Hydrolysis is a fundamental family of chemical reactions where water facilitates the cleavage of bonds. The process is ubiquitous in biological and chemical systems, owing to water's remarkable versatility as a solvent. However, accurately predicting the feasibility of hydrolysis through computational techniques is a difficult task, as subtle changes in reactant structure like heteroatom substitutions or neighboring functional groups can influence the reaction outcome. Furthermore, hydrolysis is sensitive to the pH of the aqueous medium, and the same reaction can have different reaction properties at different pH conditions. In this work, we have combined reaction templates and high-throughput ab initio calculations to construct a diverse data set of hydrolysis free energies. The developed framework automatically identifies reaction centers, generates hydrolysis products, and utilizes a trained graph neural network (GNN) model to predict ΔG values for all potential hydrolysis reactions in a given molecule. The long-term goal of the work is to develop a data-driven, computational tool for high-throughput screening of pH-specific hydrolytic stability and the rapid prediction of reaction products, which can then be applied in a wide array of applications including chemical recycling of polymers and ion-conducting membranes for clean energy generation and storage.

Activation barriers of elementary reactions are essential to predict molecular reaction mechanisms and kinetics. However, computing these energy barriers by identifying transition states with electronic structure methods (e.g., density functional theory) can be time-consuming and computationally expensive. In this work, we introduce CoeffNet, an equivariant graph neural network that predicts activation barriers using coefficients of any frontier molecular orbital (such as the highest occupied molecular orbital) of reactant and product complexes as graph node features. We show that using coefficients as features offer several advantages, such as chemical interpretability and physical constraints on the network's behaviour and numerical range. Model outputs are either activation barriers or coefficients of the chosen molecular orbital of the transition state; the latter quantity allows us to interpret the results of the neural network through chemical intuition. We test CoeffNet on a dataset of S(N)2 reactions as a proof-of-concept and show that the activation barriers are predicted with a mean absolute error of less than 0.025 eV. The highest occupied molecular orbital of the transition state is visualized and the distribution of the orbital densities of the transition states is described for a few prototype S(N)2 reactions.

Here, we show how to redesign cycloolefin resins for biorenewable circularity by addressing how bio-derived polar functionality influences ring strain, ring–chain equilibria, and likelihood for catalyst poisoning. These bio-based cycloolefin resins (bioCOR) are thermally and photochemically polymerizable, resulting in high-modulus thermosets with excellent thermal stability. Most notably, we found that the efficacy of thermoset deconstruction to the original crosslinker depended strongly on the ring-closing depolymerization catalyst.

A modified shifted-echo PIETA pulse sequence is developed to acquire natural abundance 29Si 2D J-resolved spectra in crystalline silicates. The sequence is applied to the highly siliceous zeolites Sigma-2 and ZSM-12. The 2D J-resolved spectra are used to develop a silicate framework structure refinement strategy based on Si-O, O-O, and Si-Si distance restraints and analytical relationships between local structure and 29Si chemical shifts and geminal 2JSi-O-Si couplings. The refinement of the Sigma-2 structure showed that the Si-O and O-O distances were in excellent agreement with the single-crystal X-ray diffraction (SCXRD) data. The refinement of the ZSM-12 structure, initially determined from synchrotron powder XRD data, highlighted significant improvements in Si-O and O-O distances, and better agreement between calculated and experimental chemical shifts and J-couplings.

The open-source Python package, MRSimulator, is presented as a simple-to-use, fast, versatile, and extendable package capable of simulating one- and higher-dimensional Nuclear Magnetic Resonance (NMR) spectra under static, magic-angle, and variable-angle conditions. High benchmarks in spectral simulations are achieved by assuming that there are no degeneracies in the energy eigenstates, i.e., all dipolar couplings are in the weak limit and that there are no rotational resonances during evolution periods. Under these assumptions, the symmetry pathway formalism is exploited to reduce an NMR method applied to a spin system into a sum of individual transition pathways, whose signals are more efficiently calculated individually than as part of a full-density matrix simulation. To increase numerical efficiencies further, our approach restricts coherence transfer among transitions to pure rotations about an axis in the x-y plane of the rotating frame or through an artificial total mixing operation between selected transitions of adjacent free evolution periods. The assumptions used in this approach are valid for most commonly used solid-state NMR methods. Details of the implementation, along with example code usage, are given, including a least-squares spectral analysis.

