Manuel Athènes

Free energy calculations from adaptive molecular dynamics simulations with adiabatic reweighting

We propose an adiabatic reweighting algorithm for computing the free energy along an external parameter from adaptive molecular dynamics simulations. The adaptive bias is estimated using Bayes identity and information from all the sampled configurations. We apply the algorithm to a structural transition in a cluster and to the migration of a crystalline defect along a reaction coordinate. Compared to standard adaptive molecular dynamics, we observe an acceleration of convergence. With the aid of the algorithm, it is also possible to iteratively construct the free energy along the reaction coordinate without having to differentiate the gradient of the reaction coordinate or any biasing potential.

Measurement of nonequilibrium entropy from space-time thermodynamic integration

The entropy of a system transiently driven out of equilibrium by a time-inhomogeneous stochastic dynamics is first expressed as a transient response function generalizing the nonlinear Kawasaki-Crooks response. This function is then reformulated into three statistical averages defined over ensembles of nonequilibrium trajectories. The first average corresponds to a space-time thermodynamic perturbation relation, while the two following ones correspond to space-time thermodynamic integration relations. Provided that trajectories are initiated starting from a distribution of states that is analytically known, the ensemble averages are computationally amenable to Markov chain Monte Carlo methods. The relevance of importance sampling in path ensembles is confirmed in practice by computing the nonequilibrium entropy of a driven toy system. We finally study a situation where the dynamics produces entropy. In this case, we observe that space-time thermodynamic integration still yields converged estimates, while space-time thermodynamic perturbation turns out to converge very slowly.

Waste-recycling Monte Carlo with optimal estimates: Application to free energy calculations in alloys

The estimator proposed recently by Delmas and Jourdain for waste-recycling Monte Carlo achieves variance reduction optimally with respect to a control variate that is evaluated directly using the simulation data. Here, the performance of this estimator is assessed numerically for free energy calculations in generic binary alloys and is compared to those of other estimators taken from the literature. A systematic investigation with varying simulation parameters of a simplified system, the anti-ferromagnetic Ising model, is first carried out in the transmutation ensemble using path-sampling. We observe numerically that (i) the variance of the Delmas-Jourdain estimator is indeed reduced compared to that of other estimators; and that (ii) the resulting reduction is close to the maximal possible one, despite the inaccuracy in the estimated control variate. More extensive path-sampling simulations involving an FeCr alloy system described by a many-body potential additionally show that (iii) gradual transmutations accommodate the atomic frustrations; thus, alleviating the numerical ergodicity issue present in numerous alloy systems and eventually enabling the determination of phase coexistence conditions.

Estimating time-correlation functions by sampling and unbiasing dynamically activated events

Transition path sampling is a rare-event method that estimates state-to-state time-correlation functions in many-body systems from samples of short trajectories. In this framework, it is proposed to bias the importance function using the lowest Jacobian eigenvalue moduli along the dynamical trajectory. A lowest eigenvalue modulus is related to the lowest eigenvalue of the Hessian matrix and is evaluated here using the Lanczos algorithm as in activation-relaxation techniques. This results in favoring the sampling of activated trajectories and enhancing the occurrence of the rare reactive trajectories of interest, those corresponding to transitions between locally stable states. Estimating the time-correlation functions involves unbiasing the sample of simulated trajectories which is done using the multi-state Bennett acceptance ratio (MBAR) method. To assess the performance of our procedure, we compute the time-correlation function associated with the migration of a vacancy in α-iron. The derivative of the estimated time-correlation function yields a migration rate in agreement with the one given by transition state theory. Besides, we show that the information relative to rejected trajectories can be recycled within MBAR, resulting in a substantial speed-up. Unlike original transition path-sampling, our approach does not require computing the reversible work to confine the trajectory endpoints to a reactive state.

Thermodynamic modelling of glasses at atomistic scale

Establishing the conditions of phase equilibria involves measuring the relative free energies of the various phases appearing in a given multi‐component system. Equilibrium free energy differences are traditionally computed by thermodynamic integration : one may integrate, for instance with respect to inverse temperature, the mean internal energy that has been previously estimated using a Monte Carlo sampling method. Such an approach can not be applied to a glassy system because it slowly relaxes with time. To palliate this difficulty, we focus on the glass thermodynamic properties that can be described by restricting the phase space to a relevant portion of the energy landscape. Assuming that the glassy system gets trapped into a single metabasin after it is driven out of equilibrium, the thermodynamical potential is shown to correspond to a Landau free energy that can be computed by implementing a path‐sampling Monte Carlo scheme. The present method is applied to the calculation of the Gibbs free energi...

Cluster dynamics modelling of materials: A new hybrid deterministic/stochastic coupling approach

Deterministic simulations of the rate equations governing cluster dynamics in materials are limited by the number of equations to integrate. Stochastic simulations are limited by the high frequency of certain events. We propose a coupling method combining deterministic and stochastic approaches. It allows handling different time scale phenomena for cluster dynamics. This method, based on a splitting of the dynamics, is generic and we highlight two different hybrid deterministic/stochastic methods. These coupling schemes are highly parallelizable and specifically designed to treat large size cluster problems. The proof of concept is made on a simple model of vacancy clustering under thermal ageing.

Using Bayes formula to estimate rates of rare events in transition path sampling simulations

Transition path sampling is a method for estimating the rates of rare events in molecular systems based on the gradual transformation of a path distribution containing a small fraction of reactive trajectories into a biased distribution in which these rare trajectories have become frequent. Then, a multistate reweighting scheme is implemented to postprocess data collected from the staged simulations. Herein, we show how Bayes formula allows to directly construct a biased sample containing an enhanced fraction of reactive trajectories and to concomitantly estimate the transition rate from this sample. The approach can remediate the convergence issues encountered in free energy perturbation or umbrella sampling simulations when the transformed distribution insufficiently overlaps with the reference distribution.

Estimating thermodynamic expectations and free energies in expanded ensemble simulations: Systematic variance reduction through conditioning

Markov chain Monte Carlo methods are primarily used for sampling from a given probability distribution and estimating multi-dimensional integrals based on the information contained in the generated samples. Whenever it is possible, more accurate estimates are obtained by combining Monte Carlo integration and integration by numerical quadrature along particular coordinates. We show that this variance reduction technique, referred to as conditioning in probability theory, can be advantageously implemented in expanded ensemble simulations. These simulations aim at estimating thermodynamic expectations as a function of an external parameter that is sampled like an additional coordinate. Conditioning therein entails integrating along the external coordinate by numerical quadrature. We prove variance reduction with respect to alternative standard estimators and demonstrate the practical efficiency of the technique by estimating free energies and characterizing a structural phase transition between two solid phases.