Understanding Catalytic Biomass Conversion Through Data Mining

Abstract

Catalytic conversion of biomass is a key challenge that we chemists face in the twenty-first century. Worldwide, research is conducted into obtaining bulk chemicals, polymers and fuels. Our project centres on glucose valorisation via furfural derivatives using catalytic hydrogenation. We present here new results for a set of 48 bimetallic catalysts supported on silica, and demonstrate the application of data mining tools to identify major trends in the data. These results are combined with a full factorial data set for the hydrogenation of 5-ethoxymethylfurfural over alumina-supported transition metal catalysts. All the catalysts in the combined datasets were synthesized and tested for performance under identical conditions. This, combined with the fact that no combinations of metals were left out, enables the use of advanced data mining tools. The paper describes the data and highlights the relevant trends from a chemist's viewpoint.

Production of phenols from lignin-derived slurry liquid using iron oxide catalyst

“…From the perspective of fossil fuel depletion and the need to control carbon dioxide emissions, the production of useful chemicals from lignocellulose, which is the most abundant inedible biomass in nature, is becoming increasingly important [1][2][3].…”

Machine learning for heterogeneous catalyst design and discovery

“…Data mining methods are powerful ML tools to find nontrivial insights in big data and to help build predictive models. Efforts have been made to integrate data mining methods with heterogeneous or homogeneous catalysis data to promote catalyst characterization and to build quantitative structure‐property relationship models . An early study used data mining to help make predictive models of cyclohexene epoxidation yield by mesoporous titanium‐silicate catalysts .…”

“…Besides helping to extract predictive features, data mining can find trends in catalytic reactions . For example, selective hydrogenation of 5‐ethoxymethylfurfural was examined over 96 bimetallic catalysts and 16 metal catalysts supported on either SiO 2 or Al 2 O 3 . Each catalyst was tested in two solvents (diethyl carbonate, 1,4‐dioxane) and three temperatures.…”

A Quantitative Model for Alkane Nucleophilicity Based on C−H Bond Structural/Topological Descriptors

“…On the other hand, Aires and co‐workers took advantage of statistical techniques to estimate bond dissociation energies (BDEs) for a number of organic molecules, whereas Jensen, Alsberg, and co‐workers automated the design of organometallic molecules . Predictive modeling has also been provided by the groups of Rothenberg, Bo, and Paton, while one of our groups, among others, has used multivariate regressions to find free‐energy relationships between BDEs in organometallic chemistry …”