Dynamics of N2 Dissociation upon Inner ... - ACS Publications

Linda Young , Kiyoshi Ueda , Markus Gühr , Philip H Bucksbaum , Marc Simon ... F. Lépine , A. I. Kuleff , M. J. J. Vrakking , O. Kornilov , H. Köppel ...
0 downloads 0 Views 5MB Size
Letter pubs.acs.org/JPCL

Dynamics of N2 Dissociation upon Inner-Valence Ionization by Wavelength-Selected XUV Pulses Martin Eckstein,† Chung-Hsin Yang,†,§ Markus Kubin,†,∥ Fabio Frassetto,‡ Luca Poletto,‡ Hans-Hermann Ritze,† Marc J. J. Vrakking,† and Oleg Kornilov*,† †

Max-Born-Institute, Max-Born-Straße 2A, 12489 Berlin, Germany National Research Council, Institute of Photonics and Nanotechnologies (CNR-IFN), via Trasea 7, I-35131 Padova, Italy



S Supporting Information *

ABSTRACT: Ionization of nitrogen by extreme ultraviolet (XUV) light from the Sun has recently been recognized as an important driver of chemical reactions in the atmosphere of Titan. XUV photons with energies of 24 eV and above convert inert nitrogen molecules into reactive neutral and ionic fragments that initiate chemical reactions. Understanding the XUV-induced fragmentation poses significant challenges to modern theory owing to its ultrafast time scales, complex electronic rearrangements, and strong dependence on the XUV photon energy. Here, we apply femtosecond timeresolved photoelectron and photoion spectroscopy to study dissociative ionization of nitrogen, the most abundant molecule in Titan’s atmosphere, at selected XUV photon energies using a table-top XUV time-compensating monochromator. We probe the resulting dynamics using a time-delayed infrared (IR) ionization pulse. Coupled with ab initio calculations, the results allow us to assign the major dissociation channels resulting from production of an inner-valence hole, with important implications for models of Titan’s XUV-driven atmospheric chemistry.

I

quantum chemistry methods, given that strong electron correlation, very high electronic excitations, and ultrafast nuclear dynamics all have to be accurately accounted for. In this work, we investigate XUV ionization of N2 for photon energies between 33 and 50 eV (see Figure 1). In addition to electrons with binding energies below 20 eV accompanying nondissociative ionization, the XUV photoemission spectrum in this range contains three dominant features with binding energies between 24 and 38 eV (see Figure 2a). The lowest one is attributed to formation of the predissociative C2Σu+ ionic state with electronic configuration 3σg−1 1πu−1 1πg1.5 The C2Σu+ state fragments on time scales of several picoseconds to nanoseconds, producing ground state N atoms in a 2s2 2p3 (4S) electronic configuration, accompanied by ground state N+ ions in a 2s2 2p2 (3P) electronic configuration. The next dominant feature, with a binding energy of about 28 eV, is reliably assigned to the F2Σg+ state, with dominant electronic configuration 2σu−1 1πu−1 1πg1.5 High-quality ab initio calculations show that this state dissociates along an adiabatic potential energy curve producing ground state N+ ions and excited N atoms in a 2D electronic configuration. The N(2D) atoms are far more reactive than the ground state N(4S) atoms. It has recently been suggested that N(2D) atoms play an important role in the chemistry of the Titan atmosphere by

n recent years, our understanding of extraterrestrial atmospheric chemistry has tremendously advanced owing to the spectacular success of the Cassini−Huygens space mission in the vicinity of Titan, one of the moons of Saturn.1 Titan’s atmosphere has been compared to that of the early Earth.2 The upper atmosphere consists to about 98% of nitrogen and to slightly less than 2% of methane,3 with additional small traces from other gases. Illuminated by high energy photons from the Sun, the otherwise inert nitrogen molecules fall apart to form reactive neutral and ionic atomic fragments, which initiate various chemical reactions and contribute to the formation of a mysterious haze: the orangecolored aerosols that hide the surface of Titan from visual observation.4 In this study, we investigate these dissociation processes in the laboratory. Dissociative ionization of nitrogen by XUV light has been the focus of basic research for many years, serving as a benchmark for novel theoretical and experimental methods. Nevertheless, the description of the XUV ionization of nitrogen is far from complete. The N2 molecule is bound by a triple bond, one of the strongest known chemical bonds. Nitrogen’s ionization potential is 15.6 eV, but ionization with photon energies below 24.3 eV results exclusively in the production of bound N2+ ions. Ionization above 24.3 eV leads to predissociation on picosecond to nanosecond time scales, whereas above 28 eV, the molecule can dissociate directly on a time scale of several femtoseconds to form the N+ ion and the neutral N atom. These processes involve significant electronic rearrangement, making their description challenging even for modern ab initio © XXXX American Chemical Society

