Bulletin of the American Physical Society
APS March Meeting 2010
Volume 55, Number 2
Monday–Friday, March 15–19, 2010; Portland, Oregon
Session P23: Classical and Quantum Molecular Dynamics |
Hide Abstracts |
Sponsoring Units: DCOMP Chair: Glenn Martyna, IBM T.J. Watson Research Center Room: C125-C126 |
Wednesday, March 17, 2010 8:00AM - 8:12AM |
P23.00001: Accelerating Orbital-Free DFT Calculation with a heterogeneous multi-core system Masaru Aoki, Hidekazu Tomono, Kazuo Tsumuraya Computational material design requires efficient algorithms and high-speed computers to calculate and to predict material properties. The orbital-free first principles calculation (OF- FPC) method, which is a tool for calculating and designing the properties, is an $O(N)$ method and is a powerful tool to study large-scaled systems. We implement a CUFFT routine, which is an FFT library of CUDA (Compute Unified Device Architecture) for GPGPU (General-Purpose Graphics Processing Unit), into our in- house OF-FPC code. We evaluate the computation time to optimize the electron charge density of the sodium crystal systems containing 2, 16, 128, 1024, and 6750 atoms. The GPU-CPU system reduces the time to half of that of the CPU system for the system with 6750 atoms. The GPGPU is effective in accelerating the OF-FPC code. [Preview Abstract] |
Wednesday, March 17, 2010 8:12AM - 8:24AM |
P23.00002: Stochastic Quantum Molecular Dynamics Heiko Appel, Massimiliano Di Ventra Conventional molecular dynamics (MD) approaches like Car-Parinello (CPMD), Born-Oppenheimer (BOMD) or Ehrenfest MD have in common that the electronic degrees of freedom are always described in terms of a closed quantum system. By construction, CPMD and BOMD only refer to the ground-state Born-Oppenheimer surface, whereas Ehrenfest MD can take into account electronic excitations. However, in all cases the description in terms of a closed electronic system excludes the possibility of electronic relaxation and decoherence. Based on stochastic Kohn-Sham equations, a new MD approach is presented that allows for an open quantum system description of both the electronic and the nuclear degrees of freedom. The new approach is illustrated with a MD study of vibrational and rotational relaxation in 4-(N,N-dimethylamino)benzonitrile (DMABN). Reference: cond-mat/09082411. [Preview Abstract] |
Wednesday, March 17, 2010 8:24AM - 8:36AM |
P23.00003: The Phonostat: Thermostating Phonons in Molecular Dynamics Simulations Rajamani Raghunathan, Alex P. Greaney, Jeffrey C. Grossman We present a new phonostat algorithm to regulate temperature within molecular dynamics (MD) simulations. Our technique couples vibrational degrees of freedom of a group of atoms to a thermal reservoir allowing us to regulate an athermal population of various vibrational modes. In this phonostat algorithm, first the normal modes of vibrations are obtained using the frozen phonon calculation. Then we track modal energy in various vibrational modes by projecting the velocities and displacements of the group of atoms within each and every MD time step on to the corresponding normal mode. The modal energies are then coupled to a Nose-Hoover thermostat to regulate the temperature. We then emply this technique to understand the distribution and relaxation of vibrational modes in carbon nanotubes. [Preview Abstract] |
Wednesday, March 17, 2010 8:36AM - 8:48AM |
P23.00004: Gaussian Molecular Dynamics in Imaginary and Real Time Ionut Georgescu, Vladimir Mandelshtam The variational Gaussian wavepacket (VGW) method can be used to estimate the equilibrium density matrix by propagating Gaussian wavepackets in imaginary time [1,2]. It has proven to be practically accurate and computationally less expensive than the path integral methods. We compare the VGW method to the Feynman-Kleinert approximation (FKA), which has comparable computational cost. Although both methods are variational, they utilize different variational principles: In FKA the partition function is optimized, while in VGW it is the imaginary-time-dependent wave packet. We show that the VGW method is more accurate for a wide variety of systems. The differences are particularly important when thermodynamic properties, such as heat capacity, are of main interest. Moreover, unlike the case of FKA, in the VGW method the imaginary frequencies do not arise. In the spirit of the Centroid Molecular Dynamics the VGW method has also been extended to simulate the real-time dynamics, e.g., it can be used to estimate the Kubo-transformed quantum time correlation functions. The latter are exact in the high-temperature and harmonic limits. \newline [1] P. Frantsuzov and V.A. Mandelshtam, J. Chem. Phys \textbf{121}, 9247 (2004)\newline [2] C. Predescu, P. Frantsuzov and V.A. Mandelshtam J. Chem. Phys \textbf{122}, 154305 (2005) [Preview Abstract] |
Wednesday, March 17, 2010 8:48AM - 9:00AM |
