Coordination framework compounds and artificial intelligence-assisted molecular screening for coordination framework compounds

Abstract

This specification describes methods and systems for proposing coordination framework compounds, generally including crystalline porous materials, crystalline open frameworks, network chemical compounds, metal-organic framework (MOF) compounds, covalent organic framework (COF) compounds, zeolite imidazolate framework (ZIF) compounds, and combinations thereof. Coordination framework compounds produced by the same systems, and adsorbent systems containing coordination framework compounds, are also described herein. The methods and systems described herein combine machine learning and chemistry to propose chemically valid and improved-performing coordination framework compounds that meet different objectives of materials discovery.

Description

[Technical Field]

[0001] The field of this disclosure generally relates to methods and systems for proposing coordination framework compounds such as crystalline porous materials, crystalline open frameworks, network chemical compounds, metal-organic framework (MOF) compounds, covalent organic framework (COF) compounds, zeolite imidazolate framework (ZIF) compounds, and combinations thereof. The field of this disclosure also relates to coordination framework compounds produced thereby.

[0002] Coordination framework compounds such as MOF, ZIF, and COF compounds are useful for a wide variety of purposes. For example, they can be particularly useful in carbon capture adsorbent systems or methane capture adsorbent systems, such as post-combustion carbon dioxide (CO2) capture and direct atmospheric capture. Another example is their usefulness in atmospheric water extraction (AWE), where water is generated remotely and on demand. However, carbon capture and AWE processes each require constant and iterative improvement of the materials.

[0003] Existing iterative trial-and-error methods for discovering chemically-based coordination framework compounds require considerable investment and effort to achieve innovative and effective materials. While pure machine learning (ML) methods can assist in such discoveries, they suffer from a lack of deep chemical insight. For example, the model developed by Xie et al. ("Crystal Diffusion Variational Autoencoder for Periodic Material Generation." arXiv preprint arXiv:2110.06197(2021)) advances machine learning capabilities for application to small unit cell compounds (e.g., small crystals), but lacks the necessary chemical knowledge and guidance. This model is also only applicable to small unit cell compounds and has not been demonstrated for larger structures such as MOF compounds.

[0004] Due to the shortcomings of existing methods, it would be desirable to develop a process that combines the strengths of machine learning-based and chemical knowledge-based methods to more effectively propose and discover coordination framework compounds. [Prior art documents] [Patent Documents]

[0005] [Patent Document 1] U.S. Patent Application Publication No. 2014 / 0234624 [Overview of the Initiative]

[0006] In one embodiment, a method for proposing coordination framework compounds is provided herein. In an exemplary embodiment, the method is performed using a coordination framework compound proposal (CFCP) computing device comprising a processor coupled to a memory device. The exemplary method includes generating an initial set of coordination framework compounds using a machine learning model on the CFCP computing device; subjecting at least one coordination framework compound included in the initial set of coordination framework compounds to a review of at least one chemical property; and using the CFCP computing device, generating a preliminary set of coordination framework compounds using the machine learning model based on the initial set of coordination framework compounds and the review of at least one chemical property of at least one coordination framework compound included in the initial set of coordination framework compounds. The method also includes proposing coordination framework compounds using the CFCP computing device.

[0007] In another aspect, there is provided a coordination framework compound proposal (CFCP) computing device, the device comprising a memory and a processor communicatively connected to the memory, the processor being programmed to generate an initial set of coordination framework compounds using a machine learning model and to subject at least one coordination framework compound included in the initial set of coordination framework compounds to a review of at least one chemical property. The processor is further programmed to generate a preliminary set of coordination framework compounds using the machine learning model and to propose a coordination framework compound based on a review of the initial set of coordination framework compounds and at least one chemical property of at least one coordination framework compound included in the initial set of coordination framework compounds.

[0008] In yet another aspect, a non-transitory computer-readable storage medium having computer-executable instructions embodied thereon is provided herein, which when executed by a coordination framework compound proposal (CFCP) computing device comprising at least one processor communicatively coupled to a memory, causes the computer-readable instructions to cause the CFCP computing device to generate an initial set of coordination framework compounds using a machine learning model and to subject at least one coordination framework compound included in the initial set of coordination framework compounds to a review of at least one chemical property. The computer-readable instructions also cause the CFCP computing device to generate a preliminary set of coordination framework compounds using the machine learning model and to propose a coordination framework compound based on a review of the initial set of coordination framework compounds and at least one chemical property of at least one coordination framework compound included in the initial set of coordination framework compounds.

[0009] In yet another aspect, a coordination framework compound is provided herein that includes a plurality of secondary building units (SBUs), a plurality of linkers that form connections between the plurality of SBUs, and a plurality of pores formed in the gaps between the connections. The coordination framework compound includes at least two chemically different linkers, at least two geometrically different pores, and at least one of at least two connections between two of the plurality of SBUs.

[0010] These and other features, aspects, and advantages of the present disclosure will be better understood by reading the following detailed description of the invention with reference to the accompanying drawings, in which like numerals represent like parts throughout the drawings.

Brief Description of the Drawings

[0011] [Figure 1A] An exemplary method flowchart according to the present disclosure. [Figure 1B] An exemplary method flowchart according to the present disclosure. [Figure 2] Another exemplary method flowchart according to the present disclosure. [Figure 3] A further exemplary method flowchart according to the present disclosure. [Figure 4] Yet another exemplary method flowchart according to the present disclosure. [Figure 5] An exemplary block diagram of a computer system according to the present disclosure. [Figure 6] A diagram showing an exemplary configuration of a server system, such as the computer system of FIG. 1A, according to the present disclosure. [Figure 7] A diagram showing an exemplary configuration of a client system shown in FIG. 1A according to the present disclosure. [Figure 8] A diagram showing an exemplary method according to the present disclosure. [Figure 9] An exemplary coordination framework compound obtained from an ML model according to the present disclosure. [Figure 10]This disclosure includes exemplary molecules used in substitutional linkers in coordination framework compounds. [Figure 11] This figure shows the performance of the simulated coordination framework compound and the coordination framework compound of the comparative example described herein. [Figure 12A] This is a diagram of the first coordination framework compound including a linear SBU topology according to the present disclosure. [Figure 12B] This is a second diagram of a coordination framework compound including a linear SBU topology according to the present disclosure. [Figure 13A] This is a diagram of the first coordination framework compound including a curved SBU topology according to the present disclosure. [Figure 13B] This is a second diagram of a coordination framework compound including a curved SBU topology according to the present disclosure. [Figure 14A] This is the first figure of a coordination framework compound containing an SBU topology that is neither linear nor curved, as disclosed herein. [Figure 14B] This is a second figure of a coordination framework compound including an SBU topology that is neither linear nor curved, as disclosed herein. [Figure 15A] This is a first diagram of a coordination framework compound including a crosslinked arm linker structure according to the present disclosure. [Figure 15B] Figure 2 shows a coordination framework compound including a crosslinked arm linker structure according to this disclosure. [Figure 16] This figure shows a coordination framework compound including altered pore shapes resulting from a mixture of different linkers as described in this disclosure. [Modes for carrying out the invention]

[0012] Unless otherwise indicated, the drawings provided herein are intended to illustrate features of embodiments of the disclosure. These features are considered applicable to a wide variety of systems, including one or more embodiments of the disclosure. Accordingly, the drawings are not intended to include all prior art features known to those skilled in the art that may be required to carry out the embodiments disclosed herein.

[0013] The embodiments described herein overcome at least some of the shortcomings of known methods for proposing coordination framework compounds. These embodiments combine machine learning and chemistry to propose chemically valid and improved-performing coordination framework compounds that meet different objectives of materials discovery. Machine learning leads to the conception and generation of material structures, while chemical guidance presents modifications and improvements to the materials generated by machine learning.

[0014] As used herein, a coordination framework compound comprises points for structural connections, nodes giving rise to secondary building units (SBUs), one or more linkers, and a topology defining the mode of connection between the SBUs and the linkers. Varying any of these embodiments, individually or in combination, yields novel and distinct coordination framework compounds. The methods described herein are applicable to the discovery of any crystal structure.

[0015] Examples of coordination framework compounds include metal-organic framework (MOF) compounds, covalent organic framework (COF) compounds, zeolite imidazolate framework (ZIF) compounds, crystalline porous materials, crystalline open frameworks, network chemical compounds, and combinations thereof. MOF compounds have strong bonds between metal atoms and charged ligands and may contain one or more metal atoms in the SBU and linker composition, but are not limited to these. COF compounds have strong covalent bonds between light elements (e.g., B, C, N, O, Si, P) and may contain one or more organic structures as the SBU, but are not limited to these. ZIF is a subclass of MOF compounds having a structure in which tetrahedral-coordinated transition metal ions (e.g., Fe, Co, Cu, Zn) are linked by imidazolate linkers that are topologically isomorphic to zeolites.

[0016] In general, coordination framework compounds are not limited to having only one type of SBU or one type of linker. Rather, they can have more complex structures consisting of multiple types of nodes and linkers. In other words, the components of a coordination framework compound can be combined in various ways. For example, a coordination framework compound may contain two linkers and one metal node.

