Molecular generation targeting desired electronic properties via deep generative models

As we seek to discover new functional materials, we need ways to explore the vast chemical space of precursor building blocks, not only generating large numbers of possible building blocks to investigate, but trying to find non-obvious options, that we might not suggest by chemical experience alone. Artificial intelligence techniques provide a possible avenue to generate large numbers of organic building blocks for functional materials, and can even do so from very small initial libraries of known building blocks. Specifically, we demonstrate the application of deep recurrent neural networks for the exploration of the chemical space of building blocks for a test case of donor-acceptor oligomers with specific electronic properties. The recurrent neural network learned how to produce novel donor-acceptor oligomers by trading off between selected atomic substitutions, such as halogenation or methylation, and molecular features such as the oligomer's size. The electronic and structural properties of the generated oligomers can be tuned by sampling from different subsets of the training database, which enabled us to enrich the library of donor-acceptors towards desired properties. We generated approximately 1700 new donor-acceptor oligomers with a recurrent neural network tuned to target oligomers with a HOMO-LUMO gap <2 eV and a dipole moment <2 Debye, which could have potential application in organic photovoltaics.

Organic Cage Dumbbells

Molecular dumbbells with organic cage capping units were synthesised via a multi-component imine condensation between a tri-topic amine and di- and tetra-topic aldehydes. This is an example of self-sorting, which can be rationalised by computational modelling.

Author Correction: Hydrophilic microporous membranes for selective ion separation and flow-battery energy storage

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

Hydrophilic microporous membranes for selective ion separation and flow-battery energy storage

Membranes with fast and selective ion transport are widely used for water purification and devices for energy conversion and storage including fuel cells, redox flow batteries and electrochemical reactors. However, it remains challenging to design cost-effective, easily processed ion-conductive membranes with well-defined pore architectures. Here, we report a new approach to designing membranes with narrow molecular-sized channels and hydrophilic functionality that enable fast transport of salt ions and high size-exclusion selectivity towards small organic molecules. These membranes, based on polymers of intrinsic microporosity containing Tröger's base or amidoxime groups, demonstrate that exquisite control over subnanometre pore structure, the introduction of hydrophilic functional groups and thickness control all play important roles in achieving fast ion transport combined with high molecular selectivity. These membranes enable aqueous organic flow batteries with high energy efficiency and high capacity retention, suggesting their utility for a variety of energy-related devices and water purification processes.

Conformational control of Pd2L4 assemblies with unsymmetrical ligands

With increasing interest in the potential utility of metallo-supramolecular architectures for applications as diverse as catalysis and drug delivery, the ability to develop more complex assemblies is keenly sought after. Despite this, symmetrical ligands have been utilised almost exclusively to simplify the self-assembly process as without a significant driving foa mixture of isomeric products will be obtained. Although a small number of unsymmetrical ligands have been shown to serendipitously form well-defined metallo-supramolecular assemblies, a more systematic study could provide generally applicable information to assist in the design of lower symmetry architectures. Pd2L4 cages are a popular class of metallo-supramolecular assembly; research seeking to introduce added complexity into their structure to further their functionality has resulted in a handful of examples of heteroleptic structures, whilst the use of unsymmetrical ligands remains underexplored. Herein we show that it is possible to design unsymmetrical ligands in which either steric or geometric constraints, or both, can be incorporated into ligand frameworks to ensure exclusive formation of single isomers of three-dimensional Pd2L4 metallo-supramolecular assemblies with high fidelity. In this manner it is possible to access Pd2L4 cage architectures of reduced symmetry, a concept that could allow for the controlled spatial segregation of different functionalities within these systems. The introduction of steric directing groups was also seen to have a profound effect on the cage structures, suggesting that simple ligand modifications could be used to engineer structural properties.

Structurally Diverse Covalent Triazine-Based Framework Materials for Photocatalytic Hydrogen Evolution from Water

A structurally diverse family of 39 covalent triazine-based framework materials (CTFs) are synthesized by Suzuki-Miyaura polycondensation and tested as hydrogen evolution photocatalysts using a high-throughput workflow. The two best-performing CTFs are based on benzonitrile and dibenzo[b,d]thiophene sulfone linkers, respectively, with catalytic activities that are among the highest for this material class. The activities of the different CTFs are rationalized in terms of four variables: the predicted electron affinity, the predicted ionization potential, the optical gap, and the dispersibility of the CTFs particles in solution, as measured by optical transmittance. The electron affinity and dispersibility in solution are found to be the best predictors of photocatalytic hydrogen evolution activity.

