The Effect of Chemical Representation on Active Machine Learning Towards Closed-Loop Optimization

Multivariate chemical reaction optimization involving catalytic systems is a non-trivial task due to the high number of tuneable parameters and discrete choices. Closed-loop optimization featuring active Machine Learning (ML) represents a powerful strategy for automating reaction optimization. However, the translation of chemical reaction conditions into a machine-readable format comes with the challenge of finding highly informative features which accurately capture the factors for reaction success and allow the model to learn efficiently. Herein, we compare the efficacy of different calculated chemical descriptors for a high throughput generated dataset to determine the impact on a supervised ML model when predicting reaction yield. Then, the effect of featurization and size of the initial dataset within a closed-loop reaction optimization was examined. Finally, the balance between descriptor complexity and dataset size was considered. Ultimately, tailored descriptors did not outperform simple generic representations, however, a larger initial dataset accelerated reaction optimization.

High-yield synthesis of CsPbBr3 nanoparticles: Diphenylphosphine as a reducing agent and its effect in Pb-seeding nucleation and growth

Based on the reported nucleation mechanisms for CsPbX3 and II-VI/IV-VI quantum dots, CsPbBr3 nanoparticles with a high reaction-yield, up to 393% mass-increment, were synthesized by the hot-injection method. The introduction of diphenylphosphine (DPP) as a reducing agent improved nanoparticle nucleation and growth, giving out evidence for Pb-seeding in CsPbBr3 nanoparticles formation. Additionally, a clear influence of the DPP in a CsPbBr3-Cs4PbBr6 incomplete phase transformation was observed, marked by the appearance of several PbBr2 nanoparticles, indicating the need for an improved ratio between the stabilizing agents and the precursors, due to the increased number of nucleation sites produced by the DPP. The resulting CsPbBr3 nanoparticles showed high quality, as they displayed 70%-90% photoluminescence quantum yield (PLQY), narrow size distribution with an average nanoparticle size of ~10 nm and the characteristic cubic morphology reported in previous works. This increment in CsPbBr3 nanoparticles’ reaction yield will contribute to making them a more attractive option for different optoelectronic applications.

Mechanistic Insight into the Precursor Chemistry of ZrO2 and HfO2 Nanocrystals; towards Size-Tunable Syntheses

One can nowadays readily generate monodisperse colloidal nanocrystals, but a retrosynthetic analysis is still not possible since the underlying chemistry is often poorly understood. Here, we provide insight into the reaction mechanism of colloidal zirconia and hafnia nanocrystals synthesized from metal chloride and metal isopropoxide. We identify the active precursor species in the reaction mixture through a combination of nuclear magnetic resonance spectroscopy (NMR), density functional theory (DFT) calculations, and pair distribution function (PDF) analysis. We gain insight into the interaction of the surfactant, tri-n-octylphosphine oxide (TOPO), and the different precursors. Interestingly, we identify a peculiar X-type ligand redistribution mechanism that can be steered by the relative amount of Lewis base (L-type). We further monitor how the reaction mixture decomposes using solution NMR and gas chromatography, and we find that ZrCl4 is formed as a by-product of the reaction, limiting the reaction yield. The reaction proceeds via two competing mechanisms: E1 elimination (dominating) and SN1 substitution (minor). Using this new mechanistic insight, we adapted the synthesis to optimize the yield and gain control over nanocrystal size. These insights will allow the rational design and synthesis of complex oxide nanocrystals.

Synthesis of a new furan-bearing triazine dichloride compound

Donor-acceptor compounds have been receiving increased attention lately, especially in opto-electronic applications. In this work, a novel compound named triazine-furan, comprised of a triazine acceptor moiety and a furan donor side group, was designed and synthesised for the first time. The influences of temperature and the feeding proportion of cyanuric chloride to furfurylamine were investigated to optimise the reaction yield and purification process. The synthesised compound was characterised using thin-layer chromatography (TLC), Fourier-transform infrared spectroscopy (FTIR), and proton nuclear magnetic resonance (1H-NMR) spectroscopy.

PREPARATION AND STUDY OF THE PROPERTIES OF THE COPOLYMER OF CYCLIC UNSATURATED BISMALEIMIDE WITH STYRENE

The paper is devoted to the study of regularities of radical copolymerization reactions of unsaturated cyclic N,N'-bisimides (M1) with styrene (M2), the study of dependence of copolymerization reaction rate, composition of obtained copolymers, composition and estimated molecular weight on the composition of the initial monomer mixture. With the increase of bismaleimide amount in the initial monomer mixture a decrease of copolymerization reaction yield and molecular weight of the obtained copolymers was observed. This fact can be explained by active participation of bismaleimide in chain transfer in radical copolymerization reaction. The composition of the obtained copolymers in all cases is rich in styrene links

