Quantum Simulations of Chemistry in First Quantization with any Basis Set

• Published in: arXiv.org 2024-08
• Main Authors: Georges, Timothy N, Bothe, Marius, Sünderhauf, Christoph, Berntson, Bjorn K, Izsák, Róbert, Ivanov, Aleksei V
• Format: Article
• Language: English
• Subjects: Algorithms; Asymptotic methods; Chemistry; Fault tolerance; Gate counting; Hamiltonian functions; Molecular orbitals; Plane waves; Pseudopotentials; Quantum chemistry; Quantum computers; Quantum computing; Quantum phenomena
• ISSN: 2331-8422

Quantum computation of the energy of molecules and materials is one of the most promising applications of fault-tolerant quantum computers. However, practical applications require algorithms with reduced resource requirements. Previous work has mainly represented the Hamiltonian of the system in second quantization. Existing methods in first quantization are limited to grid-based approaches that do not allow for active space calculations. In this work, we present a method to solve the generic ground-state chemistry problem in first quantization on a fault-tolerant quantum computer using any basis set. This allows for calculations in the active space using modern quantum chemistry basis sets. We derive a linear-combination-of-unitaries decomposition for a chemical Hamiltonian in first quantization and then construct an efficient block encoding, exploiting sparsity of the Hamiltonian. For active space calculations using a molecular orbital basis set, we achieve an asymptotic speed up in Toffoli-gate count compared to the equivalent method in second quantization [Berry, et. al. Quantum 3, 208 (2019)]. We also consider the dual plane waves for materials simulations and find that in physically interesting regimes we achieve orders of magnitude improvement in quantum resources compared to the second quantization counterpart. In some instances, our approach provides similar or even lower resources compared to the first quantization plane wave algorithm of Refs.[Babbush, et. al npj Quantum Inf 5(1) 92 (2019), Su et. al PRX Quantum 2(4), 040332 (2021)] that, unlike our approach, avoids loading the classical data from quantum memory. This work opens up possibilities to reduce quantum resources even further using factorization methods of a Hamiltonian or modern pseudopotentials. Furthermore, our approach can be adapted to other applications, such as the vibrational properties of chemical systems.