The special salt effect exerted by LiClO 4 (LP) on the rate of dicyanoanthracene (DCA)-photo-induced electron transfer oxygenation of tetraarylcyclopropane ( 1) and 1,1-diarylethylenes ( 3a–d) in acetonitrile, was used to determine several important parameters of the solvent-separated radical ion pairs (SSIP): (1) the efficiency α o of dissociation into solvated free radical ions, (2) the ratio of rate constants for the separation of SSIP and the back electron transfer within the SSIP ( k sep/ k back) and (3) the ratio of rate constants for scavenging DCA − from SSIP by LP (or Li +) and the back electron transfer ( k L/ k back). The efficiencies α o turned out to be relatively large, i.e. 0.66 for 1, and, increasing with increasing oxidation potentials, between 0.38 for 3a and 0.8 for 3d. Based on the assumption that k L equals the diffusion-controlled rate constant k diff in MeCN, k sep = 1.7 × 10 7 s −1 and k back = 8 × 10 6 s −1 were estimated for 1. Within the limits of error, k sep = (8±2) × 10 7 s −1 was obtained for substrates 3a–d, whereas k back was found to decrease from 1.64 × 10 8 s −1 for 3a to 2 × 10 7 s −1 for 3d, indicating that back electron transfer, associated with Δ G back values of −2.2 to −2.8 eV, increases with decreasing energy content of the SSIP. This result represents another example for back electron transfer occurring in the Marcus ‘inverted region’. However, by assuming k L ⪕ k diff, the absolute values of k back are about one to two orders of magnitude smaller than those obtained by using the generally accepted value of 5 × 10 8 s −1 for k sep. To explain the lower values of k sep and k back, it is proposed that the encounter distance for electron transfer with the electron donors used in this study may be somewhat larger (about 7–10 Å) than the distance (of about 6.5–7.5 Å) usually assumed for these processes with less bulky electron donors.