Parametrization of Non-covalent Interactions for Transition State Interrogation Applied to Asymmetric Catalysis

Catalysing (organo-)catalysis: Trends in the application of machine learning to enantioselective organocatalysis

Organocatalysis has established itself as a third pillar of homogeneous catalysis, besides transition metal catalysis and biocatalysis, as its use for enantioselective reactions has gathered significant interest over the last decades. Concurrent to this development, machine learning (ML) has been increasingly applied in the chemical domain to efficiently uncover hidden patterns in data and accelerate scientific discovery. While the uptake of ML in organocatalysis has been comparably slow, the last two decades have showed an increased interest from the community. This review gives an overview of the work in the field of ML in organocatalysis. The review starts by giving a short primer on ML for experimental chemists, before discussing its application for predicting the selectivity of organocatalytic transformations. Subsequently, we review ML employed for privileged catalysts, before focusing on its application for catalyst and reaction design. Concluding, we give our view on current challenges and future directions for this field, drawing inspiration from the application of ML to other scientific domains.

Highly acidic N-triflylphosphoramides as chiral Brønsted acid catalysts: the effect of weak hydrogen bonds and multiple acceptors on complex structures and aggregation

N-Triflylphosphoramides (NTPAs) represent an important catalyst class in asymmetric catalysis due to their multiple hydrogen bond acceptor sites and acidity, which is increased by several orders of magnitude compared to conventional chiral phosphoric acids (CPAs). Thus, NTPAs allow for several challenging transformations, which are not accessible with CPAs. However, detailed evidence on their hydrogen bonding situation, complex structures and aggregation is still lacking. Therefore, this study covers the hydrogen bonding behavior and structural features of binary NTPA/imine complexes compared to their CPA counterparts. Deviating from the single-well potential hydrogen bonds commonly observed in CPA/imine complexes, the NTPA/imine complexes exhibit a tautomeric equilibrium between two proton positions. Low-temperature NMR at 180 K supported by computer simulations indicates a OHN hydrogen bond between the phosphoramide oxygen and the imine, instead of the mostly proposed NHN H-bond. Furthermore, this study finds no evidence for the existence of dimeric NTPA/NTPA/imine complexes as previously suggested for CPA systems, both synthetically and through NMR studies.

Iridium Acylnitrenoid-Initiated Biomimetic Cascade Cyclizations: Stereodefined Access to Polycyclic δ-Lactams

Ring-fused azacyclic compounds are important building units in the synthesis of biorelevant natural products, pharmaceutical agents, and molecular materials. Herein, we present a new approach to these condensed azacycles by a biomimetic cascade cyclization of arylalkenyl dioxazolones. This cascade reaction was found to proceed with excellent stereoselectivity and a high functional group tolerance. The substrate scope of arylalkenyl dioxazolones turned out to be highly flexible and extendable to additional terminating subunits, such as heteroaryl and alkynyl moieties. This biomimetic cyclization was elucidated to be initiated by an intramolecular transfer of the in situ generated electrophilic Ir-acylnitrenoid to the tethered olefinic double bond, leading to a key N-acylaziridine intermediate, which is in turn reacted with pendant (hetero)arenes or alkynes in a highly regio- and stereoselective manner to produce ring-fused azacyclic compounds.

Molecular Machine Learning for Chemical Catalysis: Prospects and Challenges

ConspectusIn the domain of reaction development, one aims to obtain higher efficacies as measured in terms of yield and/or selectivities. During the empirical cycles, an admixture of outcomes from low to high yields/selectivities is expected. While it is not easy to identify all of the factors that might impact the reaction efficiency, complex and nonlinear dependence on the nature of reactants, catalysts, solvents, etc. is quite likely. Developmental stages of newer reactions would typically offer a few hundreds of samples with variations in participating molecules and/or reaction conditions. These "observations" and their "output" can be harnessed as valuable labeled data for developing molecular machine learning (ML) models. Once a robust ML model is built for a specific reaction under development, it can predict the reaction outcome for any new choice of substrates/catalyst in a few seconds/minutes and thus can expedite the identification of promising candidates for experimental validation. Recent years have witnessed impressive applications of ML in the molecular world, most of them aimed at predicting important chemical or biological properties. We believe that an integration of effective ML workflows can be made richly beneficial to reaction discovery.As with any new technology, direct adaptation of ML as used in well-developed domains, such as natural language processing (NLP) and image recognition, is unlikely to succeed in reaction discovery. Some of the challenges stem from ineffective featurization of the molecular space, unavailability of quality data and its distribution, in making the right choice of ML model and its technically robust deployment. It shall be noted that there is no universal ML model suitable for an inherently high-dimensional problem such as chemical reactions. Given these backgrounds, rendering ML tools conducive for reactions is an exciting as well as challenging endeavor at the same time. With the increased availability of efficient ML algorithms, we focused on tapping their potential for small-data reaction discovery (a few hundreds to thousands of samples).In this Account, we describe both feature engineering and feature learning approaches for molecular ML as applied to diverse reactions of high contemporary interest. Among these, catalytic asymmetric hydrogenation of imines/alkenes, β-C(sp³)-H bond functionalization, and relay Heck reaction employed a feature engineering approach using the quantum-chemically derived physical organic descriptors as the molecular features─all designed to predict the enantioselectivity. The selection of molecular features to customize it for a reaction of interest is described, along with emphasizing the chemical insights that could be gathered through the use of such features. Feature learning methods for predicting the yield of Buchwald-Hartwig cross-coupling, deoxyfluorination of alcohols, and enantioselectivity of N,S-acetal formation are found to offer excellent predictions. We propose a transfer learning protocol, wherein an ML model such as a language model is trained on a large number of molecules (10⁵-10⁶) and fine-tuned on a focused library of target task reactions, as an effective alternative for small-data reaction discovery (10²-10³ reactions). The exploitation of deep neural network latent space as a method for generative tasks to identify useful substrates for a reaction is demonstrated as a promising strategy.

