Design of asymmetric catalysts generally involves time- and resource-intensive heuristic endeavors. In view of the steady increase in interest toward efficient catalytic asymmetric reactions and the rapid growth in the field of machine learning (ML) in recent years, we envisaged dovetailing these two important domains. We selected a set of quantum chemically derived molecular descriptors from five different asymmetric binaphthyl-derived catalyst families with the propensity to impact the enantioselectivity of asymmetric hydrogenation of alkenes and imines. The predictive power of the random forest (RF) built using the molecular parameters of a set of 368 substrate–catalyst combinations is found to be impressive, with a root-mean-square error (rmse) in the predicted enantiomeric excess (%ee) of about 8.4 ± 1.8 compared to the experimentally known values. The accuracy of RF is found to be superior to other ML methods such as convolutional neural network, decision tree, and eXtreme gradient boosting as well as stepwise linear regression. The proposed method is expected to provide a leap forward in the design of catalysts for asymmetric transformations.

Recently, nickel phosphides (Ni x P y ) have been reported to enable selective electrochemical formation of multicarbon products (C 3 and C 4 ) via the CO 2 reduction reaction (CO 2 RR); nevertheless, their activities remain low. In order to understand the roots of their high selectivity and low activity and to direct the design of more active Ni x P y -based CO 2 RR catalysts, we investigate the CO 2 RR mechanism on Ni 2 P using density functional theory (DFT) calculations. We reveal that the reaction proceeds through the formate pathway, followed by formaldehyde (H 2 CO*) formation and self-condensation. Moreover, we demonstrate that surface hydride transfer steps, along with surface-mediated C−C coupling, are essential in order to avoid C 1 product formation and boost selectivity toward multicarbon products. In addition, we find that the thermal surface hydride transfer from the surface to the physisorbed CO 2 is one of the key rate-limiting steps, and since it is not electroactive, it cannot be accelerated by applying an overpotential. Finally, our results also show that the hydrogen affinity of the surface and the dynamic surface reconstruction via H adsorption facilitate selective CO 2 reduction and C−C coupling on Ni 2 P. These findings provide an impetus for exploring materials design space to identify the physical principles that govern the thermodynamics of rate-limiting thermal steps in electrocatalytic processes.

Gaining predictable control over various forms of selectivities, such as enantio- and/or regio-selectivities, has been a long-standing goal in chemical catalysis. Although a number of factors such as the molecular features of the reactants and catalysts, as well as the reaction conditions, can influence the outcome of a reaction, it is not quite conspicuous as to what combinations of these parameters would offer a desired form of selectivity. We use machine learning tools, such as the neural network (NN), decision tree (DT), logistic regression (LR) and Random forest algorithms, to (a) analyze the outcome of an important catalytic regio-selective difluorination reaction of alkenes, and (b) decipher the complex interplay of various molecular parameters and their non-linear dependencies. The connection between what features of alkenes will yield 1,1-difluorination and how subtle changes would steer the reaction to 1,2-difluorination under identical conditions is enunciated. The NN was able to accurately predict whether a given alkene would yield a 1,1- or 1,2-difluorinated product. A combination of DT and the random forest classifier offered important chemical insights, which could be used in making a more rational choice of the reactant alkene for the desired regioisomeric product. The results could have far reaching implications in predicting which regioisomer is likely to be formed under a given set of conditions, and thus this technique is capable of expediting the development of catalytic transformations.

In this study, we propose that the curvature of graphene can be exploited to perform directional molecular motion and provide atomistic insights into the curvature-dependent molecular migration through density functional theory calculations. We first reveal the origin of the different migration trends observed experimentally for aromatic molecules with electron-donating and -withdrawing groups on p-doped functionalized graphene. Next, we show that the kinetic barrier for migration depends on the amount and nature of the curvature, that is, positive versus negative curvature. We find that the molecular migration on a wrinkled/rippled graphene sheet preferentially happens from the valley (positive curvature) to the mountain (negative curvature) regions. To understand the origin of such curvature-dependent molecular motion and migrational kinetic barrier trends, we develop a descriptor based on the frontier orbital orientation of graphene. Finally, based on these findings, we predict that time- and space-varying curvature can drive directional molecular motion on graphene and thus further propose that efforts should focus on exploring other two-dimensional materials as active platforms for performing controlled molecular motion.

The unique properties of hybrid organic–inorganic perovskites (HOIPs) promise to open doors to next-generation flexible optoelectronic devices. Before such advances are realized, a fundamental understanding of the mechanical properties of HOIPs is required. Here, we combine ab initio density functional theory (DFT) modeling with a diverse set of experiments to study the elastic properties of (quasi)­2D HOIPs. Specifically, we focus on (quasi)­2D single crystals of phenethylammonium methylammonium lead iodide, (PEA)2PbI4(MAPbI3) n−1, and their 3D counterpart, MAPbI3. We used nanoindentation (both Hertzian and Oliver–Pharr analyses) in combination with elastic buckling instability experiments to establish the out-of-plane and in-plane elastic moduli. The effect of Van der Waals (vdW) forces, different interlayer interactions, and finite temperature are combined with DFT calculations to accurately model the system. Our results reveal a nonmonotonic dependence of both the in-plane and out-of plane elastic moduli on the number of inorganic layers (n) rationalized by first-principles calculations. We discuss how the presence of defects in as-grown crystals and macroscopic interlayer deformations affect the mechanical response of (quasi)­2D HOIPs. Comparing the in- and out-of-plane experimental results with the theory reveals that perturbations to the covalent and ionic bonds (which hold a 2D layer together) is responsible for the relative out-of-plane stiffness of these materials. In contrast, we conjecture that the in-plane softness originates from macroscopic or mesoscopic motions between 2D layers during buckling experiments. Additionally, we learn how dispersion and π interactions in organic bilayers can have a determining role in the elastic response of the materials, especially in the out-of-plane direction. The understanding gained by comparing ab initio and experimental techniques paves the way for rational design of layered HOIPs with mechanical properties favorable for strain–intensive applications. Combined with filters for other favorable criteria, e.g., thermal or moisture stability, one can systematically screen viable (quasi)­2D HOIPs for a variety of flexible optoelectronic applications.