David Keffer

Molecular Dynamic Simulations of the Effect on the Hydration of Nafion in the Presence of a Platinum Nanoparticle

Myvizhi Esai Selvan et al. — Fri, 31 Aug 2012 13:15:50 PDT

Platinum catalysts play a critical role in fuel cell technology. Current optimization efforts focus on reducing the amount of Pt in the system and optimizing the utilization of that which remains. The effect of the presence of Pt nanoparticles on the local structure and morphology of the polymer electrolyte membrane, water, and hydronium ions has been studied at molecular level in this work. Classical molecular dynamics simulation has been used to examine a system containing a 4 nm fcc cubic ({100} face) platinum nanoparticle at the center surrounded by Nafion polymer, water molecules, and hydronium ions at λ = 3, 6, 9, 15, and 22. The changes in density and orientation distribution of sulfonic acid groups in the side-chains, water, and hydronium as a function of distance from platinum surface are analyzed in this study. Sulfonic acid groups and hydronium ions show a very high increase in density near the platinum surface, and they approach the bulk value as they move away from the platinum surface. At lower water contents (λ = 3 and 6), water is strongly attracted to platinum surface, increasing the density near the platinum surface. However, at the highest humidity level studied, the density of water farthest from the platinum surface (>20 Å) is higher than the bulk value and the density of water nearest to the platinum surface. The above observed phenomenon is explained using the probability distribution of orientation computed at various distances from platinum surface. As the water content increases, sulfonic acid groups show a preferential orientation for the side-chains to align themselves horizontally on the platinum than vertically, thereby covering a larger area of platinum and pushing the water away from the platinum surface. This causes the water density to drop near the platinum surface. Water molecules and hydronium ions show preference to certain orientations near the platinum surface, but usually to none as they move away from the platinum surface.

DOI: 10.1021/jp3020436

Multi-scale Models for Sulfonated Cross-linked Poly (1, 3-cyclohexadiene) Polymer

Qifei Wang et al. — Fri, 31 Aug 2012 13:12:23 PDT

Atomistic and coarse-grained (CG) models of cross-linked sulfonated Poly (1, 3-cyclohexadiene) (xsPCHD) were developed and implemented in Molecular Dynamics (MD) simulations of PCHD chains with different architectures. In the atomistic model, PCHD chains are cross linked by a sulfur–sulfur bond. Sulfonic acid groups are evenly distributed along the chain. The architecture is specifically aimed for application as a proton exchange membrane used in fuel cells. An atomistic force field for this architecture was tested and applied in the atomistic MD simulation of xsPCHD for the first time. The atomistic simulations generate the density and cross-linker separation distribution. To further study the structural properties of longer chain systems, a CG model was proposed. The bonded structural probability distribution functions (PDFs) and non-bonded pair correlation function (PCF) of the CG beads were obtained from the atomistic simulation results. The bonded CG potentials are obtained by simple inversion of the corresponding PDFs. The CG non-bonded potential is parameterized to the PCF using the Iterative Boltzmann Inversion (IBI) method. The CGMD simulations of xsPCHD chains using potentials from above method satisfactorily reproduce the structural properties from atomistic MD simulation of the same system. The transferability of the CG potentials has been further tested through CGMD simulation of xsPCHD homopolymer with different architectures.

