Track A
Advances in Fundamentals of Theory, Computation and Simulation of Materials Systems: Classical to Quantum
ABSTRACTS
Session A-1 Ab-initio methods for bulk and reduced-dimensional materials (density functional, many-particle interacting Green’s functions, quantum Monte Carlo, quantum chemistry techniques)
A-1:IL02 Ab initio Extended Hubbard Interactions and their Applications
YOUNG-WOO SON, Korea institute for Advanced Study, Seoul, South Korea
In this talk, I will present my group’s recent efforts on systematic improvements of self-interactions error corrections in approximated exchange-correlation functionals commonly used for first-principles calculations based on density functional theory (DFT) [1-3]. I will present theories on efficient ways of computing position-dependent self-consistent on-site and inter-site Hubbard interactions and their fully relativistic extension [1-3]. Owing to a low computational cost of the new method comparable to DFT and improved accuracy to the GW approximation, the new method provides an opportunity to study the correlated solids in large scale structures and full phase space of interests. A few examples [4-7] obtained using newly implemented subroutines within conventional first-principles codes will be demonstrated to show superior performance of our new method.
[1] S.-H. Lee and Y.-W. Son, Phys. Rev. Res. 2, 043410 (2020).
[2] W. Yang et al., Phys. Rev. B 104, 104313 (2021).
[3] W. Yang and Y.-W. Son, in preparation (2023).
[4] J. Huang et al., Phys. Rev. B 102, 165157 (2020).
[5] W. Yang et al., J. Phys. Condens. Matter 34, 295601 (2022).
[6] B. G. Jang et al., Phys. Rev. Lett. 130, 136401 (2023).
[7] N. Tian et al., arXiv:2211.08114.
A-1:IL03 New Algorithms for Real-space Solutions to the Electronic Structure Problem for Confined Systems: Quantum Dots with Nearly a Million Electrons
J.R. CHELIKOWSKY, University of Texas at Austin, Austin, TX, USA
We report a density functional theory calculation for a nanocrystal with more than 200,000 atoms (800,000 electrons). Our system of choice was a 20 nm spherical quantum dot with 202,617 silicon atoms and 13,836 hydrogen atoms used to passivate surface bonds. To speed up the convergence of the eigenspace, we utilized Chebyshev-filtered subspace iteration. For sparse matrix-vector multiplications, we used blockwise Hilbert space-filling curves, implemented in the PARSEC code. We utilized all of the 8,192 nodes (458,752 processors) on the Frontera machine at the Texas Advanced Computing Center. We achieved two subspace filtering iterations, yielding a good approximation of the electronic density of states. Our work pushes the limits of the current solvers to nearly a million electrons for the ground state. We will also discuss excited state computations with the GW method for Si quantum dots. We reformulate the correlation part of the GW self-energy as a resolvent matrix element, which can be efficiently computed with a Lanczos algorithm. We will illustrate calculations of quasiparticle energies for a nanocrystal containing nearly 2,000 atoms with around 8,300 electrons.
A-1:IL06 Engineering the Properties of 2D Materials by Defect Creation, Strain and Intercalation
A. Krasheninnikov, Institute of Ion Beam Physics and Materials Research, Helmholtz-Zentrum Dresden-Rossendorf, Dresden, Germany
Following isolation of graphene, many other 2D systems, e.g., single sheets of transition metal dichalcogenides, have been manufactured. All these materials contain defects and impurities, which may govern their electronic and optical properties. Moreover, defects can intentionally be introduced using beams of energetic particles – ions and electrons. Formation of defects may also give rise to phase transformations in these materials and/or tune their properties. Mechanical strain and intercalation by, e.g., alkali metal atoms, can further be used to tailor the materials characteristics. All of these calls upon the studies on defects and their role upon intercalation, response of materials to strain and irradiation. In my talk, I will present the results of our recent theoretical studies of point and line defects in 2D materials obtained in close collaboration with several experimental groups [1-3]. I will further discuss how new 2D phases of materials can be created upon atom intercalation between graphene sheets and address the role of defects in this process [4].
1. F. Long et al., Nano Lett. 23 (2023) 8468.
2. F. Davies, et al., “2D Mater. (2023). DOI: 10.1088/2053-1583/ad00ca.
3. J. Li, et al., ACS Nano 17 (2023) 5913.
4. X. Zhang, et al., Mater. Today Ener. 34 (2023) 10129
A-1:IL07 Are Simulations and Experiments Accurate for the Lattice Energies of Molecular Crystals?
