Machine learning for efficient CO2 sequestration in cementitious materials - a data-driven method

Aug 5, 2026

Upload by Rowan

Extensive experimental evidence shows CO2 sequestration by cementitious materials can help address rising carbon emissions, but experimental-only or simple empirical approaches struggle to capture the full influence of complex cement systems. This paper develops a data-driven machine-learning framework using Decision Tree, Random Forest, and XGBoost trained on datasets built from literature and collected data. Results indicate XGBoost outperforms linear regression, and SHAP analysis highlights cement type as a key driver of carbonation depth, including high-potential CEM II/B-LL and CEM II/B-M.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages10

Reading Time25 min

AccessFree to download

Overview

Background and motivation

Challenges in conventional studies

Data-driven machine learning approach

Methods and models

Decision Tree and Random Forest

XGBoost model

SHAP interpretation

Key findings

Cement type effects on carbonation depth

High-potential cement types

Implications

This is a repository copy of Machine learning for efficient CO2 sequestration in cementitious materials: a data-driven method. White Rose Research Online URL for this paper: [https://eprints.whiterose.ac.uk/225881/](https://eprints.whiterose.ac.uk/225881/) Version: Published Version Article: Sun, Y. , Zhang, C. , Wei, Y.-H. et al. (5 more authors) (2025) Machine learning for efficient CO2 sequestration in cementitious materials: a data-driven method. npj Materials Sustainability, 3 (1) . 9. ISSN 2948-1775 [https://doi.org/10.1038/s44296-025-00053-z](https://doi.org/10.1038/s44296-025-00053-z) Reuse This article is distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs (CC BY-NC-ND) licence. This licence only allows you to download this work and share it with others as long as you credit the authors, but you can’t change the article in any way or use it commercially. More information and the full terms of the licence here: [https://creativecommons.org/licenses/](https://creativecommons.org/licenses/) Takedown If you consider content in White Rose Research Online to be in breach of UK law, please notify us by emailing [eprints@whiterose.ac.uk](eprints@whiterose.ac.uk) including the URL of the record and the reason for the withdrawal request. [eprints@whiterose.ac.uk](eprints@whiterose.ac.uk)[ ](eprints@whiterose.ac.uk)[https://eprints.whiterose.ac.uk/](https://eprints.whiterose.ac.uk/) [https://doi.org/10.1038/s44296-025-00053-z](https://doi.org/10.1038/s44296-025-00053-z) Machine learning for ef􀀁cient CO2 sequestration in cementitious materials: a data-driven method Check for updates uence of perature have tobe controlled to ensure the reliability and reproducibility of | SUNYanjie1, ZHANG Chen1,2, WEI Yuan-Hao1, JIN Haoliang3, SHEN Peiliang4, POON Chi Sun4, YAN He5 & WEI Xiao-Yong1,6 | | | --- | --- | | Extensive experimental work has proved that CO2 sequestration by cementitious materials offers a promising venue for addressing the rising carbon emissions problem. However, relying merely on experiments on speciﬁc materials or some simple empirical methods makes it difﬁcult to provide a comprehensive understanding. To address these challenges, this paper applies three advanced machine-learning techniques (Decision Tree, Random Forest, and eXtreme Gradient Boosting (XGBoost)), with existing datasets coupling with data collected from the literature. The results show that theXGBoost model signiﬁcantly outperforms traditional linear regression approaches. In addition, aiding in the SHapley Additive exPlanations(SHAP), apart from the widely recognized factors, cement type was also investigated and shown its crucial role in affecting carbonation depth. CEM II/B-LL and CEM II/B-M are two types having high carbonation potential. The results enable the identiﬁcation of key factors inﬂuencing CO2 sequestration through cement and provide insights into optimizing experimental design. | | | Greenhouse gases, particularly carbon dioxide (CO2), are major and wellagreed contributors to climate change, leading toa global push for strategies aimed at reducing carbon emissions. This has resulted in signiﬁcant efforts focused on carbon capture, utilization, and storage (CCUS), which seek to mitigate CO2 levels in the atmosphere and reduce the impacts of climate change1. Among the various CCUS strategies, one of the most promising approaches to close the carbon loop involves the sequestration of CO2 through hydration products in cementitious materials, which was previously considered the major source of CO2 emission2–4. The main advantage ofcementitious material carbonation is its favorable thermodynamics5. However, one of the main challenges is the slow kinetics ofthe carbonation reaction, which limits the efﬁciency and overall sequestration capacity of these materials.
The CO2 absorption capacity of cementitious materials is collectively inﬂuenced by various factors, inc

FAQs

Why do conventional experimental or empirical approaches have limitations for CO2 sequestration in cementitious materials?

Because carbonation behavior is governed by many interacting factors in a complex cement system, experiments or simple empirical methods often cannot generalize well to new materials or conditions.

Which machine-learning models were used in the study and what was the main performance result?

Decision Tree, Random Forest, and XGBoost were used. XGBoost significantly outperformed traditional linear regression approaches.

How did SHAP contribute to understanding carbonation depth, and what key factor was identified?

SHAP provided feature importance insights beyond widely recognized factors. Cement type was shown to play a crucial role in affecting carbonation depth, with CEM II/B-LL and CEM II/B-M exhibiting high carbonation potential.

Aug 5, 2026

Upload by Rowan