International Symposium on Molecular Spectroscopy 70 TH MEETING

International Symposium on Molecular Spectroscopy 70 TH MEETING - JUNE 22 -26, 2015 CHAMPAIGN-URBANA, ILLINOIS TOWARD THE ACCURATE SIMULATION OF TWO‐DIMENSIONAL ELECTRONIC SPECTRA Angelo Giussani, Artur Nenov, Javier Segarra‐Martí, Vishal K. Jaiswal, Ivan Rivalta, Elise Dumont, Shaul Mukamel, Marco Garavelli Dipartimento di Chimica G. Ciamician, Universit di Bologna, Via F. Selmi 2, 40126 Bologna, Italy. Angelo. Giussani 2@unibo. it

Outline § Introduction on two‐dimensional (2 D) electronic spectroscopy ‐ Basic principles and potentialities § How to compute 2 D electronic spectra? ‐ Static approach ‐ Single‐trajectory approach § Static and single‐trajectory approach at work: the 2 D spectra of pyrene § Conclusions

Introduction on two‐dimensional electronic spectroscopy

2 D electronic spectroscopy is a nonlinear optical technique that measures the full nonlinear polarization of a quantum system in third order (R 3) with respect to the field‐matter interaction. Schematic representation of experimental setup for heterodyne detected three‐pulses photon echo k 1, k 2: Pump signal split in two pulses. Varying the t 1 time and through a subsequent Fourier transformation the spectral resolution for the pump frequency is obtained. k 3: Probe signal, probing the sample after t 2 time of the pump interaction. Rivalta I, Nenov A, Cerullo G, Mukamel S, Garavelli M, Int. J. Quantum Chem. , 2014, 114, 85

2 D electronic spectra Pyrene 2 D spectrum at t 2=0 Red; excitation/de‐excitation to/from the pump reachable excited states Blue ; excitation to a higher excited state pyrene electronic states f pump Lb dark La e bright g La trace

2 D electronic spectra Pyrene 2 D spectrum at t 2=0 Red; excitation/de‐excitation to/from the pump reachable excited states Blue ; excitation to a higher excited state GS bleaching pyrene electronic states f pump Lb dark La e bright g La trace

2 D electronic spectra Pyrene 2 D spectrum at t 2=0 ES absorption Red; excitation/de‐excitation to/from the pump reachable excited states Blue ; excitation to a higher excited state GS bleaching pyrene electronic states f ES absorption pump Lb dark La e bright g La trace

2 D electronic spectroscopy; main potentialities Potentially, 2 D electronic spectroscopy provides both high spectral and temporal resolution and can clearly distinguish between different chromophores and/or bright states present in a system. This in turn will in principle allow to: • resolve chromophore interactions • disentangles de‐excitation pathways • resolves population transfer Model peptide containing benzene and phenol* 1 D 2 D benzene phenol *Nenov A, Beccara S, Rivalta I, Cerullo G, Mukamel S, Garavelli M Chem. Phys. Chem. 2014, 15, 1‐ 10

How to compute 2 D electronic spectra

Nonlinear optical spectroscopy theory* In semi‐impulsive limit (laser pulses shorter than time separation between them) the emitting signal is proportional to the third order response function, R(3) * S. Mukamel, Principles of Nonlinear Optics and Spectroscopy, Oxford Unversity Press, Oxford, UK, 1995.

Mixed quantum‐classical dynamics In mixed quantum‐classical dynamics, the following coupled equations govern the nuclei and electrons respectively: ; Making the approximation of using one single trajectory, instead of a swarm, is equivalent to neglect electronic inter‐state interactions, which translates in:

Mixed quantum‐classical dynamics In mixed quantum‐classical dynamics, the following coupled equations govern the nuclei and electrons respectively: ; Making the approximation of using one single trajectory, instead of a swarm, is equivalent to neglect electronic inter‐state interactions, which translates in: coherence dynamics

Mixed quantum‐classical dynamics In mixed quantum‐classical dynamics, the following coupled equations govern the nuclei and electrons respectively: ; Making the approximation of using one single trajectory, instead of a swarm, is equivalent to neglect electronic inter‐state interactions, which translates in: coherence dynamics

Mixed quantum‐classical dynamics In mixed quantum‐classical dynamics, the following coupled equations govern the nuclei and electrons respectively: ; Making the approximation of using one single trajectory, instead of a swarm, is equivalent to neglect electronic inter‐state interactions, which translates in: coherence dynamics probe pump f e g

Brownian harmonic oscillator model Since the calculation of the coherence dynamics using a single trajectory can cause significant problems, it is preferable to use the analytical solution provided by the Brownian harmonic oscillator model (BHO)*: BHO Where wij is adiabatic transition between the involved electronic states, * S. Mukamel, Principles of Nonlinear Optics and Spectroscopy, Oxford Unversity Press, Oxford, UK, 1995.

Brownian harmonic oscillator model Since the calculation of the coherence dynamics using a single trajectory can cause significant problems, it is preferable to use the analytical solution provided by the Brownian harmonic oscillator model (BHO)*: BHO Where wij is adiabatic transition between the involved electronic states, and g(t) is a function of the Franck‐Condon coefficients wk and Dk g(t) can be obtained using a Fourier series to fit the temporal evolution of the electronic gap Ei(t)-Ej(t) along a computed dynamics on the populated excited state. Since dynamics are expensive, the BHO model is only used for coherence dynamics among ground and first bright state (i. e. along t 1) * S. Mukamel, Principles of Nonlinear Optics and Spectroscopy, Oxford Unversity Press, Oxford, UK, 1995.

