Systematic Investigation of the Reliability of the Frozen Nuclei Approximation for Short-Pulse Excitation: The Example of HCCI+

In this work we quantitatively study the reliability of the frozen nuclei approximation for ultrafast dynamics. Specifically we study laser excitation of HCCI⁺ from its ground state to the first electronically excited state. The population of the first excited state is obtained by both the frozen nuclei approximation and by multidimensional nuclear dynamics. Detailed comparison of the results by the two methods are performed to provide quantitative criteria for the reliability of the frozen nuclei approximation for this system.

1 Introduction

The rapid advances of ultrafast science and technology have made it possible to manipulate electron dynamics in molecular systems with ultrashort laser pulses. In particular, laser induced electron density redistribution such as charge transfer (Marcus, 1956; May and Kühn, 2008) and charge migration (Weinkauf et al., 1996, 1997; Cederbaum and Zobeley, 1999; Calegari et al., 2014; Kraus et al., 2015) have been extensively investigated. In general, charge migration prefers frozen nuclei or small amplitude nuclear motions, while charge transfer is typically accompanied by large amplitude nuclear motions. The research of charge transfer processes has a relatively long history. While ultrafast charge migration emerged as a hot topic during the past two decades (Weinkauf et al., 1996, 1997; Remacle et al., 1998; Cederbaum and Zobeley, 1999; Remacle and Levine, 1999; Barth and Manz, 2006; Kanno et al., 2006; Yudin et al., 2006; Remacle et al., 2007; Kanno et al., 2010; Mineo et al., 2012; Calegari et al., 2014; Kraus et al., 2015; Li et al., 2015; Yamaki et al., 2016; Wörner et al., 2017; Mineo et al., 2021). It should be noted that the first attosecond charge migration phenomenon was already introduced in 1944 (Eyring et al., 1944) and was largely forgotten during the next decades. Surveys of the literature on ultrafast charge migration can be found in Ref. (Jia et al., 2017a; Wörner et al., 2017). Below we summarize some typical features of ultrafast charge migration and its connection to the frozen nuclei approximation (FNA).

Ultrafast charge migration typically represents quantum dynamics of a coherent superposition of more than one electronic state. The typical time scale of ultrafast charge migration ranges from several hundred attoseconds to a few femtoseconds which makes the experimental observation (Kraus et al., 2015) rather difficult. For such a short time, the frozen nuclei approximation has been widely used for theoretical work of ultrafast charge migration. There are also several theoretical investigations which include the effects of nuclear motions (Bandrauk et al., 2009; Kanno et al., 2010; Ulusoy and Nest, 2012; Mendive-Tapia et al., 2013; Mineo et al., 2014; Despré et al., 2015; Mineo et al., 2021). The amplitude of charge migration can be significantly modulated by nuclear motions, in particular for relatively long-time dynamics (Mendive-Tapia et al., 2013; Mineo et al., 2014; Jia et al., 2019a,b). In general, the FNA is widely believed to be only valid for short time pulses, but there are no quantitative criteria for how short the pulses should be. This serves as the motivation for the present work: to seek quantitative criteria for the reliability of the FNA. Specifically, we will investigate short-pulse excitations of HCCI⁺ by systematically varying the laser parameters in a sufficiently wide region.

The choice of HCCI⁺ as our model of interest is based on the availability of experimental data (Heilbronner et al., 1971; Kraus et al., 2015) and theoretical techniques (Jia et al., 2019b). The combined experimental and theoretical reconstruction of attosecond charge migration has been reported for ultrafast ionization of HCCI (Kraus et al., 2015). Coherent superposition of the ground and first excited states has been created and analyzed. Subsequent theoretical investigations of ultrafast charge migration in HCCI⁺ (Jia et al., 2017b; Ding et al., 2017) related to the experimental observation (Kraus et al., 2015) exploit the FNA. In-depth investigations of simulations and manipulations of charge migration in HCCI⁺ including multidimensional nuclear dynamics have been reported recently (Jia et al., 2019a,b). However, no comparisons between the results of multidimensional nuclear dynamics and the ones of the FNA are available.

In the present work, we will investigate the reliability of the FNA by comparing the FNA and multidimensional nuclear dynamics. The remainder parts of the paper are organized as follows. Section 2 contains the model and methods for numerical calculations. Section 3 presents the results and discussion. The conclusions are drawn in Section 4.

