Probing ultrafast dynamics during and after passing through conical intersections

Shunsuke Adachi a, Tom Schatteburg b, Alexander Humeniuk c, Roland Mitrić c and Toshinori Suzuki *a
aDepartment of Chemistry, Graduate School of Science, Kyoto University, Kitashirakawa Oiwakecho, Sakyo-ku, Kyoto 606-8502, Japan. E-mail:
bLaboratorium für Physikalische Chemie, ETH Zürich, Vladimir-Prelog-Weg 1, 8093 Zürich, Switzerland
cInstitut für Physikalische und Theoretische Chemie, Julius-Maximilians-Universität Würzburg, Emil-Fischer-Str. 42, D-97074 Würzburg, Germany

Received 13th July 2018 , Accepted 24th September 2018

First published on 24th September 2018

Time-resolved photoelectron spectroscopy using vacuum-UV probe pulses enables observing ultrafast dynamics during and after passing through conical intersections (CIs). The ring-puckering CI plays a prominent role following the ππ* photoexcitation of furan. More than 90% of the excited molecules safely return to the original ground state, while the remaining 10% transforms into isomers after passing through the puckering CI.

Ultrafast nonradiative deactivation of polyatomic molecules generally involves non-adiabatic transitions around conical intersections (CIs).1 Advances in computational chemistry have enabled identifying these CIs and predicting possible nonradiative deactivation pathways.2–5 The electronic excited states often have multiple CIs (or more properly minimum energy CIs) at different geometries, which facilitate non-adiabatic transitions to different electronic states and ultimately lead to different products.3 For example, in UV photochemistry of heterocyclic molecules, there are generally two types of CIs: the ring-puckering and ring-opening CIs. The former appears to facilitate ultrafast internal conversion to the original ground state and is often associated with high photostability of heterocycles.2,6–8 However, quantitative prediction of these branching ratios and photoproduct quantum yields (QYs) remains as a challenging problem in theoretical chemistry.

Time-resolved photoelectron spectroscopy (TRPES) provides valuable insights into the non-radiative dynamics and photochemistry. Previous TRPES studies using deep-UV (DUV, < 6 eV) probe pulses have elucidated the wavepacket dynamics from the Franck–Condon (FC) region to CIs on the excited-state potential energy surface.2,8–12 Recent introduction of vacuum-UV (VUV) probe pulses have further strengthened TRPES by enabling observation of cascaded electronic deactivations down to the ground state in pyrazine13 and photoproduct formation in a ring-opening reaction of 1,3-cyclohexadiene.14 Here we present a VUV-TRPES study of furan, a prototypical heterocyclic molecule, using 14 eV probe pulses. The main relaxation pathway, photoproducts, and their QYs become evident from the TRPES spectra, all of which have been debated in previous theoretical calculations.2,8

The experimental apparatus used in this study has been described in detail previously.14,15 Briefly, the 14 eV probe pulse was obtained by UV-driven harmonic generation,15 while the 6.0 eV pump pulse was generated using nonlinear crystals. A continuous molecular beam of furan (>99.0%, Tokyo Chemical Industry, diluted with helium carrier gas) was formed from a nozzle inside a magnetic-bottle time-of-flight photoelectron spectrometer. The photoelectrons generated were detected using a micro-channel plate at the end of a 1 m long flight tube, and the signal was acquired using an analog-to-digital converter. The pump–probe cross-correlation time was estimated to be 80 fs using photoionization of nitric oxide. The non-adiabatic dynamics simulations performed here are an extension of the work in ref. 2. We employ the time-dependent density functional theory (TDDFT) combined with the Tully's surface hopping procedure, which is described in detail in ref. 16. The energies, gradients and nonadiabatic couplings needed to carry out the nonadiabatic dynamics have been calculated “on the fly” using the hybrid PBE0 functional17 combined with the 6-311++G** basis set18 also containing diffuse functions, which is suitable for accurate description of stationary absorption properties and dynamics of furan.2 The simulations have been performed in a manifold consisting of the ground and the three lowest excited states and have been extended to 2 ps in order to follow the photochemical reactions of hot photoproducts after returning to the ground state. Along each trajectory, the vertical ionization energy from the current electronic state to the cation ground state was computed. Although the cationic excited states are also energetically accessible, their contributions were excluded based on the independent electron approximation (see ESI, for more details).