Received: December 3, 2014 Accepted: January 13, 2015

419

DOI: 10.1021/jz5025542 J. Phys. Chem. Lett. 2015, 6, 419−425

Letter

The Journal of Physical Chemistry Letters

Figure 1. Dissociative ionization of N2 molecules by extreme ultraviolet (XUV) light. Schematic view of the XUV time-compensating monochromator with a typical high-harmonic generation (HHG) spectrum, a typical XUV−IR cross-correlation profile and a typical ion velocity map imaging (VMI) measurement. The inset in the HHG spectrum lists the energies of the harmonics used in the present experiments.

Figure 2. Ionization of N2 molecules with XUV photons recorded in present experiments. (a) Photoelectron spectra indicating the contributions of the C2Σu+ state, the F2Σg+ state and the H band. (b) Photoion spectra indicating the contributions of the F2Σg+ state and the H band on the basis of the analysis elaborated in Figure 3

dissociative ionization of nitrogen with improved time resolution, suggesting that some of the observed features may arise from autoionization dynamics. All these previous time-resolved studies employed fixed XUV photon energies. However, the dissociation dynamics strongly depends on the photon energy.12 In the present work, we employ a recently constructed apparatus that allows performing time-dependent studies using femtosecond XUV pulses with a tunable photon energy. The setup consists of an HHG source coupled to a time-compensating XUV monochromator schematically depicted in Figure 1 (see Methods section for details). Individual harmonics in the HHG spectrum covering photon energies from 20 to 50 eV can be selected (inset in Figure 1), permitting an investigation of the dissociation dynamics of nitrogen both below and above the H band origin. Figure 2a shows static photoelectron spectra recorded using harmonics H21 to H31. The corresponding photon energies are listed in Figure 1 and the spectral widths are about 0.5 eV. The spectra are in very good agreement with previously published photoelectron spectra.12 Corresponding ion kinetic energy spectra are shown in Figure 2b. They predominantly contain two features. A ∼2 eV kinetic energy release (KER) results from dissociation of the F2Σg+ ionic state,13 whereas another feature with a KER of 4−6 eV changes its shape as the photon energy increases and previously could not be reliably assigned to a particular dissociation channel.

opening new routes toward the formation of HCN radicals, one of the key species involved in the formation of the Titan haze.6 The last of the intense bands mentioned above, with a binding energy of about 38 eV, is commonly labeled as the H band (or 2σg−1 band) and is attributed to ionization of the 2σg inner valence orbital.5 This band shows a rich structure of narrow photoemission peaks on top of a broad feature7 and, therefore, cannot be attributed to a single dissociative state. Many spectroscopic methods have already been applied to study dissociative ionization of the H band, but its dynamical nature has challenged spectroscopic assignments. Timeresolved methods are a natural tool to approach spectroscopic problems that are affected by dynamics. Recent developments in time-resolved XUV science, based on the availability of highharmonic generation (HHG) and XUV/X-ray free electron laser radiation, thus allow significant progress. In a seminal paper, Gagnon et al.8 used femtosecond XUV pulses with a fixed photon energy of 43 eV to initiate the dissociation of nitrogen and probed the resulting dynamics by infrared (IR) multiphoton ionization. They assigned observed time-dependent signals to one of the electron-shakeup states, as previously corroborated by Sambe and Ramaker,9 among others. Magrakvelidze et al. used a pair of 38 eV XUV pulses from the FLASH free electron laser for probing dissociative and nondissociative channels.10 And recently, Lucchini at al11 used a train of attosecond pulses for studying the XUV-induced 420

DOI: 10.1021/jz5025542 J. Phys. Chem. Lett. 2015, 6, 419−425

Letter

The Journal of Physical Chemistry Letters

Figure 3. Comparison between the photoelectron (solid line, upper graphs) and photoion spectra (solid line, lower graphs) for harmonics H27, H29, and H31. Assuming that the electronic energy of the fragments resulting from the F2Σg+ state dissociation is 26.7 eV (L3 asymptote) and that of the H band is 31.9 eV (L9 asymptote), the measured photoelectron spectra permit prediction of the ion fragment KER (gray lines, lower graphs).

