Search

Advanced potential energy surfaces are defined as theoretical models that explicitly include many-body effects that transcend the standard fixed-charge, pairwise-additive paradigm typically used in molecular simulation. However, several factors relating to their software implementation have precluded their widespread use in condensed-phase simulations: the computational cost of the theoretical models, a paucity of approximate models and algorithmic improvements that can ameliorate their cost, underdeveloped interfaces and limited dissemination in computational code bases that are widely used in the computational chemistry community, and software implementations that have not kept pace with modern high-performance computing (HPC) architectures, such as multicore CPUs and modern graphics processing units (GPUs). In this Feature Article we review recent progress made in these areas, including well-defined polarization approximations and new multipole electrostatic formulations, novel methods for solving the mutual polarization equations and increasing the MD time step, combining linear-scaling electronic structure methods with new QM/MM methods that account for mutual polarization between the two regions, and the greatly improved software deployment of these models and methods onto GPU and CPU hardware platforms. We have now approached an era where multipole-based polarizable force fields can be routinely used to obtain computational results comparable to state-of-the-art density functional theory while reaching sampling statistics that are acceptable when compared to that obtained from simpler fixed partial charge force fields.

Trifluoromethanesulfonic acid hydrates provide a well-defined system to study proton dissociation and transport in perfluorosulfonic acid membranes, typically used as the electrolyte in hydrogen fuel cells, in the limit of minimal water. The triflic acid pentahydrate crystal (CF(3)SO(3)H·5H(2)O) is sufficiently aqueous that it contains an extended three-dimensional water network. Despite it being extended, however, long-range proton transport along the network is structurally unfavorable and would require considerable rearrangement. Nevertheless, the triflic acid pentahydrate crystal system can provide a clear picture of the preferred locations of local protonic defects in the water network, which provides insights about related structures in the disordered, low-hydration environment of perfluorosulfonic acid membranes. Ab initio molecular dynamics simulations reveal that the proton defect is most likely to transfer to the closest water that has the expected presolvation and only contains water in its first solvation shell. Unlike the tetrahydrate of triflic acid (CF(3)SO(3)H·4H(2)O), there is no evidence of the proton preferentially transferring to a water molecule bridging two of the sulfonate groups. However, this could be an artifact of the crystal structure since the only such water molecule is separated from the proton by long O-O distances. Hydrogen bonding criteria, using the two-dimensional potential of mean force, are extracted. Radial distribution functions, free energy profiles, radii of gyration, and the root-mean-square displacement computed from ab initio path integral molecular dynamics simulations reveal that quantum effects do significantly extend the size of the protonic defect and increase the frequency of proton transfer events by nearly 15%. The calculated IR spectra confirm that the dominant protonic defect mostly exists as an Eigen cation but contains some Zundel ion characteristics. Chain lengths and ring sizes determined from the hydrogen bond network, counted using graph theory techniques, are only moderately sensitive to quantum effects. Deliberately introducing a structural defect into the native crystal yields a protonic defect with one hydrogen bond to a sulfonate group that was found to be metastable for at least 10 ps., (© 2011 American Chemical Society)

Intermolecular enol tautomers of Watson-Crick base pairs could emerge spontaneously via interbase double proton transfer. It has been hypothesized that their formation could be facilitated by thermal fluctuations and proton tunneling, and possibly be relevant to DNA damage. Theoretical and computational studies, assuming classical nuclei, have confirmed the dynamic stability of these rare tautomers. However, by accounting for nuclear quantum effects explicitly through Car-Parrinello path integral molecular dynamics calculations, we find the tautomeric enol form to be dynamically metastable, with lifetimes too insignificant to be implicated in DNA damage.

