Catalytic reforming is a process by which saturated hydrocarbons (alkanes and cycloalkanes) in petroleum naphtha fractions are converted as selectively as possible to aromatic hydrocarbons which have a much higher 'anti-knock' quality as automotive fuels. The recent introduction of 'lead free' gasoline regulations has caused the use of higher severities during commercial operation so as to improve octane product quality. This has resulted in faster deactivation of commercial catalysts, hence higher operating costs. The use of modern bimetallic reforming catalysts, replacing the older platinum-on-alumina catalyst, has resulted in greater stability, thus improving operating economics. However, a concise understanding of the operating characteristics of the mono- and bimetallic catalysts is needed to further science-aided reforming catalyst design. For this reason, a research program was initiated to investigate the ‘workings’ of a reforming catalyst. The transient kinetic behaviour of a fresh or pre-sulphided catalyst brought on-line, which is termed 'lineout', is found to be due to an initial deposition of coke on the metal sites of the catalyst (~1 wt%), the quantity of this deposition being essentially constant over the length of reformer operation. The coke deposition during long-term reformer operation (~20 wt%) is found to be on the alumina; however the observed deactivation in octane yields is due to the change in nature of coke (gradual graphitization) on the metal sites of the catalyst. Thus, two types of coke are distinguished on the metal sites, one being easily removed by hydrogen, the other less readily; they are labelled reversible and irreversible (graphitic) coke. Hydrogen limits the metal site deactivation rate by removal of the reversible and irreversible coke types by catalyzed hydrogenation and hydrogasification respectively. In addition to hydrogen removal, hydrogenolysis of coke precursors serves to limit catalyst deactivation. The improved hydrogenolysis, hydrogasification as well as paraffin dehydrocyclization characteristics of the iridium component of platinum-iridium catalysts, makes this catalyst the most attractive from viewpoints of aromatic yield and deactivation resistance. On the other hand, the improved deactivation resistance of platinum-rhenium catalysts compared to Pt/Al2O3 is probably due to the lower graphitization rate of metallic coke because of the steric hindrance introduced by the preferentially sulphided rhenium atoms incorporated in the platinum crystallites. A study of n-heptane reforming on presulphided and unsulphided Pt/Al2O3 and Pt-Re/Al2O3 showed typical effects of weight hour space velocity, hydrogen: hydrocarbon molar ratio, temperature and pressure on product yields and sensitivities. The liquid yield, product selectivity and deactivation resistance characteristics of sulphided Pt-Re/Al2O3 makes this the most attractive for industrial operation. Unsulphided Pt-Re/Al2O3 has a higher hydrogenolysis activity than unsulphided Pt/Al2O3 due to the rhenium function. Presulphidation causes the preferential attachment of sulphur to the rhenium atoms, lowering hydrogenolysis activity, but also lowering the activity for dehydrocyclization due to the sulphidation of some platinum atoms. The dehydrocyclization activity of this presulphided Pt-Re/Al2O3 catalyst is similar to that of a presulphided Pt/Al2O3. Presulphidation of Pt/Al2O3 gives the same activity as the addition of small amounts of sulphur in the feed to a fresh non-sulphided catalyst. Thus, due to the presence of small amounts of sulphur in industrial feeds, an industrial catalyst always operates in a partially sulphided state, where the important dehydrocyclization activity is 40-50% lower than the activity of a catalyst subjected to sulphur-free feeds. Further experiments with feed sulphur showed(i) effect of sulphur in activity attenuation is less at higher pressures (ii) selective partial presulphidation may improve aromatics yield (iii) fast reactions are more sulphur resistant (iv) reactions such as hydrogenolysis are strongly influenced by sulphur addition compared to dehydrocyclization and dehydrogenation. The activity of typical platinum-based catalysts is dependent on the surface crystallite configuration. The coordination of metal atoms in the crystallite, i.e. face or corner atoms, has some influence in reaction selectivity; the electron deficient character of corner atoms makes these atoms more active for certain reactions. The ensemble (site) requirement is the most important factor, e.g. alkane hydrogenolysis requires a minimum of three metal face atoms, while 1,6 dehydrocyclization requires one such atom of any type. A complete investigation of the principles given above has been conducted for the majority of reforming reactions. Because of this ensemble requirement, deposition of coke and sulphur has the same effect as adding an inert element such as tin to a platinum surface, since the surface is broken up into smaller atomic ensembles thus increasing the selectivity for reaction products requiring fewer platinum surface atoms in the reaction pathway. The electron deficient character of the crystallite corner atoms affects the location and influence of sulphur and coke on the crystallite in a number of ways: (i) a complete investigation of metal-site coking mechanisms showed that coke preferentially forms on the crystallite face atoms(ii) the high coke hydrogenation rates on corner atoms, compared to face atoms, keeps the corner atoms relatively coke-free(iii) sulphur is adsorbed irreversibly on crystallite face atoms in the atomic ratio Si/M = 0.5. Corner atoms are covered reversibly by sulphur depending on relative partial pressures of sulphur and hydrogen. Although the chemistry of deactivation of a reforming catalyst has been investigated, the mass transfer aspect has not, i.e., the effect of large amounts of deposited coke on the increased diffusion resistance of reactants into the catalyst pellet due to constriction of the pores. A general catalyst model is developed to quantify this effect, thus applying it to a bidisperse reforming catalyst. The major conclusions of this study are: choking or plugging of pores introduces significant mass transfer resistance.under deactivation free conditions (initial fresh catalyst) the mass transfer resistance lies in the macroporeswith large amounts of deposited coke, the diffusion resistance of the catalyst gradually incorporates in the microporesthe tortuosity factors in macro- and micropores are important in modelling effects of coking on transport and deactivation properties of a bidisperse pellet. Hence a monodisperse model is inadequate for deactivation modelling in a bidisperse pellet. A number of methods are investigated to measure the transport properties of a bidisperse catalyst. The diffusion cell and batch adsorber are seen to be the most flexible. The main source of experimental errors in the tortuosity evaluation by non-reactive techniques in the past has been effects of surface diffusion and finite rate of adsorption of tracer onto the catalyst surface. The latter has been quantified using a simple model, and it is shown that for silica gel at 50°C, it may cause an overestimation of tortuosity factor by a factor of 10. Both problems may be overcome by conducting experiments at high temperatures. Application of theory and experiment to each technique showed that the macropore tortuosity factor is in the range 2.8-3.1. However, the micropore tortuosity factor cannot be evaluated since the diffusion time in micropores is small compared to macropores. This gives low sensitivity for non-reactive evaluation of diffusivities. Increased sensitivity can be achieved by lowering the pellet radius so that macropore diffusion time is comparable to that in micropores, but this gives a transient response too fast to measure experimentally. Hence, these techniques should be used only to evaluate micropore transport properties for catalysts with microsphere radii above 40 μm and/or diffusivities less than l0-5cm2/s, such that the micropore diffusion time is sufficiently large to influence the overall uptake. The batch adsorber is shown to be the better technique for a number of reasons: (i) in a diffusion cell, the powder is pelletized therein, thus its transport properties are not the same as an industrial pellet(ii) mass transfer resistances can be eliminated totally in a batch adsorber. Finally, recommendations for further work and research needs are presented.