Session M58: Ultrafast Dynamics from Electron-Phonon Interactions I

Ab initio approach for exciton-phonon interactions and exciton dynamics

Advances in ab initio calculations for electron-phonon (eph) coupling has enabled high- accurate predictions on phenomenon rooted in the interplay between carrier and lattice vibration in materials ranging from silicon semiconductors and molecular crystals to perovskites. However, in materials of weak Coulomb screening, such as low-dimensional semiconductors and quantum dots, predicting the carrier dynamics remains an open field to explore; there, carriers are dominated by strong-bounded electron-hole paired excitation, i.e., the exciton.

In this talk, we will discuss the method combining many-body perturbation theory and density functional perturbation theory to compute the exciton-phonon (ex-ph) coupling strength. Subsequently, we will present treatments to study statical observables that can predict carrier relaxation and phonon-modulated optical response. Lastly, we introduce the approaches adopting the Boltzmann transport equation to examine the ultrafast transient signatures, such as trARPES and transient absorption.   

First-principles ultrafast phonon dynamics by solving the real-time Boltzmann transport equation with adaptive multirate time stepping

Ultrafast lattice dynamics in materials can be accessed with pump-probe spectroscopies and used to prepare novel nonequilibrium quantum phases. Therefore, quantitative modeling of ultrafast lattice dynamics would advance nonequilibrium physics. Electron dynamics can be modeled by the electron real-time Boltzmann transport equation (rt-BTE) with first-principles electron-phonon (e-ph) collisions using an efficient parallel algorithm [1]. Yet, solving the lattice (phonon) rt-BTE with e-ph and phonon-phonon (ph-ph) collisions [2] remains challenging due to the different timescales of e-ph and ph-ph interactions. In this work, we interface the PERTURBO code [1] with the SUNDIALS library [3] to efficiently advance coupled electron and phonon rt-BTEs in time. We demonstrate a significant speed-up using adaptive step size and multirate infinitesimal (MRI) methods from SUNDIALS that enable utilizing different time step sizes for fast-evolving, cheap e-ph interaction and the slow, expensive ph-ph interaction. These advances allow us to model ultrafast lattice dynamics in real materials with moderate computational cost. Application to graphene and silicon will demonstrate the capabilities of this novel first-principles tool, which expands the scope of nonequilibrium first-principles calculations.

[1] J.-J. Zhou, et al. Comput. Phys. Commun. 264,107970 (2021)

[2] X. Tong, M. Bernardi, Phys. Rev. Research, 3 (2021)

[3] D. R. Reynolds, et al. ACM Trans. on Math. Software, 49(2), pp. 1-26 (2023)   

Ultrafast dynamics and dichroism of chiral valley phonons in monolayer MoS2

Valley degrees of freedom in transition-metal dichalcogenides influence thoroughly electron-phonon coupling and its nonequilibrium dynamics. We conducted a first-principles study of the ultrafast dynamics of chiral phonons following valley-selective carrier excitation with circularly-polarized light. Our numerical investigations treat the ultrafast dynamics of electrons and phonons on equal footing within a parameter-free ab-initio framework. We report the emergence of valley-polarized phonon populations in monolayer MoS₂ that can be selectively excited at either the K or K' valleys depending on the light helicity. The resulting vibrational state is characterized by a distinctive chirality, which lifts time-reversal symmetry of the lattice on transient timescales. We show that chiral valley phonons can further lead to fingerprints of vibrational dichroism detectable by ultrafast diffuse scattering and persisting beyond 10 ps. The valley polarization of nonequilibrium phonon populations could be exploited as information carrier, thereby extending the paradigm of valleytronics to the domain of vibrational excitations.   

