The mystery of metal oxides

Previously, it was believed that, in order to cleave the C-O bond selectively on heterogeneous catalysts, one had to resort to expensive noble metals (platinum, palladium, ruthenium). Besides their high cost, such catalysts were also not very selective, i.e. they produced undesirable side products, reducing the efficiency of the overall process from the economic point of view.

However, in 2013-2014 it was discovered in CCEI that the ruthenium catalyst, while still being a noble metal, was capable of breaking the C-O bond very selectively, effectively valorizing two essential biomass "platform chemicals" - furfural and HMF.

As it often happens in science, the story turned out to be more complicated, and we (in fact, our colleagues) realized that the original ruthenium catalyst was in fact a mixture of metal and metal oxide (ruthenium dioxide), and it is the oxide that facilitates the crucial C-O bond breaking  step (usually a chemical reaction consists of several steps).

In our Nature Catalysis paper, we generalized those early findings by demonstrating that besides ruthenium oxide (which is expensive and unstable), there are many other metal oxides (cheaper and more stable) that can drive the C-O bond scission reaction.

We should note that metal oxides have been proposed and used for the C-O bond activation previously; those earlier publications focused on one material at a time. In our work, we presented the unified view of metal oxide catalytic properties and their behavior for this reaction.

Perhaps more importantly, we introduced two catalytic descriptors. The descriptor is a property of the material, which, on the one hand, can be easily calculated or measured, and, on the other hand, correlates with the catalyst performance. Descriptors replace experiments in a certain sense - if someone hypothesizes that metal oxides A, B, and C should be effective in breaking the C-O bond, they can look up the oxide descriptors first to determine whether their hypothesis is valid or not, instead of carrying out expensive and time-consuming experiments.

Metal oxides are generally not as well studied as metals; what we did here has the potential to guide discovery in all sorts of heterogeneous catalysis processes, from industrial processes to environmental catalysis. We introduced a way of thinking about oxide catalysts that had never been previously published. 

The approach

Regarding first principles calculations and microkinetic modeling, they helped us study catalytic properties of metal oxides on the atomic level, i.e. effectively at higher resolution. While first principles calculations are common in catalysis, its the microkinetic modeling which allowed us to bridge theory and experiment across different time and length scales.

Regarding experimental methods, we needed to measure the C-O bond scission reaction rate over these catalysts, so we ran liquid-phase batch reactions and measured the products formed. We  needed to keep track of the oxidation state of the catalysts (so we could be convinced that they actually stay oxidized), so we ran characterization studies before and after reaction. To cap it off, we measured the oxidation state DURING reaction (operando X-ray absorption spectroscopy) and made sure that the ex situ data corresponded to what we measured during reaction. This is important, because the measurements were quite challenging.

The future 

Since now we have understanding of so-called monometallic metal oxides , the next logical step in research is to consider combinations of metal oxides and search for those that would appear near the top of the volcano. Such so-called bimetallic oxides should be simultaneously stable and active for the C-O bond activation reaction, one of the essential steps in biomass conversion to value-added fuels and chemicals.        

To learn more about this research, click the link to read the paper: https://www.nature.com/articles/s41929-019-0234-6