[0017] The exemplary embodiments described herein include a method for proposing coordination framework compounds, which is performed using a coordination framework compound proposal (CFCP) computing device comprising a processor coupled to a memory device. The method includes generating an initial set of coordination framework compounds using a machine learning model of the CFCP computing device; subjecting at least one chemical property of at least one coordination framework compound included in the initial set of coordination framework compounds to review; generating a preliminary set of coordination framework compounds using the machine learning model based on the initial set of coordination framework compounds and the review of at least one chemical property of at least one coordination framework compound included in the initial set of coordination framework compounds; and proposing coordination framework compounds using the CFCP computing device.

[0018] The method may further include one or more additional process steps which are preferable to facilitate the method described herein. In some embodiments, the method also includes subjecting at least one coordination framework compound included in a preliminary set of coordination framework compounds to review at least one chemical property. In some embodiments, the method further includes validating at least one coordination framework compound included in a preliminary set of coordination framework compounds. In some embodiments, the method further includes performing at least one iteration of the sequence, which includes generating a further preliminary set of coordination framework compounds using a machine learning model based on at least one set of generated coordination framework compounds and a review of at least one chemical property of at least one coordination framework compound included in the preliminary set of at least one set of generated coordination framework compounds, optionally subjecting at least one coordination framework compound included in the further preliminary set of coordination framework compounds to review at least one chemical property, and optionally validating at least one coordination framework compound included in the further preliminary set of coordination framework compounds.

[0019] The illustrated embodiments also include a Coordination Framework Compound Proposal (CFCP) computing device comprising a memory and a processor communicatively connected to the memory, the processor being programmed to perform embodiments of the method herein.

[0020] The processor may be further programmed to perform one or more additional steps which are preferable to facilitate the method described herein. In some embodiments, the processor is further programmed to subject at least one coordination framework compound included in a preliminary set of coordination framework compounds to review at least one chemical property. In some embodiments, the processor is further programmed to validate at least one coordination framework compound included in a preliminary set of coordination framework compounds. In some embodiments, the processor is further programmed to perform at least one iteration of the sequence, which includes generating a further preliminary set of coordination framework compounds using a machine learning model based on at least one set of generated coordination framework compounds and a review of at least one chemical property of at least one coordination framework compound included in the preliminary set of at least one set of generated coordination framework compounds, optionally subjecting at least one coordination framework compound included in the further preliminary set of coordination framework compounds to review at least one chemical property, and optionally validating at least one coordination framework compound included in the further preliminary set of coordination framework compounds.

[0021] An exemplary embodiment also includes a non-temporary computer-readable storage medium in which computer-executable instructions are embodied, which is realized when the instructions are executed by a Coordination Framework Compound Proposal (CFCP) computing device having at least one processor communicating with memory. The computer-readable instructions cause the Coordination Framework Compound Proposal computing device to execute embodiments of the method described herein.

[0022] Computer-readable instructions may include one or more additional steps which are preferable to facilitate the methods described herein. In some embodiments, the computer-readable instructions further cause a CFCP computing device to subject at least one coordination framework compound included in a preliminary set of coordination framework compounds to review at least one chemical property. In some embodiments, the computer-readable instructions further cause a CFCP computing device to verify at least one coordination framework compound included in a preliminary set of coordination framework compounds. In some embodiments, the computer-readable instructions further cause a CFCP computing device to perform at least one iteration of the sequence, which includes generating a further preliminary set of coordination framework compounds using a machine learning model based on at least one set of generated coordination framework compounds and a review of at least one chemical property of at least one coordination framework compound included in the further preliminary set of coordination framework compounds, optionally subjecting at least one coordination framework compound included in the further preliminary set of coordination framework compounds to review at least one chemical property, and optionally verifying at least one coordination framework compound included in the further preliminary set of coordination framework compounds.

[0023] In general, any method described herein may be carried out using a Coordination Framework Compound Proposal (CFCP) computing device. As used herein, a Coordination Framework Compound Proposal computing device includes any suitable computing device known in the Art that facilitates the method described herein. Suitable computing devices may include, but are not limited to, a computer, desktop computer, handheld computer, or smartphone.

[0024] In general, a coordination framework compound may be any suitable coordination framework compound realized or proposed by the exemplary methods described herein. In some embodiments, a coordination framework compound is a coordination framework compound included in an initial set of coordination framework compounds, a coordination framework compound included in a preliminary set of coordination framework compounds, or a coordination framework compound included in a further preliminary set of coordination framework compounds. In other words, a coordination framework compound may be a hypothetical coordination framework compound, a coordination framework compound realized by chemical insight into a hypothetical coordination framework compound, or a coordination framework compound realized by iterative analysis of a coordination framework compound.

[0025] In some embodiments, the machine learning model is trained on existing coordination framework compounds, novel coordination framework compounds, and combinations thereof.

[0026] In general, the machine learning model may use any preferred technique that facilitates the exemplary methods described herein. In some embodiments, the machine learning model uses at least one of the techniques of latent space, inverse search, variational autoencoder (VAE), crystal diffusion variational autoencoder (CDVAE), inverse search of VAE latent space, graph neural network (GNN), neural network, optimization, and combinations thereof. In some embodiments, the machine learning model is programmed to learn a latent space configured to reconstruct the crystal structure of a coordination framework compound and accurately predict the relevant target properties via artificial intelligence.

[0027] In some embodiments, a machine learning model is programmed to learn via a VAE (e.g., CDVAE) that maps a coordination framework compound structure (e.g., the crystal structure of an MOF) to a single point in a latent space, and then the latent space is reconstructed from the latent space to the original coordination framework compound structure. In this case, any point in the latent space corresponds to its correct crystal structure, and exploration in the latent space can be performed to propose new coordination framework compound structures. Furthermore, since each coordination framework compound has a specific target chemical property, such as a water adsorption isotherm, this latent space can also be mapped to accurately predict the target chemical properties of all coordination framework compounds. Thus, each point in the latent space corresponds to its exact coordination framework compound and its target chemical property. Back-search and target property optimization techniques can be performed to find new coordination framework compounds with more desirable target chemical properties. This VAE method can be combined with GNN techniques that map crystal structures to machine-readable graphs that can be read and trained by a machine learning model.

[0028] In some embodiments, at least one chemical property is selected from one or more of the following: adsorbate uptake capacity, adsorbate uptake kinetics, adsorbate weight productivity, adsorbate volume productivity, adsorbate isotherms and isobars, pore size, pore volume, heat of adsorption, heat of adsorption of equal amounts, chemical stability, thermal stability, mechanical stability, synthesizability, zeta potential, surface energy, hydrophobicity, hydrophilicity, chemical performance, chemical modification for improvement, and combinations thereof.

[0029] In some embodiments, verifying at least one coordination framework compound includes experimentally verifying the structure and / or properties of at least one coordination framework compound.

[0030] In some embodiments, experimental verification of the structure and / or properties of at least one coordination framework compound includes experimental verification using techniques selected from one or more of the following: powder X-ray diffraction (PXRD), single-crystal X-ray diffraction, solid-state nuclear magnetic resonance (SS-NMR), digestion NMR, and combinations thereof.

[0031] In some embodiments, experimentally verifying the structure and / or properties of at least one coordination framework compound involves relaxing the structure (e.g., electronically relaxing and / or optimizing) using first-principles calculations to ensure that the coordination framework compound structure is stable. The pore volume and pore diameter of the coordination framework compound are then calculated. Water stability and hydrophilicity are analyzed to verify whether the compound has the potential to absorb more water molecules. Subsequently, the water adsorption capacity of the compound is estimated by calculating a computational water adsorption isotherm for the proposed coordination framework compound using Gibbs Ensemble Monte Carlo (GEMC). The proposed coordination framework compound is then verified by synthetic experiments and experimentally measured chemical properties such as water adsorption isotherms.

[0032] In some embodiments, a machine learning model generates a preliminary coordination framework compound based on at least one of the following inputs: the crystal structure of an existing coordination framework compound, target properties, target chemical properties, sorption isotherm, water sorption isotherm, pore volume, pore diameter, water stability, hydrophilicity, and combinations thereof.

[0033] In some embodiments, the review of at least one chemical property is performed by a machine, a human, or a combination thereof. In some embodiments, the review of at least one chemical property is performed by a human. In some embodiments, the review of at least one chemical property is performed using a CFCP computing device.

[0034] In general, there are several challenges to applying ML to coordination framework compounds. These challenges include the reversibility of the representation of periodic crystal structures as ML inputs, the invariance of crystal structures, and the large unit cell and atomic number.

[0035] An exemplary periodic structure can also be represented as M=(A,X,L), where N is the number of atoms. A∈A N , atomic species, X∈R NX3 , atomic position, L∈R 3X3 , periodic lattice, c∈R |A| This is the composition. Schematically, the periodic structure can be represented with atoms as nodes and bonds as edges.

[0036] Regarding invariance with respect to materials, permutation invariance includes swapping the indices of any pair of atoms. Translational invariance includes translating X by any vector. Rotational invariance includes rotating X and L together by any rotation matrix. Periodic invariance includes an infinite number of ways of unit cells having different shapes and sizes.

