sign in sign up
<<
Home , Femtochemistry

COMMENTARY COMMENTARY

Photochemistry of highly excited states R. D. Levinea,b,c,1

These days on earth we are photochemically reasonably A forceful illustration of the benefits of the synergy of well shielded. The far UV radiation is filtered by the ozone experiment and theory in understanding the detailed and other components of the atmosphere. In the iono- electronic and nuclear dynamics in highly excited states is sphere and beyond there is much challenging provided by the very recent work reported by Peters et al. induced by such higher-energy photons. The shielding by (8). Methyl azide is pumped by an 8-eV vacuum UV (VUV) the atmosphere also makes demands on such experi- pulse of 10-fs duration and probed by a second 1.6-eV in- ments in the laboratory. There is, however, recent interest frared (IR) pulse that ionizes the . The experiment largely driven by that makes higher resolution as seen through the eyes of a theorist is shown in Fig. 1. The possible. The experimental technology is advanced light pump pulse prepares a single excited electronic state, la- sources that offer far higher intensities, sharper wave- beled as S8, that theory shows has a mixed valence and length definition or, complementarily, short-duration Rydberg character. It is unusual that a single excited state pulses. The two alternatives make a whole because of is accessed at such a high energy. The high-level computa- the quantum mechanical time–energy uncertainty princi- tions reported by Peters et al. are quite clear that the ple that makes short pulses necessarily broad in energy. S8 state is relatively isolated and we will return to this point. The new light sources also enable pump and probe ex- The IR pulse ionizes the excited state through a two-photon periments, where a second light beam is used to monitor absorption and the and electrons are detected in co- the processes induced by the pump (1, 2). One can safely incidence on a few time scale, which provides foresee that in the near future three and even four light a direct probe of the ultrafast correlated electron-nuclear beams will be used in confluence (3). The other technol- dynamics in the high excited-state manifold of the neutral. ogy that played a key role is the breakthrough improve- Initially, two electronically excited states of the cation are ment of the available theoretical tools. The highly excited detected. But as the C–N = N bond opens up, theory shows dense set of electronic states takes full advantage of the that the electronic energy of S8 goes down, the bend vibra- spectacular improvements in quantum chemical method- tional motion is excited, and only the ground state of the ologies. At the same time, quantum dynamics has come cation can be detected. The experiment clearly shows that of age, being able to treat the correlated motions of elec- the ionization turns by about 25 fs after excitation. Theory trons and nuclei. This is essential because it is a charac- suggests that the quenching of the ionization is direct evi- teristic of these highly excited states that small changes in dence of the nonadiabatic interaction of the S8 state with an the positions of the can bring about qualitative optically dark S4 state, and that this occurs on a rather fast changes in the electronic states. The result is that such time scale. The S4 state is a valence-excited state that highly excited states are typically strongly coupled to has no allowed ionization to the ground state of the cation. other states. Often, this leads to so-called conical inter- It is dark to the IR probe because its electronic energy is too sections (4, 5). The progress in theory has been stimulated low. The CNN bending motion brings the S8 state to the by the increasing sophistication of the experiments. In neighborhood of the S4 state by about 10 fs. turn, theory has challenged the laboratory with new con- Very recent evidence (9) is that a similarly short time scale cepts and new issues. to reach the first conical intersection is observed in the ben- It is not only space and that provide the zene cation. The experiment detects charged products. By challenges of dynamics of highly excited states often driven keeping the count rate rather low, electrons and cations can by extreme conditions. Progress in propellants and explo- be measured in coincidence. The count vs. time is shown in sives and other energetic compounds (6, 7) is also aided by Fig. 1. Such time-resolved photoelectron spectra have been the recent developments. An explosion is, necessarily, a used before, but mostly at lower excitation energies. Appli- bulk event, so a single-molecule point of view is not the cations to valence electron dynamics and to conical intersec- only germane aspect. Nor are the typical firearm tions were reported recently (10–12). The parent signal primers. But the insights on the very sudden release of decays with a time-decay constant of about 20 fs, while the + significant amounts of kinetic energy are both valuable primary daughter ion (HCNH ) signal rises with the same and interesting in their own right. time constant.

aThe Fritz Haber Center for Molecular Dynamics and Institute of Chemistry, The Hebrew University of Jerusalem, Jerusalem 91904, Israel; bDepartment of Molecular and Medical , David Geffen School of Medicine, University of California, Los Angeles, CA 90095; and cDepartment of Chemistry and , University of California, Los Angeles, CA 90095 Author contributions: R.D.L. wrote the paper. The author declares no conflict of interest. Published under the PNAS license. See companion article on page E11072. 1Email: [email protected].

