Enhancing Molecular Representations Via Graph Transformation Layers.

Molecular representation learning is an essential component of many molecule-oriented tasks, such as molecular property prediction and molecule generation. In recent years, graph neural networks (GNNs) have shown great promise in this area, representing a molecule as a graph composed of nodes and edges. There are increasing studies showing that coarse-grained or multiview molecular graphs are important for molecular representation learning. Most of their models, however, are too complex and lack flexibility in learning different granular information for different tasks. Here, we proposed a flexible and simple graph transformation layer (i.e., LineEvo), a plug-and-use module for GNNs, which enables molecular representation learning from multiple perspectives. The LineEvo layer transforms fine-grained molecular graphs into coarse-grained ones based on the line graph transformation strategy. Especially, it treats the edges as nodes and generates the new connected edges, atom features, and atom positions. By stacking LineEvo layers, GNNs can learn multilevel information, from atom-level to triple-atoms level and coarser level. Experimental results show that the LineEvo layers can improve the performance of traditional GNNs on molecular property prediction benchmarks on average by 7%. Additionally, we show that the LineEvo layers can help GNNs have more expressive power than the Weisfeiler-Lehman graph isomorphism test.

Enantioseparation of racemic amlodipine using immobilized ionic liquid by solid-phase extraction.

In this paper, a novel l-glutamate based immobilized chiral ionic liquid (SBA-IL (Glu)) was prepared by chemical bonding method and applied as a solid sorbent for chiral separation of amlodipine. The performance of SBA-IL (Glu) was investigated for the absorption of (S)-amlodipine and separation of amlodipine enantiomer. The static experiment showed that equilibrium adsorption was achieved within 80 minutes, and the saturation adsorptions capacity was 12 mg/g. The complex was then packed in a glass chromatographic column for the separation of amlodipine and the enantiomeric excess (%ee) of (S)-amlodipine reached 24.67%. The immobilized ionic liquids exhibit good reusability, and the separation efficiency remains 18.24% after reused five times, which allows potential scale-up for the chiral separation of amlodipine.

The effects of exogenous organic acids on the growth, photosynthesis and cellular ultrastructure of Salix variegata Franch. Under Cd stress.

We studied the effects of three organic acids (citric acid, tartaric acid and malic acid) on the biomass, photosynthetic pigment content and photosynthetic parameters of Salix variegata under Cd stress and observed the ultrastructure of mesophyll cells in each treatment. Cd stress significantly reduced photosynthesis by reducing the content of pigments and disrupting chloroplast structure, which consequently decreased the biomass. However, respective addition of three organic acids greatly increased the biomass of S. variegata under Cd stress. Among them, the effect of malic acid or tartaric acid on shoot and total biomass accumulation was greater than that of citric acid. The alleviation of biomass probably related with the photosynthetic process. Results revealed that treatment with each organic acid enhanced the net photosynthesis rate under Cd stress. Malic acid promoted plant growth and biomass by increasing the chlorophyll content and mitigating damage to the photosynthetic apparatus resulting from Cd stress. Tartaric acid had little impact on the photosynthetic pigment content, but it was important in mitigating the ultrastructural damage of plants caused by Cd. Addition of citric acid significantly increased the carotenoid as well as the number and volume of chloroplasts in mesophyll cells, while the mitigation of structural damage in the photosynthetic apparatus was weaker than that in tartaric acid or malic acid treatment. It is concluded that application of tartaric acid or malic acid is effective in increasing the growth potential of S. variegata under Cd stress and thus can be a promising approach for the phytoremediation of Cd-contaminated soil.

Enantioseparation of flurbiprofen enantiomers using chiral ionic liquids by liquid-liquid extraction.

Flurbiprofen is a kind of nonsteroidal anti-inflammatory drug, which has been widely used in clinic for treatment of rheumatoid arthritis and osteoarthritis. It has been reported that S-flurbiprofen shows good performance on clinic anti-inflammatory treatment, while R-enantiomer almost has no pharmacological activities. It has important practical values to obtain optically pure S-flurbiprofen. In this work, chiral ionic liquids, which have good structural designability and chiral recognize ability, were selected as the extraction selector by the assistance of quantum chemistry calculations. The distribution behaviors of flurbiprofen enantiomers were investigated in the extraction system, which was composed of organic solvent and aqueous phase containing chiral ionic liquid. The results show that maximum enantioselectivity up to 1.20 was attained at pH 2.0, 25°C using 1,2-dichloroethane as organic solvent, 1-butyl-3-methylimidazole L-tryptophan ([Bmim][L-trp]) as chiral selector. The racemic flurbiprofen initial concentration was 0.2 mmol L-1 , and [Bmim][L-trp] concentration was 0.02 mol L-1 . Furthermore, the recycle of chiral ionic liquids has been achieved by reverse extraction process of the aqueous phase with chiral selector, which is significant for industrial application of chiral ionic liquids and scale-up of the extraction process.