[0037] ML training can be achieved as follows: First, M is encoded to z by a periodic GNN encoder. Second, a characteristic predictor predicts c, L, and N of M and predicts a target characteristic such as water adsorption isotherm or pore volume (PV) from z. Third, a periodic GNN decoder, given z,

number

number

[0038] Material optimization can be achieved in the following ways: First, start with an existing structure M (e.g., MOF-303) and encode M into z. Second, optimize the PV for z to find z' which gives a desired target property such as a desirable water adsorption isotherm or larger PV. Third, use z' to predict c, L, and N. Fourth, initialize the initial M0 randomly. Fifth, decode M0 and update it to an optimized material M' which gives a desired target property such as a desirable water adsorption isotherm or larger PV.

[0039] In some embodiments, variational autoencoder-based ML models, such as Crystal Diffusion Variational Autoencoders (CDVAEs), are used to generate novel, purpose-directed coordination framework compounds. In some preferred embodiments, CDVAEs are used to generate novel, purpose-directed MOFs after being trained only on existing coordination framework compounds.

[0040] In some embodiments, a codebase is used to degrade, modify, and / or reconstruct ML-generating coordination framework compounds. Generally, the codebase is compatible with any coordination framework compound. In some embodiments, the codebase is selected from the group consisting of ToBaCCo-based codebases, MOFid-based codebases, molfunc-based codebases, and combinations thereof. In some embodiments, a ToBaCCo-based codebase is used to perform degrade-modify-reconstruct ML-generating coordination framework compounds having rod-shaped secondary construct units (SBUs).

[0041] In many embodiments, the proposed coordination framework compound can be used according to any preferred purpose known in the art. In some embodiments, the proposed coordination framework compound is used in adsorbent systems. In some embodiments, the proposed coordination framework compound is used in carbon capture adsorbent systems. In some embodiments, the proposed coordination framework compound is used in moisture adsorbent systems. In some embodiments, the proposed coordination framework compound is used for gas capture. In some embodiments, the proposed coordination framework compound is used for post-combustion capture of CO2 and / or direct air capture of CO2. In some embodiments, the proposed coordination framework compound is used for moisture extraction from the atmosphere.

[0042] In many embodiments, the proposed coordination framework compound comprises a plurality of secondary building units (SBUs), a plurality of linkers forming links between the SBUs, and a plurality of pores formed in the gaps between the links. The coordination framework compound comprises at least two chemically distinct linkers, at least two geometrically distinct pores, and at least one of at least two links between two of the SBUs.

[0043] In some embodiments, the coordination framework compound includes at least three linkages between two of the multiple SBUs. In some embodiments, the coordination framework compound includes at least four linkages between two of the multiple SBUs. In some embodiments, the coordination framework compound includes at least five linkages between two of the multiple SBUs. In some embodiments, the coordination framework compound includes at least six linkages between two of the multiple SBUs.

[0044] In some embodiments, the coordination framework compound is selected from the group consisting of metal-organic framework (MOF) compounds, covalent organic framework (COF) compounds, zeolite imidazolate framework (ZIF) compounds, crystalline porous materials, crystalline open frameworks, network chemical compounds, and combinations thereof.

[0045] In some embodiments, at least one of the multiple SBUs is Metal atoms, Al or Mg, B, C, N, O, Si, or P, Transition metal atoms, Fe, Co, Cu, or Zn, Includes nodes that contain combinations of those.

[0046] In some embodiments, at least one of the multiple SBUs includes a coordination structure selected from the group consisting of polyhedra, tetrahedra, octahedrons, cubes, dodecahedrons, and combinations thereof.

[0047] In some embodiments, the coordination framework compound is planar symmetric. In these embodiments, the planar symmetric coordination framework compound may be of the MIL-53 type.

[0048] In some embodiments, the coordination framework compound is not planar symmetric. In these embodiments, the non-planar symmetric coordination framework compound may be non-MIL-53 type.

[0049] In some embodiments, the at least two chemically distinct linkers include at least two linkers of different lengths.

[0050] In some embodiments, at least two geometrically distinct pores differ by geometric properties selected from the group consisting of size, shape, and combinations thereof. In some embodiments, the geometric properties of the pores can be modified by changing one or more of the following: the atoms of the nodes, the coordination structure of at least one SBU, the symmetry of the coordination framework compound, and the length of at least one linker. These various modifications and their combinations result in SBUs of different shapes, different connections between SBUs, and different pore shapes.

[0051] In some embodiments, the linkers include a chain of repeating linkers. In these embodiments, the repeating linkers may include a single repeating linker or a combination of different linkers in a random or patterned configuration (e.g., a repeating block of linkers). The chain of repeating linkers may include a small number of repeating linkers (e.g., less than 10) or a large number of repeating linkers (e.g., more than 1000).

[0052] In some embodiments, the linkers include a linker comprising at least one substituted or unsubstituted aryl ring, at least one substituted or unsubstituted condensed aryl ring, at least two substituted or unsubstituted bonded aryl rings, or a combination thereof.

[0053] In some embodiments, multiple linkers are provided. Linker of formula IA: [ka] Linker of formula IA or formula IB: [ka] (Formula IB) (In the formula, n1, m1, n2, and m2 are each individually selected from the group consisting of integers of 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 100 or less, integers of 1000 or less, integers of 10,000 or less, integers of 100,000 or less, and integers of 1,000,000 or less, R1, R2, R3, R4, R8, R9, R 10 and R 11 are each individually selected from the group consisting of H, NH2, OH, and SH, R5 and R6 are each individually selected from the group consisting of at least one substituent selected from the group consisting of a direct bond, R 12 NHR 13 、R 12 OR 13 、R 12 SR 13 、NH2, OH, and SH, C1-C6 alkyl optionally substituted with at least one substituent selected from the group consisting of NH2, OH, and SH, C1-C6 alkylene optionally substituted with at least one substituent selected from the group consisting of NH2, OH, and SH, and combinations thereof, R7 is selected from the group consisting of a direct bond, ring fusion, NH, O, S, and C1-C6 alkyl, R 12 and R 13 are each individually selected from the group consisting of a direct bond, NH, O, S, and C1-C6 alkyl, A1, A2, A3, A4, A5, A6, A7, and A8 are each individually selected from the group consisting of C, N, O, and S) Linker of formula IIA:

Chemical formula

Chemical formula

[0054] In some embodiments, multiple linkers are provided. [ka] [ka] [ka] [ka] , and This includes a linker selected from a group consisting of combinations of those linkers.

[0055] In some embodiments, multiple linkers are provided. [ka] , and This includes a linker selected from a group consisting of combinations of those linkers.

[0056] In some embodiments, the adsorbent system includes a coordination framework compound.

[0057] Referring here to the drawings, Figure 1A is an exemplary method flowchart 110. In this exemplary embodiment, the method flowchart 110 illustrates exemplary steps of an embodiment of the method described herein and is not intended to limit the method embodiments. In the exemplary embodiment, an initial set of coordination framework compounds is generated using a machine learning model on a CFCP computing device 112. At least one coordination framework compound included in the initial set of coordination framework compounds is subjected to a review of at least one chemical property 114. Based on the initial set of coordination framework compounds and the review of at least one chemical property of at least one coordination framework compound included in the initial set of coordination framework compounds, a preliminary set of coordination framework compounds is generated using a machine learning model on the CFCP computing device 116. The coordination framework compounds are proposed using the CFCP computing device 118.

[0058] The method flowchart 110 in Figure 1A may be recursive and / or iterative, and any exemplary step may be performed multiple times and / or used to inform another step. One such exemplary embodiment is the hybrid ML-chemical coordination framework compound presentation workflow shown in Figure 1B. In this embodiment of the active learning workflow, the method flowchart 120 illustrates exemplary steps of the method embodiment herein and is not intended to limit the method embodiments. In the exemplary embodiment, the machine learning model 122 outputs a proposal 124 of coordination framework compounds. At least one of the coordination framework compounds of the proposal is subjected to chemical calculations and / or experiments 126. The results of the chemical calculations and / or experiments may result in a presentation 134 of modifications to the proposed coordination framework compound. This presentation 134 of modifications can be directed to the machine learning model 122 or output as the proposal 124 of coordination framework compounds. The experimental feasibility and calculated and / or measured chemical properties of the coordination framework compound are determined 128. Proposed compounds with desirable properties can be marked as targets for laboratory-scale synthesis.136 Other proposed compounds can be compared with experimental coordination framework compounds and hypothetical coordination framework compounds.130 The comparison may further include structural and isotherm inputs.132 The learning of this method can be reapplied to a machine learning model until satisfactory results are achieved.

[0059] Figure 2 is an exemplary flowchart 210 for use when proposing novel coordination framework compounds based on known coordination framework compounds. Object 212 includes starting with a known coordination framework compound. Object 214 includes changing the SBU, linker, and / or topology of the known coordination framework compound. Object 216 includes proposing a new coordination framework compound structure based on the changes in the known coordination framework compound. This scheme is more targeted and allows for the immediate presentation of coordination framework compounds. However, this requires considerable chemical insight.

[0060] Figure 3 is an exemplary objective flowchart 310. Object 312 contains inputs for coordination framework compounds. Object 314 contains an ML model, such as a VAE, which is updated by the inputs. Object 316 contains predictions of coordination framework compound properties, such as water adsorption isotherms. Object 318 contains a reverse search latent space for presenting and validating new coordination framework compounds. These new coordination framework compounds can be used as inputs for object 312 in subsequent iterations of the exemplary objective flowchart 310. The exemplary objective flowchart 310 enables the learning of construction information for coordination framework compounds based on known coordination framework compounds. It contains a good latent space representing the coordination framework compound space. This requires a large number of coordination framework compound structures with chemical properties such as pore volume, H-bond site density, and water adsorption isotherms. In some embodiments, the model may include a large number of coordination framework compound structures having DFT / GCMC or DFT / GEMC water adsorption isotherms calculated using techniques such as density functional theory (DFT) and / or grand canonical Monte Carlo (GCMC) calculations and / or Gibbs ensemble Monte Carlo (GEMC) calculations. This allows for the exploration of coordination framework compounds that do not have rod-shaped metal nodes.