13594–13596 | PNAS | December 26, 2017 | vol. 114 | no. 52 www.pnas.org/cgi/doi/10.1073/pnas.1718814114 Downloaded by guest on September 28, 2021 Fig. 1. The early time dynamics in the high-energy states of methyl azide. It is an ID plot vs. time. The very short, 10-fs VUV pulse accesses a single, relatively isolated highly excited singlet state, the eighth in order of increasing energy. The electronic density of the state is shown. The molecule responds by opening the CNN bond angle or it is ionized to one of two energetically allowed states of the cation, but an intense IR probe pulse. The ions and electrons are detected in coincidence. By about 20 fs, the S8 state approaches a conical intersection with an S4 state, a state that, due to selection rules, cannot be ionized. The quenching of the charged products sets the time scale for the S8 state to cross the conical intersection to the S4 state.

The reported theoretical simulations were carried out at several of pumped methyl azide can be described in terms of motion of nuclei levels. The highest, coupled clusters, were used to identify the optically along a potential energy, because the pump pulse duration, 10 fs, is absorbing state as a single, relatively isolated state, the eighth in relatively long. This means that its span in frequency, about 0.4 eV, is succession of singlets, hence S8. This electronic state is a superposition smaller than the spacings of the accessible excited electronic state in the of two Rydberg states, each one correlating to a different state of the 8-eV energy range. Thereby, only one excited electronic state is accessed. cation. It is the superposition that allows the probe IR transitions to the In such a case, a Born–Oppenheimer separation is valid at least immedi- two states of the cation, as illustrated in Fig. 1. Theory shows that ately past excitation. A description in terms of a single potential energy the S8 state is diabatically coupled to the S6 and S4 states, both of surface is then realistic. Using a shorter pulse, what is today becoming which are also a Rydberg–valence mix. (It is customary to characterize possible, provides a wider span in frequency; with such a pulse, two or the coupling as “nonadiabatic,” but strictly speaking “adiabatic” means more distinct electronic states can be reached. Such an excited initial state not diabatic and a theorist is expected to be careful with words.) There can be technically described as a coherent superposition of stationary was no experimental evidence that the S6 state is populated by the electronic states, otherwise known as an electronic wavepacket. The de- pump to any measurable extent. A key point of the theory is that the velopment of shorter pulses opens for exploration a new regime in correlated electron-nuclear motion subsequent to the excitation was dynamics. The previous analogous development is the introduction fully described as correlated. This means that each electronic state of femtosecond pulses with frequency spans wider than a vibrational was assigned its own wavepacket. The overall amplitudes of these spacing. These allow pumping of coherent superpositions of vibrational separate wavepackets were changed on the fly, as indicated by the states, states that are localized in space. As such states evolve in time, dynamics (13, 14). The dynamics shows that the motion out of the the geometry of the system changes. Therefore, such evolution corre- Franck–Condon regime on S8 to the conical intersection with S4 takes sponds to our intuitive idea of a molecular view of a . about a quarter bend of the CNN angle, or about 10 fs. In the hands of (17, 18) and others, pumping with For lower excited states, there were earlier indications of the fast femtosecond lasers allowed monitoring the unfolding of a chemical time scale for the relaxation of the optically excited state. Indeed, it is rearrangement in time. The new, sub–10-fs lasers allow monitoring the implied by Kasha’s rule that higher accessed states will relax to the electronic rearrangements that take place during a chemical event. On lowest singlet excited states faster than the time scale for fluores- such a short time scale, the nuclei barely have time to move. So the cence. More recent indirect evidence is based on the physical assump- early time dynamics is purely electronic. In the Born–Oppenheimer tion that ionization is electronically resonant, meaning Franck– approximation, at a given geometry the electrons are in a stationary Condon. Then, ionization occurs when the molecular configuration state. This state can change, a so called nonadiabatic transition, but has adjusted such that the electronic gap equals the energy of the this can happen only when the nuclei move. In the sub–10-fs early time photon. In a series of , it was then observed (15, 16) that the dynamics, the electrons are not stationary. The electrons can perform motion out of the Franck–Condon region of the initially excited state quite complex motions. This is often the case because the higher- and into a dark state occurs rather rapidly, often in below 50 fs. energy excited states are typically superpositions of states of different The results can be described in each temporal stage with a well bonding character, for example Rydberg states and valence states. Or, defined, simplified, dynamical characterization. The early time dynamics as in Remacle and Levine (19), the superposition is of states that are

