Lewis Base Catalysis in Organic Synthesis

By Scott E. Denmark

This three-volume set represents the first comprehensive coverage of the rapidly expanding field of Lewis base catalysis that has attracted enormous attention in recent years. Lewis base catalysis is a conceptually novel paradigm that encompasses an extremely wide variety of preparatively useful transformations and is particularly effective for enantioselectively constructing new stereogenic centers. As electron-pair donors, Lewis bases can influence the rate and stereochemical course of myriad synthetic organic reactions. The book presents the conceptual/mechanistic principles that underlie Lewis base catalysis, and then builds upon that foundation with a thorough presentation of many different reaction types. And last but not least, the editors, Prof. Edwin Vedejs and Prof. Scott E. Denmark, are without doubt the leaders in this emerging field and have compiled high quality contributions from an impressive collection of international experts.

• Language: English
• Publisher: Wiley
• Release date: Aug 3, 2016
• ISBN: 9783527675173

Book preview

Preface for Volumes 1–3

This three-volume book originates from a widely cited 2008 review with the same title, Lewis Base Catalysis in Organic Synthesis, coauthored by Denmark and Beutner. Given the interest generated by that article, as well as the explosion of related topics in the literature, a more comprehensive treatment was desired by Wiley-VCH. Scott Denmark declined taking on the current project as sole editor due to extensive prior commitments, but did agree to serve as coeditor in planning the project and determining scientific content. In addition, he edited Chapter 1, authored several of the later chapters, and wrote the Introduction that traces definitions of catalysis from Ostwald to the current era and presents an updated, broadly inclusive definition that is used in the current volumes.

After extensive discussion by both coeditors during the planning stages, the decision was made to emphasize mechanistic aspects of Lewis base catalysis where possible, and to provide broad coverage of the most important preparative advances with sufficient commentary and explanation to facilitate graduate instruction as well as to stimulate new research initiatives. Another important objective was to remind the current generation of the remarkable insight and contributions of G.N. Lewis. He was the first to recognize the possibility of catalysis by electron pair donors, and did so two decades before independent attempts to classify this family of reactions resulted in the alternative terminology nucleophilic catalysis. For historical as well as heuristic and conceptual reasons, it is better and more correct to regard this chemistry as Lewis base catalysis.

All of the examples of Lewis base catalysis in these volumes feature activation by a key bonding event between a substrate acceptor orbital (classified as n*, π*, or σ* in chapter headings) and two electrons from a donor orbital in the Lewis base catalyst, but this donor–acceptor interaction is only the appetizer. The main course consists of the stages that follow the Lewis base activation step, and the menu of mechanistic options can be incredibly rich. The options can be very simple, as in halide catalysis (Chapter 1) where a single activation stage by the halide Lewis base is usually followed by a single product-forming stage. However, such mechanistic simplicity is the exception. More often, the mechanisms are deceptively simple, multifaceted, and amazingly subtle. Even that familiar undergraduate-level example of Lewis base catalysis, the venerable benzoin condensation, can be challenging for students who must confront multiple conceptual layers (reversible nucleophilic addition of cyanide; acid–base concepts; carbanion delocalization; leaving group ability) and decipher several steps following the activation stage. It is worth recalling that an earlier mechanism for the benzoin condensation proposed the dimerization of PhC(OH) (yes, the hydroxyl carbene tautomer of benzaldehyde!) to the intermediate enediol PhC(OH)=C(OH)Ph (Bredig, 1904). This suggestion was perfectly logical, concise, and plausible at the time, but lasted only until the alternatives were considered and the mechanism was studied. Perhaps a similar fate awaits other plausible mechanisms, a phrase that appears often in these volumes.

By now, many of the fundamental principles underlying Lewis base catalysis have indeed been studied, and several of the most extensively investigated topics are featured in Volume 1. Chapter 1 begins with a historical account tracing key highlights in the development of catalysis, including important contributions by Berzelius, Liebig, Ostwald, and other major figures of nineteenth century chemistry. This chapter also mentions milestones in Lewis base catalysis from 1834 to 1970, and briefly comments on a few more recent developments that await detailed investigation.

Lewis was the first to recognize the electronic features that define Lewis base catalysis (Introduction and Chapter 2). An overview of his profound insight is presented in Chapter 2, which traces the evolution of Lewis's landmark formulation of the electronic theory of structure and bonding to a clear assertion that his (Lewis's) bases possess every property ascribed to Brønsted bases, including their ability to act as catalysts. The Lewis concepts benefited greatly from refinement and popularization by Mulliken and Jensen, who helped to develop the unifying conceptual basis, a classification scheme of reaction types according to relevant orbital interactions, and a generally applicable terminology that serves as the organizational framework for these volumes.

The next two chapters focus on the thermodynamic and kinetic aspects of Lewis base catalysis, respectively. Chapter 3 presents the classical methods that have been used to quantify Lewis basicity of the most important Lewis bases, and defines the concepts of Lewis Affinity and Basicity. Extensive discussion and tables compare Lewis bases using representative affinity parameters, including those for various cations (proton, methyl, lithium) and neutral Lewis acids (BF3, iodine, 4-fluorophenol). Similarly, Chapter 4 quantifies the corresponding kinetic component (nucleophilicity) using the Mayr Scale, introduces the related concepts of electrofugality and nucleofugality, and provides examples of how these concepts are used by synthetic chemists.

The selection of topics for the subsequent chapters of Volume 1 was made according to several criteria: (i) extensive in-depth mechanistic study, (ii) preparative importance, and (iii) mechanistic diversity following attack by the Lewis base. The first of these chapters (Chapter 5) takes on acyl transfer catalysis by pyridine derivatives, a topic that has been studied in sufficient depth to develop a mechanism that is well understood and widely accepted. Perhaps the same can now be said for much of Chapter 6, involving the mechanism for proline-catalyzed carbonyl activation in enantioselective synthesis, but this is complex, broadly applicable chemistry and the evaluation of models for enantioselection often depends on computational methods that are still undergoing refinement. Similar concerns regarding computations arise in reactions where complexity is associated with the timing and nature of proton transfer events, or with the role of various additives. Those scenarios have long confounded attempts to fully understand the mechanism of the Morita–Baylis–Hillman reaction, a topic that is summarized in Chapter 7. Progress has been made using sophisticated mechanistic tools based on kinetics, mass spectroscopy, computation, and acid–base relationships, but developing a generally applicable mechanism has proven to be difficult.

