Photoinitiated Quantum Molecular Dynamics: Concluding Remarks Michael N. R. Ashfold School of Chemistry, University of Bristol, Bristol, UK, BS8 1TS Photochemistry and molecular photophysics have long been very active fields of research. Recent advances in both experimental technology and computational methodology have triggered further spectacular progress in these areas – exploring, understanding and even controlling photochemistry and photophysics at the quantum level, and greatly expanding the range and complexity of photochemical systems amenable to study. The frequency of recent Faraday Discussions on related themes, e.g. Molecular Reaction Dynamics in Gases, Liquids and Interfaces (FD157, Assisi, 6/2012), Coherence and Control (FD153, Leeds, 7/2011), Frontiers in Spectroscopy (FD150, Basel) and Cold and Ultracold Molecules (FD142, Durham, 4/2009) is testimony to this burgeoning interest. Faraday Discussions addressing photochemical processes have a long and distinguished tradition. The first, on Photochemical Reactions in Liquids and Gases, was held at Jesus College, Oxford, in 1925, and published in the Transactions of the Faraday Society in 1926.1 That Discussion featured two introductory addresses, by Allmand (on Einstein’s Law of Photochemical Equivalence) and by Bodenstein (on the Mechanisms of Photochemical Reactions) and a summary presentation by Rideal, which opened with the sentence ‘It must be admitted that there is more divergence in opinion than unanimity in views as to the mode of operation of photochemical processes’. The field had advanced considerably by the time of the 1931 Discussion on Photochemical Processes, held in Liverpool.2 The programme of this Discussion sub-divided into four parts, with sessions on Molecular Spectra in Relation to Photochemical Change, Photochemical Kinetics in Gaseous Systems, Photochemical Change in Liquids and Solids and Photosynthesis. Mecke gave the opening lecture in the first session. His paper contains sentences like ‘But we can go still further now and make the general statement, that the appearance of a continuous, i.e. non-quantised, absorption spectrum of a gas is always a sign of photochemical dissociation taking place in the gas’ and ‘It still remains to explain why in 1 one spectrum (e.g., the ultraviolet absorption spectrum of oxygen) the convergence limit is directly observed, whilst in another absorption spectrum of the same gas (i.e., red absorption spectrum of O2) it is missing. The explanation is due to Condon-Franck’ – whose ideas he then proceeded to describe. Such interpretations still feature prominently in undergraduate textbooks of today. One other Discussion certainly deserves reference in any retrospective section such as this – the 1962 Discussion at St Andrews University in Dundee on The Structure of Electronically Excited Species in the Gas Phase.3 Herzberg presented the Introductory Lecture, and the list of those contributing papers (e.g. Buckingham, Coulson, Dixon, Douglas, Kuppermann, Linnett, Longuet-Higgins, Price, Ramsay, Thrush, van der Waals, Walsh, Watson, etc) reads like a who’s who of the leading proponents of molecular spectroscopy, structure and bonding at that time. This 1962 Discussion featured the seminal paper by Herzberg and LonguetHiggins describing the intersection of potential energy surfaces (PESs) in polyatomic molecules 4 that was highlighted in the Introductory Lecture to the present Discussion.5 As with the 1931 Discussion, concepts aired and honed in the 1962 Discussion now feature in contemporary textbooks. In concluding this retrospective, it is marvellous to note the sense of continuity provided by Dr Mike Hollas – a young contributor at the 1962 Discussion 6 and a participant at the present Discussion. FD163 has also focussed on core fundamental science, but the potential impact of photochemistry and molecular photophysics now stretch much more widely. There are myriad reasons for developing a better understanding of the coupling and flow of energy between light, electrons and chemical bonds in molecules, not least to enable this community to contribute to the challenges of designing light harvesting systems for clean energy generation, and to diverse fields like photocatalysis, the design of efficient light-driven molecular devices for data storage and processing, and photomedicine. Four over-arching themes were identified when planning the present Discussion: Single molecules: photochemistry and photophysics in isolated molecular systems; Extended systems: photochemistry and photophysics of chromophores in proteins, solution or clusters; Controlling molecular dynamics: controlling photochemistry using sequences of light pulses, shaped light pulses or bond selection prior to photoexcitation; and Applications of molecular dynamics to global challenges: photovoltaic cells, photodynamic therapy, imaging. It is too early to predict the likely lasting legacy of the work presented at this Discussion, and I do not intend to provide a blow-by-blow