Metal-organic frameworks: Role of artificial intelligence and machine learning algorithms for efficient discovery, design, synthesis and prediction of CO2 capture capacity - A state of art review

Authors

• Aayzaz Ahmed, Universiti Malaya
• Mirza Aamish Hassan Khan, The University of Texas at El Paso College of Engineering
• George Uwadiegwu Alaneme, Kampala International University
• Tariq Ali, Swedish College of Engineering and Technology
• Muhammad Zeeshan Qureshi, University of Engineering and Technology Taxila
• Ayed Eid Alluqmani, Islamic University of Madinah
• Ali Ajwad, University of Management and Technology Lahore

Document Type

Review

Publication Date

4-22-2026

Abstract

Metal-organic frameworks (MOFs) have emerged as promising materials for CO2 capture, owing to their high surface areas and diverse chemical functionalities. However, their practical application is often hindered by stability concerns, particularly in water-rich environments and under fluctuating thermal and chemical conditions. This paper provides a state-of-the-art review of MOFs, focusing on their water, thermodynamic, kinetic and chemical stabilities and functions. Additionally, comparative analysis of MOF stability types and key design rules for their stability enhancement are discussed. Furthermore, role of artificial intelligence (AI) and machine learning (ML) in revolutionizing MOF research for their optimized formation (discovery, screening, identification, design, synthesis) and prediction of CO2 capture capacity has been reviewed. Moreover, this paper outlines the challenges and limitations like lack of datasets, feature interpretability, model transferability, scalability and practical applicability. The findings revealed that High-throughput screening, AI-ML-driven optimized design and synthesis of Aluminum Porphyrin-MOF (Al-PMOF) which produced with over 80% yield in only 50 min instead of 16-h traditional method, making it 20 times faster than conventional approaches, while examining 320,000 hMOFs database. Additionally, AI-ML driven, Least Squares Support Vector Machine- genetic optimization (LSSVM-GO) a hybrid ML model has shown prediction accuracy up to R-square value of 0.9797 for CO2 adsorption capacity of MOFs. In conclusion, AI-ML driven frameworks significantly outperform conventional approaches in the design, discovery, synthesis and prediction of CO2 adsorption in MOFs. Looking ahead, the integration AI-ML models with autonomous synthesis by utilizing robotic/additive manufacturing to reduce costs, time and expedite the creation of innovative MOFs.

Keywords

Artificial intelligence (AI), CO2 adsorption, CO2 capture, Gas separation, Machine learning (ML), Metal-organic frameworks (MOFs)

Publication Title

Separation and Purification Technology

ISSN

1383-5866

DOI

10.1016/j.seppur.2026.136737

Recommended Citation

Ahmed, Aayzaz; Khan, Mirza Aamish Hassan; Alaneme, George Uwadiegwu; Ali, Tariq; Qureshi, Muhammad Zeeshan; Alluqmani, Ayed Eid; and Ajwad, Ali, "Metal-organic frameworks: Role of artificial intelligence and machine learning algorithms for efficient discovery, design, synthesis and prediction of CO2 capture capacity - A state of art review" (2026). Research Publications (2026 to 2030). 125.
https://knova.um.edu.my/research_publications_2026_2030/125

Volume

388

Publisher

Elsevier