Schlegel H Bernhard, Smith Stanley M, Li Xiaosong
Department of Chemistry, Wayne State University, Detroit, Michigan 48202, USA.
J Chem Phys. 2007 Jun 28;126(24):244110. doi: 10.1063/1.2743982.
Time-dependent Hartree-Fock (TD-HF) and time-dependent configuration interaction (TD-CI) methods with Gaussian basis sets have been compared in modeling the response of hydrogen molecule, butadiene, and hexatriene exposed to very short, intense laser pulses (760 nm, 3 cycles). After the electric field of the pulse returns to zero, the molecular dipole continues to oscillate due to the coherent superposition of excited states resulting from the nonadiabatic excitation caused by the pulse. The Fourier transform of this residual dipole gives a measure of the nonadiabatic excitation. For low fields, only the lowest excited states are populated, and TD-CI simulations using singly excited states with and without perturbative corrections for double excitations [TD-CIS(D) and TD-CIS, respectively] are generally in good agreement with the TD-HF simulations. At higher field strengths, higher states are populated and the methods begin to differ significantly if the coefficients of the excited states become larger than approximately 0.1. The response of individual excited states does not grow linearly with intensity because of excited state to excited state transitions. Beyond a threshold in the field strength, there is a rapid increase in the population of many higher excited states, possibly signaling an approach to ionization. However, without continuum functions, the present TD-HF and TD-CI calculations cannot model ionization directly. The TD-HF and TD-CIS simulations are in good accord because the excitation energies obtained by linear response TD-HF [also known as random phase approximation (RPA)] agree very well with those obtained from singly excited configuration interaction (CIS) calculations. Because CIS excitation energies with the perturbative doubles corrections [CIS(D)] are on average lower than the CIS excitation energies, the TD-CIS(D) response is generally stronger than TD-CIS.
已对采用高斯基组的含时哈特里 - 福克(TD - HF)方法和含时组态相互作用(TD - CI)方法进行了比较,以模拟氢气分子、丁二烯和己三烯在极短强激光脉冲(760 nm,3个周期)作用下的响应。脉冲电场归零后,由于脉冲引起的非绝热激发导致的激发态相干叠加,分子偶极矩继续振荡。该残余偶极矩的傅里叶变换给出了非绝热激发的量度。对于低场,仅填充最低激发态,使用单激发态且对双激发进行微扰校正[分别为TD - CIS(D)和TD - CIS]的TD - CI模拟通常与TD - HF模拟吻合良好。在较高场强下,更高的态被填充,如果激发态系数变得大于约0.1,则这些方法开始出现显著差异。由于激发态到激发态的跃迁,单个激发态的响应并非随强度线性增长。超过场强阈值后,许多更高激发态的布居迅速增加,这可能预示着接近电离。然而,由于没有连续函数,目前的TD - HF和TD - CI计算无法直接模拟电离。TD - HF和TD - CIS模拟吻合良好,因为通过线性响应TD - HF[也称为随机相位近似(RPA)]获得的激发能与从单激发组态相互作用(CIS)计算得到的激发能非常吻合。由于带有微扰双激发校正[CIS(D)]的CIS激发能平均低于CIS激发能,所以TD - CIS(D)响应通常比TD - CIS更强。