulations of more concentrated glucose solutions. It is possible to obtain much more accurate densities and viscosities at high glucose concentrations by uniformly reducing the partial charges on glucose by 20%, which attenuates the self-association tendencies of glucose.
keywords
NEUTRON FIBER DIFFRACTION; HYDROGEN-BONDING SYSTEM; SYNCHROTRON X-RAY; IONIC LIQUIDS; 1-ETHYL-3-METHYLIMIDAZOLIUM CHLORIDE; CRYSTAL-STRUCTURE; SIMULATIONS; WATER; MODELS; CONFIGURATIONS
subject category
Chemistry
authors
Batista, MLS; Perez-Sanchez, G; Gomes, JRB; Coutinho, JAP; Maginn, EJ
our authors
Germán Pérez-Sánchez
Junior Researcher
José R. B. Gomes
Principal Researcher
Marta Luísa Salsas Batista
PhD StudentGroups
G4 - Renewable Materials and Circular Economy
G6 - Virtual Materials and Artificial Intelligence
acknowledgements
The authors thank Fundacao para a Ciencia e Tecnologia (FCT) for funding, for project PEst-C/CTM/LA0011/2013 and for Programa Investigador FCT. FCT is also recognized for the PhD grant SFRH/BD/74551/2010 of M.L.S.B. Thanks are due as well to the Center for Research Computing at the University of Notre Dame for providing access to computational resources. The authors acknowledge the help provided by Akash Sharma, who unfortunately passed away before this work was completed.