Influence Functional Approach to Non-Perturbative Exciton Binding Renormalization from Phonons
Rohit Rana, Eric R. Heller, Antonios M. Alvertis, Jeffrey B. Neaton, David T. Limmer · 2026
Read Influence Functional Approach to Non-Perturbative Exciton Binding Renormalization from Phonons by Rohit Rana, Eric R. Heller, Antonios M. Alvertis, Jeffrey B. Neaton, David T. Limmer (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-22575.txt.
Introduction
Rana and co-authors model how phonons shift exciton binding energies with temperature. They build a many-body Hamiltonian whose inputs all come from first-principles methods. To treat the binding without perturbation shortcuts, they evolve the model in imaginary time with path integral Monte Carlo. For Wannier-Mott excitons the computed binding matches experiments closely. Optical phonons do the real work of shifting exciton binding. Sound-like and transverse modes matter more for the separate electron and hole polarons.
Catalog record
| CREATOR | Rohit Rana, Eric R. Heller, Antonios M. Alvertis, Jeffrey B. Neaton, David T. Limmer |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 71114 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.22575
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Influence Functional Approach to Non-Perturbative Exciton Binding Renormalization from Phonons Rohit Rana, Eric R. Heller, Antonios M. Alvertis, Jeffrey B. Neaton, David T. Limmer Abstract We construct a many-body model Hamiltonian to capture how phonons renormalize exciton binding as a function of temperature. By using the GW approximation and density functional perturbation theory, we are able to parameterize this Hamiltonian completely from first principles. To capture static quasiparticle properties non-perturbatively, we evolve this Hamiltonian in imaginary time with path integral Monte Carlo using an influence functional based approach. For a class of Wannier–Mott type excitons, our binding energies are in quantitative agreement with experiment. We find that in addition to long-range dipolar interactions from longitudinal optical modes, short-ranged deformation potentials from acoustic modes and transverse optical modes can significantly renormalize electron and hole polaron binding energies at elevated temperature. However, exciton binding energies are only appreciably renormalized by coupling to optical phonons. I Intro The evaluation of optical excitation energies in semiconductors is complicated by a myriad of physical factors. Excitations in extended systems are often strongly correlated, forming quasiparticle excitonic states, whose charged distributions couple to the surrounding lattice. 50 As a consequence, to achieve accurate estimates of these energies, numerical techniques must be developed that capture the interplay between electronic and nuclear degrees of freedom while maintaining computational efficiency. 13 In this work, we develop a downfolded Hamiltonian approach to evaluate exciton binding energies using path integral Monte Carlo (PIMC). 9 The Hamiltonian is parameterized by ab initio electronic structure methods, such as density functional theory (DFT) and many-body perturbation theory within the GW approximation 34 ; 68 , which provide an accurate representation of effective electronic degrees of freedom. An influence functional calculation within PIMC is used to compute non-perturbative and temperature-dependent exciton binding energies renormalized by coupling to phonons across a variety of polar semiconductors. 19 ; 20 ; 45 Using this approach, we find that although acoustic and transverse optical phonons can significantly affect single charge polaron binding energies, they play only a minor role in renormalizing exciton binding compared to longitudinal optical phonons. Our estimates of exciton binding energies are systematically in better agreement than those reported using standard approaches. 5 ; 6 An exciton is a quasiparticle in which an electron and hole interact electrostatically with each other creating a correlated, long-lived bound state. The exciton binding energy, the reversible work to separate the electron–hole pair, renormalizes the optical gap of a semiconductor, lowering it relative to expectations from the single-particle states forming the material’s band structure. For this reason, significant effort has been directed towards developing an accurate theoretical framework for this excited state problem. The favorable system size scaling of DFT makes it a good candidate for obtaining band energies in solids. 41 ; 62 However, DFT is fundamentally limited in its ability to capture many-body electronic correlations, as it cannot, by construction, describe two-particle excitations. 64 ; 72 The GW approximation provides a correction to the DFT band energies. 34 ; 36 The Bethe-Salpeter Equation (BSE) approach then builds upon this approximation by including screened electron–hole interactions, typically accounting for electronic screening in the random phase approximation and static limit. 68 ; 57 Nonetheless, the standard ab initio GW–BSE method tends to overestimate the magnitude of exciton binding energies in comparison with experiment. 