This work addresses mechanistic issues in halogen-assisted diamond growth via ab initio molecular orbital calculations on halogenated and hydrohalogenated carbon clusters (C(9)H(12)X(i), C(9)H(13)X, X = F Or Cl; i = 0, 1, or 2) as models of the dimer-reconstructed diamond (100)2x1 surface. The bond lengths (XC-CX, XC-CH, XC-C*, C=C, C-X, XCC-H) and the C-X and XCC-H bond energies have been determined along with the effects of lattice constraints. The dimer bonds are highly strained and are longer in clusters where only the topmost carbon layer is allowed to relax and the remaining carbon atoms are constrained to lie at diamond lattice positions than in fully relaxed clusters. The first C-F and C-Cl bond energies in monohalide structures are calculated to be 502 and 366 kJ/mol, respectively, and are approximately the same in constrained and fully relaxed clusters. The second C-X bond is distinctly weaker due to rr bond formation, particularly in fully relaxed clusters. The pairing energies for F + F, Cl + Cl, H + F, and H + Cl were each found to be very nearly equal to that for H + H, calculated as 64 and 117 kJ/mol for constrained and relaxed clusters, respectively, affirming the identification of the pairing energy with the pi bond strength. The optimized structures of the transition states (TS) and the activation energies for adsorption and desorption of HF and HCl on diamond (100)2x1 have also been determined. The TS calculations suggest that HF adsorbs and desorbs on diamond (100)2x1 via an asymmetric four-centered transition state, whereas HCl prefers a two-centered (Cl**H**C) electrophilic addition. Both transition states are reactant-like (C9H12 + HX), i.e., occur early in the reaction coordinate. Desorption of HX from diamond (100)2x1 is predicted to be approximately as endothermic as that of H-2, ca 360 and 310 kJ/mol for constrained and relaxed clusters, respectively, and therefore should be unimportant during halogen-assisted chemical vapor deposition (CVD) and atomic layer epitaxy (ALE) growth of diamond. Adsorbed F and Cl are predicted to be stable at substantially higher temperatures than observed experimentally. The discrepancy is attributed to electronic repulsion effects at near-monolayer coverage that cannot be accounted for properly in small cluster models.