[0061] Figure 4 is an exemplary ML model flowchart 410. Object 412 contains the input of a coordination framework compound. Object 414 contains a latent space, which is lower dimensional compared to the input, but allows for the exploration of desired coordination framework compound properties, as well as mapping to reconstruction and new coordination framework compounds. Object 416 contains a first output containing the reconstructed coordination framework compound. Object 418 contains a second output containing coordination framework compound properties such as water adsorption isotherms.

[0062] Figure 5 is a block diagram of an exemplary embodiment of a computer system 500 used in proposing a coordination framework compound, including a computing device 502, according to an exemplary embodiment of the present disclosure. The computing device 502 may also be referred herein to as a coordination framework compound proposal (CFCP) computing device. In the exemplary embodiment, the system 500 is used to propose a coordination framework compound as described herein. The computer system 500 can be used to carry out one or more of the methods described herein.

[0063] More specifically, in exemplary embodiments, system 500 includes a computing device 502 and a plurality of client subsystems, also called client systems 504, connected to the computing device 502. In one embodiment, client system 504 is a computer including a web browser, and computing device 502 is accessible from client system 504 using the Internet and / or network 506. Client system 504 is interconnected to the Internet via many interfaces, including network 506 such as a local area network (LAN) or wide area network (WAN), dial-in connection, cable modem, dedicated high-speed integrated digital communication network (ISDN) line, and RDT network. Client system 504 may include external systems used for data storage. Computing device 502 is also in a state of communicating with one or more data sources 514 using network 506. Furthermore, client system 504 can communicate further with data sources 514 using network 506. Furthermore, in some embodiments, as described herein, one or more client systems 504 can function as data sources 514. The client system 504 may be any device that can interconnect to the Internet, including a web-based telephone, PDA, or other web-based connectable device.

[0064] The database server 508 is connected to a database 512 that contains information on various matters, as will be described in more detail below. In one embodiment, the centralized database 512 is stored on device 502 and can be accessed by a potential user in any one of the client systems 504 by logging on to computing device 502 via one of the client systems 504. In an alternative embodiment, the database 512 may be stored remotely from device 502 and be decentralized. The database 512 may be a database configured to store information used by computing device 502, including, for example, transaction records, as described herein.

[0065] Database 512 may include a single database having separate sections or partitions, or it may include multiple databases, each separated from the others. Database 512 can store data received from data source 514 and data generated by computing device 502. For example, database 512 may store coordination framework compound data, as described in detail herein.

[0066] In exemplary embodiments, the client system 504 may be associated with any party that can use the system 500 as described herein. In exemplary embodiments, at least one of the client systems 504 includes a user interface 510. For example, the user interface 510 may include a graphical user interface having interactive capabilities, in which coordination framework compound data and suggestions transmitted from the computing device 502 to the client system 504 may be displayed in a graphical format. A user of the client system 504 can interact with the user interface 510 to view, explore, or otherwise interact with the displayed information.

[0067] In an exemplary embodiment, a computing device 502 receives data from multiple data sources 514, aggregates and analyzes the received data (e.g., using machine learning) to propose a coordination framework compound, as described in detail herein.

[0068] Figure 6 illustrates an exemplary configuration of a server system 602, such as a computing device, according to one exemplary embodiment of the present disclosure. The server system 602 can be used to carry out one or more of the methods described herein. The server system 602 may also include, but is not limited to, a database server (not shown). In the exemplary embodiment, the server system 602 proposes a coordination framework compound as described herein.

[0069] The server system 602 includes a processor 606 for executing instructions. Instructions can be stored, for example, in a memory area 610. The processor 606 may include one or more processing units for executing instructions (for example, in a multi-core configuration). Instructions can be executed within various different operating systems on the server system 602, such as UNIX®, LINUX, and Microsoft Windows®. It should also be understood that various instructions can be executed during initialization at the start of the computer-based method. While some operations may be required to execute one or more processes described herein, other operations may be more general and / or specific to a particular programming language (e.g., C, C#, C++, Java, or other preferred programming languages).

[0070] The processor 606 is operably connected to a communication interface 604 so that the server system 602 can communicate with a user system or a remote device such as another server system 602. For example, the communication interface 604 may receive requests from a client system via the internet (not shown).

[0071] The processor 606 can also be operably coupled to a storage device 612. The storage device 612 is hardware operated and controlled by any computer suitable for storing and / or retrieving data. In some embodiments, the storage device 612 is integrated into a server system 602. For example, the server system 602 may have one or more hard disk drives as the storage device 612. In other embodiments, the storage device 612 is external to the server system 602 and may be accessed by multiple server systems 602. For example, the storage device 612 may include multiple storage units, such as hard disks or solid-state disks, in a RAID (Redundant Array of Inexpensive Disks) configuration. The storage device 612 may include a storage area network (SAN) and / or network-attached storage (NAS) system.

[0072] In some embodiments, the processor 606 is operably coupled to the storage device 612 via a storage interface 608. The storage interface 608 is an optional component that allows the processor 606 to access the storage device 612. The storage interface 608 may include, for example, an Advanced Technology Attachment (ATA) adapter, a Serial ATA (SATA) adapter, a Small Computer System Interface (SCSI) adapter, a RAID controller, a SAN adapter, a network adapter, and / or any other component that provides the processor 606 with access to the storage device 612.

[0073] Memory area 610 may include, but is not limited to, random access memory (RAM) such as dynamic RAM (DRAM) or static RAM (SRAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), and non-volatile RAM (NVRAM). The above memory types are merely examples and are therefore not limited to the types of memory that can be used to store computer programs.

[0074] Figure 7 shows an exemplary configuration of a client computing device 702. The client computing device 702 may be used to implement one or more of the methods described herein. The client computing device 702 may, but is not limited to, include a client system ("client computing device") 504. The client computing device 702 includes a processor 704 for executing instructions. In some embodiments, executable instructions are stored in a memory area 706. The processor 704 may include one or more processing units (for example, in a multi-core configuration). The memory area 706 is any device capable of storing and retrieving information such as executable instructions and / or other data. The memory area 706 may include one or more computer-readable media.

[0075] The client computing device 702 also includes at least one media output component 708 for presenting information to the user 714. The media output component 708 is any component capable of conveying information to the user 714. In some embodiments, the media output component 708 includes an output adapter, such as a video adapter and / or an audio adapter. The output adapter is operably coupled to the processor 704 and operably coupled to an output device, such as a display device (e.g., a liquid crystal display (LCD), an organic light-emitting diode (OLED) display, a cathode ray tube (CRT), or an "electronic ink" display) or an audio output device (e.g., a speaker or headphones).

[0076] In some embodiments, the client computing device 702 includes an input device 710 for receiving input from the user 714. The input device 710 may include, for example, a keyboard, a pointing device, a mouse, a stylus, a touch-sensitive panel (e.g., a touchpad or touchscreen), a camera, a gyroscope, an accelerometer, a position detector, and / or an audio input device. A single component, such as a touchscreen, may function as both an output device and an input device 710 for the media output component 708.

[0077] The client computing device 702 may also include a communication interface 712 that can be communicatively coupled to a server system 301 or a remote device such as a web server. The communication interface 712 may include a wired or wireless network adapter, or a wireless data transceiver for use with, for example, a cellular network (e.g., Global Mobile Communication System (GSM), 3G, 4G, 5G, or Bluetooth) or another mobile data network (e.g., Worldwide Interoperability for Microwave Access (WiMAX)).

[0078] Memory area 706 stores computer-readable instructions for providing a user interface to user 714, for example, via media output component 708, and optionally for receiving and processing input from input device 710. The user interface may include, among other possibilities, a web browser and a client application. The web browser allows user 714 to view and interact with media and other information embedded on web pages or websites, typically from a web server. The client application allows user 714 to interact with a server application. The user interface facilitates the display of information provided by computing device 502, either through the web browser and / or the client application. The client application can operate in both online mode (where the client application is communicating with computing device 502) and offline mode (where the client application is not communicating with computing device 502).

[0079] Figure 8 shows an exemplary embodiment illustrating a proposed novel coordination framework compound based on a known coordination framework compound. This is an example of the method in Figure 2, illustrating how chemical insight is incorporated into the method. In Figure 8, a known MOF, MOF303;Al(OH)(1H-pyrazole-3,5-dicarboxylate) is considered. This known MOF has a rod-shaped SBU with a cis-trans alternating vertex (corner) shared aluminum octahedron. The SBU is set as a bent aluminum-SBU, and the topology is maintained. By changing the linker, a novel MOF is created that is a variant of the known MOF-303 MOF. In this example, changing any of the SBU, linker, or topology yields a novel MOF. This approach is more targeted by immediately presenting the MOF, but requires chemical intuition and knowledge. However, by focusing on MOFs with Al-SBUs, for example, the data requirements can be reduced in this method where one variable is limited.

