Reaction performance prediction with an extrapolative and interpretable graph model based on chemical knowledge.

Accurate prediction of reactivity and selectivity provides the desired guideline for synthetic development. Due to the high-dimensional relationship between molecular structure and synthetic function, it is challenging to achieve the predictive modelling of synthetic transformation with the required extrapolative ability and chemical interpretability. To meet the gap between the rich domain knowledge of chemistry and the advanced molecular graph model, herein we report a knowledge-based graph model that embeds the digitalized steric and electronic information. In addition, a molecular interaction module is developed to enable the learning of the synergistic influence of reaction components. In this study, we demonstrate that this knowledge-based graph model achieves excellent predictions of reaction yield and stereoselectivity, whose extrapolative ability is corroborated by additional scaffold-based data splittings and experimental verifications with new catalysts. Because of the embedding of local environment, the model allows the atomic level of interpretation of the steric and electronic influence on the overall synthetic performance, which serves as a useful guide for the molecular engineering towards the target synthetic function. This model offers an extrapolative and interpretable approach for reaction performance prediction, pointing out the importance of chemical knowledge-constrained reaction modelling for synthetic purpose.

Unsymmetric N-heterocyclic carbene ligand enabled nickel-catalysed arylation of bulky primary and secondary amines.

The arylation of sterically hindered amines represents one of the long-standing challenges in synthetic chemistry. Herein, we report a highly efficient Ni-catalysed arylation of sterically hindered primary and secondary amines with aryl chlorides or phenol derivatives enabled by an unsymmetric N-heterocyclic carbene (NHC) ligand. The protocol provides general, efficient, and scalable access to various sterically demanding anilines in excellent yields under mild conditions. A wide range of functional groups and heterocycles are compatible (>50 examples), including those present in biologically relevant molecules. Computational studies suggest that the unsymmetric bulky and flexible NHC ligand was critical to balance the oxidative addition and reductive elimination elementary steps, thus promoting this challenging transformation.

Data-driven design of new chiral carboxylic acid for construction of indoles with C-central and C-N axial chirality via cobalt catalysis.

Challenging enantio- and diastereoselective cobalt-catalyzed C-H alkylation has been realized by an innovative data-driven knowledge transfer strategy. Harnessing the statistics of a related transformation as the knowledge source, the designed machine learning (ML) model took advantage of delta learning and enabled accurate and extrapolative enantioselectivity predictions. Powered by the knowledge transfer model, the virtual screening of a broad scope of 360 chiral carboxylic acids led to the discovery of a new catalyst featuring an intriguing furyl moiety. Further experiments verified that the predicted chiral carboxylic acid can achieve excellent stereochemical control for the target C-H alkylation, which supported the expedient synthesis for a large library of substituted indoles with C-central and C-N axial chirality. The reported machine learning approach provides a powerful data engine to accelerate the discovery of molecular catalysis by harnessing the hidden value of the available structure-performance statistics.

An Asynchronous Concerted Mechanism and Its Origin in Lewis Acid-Mediated Carbonyl-Olefin [2+2] Cycloaddition.

The concerted mechanism of thermal BBr3 -mediated [2+2] carbonyl-olefin cycloaddition is intriguing considering the conflict against the Woodward-Hoffmann rule. In this work, we report a mechanistic study of the titled reaction using density functional theory calculations. The concerted [2+2] cycloaddition mechanism is operative even for the truncated model system of 2-methyl-2-butene and butanone. Intrinsic reaction coordinate analysis and potential surface mapping showed the asynchronous nature of the apparent concerted [2+2] cycloaddition, resulting from the proximity of carbocation-oxyanion in the transient zwitterionic structure. Solvent effects can change the reaction to a stepwise mechanism, highlighting the importance of zwitterion stability. Building upon this finding of the asynchronous concerted or stepwise [2+2] cycloaddition mechanism, two mechanistic classifications were elaborated based on substrates containing varied substituents and Lewis acids, with a focus on manipulating the stability of positive carbocation and negative oxyanion in the transient zwitterionic species.

Exploring Spectrum-based Molecular Descriptors for Reaction Performance Prediction.