P23.00005: Interatomic Potentials for Ultra High Temperature Ceramics (UHTC): ZrB2 and HfB2 John Lawson, Murray Daw, Charles Bauschlicher Ultra high temperature ceramics including ZrB2 and HfB2 are characterized by high melting point, good strength, and reasonable oxidation resistance. These materials are of interest for use as sharp leading edges for hypersonic vehicles among other applications. Progress in computational modeling of UHTCs has been limited in part due to the absence of suitable interatomic potentials. We present a Tersoff style parametrization of such potentials for ZrB2 and HfB2 appropriate for atomistic simulations. Parameters are fit to data generated from DFT based ab initio calculations. The accuracy of the potentials is assessed against further ab initio data. [Preview Abstract] |
Wednesday, March 17, 2010 9:00AM - 9:12AM |
P23.00006: A polarisable atomistic force-field for alumina parametrised using density functional theory Joanne Sarsam, Mike Finnis, Paul Tangney We present an effective potential for bulk alumina (Al$_2$O$_3$), which has been parametrised by fitting the energies, forces, and stresses of a large database of reference configurations to those calculated with density functional theory (DFT). This approach, pioneered by Ercolessi and Adams [1], allows the construction of accurate force-fields without reference to experimental data, given a functional form which is capable of mimicking those electronic effects which dominate interatomic forces. Our functional form is simpler and less expensive than previous models of alumina parametrised by this technique. We compare our potential to those existing models [2], and to experimental and {\em ab initio} data on crystal structures and energies, elastic constants, phonon spectra, and thermal expansion. We demonstrate an overall accuracy that is close to that of DFT for these quantities. Some applications to the structure and diffusion of defects in corundum are discussed.\\[4pt] [1] F. Ercolessi and J. B. Adams. Europhys. Lett. {\bf 26}, 583 (1994). [2] S. Jahn, P. A. Madden and M. Wilson. Phys. Rev. B {\bf 74}(2), 024112 (2006). [Preview Abstract] |
Wednesday, March 17, 2010 9:12AM - 9:24AM |
P23.00007: Development of an interatomic potential for aluminum oxynitride N. Scott Weingarten, Iskander G. Batyrev, Betsy M. Rice Aluminum oxynitride (AlON, or Al$_{23}$N$_{5}$O$_{27})$ is a ceramic whose transparency and high strength make it a potentially useful material for many structural engineering applications. AlON is a cubic spinel, with anions forming a close-packed structure, and aluminum atoms occupying the tetrahedral and octahedral interstitial sites, with one site remaining vacant. However, the location of the vacancy has not been determined experimentally, nor have the positions of the nitrogen atoms, which replace oxygen atoms in the close-packed structure. We have developed an interatomic potential, based on the Buckingham model, for use in classical molecular dynamics (MD) simulations of AlON. The adiabatic shell model can be used to polarize atoms, which allows for a more accurate description of dielectric and defect properties of the material. Using crystal structures determined from first principles calculations, we have calculated a number of material properties using this model, including the lattice parameter and elastic constants, both with and without the shell model, and we compare these to experimental values. [Preview Abstract] |
Wednesday, March 17, 2010 9:24AM - 9:36AM |
P23.00008: A Molecular Dynamics Study of Water Confined Between Hydrophobic Plates Josh Layfield, Diego Troya The interactions between water and hydrophobic surfaces are ubiquitous in nature and have warranted a wealth of research effort studying their properties. We present a molecular dynamics (MD) simulation of the behavior of water molecules when confined between two hydrophobic self-assembled monolayers of varying size. We have performed MD simulations with simulation times $>$10ns in an effort to understand the nature of water in hydrophobic environments. Analysis of the water between the hydrophobic plates elucidates the role that hydrogen bonding, water structuring, and orientation play in hydrophobic interactions. We compare the simple point charge-extended (SPC/E) and TIP5P molecular-mechanics force fields and their ability to model the water in a hydrophobic environment. We also present the preliminary results from an investigation into the role that dissolved gases have in the ordering of water between hydrophobic plates. [Preview Abstract] |
Wednesday, March 17, 2010 9:36AM - 9:48AM |
P23.00009: Study of metal oxide nanostructures using molecular dynamics simulation Wun Chet Davy Cheong The nanostructures of metal oxides such as TiO2 and ZnO have a wide range of potential applications such as in catalyst, photoelectronics, photochemistry, surface coatings and power generation. However, experimental methods to characterize these nanostructures are limited and the growth mechanisms of these nanostructures are still not well understood. In this study, molecular dynamics is used to elucidate some interesting mechanical properties of TiO2 and ZnO nanowires on a size scale that is currently still not possible to investigate experimentally. It is found that the mechanical properties of these nanowires vary drastically and differently depending on the size of the wires. Classical molecular dynamics is able to clearly explain the different mechanisms and their causes. Simulation results show that while large surface-to-volume ratio is responsible for their size effects, ZnO and TiO2 wires displayed opposite trends. It was also found that when crystalline ZnO nanowires are stretched, necking initiated at localized amorphous regions to eventually form single-atom chains which can sustain strains above 100{\%}. Such large elongations are not observed in TiO2 nanowires. Molecular dynamics is also used to explain the formation of nanorings, helices and bows by ZnO nanobelts. [Preview Abstract] |
