Unification of Ewald and shifted force methods to calculate Coulomb interactions in molecular simulations.

Three new Ewald series are derived using a new strategy that does not start with a proposed charge spreading function. Of these, the Ewald series produced using shifted potential interactions for the Ewald real space series converges relatively slowly, while the corresponding expression using a shifted force (SF) interaction does not converge. A comparison is made between several approximations of the Ewald method and the SF route to include Coulomb interactions in molecular dynamics (MD) computer simulations. MD simulations of a model bulk molten salt and water were carried out. The recently derived α' variant of Ewald, by K. D. Hammonds and D. M. Heyes [J. Chem. Phys. 157, 074108 (2022)], has been developed analytically and found to be more accurate and computationally efficient than SF in part due to the smaller real space truncation distance that can be used. In addition, with α', the number of reciprocal lattice vectors required is reduced considerably compared with the standard Ewald implementations to give the same accuracy. The invention of the α' method shifts the computational balance back toward using an Ewald construction. The SF method shows greater errors in the Coulomb pressure and time dependent fluctuation properties compared to α'. It does not conserve the shadow Hamiltonian in a microcanonical MD simulation, whereas the α' method does, which facilitates long time stability and insignificant drift of properties over time. The speed of the Ewald computer code is improved by using a new lookup table method.

Harmonic models and molecular dynamics simulations of isomorph behavior of Lennard-Jones fluids: Excess entropy and high temperature limiting behavior.

Henchman's approximate harmonic model of liquids is extended to predict the thermodynamic behavior along lines of constant excess entropy ("isomorphs") in the liquid and supercritical fluid regimes of the Lennard-Jones (LJ) potential phase diagram. Simple analytic expressions based on harmonic cell models of fluids are derived for the isomorph lines, one accurate version of which only requires as input parameters the average repulsive and attractive parts of the potential energy per particle at a single reference state point on the isomorph. The new harmonic cell routes for generating the isomorph lines are compared with those predicted by the literature molecular dynamics (MD) methods, the small step MD method giving typically the best agreement over a wide density and temperature range. Four routes to calculate the excess entropy in the MD simulations are compared, which includes employing Henchman's formulation, Widom's particle insertion method, thermodynamic integration, and parameterized LJ equations of state. The thermodynamic integration method proves to be the most computationally efficient. The excess entropy is resolved into contributions from the repulsive and attractive parts of the potential. The repulsive and attractive components of the potential energy, excess Helmholtz free energy, and excess entropy along a fluid isomorph are predicted to vary as ∼T-1/2 in the high temperature limit by an extension of classical inverse power potential perturbation theory statistical mechanics, trends that are confirmed by the MD simulations.

Departures from perfect isomorph behavior in Lennard-Jones fluids and solids.

Isomorphs are lines on a fluid or solid phase diagram along which the microstructure is invariant on affine density scaling of the molecular coordinates. Only inverse power (IP) and hard sphere potential systems are perfectly isomorphic. This work provides new theoretical tools and criteria to determine the extent of deviation from perfect isomorphicity for other pair potentials using the Lennard-Jones (LJ) system as a test case. A simple prescription for predicting isomorphs in the fluid range using the freezing line as a reference is shown to be quite accurate for the LJ system. The shear viscosity and self-diffusion coefficient scale well are calculated using this method, which enables comments on the physical significance of the correlations found previously in the literature to be made. The virial-potential energy fluctuation and the concept of an effective IPL system and exponent, n', are investigated, particularly with reference to the LJ freezing and melting lines. It is shown that the exponent, n', converges to the value 12 at a high temperature as ∼T-1/2, where T is the temperature. Analytic expressions are derived for the density, temperature, and radius derivatives of the radial distribution function along an isomorph that can be used in molecular simulation. The variance of the radial distribution function and radial fluctuation function are shown to be isomorph invariant.

Non-equilibrium molecular simulations of thin film rupture.

