Democratizing Reaction Kinetics through Machine Vision and Learning
Mitchell Baumer, Liliana Gallegos, Dylan Anstine, Andrew Kubaney, Jose Regio, Olexandr Isayev, Stefan Bernhard, Gabe Gomes
Highlight
Democratizing Reaction Kinetics through Machine Vision and Learning.
Abstract
We present an innovative methodology for measuring amide coupling reaction rates by monitoring pH changes via indicator dyes, achieving precision comparable to traditional NMR techniques, called PRISM (Parallelized Reaction-rates via Indicator Spectrometry using Machine-vision) The experimental design, enabled by a serial dilution, allowed for measuring twelve rate constants concurrently, spanning more than four orders of magnitude using 96-well plates, with 1,162 total rate constants collected. Moreover, the instrumentation is 3D-printed, with the remaining components comprising readily available and cost-effective hardware, promoting the democratized use of this technique to generate uniform data sets. Validation with 19F-NMR confirmed PRISM’s reliability. Computational investigations reveal a concerted asynchronous SN2 mechanism, with base-catalyzed pathways exhibiting the lowest energy barriers. To complement the PRISM rate dataset, we developed a classification model that achieves high accuracy for out-of-distribution reactants in determining rate measurability, and a chemically rich graph neural network regression model for predicting quantitative reaction rates. This approach provides a framework that offers a resource-efficient strategy for studying reaction kinetics, which can be applied to other reaction classes.
Keywords
Cite This Paper
@article{Baumer2025,
author = {Baumer, Mitchell and Gallegos, Liliana and Anstine, Dylan and Kubaney, Andrew and Regio, Jose and Isayev, Olexandr and Bernhard, Stefan and Gomes, Gabe},
title = {Democratizing Reaction Kinetics through Machine Vision and Learning},
year = {2025},
doi = {10.26434/chemrxiv-2025-4tk40},
url = {http://dx.doi.org/10.26434/chemrxiv-2025-4tk40},
publisher = {American Chemical Society (ACS)},
keywords = {neural network, graph neural},
researchAreas = {reactions-reactivity, ml-potentials},
highlight = {Democratizing Reaction Kinetics through Machine Vision and Learning.}
} Copied to clipboard!
Related Research Areas
Related Publications
Accurate Ring Strain Energy Predictions with Machine Learning and Application in Strain-Promoted Reactions
(2024)
Ring strain energy (RSE) is crucial for understanding molecular reactivity.
Teaching a neural network to attach and detach electrons from molecules
Nature Communications, 12 (2021)
Abstract Interatomic potentials derived with Machine Learning algorithms such as Deep-Neural Networks (DNNs), achieve the accuracy of high-fidelity quantum mechanical (QM) methods in areas traditionally dominated by empirical force fields and allow performing massive simulations.
Transferable Machine Learning Interatomic Potential for Pd-Catalyzed Cross-Coupling Reactions
(2025)
Finding efficient substrate-catalyst combinations for palladium-catalyzed cross-coupling reactions remains a critical challenge in synthetic chemistry, with broad implications for pharmaceutical and materials manufacturing.
AIMNet2‐NSE: A Transferable Reactive Neural Network Potential for Open‐Shell Chemistry
Angewandte Chemie International Edition (2025)
Abstract Open‐shell systems such as radical intermediates are central to radical polymerization (RP), combustion, catalysis, and many other chemical and industrial processes, yet their accurate modeling presents significant computational challenges.
AIMNet2‐NSE: A Transferable Reactive Neural Network Potential for Open‐Shell Chemistry
Angewandte Chemie (2025)
Abstract Open‐shell systems such as radical intermediates are central to radical polymerization (RP), combustion, catalysis, and many other chemical and industrial processes, yet their accurate modeling presents significant computational challenges.