UV-vis subpicosecond spectroscopy of 4-(9-Anthryl)-N,N'-dimethylaniline in polar and nonpolar solvents:: A two-dimensional view of the photodynamics

Transient absorption and gain spectra are reported for 4-(9-anthryl)-N,N'-dimethylaniline (ADMA) in polar and nonpolar solvents in the 330-780 nm wavelength range after excitation with a 500-fs laser at 370 nm. In acetonitrile, the initial transient spectra can be interpreted by the superposition of the dimethylaniline cation radical and the anthracene anion radical absorption bands, resulting from the well-known ultrafast photoinduced charge separation. In benzonitrile, similar transient spectra are reached 20-30 ps after excitation. In this solvent, as well as in alcohols, tetrahydrofuran and cyclohexane, the initial UV and visible absorption bands exhibit the spectral characteristics of the locally excited (LE) state. With increasing time, while these bands decay, the signature of the charge-transfer (CT) state appears and one observes a red shift of the region of minimum differential absorption, where gain is expected. These changes occur at a solvent-dependent rate, except in acetonitrile in which the spectra show little evolution because of our limited time resolution. Discrepancies with previously published models for ADMA photodynamics are discussed. In polar solvents except acetonitrile, a quasi-barrierless or slightly activated electron transfer is proposed to explain the longtime decay of the UV absorption band attributed to the LE state. On the other hand, inertial internal torsion around the bond linking the aniline and anthracene moieties in the LE state followed by solvation-induced charge separation similar to that in bianthryl is proposed to explain differences between acetonitrile and benzonitrile or other solvents in the UV range. A two-dimensional picture in which both internal torsion and solvation dynamics determine the excited-state reaction path is discussed, taking into account the initial proposal made by Mataga's group. In cyclohexane, in which solvation effects are not expected, the spectral changes in the picosecond time range are attributed to geometrical relaxation within the S-1 state.