Mapping the Dissociative Ionization Dynamics of Molecular Nitrogen with Attosecond Time Resolution

Studying the interaction of molecular nitrogen with extreme ultraviolet (XUV) radiation is of prime importance to understand radiation-induced processes occurring in Earth’s upper atmosphere. In particular, photoinduced dissociation dynamics involving excited states of ${{\mathrm{N}}_2}^{+}$ leads to N and ${\mathrm{N}}^{+}$ atomic species that are relevant in atmospheric photochemical processes. However, tracking the relaxation dynamics of highly excited states of ${{\mathrm{N}}_2}^{+}$ is difficult to achieve, and its theoretical modeling is notoriously complex. Here, we report on an experimental and theoretical investigation of the dissociation dynamics of ${{\mathrm{N}}_2}^{+}$ induced by isolated attosecond XUV pulses in combination with few-optical-cycle near-infrared/visible (NIR/VIS) pulses. The momentum distribution of the produced ${\mathrm{N}}^{+}$ fragments is measured as a function of pump-probe delay with subfemtosecond resolution using a velocity map imaging spectrometer. The time-dependent measurements reveal the presence of NIR/VIS-induced transitions between ${{\mathrm{N}}_2}^{+}$ states together with an interference pattern that carries the signature of the potential energy curves activated by the XUV pulse. We show that the subfemtosecond characterization of the interference pattern is essential for a semiquantitative determination of the repulsive part of these curves.

Popular Summary

Molecular nitrogen is one of the major constituents of the upper atmospheres of Earth, Jupiter, Saturn, and Saturn’s moon Titan. Studying the interaction between molecular nitrogen and extreme ultraviolet (XUV) radiation is critically important to understanding radiation-induced atmospheric photochemical processes. In particular, high-energy photons induce ionization and subsequent dissociation of ${\mathrm{N}}_2$, leading to the formation of N and ${\mathrm{N}}^{+}$ atomic species that are relevant to the photochemistry of the upper atmosphere. However, tracking the earliest stages of the coupled electron and nuclear dynamics of a multielectron system such as ionized molecular nitrogen is difficult to investigate experimentally and notoriously complex to model theoretically. Here, we investigate, for the first time, the dissociation dynamics of molecular nitrogen with attosecond (${10}^{-18}$ s) time resolution.

Our time-to-space mapping procedure, based on the use of an isolated XUV attosecond pulse (pump) in combination with a visible few-femtosecond laser pulse (probe), allows us to obtain substantial information about the molecular dynamics triggered by the XUV pulse that disassociates the molecule. We study the dissociative mechanisms of molecular nitrogen by measuring the angular momentum distribution of the ${\mathrm{N}}^{+}$ fragments as a function of the pump-probe delay. In addition, we achieve a coherent control of the molecular response immediately following the interaction with high-energy photons. Finally, our experimental results allow benchmarking a complex theoretical model, which predicts the production of various excited N and ${\mathrm{N}}^{+}$ atomic species in significant amounts.

Our work opens up new perspectives for a more complete understanding of the radiative-transfer processes that occur in the upper atmospheres of planets and moons where XUV solar radiation is largely attenuated by the presence of molecular nitrogen.

Figure 1

Potential energy curves of the ground state of ${\mathrm{N}}_2$ (black curve) and the relevant electronic states of ${{\mathrm{N}}_2}^{+}$ (gray curves): the most relevant electronic states in the ionization process with ${\mathrm{\Sigma }}_g$ symmetry, namely, the $X{{}^2\mathrm{\Sigma }}_g$, the $F{{}^2\mathrm{\Sigma }}_g$, and the $3{{}^2\mathrm{\Sigma }}_g$ states, are shown in purple, red, and green lines, respectively. Left: Relative initial populations of the $X{{}^2\mathrm{\Sigma }}_g$, the $F{{}^2\mathrm{\Sigma }}_g$, and the $3{{}^2\mathrm{\Sigma }}_g$ states. Right: Measured XUV spectrum.