Synthesis of a Large, Shape-Flexible, Solvatomorphic Porous Organic Cage

Porous organic cages have emerged over the last 10 years as a subclass of functional microporous materials. However, among all of the organic cages reported, large multicomponent organic cages with 20 components or more are still rare. Here, we present an [8 + 12] porous organic imine cage, CC20, which has an apparent surface area up to 1752 m² g^-1, depending on the crystallization and activation conditions. The cage is solvatomorphic and displays distinct geometrical cage structures, caused by crystal-packing effects, in its crystal structures. This indicates that larger cages can display a certain range of shape flexibility in the solid state, while remaining shape persistent and porous.

Mapping binary copolymer property space with neural networks

The extremely large number of unique polymer compositions that can be achieved through copolymerisation makes it an attractive strategy for tuning their optoelectronic properties. However, this same attribute also makes it challenging to explore the resulting property space and understand the range of properties that can be realised. In an effort to enable the rapid exploration of this space in the case of binary copolymers, we train a neural network using a tiered data generation strategy to accurately predict the optical and electronic properties of 350 000 binary copolymers that are, in principle, synthesizable from their dihalogen monomers via Yamamoto, or Suzuki-Miyaura and Stille coupling after one-step functionalisation. By extracting general features of this property space that would otherwise be obscured in smaller datasets, we identify simple models that effectively relate the properties of these copolymers to the homopolymers of their constituent monomers, and challenge common ideas behind copolymer design. We find that binary copolymerisation does not appear to allow access to regions of the optoelectronic property space that are not already sampled by the homopolymers, although it conceptually allows for more fine-grained property control. Using the large volume of data available, we test the hypothesis that copolymerisation of 'donor' and 'acceptor' monomers can result in copolymers with a lower optical gap than their related homopolymers. Overall, despite the prevalence of this concept in the literature, we observe that this phenomenon is relatively rare, and propose conditions that greatly enhance the likelihood of its experimental realisation. Finally, through a 'topographical' analysis of the co-polymer property space, we show how this large volume of data can be used to identify dominant monomers in specific regions of property space that may be amenable to a variety of applications, such as organic photovoltaics, light emitting diodes, and thermoelectrics.

The influence of nitrogen position on charge carrier mobility in enantiopure aza[6]helicene crystals

A computational study exploring the influence of the nitrogen position on charge carrier mobility in enantiopure aza[6]helicene crystals

Computationally-inspired discovery of an unsymmetrical porous organic cage

A completely unsymmetrical porous organic cage was synthesised from a C_2v symmetrical building block that was identified by a computational screen. The cage was formed through a 12-fold imine condensation of a tritopic C_2v symmetric trialdehyde with a ditopic C₂ symmetric diamine in a [4 + 6] reaction. The cage was rigid and microporous, as predicted by the simulations, with an apparent Brunauer-Emmett-Teller surface area of 578 m² g^-1. The reduced symmetry of the tritopic building block relative to its topicity meant there were 36 possible structural isomers of the cage. Experimental characterisation suggests a single isomer with 12 unique imine environments, but techniques such as NMR could not conclusively identify the isomer. Computational structural and electronic analysis of the possible isomers was used to identify the most likely candidates, and hence to construct a 3-dimensional model of the amorphous solid. The rational design of unsymmetrical cages using building blocks with reduced symmetry offers new possibilities in controlling the degree of crystallinity, porosity, and solubility, of self-assembled materials.

An evolutionary algorithm for the discovery of porous organic cages

The chemical and structural space of possible molecular materials is enormous, as they can, in principle, be built from any combination of organic building blocks. Here we have developed an evolutionary algorithm (EA) that can assist in the efficient exploration of chemical space for molecular materials, helping to guide synthesis to materials with promising applications. We demonstrate the utility of our EA to porous organic cages, predicting both promising targets and identifying the chemical features that emerge as important for a cage to be shape persistent or to adopt a particular cavity size. We identify that shape persistent cages require a low percentage of rotatable bonds in their precursors (<20%) and that the higher topicity building block in particular should use double bonds for rigidity. We can use the EA to explore what size ranges for precursors are required for achieving a given pore size in a cage and show that 16 Å pores, which are absent in the literature, should be synthetically achievable. Our EA implementation is adaptable and easily extendable, not only to target specific properties of porous organic cages, such as optimal encapsulants or molecular separation materials, but also to any easily calculable property of other molecular materials.

