Machine learning speeds up heterogeneous catalysis simulations dramatically

Good news! This is only the beginning!

"In heterogeneous catalysis, calculations require complicated computational processes due to the variability of reaction pathways and possibilities. Now, thanks to a clever combination of programming and machine learning, researchers have achieved a dramatic increase in the speed of simulations, enhancing the energy efficiency of an otherwise resource-hungry process. The results, reported for reactions to convert carbon dioxide into fuels, could rapidly translate into other industrially relevant reactions, such as depolymerisation and biomass valorisation. ...

Whereas traditional density functional theory (DFT) programmes predict up to 500 steps in 100 processing hours, this new solution speeds the search by orders of magnitude – simulating 370,000 possible pathways in a similar time. ..."

From the abstract:

"The rationalization of catalytic processes relies on the fundamental understanding of competing reaction mechanisms driving reactants to products. The list of elementary steps composing the reaction networks is proposed based on chemical intuition and evaluated via density functional theory. This approach is limited by the size of the network and disregards alternative paths.

Here we present the Catalytic Automated Reaction Evaluator (CARE), a flexible end-to-end framework for heterogeneous catalysis composed of

(1) a rule-based reaction network generator,

(2) a thermodynamic and kinetic parameter evaluator powered by state-of-the-art machine learning models and

(3) a fast microkinetic solver.

CARE reproduces the experimental activity trends in methanol decomposition, identifies the selectivity to C3 products in CO2 electroreduction and generates the Fischer–Tropsch synthesis mechanism including 370,000 reactions reaching C6 products. This comprehensive framework enables the exploration of thermal and electrocatalytic reactions previously not amenable to atomistic simulations."

Machine learning speeds up heterogeneous catalysis simulations dramatically | Chemistry World

An end-to-end framework for reactivity in heterogeneous catalysis (open access)

Fig. 1: CARE workflow from network generation to reactivity analysis.