INFLUENCE OF SOLVENT ON THE SPECTROSCOPIC PROPERTIES OF CYANO COMPLEXES OF RUTHENIUM(II)

UV-visible spectra, emission spectra, and RU(III/II) reduction potenti als have been measured for cis[Ru(bpy)(2)(py)(CN)](+) (bpy is 2,2'-bip yridine; py is pyridine), cis-Ru(bpy)(2)(CN)(2), [Ru(tpy)(CN)(3)](-) ( tpy is 2,2': 6',2''-terpyridine), [Ru(bpy)(CN)(4)](2-), and [Ru(MQ(+)) (CN)(5)](2-) (MQ(+) is N-methyl-4,4'-bipyridinium cation) in twelve so lvents. The shifts in the metal-to-ligand charge transfer (MLCT) absor ption (E(abs)) or emission (E(em)) band energies with solvent increase linearly with the number of cyano ligands and correlate well with the Gutmann ''acceptor number'' of the solvent. Intraligand pi --> pi ba nd energies also correlate with acceptor number, but with only similar to 30% of the shifts for the MLCT bands, The solvent dependence arise s through mixing of the pi --> pi transitions with lower energy MLCT transitions. MLCT absorption and emission spectra are convolutions of overlapping vibronic components, and a Franck-Condon analysis of emiss ion spectral profiles for cis-Ru(bpy)(2)(CN)(2) has been used to eval uate the energy gap, E(0), and chi'(0.gs), where chi'(0.gs) is the sum of the solvent reorganizational energy for the ground state below the excited state and the inner-sphere reorganizational energy of the low -frequency modes, chi(i,L), is treated classically. Both E(0) and chi' (0.gs) correlate well with acceptor number with Delta E(0)/Delta AN = 44 +/- 2 cm(-1)/AN unit and Delta chi(0.gs)/AN 21 +/- 3 cm(-1)/AN unit if it is assumed that chi(i,L) is solvent independent. From electroch emical measurements and the difference in E(1/2) values for metal oxid ation and bpy reduction, Delta Delta G(es)(0)/Delta AN similar or equa l to 70 +/- 7 cm(-1)/AN unit with Delta G(es)(0) the free energy of th e excited state above the ground state, These correlations show that t he energy gap is far more sensitive to solvent than chi(0,gs). Delta c hi(0,gs)/Delta AN can also be estimated from the relation Delta Delta G(es)(0)/Delta AN = Delta E(0)/AN + Delta chi'(0,gs)/Delta AN, which g ives Delta chi(0,gs)/Delta AN = 26 +/- 7 cm(-1)/AN unit. The solvent r eorganizational energy of the excited state above the ground state is chi(0,es). Its variation with acceptor number can be estimated from th e relation Delta E(abs)/Delta AN - Delta G(es)(0)/Delta AN = Delta chi (0,es) (=13 +/- 8 cm(-1)/AN) or from Delta E(abs)/Delta AN - Delta E(e m)/Delta AN - Delta chi(0,gs)/Delta AN (=14 +/- 8 cm(-1)/AN), if chi(i ,L) is solvent independent. These results suggest that chi(0,gs) is mo re sensitive to solvent than chi(0,es) by as much as a factor of 2. De lta E(em)/Delta AN = 45 +/- 3 cm(-1)/AN unit similar or equal to Delta E(0)/Delta AN = 44 +/- 2 cm(-1)/AN unit, showing that the emission ma ximum gives accurate information about the solvent dependence of the e nergy gap. A model is invoked to explain the acceptor number dependenc e. It is based on electron pair donation from the lone pair on cyanide to individual solvent molecules through donor-acceptor interactions. The model is consistent with variations in v(CN) with acceptor number in cis-[Ru(bpy)(2)(py)(CN)](+) and in E(1/2)(Ru-III/II) for the series of complexes. In this model it is assumed that donor-acceptor interac tions are more important in the ground state than in the excited state , consistent with pK(a) measurements, and that they are additive in th e number of cyanide ligands. These and I-I-bonding interactions in rel ated ammine complexes perturb the internal electronic structure of the solute with important consequences. One is that chi(0,gs) not equal c hi(0,es) although they are commonly assumed to be equal.