To gain more insight into the dissociation dynamics we can exploit energy conservation. Following XUV absorption, the available energy is distributed over the ion KER, the photoelectron energy and the electronic energy of the N/N+ fragments. Thus, for a given electronic energy of the fragments (i.e., a given dissociation channel), the photoelectron spectrum is a mirror image of the ion KER spectrum. In Figure 3, the photoelectron spectra of Figure 2a measured for harmonics H27, H29, and H31 are overlapped with the corresponding fragment ion KER distributions of Figure 2b, assuming the wellknown N+ (3P) + N (2D) dissociation channel for the F2Σg+ state (L3 asymptote) and a newly assigned N+ (1S) + N (2P) channel for the H band (the L9 asymptote, see Supporting Information for a table of N2+ dissociation asymptotes). For harmonics H27 to H31, the positions and widths of the features in the photoelectron spectrum and in the ion KERs are consistent with this assumption, indicating that dissociation of the H band follows a pathway that maximizes the electronic energy in the fragments, considering the occupation of the 2s and 2p shells of the nitrogen atoms and ions. This assignment contradicts recent calculations of Aoto et al.,14 which assign the H band to the fifth and sixth adiabatic potential energy curves of Σg symmetry (the L6 asymptote). To clarify the discrepancy between our experimental results and the calculations of Aoto et al.,14 we have carried out stateof-the-art ab initio calculations for 13 excited states of N2+ for the case of Σg symmetry. Expecting that Rydberg states may be significant in the energy range of the H band,9 the main difference between our calculations and those of ref 14 is the use of an extended basis set augmented with diffuse functions for a proper description of Rydberg-type excited states (aug-ccpVQZ15). We employ the novel density matrix renormalization group (DMRG) technique16−19 that was recently implemented in quantum chemistry packages and that is shown to be highly efficient for strongly correlated systems and very large active spaces (see Methods section for details). The results of the ab initio calculations are shown in Figure 4. The electronic state that adiabatically connects to the experimentally observed L9 dissociation channel is shown by a blue line. In the Franck−Condon region, this state (36.9 eV) is

Figure 4. Ab initio calculations of N2+ excited state potential energy curves calculated using the DMRG method; yellow line, the potential energy curve corresponding to the F2Σg+ state; blue line, the major dissociation channel of the H band connecting to the L9 asymptote; red lines, dissociation curves leading to 3s excited fragments (L10 and L11 asymptotes). The contribution of the inner-valence 2σg−1 hole to a particular state is indicated by the size and color of the symbol superimposed on the solid curves. The black lines show the N2+ and the N22+ ground state potential energy curves. Dissociation asymptotes are labeled on the right.

421

DOI: 10.1021/jz5025542 J. Phys. Chem. Lett. 2015, 6, 419−425

Letter

The Journal of Physical Chemistry Letters

Figure 5. Transient photoion KER maps, showing the variation of the KER as a function of the XUV-IR pump−probe delay. The solid lines indicate the results from classical trajectory calculation and reveal the role of Coulomb repulsion in the final state. The dashed lines indicate the KER of highly excited neutral atoms or molecules ionized by IR. The color scale for each map is chosen to emphasize the dynamics.

Figure 6. Transient photoelectron maps, showing the variation of the photoelectron energy as a function of the XUV-IR pump−probe delay. The solid lines are photoelectron energies resulting from the classical trajectory calculation as presented in Figure 4. (see also main text). The color scale for each map is chosen to emphasize the dynamics.