Some of the mechanistically most intriguing examples of Lewis base catalysis are featured in Chapters 8–11 of Volume 1. These chapters describe reactions that begin with a bonding interaction between the Lewis base and the σ* or n* (unoccupied p) orbitals of the electrophile, reactions that proceed with astonishing mechanistic diversification, even in the relatively simple context of Lewis base activation of silicon nucleophiles (Chapter 8). One take-home message is that only by extensive mechanistic investigation of each substrate category is it possible to classify the reactive intermediates as carbon-bound siliconates or as free carbanions. This conclusion would not surprise authors from an earlier era when physical organic chemistry was the central focus of organic chemistry, and it is underscored by the content of Chapters 9–11. Massive mechanistic study and correlation of enantioselectivity data were required to reveal details of how a chiral Lewis base induces the catalytic formation of cationic silicon electrophiles in aldol and related reactions (Chapter 9), or how Noyori's bifunctional Lewis base catalyst converts dimeric organozinc reagents into intermediates having both Lewis base and Lewis acid character (Chapter 10). In the case of Chapter 11, the combination of kinetic isotope effects, computation, and correlation of extensive enantioselectivity data are shown to confirm Corey's insightful dual activation mechanism for borane reduction of ketones catalyzed by oxazaborolidines, catalytically active intermediates that rely on a single B—N subunit in the key role of both Lewis base and Lewis acid. An unexpected bonus from these studies is the entertaining conclusion that those who favored a boat-like six-center transition state were right, but those who preferred a chair-like transition state were also right. The two transition state geometries are similar in terms of free energy, and both predict the same major enantiomer.

In some cases, the topics selected for Volume 1 were so large that most of the applications and preparative chemistry were split into separate chapters and placed in Volume 2 or Volume 3. Several of the Volume 2 and 3 chapters, such as the N-heterocyclic carbene (NHC) chemistry of Chapter 27, could easily have worked as chapters in Volume 1. Indeed, most of the Volume 2 and 3 chapters contain substantial mechanistic discussion, but the primary consideration is the preparative chemistry and, in particular, the enantioselectivity. No simple generalizations can prepare the reader for the exceptional scope of applications that are covered in Volumes 2 and 3, but some examples will be mentioned below due to their historical, preparative, or mechanistic importance.

Chapters 12–17 of Volume 2 feature the activation of various substrate π* orbitals by nitrogen or phosphorus Lewis bases. These chapters might also be loosely classified as logical descendants of several historically important reactions: amine-catalyzed acylation (Chapters 12 and 13), Morita–Baylis–Hillman Reaction (Chapters 14 and 15), and Knoevenagel condensation (Chapters 16 and 17). However, a closer look will reveal profound differences compared to the historical precedents. There are striking, perhaps even revolutionary, practical advances (e.g., Birman's enantioselective heterocyclic acylation catalysts, Chapter 12), unexpected mechanistic variations (as in the activation of ketenes and ketene equivalents by Lewis bases described in Chapter 13, or in the postactivation proton transfer events enabling phosphine-catalyzed formation of Cl—C or Cl—N bonds in Chapter 15), and astonishing diversity in preparative outcomes. The many variations of Lewis base-catalyzed annulation chemistry are noteworthy for their potential to access heterocycles including β-lactones and β-lactams (Chapter 13), as well as a variety of five- and six-membered rings containing nitrogen or oxygen (Chapters 15 and 26). Zwitterionic ammonium enolates (Chapter 13) or phosphonium enolates or allenolates (Chapter 15) play a key role as the reactive intermediates in the annulation chemistry and in the related Rauhut–Currier reaction. Similar zwitterionic species are also responsible for the aldol-like condensation step in the Morita–Baylis–Hillman synthesis of allylic alcohols and amines (Chapter 14).

Especially important are the preparative aspects described in Chapters 16 and 17, featuring iminium or enamine intermediates generated from carbonyl substrates. Catalysis using MacMillan's chiral imidazolidinones and analogs has revolutionized the activation of α,β-unsaturated aldehydes for enantioselective Michael addition and Diels–Alder applications (Chapter 16). The related topic of enamine catalysis is described in Chapter 17, and the most practical examples use prolinol or its analogs to generate enamines that serve as chiral enolate equivalents in the reactions of aldehydes with a broad range of electrophiles. The latter methods enable formation of C—X bonds (X = O, N, halide) as well as C—C bonds with exceptional enantiocontrol.

Chapters 18–23 of Volume 3 explore the generation of ate intermediates and their role in subsequent events initiated by the interaction of a Lewis base with substrates having an empty σ* or n* orbital. In Chapter 18, the Lewis base attacks organosilane or organotin substrates to form pentacoordinate ate species that are structurally similar to the siliconates considered at the mechanistic level in Chapter 8 (Volume 1). Numerous examples are shown where the siliconate or stannate intermediates behave as carbon nucleophiles (carbanion equivalents) toward various acceptors, and in particular, toward C=O groups (1,2-addition) or C=C—C=O groups (1,2- or 1,4-addition). Some of the most useful cases feature substrates having heteroatom functionality (halide, N, O) or unsaturation attached to the C—Si or C—Sn bond, including such familiar applications as Peterson olefination, allylation, or Prasad trifluoromethylation of carbonyl groups. In the special case of substrates containing a Cl—C—X—C—Si linkage, the Lewis base activation step results in labile intermediates that release transient 1,3-dipoles C=X+—C− that are suitable for 1,3-dipolar cycloaddition chemistry. For completeness, the examples illustrated in Chapter 18 include stoichiometric as well as catalytic uses of the Lewis base.

Nucleophilic ate intermediates are also featured in Chapter 19, but the typical substrates contain trivalent boron as well as a B—B linkage (e.g., bis(pinacolato)diboron, B2pin2), and the activation stage involves attack by the Lewis base at an n* orbital (i.e., empty p-orbital). The resulting boronate species have remarkable reactivity toward Michael acceptors, and use of a chiral N-heterocyclic carbene or chiral phosphine as the Lewis base catalyst triggers highly enantioselective borylation by 1,4-addition of a tricoordinate Bpin subunit to the conjugated substrate. Analogous R3Si—Bpin reagents are also shown to effect 1,4-addition, but it is the silicon subunit that is transferred as might be predicted assuming a boronate intermediate. On the other hand, alkoxide activation of Si—B generates a reagent that transfers boron in the case of aryl halide substrates, resulting in replacement of the halogen! Furthermore, the treatment of B2pin2 with Lewis basic alkoxides generates a reagent capable of transferring both Bpin subunits to both sp² carbons of simple alkenes! These intriguing observations remain to be understood at the mechanistic level.

Unusual mechanisms and ate intermediates are also involved in the reactions of halogenated silane reagents considered in Chapters 20–23, but here the unique reactivity profiles are due to the presence of chloride substituents at silicon. The electronegative chlorides activate silicon for adduct formation with Lewis bases containing polarized P+—O−, N+—O−, or S+—O− subunits that function as nucleophilic oxygen electron pair donors. Chapter 20 illustrates enantioselective allylation and propargylation reactions of carbonyl or imine substrates catalyzed by chiral phosphine oxide or pyridine N-oxide derivatives, while Chapter 22 explores related Lewis base catalysts in the enantioselective reduction or reductive amination of ketones by trichlorosilane. Conceptually, these reactions are easy to understand as examples of Lewis base-induced nucleophilicity at C—SiCl3 or H—SiCl3 bonds in the ate complexes, but things are not so simple in the closely related adducts of SiCl4. Thus, Chapter 21 details the consequences of (n → σ*) activation with silicon tetrachloride. Herein, the polarization of electron density leads to the phenomenon of Lewis base activation of Lewis acids via Si—Cl heterolysis and the resulting catalysis of reactions that are mediated by strong, in situ generated, Lewis acidic siliconium agents. These unique Lewis acids promote myriad, classical carbonyl addition reactions (allylation, aldol, vinylogous aldol) with exceptionally high stereochemical fidelity. This chapter also details extensive kinetic and spectroscopic investigations on the reactive intermediates. Chapter 23 also involves the Lewis base activation of silicon tetrachloride, but in this instance, chloride becomes incorporated into the product in the formation of chlorohydrins from epoxides. Both desymmetrization of meso-epoxides and kinetic resolution of chiral epoxides are demonstrated. Here, the mechanistic underpinnings of the process are also described.

