Time-resolved inner-shell photoelectron spectroscopy: From a bound molecule to an isolated atom

Due to its element and site specificity, inner-shell photoelectron spectroscopy is a widely used technique to probe the chemical structure of matter. Here, we show that time-resolved inner-shell photoelectron spectroscopy can be employed to observe ultrafast chemical reactions and the electronic response to the nuclear motion with high sensitivity. The ultraviolet dissociation of iodomethane (${\mathrm{CH}}_3\mathrm{I}$) is investigated by ionization above the iodine $4d$ edge, using time-resolved inner-shell photoelectron and photoion spectroscopy. The dynamics observed in the photoelectron spectra appear earlier and are faster than those seen in the iodine fragments. The experimental results are interpreted using crystal-field and spin-orbit configuration interaction calculations, and demonstrate that time-resolved inner-shell photoelectron spectroscopy is a powerful tool to directly track ultrafast structural and electronic transformations in gas-phase molecules.

Figure 1

Sketch of the experimental setup. The 272-nm laser beam and the 11.6-nm FEL beam are collinearly overlapped and focused inside a target sample of iodomethane molecules at the center of a double-sided velocity map imaging spectrometer. Ion and electron momentum distributions are recorded at opposite ends of the spectrometer. The ion detector side is equipped with the PImMS2 camera that allows the arrival time and position of all ions to be recorded simultaneously, while the electron detector incorporates a MCP/phosphor screen assembly followed by a CCD camera. P: prism; CW beam: continuous molecular beam; DL: drilled mirror; BBO: beta barium borate crystal; WP: wave plate.

Figure 2

Potential energy curves of the ground and selected neutral excited states of ${\mathrm{CH}}_3\mathrm{I}$ as a function of C-I distance. The remaining structural parameters [R(C-H), (H-C-I)] are optimized for the ground ($X{}^{1}A_1$) state. The symbols represent selected internuclear distances for which the calculation was performed.

Figure 3

Electronic structure of the ${\mathrm{CH}}_3{\mathrm{I}}^{+}$ molecular cation as a function of dissociation coordinate [(a) and (b)] or time after the initial excitation [(c) and (d)]. Lines connect final cation states at each C-I distance ordered by energy, and do not imply continuity of electronic character of the state. Some of the final states may be inaccessible from the chosen initial state due to the selection rules, which were not taken into account. (a), (b) Calculated electron removal energy from the valence and I $4d$ shells, respectively, as a function of distance. (c), (d) $4d$ electron removal energies as a function of time after initial excitation. (a), (c) The initial state is ${}^{3}Q_1$. (b), (d) The initial state is ${}^{3}Q_{0^{+}}$.

Figure 4

Ion TOF spectrum of ${\mathrm{CH}}_3\mathrm{I}$ following ionization by an 11.6-nm FEL pulse. ${\mathrm{I}}^{n+}$ and ${{\mathrm{CH}}_{\mathrm{x}}}^{+}$ fragments are observed. We note the presence of “ghost peaks,” labeled as GP, that are observed when the ion drift tube voltage was higher than the front voltage on the MCP detector.

Figure 5

Time-dependent ${\mathrm{I}}^{n+}$ ion momentum distributions recorded in ${\mathrm{CH}}_3\mathrm{I}$ (a) as a function of the UV pump-FEL probe delay and (b) the corresponding kinetic energy spectra for two delays: (dashed blue line) $\tau =-1$ ps, i.e., FEL pulse comes first; (orange solid line) $\tau =+1$ ps, i.e., UV pulse comes first. The blue and orange areas emphasize the increase and the depletion, respectively, of the signal when the UV pulse arrives before the FEL pulse. Channel $\mathrm{A}$ (dotted blue lines): ionization of the wave packet that propagates on the dissociative $Q$-state manifold of ${\mathrm{CH}}_3\mathrm{I}$; channel $\mathrm{B}$ (dashed-dotted red lines): Coulomb explosion of the ground-state molecules.

Figure 6

Normalized integrated yield of the low-energy channel (integrated between 0 and 0.4 eV) in the multiply charged iodine ions plotted as a function of the delay between the UV pump and FEL probe pulses for several charge states of iodine (open circles), and the corresponding fit using a Gaussian cumulative distribution function (line). The centers of the fitted functions are indicated in parentheses, together with the standard deviations retrieved from the fits.

Figure 7

Comparison between the critical internuclear distances predicted using Eq. (3) (orange line) and the internuclear distances obtained from adiabatic propagation of the wave packet on the ${}^{3}Q_{0^{+}}$ potential energy curve using the reaction times measured experimentally. The reaction times are used without including an additional time offset (full squares) and with an additional time offset (open squares).

Figure 8

Projected 2D electron momentum distributions (left half) and slices through the 3D photoelectron momentum distribution obtained after Abel inversion (right half) following ionization of ${\mathrm{CH}}_3\mathrm{I}$ (a) and ${\mathrm{CH}}_2\mathrm{ICl}$ (b) at a photon energy of 107 eV. (c) Corresponding electron spectra displayed as a function of the binding energy.

Figure 9

Comparison between the experimental photoelectron spectrum measured in the ${\mathrm{CH}}_3\mathrm{I}$ molecule following irradiation by the 11.6-nm FEL pulse (open blue circle), and a convolution of the reported $4d$ photoline components (${\mathrm{\Sigma }}_{1/2},{\mathrm{\Delta }}_{5/2},{\mathrm{\Pi }}_{3/2},{\mathrm{\Sigma }}_{3/2},{\mathrm{\Pi }}_{1/2},\nu =1$) with the experimental resolution (blue solid line). The experimental resolution is a sum of the instrumental function of the VMI spectrometer and the FEL bandwidth (2.2 eV combined). The peak positions and Lorentzian widths are taken from Cutler et al. [43] and are displayed in the figure. A comparison with a fit using a single Gaussian function is also included (dashed orange line).

Figure 10

(a) Time-dependent difference photoelectron spectra (see text) recorded in ${\mathrm{CH}}_3\mathrm{I}$ as a function of the pump-probe delay. (b) Time-average difference spectrum calculated for positive delays near the iodine $4d$ line. Gaussian fits to the molecular and atomic components of the I $4d$ photoelectron peak, shown for time delays of $-1$ ps (c) and $+1$ ps (d). The symbols in (e) and (f) show the time evolution of the intensity of the two Gaussians used to fit the depletion of the molecular $4d$ iodine contribution (e) and the rise of the atomic iodine $4d$ contribution (f). The lines are obtained from a fit using a cumulative Gaussian distribution function. The fit parameters are summarized in Table 2 along with the parameters obtained from the ion data.

Figure 11

(Upper panel) Calculated evolution of the crystal-field binding energy of the I $4d_{5/2}$ and $4d_{3/2}$ states along the C-I bond coordinate. $J$ and $J_z$ are the total angular momentum and the component of the total angular momentum along the principal molecular axis, respectively. (Lower panel) Time evolution of the binding energy of the I $4d_{5/2}$ and $4d_{3/2}$ states, along the ${}^{3}Q_{0^{+}}$ dissociation pathway, treated classically. Similar results are observed for the ${}^{3}Q_1$ state. Absolute binding energies are overestimated due to the incomplete treatment of the electron-shell relaxation.