Can Machine Learning Predict the Phase Behavior of Surfactants?

We explore the prediction of surfactant phase behavior using state-of-the-art machine learning methods, using a data set for twenty-three nonionic surfactants. Most machine learning classifiers we tested are capable of filling in missing data in a partially complete data set. However, strong data bias and a lack of chemical space information generally lead to poorer results for entire de novo phase diagram prediction. Although some machine learning classifiers perform better than others, these observations are largely robust to the particular choice of algorithm. Finally, we explore how de novo phase diagram prediction can be improved by the inclusion of observations from state points sampled by an analogy to commonly used experimental protocols. Our results indicate what factors should be considered when preparing for machine learning prediction of surfactant phase behavior in future studies.

Partial osmotic pressures of ions in electrolyte solutions and the Gibbs-Guggenheim uncertainty principle.

The concept of the partial osmotic pressure of ions in an electrolyte solution is critically examined. In principle these can be defined by introducing a solvent-permeable wall and measuring the force per unit area which can certainly be attributed to individual ions. Here I demonstrate that although the total wall force balances the bulk osmotic pressure as required by mechanical equilibrium, the individual partial osmotic pressures are extrathermodynamic quantities dependent on the electrical structure at the wall, and as such they resemble attempts to define individual ion activity coefficients. The limiting case where the wall is a barrier to only one species of ion is also considered, and with ions on both sides the classic Gibbs-Donnan membrane equilibrium is recovered, thus providing a unifying treatment. The analysis can be extended to illustrate how the electrical state of the bulk is affected by the nature of the walls and the container handling history, thus supporting the "Gibbs-Guggenheim uncertainty principle," namely the notion that the electrical state is unmeasurable and usually accidentally determined. Since this uncertainty is conferred also onto individual ion activities, it has implications for the current (2002) IUPAC definition of pH.

Phase Behavior of Alkyl Ethoxylate Surfactants in a Dissipative Particle Dynamics Model.

We present a dissipative particle dynamics (DPD) model capable of capturing the liquid state phase behavior of nonionic surfactants from the alkyl ethoxylate (CnEm) family. The model is based upon our recent work [Anderson et al. J. Chem. Phys. 2017, 147, 094503] but adopts tighter control of the molecular structure by setting the bond angles with guidance from molecular dynamics simulations. Changes to the geometry of the surfactants were shown to have little effect on the predicted micelle properties of sampled surfactants, or the water-octanol partition coefficients of small molecules, when compared to the original work. With these modifications the model is capable of reproducing the binary water-surfactant phase behavior of nine surfactants (C8E4, C8E5, C8E6, C10E4, C10E6, C10E8, C12E6, C12E8, and C12E12) with a good degree of accuracy.

Modeling Alkyl Aromatic Hydrocarbons with Dissipative Particle Dynamics.

Building on previous work studying alkanes, we develop a dissipative particle dynamics (DPD) model to capture the behavior of the alkyl aromatic hydrocarbon family under ambient conditions of 298 K and 1 atmosphere. Such materials are of significant worldwide industrial importance in applications such as solvents, chemical intermediates, surfactants, lubricating oils, hydraulic fluids, and greases. We model both liquids and waxy solids for molecules up to 36 carbons in size and demonstrate that we can correctly capture both the freezing transition and liquid-phase densities in pure substances and mixtures. We also demonstrate the importance of including specialized bead types into the DPD model (rather than solely relying on generic bead types) to capture specific local geometrical constructs such as the benzene ring found in the benzyl chemical group; this can be thought of as representing subtle real-world many-body effects via customized pairwise non-bonded potentials.

Systematic Parameterization of Ion-Surfactant Interactions in Dissipative Particle Dynamics Using Setschenow Coefficients.

Dissipative particle dynamics (DPD) simulations of nonionic surfactants with an added salt show that the Setschenow relationship is reproduced; that is, the critical micelle concentration is log-linearly dependent on the added salt concentration. The simulated Setschenow coefficients depend on the DPD bead-bead repulsion amplitudes, and matching to the experimentally determined values provides a systematic method to parameterize the interactions between salt ion beads and surfactant beads. The optimized ion-specific interaction parameters appear to be transferrable and follow the same trends as the empirical Hofmeister series.

Diffusion, phase behavior, and gelation in a two-dimensional layer of colloids in osmotic equilibrium with a polymer reservoir.

