Variational adiabatic eigensolver for band structure excited states
Addressing the critical challenge of excited-state preparation in semiconductor band structure calculations, this work presents a variational adiabatic eigensolver (VAE) protocol that combines adiabatic evolution with variational optimization to prepare high-fidelity eigenstates on noisy intermediate-scale quantum (NISQ) devices. Our approach utilizes multi-qubit parameterized circuits and a phase-based loss function to circumvent the limitations of conventional methods, such as the difficulty of constructing quantum circuits in traditional adiabatic methods and the reduced accuracy of variational quantum eigensolvers for excited states. Through rigorous numerical simulations and experimental implementation on a superconducting quantum processor, we successfully prepare the valence-band and conduction-band eigenstates of silicon. Experimental results demonstrate state fidelities exceeding 96% in single-shot readout, while energy expectations agree with theoretical band energies within 0.5 eV. Subsequent refinement through single-frequency oscillation fitting further reduces the energy deviation to below 0.01 eV. This framework offers a valuable pathway for accurate electronic structure determinations in quantum materials.
Authors
- Yicong Zheng (ORCID: https://orcid.org/0000-0003-4925-6069)
- Zhiwen Zong (ORCID: https://orcid.org/0009-0009-9456-8089)
- Tianqi Cai (ORCID: https://orcid.org/0000-0003-2951-0744)
- Liyang Sui
- Wenyan Jin
- Zhihao Gong (ORCID: https://orcid.org/0000-0002-9643-633X)
- Shengyu Zhang (ORCID: https://orcid.org/0000-0001-5907-2277)
- Yufan Li (ORCID: https://orcid.org/0000-0003-3595-9089)
- Kunliang Bu (ORCID: https://orcid.org/0009-0003-5359-9598)
- Xutao Zhang
- Xingrui Liu
- Yi Yin
- Bowen Chen
- Jianlan Wu
Publication Details
- Journal
- Physical Review Applied
- Published
- 2026-11-17
- DOI
- https://doi.org/10.1103/zyln-zqn7
- Primary Topic
- Quantum Computing Algorithms and Architecture
- Type
- article
- Field-Weighted Citation Impact
- 0.00