Despite the availability and accuracy of modern spectroscopic characterization, the utilization of spectral information in chemical machine learning is still primitive. Here, we report an optical character recognition-based automatic process to utilize spectral information as molecular descriptors, which directly transforms experimental spectrum images to readable vectors. We demonstrate its machine learning application in the reaction yield dataset of Pd-catalyzed Buchwald-Hartwig cross-coupling with aryl halides. In addition, we also show that the predicted spectrum can serve as an alternative encoding source to support the model training.

Direct Incorporation of Dinitrogen into an Aliphatic C-H Bond.

The activation of dinitrogen (N2) and direct incorporation of its N atom into C-H bonds to create aliphatic C-N compounds remains unresolved. Incompatible conditions between dinitrogen reduction and C-H functionalization make this process extremely challenging. Herein, we report the first example of dinitrogen insertion into an aliphatic Csp3-H bond on the ligand scaffold of a 1,3-propane-bridged [N2N]2--type dititanium complex. Mechanistic investigations on the behaviors of dinuclear and mononuclear Ti complexes indicated the intramolecular synergistic effect of two Ti centers at a C-N bond-forming step. Computational studies revealed the critical isomerization between the inactive side-on N2 complex and the active nitridyl complex, which is responsible for the Csp3-H amination. This strategy maps an efficient route toward the future synthesis of aliphatic amines directly from N2.

Identification of Alkoxy Radicals as Hydrogen Atom Transfer Agents in Ce-Catalyzed C-H Functionalization.

The intermediacy of alkoxy radicals in cerium-catalyzed C-H functionalization via H-atom abstraction has been unambiguously confirmed. Catalytically relevant Ce(IV)-alkoxide complexes have been synthesized and characterized by X-ray diffraction. Operando electron paramagnetic resonance and transient absorption spectroscopy experiments on isolated pentachloro Ce(IV) alkoxides identified alkoxy radicals as the sole heteroatom-centered radical species generated via ligand-to-metal charge transfer (LMCT) excitation. Alkoxy-radical-mediated hydrogen atom transfer (HAT) has been verified via kinetic analysis, density functional theory (DFT) calculations, and reactions under strictly chloride-free conditions. These experimental findings unambiguously establish the critical role of alkoxy radicals in Ce-LMCT catalysis and definitively preclude the involvement of chlorine radical. This study has also reinforced the necessity of a high relative ratio of alcohol vs Ce for the selective alkoxy-radical-mediated HAT, as seemingly trivial changes in the relative ratio of alcohol vs Ce can lead to drastically different mechanistic pathways. Importantly, the previously proposed chlorine radical-alcohol complex, postulated to explain alkoxy-radical-enabled selectivities in this system, has been examined under scrutiny and ruled out by regioselectivity studies, transient absorption experiments, and high-level calculations. Moreover, the peculiar selectivity of alkoxy radical generation in the LMCT homolysis of Ce(IV) heteroleptic complexes has been analyzed and back-electron transfer (BET) may have regulated the efficiency and selectivity for the formation of ligand-centered radicals.

Bridging Chemical Knowledge and Machine Learning for Performance Prediction of Organic Synthesis.

Recent years have witnessed a boom of machine learning (ML) applications in chemistry, which reveals the potential of data-driven prediction of synthesis performance. Digitalization and ML modelling are the key strategies to fully exploit the unique potential within the synergistic interplay between experimental data and the robust prediction of performance and selectivity. A series of exciting studies have demonstrated the importance of chemical knowledge implementation in ML, which improves the model's capability for making predictions that are challenging and often go beyond the abilities of human beings. This Minireview summarizes the cutting-edge embedding techniques and model designs in synthetic performance prediction, elaborating how chemical knowledge can be incorporated into machine learning until June 2022. By merging organic synthesis tactics and chemical informatics, we hope this Review can provide a guide map and intrigue chemists to revisit the digitalization and computerization of organic chemistry principles.

(n-Bu)4NBr-Promoted N2 Splitting to Molybdenum Nitride.