F. Della Pia, A. Zen, D. Alfè, A. Michaelides, Department of Earth Sciences, University College London, London, UK
The lattice energies of molecular crystals are relevant to a wide range of scientific areas, extending from benchmarks of electronic structure theory to structure prediction and its application to pharmaceuticals and organic semiconductor devices. Due to a delicate interplay of intermolecular interactions, they are challenging systems to describe accurately with computational approaches. Faithful predictions of e.g., new molecular crystals with defined properties, require simulations that describe lattice energies of molecular crystals within the 4 kJ/mol chemical accuracy limit; a target that more often than not is not achieved with current methods. In this work, we perform diffusion Monte Carlo calculations on the entire X23 dataset, comprising 23 molecular crystals, providing valuable reference values. We conduct a rigorous analysis on both experimental and state-of-the-art computational lattice energies and show that high-accuracy computational methods agree within the chemical accuracy limit, as opposed to an actual uncertainty on the experimental lattice energies which is larger than the target threshold. We conclude that computational high-accuracy methods have become at least as reliable as (computationally corrected) experiments for electronic lattice energies and make a further step towards the application of explicitly correlated electronic structure calculations to more complex condensed phase systems.
A-1:IL08 Many-body Effects on the Photophysics of 2D Materials
D.Y. QIU, Yale University, New Haven, CT, USA
In low-dimensional and nanostructured materials, the optical response is dominated by correlated electron-hole pairs---or excitons---bound together by the Coulomb interaction. Understanding the energetics and dynamics of these excitons is essential for diverse applications across optoelectronics, quantum information and sensing, as well as energy harvesting and conversion. By now, it is well-established that these large excitonic effects in low dimensional materials are a combined consequence of quantum confinement and inhomogeneous screening. However, many challenges remain in understanding their dynamical processes, especially when it comes to correlating complex experimental signatures with underlying physical phenomena through the use of quantitatively predictive theories. In this talk, I will discuss three different frontiers related to the first principles understanding of exciton dynamics. Firstly, we will explore the relationship between exciton dispersion and exciton-phonon interactions. Secondly, we will look at how the electron-hole exchange interaction drives dynamical excitons processes for both bound and resonant exciton states in halide perovskites, transition metal dichalcogenides (TMDs), and topological insulators. Finally, we will look at many-body effects in nonlinear optics in the nonperturbative regime and show that excitons contribute to an enhancement of high harmonic generation in solids.
A-1:IL10 Auxiliary-field Quantum Monte Carlo Beyond Hartree-Fock Trial Wavefunctions
JOONHO LEE, Harvard University, Cambridge, MA, USA
This talk will present recent progress in using sophisticated trial wavefunctions in auxiliary-field quantum Monte Carlo (AFQMC). We will also discuss the use of quantum circuit wavefunctions as a trial wavefunction in AFQMC. We will highlight success and failure along with future research directions.
Session A-2 Quantum many-body methods for study of electron-electron and electron-phonon interactions
A-2:IL01 Correlation-enhanced Electron-phonon Interaction in Oxide Superconductors from GW Perturbation Theory
ZHENGLU LI, Mork Family Department of Chemical Engineering and Materials Science, University of Southern California, Los Angeles, CA, USA
Accurate and practical ab initio treatment of electron-phonon (e-ph) coupling is essential to the understanding of many condensed-matter phenomena. In this talk, I will present a recently developed ab initio linear-response method named GW perturbation theory (GWPT) that computes the e-ph interaction with the inclusion of the GW nonlocal, energy-dependent self-energy effects. GWPT goes beyond the commonly used density-functional perturbation theory (DFPT), which becomes inadequate in some materials when correlation effects are non-negligible. We demonstrate the GWPT method by showing that the e-ph coupling in Ba1-xKxBiO3 is significantly enhanced by many-electron correlation, strong enough to explain its high superconducting Tc of 32 K. Furthermore, GW-level anisotropic Eliashberg equation calculations suggest that infinite-layer nickelate superconductor Nd1-xSrxNiO2 may host a strong phonon-mediated two-gap s-wave superconductivity. I will also present studies on the e-ph coupling in cuprates and discuss new understanding in phenomena such as the ubiquitous 70-meV nodal dispersion kink.