Catalytic asymmetric C-H insertion reactions of vinyl carbocations

From the preparation of pharmaceuticals to enzymatic construction of natural products, carbocations are central to molecular synthesis. Although these reactive intermediates are engaged in stereoselective processes in nature, exerting enantiocontrol over carbocations with synthetic catalysts remains challenging. Many resonance-stabilized tricoordinated carbocations, such as iminium and oxocarbenium ions, have been applied in catalytic enantioselective reactions. However, their dicoordinated counterparts (aryl and vinyl carbocations) have not, despite their emerging utility in chemical synthesis. We report the discovery of a highly enantioselective vinyl carbocation carbon-hydrogen (C-H) insertion reaction enabled by imidodiphosphorimidate organocatalysts. Active site confinement featured in this catalyst class not only enables effective enantiocontrol but also expands the scope of vinyl cation C-H insertion chemistry, which broadens the utility of this transition metal-free C(sp³)-H functionalization platform.

Recognition in the Domain of Molecular Chirality: From Noncovalent Interactions to Separation of Enantiomers

It is not a coincidence that both chirality and noncovalent interactions are ubiquitous in nature and synthetic molecular systems. Noncovalent interactivity between chiral molecules underlies enantioselective recognition as a fundamental phenomenon regulating life and human activities. Thus, noncovalent interactions represent the narrative thread of a fascinating story which goes across several disciplines of medical, chemical, physical, biological, and other natural sciences. This review has been conceived with the awareness that a modern attitude toward molecular chirality and its consequences needs to be founded on multidisciplinary approaches to disclose the molecular basis of essential enantioselective phenomena in the domain of chemical, physical, and life sciences. With the primary aim of discussing this topic in an integrated way, a comprehensive pool of rational and systematic multidisciplinary information is provided, which concerns the fundamentals of chirality, a description of noncovalent interactions, and their implications in enantioselective processes occurring in different contexts. A specific focus is devoted to enantioselection in chromatography and electromigration techniques because of their unique feature as "multistep" processes. A second motivation for writing this review is to make a clear statement about the state of the art, the tools we have at our disposal, and what is still missing to fully understand the mechanisms underlying enantioselective recognition.

Molecular field analysis for data-driven molecular design in asymmetric catalysis

This review highlights the recent advances (2019-present) in the use of MFA (molecular field analysis) for data-driven catalyst design, enabling to improve selectivities/reaction outcomes in asymmetric catalysis. Successful examples of MFA-based molecular design and how to design molecules by MFA are described, including how to generate and evaluate MFA-based regression models, and future challenges in MFA-based molecular design in molecular catalysis.

Assessing the persistence of chalcogen bonds in solution with neural network potentials

