arXiv:2610.13246
Theory of exciton-exciton interactions with nontrivial electronic bands
L. Maisel Licerán, H. T. C. Stoof
Abstract
Excitons are bound electron-hole pairs in semiconductors whose interactions influence optical properties and determine their collective behavior and emergent correlated phases. Despite the theoretical and experimental significance of these phenomena, a microscopic theory of exciton-exciton interactions that takes into account the full composite nature of these quasiparticles in systems with an arbitrary single-particle Hamiltonian is lacking. Here, we present a general theory of exciton-exciton interactions that builds upon our previous publication [Physical Review B 113, 155440 (2026)], which applied only to conventional semiconductors with weak band hybridization. By contrast, the theory presented here is valid in virtually any crystalline semiconducting system and can be applied to the study of biexciton formation and correlated excitonic phases in systems where band topology, quantum geometry, and moiré effects play a significant role. In addition to the analytical theory, we also present results in simple systems with flat bands with a nontrivial quantum metric and show that the biexcitonic spectrum differs strongly from that in the case without quantum geometry. Our theory paves the way for accurate quantitative calculations of biexciton spectra and excitonic many-body phases in modern condensed-matter systems.
AI Summary
The paper addresses microscopic exciton–exciton interactions in semiconductors with momentum-dependent electronic eigenstates, including band hybridization and quantum geometry. It extends an earlier conventional-semiconductor treatment by incorporating Bloch-state form factors into direct and exchange scattering amplitudes while retaining the composite fermionic structure of excitons. A variational two-exciton ansatz yields a generalized biexciton eigenproblem with a nontrivial overlap matrix enforcing electron and hole antisymmetry. Exchange symmetries reduce the required interaction terms; restricting the electron–hole basis permits an effective exciton–exciton potential to be extracted. Spin-degenerate systems admit a further decomposition into coupled-electron-spin sectors. An appendix extends the interaction vertex beyond flavor-conserving density interactions. Numerical examples use contact interactions in bilayer lattice models. A square-lattice model supports singlet biexcitons within the ground-exciton approximation, but none for hopping-to-attraction ratio $t/U\gtrsim0.9$; binding also vanishes in its geometrically trivial flat-band limit. Flat-band Lieb and two-orbital square-lattice models instead support both singlet and triplet biexcitons. Their binding varies with quantum geometry, and the even-parity triplet states are independent of on-site intralayer repulsion through Pauli exclusion. Calculations use momentum grids up to $50\times50$, with grid-convergence and eigensolver-residual checks. The examples retain only the lowest exciton mode and omit scattering states. This approximation is unsuitable near weak binding or small intermode gaps, including small Lieb-lattice imbalance and square-lattice geometry parameter $\zeta\gtrsim1.2$. Excited modes could introduce additional shallow triplet binding. Realistic Coulomb interactions, material-specific spectra, and correlated many-exciton phases are not calculated.
Error found: The claimed absence of interactions and biexcitons in the trivial t = 0 contact-interaction model is false without additional assumptions: an opposite-spin on-site biexciton lies 2(U − U_rep) below two separated excitons when U_rep < U. Restricting the exciton basis does not necessarily make the overlap matrix R invertible, as the paper's localized ground-exciton limit retains a Pauli-forbidden null vector in the triplet sector.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 19 pages, 8 figures