summary of all the presentations and the accompanying 2 debate. This can be gained by reading the Discussion volume. I am very happy to offer a few personal observations however. The Introductory Lecturer (Professor Stolow, NRC Ottawa) provided a masterly overview of the current state-of-the-art couched in terms of ‘three pillars’ of light-matter interaction: structure, function and dynamics, viewed from the perspectives of energy/time, phase/coherence and intensity.5 Stolow posed and offered answers to many of the current ‘hot’ issues in the field, and set the tone for much of the ensuing discussion. For example, are there compelling reasons why most natural photoprocesses (e.g. vision, photostability and/or photoprotection mechanisms in DNA and DNA bases, etc) occur on ultrafast timescales? How applicable is the quasi-static model (developed to account for strong field ionisation in atoms) to polyatomic molecules? What are the relevant timescales when thinking about electron dynamics in regions of conical intersection (CI) between PESs? The contributed papers that followed spanned all of the four target themes though, perhaps inevitably, those addressing applications of molecular dynamics to global challenges like light harvesting, phototherapy, etc. could only offer a taste of (selected aspects of) the current state-of-the-art in these fields; each could readily form the basis of a Faraday Discussion in their own right. Were there topics that I had expected to see represented more strongly? Yes, I had anticipated more contributions addressing quantum coherence effects,7 and one or more papers reporting dynamical studies initiated by short pulse XUV radiation. In the event, the late-breaking contribution from Professor Scholes (Toronto) describing 2-D spectroscopy studies of two phycobiliproteins (both of which act as successful light harvesting complexes in nature, but only one of which is found to show measurable coherences between exciton states) served to provoke much interesting discussion of quantum coherence effects on the final morning. I am also aware that the Faraday Standing Committee on Conferences is already addressing my other perceived ‘omission’: FD171 (Sheffield, July 2014) will be devoted to Emerging Photon Technologies for Chemical Dynamics. As at any Discussion, the contributed papers triggered many complementary short presentations and a wealth of specific questions and comments, amongst which I identified several recurring issues. The transient spectra obtained in condensed phase pump-probe experiments (even those involving relatively small molecules) are generally comprised of many broad, overlapped features. Interpreting such data usually involves some form of global analysis, but is made challenging by our imperfect knowledge of the number, nature and time evolution of the constituent basis functions used in the decomposition. Riedle and co-workers 3 8 showed the clear benefits of recording such broadband spectra over the widest possible range of probe wavelengths (which, ideally, would stretch into the infrared (IR) fundamental region also), but rigorous analysis and interpretation of such transient spectra remains a challenge. Challenges are by no means limited to the experimental domain, however! The photochemistry/physics community is attempting to tackle ever bigger, more complex systems, and the theorists in the audience were repeatedly quizzed regarding the most appropriate form of theory to address a particular problem, and what confidence any nonexpert should have in the results of different calculation types. The relentless progress in experiment and theory is enabling study of an ever greater range of systems; indeed, the range of chemistry now accessible to the photophysics community far outstrips that which can realistically be investigated. Hence the suggestion of a consortium approach, wherein a limited number of key systems are identified for detailed investigation using the full armoury of available methods (experimental and theoretical) – a concept pioneered to great effect by Delbrück’s promotion of the Phage Group and the exploration of genetics through researches on bacterial viruses. Two further themes of the Discussion relate closely to recent work in our own group, and I claim summariser’s prerogative to conclude this presentation by highlighting some such links. One discussant enquired ‘passive control – is that not just chemistry?’ to which the answer is surely yes. This Discussion provides several illustrations of just how effective chemistry can be at controlling photophysical outcomes. For example, Rothlisberger and coworkers 9 provide a theoretical rationale for the observation 10 that C–NH3+ bond fission following UV excitation of the tryptophan chromophore in a ‘dry’ Lys +-Trp pair can be completely suppressed by solvating the amino group with just two water molecules. They also find that this photo-protection mechanism (which can be traced to a relative increase in the energy of the dissociative states formed by σCN*π excitation) is maintained under fully solvated conditions. In a similar vein, Meech and co-workers 11 find that the excited state photophysics of both the neutral and anionic states of the chromophores of green fluorescent protein (GFP) and the kindling fluorescent protein in methanol solution are very similar, yet their photophysical behaviour in the two proteins (i.e. in the presence of the respective protein scaffolds) are strikingly different. These examples highlight the potential sensitivity of excited states, and their photophysics, to intermolecular interactions with the local environment. Photophysics can also be ‘tuned’ by intramolecular modifications – as illustrated in Professor Stolow’s overview of the 4 contrasting near UV photochemistry of acrolein and other ,-enone isomers 5 and in the paper by Sension and co-workers on the photo-induced ring-opening of various substituted 1,3-cyclohexadienes.12 The data shown in fig. 1 provide another example. Excitation to the S1(1*) state of phenol in the gas phase results in O–H bond fission, by tunnelling through the barrier under the CI between the 1* and the dissociative 1* PESs – as discussed by Ramesh and Domcke in this Discussion.13 4-substitution affects the relative energies of the two PESs, and thus the magnitude of the barrier under the CI and the tunnelling probability from the 1* state. Substituting a strong electron withdrawing group like CN completely suppresses this O–H bond fission channel.14 The choice of substitution site can also be important. O–H bond fission is just one of several possible decay routes from the 1* state of phenol – others include internal conversion (IC, to the ground (S0) state) and intersystem crossing. Each of these non-radiative processes must be relatively slow in phenol, given the measured fluorescence lifetime of the 1* state (~2.2 ns).15 The corresponding 1* state of 4-chlorophenol has a similar lifetime (~1.6 ns 16) yet, as fig. 2 shows, recent pump-probe studies confirm earlier suggestions 17 that the same 1* state in 2-chlorophenol decays on a sub-ps timescale. This orders of magnitude increase in loss rate may be attributable to a much enhanced IC rate, enabled by a reduction in the energy of the lowest CI between the 1* and S0 states. Calculations show this to involve a prefulvenic configuration which, in 2chlorophenol, is stabilised by intramolecular H-bonding between the hydroxyl H atom and the neighbouring Cl atom.18 Substitution can also affect the branching into different product channels. Consider thiophenol. The PESs for the S0, 1* and 1* states are qualitatively similar to those for phenol, but the barrier to S–H bond fission by tunnelling under the 1*/1* CI is much smaller. Thus the H atom products, even those formed when exciting at the 1*S0 origin, display anisotropic recoil velocity distributions; i.e. the timescale for fragmentation is shorter than the parent rotational period.19 The 1* PES of thiophenol (and phenol) exhibits another CI, with the S0 PES, at larger S–H(O–H) bond lengths. The energy separation between the ground (X) and first excited (A) electronic states of the resulting radicals is much smaller in the case of thiophenoxyl, and the radical products from near UV photolysis of thiophenol are formed in both of these states. 4-substitution can have a dramatic effect on the A/X product branching ratio, which decreases from ~0.8 to ~0.6 as the excitation wavelength is tuned from that of the 1*S0 origin to ~250 nm in the case of 4-methylthiophenol, but increases from 5 ~1.3 to >3 (i.e. a population inversion throughout) across the same wavelength range when the methyl is replaced by a methoxy substituent. Branching is established in the vicinity of the 1*/S0 CI, but is largely controlled by the orientation of the S–H bond (relative to the ring plane) as the dissociating molecules approach this CI.19 Such ‘passive’ control may well be ‘just chemistry’ but, as with so much of chemistry, it can be very influential! We end by revisiting the question ‘to what extent can dynamical insights gleaned from detailed studies of a photo-induced process in the gas phase inform our understanding of the equivalent process in solution?’ As noted above, Rothlisberger and coworkers 9 find that the excited state photophysics displayed by the Lys +-Trp pair in the gas phase and in aqueous solution are very different, but the discussion of the paper by Bochenkova and Andersen noted evident similarities in the decay kinetics measured for the anionic GFP chromophore in the gas and solution phases.20,21 Much of our recent activity has focussed on comparing selected benchmark photodissociation processes and bimolecular reactions in the gas and solution phases. For example, investigations of the near UV photodissociation of phenol and of various thiophenols and thioanisoles, in the gas phase and in weakly interacting solvents like acetonitrile and cyclohexane, show that the gas phase results provide an excellent guide to the early time