Understanding the modulation mechanism in resonance-enhanced multiphoton probing of molecular dynamics

Time-resolved spectroscopy on isolated molecules gives fundamental insight into the conversion of light energy to other degrees of freedom. Probing of the photoinduced dynamics can be accomplished by ionization, via a single-photon or multiphoton transition. In this Rapid Communication we directly contrast transient spectra on the molecule perylene obtained with multiphoton ionization (MPI) to single-photon ionization (SPI). The photoinduced nuclear geometry relaxation modulates the MPI transient with a decay time constant of $0.9\pm 0.2\mathrm{ps}$. In contrast, the SPI transient completely lacks any indication for relaxation. We attribute this difference to a change in resonance enhancement of the MPI probe as the molecular geometry changes. Our results underline the importance of a detailed knowledge about these resonances for a proper interpretation of transient signals of molecular dynamics subject to nuclear and electronic relaxation effects. At the same time, the direct comparison to SPI directly demonstrates the higher sensitivity of resonance-enhanced MPI as a probe in time-resolved dynamical studies.

Figure 1

Sketch of the electronic states of perylene and excitation-ionization schemes. Shown are the ground state $S_0$, the first excited state $S_1$, resonant intermediate states, and the lowest ionic state ${D_0}^{+}$ for $S_0$ and $S_1$ minimum geometries. After 400-nm excitation the molecule is either ionized by SPI with a single 89-nm VUV photon or via 800-nm MPI. During vibrational relaxation in the excited state the ionization potential ($S_1\text{−}{D_0}^{+}$ distance) remains nearly constant, while the energy distance to states that lead to resonance enhancement of the MPI probe increases.

Figure 2

Comparison of H9 (89-nm) SPI probing with 800-nm MPI probing for perylene pump-probe photoionization. (a) PE spectra obtained with the 400-nm and H9 pump-probe spectrum (gray), H9 and 400-nm probe-pump signal (black), 400-nm pump pulse only (red), and H9 probe only (blue). Traces are additionally shown with a magnification of 10 for better visibility. (b) Difference spectrum obtained as the pump-probe spectrum minus the probe-pump signal (black) together with a Gaussian fit of the PE peak at 9.8 eV (red). (c) Position of the 9.8-eV PE peak as obtained from the Gaussian fit as a function of the pump-probe time delay. A fit with the product of an error function and an exponential term with variable amplitude (red curve) [31] (see also [32], Table 2) reveals a negligible PE signal shift of $-0.2\pm 0.5\%$. The red dashed lines mark the 95% confidence interval of the exponential term amplitude. (d) Same as in (c) for the PE peak amplitude. The fit gives a $15\pm 10\%$ increase with a 1.2-ps time constant. (e) Transient PI signals for 800-nm MPI probing observed at the perylene parent ion mass at 252 u (red curve) and for the total ion yield (blue curve, vertically offset by 0.5). Fits to the same function as in (c) and (d) reveal decays of $32\pm 3\%\left(252\mathrm{u}\right)$ and $34\pm 3\%$ (total ion yield), both with a $0.9\pm 0.2\mathrm{ps}$ time constant (see [32], Table 2).

Figure 3

(a) Laser intensity scans for 800-nm (red) (252-u detection and slope equal to $3.60\pm 0.18$), as well as 400- and 800-nm pump-probe MPI (blue) (252-u detection, time delay of 0.3 ps, and slope equal to $1.67\pm 0.13$). Due to saturation of the process, the data points (crosses) for the highest laser intensities deviate from the expected linear dependence. Only encircled data points are included in the fit. (b) Slopes for 400- and 800-nm pump-probe MPI at different time delays.