DOI: http://dx.doi.org/10.1016/j.polymer.2012.02.005

A Coarse-grained Model for Polyethylene Glycol (PEG) Polymer

Qifei Wang et al. — Fri, 31 Aug 2012 13:09:47 PDT

A coarse-grained (CG) model of polyethylene glycol (PEG) was developed and implemented in CG molecular dynamics (MD) simulations of PEG chains with degree of polymerization (DP) 20 and 40. In the model, two repeat units of PEG are grouped as one CG bead. Atomistic MD simulation of PEG chains with DP = 20 was first conducted to obtain the bonded structural probability distribution functions (PDFs) and nonbonded pair correlation function (PCF) of the CG beads. The bonded CG potentials are obtained by simple inversion of the corresponding PDFs. The CG nonbonded potential is parameterized to the PCF using both an inversion procedure based on the Ornstein-Zernike equation with the Percus-Yevick approximation (OZPY−1) and a combination of OZPY−1 with the iterative Boltzmann inversion (IBI) method (OZPY−1+IBI). As a simple one step method, the OZPY−1 method possesses an advantage in computational efficiency. Using the potential from OZPY−1 as an initial guess, the IBI method shows fast convergence. The coarse-grained molecular dynamics (CGMD) simulations of PEG chains with DP = 20 using potentials from both methods satisfactorily reproduce the structural properties from atomistic MD simulation of the same systems. The OZPY−1+IBI method yields better agreement than the OZPY−1 method alone. The new CG model and CG potentials from OZPY−1+IBI method was further tested through CGMD simulation of PEG with DP = 40 system. No significant changes are observed in the comparison of PCFs from CGMD simulations of PEG with DP = 20 and 40 systems indicating that the potential is independent of chain length.

DOI: http://dx.doi.org/10.1063/1.3664623

Molecular Simulations of H2 Adsorption in Metal-Porphyrin Frameworks (MPFs): A Potential New Material Evaluation

Ruichang Xiong et al. — Fri, 31 Aug 2012 13:08:04 PDT

Path integral grand canonical Monte Carlo (PI-GCMC) simulations using standard force fields are carried out to calculate the adsorption of H2 in five metal-porphyrin frameworks (MPFs), a new class of metal organic framework (MOF)-type materials. These simulations are performed at 77 K and room temperature (300 K). The adsorption isotherms of H2 in IRMOF-1 and IRMOF-10 are also calculated as a comparison. All calculations indicate that all MPFs adsorbed a higher weight fraction of H2 than both IRMOF-1 and IRMOF-10, with one exception (MPF-2). The gravimetric hydrogen capacities are still well short of practical goals. The MPFs provide additional adsorption sites due to the porphyrin. A statistical mechanical lattice model predicts the adsorption well at room temperature. The prediction by this model showed that a weight fraction of hydrogen of 6 wt. % adsorbed in pores of the size found in IRMOF-1 at ambient temperature and modest pressures required a binding energy of about 17 kJ/mole, which is consistent with other findings.

DOI: http://dx.doi.org/10.1063/1.3655373

Reactive Molecular Dynamics Study of Proton Transport in Polymer Electrolyte Membranes

Myvizhi Esai Selvan et al. — Fri, 31 Aug 2012 13:05:56 PDT

Dynamical properties of water and protons in Nafion with an equivalent weight of 1144 are studied using the recently developed reactive molecular dynamics (RMD) algorithm at various water contents. The structural diffusion of a proton along the aqueous domains is modeled via a mechanism similar to that observed in bulk aqueous systems. The algorithm implements reactivity in classical MD simulations by three steps: (i) satisfaction of the trigger, (ii) instantaneous reaction, and (iii) local equilibration. Two different schemes (Method 1 and Method 2) of execution of the algorithm are investigated, which differ in terms of the range of the local environment involved in the reaction. Both methods are parametrized to the experimental rate constant of proton transport in bulk water. The mean lifetime of the protons increased with the water content in Nafion. Water diffusivities of the reactive systems using the two schemes increased with the hydration level and were higher than the diffusion coefficients in the nonreactive system. The more detailed scheme in Method 2 which included the relaxation of the extended hydrogen bonding network around the proton-hopping reaction lowered the water diffusivity compared to that of Method 1 and also affected both the structural and vehicular components of diffusion. The total charge diffusion, vehicular component, and structural component increased with water content, and the two components of the charge diffusion were found to be negatively correlated. The negative correlation is due to preferential structural diffusion of the proton toward the sulfonate group, whereas vehicular diffusion tends to move the H3O+ ion away from the sulfonate group.

