This review is a summary of the mechanisms of catalytic cracking of small (C-3-C-6) alkanes. Most of the evidence has arisen from product distributions and kinetics of cracking of these alkanes, interpreted on the basis of solution carbocation chemistry and theoretical chemistry. Cracking of small alkanes catalyzed by solid acids such as the zeolite HZSM-5 proceeds by two mechanisms: (1) The unimolecular (protolytic cracking) mechanism, which proceeds via an alkanium ion formed by protonation of the alkane by the catalyst. This supposed transition state collapses to give either H-2 and a carbenium ion or an alkane and a carbenium ion; the carbenium ions give up protons to the catalyst to form alkenes. The cracking products include methane and ethane as well as H-2. (2) The classical (bimolecular) cracking mechanism, which involves carbenium ion chain carriers that react with the alkane reactant to abstract hydrides and generate carbenium ions that undergo beta-scission. The products include alkanes and alkenes, but not methane, ethane, or H-2. Because protolytic cracking gives alkene products, which are much stronger bases than alkanes, the alkenes become the predominant proton accepters as conversions increase, and thus bimolecular cracking prevails at all but the lowest conversions. Protolytic cracking in the near absence of secondary reactions has been observed only for propane and n-butane at low conversions; secondary reactions appear to be generally significant for other alkanes. Although the product distributions are qualitatively understood, there are still inconsistencies in the literature of quantitative product distributions and kinetics, and more experimental work is needed with standard catalysts such as HZSM-5. Theoretical chemistry is leading to deeper understanding of the transition states, showing that cracking mechanisms involving bare carbocations are oversimplified. Rather, the catalyst surface must be included, and it has been simulated by clusters that are zeolite fragments. Surface alkoxides are more stable than surface carbenium ions, and cracking takes place by concerted bond breaking and formation. Theoretical activation energies for protolytic cracking of alkanes are close to experimental activation energies that have been corrected for the adsorption energy of the reactant, but it appears that more theoretical work (as well as better data) is required for satisfactory agreement of theory and experiment.