Levine PNAS | December 26, 2017 | vol. 114 | no. 52 | 13595 Downloaded by guest on September 28, 2021 localized at other ends of a small peptide. As this superposition evolves, diagonalizing the electronic Hamiltonian so as to generate the possible the charge moves from one end of the peptide to the other. What it takes potential energy surfaces on which the atoms will move. Here, too, we to access such a superposition is a fast enough perturbation of the elec- can diagonalize the electronic Hamiltonian, but together with the in- tronic states, typically excitation or ionization by a sub–10-fs laser: fast teraction of the electrons with the electric field of the laser. Such po- enough, meaning that the energy span allows accessing two or more tential energy surfaces will undulate with time and may bring down states. In the simpler cases, when only two electronic states are super- static barriers, thereby allowing new modes of laser-dressed reac- posed, one has an electronic version of the Schroedinger cat. The system tivity. When the electric field varies slowly enough, such an adia- is neither in one definite electronic state nor in the other; it is (coherently) in batic approach is realistic. We have seen it working, for example, in both states. A particular state will be selected when a reaction or dissipa- C60 excited by an IR pulse where a collective motion of electrons tion occurs. Such decoherence is typically driven by the motion of the responds to the laser. In molecules with a less-regular level structure nuclei. As Born and Oppenheimer first observed, it takes time for the heavy and for higher carrier frequencies, the electrons will not quite track nuclei to notice that the light electrons are up to something. So there is the laser field so that the dynamics are not adiabatic (20, 21). At the some short delay before the nuclei respond to an abrupt onset of elec- higher energies that we deal with, it is typical for more than one tronic disequilibrium. What happens next must be analyzed within a frame- product channel to be open. In such reactions it is to be expected work that takes proper account of the correlations between electronic and that one can control the branching ratios. The polarization of the nuclear motions. Otherwise one can identify a decoherence that is due to pulse is an obvious parameter. But also the orientation of the electric approximate dynamics rather than the correct physics. When proper care is field matters because the transition is most driven at the peak of taken, the coherence is often maintained for many vibrational periods. This the field and at the peak the field can be parallel or antiparallel to is so even in a condensed medium, which is not fully unexpected because the molecular transition dipole. An electronic transition dipole is it takes time for a solvent to respond to changes in an embedded system. the distance that charge moves upon the electronic excitation. One Can there be applications in physics that take advantage of the can therefore send the charge in a molecule toward different high fields and short times? For example, coherent transfer between directions. molecules might in the future allow ultrafast information transmission Will we be able to control electrophilicity of reagents? An organic on a massive scale. The molecules can be randomly arranged or in an will welcome such a tool. The electric field can also align the ordered array. Indeed, understanding 2D layers (e.g., grapheme) is molecule and thereby affect steric requirements of reactions involving fast becoming a very active field of research. this molecule. One day we may generalize proton-coupled electron is originally about driving chemical reactions that transfer to processes where even heavier nuclei are made to move. At are not possible by thermal means. But current sub–10-fs pulses pro- the higher energies there are many conical intersections. Can we learn vide too few photons to make preparative chemistry of interest. In part, from Peters et al. (8) and guide the molecule to those intersections that this is because delivering many photons in a short time will make the are of primary interest? power of the pump so high that molecules will be extensively ionized. Still, can one conceive of a new chemistry that can be made possible Acknowledgments using initial states that are electronically not stationary? A simple ap- I thank Benoit Mignolet, Universit ´eof Liège, for the figure. My work was sup- proach that is very useful, except that it does not always work, is to learn ported by the US Department of Energy, Office of Science, Basic Energy Sci- from reaction dynamics. Toward understanding reactivity, we begin by ences, under Award DE-SC0012628.