[0080] Figure 9 shows exemplary coordination framework compounds, particularly MOFs, obtained from the ML model. The ML model analyzed the structure and composition of MOFs and suggested the use of a mixture of short-chain and long-chain linkers. Chemical studies of this presentation revealed that replacing the biphenyl-4,4'-dicarboxylate linker with a shorter linker improved hydrophilicity, and replacing the biphenyl-4,4'-dicarboxylate linker with a longer, more hydrophilic linker increased pore volume and potentially improved water adsorption isotherms.

[0081] Figure 10 shows exemplary molecules used in substitutional linkers in coordination framework compounds according to this disclosure. As can be seen from the figure, each of the 15 substitutional base molecules has two carboxylate moieties. Preliminary isotherms were obtained for five of the substitutional base molecules.

[0082] Figure 11 includes a comparison of the properties of a comparative coordination framework compound with one of the coordination framework compounds designed according to this disclosure. As can be seen, the coordination framework compounds according to this disclosure demonstrate equivalent or significantly improved water adsorption.

[0083] Figures 12–16 include exemplary coordination framework compounds proposed by the hybrid ML-Chemistry MOF proposed workflow.

[0084] Further aspects of this disclosure are provided by the subject matter of the following clauses.

[0085] 1. A method for proposing a coordination framework compound, The method is performed using a Coordination Framework Compound Proposal (CFCP) computing device, which includes a processor coupled to a memory device. The method involves using a machine learning model on a CFCP computing device to generate an initial set of coordination framework compounds, The initial set of coordination framework compounds is subjected to a review of at least one chemical property of at least one coordination framework compound, Using a CFCP computing device, a preliminary set of coordination framework compounds is generated using a machine learning model based on an initial set of coordination framework compounds and a review of at least one chemical property of at least one coordination framework compound included in the initial set of coordination framework compounds. A method comprising proposing a coordination framework compound using a CFCP computing device.

[0086] 2. The method according to the preceding paragraph, further comprising subjecting at least one coordination framework compound included in a preliminary set of coordination framework compounds to review at least one chemical property.

[0087] 3. The method according to any of the preceding paragraphs, further comprising verifying at least one coordination framework compound included in a preliminary set of coordination framework compounds.

[0088] 4. Using a machine learning model to generate a further preliminary set of coordination framework compounds based on at least one set of generated coordination framework compounds and a review of at least one chemical property of at least one coordination framework compound included in the set of at least one generated coordination framework compounds, In some cases, at least one coordination framework compound included in a further preliminary set of coordination framework compounds may be subjected to review of at least one chemical property, In some cases, verify at least one coordination framework compound included in a further preliminary set of coordination framework compounds. The method of any of the preceding items, further comprising performing at least one iteration of the sequence, including the following.

[0089] 5. The method according to any of the preceding paragraphs, wherein the coordination framework compound is a coordination framework compound included in an initial set of coordination framework compounds, a coordination framework compound included in a reserve set of coordination framework compounds, or a coordination framework compound included in a further reserve set of coordination framework compounds.

[0090] 6. The method according to any of the preceding items, wherein the coordination framework compound comprises a secondary construct unit (SBU), a linker, and topology.

[0091] 7. The method according to any of the preceding items, wherein the coordination framework compound is a metal-organic framework (MOF) compound or a covalent organic framework (COF) compound.

[0092] 8. The method described in any of the preceding paragraphs, wherein a machine learning model is trained on existing coordination framework compounds.

[0093] 9. The method described in any of the preceding paragraphs, wherein the machine learning model uses at least one of the following techniques: latent space, inverse search, variational autoencoder (VAE), crystal diffusion variational autoencoder (CDVAE), inverse search of VAE latent space, graph neural network (GNN), neural network, optimization, and combinations thereof.

[0094] 10. The method described in any of the preceding paragraphs, wherein a machine learning model is trained to learn a latent space configured to reconstruct the crystal structure of a coordination framework compound and accurately predict the relevant target properties.

[0095] 11. The method according to any of the preceding paragraphs, wherein at least one chemical property is selected from the group consisting of one or more of the following: adsorbate uptake capacity, adsorbate uptake kinetics, adsorbate weight productivity, adsorbate volume productivity, adsorbate isotherms and isobars, pore size, pore volume, heat of adsorption, heat of adsorption of equal amounts, chemical stability, thermal stability, mechanical stability, synthesizability, zeta potential, surface energy, hydrophobicity, hydrophilicity, chemical performance, chemical modification for improvement, and combinations thereof.

[0096] 12. The method according to any of the preceding paragraphs, wherein a machine learning model generates a preliminary coordination framework compound based on at least one input selected from the group consisting of the crystal structure, target properties, target chemical properties, sorption isotherm, water sorption isotherm, pore volume, pore diameter, water stability, hydrophilicity, and combinations thereof of an existing coordination framework compound.

[0097] 13. The method described in any of the preceding paragraphs, wherein the review of at least one chemical property is performed by machine, human, or a combination thereof.

[0098] 14. Coordination Framework Compound (CFCP) Proposal for Computing Device, Memory and A processor that is communicatively connected to memory, We generate an initial set of coordination framework compounds using a machine learning model. At least one coordination framework compound included in the initial set of coordination framework compounds is subjected to review of at least one chemical property. Based on a review of an initial set of coordination framework compounds and at least one chemical property of at least one coordination framework compound included in the initial set of coordination framework compounds, a preliminary set of coordination framework compounds is generated using a machine learning model. We propose a coordination framework compound. A processor programmed to do so, A computing device equipped with [a certain feature].

[0099] 15. A CFCP computing device as described in any of the preceding paragraphs, wherein the processor is further programmed to provide at least one coordination framework compound from a preliminary set of coordination framework compounds for review of at least one chemical property.

[0100] 16. A CFCP computing device as described in any of the preceding paragraphs, wherein the processor is further programmed to validate at least one coordination framework compound included in a preliminary set of coordination framework compounds.

[0101] 17. The processor, Based on a set of at least one generated coordination framework compounds and a review of at least one chemical property of at least one coordination framework compound included in the set of at least one generated coordination framework compounds, a machine learning model is used to generate a further preliminary set of coordination framework compounds. In some cases, at least one coordination framework compound included in a further preliminary set of coordination framework compounds may be subjected to review of at least one chemical property, In some cases, verify at least one coordination framework compound included in a further preliminary set of coordination framework compounds. A CFCP computing device as described in any of the preceding paragraphs, further programmed to perform at least one iteration of the sequence, including the above.

[0102] 18. A CFCP computing device according to any of the preceding paragraphs, wherein the coordination framework compound is a coordination framework compound included in an initial set of coordination framework compounds, a coordination framework compound included in a reserve set of coordination framework compounds, or a coordination framework compound included in a further reserve set of coordination framework compounds.

[0103] 19. A CFCP computing device according to any of the preceding items, wherein the coordination framework compound comprises a secondary construction unit (SBU), a linker, and a topology.

[0104] 20. A CFCP computing device according to any of the preceding items, wherein the coordination framework compound is a metal-organic framework (MOF) compound or a covalent organic framework (COF) compound.

[0105] 21. A CFCP computing device as described in any of the preceding paragraphs, in which a machine learning model is trained on an existing coordination framework compound.

[0106] 22. A CFCP computing device as described in any of the preceding paragraphs, wherein the machine learning model uses at least one of the following techniques: latent space, inverse search, variational autoencoder (VAE), crystal diffusion variational autoencoder (CDVAE), inverse search of VAE latent space, graph neural network (GNN), neural network, optimization, and combinations thereof.

[0107] 23. A CFCP computing device as described in any of the preceding paragraphs, in which a machine learning model is trained to learn a latent space configured to reconstruct the crystal structure of a coordination framework compound and accurately predict the associated target properties.

[0108] 24. A CFCP computing device according to any of the preceding paragraphs, wherein at least one chemical property is selected from the group consisting of one or more of the following: adsorbate uptake capacity, adsorbate uptake kinetics, adsorbate weight productivity, adsorbate volume productivity, adsorbate isotherms and isobars, pore size, pore volume, heat of adsorption, heat of adsorption of equal amounts, chemical stability, thermal stability, mechanical stability, synthesizability, zeta potential, surface energy, hydrophobicity, hydrophilicity, chemical performance, chemical modification for improvement, and combinations thereof.

[0109] 25. A CFCP computing device according to any of the preceding paragraphs, wherein a machine learning model generates a preliminary coordination framework compound based on at least one input selected from the group consisting of the crystal structure, target properties, target chemical properties, sorption isotherm, water sorption isotherm, pore volume, pore diameter, water stability, hydrophilicity, and combinations thereof of an existing coordination framework compound.

[0110] 26. A CFCP computing device as described in any of the preceding paragraphs, wherein a review of at least one chemical property is performed by machine, human, or a combination thereof.

[0111] 27. A non-temporary computer-readable storage medium in which a computer-executable instruction is embodied, and when executed by a proposed Coordination Framework Compound (CFCP) computing device including at least one processor communicating with memory, the computer-readable instruction is transmitted to the CFCP computing device. We use a machine learning model to generate an initial set of coordination framework compounds. At least one coordination framework compound included in the initial set of coordination framework compounds is subjected to review of at least one chemical property. Based on a review of an initial set of coordination framework compounds and at least one chemical property of at least one coordination framework compound included in the initial set of coordination framework compounds, a machine learning model is used to generate a preliminary set of coordination framework compounds, and A non-temporary computer-readable storage medium that prompts the proposal of coordination framework compounds.