5 ; 70 The main reason for this is that phonons can renormalize the electron-hole interaction. 22 ; 1 ; 61 ; 13 ; 12 The polarization generated from phonons can interact with the electron and hole, introducing a dynamical, retarded interaction that goes beyond static lattice screening. To account for the charge-phonon interaction in the GW–BSE framework, one can introduce a phonon-screened electron-hole kernel. The limitation of this approach is that the kernel is typically introduced perturbatively. 22 ; 1 ; 12 Path integral methods like PIMC have shown promise in capturing exciton–phonon interactions. 60 ; 58 ; 59 ; 67 ; 61 Consequently, computed exciton binding energies and radiative recombination rates from these methods tend to be in good agreement with experiment. However, PIMC is typically applied to simple model Hamiltonians parameterized empirically from experimental measurements of the dielectric constants and effective masses limiting their predictive capability. Here, we present an ab initio framework to parameterize an excitonic Hamiltonian complete with interactions with arbitrary phonon modes. We then exactly integrate out those phonon modes using an imaginary time influence functional that is sampled with PIMC. This allows us to capture non-perturbative static exciton properties as a function of temperature, which for strongly coupled systems like MgO is essential to achieve close agreement with experiment. The resultant zero temperature exciton binding energies validate this approach, and the temperature dependence of these energies reveals the role of thermal phonon screening on exciton stability, allowing us to identify conditions for the onset of free carrier generation. II Theory and Implementation We aim to model the energetics of excitons in bulk, three-dimensional semiconductors and how these energies are renormalized by interactions with phonons. To proceed, we consider a Hamiltonian decomposable as \mathcal{H}=\mathcal{H}_{\text{e}}+\mathcal{H}_{\text{ph}}+\mathcal{H}_{\text{int}} (1) where \mathcal{H}_{\text{e}} represents the Hamiltonian for the electronic degrees of freedom, \mathcal{H}_{\text{ph}} that for the phonon degrees of freedom, and \mathcal{H}_{\text{int}} the interaction between the two. Generically, the electronic Hamiltonian can be expressed as a sum over bands, \lambda , and wavevectors \mathbf{k} , as \mathcal{H}_{\text{e}}=\sum_{\mathbf{k},\lambda}E_{\lambda,\mathbf{k}}\hat{c}^{\dagger}_{\lambda,\mathbf{k}}\hat{c}_{\lambda,\mathbf{k}}+\\ \sum_{\mathbf{k},\mathbf{k}^{\prime},\Delta\mathbf{k},\lambda,\lambda^{\prime}}U_{\Delta\mathbf{k}}^{\lambda,\lambda^{\prime}}\hat{c}^{\dagger}_{\lambda,\mathbf{k}+\Delta\mathbf{k}}\hat{c}^{\dagger}_{\lambda^{\prime},\mathbf{k}^{\prime}-\Delta\mathbf{k}}\hat{c}_{\lambda^{\prime},\mathbf{k}^{\prime}}\hat{c}_{\lambda,\mathbf{k}} (2) where E_{\lambda,\mathbf{k}} is the energy of the \lambda ’th band at wavevector \mathbf{k} , U_{\Delta\mathbf{k}}^{\lambda,\lambda^{\prime}} is the interaction between electrons in different bands \lambda and \lambda^{\prime} with an exchange of momentum \Delta\mathbf{k} . The operators \hat{c}^{\dagger}_{\lambda,\mathbf{k}} and \hat{c}_{\lambda,\mathbf{k}} are creation and annihilation operators for electrons in the \lambda ’th band at wavevector \mathbf{k} , respectively. The basis of single particle plane wave states is appropriate for the class of semiconductors we consider here, which are weakly correlated in their ground state. While Eq. 2 is general, we will restrict our attention to band edge transitions such that \lambda and \lambda^{\prime} include only the valence and conduction bands, \lambda=\{\mathrm{c},\mathrm{v}\} . Further, we will consider low energy excitons such that E_{\lambda,\mathbf{k}} is well approximated by an effective mass model, E_{\lambda,\mathbf{k}}=E_{\lambda}+\frac{\hbar^{2}k^{2}}{2m_{\lambda}} (3) where m_{\lambda} is the effective mass of the valence or conduction band, \hbar is Planck’s reduced constant, and E_{\lambda} the energy of band \lambda at the zone center. Since we consider charge neutral systems, we identify the conduction band mass as an electron mass, m_{\mathrm{c}}=m_{e} , while that for the valence band as the mass of a hole, m_{\mathrm{v}}=-m_{h} . The interaction matrix element, U_{\mathbf{k}}^{\lambda,\lambda^{\prime}} , is assumed to be Coulombic, U_{\mathbf{k}}^{\lambda,\lambda^{\prime}}=\frac{e^{2}}{\epsilon_{0}\epsilon_{\infty}k^{2}}\left(1-\delta_{\lambda,\lambda^{\prime}}\right) (4) screened by a single frequency independent dielectric constant \epsilon_{\infty} , where e is the charge of the electron and \epsilon_{0} is the permittivity of free space. The Kronecker delta \delta_{\lambda,\lambda^{\prime}} ensures that the interactions are only across the band, between the electron and hole, and attractive. These two simplifications encode a Wannier–Mott exciton model. 40 ; 84 ; 54 In the absence of interactions with phonons, this model is exactly solvable yielding hydrogenic bound states. 40 We assume that the lattice is well approximated by a collection of phonons that couple linearly with the electronic degrees of freedom. This allows us to write the phonon Hamiltonian