The broad applicability of Lewis base activation of Lewis acidity is illustrated in the extension to the chemistry of elements in Groups 16 and 17, as described in Chapter 24. Here again, the (n → σ*) interaction generates hyperactive chalcogenic and halogenic electrophiles from stable precursors to effect the vicinal difunctionalization of isolated alkenes with high and predictable levels of stereoselectivity. Not surprisingly, the rules that govern the activation of Group 16 and 17 electrophiles are different from those discovered in the foregoing investigations of silicon electrophiles (Group 14). Thus, catalysts for seleno-, sulfeno, and halofunctionalization of double bonds require different electronic features, which have been subject to extensive mechanistic investigations that reveal the origins of both catalytic reactivity and enantioselectivity.

The final chapter on activation of organosilicon compounds, Chapter 25, concerns the phenomenon of double activation catalysis wherein both Lewis acidic and Lewis basic catalysts are simultaneously involved. Strategies to prevent the obvious self-neutralization of the catalysts involve both spatial separation in a single catalytic entity (bifunctional catalysis) and the careful engineering of mutually compatible, but independent components (synergistic catalysis). Many different catalytic systems have been developed for silylcyanation, conjugate cyanation, and Reissert reactions. A major effort has led to the development of catalytic systems containing fluoride as the Lewis basic component with many different metal complexes (Sn, Ag, Cu, and Zn) in a variety of chiral ligand scaffolds.

The last two chapters extend the coverage to (n → π*) activation sequences involving specialized bifunctional Lewis base catalysts. In some respects, these chapters present close parallels to the preparative and mechanistic chemistry already seen in Chapters 12–17, but they focus on applications where proton sharing or proton transfer after the activation stage plays an increasingly central role. In Chapter 26, the Lewis base catalysts contain NH, OH, or SH functionality that may participate in hydrogen bonding to stabilize intermediates and transition states, or that may promote formal deprotonation and proton transfer events along the reaction pathway. The chapter begins with the development of peptide-derived Lewis base catalysts for enantioselective acyl transfer and extends the concepts to fascinating examples of phosphorylation, sulfonylation, and sulfinylation, including the use of cinchona-derived catalysts. Relatively simple amino acid-derived Lewis base catalysts are also described, including bifunctional or polyfunctional thiols and phosphines. These and related catalysts are shown to effect representative allenoate–imine annulations and Morita–Baylis–Hillman or Rauhut Currier reactions with high enantioselectivity.

In Chapter 27, an unusual form of bifunctionality is built into the N-heterocyclic carbene catalysts because the Lewis base activation stage by the neutral NHC reveals a cationic azolium intermediate that is activated for subsequent deprotonation to form an enamine. Subsequent events exploit the nucleophilic properties of the enamine to generate new C—C or C—X bonds by reaction with electrophiles, but these events also form a new azolium subunit, at which point the plot thickens. Additional proton transfer may occur from the azolium species, but it is also capable of heterolytic C—C cleavage to regenerate the original NHC Lewis base. These are some of the fascinating mechanistic options, and other options are also possible, depending on substrate functionality.

These volumes offer coverage of many topics under the very broad umbrella of Lewis base catalysis, but they are by no means all-inclusive. Decisions had to be made about which topics from neighboring fields should be left out. Thus, coverage does not include examples where a Lewis basic solvent accelerates a chemical transformation, even if that effect can be attributed to an interaction between a solvent electron pair and a substrate acceptor orbital. Similarly, the mechanistically analogous Lewis base-catalyzed oligomerizations are not treated beyond brief mention in a historical context (Chapter 15). As in the 2008 review, coverage also does not include those examples where the Lewis base functions as an activating ligand for transition metals. There is no conceptual reason to exclude these cases, but they belong to a separate universe that is best treated in specialized reviews. However, selected transition metal reactions are mentioned in some chapters in the context of cocatalysis, or if interesting analogies can be made with the main group examples. Another topic that is not covered is the emerging field of frustrated Lewis pair chemistry (FLP). In part, this is another separate universe situation, but much of the FLP chemistry is triggered by the electrophile, enabled by what the Lewis base does not do (deactivate the electrophile by premature binding). The Lewis base-catalyzed reactions featured in Volumes 1–3 begin with a bonding interaction between the Lewis base catalyst and an electrophile orbital.

The editors are indebted to Elke Maase for her central role in stimulating and facilitating this project during its inception. Special thanks are also due to Stefanie Volk, who guided the complex stages up to and beyond copy editing, and to Dr. Ulf Scheffler for facilitating the cover design.

EV is grateful to the Chemistry Library at the University of Illinois for easy access to hard copy Beilstein indices, and to Mary Schlembach and SD for their hospitality amid ice and snow in Champaign-Urbana. Also appreciated is the granting of emeritus status by former colleagues at the University of Wisconsin, a favor that facilitated access to historical texts and journals.

Finally, EV thanks his wife Pat Anderson for caring, warmth, patience, and acceptance of yet another chemistry project. After many years and many projects, her understanding and support were always there.

Edwin Vedejs

University of Michigan

Ann Arbor, MI, USA

Scott E. Denmark

University of Illinois

Urbana, IL, USA

Introduction: Definitions of Catalysis

The assumption of this new force is detrimental to the progress of science, since it appears to satisfy the human spirit, and thus provides a limit to further research.

–Leibig, on Berzelius' description of catalysis

This easy kind of physiological chemistry is created at the writing desk and is the more dangerous, the more genius goes into its execution.

–Berzelius, on Leibig's Organic Chemistry in Its Applications to Physiology and Pathology

Catalysis is ubiquitous. It has been practiced and refined by Nature to extraordinary levels since the emergence of life on this planet. It has been practiced and refined by Man also to extraordinary levels over the past two centuries. Yet this profound and central concept in chemistry has been notoriously difficult to define, and thus evokes the infamous quote from US Supreme Court Justice Potter Stewart to describe his threshold test for obscenity, I know it when I see it. Nevertheless, catalysis deserves better! Given that every chapter in this compilation describes a manifestation of this phenomenon, the editors feel compelled to formulate a definition of catalysis and in particular, Lewis base catalysis, that can be universally applied and that is rooted in the recommendations of the cognoscenti from the past 100+ years. (Chapter 1 chronicles the historical evolution of the concepts of catalysis and Lewis base catalysis.)

