Accelerating Discovery of Metal–Organic Frameworks for Methane Adsorption with Hierarchical Screening and Deep Learning

Abstract

In recent years, machine learning (ML) methods have made significant progress, and ML models have been adopted in virtually all aspects of chemistry. In this study, based on the crystal graph convolutional neural networks algorithm, an end-to-end deep learning model was developed for predicting the methane adsorption properties of metal–organic frameworks (MOFs). High-throughput grand canonical Monte Carlo calculations were carried out on the computation-ready, experimental MOF database, which contains approximately 11 000 MOFs, to construct the data set. An area under the curve of 0.930 for the test set proved the reliability of the developed deep learning model. To assess the transferability of the model, we applied it to predict the methane adsorption volume for some randomly selected covalent organic frameworks and zeolitic imidazolate framework materials. The results indicated that the model could also be suitable for other porous materials. We also applied it to the hierarchical screening of a hypothetical MOFs database (∼330 000 MOFs). Four hypothetical MOFs were demonstrated to have the highest performance in methane adsorption. A calculated maximum working capacity of 145 cm3/cm3 at 5–35 bar and 298 K indicated that the hypothetical MOF is close to the Department of Energy’s 2015 target of 180 cm3/cm3. Further analyses on all screened out MOFs established correlations between some structural features with the working capacity. The successful incorporation of ML and hierarchical screening can accelerate the discovery of new materials not just for gas adsorption, but also other areas involving interactions in materials and molecules.

High‐throughput computational screening of porous polymer networks for natural gas sweetening based on a neural network

“…It is indicated that the screened 390 PPNs with the high‐performance for acid gas capture exhibit the similar structures. CGCNN algorithms can high‐efficiently extracted feature information from periodic crystal systems and converted it into graph inputs of the convolutional neural network, which can be trained to predict various targeted properties of the crystals 47,51 . Figure 8B shows the parity plots of CGCNN predicted vs. GCMC simulated APS.…”

The application of machine learning for predicting the methane uptake and working capacity of MOFs

“…Multiple linear regression analysis has been employed to develop equations to describe methane working capacities at 35–5.8 bar considering the database obtained from high-throughput GCMC calculations 61 of the CoRE MOF database (11 000 MOFs). The details of the methane sorption simulation via GCMC simulations are presented in that work.…”

“…The accuracy of the prediction may be further enhanced by implementing in the equation some more geometrical descriptors: 61 void fraction, LCD (largest cavity diameter) and PLD (pore limiting diameter)Working_capacity (35–5.8 bar) = 36.191 + 0.023 × Sa − 3.405 × Pv − 14.153 × Dc + 35.154 × Vf + 0.689 × LCD − 0.695 × PLD R 2 = 0.899; MAE = 9.26 cm 3 g −1 ; MSE = 159.78; RMSE = 12.63 cm 3 g −1 .…”

Machine Learning Assisted Prediction of Cathode Materials for Zn‐Ion Batteries

“…of thermodynamic, mechanical, and electronic properties for 2-D materials, [31] the formation energy of compounds, [32] and the methane adsorption properties of metal-organic frameworks, [33] etc.…”

“…One particular example is the recent development of the crystal graph convolutional neural network (CGCNN), [ 29 ] an ML tool that encodes the crystal structure as a "graph," has enabled the accelerated design of crystalline materials with desired properties, while also providing fundamental insights on the elementary/structural contributions to the energetics and properties of the materials. It has been successfully applied to the design of various crystalline materials for specific properties and applications, including the screening of solid electrolytes for dendrite suppression in lithium battery, [ 30 ] prediction of thermodynamic, mechanical, and electronic properties for 2‐D materials, [ 31 ] the formation energy of compounds, [ 32 ] and the methane adsorption properties of metal‐organic frameworks, [ 33 ] etc.…”