Splitting of N2 via six-electron reduction and further functionalization to value-added products is one of the most important and challenging chemical transformations in N2 fixation. However, most N2 splitting approaches rely on strong chemical or electrochemical reduction to generate highly reactive metal species to bind and activate N2, which is often incompatible with functionalizing agents. Catalytic and sustainable N2 splitting to produce metal nitrides under mild conditions may create efficient and straightforward methods for N-containing organic compounds. Herein, we present that a readily available and nonredox (n-Bu)4NBr can promote N2-splitting with a Mo(III) platform. Both experimental and theoretical mechanistic studies suggest that simple X- (X = Br, Cl, etc.) anions could induce the disproportionation of MoIII[N(TMS)Ar]3 at the early stage of the catalysis to generate a catalytically active {MoII[N(TMS)Ar]3}- species. The quintet MoII species prove to be more favorable for N2 fixation kinetically and thermodynamically, compared with the quartet MoIII counterpart. Especially, computational studies reveal a distinct heterovalent {MoII-N2-MoIII} dimeric intermediate for the N≡N triple bond cleavage.

Towards Data-Driven Design of Asymmetric Hydrogenation of Olefins: Database and Hierarchical Learning.

Asymmetric hydrogenation of olefins is one of the most powerful asymmetric transformations in molecular synthesis. Although several privileged catalyst scaffolds are available, the catalyst development for asymmetric hydrogenation is still a time- and resource-consuming process due to the lack of predictive catalyst design strategy. Targeting the data-driven design of asymmetric catalysis, we herein report the development of a standardized database that contains the detailed information of over 12000 literature asymmetric hydrogenations of olefins. This database provides a valuable platform for the machine learning applications in asymmetric catalysis. Based on this database, we developed a hierarchical learning approach to achieve predictive machine leaning model using only dozens of enantioselectivity data with the target olefin, which offers a useful solution for the few-shot learning problem and will facilitate the reaction optimization with new olefin substrate in catalysis screening.

Mechanism and Selectivity Control in Ni- and Pd-Catalyzed Cross-Couplings Involving Carbon-Oxygen Bond Activation.

Transition-metal-catalyzed C-O bond activation provides a useful strategy for utilizing alcohol- and phenol-derived electrophiles in cross-coupling reactions, which has become a research field of active and growing interest in organic chemistry. The synergy between computation and experiment elucidated the mechanistic model and controlling factors of selectivities in these transformations, leading to advances in innovative C-O bond activation and functionalization methods.Toward the rational design of C-O bond activation, our collaborations with the Jarvo group bridged the mechanistic models of C(sp2)-O and C(sp3)-O bond activations. We found that the nickel catalyst cleaves the benzylic and allylic C(sp3)-O bonds via two general mechanisms: the stereoinvertive SN2 back-side attack model and the stereoretentive chelation-assisted model. These two models control the stereochemistry in a wide array of stereospecific Ni-catalyzed cross-coupling reactions with benzylic or allylic alcohol derivatives. Because of the catalyst distortion, the ligands can differentiate the competing stereospecific C(sp3)-O bond activations. The PCy3 ligand interacts with nickel mainly through σ-donation, and the Ni(PCy3) catalyst can undergo facile bending of the substrate-nickel-ligand angle, which favors the stereoretentive benzylic C-O bond activation. The N-heterocyclic carbene SIMes ligand has additional d(metal)-p(ligand) back-donation with nickel, which leads to an extra energy penalty for the same angle bending. This results in the preference of stereoinvertive benzylic C-O bond activation under Ni/SIMes catalysis. In addition to ligand control, a Lewis acid can increase the selectivity for stereoinvertive C(sp3)-O activation by stabilizing the SN2 back-side attack transition state. The oxygen leaving group complexes with the MgI2 Lewis acid in the stereoinvertive activation, leading to the exclusive stereoinvertive Kumada coupling of benzylic ethers. We also identified that the competing C(sp3)-O bond activation models have noticeable differences in charge separation. This leads to the solvent polarity control of the stereospecificity in C(sp3)-O activations. Low-polarity solvents favor the neutral stereoretentive C-O bond activation, while high-polarity solvents favor the zwitterionic stereoinvertive cleavage.In sharp contrast to the nickel catalysts, the C(sp2)-O bond activation under palladium catalysis mainly proceeds via the classic three-membered ring oxidative addition mechanism instead of the chelation-assisted mechanism. This is due to the lower oxophilicity of palladium, which disfavors the oxygen coordination in the chelation-assisted-type activation. The three-membered ring activation model selectively cleaves the weak C-O bond, resulting in the exclusive chemoselectivity of acyl C-O bond activation in Pd-catalyzed cross-coupling reactions with aryl carboxylic acid derivatives. This explains the overall acylation in the Pd-catalyzed Suzuki-Miyaura coupling with aryl esters. In collaboration with the Szostak group, we revealed that the three-membered ring model applies in the Pd-catalyzed C-O bond activation of carboxylic acid anhydride, which stimulated the development of a series of Pd-catalyzed decarbonylative functionalizations of aryl carboxylic acids.