2 Model and Methods

We focus on laser excitations of HCCI⁺ from its ground state. Full dimensional simulations of the system involve sets of electronic coordinates r = {r₁, r₂, … } and nuclear coordinates R = {R₁, R₂, … }. Here r_i and R_j are the spatial coordinates of the i-th electron and the j-th nucleus, respectively. It is convenient to use the Dirac notation for the electronic degrees of freedom. The total wavefunction of the system is thus

$\mathrm{\Psi }\left(\mathit{r},\mathit{R},t\right)=⟨\mathit{r}|\mathrm{\Psi }\left(\mathit{R},t\right)⟩.(1)$

Using the Born-Huang expansion (Born and Oppenheimer, 1927; Born and Huang, 1954), the total wavefunction can be expressed in terms of the electronic eigenstates |k(R)⟩ which are the solutions of the standard time-independent electronic schrödinger equation

$H_{\text{el}}\left(\mathit{R}\right)|k\left(\mathit{R}\right)\rangle =V_k\left(\mathit{R}\right)|k\left(\mathit{R}\right)\rangle .(2)$

The corresponding electronic eigenenergy V_k(R) is the k-th potential energy surface (PES). Here V_k(R) and ⟨r|k(R)⟩ are the same as the ones used in Refs. (Jia et al., 2019a,b) which are calculated by Molpro (Werner et al., 2012) using the state-averaged CASSCF(15,13) with cc-pVQZ basis set (cc-pVQZ-pp for iodine).

According to (Kraus et al., 2015; Jia et al., 2019a,b), nonadiabatic couplings between different electronic states |k(R)⟩ are negligible. The total Hamiltonian for HCCI⁺ in an external laser field E(t) can be approximated as

$H\left(t\right)=\sum _{k{k}^{\prime }}|k\left(\mathit{R}\right)\rangle \left[T\left(\mathit{R}\right){\delta }_{k{k}^{\prime }}+V_k\left(\mathit{R}\right){\delta }_{k{k}^{\prime }}-{\mathit{\mu }}_{k{k}^{\prime }}\left(\mathit{R}\right)\cdot \mathit{E}\left(t\right)\right]\langle k^{\prime }\left(\mathit{R}\right)|,(3)$

where T(R) is the nuclear kinetic energy and μ_kk′(R) = ⟨k(R)|μ|k′(R)⟩ is the transition (or permanent) dipole moment. The laser pulse has a Gaussian shape with maximum amplitude E_max and carrier frequency ω

$\begin{array}{c}\hfill \mathit{E}\left(t\right)={\mathit{e}}_z{E}_{\text{max}}s\left(t\right)\sin \left(\omega t\right),\\ \hfill s\left(t\right)=e^{-a{t}^2/T^2},\end{array}(4)$

where e_z is the direction of the electric field. For convenience, the electric field and the molecules are oriented along the z-axis. In the literature, there are different choices of the parameter a in Eq. 4. For the present work we set

$a=4\ln 2(5)$

for easier characterization of the pulse duration. We define the pulse duration as the full width at half maximum (FWHM) of s(t), which is just T in Eq. 4.

The quantum dynamics of the system can be simulated by the time-dependent schrödinger equation subject to initial condition at t = −∞

$\begin{array}{cc}\hfill i\hslash \frac{\mathrm{d}}{\mathrm{d}t}|\mathrm{\Psi }\left(\mathit{R},t\right)\rangle & =H|\mathrm{\Psi }\left(\mathit{R},t\right)\rangle ,\hfill \\ \hfill |\mathrm{\Psi }\left(\mathit{R},t=-\infty \right)\rangle & ={\chi }_{g,v=0}\left(\mathit{R}\right)|g\left(\mathit{R}\right)\rangle ,\hfill \end{array}(6)$

where χ_g,v=0(R) is the vibrational ground state wavefunction of the lowest potential energy surface V_g(R). For convenience we use k = g, e to represent the lowest and first excited electronic states, respectively. The wave packet is numerically propagated by means of the split operator method (Leforestier et al., 1991).

Subsequently, we can obtain the population of the electronic state |k(R)⟩ according to

$\begin{array}{cc}\hfill P_k\left(t\right)& =\int \langle \mathrm{\Psi }\left(\mathit{R},t\right)|k\left(\mathit{R}\right)\rangle \langle k\left(\mathit{R}\right)|\mathrm{\Psi }\left(\mathit{R},t\right)\rangle \mathrm{d}\mathit{R}\hfill \\ \hfill & \equiv \int {\chi }_k^{*}\left(\mathit{R},\mathit{t}\right){\chi }_k\left(\mathit{R},\mathit{t}\right)\mathrm{d}\mathit{R},\hfill \end{array}(7)$

where χ_k (R,t) = ⟨k(R)|Ψ(R, t)⟩ is the nuclear wave packet on the k-th PES V_k(R). It contains seven vibrational coordinates. According to Ref. (Jia et al., 2019a,b), one-dimensional (1D), three-dimensional (3D), and seven-dimensional (7D) calculations lead to essentially the same results. In the 3D calculations, the H-C, C-C and C-I bond lengths are explicitly taken into account and the four bending degrees of freedom are neglected. This kind of approximation is reasonable for linear molecules, such as HCCI⁺. In the present work we use the same 3D calculations for the nuclear wave packet χ_k (R,t) as in Ref. (Jia et al., 2019b). Then we mainly focus on the population of the first electronically excited state P_k(t) for k = e.