Such an undertaking requires the assembly of authoritative writings on the subject. The following (necessarily incomplete) list represents a wide spectrum of definitions of catalysis ranging from accredited sources (IUPAC) to personal communications (Beak) to quotes from several monographs dealing with catalysis, including two authored by Chemistry Nobel Prize winners (Ostwald and Noyori). The range of emphasis in these definitions highlights the difficulty in formulating an all-encompassing, universal definition. The reader is encouraged to review these definitions and decide which (if any) resonates with their own interpretation.

IUPAC Recommendations 1994http://www.chem.qmul.ac.uk/iupac/gtpoc/

Catalysis

The action of a catalyst.

Catalyst

A substance that participates in a particular chemical reaction and thereby increases its rate but without a net change in the amount of that substance in the system. At the molecular level, the catalyst is used and regenerated during each set of microscopic chemical events leading from a molecular entity of reactant to a molecular entity of product.

IUPAC Gold Book: Compendium of Chemical Terminology,goldbook.iupac.org/C00876.html

Catalyst

A substance that increases the rate of a reaction without modifying the overall standard Gibbs energy change in the reaction; the process is called catalysis. The catalyst is both a reactant and product of the reaction.

Wilhelm Ostwald

Zeitschrift für Elektrochemie1901, 7, 995–1004

A catalyst is any substance that changes the rate of a chemical reaction without appearing in the final products.

This definition explicitly avoids any speculation about the reasons for such influences. In fact, one should avoid saying that all catalytic effects have the same relevant causes. The origin of the effect is another question altogether; the immediate objective is to formulate a definition that allows one to scientifically analyze this question. As will become apparent, the given definition serves this purpose, because it implicitly questions the quantitative value of such accelerations and decelerations, respectively, and their dependence on the nature and concentration of the catalysts as well as on the temperature and the presence of other substances. Obviously, it should be stressed that all attempts to develop theories for the reasons for such catalytic activities are worthless until such quantitative measurements are carried out.

Katalytische Reaktionen, Lehrbuch der Allgemeinen Chemie, Verlag von Wilhelm Engelmann, Leipzig, 1896–1902, p 262

The nature of catalytic processes, insofar as they are described herein, has already been defined as an acceleration1 of a reaction rate by one of the substances present without a change in the amount of that substance. As long as the amount of the catalytically active substance or the catalyst is not increased or decreased by the starting materials or the products, its influence is found only in the rate constant, not in the rate law of the reaction.

(Ostwald's definition allows for negative catalysis whereby the catalyst slows the reaction. It does not allow any conversion of the catalysts and does not account for autocatalysis.)

Ryoji Noyori, Asymmetric Catalysis in Organic Synthesis, Wiley-Interscience, 1994, p. 4

Catalysis is the process by which a relatively small amount of a foreign material, called a catalyst, increases the rate of chemical reaction without itself being consumed in the reaction.

Bruce C. Gates, Catalytic Chemistry, Wiley, New York, 1992, p. 2

A catalyst is by definition a substance that increases the rate of approach to equilibrium of a chemical reaction without being substantially consumed in the reaction. A catalyst usually works by forming chemical bonds to one or more reactants and thereby facilitating their conversion – it does not significantly affect the reaction equilibrium.

Catalysis always involves a cycle of reaction steps, and the catalyst is converted from one form to the next, ideally without being consumed in the overall process. The occurrence of a cyclic reaction sequence is a requirement for catalysis, and catalysis may even be defined as such an occurrence. One might be tempted to designate the compound that enters the cycle as the catalyst for the reaction. However, this choice would be arbitrary; any of the compounds in the cycle could with equal correctness be designated as the catalyst. The ambiguity in catalyst designation is typical, and the identification of one unique catalytic species is usually not possible. Reference to one species as the catalyst is imprecise but common because it is convenient.

George Parshall. Homogeneous Catalysis, Wiley, New York, 1980, p. 17

The essential characteristics of a catalytic reaction are the same whether the catalyst is soluble or insoluble:

The effect of a catalyst is purely kinetic. It does not make a thermodynamically forbidden reaction favorable, but it can dramatically accelerate an allowed reaction by providing a pathway for a low energy of activation.

The microscopic catalyst site operates in a cyclic fashion through a series of reactions that are repeated each time a molecule of substrate is transformed.

The active catalytic species is not necessarily the same compound that is put into the reaction mixture as a catalyst. Many transformations of the nominal catalyst may occur. The precatalytic reactions often give rise to an induction period before catalysis begins.

Peter Beak, Personal Communication

A catalyst is a species that appears in the rate equation, but not the balanced equation of a given reaction. (This is basically a restatement of Ostwald's second definition, but couches the behavior of a catalyst in both kinetic and thermodynamic terms).

Careful inspection of these definitions reveals the crux of the problem; definitions can be phenomenological or mechanistic. The former is broader but less informative, whereas the latter is less general but more informative! Other problems encountered in defining catalysts arise from (i) any stipulation of the amount of the substance required to qualify as a catalyst (Noyori) and (ii) requiring the ability to recover the substance unchanged at the end of the reaction (IUPAC Gold Book). It should be apparent that the amount of the substance is irrelevant to its ability to change the rate of a reaction and not be consumed; the amount used is a qualitative reflection of its cost, availability, and activity. The recoverability criterion also becomes problematic if the catalytically active species is not the substance added to the reaction and is unstable outside of the catalytic cycle.

The best example of a phenomenological definition is that from Ostwald in 1901 (restated by Beak). There are no exceptions to this definition nor does it lead to confusion because it is devoid of mechanistic implications (as Ostwald explicitly intended). However, since the structure and action of the catalyst or catalytically active species are ignored, it is less informative.

The best example of a mechanistic definition is that from Parshall that forms the basis for the IUPAC definition. These definitions address the rate-accelerating effect (and should also allow for deceleration), but go on to specify that (i) catalysts operate in a cycle that connects the reactants and products, (ii) the catalyst is regenerated during each cycle, and (iii) that the compound added is not necessarily a catalytically active species. These important insights come at a cost however, because for many catalytic reactions this information is not known. That is not to say, however, that these are not knowable, but rather that to satisfy the definition at this level of precision (i.e., the structure and action of the catalytically active species) is not currently possible.

Thus, the following definitions of a catalyst are recommended:

Phenomenological

A catalyst is any substance that changes the rate of a chemical reaction without appearing in the final products, or

a catalyst is a species that appears in the rate equation (constant), but not in the balanced equation of a given reaction.

Mechanistic

A catalyst is any substance that changes the rate of a chemical reaction but without a significant change in the amount of that substance in the system. The active catalytic species is not necessarily the same compound that is put into the reaction mixture. At the molecular level, the catalyst is used and regenerated during each set of microscopic chemical events leading from a molecular entity of reactant to a molecular entity of product.

