Research in the Estes Laboratory

While homogeneous catalysts are widely acknowledged to have very high specific activities and high selectivities in a variety of potentially industrially relevant reactions, they are employed much less often than heterogeneous catalysts in industry. One solution to this is to heterogenize molecular catalysts on a solid support via covalent immobilization strategies, which allows molecular catalysts to be easily applied in flow reactors. Immobilized catalysts combine the benefits of both homogeneous and heterogeneous catalysts in that they are highly active and selective but also have increased stability and easier separation from the product phase. However, such immobilizations often drastically change the catalytic properties of the complex itself in ways that are not well understood. The goal of our work in the Estes group is to understand the physical and chemical origins of these reactivity differences in immobilized catalysts and to optimize the synergy between catalyst and support such that catalysis is improved upon immobilization.

We use three different techniques to examine how surfaces change the reactivity of molecular catalysts: 1) surface organometallic chemistry (SOMC), 2) molecular heterogeneous catalysis (MHC), and 3) simple wet impregnation. In SOMC, a metal oxide surface is prepared such that is covered by a homogeneous distribution of chemically similar OH groups. These hydroxyl groups are then allowed to react with basic metal complexes to deprotonate the OH group and thereby create a new M-O bond to the surface. In this strategy, the surface acts as a ligand in the coordination complex. In MHC, on the other hand, the covalent linkage between surface and support is established by adding pendant linker groups such as trialkoxysilanes or phosphonate esters to molecular catalysts. Upon reaction with the surface, these linker groups bind through one or more bonds to the metal oxide, thus keeping the metal closely tied to the surface and more closely preserving the structure of the precursor complex. Thirdly, we have also found that simple physisorption on surfaces already imparts certain catalytic benefits for gas-phase catalytic reactions such as increased activity and stability. In these cases, the support material can be thought of as a solid solvent, which for gas-phase small molecule activation lead to longer catalyst lifetime, higher rates of gas diffusion, and higher gas adsorption than in normal liquid solvents, while being simultaneously much more environmentally friendly for large scale processes. These methods are summarized in figure 1 below.

We use these methods to understand the role that the supports play in catalytic processes of both immobilized molecular and mainstream heterogeneous catalyst types by understanding how molecules interact with catalyst surfaces. Below we show three main areas of research that we are investigating in our group, including 1) Understanding the thermodynamics and kinetics of hydrogen spillover and its role in catalytic processes; 2) Understanding how Lewis acidic surfaces affect CO2 hydrogenation Catalysis; and 3) How adsorption and mass transport phenomena affect immobilized catalysts. 

Understanding the thermodynamics and kinetics of hydrogen spillover and its role in catalytic processes

One way in which organometallic catalysts, in particular metal hydrides, can be affected by the support material is cases in which they undergo hydrogen spillover. Hydrogen spillover is a phenomenon in which hydrogen atoms adsorbed on a metal nanoparticle are transferred onto a reducible metal oxide or other reducible support via a proton-coupled electron transfer (PCET) reaction. Hydrogen spillover results in formation of reactive sites such as reduced metal centers, acidic OH groups, and oxygen vacancies on metal oxide surfaces that are often important for catalysis. However, hydrogen spillover is still quite a mystery even today, as its basic reactivity (thermodynamics and kinetics) are still not well understood. This makes it difficult to predict when hydrogen spillover will occur to a greater or lesser extent.

In the Estes group, we study hydrogen spillover by replacing the metal nanoparticle with molecular metal hydrides that have similar reactivity. This allows us to study both the thermodynamics (via M-H titration) and the kinetics (via ReactIR spectroscopy) of the PCET reactions on metal oxide surfaces. Studying these model reactions with a variety of different metal oxides and even the same metal oxide with different crystal structures or particle morphologies allows us to understand why some catalysts undergo more hydrogen spillover and why others undergo less or no hydrogen spillover at all.

Furthermore, we can correlate the thermodynamics and rate of hydrogen spillover with the catalytic reactivity of the individual samples. For example, we found that hydrogen spillover from Ru to molybdenum oxide results in formation of Mo(V)-OH groups that accelerate the hydrodeoxygenation of phenol to benzene. The mildly acidic OH groups catalyze the tautomerization of phenol to its keto-tautomer, which can then be hydrogenated by the Ru, ultimately forming benzene. This transformation is potentially useful as a way of transforming biomass into jet fuels.  