To check the reliability of the FNA, we further calculated the population of the first electronically excited state $P_e^{\text{FNA}}$ using the FNA. Accordingly, the molecular structure is fixed at the minimum of the lowest PES V_g(R). This structure is called equilibrium structure R_eq. The corresponding transition dipole moment is μ_eq ≡μ_ge (R = R_eq). The electronic wavefunction of the FNA is expanded as

$|{\mathrm{\Psi }}^{\text{FNA}}\left(t\right)\rangle =\sum _k{c}_k\left(t\right)|k\left(\mathit{R}={\mathit{R}}_{\text{eq}}\right)\rangle .(8)$

the time dependent coefficient c_k(t) can be obtained subject to the initial condition c_k(t = −∞) = δ_kg. The corresponding population is

$P_k^{\text{FNA}}\left(t\right)=|c_k\left(t\right){|}^2.(9)$

Throughout this work we fix the carrier frequency of the laser in Eq. 4 as ℏω = V_e (R_eq) − V_g (R_eq). We only focus on the final population at t = t_f when the laser pulse is off. This leads to the following analytical expression (Jia et al., 2017a)

$P_e^{\text{FNA}}\left(t=t_f\right)\approx {\sin }^2\left(\frac{\sqrt{\pi }{E}_{\text{max}}{\mu }_{\text{eq}}^{z}T}{\sqrt{a}\hslash }\right),(10)$

where ${\mu }_{\text{eq}}^z$ denotes the z component of the transition dipole at R_eq. The relative error of the FNA with respect to multi-dimensional nuclear dynamic is defined as

$\frac{\delta P_e}{P_e}=\frac{\left|P_e^{\text{FNA}}\left(t=t_f\right)-P_e\left(t=t_f\right)\right|}{P_e\left(t=t_f\right)}\times 100\%.(11)$

For all the subsequent numerical calculations we set t_f = 5T. However, this should not be wrongly interpreted as the FNA is valid even for t = 5T. We choose t_f = 5T just to make use of the property that the results presented in this work do not depend on different choices of t_f as long as t_f ≥ T.

3 Results and Discussion

The equilibrium structure R_eq of HCCI⁺ is linear with bond lengths R_HC = 1.06 Å, R_CC = 1.21 Å and R_CI = 1.95 Å. The corresponding vertical excitation energy from ground state |g(R_eq)⟩ to the first excited state |e(R_eq)⟩ is ℏω = 2.41 eV. For typical pulse durations, there are sufficient numbers of cycles in E(t) to make the electronic transition resonant. The corresponding transition dipole has only a z-component, which is ${\mu }_{\text{eq}}^z=$ 3.21 Debye. Subsequently we calculated the population of the first electronically excited state $P_e^{\text{FNA}}(t=t_f)$ according to Eq. 10 by the FNA. Convergence tests are performed for nuclear dynamics simulations such that the corresponding population of the first electronically excited state P_e(t) does not change subject to further increase of the grid-region or decrease of the spatial or temporal steps. We first analyze the dependence of the results on the pulse durations with the other parameters fixed. Specifically, the maximum amplitude of the electric field E_max is fixed at 2.0 × 10⁹ V/m. The detailed comparison between P_e(t = t_f) and $P_e^{\text{FNA}}(t=t_f)$ is shown in Figure 1 for T ≤ 20 fs.

FIGURE 1

FIGURE 1. Population of the first electronically excited state of HCCI⁺ versus the pulse duration T, at t = t_f when the laser pulse is switched off. The maximum-amplitude of the electric field E_max is 2 GV/m. (A) P_e (t_f) and $P_e^{\text{FNA}}(t_f)$ obtained by multidimensional nuclear dynamics (green) and by the frozen nuclei approximation (blue), respectively. (B) The relative error $\frac{\delta P_e}{P_e}$ of $P_e^{\text{FNA}}(t_f)$ with respect to P_e (t_f). Vertical lines show the positions of $\frac{\delta P_e}{P_e}=10\%$, 20%, and 40%, respectively.

