ActMOF: Mapping Reticular Synthesis through Active Learning
Abstract
Reticular chemistry allows metal-organic frameworks (MOFs) to be designed from their building units, yet their synthesis remains constrained by sparse experimental data, noisy crystallization outcomes, and the absence of benchmarks for comparing active-learning strategies. While data-driven approaches such as Bayesian optimization (BO) can improve synthesis success rates, selecting appropriate optimizers beforehand for mapping MOF crystallization landscapes remains largely empirical. Here we present ActMOF, an experiment-grounded benchmark for comparing search strategies in aluminum-carboxylate MOF synthesis. ActMOF is emulated from experimentally measured microwave syntheses with each encoded condition assigned a crystallinity index calculated based on diffraction peak intensity and full width at half-maximum. By benchmarking policies guided by Gaussian processes (GP), tree ensembles and large language models (LLM), ActMOF placed the GP family among the leading policies for active learning in sparse synthesis landscapes. It then demonstrated transferability by guiding the microwave synthesis of six reticularly related aluminum MOFs. All six campaigns gave crystalline products within a fixed budget of 12 experiments each, and established a reusable testbed for AI-driven MOF synthesis.
Related
- Reticular Synthesis of Ultrastable Dynamic Zinc-Pyrazolate Frameworks for Efficient CO₂ Photoreduction.
- Exposure States, Environmental Fate, and Functional Effects of Microplastics in Soil: Bibliometric Mapping and Critical Mechanistic Synthesis
- LinkerGen: Guided Autoregressive Generation of Organic Linkers for Reticular Chemistry
- A PN adamantanoid ligand for reticular materials
- Agentic Workflow for High-Throughput Synthesis Campaigns: Mapping the Persistent Phosphor Ca2ZnSi2O7:Eu2+,Nd3+
BibTeX
@misc{sheng2026actmof,
title = {{ActMOF: Mapping Reticular Synthesis through Active Learning}},
author = {Hongrui Sheng and Shutian Liang and Reid A. Coyle and Shyam Chand Pal and Peter Walther and Bin Feng and Xinyu Ding and Zhiling Zheng},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.19404},
eprint = {2610.19404},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.19404},
}
RIS
TY - UNPB
TI - ActMOF: Mapping Reticular Synthesis through Active Learning
AU - Hongrui Sheng
AU - Shutian Liang
AU - Reid A. Coyle
AU - Shyam Chand Pal
AU - Peter Walther
AU - Bin Feng
AU - Xinyu Ding
AU - Zhiling Zheng
PY - 2026
DA - 2026/10/14
PB - arXiv
DO - 10.48550/arxiv.2610.19404
UR - https://arxiv.org/abs/2610.19404
AB - Reticular chemistry allows metal-organic frameworks (MOFs) to be designed from their building units, yet their synthesis remains constrained by sparse experimental data, noisy crystallization outcomes, and the absence of benchmarks for comparing active-learning strategies. While data-driven approaches such as Bayesian optimization (BO) can improve synthesis success rates, selecting appropriate optimizers beforehand for mapping MOF crystallization landscapes remains largely empirical. Here we present ActMOF, an experiment-grounded benchmark for comparing search strategies in aluminum-carboxylate MOF synthesis. ActMOF is emulated from experimentally measured microwave syntheses with each encoded condition assigned a crystallinity index calculated based on diffraction peak intensity and full width at half-maximum. By benchmarking policies guided by Gaussian processes (GP), tree ensembles and large language models (LLM), ActMOF placed the GP family among the leading policies for active learning in sparse synthesis landscapes. It then demonstrated transferability by guiding the microwave synthesis of six reticularly related aluminum MOFs. All six campaigns gave crystalline products within a fixed budget of 12 experiments each, and established a reusable testbed for AI-driven MOF synthesis.
ER -