Non-covalent bonding patterns are commonly harvested as a design principle in the field of catalysis, supramolecular chemistry and functional materials to name a few. Yet, their computational description generally neglects finite temperature and environment effects, which promote competing interactions and alter their static gas-phase properties. Recently, neural network potentials (NNPs) trained on Density Functional Theory (DFT) data have become increasingly popular to simulate molecular phenomena in condensed phase with an accuracy comparable to ab initio methods. To date, most applications have centered on solid-state materials or fairly simple molecules made of a limited number of elements. Herein, we focus on the persistence and strength of chalcogen bonds involving a benzotelluradiazole in condensed phase. While the tellurium-containing heteroaromatic molecules are known to exhibit pronounced interactions with anions and lone pairs of different atoms, the relevance of competing intermolecular interactions, notably with the solvent, is complicated to monitor experimentally but also challenging to model at an accurate electronic structure level. Here, we train direct and baselined NNPs to reproduce hybrid DFT energies and forces in order to identify what are the most prevalent non-covalent interactions occurring in a solute-Cl$^-$-THF mixture. The simulations in explicit solvent highlight competition with chalcogen bonds formed with the solvent and the short-range directionality of the interaction with direct consequences for the molecular properties in the solution. The comparison with other potentials (e.g., AMOEBA, direct NNP and continuum solvent model) also demonstrates that baselined NNPs offer a reliable picture of the non-covalent interaction interplay occurring in solution.

H-Bonded Counterion-Directed Enantioselective Au(I) Catalysis

A new strategy for enantioselective transition-metal catalysis is presented, wherein a H-bond donor placed on the ligand of a cationic complex allows precise positioning of the chiral counteranion responsible for asymmetric induction. The successful implementation of this paradigm is demonstrated in 5-exo-dig and 6-endo-dig cyclizations of 1,6-enynes, combining an achiral phosphinourea Au(I) chloride complex with a BINOL-derived phosphoramidate Ag(I) salt and thus allowing the first general use of chiral anions in Au(I)-catalyzed reactions of challenging alkyne substrates. Experiments with modified complexes and anions, ¹H NMR titrations, kinetic data, and studies of solvent and nonlinear effects substantiate the key H-bonding interaction at the heart of the catalytic system. This conceptually novel approach, which lies at the intersection of metal catalysis, H-bond organocatalysis, and asymmetric counterion-directed catalysis, provides a blueprint for the development of supramolecularly assembled chiral ligands for metal complexes.

Harvesting the fragment-based nature of bifunctional organocatalysts to enhance their activity

Enhancing the activity of bifunctional organocatalysts: a fragment-based approach coupled with activity maps helps identifying better-performing catalytic motifs.

A DFT Study on Gold-Catalyzed Domino Cyclization for Post-Ugi Synthesis of Spiroindolines: Insights on the Origin of Remarkable Diastereoselectivity

We report a comprehensive DFT study on gold-catalyzed domino cyclization to spiroindolines. The diastereoselectivity was analyzed based on the established coordination spheres. These computational results not only explain the origin...

Data Science Meets Physical Organic Chemistry

ConspectusAt the heart of synthetic chemistry is the holy grail of predictable catalyst design. In particular, researchers involved in reaction development in asymmetric catalysis have pursued a variety of strategies toward this goal. This is driven by both the pragmatic need to achieve high selectivities and the inability to readily identify why a certain catalyst is effective for a given reaction. While empiricism and intuition have dominated the field of asymmetric catalysis since its inception, enantioselectivity offers a mechanistically rich platform to interrogate catalyst-structure response patterns that explain the performance of a particular catalyst or substrate.In the early stages of an asymmetric reaction development campaign, the overarching mechanism of the reaction, catalyst speciation, the turnover limiting step, and many other details are unknown or posited based on related reactions. Considering the unclear details leading to a successful reaction, initial enantioselectivity data are often used to intuitively guide the ultimate direction of optimization. However, if the conditions of the Curtin-Hammett principle are satisfied, then measured enantioselectivity can be directly connected to the ensemble of diastereomeric transition states (TSs) that lead to the enantiomeric products, and the associated free energy difference between competing TSs (ΔΔG⧧ = -RT ln[(S)/(R)], where (S) and (R) represent the concentrations of the enantiomeric products). We, and others, speculated that this important piece of information can be leveraged to guide reaction optimization in a quantitative way.Although traditional linear free energy relationships (LFERs), such as Hammett plots, have been used to illuminate important mechanistic features, we sought to develop data science derived tools to expand the power of LFERs in order to describe complex reactions frequently encountered in modern asymmetric catalysis. Specifically, we investigated whether enantioselectivity data from a reaction can be quantitatively connected to the attributes of reaction components, such as catalyst and substrate structural features, to harness data for asymmetric catalyst design.In this context, we developed a workflow to relate computationally derived features of reaction components to enantioselectivity using data science tools. The mathematical representation of molecules can incorporate many aspects of a transformation, such as molecular features from substrate, product, catalyst, and proposed transition states. Statistical models relating these features to reaction outputs can be used for various tasks, such as performance prediction of untested molecules. Perhaps most importantly, statistical models can guide the generation of mechanistic hypotheses that are embedded within complex patterns of reaction responses. Overall, merging traditional physical organic experiments with statistical modeling techniques creates a feedback loop that enables both evaluation of multiple mechanistic hypotheses and future catalyst design. In this Account, we highlight the evolution and application of this approach in the context of a collaborative program based on chiral phosphoric acid catalysts (CPAs) in asymmetric catalysis.

