Hypothetical Metal-organic Frameworks Research Articles

MOFormer: Self-Supervised Transformer Model for Metal-Organic Framework Property Prediction.

Metal-organic frameworks (MOFs) are materials with a high degree of porosity that can be used for many applications. However, the chemical space of MOFs is enormous due to the large variety of possible combinations of building blocks and topology. Discovering the optimal MOFs for specific applications requires an efficient and accurate search over countless potential candidates. Previous high-throughput screening methods using computational simulations like DFT can be time-consuming. Such methods also require the 3D atomic structures of MOFs, which adds one extra step when evaluating hypothetical MOFs. In this work, we propose a structure-agnostic deep learning method based on the Transformer model, named as MOFormer, for property predictions of MOFs. MOFormer takes a text string representation of MOF (MOFid) as input, thus circumventing the need of obtaining the 3D structure of a hypothetical MOF and accelerating the screening process. By comparing to other descriptors such as Stoichiometric-120 and revised autocorrelations, we demonstrate that MOFormer can achieve state-of-the-art structure-agnostic prediction accuracy on all benchmarks. Furthermore, we introduce a self-supervised learning framework that pretrains the MOFormer via maximizing the cross-correlation between its structure-agnostic representations and structure-based representations of the crystal graph convolutional neural network (CGCNN) on >400k publicly available MOF data. Benchmarks show that pretraining improves the prediction accuracy of both models on various downstream prediction tasks. Furthermore, we revealed that MOFormer can be more data-efficient on quantum-chemical property prediction than structure-based CGCNN when training data is limited. Overall, MOFormer provides a novel perspective on efficient MOF property prediction using deep learning.

High-throughput screening of hypothetical metal-organic frameworks for thermal conductivity

Thermal energy management in metal-organic frameworks (MOFs) is an important, yet often neglected, challenge for many adsorption-based applications such as gas storage and separations. Despite its importance, there is insufficient understanding of the structure-property relationships governing thermal transport in MOFs. To provide a data-driven perspective into these relationships, here we perform large-scale computational screening of thermal conductivity k in MOFs, leveraging classical molecular dynamics simulations and 10,194 hypothetical MOFs created using the ToBaCCo 3.0 code. We found that high thermal conductivity in MOFs is favored by high densities (> 1.0 g cm−3), small pores (< 10 Å), and four-connected metal nodes. We also found that 36 MOFs exhibit ultra-low thermal conductivity (< 0.02 W m−1 K−1), which is primarily due to having extremely large pores (~65 Å). Furthermore, we discovered six hypothetical MOFs with very high thermal conductivity (> 10 W m−1 K−1), the structures of which we describe in additional detail.

Design and Screening of Metal-Organic Frameworks for Ethane/Ethylene Separation.

Separation of ethane and ethylene is considered to be industrially important for various chemical processes, but their similarities make the process expensive. In this study, we integrated computational screening with machine learning to find optimal metal-organic frameworks (MOFs) with high ethane/ethylene selectivity. Using our algorithm, a hypothetical MOF structure with an ideal adsorption solution theory selectivity of 3.6 at 298 K and 1 bar was discovered. Furthermore, structural analysis was implemented, and the full adsorption isotherm of some of the top structures was obtained.

Screening of Hypothetical Metal–Organic Frameworks for Xylene Isomers and Ethylbenzene Separation

A database of hypothetical metal–organic frameworks (hMOFs) was investigated in search of top-performing adsorbents for the separation process of xylene isomers and ethylbenzene mixture. The structural properties of individual hMOF are calculated to understand the correlation between the properties with adsorption capacity or selectivity. The high-throughput staged screening of the hMOF database is performed and screened. The simulations of hMOF with a gas mixture were conducted at 373 K and 30 kPa using a multicomponent grand canonical Monte Carlo method to obtain the para-xylene selectivity. The top five hMOFs showed significantly better performance of all and compared to previously discovered materials. The top hMOFs were further simulated at different pressures to obtain the adsorption isotherms. Further, the energetic contribution of organic linkers and inorganic metal nodes toward competitive adsorption was studied to further understand the enhanced interaction between hMOF and para-xylene.

