Parameters affecting electron injection dynamics from ruthenium dyes to titanium dioxide nanocrystalline thin film

Electron injection rates from Ru(dcbpy)₂(X)₂ [X = 2NCS, 2CN, and dcbpy; dcbpy = 2,2′-bipyridine-4,4′-carboxylate] (called Ru535 or Ru N3, Ru505, and Ru470) to TiO₂ nanocrystalline thin films are examined as a function of adsorbate redox potential, pH of the solution, excitation wavelength, and solvent. For all three dyes, the injection kinetics are biphasic, consisting of a distinct ultrafast component (<100 fs) and slower components. Under different experimental conditions, the partitioning between these two components and the rate of the slow components change, but the rate of the fast component shows no noticeable variations within the ∼200 fs time resolution of the measurement. When Ru535, Ru505, and Ru470 were compared at the same pH, increasing amplitude and decreasing rate of slow component were observed, correlating with less negative excited-state redox potentials in these dyes. An analogous trend was seen for RuN3/TiO₂ by increasing the pH of the solution from pH = 2 to 8 and changing from pH = 2 aqueous solution to a (1:1) ethylene/propylene carbonate mixture. The injection dynamics are also dependent on excitation wavelength. The relative amplitude of the slow component increases when the excitation wavelength is changed from 400 to 630 nm. All data can be described by a two-state injection model, which attributes the fast (<100 fs) component to injection from a nonthermalized excited state and the slow component to injection from the thermalized excited state. The partitioning between these two components and the rate of the slow components depend on the relative energetics between dye excited states and the conduction band edge.