Divergent rhodium-catalyzed electrochemical vinylic C-H annulation of acrylamides with alkynes.

α-Pyridones and α-pyrones are ubiquitous structural motifs found in natural products and biologically active small molecules. Here, we report an Rh-catalyzed electrochemical vinylic C-H annulation of acrylamides with alkynes, affording cyclic products in good to excellent yield. Divergent syntheses of α-pyridones and cyclic imidates are accomplished by employing N-phenyl acrylamides and N-tosyl acrylamides as substrates, respectively. Additionally, excellent regioselectivities are achieved when using unsymmetrical alkynes. This electrochemical process is environmentally benign compared to traditional transition metal-catalyzed C-H annulations because it avoids the use of stoichiometric metal oxidants. DFT calculations elucidated the reaction mechanism and origins of substituent-controlled chemoselectivity. The sequential C-H activation and alkyne insertion under rhodium catalysis leads to the seven-membered ring vinyl-rhodium intermediate. This intermediate undergoes either the classic neutral concerted reductive elimination to produce α-pyridones, or the ionic stepwise pathway to produce cyclic imidates.

Understanding the mechanism and reactivity of Pd-catalyzed C-P bond metathesis of aryl phosphines: a computational study.

Transition metal-catalyzed single bond metathesis has recently emerged as a useful strategy for functional group transfer. In this work, we explored the mechanism and reactivity profile of Pd/PhI-cocatalyzed C-P bond metathesis between aryl phosphines using density functional theory (DFT) calculations. The overall single bond metathesis involves two Pd(ii)-catalyzed C-P reductive eliminations and two Pd(0)-catalyzed C-P oxidative additions, which allows the reversible C-P bond cleavage and formation of the phosphonium cation. Distortion/interaction analysis indicates that the facile C-P bond cleavage and formation of the phosphonium cation are due to the involvement of coordinating aryl phosphine in the process. In addition, the substituent effects on the reaction kinetics and thermodynamics of metathesis were computed, which provides helpful mechanistic information for the design of related single bond metathesis reactions.

Catalytic and Photochemical Strategies to Stabilized Radicals Based on Anomeric Nucleophiles.

Carbohydrates, one of the three primary macromolecules of living organisms, play significant roles in various biological processes such as intercellular communication, cell recognition, and immune activity. While the majority of established methods for the installation of carbohydrates through the anomeric carbon rely on nucleophilic displacement, anomeric radicals represent an attractive alternative because of their functional group compatibility and high anomeric selectivities. Herein, we demonstrate that anomeric nucleophiles such as C1 stannanes can be converted into anomeric radicals by merging Cu(I) catalysis with blue light irradiation to achieve highly stereoselective C(sp3)-S cross-coupling reactions. Mechanistic studies and DFT calculations revealed that the C-S bond-forming step occurs via the transfer of the anomeric radical directly to a sulfur electrophile bound to Cu(II) species. This pathway complements a radical chain observed for photochemical metal-free conditions where a disulfide initiator can be activated by a Lewis base additive. Both strategies utilize anomeric nucleophiles as efficient radical donors and achieve a switch from an ionic to a radical pathway. Taken together, the stability of glycosyl nucleophiles, a broad substrate scope, and high anomeric selectivities observed for the thermal and photochemical protocols make this novel C-S cross coupling a practical tool for late-stage glycodiversification of bioactive natural products and drug candidates.

Predicting Regioselectivity in Radical C-H Functionalization of Heterocycles through Machine Learning.

Radical C-H bond functionalization provides a versatile approach for elaborating heterocyclic compounds. The synthetic design of this transformation relies heavily on the knowledge of regioselectivity, while a quantified and efficient regioselectivity prediction approach is still elusive. Herein, we report the feasibility of using a machine learning model to predict the transition state barrier from the computed properties of isolated reactants. This enables rapid and reliable regioselectivity prediction for radical C-H bond functionalization of heterocycles. The Random Forest model with physical organic features achieved 94.2 % site accuracy and 89.9 % selectivity accuracy in the out-of-sample test set. The prediction performance was further validated by comparing the machine learning results with additional substituents, heteroarene scaffolds and experimental observations. This work revealed that the combination of mechanism-based computational statistics and machine learning model can serve as a useful strategy for selectivity prediction of organic transformations.