Enantioselective α-Arylation of Ketones via a Novel Cu(I)-Bis(phosphine) Dioxide Catalytic System

A novel catalytic system based on copper(I) and chiral bis(phosphine) dioxides is described. This allows the arylation of silyl enol ethers to access enolizable α-arylated ketones in good yields and enantiomeric excess up to 95%. Noncyclic ketones are amenable substrates with this method, which complements other approaches based on palladium catalysis. Optimization of the ligand structure is accomplished via rational design driven by correlation analysis. Preliminary mechanistic hypotheses are also evaluated in order to identify the role of chiral bis(phosphine) dioxides.

Nucleophilicity Prediction via Multivariate Linear Regression Analysis

The concept of nucleophilicity is at the basis of most transformations in chemistry. Understanding and predicting the relative reactivity of different nucleophiles is therefore of paramount importance. Mayr’s nucleophilicity scale likely represents the most complete collection of reactivity data, which currently includes over 1200 nucleophiles. Several attempts have been made to theoretically predict Mayr’s nucleophilicity parameters N based on calculation of molecular properties, but a general model accounting for different classes of nucleophiles could not be obtained so far. We herein show that multivariate linear regression analysis is a suitable tool for obtaining a simple model predicting N for virtually any class of nucleophiles in different solvents for a set of 341 data points. The key descriptors of the model were found to account for the proton affinity, solvation energies, and sterics.

John PC, Kim S, Paton RS. Importance of Engineered and Learned Molecular Representations in Predicting Organic Reactivity, Selectivity, and Chemical Properties

Machine-readable chemical structure representations are foundational in all attempts to harness machine learning for the prediction of reactivities, selectivities, and chemical properties directly from molecular structure. The featurization of discrete chemical structures into a continuous vector space is a critical phase undertaken before model selection, and the development of new ways to quantitatively encode molecules is an active area of research. In this Account, we highlight the application and suitability of different representations, from expert-guided "engineered" descriptors to automatically "learned" features, in different prediction tasks relevant to organic and organometallic chemistry, where differing amounts of training data are available. These tasks include statistical models of stereo- and enantioselectivity, thermochemistry, and kinetics developed using experimental and quantum chemical data.The use of expert-guided molecular descriptors provides an opportunity to incorporate chemical knowledge, domain expertise, and physical constraints into statistical modeling. In applications to stereoselective organic and organometallic catalysis, where data sets may be relatively small and 3D-geometries and conformations play an important role, mechanistically informed features can be used successfully to obtain predictive statistical models that are also chemically interpretable. We provide an overview of several recent applications of this approach to obtain quantitative models for reactivity and selectivity, where topological descriptors, quantum mechanical calculations of electronic and steric properties, along with conformational ensembles, all feature as essential ingredients of the molecular representations used.Alternatively, more flexible, general-purpose molecular representations such as attributed molecular graphs can be used with machine learning approaches to learn the complex relationship between a structure and prediction target. This approach has the potential to out-perform more traditional representation methods such as "hand-crafted" molecular descriptors, particularly as data set sizes grow. One area where this is particularly relevant is in the use of large sets of quantum mechanical data to train quantitative structure-property relationships. A general approach toward curating useful data sets and training highly accurate graph neural network models is discussed in the context of organic bond dissociation enthalpies, where this strategy outperforms regression using precomputed descriptors.Finally, we describe how graph neural network predictions can be incorporated into mechanistically informed statistical models of chemical reactivity and selectivity. Once trained, this approach avoids the expensive computational overhead associated with quantum mechanical calculations, while maintaining chemical interpretability. We illustrate examples for which fast predictions of bond dissociation enthalpy and of the identities of radicals formed through cleavage of a molecule's weakest bond are used in simple physical models of site-selectivity and reactivity.

The origin of the regiospecificity of acrolein dimerization

Acrolein dimerization is a intriguing case since the reaction does not occur to form the electronically preferred regioisomeric adduct. Various explanations have been suggested to rationalize this experimental regioselectivity, however, none of these arguments had been convincing enough. In this work, the hetero Diels-Alder acrolein dimerization was theoretically investigated using DFT and MP2 methods. The influence of nucleophilic/electrophilic interactions and non-covalent interactions (NCI) in the regiospecificity of the reaction were analyzed. Our results show that the NCI at the transition state are the key factor controlling the regiospecificity in this reaction. Besides, we found that the choice of calculation method can have an effect on the prediction of the mechanism in the reaction, as all DFT methods forecast a one-step hetero Diels-Alder acrolein dimerization, while MP2 predicts a stepwise description for the lower energy reaction channel.