The addition of enough non-adsorbing polymers to an otherwise stable colloidal suspension gives rise to a variety of phase behaviors and kinetic arrest due to the depletion attraction induced between the colloids by the polymers. We report a study of these phenomena in a two-dimensional layer of colloids. The three-dimensional phenomenology of crystal-fluid coexistence is reproduced, but gelation takes a novel form, in which the strands in the gel structure are locally crystalline. We compare our findings with a previous simulation and theory and find substantial agreement.

Critical Micelle Concentrations in Surfactant Mixtures and Blends by Simulation.

We explore the use of coarse-grained dissipative particle dynamics simulations to predict critical micelle concentrations (CMCs) in polydisperse surfactant mixtures and blends. By fitting pseudo-phase separation models (PSMs) to aqueous solutions of binary surfactant mixtures at selected compositions above the CMC, we avoid the need for expensive simulations of more complex multicomponent mixtures performed as a function of dilution. The approach is demonstrated for sodium laureth sulfate (SLES) surfactants with polydispersity in the ethoxylate spacer. For this system, we find a modest degree of cooperativity in micelle formation, which we attribute to the reduced repulsion between charged headgroups for surfactants with dissimilar ethoxylate spacer lengths. However, this is insufficient to explain the lowered CMC often observed in commercial SLES samples, which we attribute to the presence of small amounts of unsulfated alkyl ethoxylates and/or traces of salt.

Translation of Chemical Structure into Dissipative Particle Dynamics Parameters for Simulation of Surfactant Self-Assembly.

Dissipative particle dynamics (DPD) can be used to simulate the self-assembly properties of surfactants in aqueous solutions, but in order to simulate a new compound, a large number of new parameters are required. New methods for the calculation of reliable DPD parameters directly from chemical structure are described, allowing the DPD approach to be applied to a much wider range of organic compounds. The parameters required to describe the bonded interactions between DPD beads were calculated from molecular mechanics structures. The parameters required to describe the nonbonded interactions were calculated from surface site interaction point (SSIP) descriptions of molecular fragments that represent individual beads. The SSIPs were obtained from molecular electrostatic potential surfaces calculated using density functional theory and used in the SSIMPLE algorithm to calculate transfer free energies between different bead liquids. This approach was used to calculate DPD parameters for a range of different types of surfactants, which include ester, amide, and sugar moieties. The parameters were used to simulate the self-assembly properties in aqueous solutions, and comparison of the results for 27 surfactants with the available experimental data shows that these DPD simulations accurately predict critical micelle concentrations, aggregation numbers, and the shapes of the supramolecular assemblies formed. The methods described here provide a general approach to determining DPD parameters for neutral organic compounds of arbitrary structure.

Wax Formation in Linear and Branched Alkanes with Dissipative Particle Dynamics.

We present a dissipative particle dynamics (DPD) model for wax formation (i.e., the freezing transition) in linear and branched alkanes at room temperature (298 K) and atmospheric pressure. We parametrize the model using pure liquid phase densities and the onset of wax formation as a function of alkyl chain length. Significant emphasis is placed on building an accurate representation of the underlying molecular architecture by careful consideration of bond lengths and angles, aided by distributions obtained from molecular dynamics simulation. Using the derived model, we observe wax formation in n-alkanes when the alkyl chain length is greater than 18 (n-octadecane), in excellent agreement with experimental observations. Further, we reproduce the behavior of branched alkanes and mixtures including solubilities of heavy alkanes in light alkane solvents.

Non-Faradaic Electric Currents in the Nernst-Planck Equations and Nonlocal Diffusiophoresis of Suspended Colloids in Crossed Salt Gradients.

In the Nernst-Planck equations in two or more dimensions, a non-Faradaic solenoidal current can arise as a consequence of connecting patches with different liquid junction potentials. Whereas this current vanishes for binary electrolytes or in one-dimensional problems, it is in general nonvanishing, for example, in crossed salt gradients. For a suspended colloidal particle, chemiphoresis in the concentration gradients is generally vectorially misaligned with electrophoresis in the electrostatic potential gradient, and there is a nonlocal contribution to the latter deriving from the Ohmic electric field associated with the current; in a case study this contributes up to 20%-30% of the total effect.

A Surface Site Interaction Point Method for Dissipative Particle Dynamics Parametrization: Application to Alkyl Ethoxylate Surfactant Self-Assembly.