stk: A python toolkit for supramolecular assembly

A tool for the automated assembly, molecular optimization and property calculation of supramolecular materials is presented. stk is a modular, extensible and open‐source Python library that provides a simple Python API and integration with third party computational codes. stk currently supports the construction of linear polymers, small linear oligomers, organic cages in multiple topologies and covalent organic frameworks (COFs) in multiple framework topologies, but is designed to be easy to extend to new, unrelated, supramolecules or new topologies. Extension to metal–organic frameworks (MOFs), metallocycles or supramolecules, such as catenanes, would be straightforward. Through integration with third party codes, stk offers the user the opportunity to explore the potential energy landscape of the assembled supramolecule and then calculate the supramolecule's structural features and properties. stk provides support for high‐throughput screening of large batches of supramolecules at a time. The source code of the program can be found at https://github.com/supramolecular-toolkit/stk. © 2018 The Authors. Journal of Computational Chemistry published by Wiley Periodicals, Inc.

pywindow: Automated Structural Analysis of Molecular Pores

Structural analysis of molecular pores can yield important information on their behavior in solution and in the solid state. We developed pywindow, a python package that enables the automated analysis of structural features of porous molecular materials, such as molecular cages. Our analysis includes the cavity diameter, number of windows, window diameters, and average molecular diameter. Molecular dynamics trajectories of molecular pores can also be analyzed to explore the influence of flexibility. We present the methodology, validation, and application of pywindow for the analysis of molecular pores, metal-organic polyhedra, and some instances of framework materials. pywindow is freely available from github.com/JelfsMaterialsGroup/pywindow .

High-throughput discovery of organic cages and catenanes using computational screening fused with robotic synthesis

Supramolecular synthesis is a powerful strategy for assembling complex molecules, but to do this by targeted design is challenging. This is because multicomponent assembly reactions have the potential to form a wide variety of products. High-throughput screening can explore a broad synthetic space, but this is inefficient and inelegant when applied blindly. Here we fuse computation with robotic synthesis to create a hybrid discovery workflow for discovering new organic cage molecules, and by extension, other supramolecular systems. A total of 78 precursor combinations were investigated by computation and experiment, leading to 33 cages that were formed cleanly in one-pot syntheses. Comparison of calculations with experimental outcomes across this broad library shows that computation has the power to focus experiments, for example by identifying linkers that are less likely to be reliable for cage formation. Screening also led to the unplanned discovery of a new cage topology-doubly bridged, triply interlocked cage catenanes.

High-Throughput Screening Approach for the Optoelectronic Properties of Conjugated Polymers

We propose a general high-throughput virtual screening approach for the optical and electronic properties of conjugated polymers. This approach makes use of the recently developed xTB family of low-computational-cost density functional tight-binding methods from Grimme and co-workers, calibrated here to (Time-Dependent) Density Functional Theory ((TD)DFT) data computed for a representative diverse set of (co)polymers. Parameters drawn from the resulting calibration using a linear model can then be applied to the xTB derived results for new polymers, thus generating near DFT-quality data with orders of magnitude reduction in computational cost. As a result, after an initial computational investment for calibration, this approach can be used to quickly and accurately screen on the order of thousands of polymers for target applications. We also demonstrate that the (opto)electronic properties of the conjugated polymers show only a very minor variation when considering different conformers and that the results of high-throughput screening are therefore expected to be relatively insensitive with respect to the conformer search methodology applied.

Correction: A computational exploration of the crystal energy and charge-carrier mobility landscapes of the chiral [6]helicene molecule

Correction for 'A computational exploration of the crystal energy and charge-carrier mobility landscapes of the chiral [6]helicene molecule' by Beth Rice et al., Nanoscale, 2018, 10, 1865-1876.

A computational exploration of the crystal energy and charge-carrier mobility landscapes of the chiral [6]helicene molecule

The potential of a given π-conjugated organic molecule in an organic semiconductor device is highly dependent on molecular packing, as it strongly influences the charge-carrier mobility of the material. Such solid-state packing is sensitive to subtle differences in their intermolecular interactions and is challenging to predict. Chirality of the organic molecule adds an additional element of complexity to intuitive packing prediction. Here we use crystal structure prediction to explore the lattice-energy landscape of a potential chiral organic semiconductor, [6]helicene. We reproduce the experimentally observed enantiopure crystal structure and explain the absence of an experimentally observed racemate structure. By exploring how the hole and electron-mobility varies across the energy-structure-function landscape for [6]helicene, we find that an energetically favourable and frequently occurring packing motif is particularly promising for electron-mobility, with a highest calculated mobility of 2.9 cm² V^-1 s^-1 (assuming a reorganization energy of 0.46 eV). We also calculate relatively high hole-mobility in some structures, with a highest calculated mobility of 2.0 cm² V^-1 s^-1 found for chains of helicenes packed in a herringbone fashion. Neither the energetically favourable nor high charge-carrier mobility packing motifs are intuitively obvious, and this demonstrates the utility of our approach to computationally explore the energy-structure-function landscape for organic semiconductors. Our work demonstrates a route for the use of computational simulations to aid in the design of new molecules for organic electronics, through the a priori prediction of their likely solid-state form and properties.

