BonDNet: a graph neural network for the prediction of bond dissociation energies for charged molecules

A broad collection of technologies, including e.g. drug metabolism, biofuel combustion, photochemical decontamination of water, and interfacial passivation in energy production/storage systems rely on chemical processes that involve bond-breaking molecular reactions. In this context, a fundamental thermodynamic property of interest is the bond dissociation energy (BDE) which measures the strength of a chemical bond. Fast and accurate prediction of BDEs for arbitrary molecules would lay the groundwork for data-driven projections of complex reaction cascades and hence a deeper understanding of these critical chemical processes and, ultimately, how to reverse design them. In this paper, we propose a chemically inspired graph neural network machine learning model, BonDNet, for the rapid and accurate prediction of BDEs. BonDNet maps the difference between the molecular representations of the reactants and products to the reaction BDE. Because of the use of this difference representation and the introduction of global features, including molecular charge, it is the first machine learning model capable of predicting both homolytic and heterolytic BDEs for molecules of any charge. To test the model, we have constructed a dataset of both homolytic and heterolytic BDEs for neutral and charged (−1 and +1) molecules. BonDNet achieves a mean absolute error (MAE) of 0.022 eV for unseen test data, significantly below chemical accuracy (0.043 eV). Besides the ability to handle complex bond dissociation reactions that no previous model could consider, BonDNet distinguishes itself even in only predicting homolytic BDEs for neutral molecules; it achieves an MAE of 0.020 eV on the PubChem BDE dataset, a 20% improvement over the previous best performing model. We gain additional insight into the model''s predictions by analyzing the patterns in the features representing the molecules and the bond dissociation reactions, which are qualitatively consistent with chemical rules and intuition. BonDNet is just one application of our general approach to representing and learning chemical reactivity, and it could be easily extended to the prediction of other reaction properties in the future.

Prediction of bond dissociation energies for charged molecules with a graph neural network enabled by global molecular features and reaction difference features between products and reactants.     

Katalytische isomerisierung von heptenen mit IrX(CO)L2

1-, 2-cis-, 2-trans-, and 3-trans-heptenes (C₇)are isomerized either very slowly or not at all with IrX(CO)L₂ at 80°C in toluene and under N₂. However, under the conditions of hydrogenation fast isomerisation takes place. With IrCl(CO)L₂ as catalyst the rate of isomerisation decreases the order: 1-C₇ ∼ 2-cis-C₇ > 3-trans-C₇ > 2-trans-C₇. This sequence is independent of the ligand L in lrCl(CO)L₂, however, with a particular isomer the rate of isomerisation is a function of L in the order L = PPh₃ > P(C₆H₁₁)3 > P(OPh)₃.     

Dendrophanes: Water-soluble dendritic receptors as models for buried recognition sites in globular proteins

Water-soluble dendritic cyclophanes (dendrophanes) of first ( 1 , 4 ), second ( 2 5 ), and third generation ( 3 6 ) with poly(ether amide) branching and 12, 36, and 108 terminal carboxylate groups, respectively, were prepared by divergent synthesis, and their molecular recognition properties in aqueous solutions were investigated. Dendrophanes 1 – 3 incorporate as the initiator core a tetraoxa[6.1.6.1]paracyclophane 7 with a suitably sized cavity for inclusion complexation of benzene or naphthalene derivatives. The initiator core in 4 – 6 is the [6.1.6.1]cyclo-phane 8 shaped by two naphthyl(phenyl) methane units with a cavity suitable for steroid incorporation. The syntheses of 1 – 6 involved sequential peptide coupling to monomer 9 , followed by ester hydrolysis (Schemes 1 and 4), Purification by gel-permeation chromatography (GPC; Fig. 3) and full spectral characterization were accomplished at the stage of the intermediate poly(methyl carboxylates) 10 – 12 and 23 – 25 , respectively. The third-generation 108-ester 25 was also independently prepared by a semi-convergent synthetic strategy, starting from 4 (Scheme 5). All dendrophanes with terminal ester groups were obtained in pure form according to the ¹³C-NMR spectral criterion (Figs, 1 and 5). The MALDI-TOF mass spectra of the third-generation derivative 25 (mol. wt. 19328 D) displayed the molecular ion as base peak, accompanied by a series of ions [M – n(1041 ± 7)]⁺, tentatively assigned as characteristic fragment ions of the poly(ether amide) cascade. A similar fragmentation pattern was also observed in the spectra of other higher-generation poly(ether amide) dendrimers. Attempts to prepare monodisperse fourth-generation dendrophanes by divergent synthesis failed. ¹H-NMR and fluorescence binding titrations in basic aqueous buffer solutions showed that dendrophanes 1 – 3 complexed benzene and naphthalene derivatives, whereas 4 – 6 bound the steroid testosterone. Complexation occurred exclusively at the cavity-binding site of the central cyclophane core rather than in fluctuating voids in the dendritic branches, and the association strength was similar to that of the complexes formed by the initiator cores 7 and 8 , respectively (Tables 1 and 3). Fluorescence titrations with 6-(p-toluidino)naphthalene-2-sulfonate as fluorescent probe in aqueous buffer showed that the micropolarity at the cyclophane core in dendrophanes 1 - 3 becomes increasingly reduced with increasing size and density of the dendritic superstructure; the polarity at the core of the third-generation compound 3 is similar to that of EtOH (Table 2). Host-guest exchange kinetics were remarkably fast and, except for receptor 3 , the stabilities of all dendrophane complexes could be evaluated by ¹H-NMR titrations. The rapid complexation-decomplexation kinetics are explained by the specific attachment of the dendritic wedges to large, nanometer-sized cyclophane initiator cores, which generates apertures in the surrounding dendritic superstructure.     

