Bulletin of the American Physical Society
APS March Meeting 2019
Volume 64, Number 2
Monday–Friday, March 4–8, 2019; Boston, Massachusetts
Session R31: Quantum Monte Carlo (ES2)Focus
|
Hide Abstracts |
Sponsoring Units: DCP Chair: Can Ataca, Univ of Maryland-Baltimore County Room: BCEC 203 |
Thursday, March 7, 2019 8:00AM - 8:36AM |
R31.00001: Auxiliary-field quantum Monte Carlo calculations for periodic solids with Brillouin zone sampling Invited Speaker: Mario Motta Modern electronic structure methods are reaching a level of accuracy and efficiency, that enables simulation and prediction of important phenomena in solid-state chemistry. In order to allow for direct comparison with experiments, result of these simulations have to reach the thermodynamic limit of infinite system size, which substantially increases the cost of many-electron wave-function theories. |
Thursday, March 7, 2019 8:36AM - 8:48AM |
R31.00002: Auxiliary Field Quantum Monte Carlo Simulations of Real Strongly Correlated Materials Fionn Malone, Shuai Zhang, Miguel Morales Auxiliary field quantum Monte Carlo has demonstrated itself as one of the most reliable methods of simulating strongly correlated model systems. However, its application to more realistic strongly correlated solids has been limited. Here we present phaseless AFQMC results for a range of transition metal oxides. Using recent algorithmic developments, we compute properties other than the total energy including the magnetic moment and also many-particle energy gaps. Finally, we investigate the importance of the trial wavefunction through multi determinant expansions and show how interpolative separable density fitting significantly extends the size of system which can be tackled using AFQMC. |
Thursday, March 7, 2019 8:48AM - 9:00AM |
R31.00003: Diffusion Monte Carlo study of hydrogen adsorption on silicon carbide nanotube Genki Imam Prayogo, Kenta Hongo, Hyeondeok Shin, Ryo Maezono, Anouar Benali Hydrogen is one of the candidates for environmentally friendly energy carriers. Although it has a high energy density per unit weight, its volumetric energy density is rather low, making its compact storage difficult. This is rather important when storage volume is paramount, such as in automobile and aviation industries. Physisorption of hydrogen molecules on materials with high surface area to volume ratio like nanotubes is one of the strategies to increase this volumetric efficiency. Along with carbon nanotube (CNT) and boron nitride nanotube (BNNT), silicon carbide nanotube (SiCNT) is one of the candidate materials for this use. Although it has yet to be experimentally sythesized in single-walled form, larger silicon carbide nanotubes have shown promising gains compared to carbon nanotubes in terms of storage capacity and lack of sorption hysteresis. Theoretical studies points to a stronger binding energy and the existence of point charges naturally occurring on alternating Si-C surface. We present our initial Diffusion Monte Carlo (DMC) results of the adsorption of molecular hydrogen on single walled SiCNT surface. DMC is a stochastic method solving the many-body Schrödinger equation and has demonstrated accuracies below the Kcal/mol for a wide range of materials. |
Thursday, March 7, 2019 9:00AM - 9:12AM |
R31.00004: Band gaps and excitons in quantum Monte Carlo Michael Bennett, Matej Ditte, Cody Melton, Matus Dubecky, Lubos Mitas Using the fixed-node approximation, quantum Monte Carlo (QMC) allows for calculations of both ground and excited energy levels with explicit electron correlations. For example, in periodic systems, one can estimate band gaps by promoting an electron from the valence to conduction band or alternatively from the ionization potential and electron affinity. Since fixed-node QMC's accuracy depends on the quality of trial functions this tacitly assumes that the relevant states are close to single reference states and well represented by single orbital promotions in the Slater determinant. However, this becomes more challenging for excitations that appear in the gap and are not representative of typical conduction states such as significantly localized excitonic states, impurity or defect levels that are not captured by ordinary band theories. Here the construction of the corresponding trial functions requires more elaborate schemes that might involve multi-reference forms, geometry relaxation and/or additional considerations depending on experiment that one compares with such as optical, conductivity or photoemission measurements. We study a few such cases, eg, benzene molecule and crystal and other systems to capture and illustrate these effects using QMC calculations. |
Thursday, March 7, 2019 9:12AM - 9:48AM |
R31.00005: Ab Initio Finite Temperature Auxiliary Field Quantum Monte Carlo for Solids Invited Speaker: Brenda Rubenstein Predicting the finite temperature properties of molecules, and especially, solids is critical for understanding many physical phenomena. Nevertheless, developing accurate, yet efficient methodologies for finite temperature applications remains an outstanding challenge. In this work, we present an Auxiliary Field Quantum Monte Carlo method with an O(N3) scaling for studying the finite temperature electronic structure of any system that can described by an ab initio Hamiltonian. The algorithm marries the ab initio phaseless auxiliary field quantum Monte Carlo algorithm known to produce high accuracy ground state energies of molecules and solids with its finite temperature variant, long used by condensed matter physicists for studying model Hamiltonian phase diagrams, to yield a phaseless, ab initio finite temperature method. We demonstrate the accuracy of this approach for benchmark solids, including hydrogen chains and networks, and compare to it more popular mean field treatments of real materials. Our method serves as a new, robust tool for studying low, but finite temperature phase transitions in models and solids, ultracold chemistry, and warm dense matter. |
