(375a) Synthesis and Characterization of Bimetallic Overlayer Aqueous Phase Reforming Catalysts

Introduction:  A great deal of today’s research seeks to create new and sustainable fuels to meet our world’s ever growing energy demands.  Recently, aqueous phase reforming (APR) has emerged as a process capable of turning renewable feedstocks into hydrogen gas or alkanes.  Much of the current hydrogen supply is derived from light hydrocarbons, and a sustainable method of production will prove imperative as demand increases.  The majority of past APR studies have used feedstocks that are derived from biomass (methanol, ethylene glycol, glycerol, sorbitol, and glucose) but in order to prove the full potential of the process, further studies need to be done using a non-derived biomass feed.  Lactose, a sugar waste product of the cheese and dairy industry, has shown promise as a biomass product capable of undergoing the APR reaction in a single step.

Past APR studies have looked at a variety of different supported metal catalysts for the reaction.  These studies have shown that bimetallic Pt/Ni and Pt/Co make ideal catalysts for production of hydrogen through APR.  Building upon the success of bimetallic catalysts for APR, we have synthesized pseudomorphic overlayer type bimetallic catalysts which aim to increase the activity of the catalyst while reducing the necessary loading of costly platinum.  An ideal pseudomorphic overlayer catalysts consist of a monatomic layer deposited atop a different bulk metal.  These catalysts have received widespread attention from computational and single crystal studies because of their notably different adsorption properties when compared to their constituent metals.  These changes are likely due to electronic interactions between the two metals causing a shift in the d-band center of the overlayer metal.  Computational work has demonstrated a linear relationship between the center of the d-band and heats of adsorption.  However, there is still a need to develop effective broad-based synthesis methods for overlayer catalysts on high surface area supports that will empirically demonstrate predicted behavior.  Additionally, it would be beneficial to correlate this behavior with reactivity, which cannot be predicted computationally.

In order to capitalize on the behavior of pseudomorphic overlayers, we have previously synthesized Re@Pd (base@overlayer) supported catalysts.  A decrease in heat of hydrogen adsorption was seen for Re@Pd catalysts when compared to both the Re and Pd parents; these results were consistent with computational predictions found in literature.  Re@Pd overlayer catalysts also demonstrated a decrease in activity for ethylene hydrogenation when compared to the activity of Pd.  Thus reactivity was correlated with a computationally predicted parameter.  We have built upon these results and have extended the synthesis technique toward creating Ni@Pt and Co@Pt catalysts.  Ni/Al2O3, Co/Al2O3, and Pt/Al2O3 monometallic baseline catalysts were prepared using standard techniques.  The directed deposition technique was used to selectively deposit the overlayer (platinum) metal on the base metal without it depositing onto the support surface. 

Experimental:  Hydrogen chemisorption studies for isotherm analysis and to determine hydrogen heats of adsorption were done using a Micromeritics ASAP 2020.  Isotherms for each sample were measured from 35 to 300°C and pressures from 1 mtorr to 900 torr.  The reactivity of each sample was studied through the ethylene hydrogenation reaction.  The ethylene hydrogenation reaction was performed using a plug flow reactor at temperatures from -10 to 300°C at atmospheric pressure with a typical gas flow rate of 750 mL/min.  The reactor effluent was monitored using a GCMS. 

Results and Discussion:  Chemisorption results showed that hydrogen adsorption isotherms were distinct in shape and temperature dependence for each of the Ni/Al2O3 and Co/Al2O3 parent catalysts.  Isotherms for Ni@Pt/Al2O3 were extremely similar to those of the Ni/Al2O3; giving no indication of isolated Pt metal formation.  Adsorption measurements done at multiple temperatures allowed for the determination of isosteric heat of hydrogen adsorption.  Measured hydrogen heats of adsorption for Pt/Al2O3 and Ni/Al2O3 were 57 kJ/mol and 36 kJ/mol respectively at high coverage.  Platinum data shows similar results to single crystal and computational predictions.  The heat of adsorption for Ni@Pt/Al2O3 was 47 kJ/mol at high coverage.  As predicted computationally, the heat of adsorption for hydrogen decreased for the overlayer compared to Pt.  The weaker hydrogen adsorption for Ni@Pt when compared to monometallic Pt should aid in increasing APR reaction rates because the less tightly bound hydrogen product will more quickly desorb from the metal surface, thereby freeing up active sites for adsorption of reactants.

Pt/Al2O3 is highly active for ethylene hydrogenation and both of the Ni/Al2O3 and Co/Al2O3 baseline catalysts show lower activities.  Addition of a platinum overlayer to make Ni@Pt and Co@Pt overlayer catalysts increased the activity of each catalyst but still showed reduced activity when compared to a pure platinum catalyst.  These results agree with predicted d-band shifts that would cause weaker hydrogen adsorption on the metal surface and ultimately reduced activity for ethylene hydrogenation when compared to platinum metal alone.  By using the ethylene hydrogenation reaction as a characterization tool we have been able to directly correlate reactivity with hydrogen heat of adsorption.  We have previously shown that as hydrogen heat of adsorption decreases, the turnover frequency for ethylene hydrogenation decreases in the Re@Pd system.  This activity decrease is consistent with literature results that show ethylene hydrogenation activity is directly related to the amount of hydrogen adsorbed on the metal surface.  Thus, the relationship between hydrogen heat of adsorption and ethylene hydrogenation activity has been extended to multiple systems.  Additional reactivity studies are underway to further characterize the Ni@Pt and Co@Pt catalysts using apparent activation energies and reaction orders.  APR studies will also be initiated for each catalyst and lactose feed in order to determine possible effects that overlayer catalysts can have on the APR reaction.