The determination of chromium, copper and nickel in groundwater using axial plasma inductively coupled plasma atomic emission spectrometry and proportional correction matrix effect reduction

The performance of a proportional correction matrix effect reduction procedure was investigated for an axially viewed inductively coupled plasma. It was shown that the proportional correction factor (ratio of analyte matrix effect and internal standard matrix effect) was sufficiently stable over the investigated matrix element concentration ranges (0–2000 mg/L of Na and 0–400 mg/L of Ca) for the procedure to be successful. Proportional correction results in the best correction for matrix effects compared to the classical 1?:?1 intensity ratio correction procedure or the approach without any correction, as was shown in recovery experiments using analyte spiked groundwater samples. Matrix effects as high as 18% without correction were reduced to less than 4% applying proportional correction.     

Synthesis and structural characterization of acetatobis(triphenylphosphine)dicarbonylmanganese(I), (CH3CO2)Mn(CO)2[P(C6H5)3]2: An organometallic complex containing a chelating acetate ligand

The accidental but intriguing synthesis of acetatobis(triphenylphosphine)dicarbonylmanganese(I), (CH₃CO₂)Mn(CO)₂[P(C₆H₅)₃]₂, has been accomplished by the reaction of NaMn(CO)₅ with (CH₃)₃SiCl followed by the addition of triphenylphosphine and acetic acid. A three-dimensional single-crystal X-ray diffraction analysis has shown an octahedral-like molecule containing a symmetrically oxygen-chelating acetate group, the first such group to be reported in a metal carbonyl complex. The two triphenylphosphine ligands occupy mutually trans positions with the two carbonyl ligands possessing the remaining cis sites in the octahedral complex. The compound crystallizes with four molecules in a monoclinic unit cell of space group symmetry $\text{P}\text{2}{}_1\text{c}$ and of dimensions a = 17.744(2) Å, b = 9.692(1) Å, c = 20.004(2) Å, and β = 106.195(4)°. The relatively long MnO(acetate) bond lengths [2.066(6) and 2.069(7) Å] and the relatively short MnCO bond lengths [1.701(12) and 1.760(13) Å] and the relatively short MnP(C₆H₅)₃ bond lengths [2.260(3) and 2.275(3) Å], compared to the corresponding MnCO and MnP(C₆H₅)₃ bond lengths in other manganese carbonyl triphenylphosphine complexes, are rationalized on the basis that the acetate ligand in this molecule functions primarily as a σ-donor.     

Deactivation of vibrationally excited CD3H using laser-induced fluorescence

Infrared fluorescence observed after exciting to ν₆ (ν=1) of CD₃H with a Q-switched CO₂ laser yields the exponential deactivation rate constant of 0.84 ms^?1 torr^?1. Rate constants for deactivation of CD₃H by rare gases vary from 1.4 (for He) to 0.029 (for Xe) ms^?1 torr^?1.     

Reference Energies in Semiempirical Parametrizations

The MNDO scheme has been reparametrized for hydrocarbons using both molecular binding energies and heats of formation at 0 K as experimental reference data. Compared with MNDO, there are only minor changes in the optimized parameters, and the results are essentially of the same accuracy. These tests justify the use of heats of formation at 298 K as reference data in the original MNDO parametrization.     

Vinylsulfonylethyl derivatives of nucleic acid bases and their graft polymers on polyethyleneimine

Vinylsulfonylethyl (VSE) derivatives of the nucleic acid bases adenine, thymine, cytosine, and the nucleosides inosine and uridine have been prepared via a simple Michael reaction with divinyl sulfone. The VSE derivatives were grafted on a polyethyleneimine (PEI) backbone. PEI of different molecular weights (1400, 1800 and 50,000–100,000) were used and also two different molar ratios (1:1 and 1:2) of monomer to PEI were employed. From the ¹H-NMR and elemental analysis, it appeared that in almost all instances the grafting was quantitative. In one case, both 1-VSE-thymine and 9-VSE-adenine were grafted on the same PEI backbone. Interactions between some of these polymers were investigated by UV spectroscopy. The expected complementary base pairing was observed only in DMSO–ethylene glycol solvent system but not in DMSO. The adenine polymer showed a one-to-one interaction with the thymine polymer.     

Speciation analysis — why and how?

Summary The different aspects of speciation analysis are reviewed. Species-specific instrumental techniques as well as various speciation schemes are considered for the determination of species of metals and metalloids, including organometallic compounds. The application of the methods are discussed in some detail for the analysis of natural waters, air, soil, sediment and biological samples. The relationship between metal species and bioavailability is also briefly dealt with.