Bulletin of the American Physical Society
APS March Meeting 2018
Volume 63, Number 1
Monday–Friday, March 5–9, 2018; Los Angeles, California
Session B38: Advances in Computational Statistical Mechanics and their Applications: Part 2Focus
|
Hide Abstracts |
Sponsoring Units: DCOMP DCMP GSNP Chair: Ying Wai Li, Oak Ridge National Lab Room: LACC 501A |
Monday, March 5, 2018 11:15AM - 11:51AM |
B38.00001: Nested sampling for computational thermodynamics Invited Speaker: Livia Bartok-Partay In recent years we have been working on adapting a novel computational sampling technique, called nested sampling, to study the potential energy surface of atomistic systems. Nested sampling automatically generates all the relevant atomic configurations, unhindered by high barriers, and one of its most appealing advantages is that the global partition function can be calculated very easily as a simple post-processing step, thus thermodynamic properties become accessible. Nested sampling explores the potential energy surface starting from the high energy region, hence no prior knowledge of the potentially stable structures is needed. This means that unlike other methods, nested sampling may be fully automated, allowing high-throughput calculations of phase transformations and phase diagrams of different materials. |
Monday, March 5, 2018 11:51AM - 12:03PM |
B38.00002: Phase diagrams of complex systems with nested sampling Noam Bernstein, K. Michael Salerno, Gabor Csanyi The nested sampling method provides an efficient way of computing the classical partition function, and from it thermodynamic quantities and other observables as a function of temperature and pressure. It can therefore be used to calculate temperature dependent properties and phase diagrams. We show how nested sampling can be applied to two systems with chemical and structural complexity. The first is polymers, which have both short range (bond length, angle) and long range (conformation, dense packing in the melt/solid) structure, which we model with chemicaly specific coarse-grained potentials. We show how to explore configuration space given the form of typical coarse-grained models, and present results for their phase diagrams. The second is multicomponent systems, which can exhibit phase separation in some parts of the phase diagram, e.g. a two-phase coexistence region delimited by a liquidus at high T and solidus at low T for the liquid-solid transition in a binary alloy. We discuss several approaches to handling this tendency to phase separation, and demonstrate that we can efficiently compute phase diagrams for multicomponent systems. |
Monday, March 5, 2018 12:03PM - 12:15PM |
B38.00003: Monte Carlo Simulations of Coarse-grained Protein Models for Crambin Alfred Farris, Ying Wai Li, Daniel Seaton, David Landau In order to study interactions of primary importance in protein folding, we apply Wang-Landau sampling [1] and a novel histogram-free multicanonical sampling algorithm [2] to the hydrophobic-polar (HP) [3] and H0P [4] lattice protein models, as well as their continuum versions inspired by the AB model [5]. These models offer significant computational advantage over all-atom models, due to simplified interactions and the reduction of the 20 amino acids to a limited number of types. In this work, we compare and contrast folding behavior between such models for Crambin - a 46 amino acid plant protein, and for a homopolymer of the same length. Our results highlight the strengths of the two simulational frameworks and the differences in folding between models. |
Monday, March 5, 2018 12:15PM - 12:27PM |
B38.00004: Amyloid Fibril Formation Studied by Replica-Exchange Wang-Landau Simulations Matthew Wilson, Guangjie Shi, David Landau, Thomas Wuest, Friederike Schmid As neurological diseases associated with toxic peptide aggregates emerge, the research of aggregate and fibril formation has become a prominent, yet extremely challenging problem in the biological and physical sciences. The ability to form an amyloid state has been posited as a general feature of peptide systems, which is utilized in this study using a collection of coarse-grained, generic model peptides. The H0P model1 adds an additional neutral polarity group to the classic hydrophobic-polar (HP) model2 , and is used for simplicity and efficiency. With the parallelized framework of the replica-exchange Wang-Landau (REWL)3 algorithm, we determined the density of states for all possible energies of multiple interacting model peptides. Average thermodynamic quantities are studied as the system transitions between states as dissolved peptides, disordered oligomers, and crystalline aggregate structures. Additional structural observables are calculated in a post-simulation production run, and are used to further elucidate the physical behavior during the observed transitions. |
Monday, March 5, 2018 12:27PM - 12:39PM |