1 Ramasesha K, Leone SR, Neumark DM (2016) Real-time probing of electron dynamics using attosecond time-resolved . Annu Rev Phys Chem 67:41–63. 2 Dorney KM, et al. (2017) Helicity-selective enhancement and polarization control of attosecond high harmonic waveforms driven by bichromatic circularly polarized laser fields. Phys Rev Lett 119:063201. 3 Cao W, Warrick ER, Fidler A, Neumark DM, Leone SR (2016) Noncollinear wave mixing of attosecond XUV and few-cycle optical laser pulses in gas-phase atoms: Toward multidimensional spectroscopy involving XUV excitations. Phys Rev A 94:053846. 4 Matsika S, Krause P (2011) Nonadiabatic events and conical intersections. Annu Rev Phys Chem 62:621–643. 5 Domcke W, Yarkony DR (2012) Role of conical intersections in molecular spectroscopy and photoinduced chemical dynamics. Annu Rev Phys Chem 63:325–352. 6 Bernstein ER (2014) On the release of stored energy from energetic materials. Adv Quantum Chem 69:31–69. 7 Dlott DD (2003) Fast molecular processes in energetic materials. Energetic Materials Part 2. Detonation, Combustion, Theoretical and , eds Politzer P, Murray J (Elsevier, Amsterdam), pp 125–191. 8 Peters WK, et al. (2017) Ultrafast 25-fs relaxation in highly excited states of methyl azide mediated by strong nonadiabatic coupling. Proc Natl Acad Sci USA 114:E11072–E11081. 9 Galbraith MCE, et al. (2017) Few-femtosecond passage of conical intersections in the benzene cation. Nat Commun 8:1018. 10 Hockett P, Bisgaard CZ, Clarkin OJ, Stolow A (2011) Time-resolved imaging of purely valence-electron dynamics during a chemical reaction. Nat Phys 7:612–615. 11 Suzuki T (2014) Ultrafast electronic dynamics in polyatomic molecules studied using femtosecond vacuum ultraviolet and X-ray pulses. J Phys At Mol Opt Phys 47:124001. 12 Wu G, et al. (2015) Excited state non-adiabatic dynamics of pyrrole: A time-resolved photoelectron spectroscopy and quantum dynamics study. J Chem Phys 142:074302. 13 Levine BG, Mart´ınez TJ (2007) Isomerization through conical intersections. Annu Rev Phys Chem 58:613–634. 14 Martinez TJ, Ben-Nun M, Levine RD (1996) Multi-electronic-state molecular dynamics: A wave function approach with applications. J Phys Chem 100:7884–7895. 15 Fuss W, et al. (1998) Pathway approach to ultrafast photochemistry: Potential surfaces, conical intersections and isomerizations of small polyenes. Chem Phys 232:161–174. 16 Fuss W, Schmid WE, Trushin SA (2000) Time-resolved dissociative intense- laser field ionization for probing dynamics: Femtosecond photochemical ring opening of 1,3-cyclohexadiene. J Chem Phys 112:8347–8362. 17 Zewail AH (1996) : Recent progress in studies of dynamics and control of reactions and their transition states. J Phys Chem 100:12701–12724. 18 Zewail AH, Bernstein RB (1988) Special report: Real-time laser femtochemistry—Viewing the transition from reagents to products. Chem Eng News 66:24–43. 19 Remacle F, Levine RD (2006) An electronic time scale in chemistry. Proc Natl Acad Sci USA 103:6793–6798. 20 Fan L, et al. (2017) A new electron-ion coincidence 3D momentum-imaging method and its application in probing strong field dynamics of 2-phenylethyl-N, N-dimethylamine. J Chem Phys 147:013920. 21 Mignolet B, Levine RD, Remacle F (2014) Charge migration in the bifunctional PENNA cation induced and probed by ultrafast ionization: A dynamical study. J Phys At Mol Opt Phys 47:6721–6729.

13596 | www.pnas.org/cgi/doi/10.1073/pnas.1718814114 Levine Downloaded by guest on September 28, 2021