A-2:IL02 Fundamental Theory of Geometric Phase and Non-adiabatic Phenomena
R. REQUIST, Fritz Haber Center for Molecular Dynamics, Hebrew University of Jerusalem, Jerusalem, Israel
Geometric phase and other quantum geometric concepts are important because they help in understanding complex phenomena and often lead to more efficient computational methods. It is well known that the adiabatic geometric phase associated with conical intersections of potential energy surfaces is quantized to 0 or pi in the absence of spin-orbit interactions. Even if the nuclear configuration never approaches the conical intersection, it makes its presence known by changing the energetic ordering of vibrational states. Non-adiabatic effects lift the quantization of the geometric phase, making it a path-dependent quantity [1]; however, the numerical methods used to show this were not extensible to large systems. In this talk, I introduce a new perturbation theory of the large nuclear mass limit. Non-adiabatic corrections to the geometric phase and non-adiabatic phenomena can be calculated efficiently with this theory. I will discuss the implications for density function perturbation theory and calculations in electron-phonon systems [2].
[1] Requist Tandetzky Gross, Phys. Rev. A 93, 042108 (2016).
[2] Requist Proetto Gross, Phys. Rev. B 99, 165136 (2019).
Session A-3 Molecular dynamics, Langevin dynamics, stochastic and finite element methods
A-3:IL01 Universal First Principles Force-fields for Materials Simulations based on Sparse Gaussian Process Regression
KWANG S. KIM, Department of Chemistry, Ulsan National Institute of Science and Technology (UNIST), 50 UNIST-gil, Ulsan, South Korea
The sparse Gaussian process regression (SGPR) algorithm is utilized for machine learning of ab initio interatomic potentials/forces, improving scalability by using local chemical environments based on characteristic geometric features selected via active learning (arXiv:200913179 (2020); Phys Rev B 103, 214102 (2021)). The local models can be combined to generate global, transferable, scalable, and precise universal potentials for complex systems (J. Phys. Chem. Lett. 12, 8115 (2021)). SGPR enables faster computing speeds, better accuracy, and smaller data sets than existing methods and has been applied to a variety of systems, including hydrocarbons (J. Phys. Chem. A 125, 9414 (2021)), metal clusters (J. Phys.: Condensed Matter. 34, 344007 (2022)), ternary solid electrolytes (J. Phys. Chem. Lett. 12, 8115 (2021)), lithium batteries (Adv. Energy Mater. 12, 2201497 (2022)), and perovskite solar cells (Adv. Energy Mater. 12, 2202279 (2022); Nature 598, 444 (2021)). This method provides new physical and chemical insights into complex molecular and material systems and paves the way for large-scale complexity modeling through a combinatorial approach.
A-3:IL03 Prediction of a Supersolid Phase in High-pressure Deuterium
CHANG WOO MYUNG, Department of Energy Science, Sungkyunkwan University, Suwon, South Korea
Supersolid is a mysterious and puzzling state of matter whose possible existence has stirred a vigorous debate among physicists for over 60 years. Its elusive nature stems from the coexistence of two seemingly contradicting properties, long-range order and superfluidity. We report computational evidence of a supersolid phase of deuterium under high pressure (p > 800 GPa) and low temperature (T < 1.0 K). In our simulations, that are based on bosonic path integral molecular dynamics and machine learning potential, we observe a highly concerted exchange of atoms while the system preserves its crystalline order. The exchange processes are favoured by the soft-core interactions between deuterium atoms that form a densely packed metallic solid. At the zero-temperature limit, Bose-Einstein condensation is observed as the permutation probability of N deuterium atoms approaches 1/N with a finite superfluid fraction. Our study provides concrete evidence for the existence of a supersolid phase in high-pressure deuterium and could provide insights on the future investigation of supersolid phases in real materials.
A-3:L04 Graph Theory Ideas Reveal Long Range Conduction Pathways
m.a. gomez, Department of Chemistry, Mount Holyoke College, South Hadley, MA, USA
Finding long range conduction pathways is challenging when the system includes both fast frequency modes needing short time step force integration and slow frequency modes requiring long time samples. Methods such as kinetic Monte Carlo avoid the integration of steps and instead use probabilities to choose the next moves and advanced the clock based on the move chosen. Nevertheless, extracting fast conduction mechanisms between traps can be challenging. Centrality methods based on the number of steps to return to key sites or vertices in a graph have been used to identify the most central areas in a graph. However, physical systems with traps and highways have steps with distinctly different barriers, making steps non-equivalent. This contribution reviews both traditional and new graph theory schemes for finding long range pathways, with a special focus on time based centrality measures [1] and applications of these to proton conduction in doped perovskites where correlated proton motion is important.[2]
[1] Gomez-Haibach, KS; Gomez, MA, Revised Centrality Measures Tell a Robust Story of Ion Conduction in Solids, J. Chem Phys B, in press.