significantly closer to the origin of the H band (36.5 eV) than the corresponding state in the calculations of Aoto et al. (40.5 eV). Thus, the choice of the augmented basis set indeed leads to substantial improvement and confirms that Rydberg-like states play a role in the H band. Approximate populations of the adiabatic states upon photoemission can be estimated from orbital occupation numbers, which are extracted from the DMRG reduced density matrix. The hole occupations are included in Figure 4 as round circles for each calculated point with their size and color indicating the contribution of the 2σg−1 hole. At the N2 equilibrium distance of 1.12 Å, the 2σg−1 hole occupation is dominant in an electronic state that adiabatically connects to the L10 asymptote. At larger internuclear distances the 2σg−1

hole is distributed among several excited states, including the aforementioned state connected to the L9 asymptote (blue line). The evolution of the hole occupation number as a function of internuclear distance indicates strong nonadiabatic couplings between the closely spaced adiabatic states. Our experimental findings thus suggest that part of the population is efficiently and nonadiabatically transferred from the initially excited state to the state connected to the L9 asymptote. This result has direct consequences for the role of nitrogen reactivity in the Titan atmosphere: at photon energies above 40 eV (31 nm), the dissociation produces a significant fraction of neutral atoms in the excited 2 P electronic configuration (L9 asymptote), whereas the contribution of atoms in the 2D electronic configuration (from the F2Σg+ state) is lower than 422

DOI: 10.1021/jz5025542 J. Phys. Chem. Lett. 2015, 6, 419−425

Letter

The Journal of Physical Chemistry Letters

asymptote.21,22 The latter channel is not strongly expressed in our measurements because for this state three photon ionization is only possible at delays greater than 300 fs. Similar classical trajectories are plotted as solid lines in Figure 5 assuming the initial potential energy of the molecular ion corresponds to the maximum of the H band, except for harmonic H23, where the resonant photon energy (slightly below the H band origin) is taken for the initial conditions. A second class of features observed in the maps of Figures 5 and 6 are a number of delay-independent KER peaks and corresponding photoelectron peaks that are best visible in the maps obtained with harmonics H21 and H23 (dashed lines). Given that they show no delay dependence, these signals must arise from IR dissociation of bound N2+ ions, from dissociative ionization of neutral molecules or from ionization of atomic fragments. The first possibility can be ruled out, because the IRinduced dissociation of bound N2+ ions is unlikely to lead to a channel with a well-defined KER of 5.2 eV (H21) or 6.8 eV (H23). Moreover, the intensity of the IR used in the experiment is too low to dissociate bound N2+ molecules. The features, thus, must be related to neutral molecules or fragments. The fragment ion KER changes from one harmonic to the next, suggesting that excited states of the N2 molecule are involved that are selected by the photon energy. Such neutral excited states can be formed when an electron is removed from the 2σg orbital and placed in a Rydberg orbital.23 Using the Rydberg energy for an electron in the 4s state of 3.47 eV9 (we adopt the united atom notation of Rydberg states, as found in Table 8 of ref 9) and the experimentally observed origin of the H band at 36.5 eV, we can estimate the binding energy of the Rydberg state attached to the 2σg−1 core to be around 36.5 − 3.5 = 33.0 eV, close to the photon energy of harmonic H21. Following the same recipe the neutral channel observed in the map of harmonic H23 could be due to Rydberg states in the n = 7 manifold. In conclusion, we have experimentally investigated XUV dissociative ionization of N2 molecules using a newly constructed time-compensating XUV monochromator beamline. The XUV photon energy was tuned over the energy range corresponding to ionization of the 2σg inner-valence orbital (the H band), and it was observed that the fragmentation processes strongly depend on the XUV photon energies. For photon energies more than 10 eV above the H band binding energy, the dissociation mainly leads to the N+ (1S) + N (2P) channel, which is characterized by the highest excitation energy of the fragments within the n = 2 manifold of states: the 2P state of the neutral N atom and the 1S state of the N+ ion. We confirm this assignment by ab initio calculations using the DMRG method. This result has important implications in the context of atmospheric chemistry, in particular that of Titan’s atmosphere, where electronic states of N atomic fragments produced by XUV ionization play important roles in existing chemical models. The recent model of Krasnopolsky4 assumes that all excited N* atoms produced by radiation from the Sun below 51 nm (24.3 eV) are in the 2D state. Thus, the reaction rates of these excited atoms with methane and other molecules are taken to be those for N(2D) atoms. However, our results indicate that below 34 nm (above 36.5 eV), excitation of the H band dominates over that of the F2Σg+ state,5 and the production of N(2D) atoms by direct photodissociation (about 22% of all N(2D) atoms, the rest coming from neutral dissociation4,24) decreases at the expense of N(2P), which has almost 50 times lower reaction rates.

