Quantum Solver Achieves Efficient Solution of Single-Impurity Anderson Models with Particle-Hole Symmetry

Key Takeaways

• VQE solves Anderson impurity model with shallow circuits
• Unified ansatz prepares ground and excited states efficiently
• Benchmarks show accurate DOS reconstruction versus classical methods
• L‑BFGS‑B outperforms COBYLA and Adam in convergence
• Demonstrated feasibility on noisy near‑term quantum devices

Summary

Researchers from Middle Tennessee State University and Oak Ridge National Laboratory introduced a quantum‑classical hybrid solver that uses the variational quantum eigensolver (VQE) to tackle the Anderson impurity model (AIM) within dynamical mean‑field theory (DMFT). The solver employs a unified shallow‑circuit ansatz to prepare both ground‑state and particle‑hole excitations, enabling reconstruction of the impurity Green’s function via a continued‑fraction expansion. Benchmarking against classical pipelines across multiple bath sizes and interaction strengths shows accurate density‑of‑states (DOS) results and comparable or faster convergence, even under realistic noise and shot‑limited conditions. Optimization comparisons (COBYLA, Adam, L‑BFGS‑B) and a quantum‑computed moment correction further improve fidelity, establishing practical benchmarks for future DMFT integrations.

Pulse Analysis

Dynamical mean‑field theory has become the workhorse for modeling strongly correlated electrons, yet its computational bottleneck remains the exact solution of the Anderson impurity model. Classical solvers scale poorly as the bath size grows, forcing researchers to rely on costly approximations that limit predictive power. The recent surge in quantum hardware, especially noisy intermediate‑scale quantum (NISQ) devices, sparked interest in quantum embedding techniques that could alleviate this scaling barrier.

The new hybrid solver leverages the variational quantum eigensolver to prepare the AIM ground state with a compact, unified ansatz capable of generating both particle and hole excitations. By applying a parameter‑shifted circuit and a continued‑fraction expansion, the team reconstructs the impurity Green’s function and extracts the density of states. Extensive benchmarks across varying interaction strengths and bath sizes demonstrate that the quantum approach matches or exceeds classical accuracy, even when simulated noise mimics current hardware limitations. Moreover, the study compares three optimization strategies—COBYLA, Adam, and L‑BFGS‑B—finding that gradient‑based L‑BFGS‑B, combined with a quantum‑computed moment correction, yields the fastest convergence.

The implications extend beyond academic curiosity. Embedding this VQE‑based impurity solver into full DMFT loops could dramatically reduce simulation times for materials where electron correlation drives functionality, such as high‑temperature superconductors and Mott insulators. As quantum processors mature, the shallow‑circuit design ensures scalability while keeping error rates manageable. Industry players focused on materials discovery stand to benefit from faster, more accurate predictions, accelerating the pipeline from theory to prototype. This work therefore marks a concrete step toward practical quantum‑enhanced computational materials science.