[2] Pan, Y.; Hoang, M.T.; Mansoor, S.; Gomez, M.A. Inorganics 2023, 11, 160. https://doi.org/10.3390/inorganics11040160
Session A-4 Advances in multiscale computation methods, from the atomistic to the mesoscopic and continuum levels
A-4:IL01 Microstructure Prediction of High Temperature Alloys by a First-principles Phase Field Method
RYOJI SAHARA1, T.N. Pham2, S. Bhattacharyya2, 3, R. Kuwahara4, K. Ohno1, 2, 1National Institute for Materials Science, Japan; 2Yokohama National University, Japan; 3Birla Institute of Technology and Science Pilani, India; 4Dassault Systèmes K.K., Japan
It is important to predict microstructures by computer simulation for effective design of high temperature alloys. However, since conventional phase field method has no ability of prediction because it uses experimental data and empirical parameter. To overcome the problem, in this presentation, we will show the results of microstructure prediction of alloys by a first-principles phase field method (FPPF), which is originally developed by our research group. In order to combine computational methods at different scales from nanoscale (first-principles calculations) to macroscale (phase field method), the potential renormalization theory and cluster expansion theory are used to obtain the free energy function as a step-wise function of the concentration φX of element X. Then solve the macroscale Cahn-Hilliard equation. We will show that the FPPF can predict microstructures of alloys such as Ni-Al alloy (1), Ti-6Al-4V (2) alloy and so on at high temperature region without any empirical parameter.
(1) S. Bhattacharyya, R. Sahara, and K. Ohno, Nature Communications 10 (2019) 3451.
(2) T. N. Pham, K. Ohno, R. Sahara, R. Kuwahara, and S. Bhattacharyya, J. Phys.: Cond. Matt. 32 (2020) 264001.
A-4:L02 Ab Initio Informed Microstructure and Process Modelling of Metals
D. SCHEIBER, Materials Center Leoben Forschung GmbH, Leoben, Austria
The steel industry on its own is responsible for an estimated 11% of the total CO2 emissions worldwide showing the need for increased recycling. But plain circular economy leads to an accumulation of unwanted impurities which increases the scatter in microstructure formation and deteriorates material properties being the root-cause of downcycling. While available models for process design interpolate within known parameter spaces, they fail at extrapolative modelling and do not consider impurities. Yet, extrapolative models are needed to allow for high recycling rates also in microstructure-sensitive high-performance alloys. This talk will present the outline of a predictive multi-scale modelling framework for microstructure evolution that combines ab initio data with mean field models. Such a multi-scale modelling framework can be used to optimize process parameters to mitigate impurity effects from recycling inline in production. Specific examples will be shown for connecting ab initio data with mean field models of recrystallization, precipitation, grain growth, and segregation. The examples show the feasibility of such an approach and point the way towards needed future research.
Session A-5 Ultrafast excitation and decay processes in materials
A-5:IL01 Quantum Dynamics of Charge Carriers in Optoelectronic Materials
O. PREZHDO, University of Southern California, Los Angeles, CA, USA
Excited state dynamics play key roles in numerous molecular and nanoscale materials designed for energy conversion. Controlling these far-from-equilibrium processes and steering them in desired directions require understanding of material’s dynamical response on the nanometer scale and with fine time resolution. We couple real-time time-dependent density functional theory for the evolution of electrons with non-adiabatic molecular dynamics for atomic motions to model such non-equilibrium response in the time-domain and at the atomistic level. The talk will introduce the simulation methodology and discuss exciting applications, e.g., metal halide perovskites, transition metal dichalcogenides, semiconducting and metallic quantum dots and films, polymers, molecular crystals, nanocarbon, etc. Photo-induced charge and energy transfer, plasmonic excitations, Auger-type processes, energy losses and charge recombination create many challenges due to qualitative differences between molecular and periodic, and organic and inorganic matter. Our simulations provide a unifying description of quantum dynamics on the nanoscale, characterize the timescales and branching ratios of competing processes, resolve debated issues, and generate theoretical guidelines for development of novel systems.
A-5:IL02 Correlated Electron-nuclear Dynamics of Extended Systems Based on Exact Factorization
SEUNG KYU MIN, Ulsan National Institute of Science and Technology, Ulsan, South Korea
Correlated electron-nuclear dynamics is crucial in various phenomena, including photosynthesis, photovoltaics, photocatalysis, radiation chemistry, polaron chemistry, and quantum computing. Electron-nuclear correlations are handled using mixed quantum-classical approaches for small and intermediate molecular systems. However, for extended molecular systems, real-time time-dependent density functional theory (RT-TDDFT) is used frequently. RT-TDDFT propagates electronic density with time-dependent Kohn-Sham orbitals and time-dependent external potentials while the classical nuclei move according to the Ehrenfest-type equation of motion. While the electronic equation of motion can describe nonadiabatic transitions, it cannot account for nuclear wave packet splitting and quantum decoherence. To overcome this limitation, we propose a Hermitian form of an electron-nuclear correlation operator, which is equivalent to the original non-Hermitian operator. This approach obtains a stable real-time and real-space electronic propagation with quantum decoherence in correlated electron-nuclear dynamics, which is essential for condensed phase simulations [1].
