超越密度泛函理论的共振拉曼光谱学
Resonant Raman spectroscopies beyond density-functional theory
浏览论文内容
中文总结 AI 辅助
本文提出通用有限差分框架，可计算任意电子结构方法的共振拉曼张量，将其应用于石墨烯与单层MoS₂发现杂化泛函等方法的电子-声子耦合强于半局域DFT，且杂化泛函与实验一致性最佳，为二维材料拉曼表征提供新途径。
中文摘要 AI 辅助
共振拉曼光谱可在单次测量中探测材料内电子与声子的耦合情况。密度泛函理论（DFT）通常能较好地重现声子频率，但共振拉曼强度取决于电子-声子矩阵元与电子跃迁，对基础的交换-关联近似更为敏感。然而，电子-声子耦合迄今仅能通过针对少数半局域DFT方法开发的线性响应理论获取，共振拉曼强度对电子结构近似的敏感性基本未被探索。本文提出一种通用有限差分框架，可计算任何能提供原始及位移构型的力、本征值和波函数的电子结构方法的共振拉曼张量。将该形式应用于石墨烯与单层MoS₂，采用杂化泛函或meta-GGA，结果显示这些方法相比半局域DFT系统性增强了电子-声子耦合，反映出介电过屏蔽的减弱。拉曼张量分解表明，准确的强度要求电子本征值与电子-声子矩阵元在同一理论水平上一致处理。在测试的方法中，杂化泛函与实验的整体一致性最佳。由于该框架仅需各电子结构代码已产生的量，它为超越DFT的系统拉曼表征或针对实验的基准测试打开了大门，尤其适用于二维材料。
英文摘要
Resonant Raman spectroscopy probes, in a single measurement, how electrons and phonons couple in a material. Density-functional theory (DFT) typically reproduces well phonon frequencies, but resonant Raman intensities hinge on electron-phonon matrix elements and electronic transitions that are far more sensitive to the underlying exchange-correlation approximation. However, electron-phonon coupling has so far been accessible only through linear-response theories developed for a handful of semilocal DFT methods, leaving the sensitivity of resonant Raman intensities to the electronic-structure approximation essentially unexplored. Here, we introduce a general finite-difference framework that can compute resonant Raman tensors for any electronic-structure method capable of delivering forces, eigenvalues, and wavefunctions of pristine and displaced configurations. We apply the formalism to graphene and monolayer MoS$_2$, using hybrid functionals or meta-GGAs, and show that these approaches systematically enhance electron-phonon couplings relative to semilocal DFT, reflecting reduced dielectric overscreening. A decomposition of the Raman tensor shows that accurate intensities require electronic eigenvalues and electron-phonon matrix elements to be treated consistently at the same level of theory. Among the approaches tested, hybrid functionals provide the best overall agreement with experiment. Because the framework needs only quantities every electronic-structure code already produces, it opens the door to systematic, beyond-DFT Raman characterization or benchmarking against experiments, especially for 2D materials.