High-Throughput Screening of Metal–Organic Frameworks Assisted by Machine Learning: Propane/Propylene Separation

The separation of a propane (C3H8)/propylene(C3H6) mixture is of paramount importance in the petrochemical industry. Metal–organic frameworks (MOFs), as a class of promising alternative to the traditional adsorbents, have garnered extensive interest. This study proposes a machine learning-assisted high-throughput screening strategy for the identification of suitable MOFs for C3H8/C3H6 separation, striving to accelerate the discovery of high-performance MOF candidates for this particular application. First, a chemical/geometric analysis-based prescreening is applied to a data set of 146 203 MOFs composed of an experimentally synthesized MOF database and a hypothetical MOF database, and MOFs with undesirable chemical/geometric features were excluded. Six structural and nine chemical descriptors were calculated for the remaining MOFs. Random Forest regression algorithm was applied to “learn” the relationship correlations between the features (chemical and/or structural) of MOFs and their C3H8/C3H6 separation capacity. Grand Canonical Monte Carlo (GCMC) simulations were applied to evaluate the C3H8/C3H6 separation performances of the randomly selected training and testing MOF samples. A performance prediction model based on chemical and structural descriptors was obtained with R2 equal to 0.96, which was employed for a separation performance prediction of the remaining MOFs. 2500 MOFs with potential to possess high C3H8/C3H6 separation performance were shortlisted by the prediction model. GCMC simulations were applied to calibrate the prediction results and validate of the machine learning model. MOFs with competitively high C3H8/C3H6 separation potential and regenerability were identified, and a comparison with MOFs reported in the literature was made, indicating that the proposed machine learning-assisted high-throughput screening approach is efficient and effective. Furthermore, structure–property correlation analysis was conducted.

The HEALED SBU Library of Chemically Realistic Building Blocks for Construction of Hypothetical Metal-Organic Frameworks.

Advancements in hypothetical metal-organic framework (hMOF) databases and construction tools have resulted in a rapidly expanding chemical design space for nanoporous materials. The bulk of these hypothetical structures are constructed using structural building units (SBUs) derived from experimental MOF structures, often collected from the CoRE-MOF database. Recent investigations into the state of these deposited experimental structures' chemical accuracy identified an array of common structural errors, including omitted protons, missing counterions, and disordered structures. These structural errors propagate into the SBUs mined from experimental MOFs, culminating in inaccurate hMOF structures possessing net charges or missing atoms which were not accounted for previously. This work demonstrates how manual investigation was applied to diagnose structural errors in SBUs obtained from several popular hMOF construction tools and databases. An analysis of the prevailing errors discovered during the examination process is provided along with representative cases to aid with error detection in future studies involving SBU extraction and hMOF construction. A novel repair protocol was established and employed to generate a library of SBUs that are hand-examined and labeled with enhanced detail (HEALED). This repaired library of SBUs contains 952 inorganic SBUs and 568 organic SBUs ideally suited for the generation of hypothetical frameworks that are chemically accurate and properly charge labeled. Additionally, case studies following the effects of SBU errors on electrostatic potential-fitted charges and GCMC-simulated gas adsorption predictions are presented to highlight the significance of using chemically accurate hMOF structures exclusively in all screening efforts going forward.

Combining Machine Learning and Molecular Simulations to Unlock Gas Separation Potentials of MOF Membranes and MOF/Polymer MMMs.