Bifunctional squaramides with benzyl-like fragments: analysis of CH⋯π interactions by a multivariate linear regression model and quantum chemical topology

A multivariate linear regression model and quantum chemical topology are used for the quantitative description of non-covalent interactions in the transition state of the Michael addition catalyzed by bifunctional squaramides.

Catalytic Enantioselective Synthesis of Difluorinated Alkyl Bromides

We report an iodoarene-catalyzed enantioselective synthesis of β,β-difluoroalkyl bromide building blocks. The transformation involves an oxidative rearrangement of α-bromostyrenes, utilizing HF-pyridine as the fluoride source and m-CPBA as the stoichiometric oxidant. A catalyst decomposition pathway was identified, which, in tandem with catalyst structure-activity relationship studies, facilitated the development of an improved catalyst providing higher enantioselectivity with lower catalyst loadings. The versatility of the difluoroalkyl bromide products was demonstrated via highly enantiospecific substitution reactions with suitably reactive nucleophiles. The origins of enantioselectivity were investigated using computed interaction energies of simplified catalyst and substrate structures, providing evidence for both CH-π and π-π transition state interactions as critical features.

Enantioselective Intramolecular Allylic Substitution via Synergistic Palladium/Chiral Phosphoric Acid Catalysis: Insight into Stereoinduction through Statistical Modeling

The mode of asymmetric induction in an enantioselective intramolecular allylic substitution reaction catalyzed by a combination of palladium and a chiral phosphoric acid was investigated by a combined experimental and statistical modeling approach. Experiments to probe nonlinear effects, the reactivity of deuterium-labeled substrates, and control experiments revealed that nucleophilic attack to the π-allylpalladium intermediate is the enantio-determining step, in which the chiral phosphate anion is involved in stereoinduction. Using multivariable linear regression analysis, we determined that multiple noncovalent interactions with the chiral environment of the phosphate anion are integral to enantiocontrol in the transition state. The synthetic protocol to form chiral pyrrolidines was further applied to the asymmetric construction of C-O bonds at fully substituted carbon centers in the synthesis of chiral 2,2-disubstituted benzomorpholines.

Molecular-level insight in supported olefin metathesis catalysts by combining surface organometallic chemistry, high throughput experimentation, and data analysis

A combination of high-throughput experimentation (HTE), surface organometallic chemistry (SOMC) and statistical data analysis provided the platform to analyze in situ silica-grafted Mo imido alkylidene catalysts based on a library of 35 phenols. Overall, these tools allowed for the identification of σ-donor electronic effects and dispersive interactions and as key drivers in a prototypical metathesis reaction, homodimerization of 1-nonene. Univariate and multivariate correlation analysis confirmed the categorization of the catalytic data into two groups, depending on the presence of aryl groups in ortho position of the phenol ligand. The initial activity (TOFin) was predominantly correlated to the σ-donor ability of the aryloxy ligands, while the overall catalytic performance (TON1 h) was mainly dependent on attractive dispersive interactions with the used phenol ligands featuring aryl ortho substituents and, in sharp contrast, repulsive dispersive interactions with phenol free of aryl ortho substituents. This work outlines a fast and efficient workflow of gaining molecular-level insight into supported metathesis catalysts and highlights σ-donor ability and noncovalent interactions as crucial properties for designing active d0 supported metathesis catalysts.

Basic principles of substrate activation through non-covalent bond interactions

In the last twenty years, chiral Brønsted acid and chiral counteranion catalysis have emerged as a fundamental area of organocatalysis. The development of chiral acidic catalysts has allowed extending many known Brønsted catalyzed reactions to the stereoselective domain. Moreover, the controlled conditions under which these catalysts can be used, allowed accessing reactivity of increasing complexity with extraordinary selectivity levels. However, compared to the explosion of this branch of organocatalysis in an applicative direction, only little has been done to understand and rationalize the observed reaction outcomes. This is due, in part, to the complex nature of the weak interactions (H-bonds, electrostatic, and dispersion interactions) governing this class of reactions. Here we review relevant mechanistic analyses from both chiral Brønsted acid and chiral counteranion directed catalysis. Both experimental and computational work is included that aimed at unveiling the nature of the interactions governing the a number of reactions. These include the: enantioselective reduction of ketoimines with Hantzsch esters; ring opening reactions of epoxides, oxetanes, aziridinium, and sulfonium ions; stereoselective fluorination of allylic alcohols; oxidative aminations of benzylic thioethers (enantioselective Pummerer reaction). These case studies are analyzed and discussed in order to highlight key features and similarities across the different catalytic systems.