In this study, a pump pulse (pump = 6.0 eV) creates a vibrational wavepacket in the FC region of the lowest lying 1B2(ππ*) electronic excited state of furan. As shown in Fig. 1, the subsequent electronic deactivation may proceed through dual pathways of (1) ring-puckering of C–O–C subunit and (2) ring-opening of C–O bond. In (1), the wavepacket moves on the ππ* excited-state surface toward a ring-puckering CI (CIP),2 while in (2), the ππ* state changes its character to a πσ* state, on which the wavepacket moves toward a ring-opening CI (CIO).8 Wavepacket dynamics in the excited states manifest themselves in lower electron binding energy (eBE) regions of VUV-TRPES spectra.14,19 Here eBE is the energy required to remove an electron from a molecule and is given by the difference between the probe photon energy (probe) and observed photoelectron kinetic energy. Fig. 2a shows a lower eBE region (<7 eV) of the measured TRPES spectra (see ESI, for the entire TRPES spectra). The time profile measured at each eBE has been normalized to highlight the changes in spectral shape. In Fig. 2a, it is evident that the spectral peak (see dashed curve for eye guide) gradually shifts toward higher eBE by several electronvolts within initial 150 fs. Since CIs accessible to the excited-state wavepacket lie energetically lower than the FC region, eBE increases with the wavepacket motions toward them on the excited-state surface. [To be more precise, discussion on lower (i.e. neutral) and upper (i.e. cationic) state potential energy surfaces is needed. See ESI, for more details.]

image file: c8cp04426k-f1.tif
Fig. 1 Relaxation scheme of furan from the ππ* excited state. The reaction time scales and branching ratios were obtained in the present study.

image file: c8cp04426k-f2.tif
Fig. 2 Comparison between (a) experiment and (b) simulation. Dash curves are for eye guides. In (a), the time profile measured at each eBE has been normalized to highlight the changes in spectral shape, and the result we reported previously2 using DUV probe pulses (probe = 4.7 eV) is also shown in dotted line.

Interestingly enough, there exists a kink at τ ∼ 90 fs (eBE ≈ 5 eV) in Fig. 2a. The kink is indicative of an abrupt change of the potential gradient, and it appears most natural to assign it to a CI mediating the nonadiabatic transition to the ground state. Also shown in dotted line is the result we reported previously2 using DUV probe pulses (probe = 4.7 eV). It is noted that the kink at eBE ≈ 5 eV was outside of the observation window using the DUV probe pulses, demonstrating the usefulness of the VUV probe pulses employed in the present study. Fig. 2b shows the simulated TRPES spectra by the TDDFT, in which the vertical ionization energies from the neutral excited state to the cationic ground state were convolved with a two-dimensional Gaussian of width σtime = 10 fs and σenergy = 0.5 eV. Here each eBE slice has been normalized to unity, as well as in Fig. 2a, to reveal the spectral peak shift toward higher eBE, and the convolution widths were chosen so that the peak shift becomes conspicuous. It is notable that in our dynamical simulations the nonadiabatic transition mainly proceeds through the puckering pathway. The experimental kink is well reproduced by the simulations, and thus assignable to the puckering CI.