As can be seen from Figure 1A, the deviation between $P_e^{\text{FNA}}(t_f)$ and P_e(t_f) gradually increases with the pulse duration T, for the region of short pulses. For T ≤ 20 fs, P_e(t) keeps increasing with T. However, $P_e^{\text{FNA}}(t)$ first increases and then decreases for T ≤ 20 fs. This kind of qualitative deviation will be further discussed below. According to Eq. 10 $P_e^{\text{FNA}}(t_f)$ will oscillate periodically with the pulse duration T. In Figure 1A, $P_E^{\text{FNA}}(t_f)$ reaches its maximum $P_{e,\mathrm{m}\mathrm{a}\mathrm{x}}^{\text{FNA}}=1$ for T = 15.45 fs. However, P_e(t_f) is still substantially below one even for T = 20 fs.

To quantitatively compare $P_e^{\text{FNA}}(t_f)$ and P_e(t_f), the relative error $\frac{\delta P_e}{p_e}$ defined in Eq. 11 is shown in Figure 1B. The relative error increases relatively slowly when the pulse duration T is smaller than 5 fs, and increases rapidly when T is larger than 5 fs. For long pulses, say T ≥ 15 fs, Figure 1B shows significant decrease of $\frac{\delta P_e}{p_e}$. However, this is pure coincidence. As can be identified from Figure 1A, the trend of $P_e^{\text{FNA}}(t_f)$ is already qualitatively wrong for T > 15.45 fs. Smaller relative error in this region does not imply better agreement between the frozen nuclei approximation and real physics. Consequently, we focus on short pulses for which the FNA is expected to be reasonable. Accordingly, we add three vertical lines in Figure 1B for relative errors of 40, 20, and 10%, respectively. The corresponding pulse durations with fixed value of E_max = 2.0×10⁹ V/m are T = 6.09, 4.65, and 1.97 fs, respectively.

Figure 2 shows the color-coded contour plots for the dependence of $P_e^{\text{FNA}}(t_f)$ and P_e (t_f) on the amplitude of the electric field E_max and the pulse duration T. The full set of the involved parameters span the region 0.5 GV/m ≤ E_max ≤ 4.0 GV/m and T ≤ 120 fs. The region of E_max more or less covers the reported amplitudes of lasers exploited in typical applications in the literature. The region of T reaches the first revival of charge migration in HCCI⁺ reported in Ref. (Jia et al., 2019b). As can be seen from Figure 2A, $P_e^{\text{FNA}}(t_f)$ oscillates between 1 and 0 periodically with E_max or T. Larger values of E_max or T corresponds to smaller oscillation period of $P_e^{\text{FNA}}(t_f)$, c.f., Eq. 10. Figure 2B is the same as Figure 2A except that $P_e^{\text{FNA}}(t_f)$ is replaced by P_e(t_f) which is obtained by performing 3D nuclear dynamics simulations. We can immediately identify that Figures 2A,B are qualitatively different for relatively long laser lulses (e.g., for T ≥ 20 fs). We therefore consider the results for the FNA are not meaningful for relatively long laser pulses. This is quite natural. Even if $P_e^{\text{FNA}}(t_f)$ is coincidentally very close to P_e(t_f) for T ≥ 20 fs, the nuclear wave packet is quite different from the initial one as has been reported in Ref. (Jia et al., 2019b). Accordingly, the basic assumption of the FNA breaks down for T ≥ 20 fs. We further plot Figure 2C which is the same with Figure 2B except that the population of the first electronically excited state is obtained by performing 1D nuclear dynamics simulations which explicitly treats the C-I stretch (Jia et al., 2019a). The results of 1D and 3D simulations agree well with each other, which confirms the findings in Refs. (Jia et al., 2019a,b).

FIGURE 2

FIGURE 2. Population of the first electronically excited state of HCCI⁺ versus E_max and T illustrated by color-coded contour plots. (A) $P_e^{\text{FNA}}(t_f)$ calculated by the FNA. (B) P_e (t_f) obtained by 3D nuclear dynamics simulations. (C) Same as (B) but the nuclear dynamics is one-dimensional.

