RaspAssistant automates adsorption simulation workflows in porous materials
RaspAssistant as an Agentic Framework for Automating Adsorption Simulation Research in Porous Materials
Artificial Intelligence
Summary
Adsorption studies in porous materials need many complex steps like preparing structures, choosing force fields, running simulations, and checking results. The authors tested RaspAssistant, a system of coordinated agents, to manage all these steps automatically from start to finish. They found it could run workflows correctly, validate simulations, and analyze results consistently across various tests. The system also helped refine models and screen many materials automatically, showing it can handle detailed adsorption research with minimal human input.
What this means in practice
- •For chemical engineers: Automate end-to-end adsorption simulation workflows to accelerate porous material analysis and design.
- •For materials informatics teams: Automatically prepare, validate, and analyze large porous material datasets for adsorption and separation performance screening.
Authors
Marko Petković, Susanne Gooijer, José Manuel Vicent-Luna, Vlado Menkovski, Sofía Calero
Abstract
Porous-material adsorption studies require coordinated structure preparation, force-field selection, simulation, validation, and interpretation, making end-to-end automation difficult. We test whether RaspAssistant, a hierarchical multi-agent framework, can carry an adsorption study from force-field preparation through validation and downstream separation analysis. Across twelve multi-agent workflow-reproduction runs spanning four adsorption and separation benchmarks, overall scores ranged from 8.50 to 10.00 on a rubric evaluating workflow correctness, simulation validity, and analysis consistency. In a separate eight-case force-field extraction benchmark, the mean weighted score was 0.967, combining parameter agreement and RASPA readiness. In open-ended studies, the system diagnosed and refined an acetylene force field and, after validating a CO$_2$ force field against MFI experiments, screened 197 pure-silica zeolites that passed automated input checks. These results demonstrate autonomous workflow execution after task specification and input provision: RaspAssistant coordinates, validates, adapts, and documents extended adsorption-simulation workflows while exposing model sensitivity and methodological limitations for expert review.