Machine learning methods are becoming integral to scientific inquiry in numerous disciplines. We demonstrated that machine learning can be used to predict the performance of a synthetic reaction in multidimensional chemical space using data obtained via high-throughput experimentation. We created scripts to compute and extract atomic, molecular, and vibrational descriptors for the components of a palladium-catalyzed Buchwald-Hartwig cross-coupling of aryl halides with 4-methylaniline in the presence of various potentially inhibitory additives. Using these descriptors as inputs and reaction yield as output, we showed that a random forest algorithm provides significantly improved predictive performance over linear regression analysis. The random forest model was also successfully applied to sparse training sets and out-of-sample prediction, suggesting its value in facilitating adoption of synthetic methodology.

Over the past 40 years, transition metal catalysis has enabled bond formation between aryl and olefinic (sp2) carbons in a selective and predictable manner with high functional group tolerance. Couplings involving alkyl (sp3) carbons have proven more challenging. Here, we demonstrate that the synergistic combination of photoredox catalysis and nickel catalysis provides an alternative cross-coupling paradigm, in which simple and readily available organic molecules can be systematically used as coupling partners. By using this photoredox-metal catalysis approach, we have achieved a direct decarboxylative sp3–sp2 cross-coupling of amino acids, as well as α-O– or phenyl-substituted carboxylic acids, with aryl halides. Moreover, this mode of catalysis can be applied to direct cross-coupling of Csp3–H in dimethylaniline with aryl halides via C–H functionalization.

Through fine-tuning of reagent and base structure, sulfonyl fluorides can efficiently fluorinate diverse classes of alcohols. We show that machine learning can map the intricate reaction landscape and enable accurate prediction of high-yielding conditions for untested substrates.

We demonstrate that the chemical-feature model described in our original paper is distinguishable from the nongeneralizable models introduced by Chuang and Keiser. Furthermore, the chemical-feature model significantly outperforms these models in out-of-sample predictions, justifying the use of chemical featurization from which machine learning models can extract meaningful patterns in the dataset, as originally described.

We describe the functionalization of α-amino C–H bonds with aryl halides using a combination of nickel and photoredox catalysis. This direct C–H, C–X coupling uses inexpensive and readily available starting materials to generate benzylic amines, an important class of bioactive molecules. Mechanistically, this method features the direct arylation of α-amino radicals mediated by a nickel catalyst. This reactivity is demonstrated for a range of aryl halides and N-aryl amines, with orthogonal scope to existing C–H activation and photoredox methodologies. We also report reactions with several complex aryl halides, demonstrating the potential utility of this approach in late-stage functionalization.