Dissipative particle dynamics (DPD) is a coarse-grained approach to the simulation of large supramolecular systems, but one limitation has been that the parameters required to describe the noncovalent interactions between beads are not readily accessible. A first-principles computational method has been developed so that bead interaction parameters can be calculated directly from ab initio gas-phase molecular electrostatic potential surfaces of the molecular fragments that represent the beads. A footprinting algorithm converts the molecular electrostatic potential surfaces into a discrete set of surface site interaction points (SSIPs), and these SSIPs are used in the SSIMPLE (surface site interaction model for the properties of liquids at equilibrium) algorithm to calculate the free energies of transfer of one bead into a solution of any other bead. The bead transfer free energies are then converted into the required DPD interaction parameters for all pairwise combinations of different beads. The reliability of the parameters was demonstrated using DPD simulations of a range of alkyl ethoxylate surfactants. The simulations reproduce the experimentally determined values of the critical micelle concentration and mean aggregation number well for all 22 surfactants studied.

Diffusiophoresis in ionic surfactants: effect of micelle formation.

We explore the consequences of micelle formation for diffusiophoresis of charged colloidal particles in ionic surfactant concentration gradients, using a quasi-chemical association model for surfactant self assembly. The electrophoretic contribution to diffusiophoresis is determined by re-arranging the Nernst-Planck equations, and the chemiphoretic contribution is estimated by making plausible approximations for the density profiles in the electrical double layer surrounding the particle. For sub-micellar solutions we find that a particle will typically be propelled down the concentration gradient, although electrophoresis and chemiphoresis are finely balanced and the effect is sensitive to the detailed parameter choices and simplifying assumptions in the model. Above the critical micelle concentration (CMC), diffusiophoresis becomes much weaker and may even reverse sign, due to the fact that added surfactant goes into building micelles and not augmenting the monomer or counterion concentrations. We present detailed calculations for sodium dodecyl sulfate (SDS), finding that the typical drift speed for a colloidal particle in a ∼100 µm length scale SDS gradient is ∼1 µm s-1 below the CMC, falling to ⪅0.2 µm s-1 above the CMC. These predictions are broadly in agreement with recent experimental work.

Why Clothes Don't Fall Apart: Tension Transmission in Staple Yarns.

The problem of how staple yarns transmit tension is addressed within abstract models in which the Amontons-Coulomb friction laws yield a linear programing (LP) problem for the tensions in the fiber elements. We find there is a percolation transition such that above the percolation threshold the transmitted tension is in principle unbounded. We determine that the mean slack in the LP constraints is a suitable order parameter to characterize this supercritical state. We argue the mechanism is generic, and in practical terms, it corresponds to a switch from a ductile to a brittle failure mode accompanied by a significant increase in mechanical strength.

Micelle Formation in Alkyl Sulfate Surfactants Using Dissipative Particle Dynamics.

We use dissipative particle dynamics (DPD) to study micelle formation in alkyl sulfate surfactants, with alkyl chain lengths ranging from 6 to 12 carbon atoms. We extend our recent DPD force field [ J. Chem. Phys. 2017 , 147 , 094503 ] to include a charged sulfate chemical group and aqueous sodium ions. With this model, we achieve good agreement with the experimentally reported critical micelle concentrations (CMCs) and can match the trend in mean aggregation numbers versus alkyl chain length. We determine the CMC by fitting a charged pseudophase model to the dependence of the free surfactant on the total surfactant concentration above the CMC and compare it with a direct operational definition of the CMC as the point at which half of the surfactant is classed as micellar and half as monomers and submicellar aggregates. We find that the latter provides the best agreement with experimental results. Finally, with the same model, we are able to observe the sphere-to-rod morphological transition for sodium dodecyl sulfate (SDS) micelles and determine that it corresponds to SDS concentrations in the region of 300-500 mM.

Diffusiophoresis in one-dimensional solute gradients.

The diffusiophoretic motion of suspended colloidal particles under one-dimensional solute gradients is solved using numerical and analytical techniques. Similarity solutions are developed for the injection and withdrawal dynamics of particles into semi-infinite pores. Furthermore, a method of characteristics formulation of the diffusion-free particle transport model is presented and integrated to realize particle trajectories. Analytical solutions are presented for the limit of small particle diffusiophoretic mobility Γp relative to the solute diffusivity Ds for particle motions in both semi-infinite and finite domains. Results confirm the build up of local maxima and minima in the propagating particle front dynamics. The method of characteristics is shown to successfully predict particle motions and the position of the particle front, although it fails to accurately predict suspended particle concentrations in the vicinity of sharp gradients, such as at the particle front peak seen in some injection cases, where particle diffusion inevitably plays an important role. Results inform the design of applications in which the use of applied solute gradients can greatly enhance particle injection into and withdrawal from pores.