With clear definitions of catalysts in hand, the next important challenge is to similarly define Lewis base catalysts. The definition used throughout these volumes is a hybrid of the two described above to allow for both scenarios (i.e., both phenomenological and mechanistic) depending upon how much mechanistic information is available. Thus, on a phenomenological level, Lewis base catalysts cause the acceleration (or deceleration) of a reaction by addition of an electron pair donor that appears in the rate equation (constant) but not in the balanced equation. However, the electron pair donor can operate by enhancing nucleophilic or electrophilic character of the reacting species (or both) and all three types of catalysis are presented in the various chapters that follow, thanks to detailed mechanistic studies. Thus,

A Lewis base catalyst is an electron pair donor that changes the rate of a given chemical reaction by interacting with an acceptor atom in one of the reagents or substrates. The binding event may enhance either the electrophilic or nucleophilic character of the bound species. Furthermore, the Lewis base should not be substantially consumed during the course of the reaction, a hallmark of any catalytic process.

and by analogy….

Lewis base catalysis is the process by which an electron pair donor changes the rate of a given chemical reaction by interacting with an acceptor atom in one of the reagents or substrates. The binding event may enhance either the electrophilic or nucleophilic character of the bound species. Furthermore, the Lewis base should not be substantially consumed during the course of the reaction, a hallmark of any catalytic process.

When one considers the enormous number of catalytic reactions that have been discovered and studied, it becomes undeniably apparent that no single definition can encompass all examples. Nevertheless, the phenomenological/mechanistic hybrid recommended above is sufficiently general as to be useful for the large majority of Lewis base-catalyzed reactions described in this treatise. Problems arise with the recoverability criterion in cases wherein the Lewis base added is irreversibly changed and in the process generates a surrogate species that becomes the catalyst even though it may become part of the product.

Two illustrations of these challenges are provided in the following schemes. In Scheme 1, the catalytically active species is a complex formed in a pre-equilibrium between HMPA and silicon tetrachloride, thus making HMPA the precatalyst. Of course, silicon tetrachloride is a stoichiometric reactant. In Scheme 2, TBAF is irreversibly converted into a dithianyl ion pair and fluorotrimethylsilane. Thus, either the ion pair or the enolate adduct can be considered the catalyst for this transformation, but not fluoride.

Scheme 1 Lewis base catalysis definition in action, added Lewis base is a precatalyst.

Scheme 2 Lewis base catalysis definition in action, added Lewis base is an initiator.

The editors hope that readers of these chapters and practitioners in the field of catalysis will adopt these definitions and introduce them into their classes and lectures along with a clear explanation of the principles on which they are founded.

Note

1. Wherein both positive and negative accelerations, that is, decelerations, are implied.

1

From Catalysis to Lewis Base Catalysis with Highlights from 1806 to 1970

Edwin Vedejs¹,²,³

¹Leading Investigator, Latvian Institute of Organic Synthesis, Aizkraukles iela 21, Riga LV-1006, Riga, Latvia

²Professor Emeritus, University of Michigan, Department of Chemistry, 930 N. University Ave., Ann Arbor, MI 48109, USA

³Professor Emeritus, University of Wisconsin, Department of Chemistry, 1102 University Ave., Madison, WI 53706, USA

1.1 Introduction

This chapter tracks several of the most informative developments involving Lewis base catalysis from the time of their discovery to the period when general features of the underlying mechanistic principles had been identified and had stimulated systematic investigation. By roughly 1960–1970, it had been widely recognized that many important reactions are triggered by new bonding interactions involving a pair of electrons from a Lewis base catalyst. A partial list of examples includes such classical reactions as the benzoin and Knoevenagel condensations, pyridine-catalyzed acylations, Walden's autoracemization, the Dakin–West reaction, cyanide- or thiamine-catalyzed Stetter reaction, and so on. It was also becoming clear that the events subsequent to electron pair donation could be exceptionally diverse, ranging from very simple scenarios involving Lewis base-promoted C—X bond heterolysis to other situations involving intriguing complexity over several stages, some of which feature the Lewis base in more than one role. These mechanistic insights did not begin to develop until the twentieth century, following a protracted period of confusing formulas and structures, fascinating discoveries, false starts, and controversies that began roughly at the same time as the science of organic chemistry.

1.2 Catalysis

Much of modern organic chemistry can be traced to the period immediately after the revolution in structural understanding that occurred between 1860 and 1875, but some important stories began earlier, before it was known that carbon is tetravalent and tetrahedral, and before consensus was reached regarding the atomic weights of the most common main group elements. Catalysis is one of those older stories, and the benzoin condensation will be our connection between the first attempts to classify the phenomenon of catalysis, the first example of catalysis by a Lewis base, and (nearly 70 years later!) the first example to be understood mechanistically.

1.2.1 Berzelius Defines Catalysis

By 1835–1836, Berzelius had recognized a common feature among five very different chemical transformations that were initiated by another substance that could be recovered unchanged, including such familiar examples as the hydrolysis of starch or the ignition of a mixture of hydrogen and oxygen by a platinum surface [1]. Some sources also credit Berzelius with coining the word catalysis from Greek kata (down) and lysis (dissolution of an object or a group, among other implied meanings), but certainly the word is older and was used in various contexts by the mid-seventeenth century [2]. One interesting source is Alchimia, authored by Libavius (1597), and sometimes referred to as the first chemistry textbook. However, Libavius connected the term catalysis with a breaking down of base metals in procedures intended to produce gold. Why this term resonated with Berzelius is not entirely clear, but a common theme in his writings and lectures over the years 1820–1843 is the notion that catalysis results from an unknown force exerted by the presence of the catalyst, and that this force acts upon the substrate(s) to cause a chemical transformation. Breaking down or chemically transforming the substrate is understandable in that context, but Berzelius was careful to reject any breaking down of the catalyst itself. It was not to be seen as a participant in the chemical changes initiated by catalysis, but as the source of an unknown catalytic power.

1.2.2 Early Proposals for Intermediates in Catalytic Reactions

At about the same time that Berzelius formulated his definition of catalysis, Mitscherlich demonstrated conversion of ethanol to diethyl ether upon heating in the presence of a substoichiometric amount of sulfuric acid. He also made the connection with known cases in which a small amount of what he called a contact substance (Contactsubstanz) would cause chemical transformation of a large amount of substrate [3]. About two decades later, Williamson revisited the same method of ether synthesis and observed that monoethyl sulfate is formed initially and reacts with ethanol to form the final product [4]. These are probably the first systematic investigations of catalysis involving an organic substrate. Williamson's study is also important for documenting the first example in which catalysis of an organochemical reaction is associated with the formation of an intermediate (monoethyl sulfate) that is capable of reacting with the substrate (ethanol) to give the product (ether). An earlier (1806) study by Désormes and Clément had already implicated reactive intermediates in an inorganic reaction, the catalytic lead chamber process for manufacture of sulfuric acid from SO2 and oxygen in the presence of potassium nitrate, and had attributed the formation of SO3 to the involvement of intermediate nitrogen oxides [5]. Together with Williamson's result, these findings could have helped open the door to a general understanding of catalysis via reactive intermediates, but this did not happen. Instead, chemists interested in catalysis wavered between the influence of Berzelius, who had spoken out against any modification of catalysts during catalyzed reactions, and the skepticism of Liebig, who dismissed Berzelius' unknown catalytic force with the words …creation of a new force with a new word that does not clarify the phenomenon… [6].