To systematically study the reliability of the FNA for short pulses, the difference between $P_e^{\text{FNA}}(t_f)$ in Figure 2A and P_e (t_f) in Figure 2B is shown in Figure 3A, for the region T ≤ 20 fs when the FNA may be expected to work. The dependence of $P_e^{\text{FNA}}(t_f)-P_e(t_f)$ on E_max and T is shown by color-code contour plots. Results for which $|P_e^{\text{FNA}}(t_f)-P_e(t_f)|\ge 0.5$ are not resolved, since results with errors larger than 50 percent are in general not helpful. For any fixed value of E_max, the deviation between $P_e^{\text{FNA}}(t_f)$ and P_e (t_f) first increases then decreases with T after reaching a maximum. More complicated features can be found for relatively large E_max combined with relatively long T. However, as discussed above for Figure 1, the results of $P_e^{\text{FNA}}(t_f)$ in this complicated region in Figure 3A (specifically, after reaching maximum deviations) can agree better with P_e(t_f) by coincidence. In the following we only focus on the left bottom region of Figure 3A before the corresponding deviation $P_e^{\text{FNA}}(t_f)-P_e(t_f)$ reaches its maximum for any fixed value of E_max. As can be seen from Figure 3A, the deviations are rather small for T ≤ 5 fs for all the different values of E_max involved in the present work. This implies that we may roughly use T ≤ 5 fs as the criterion for the reliability of the FNA. The deviation not only increases with T for any given E_max, but also increases with E_max for any given T. According to Eq. 10, $P_e^{\text{FNA}}(t_f)$ increases with the product of E_max and T before reaching its maximum. From Figure 3A we can also find that the deviation $P_e^{\text{FNA}}(t_f)-P_e(t_f)$ essentially increases with the product of E_max and T before reaching its maximum. Better criteria for the reliability of the FNA can be obtained by analyzing the relative error of $P_e^{\text{FNA}}(t_f)$ with respect to P_e (t_f) shown in Figure 3B. In the same spirit, we only need to focus on the left region of Figure 3B indicated by the dashed curve in which the relative error $\frac{\delta P_e}{P_e}$ never exceeds 50%. For relatively long pulses, say T = 7 fs, the relative error of the FNA is already larger than 50% for any value of E_max shown in Figure 3B. In this case, the FNA is no longer reliable for T ≥ 7 fs.

FIGURE 3

FIGURE 3. (A) Color-coded contour plots for the deviation $P_e^{\text{FNA}}(t_f)-P_e(t_f)$ between $P_e^{\text{FNA}}(t_f)$ in Panel 2A and P_e (t_f) in Panel 2B for T ≤20 fs. (B) Color-coded contour plots for the relative error $\frac{\delta P_e}{P_e}$ of $P_e^{\text{FNA}}(t_f)$ with respect to P_e (t_f). The black dashed curve indicates $\frac{\delta P_e}{P_e}=50\%$.

In the short pulse region, the relative error $\frac{\delta P_e}{P_e}$ increases with the pulse duration T. To give quantitative criteria for the reliability of the FNA, we define certain characteristic pulse durations as follows:

$TN:\frac{\delta P_e}{P_e}\le \left(100-N\right)\%\mathrm{i}\mathrm{f}T\le TN.\left(N=\mathrm{60,80,90,95}\right)$

For example, if we want to use the FNA to obtain results with relative errors smaller than 5%, we need to set the pulse durations of the lasers to be smaller than T95. Similarly for the meanings of T90, T80, and T60. According to our model the characteristic pulse durations T95, T90, T80, and T60 only depend on one parameter E_max, which will be investigated subsequently.

The detailed dependence of T95, T90, T80, and T60 on the maximum amplitude of the electric field E_max is shown in Figure 4. A quite good property for the characteristic pulse durations is that T95, T90, and T80 almost do not depend on E_max. The corresponding values are T95 = 1.32 fs, T90 = 1.97 fs, and T80 = 4.65 fs respectively. The value of T60 increases with E_max extremely slightly from 6.09 to 6.14 fs. For relatively high standard criteria, say relative errors below 20%, the corresponding characteristic pulse durations are quite robust with respect to different amplitudes of lasers. This greatly simplifies the criteria for choosing proper lasers for applications of short pulse excitations of HCCI⁺. Essentially, we only need to care about the durations of the laser pulses with quantitative guidance derived from Figure 4 for the reliability of the FNA.

FIGURE 4

FIGURE 4. Characteristic pulse durations T95, T90, T80, and T60 for relative error $\frac{\delta P_e}{P_e}=$ 5, 10, 20, and 40%, respectively. See the text for more details.

