This research presents an experimental implementation of the Contextual Subspace Variational Quantum Eigensolver (CSVQE) using superconducting quantum hardware. The study focuses on calculating the potential energy curve of molecular nitrogen (N₂), a challenging problem in quantum chemistry due to strong static correlation effects during molecular dissociation.
The experiment demonstrates that their quantum approach achieves better accuracy than traditional single-reference wavefunction methods when modeling bond-breaking in N₂. Their methodology proves competitive with multiconfigurational approaches while requiring fewer quantum resources, enabling the treatment of larger active spaces with a fixed number of qubits.
To achieve these results, the researchers implemented a comprehensive error management strategy combining three key techniques: Dynamical Decoupling, Measurement-Error Mitigation, and Zero-Noise Extrapolation. They also enhanced their approach through circuit parallelization, which provides passive noise-averaging and improves the effective shot yield, thereby reducing measurement overhead. A notable innovation is their modified adaptive ansatz construction algorithm, which optimizes variational circuits for specific qubit topologies, minimizing transpilation costs.
The work positions itself within the broader context of quantum chemistry applications in quantum computing, a field that has seen two decades of development. While current noisy intermediate-scale quantum (NISQ) devices limit demonstrations to small molecules and basis sets, these implementations serve as important benchmarks for both algorithm practicality and hardware capability. The researchers acknowledge that current quantum simulations cannot yet challenge classical computers, as evidenced by previous VQE implementations detailed in their supplementary information.
The paper addresses the concept of quantum advantage in chemistry applications, noting the challenge of finding problems where quantum algorithms can definitively outperform all classical approaches. They discuss the conventional benchmark of “chemical accuracy” (43 meV error threshold), though they point out this terminology can be misleading when working with small basis sets. Instead, they suggest using “algorithmic accuracy” as a more appropriate term.
Their approach employs a hybrid quantum-classical method, where part of the calculation is performed classically while a quantum correction is computed on the NISQ device. This strategy reduces quantum resource requirements while ensuring the quantum portion of the calculation remains fundamentally quantum mechanical, as verified through contextuality analysis.
The choice of molecular nitrogen as a test case is significant, as N₂ serves as a standard benchmark problem in quantum chemistry, particularly challenging during bond-breaking calculations. This selection allows for direct comparison with classical methods and demonstrates the potential advantages of their quantum approach in handling strongly correlated systems.
The research represents a step forward in practical quantum chemistry calculations on NISQ devices, though still operating within current hardware limitations. Their innovations in error mitigation and circuit optimization, combined with the hybrid quantum-classical approach, suggest a pathway toward handling larger chemical systems as quantum hardware continues to improve. The work maintains scientific rigor while addressing practical implementation challenges, contributing to the ongoing development of quantum computing applications in chemistry.
npj Quantum Information, Published online: 12 February 2025; doi:10.1038/s41534-024-00952-4