Enantioselective Allenoate-Claisen Rearrangement Using Chiral Phosphate Catalysts

Herein we report the first highly enantioselective allenoate-Claisen rearrangement using doubly axially chiral phosphate sodium salts as catalysts. This synthetic method provides access to β-amino acid derivatives with vicinal stereocenters in up to 95% ee. We also investigated the mechanism of enantioinduction by transition state (TS) computations with DFT as well as statistical modeling of the relationship between selectivity and the molecular features of both the catalyst and substrate. The mutual interactions of charge-separated regions in both the zwitterionic intermediate generated by reaction of an amine to the allenoate and the Na⁺-salt of the chiral phosphate leads to an orientation of the TS in the catalytic pocket that maximizes favorable noncovalent interactions. Crucial arene-arene interactions at the periphery of the catalyst lead to a differentiation of the TS diastereomers. These interactions were interrogated using DFT calculations and validated through statistical modeling of parameters describing noncovalent interactions.

Strategies for remote enantiocontrol in chiral gold(iii) complexes applied to catalytic enantioselective γ,δ-Diels-Alder reactions

The use of chiral square planar gold(iii) complexes to access enantioenriched products has rarely been applied in asymmetric catalysis. In this context, we report a mechanistic and synthetic investigation into the use of N-heterocyclic (NHC) gold(iii) complexes in γ,δ-Diels–Alder reactions of 2,4-dienals with cyclopentadiene. The optimal catalyst bearing a unique 2-chloro-1-naphthyl substituent allowed efficient synthesis of functionally rich carbocycles in good yields, diastereo- and enantioselectivities. Transition state and multivariate linear regression (MLR) analysis of both catalyst and substrate trends using molecular descriptors derived from designer parameter acquisition platforms, reveals attractive non-covalent interactions (NCIs) to be key selectivity determinates. These analyses demonstrate that a putative π–π interaction between the substrate proximal double bond and the catalyst aromatic group is an essential feature for high enantioselectivity.

Chiral square planar gold(iii) complexes are employed as catalysts in asymmetric Diels–Alder reactions. The non-covalent interactions responsible for the enantioselectivity are revealed through multivariate linear regression analysis.

What Does the Machine Learn? Knowledge Representations of Chemical Reactivity

In a departure from conventional chemical approaches, data-driven models of chemical reactions have recently been shown to be statistically successful using machine learning. These models, however, are largely black box in character and have not provided the kind of chemical insights that historically advanced the field of chemistry. To examine the knowledgebase of machine-learning models-what does the machine learn-this article deconstructs black-box machine-learning models of a diverse chemical reaction data set. Through experimentation with chemical representations and modeling techniques, the analysis provides insights into the nature of how statistical accuracy can arise, even when the model lacks informative physical principles. By peeling back the layers of these complicated models we arrive at a minimal, chemically intuitive model (and no machine learning involved). This model is based on systematic reaction-type classification and Evans-Polanyi relationships within reaction types which are easily visualized and interpreted. Through exploring this simple model, we gain deeper understanding of the data set and uncover a means for expert interactions to improve the model's reliability.

Quantitative Structure-Selectivity Relationships in Enantioselective Catalysis: Past, Present, and Future

The dawn of the 21st century has brought with it a surge of research related to computer-guided approaches to catalyst design. In the past two decades, chemoinformatics, the application of informatics to solve problems in chemistry, has increasingly influenced prediction of activity and mechanistic investigations of organic reactions. The advent of advanced statistical and machine learning methods, as well as dramatic increases in computational speed and memory, has contributed to this emerging field of study. This review summarizes strategies to employ quantitative structure-selectivity relationships (QSSR) in asymmetric catalytic reactions. The coverage is structured by initially introducing the basic features of these methods. Subsequent topics are discussed according to increasing complexity of molecular representations. As the most applied subfield of QSSR in enantioselective catalysis, the application of local parametrization approaches and linear free energy relationships (LFERs) along with multivariate modeling techniques is described first. This section is followed by a description of global parametrization methods, the first of which is continuous chirality measures (CCM) because it is a single parameter derived from the global structure of a molecule. Chirality codes, global, multivariate descriptors, are then introduced followed by molecular interaction fields (MIFs), a global descriptor class that typically has the highest dimensionality. To highlight the current reach of QSSR in enantioselective transformations, a comprehensive collection of examples is presented. When combined with traditional experimental approaches, chemoinformatics holds great promise to predict new catalyst structures, rationalize mechanistic behavior, and profoundly change the way chemists discover and optimize reactions.