Thursday, March 7, 2019 9:48AM - 10:00AM |
R31.00006: Multi-determinant Diffusion Monte Carlo in solids at the thermodynamic limit Kevin Gasperich, Thomas Applencourt, Ye Luo, Qiming Sun, Kenneth D Jordan, Luke Shulenburger, Anthony Scemama, Michel Caffarel, Anouar Benali In the past decade, fixed-node diffusion Monte Carlo using a single determinant Slater-Jastrow trial wavefunction as proven to systematically reproduce (within 5-10Kcal/mol) the energies of a wide range of molecules and solids. While it has been demonstrated for molecular systems that the path to chemical accuracy (<1Kcal/mol) can be realized by improving the nodal surface using for example multidetermiant trial wavefunctions generated with MCSCF or Selected CI. Reaching the same level of accuracy for solids in the thermodynamic limit has proven harder due to the exponential scaling of these wavefunction generation methods. In this talk, we will discuss the effect of using a multi-determinant Slater-Jastrow trial wavefunction with complex k-points and twist averaging in combination to reach chemical accuracy in the thermodynamic limit for solid carbon. |
Thursday, March 7, 2019 10:00AM - 10:12AM |
R31.00007: Progress towards finite temperature density matrix quantum Monte Carlo calculations on solids Hayley Petras, James Shepherd There has been a recent surge in interest in finite temperature calculations on molecules driven by methods such as FCI, CC, MP2, and AFQMC. By means of a follow-up to recent work on the uniform electron gas, we here present density matrix quantum Monte Carlo (DMQMC) calculations on molecular Hamiltonians. We compare and contrast molecular diatomics with the electron gas to make comment on the scope for performing periodic calculations with DMQMC. We also describe how DMQMC can contribute to the study of finite temperature ab initio Hamiltonians with other methods. |
Thursday, March 7, 2019 10:12AM - 10:24AM |
R31.00008: Difficulty to capture non-additive enhancement of stacking energy by conventional ab initio methods Ken Qin, Tomohiro Ichibha, Kenta Hongo, Ryo Maezono We evaluated the non-additive contributions in the inter-molecular interactions in B-DNA stacking by using diffusion Monte Carlo (DMC) methods. It is found that only DMC can capture the sign inversion in the contribution (i.e., the non-additivity enhances or reduces the interaction depending on the base pairs of DNA), which is never predicted by any other ab initio methods. Even by CCSD(T), the inversion is found to be difficult to be captured because of the practical handling of CBS (complete basis set correction) at the feasible level with MP2. The predicted non-additivity turns out to be several times larger (~8 kcal/mol) than those by other methods. The importance of the hydrogen bondings between the bases horizontally on the inversion is also clarified. |
Thursday, March 7, 2019 10:24AM - 10:36AM |
R31.00009: A simpler twist averaging for the uniform electron gas designed for finite basis set calculations such as coupled cluster and full configuration interaction quantum Monte Carlo Tina Mihm, Alexandra McIsaac, James Shepherd As a long term aim, we would like to improve the efficiency of wavefunction calculations for solids. An important step is the elimination of finite size error, and it is common for continuum quantum Monte Carlo to use twist averaging. Here, we describe and analyze a simple twist averaging scheme as applied to coupled cluster and full configuration interaction quantum Monte Carlo calculations of the uniform electron gas. |
Thursday, March 7, 2019 10:36AM - 10:48AM |
R31.00010: A diffusion Monte Carlo study of point defect diffusion in rutile TiO2 bulk Tom Ichibha, Anouar Benali, Kenta Hongo, Ryo Maezono TiO2 is one of the most popular photocatalysts. The diffusion of point defects (Ti interstitial/O vacancy) is responsible for charge transport. We investigated the diffusion mechanism of the point defects in rutile TiO2 bulk using diffusion Monte Carlo (DMC). There are two issues related to the point defect diffusion: (a) The diffusion of positively charged defects from the bulk inside to the (101) surface promotes the surface oxidation reaction. Then, which type of defect transfers positive charges faster? (b) There are two known diffusion paths for Ti interstitial defects, one parallel to the c-axis and the other perpendicular. In which direction are the defects easier to diffuse? Our DMC calculations established that (a) Ti interstitial defects transfer a larger amount of positive charges to the surface and (b) the primary diffusion direction of Ti interstitial defects changes from `perpendicular to c-axis' to `parallel', along with the change of defect charge, 0 to +4. Both conclusions qualitatively support the previous GGA-DFT work. Yet, the barrier energy prediction is quantitatively much different by ~2eV at the maximum. We will discuss this huge difference and justify the barrier energies predicted by DMC, based on Bader charge analysis and electronic density prediction. |
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