Optimizing Metal Organic Framework (MOF) Material Properties for Atmospheric Water Harvesting: an ML Approach with xAI

Overview

Water scarcity is an important issue for our global population. Natural sources of water are insufficient to meet this demand. However, the atmosphere is an untapped resource which at any point contains 13,000 trillion liters of water. Given this, in recent years, atmospheric water harvesting has become an important approach; however, many existing strategies require significant costs and are ineffective. Metal organic frameworks (MOFs) are novel materials that have extremely high surface area per volume and tremendous chemical variability and tunability. Given these properties, they have emerged as optimal materials for gas adsorption, especially water vapor, and have been demonstrated to be effective in atmospheric water harvesting. Given there are >100,000 MOF structures, finding optimal MOFs for atmospheric water harvesting is challenging, and machine learning approaches have emerged as a technique to better identify optimal materials, including MOFs. In this project, a Random Forest Classifier machine learning model was developed to help predict water adsorption (and indirectly desorption) behavior for MOFs, based on structural and chemical characteristics. Pore limiting diameter emerged as an impactful structural parameter, and polar atom fraction was notably an important chemical parameter in model performance. Based on a testing dataset, the model also achieved 70% accuracy in identifying whether a MOF had type V water adsorption behavior and/or was an early high adsorber of water at low relative humidity levels. Overall, this project demonstrated the potential of a machine learning classifier model to enable the identification of MOFs that could be optimal for water harvesting applications.

Competition history

• CSEF 2026 Environmental Engineering (Senior Division) · Entry S-11-15

Related projects

Closest projects by meaning, across every fair and year in the corpus.

Browse more like this

• All Environmental Engineering projects
• CSEF Environmental Engineering projects
• Every CSEF 2026 project