Due to the enormous increase in the number of metal-organic frameworks (MOFs), combining molecular simulations with machine learning (ML) would be a very useful approach for the accurate and rapid assessment of the separation performances of thousands of materials. In this work, we combined these two powerful approaches, molecular simulations and ML, to evaluate MOF membranes and MOF/polymer mixed matrix membranes (MMMs) for six different gas separations: He/H2, He/N2, He/CH4, H2/N2, H2/CH4, and N2/CH4. Single-component gas uptakes and diffusivities were computed by grand canonical Monte Carlo (GCMC) and molecular dynamics (MD) simulations, respectively, and these simulation results were used to assess gas permeabilities and selectivities of MOF membranes. Physical, chemical, and energetic features of MOFs were used as descriptors, and eight different ML models were developed to predict gas adsorption and diffusion properties of MOFs. Gas permeabilities and membrane selectivities of 5249 MOFs and 31,494 MOF/polymer MMMs were predicted using these ML models. To examine the transferability of the ML models, we also focused on computer-generated, hypothetical MOFs (hMOFs) and predicted the gas permeability and selectivity of 1000 hMOF/polymer MMMs. The ML models that we developed accurately predict the uptake and diffusion properties of He, H2, N2, and CH4 gases in MOFs and will significantly accelerate the assessment of separation performances of MOF membranes and MOF/polymer MMMs. These models will also be useful to direct the extensive experimental efforts and computationally demanding molecular simulations to the fabrication and analysis of membrane materials offering high performance for a target gas separation.

Molecular-fingerprint machine-learning-assisted design and prediction for high-performance MOFs for capture of NMHCs from air

The capture of trace amounts of non-methane hydrocarbons (NMHCs) from air due to the toxicity of volatile organic compounds is a significant challenge. A total of 31399 hydrophobic metal–organic frameworks (MOFs) were first screened from 137953 hypothetical MOFs using high-throughput computational screening (HTCS), and their performance indices (adsorption capacity and selectivity) for the adsorption of NMHCs (C 3 –C 6 ) were obtained by molecular simulations. The discovery of a “second peak” near twice the kinetic diameter of the corresponding NMHC provided more choices for excellent MOFs that adsorb NMHCs. Four machine learning (ML) classification and regression algorithms predicted the performance of MOFs, and the relative importance values of the six descriptors were determined. The combination of the Random Forests algorithm and Molecular ACCess Systems molecular fingerprint (MF) had an excellent predictive ability for MOFs. According to the performance, the fingerprint commonalities of the 100 top-performing MOFs were counted, and the excellent bits (EBs) that could promote the performance were defined. Finally, new substructures containing all of the EBs were designed for each NMHC to build a new MOF database. This work combined the HTCS, ML, and MF to provide a detailed insight into the design of efficient MOFs for adsorbing NMHCs.

Hydrogen storage metal-organic framework classification models based on crystal graph convolutional neural networks

Metal-organic frameworks (MOFs) have been considered as promising physical adsorbents for hydrogen storage due to their high porosity and structural tunability. We selected 7643 real MOFs from the computation-ready MOF 2019 database to screen high-performance materials for hydrogen storage based on the grand canonical Monte Carlo (GCMC) simulations. Based on the obtained data set, we proposed a deep learning classification model powered by the crystal graph convolutional neural networks (CGCNN) for the discovery of the optimal hydrogen storage MOFs. It is demonstrated that our classification model based on CGCNN algorithms exhibits a high prediction accuracy with the area under the ROC-plot curve (AUC) of 0.9208 for hydrogen storage performance of volumetric deliverable capacity (VDC). Compared with the classification models based on other machine learning algorithms such as random forest, our CGCNN model demonstrates the advantages of fast prediction with no feature extraction and little accuracy loss. When the trained CGCNN model was used to predict the classification of the unfamiliar samples (the randomly selected 1000 MOFs from the 137,953 hypothetical MOF database), we also obtain high accuracy with AUC > 0.81, indicating that this model exhibits reliable transferability for other types of MOFs. Meanwhile, we also elucidated the relationship between structure and performance of MOFs for hydrogen storage using the decision tree algorithms and quantitative structure–property analysis. Furthermore, the hydrogen adsorption performance and mechanism of top-performance MOFs were analyzed by adsorption isotherms, radial distribution functions, and mass center density distribution of equilibrium configurations. All those insights from atomic simulations and machine learnings can accelerate the discovery of new nanoporous materials not only for gas adsorption in MOFs but also for gas separation in other types of porous materials.