Computation-Guided Development of the "Click" ortho-Quinone Methide Cycloaddition with Improved Kinetics.

We report here a deep mechanistic study of the "click" ortho-quinone methide (oQM) cycloaddition between ortho-quinolinone quinone methide (oQQM) and thio-vinyl ether (TV), named as TQ-ligation. DFT calculations revealed the unexpected fact that dehydration of oQQM precursors is the rate-determining step of this transformation, and two highly reactive oQQM precursors were predicted. Guided by the calculations, a new "click" oQM cycloaddition which shows significantly improved kinetics and remarkable efficiency on protein labeling was developed.

Carboxylate breaks the arene C-H bond via a hydrogen-atom-transfer mechanism in electrochemical cobalt catalysis.

Combined computational and experimental studies elucidated the distinctive mechanistic features of electrochemical cobalt-catalyzed C-H oxygenation. A sequential electrochemical-chemical (EC) process was identified for the formation of an amidylcobalt(iii) intermediate. The synthesis, characterization, cyclic voltammetry studies, and stoichiometric reactions of the related amidylcobalt(iii) intermediate suggested that a second on-cycle electro-oxidation occurs on the amidylcobalt(iii) species, which leads to a formal Co(iv) intermediate. This amidylcobalt(iv) intermediate is essentially a cobalt(iii) complex with one additional single electron distributed on the coordinating heteroatoms. The radical nature of the coordinating pivalate allows the formal Co(iv) intermediate to undergo a novel carboxylate-assisted HAT mechanism to cleave the arene C-H bond, and a CMD mechanism could be excluded for a Co(iii/i) catalytic scenario. The mechanistic understanding of electrochemical cobalt-catalyzed C-H bond activation highlights the multi-tasking electro-oxidation and the underexplored reaction channels in electrochemical transition metal catalysis.

Divergent pathway and reactivity control of intramolecular arene C-H vinylation by vinyl cations.

Vinyl cations exhibit remarkable reactivity towards arene C-H functionalizations. This computational study revealed the key mechanistic details of intramolecular C-H vinylation through a vinyl cation intermediate. Based on the reaction mechanism, the effects of substitution, ring strain and tether length on the reactivity of the vinyl cation were elucidated.

Computational studies on Ni-catalyzed amide C-N bond activation.

Recently, remarkable advances in the mechanistic understanding of Ni-catalyzed amide C-N bond activation have been achieved by computational chemists. These computational studies revealed the key mechanistic processes of Ni-catalyzed amide C-N bond cleavage, providing the molecular basis for the rationalization of reactivities and selectivities. This review summarizes the general mechanistic models of Ni-catalyzed amide C-N bond activation, and discusses their applications in the understanding and design of Ni-catalyzed cross coupling reactions involving amides.

Rhodium(III)-Catalyzed Asymmetric Borylative Cyclization of Cyclohexadienone-Containing 1,6-Dienes: An Experimental and DFT Study.

Because of the inherent difficulty in differentiating two olefins, the development of metal-catalyzed asymmetric cyclization of 1,6-dienes remains challenging. Herein, we describe the first rhodium(III)-catalyzed asymmetric borylative cyclization of cyclohexadienone-tethered mono-, 1,1-di-, and (E)-1,2-disubstituted alkenes (1,6-dienes), affording optically pure cis-bicyclic skeletons bearing three or four contiguous stereocenters with high yields (25-93%), and excellent diastereoselectivities (>20:1 dr) and enantioselectivities (90-99% ee). This mild catalytic approach is generally compatible with a wide range of functional groups, which allows several facile conversions of the cyclization products. Furthermore, on the basis of our SAESI-MS experiment and computational study, a Rh(I)/(III) catalytic cycle is proposed in this tandem reaction, and the Rh(I) active species catalyzes the overall transformation via sequential oxidative addition of B2pin2, olefin insertion, cyclizing conjugate addition, and reductive elimination. The irreversible conjugate addition determines the overall regioselectivity of borylative cyclization, and the ring strain favors the formation of 5,6-bicyclic structure. This highlights the control of ring strain in diene cyclizations, which provides a useful basis for future reaction designs.