Continuous synthesis of hollow silver-palladium nanoparticles for catalytic applications.

Hollow bimetallic nanoparticles exhibit unique surface plasmonic properties, enhanced catalytic activities and high photo-thermal conversion efficiencies amongst other properties, however, their research and further deployment are currently limited by their complicated multi-step syntheses. This paper presents a novel approach for their continuous synthesis with controllable and tuneable sizes and compositions. This robust manufacturing tool, consisting of coiled flow inverter (CFI) reactors connected in series, allows for the first time the temporal and spatial separation of the initial formation of silver seeds and their subsequent galvanic displacement reaction in the presence of a palladium precursor, leading to the full control of both steps separately. We have also demonstrated that coupling the galvanic replacement and co-reduction leads to a great kinetic enhancement of the system leading to a high yield process of hollow bimetallic nanoparticles, directly applicable to other metal combinations.

Force field-inspired molecular representation learning for property prediction.

Molecular representation learning is a crucial task to accelerate drug discovery and materials design. Graph neural networks (GNNs) have emerged as a promising approach to tackle this task. However, most of them do not fully consider the intramolecular interactions, i.e. bond stretching, angle bending, torsion, and nonbonded interactions, which are critical for determining molecular property. Recently, a growing number of 3D-aware GNNs have been proposed to cope with the issue, while these models usually need large datasets and accurate spatial information. In this work, we aim to design a GNN which is less dependent on the quantity and quality of datasets. To this end, we propose a force field-inspired neural network (FFiNet), which can include all the interactions by incorporating the functional form of the potential energy of molecules. Experiments show that FFiNet achieves state-of-the-art performance on various molecular property datasets including both small molecules and large protein-ligand complexes, even on those datasets which are relatively small and without accurate spatial information. Moreover, the visualization for FFiNet indicates that it automatically learns the relationship between property and structure, which can promote an in-depth understanding of molecular structure.

Force field-inspired transformer network assisted crystal density prediction for energetic materials.

Machine learning has great potential in predicting chemical information with greater precision than traditional methods. Graph neural networks (GNNs) have become increasingly popular in recent years, as they can automatically learn the features of the molecule from the graph, significantly reducing the time needed to find and build molecular descriptors. However, the application of machine learning to energetic materials property prediction is still in the initial stage due to insufficient data. In this work, we first curated a dataset of 12,072 compounds containing CHON elements, which are traditionally regarded as main composition elements of energetic materials, from the Cambridge Structural Database, then we implemented a refinement to our force field-inspired neural network (FFiNet), through the adoption of a Transformer encoder, resulting in force field-inspired Transformer network (FFiTrNet). After the improvement, our model outperforms other machine learning-based and GNNs-based models and shows its powerful predictive capabilities especially for high-density materials. Our model also shows its capability in predicting the crystal density of potential energetic materials dataset (i.e. Huang & Massa dataset), which will be helpful in practical high-throughput screening of energetic materials.

Transition State Theory-Inspired Neural Network for Estimating the Viscosity of Deep Eutectic Solvents.

The lack of accurate methods for predicting the viscosity of solvent materials, especially those with complex interactions, remains unresolved. Deep eutectic solvents (DESs), an emerging class of green solvents, have a severe lack of viscosity data, resulting in their application still staying at the stage of random trial and error, and it is difficult for them to be implemented on an industrial scale. In this work, we demonstrate the successful prediction of the viscosity of DESs based on the transition state theory-inspired neural network (TSTiNet). The TSTiNet adopts multilayer perceptron (MLP) for the transition state theory-inspired equation (TSTiEq) parameters calculation and verification using the most comprehensive DESs viscosity data set to date. For the energy parameters of the TSTiEq, the constant assumption and the fast iteration with the help of MLP can allow TSTiNet to achieve the best performance (the average absolute relative deviation on the test set of 6.84% and R 2 of 0.9805). Compared with the traditional machine learning methods, the TSTiNet has better generalization ability and dramatically reduces the maximum relative deviation of prediction under the constraints of the thermodynamic formulation. It requires only the structural information on DESs and is the most accurate and reliable model available for DESs viscosity prediction.

