Parag Adhangale

A generalized Hamiltonian-based algorithm for rigorous equilibrium molecular dynamics simulation in the isobaric–isothermal ensemble

In this work, we employ a Hamiltonian-based procedure to derive a generalized Nosé barostat, which generates trajectories that rigorously satisfy the statistical mechanical isobaric–isothermal (NpT) ensemble. This generalized algorithm, unlike Nosés original NpT algorithm, maintains rigor in the presence of (i) a non-zero system momentum and (ii) non-negligible external forces. The generalized algorithm reduces to the conventional Nosé NpT algorithm when neither condition is satisfied. The key element of the generalized algorithm is that the thermostat and barostat are applied only to those degrees of freedom that contribute to the temperature and pressure, which excludes, e.g. the total system momentum. We show that the generalized algorithm satisfies the two criteria for rigor (Hamiltonian and non-Hamiltonian) that exist in the literature. Finally, we provide some numerical examples demonstrating the success of the generalized algorithm.

Equation of state for a coarse-grained DPPC monolayer at the air/water interface

Pulmonary surfactant, a complex mixture of phospholipids and proteins, secreted by the type II epithelial cells in the lungs, is crucial to reducing the effort required for breathing. A lack of adequate amounts of pulmonary surfactant in underdeveloped lungs of premature infants results in infant respiratory distress syndrome (RDS). Surfactant replacement therapy (SRT) is the approved method of mitigating the effects of RDS. The development of new SRT regimens requires a fundamental understanding of the links between surfactant biochemistry and functionality. We use a coarse-grained (CG) model to predict the surface pressure–surface concentration relationship (equation of state) for pure DPPC, which is a major component of endogenous and synthetic pulmonary surfactant mixtures. We show that the model can be efficiently used to predict the equation of state in excellent agreement with experimental results. We also study the structure of the monolayer as a function of the surface tension of the system. We show that a decrease in the applied surface tension results in an increase in order in the head group region and a decrease in order in the tail region of DPPC. This technique may be useful in the prediction of equations of state for surfactant replacements.

Exploiting single-file motion in one-dimensional nanoporous materials for hydrocarbon separation

The mobility of fluids adsorbed in nanoporous materials is a strong function of the size, shape, and dimensionality of the porous network. Nowhere is this dependence demonstrated more drastically than by fluids adsorbed in one-dimensional (cylinder-like) nanopores. It has been demonstrated theoretically, computationally, and experimentally that the mobility of a fluid adsorbed in one-dimensional nanopores varies with adsorbate size not only quantitatively (over several orders of magnitude) but also qualitatively.1-4 When the pore size is small enough to prohibit passing of individual fluid molecules in the pore, the ordinary diffusion (where the mean square displacement is proportional to the observation time, and the proportionality constant is the diffusion coefficient) gives way to single-file motion (where the mean square displacement is proportional to the square root of the observation time, and the proportionality constant no longer has units of diffusivity). This difference in qualitative modes of motion results in a drastic quantitative difference in mobility; the single-file mode is much slower. Using molecular dynamics simulations of methane and ethane in the one-dimensional molecular sieve, AlPO4-5, this work demonstrates that the transition from ordinary diffusion to single-file motion can be exploited to effect a kinetic separation. In this case, methane molecules are small enough to pass each other in the pores of AlPO4-5. The ethane molecules are too large to pass and undergo single-file motion. When a mixture of these two fluids is adsorbed in AlPO4-5, the methane can still pass ethane and retains its fast, ordinary mode of diffusion. Thus, by careful selection of the adsorbent, we create an environment where these two fluids, with roughly the same bulk diffusivities, exhibit mobilities differing by several orders of magnitude. This transport phenomenon has no bulk analog; it is a novel characteristic of fluids confined in nanoscale channels.

Single-file Motion of Polyatomic Molecules in One-dimensional Nanoporous Materials

In a continuation of the study of the mobility of fluids adsorbed in nanoporous materials, molecular dynamics simulations are used to investigate the behaviour of polyatomic ethane molecules adsorbed in AlPO4-5. The current work is based on the use of the united atom approach as a better model than the single-centre ethane used to date. Ethane molecules are modelled as rigid diatoms, and as a result the molecules have more degrees of freedom in the form of the rotational components that are absent in the single-centre ethane model. This represents a more sophisticated model for ethane and is used in the simulations to test earlier findings. Simulations with binary mixtures of methane and ethane also have been conducted with three mixture compositions. The transition from ordinary diffusion to single-file motion for a finite residence time is found to occur at a methyl group diameter of 4.75 Å. This is identical to the ethane diameter in the earlier study. Thus, only the minimum dimension determines the transition size. Also it is shown that the diatomic molecules undergo free rotation within the channel even when they are in the single-file mode of motion. In the case of binary mixtures, the methane molecules still undergo ordinary diffusion. Ethane molecules exhibit single-file motion at a methyl group diameter of 4.75 Å. The diffusion coefficient of methane decreases with increasing ethane size, while the trends in the single-file mobility of ethane as a function of methyl group diameter are nonlinear.

Obtaining transport diffusion coefficients from self-diffusion coefficients in nanoporous adsorption systems

In a continuation of the study of the adsorption and transport properties in a nanoporous adsorption system, transport diffusion coefficients are predicted from the self-diffusion coefficients using the Darken equation, modified for use in the adsorbed phase. We obtain self-diffusion coefficients using equilibrium molecular dynamics (MD) simulations. Adsorption equilibrium data are required in this modified form, which are obtained using grand canonical Monte Carlo (GCMC) simulations. We show that principal component regression provides an elegant and robust method to integrate the data available from the MD and GCMC simulations for the prediction of transport diffusivities. We investigate the effect of the adsorbed phase concentration, mole fraction and temperature on the transport diffusion coefficient thus predicted. We show that the self- and the transport diffusion coefficients decrease with increasing adsorbed phase concentration. We also show that the transport diffusion coefficients decrease with increasing methane mole fraction, in contrast to the behaviour of the self-diffusion coefficients.