After passing through the puckering CI, the wavepacket returns to the ground state, and thus the ground state population recovers. This can be observed in a higher eBE region (>8.5 eV) of the TRPES spectra (Fig. S1b in ESI). Furan molecules unexcited by the pump pulses provide a one-color background signal (see the one-color spectrum of furan in Fig. S1c, ESI) in this region. Thus, we evaluated the one-color signal from the photoelectron spectra averaged for negative delays (−1450 to −490 fs) and subtracted to obtain Fig. S1b (ESI). Consequently, as is shown in Fig. 3a, the time profile at eBE = 9.1 eV (a vertical ionization energy of ground state furan) shows a “bleach” feature, negative signal intensity frequently observed in transient absorption spectroscopy due to depletion of ground state population.20 With increasing the delay, the bleach signal recovers owing to the ground state repopulation. However, the ground state bleach does not recover fully, suggesting permanent conversion into photoproducts. The observed incomplete bleach recovery is not due to ro-vibrational excitations. It is possible that the TRPES spectra of ro-vibrationally excited molecules after returning to the ground state deviate from those of molecules in thermal equilibrium in terms of spectral shape and intensity. However, as we have shown previously,14,19,21 the influence of ro-vibrational excitation on the shape of a photoelectron spectrum is usually small.

image file: c8cp04426k-f3.tif
Fig. 3 (a) Time profile of photoelectron intensity at eBE = 9.1 eV (symbols) and the result of least-squares fitting (dashed curve). (b) Ground state population determined from the fitting. The region painted by red represents the time of the ground state recovery (τrecovery = 150–260 fs).

The ground state populations were well expressed by the following kinetic equation:2,8

image file: c8cp04426k-t1.tif
As shown in Fig. 3b, the bleach recovery starts and finishes at τ = τ1 and τ1 + τ2, respectively, while η represents QY of photoproducts formation. The result of least-square fitting with this kinetic model convoluted with our apparatus function (Gaussian with 80 fs FWHM) is shown in Fig. 3a (dashed curve), and the ground state population determined from the fitting is shown in Fig. 3b. The fitting result of the ground state time profile appears consistent with the excited state dynamics; the excited state wavepacket reaches the puckering CI at τCI = 90 fs, and returns to the original ground state at τrecovery = (150–260) fs. In addition, the photoproduct QY was determined to be η = 0.09 from the fitting. In other words, >90% of the excited furan molecules safely decays back to the original ground state geometry.

So far we have presented the reaction time scales and branching ratios, as summarized in Fig. 1, and shown that the puckering CI mediates efficient ultrafast nonradiative deactivation to the original ground state. In what follows, we shed light on the products. Fig. 4a (solid curve) shows an experimental pump–probe spectrum averaged for longer delay times (+700–+1500 fs). The spectral intensities were calibrated using reference spectra22,23 (see ESI, for more details on the calibration). The averaged spectrum consists of positive and negative bands that are respectively attributed to photoproducts and the ground state bleach of furan. Thus subtracting the latter [dashed curve in Fig. 4a normalized at lower eBE region (<9 eV) of the spectrum] yields the product spectrum (Fig. 4b). For comparison, calculated vertical ionization energies for possible photoproducts via the puckering CI (oxabicyclopentene and cyclopropene–carbaldehyde) and the ring-opening CI (1,3-, 1,2-, and 2,3-butadienals) are presented with colored symbols in Fig. 4c. Our calculations have reproduced the experimental vertical ionization energies of furan and 1,3-butadienal within 0.3 eV (see Fig. S5 in ESI). In Fig. 4c, expected band positions of photoproducts via the puckering and ring-opening pathways are painted by red and blue, respectively. The experimental product spectrum (Fig. 4b) is reasonably explainable by photoproducts via the puckering pathway (oxabicyclopentene and cyclopropene–carbaldehyde). Meanwhile, Oesterling et al. highlighted the significance of ring-opening trajectories;8 they argued that a good half of the trajectories passing close to the ring-opening CI stays in the excited state lead to ring opening reaction. However, no evidence of the ring-opening trajectories was found neither in the excited-state TRPES spectra (Fig. 2a) nor in the product spectrum (Fig. 4b). Presumably, as they described,8 their complete active space self-consistent field calculations overestimated QY for the ring-opened pathway due to a vanishing energy barrier along the path. In our dynamical simulations, oxabicyclopentene and cyclopropene–carbaldehyde can be formed in a few ps by subsequent ground-state reactions after passing through the puckering CI. Though there exists fairly large uncertainty in the simulated photoproduct QY (η = 0.03–0.3, see ESI, for more details), it appears qualitatively consistent with the experimental QY (η = 0.09).