Continuous Synthesis of Uniformly Dispersed Mesoporous SBA-15 Supported Silver Nanoparticles in a Coiled Flow Inverter Reactor.

Mesoporous silica supported nanocatalysts have shown great potential in industrial processes due to their unique properties, such as high surface area, large pore volume, good chemomechanical stability and so on. Controllable and tunable synthesis of supported nanocatalysts is a crucial problem. Continuous synthesis of supported nanoparticles has been reported to get uniformly dispersed nanomaterials. Here, a method for continuous synthesis of uniformly dispersed mesoporous SBA-15 supported silver nanoparticles in a coiled flow inverter (CFI) microreactor is described. Compared to Ag/SBA-15 synthesized in the conventional batch reactor and Ag synthesized in continuous flow, mesoporous silica nanocatalysts synthesized in continuous flow are found to have smaller average size (7-11 nm) and narrower size distribution. The addition of capping agents can effectively change the characteristic of catalysts. Moreover, two kinds of support with different surface area and pore size have been added into the continuous synthesis. This method can provide further understandings for the synthesis of uniformly dispersed supported nanocatalysts in continuous flow, especially for mesoporous nanomaterials, which provides the possibilities of large-scale yield process of supported nanocatalysts in industry.

Continuous microfluidic synthesis of zirconium-based UiO-67 using a coiled flow invertor reactor.

Metal-organic frameworks (MOFs), particularly Zirconium based, have a wide variety of potential applications, such as catalysis and separation. However, these are held back by traditionally only being synthesised in long batch reactions, which causes the process to be expensive and limit the amount of reaction control available, leading to potential batch to batch variation in the products, such as particle size distributions. Microfluidics allows for batch reactions to be performed with enhanced mass/heat transfer, with the coiled flow inverter reactor (CFIR) setup narrowing the residence time distribution, which is key in controlling the particle size and crystallinity. In this work, a Zirconium based MOF, UiO-67, has been synthesised continuously using a microfluidic CFIR, which has allowed for the product to be formed in 30 min, a fraction of the traditional batch heating time of 24 h. The microfluidicially synthesised UiO-67 is also smaller product with a narrower particle size distribution (≈200 nm to ≈400 nm) than its batch counterpart (~500 nm to over 3 µm).

catena-Poly[[[tetra-aqua-samarium(III)]-di-µ-isonicotinato-κO:O'] chloride].

In the structure of the title compound, {[Sm(C(6)H(4)NO(2))(2)(H(2)O)(4)]Cl}(n), the unique Sm(III) atom lies on a crystallographic twofold axis and is eight-coordinated by four O atoms from four isonicotinate ligands and four water mol-ecules in a slightly distorted square-anti-prismatic coodination environment. The Sm(III) atoms are bridged by two carboxyl-ate groups of two isonicotinate ligands, forming an extended chain along the c-axis direction. These chains are cross-linked through hydrogen bonds, forming a three-dimensional framework, with channels which accommodate the chloride anions.

Synthesis, crystal and band structures, and optical properties of a new mixed-framework mercury selenide diselenite, (Hg3Se2)(Se2O5).

A new mixed-framework mercury selenide diselenite, (Hg(3)Se(2))(Se(2)O(5)) (1), has been prepared by a solid-state reaction and structurally characterized by single-crystal X-ray diffraction analysis. The crystal structure of 1 consists of parallel stair-like cationic (Hg(3)Se(2))(2+) chains, which are bridged by (Se(2)O(5))(2-) anionic groups to form a novel 2-D layered mixed-framework. The optical properties were investigated in terms of the diffuse reflectance and microscopic infrared spectra. The electronic band structure along with density of states (DOS) calculated by the DFT method indicates that compound 1 is a semiconductor, and that the optical absorption of 1 is mainly ascribed to the charge transitions from the O-2p and Se(-II)-4p states to the Se(IV)-4p and Hg-6s states.