DOI: 10.1021/jp203443c

Effective potentials between nanoparticles in suspension

Gary S. Grest et al. — Fri, 31 Aug 2012 13:00:44 PDT

Results of molecular dynamics simulations are presented for the pair distribution function between nanoparticles in an explicit solvent as a function of nanoparticle diameter and interaction strength between the nanoparticle and solvent. The effect of including the solvent explicitly is demonstrated by comparing the pair distribution function of nanoparticles to that in an implicit solvent. The nanoparticles are modeled as a uniform distribution of Lennard-Jones particles, while the solvent is represented by standard Lennard-Jones particles. The diameter of the nanoparticle is varied from 10 to 25 times that of the solvent for a range of nanoparticle volume fractions. As the strength of the interactions between nanoparticles and the solvent increases, the solvent layer surrounding the nanoparticle is formed which increases the effective radii of the nanoparticles. The pair distribution functions are inverted using the Ornstein–Zernike integral equation to determine an effective pair potential between the nanoparticles mediated by the introduction of an explicit solvent.

DOI: http://dx.doi.org/10.1063/1.3578181

On the Relationship between the Structure of Metal-Organic Frameworks and the Adsorption and Diffusion of Hydrogen

Nethika S. Suraweera et al. — Fri, 31 Aug 2012 12:57:18 PDT

In this work, the adsorptive and diffusive behaviours of molecular hydrogen in 10 different isoreticular metal–organic frameworks (IRMOFs) are studied using molecular-level simulation. Hydrogen adsorption isotherms and heats of adsorption at 77 and 300 K were generated for 10 MOFs at low-pressure conditions (up to 10 bar) using Path Integral Grand Canonical Monte Carlo simulations. Self-diffusivities and activation energies for diffusion were generated using molecular dynamics simulation. Density distributions showing the location and the shape of the adsorption sites are also provided. Statistical correlations for all of the properties as a function of surface area (SA), accessible volume (AV) and binding energy are provided. Based on this work, we observe that at pressures up to 10 bar at 300 K, the adsorption process is virtually completely governed by entropic considerations, resulting in a strong correlation between the amount of hydrogen adsorbed and the AV of the adsorbent. At 77 K, we observe more than one adsorption regime. At low pressures, the adsorption process is governed by energetic considerations, resulting in a strong correlation between the amount of hydrogen adsorbed and the energy of adsorption. At the high end of the pressure range, the adsorption becomes a process dominated by entropic considerations, again resulting in a strong correlation between the amount of hydrogen adsorbed and the AV. Only in the intermediate regime does one observe that an increase in SA results in an increase in the amount of hydrogen adsorbed. The self-diffusivity of hydrogen at infinite dilution is highly correlated with both the energy of adsorption and the AV. The diffusion in larger IRMOFs is faster because of an entropic advantage and specifically not because of a lower activation energy for diffusion.

DOI: 10.1080/08927022.2011.561432

Toward a Predictive Understanding of Water and Charge Transport in Proton Exchange Membranes

Myvizhi Esai Selvan et al. — Fri, 31 Aug 2012 12:52:54 PDT

An analytical model for water and charge transport in highly acidic and highly confined systems such as proton exchange membranes of fuel cells is developed and compared to available experimental data. The model is based on observations from both experiment and multiscale simulation. The model accounts for three factors in the system including acidity, confinement, and connectivity. This model has its basis in the molecular-level mechanisms of water transport but has been coarse-grained to the extent that it can be expressed in an analytical form. The model uses the concentration of H3O+ ion to characterize acidity, interfacial surface area per water molecule to characterize confinement, and percolation theory to describe connectivity. Several important results are presented. First, an integrated multiscale simulation approach including both molecular dynamics simulation and confined random walk theory is capable of quantitatively reproducing experimentally measured self-diffusivities of water in the perfluorinated sulfonic acid proton exchange membrane material, Nafion. The simulations, across a range of hydration conditions from minimally hydrated to fully saturated, have an average error for the self-diffusivity of water of 16% relative to experiment. Second, accounting for three factors—acidity, confinement, and connectivity—is necessary and sufficient to understand the self-diffusivity of water in proton exchange membranes. Third, an analytical model based on percolation theory is capable of quantitatively reproducing experimentally measured self-diffusivities of both water and charge in Nafion across a full range of hydration.