Ultrafast photo-induced lattice dynamics in MoTe_2 studied using ab-initio MD simulations and MeV Ultrafast Electron Diffraction (MUED)

Transition metal dichalcogenides (TMDs) like MoTe_2 show promising high electron mobilities and light-induced phase transitions involving their metastable trigonal 1T’ and Td phases could be used for photo-induced fast switching in electronic devices. Absorbed laser energy alters electronic properties and initiates ultrafast lattice dynamics, but the detailed mechanisms of the dynamics and under what conditions they lead to phase transitions remains unclear. Here, we performed ab-initio MD simulations of laser-excited MoTe_2 in the 1T' and T_d phases (which differ by stacking) and some mixed phases involving stacking faults. We compute the time evolution of Bragg peak intensities to identify contributions of each phase after the excitation and determine intensity thresholds for which nonthermal phase transitions are initiated. Moreover, by comparing computed Bragg peak intensities to time-dependent Bragg peak intensities measured via MUED experiments, we identify contributions in the experiment from thermal and nonthermal lattice dynamics, stacking faults, and possible phase transitions. Understanding these processes gives insight into the ultrafast control of structural and electronic properties in metastable 2D TMDs.   

First-principles characterization for the nitrogen dimer electronic free energy surface within the free-energy Born—Oppenheimer approximation.

We investigate the free energy surface for the nitrogen dimer within the free-energy Born—Oppenheimer (FEBO) approximation. Our initial investigations calculated the exact internal energy (U), entropy (S), and free energy (F) using finite-temperature full configuration interaction. Considering only the electronic degree of freedom we observed the free energy surface predicts a dissociation at temperatures below the experimental dissociation energy. To explain this behavior, we compared the contributions from U and S to F and found that the two quantities have opposing effects on the free energy, with the entropic contribution promoting dissociation. We also explored how the symmetries used within the Hamiltonian impact the temperature for dissociation and found that more symmetries resulted in lower temperatures for dissociation. To test the impact of basis set size, we used the density matrix quantum Monte Carlo (DMQMC) family of methods to simulate the finite temperature density matrix in a cc-pVDZ basis set. We found that the larger basis set resulted in the temperature for dissociation increasing. Finally, we explored the origins of this behavior using histograms for the density matrix elements. We observed that the histogram for longer bond lengths showed more variation across temperature compared to shorter bonds. Additionally, the longer bond length histograms shared many similarities with ground state histograms for multi-reference systems.   

Light induced phase transitions in GeTe and SnSe monochalcogenides.

In this talk I will demonstrate the capability of constrained density functional perturbation theory (1) coupled with a non-perturbative treatment of quantum anharmonicity (2) to describe light-induced phase transition in GeTe and SnSe monochalcogenides.

In SnSe I will show that ultrafast lasers can permanently transform the topologically-trivial orthorhombic structure of SnSe into the topological crystalline insulating rocksalt phase via a first-order non-thermal phase transition (3). I will describe the reaction path and evaluate the critical fluence and the possible decay channels after photoexcitation.

Our simulations of the photoexcited structural and vibrational properties are in excellent agreement with recent pump-probe data in the intermediate fluence regime below the transition with an error on the curvature of the quantum free energy of the photoexcited state that is smaller than $2\%$.

In GeTe I will investigate the non-thermal phase transition from a rhombohedral to a rocksalt crystalline phase. The microscopic mechanism and nature of the transition are unclear. I will show that the non-thermal phase transition is strongly first order and does not involve phonon softening, in contrast to the thermal one. The transition is driven by the closure of the single-particle gap in the photoexcited rhombohedral phase. Finally, I will show that ultrafast XRD data are consistent with the coexistence of the two phases, as expected in a first order transition.

The high accuracy of our results demonstrates that an approach based on a complete self-consistent constrained density functional perturbation theory in the presence of an electron-hole plasma captures the light-induced structural deformations and transitions extremely accurately.   

Designing Active Spaces with Periodic-like Properties for Charge Transport in Multiconfigurational Nanoscale Systems

Fundamentally quantum charge transport systems such as molecular electronics require a fully quantum transport approach that consistently captures both dynamic and multireference electron correlation effects. We previously introduced a unique methodology for integrating multiconfigurational electronic structure methods, particularly multiconfiguration pair density functional theory (MC-PDFT), within a non-equilibrium Green's function (NEGF) formalism. This approach uses an active space to capture both electronic correlation types for a set of transport-relevant orbitals while treating the non-active orbitals with density functional theory. This approach creates a new theoretical challenge as limited to no design principles exist for selecting active spaces for nanoscale charge transport problems. In this talk, I will demonstrate design principles for creating active spaces for charge transport that maintain periodic-like characteristics and that can capture quantum transport phenomena in nanoscale electronics such as conductance decay reversals.   