[1] Han, D.; Ha, J.-K.; Min, S.K., J. Chem. Theory Computat., 2023, 19, 2186-2197.
A-5:IL03 Ab initio Studies of Field-driven Ultrafast Excitations and Time-dependent Phenomena
YANG-HAO CHAN, D.Y. Qiu, F.H. da Jornada, S.G. Louie, Institute of Atomic and Molecular Sciences, Academia Sinica and Physics Division, National Center for Theoretical Sciences, Taipei, Taiwan; Department of Physics, University of California at Berkeley, CA, USA and Materials Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, CA, USA
Atomically thin quasi two-dimensional (2D) insulating materials exhibit novel exciton physics due to ineffective screening, quantum confinement, and topological effects. Going beyond near-equilibrium set-up, one expects that excitonic effects also dominate the responses of out-of-equilibrium systems and can lead to interesting phenomena in optically-driven 2D materials. Using a newly developed real-time, non-equilibrium Green function method within the adiabatic GW approximation, we show that, for non-centrosymmetric 2D semiconductors, excitonic effects give rise to a strong DC current, the so-called shift current, upon even sub-bandgap frequency CW light illumination through a second-order nonlinear optical process, which is of promise for applications with appropriate materials. Furthermore, we show that, in optical-field-driven angle-resolved photoemission spectroscopy (ARPES) experiments, the energy and wavefunction of excitons may be measured directly under achievable laboratory conditions. We also applied our method to simulate the transient absorption spectrum in a transition metal dichalcogenide heterostructure and reveal the importance of excitonic effects in ultrafast charge transfer. Developments to include electron-phonon couplings in the simulation will be discussed.
A-5:IL05 Excitons in Complex Materials from First Principles
J.B. NEATON, Department of Physics, University of California, Berkeley Materials Sciences Division, Lawrence Berkeley National Laboratory Kavli Energy Nanosciences Institute at Berkeley, Berkeley, CA, USA
The ability to synthesize and probe new classes of complex photoactive materials with tunable structure and composition – such as halide perovskites, molecular solids, few-layer van der Waals heterostructures, and more – has driven the development of new theory, computational methods, and intuition for predicting their photophysics. In these novel semiconductors, photoexcited correlated electron-hole pairs, or excitons, can be strongly bound and do not conform to simple models, and new understanding is needed to interpret and predict their behavior. Here, I will discuss recent advances in density functional theory and ab initio many-body perturbation theory within the GW approximation and the Bethe-Salpeter equation framework that enable predictive understanding of excitons in these complex systems, including how they are influenced by lattice structure and dynamics, temperature, dielectric screening, and carrier concentration. I will share recent calculations for multiple systems, comparing them with experiments, and discuss new intuition for the nature and fate of excitons, as well as prospects for future predictive calculations.
A-5:IL07 Atomistic Modeling of Laser-induced Melting and Ablation of Thin Films and Nanoparticles
L.V. Zhigilei, C. Chen, M.I. Arefev, H. Huang, A.S. Valavanis, Department of Materials Science and Engineering, University of Virginia, Charlottesville, VA, USA
The interactions of ultrashort pulse lasers with thin metal films and colloidal nanoparticles has high practical importance, with applications ranging from the generation of chemically clean nanoparticles to high-precision micro/nanomanufacturing of multilayered systems in microelectronics. At the same time, strong laser excitation enables access the extreme states of electronic, mechanical, and thermodynamic nonequilibrium, thus providing unique opportunities for investigation of material behavior and properties far from the equilibrium conditions. Large-scale atomistic simulations of ultrashort pulse laser interactions with thin metal films and colloidal nanoparticles provide insights into the kinetics and microscopic mechanisms of laser-induced melting, the dynamics of photomechanical fragmentation, and the nature of a rapid (explosive) phase decomposition of films and nanoparticles superheated up to the limit of thermodynamic of the molten material. The calculation of electron diffraction and small-angle X-ray scattering profiles, as well as the transient optical properties of the irradiated targets, enables direct validation of the computational predictions against the results of pump-probe optical and diffraction probing of the laser-induced phase transformations.