The retraction of thin films, as described by the Taylor-Culick (TC) theory, is subject to widespread debate, particularly for films at the nanoscale. We use non-equilibrium molecular dynamics simulations to explore the validity of the assumptions used in continuum models by tracking the evolution of holes in a film. By deriving a new mathematical form for the surface shape and considering a locally varying surface tension at the front of the retracting film, we reconcile the original theory with our simulation to recover a corrected TC speed valid at the nanoscale.

Bulk viscosity of hard sphere fluids by equilibrium and nonequilibrium molecular dynamics simulations.

The bulk viscosity, ηb, of the hard sphere (HS) fluid is computed by equilibrium and nonequilibrium molecular dynamics (NEMD) simulations, the latter using an adaptation of the time-stepping method for continuous potential systems invented by Hoover et al. [Phys. Rev. A 21, 1756 (1980)], which employs an imposed cyclic density variation on the system by affine scaling of the particle coordinates. The time-stepping method employed for HS is validated against exact event-driven hard sphere methodology for a series of equilibrium quantities over a wide density range, including the pressure, singular parts of the hard sphere viscosities, and the nonsingular parts of the shear viscosity time correlation functions. The time steps used are typically only a little smaller than those employed in continuous potential simulations. Exact pressure tensor fluctuation expressions are derived for the singular (or infinite limiting frequency) equilibrium parts of the viscosities, which were employed in the simulations. The values obtained agree well with the predictions of the Enskog theory for all densities considered. The bulk viscosity obtained by NEMD is shown to be noticeably frequency dependent for densities in excess of ∼0.8, decaying approximately exponentially to the Enskog and equilibrium simulation values at all densities considered for frequencies in excess of ∼5 in hard sphere units. Temperature profiles during the cycle and the effects of strain amplitude on the computed frequency dependent bulk viscosity are presented. The bulk viscosity increases with the maximum density amplitude.

Optimization of the Ewald method for calculating Coulomb interactions in molecular simulations.

Practical implementations of the Ewald method used to compute Coulomb interactions in molecular dynamics simulations are hampered by the requirement to truncate its reciprocal space series. It is shown that this can be mitigated by representing the contributions from the neglected reciprocal lattice vector terms as a simple modification of the real space expression in which the real and reciprocal space series have slightly different charge spreading parameters. This procedure, called the α' method, enables significantly fewer reciprocal lattice vectors to be taken than is currently typical for Ewald, with negligible additional computational cost, which is validated on model systems representing different classes of charged system, a CsI crystal and melt, water, and a room temperature ionic liquid. A procedure for computing accurate energies and forces based on a periodic sampling of an additional number of reciprocal lattice vectors is also proposed and validated by the simulations. The convergence characteristics of expressions for the pressure based on the forces and the potential energy are compared, which is a useful assessment of the accuracy of the simulations in reproducing the Coulomb interaction. The techniques developed in this work can reduce significantly the total computer simulation times for medium sized charged systems, by factors of up to ∼5 for those in the classes studied here.

Intrinsic viscuit probability distribution functions for transport coefficients of liquids and solids.

A reformulation of the Green-Kubo expressions for the transport coefficients of liquids in terms of a probability distribution function (PDF) of short trajectory contributions, which were named "viscuits," has been explored in a number of recent publications. The viscuit PDF, P, is asymmetric on the two sides of the distribution. It is shown here using equilibrium 3D and 2D molecular dynamics simulations that the viscuit PDF of a range of simple molecular single component and mixture liquid and solid systems can be expressed in terms of the same intrinsic PDF (P0), which is derived from P with the viscuit normalized by the standard deviation separately on each side of the distribution. P0 is symmetric between the two sides and can be represented for not very small viscuit values by the same gamma distribution formulated in terms of a single disposable parameter. P0 tends to an exponential in the large viscuit wings. Scattergrams of the viscuits and their associated single trajectory correlation functions are shown to distinguish effectively between liquids, solids, and glassy systems. The so-called viscuit square root method for obtaining the transport coefficients is shown to be a useful probe of small and statistically zero self-diffusion coefficients of molecules in the liquid and solid states, respectively. The results of this work suggest that the transport coefficients have a common underlying physical origin, reflecting at a coarse-grained level the traversal statistics of the system through its high-dimensioned potential energy landscape.

