On the role of conical intersections in photodissociation. V. Conical intersections and the geometric phase in the photodissociation of methyl mercaptan

Machine Learning for Electronically Excited States of Molecules

Electronically excited states of molecules are at the heart of photochemistry, photophysics, as well as photobiology and also play a role in material science. Their theoretical description requires highly accurate quantum chemical calculations, which are computationally expensive. In this review, we focus on not only how machine learning is employed to speed up such excited-state simulations but also how this branch of artificial intelligence can be used to advance this exciting research field in all its aspects. Discussed applications of machine learning for excited states include excited-state dynamics simulations, static calculations of absorption spectra, as well as many others. In order to put these studies into context, we discuss the promises and pitfalls of the involved machine learning techniques. Since the latter are mostly based on quantum chemistry calculations, we also provide a short introduction into excited-state electronic structure methods and approaches for nonadiabatic dynamics simulations and describe tricks and problems when using them in machine learning for excited states of molecules.

Machine Learning for Electronically Excited States of Molecules

Electronically excited states of molecules are at the heart of photochemistry, photophysics, as well as photobiology and also play a role in material science. Their theoretical description requires highly accurate quantum chemical calculations, which are computationally expensive. In this review, we focus on not only how machine learning is employed to speed up such excited-state simulations but also how this branch of artificial intelligence can be used to advance this exciting research field in all its aspects. Discussed applications of machine learning for excited states include excited-state dynamics simulations, static calculations of absorption spectra, as well as many others. In order to put these studies into context, we discuss the promises and pitfalls of the involved machine learning techniques. Since the latter are mostly based on quantum chemistry calculations, we also provide a short introduction into excited-state electronic structure methods and approaches for nonadiabatic dynamics simulations and describe tricks and problems when using them in machine learning for excited states of molecules.

Vibration mediated photodissociation dynamics of CH3SH: manipulation of the dynamic energy disposal into products

The S-H bond dissociation dynamics of CH3SH have been investigated for the S1-S0 transition mediated by either the S-H stretching (2608 cm-1) or CH3 symmetric stretching (2951 cm-1) mode excitation in the S0 state. The S-H and C-S bond extensions are strongly coupled in the S1 state through the S1/S2 same-symmetry conical intersection, giving the C-S stretching mode excitation of the CH3S˙ fragment during the prompt S-H bond rupture on S1. In the IR + UV transition mediated by the S-H stretching mode, the vertical transition seems to access the Franck-Condon region where the S-H bond is shortened while the coupling to the C-S bond stretching becomes stronger compared to the case of one-photon UV transition, indicating that the intramolecular vibrational redistribution (IVR) is little activated in S0. When the IR + UV excitation is mediated by the CH3 symmetric stretching mode, on the other hand, the Franck-Condon region in S1 encompasses the enlarged molecular structures with respect to both S-H and C-S bond extensions, presumably due to the rapid IVR in S0 prior to the vertical transition. This leads to the inverted vibrational state population of the C-S bond stretching mode of the CH3S˙ fragment. This work demonstrates that the reaction dynamics upon the IR + UV excitation of CH3SH is highly mode dependent and the energy disposal dynamics could be controlled by the manipulation of the Franck-Condon region through the particular vibrational-state mediation in the ground state, shedding new light on the structure-dynamics relationship.

Photodissociation dynamics of ethanethiol in clusters: complementary information from velocity map imaging, mass spectrometry and calculations

We investigate the solvent effects on photodissociation dynamics of the S–H bond in ethanethiol CH₃CH₂SH (EtSH).

Imaging CH3SH photodissociation at 204 nm: the SH + CH3 channel

The SH + CH(3) product channel for the photodissociation of CH(3)SH at 204 nm was investigated using the sliced velocity map ion imaging technique with the detection of CH(3) products using state selective (2+1) resonance enhanced multiphoton ionization (REMPI). Images were measured for CH(3) formed in the ground and excited vibrational states (v(2) = 0, 1, and 2) of the umbrella mode from which the correlated SH vibrational state distributions were determined. The vibrational distribution of the SH fragment in the SH + CH(3) channel at 204 nm is clearly inverted and peaks at v = 1. The highly negative anisotropy parameter of the CH(3) (v(2) = 0, 1, and 2) products is indicative of a fast dissociation process for C-S bond cleavage. Two kinds of slower CH(3) products were also observed (one of which was partly vibrationally resolved) that are assigned to a two-step photodissociation processes, in which the first step is the production of the CH(3)S (X(2)E) radical via cleavage of the S-H bond in CH(3)SH, followed by probe laser photodissociation of nascent CH(3)S radicals yielding CH(3)(X(2)A(1), v(2) = 0-2) + S((3)P(j)/(1)D) products.

Imaging the photodissociation of CH3SH in the first and second absorption bands: The CH3(X̃A12)+SH(XΠ2) channel

The CH3(X2A1)+SH(X2Pi) channel of the photodissociation of CH3SH has been investigated at several wavelengths in the first 1 1A"<--X 1A' and second 2 1A"<--X1A' absorption bands by means of velocity map imaging of the CH3 fragment. A fast highly anisotropic (beta=-1+/-0.1) CH3(X2A1) signal has been observed in the images at all the photolysis wavelengths studied, which is consistent with a direct dissociation process from an electronically excited state by cleavage of the C-S bond in the parent molecule. From the analysis of the CH3 images, vibrational populations of the SH(X2Pi) counterfragment have been extracted. In the second absorption band, the SH fragment is formed with an inverted vibrational distribution as a consequence of the forces acting in the crossing from the bound 2 1A" second excited state to the unbound 1 1A" first excited state. The internal energy of the SH radical increases as the photolysis wavelength decreases. In the case of photodissociation via the first excited state, the direct production of CH3 leaves the SH counterfragment with little internal excitation. Moreover, at the longer photolysis wavelengths corresponding to excitation to the 1 1A" state, a slower anisotropic CH3 channel has been observed (beta=-0.8+/-0.1) consistent with a two step photodissociation process, where the first step corresponds to the production of CH3S(X2E) radicals via cleavage of the S-H bond in CH3SH, followed by photodissociation of the nascent CH3S radicals yielding CH3(X2A1)+S(X3P0,1,2).

Conical intersection between the lowest spin-aligned Li3(A′4) potential-energy surfaces

We have calculated new potential-energy surfaces for the lowest two spin-aligned (4)A(') states of the Li(3) trimer. This calculation shows a seam of conical intersections between these states resulting from the extra symmetry of the system when the atoms are in a collinear arrangement. This seam is especially important because of its proximity to the three-body dissociation limit of the system; ultracold scattering calculations and the bound-state energies of the system will be affected by the presence of this conical intersection. In this paper we discuss the calculation of the potential-energy surface and the location of the conical intersection seam.