Figure 2

(a) ${\mathrm{N}}^{+}$ kinetic energy spectrum obtained by integrating the retrieved 3D momentum distribution (inset) within 20° around the laser polarization axis (white dashed line). (b) ${\mathrm{N}}^{+}$ kinetic energy spectra as a function of the delay between the XUV pump pulse and the NIR/VIS probe pulse.

Figure 3

(a) Time-dependent ${\mathrm{N}}^{+}$ kinetic energy spectra acquired within the pump-probe delay interval 5–16 fs. (b) ${\mathrm{N}}^{+}$ yield integrated in a 0.3-eV-wide energy band around 0.8 eV (black curve), 1.6 eV (blue curve), 1.9 eV (red curve), and 2.2 eV (green curve). An arbitrary offset has been added to the curves for better visualization.

Figure 4

Theoretical ${\mathrm{N}}^{+}$ kinetic energy spectrum (black dashed line) and the partial contributions of groups of states leading to the different dissociation channels (in increasing order of energy of the dissociative limit: magenta, red, green, blue). The inset shows the time evolution of the initial wave packet ultimately resulting in the given kinetic energy spectrum.

Figure 5

(a) Time-dependent ${\mathrm{N}}^{+}$ kinetic energy spectra calculated by including all states within the pump-probe delay interval 5–16 fs. (b) ${\mathrm{N}}^{+}$ yield integrated over a 0.3-eV-wide energy band around 0.8 eV (black curve), 1.6 eV (blue curve), 1.9 eV (red curve), and 2.2 eV (green curve). An arbitrary offset has been added to the curves for better visualization.

Figure 6

Oscillatory pattern obtained after subtraction of a fifth-order polynomial fitting curve to filter out the slow dynamics in both the experimental (a) and the theoretical (b) data around 0.8 eV (black curve), 1.6 eV (blue curve), 1.9 eV (red curve), and 2.2 eV (green curve). (c) Fourier transform power spectrum of the experimental curve (red line) and the theoretical curve (black line) around 2.2 eV.

Figure 7

PECs with schematic depiction of initial populations shortly after ionization. Red transparent area: Franck-Condon region. Black striped area: region with significant population transfer between the $3{{}^2\mathrm{\Sigma }}_g$ and $F{{}^2\mathrm{\Sigma }}_g$ state for time delays where the interference is observed. The magenta single-headed arrows depict the two-photon process populating the $3{{}^2\mathrm{\Sigma }}_g$ state using the $5{{}^2\mathrm{\Sigma }}_u$ state as a virtual intermediate state. The black double-headed arrows show points of resonant single-photon transitions.

Figure 8

Time-dependent ${\mathrm{N}}^{+}$ kinetic energy (KE) spectra calculated (a) within the four-state model and (b) for each individual state of this model. To clearly show the effect of the probe pulse, the KE spectrum corresponding to a calculation without probe pulse is subtracted from each of the panels reported in (b).

Figure 9

(a)–(g) Time-dependent ${\mathrm{N}}^{+}$ kinetic energy spectra corresponding to the set of gradients of the $3{{}^2\mathrm{\Sigma }}_g$ state displayed in the bottom right-hand panel [(d) corresponds to the unaltered $3{{}^2\mathrm{\Sigma }}_g$ state]. As in Fig. 3, the results of a reference calculation without the probe pulse are subtracted.

Figure 10

(a) Scheme of the experimental setup. In the figure, BS is used for Beam Splitter, FS wedges is used for Fused Silica wedges, PG plates is used for Polarization Gating plates and VMIS is used for Velocity Map Imaging Spectrometer. Acquired spectrum of the NIR/VIS (b) and XUV (c) pulses.

Figure 11

Potential energy curves of symmetries including ${{}^2\mathrm{\Sigma }}_g$ (left) and ${{}^2\mathrm{\Sigma }}_u$ (right) electronic states. The main panels show the adiabatic states whose orbitals are optimized and the insets show the diabatic states used for propagation, including also some higher lying states (faint lines) obtained with the method outlined in Sec. 5.