Digital Chemical Engineering Across Scales and Domains: Case studies in electrochemistry, automated labs, and reticular materials

This thesis develops a vision for digital chemical engineering that connects modelling, automation, and data integration across scales and domains to accelerate research and enable chemical innovation. Representative case studies in electrochemistry, laboratory automation, and reticular materials were selected for their relevance to key challenges in sustainable technology development. The first part focuses on mechanistic understanding of electrochemical carbon dioxide reduction, a reaction central to carbon utilisation. It begins by analysing the challenges of connecting atomistic reaction mechanisms to reactor-scale behaviour under industrial conditions, motivating the need for integrated modelling and standardised data reporting. Building on this foundation, we develop a fully elementary microkinetic model calibrated against experimental data, and complement it with ultra-sensitive product analysis of high-current gas-diffusion electrodes to independently probe mechanistic complexity. Together, these approaches reveal complex branching behaviour and previously unreported multi-carbon products, refining our mechanistic picture of CO2 electroreduction. The second part demonstrates how The World Avatar infrastructure can accelerate research by automating tasks traditionally performed by human researchers. We show how ontology-based knowledge graphs improve accessibility and enable the integration of heterogeneous data, from reaction mechanisms to material structures. A natural language interface facilitates intuitive exploration of complex datasets, reducing barriers to data reuse and interpretation. Building on this semantic foundation, we develop a Digital Lab Framework that represents laboratory equipment and infrastructure as interconnected digital twins. Case studies demonstrate automation of resource optimisation, asset tracking, and inventory management in scalable, interoperable, and FAIR-compliant research environments. The final part addresses the autonomous design and synthesis of reticular materials. We present a pipeline that extracts synthesis procedures for metal-organic polyhedra from the literature using large language models, formalises them using ontologies, and integrates them into a machine-readable dataset. This knowledge base supports a rule-based method for predicting synthesis routes for novel materials by analogy to known protocols, offering a foundation for automated material discovery.

Name dc:type.qualificationname

Doctor of Philosophy (PhD)

Level dc:type.qualificationlevel

Doctoral

Grantor dc:publisher.institution

University of Cambridge

Year dc:date.issued

2025

Author dc:creator

• Rihm, Simon

Advisor dc:contributor.advisor

• Kraft, markus

• Multi-scale modelling
• Knowledge graphs
• Research automation
• Sustainability
• Electrochemistry
• Metal-organic polyhedra

Licence

• https://www.repository.cam.ac.uk/bitstreams/c9f595ab-be44-47f1-95ef-a007bf3e0f6e/download
• https://creativecommons.org/licenses/by/4.0/

Language dc:language

eng

Author Identifier

0000-0001-8342-7269

OAI identifier oai:identifier

oai:www.repository.cam.ac.uk:1810/391941