量子-经典碎片化与有效碎片分子轨道方法
Quantum-Classical Fragmentation with the Effective Fragment Molecular Orbital Method
浏览论文内容
中文总结 AI 辅助
提出Q-EFMO混合方法，结合VQE与EFMO框架计算分子碎片相关能，通过轨道缩减将量子资源需求降至碎片规模，在LiH基准上达到接近CCSD精度，并显著降低计算成本。
中文摘要 AI 辅助
我们提出了量子有效碎片分子轨道（Q-EFMO）方法，这是一种混合框架，利用变分量子本征求解器评估选定分子碎片的相关能，并通过尺寸一致的EFMO能量表达式进行组装。Hartree-Fock EFMO计算提供单体参考、多体极化以及远距离对EFP相互作用；独立的VQE/UCCSD计算为单体和选定的近场二聚体提供相关增量。因此，最大量子寄存器由最大的活性碎片或碎片对决定，而非由总系统大小决定。对于STO-3G基组下的三层LiH基准，完整的2.5至3.5埃分离度扫描显示，两种轨道缩减方案单调逼近全CCSD参考值。更激进的方案在分离度2.6埃及以上时误差保持在1 kcal/mol以下，在3.5埃时达到0.10 kcal/mol，单体使用6个量子比特，最大二聚体使用12个量子比特，而冻结核二聚体基线需要14个量子比特。实现特定的成本代理从1000降至40，降低了25倍。这些无噪声结果确立了方程、软件路径和资源扩展性；更广泛的化学验证、更大基组和硬件测试仍有必要。
英文摘要
We present the quantum effective fragment molecular orbital (Q-EFMO) method, a hybrid framework that evaluates selected molecular-fragment correlation energies with the variational quantum eigensolver and assembles them through the size-consistent EFMO energy expression. A Hartree-Fock EFMO calculation supplies monomer references, many-body polarization, and distant-pair EFP interactions; independent VQE/UCCSD calculations provide correlation increments for monomers and selected near-field dimers. Consequently, the maximum quantum register is determined by the largest active fragment or fragment pair rather than by total system size. For a three-layer LiH benchmark in STO-3G, a full 2.5 to 3.5 Angstrom separation scan shows that two orbital-reduction schemes approach the full CCSD reference monotonically. The more aggressive scheme remains below 1 kcal/mol for separations of 2.6 Angstrom and greater and reaches 0.10 kcal/mol at 3.5 Angstrom, using 6 qubits for monomers and 12 for the largest dimers, compared with 14 qubits for the frozen-core dimer baseline. An implementation-specific cost proxy decreases from 1000 to 40, a factor of 25. These noise-free results establish the equations, software path, and resource scaling; broader chemical validation, larger bases, and hardware tests remain necessary.