Integrated metal‐organic framework (MOF) and pressure/vacuum swing adsorption process design:MOFmatching

AbstractA two‐step integrated metal‐organic framework (MOF) and pressure/vacuum swing adsorption (P/VSA) process design has been recently established for gas separation. In the first step, selected MOF descriptors and process operating conditions are simultaneously optimized to maximize the process performance. Based on the obtained results, the second step (i.e., MOF matching) is addressed and exemplified by propene/propane separation in this work. Computational MOF synthesis and screening are explicitly carried out to find new advanced material candidates for enhancing the separation process efficiency. First, model‐based property‐performance relationships are developed for fast MOF screening. Then, MOF building blocks are extracted from 471 MOFs contained in the Computation‐Ready Experimental (CoRE) MOF database. With these building blocks, 45,472 hypothetical MOFs are created. After model‐based and molecular simulation‐based screening, six candidates are left and sent to P/VSA process optimization. Finally, three candidates are found to meet the predefined separation specifications and one candidate shows a better process performance than the best out of the 471 CoRE MOFs.

Minimal crystallographic descriptors of sorption properties in hypothetical MOFs and role in sequential learning optimization

We focus on gas sorption within metal-organic frameworks (MOFs) for energy applications and identify the minimal set of crystallographic descriptors underpinning the most important properties of MOFs for CO2 and H2O. A comprehensive comparison of several sequential learning algorithms for MOFs properties optimization is performed and the role played by those descriptors is clarified. In energy transformations, thermodynamic limits of important figures of merit crucially depend on equilibrium properties in a wide range of sorbate coverage values, which is often only partially accessible, hence possibly preventing the computation of desired objective functions. We propose a fast procedure for optimizing specific energy in a closed sorption energy storage system with only access to a single water Henry coefficient value and to the specific surface area. We are thus able to identify hypothetical candidate MOFs that are predicted to outperform state-of-the-art water-sorbent pairs for thermal energy storage applications.

Multi-Level Computational Screening of in Silico Designed MOFs for Efficient SO2 Capture.

SO2 presence in the atmosphere can cause significant harm to the human and environment through acid rain and/or smog formation. Combining the operational advantages of adsorption-based separation and diverse nature of metal–organic frameworks (MOFs), cost-effective separation processes for SO2 emissions can be developed. Herein, a large database of hypothetical MOFs composed of >300,000 materials is screened for SO2/CH4, SO2/CO2, and SO2/N2 separations using a multi-level computational approach. Based on a combination of separation performance metrics (adsorption selectivity, working capacity, and regenerability), the best materials and the most common functional groups in those most promising materials are identified for each separation. The top bare MOFs and their functionalized variants are determined to attain SO2/CH4 selectivities of 62.4–16899.7, SO2 working capacities of 0.3–20.1 mol/kg, and SO2 regenerabilities of 5.8–98.5%. Regarding SO2/CO2 separation, they possess SO2/CO2 selectivities of 13.3–367.2, SO2 working capacities of 0.1–17.7 mol/kg, and SO2 regenerabilities of 1.9–98.2%. For the SO2/N2 separation, their SO2/N2 selectivities, SO2 working capacities, and SO2 regenerabilities span the ranges of 137.9–67,338.9, 0.4–20.6 mol/kg, and 7.0–98.6%, respectively. Besides, using breakdowns of gas separation performances of MOFs into functional groups, separation performance limits of MOFs based on functional groups are identified where bare MOFs (MOFs with multiple functional groups) tend to show the smallest (largest) spreads.