Liebig's interest in catalysis had begun at least by 1837 when he and Wöhler reported the hydrolytic cleavage of amygdalin to benzaldehyde by emulsin, a biological catalyst that is present in bitter almonds (Eq. (1.1)) [7]. Over the following years, Liebig proposed alternative explanations for catalysis that invoked the disturbance of existing bonds by heat and by physical contacts between molecules [8]. However, the argument between Berzelius and Liebig was not resolved, and other ideas regarding catalysis had limited impact.

(1.1)

equation

1.3 Progress with Catalysis in Organic Chemistry

The decades immediately preceding and following the Berzelius definition of catalysis produced a number of examples of lasting importance, starting with sulfuric acid production and continuing with other industrially important inorganic processes by the end of the nineteenth century. Even the public became aware of one of the earliest discoveries, the 1823 Döbereiner lamp, consisting of a Zn/H2SO4 chamber as hydrogen source and a Pt ignition catalyst [9]. Several important biological catalysts were also recognized by 1850, including diastase (hydrolysis of starch), emulsin (hydrolysis of glycosides), yeast, and various other ferments (later called enzymes following the 1878 suggestion of W. Kühne) [10] that attracted interest due to their connection with metabolism, nutrition, and, of course, brewing and wine making. The examples of nonbiological, organic catalysis were not at that level of visibility and the topic lagged behind other developments in organic chemistry. The work of Mitscherlich and Williamson is unusual because both authors were clearly aware of the catalytic aspects, and because they happened to combine a practical problem (preparation of ether) and a conceptual advance in organic reactivity (C—O bond formation). Curiously, the generation of organic chemists after Williamson made many discoveries involving catalysis but the topic was not mentioned, may not have been recognized, and might not have been a concern to the authors. The focus of organic chemistry had moved on to identifying new reactions, expanding the understanding of functional groups, and developing useful procedures.

By 1860, revolutionary advances in the understanding of organic structures were under way as a result of Kekule's 1858 structural proposals, and uncertainties regarding carbon valency, relative atomic weights for C, H, and O, and empirical formulas had been resolved [11]. However, almost none of the emerging classical organic reactions (acid, base, or metal-catalyzed) were subjected to mechanistic study until 1900. The timing may have been stimulated in part by Ostwald's redefinition of catalysis, based on his work beginning in the 1880s and finalized in 1901 [12]. Like Berzelius, Ostwald was initially skeptical of reaction intermediates in catalysis, so it is no surprise that intermediates were hardly mentioned in organic manuscripts appearing between 1850 and 1900. More surprising is the frequent absence of any mention of catalysis in manuscripts from this era.

One of the few publications to comment on catalysis is Williamson's paper on catalytic formation of ether from ethanol as already discussed [4]. Another example is Walden's 1898 publication describing what proved to be the first case of Lewis basic halide catalysis [13]. Walden reports the apparently spontaneous decrease in optical rotation (about 40%) for enantioenriched dimethyl 2-bromosuccinate, and the total racemization for d-C6H5CHBrCO2H, both after 3 years of storage. Walden was Ostwald's student in Riga until 1887, and later (briefly) also in Leipzig, and shared his mentor's reluctance regarding hypotheses based on incomplete knowledge. His 1898 text largely avoids any rationale or comment beyond experimental observations, although the closing sentence cautiously states It is not ruled out that a small amount of cleaved HBr as catalyst influences or controls the reaction. On the other hand, Walden is less cautious in a sentence supporting autoracemization, a word that also appears in the title of the paper. Walden implies that racemization is an intramolecular event because no evidence was seen for decomposition of the substrate. However, subsequent studies indicate that the racemization is in fact catalyzed, not spontaneous [14]. In 1929, Kuhn ruled out autoracemization (i.e., self-racemization via reversible heterolysis, homolysis, or elimination) by showing that highly purified bromosuccinate ester is configurationally stable [14b]. Racemization occurs in the presence of KBr and related halide salts, and appears to be a simple example of catalysis by a Lewis base (bromide). The most likely explanation for the partial loss of configurational homogeneity observed by Walden in 1898 involves reversible SN2 displacement by traces of bromide ion, and not a new phenomenon such as autoracemization.

Many other now famous nineteenth century publications described reactions that must have been catalytic, but the authors paid attention mostly to reporting the procedures. Reports in which smaller, nonstoichiometric amounts of a key reagent were advantageous for yield, purification, or convenience certainly included suitable comments to alert the reader. Notable examples include Fischer's classical glycosidation [15] and esterification methods [16], an earlier (first) example of acid-catalyzed acetal formation [17], and Claisen's crossed aldol condensation to form benzylideneacetone in the presence of a little ZnCl2 [18]. Other examples of catalysis attracted attention if they made a good story. For example, Kekulé's 1870 manuscript describing the conversion of acetaldehyde to crotonaldehyde begins with commentary on prior work in which this reaction had been achieved by simply heating the aldehyde in ethylidenechloride [19]. However, Kekulé observed that no condensation occurred using purified solvent, but did take place in the presence of traces of HCl in control experiments. Apparently, the earlier report had used contaminated ethylidenechloride, and Kekulé made sure to start his story with this teaser. Finally, a number of reports describe catalytic reactions in which the amounts of reagents are given but not discussed, as often happened in various base-induced carbonyl condensations [20], or in which the amounts are left unspecified, as happened in the first report of hydrogenation over a platinum catalyst [21].

One common feature among all of the classical publications cited in the prior paragraph is that none of them mention catalysis, catalyst, or analogous terms. Was the concept so familiar that mention was not necessary? Were the authors avoiding pointless conflicts over unknown forces? Or was the topic simply not in fashion? Most likely, all of these reasons played some small role, but a larger factor may have been the scientific culture during the second half of the nineteenth century. During this period, a massive influx of seemingly endless and often disconnected new facts was reported along with unsatisfactory, short-lived hypotheses and explanations. Many years later, Ostwald would characterize the controversies between supporters of Berzelius and supporters of Liebig and their role in the development of catalysis as part of his 1909 Nobel Prize lecture as follows:1 Although neither Berzelius' good definition nor Liebig's bad definition promoted in any way this scientifically interesting and technically highly important field, the new definition (i.e., Ostwald's own definition) had this effect at once. Apparently, Liebig's bad definition was too familiar to merit any descriptive comment. From a current perspective, Liebig's writings about catalysis come across more as commentary than definition [6,8], and connect catalysis with physical phenomena as already mentioned. However, Ostwald did provide an evaluation of the Berzelius interpretation: Only the statement that the catalysts acted by their mere presence can be criticized. but proceeded to dismiss unknown catalytic forces or catalytic powers and to connect catalysis with the emerging fundamental principles of reaction kinetics and Gibbs energy.