Application of cell models to the melting and sublimation lines of the Lennard-Jones and related potential systems.

Harmonic cell models (HCMs) are shown to predict the melting line of the Lennard-Jones (LJ) but not the sublimation line. In addition, even for the melting line, the HCMs are found to be physically unrealistic for inverse power potential systems near the hard-sphere limit, and for the Weeks-Chandler-Andersen system at extremely low temperatures. Despite this, the HCM accurately predicts the LJ mean-square displacement (MSD) from molecular-dynamics (MD) simulations along both lines after simple scaling corrections, to include the effects of anharmonicity and correlated dynamics of the atoms, are applied. Single caged atom molecular dynamics and Monte Carlo simulations provide further quantitative characterization of these additional effects, which go beyond harmonicity. The melting indicator and a modification of the cell model in a similar form are shown to be approximately constant along the melting line, which indicates an isomorph. The less well studied LJ sublimation line is shown not to be an isomorph, yet it still can be represented analytically very accurately by the relationship k_{B}T=aρ^{4}+bρ^{2}, where a and b are constants (k_{B} is Boltzmann's constant, T is the temperature, and ρ is the number density). This relationship has been found previously for the melting line, but the two constants have opposite signs for the sublimation and melting lines. This simple formula is also predicted using a nonharmonic static lattice expression for the pressure. The probability distribution function of the melting factor indicates departures from harmonic or Gaussian behavior in the lower wing. Nevertheless, the mean melting factor is shown to follow a simple MSD Debye-Waller factor dependence along both the melting and sublimation lines. This work combining HCM and MD simulations provides a comparison of the melting and sublimation lines of the LJ system, which could provide the foundations for a more unified statistical mechanical description of these two solid boundary lines.

Shadow Hamiltonian in classical NVE molecular dynamics simulations involving Coulomb interactions.

Microcanonical ensemble (NVE) Molecular Dynamics (MD) computer simulations are performed with negligible energy drift for systems incorporating Coulomb interactions and complex constraint schemes. In principle, such systems can now be simulated in the NVE ensemble for millisecond time scales, with no requirement for system thermostatting. Numerical tools for assessing drift in MD simulations are outlined, and drift rates of 10-6 K/µs are demonstrated for molten salts, polar liquids, and room temperature ionic liquids. Such drift rates are six orders of magnitude smaller than those typically quoted in the literature. To achieve this, the standard Ewald method is slightly modified so the first four derivatives of the real space terms go smoothly to zero at the truncation distance, rc. New methods for determining standard Ewald errors and the new perturbation errors introduced by the smoothing procedure are developed and applied, these taking charge correlation effects explicitly into account. The shadow Hamiltonian, Es, is shown to be the strictly conserved quantity in these systems, and standard errors in the mean of one part in 1010 are routinely calculated. Expressions for the shadow Hamiltonian are improved over previous work by accounting for O(h4) terms, where h is the MD time step. These improvements are demonstrated by means of extreme out-of-equilibrium simulations. Using the new methodology, the very low diffusion coefficients of room temperature 1-hexyl-3-methyl-imidazolium chloride are determined from long NVE trajectories in which the equations of motion are known to be integrated correctly, with negligible drift.

Viscuit and the fluctuation theorem investigation of shear viscosity by molecular dynamics simulations: The information and the noise.