ESTIMATION OF GROSS-STRUCTURE PARAMETERS OF GIANT DIPOLE RESONANCE: 2. EXPERIMENTAL TESTING

Experimental testing of a novel technique for determination of width and maximum of excitation function of a photonuclear reaction with dominant giant dipole resonance is conducted. The method is based on measurement of normalized reaction yield in a thin target, overlapping entirely a flux of X-rays and on processing of data obtained with the use of a developed analytical model. For the checking of method, the nickel and molybdenum foils of natural isotopic composition were activated by bremsstrahlung radiation at four energies of the electron beam in the range 40…95 MeV. The obtained parameters of cross-section of the reference reactions 58Ni(γ,n)57Ni and 100Mo(γ,n)99Mo are in good agreement with those presented in the available databases.

Improved Synthesis of a Novel Biodegradable Tunable Micellar Polymer Based on Partially Hydrogenated Poly(β-malic Acid-co-benzyl Malate)

Poly(benzyl malate) (PBM), together with its derivatives, have been studied as nanocarriers for biomedical applications due to their superior biocompatibility and biodegradability. The acquisition of PBM is primarily from chemical routes, which could offer polymer-controlled molecular weight and a unique controllable morphology. Nowadays, the frequently used synthesis from L-aspartic acid gives an overall yield of 4.5%. In this work, a novel synthesis route with malic acid as the initiator was successfully designed and optimized, increasing the reaction yield up to 31.2%. Furthermore, a crystalline form of PBM (PBM-2) that polymerized from high optical purity benzyl-β-malolactonate (MLABn) was discovered during the optimization process. X-ray diffraction (XRD) patterns revealed that the crystalline PBM-2 had obvious diffraction peaks, demonstrating that its internal atoms were arranged in a more orderly manner and were different from the amorphous PBM-1 prepared from the racemic MLABn. The differential scanning calorimetry (DSC) curves and thermogravimetric curves elucidated the diverse thermal behaviors between PBM-1 and PBM-2. The degradation curves and scanning electron microscopy (SEM) images further demonstrated the biodegradability of PBM, which have different crystal structures. The hardness of PBM-2 implied the potential application in bone regeneration, while it resulted in the reduction of solubility when compared with PBM-1, which made it difficult to be dissolved and hydrogenated. The solution was therefore heated up to 75 °C to achieve benzyl deprotection, and a series of partially hydrogenated PBM was sequent prepared. Their optimal hydrogenation rates were screened to determine the optimal conditions for the formation of micelles suitable for drug-carrier applications. In summary, the synthesis route from malic acid facilitated the production of PBM for a shorter time and with a higher yield. The biodegradability, biosafety, mechanical properties, and adjustable hydrogenation widen the application of PBM with tunable properties as drug carriers.

A theoretical study on Ir(III)-catalyzed intermolecular branch-selective allylic C−H amidation

Herein, we report the mechanism of Ir(III)-catalyzed intermolecular branch-selective allylic C−H amidation, including the influence of substituent effect on yield and regioselectivity. The sequence of amidation reaction is alkene coordination, allylic C−H activation, oxidative addition of methyl dioxazolone, reductive elimination of allyl-Ir-nitrenoid complex, amine protonation and proto-demetallation. The apparent activation energy of amidation between hexene and methyl dioxazolone is 17.8 kcal/mol, and the energy difference between two transition state for formation amide is only 2.8 kcal/mol. The introduction of more electron-deficient groups at the allyl terminal increases the apparent activation energy, conversely, the introduction of electron-donating groups significantly reduces the apparent activation energy. Among them, the apparent activation energy of the reaction between aniline group substituted allyl and methyl dioxazolone is only 13.8 kcal/mol, which further improves the reaction yield. In addition, the introduction of more electron-withdrawing groups on dioxazolone can significantly improve the regioselectivity. When 3,4,5,-trifluorophenyl substituted dioxazolone and hexene occur C−N bond coupling reaction, the energy difference of the two transition states is as high as 9.0 kcal/mol, indicating that the regioselectivity is greatly improved. The mechanism explanation of allylic C−H amidation will provide strong theoretical support for streamlined synthesis of allyl branched amides.

Solid-state mechanochemical technology for deep processing of brown coal: energy efficiency improvement and dust formation control

Energy efficiency – a compromise between reaction depth and expended energy - is an actual question for any technology implementation. Mechanochemical technology for brown coal deep processing has one more compromise - between reaction depth and dust formation. Indeed, the depth of mechanochemical reactions usually correlates with grinding efficiency, but for coal cases dust formation is an unwanted process. Here we consider a solid-state mechanochemical reaction of humic acid oxidation by sodium percarbonate in one laboratory mill at different conditions. The ratio between the grinding bodies load and the payload was varied, the reaction yield and the ground samples characteristics were controlled.