Graph neural network predictions of metal organic framework CO[formula omitted] adsorption properties

The increasing CO2 level is a critical concern and suitable materials are needed to capture such gases from the environment. While experimental and conventional computational methods are useful in finding such materials, they are usually slow and there is a need to expedite such processes. We use Atomistic Line Graph Neural Network (ALIGNN) method to predict CO2 adsorption in metal organic frameworks (MOF), which are known for their high functional tunability. We train ALIGNN models for hypothetical MOF (hMOF) database with 137953 MOFs with grand canonical Monte Carlo (GCMC) based CO2 adsorption isotherms. We develop high accuracy and fast models for pre-screening applications. We apply the trained model on CoREMOF database and computationally rank them for experimental synthesis. In addition to the CO2 adsorption isotherm, we also train models for electronic bandgaps, surface area, void fraction, lowest cavity diameter, and pore limiting diameter, and illustrate the strength and limitation of such graph neural network models. For a few candidate MOFs we carry out GCMC calculations to evaluate the deep-learning (DL) predictions.

MOF-VR: A Virtual Reality Program for Performing and Visualizing Immersive Molecular Dynamics Simulations of Guest Molecules in Metal-Organic Frameworks.

Molecular dynamics simulations are useful to study diffusion of guest molecules in metal-organic frameworks. The interpretation of the generated three-dimensional trajectories is often difficult, because most visualization tools only allow two-dimensional projections. To facilitate interpretation, we present MOF-VR: a virtual reality program for performing interactive molecular dynamics simulations in metal-organic frameworks and visualizing atomic or molecular trajectories. MOF-VR consists of three subroutines: a construction routine to create hypothetical metal-organic frameworks by hand, a molecular dynamics suite, and a trajectory visualizer. To the best of our knowledge, MOF-VR is the first virtual reality program that allows hypothetical metal-organic frameworks to be constructed and tested in molecular dynamics simulations of guest molecules. We further show that MOF-VR is capable of performing state-of-the-art molecular dynamics simulations of guest molecules in rigid metal-organic frameworks in virtual reality and provides reliable simulation results.

Hypothetical yet effective: Computational identification of high-performing MOFs for CO2 capture

With the advances in computational resources and algorithms, computer simulations are being increasingly used to tackle the most challenging problems of the world. Among them, CO2 capture is a topic that needs imminent attention as the presence of high levels of CO2 in the air can lead to drastic shifts in global climate. Here, a recently developed hypothetical metal-organic framework (MOF) database comprised of anion-pillared (AP) MOFs is computationally screened for the separation of CO2/CO, CO2/H2, and CO2/N2 gas mixtures at room temperature. The best performing MOFs are identified using three performance metrics, adsorption selectivity, working capacity, and regenerability, in conjunction. In these top materials, the preferential adsorption sites are illustrated, which will be useful in guiding the experimental design of new MOFs with extraordinarily high CO2 selectivities. The favorable separation performances of AP MOFs suggest that efficient gas separations can be conducted using MOFs without open metal sites.

Accelerated Discovery of CH4 Uptake Capacity Metal–Organic Frameworks Using Bayesian Optimization

AbstractHigh‐throughput computational studies for discovery of metal–organic frameworks (MOFs) for separations and storage applications are often limited by the costs of computing thermodynamic quantities. Recent such studies at the time of writing may use ab initio results for a narrow selection of MOFs or empirical force‐field methods for larger selections. Here, a proof‐of‐concept study is conducted using Bayesian optimization on CH4 uptake capacity of hypothetical MOFs for an existing dataset (Wilmer et al., Nature Chem. 2012, 4, 83). It is shown that less than 0.1% of the database needs to be screened with the Bayesian optimization approach to recover the top candidate MOFs. This opens the possibility for efficient screening of MOF databases using accurate ab initio calculations for future adsorption studies on a minimal subset of MOFs. Furthermore, Bayesian optimization and the surrogate model presented here can offer interpretable material design insights and the framework will be applicable in the context of other target properties.

High-throughput computational screening of hypothetical metal-organic frameworks with open copper sites for CO2/H2 separation.