Emergent Properties of an Organic Semiconductor Driven by its Molecular Chirality

Chiral molecules exist as pairs of nonsuperimposable mirror images; a fundamental symmetry property vastly underexplored in organic electronic devices. Here, we show that organic field-effect transistors (OFETs) made from the helically chiral molecule 1-aza[6]helicene can display up to an 80-fold difference in hole mobility, together with differences in thin-film photophysics and morphology, solely depending on whether a single handedness or a 1:1 mixture of left- and right-handed molecules is employed under analogous fabrication conditions. As the molecular properties of either mirror image isomer are identical, these changes must be a result of the different bulk packing induced by chiral composition. Such underlying structures are investigated using crystal structure prediction, a computational methodology rarely applied to molecular materials, and linked to the difference in charge transport. These results illustrate that chirality may be used as a key tuning parameter in future device applications.

Chirality as a tool for function in porous organic cages

The control of solid state assembly for porous organic cages is more challenging than for extended frameworks, such as metal-organic frameworks. Chiral recognition is one approach to achieving this control. Here we investigate chiral analogues of cages that were previously studied as racemates. We show that chiral cages can be produced directly from chiral precursors or by separating racemic cages by co-crystallisation with a second chiral cage, opening up a route to producing chiral cages from achiral precursors. These chiral cages can be cocrystallized in a modular, 'isoreticular' fashion, thus modifying porosity, although some chiral pairings require a specific solvent to direct the crystal into the desired packing mode. Certain cages are shown to interconvert chirality in solution, and the steric factors governing this behavior are explored both by experiment and by computational modelling.

Topological landscapes of porous organic cages

We define a nomenclature for the classification of porous organic cage molecules, enumerating the 20 most probable topologies, 12 of which have been synthetically realised to date. We then discuss the computational challenges encountered when trying to predict the most likely topological outcomes from dynamic covalent chemistry (DCC) reactions of organic building blocks. This allows us to explore the extent to which comparing the internal energies of possible reaction outcomes is successful in predicting the topology for a series of 10 different building block combinations.

Application of computational methods to the design and characterisation of porous molecular materials

Composed from discrete units, porous molecular materials (PMMs) possess properties not observed for conventional, extended solids. Molecular simulations provide crucial understanding for the design and characterisation of these unique materials.

Polymer nanofilms with enhanced microporosity by interfacial polymerization

Highly permeable and selective membranes are desirable for energy-efficient gas and liquid separations. Microporous organic polymers have attracted significant attention in this respect owing to their high porosity, permeability and molecular selectivity. However, it remains challenging to fabricate selective polymer membranes with controlled microporosity that are stable in solvents. Here we report a new approach to designing crosslinked, rigid polymer nanofilms with enhanced microporosity by manipulating the molecular structure. Ultrathin polyarylate nanofilms with thickness down to 20 nm are formed in situ by interfacial polymerization. Enhanced microporosity and higher interconnectivity of intermolecular network voids, as rationalized by molecular simulations, are achieved by using contorted monomers for the interfacial polymerization. Composite membranes comprising polyarylate nanofilms with enhanced microporosity fabricated in situ on crosslinked polyimide ultrafiltration membranes show outstanding separation performance in organic solvents, with up to two orders of magnitude higher solvent permeance than membranes fabricated with nanofilms made from non-contorted planar monomers.

Predicting solvent effects on the structure of porous organic molecules

A computational approach for the prediction of the open, metastable, conformations of porous organic molecules in the presence of solvent is developed.

Understanding static, dynamic and cooperative porosity in molecular materials

The practical adsorption properties of molecular porous solids can be dominated by dynamic flexibility but these effects are still poorly understood. Here, we combine molecular simulations and experiments to rationalize the adsorption behavior of a flexible porous organic cage.

The practical adsorption properties of molecular porous solids can be dominated by dynamic flexibility but these effects are still poorly understood. Here, we combine molecular simulations and experiments to rationalize the adsorption behavior of a flexible porous organic cage.

Porous Organic Cages for Sulfur Hexafluoride Separation

A series of porous organic cages is examined for the selective adsorption of sulfur hexafluoride (SF6) over nitrogen. Despite lacking any metal sites, a porous cage, CC3, shows the highest SF6/N2 selectivity reported for any material at ambient temperature and pressure, which translates to real separations in a gas breakthrough column. The SF6 uptake of these materials is considerably higher than would be expected from the static pore structures. The location of SF6 within these materials is elucidated by X-ray crystallography, and it is shown that cooperative diffusion and structural rearrangements in these molecular crystals can rationalize their superior SF6/N2 selectivity.

