构建模型哈密顿量中的轨道选择
Orbital choice in constructing model Hamiltonians
- Pennsylvania State University（宾夕法尼亚州立大学）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
本文研究模型哈密顿量构建中局域化轨道选择的影响，发现不同化学合理的轨道选择即使能量相近也会导致参数显著不同，并利用密度矩阵降维系统分析其效应。
AI中文摘要：
模型哈密顿量（如哈伯德或海森堡模型）通过少量参数（如电子跳跃积分或磁交换耦合）为复杂电子现象提供了简单而物理上有意义的描述。其构建通常利用电子关联效应的局域性质，因此依赖于局域化轨道。尽管有许多局域化方案可用，但所得局域化轨道不仅取决于所选的局域化泛函，还取决于局域化过程所应用的轨道空间。这些选择常被视为技术细节，很少被明确讨论。在此，我们证明，不同的、化学上合理的局域化轨道选择可导致模型哈密顿量参数显著不同，即使它们产生几乎相同的电子能量。利用密度矩阵降维（Density Matrix Downfolding），我们系统研究了局域化泛函以及更重要的模型轨道空间选择如何影响从代表性π共轭体系和过渡金属配合物的从头算计算中导出的有效哈密顿量。
英文摘要：
Model Hamiltonians (e.g., the Hubbard or Heisenberg models) provide a simple yet physically meaningful description of complex electronic phenomena through a small number of parameters, such as electron hopping integrals or magnetic exchange couplings. Their construction typically exploits the local nature of electron correlation effects and therefore relies on localized orbitals. Although many localization schemes are available, the resulting localized orbitals depend not only on the chosen localization functional but also on the orbital space to which the localization procedure is applied. These choices are often treated as technical details and are rarely discussed explicitly. Here, we demonstrate that different, chemically reasonable choices of localized orbitals can lead to substantially different model Hamiltonian parameters, even when they produce nearly identical electronic energies. Using Density Matrix Downfolding, we systematically investigate how the localization functional and, more importantly, the choice of model orbital space affect effective Hamiltonians derived from ab initio calculations for representative $π$-conjugated systems and transition metal complexes.