The FNA only considers the electronic degrees of freedom and neglects the nuclear motions. Mathematically this corresponds to a large overlap of the time-dependent and the initial nuclear wave packets. The overlap can be estimated as the product of the corresponding overlap for each normal mode. The overlap for a normal mode may be approximated as $e^{-f^2(t)}$ for short time dynamics. Here f(t) is the displacement of the normal coordinate with respect to its initial value in units of the standard deviation of the initial wave packet for this normal mode. Typically, there will be only one or a few modes with f²(t) substantially above zero, which are called active modes. The overall overlap is thus mainly determined by the active modes. For the present case, there is only one active mode which is the C-I stretch with period 86 fs (Jia et al., 2019b). Due to the relatively large amplitude of the C-I stretch mode, the function $e^{-f^2(t)}$ quickly decreases to zero (Jia et al., 2019a,b). In this case, the duration of the pulse must be much shorter than a vibrational period to keep the overlap large enough. According to the results of Figure 4, all the characteristic pulse durations of HCCI⁺ are smaller than $\frac{1}{10}$ of a vibrational period. However, for some molecules with sufficiently small vibrational amplitudes for all the active modes, the overlap can be relatively large for a rather long time. For such cases, the effects of nuclear dynamics can be neglected for a longer time than just a few femtoseconds (Kanno et al., 2010; Ulusoy and Nest, 2012; Despré et al., 2015).

4 Conclusion

We have systematically investigated the population of the first electronically excited state of HCCI⁺ excited by different laser pulses. The amplitudes and durations of the laser pulses span a rather large domain for typical applications. The deviations between the results obtained by the frozen nuclei approximation and the ones obtained by multidimensional nuclear dynamics are calculated and analyzed in detail to check the reliability of the FNA. As expected the validity of the FNA can be admitted for sufficiently short laser pulses. Quantitative criteria for the reliability of the FNA are obtained. Specifically if we want to limit the relative errors of the FNA within 5% (or 10, or 20, or 40%), the durations of the laser pulses should be less than T95 = 1.3 fs (or T90 = 2.0 fs, or T80 = 4.7 fs, or T60 = 6.1 fs). For example, ultrafast charge migration in HCCI⁺ is reconstructed in Ref. (Kraus et al., 2015). for the first period of 1.85 fs. By extrapolation of our results, the error of the reported charge migration in HCCI⁺ for the first period is less than 10%. For short pulses with durations up to T60, the relative errors of the FNA are found to be almost independent of the amplitudes of the laser pulses. The results of the present work are expected to provide valuable guidance to future investigations of short pulse excitations of HCCI⁺.

Data Availability Statement

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

Author Contributions

YY proposed the conception and design of the study. DJ carried out all the quantum chemical calculations and the quantum dynamics simulations, and prepared all Figures. YY wrote the zero-order draft. All the authors contributed to the submitted version of the manuscript.

Funding

This work was supported by the National Key Research and Development Program of China (2017YFA0304203), the Program for Changjiang Scholars and Innovative Research Team (IR_17R70), the National Natural Science Foundation of China (11904215), the 111 project (Grant No. D18001), the Fund for “Shanxi 1331 Project,” and the Hundred Talent Program of Shanxi Province.

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher’s Note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors, and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Acknowledgments

We would like to express our gratitude to Professor Jörn Manz (Berlin) for stimulating discussions and careful reading of the manuscript.

References

Bandrauk, A. D., Chelkowski, S., Corkum, P. B., Manz, J., and Yudin, G. L. (2009). Attosecond Photoionization of a Coherent Superposition of Bound and Dissociative Molecular States: Effect of Nuclear Motion. J. Phys. B: Mol. Opt. Phys. 42, 134001. doi:10.1088/0953-4075/42/13/134001

CrossRef Full Text | Google Scholar

Barth, I., and Manz, J. (2006). Periodic Electron Circulation Induced by Circularly Polarized Laser Pulses: Quantum Model Simulations for Mg Porphyrin. Angew. Chem. Int. Ed. 45, 2962–2965. doi:10.1002/anie.200504147

CrossRef Full Text | Google Scholar

Born, M., and Huang, K. (1954). Dynamical Theory of crystal Lattices. Clarendon Press, Oxford.

Google Scholar

Born, M., and Oppenheimer, R. (1927). Zur Quantentheorie der Molekeln. Ann. Phys. 389, 457–484. doi:10.1002/andp.19273892002

CrossRef Full Text | Google Scholar

Calegari, F., Ayuso, D., Trabattoni, A., Belshaw, L., De Camillis, S., Anumula, S., et al. (2014). Ultrafast Electron Dynamics in Phenylalanine Initiated by Attosecond Pulses. Science 346, 336–339. doi:10.1126/science.1254061

PubMed Abstract | CrossRef Full Text | Google Scholar

Cederbaum, L. S., and Zobeley, J. (1999). Ultrafast Charge Migration by Electron Correlation. Chem. Phys. Lett. 307, 205–210. doi:10.1016/S0009-2614(99)00508-4

CrossRef Full Text | Google Scholar

Despré, V., Marciniak, A., Loriot, V., Galbraith, M. C. E., Rouzée, A., Vrakking, M. J. J., et al. (2015). Attosecond Hole Migration in Benzene Molecules Surviving Nuclear Motion. J. Phys. Chem. Lett. 6, 426–431. doi:10.1021/jz502493j

CrossRef Full Text | Google Scholar

Ding, H., Jia, D., Manz, J., and Yang, Y. (2017). Reconstruction of the Electronic Flux during Adiabatic Attosecond Charge Migration in HCCI+. Mol. Phys. 115, 1813–1825. doi:10.1080/00268976.2017.1287967

CrossRef Full Text | Google Scholar

Eyring, H., Walter, J., and Kimball, G. E. (1944). Quantum Chemistry. New York: Wiley.