hitherto thought. The reactive rate constants of 2P atoms, for example, in reactions with methane, are almost 50 times lower than those of 2D atoms (8.8 × 10−14 cm3 molecule−1 s−1 vs 2.0 × 10−12 cm3 molecule−1 s−1, respectively).20 Similar to Gagnon et al.,8 we employ moderately strong IR (800 nm) pulses with peak intensities of about 2 TW/cm2 for multiphoton ionization of electronically excited fragment atoms and observe changes induced in the fragment ion KER and the photoelectron spectra by the IR probe pulse. Transient KER maps for N+ ions are shown in Figure 5. These maps are obtained by recording KER spectra as a function of the time delay between the XUV and IR pulses (abscissa) and subtracting the signal recorded when only the XUV pulses are present. The maps show the dynamics induced by the IR pulses as the N2+ molecular ion produced by the XUV pulse dissociates. The red color corresponds to new N+ ions created by IR pulses as well as to ions with KER changed by the IR pulse from its XUV-only value (blue colored areas). These results clearly demonstrate that the dissociation dynamics changes drastically when the XUV photon energy is varied from 32.8 eV (H21) to 48.4 eV (H31). The maps in Figure 5 show two general classes of features. For harmonics H23−H31 they contain transient signals where the KER changes by a few eV over time scales of several hundreds of femtoseconds. This long time scale behavior is explained by an IR-induced Coulomb explosion of the dissociating molecular ion. The fragment pair KER increases by an amount that depends on the separation between the fragments at the moment of secondary ionization by the IR pulse. This leads to enhanced yield at higher KER (yellow-red areas) and a reduced yield at the KER corresponding to XUVonly dissociation (blue areas). The multiphoton IR ionization leads to photoelectrons with kinetic energies up to the IR photon energy (1.56 eV), which are observed in the transient photoelectron maps in Figure 6 (recorded in a manner similar to the ion measurements). The measurements performed with harmonics H27 to H31 show the appearance of a photoelectron band from a pump−probe delay of about 200 fs onward, initially with zero kinetic energy and then increasing in energy for longer delays. This behavior is consistent with the Coulomb repulsion energy decreasing at larger internuclear separations. A classical trajectory calculation of the Coulomb repulsion is included in Figures 5 and 6 in the form of black lines. The photoelectron energy is calculated as E = Ethr − E2N+ + 3ℏω − kee2/(ντ + r0), where Ethr is the electronic energy of the fragments in the dissociating N2+, E2N+ = 38.85 eV is the double ionization threshold, and 3 ℏω is the energy of the three IR photons required to ionize the neutral fragment; ke is Coulomb’s constant, r0 is the equilibrium bond length of N2. The relative velocity v of the fragments prior to Coulomb explosion is assumed to be nearly constant because the range of dissociative N+ + N potential energy curves is very short and the fragments reach their terminal velocity on a time scale close to the resolution of the present measurement. The lines plotted in Figure 6 arise from the ionization of molecules that have not relaxed to the potential energy curve that adiabatically connects to the L9 asymptote, and correspond to ionization of N atoms in a 2s2 2p2 (3P) 3s (2P) configuration (the L11 asymptote). These excited state atoms are responsible for fluorescence at 149 nm that is observed upon ionization of N2 to the H band.21 This is the second most intense fluorescence line next to the more intense 120 nm line corresponding to relaxation of N atoms in a 2s2 2p2 (3P) 3s (4P) configuration, the L10 423

