Screening properties of Gaussian electrolyte models, with application to dissipative particle dynamics

Abstract

We investigate the screening properties of Gaussian charge models of electrolyte solutions by analysing the asymptotic behaviour of the pair distribution functions. We use a combination of Monte-Carlo simulations with the hyper-netted chain integral equation closure, and the random phase approximation, to establish the conditions under which a screening length is well defined and the extent to which it matches the expected Debye length. For practical applications, for example in dissipative particle dynamics, we are able to summarise our results in succinct rules-of-thumb which can be used for mesoscale modeling of electrolyte solutions. We thereby establish a solid foundation for future work, such as the systematic incorporation of specific ion effects.

Screening properties of four mesoscale smoothed charge models, with application to dissipative particle dynamics

“…Later González-Melchor et al examined an Ewald-based method with exponential charge smoothing [5]. Most recently we have studied a related Ewald method with Gaussian charge smoothing [6]. This last choice connects with recent work on the so-called ultrasoft restricted primitive model (URPM) [7][8][9].…”

“…The DPD particle size is conventionally used to nondimensionalise the densities, with ρr 3 c = 3 being widely adopted. Making this assumption, r c and l B are fixed by physical arguments and the mapping to the underlying physical system, for example r c = 0.645 nm and l B = 0.7 nm for a room temperature 1:1 aqueous electrolyte [6]. The ion density is then also set by the mapping to the underlying system, for example ρ z r 3 c = 0.32 for a 1 M solution [6].…”

“…More details and justification for particular parameter choices can be found in previous work [6]. We represent the electrolyte as a multicomponent soft sphere fluid of DPD particles, containing positive and negative ions of valencies z ± at densities ρ ± , and a neutral (z 0 = 0) solvent species at a density ρ 0 .…”

“…Making this assumption, r c and l B are fixed by physical arguments and the mapping to the underlying physical system, for example r c = 0.645 nm and l B = 0.7 nm for a room temperature 1:1 aqueous electrolyte [6]. The ion density is then also set by the mapping to the underlying system, for example ρ z r 3 c = 0.32 for a 1 M solution [6]. These considerations thus far fix everything except the size of the charge clouds, which is the central issue.…”

“…6 to identify regions where the models exhibit a real-valued screening length, and the extent to which this agrees between models and with the Debye length in the physical system. The problem will be addressed using liquid state integral equation theory [11].…”