Google Scholar

Heilbronner, E., Muszkat, K. A., and Schäublin, J. (1971). An Estimate of the Interatomic Distances in Monohaloacetylene Radical Cations from Photoelectron-Spectroscopic Data. Hca 54, 58–76. doi:10.1002/hlca.19710540107

CrossRef Full Text | Google Scholar

Jia, D., Manz, J., Paulus, B., Pohl, V., Tremblay, J. C., and Yang, Y. (2017a). Quantum Control of Electronic Fluxes during Adiabatic Attosecond Charge Migration in Degenerate Superposition States of Benzene. Chem. Phys. 482, 146–159. doi:10.1016/j.chemphys.2016.09.021

CrossRef Full Text | Google Scholar

Jia, D., Manz, J., and Yang, Y. (2019a). De- and Recoherence of Charge Migration in Ionized Iodoacetylene. J. Phys. Chem. Lett. 10, 4273–4277. doi:10.1021/acs.jpclett.9b01687

PubMed Abstract | CrossRef Full Text | Google Scholar

Jia, D., Manz, J., and Yang, Y. (2017b). Generation of Electronic Flux during the Femtosecond Laser Pulse Tailored to Induce Adiabatic Attosecond Charge Migration in. J. Mod. Opt. 64, 960–970. doi:10.1080/09500340.2016.1269216

CrossRef Full Text | Google Scholar

Jia, D., Manz, J., and Yang, Y. (2019b). Timing the Recoherences of Attosecond Electronic Charge Migration by Quantum Control of Femtosecond Nuclear Dynamics: A Case Study for HCCI+. J. Chem. Phys. 151, 244306. doi:10.1063/1.5134665

CrossRef Full Text | Google Scholar

Kanno, M., Kono, H., and Fujimura, Y. (2006). Control of π-Electron Rotation in Chiral Aromatic Molecules by Nonhelical Laser Pulses. Angew. Chem. Int. Ed. 45, 7995–7998. doi:10.1002/anie.200602479

CrossRef Full Text | Google Scholar

Kanno, M., Kono, H., Fujimura, Y., and Lin, S. H. (2010). Nonadiabatic Response Model of Laser-Induced Ultrafastπ-Electron Rotations in Chiral Aromatic Molecules. Phys. Rev. Lett. 104, 108302. doi:10.1103/physrevlett.104.108302

PubMed Abstract | CrossRef Full Text | Google Scholar

Kraus, P. M., Mignolet, B., Baykusheva, D., Rupenyan, A., Horný, L., Penka, E. F., et al. (2015). Measurement and Laser Control of Attosecond Charge Migration in Ionized Iodoacetylene. Science 350, 790–795. doi:10.1126/science.aab2160

PubMed Abstract | CrossRef Full Text | Google Scholar

Leforestier, C., Bisseling, R. H., Cerjan, C., Feit, M. D., Friesner, R., Guldberg, A., et al. (1991). A Comparison of Different Propagation Schemes for the Time Dependent Schrödinger Equation. J. Comput. Phys. 94, 59–80. doi:10.1016/0021-9991(91)90137-a

CrossRef Full Text | Google Scholar

Li, H., Mignolet, B., Wachter, G., Skruszewicz, S., Zherebtsov, S., Süßmann, F., et al. (2015). Coherent Electronic Wave Packet Motion inC60Controlled by the Waveform and Polarization of Few-Cycle Laser Fields. Phys. Rev. Lett. 114, 123004. doi:10.1103/PhysRevLett.114.123004

PubMed Abstract | CrossRef Full Text | Google Scholar

Marcus, R. A. (1956). On the Theory of Oxidation‐Reduction Reactions Involving Electron Transfer. I. J. Chem. Phys. 24, 966–978. doi:10.1063/1.1742723

CrossRef Full Text | Google Scholar

May, V., and Kühn, O. (2011). Charge and Energy Transfer Dynamics in Molecular Systems, Third, Revised and Enlarged Edition. Wiley-VCH, Weinheim.