DOI: 10.1021/jz5025542 J. Phys. Chem. Lett. 2015, 6, 419−425

Letter

The Journal of Physical Chemistry Letters

high number density only few millimeters away from the molecular beam source. The recorded VMI images were reconstructed using the BASEX inversion routine,26 yielding the kinetic energy spectra. To improve the quality of the data, the photoelectron spectra at each temporal delay were normalized to the total photoelectron signal from ionization to the X, A, and B states of N2+ (binding energy below 20 eV in Figure 2a). The corresponding molecular ions cannot be ionized by the IR probe pulse at the present experimental conditions. Therefore, the total signal in this range of energies can be assumed to be constant. A similar procedure cannot be carried out for the N+ measurements because the IR-induced spectral changes cover the complete KER spectra. Therefore, the spectra were normalized to the total counts and then corrected for the additional ionization events using the photoelectron data and assuming that secondary photoelectrons added by the IR have energies below one IR photon energy (1.56 eV). Theoretical Methods. The ab initio calculations are carried out using the ORCA and MOLPRO quantum chemistry packages27,28 and the BLOCK quantum chemistry DMRG code.16−19 They are performed using the augmented ccpVQZ basis set.15 This ansatz is responsible for most of the state energy differences between the present calculations and those of ref 14. Further improvement is achieved by optimizing the molecular orbitals at the CASSCF(8,9) level of theory and then performing an independent DMRG run retaining the M = 1000 states with Fiedler orbital ordering. The convergence threshold of the calculation was set at 1 mEh. The multiple excited states were calculated using the Davidson stateaveraging algorithm implemented in the BLOCK package. The N2+ molecule belongs to the D∞h symmetry group. The closest group implemented in the quantum chemistry packages is the D2h symmetry group, which was used in the present calculations to restrict the number of excited states. The desired Σ g + states could be calculated in the A g irreducible representation; however, care has to be taken to remove the states of Δg symmetry, which also fall in this irreducible representation. This was achieved by performing calculations in the Ag and B1g irreducible representations and removing points which have the same energies in both calculations. The results of the calculations are presented in Figure 4. For reference, the figure also includes the N22+ ground state potential energy curve calculated at the CASSCF level of theory.

In addition to the novel information on the XUV-only dissociative ionization dynamics, moderately strong IR pulses at 800 nm have been used to probe the dissociation dynamics, demonstrating the involvement of highly excited neutral species over the whole range of XUV photon energies. At energies above the H band origin, dissociation channels have been inferred with the neutral N atom excited to the 3s state. Dissociation in these channels is strongly nonadiabatically coupled to the main dissociation pathway leading to production of N+ (1S) ions and N (2P) atoms. For energies below the H band origin, the excitation of Rydberg states attached to an N2+ core with a 2σg−1 hole is observed. The pronounced manifestation of the 2σg−1 hole over a wide range of XUV energies suggests that a time-domain description of the photoionization process may give deeper insight into the dissociation of N2. Hence, the present results will serve as a solid base for future experiments using attosecond pulses.



METHODS Experimental Methods. Femtosecond XUV pulses are produced using the high-order harmonic generation technique (HHG). A commercial Ti:sapphire laser system (Aurora, Amplitude Technologies) delivers 25 fs IR pulses with 5 mJ (up to 20 mJ) pulse energy at a central wavelength of 795 nm and at a repetition rate of 1 kHz. The pulses are equally split between the pump (XUV) and probe arms of the setup. The pump beam is focused by a f = 62.5 cm focusing mirror into a gas cell situated in a vacuum chamber and filled with Ar to a pressure of 45 mbar generating high-order harmonics. The generating IR pulses are separated from the XUV pulses by a 100 nm thin Al filter. The HHG vacuum chamber is connected to a beamline housing the optics of the time-compensating XUV monochromator (Figure 1). The design of the monochromator is similar to that implemented in ref 25. In short, the monochromator consist of two consecutive sections symmetric in reflection. In the first section, the XUV beam is collimated by a toroidal mirror and is reflected from a diffraction grating mounted in a conical diffraction configuration (grooves parallel to the reflection plane). Such mounting greatly improves the efficiency of the XUV diffraction. Different XUV wavelengths are selected by rotating the grating around an axis in the reflection plane (Figure 1). The diffracted beam is focused by the second toroidal mirror onto a 200 μm wide slit. Phase-front tilt induced by the first grating is compensated by the grating in the second section. A choice of three gratings with different groove densities (150, 300, and 600 lines/mm) allows optimization of the reflectance over the full range of wavelengths. The transmission of the monochromator varies from 3% to 16% for harmonics between H13 (20.3 eV) and H31 (48.4 eV), which results in photon fluxes of 106 to 107 photons per pulse per harmonic. The rotation of the gratings in both sections is fully automated. Harmonic orders from 21 to 31 were selected for the present experiments because they cover the complete energy range of the H band of nitrogen. The wavelength-selected XUV beam is recombined with the IR beam by means of a holey mirror. Both beams are focused in the interaction region of a velocity map imaging spectrometer (VMIS) used to detect the products of ionization: ions and electrons. In the ion detection mode the voltages of the MCP detector of the VMIS are gated to select the ions of interest (N+ ions in this work). A source of N2 molecules is built into the repeller plate of the VMIS, which allows experiments at very