Dissipative particle dynamics: Systematic parametrization using water-octanol partition coefficients.

We present a systematic, top-down, thermodynamic parametrization scheme for dissipative particle dynamics (DPD) using water-octanol partition coefficients, supplemented by water-octanol phase equilibria and pure liquid phase density data. We demonstrate the feasibility of computing the required partition coefficients in DPD using brute-force simulation, within an adaptive semi-automatic staged optimization scheme. We test the methodology by fitting to experimental partition coefficient data for twenty one small molecules in five classes comprising alcohols and poly-alcohols, amines, ethers and simple aromatics, and alkanes (i.e., hexane). Finally, we illustrate the transferability of a subset of the determined parameters by calculating the critical micelle concentrations and mean aggregation numbers of selected alkyl ethoxylate surfactants, in good agreement with reported experimental values.

Low-Cost Zeta Potentiometry Using Solute Gradients.

The zeta potential is an electric potential in the Debye screening layer of an electrolyte, which represents a key physicochemical surface property in various fields ranging from electrochemistry to pharmaceuticals. Thus, characterizing the zeta potential is essential for many applications, but available measurement techniques are limited. Electrophoretic light scattering is typically used to measure the zeta potential of particles in suspension, whereas zeta potential measurements of a solid wall in solution rely on either streaming potential or electroosmotic mobility measurement techniques, both of which are expensive and sophisticated. Here, a simple, robust method to simultaneously measure the zeta potential of particles in suspension and solid walls is presented. The method uses solute gradients to induce particle and fluid motions via diffusiophoresis and diffusioosmosis, respectively, which are both sensitive to the zeta potential of the particle and the wall. By visualizing the particle dynamics, both zeta potentials can be determined independently. Finally, a compact microscope is used to demonstrate low-cost zeta potentiometry that allows measurement of both particle and wall zeta potentials, which suggests a cost-effective tool for pharmaceuticals as well as for educational purposes.

Membraneless water filtration using CO2.

Water purification technologies such as microfiltration/ultrafiltration and reverse osmosis utilize porous membranes to remove suspended particles and solutes. These membranes, however, cause many drawbacks such as a high pumping cost and a need for periodic replacement due to fouling. Here we show an alternative membraneless method for separating suspended particles by exposing the colloidal suspension to CO2. Dissolution of CO2 into the suspension creates solute gradients that drive phoretic motion of particles. Due to the large diffusion potential generated by the dissociation of carbonic acid, colloidal particles move either away from or towards the gas-liquid interface depending on their surface charge. Using the directed motion of particles induced by exposure to CO2, we demonstrate a scalable, continuous flow, membraneless particle filtration process that exhibits low energy consumption, three orders of magnitude lower than conventional microfiltration/ultrafiltration processes, and is essentially free from fouling.

Perspective: Dissipative particle dynamics.

Dissipative particle dynamics (DPD) belongs to a class of models and computational algorithms developed to address mesoscale problems in complex fluids and soft matter in general. It is based on the notion of particles that represent coarse-grained portions of the system under study and allow, therefore, reaching time and length scales that would be otherwise unreachable from microscopic simulations. The method has been conceptually refined since its introduction almost twenty five years ago. This perspective surveys the major conceptual improvements in the original DPD model, along with its microscopic foundation, and discusses outstanding challenges in the field. We summarize some recent advances and suggest avenues for future developments.

Diffusiophoresis in nonadsorbing polymer solutions: The Asakura-Oosawa model and stratification in drying films.

A colloidal particle placed in an inhomogeneous solution of smaller nonadsorbing polymers will move towards regions of lower polymer concentration in order to reduce the free energy of the interface between the surface of the particle and the solution. This phenomenon is known as diffusiophoresis. Treating the polymer as penetrable hard spheres, as in the Asakura-Oosawa model, a simple analytic expression for the diffusiophoretic drift velocity can be obtained. In the context of drying films we show that diffusiophoresis by this mechanism can lead to stratification under easily accessible experimental conditions. By stratification we mean spontaneous formation of a layer of polymer on top of a layer of the colloid. Transposed to the case of binary colloidal mixtures, this offers an explanation for the stratification observed recently in these systems [A. Fortini et al., Phys. Rev. Lett. 116, 118301 (2016)PRLTAO0031-900710.1103/PhysRevLett.116.118301]. Our results emphasize the importance of treating solvent dynamics explicitly in these problems and caution against the neglect of hydrodynamic interactions or the use of implicit solvent models in which the absence of solvent backflow results in an unbalanced osmotic force that gives rise to large but unphysical effects.