UC Berkeley

Steam methane reforming (CH4-H2O) is the most widely used process for the production of H2 and is also a key reaction in the production of a wide variety of value-added chemicals and products from natural gas resources. Ni-based materials are the most commonly used catalysts in these processes but suffer from the undesired formation of carbonaceous deposits, most often in the form of filaments that can lead not only to deactivation but also catalyst disintegration. Low CH4/H2O feed ratios are frequently used to mitigate such carbon formation. These strategies reflect the wider effects and consequences of carbon thermodynamic activities on the catalyst surface during steady-state catalysis. The dynamics of carbon deposition during CH4-H2O and CH4-CO2 reactions (800-1000 K) on Ni-based catalysts and kinetic treatments show that carbon activities and thus carbon formation rates are uniquely determined by a ratio of pressures ψ ((CH4)(H2)/(H2O)) for a given temperature. These carbon activities provide the thermodynamic driving force for the diffusion of carbon through Ni nanoparticles and subsequent formation of filaments. The diameters of carbon filaments are correlated with the diameters of the Ni particles (5-11 nm). Smaller filaments are less stable and exhibit higher thermodynamic carbon activities, rendering their formation more difficult and decreasing thermodynamic driving force for their formation, thus decreasing also rates of carbon deposition. The conclusions from these carbon formation studies inform subsequent strategies for encapsulating metal (Pt) clusters in microporous materials.

The encapsulation of metals in zeolites provides many advantages over conventional metal-oxide supported metal nanoparticles, including the retention of small nanoparticles even at high temperatures. Such microporous materials are frequently used in catalytic applications to selectively sieve molecules based on their size and shape and can therefore be used to selectively impose intracrystalline concentration gradients, leading to changes in reactant ratios within the channels of molecular dimensions. These properties can be exploited for CH4-H2O reactions, where the diffusivity of smaller H2O molecules is expected to be much greater than for CH4 in pores of molecular dimensions. The extrapolation of diffusion data to reaction temperatures, however, can involve significant error. Furthermore, the measurement of H2O and CH4 can (R^2/De) at reaction temperatures by conventional transient uptake measurements is not feasible because of the extremely short timescales of these processes and low uptakes at such temperatures. The measurement of diffusion time constants (R^2/De) at reaction temperatures is therefore accomplished here by analyzing effective reaction rates, specifically isotopic exchange rates, in mass transport limited systems using reaction-diffusion models.

The extraction of diffusion time constants (R^2/De) from reaction-diffusion models can only be rigorously performed with accurate kinetic models. The kinetics of H2O-D2 and H2-D2 exchange and their mechanistic interpretations are therefore addressed here to allow for such calculations of diffusion time constants. H2-D2 isotopic exchange rates (5-80 kPa H2, 5-80 kPa D2; 383 K; H2/D2 = 0.0625-16) on Pt-based catalysts show monotonic increases in rate with H2 and D2 pressures, in contrast with the kinetics expected from the commonly cited recombinative desorption (H* + D*) pathway. Such recombination events only become significant at temperatures greater than 700 K and at low combined pressures (<10 kPa) of H2 and D2. H2-D2 exchange is instead shown to proceed via the reaction of H2 with D* and D2 with H* in a single-site mechanism that does not exhibit a kinetic isotope effect. These exchange events are shown, using theoretical calculations, to occur via the dissociative adsorption of H2 (or D2) at entropically-favorable vacancies that arise from fluctuations in mobile H*-adlayers. These reaction pathways circumvent desorption/recombination reactions; such exchange reactions therefore cannot be used to demonstrate reversibility of H2 adsorption on catalytic surfaces at conditions of practice for hydrogenation reactions, as is commonly practiced. The kinetics of H2O-D2 isotopic exchange reactions (473 K, 5-80 kPa D2, 2.5-40 kPa H2O) are consistent with the reaction between D2 and molecularly adsorbed H2O*. This reaction pathway remains the predominant pathway for temperatures below 900 K, at which point the dissociation of H2O (to form OH* and H*) and recombination of OH* and D* likely becomes the primary pathway for exchange. These kinetic studies also show that H2O irreversibly titrates Pt surface sites by the formation of OH* species that can only be fully removed by reductive treatments at temperatures greater than 700 K.

The kinetics of exchange inform the rigorous calculation of diffusion time constants from isotopic exchange events in mass transport limited materials. CH4-D2 (5-35 kPa CH4, 5-30 kPa D2) H2O-D2 (5-30 kPa H2O, 5-30 D2), and H2-D2 (10 kPa H2, 10 kPa D2) isotopic exchange rates are measured here on Pt/SiO2, Pt/γ-Al2O3, Pt/Na-LTA, and Pt/Ca-LTA samples at temperatures (573-900 K) relevant for CH4-H2O reforming reactions. Effectiveness factors for CH4-D2 and H2O-D2 exchange on Pt/Na-LTA and Pt/Ca-LTA are used in classical reaction-diffusion models to calculate diffusion time constants for CH4 and H2O. H2-D2 exchange reactions did not exhibit mass transport limitations in these materials. CH4 exhibited significantly larger diffusion time constants (by factors of more than 102) than H2O in both Pt/Na-LTA and Pt/Ca-LTA throughout these temperatures (573-900 K), indicating that H2O diffuses more readily in these materials, as required for elevated intracrystalline H2O/CH4 ratios during CH4-H2O reactions. The measured CH4 and H2O diffusion time constants are used in the interpretation of deactivation behaviors during CH4-H2O reactions at 873 K on Pt/γ-Al2O3, Pt/Na-LTA, and Pt/Ca-LTA samples. Deactivation rates on Pt/γ-Al2O3 were linearly dependent on a ψ ((CH4)(H2)/(H2O)), as expected from carbon formation studies. Deactivation rates on Pt/Na-LTA and Pt/Ca-LTA were undetectable for ψ ((CH4)(H2)/(H2O)) values below 10 (including stoichiometric CH4/H2O ratios) and exhibited deactivation rates that were 3-8 times slower than on Pt/γ-Al2O3 for ψ values between 10 and 40. These improvements reflect the high H2O/CH4 ratios within the zeolite pores, as indicated by extensive numerical models. Simulations of deactivation behavior are also used to provide additional insight into further optimizing these materials. The strategies for exploiting differences in diffusivity in zeolites, demonstrated here for CH4 reforming reactions, are generally applicable to reactions where the selective access for smaller reactants is desirable.