NMR Quantification of the Effects of Ligands and Counterions on Lewis Acid Catalysis

The relative Lewis acidity of a variety of metal-ligand catalyst complexes is quantified using ³¹P NMR spectroscopy. Three ³¹P NMR probes, including two new bidentate binding probes, are compared on the basis of different binding modes (i.e., monodentate vs bidentate) and the relative scale of their downfield shift upon binding to Lewis acid complexes. Bidentate coordination of catalyst complexes including metal catalysts, ligands, and counterions were assessed due to their importance to asymmetric catalysis. The effect of ligands, counterions, and additives on Lewis acidity is quantified and correlated to reaction yield at an early time point as an approximation for catalytic activity/efficiency and chelation mode in two organic transformations. Binding studies were performed under catalytically relevant conditions, giving further applicability to synthesis. Insight into activation modes are revealed through this analysis.

Lewis Base/Brønsted Acid Co-Catalyzed Asymmetric Thiolation of Alkenes with Acid-Controlled Divergent Regioselectivity

A divergent strategy for the facile preparation of various enantioenriched phenylthio-substituted lactones was developed based on Lewis base/Brønsted acid co-catalyzed thiolation of homoallylic acids. The acid-controlled regiodivergent cyclization (6-endo vs. 5-exo) and acid-mediated stereoselective rearrangement of phenylthio-substituted lactones were explored. Experimental and computational studies were performed to clarify the origins of the regioselectivity and enantioselectivity. The calculation results suggest that C-O and C-S bond formation might occur simultaneously, without formation of a commonly supposed catalyst-coordinated thiiranium ion intermediate and the potential π-π stacking between substrate and SPh as an important factor in the enantio-determining step. Finally, this methodology was applied in the rapid syntheses of the bioactive natural products (+)-ricciocarpin A and (R)-dodecan-4-olide.

Disulfonimides versus Phosphoric Acids in Brønsted Acid Catalysis: The Effect of Weak Hydrogen Bonds and Multiple Acceptors on Complex Structures and Reactivity

In Brønsted acid catalysis, hydrogen bonds play a crucial role for reactivity and selectivity. However, the contribution of weak hydrogen bonds or multiple acceptors has been unclear so far since it is extremely difficult to collect experimental evidence for weak hydrogen bonds. Here, our hydrogen bond and structural access to Brønsted acid/imine complexes was used to analyze BINOL-derived chiral disulfonimide (DSI)/imine complexes. 1H and 15N chemical shifts as well as 1JNH coupling constants revealed for DSI/imine complexes ion pairs with very weak hydrogen bonds. The high acidity of the DSIs leads to a significant weakening of the hydrogen bond as structural anchor. In addition, the five hydrogen bond acceptors of DSI allow an enormous mobility of the imine in the binary DSI complexes. Theoretical calculations predict the hydrogen bonds to oxygen to be energetically less favored; however, their considerable population is corroborated experimentally by NOE and exchange data. Furthermore, an N-alkylimine, which shows excellent reactivity and selectivity in reactions with DSI, reveals an enlarged structural space in complexes with the chiral phosphoric acid TRIP as potential explanation of its reduced reactivity and selectivity. Thus, considering factors such as flexibility and possible hydrogen bond sites is essential for catalyst development in Brønsted acid catalysis.

Site-selective acylation of natural products with BINOL-derived phosphoric acids

The site-selective acylation of a steroidal natural product 19-hydroxydehydroepiandrosterone catalyzed by 1,1'-Bi(2-napthol)-derived (BINOL) chiral phosphoric acids (CPA's) is described. Systematic variation and multivariate linear regression analysis reveal that the same steric parameters typically needed for high enantioselectivity with this class of CPAs are also required for site-selectivity in this case. Density functional theory calculations identify additional weak CH-π interactions as contributors to site discrimination. We further report a rare example of site-selective acylation of phenols through the evaluation of naringenin, a flavonoid natural product, using CPA catalysis. These results suggest that BINOL-derived CPA's may have broader applications in site-selective catalysis.

Internal acidity scale and reactivity evaluation of chiral phosphoric acids with different 3,3'-substituents in Brønsted acid catalysis

NMR H-bond analysis reveals an offset of internal and external acidities of catalysts and allows for a detailed reactivity analysis.

