Generalised dissipative particle dynamics with energy conservation: density- and temperature-dependent potentials

Single-Site Local-Density Potentials for the Mesoscopic Representation of Water Based on the SAFT-VR Mie Equation of State

In this article, we present three mesoscopic models for water. All three models make use of local density-dependent interaction potentials, as employed within the Pagonabarraga-Frenkel framework [Pagonabarraga, I.; Frenkel, D. J. Chem. Phys. 2001, 115, 5015-5026]. The forms of these three interaction potentials are based on the free energy function of the SAFT-VR Mie equation of state (EoS) [Lafitte, T. J. Chem. Phys. 2013, 139, 154504]. Two of these models represent the water-water interaction as a spherically symmetric Mie interaction with temperature-dependent parameters, while the third model works with a temperature-independent Mie potential and explicitly models the effect of hydrogen bonding using an association term. All three models provide good predictions of the vapor-liquid equilibrium of water over a wide temperature range. They also give accurate predictions of the isothermal compressibility for both sub- and supercritical conditions. To model the interfacial tension of the vapor-liquid interface with our mesoscale simulations, we added a square-gradient term to our potential energy function. We show that the addition of this term has a minimal effect on the bulk properties of water. However, by parametrizing the coefficient of this term as a function of temperature, all three models again provide excellent predictions of water's interfacial tension over a wide temperature range. Of the three models, our preference is for the model that includes an association term, as this model can operate successfully over a wider range of conditions.

Non-Newtonian dynamics modelled with non-linear transport coefficients at the mesoscale by using dissipative particle dynamics

We derive the algorithms for the dynamics of the standard dissipative particle dynamics model (DPD) for a velocity-dependent friction coefficient. By introducing simple estimators of the local rate of strain we propose an interparticle friction coefficient that decreases for high deformation rates, eventually leading to the macroscopic shear-thinning behaviour. We have derived the appropriate fluctuation-dissipation theorems that include the correction of the spurious behaviour due to the coupling of the non-linear friction and the fluctuations. The consistency of the model has been numerically investigated, including the Maxwell-Boltzmann distribution for the particle velocities as well as the comparison with the standard linear model for various stresses. The shear-thinning behaviour is clearly reported. Finally, along with the important methodological aspects related to the derivation of the algorithms for non-linear interparticle friction, we introduce a novel two-step algorithm that permits us the integration of the dynamic equations of the DPD model without the explicit derivation of the corrective terms due to the spurious behaviour.

Modeling temperature-dependent transport properties in dissipative particle dynamics: A top-down coarse-graining toward realistic dynamics at the mesoscale

Dissipative particle dynamics (DPD) is a widespread computational tool to simulate the behavior of soft matter and liquids in and out of equilibrium. Although there are many applications in which the effect of temperature is relevant, most of the DPD studies have been carried out at a fixed system temperature. Therefore, this work investigates how to incorporate the effect of system temperature variation within the DPD model to capture realistic temperature-dependent system properties. In particular, this work focuses on the relationship between temperature and transport properties, and therefore, an extended DPD model for transport properties prediction is employed. Transport properties, unlike the equilibrium properties, are often overlooked despite their significant influence on the flow dynamics of non-isothermal mesoscopic systems. Moreover, before simulating the response of the system induced by a temperature change, it is important to first estimate transport properties at a certain temperature. Thus here, the same fluid is simulated across different temperature conditions using isothermal DPD with the aim to identify a temperature-dependent parametrization methodology, capable of ensuring the correctness of both equilibrium and dynamical properties. Liquid water is used as a model system for these analyses. This work proposes a temperature-dependent form of the extended DPD model where both conservative and non-conservative interaction parameters incorporate the variation of the temperature. The predictions provided by our simulations are in excellent agreement with experimental data.

Rigorous Progress in Coarse-Graining

Low-resolution coarse-grained (CG) models provide remarkable computational and conceptual advantages for simulating soft materials. In principle, bottom-up CG models can reproduce all structural and thermodynamic properties of atomically detailed models that can be observed at the resolution of the CG model. This review discusses recent progress in developing theory and computational methods for achieving this promise. We first briefly review variational approaches for parameterizing interaction potentials and their relationship to machine learning methods. We then discuss recent approaches for simultaneously improving both the transferability and thermodynamic properties of bottom-up models by rigorously addressing the density and temperature dependence of these potentials. We also briefly discuss exciting progress in modeling high-resolution observables with low-resolution CG models. More generally, we highlight the essential role of the bottom-up framework not only for fundamentally understanding the limitations of prior CG models but also for developing robust computational methods that resolve these limitations in practice.