DOI: 10.1021/jp1115004

Applications of a General Random Walk Theory for Confined Diffusion

Elisa M. Calvo-Muñoz et al. — Fri, 31 Aug 2012 12:47:01 PDT

A general random walk theory for diffusion in the presence of nanoscale confinement is developed and applied. The random-walk theory contains two parameters describing confinement: a cage size and a cage-to-cage hopping probability. The theory captures the correct nonlinear dependence of the mean square displacement (MSD) on observation time for intermediate times. Because of its simplicity, the theory also requires modest computational requirements and is thus able to simulate systems with very low diffusivities for sufficiently long time to reach the infinite-time-limit regime where the Einstein relation can be used to extract the self-diffusivity. The theory is applied to three practical cases in which the degree of order in confinement varies. The three systems include diffusion of (i) polyatomic molecules in metal organic frameworks, (ii) water in proton exchange membranes, and (iii) liquid and glassy iron. For all three cases, the comparison between theory and the results of molecular dynamics (MD) simulations indicates that the theory can describe the observed diffusion behavior with a small fraction of the computational expense. The confined-random-walk theory fit to the MSDs of very short MD simulations is capable of accurately reproducing the MSDs of much longer MD simulations. Furthermore, the values of the parameter for cage size correspond to the physical dimensions of the systems and the cage-to-cage hopping probability corresponds to the activation barrier for diffusion, indicating that the two parameters in the theory are not simply fitted values but correspond to real properties of the physical system.

Energetic and Entropic Elasticity of Nonisothermal Flowing Polymers: Experiment, Theory, and Simulation

David Keffer et al. — Wed, 19 May 2010 07:55:18 PDT

The thermodynamical aspects of polymeric liquids subjected to nonisothermal flow are examined from the complementary perspectives of theory, experiment, and simulation. In particular, attention is paid to the energetic effects, in addition to the entropic ones, that occur under conditions of extreme deformation. Comparisons of experimental measurements of the temperature rise generated under elongational flow at high strain rates with macroscopic finite element simulations offer clear evidence of the persistence and importance of energetic effects under severe deformation. The performance of various forms of the temperature equation are evaluated with regard to experiment, and it is concluded that the standard form of this evolution equation, arising from the concept of purely entropic elasticity, is inadequate for describing nonisothermal flow processes of polymeric liquids under high deformation. Complete temperature equations, in the sense that they possess a direct and explicit dependence on the energetics of the microstructure of the material, provide excellent agreement with experimental data.

Atomistic Simulation of Energetic and Entropic Elasticity in Short-chain Polyethylenes

David Keffer et al. — Wed, 19 May 2010 07:28:15 PDT

The thermodynamical aspects of polymeric liquids subjected to uniaxial elongational flow are examined using atomistically detailed nonequilibrium Monte Carlo simulations. In particular, attention is paid to the energetic effects, in addition to the entropic ones, which occur under conditions of extreme deformation. Atomistic nonequilibrium Monte Carlo simulations of linear polyethylene systems, ranging in molecular length from C₂₄ to C₇₈ and for temperatures from 300 to 450 K, demonstrate clear contributions of energetic effects to the elasticity of the system. These are manifested in a conformationally dependent heat capacity, which is significant under large deformations. Violations of the hypothesis of purely entropic elasticity are evident in these simulations, in that the free energy of the system is demonstrated to be composed of significant energetic effects under high degrees of orientation. These arise mainly from favorable intermolecular side-to-side interactions developing in the process of elongation due to chain uncoiling and alignment in the direction of extension.