ASSOCIATED CONTENT

* Supporting Information S

Table of dissociation asymptotes of the N2+ molecular ion. This material is available free of charge via the Internet at http:// pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

* E-mail: [email protected]. Present Addresses §

(C.-H.Y.) Academia Sinica, 128 Academia Road, Section 2, Nankang, Taipei 115, Taiwan. ∥ (M.K.) Helmholtz-Zentrum Berlin, Albert-Einstein-Straße 15, 12489 Berlin, Germany. Notes

The authors declare no competing financial interest. 424

DOI: 10.1021/jz5025542 J. Phys. Chem. Lett. 2015, 6, 419−425

Letter

The Journal of Physical Chemistry Letters



(19) Sharma, S.; Chan, G. K-.L. Spin-Adapted Density Matrix Renormalization Group Algorithms for Quantum Chemistry. J. Chem. Phys. 2012, 136, 124121. (20) Herron, J. T. Evaluated Chemical Kinetics Data for Reactions of N(2D), N(2P), and N2(A3Σu+) in the Gas Phase. J. Phys. Chem. Ref. Data 1999, 28, 1453−1483. (21) Samson, J. A. R.; Chung, Y.; Lee, E.-M. Excited Ionic and Neutral Fragments Produced by Dissociation of the N2+* H Band. J. Chem. Phys. 1991, 95, 717−719. (22) Ehresmann, A.; Machida, S.; Kitajima, M.; Ukai, M.; Kameta, K.; Kouchi, N.; Hatano, Y.; Shigemasa, E.; Hayaishi, T. Dissociative Single and Double Photoionization with Excitation Between 37 and 69 eV in N2. J. Phys. B: Atom. Mol. Opt. Phys. 2000, 33, 473−490. (23) Strasser, D.; Haber, L.; Doughty, B.; Leone, S. R. Ultrafast Predissociation of Superexcited Nitrogen Molecules. Mol. Phys. 2008, 106, 275−280. (24) Helm, H.; Hazell, I.; Bjerre, N. Lifetimes and Rydberg-Valence State Mixing of the c’ 1Σg+(ν = 4) and c 1Πu(ν = 4) States of N2. Phys. Rev. A 1993, 48, 2762−2771. (25) Poletto, L.; Villoresi, P.; Benedetti, E.; Ferrari, F.; Stagira, S.; Sansone, G.; Nisoli, M. Temporal Characterization of a TimeCompensated Monochromator for High-Efficiency Selection of Extreme-Ultraviolet Pulses Generated by High-Order Harmonics. J. Opt. Soc. Am. B 2008, 25, B44−B49. (26) Dribinski, V.; Ossadtchi, A.; Mandelshtam, V. A.; Reisler, H. Reconstruction of Abel-Transformable Images: The Gaussian Basis-Set Expansion Abel Transform Method. Rev. Sci. Instrum. 2002, 73, 2634− 2642. (27) Neese, F. The ORCA Program System. WIREs Comput. Mol. Sci. 2012, 2, 73−78. (28) Werner, H.-J.; Knowles, P. J.; Knizia, G.; Manby, F. R.; Schütz, M. Molpro: A General-Purpose Quantum Chemistry Program Package. Wiley Interdiscip. Rev.: Comput. Mol. Sci. 2012, 2, 242−253.

ACKNOWLEDGMENTS We gratefully acknowledge Serguei Patchkovski for valuable discussions.