B38.00005: A Multicanonical Monte Carlo Ensemble Growth method Graziano Vernizzi, Trung Nguyen, Henri Orland, Monica Olvera De La Cruz In this work, we extend the chain-growth algorithm by T. Garel and H. Orland (J. Phys A, 23.12, L621, 1990) to the multicanonical ensemble. The method belongs to the general class of Population Monte Carlo algorithms, were multiple copies of a statistical system are considered in parallel. Such a stochastic sampling differs from more traditional approaches where one copy of the statistical system is considered at a time. This method produces the density of states of a statistical system, which can be used to produce canonical ensemble distributions and averages by standard re-weighting techniques. It is complementary to powerful and popular Monte Carlo growth methods such as the pruned-enriched Rosenbluth method (PERM), or its multicanonical extension (MuCa PERM), or its flat histogram version (FlatPERM). We discuss its implementation on simple statistical systems, such as the single-chain or multiple-chain growth problems, and its application to the case of polymers adsorbed onto the surface of a protein. |
Monday, March 5, 2018 12:39PM - 12:51PM |
B38.00006: Thermodynamically stable phases for semiflexible polymers characterized by knots of specific topologies Wolfhard Janke, Martiin Marenz We investigate the influence of bending stiffness on the conformational phases of a semiflexible bead-stick homopolymer and present the pseudo-phase diagram for the complete range of semiflexibility, from flexible to stiff. By varying the bending stiffness, the model exhibits different pseudo phases governed by bent, hairpin or toroidal conformations. In particular, we find thermodynamically stable phases characterized by knots of specific topologies. The transitions into these ``knotted'' phases from other ordered phases are quite unusual in that they display clear phase coexistence but almost no change in the mean total energy and hence no latent heat. It will be explained how we arrive at these intriguing results by computer simulations based on a combination of the replica-exchange Monte Carlo algorithm and the multicanonical method and discussed how one can understand these effects by basic statistical physics reasoning. |
Monday, March 5, 2018 12:51PM - 1:03PM |
B38.00007: A Systematic Coarse-graining Approach for High-molecular Weight Polymers Michael Webb, Jean-Yves Delannoy, Juan De Pablo A major challenge in the study of both natural and synthetic polymer systems is the theory-based design of polymers with desired properties. While computational statistical mechanics provides a framework to establish structure-property relationships that facilitate design, many important phenomena require the study of time- and length scales that are beyond the capabilities of atomistic molecular dynamics and thus require different simulation strategies. In this talk, we discuss a systematic coarse-graining approach that enables study of systems at targeted resolution. The approach, which utilizes basic concepts from graph theory, provides an unambiguous and automated way to generate coarse-grained representations that preserve chemical topology and are thus sensitive to chemical substitutions. We highlight this approach in application to several high-molecular weight polysaccharides and examinethe properties for several polysaccharides at different degrees of resolution. We further assess the performance of both traditional and machine-learning parameterization schemes. The methodology described provides a framework that should be useful for coarse-grained study of various soft matter systems. |
Monday, March 5, 2018 1:03PM - 1:15PM |
B38.00008: First Principle Free Energy Calculations Made Simple: The Example Case of Alanine Dipeptide, from Classical Force-field to Hybrid Functional. Emre Sevgen, Federico Giberti, Hythem Sidky, Jonathan Whitmer, Giulia Galli, Francois Gygi, Juan De Pablo Ab initio molecular dynamics (AIMD) is a powerful method to study the microscopic properties of broad classes of materials, but its predictive power is limited by the short timescales that can be investigated. Here, we present a scheme coupling SSAGES1, a plug-in for enhanced sampling calculations, and Qbox2, a plane-wave DFT MD code, that allows for the efficient use of several enhanced sampling techniques in AIMD. We demonstrate the power of the framework by computing the free energy surface (FES) of alanine-dipeptide in vacuum with Adaptive-Bias-Force using classical potentials and two density functionals: a semi-local (PBE) and a hybrid functional (PBE0). We find substantial differences between the results of DFT and those of the classical potentials: the latter accurately reproduce the energy minima of the system, however, they describe inaccurately high free energy states. The coupling scheme introduced here is general and allows for simple and straightforward calculations of FES from first principles. |
Monday, March 5, 2018 1:15PM - 1:27PM |
B38.00009: A Rigorous and Self-contained Algorithm for Evaluating Free Energy of Liquid and Solid Phases of an Alloy System Lin Yang, Feng Zhang, Yang Sun, Mikhail Mendelev, Cai-Zhuang Wang, Kai-Ming Ho In this work, we describe a self-contained procedure to evaluate the free energy of liquid and solid phases of an alloy system. The free energy of a pure solid phase is calculated with thermodynamic integration using the Einstein crystal as the reference system. The free energy of pure liquid can be obtained by integrating ΔH/T2 with temperature, where ΔH is the latent heat during melting. The central part of our method is the construction of a reversible path from pure liquid to liquid alloy to calculate the mixing enthalpy and entropy. We have applied the method to calculate the free energy of solid and liquid phases in the Al-Sm system. The driving force for fcc-Al nucleation in Al-Sm liquids and the melting curve for fcc-Al and Al3Sm are also calculated. |