Dynamics of Individual Molecules of Linear Polyethylene Liquids under Shear: Atomistic Simulation and Comparison with a Free-draining Bead-rod Chain

David Keffer et al. — Wed, 19 May 2010 06:57:14 PDT

Nonequilibrium molecular dynamics (NEMD) simulations of a dense liquid composed of linear polyethylene chains were performed to investigate the chain dynamics under shear. Brownian dynamics (BD) simulations of a freely jointed chain with equivalent contour length were also performed in the case of a dilute solution. This allowed for a close comparison of the chain dynamics of similar molecules for two very different types of liquids. Both simulations exhibited a distribution of the end-to-end vector, |R_ete|, with Gaussian behavior at low Weissenberg number (Wi). At high Wi, the NEMD distribution was bimodal, with two peaks associated with rotation and stretching of the individual molecules. BD simulations of a dilute solution did not display a bimodal character; distributions of |R_ete| ranged from tightly coiled to fully stretched configurations. The simulations revealed a tumbling behavior of the chains and correlations between the components of R_ete exhibited characteristic frequencies of tumbling, which scaled as Wi^−0.75. Furthermore, after a critical Wi of approximately 2, another characteristic time scale appeared which scaled as Wi^−0.63. Although the free-draining solution is very different than the dense liquid, the BD simulations revealed a similar behavior, with the characteristic time scales mentioned above scaling as Wi^−0.68 and Wi^−0.66.

Theoretical Calculation of Thermodynamic Properties of Naphthalene, Methylnaphthalenes, and Dimethylnaphthalenes

David Keffer et al. — Wed, 19 May 2010 06:18:17 PDT

For this work we performed quantum mechanical (QM) and statistical mechanical (SM) calculations to generate the entropy of 13 aromatic compoundsnaphthalene, 2 methylnaphthalene isomers, and 10 dimethylnaphthalene isomersin the ideal gas state. Density functional theory (DFT) was used to calculate the equilibrium structure and perform a full normal-mode analysis. The DFT level of theory used in this paper is B3LYP/6-31G(d,p). DFT has also been used to determine barriers for the internal rotation contribution to the entropy. For four compounds for which experimental data are available, the calculated entropies have been compared to the experimental values. The calculated entropies match experiment very well, with the percentage errors close to the experimental uncertainty, less than 0.4 %. The equilibrium distribution of dimethylnaphthalene isomers in the mixture is predicted using the calculated entropies and energies from QM and SM calculations in the 300 K to 740 K temperature range.

Surfactant and Electric Field Strength Effects on Surface Tension at Liquid/Liquid/Solid Interfaces

David Keffer et al. — Wed, 19 May 2010 06:10:14 PDT

We performed a series of experiments designed to elucidate the effects of the presence of sodium dodecyl sulfate (SDS) surfactant and an applied electrical field on the wetting behavior in a system containing a sessile droplet of phenylmethyl polysiloxane (PMPS) oil on a polished stainless steel surface submersed in aqueous solution. The voltage difference ranged from −3 to +3 V, which is at least 3 orders of magnitude smaller than from comparable recent work. We report the measured equilibrium contact angle of the droplet as a function of surfactant concentration and field strength. We then modeled the system. We solved the Laplace equation to obtain the 3D field within our system. We expanded the three surface tensions (oil droplet−aqueous solution (oa), oil droplet−metal surface (os), and aqueous solution−metal surface (as)) in a Taylor series with respect to surfactant concentration and local field strength. We use these three surface tensions in Young's equation to obtain the theoretical contact angle of the organic droplet. We demonstrate that the large changes in contact angle due to the simultaneous presence of small concentrations of surfactant and small voltage differences can be accounted for by changes in the oa and as surface tensions.

Comparison of the Hydration and Diffusion of Protons in Perfluorosulfonic Acid Membranes with Molecular Dynamics Simulations

David Keffer et al. — Wed, 19 May 2010 05:58:16 PDT

Classical molecular dynamics (MD) simulations were performed to determine the hydrated morphology and hydronium ion diffusion coefficients in two different perfluorosulfonic acid (PFSA) membranes as functions of water content. The structural and transport properties of 1143 equivalent weight (EW) Nafion, with its relatively long perfluoroether side chains, are compared to the short-side-chain (SSC) PFSA ionomer at an EW of 977. The separation of the side chains was kept uniform in both ionomers consisting of −(CF₂)₁₅− units in the backbone, and the degree of hydration was varied from 5 to 20 weight % water. The MD simulations indicated that the distribution of water clusters is more dispersed in the SSC ionomer, which leads to a more connected water-channel network at the low water contents. This suggests that the SSC ionomer may be more inclined to form sample-spanning aqueous domains through which transport of water and protons may occur. The diffusion coefficients for both hydronium ions and water molecules were calculated at hydration levels of 4.4, 6.4, 9.6, and 12.8 H₂O/SO₃H for each ionomer. When compared to experimental proton diffusion coefficients, this suggests that as the water content is increased the contribution of proton hopping to the overall proton diffusion increases.