It is challenging to identify the optimal metal-organic framework (MOF) adsorbents for gas adsorption and membrane-based separation from the large-scale material databases. The high-throughput computational screening (HTCS) method was adopted to discover the optimal materials for CO2/H2 separation from thousands of MOFs. First, a hierarchical strategy was used to select 1092 MOFs from 13 512 MOFs, and their adsorption capacity towards the equimolar CO2/H2 mixture at 298 K and 10 bar was further calculated using the grand canonical Monte Carlo (GCMC) simulations. The results show that those MOFs withlvtbtopology and organic linker 1,2,4,5-tetrazine are conducive to exhibiting high performance CO2/H2adsorption separation among top-100 MOFs with high performance. The MOFs with pore limited diameter (PLD), largest cavity diameter (LCD), gravimetrical surface area (GSA), and void fraction in the range of 4-12 Å, 5-12 Å, 5500-6500 m2 g-1 and 0.80-0.85, respectively, have high adsorption capacity towards CO2. Second, the dynamic adsorption properties of the top-4 MOFs were simulated by the breakthrough curves of the binary (CO2/H2) and quinary (CO2/H2/CH4/CO/N2) mixtures in the fixed adsorption bed. MOF-4641 exhibits a high breakthrough time of 130 for the quinary mixture. Finally, the adsorption mechanism of CO2 in the top-4 MOFs was investigated by the radial distribution function (RDF), the mass center probability density distribution, etc. The atomic insights from HTCS and breakthrough curve predictions in this work will be helpful in developing novel porous materials and obtaining superior CO2 separation performance.

High-Throughput Computational Screening of Hypothetical Metal-Organic Frameworks with Open Copper Sites for Co2/H2 Separation

High-Throughput Computational Screening of Hypothetical Metal-Organic Frameworks with Open Copper Sites for Co2/H2 Separation

Prediction of O2/N2 Selectivity in Metal-Organic Frameworks via High-Throughput Computational Screening and Machine Learning.

Machine learning (ML), which is becoming an increasingly popular tool in various scientific fields, also shows the potential to aid in the screening of materials for diverse applications. In this study, the computation-ready experimental (CoRE) metal-organic framework (MOF) data set for which the O2 and N2 uptakes, self-diffusivities, and Henry's constants were calculated was used to fit the ML models. The obtained models were subsequently employed to predict such properties for a hypothetical MOF (hMOF) data set and to identify structures having a high O2/N2 selectivity at room temperature. The performance of the model on known entries indicated that it would serve as a useful tool for the prediction of MOF characteristics with r2 correlations between the true and predicted values typically falling between 0.7 and 0.8. The use of different descriptor groups (geometric, atom type, and chemical) was studied; the inclusion of all descriptor groups yielded the best overall results. Only a small number of entries surpassed the performance of those in the CoRE MOF set; however, the use of ML was able to present the structure-property relationship and to identity the top performing hMOFs for O2/N2 separation based on the adsorption and diffusion selectivity.

Diversifying Databases of Metal Organic Frameworks for High-Throughput Computational Screening.

By combining metal nodes and organic linkers, an infinite number of metal organic frameworks (MOFs) can be designed in silico. Therefore, when making new databases of such hypothetical MOFs, we need to ensure that they not only contribute toward the growth of the count of structures but also add different chemistries to the existing databases. In this study, we designed a database of ∼20,000 hypothetical MOFs, which are diverse in terms of their chemical design space—metal nodes, organic linkers, functional groups, and pore geometries. Using machine learning techniques, we visualized and quantified the diversity of these structures. We find that on adding the structures of our database, the overall diversity metrics of hypothetical databases improve, especially in terms of the chemistry of metal nodes. We then assessed the usefulness of diverse structures by evaluating their performance, using grand-canonical Monte Carlo simulations, in two important environmental applications—post-combustion carbon capture and hydrogen storage. We find that many of these structures perform better than widely used benchmark materials such as Zeolite-13X (for post-combustion carbon capture) and MOF-5 (for hydrogen storage). All the structures developed in this study, and their properties, are provided on the Materials Cloud to encourage further use of these materials for other applications.