image file: c8cp04426k-f4.tif
Fig. 4 (a) Experimental TRPES spectrum averaged for later delays (+700–+1500 fs, solid) and ground state bleach spectrum of furan (dashed). (b) Product spectrum. (c) Calculated vertical ionization energies for furan isomers that are potentially formed after passing through the puckering CI (oxabicyclopentene and cyclopropene–carbaldehyde) and the ring-opening CI (1,3-, 1,2-, and 2,3-butadienals). Expected band positions of photoproducts via the puckering and ring-opening pathways are painted by red and blue, respectively.

In summary, when the electronic excited states have multiple CIs, quantitative prediction of branching ratios and photoproduct QYs still remains as a challenging problem in theoretical chemistry. In the TRPES spectra of furan probed by VUV pulses, ultrafast nonradiative deactivation to the ground state, ground state repopulation, and formation of photoproducts were observed simultaneously. The electronic character changed drastically at τCI ∼ 90 fs, which was assigned to the puckering CI mediating the nonadiabatic transition to the ground state. >90% of the excited furan molecules safely decayed back to the original ground state at τrecovery = (150–260) fs. Given the product spectrum, the remaining 10% is supposed to transform into isomers by subsequent reactions after passing through the puckering CI.

Conflicts of interest

There are no conflicts to declare.