1.4 Ostwald's Redefinition of Catalysis

Ostwald has been called the founding figure in physical chemistry, having contributed to several distinct and important advances and subdisciplines [22], but he received the Nobel Prize in Chemistry in 1909 specifically for his work on catalysis and for his investigations into the fundamental principles governing chemical equilibria and rates of reaction. It was Ostwald who developed the modern definition of a catalyst in his publications and lectures between 1894 and 1901.

1.4.1 The Evolution of Ostwald's Views and Their Subsequent Refinement

Ostwald's views about catalysis developed during a decade of work involving acid-catalyzed hydrolysis, and his first definition was published under unusual circumstances in 1894. Quite remarkably, the definition appeared not in a typical research contribution, but in a one-page critique of a manuscript by F. Stohmann, as follows. If this reviewer faced the task of characterizing the general phenomenon of catalysis, he would envision the following expression: the acceleration of a slowly occurring chemical change by the presence of a foreign substance. [23]. Together with the appended explanation saying (in part) that the foreign substance is not necessary for the reaction, this definition encountered some resistance. If the catalyst is not necessary, then apparently a slowly occurring chemical change would have a finite rate in the absence of catalyst. Is that different from a reaction that does not occur at all or is too slow to observe on some arbitrary laboratory timescale? Ostwald thought not, but others had doubts. Some took issue with Ostwald's earlier study of HCl-catalyzed hydrolysis of methyl acetate that began with a restatement of the Berzelius criterion for catalysis [24]. Having carefully proved that HCl is not consumed during the reaction, and that weaker acids catalyze hydrolysis more slowly, the paper closes with a statement that the catalytic acids …remain in the free state throughout the course of the reaction…. This was taken by some to imply that the acid catalyst acts by its mere presence, and does not participate directly. Whether or not that was Ostwald's intent can be questioned since he would not have speculated beyond the observable facts. In any event, his paper makes no specific proposal about how the catalyzed hydrolysis takes place.

Presumably, the mixed responses to the 1894 critique/definition may have stimulated Ostwald to take a much more careful approach in his definition published in 1901. This manuscript is the text of a lecture that summarizes the history of catalysis in depth, mentions alternative explanations, and comments on a possible catalytic role for reaction intermediates if they can be detected and their rate effects confirmed. The following simplified definition was given:

A catalyst is any substance that changes the rate of a chemical reaction without appearing in the final products. [25]

By the time of his Nobel Prize lecture in 1909,¹) Ostwald's opinion of intermediates had evolved further. As he also did in 1901, he called attention to the 1806 work by Désormes and Clément [5], but refers to it with more weight as the theory of intermediate reactions followed by the words …no other equally effective principle has hitherto been found in the theory of catalysis. Ostwald cautions against assuming that intermediates are always required, but does not comment on recent developments that had already implicated intermediates in heterogeneous as well as homogeneous catalysis, as discussed in Section 1.4.2 and Section 1.5.

1.4.2 Sabatier and Temporary Compounds in Heterogeneous Catalysis

During the evolution of Ostwald's views, Sabatier had begun his classical studies on heterogeneous catalysis. In 1897, he had already observed that changes in catalysts could change the outcome of reactions at a metal surface, and spoke out against contemporary ideas focusing on a role for physical proximity due to adsorption of gaseous reactants in porous catalysts [26]. Instead, Sabatier explained the conversion of ethylene into methane and carbon upon heating with nickel (obtained from the oxide and hydrogen) by saying that It is possible that an unstable compound of nickel and ethylene is first formed and afterwards splits up into nickel, carbon, and methane. In other publications, Sabatier suggested the formation of distinct temporary compounds (i.e., intermediates) to account for differing catalyst-dependent outcomes [27]. The importance of intermediates was further underscored by Sabatier's 1912 Nobel Prize, awarded for his method of hydrogenating organic compounds in the presence of finely disintegrated metals whereby the progress of organic chemistry has been greatly advanced in recent years (in other words, for major developments in heterogeneous catalysis, including the hydrogenation of alkenes, alkynes, and even benzene derivatives). Sabatier's award address comments on his views regarding catalysis from 1897 onward: I thought and I still think…that the decisive cause of the catalytic activity of porous platinum is not a simple process of physical condensation producing a local rise in temperature but that it is a real chemical combination of the surface of the metal with the surrounding gas.¹) Later in the same lecture, Sabatier mentions the temporary formation of nickel hydrides under conditions of catalytic hydrogenation. He also draws an analogy between (i) Williamson's monoethyl sulfate intermediate in the reaction of ethanol with sulfuric acid to form ether and (ii) hypothetical alkoxide intermediates formed at the surface of metal oxide catalysts (Al2O3, ThO2, or WO2) acting as displaceable groups in high-temperature reaction of simple alcohols to form ethers.

1.4.3 A Curious Tangent: The Radiation Hypothesis for Catalysis

Ostwald rejected a role for any special catalytic force. Nevertheless, unusual candidates for the origins of catalytic reactivity continued to enter the literature until the 1920s, culminating in the so-called radiation theory [28]. This theory proposed, among other things, that infrared radiation emitted by catalysts, solvents, walls of reaction vessels, and so on was an energy source involved in reactivity, and thus also in catalysis. In one paper by Lamble and McCudmore Lewis (no connection with G.N. Lewis), infrared radiation by the hydrogen chloride molecule was discussed as an explanation for the acid-catalyzed inversion (i.e., hydrolysis) of sucrose [29], while the abstract of a publication by Barendrecht states Since the change of activity of invertase with the change of pH corresponds to that expected on the radiation hypothesis, it is concluded that invertase also works by a radiation. [30]. The radiation theory was originally phrased using the language and mathematics of early physical chemistry and attracted sufficiently serious attention to warrant a 1928 article by Daniels in Chemical Reviews that begins with the sentence Few hypotheses in science have suffered such a rapid rise and fall as the radiation hypothesis. [31]. To be fair to the initiators of this idea (Trautz, Perrin, McCudmore Lewis), it arose partly over real concerns regarding some aspects of early collisional activation theory. In contrast to various attempts to explain catalysis during the nineteenth century, the radiation theory was presented in a scientifically testable context, was duly tested, and was soon refuted.

No one claimed to have identified the original catalytic force of Berzelius as part of the radiation episode, but the demise of this hypothesis may have helped to finally put an end to the quest for unknown forces. The new understanding of reaction kinetics and thermodynamics in the context of catalysis did not begin with Ostwald [32], nor did Ostwald bring it to the modern level. What he did do was to convince many chemists that no special forces were needed to understand catalytic phenomena, and that explanations could be found in physical chemistry.

Further discussion of catalysis is beyond the scope of this chapter, but more detailed accounts of the early developments are available in recent historical reviews [9] and the kinetic treatments are given in many textbooks. The account by Lindström and Pettersson also provides a summary of milestones in industrial catalysis from the nineteenth century up to 1970, including Ostwald's process for oxidation of ammonia to nitric acid, Haber's conversion of nitrogen into ammonia, the Fischer–Tropsch method for production of hydrocarbons by coal gasification, Houdry's petroleum cracking process, the Ziegler–Natta alkene polymerization chemistry, Wacker oxidation of ethylene to acetaldehyde, and many other developments that profoundly affect chemical technology [9a].