Google Scholar

Mendive-Tapia, D., Vacher, M., Bearpark, M. J., and Robb, M. A. (2013). Coupled Electron-Nuclear Dynamics: Charge Migration and Charge Transfer Initiated Near a Conical Intersection. J. Chem. Phys. 139, 044110. doi:10.1063/1.4815914

CrossRef Full Text | Google Scholar

Mineo, H., Lin, S. H., and Fujimura, Y. (2014). Vibrational Effects on UV/Vis Laser-Driven π-electron Ring Currents in Aromatic Ring Molecules. Chem. Phys. 442, 103–110. doi:10.1016/j.chemphys.2014.02.011

CrossRef Full Text | Google Scholar

Mineo, H., Phan, N.-L., La, D.-K., and Fujimura, Y. (2021). Theoretical Study of Dynamic Stark-Induced π-Electron Rotations in Low-Symmetry Aromatic Ring Molecules beyond the Frozen Nuclear Approximation. J. Phys. Chem. A. 125, 1476–1489. doi:10.1021/acs.jpca.0c10216

CrossRef Full Text | Google Scholar

Mineo, H., Yamaki, M., Teranishi, Y., Hayashi, M., Lin, S. H., and Fujimura, Y. (2012). Quantum Switching of π-Electron Rotations in a Nonplanar Chiral Molecule by Using Linearly Polarized UV Laser Pulses. J. Am. Chem. Soc. 134, 14279–14282. doi:10.1021/ja3047848

CrossRef Full Text | Google Scholar

Remacle, F., and Levine, R. D. (1999). Charge Migration and Control of Site Selective Reactivity: The Role of Covalent and Ionic States. J. Chem. Phys. 110, 5089–5099. doi:10.1063/1.478406

CrossRef Full Text | Google Scholar

Remacle, F., Levine, R. D., and Ratner, M. A. (1998). Charge Directed Reactivity:. Chem. Phys. Lett. 285, 25–33. doi:10.1016/S0009-2614(97)01314-6

CrossRef Full Text | Google Scholar

Remacle, F., Nest, M., and Levine, R. D. (2007). Laser Steered Ultrafast Quantum Dynamics of Electrons in LiH. Phys. Rev. Lett. 99, 183902. doi:10.1103/PhysRevLett.99.183902

PubMed Abstract | CrossRef Full Text | Google Scholar

Ulusoy, I. S., and Nest, M. (2012). Remarks on the Validity of the Fixed Nuclei Approximation in Quantum Electron Dynamics. J. Phys. Chem. A. 116, 11107–11110. doi:10.1021/jp304140r

CrossRef Full Text | Google Scholar

Weinkauf, R., Schanen, P., Metsala, A., Schlag, E. W., Bürgle, M., and Kessler, H. (1996). Highly Efficient Charge Transfer in Peptide Cations in the Gas Phase: Threshold Effects and Mechanism. J. Phys. Chem. 100, 18567–18585. doi:10.1021/jp960926m

CrossRef Full Text | Google Scholar

Weinkauf, R., Schlag, E. W., Martinez, T. J., and Levine, R. D. (1997). Nonstationary Electronic States and Site-Selective Reactivity. J. Phys. Chem. A. 101, 7702–7710. doi:10.1021/jp9715742

CrossRef Full Text | Google Scholar

Werner, H.-J., Knowles, P. J., Knizia, G., Manby, F. R., Schütz, M., Celani, P., et al. (2012). Molpro, Version 2012.1, a Package of Ab Initio Programs. Available at http://www.molpro.net (accessed Feb 18, 2016)

Google Scholar

Wörner, H. J., Arrell, C. A., Banerji, N., Cannizzo, A., Chergui, M., Das, A. K., et al. (2017). Charge Migration and Charge Transfer in Molecular Systems. Struct. Dyn. 4, 061508. doi:10.1063/1.4996505

PubMed Abstract | CrossRef Full Text | Google Scholar

Yamaki, M., Mineo, H., Teranishi, Y., Lin, S. H., and Fujimura, Y. (2016). Quantum Control of Coherent ρρ‐Electron Dynamics in Chiral Aromatic Molecules. Jnl Chin. Chem. Soc 63, 87–92. doi:10.1002/jccs.201500043

CrossRef Full Text | Google Scholar

Yudin, G. L., Bandrauk, A. D., and Corkum, P. B. (2006). Chirped Attosecond Photoelectron Spectroscopy. Phys. Rev. Lett. 96, 063002. doi:10.1103/PhysRevLett.96.063002

PubMed Abstract | CrossRef Full Text | Google Scholar