A Reactive Molecular Dynamics Study of the Thermal Decomposition of Perfluorodimethyl Ether

David Keffer et al. — Wed, 19 May 2010 05:36:52 PDT

Classical reactive molecular dynamics (RMD) simulation is used to model the thermal decomposition of perfluorodimethyl ether (CF3OCF3), which is relevant as a simple molecule containing the necessary architectural elements to study the chemical stability of perfluoropolyether lubricants. The RMD algorithm employs nonreactive interaction potentials for the reactants and products. The reactivity is implemented through a coarse-grained simulation algorithm, incorporating elements from both the quantum and macroscopic descriptions of the reaction. The RMD scheme maps the quantum mechanically determined transition state onto a set of geometric triggers. When a configuration matching those triggers is found in the RMD simulation, the reaction instantaneously occurs. A brief, local equilibration process stabilizes the configuration, and the simulation continues. Using two geometric triggers, the RMD simulation can describe quantitatively the temperature dependence of the thermal decomposition of CF3OCF3, when compared to the quantum mechanical standard.

Molecular Dynamics Simulation of Poly(ethylene terephthalate) Oligomers

David Keffer et al. — Fri, 14 May 2010 08:46:37 PDT

Molecular dynamics simulations of poly(ethylene terephthalate) (PET) oligomers are performed in the isobaric−isothermal (NpT) ensemble at a state point typical of a finishing reactor. The oligomer size ranges from 1 to 10 repeat units. We report thermodynamic properties (density, potential energy, enthalpy, heat capacity, isothermal compressibility, and thermal expansivity), transport properties (self-diffusivity, zero-shear-rate viscosity, thermal conductivity), and structural properties (pair correlation functions, hydrogen bonding network, chain radius of gyration, chain end-to-end distance) as a function of oligomer size. We compare the results with existing molecular-level theories and experimental data. Scaling exponents as a function of degree of polymerization are extracted. The distribution of the end-to-end distance is bimodal for the dimer and gradually shifts to a single peak as the degree of polymerization increases. The scaling exponents for the average chain radius of gyration and end-to-end distance are 0.594 and 0.571, respectively. The values of the heat capacity, isothermal compressibility, and thermal expansivity agree well with the available experimental data, which are of much longer PET chains. The scaling exponents for the self-diffusivity and zero-shear-rate viscosity are, respectively, −2.01 and 0.96, with the latter one being close to the theoretical prediction 1.0 for short-chain polymers.

Molecular Simulation Images

David Keffer — Tue, 09 Mar 2010 09:00:59 PST

These animations and interactive structures are created from various molecular dynamics simulations and quantum calculations. In order to view the interactive structures, you need the free "Chime" Plug-in. In order to view the movie files (in avi format), you will require the following codec: TSCC codec. This work has been supported by DOE BES, AFOSR, NSF and ACS PRF.

Self-Consistent Multiscale Modeling in the Presence of Inhomogeneous Fields

David Keffer — Tue, 09 Mar 2010 08:35:27 PST

Molecular dynamics (MD) simulations of a Lennard–Jones fluid in an inhomogeneous external field generate steady-state profiles of density and pressure with nanoscopic heterogeneities. The continuum level of mass, momentum, and energy transport balances is capable of reproducing the MD profiles only when the equation of state for pressure as a function of density is extracted directly from the molecular level of description. We show that the density profile resulting from simulation is consistent with both a molecular-level theoretical prediction from statistical mechanics as well as the solution of the continuum-level set of differential equations describing the conservation of mass and momentum.