Ab initio excited state forces in the study of self-trapped excitons and coherent phonon generation.

In transition metal dichalcogenides (TMD) the exciton-phonon interaction (EPI) plays an important role in optical and Raman spectra. One manifestation of EPI is the excited state (ES) forces, which are the gradient of the ES potential energy surface. Those forces indicate the direction atoms tend to move due to light absorption and are useful in the microscopic understanding of exciton self-trapping. We implemented an ES forces code that combines GW/BSE results from BerkeleyGW and electron-phonon coefficients from Quantum Espresso. With our results, we can relax the system towards the minimum of the ES potential energy surface, using gradient descent or a harmonic extrapolation, and avoiding the need for 3N GW/BSE calculations for finite differences. First, we study LiF as a test system for self-trapped excitons (STEs). By relaxing from polaronic-like distortions we found an STE that shows absorption peaks redshifted about a few eV. This illustrates a novel and efficient approach to study STEs at an ab initio level. Then we move to MoS2. For the monolayer, we observe the coupling of A exciton with A’₁ phonon, while C exciton couples with the A’₁ and E’ modes. In the bilayer case, we also observe coupling with the layer breathing mode. Our results agree with experimental data on the generation of coherent phonons by ultrafast light pulses. This approach can be used as a guide to future investigations of EPI of diverse TMDs.   

Fully relativistic first-principles calculation by using extended Hubbard energy functionals

The extended Hubbard energy functional has been shown the significantly improvement for the description of electron localization and orbital hybridization on the equal footing. To validate this correction in heavy element systems, we extended the energy functionals to account for noncollinear spin states to include spin-orbit coupling effects. We present band structure results, particularly focusing on Bi₂Se₃, which exhibits substantial spin-orbit coupling. This yields a dramatic change in the low-energy band structure compared to conventional exchange-correlational functionals. Parabolic dispersion in valence and conduction bands is observed as similar as the GW approximation. We also confirm gap reduction depending on the number of layers, in line with topological insulator characteristics. Consequently, accurate high-throughput calculations are expected, even in systems that both spin-orbit coupling and strong interactions are necessarily considered.   

Progress towards a more efficient denisty matrix quantum Monte Carlo method

In this talk, we will present our work in modifying the density matrix quantum Monte Carlo (DMQMC) algorithm, which samples the density matrix for the electronic Hamiltonian across a range of temperatures. The DMQMC algorithm is modified to propagate the density matrix in a semi-stochastic (partly stochastic and partly deterministic) fashion. This work is based on and inspired by the semi-stochastic projector modification to full configuration interaction quantum Monte Carlo (FCIQMC). This work was motivated to provide high accuracy finite temperature electronic structure results with fewer required simulations, which we refer to as increasing the method's efficiency. We evaluated the accuracy of this modification through comparisons with finite temperature full configuration interaction (ft-FCI), as well as the original DMQMC and its interaction picture variant (IP-DMQMC). We investigated the impact the size of and selection scheme for the deterministic space has on the efficiency of the semi-stochastic DMQMC algorithm. We believe that this modification will allow for increases in the systems sizes with which we can treat with finite temperature.   

Towards combining dynamical mean-field theory with diagrammatic Monte Carlo

Dynamical mean-field theory (DMFT) has been successfully combined with Density Functional Theory and other weak coupling diagrammatic methods such as GW, Hartree-Fock, and second order perturbation theory. But all these methods are uncontrolled and their success is sensitive to the choices made in the approximations or the starting point. Here, we explore how DMFT can be turned into a numerically controlled method by systematically adding to the DMFT approximation the higher order Feynman diagrams order by order using the diagrammatic Monte Carlo method.

The intersection of DMFT and the weak coupling method needs to be subtracted carefully as an exact double counting term, which is here rigorously defined and identified. Since the building blocks of the diagrammatic theory are the DMFT Green's functions, the series is expected to converge faster than with a weak coupling starting point, as the local correlations are already accounted for exactly.  The scheme is tested on the H₂ molecule in a Gaussian basis set, using second- and third-order diagrams added to DMFT, and iterating to charge self-consistency.