Notes and references

  1. W. Domcke, D. Yarkony and H. Köppel, Conical Intersections: Theory, Computation and Experiment, World Scientific, Singapore, 2011 Search PubMed.
  2. T. Fuji, Y.-I. Suzuki, T. Horio, T. Suzuki, R. Mitrić, U. Werner and V. Bonačić-Koutecký, Ultrafast photodynamics of furan, J. Chem. Phys., 2010, 133, 234303 CrossRef.
  3. A. L. Thompson and T. J. Martínez, Time-resolved photoelectron spectroscopy from first principles: excited state dynamics of benzene, Faraday Discuss., 2011, 150, 293 RSC.
  4. S. P. Neville, Y. Wang, A. E. Boguslavskiy, A. Stolow and M. S. Schuurman, Substituent effects on dynamics at conical intersections: allene and methyl allenes, J. Chem. Phys., 2016, 144, 014305 CrossRef PubMed.
  5. A. Humeniuk, M. Wohlgemuth, T. Suzuki and R. Mitrić, Time-resolved photoelectron imaging spectra from non-adiabatic molecular dynamics simulations, J. Chem. Phys., 2013, 139, 134104 CrossRef PubMed.
  6. S. Perun, A. L. Sobolewski and W. Domcke, Ab initio studies on the radiationless decay mechanisms of the lowest excited singlet states of 9H-adenine, J. Am. Chem. Soc., 2005, 127, 6257–6265 CrossRef CAS PubMed.
  7. M. Barbatti, A. J. A. Aquino, J. J. Szymczak, D. Nachtigallova, P. Hobza and H. Lischka, Relaxation mechanisms of UV-photoexcited DNA and RNA nucleobases, Proc. Natl. Acad. Sci. U. S. A., 2010, 107, 21453–21458 CrossRef CAS PubMed.
  8. S. Oesterling, O. Schalk, T. Geng, R. D. Thomas, T. Hansson and R. de Vivie-Riedle, Substituent effects on the relaxation dynamics of furan, furfural and β-furfural: a combined theoretical and experimental approach, Phys. Chem. Chem. Phys., 2017, 19, 2025–2035 RSC.
  9. H. Kang, K. T. Lee, B. Jung, Y. J. Ko and S. K. Kim, Intrinsic lifetimes of the excited state of DNA and RNA bases, J. Am. Chem. Soc., 2002, 124, 12958–12959 CrossRef CAS PubMed.
  10. S. Ullrich, T. Schultz, M. Z. Zgierski and A. Stolow, Electronic relaxation dynamics in DNA and RNA bases studied by time-resolved photoelectron spectroscopy, Phys. Chem. Chem. Phys., 2004, 6, 2796 RSC.
  11. H. Satzger, D. Townsend, M. Z. Zgierski, S. Patchkovskii, S. Ullrich and A. Stolow, Primary processes underlying the photostability of isolated DNA bases: adenine, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 10196–10201 CrossRef CAS.
  12. H. R. Hudock, B. G. Levine, A. L. Thompson, H. Satzger, D. Townsend, N. Gador, S. Ullrich, A. Stolow and T. J. Martínez, Ab initio molecular dynamics and time-resolved photoelectron spectroscopy of electronically excited uracil and thymine, J. Phys. Chem. A, 2007, 111, 8500–8508 CrossRef CAS PubMed.
  13. T. Horio, R. Spesyvtsev, K. Nagashima, R. A. Ingle, Y. Suzuki and T. Suzuki, Full observation of ultrafast cascaded radiationless transitions from S2(ππ*) state of pyrazine using vacuum ultraviolet photoelectron imaging, J. Chem. Phys., 2016, 145, 044306 CrossRef PubMed.
  14. S. Adachi, M. Sato and T. Suzuki, Direct observation of ground-state product formation in a 1,3-cyclohexadiene ring-opening reaction, J. Phys. Chem. Lett., 2015, 6, 343–346 CrossRef CAS PubMed.
  15. S. Adachi, T. Horio and T. Suzuki, Generation of intense single-order harmonic pulse in the vacuum ultraviolet region using a deep ultraviolet driving laser, Opt. Lett., 2012, 37, 2118–2120 CrossRef CAS.
  16. R. Mitrić, U. Werner and V. Bonačić-Koutecký, Nonadiabatic dynamics and simulation of time resolved photoelectron spectra within time-dependent density functional theory: ultrafast photoswitching in benzylideneaniline, J. Chem. Phys., 2008, 129, 164118 CrossRef.
  17. C. Adamo and V. Barone, Toward reliable density functional methods without adjustable parameters: the PBE0 model, J. Chem. Phys., 1999, 110, 6158–6170 CrossRef CAS.
  18. R. Krishnan, J. S. Binkley, R. Seeger and J. A. Pople, Self-consistent molecular orbital methods. XX. A basis set for correlated wave functions, J. Chem. Phys., 1980, 72, 650–654 CrossRef CAS.
  19. S. Adachi, H. Kohguchi and T. Suzuki, Unravelling the Electronic State of NO2 Product in Ultrafast Photodissociation of Nitromethane, J. Phys. Chem. Lett., 2018, 9, 270–273 CrossRef CAS.
  20. R. Berera, R. van Grondelle and J. T. M. Kennis, Ultrafast transient absorption spectroscopy: principles and application to photosynthetic systems, Photosynth. Res., 2009, 101, 105–118 CrossRef CAS.
  21. M. Sato, S. Adachi and T. Suzuki, Photoisomerization of Vibrationally Hot Tetramethylethylene Produced by Ultrafast Internal Conversion from the Excited State, J. Phys. Chem. A, 2016, 120, 5099–5102 CrossRef CAS.
  22. K. Kimura, S. Katsumata, Y. Achiba, T. Yamazaki and S. Iwata, Handbook of HeI Photoelectron Spectra of Fundamental Organic Compounds, Japan Scientific Soc. Press, Tokyo, 1981 Search PubMed.
  23. A. Ponzi, M. Sapunar, C. Angeli, R. Cimiraglia, N. Došlić and P. Decleva, Photoionization of furan from the ground and excited electronic states, J. Chem. Phys., 2016, 144, 084307 CrossRef.


Electronic supplementary information (ESI) available: Entire pump–probe photoelectron spectra, discussion on lower- and upper-state potential energy surfaces, contribution of cationic excited states in the simulated TRPES spectra, spectral intensity calibration using a He(I) reference spectrum, experimental and calculated vertical ionization energies of furan and 1,3-butadienal, and subsequent ground-state reactions after passing through the puckering CI. See DOI: 10.1039/c8cp04426k

This journal is © the Owner Societies 2019