1.5 The First Example of Lewis Base Catalysis

Shortly before Berzelius' comments regarding catalysis were published in 1836, the first example that can be recognized as Lewis base catalysis appeared as a very small part of the famous 1832 paper by Wöhler and Liebig describing their investigation of oil of bitter almonds (benzaldehyde) [33]. Treatment of the naturally derived oil with KOH resulted in modest conversion to a crystalline product that was called benzoin (after gum benzoin, a source of benzoic acid, as well as the origin of traditional nomenclature of benzene derivatives). Elemental analysis gave the same % composition as determined for the starting material. The original Wöhler–Liebig publication was instrumental in correctly defining oil of bitter almonds as a benzoyl hydride, although the exact identity of benzoyl remained elusive for some years. Of course, the benzoyl hydride was later identified as benzaldehyde, but benzoin came to be described as an isomer (not a dimer) and the correct empirical formulas remained uncertain [isomerism had already been encountered by Wöhler in 1827–1828 and was defined by Berzelius between 1830 and 1831 [34]. The formula of benzaldehyde given by Liebig happened to be incorrect (C14H12O2) because the work was done well before the 1860 consensus regarding correct atomic weights, but this meant that the formula fits benzoin. Such formulas were in flux until 1869 [35–38], as shown in Scheme 1.1, but the more intriguing conclusion was entirely correct: benzoin is a different substance, but has the same elemental composition as the starting material, benzaldehyde.

Scheme depicting various chemical reactions in a time span of 40 years (1832–1869) that led to finalization of formula and structure of Benzoin.

Scheme 1.1 Benzoin condensation; 40 years to finalize formula and structure.

One other development having special significance for Lewis base catalysis occurred during the first decade following discovery of the benzoin reaction. In 1840, Zinin reported that formation of benzoin from benzaldehyde can be effected using potassium cyanide without any potassium hydroxide [35]. The original experiments of Liebig and Wöhler had unknowingly generated the true catalyst in situ because oil of bitter almonds is contaminated with HCN. Zinin must have understood this because he was working in Liebig's laboratory at the time. He comments that good conversion to benzoin depends on how much HCN is present in the naturally derived oil, and also mentions successful experiments using pure benzoyl hydride (presumably, distilled benzaldehyde) and dilute ethanolic KCN (no KOH), or ethanolic KOH plus a few drops of HCN. The manuscript does not specify how much cyanide was used, so we cannot be certain whether Zinin knew what would happen using a substoichiometric quantity of cyanide. On the other hand, he certainly knew what would happen with an excess of HCN because the same paper describes isolation of the HCN adduct (cyanohydrin) of benzoyl hydride and comments on its somewhat variable presence in samples of oil of bitter almonds. Zinin also knew that cyanide did not become part of the benzoin product, a key criterion for catalysis.

Here was a case where a natural substrate contained at least two potential precatalysts (HCN and the cyanohydrin) for its own conversion to a new structure. A number of other phenomena were already recognized where chemical transformations were affected by the action of natural contaminants, but these examples would prove to be the result of contaminating enzymes (leavening of bread, fermentation of grapes, etc.). One of these, the 1837 report by Wöhler and Liebig describing the hydrolytic cleavage of amygdalin using emulsin (Eq. (1.1)), has already been discussed in connection with catalysis in Section 1.2.2. Perhaps these analogies between cyanide and biological catalysts escaped Zinin's notice, but more likely catalysis was not his concern and not his problem. Certainly, the possibility of biological as well as chemical catalysis was already explicit in the Berzelius summation of known catalytic events in 1836, but Zinin did not raise that issue and the literature makes no connection between catalysis and benzoin condensation until much later.

In 1861, the correct formula of benzoin appeared in a Zinin publication [37], without comment about a different formula in a paper by the same author a few years earlier (Scheme 1.1) [36]. On the other hand, the structural problem had not been solved. By then, benzoin had been correlated chemically with benzil (originally, benzoyl) as well as diphenylethane, both of which were recognized to contain two phenyl substituents. However, the connectivity of benzil was clouded by its controversial base-induced rearrangement to benzylic acid, as discussed in the broader historical context by Berson [39]. At the time, many chemists followed Kekule's view that carbon skeletons could be degraded into simpler carbon segments, but could not be rearranged [39]. Nevertheless, these concerns were largely overcome by 1870 and all of the key structures relevant to the benzoin problem were understood by at least some of the principal players. In 1874, the structures were already correctly shown in a leading textbook of that era [40]. That is not to say that they were universally accepted [41], but consensus was near.

1.6 The Road to Mechanistic Comprehension; Multistage Catalysis by Lewis Base

The structural problems involving benzoin were solved, but little progress was forthcoming regarding the nature of the C—C bond-forming step. Further progress would have to await the emergence of mechanistic thinking in connection with the role of intermediates in catalysis, and also with amine catalysis as an important advance in preparative organic chemistry. This topic would eventually become one of the most broadly applicable categories of Lewis base catalysis, and currently is probably the most important one (see Chapters 6, 12, 13, 16, and 17). The point of entry in this account will be the 1894–1898 study by Knoevenagel describing the condensation between aldehydes and acetoacetate or malonate esters [42].

1.6.1 The Knoevenagel Condensation

In his first publications on the topic, Knoevenagel showed that primary or secondary amines promote the condensation of β-dicarbonyl compounds with aldehydes to give amine-free products (Scheme 1.2). His 1896 paper is especially significant because it shows that the reaction of ethyl acetoacetate with benzaldehyde in the presence of about 1% of piperidine can be controlled to give either a 1 : 1 adduct (1) or a 2 : 1 adduct (2), depending on the temperature [43]. The text shows the reaction scheme and comments As one can see, the piperidine is regenerated during the reaction. That explains how 1% of piperidine effects the condensation of a large amount of acetoacetate and aldehyde. Importantly, this paper also offers a rationale (chemismus) suggesting that benzaldehyde is converted initially into benzylidenebispiperidine (the benzaldehyde aminal), which then condenses with acetoacetate by replacement of the mobile (bewegliche) methylene hydrogens. Two years later, another publication in the series describes the same events, "…it is not yet clarified how the amine plays the role of contactsubstanz whereby a small amount is sufficient to convert a large amount of aldehyde and acetoacetate…" [44]. However, later that year (1898) the most definitive paper in the series appeared demonstrating that the previously known crystalline aminals CH2(NHPh)2 and CH2(NC5H10)2 react upon heating with an excess of dimethyl malonate to give the 2 : 1 adduct (5) [45]. The sequence of three publications from 1896–1898 is noteworthy not only for its preparative impact but also for its scientific approach: (i) catalysis was clearly recognized and emphasized, (ii) a mechanistic hypothesis based on a precedented intermediate