Wednesday, March 17, 2010 9:48AM - 10:00AM |
P23.00010: Dynamics of SiO$_2$: A Computer Simulation Landon Chambers, Katharina Vollmayr-Lee, Robin Bjorkquist Using molecular dynamics simulations, we study the dynamics of SiO$_2$ which undergoes a temperature quench from a higher temperature, $T_{\rm i} \in \{5000$\,K$, 3760$\,K\} to a lower temperature $T_{\rm f} \in \{ 2500$\,K, 2750\,K, 3000\,K, 3250\,K$\}$. We observe the system at the lower temperature as its dynamics changes from out-of equilibrium to equilibrium dynamics. Using single particle trajectories we identify ``jumps'' and ``drifts'' (particle motion between jumps). To characterize the dynamics we determine the size and duration of jumps and drifts as a function of waiting time, which is the elapsed time since the temperature quench until the time of the measurement. We find that out-of-equilibrium all investigated quantities are dependent on waiting time. With increasing waiting time the size of the drifts increases with increasing waiting time and becomes more comparable to jump sizes at larger waiting times. For large waiting times the system reaches equilibrium and all investigated quantities become independent of waiting time. [Preview Abstract] |
Wednesday, March 17, 2010 10:00AM - 10:12AM |
P23.00011: Hybrid QM/MM simulations of liquids with molecule exchange Noam Bernstein, Csilla Varnai, Ivan Solt, Monika Fuxreiter, Gabor Csanyi Many chemical reactions occur in a solvent, including essentially all biologically relevant ones. To describe these reactions accurately, one needs both a quantum mechanical (QM) description of the reaction site, as well as a large number of solvent molecule which affect the reaction via their electrostatic fields and free energy effects of their long-range structure. We present results of hybrid simulations that embed a density-functional theory QM region in a solvent region described with the CHARMM interatomic potential. Forces in the QM region are accurate, due to the use of force mixing and buffer regions. The methods allows for diffusion of molecules into and out of the QM region, a capability that improves the description of the local structure, and is essential for reactions where the transport of products is important for determining the reaction rate. We present results of the method on the structure of pure water, and dissolved ions. For pure water we show that the structure in the QM region reproduces the full QM results, and that it changes smoothly to the full MM result away from the QM region. We also show the structure of water around a dissolved Cl$^-$ ion. [Preview Abstract] |
Wednesday, March 17, 2010 10:12AM - 10:24AM |
P23.00012: ABSTRACT WITHDRAWN |
Wednesday, March 17, 2010 10:24AM - 10:36AM |
P23.00013: High-resolution O(N) DFT method and its application to large-scale nanowire simulations Jean-Luc Fattebert, Sebastien Hamel, Giulia Galli Using a real-space finite difference discretization and orbitals localization techniques, accurate O(N) Density Functional Theory calculations of systems made of thousands of atoms are now possible [1]. Using that methodology, we have investigated the static dielectric properties of silicon nanorods for diameters as large as 5 nm. We used a finite electric field method with non-periodic boundary conditions to calculate the dielectric response of the system, extending a previous study [2] to larger nanowires. \\[4pt] [1] J.-L. Fattebert and F. Gygi, Phys. Rev. B 73, 115124 (2006)\\[0pt] [2] S. Hamel et al., Appl. Phys. Lett. 92, 043115 (2008) [Preview Abstract] |
Wednesday, March 17, 2010 10:36AM - 10:48AM |
P23.00014: GPU implementation of Car-Parrinello molecular dynamics calculations Jacob H. Miner, Rob Farber, John J. Rehr General Purpose Graphics Processing Units (GPUs) have great potential for speeding up scientific computing applications. With modern cards offering over one teraFLOP performance and the increasing availability of libraries and compilers, the utilization of these tools is spreading rapidly. Here we investigate how the increased computational power of GPUs makes it possible to include efficiently both nuclear and electronic degrees of freedom in complex systems. We demonstrate the speed gains and the scaling with system size of the CPMD code on GPUs, compared to conventional CPU-based computer systems on several examples. We also discuss some of the challenges in implementing GPUs on these codes. [Preview Abstract] |
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