Neufeld Ofer, Tancogne-Dejean Nicolas, Rubio Angel
Max Planck Institute for the Structure and Dynamics of Matter and Center for Free-Electron Laser Science, Hamburg 22761, Germany.
Center for Computational Quantum Physics (CCQ), The Flatiron Institute, New York, New York 10010, United States.
J Phys Chem Lett. 2024 Jul 18;15(28):7254-7264. doi: 10.1021/acs.jpclett.4c01383. Epub 2024 Jul 8.
In recent years, time-dependent density functional theory (TDDFT) has been extensively employed for highly nonlinear optics in molecules and solids, including high harmonic generation (HHG), photoemission, and more. TDDFT exhibits a relatively low numerical cost while still describing both light-matter and electron-electron interactions ab initio, making it highly appealing. However, the majority of implementations of the theory utilize the simplest possible approximations for the exchange-correlation (XC) functional-either the local density or generalized gradient approximations, which are traditionally considered to have rather poor chemical accuracy. We present the first systematic study of the XC functional effect on molecular HHG, testing various levels of theory. Our numerical results suggest justification for using simpler approximations for the XC functional, showing that hybrid and meta functionals (as well as Hartree-Fock) can, at times, lead to poor and unphysical results. The specific source of the failure in more elaborate functionals should be topic of future work, but we hypothesize that its origin might be connected to the adiabatic approximation of TDDFT.
近年来，含时密度泛函理论（TDDFT）已被广泛应用于分子和固体中的高非线性光学领域，包括高次谐波产生（HHG）、光发射等。TDDFT在从头算描述光与物质以及电子与电子相互作用的同时，展现出相对较低的数值计算成本，这使其极具吸引力。然而，该理论的大多数实现方式都采用了最简单的交换关联（XC）泛函近似——要么是局域密度近似，要么是广义梯度近似，传统上认为这些近似的化学精度相当差。我们首次对XC泛函对分子高次谐波产生的影响进行了系统研究，并测试了不同的理论水平。我们的数值结果表明，使用更简单的XC泛函近似是合理的，结果显示杂化泛函和元泛函（以及哈特里-福克方法）有时会导致较差的、不符合实际的结果。在更精细的泛函中出现这种失败的具体原因应是未来工作的主题，但我们推测其根源可能与TDDFT的绝热近似有关。