Green-Kubo expressions for transport coefficients from dissipative particle dynamics simulations revisited

This article addresses the debate about the correct application of Green-Kubo expressions for transport coefficients from dissipative particle dynamics simulations. We demonstrate that the Green-Kubo expressions are valid provided that (i) the dynamic model conserves the physical property, whose transport is studied, and (ii) the fluctuations satisfy detailed balance. As a result, the traditional expressions used in molecular dynamics can also be applied to dissipative particle dynamics simulations. However, taking the calculation of the shear viscosity as a paradigmatic example, a random contribution, whose strength scales as 1/δt1/2, with δt the time-step, can cause difficulties if the stress tensor is not separated into the different contributions. We compare our expression to that of Ernst and Brito (M. H. Ernst and R. Brito, Europhys. Lett., 2006, 73, 183-189), which arises from a diametrically different perspective. We demonstrate that the two expressions are completely equivalent and find exactly the same result both analytically and numerically. We show that the differences are not due to the lack of time-reversibility but instead from a pre-averaging of the random contributions. Despite the overall validity of Green-Kubo expressions, we find that the Einstein-Helfand relations (D. C. Malaspina et al. Phys. Chem. Chem. Phys., 2023, 25, 12025-12040) do not suffer from the need to decompose the stress tensor and can readily be used with a high degree of accuracy. Consequently, Einstein-Helfand relations should be seen as the preferred method to calculate transport coefficients from dissipative particle dynamics simulations.

Local density dependent potentials for an underlying van der Waals equation of state: A simulation and density functional theory analysis

There is an ever increasing use of local density dependent potentials in the mesoscale modeling of complex fluids. Questions remain, though, about the dependence of the thermodynamic and structural properties of such systems on the cutoff distance used to calculate these local densities. These questions are particularly acute when it comes to the stability and structure of the vapor/liquid interface. In this article, we consider local density dependent potentials derived from an underlying van der Waals equation of state. We use simulation and density functional theory to examine how the bulk thermodynamic and interfacial properties vary with the cutoff distance, rc, used to calculate the local densities. We show quantitatively how the simulation results for bulk thermodynamic properties and vapor-liquid equilibrium approach the van der Waals limit as rc increases and demonstrate a scaling law for the radial distribution function in the large rc limit. We show that the vapor-liquid interface is stable with a well-defined surface tension and that the interfacial density profile is oscillatory, except for temperatures close to critical. Finally, we show that in the large rc limit, the interfacial tension is proportional to rc and, therefore, unlike the bulk thermodynamic properties, does not approach a constant value as rc increases. We believe that these results give new insights into the properties of local density dependent potentials, in particular their unusual interfacial behavior, which is relevant for modeling complex fluids in soft matter.

A coarse-grain reactive model of RDX: Molecular resolution at the μm scale

Predictive models for the thermal, chemical, and mechanical response of high explosives at extreme conditions are important for investigating their performance and safety. We introduce a particle-based, reactive model of 1,3,5-trinitro-1,3,5-triazinane (RDX) with molecular resolution utilizing generalized energy-conserving dissipative particle dynamics with reactions. The model is parameterized with respect to the data from atomistic molecular dynamics simulations as well as from quantum mechanical calculations, thus bridging atomic processes to the mesoscales, including microstructures and defects. It accurately captures the response of RDX under a range of thermal loading conditions compared to atomistic simulations. In addition, the Hugoniot response of the CG model in the overdriven regime reasonably matches atomistic simulations and experiments. Exploiting the model's high computational efficiency, we investigate mesoscale systems involving millions of molecules and characterize size-dependent criticality of hotspots in RDX. The combination of accuracy and computational efficiency of our reactive model provides a tool for investigation of mesoscale phenomena, such as the role of microstructures and defects in the shock-to-deflagration transition, through particle-based simulation.

Generalized Energy-Conserving Dissipative Particle Dynamics with Mass Transfer. Part 1: Theoretical Foundation and Algorithm

An extension of the generalized energy-conserving dissipative particle dynamics method (GenDPDE) that allows mass transfer between mesoparticles via a diffusion process is presented. By considering the concept of the mesoparticles as property carriers, the complexity and flexibility of the GenDPDE framework were enhanced to allow for interparticle mass transfer under isoenergetic conditions, notated here as GenDPDE-M. In the formulation, diffusion is described via the theory of mesoscale irreversible processes based on linear relationships between the fluxes and thermodynamic forces, where their fluctuations are described by Langevin-like equations. The mass exchange between mesoparticles is such that the mass of the mesoparticle remains unchanged after the transfer process and requires additional considerations regarding the coupling with other system properties such as the particle internal energy. The proof-of-concept work presented in this article is the first part of a two-part article series. In Part 1, the development of the GenDPDE-M theoretical framework and the derivation of the algorithm are presented in detail. Part 2 of this article series is targeted for practitioners, where applications, demonstrations, and practical considerations for implementing the GenDPDE-M method are presented and discussed.

