A Computational Workflow for Cost-Effective Synthesis of Inorganic Materials: Integrating Thermodynamics, Cellular Automata, Machine Learning, and Commercial Databases

David Garcia Valcarce, Eduardo Abenza, Roberto Gómez-Espinosa Martín

Keywords: Material synthesis, Thermodynamic modeling, Machine learning, Cellular automata, Process optimization, Economic efficiency, Inorganic materials, Data-driven workflows

Abstract:

This work presents a workflow designed to enhance economic efficiency in indus-
trial processes by optimizing material synthesis through thermodynamic modeling,
cellular automata, machine learning (ML), and commercial databases. The method-
ology enables the systematic identification and evaluation of synthesis routes that
balance thermodynamic performance with economic profitability. As a proof of
concept, the workflow was applied to the production of Ca2SiO4, a key refrac-
tory material for boilers and ship engines. Three novel synthesis recipes and four
optimal commercial reagents (CaO, SiO2, Ca3SiO5, and Si3N4) were identified,
demonstrating both improved process efficiency and cost-effectiveness. Beyond
this case study, the approach is broadly applicable to a wide range of inorganic sys-
tems, offering a scalable path toward maximizing economic efficiency in material
production. These results highlight the potential of data-driven workflows to accel-
erate the development of sustainable and competitive industrial manufacturing.

Submission Track: Findings, Tools & Open Challenges

Submission Category: AI-Guided Design

Institution Location: Madrid, Spain

AI4Mat Journal Track: Yes

AI4Mat RLSF: Yes

Submission Number: 97