[0112] 28. A non-temporary computer-readable storage medium as described in the preceding paragraph, further comprising a computer-readable instruction that provides a CFCP computing device with at least one coordination framework compound included in a preliminary set of coordination framework compounds for review of at least one chemical property.

[0113] 29. A non-temporary computer-readable storage medium as described in the preceding paragraph, wherein a computer-readable instruction causes a CFCP computing device to further verify at least one coordination framework compound included in a preliminary set of coordination framework compounds.

[0114] 30. Computer-readable instructions are sent to a CFCP computing device. Based on a set of at least one generated coordination framework compounds and a review of at least one chemical property of at least one coordination framework compound included in the set of at least one generated coordination framework compounds, a machine learning model is used to generate a further preliminary set of coordination framework compounds. In some cases, at least one coordination framework compound included in a further preliminary set of coordination framework compounds may be subjected to review of at least one chemical property, In some cases, verify at least one coordination framework compound included in a further preliminary set of coordination framework compounds. A non-temporary computer-readable storage medium as described in the preceding paragraph, which causes at least one further iteration of the sequence, including the sequence.

[0115] 31. A non-temporary computer-readable storage medium as described in any of the preceding paragraphs, wherein the coordination framework compound is a coordination framework compound included in an initial set of coordination framework compounds, a coordination framework compound included in a reserve set of coordination framework compounds, or a coordination framework compound included in a further reserve set of coordination framework compounds.

[0116] 32. A non-temporary computer-readable storage medium according to any of the preceding items, wherein the coordination framework compound includes secondary construction units (SBUs), linkers, and topology.

[0117] 33. A non-temporary computer-readable storage medium according to any of the preceding items, wherein the coordination framework compound is a metal-organic framework (MOF) compound or a covalent organic framework (COF) compound.

[0118] 34. A non-temporary computer-readable storage medium described in any of the preceding paragraphs, on which a machine learning model is trained using existing coordination framework compounds.

[0119] 35. A non-temporary computer-readable storage medium as described in any of the preceding paragraphs, wherein the machine learning model uses at least one of the following techniques: latent space, inverse search, variational autoencoder (VAE), crystal diffusion variational autoencoder (CDVAE), inverse search of VAE latent space, graph neural network (GNN), neural network, optimization, and combinations thereof.

[0120] 36. A non-temporary computer-readable storage medium as described in any of the preceding paragraphs, in which a machine learning model is trained to learn a latent space configured to reconstruct the crystal structure of a coordination framework compound and accurately predict the relevant target properties.

[0121] 37. A non-temporary computer-readable storage medium as described in any of the preceding paragraphs, wherein at least one chemical property is selected from the group consisting of one or more of the following: adsorbate uptake capacity, adsorbate uptake kinetics, adsorbate weight productivity, adsorbate volume productivity, adsorbate isotherms and isobars, pore size, pore volume, heat of adsorption, heat of adsorption of equal amounts, chemical stability, thermal stability, mechanical stability, synthesizability, zeta potential, surface energy, hydrophobicity, hydrophilicity, chemical performance, chemical modification for improvement, and combinations thereof.

[0122] 38. A non-temporary computer-readable storage medium according to any of the preceding paragraphs, wherein a machine learning model generates a preliminary coordination framework compound based on at least one input selected from the group consisting of the crystal structure, target properties, target chemical properties, sorption isotherm, water sorption isotherm, pore volume, pore diameter, water stability, hydrophilicity, and combinations thereof of an existing coordination framework compound.

[0123] 39. A non-temporary computer-readable storage medium as described in any of the preceding paragraphs, wherein a review of at least one chemical property is performed by machine, human, or a combination thereof.

[0124] 40. Coordination framework compounds, Coordination framework compounds, Multiple secondary building units (SBUs), Multiple linkers that form connections between multiple SBUs, It includes multiple pores formed in the gaps between the connections, At least two chemically different linkers, At least two geometrically different pores, and / or A coordination framework compound comprising at least one of at least two links between at least two SBUs among a plurality of SBUs.

[0125] 41. Coordination framework compounds as described in the preceding paragraphs, selected from the group consisting of metal-organic framework (MOF) compounds, covalent organic framework (COF) compounds, zeolite imidazolate framework (ZIF) compounds, crystalline porous materials, crystalline open frameworks, network chemical compounds, and combinations thereof.

[0126] 42. At least one of the multiple SBUs, Metal atoms, Al or Mg, B, C, N, O, Si, or P, Transition metal atoms, Fe, Co, Cu, or Zn, A coordination framework compound according to any of the preceding items, comprising a node containing an atom selected from the group consisting of those combinations.

[0127] 43. A coordination framework compound according to any of the preceding items, wherein at least one of the multiple SBUs includes a coordination structure selected from the group consisting of polyhedra, tetrahedra, octahedrons, cubes, dodecahedrons, and combinations thereof.

[0128] 44. A coordination framework compound described in any of the preceding items, which is symmetrical in the plane.

[0129] 45. A coordination framework compound described in any of the preceding items, which is not plane symmetric.

[0130] 46. ​​A coordination framework compound according to any of the preceding items, wherein at least two chemically distinct linkers comprise at least two linkers of different lengths.

[0131] 47. A coordination framework compound according to any of the preceding items, wherein at least two geometrically distinct pores are distinguished by geometric properties selected from the group consisting of size, shape, and combinations thereof.

[0132] 48. Multiple linkers, Linker of formula IA: [ka] Linker of formula IA or formula IB: [ka] (Formula IB) (In the formula, n1, m1, n2, and m2 are individually selected from the groups consisting of integers less than or equal to 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, and 100, integers less than or equal to 1,000, integers less than or equal to 10,000, integers less than or equal to 100,000, and integers less than or equal to 1,000,000, respectively. R1, R2, R3, R4, R8, R9, R 10 and R 11 Each is individually selected from the group consisting of H, NH2, OH, and SH. R5 and R6 are directly bonded, R 12 NHR 13 , R 12 Ure 13 , R 12 SR 13 Individually selected from the group consisting of C1-C6 alkyls optionally substituted with at least one substituent selected from the group consisting of NH2, OH, and SH, C1-C6 alkylenes optionally substituted with at least one substituent selected from the group consisting of NH2, OH, and SH, and combinations thereof, R7 is selected from the group consisting of direct bonds, ring fusions, NH, O, S, and C1-C6 alkyl groups. R 12 and R 13 Each is individually selected from the group consisting of direct bonds, NH, O, S, and C1-C6 alkyl groups. A1, A2, A3, A4, A5, A6, A7, and A8 are each individually selected from the group consisting of C, N, O, and S. Linker of formula IIA: [ka] Linker of formula IIA or formula IIB: [ka] (Formula IIB) (In the formula, n3, m3, n4, and m4 are individually selected from the group consisting of integers less than or equal to 1, 2, 3, 4, 5, 6, 7, 8, 9, and 10, integers less than or equal to 100, integers less than or equal to 1000, integers less than or equal to 10,000, integers less than or equal to 10,000, and integers less than or equal to 1,000,000, respectively.) R 14 , R 15 , R 16 , R 20 , R 21 and R 22 Each is individually selected from the group consisting of H, NH2, OH, and SH. R 17 and R 18 These are direct bonding and R, respectively. 23 NHR 24 , R 23 Ure 24 , R 23 SR 24 Individually selected from the group consisting of C1-C6 alkyls optionally substituted with at least one substituent selected from the group consisting of NH2, OH, and SH, C1-C6 alkylenes optionally substituted with at least one substituent selected from the group consisting of NH2, OH, and SH, and combinations thereof, R 19 These are selected from the group consisting of direct bonds, ring fusions, NH, O, S, and C1-C6 alkyl groups. R 23 and R 24 Each is individually selected from the group consisting of direct bonds, NH, O, S, and C1-C6 alkyl; and B1, B2, B3, B4, B5, and B6 are each individually selected from the group consisting of C, N, O, and S. A coordination framework compound selected from the group consisting of those combinations, as described in any of the preceding items.

[0133] 49. Multiple linkers, [ka] [ka] [ka] [ka] , and A coordination framework compound as described in any of the preceding items, comprising a linker selected from the group consisting of those combinations.

[0134] 50. Multiple linkers, [ka] , and A coordination framework compound as described in any of the preceding items, comprising a linker selected from the group consisting of those combinations.

[0135] The reference to “several embodiments” in the above description is not intended to be interpreted as excluding the existence of additional embodiments that also incorporate the features described. [Examples]

[0136] Without further detail, it is assumed that those skilled in the art using the foregoing description can make full use of the present invention. Therefore, the following embodiments should be construed as merely illustrative and not in any way limiting the disclosure. The starting materials for the following embodiments do not necessarily have to be prepared by the specific preparation procedures described in other embodiments. Furthermore, any numerical ranges listed herein should be understood to include all values ​​from the lower to the upper values. For example, if the range is stated as 10–50, it is intended that values ​​such as 12–30, 20–40, or 30–50 are explicitly listed herein. These are merely examples of what is specifically intended, and all possible combinations of numbers between the listed lowest and highest values ​​should be considered as explicitly stated in this application. The following embodiments can be performed using the CFCP computing devices and / or non-temporary computer-readable storage media described herein. [Examples]

[0137] As detailed herein, we followed the proposed hybrid ML-Chemistry MOF workflow.