The concept of hydrogen bonding for enhancing substrate binding and controlling selectivity and reactivity is central in catalysis. However, the properties of these key hydrogen bonds and their catalyst-dependent variations are extremely difficult to determine directly by experiments. Here, for the first time the hydrogen bond properties of a whole series of BINOL-derived chiral phosphoric acid (CPA) catalysts in their substrate complexes with various imines were investigated to derive the influence of different 3,3′-substituents on the acidity and reactivity. NMR ¹H and ¹⁵N chemical shifts and ¹J_NH coupling constants of these hydrogen bonds were used to establish an internal acidity scale corroborated by calculations. Deviations from calculated external acidities reveal the importance of intermolecular interactions for this key feature of CPAs. For CPAs with similarly sized binding pockets, a correlation of reactivity and hydrogen bond strengths of the catalyst was found. A catalyst with a very small binding pocket showed significantly reduced reactivities. Therefore, NMR isomerization kinetics, population and chemical shift analyses of binary and ternary complexes as well as reaction kinetics were performed to address the steps of the transfer hydrogenation influencing the overall reaction rate. The results of CPAs with different 3,3′-substituents show a delicate balance between the isomerization and the ternary complex formation to be rate-determining. For CPAs with an identical acidic motif and similar sterics, reactivity and internal acidity correlated inversely. In cases where higher sterical demand within the binary complex hinders the binding of the second substrate, the correlation between acidity and reactivity breaks down.

Palladium-catalyzed enantioselective alkenylation of alkenylbenzene derivatives

A regioselective and enantioselective palladium-catalyzed relay Heck alkenylation of alkenylbenzene derivatives to construct remote stereocenters is disclosed. Various β-substituted styrenes were readily obtained in moderate yields with good to excellent levels of enantioselectivity. This strategy provides rapid access to enantioenriched δ, ε, ζ, and η-alkenyl aryl compounds from simple starting materials. Mechanistic studies suggest that termination of the relay reaction is controlled by affinity of the arene for the Pd complex during migration.

Mechanistically Guided Predictive Models for Ligand and Initiator Effects in Copper-Catalyzed Atom Transfer Radical Polymerization (Cu-ATRP)

Copper-catalyzed atom transfer radical polymerization (Cu-ATRP) is one of the most widely used controlled radical polymerization techniques. Notwithstanding the extensive mechanistic studies in the literature, the transition states of the activation/deactivation of the growing polymer chain, a key equilibrium in Cu-ATRP, have not been investigated computationally. Therefore, the understanding of the origin of ligand and initiator effects on the rates of activation/deactivation is still limited. Here, we present the first computational analysis of Cu-ATRP activation transition states to reveal factors that affect the rates of activation and deactivation. The Br atom transfer between the polymer chain and the Cu catalyst occurs through an unusual bent geometry that involves pronounced interactions between the polymer chain end and the ancillary ligand on the Cu catalyst. Therefore, the rates of activation/deactivation are determined by both the electronic properties of the Cu catalyst and the ligand-initiator steric repulsions. In addition, our calculations revealed the important role of ligand backbone flexibility on the activation. These theoretical analyses led to the identification of three chemically meaningful descriptors, namely HOMO energy of the catalyst ( EHOMO), percent buried volume ( Vbur%), and distortion energy of the catalyst (Δ Edist), to describe the electronic, steric, and flexibility effects on reactivity, respectively. A robust and simple predictive model for ligand effect on reactivity is thereby established by correlating these three descriptors with experimental activation rate constants using multivariate linear regression. Validation using a structurally diverse set of ligands revealed the average error is less than ±2 kcal/mol compared to the experimentally derived activation energies. The same approach was also applied to develop a predictive model for reactivity of different alkyl halide initiators using R-X bond dissociation energy (BDE) and Cu-X halogenophilicity as descriptors.

Brønsted acid catalysis - the effect of 3,3'-substituents on the structural space and the stabilization of imine/phosphoric acid complexes

BINOL derived chiral phosphoric acids (CPAs) are widely known for their high selectivity. Numerous 3,3'-substituents are used for a variety of stereoselective reactions and theoretical models of their effects are provided. However, experimental data about the structural space of CPA complexes in solution is extremely rare and so far restricted to NMR investigations of binary TRIP/imine complexes featuring two E- and two Z-imine conformations. Therefore, in this paper the structural space of 16 CPA/imine binary complexes is screened and 8 of them are investigated in detail by NMR. For the first time dimers of CPA/imine complexes in solution were experimentally identified, which show an imine position similar to the transition state in transfer hydrogenations. Furthermore, our experimental and computational data revealed an astonishing invariance of the four core structures regardless of the different steric and electronic properties of the 3,3'-substituent. However, a significant variation of E/Z-ratios is observed, demonstrating a strong influence of the 3,3'-substituents on the stabilization of the imine in the complexes. These experimental E/Z-ratios cannot be reproduced by calculations commonly applied for mechanistic studies, despite extensive conformational scans and treatment of the electronic structure at a high level of theory with various implicit solvent corrections. Thus, these first detailed experimental data about the structural space and influence of the 3,3'-substituent on the energetics of CPA/imine complexes can serve as basis to validate and improve theoretical predictive models.