Computational Studies of Ion-Pair Separation of Benzylic Organolithium Compounds in THF: Importance of Explicit and Implicit Solvation

Abstract

Ion-pair separation (IPS) of THF-solvated fluorenyl (1(C)), diphenylmethyl (2(C)), and trityl (3(C)) lithium was studied computationally. Minimum energy equilibrium geometries of explicit bis- and tris-solvated contact ion pairs (CIPs) and tetrakis-solvated separated ion pairs (SIPs) were located at B3LYP/6-31G*. Associative transition structures linking the tris-solvated CIPs and tetrakis-solvated SIPs were also located. Based on MP2/6-31G*//B3LYP/6-31G* energies, the resting states of the CIPs are predicted to be trisolvates. Calculated enthalpies of IPS (DeltaH(IPS)) at 298 K were compared to experimental (UV-vis spectroscopy) solution values reported in the literature. In vacuum, B3LYP/6-31G* DeltaH(IPS) values for 1(C) x (THF)(3)-3(C) x (THF)(3) are 6-8 kcal/mol less exothermic than the experimentally determined values in THF solution. Closer examination of the individual steps of ion-pair separation (ionization, solvation, ion-pair recombination), as well as comparison of calculated structures with the published X-ray structures of 1(C) x (THF)(3) and 3(S) x (THF)(4), suggested that in vacuo modeling of the SSIPs was problematic. Incorporation of secondary solvation in the form of Onsager and PCM single-point calculations showed an increase in exothermicity of IPS. Application of a continuum solvation model (Onsager) during optimization at the B3LYP/6-31G* level of theory produced significant changes in the C(alpha)-Li contact distances in the SSIPs, and B3LYP/6-31G* (PCM)//B3LYP/6-31G* (Onsager) energies bring DeltaH(IPS) within 5-6 kcal/mol of experiment. Possible strategies to achieve closer agreement with experiment are discussed.