Generalized Energy-Conserving Dissipative Particle Dynamics with Mass Transfer. Part 2: Applications and Demonstrations

We present the second part of a two-part paper series intended to address a gap in computational capability for coarse-grain particle modeling and simulation, namely, the simulation of phenomena in which diffusion via mass transfer is a contributing mechanism. In part 1, we presented a formulation of a dissipative particle dynamics method to simulate interparticle mass transfer, termed generalized energy-conserving dissipative particle dynamics with mass transfer (GenDPDE-M). In the GenDPDE-M method, the mass of each mesoparticle remains constant following the interparticle mass exchange. In part 2 of this series, further verification and demonstrations of the GenDPDE-M method are presented for mesoparticles with embedded binary mixtures using the ideal gas (IG) and van der Waals (vdW) equation-of-state (EoS). The targeted readership of part 2 is toward practitioners, where applications and practical considerations for implementing the GenDPDE-M method are presented and discussed, including a numerical discretisztion algorithm for the equations-of-motion. The GenDPDE-M method is verified by reproducing the particle distributions predicted by Monte Carlo simulations for the IG and vdW fluids, along with several demonstrations under both equilibrium and non-equilibrium conditions. GenDPDE-M can be generally applied to multi-component mixtures and to other fundamental EoS, such as the Lennard-Jones or Exponential-6 models, as well as to more advanced EoS models such as Statistical Associating Fluid Theory.

Generalized Energy-Conserving Dissipative Particle Dynamics with Reactions

We present an extension of the generalized energy-conserving dissipative particle dynamics method (J. Bonet Avalos, et al., Phys Chem Chem Phys, 2019, 21, 24891-24911) to include chemical reactivity, denoted GenDPDE-RX. GenDPDE-RX provides a means of simulating chemical reactivity at the micro- and mesoscales, while exploiting the attributes of density- and temperature-dependent many-body force fields, which include improved transferability and scalability compared to two-body pairwise models. The GenDPDE-RX formulation considers intra-particle reactivity via a coarse-grain reactor construct. Extent-of-reaction variables assigned to each coarse-grain particle monitor the temporal evolution of the prescribed reaction mechanisms and kinetics assumed to occur within the particle. Descriptions of the algorithm, equations of motion, and numerical discretization are presented, followed by verification of the GenDPDE-RX method through comparison with reaction kinetics theoretical model predictions. Demonstrations of the GenDPDE-RX method are performed using constant-volume adiabatic heating simulations of three different reaction models, including both reversible and irreversible reactions, as well as multistep reaction mechanisms. The selection of the demonstrations is intended to illustrate the flexibility and generality of the method but is inspired by real material systems that span from fluids to solids. Many-body force fields using analytical forms of the ideal gas, Lennard-Jones, and exponential-6 equations of state are used for demonstration, although application to other forms of equation of states is possible. Finally, the flexibility of the GenDPDE-RX framework is addressed with a brief discussion of other possible adaptations and extensions of the method.

Structural properties of cationic surfactant-fatty alcohol bilayers: insights from dissipative particle dynamics

Bilayers, self-assembled by cationic surfactants and fatty alcohols in water, are the basic units of lamellar gel networks - creamy formulations extensively used in cosmetics and pharmaceutics. Mesoscopic modelling and study of the bilayers formed by single- or double-tail cationic surfactants (CTAC or DHDAC), and fatty alcohols (FAs) in the lamellar fluid and gel phases were employed. Fatty alcohols with alkyl tail equal to or greater than the surfactant alkyl tail, i.e., C₁₆FA or C₁₈FA and C₂₂FA, were considered. A model formulation was explored with the FA concentration greater than that of the surfactant and the structure of the fluid and gel bilayers in tensionless state characterised via the density profiles across the bilayers, orientational order parameters of the surfactant and FA chains, intrinsic analysis of the bilayer interfaces, and bending rigidity. The intrinsic analysis allows identification and quantification of the coexistence of the interdigitated and non-interdigitated phases present within the gel bilayers. The FA chains were found to conform the primary scaffolding of the bilayers while the surfactant chains tessellate bilayer monolayers from their water-hydrophobic interface. Further, the overlap of the FA chains from the apposed monolayers of the fluid bilayers rises with increasing FA length. Finally, the prevalence of the non-interdigitated phase over the interdigitated phase within the gel bilayers becomes enhanced upon the FA length increase with a preference of the surfactant chains to reside in the non-interdigitated phase rather than the interdigitated phase.