[0138] First, training data from an MOF database, including the MOF-303 variant, was provided to the CDVAE model. Starting with the MOF-303 variant, the MOF structure was optimized to maximize pore volume (PV). Approximately 10,000 MOFs were generated. After algorithmic filtering, approximately 100 MOFs were filtered out. An example MOF from the ML model is shown in Figure 9.

[0139] Next, the filtered MOFs were subjected to chemical analysis. CDVAE proposed using a mixture of long linkers and several short linkers. This proposal was further investigated by running density functional theory by Perdew-Burke-Ernzerhof (PBE) in the Vienna Ab Initio Simulation Package (VASP) program. For plane waves, an energy cutoff of 520 eV was used. Soft PAW pseudopotentials were used for atoms (H, C, N, O) where available. The k-point mesh contained only Γ points. The convergence threshold for the self-consistent field cycle was 10 -6 The force threshold for geometric structure optimization was 0.05 eV / Å. There were 39 candidate density functional theories that were optimized. Different SBUs and different linkers were considered. This chemical analysis revealed that replacing the biphenyl linker with a shorter linker improved the hydrophilicity of the MOF.

[0140] Following this chemical analysis, 15 different compounds optimized by density functional theory were proposed as MOF compounds. These 15 different compounds contained shorter linkers, underwent chemical evaluation, and were proposed as MOF compounds. The molecules used as substitute linkers in these 15 compounds are shown in Figure 10. Their corresponding pore volumes are shown in Table 1 below. As referenced below, MOF-303 corresponds to Al(OH)(1H-pyrazole-3,5-dicarboxylate). Finally, several hypothetical MOFs with larger pore volumes than MOF-303 were identified, and these compounds have the potential to improve water adsorption isotherms compared to MOF-303.

[0141] [Table 1] [Examples]

[0142] The coordination framework compounds proposed by the hybrid ML-Chemistry MOF proposed workflow were simulated using GEMC simulations, and their simulated properties were compared with existing coordination framework compounds.

[0143] Exemplary coordination framework compounds were simulated according to the following description. For each coordination framework compound, the SBU contained Al(μ2-OH rods) arranged alternately in a cis-trans configuration.

[0144] AWE-MOF-2: A coordination framework compound containing a 50:50 mixture of the following linkers: [ka]

[0145] AWE-MOF-3: A coordination framework compound containing a 50:50 mixture of the following linkers: [ka]

[0146] AWE-MOF-4: A coordination framework compound containing a 50:50 mixture of the following linkers: [ka]

[0147] AWE-MOF-5: A coordination framework compound containing a 50:50 mixture of the following linkers: [ka]

[0148] AWE-MOF-6: Comparative example coordination framework compound containing the following linker: [ka]

[0149] MOF-LA2-1: Coordination framework compound of comparative example containing the following linker: [ka]

[0150] The results are shown in Figure 11. Compared with the coordination framework compounds of the comparative examples, the coordination framework compounds according to this disclosure showed equivalent or significantly improved water adsorption. [Examples]

[0151] The coordination framework compounds proposed by the hybrid ML-Chemistry MOF proposed workflow are shown in Figures 12 to 16.

[0152] The coordination framework compounds according to this disclosure can exhibit various topologies. Figures 12A and 12B show coordination framework compounds including a linear SBU topology. Figures 13A and 13B show coordination framework compounds including a curved SBU topology. Figures 14A and 14B show coordination framework compounds including an SBU topology that is neither linear nor curved. It is observed that a single linker can connect to three or more connection points on an SBU.

[0153] The coordination framework compounds described herein may exhibit a variety of linkers and linkages. Figure 15 shows a coordination framework compound including a crosslinked arm linker structure. Figure 16 shows a coordination framework compound including a pore shape altered as a result of a mixture of different linkers.

[0154] Conclusion.

[0155] This specification describes a process that combines the strengths of machine learning-based and chemical knowledge-based methods to more effectively propose and discover coordination framework compounds. This process is broadly applicable to coordination framework compounds, and exemplary coordination framework compounds are demonstrated.

[0156] Unless otherwise indicated, the approximation terms used herein, such as “generally,” “substantially,” and “about,” indicate that the terms thus modified may apply only to an approximate degree as recognized by those skilled in the art, and not to an absolute or complete degree. Therefore, values ​​modified with one or more terms such as “about,” “approximately,” and “substantially” are not limited to the exact values ​​specified. In at least some cases, the approximation terms may correspond to the precision of the instrument used to measure the value. In addition, unless otherwise indicated, terms such as “first,” “second,” etc., are used herein solely as labels and are not intended to impose any order, position, or hierarchical requirements on the items referred to by these terms. Furthermore, a reference to, for example, a “second” item does not require or exclude, for example, the presence of a “first” item or an item with a lower number, or a “third” item or an item with a higher number.

[0157] Certain features of various embodiments of the present invention may be shown in some drawings and not in others, but this is for convenience only. Furthermore, the reference to “some embodiments” in the above description is not intended to be construed as excluding the existence of additional embodiments that also incorporate the described features. According to the principles of the present invention, any feature in the drawings may be referenced and / or claimed in combination with any feature in any other drawing.

[0158] This specification discloses the present invention, including its best mode, using examples, and enables any person skilled in the art to practice the invention, including the fabrication and use of any device or system, and the execution of any incorporated method. The patentable scope of the present invention is defined by the claims and may include other examples that a person skilled in the art could conceive. Such other examples are intended to be within the claims if they have structural elements that do not differ from the language of the claims, or if they include equivalent structural elements that do not substantially differ from the language of the claims. [Explanation of Symbols]

[0159] 110 Method Flowchart 120 Method Flowcharts 122 Machine Learning Models 210 Flowchart 301 Server System 310 Objective Flowchart 312 Coordination Framework Compounds 314 ML model (VAE) 318 Back-searching of Latent Space 410 ML Model Flowchart 414 Latent space 500 Computer Systems 502 Computing Devices 504 Client System 506 Network 508 Database Server 510 User Interface 512 Databases 514 Data Sources 602 Server System 604 Communication Interface 606 Processor 608 Storage Interfaces 610 memory, memory area 612 storage devices 702 Client Computing Devices 704 Processor 706 memory area 708 Media Output Components 710 Input Devices 712 Communication Interface 714 users

Claims

1. A method for proposing a coordination framework compound (312), wherein the method is performed using a coordination framework compound proposal (CFCP) computing device (502) which includes a processor (606) coupled to a memory device (610), The method involves generating an initial set of coordination framework compounds (312) using the machine learning model (122) of the CFCP computing device (502), Subject at least one coordination framework compound (312) included in the initial set of coordination framework compounds (312) to review at least one chemical property, Using the CFCP computing device (502), a preliminary set of coordination framework compounds (312) is generated using the machine learning model (122) based on the initial set of coordination framework compounds (312) and a review of at least one chemical property of at least one of the coordination framework compounds (312) included in the initial set of coordination framework compounds (312), Using the CFCP computing device (502), the coordination framework compound (312) is proposed. Methods that include...

2. The method according to claim 1, further comprising subjecting at least one of the coordination framework compounds (312) included in the preliminary set of coordination framework compounds (312) to review at least one chemical property.

3. The method according to claim 1, further comprising verifying at least one coordination framework compound (312) included in the preliminary set of coordination framework compounds (312).

4. Based on a set of at least one generated coordination framework compounds (312) and a review of at least one chemical property of the at least one coordination framework compound (312) included in the set of at least one generated coordination framework compounds (312), a further preliminary set of coordination framework compounds (312) is generated using the machine learning model (122), In some cases, at least one of the coordination framework compounds (312) included in a further reserve set of the coordination framework compounds (312) is subjected to review of at least one chemical property, Depending on the circumstances, verify at least one of the coordination framework compounds (312) included in a further reserve set of the coordination framework compounds (312) and The method according to claim 1, further comprising performing at least one iteration of the sequence, including the following:

5. The method according to claim 1, wherein the coordination framework compound (312) comprises a secondary construction unit (SBU), a linker, and a topology.

6. The method according to claim 1, wherein the coordination framework compound (312) is a metal-organic framework (MOF) compound or a covalent organic framework (COF) compound.

7. The method according to claim 1, wherein the machine learning model (122) is trained with an existing coordination framework compound (312).

8. The method according to claim 1, wherein the machine learning model (122) uses at least one of the techniques of latent space (414), inverse search, variational autoencoder (VAE), crystal diffusion variational autoencoder (CDVAE), inverse search of VAE latent space, graph neural network (GNN), neural network, optimization, and combinations thereof.

9. The method according to claim 1, wherein the machine learning model (122) is trained to learn a latent space (414) configured to reconstruct the crystal structure of the coordination framework compound (312) and accurately predict the relevant target properties.