The shear viscosity, η, of model liquids and solids is investigated within the framework of the viscuit and Fluctuation Theorem (FT) probability distribution function (PDF) theories, following Heyes et al. [J. Chem. Phys. 152, 194504 (2020)] using equilibrium molecular dynamics (MD) simulations on Lennard-Jones and Weeks-Chandler-Andersen model systems. The viscosity can be obtained in equilibrium MD simulation from the first moment of the viscuit PDF, which is shown for finite simulation lengths to give a less noisy plateau region than the Green-Kubo method. Two other formulas for the shear viscosity in terms of the viscuit and PDF analysis are also derived. A separation of the time-dependent average negative and positive viscuits extrapolated from the noise dominated region to zero time provides another route to η. The third method involves the relative number of positive and negative viscuits and their PDF standard deviations on the two sides for an equilibrium system. For the FT and finite shear rates, accurate analytic expressions for the relative number of positive to negative block average shear stresses is derived assuming a shifted Gaussian PDF, which is shown to agree well with non-equilibrium molecular dynamics simulations. A similar treatment of the positive and negative block average contributions to the viscosity is also shown to match the simulation data very well.

Reentrant melting and multiple occupancy crystals of bounded potentials: Simple theory and direct observation by molecular dynamics simulations.

Aspects of the phase coexistence behavior of the generalized exponential model (GEM-m) and bounded versions of inverse power potentials based on theory and molecular dynamics (MD) simulation data are reported. The GEM-m potential is ϕ(r)=exp(-r^{m}), where r is the pair separation and m is an adjustable exponent. A simple analytic formula for the fluid-solid envelope of the Gaussian core model which takes account of the known low- and high-density limiting forms is proposed and shown to represent the simulation data well. The bounded inverse power (BIP) potential is ϕ(r)=1/(a^{q}+r^{q})^{n/q}, where a, n, and q are positive constants. The BIP potential multiple occupancy crystal or cluster crystals are predicted to form when q>2 and a>0, for n>3, which compares with the corresponding GEM-m condition of m>2. Reentrant melting should occur for the BIP potential when q≤2 and a>0. MD simulations in which the system was gradually compressed at constant temperature using the BIP potential produced cluster states in the parameter domain expected but it was not possible to establish conclusively whether a multiply occupied crystal or a cluster fluid had formed owing to assembly structural fluctuations. The random phase approximation reproduces very well the BIP MD energy per particle without any discontinuities at the phase boundaries. The Lindemann melting rule is shown analytically to give a more rapidly decaying reentrant melting curve boundary than the so-called melting indicator (MI) empirical melting criterion which has also been investigated in this study. The MI model gives a better match to the high-density phase boundary for small m and q values.

Shadow Hamiltonian in classical NVE molecular dynamics simulations: A path to long time stability.

The shadow energy, Es, is the conserved quantity in microcanonical ensemble (NVE) molecular dynamics simulations carried out with the position Verlet central-difference algorithm. A new methodology for calculating precise and accurate values of Es is presented. It is shown for the first time that Es rather than E is constant during structural changes occurring within a supercooled liquid. It is also explained how to prepare and conduct microsecond range bulk-phase NVE simulations with essentially zero energy drift without the need for thermostating. The drift is analyzed with block averaging and new drift functions of the shadow energy. With such minimal drift, extremely small and accurate standard errors in the mean for quantities like Es, E, and temperature, T, can be obtained. Values of the standard error for Es of ≈10-10 in molecule-based reduced units can be routinely achieved for simulations of 108 time steps. This corresponds to a simulation temperature drift of ≈10-6 K/µs, six orders of magnitude smaller than generally considered to be acceptable for protein simulations. We also show for the first time how these treatments can be extended with no loss of accuracy to polyatomic systems with both flexible degrees of freedom and arbitrary geometric constraints imposed via the SHAKE algorithm. As a bonus, estimates of simulation-average kinetic and total energies from high order velocity expressions can be obtained to a good approximation from 2nd order velocities and the average mean square force (for polyatomics, this refers to per site, including any constraint forces).

Single trajectory transport coefficients and the energy landscape by molecular dynamics simulations.