dynamics prevailing in solution though, clearly, additional processes (e.g. vibrational relaxation, geminate recombination, adduct formation, etc) specific to the condensed phase reveal themselves on longer timescales.15,22,23 Similar conclusions have been reached in photo-initiated studies of F and Cl atoms, and of CN radicals, with alkanes. For example, transient broadband IR probe measurements of the HCN products formed in the reaction of CN radicals with cyclohexane in a range of chlorinated organic solvents reveal preferential excitation of one quantum of the CH stretch and up to two quanta in the bending mode, which is progressively quenched through solvent interaction on the longer timescale.24 Such state-specific vibrational energy disposal at early times parallels that found in similar CN + alkane reactions in the gas phase,25 though the extent of vibrational relaxation is less – consistent with partial damping of the developing HCN vibrational motion in the posttransition state region.26 Many more such comparisons involving larger, more complex systems and a wider range of solvents (particularly water) can be anticipated in the coming months and years, which might well form the basis for a future Faraday Discussion! I conclude by congratulating, and thanking, the organizers of FD163 – Helen Fielding, Steve Leone, Christoph Meier, Andrew Orr-Ewing, Katharine Reid and Graham Worth – for proposing and coordinating such an interesting and thought provoking meeting. 6 Figure 1 Total kinetic energy release (TKER) spectra obtained from time-of-flight spectra of H atoms formed in the photolysis of (a) phenol at = 275.113 nm and (b) 4-cyanophenol at = 283.998 nm. The peaks in (a) are attributable to formation of phenoxyl radicals, in selected vibrational levels of the ground electronic state, as a result of O–H bond fission by tunnelling through the barrier under the CI between the 1* and 1* PESs. 7 Figure 2 Time-resolved parent ion signal obtained following excitation of 2-chlorophenol at the origin of the 1*S0 transition of the cis-isomer and probing (ionization) at 243.1 nm. ‘One colour’ contributions from the pump and probe lasers alone have been subtracted prior to display. The solid curve (obtained by convoluting a 120 fs (Gaussian) instrument response function and a single exponential decay with time constant 260 fs) is intended simply to illustrate the rapidity of the excited state decay (cf. 1.6 ns for the corresponding state of 4-chlorophenol). The calculated prefulvenic structure of the lowest energy CI between the 1* and S0 PESs is shown on the right. 8 References 1 Trans. Faraday Soc., 1926, 21, 437. 2 Trans. Faraday Soc., 1931, 27, 357. 3 Discuss. Faraday Soc., 1963, 35, 1. 4 G. Herzberg and H.C. Longuet-Higgins, Discuss. Faraday Soc., 1963, 35, 77. 5 A. Stolow, paper 1 of FD163. 6 J.M. Hollas, Discuss. Faraday Soc., 1963, 35, 233. 7 See, for example, G.D. Scholes, G.R. Fleming, A. Olaya-Castro and R. von Grondelle, Nature Chem. 2011, 3, 763 and references therein. 8 E. Riedle, M. Bradler, M. Wenninger, C.F. Sailer and I. Pugliesi, paper 6 of FD163. 9 M. Guglielmi, M. Doemer, I. Tavernelli and U. Rothlisberger, paper 9 of FD163. 10 S. Mercier, O.V. Boyarkin, A. Kamariotis, M. Guglielmi, I. Tavernelli, M. Cascella, U. Rothlisberger and T.R. Rizzo, J. Am. Chem. Soc. 2006, 128, 16938. 11 K. Addison, I.A. Heisler, J. Conyard, T. Dixon, P.C. Bulman Page and S.R. Meech, paper 12 of FD163. 12 B.C. Arruda, B. Smith, K.G. Spears and R.J. Sension, paper 7 of FD163. 13 S.G. Ramesh and W. Domcke, paper 4 of FD163. 14 T.N.V. Karsili, A.M. Wenge, S.J. Harris, D. Murdock, J.N. Harvey, R.N. Dixon and M.N.R. Ashfold, Chem. Sci. 2013, DOI: 10.1039/C3SC50296A. 15 S.J. Harris, D. Murdock, Y. Zhang, T.A.A. Oliver, M.P. Grubb, A.J. Orr-Ewing, G.M. Greetham, I.P. Grant, M. Towrie, S.E. Bradforth and M.N.R. Ashfold, Phys. Chem. Chem. Phys., 2013, 15, 6567 16 M. Böhm, C. Ratzer and M. Schmitt, J. Mol. Struc., 2006, 800, 55 17 S. Yamamoto, T. Ebata and M. Ito, J. Phys. Chem. 1989, 93, 6340. 18 S.J. Harris, T.N.V. Karsili, D.J. Hadden, G.M. Roberts, V.G. Stavros and M.N.R. Ashfold, unpublished results. 19 T.A.A. Oliver, G.A. King, D.P. Tew, R.N. Dixon and M.N.R. Ashfold, J. Phys. Chem. A 2012, 116, 12444. 20 21 A.V. Bochenkova and L.H. Andersen, paper 13 of FD163. C.R.S. Mooney, D.A. Horke, A.S. Chatterley, A. Simperler, H.H. Fielding and J.R.R. Verlet, Chem. Sci., 2013, 4 921. 22 T.A.A. Oliver, Y. Zhang, M.N.R. Ashfold and S.E. Bradforth, Farad. Disc. Chem. Soc. 2011, 150, 439. 9 23 Y. Zhang, T.A.A. Oliver, M.N.R. Ashfold and S.E. Bradforth, Farad. Disc. Chem. Soc. 2012, 157, 141. 24 S.J. Greaves, R.A. Rose, T.A.A. Oliver, D.R. Glowacki, M.N.R. Ashfold, J.N. Harvey, I.P. Clark, G.M. Greetham, A.W. Parker, M. Towrie and A.J. Orr-Ewing, Science 2011, 331, 1423. 25 C. Huang, W. Li, A.D. Estillore and A.G. Suits, J. Chem. Phys. 2008, 129, 074301 and references therein. 26 D.R. Glowacki, R.A. Rose, S.J. Greaves, A.J. Orr-Ewing and J.N. Harvey, Nature Chem. 2011, 3, 850. 10