结合局域范围分离的线性响应含时密度泛函理论：氖原子的芯价共振
Linear-response time-dependent density-functional theory with local range separation: Core and valence resonances of the neon atom
浏览论文内容
中文总结 AI 辅助
本研究针对氖原子光电离谱，评估双参数范围分离函数的局域范围分离杂化LRSH方法，发现其能合理描述价/芯共振能量，仅芯共振寿命因双激发问题被严重高估。
中文摘要 AI 辅助
我们针对氖原子光电离谱的线性响应含时密度泛函理论（TDDFT）Sternheimer计算，研究了范围分离杂化（RSHs）和局域范围分离杂化（LRSHs）。该体系对近似交换-关联处理构成严格测试，因为它同时呈现能量尺度差异极大的价共振和芯共振。基于此前采用简单单参数局域范围分离函数的工作，我们在此评估一种更灵活的双参数范围分离函数，旨在改善高密度极限。我们对比了RSHs和LRSHs得到的光电离谱及共振参数。研究发现，采用双参数范围分离函数的LRSH方法能对光电离谱给出整体令人满意的描述，包括价共振和芯共振的能量。由于2s→np价共振的衰变不涉及双激发，其寿命也得到了合理重现。相比之下，1s→np芯共振的寿命被高估了数个数量级，因为它们的俄歇衰变通道涉及绝热单行列式TDDFT中不存在的双激发。要在范围分离线性响应TDDFT框架内获得更准确的共振宽度，需采用多行列式方案和/或添加频率依赖的响应核。
英文摘要
We investigate range-separated hybrids (RSHs) and locally range-separated hybrids (LRSHs) for linear-response time-dependent density-functional theory (TDDFT) Sternheimer calculations of the photoionization spectrum of the Ne atom. This system constitutes a stringent test for approximate exchange-correlation treatments because it exhibits both valence and core resonances with very different energy scales. Building on previous work employing a simple one-parameter local range-separation function, we assess here a more flexible two-parameter range-separation function designed to improve the high-density limit. We compare photoionization spectra and resonance parameters obtained with RSHs and LRSHs. We find that the LRSH approach with the two-parameter range-separation function provides an overall satisfactory description of the photoionization spectrum, including energies for both valence and core resonances. The lifetimes of the 2s $\rightarrow$ np valence resonances are also reasonably reproduced since their decay does not involve double excitations. In contrast, the lifetimes of the 1s $\rightarrow$ np core resonances are overestimated by orders of magnitude because their Auger decay channels involve double excitations that are absent in adiabatic, single-determinant TDDFT. Obtaining more accurate resonance widths within linear-response range-separated TDDFT would require using multideterminant schemes and/or adding a frequency-dependent response kernel.