The Green-Kubo (GK) method is widely used to calculate the transport coefficients of model liquids by Molecular Dynamics (MD) simulation. A reformulation of GK was proposed by Heyes et al. [J. Chem. Phys. 150, 174504 (2019)], which expressed the shear viscosity in terms of a probability distribution function (PDF) of "single trajectory (ST) viscosities," called "viscuits." This approach is extended here to the bulk viscosity, thermal conductivity, and diffusion coefficient. The PDFs of the four STs expressed in terms of their standard deviations (calculated separately for the positive and negative sides) are shown by MD to be statistically the same for the Lennard-Jones fluid. This PDF can be represented well by a sum of exponentials and is independent of system size and state point in the equilibrium fluid regime. The PDF is not well reproduced by a stochastic model. The PDF is statistically the same as that derived from the potential energy, u, and other thermodynamic quantities, indicating that the transport coefficients are determined quantitatively by and follow closely the time evolution of the underlying energy landscape. The PDFs of out-of-equilibrium supercooled high density states are quite different from those of the equilibrium states.

Transport coefficients of the Lennard-Jones fluid close to the freezing line.

Molecular dynamics simulations have been carried out along four Lennard-Jones (LJ) fluid isomorphs close to the freezing line, covering a temperature, T, in the range of 0.8-350 and a number density, ρ, in the range of 1.1-3.0 in LJ units. Analysis of the transport coefficients is via the Green-Kubo time correlation function method. The radial distribution function, percolation threshold connectivity distance, self-diffusion coefficient, and shear viscosity are shown to be invariant along an isomorph to a very good approximation when scaled with Rosenfeld's macroscopic units, although there are some small departures for T ≃ 1 and lower temperatures. The thermal conductivity is shown for the first time also to be isomorph invariant. In contrast, the Einstein and moment-based frequencies, and especially the bulk viscosity, ηb, show poor isomorphic collapse at low T but not surprisingly tend to an "inverse power" potential limiting value in the high T limit. In the case of the bulk viscosity, the significant departures from invariance arise from oscillations in the pressure autocorrelation function at intermediate times, which scale for inverse power potential systems but not for the LJ case, at least in part, as the pressure and bulk elastic moduli are not isomorph invariant.

Shear stress relaxation and diffusion in simple liquids by molecular dynamics simulations: Analytic expressions and paths to viscosity.

The results are reported of an equilibrium molecular dynamics simulation study of the shear viscosity, η, and self-diffusion coefficient, D, of the Lennard-Jones liquid using the Green-Kubo (GK) method. Semiempirical analytic expressions for both GK time correlation functions were fitted to the simulation data and used to derive analytic expressions for the time dependent diffusion coefficient and shear viscosity, and also the correlation function frequency transforms. In the case of the shear viscosity for a state point near the triple point, a sech function was found to fit the correlation function significantly better than a gaussian in the ballistic short time regime. A reformulation of the shear GK formula in terms of a time series of time integrals ("viscuits") and contributions to the viscosity from components based on the initial stress ("visclets") enable the GK expressions to be recast in terms of probability distributions which could be used in coarse grained stochastic models of nanoscale flow. The visclet treatment shows that stress relaxation is statistically independent of the initial stress for equilibrium and metastable liquids, suggesting that shear stress relaxation in liquids is diffusion controlled. By contrast, the velocity autocorrelation function is sensitive to the initial velocity. Weak oscillations and a plateau at intermediate times originate to a greater extent from the high velocity tail of the Maxwell-Boltzmann velocity distribution. Simple approximate analytic expressions for the mean square displacement and the self Van Hove correlation function are also derived.

Translational and rotational diffusion of rod shaped molecules by molecular dynamics simulations.

The results of molecular dynamics simulations of the dynamical evolution of assemblies of linear rigid rods of variable aspect ratio, a, and number density, ρ, in the isotropic phase are reported. The rods consist of m equally spaced sites interacting with the Weeks-Chandler-Andersen repulsive pair potential, where 2 < m < 16. With increasing m, features specific to long rods, such as anisotropic self-diffusion, become apparent. There is also an increasing separation between the characteristic relaxation times of the torque, angular velocity, and reorientational time correlation functions with increasing density. The latter is exponential at high densities even for dimers. The isotropic translational diffusion coefficient, Di, and rotational diffusion coefficient, Dr, are reported as a function of m and ρ or volume fraction, ξ. The mDi data scale with ξ throughout much of the simulated range, while the rotational diffusion coefficients scale approximately as m3Dr against ρ at low densities but as ∼m6Dr at high ρ, consistent with theories of colloidal and noncolloidal rod-containing liquids. The crossover density between the two regimes is parameterized in analytic form. The probability distribution functions for displacements and angular jumps in a given time show evidence of non-Gaussian behavior with increasing density. The shear viscosity and Di scale approximately as m and m-1, respectively, in the semidilute regime, which is consistent with a Stokes-Einstein-like relationship. At high concentrations, a frustrated or glassy structure formed in which the rods were randomly oriented.