REFERENCES

(1) Cable, M. L.; Hörst, S. M.; Hodyss, R.; Beauchamp, P. M.; Smith, M. A.; Willis, P. A. Titan Tholins: Simulating Titan Organic Chemistry in the Cassini−Huygens Era. Chem. Rev. 2012, 112, 1882−1909. (2) Trainer, M. G.; Pavlov, A. A.; DeWitt, H. L.; Jimenez, J. L.; McKay, C. P.; Toon, O. B.; Tolbert, M. A. Organic Haze on Titan and the Early Earth. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 18035−18042. (3) Niemann, H. B.; Atreya, S. K.; Bauer, S. J.; Carignan, G. R.; Demick, J. E.; Frost, R. L.; Gautier, D.; Haberman, J. A.; Harpold, D. N.; Hunten, D. M.; et al. The Abundances of Constituents of Titan’s Atmosphere from the GCMS Instrument on the Huygens Probe. Nature 2005, 438, 779−784. (4) Krasnopolsky, V. A. A Photochemical Model of Titan’s Atmosphere and Ionosphere. Icarus 2009, 201, 226−256. (5) Langhoff, P.; Langhoff, S.; Rescigno, T.; Schirmer, J.; Cederbaum, L.; Domcke, W.; Niessen, W. V. Theoretical Studies of Inner-ValenceShell Photoionization Cross Sections in N2 and CO. Chem. Phys. 1981, 58, 71−91. (6) Imanaka, H.; Smith, M. A. Formation of Nitrogenated Organic Aerosols in the Titan Upper Atmosphere. Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 12423−12428. (7) Liu, Z.; Bancroft, G.; Coatsworth, L.; Tan, K. Fine Structure in the Inner Valence Photoelectron Spectra of N2 and CO. Chem. Phys. Lett. 1993, 203, 337−343. (8) Gagnon, E.; Ranitovic, P.; Tong, X.-M.; Cocke, C. L.; Murnane, M. M.; Kapteyn, H. C.; Sandhu, A. S. Soft X-Ray-Driven Femtosecond Molecular Dynamics. Science 2007, 317, 1374−1378. (9) Sambe, H.; Ramaker, D. E. Rydberg States Converging to the N22+ Ionized States. Chem. Phys. 1986, 107, 351−360. (10) Magrakvelidze, M.; Herrwerth, O.; Jiang, Y. H.; Rudenko, A.; Kurka, M.; Foucar, L.; Kühnel, K. U.; Kübel, M.; Johnson, N. G.; Schröter, C. D.; et al. Tracing Nuclear-Wave-Packet Dynamics in Singly and Doubly Charged States of N2 and O2 with XUV-Pump XUV-Probe Experiments. Phys. Rev. A 2012, 86, 013415. (11) Lucchini, M.; Kim, K.; Calegari, F.; Kelkensberg, F.; Siu, W.; Sansone, G.; Vrakking, M. J. J.; Hochlaf, M.; Nisoli, M. Autoionization and Ultrafast Relaxation Dynamics of Highly Excited States in N2. Phys. Rev. A 2012, 86, 043404. (12) Krummacher, S.; Schmidt, V.; Wuilleumier, F. Inner-Shell Photoionisation in Molecules: the Nitrogen Case. J. Phys. B: Atom. Mol. Phys. 1980, 13, 3993−4005. (13) Jiang, Y. H.; Rudenko, A.; Kurka, M.; Kühnel, K. U.; Ergler, T.; Foucar, L.; Schöffler, M.; Schössler, S.; Havermeier, T.; Smolarski, M.; et al. Few-Photon Multiple Ionization of N2 by Extreme Ultraviolet Free-Electron Laser Radiation. Phys. Rev. Lett. 2009, 102, 123002. (14) Aoto, T.; Ito, K.; Hikosaka, Y.; Shibasaki, A.; Hirayama, R.; Yamamono, N.; Miyoshi, E. Inner-Valence States of N2+ and the Dissociation Dynamics Studied by Threshold Photoelectron Spectroscopy and Configuration Interaction Calculation. J. Chem. Phys. 2006, 124, 234306. (15) Dunning, T. H., Jr. Gaussian Basis Sets for Use in Correlated Molecular Calculations. I. The Atoms Boron Through Neon and Hydrogen. J. Chem. Phys. 1989, 90, 1007−1023. (16) Chan, G. K.-L.; Head-Gordon, M. Highly Correlated Calculations with a Polynomial Cost Algorithm: A Study of the Density Matrix Renormalization Group. J. Chem. Phys. 2002, 116, 4462−4476. (17) Chan, G. K.-L. An Algorithm for Large Scale Density Matrix Renormalization Group Calculations. J. Chem. Phys. 2004, 120, 3172− 3178. (18) Ghosh, D.; Hachmann, J.; Yanai, T.; Chan, G. K.-L. Orbital Optimization in the Density Matrix Renormalization Group with Applications to Polyenes and β-Carotene. J. Chem. Phys. 2008, 128, 144117. 425

DOI: 10.1021/jz5025542 J. Phys. Chem. Lett. 2015, 6, 419−425