KIT | KIT-Bibliothek | Impressum | Datenschutz

Bayesian Optimisation of Multi-Component Promotion: A Case Study on Pt/Al$_2$O$_3$ Catalysts for Perhydro Benzyltoluene Dehydrogenation

Biesinger, Jakob 1; Tiezema, Torben 2; Helleckes, Laura M.; Wolf, Moritz ORCID iD icon 1,3
1 Institut für Katalyseforschung und -technologie (IKFT), Karlsruher Institut für Technologie (KIT)
2 Karlsruher Institut für Technologie (KIT)
3 Engler-Bunte-Institut (EBI), Karlsruher Institut für Technologie (KIT)

Abstract:

The rational design of multi-promoter catalysts is challenged by the large combinatorial design space and the high experimental costs associated with catalyst synthesis and testing. Bayesian optimisation (BO) has emerged as a promising machine-learning framework for data-efficient optimisation of expensive black-box problems. BO provides immense potential for accelerating discovery in catalyst promotion due to the combinatorial explosion of vast multi-variable spaces, in particular for multi-component promotion. In this work, BO is introduced as a practical decision-support tool for the optimisation of promoter compositions of a 0.30 wt.% Pt/Al₂O₃ catalyst for the dehydrogenation of the liquid organic hydrogen carrier (LOHC) perhydro benzyltoluene. Beyond demonstrating the practical application of BO, the presented framework is systematically analysed with respect to surrogate model selection and acquisition strategy, providing general guidance for its application in catalyst optimisation. Conventional analysis of an initial experimental dataset reveals characteristic trends for individual promoter effects on catalyst activity and selectivity, which highlights the complex trade-offs governing catalyst design. ... mehr


Postprint §
DOI: 10.5445/IR/1000199244
Veröffentlicht am 18.11.2026
Originalveröffentlichung
DOI: 10.1039/D6CY01164K
Cover der Publikation
Zugehörige Institution(en) am KIT Engler-Bunte-Institut (EBI)
Institut für Katalyseforschung und -technologie (IKFT)
Publikationstyp Zeitschriftenaufsatz
Publikationsjahr 2026
Sprache Englisch
Identifikator ISSN: 2044-4753, 2044-4761
KITopen-ID: 1000199244
HGF-Programm 38.03.02 (POF IV, LK 01) Power-based Fuels and Chemicals
Erschienen in Catalysis Science & Technology
Verlag Royal Society of Chemistry (RSC)
Nachgewiesen in OpenAlex
Relationen in KITopen
KIT – Die Universität in der Helmholtz-Gemeinschaft
KITopen Landing Page