Understanding how Lewis acidic surfaces affect CO2 hydrogenation catalysis

The unmitigated carbon emissions caused by burning fossil resources has dire environmental consequences and represents a breakdown of the natural carbon cycle, in which equal amounts of carbon dioxide are emitted and taken up. An elegant solution to this problem is to make fuels (MeOH) directly from CO2 and renewable energy, as is done by nature. If the CO2 comes from renewable sources, such as biomass or directly out of the atmosphere, the fuels so derived would be overall nearly carbon-neutral. Common catalysts for hydrogenation of CO2 to methanol include mixtures of Cu metal nanoparticles dispersed on Lewis acidic supports, such as ZrO2. The metal oxide is critical for the reactivity and is thought to activate CO2 and especially formate groups on the surface toward further hydrogenation, while the hydrogen atoms are transferred from nearby metal nanoparticles. 

Using molecular catalysts rather than supported metal catalysts has the distinct advantage that the molecular catalysts are often active at much lower temperatures than the metal nanoparticle catalysts. Therefore, immobilize CO2 hydrogenation catalysts on Lewis acidic supports and examine their activity for CO2 hydrogenation, especially to methanol. For example, we demonstrated that amine-assisted CO2 hydrogenation to formamides was six times more productive when the RuH2 catalysts were immobilized on ZnO rather than on non-Lewis acidic supports. We also demonstrated that Cu(I) phosphine catalysts are much more active inside of metalorganic frameworks (MOFs) containing Lewis acidic Zr nodes. The activation of CO2 on the Lewis acid sites led to a catalyst that converts CO2 to MeOH at as low as 110 °C in the gas phase. 

How adsorption and mass transport phenomena affect immobilized catalysts

One of the biggest differences for immobilized catalysts that often goes overlooked is that on the surface of a material the surroundings of a catalyst are much different than in a bulk liquid solution. For example, the materials can have vastly different polarity, pore sizes, pore shapes, and coordinating capability than a liquid solvent that works particularly well. Given that these factors all also control how the reactants and products of the catalytic reaction interact with the surface, it is possible that the conditions inside of the pore differ greatly from those in solution. This is often seen in confined systems, or systems in which the catalyst is not only immobilized on a surface but completely encased inside of small mesopores (2 – 8 nm diameter). In such cases, favorable interactions between reactants and the pore wall could lead to selective adsorption of the reactant, which would result in much higher local concentrations of the reactants in the material pores than in solution. This was the case when we found that the rate of formic acid dehydrogenation using confined systems was ca. 20 x slower under confinement than in solution. In collaboration with the groups of Prof. Thomas Sottmann (neutron scattering) and Prof. Niels Hansen (MD simulations), we were able to show that formic acid selectively adsorbs within the pores of the material to give a local concentrations that is ca. 6x higher than in solution. Since formic acid dehydrogenation is self-inhibited at high concentrations, this results in the confined catalyst being much less active than unconfined catalysts. 

Simple wet impregnation of molecular catalysts inside of solids also functions as a way of improving the catalytic activity for gas-phase catalytic reactions. For example, Crabtree’s catalyst can be impregnated inside of various porous solids very simply and in every case shows higher activity and recyclability than in DCM solution. This is because in the solids the gases are present inside the pores in much higher concentrations and are transported into the pores at higher rates than in DCM under typical reaction conditions. In addition, Crabtree’s catalyst in SBA-15 shows almost no deactivation over multiple recycling stages, giving TON > 17,000, an astounding improvement over the homogeneous catalyst. In these impregnated molecular catalysts, the support material can be thought of as a solid solvent, which has a higher solubility of the gases and higher rates of gas dissolution than typical organic solvents. Therefore, optimizing the materials properties for a given gas-phase reaction makes very high activities for industrially relevant gas-phase reactions possible at much higher temperatures than in organic solvents, in flow, with a much lower environmental footprint.