NMR spectroscopy is an important analytical technique in synthetic organic chemistry, but its integration into high-throughput experimentation workflows has been limited by the necessity to manually analyze NMR spectra of new chemical entities. Current efforts to automate the analysis of NMR spectra rely on comparisons to databases of reported spectra for known compounds, and, therefore, are incompatible with the exploration of new chemical space. By reframing the NMR spectrum of a reaction mixture as a joint probability distribution, we have used Hamiltonian Monte Carlo Markov Chain (HMCMC) and density functional theory (DFT) to fit predicted NMR spectra to those of crude reaction mixtures. This approach enables the deconvolution and analysis of spectra of containing mixtures of compounds, without relying on reported spectra. The utility of our approach to analyze crude reaction mixtures is demonstrated with experimental spectra of reactions that generate a mixture of isomers, such as Wittig olefination and C--H functionalization reactions. The correct identification of compounds in a reaction mixture and their relative concentrations is achieved with mean absolute error as low as 1%.

The nuclear magnetic resonance (NMR) chemical shift tensor is a highly sensitive probe of the electronic structure of an atom and furthermore its local structure. Recently, machine learning has been applied to NMR in the prediction of isotropic chemical shifts from a structure. Current machine learning models, however, often ignore the full chemical shift tensor for the easier-to predict isotropic chemical shift, effectively ignoring a multitude of structural information available in the NMR chemical shift tensor. Here we use an equivariant graph neural network (GNN) to predict full 29Si chemical shift tensors in silicate materials. The equivariant GNN model predicts full tensors to a mean absolute error of 1.05 ppm and is able to accurately determine the magnitude, anisotropy, and tensor orientation in a diverse set of silicon oxide local structures. When compared with other models, the equivariant GNN model outperforms the state-of-the-art machine learning models by 53%. The equivariant GNN model also outperforms historic analytical models by 57% for isotropic chemical shift and 91% for anisotropy. The software is available as a simple-to-use open-source repository, allowing similar models to be created and trained with ease.

Hydrolysis reactions are ubiquitous in biological, environmental, and industrial chemistry. Density functional theory (DFT) is commonly employed to study the kinetics and reaction mechanisms of hydrolysis processes. Here, we present a new data set, Barrier Heights for HydrOlysis - 36 (BH2O-36), to enable the design of density functional approximations (DFAs) and the rational selection of DFAs for applications in aqueous chemistry. BH2O-36 consists of 36 diverse organic and inorganic forward and reverse hydrolysis reactions with reference energy barriers ΔE‡ calculated at the CCSD(T)/CBS level. Using BH2O-36, we evaluate 63 DFAs. In terms of mean absolute error (MAE) and mean relative absolute error (MRAE), ωB97M-V is the best-performing DFA tested, while MN12-L-D3(BJ) is the best-performing pure (nonhybrid) DFA. Broadly, we find that range-separated hybrid DFAs are necessary to approach chemical accuracy (0.043 eV). Although the best-performing DFAs include a dispersion correction to account for long-range interactions, we find that dispersion corrections do not generally improve MAE or MRAE for this data set.

The nuclear shielding anisotropy, zeta, is a useful nuclear magnetic resonance (NMR) shielding tensor parameter in describing the extent of electron cloud distortion about an atom. Despite the advantages afforded by NMR in structural characterization, the relationship between. and local structure of an atom in high-symmetry environments, such as Si-Q4 sites, is poorly understood. Here, we use a data-driven approach combining random forest feature ranking and the Sure Independence Screening and Sparsifying Operator (SISSO) approach to derive a simple and accurate geometric descriptor for zeta with a root-meansquared prediction error of 6.77 ppm and an R-2 of 0.761. We then apply this descriptor to describe the local geometric distortion of zeolites Sigma-2 and silica-ZSM-5 whose chemical shift anisotropy tensor has been reported. We envision that this geometric descriptor will allow for structural description and refinement in previously difficult-to-describe materials.

The facile synthesis of well-aligned, main-chain liquid crystalline elastomers that retain the cholesteric phase (CLCEs) is reported. The selective reflection inherent to this phase is thermally tuned more than 200 nm in these solid films, across the visible spectrum. The optical response is directly correlated to thermomechanical expansion of the CLCE film thickness. The bandwidth of the selective reflection of the CLCEs is increased to more than 200 nm by the incorporation of photosensitive chiral dopants that introduce heterogeneity in the pitch distribution. The mirror-like reflection of this CLCE film is also thermochromic, shifting from the visible to infrared. Reminiscent of cephalopods, when combined with the mechanical deformation of voxelated nematic liquid crystal elastomer, the thermochromic response of the CLCE produces solid-state elements with concurrent variation of specular and diffuse reflectance. These results demonstrate distinctive potential opportunities for liquid crystal elastomers to control light enabling new application in textiles, optics, and architecture.