Metal-organic framework (MOFs) are nanoporous materials that could be used to capture carbon dioxide from the exhaust gas of fossil fuel power plants to mitigate climate change. In this work, we design and train a message passing neural network (MPNN) to predict simulated CO$_2$ adsorption in MOFs. Towards providing insights into what substructures of the MOFs are important for the prediction, we introduce a soft attention mechanism into the readout function that quantifies the contributions of the node representations towards the graph representations. We investigate different mechanisms for sparse attention to ensure only the most relevant substructures are identified.
Symmetry-Informed Graph Neural Networks for Carbon Dioxide Isotherm and Adsorption Prediction in Aluminum-Substituted Zeolites
Eindhoven University of Technology（埃因霍温理工大学）；Eindhoven Artificial Intelligence Systems Institute（埃因霍温人工智能系统研究所）；Data and AI, Department of Mathematics and Computer Science, Eindhoven University of Technology（埃因霍温理工大学数学与计算机科学系数据与人工智能部）；Materials Simulation and Modelling（材料模拟与建模中心）
In recent years, the scientific community has given more and more attention to the issue of climate change and global warming, which is largely attributed to the massive quantity of carbon dioxide emissions. Thus, the demand for a carbon dioxide capture material is massive and continuously increasing. In this study, we perform first-principle calculations based on density functional theory to investigate the carbon dioxide capture ability of pristine and doped beryllonitrene. Our results show that carbon dioxide had an adsorption energy of -0.046 eV on pristine beryllonitrene, so it appears th…
We study numerically the adsorption of a mixture of CO$_2$ and CH$_4$ on a graphite substrate covered by graphene nanoribbons (NRs). The NRs are flat and parallel to the graphite surface, at a variable distance ranging from 6 Å to 14 Å. We show that the NRs-graphite substrate acts as an effective filter for CO$_2$. Our study is based on Molecular Dynamics (MD) simulations. Methane is considered a spherical molecule, and carbon dioxide is represented as a linear rigid body. Graphite is modeled as a continuous material, while the NRs are approached atomistically. We observe that when the NRs are…
The large number of possible structures of metal-organic frameworks (MOFs) and their limitless potential applications has motivated molecular modelers and researchers to develop methods and models to efficiently assess MOF performance. Some of the techniques include large-scale high-throughput molecular simulations and machine learning models. Despite those advances, the number of possible materials and the potential conditions that could be used still pose a formidable challenge for model development requiring large data sets. Therefore, there is a clear need for algorithms that can efficient…
Charge transfer between molecules and catalysts plays a critical role in determining the efficiency and yield of photo-chemical catalytic processes. In this paper, we study light-induced electron transfer between transition metal doped aluminium clusters and CO$_2$ molecules using first-principles time-dependent density-functional theory. Specifically, we carry out calculations for a range of dopants (Zr, Mn, Fe, Ru, Co, Ni and Cu) and find that the resulting systems fall into two categories: Cu- and Fe-doped clusters exhibit no ground state charge transfer, weak CO$_2$ adsorption and light-in…
Carbon dioxide, CO2, capture by room-temperature ionic liquids (RTILs) is a vivid research area featuring both accomplishments and frustrations. This work employs the PM7-MD method to simulate adsorption of CO2 by 1,3-dimethylimidazolium thiocyanate at 300 K. The obtained result evidences that the thiocyanate anion plays a key role in gas capture, whereas the impact of the 1,3-dimethylimidazolium cation is mediocre. Decomposition of the computed wave function on the individual molecular orbitals confirms that CO2-SCN binding extends beyond just expected electrostatic interactions in the ion-mo…
Ubiquitous graphene is a stricly 2D material representing an ideal adsorbing platform due to its large specific surface area as well as its mechanical strength and resistance to both thermal and chemical stresses. However, graphene as a bulk material has the tendency to form irreversible agglomerates leading to 3D graphitic structures with a significant decrease of the area available for adsorption and no room for gas intercalation. In this paper a novel nanoporous graphite formed by graphtriyne sheets is introduced: its 3D structure is theoretically assessed by means of electronic structure a…
In this study, we demonstrate that defected h--BN (1B-3N and 1N-3B defects) can be used as a suitable membrane for hydrogen purification and helium separation using density functional theory (DFT) calculations and molecular dynamics simulations (MD). At 300 K, DFT calculations show that the selectivity of $H_2/CO_2$, $H_2/N_2$, $H_2/CO$, and $H_2/CH_4$ are $1 \times 10^{26}$, $6 \times 10^{49}$, $2 \times 10^{40}$, and $1 \times 10^{94}$ respectively for 1B-3N, while they are $9 \times 10^{13}$, $1 \times 10^{21}$, $1 \times 10^{19}$, and $1 \times 10^{42}$ for 1N-3B. Although selectivity of 1…
Lawrence Berkeley National Laboratory（劳伦斯伯克利国家实验室）；Toyota Research Institute（丰田研究院）；IMDEA Materials Institute（IMDEA材料研究所）
Machine learning has emerged as a powerful approach in materials discovery. Its major challenge is selecting features that create interpretable representations of materials, useful across multiple prediction tasks. We introduce an end-to-end machine learning model that automatically generates descriptors that capture a complex representation of a material's structure and chemistry. This approach builds on computational topology techniques (namely, persistent homology) and word embeddings from natural language processing. It automatically encapsulates geometric and chemical information directly…
Hydrogen production as a clean, sustainable replacement for fossil fuels is gathering pace. Doubling the capacity of Paris-CDG airport has been halted, even with the upcoming Olympic Games, until hydrogen powered planes can be used. It is thus timely to work on catalytic selective hydrogen production and optimise catalyst structure. Over 90 % of all chemical manufacture uses a solid catalyst. This work describes adsorption of carbon-monoxide (CO) on platinum thin films, supported by cheap Al(111). CO reacts with water to produce hydrogen (water-gas shift). Quantum Monte Carlo methods are the o…