Monday, March 5, 2018 1:27PM - 1:39PM |
B38.00010: First-principles study of multicomponent solid-solution alloys using Wang-Landau Monte Carlo method Zongrui Pei, Markus Eisenbach, Ying Wai Li, G. Malcolm Stocks Ordering transitions have an important influence on the physical properties of multicomponent alloys. Here, we present results of first-principle studies of ordering transitions and short-range order parameters in multicomponent concentrated solid-solution alloys using the Wang-Landau Monte Carlo method. As a first step, we tested the performance (accuracy and efficiency) of the combined Cluster-Expansion and Wang-Landau Monte Carlo (WL-CE) method. Secondly, the WL-CE method was used to calculate the transition temperature and short-range order parameters as a function of temperature. Finally, the possible correlations between the finite-temperature short-range order parameters and physical properties are explored. |
Monday, March 5, 2018 1:39PM - 1:51PM |
B38.00011: Feasibility of Direct Use of Ab Initio Energies in Replica Exchange Monte Carlo Simulation of Ion Disorder in Solids Shusuke Kasamatsu, Osamu Sugino The Metropolis Monte Carlo (MMC) algorithm has seen much application in the simulation of order-disorder transitions and prediction of phase diagrams in systems such as alloys and disordered oxides. Usually, ab initio total energy methods are deemed too expensive to provide sufficient configuration sampling, and the usual approach is to use lightweight models derived via, e.g., cluster expansion fitting of ab initio energies. However, deriving quantitative models can be quite challenging for systems with complex long-range interactions such as multivalent ion oxides and heterointerfaces [1]. Here, we reexamine the feasibility of directly calculating ab initio energies at every MMC trial step. We also consider the replica exchange method [2] to speed up the equilibration and sampling. The feasibility of this idea, especially using modern-day supercomputers, is demonstrated on the calculation of temperature-dependent cation disorder in MgAl2O4 spinel. [1] Seko and Tanaka, J. Phys. Condens. Matter 26, 115403 (2014). [2] Hukushima and Nemoto, J. Phys. Soc. Jpn. 65, 1604 (1996). |
Monday, March 5, 2018 1:51PM - 2:03PM |
B38.00012: Heat Conductivity in Amorphous Solids from Equilibrium ab initio Molecular Dynamics Loris Ercole, Aris Marcolongo, Stefano Baroni Until very recently, the Green-Kubo (GK) theory of linear response was not deemed compatible with quantum simulation techniques based on density functional theory because the concepts of energy density and current are not well defined at the atomic scale, and because the study of transport coefficients using the GK theory is known to require so long molecular dynamics simulations as to make ab initio techniques unaffordable. Recently, the first of these hurdles was overcome thanks to a paradigm shift based on the concept of gauge invariance of transport coefficients, [1, 2] while the latter was crossed using a novel data analysis technique based on the so-called cepstral analysis of stationary time series. [3] |
Monday, March 5, 2018 2:03PM - 2:15PM |
B38.00013: Efficient and Accurate Modeling of Quantum Nuclear Effects in Molecules and Materials Igor Poltavskyi, Alexandre Tkatchenko Nuclear quantum effects (NQE), which includes both zero-point motion and tunneling, exhibits an important influence on equilibrium and dynamical properties of molecules and materials. These effects can be taken into account using the Feynman-Kac imaginary-time path integral molecular dynamics (PIMD) calculations. Efficient PIMD schemes require the knowledge about high-order derivatives of the potential energy, which limits their practical applicability in ab initio simulations. Recently it was shown [V. Kapil et al., J. Chem. Phys. (2016)] that the finite differences method can be successfully employed to compute these derivatives, considerably decreasing the computational costs of ab initio PIMD simulations. Here we demonstrate how the efficiency of such simulations can be further improved by combining PIMD approach and thermodynamic perturbation theory [I. Poltavsky and A. Tkatchenko, Chem. Sci. (2016)]. The developed method allows the calculation of NQE in realistic molecules and materials, and paves the way to converged PIMD simulations even at relatively low temperature. |
Follow Us |
Engage
Become an APS Member |
My APS
Renew Membership |
Information for |
About APSThe American Physical Society (APS) is a non-profit membership organization working to advance the knowledge of physics. |
© 2024 American Physical Society
| All rights reserved | Terms of Use
| Contact Us
Headquarters
1 Physics Ellipse, College Park, MD 20740-3844
(301) 209-3200
Editorial Office
100 Motor Pkwy, Suite 110, Hauppauge, NY 11788
(631) 591-4000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 20045-2001
(202) 662-8700