Higher order factorization schemes are developed for path integral molecular dynamics in order to improve the convergence of estimators for physical observables as a function of the Trotter number. The methods are based on the Takahashi-Imada and Susuki decompositions of the Boltzmann operator. The methods introduced improve the averages of the estimators by using the classical forces needed to carry out the dynamics to construct a posteriori weighting factors for standard path integral molecular dynamics. The new approaches are straightforward to implement in existing path integral codes and carry no significant overhead. The Suzuki higher order factorization was also used to improve the end-to-end distance estimator in open path integral molecular dynamics. The new schemes are tested in various model systems, including an ab initio path integral molecular dynamics calculation on the hydrogen molecule and a quantum water model. The proposed algorithms have potential utility for reducing the cost of path integral molecular dynamics calculations of bulk systems. [ABSTRACT FROM AUTHOR]

The problems of ergodicity and internal consistency in the centroid and ring-polymer molecular dynamics methods are addressed in the context of a comparative study of the two methods. Enhanced sampling in ring-polymer molecular dynamics (RPMD) is achieved by first performing an equilibrium path integral calculation and then launching RPMD trajectories from selected, stochastically independent equilibrium configurations. It is shown that this approach converges more rapidly than periodic resampling of velocities from a single long RPMD run. Dynamical quantities obtained from RPMD and centroid molecular dynamics (CMD) are compared to exact results for a variety of model systems. Fully converged results for correlations functions are presented for several one dimensional systems and para-hydrogen near its triple point using an improved sampling technique. Our results indicate that CMD shows very similar performance to RPMD. The quality of each method is further assessed via a new χ2 descriptor constructed by transforming approximate real-time correlation functions from CMD and RPMD trajectories to imaginary time and comparing these to numerically exact imaginary time correlation functions. For para-hydrogen near its triple point, it is found that adiabatic CMD and RPMD both have similar χ2 error. [ABSTRACT FROM AUTHOR]

A frequently encountered problem in molecular dynamics is how to treat the long times that are required to simulate condensed systems consisting of particles interacting through long range forces. Standard methods require the calculation of the forces at every time step. Because each particle interacts with all particles within the interaction range of the potential the longer the range of the potential the larger the number forces that must be calculated at each time step. In this note we present a variant of the RESPA (reference system propagator algorithm), which we developed for handling systems with multiple time scales like disparate mass mixtures. This version of RESPA greatly reduces the number of forces that must be computed at each time step and thereby leads to a dramatic acceleration of such simulations. The RESPA method uses ideas similar to NAPA, an algorithm we invented to treat high frequency oscillators interacting with low frequency bath. The method is based on a choice of a reference system in which the particles interact through short range forces. The reference system is numerically integrated for n time steps δt and the error incurred by using short range forces is corrected by solving a rigorous set of equations once every Δt=nδt. This method reduces the cpu time dramatically. It is shown that this approach and suitable generalizations should be very useful for future simulations of quantum and classical condensed matter systems. [ABSTRACT FROM AUTHOR]

A frequently encountered problem in molecular dynamics simulations is the long runs required to study condensed systems consisting of both high frequency and low frequency degrees of freedom. Standard integrators require the choice of time step sufficiently small to guarantee stable solution of the highest frequency motion with the consequence that simulations require a very large number of central processing unit (CPU) cycles. In this note we present a new integrator that allows one to use a time step appropriate for the low frequency degrees of freedom without making any approximations related to the separation of time scales. This method is based on a choice of an analytically solvable reference system for the high frequency motion. We show how the analytical solution can be incorporated into a numerical integrator. The method is applied to two cases which are paradigms for this problem and it is shown that this approach and suitable generalizations should be very useful for future simulations of quantum and classical condensed matter systems. [ABSTRACT FROM AUTHOR]

The article examines the process of characterizing the catalic mechanism of histone deacetylases by usingdensity functional theory quantum mechanical and molecular mechanical (QM/MM) studies by histone-deacetylase-like protein. The calculations are based on the pseudobond ab initio QM/MM approach. The quantum mechanical subsystem is treated by the hybrid B3LYP functional with a Stuttgart ECP/basis set.