10. The method according to claim 1, wherein the at least one chemical property is selected from the group consisting of one or more of the following: adsorbate uptake capacity, adsorbate uptake kinetics, adsorbate weight productivity, adsorbate volume productivity, adsorbate isotherms and isobars, pore size, pore volume, heat of adsorption, heat of adsorption of equal amounts, chemical stability, thermal stability, mechanical stability, synthesizability, zeta potential, surface energy, hydrophobicity, hydrophilicity, chemical performance, chemical modification for improvement, and combinations thereof.

11. The method according to claim 1, wherein the machine learning model (122) generates a preliminary coordination framework compound based on at least one input selected from the group consisting of the crystal structure of an existing coordination framework compound (312), target properties, target chemical properties, sorption isotherm, water sorption isotherm, pore volume, pore diameter, water stability, hydrophilicity, and combinations thereof.

12. The method according to claim 1, wherein the review of the at least one chemical property is performed by machine, human, or a combination thereof.

13. A coordination framework compound proposal (CFCP) computing device (502), Memory (610) and, A processor (606) that is communicatively connected to the memory (610), An initial set of coordination framework compounds (312) is generated using a machine learning model (122). The at least one coordination framework compound (312) included in the initial set of coordination framework compounds (312) is subjected to review of at least one chemical property. Based on a review of the initial set of coordination framework compounds (312) and the at least one chemical property of the at least one coordination framework compound (312) included in the initial set of coordination framework compounds (312), a preliminary set of coordination framework compounds (312) is generated using the machine learning model. A processor (606) programmed to propose the aforementioned coordination framework compound (312), A computing device (502) equipped with the following:

14. The CFCP computing device (502) according to claim 13, wherein the processor (606) is further programmed to subject at least one of the coordination framework compounds (312) included in the preliminary set of coordination framework compounds (312) to review at least one chemical property.

15. The CFCP computing device (502) according to claim 13, wherein the processor (606) is further programmed to verify at least one coordination framework compound (312) included in the preliminary set of coordination framework compounds (312).

16. The aforementioned processor (606) Based on a set of at least one generated coordination framework compounds (312) and a review of at least one chemical property of the at least one coordination framework compound (312) included in the set of at least one generated coordination framework compounds (312), a further preliminary set of coordination framework compounds (312) is generated using the machine learning model (122), In some cases, at least one of the coordination framework compounds (312) included in the further set of coordination framework compounds (312) is subjected to review of the at least one chemical property, In some cases, verify at least one of the coordination framework compounds (312) included in the further set of coordination framework compounds (312) A CFCP computing device (502) according to claim 13, further programmed to perform at least one iteration of a sequence, including the above.

17. A non-temporary computer-readable storage medium in which a computer-executable instruction is embodied, and when executed by a Coordination Framework Compound Proposal (CFCP) computing device including at least one processor communicating with memory, the computer-readable instruction is transmitted to the CFCP computing device (502), An initial set of coordination framework compounds (312) is generated using a machine learning model (122). At least one of the coordination framework compounds (312) included in the initial set of coordination framework compounds (312) is subjected to review of the chemical properties of the at least one. Based on a review of the initial set of coordination framework compounds (312) and the at least one chemical property of the at least one coordination framework compound (312) of the initial set of coordination framework compounds (312), a preliminary set of coordination framework compounds (312) is generated using the machine learning model (122), and A non-temporary computer-readable storage medium that prompts the proposal of coordination framework compounds.

18. The non-temporary computer-readable storage medium according to claim 17, wherein the computer-readable instruction further provides the CFCP computing device (502) with the at least one coordination framework compound (312) included in the spare set of coordination framework compounds (312) for review of the at least one chemical property.

19. The non-temporary computer-readable storage medium according to claim 17, wherein the computer-readable instruction causes the CFCP computing device (502) to further verify at least one of the coordination framework compounds (312) included in the spare set of coordination framework compounds (312).

20. The computer-readable instruction is sent to the CFCP computing device (502), Based on the set of at least one generated coordination framework compound (312) and a review of at least one chemical property of the at least one coordination framework compound (312) included in the set of at least one generated coordination framework compound (312), a further preliminary set of coordination framework compounds (312) is generated using the machine learning model (122), In some cases, at least one of the coordination framework compounds (312) included in the further set of coordination framework compounds (312) is subjected to review of the at least one chemical property, In some cases, verify at least one of the coordination framework compounds (312) included in the further set of coordination framework compounds (312) A non-temporary computer-readable storage medium according to claim 17, which causes at least one more iteration of the sequence to be performed.

21. Coordination framework compound (312), The coordination framework compound (312) comprises a plurality of secondary construct units (SBUs), Multiple linkers that form connections between the multiple SBUs, The gap between the connections includes a plurality of pores formed therein, At least two chemically different linkers, At least two geometrically distinct pores, and / or At least two connections between two of the aforementioned SBUs A coordination framework compound (312) comprising at least one of the following.

22. The coordination framework compound (312) according to claim 21, selected from the group consisting of metal-organic framework (MOF) compounds, covalent organic framework (COF) compounds, zeolite imidazolate framework (ZIF) compounds, crystalline porous materials, crystalline open frameworks, network chemical compounds, and combinations thereof.

23. At least one of the aforementioned plurality of SBUs, Metal atoms, Al or Mg, B, C, N, O, Si, or P A transition metal atom, Fe, Co, Cu, or Zn, The coordination framework compound (312) according to claim 21, comprising a node containing an atom selected from the group consisting of such combinations.

24. The coordination framework compound (312) according to claim 21, wherein at least one of the plurality of SBUs includes a coordination structure selected from the group consisting of polyhedra, tetrahedra, octahedrons, cubes, dodecahedrons, and combinations thereof.

25. The coordination framework compound (312) according to claim 21, wherein the coordination framework compound (312) is symmetrical in the plane.

26. The coordination framework compound (312) according to claim 21, which is not planar symmetric.

27. The coordination framework compound (312) according to claim 21, wherein the at least two chemically distinct linkers comprise at least two linkers of different lengths.

28. The coordination framework compound (312) according to claim 21, wherein the at least two geometrically distinct pores are distinguished by geometric properties selected from the group consisting of size, shape, and combinations thereof.

29. Multiple linkers, Linker of formula IA: 【Chemistry 1】 Linker of formula IA or formula IB: 【Chemistry 2】 (Formula IB) (In the formula, n １ , m １ , n ２ , and m ２ Each of these is individually selected from the groups consisting of integers up to 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 100, integers up to 1000, integers up to 10,000, integers up to 10,000, and integers up to 1,000,000. R １ 、 R ２ 、 R ３ 、 R ４ 、 R ８ 、 R ９ 、 R １０ and R １１ are each independently selected from the group consisting of H, NH ２ , OH and SH R ５ and R ６ These are direct bonding and R, respectively. １２ NHR １３ , R １２ OR １３ , R １２ SR １３ NH ２ C optionally substituted with at least one substituent selected from the group consisting of , OH, and SH １ ~C ６ Alkyl, NH ２ C optionally substituted with at least one substituent selected from the group consisting of , OH, and SH １ ~C ６ Alkylenes and their combinations are individually selected from the group, R ７ These are direct bonds, ring fusions, NH, O, S and C １ ~C ６ Selected from the group consisting of alkyl groups, R １２ and R １３ These are direct bonds, NH, O, S, and C, respectively. １ ~C ６ Individually selected from the group consisting of alkyl groups, A １ A ２ A ３ A ４ A ５ A ６ A ７ , and A ８ (Each of these is individually selected from the group consisting of C, N, O, and S.) Linker of formula IIA: 【Transformation 3】 Linker of formula IIA or formula IIB: 【Chemistry 4】 (Formula IIB) (In the formula, n ３ , m ３ , n ４ , and m ４ Each of these is individually selected from the groups consisting of integers up to 1, 2, 3, 4, 5, 6, 7, 8, 9, and 10, integers up to 100, integers up to 1000, integers up to 10,000, integers up to 10,000, and integers up to 1,000,000. R １４ , R １５ , R １６ , R ２０ , R ２１ and R ２２ H and NH, respectively. ２ Individually selected from the group consisting of OH and SH, R １７ and R １８ These are direct bonding and R, respectively. ２３ NHR ２４ , R ２３ OR ２４ , R ２３ SR ２４ NH ２ C optionally substituted with at least one substituent selected from the group consisting of , OH, and SH １ ~C ６ Alkyl, NH ２ C optionally substituted with at least one substituent selected from the group consisting of , OH, and SH １ ~C ６ Alkylenes and their combinations are individually selected from the group, R １９ These are direct bonds, ring fusions, NH, O, S and C １ ~C ６ Selected from the group consisting of alkyl groups, R ２３ and R ２４ These are direct bonds, NH, O, S, and C, respectively. １ ~C ６ Individually selected from the group consisting of alkyl groups, B １ , B ２ , B ３ , B ４ , B ５ and B ６ (Each is individually selected from the group consisting of C, N, O, and S) and A coordination framework compound (312) according to claim 21, selected from the group consisting of those combinations.

30. The aforementioned plurality of linkers 【Chemistry 5-1】 【Chemistry 5-2】 【Chemistry 5-3】 【Chemistry 5-4】 and The coordination framework compound (312) according to claim 21, comprising a linker selected from the group consisting of those combinations.

31. The aforementioned plurality of linkers 【Transformation 6】 and The coordination framework compound (312) according to claim 21, comprising a linker selected from the group consisting of those combinations.

32. An adsorbent system comprising the coordination framework compound (312) according to claim 21.

Images