Bounded inverse power potentials: Isomorphism and isosbestic points.

The bounded inverse power (BIP) interaction pair potential, ϕ(r)=1/(aq+rq)n/q, where a and the exponent, n, are constants which control the interaction softness, q is a positive integer, and r is the pair separation, is shown to exhibit isomorphic scaling as does the well-known inverse power potential, i.e., where a = 0. If T is the temperature and ρ is the number density of particles, two state points are isomorphic if a reference state, ρ0, T0, a0 and another state, ρ, T, a are related through the relationships ρn/3/T=ρ0 n/3/T0 and a=a0ρ0/ρ1/3=a0T0/T1/n. The potential form is therefore density dependent along an isomorph. Molecular dynamics simulations and solutions of the Ornstein-Zernike integral equation for q = 2 demonstrate the existence of isosbestic points (IBPs) in the radial distribution function and structure factor for 6 ≤ n ≤ 18 and a wide range of a and ρ values. For the BIP potentials with not too small a values and over a wide density range, the IBP distance is insensitive to the number density and is equal to the distance, rT, defined through ϕ(rT) = T. For exponential potentials of the general form, ϕ(r) = C exp(-rm) with 1 ≤ m ≤ 3, there are also IBPs which are at r values that are typically ∼10-15% larger than predicted by the formula for rT.

Incremental viscosity by non-equilibrium molecular dynamics and the Eyring model.

The viscoelastic behavior of sheared fluids is calculated by Non-Equilibrium Molecular Dynamics (NEMD) simulation, and complementary analytic solutions of a time-dependent extension of Eyring's model (EM) for shear thinning are derived. It is argued that an "incremental viscosity," η i , or IV which is the derivative of the steady state stress with respect to the shear rate is a better measure of the physical state of the system than the conventional definition of the shear rate dependent viscosity (i.e., the shear stress divided by the strain rate). The stress relaxation function, C i (t), associated with η i is consistent with Boltzmann's superposition principle and is computed by NEMD and the EM. The IV of the Eyring model is shown to be a special case of the Carreau formula for shear thinning. An analytic solution for the transient time correlation function for the EM is derived. An extension of the EM to allow for significant local shear stress fluctuations on a molecular level, represented by a gaussian distribution, is shown to have the same analytic form as the original EM but with the EM stress replaced by its time and spatial average. Even at high shear rates and on small scales, the probability distribution function is almost gaussian (apart from in the wings) with the peak shifted by the shear. The Eyring formula approximately satisfies the Fluctuation Theorem, which may in part explain its success in representing the shear thinning curves of a wide range of different types of chemical systems.

Thermodynamic curvature of soft-sphere fluids and solids.

The influence of the strength of repulsion between particles on the thermodynamic curvature scalar R for the fluid and solid states is investigated for particles interacting with the inverse power (r^{-n}) potential, where r is the pair separation and 1/n is the softness. Exact results are obtained for R in certain limiting cases, and the R behavior determined for the systems in the fluid and solid phases. It is found that in such systems the thermodynamic curvature can be positive for very soft particles, negative for steeply repulsive (or large n) particles across almost the entire density range, and can change sign between negative and positive at a certain density. The relationship between R and the form of the interaction potential is more complex than previously suggested, and it may be that R is an indicator of the relative importance of energy and entropy contributions to the thermodynamic properties of the system.