Static and single‐trajectory approaches Using the presented approximations, (i. e. quantum‐classical description of the system, neglecting electronic inter‐state interactions, Brownian harmonic oscillator model along t 1) the third order response is written as follows, which is what we referred as the single-trajectory approach:

Static and single‐trajectory approaches Using the presented approximations, (i. e. quantum‐classical description of the system, neglecting electronic inter‐state interactions, Brownian harmonic oscillator model along t 1) the third order response is written as follows, which is what we referred as the single-trajectory approach: Making the further assumption that the coherence dynamics is slower than the duration of the experiment (i. e. that during t 1 and t 3 the system has no enough time to evolve) leads to the following simplification, which constitutes the static approach:

Static and single‐trajectory approaches at work: the 2 D spectra of pyrene

Pyrene photophysics Pyrene is characterized by two low‐lying excited states: the dark 1 pp* Lb state (S ) and the bright 1 pp* La state (S ). Experimental 1 2 evidences show that from the La state the system decays with a time constant of 85 fs to the Lb states, where it remains trapped in the Lb minimum up to the picosecond time scale. * 85 fs Lb pp* hn 1 ps La pp* Pyrene gs * N. Krebs, I. Pugliesi, J. Hauer and E. Riedle, New J. Phys. , 2013, 15, 085016.

Pyrene 2 D electronic spectrum at t 2=0 3° pulse 1° pulse 2°, 3°, pulses 2° pulse During t 1 the system is excited into the bright La 85 fs La pp* state and evolves along it. Setting t 2=0 means that the third (probe) pulse arrive at the same 1 ps time that the second pumping pulse, which in Lb pp* this case translate in probing the system along its evolution on the La potential energy 1° pulse hypersurface. gs

Pyrene 2 D electronic spectrum at t 2=0 Consequently, in order to compute the spectrum a t 2=0, a CASSCF dynamics on the La state from a previously optimized ground state geometry has been performed, and along it the high‐lying excited states have been evaluated performing SA‐ 60‐RASSCF(4, 8|0, 0|4, 8)/SS‐RASPT 2 calculations at selected points. (dynamics performed with Gaussian‐ Cobram, RASSCF‐RASPT 2 computation with Molcas) 2°, 3°, pulses 85 fs 1 ps 1° pulse gs La pp* Lb pp* La CASSCF dynamics SA-60 -RASSCF(4, 8|0, 0|4, 8) SS-RASPT 2 f e

Pyrene 2 D electronic spectrum at t 2=0 Using a development version of Spectron 2. 7 for simulating the spectra: Static approach Singletrajectory approach Passing from the static to the single‐trajectory approach description (which accounts for the evolution of the system during times t 1 and t 3) causes:

Pyrene 2 D electronic spectra at t 2=0 Using a development version of Spectron 2. 7 for simulating the spectra: Static approach Singletrajectory approach Passing from the static to the single‐trajectory approach description (which accounts for the evolution of the system during times t 1 and t 3) causes: 1‐ appearing of the vibronic structure of the GS bleaching

Pyrene 2 D electronic spectra at t 2=0 Using a development version of Spectron 2. 7 for simulating the spectra: Static approach Singletrajectory approach Passing from the static to the single‐trajectory approach description (which accounts for the evolution of the system during times t 1 and t 3) causes: 1‐ appearing of the vibronic structure of the GS bleaching 2‐ significant spectral shifts of the peaks

Pyrene 2 D electronic spectrum at t 2=1 ps 3° pulse 1° pulse 2° pulse There is no need to account for the dynamics evolution during t 3, since the system doesn’t 85 fs La pp* evolve, being trapped in the Lb minimum. The two approaches provide the same description for the 1 ps 3° pulse response of the system with the third pulse, so Lb pp* the only missing data for the construction of the spectrum t 2=1 ps are the higher excited states at 1° pulse the Lb minimum. gs 2° pulse

Pyrene 2 D electronic spectrum at t 2=1 ps Singletrajectory approach* (gas phase) Experimentally recorded** (methanol) The comparison between theory and experiment shows a remarkable agreement, validating the approximations in the single‐trajectory approach. *Nenov A, Giussani A, Fingerhut B P, Rivalta I, Dumont E, Mukamel S, Garavelli M, Phys. Chem. Phys. DOI: 10. 1039/c 5 cp 01167 a ** Krebs N, Pugliesi I, Hauer J, Riedle E, New J. Phys. , 2013, 15, 085016.

Pyrene 2 D electronic spectrum at t 2=1 ps Theoretical * Experimental** : The simulation of the spectrum provides important information not accessible from experiments, as the characterization of the electronic nature of the excited states involved in the appearing peaks. *Nenov A, Giussani A, Fingerhut B P, Rivalta I, Dumont E, Mukamel S, Garavelli M, Phys. Chem. Phys. DOI: 10. 1039/c 5 cp 01167 a ** Krebs N, Pugliesi I, Hauer J, Riedle E, New J. Phys. , 2013, 15, 085016.

Conclusions • Two strategies have been presented in order to compute 2 D spectra: the static and the single‐trajectory approaches, which respectively not consider and consider the evolution of the system during time t 1 and t 3. • The computation of the pyrene 2 D spectrum for t 2=0 has shown that the static approximation reproduces the positions of the peaks only qualitatively. • Through the comparison with the experimental spectrum of pyrene at t 2=1 ps recorded by Krebs et al. , the validity of the single‐trajectory approach has been proven. • The simulation of 2 D spectra provides fundamental information on the origin of the emerging peaks, that can not be extracted from the experiments.

Acknowledgments Prof. Marco Garavelli ENS-Lyon, FR Universita di Bologna, IT Prof. Shaul Mukamel University of California Irvine, USA Prof. Giulio Cerullo Politecnico di Milano, IT Dr. Artur Nenov Universita di Bologna, IT Dr. Ivan Rivalta ENS-Lyon, FR Universita di Bologna, IT Support and funding:

Thanks for your attention