Generalized energy-conserving dissipative particle dynamics revisited: Insight from the thermodynamics of the mesoparticle leading to an alternative heat flow model

Recently we introduced the generalized energy-conserving dissipative particle dynamics method (GenDPDE) [J. Bonet Avalos, M. Lísal, J. P. Larentzos, A. D. Mackie, and J. K. Brennan, Phys. Chem. Chem. Phys. 21, 24891 (2019)]PPCPFQ1463-907610.1039/C9CP04404C, which has been formulated for an emerging class of density- and temperature-dependent coarse-grain models. In the original work, GenDPDE was formulated to ensure a fundamental link is maintained with the underlying physical system at the higher resolution scale. In this paper, we revisit the formulation of the GenDPDE method, and rederive the particle thermodynamics to ensure consistency at the opposing scale extreme, i.e., between the local thermodynamics in the mesoscopic systems and the corresponding macroscopic properties. We demonstrate this consistency by introducing unambiguous, physically meaningful definitions of the heat and work, which lead to the formulation of an alternative heat flow model that is analogous to Fourier's law of heat conduction. We present further analysis of the internal, unresolved degrees-of-freedom of the mesoparticles by considering the thermodynamics of an individual mesoparticle within the GenDPDE framework. Several key outcomes of the analysis include: (i) demonstration that the choice of the independent variables alters the particle thermodynamic description; (ii) demonstration that the mesoscopic thermodynamic transformations introduce additional terms of the order of the size of the local fluctuations, which prevent an unambiguous definition of both the heat and work; (iii) an emphasis on the importance of the choice of the proper estimators of the thermodynamic properties that are embedded in the chosen thermodynamic description; and (iv) a clearly defined path for determining any thermodynamic quantity dressed by the fluctuations. The further insight provided by this deeper analysis is useful for both readers interested in the GenDPDE theoretical framework, as well as readers interested in the practical ramifications of the analysis, namely, the alternative heat flow model.

Operator learning for predicting multiscale bubble growth dynamics

Simulating and predicting multiscale problems that couple multiple physics and dynamics across many orders of spatiotemporal scales is a great challenge that has not been investigated systematically by deep neural networks (DNNs). Herein, we develop a framework based on operator regression, the so-called deep operator network (DeepONet), with the long-term objective to simplify multiscale modeling by avoiding the fragile and time-consuming "hand-shaking" interface algorithms for stitching together heterogeneous descriptions of multiscale phenomena. To this end, as a first step, we investigate if a DeepONet can learn the dynamics of different scale regimes, one at the deterministic macroscale and the other at the stochastic microscale regime with inherent thermal fluctuations. Specifically, we test the effectiveness and accuracy of the DeepONet in predicting multirate bubble growth dynamics, which is described by a Rayleigh-Plesset (R-P) equation at the macroscale and modeled as a stochastic nucleation and cavitation process at the microscale by dissipative particle dynamics (DPD). First, we generate data using the R-P equation for multirate bubble growth dynamics caused by randomly time-varying liquid pressures drawn from Gaussian random fields (GRFs). Our results show that properly trained DeepONets can accurately predict the macroscale bubble growth dynamics and can outperform long short-term memory networks. We also demonstrate that the DeepONet can extrapolate accurately outside the input distribution using only very few new measurements. Subsequently, we train the DeepONet with DPD data corresponding to stochastic bubble growth dynamics. Although the DPD data are noisy and we only collect sparse data points on the trajectories, the trained DeepONet model is able to predict accurately the mean bubble dynamics for time-varying GRF pressures. Taken together, our findings demonstrate that DeepONets can be employed to unify the macroscale and microscale models of the multirate bubble growth problem, hence providing new insight into the role of operator regression via DNNs in tackling realistic multiscale problems and in simplifying modeling with heterogeneous descriptions.

Bottom-up coarse-grained models for external fields and interfaces

Bottom-up coarse-grained (CG) models accurately describe the structure of homogeneous systems but sometimes provide limited transferability and a poor description of thermodynamic properties. Consequently, inhomogeneous systems present a severe challenge for bottom-up models. In this work, we examine bottom-up CG models for interfaces and inhomogeneous systems. We first analyze the effect of external fields upon the many-body potential of mean force. We also demonstrate that the multiscale CG (MS-CG) variational principle for modeling the external field corresponds to a generalization of the first Yvon-Born-Green equation. This provides an important connection with liquid state theory, as well as physical insight into the structure of interfaces and the resulting MS-CG models. We then develop and assess MS-CG models for a film of liquid methanol that is adsorbed on an attractive wall and in coexistence with its vapor phase. While pair-additive potentials provide unsatisfactory accuracy and transferability, the inclusion of local-density (LD) potentials dramatically improves the accuracy and transferability of the MS-CG model. The MS-CG model with LD potentials quite accurately describes the wall-liquid interface, the bulk liquid density, and the liquid-vapor interface while simultaneously providing a much improved description of the vapor phase. This model also provides an excellent description of the pair structure and pressure-density equation of state for the bulk liquid. Thus, LD potentials hold considerable promise for transferable bottom-up models that accurately describe the structure and thermodynamic properties of both bulk and interfacial systems.