ZeoSyn - A Comprehensive Zeolite Synthesis Dataset - Enabling Machine-Learning Rationalization of Hydrothermal Parameters

Aug 5, 2026

Upload by Eden

Zeolites are nanoporous aluminosilicates with tailored porous structures, widely used in catalysis, gas separation, and ion exchange. Hydrothermal synthesis enables control over composition, crystallinity, and pore size, yet the complex interactions among synthesis parameters make synthesis–structure optimization challenging. ZeoSyn provides 23,961 hydrothermal synthesis routes covering 233 zeolite topologies and 921 OSDAs, linking gel composition, reaction conditions, OSDAs, and resulting products. The work trains a >70% accurate ML classifier and applies SHAP, including an aggregation strategy, to identify key parameters across over 200 frameworks and demonstrate phase-selective and intergrowth synthesis guidance.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages15

Reading Time38 min

AccessFree to download

Abstract

Introduction

Zeolite synthesis background and challenges

Prior zeolite synthesis datasets and limitations

Dataset overview and methodology

ZeoSyn dataset scope and coverage

Machine-learning prediction and model interpretability

Applications

Phase-selective synthesis and intergrowth synthesis

This article is licensed under CC-BY 4.0 [http://pubs.acs.org/journal/acscii](http://pubs.acs.org/journal/acscii) Article ZeoSyn: A Comprehensive Zeolite Synthesis Dataset Enabling Machine-Learning Rationalization of Hydrothermal Parameters Elton Pan, Soonhyoung Kwon, Zach Jensen, Mingrou Xie, Rafael Gómez-Bombarelli, Manuel Moliner, Yuriy Román-Leshkov, and Elsa Olivetti * Cite This: ACS Cent. Sci. 2024, 10, 729−743 Read Online Downloaded via UNIV POLITECNICA DE VALENCIA on July 17, 2024 at 07:26:03 (UTC) . See [https://pubs.acs.org/sharingguidelines](https://pubs.acs.org/sharingguidelines) for options on how to legitimately share published articles. ACCESS Metrics & More Article Recommendations *sı Supporting Information ABSTRACT: Zeolites, nanoporous aluminosilicates with welldefined porous structures, are versatile materials with applications in catalysis, gas separation, and ion exchange. Hydrothermal synthesis is widely used for zeolite production, offering control over composition, crystallinity, and pore size. However, the intricate interplay of synthesis parameters necessitates a comprehensive understanding of synthesis−structure relationships to optimize the synthesis process. Hitherto, public zeolite synthesis databases only contain a subset of parameters and are small in scale, comprising up to a few thousand synthesis routes. We present ZeoSyn, a dataset of 23,961 zeolite hydrothermal synthesis routes, encompassing 233 zeolite topologies and 921 organic structure-directing agents (OSDAs). Each synthesis route comprises comprehensive synthesis parameters: 1) gel composition, 2) reaction conditions, 3) OSDAs, and 4) zeolite products. Using ZeoSyn, we develop a machine learning classifier to predict the resultant zeolite given a synthesis route with >70% accuracy. We employ SHapley Additive exPlanations (SHAP) to uncover key synthesis parameters for >200 zeolite frameworks. We introduce an aggregation approach to extend SHAP to all building units. We demonstrate applications of this approach to phase-selective and intergrowth synthesis. This comprehensive analysis illuminates the synthesis parameters pivotal in driving zeolite crystallization, offering the potential to guide the synthesis of desired zeolites. The dataset is available at [https://github.com/eltonpan/zeosyn](https://github.com/eltonpan/zeosyn)_dataset. ■ INTRODUCTION Zeolites are nanoporous, crystalline aluminosilicate materials with a wide range of industrial applications including catalysis, separations, and ion exchange. 1−3 In addition to composition, the crystalline structure and corresponding porous network are crucial in determining a zeolite’s suitability for a target application.4,5 While thousands of potential zeolite structures are thought to be thermodynamically accessible,6 only 264 have been synthesized7 highlighting a synthesis bottleneck to zeolite discovery and deployment. Zeolite synthesis has typically been based on trial-and-error methods guided by accumulated domain knowledge.8 The synthesis of zeolites is intricate, with numerous variables influencing the resultant zeolite structure.9 These factors include framework heteroatoms, the presence of inorganic and organic cations, between these factors across the entire field is lacking. Data science and machine learning have shown promise in generalizing some synthesis−structure relationships26−29 but have been limited to subsections of the zeolite design space due to a lack of data, which implies that larger datasets may generalize learning more broadly across the zeolite space. Previous works have curated zeolite synthesis datasets. Specifically, a dataset consisting of 1,200 unique synthetic routes for Ge-containing zeolites has been reported by Jensen et al.28 In the same vein, Yan et al. compiled a database of 1,600 synthetic records of open-framework aluminophosphate (AlPO) syntheses.30 However, these datasets cover a subset of frameworks, g

FAQs

What does the ZeoSyn dataset contain?

ZeoSyn contains 23,961 zeolite hydrothermal synthesis routes, including gel composition, reaction conditions, OSDA information, and the resultant zeolite products across 233 topologies and 921 OSDAs.

How is machine learning used in the study?

A machine-learning classifier predicts the resultant zeolite from a synthesis route with over 70% accuracy, learning synthesis–structure relationships from the dataset.

How do the authors identify which synthesis parameters matter most?

SHAP (SHapley Additive exPlanations) is used to uncover key synthesis parameters for more than 200 zeolite frameworks, and an aggregation approach extends SHAP insights across all building units.

Aug 5, 2026

Upload by Eden