Separation of rare gases and chiral molecules by selective binding in porous organic cages

The separation of molecules with similar size and shape is an important technological challenge. For example, rare gases can pose either an economic opportunity or an environmental hazard and there is a need to separate these spherical molecules selectively at low concentrations in air. Likewise, chiral molecules are important building blocks for pharmaceuticals, but chiral enantiomers, by definition, have identical size and shape, and their separation can be challenging. Here we show that a porous organic cage molecule has unprecedented performance in the solid state for the separation of rare gases, such as krypton and xenon. The selectivity arises from a precise size match between the rare gas and the organic cage cavity, as predicted by molecular simulations. Breakthrough experiments demonstrate real practical potential for the separation of krypton, xenon and radon from air at concentrations of only a few parts per million. We also demonstrate selective binding of chiral organic molecules such as 1-phenylethanol, suggesting applications in enantioselective separation.

Shape selectivity by guest-driven restructuring of a porous material

A flexible metal-organic framework selectively sorbs para- (pX) over meta-xylene (mX) by synergic restructuring around pX coupled with generation of unused void space upon mX loading. The nature of the structural change suggests more generally that flexible structures which are initially mismatched in terms of fit and capacity to the preferred guest are strong candidates for effective molecular separations.

Predicted crystal energy landscapes of porous organic cages

Computational methods predict the crystal packing of porous organic cage molecules, allowing crystal structure and porosity to be predicted starting from the chemical diagram alone.

In silico design of supramolecules from their precursors: odd-even effects in cage-forming reactions

We synthesize a series of imine cage molecules where increasing the chain length of the alkanediamine precursor results in an odd-even alternation between [2 + 3] and [4 + 6] cage macrocycles. A computational procedure is developed to predict the thermodynamically preferred product and the lowest energy conformer, hence rationalizing the observed alternation and the 3D cage structures, based on knowledge of the precursors alone.

Dimensionality transformation through paddlewheel reconfiguration in a flexible and porous Zn-based metal-organic framework

The reaction between Zn and a pyrene-based ligand decorated with benzoate fragments (H(4)TBAPy) yields a 2D layered porous network with the metal coordination based on a paddlewheel motif. Upon desolvation, the structure undergoes a significant and reversible structural adjustment with a corresponding reduction in crystallinity. The combination of computationally assisted structure determination and experimental data analysis of the desolvated phase revealed a structural change in the metal coordination geometry from square-pyramidal to tetrahedral. Simulations of desolvation showed that the local distortion of the ligand geometry followed by the rotation and displacement of the pyrene core permits the breakup of the metal-paddlewheel motifs and the formation of 1D Zn-O chains that cross-link adjacent layers, resulting in a dimensionality change from the 2D layered structure to a 3D structure. Constrained Rietveld refinement of the powder X-ray diffraction pattern of the desolvated phase and the use of other analytical techniques such as porosity measurements, (13)C CP MAS NMR spectroscopy, and fluorescence spectroscopy strongly supported the observed structural transformation. The 3D network is stable up to 425 °C and is permanently porous to CO(2) with an apparent BET surface area of 523(8) m(2)/g (p/p° = 0.02-0.22). Because of the hydrophobic nature, size, and shape of the pores of the 3D framework, the adsorption behavior of the structure toward p-xylene and m-xylene was studied, and the results indicated that the shape of the isotherm and the kinetics of the adsorption process are determined mainly by the shape of the xylene isomers, with each xylene isomer interacting with the host framework in a different manner.

Porous organic cage nanocrystals by solution mixing

We present here a simple method for the bottom-up fabrication of microporous organic particles with surface areas in the range 500-1000 m(2) g(-1). The method involves chiral recognition between prefabricated, intrinsically porous organic cage molecules that precipitate spontaneously upon mixing in solution. Fine control over particle size from 50 nm to 1 μm can be achieved by varying the mixing temperature or the rate of mixing. No surfactants or templates are required, and the resulting organic dispersions are stable for months. In this method, the covalent synthesis of the cage modules can be separated from their solution processing into particles because the modules can be dissolved in common solvents. This allows a "mix and match" approach to porous organic particles. The marked solubility change that occurs upon mixing cages with opposite chirality is rationalized by density functional theory calculations that suggest favorable intermolecular interactions for heterochiral cage pairings. The important contribution of molecular disorder to porosity and surface area is highlighted. In one case, a purposefully amorphized sample has more than twice the surface area of its crystalline analogue.