Emerging Trends in Cross-Coupling: Twelve-Electron-Based L1Pd(0) Catalysts, Their Mechanism of Action, and Selected Applications

Monoligated palladium(0) species, L1Pd(0), have emerged as the most active catalytic species in the cross-coupling cycle. Today, there are methods available to generate the highly active but unstable L1Pd(0) catalysts from stable precatalysts. While the size of the ligand plays an important role in the formation of L1Pd(0) during in situ catalysis, the latter can be precisely generated from the precatalyst by various technologies. Computational, kinetic, and experimental studies indicate that all three steps in the catalytic cycle—oxidative addition, transmetalation, and reductive elimination—contain monoligated Pd. The synthesis of precatalysts, their mode of activation, application studies in model systems, as well as in industry are discussed. Ligand parametrization and AI based data science can potentially help predict the facile formation of L1Pd(0) species.


INTRODUCTION
The palladium-catalyzed cross-coupling is an exceptionally important area within the field of homogeneous catalysis in modern organic synthesis. This is attested to by the awarding of the 2010 Nobel Prize in chemistry to Richard Heck, Akira Suzuki, and Ei-ichi Negishi, three pioneers in the 1970s of different types of Pd catalyzed carbon-carbon cross-coupling reactions that bear their names. 1 Other pioneers, such as Murahashi, Stille, and Mizoroki, have also made fundamental contributions that, along with those of Hiyama, Tamao, and Miyaura, have made it possible for the technology to develop and mature into what it is today. The significant discoveries and applications of the different types of cross-coupling, which have propelled these reactions to become a leading synthesis technology of the 21st century, have been reviewed by Colacot, Snieckus, and coauthors. 2 Cross-coupling is an example of a technology where incremental rather than breakthrough innovations led to the major advances in the field and, consequently, the awarding of the Nobel Prize, with a focus on C−C bond forming cross-coupling reactions. Since the mid-1990s, independent pioneering work by Hartwig and Buchwald has expanded the scope of the cross-coupling reaction to carbon−heteroatom bond-forming reactions, known today collectively as the Buchwald−Hartwig cross-coupling, which has seen so far a tremendous growth in its applications. Until the late 1990s, the improvement and expansion of crosscoupling reactions had focused on: (i) switching from Ni to Pd in order to utilize the well-defined, two-electron process, 3 and (ii) changing the nucleophilic coupling partner from an organomagnesium nucleophile to an organozinc, tin, silicon, or boron guided by Pauling's electronegativity scale. 4 One of the milestones in the development of cross-coupling was disclosed by Littke and Fu in 1998, 5 in which the effective coupling of an aryl chloride in a C−C coupling was achieved using P(t-Bu) 3 in conjunction with a palladium catalyst precursor. 6 In the same year, Koie reported the use of a P(t-Bu) 3 /Pd system for the amination of aryl chlorides with diarylamines. 7,8 Concurrently with the disclosure of Littke and Fu, reports from Buchwald's 9 and Hartwig's 10 laboratories also helped to significantly propel this area of research to a whole new level with the introduction of novel privileged ligands, newer applications, and better processes. 11 N-Heterocyclic carbenes (NHCs), 12 which had been investigated by Herrmann 13,14 and co-workers in conjunction with palladium, also emerged as a new class of ligand for cross-coupling, with significant successive contributions by Nolan, 15−20 Organ, 21,22 and Glorius. 23 A consensus then emerged that designed ligands with appropriate steric bulk and electronic parameters were the most important component in cross-coupling, 11 an observation that had also been implied earlier independently by Osborn 24 and Milstein 25 for the palladium-catalyzed carbonylations of aryl chlorides. Key observations led to important conclusions regarding the relationship of ligand properties to its overall effect on the catalytic cycle: oxidative addition, transmetalation, and reductive elimination. 26 Although monoligated L 1 Pd(0) complexes constitute the active species in the cross-coupling catalytic cycle, these complexes are practically impossible to synthesize and isolate due to their high reactivity. Hence, precursors such as Pd 2 (dba) 3 and Pd(OAc) 2 have been employed in conjunction with suitable ligands to generate the "active" Pd(0) in situ. The inherent purity issues of both Pd 2 (dba) 3 and Pd(OAc) 2 have been highlighted 27,28 based on the work of Ananikov 29 and Colacot,30,31 respectively. When one uses in situ catalysis (ligand plus a Pd precursor), the precise formation of the desired catalytic species is difficult to achieve, although the size and geometry of the ligand may help to a certain extent. Hence, the in situ technology can result in poor efficiency of the overall catalytic process in terms of metal loading, conversion, and selectivity. It is well understood that even when utilizing the same ligand, there could be a significant difference in activity between monoligated and bisligated complexes ( Figure 1). 32 As far as the preformed catalysts are concerned, the 14-electronbased L 2 Pd(0) complexes are typically utilized directly or generated in situ from a Pd precursor, L 2 PdX 2 . 33 Although many examples of bis-coordinated L 2 Pd(0) complexes are commercially available, even in bulk quantities, 34 monoligated L 1 Pd(0) complexes have yet to be isolated or fully characterized as stated above. The high activity of monoligated L 1 Pd(0) is related to its unsaturated coordination sphere, based on the well-known 18electron rule postulated by Langmuir. 35,36 Fortunately, several new technologies do exist today that permit the generation of the air-sensitive and highly reactive L 1 Pd(0) species through activation of suitable precatalysts. Many of these new-generation precatalysts are air-and moisturestable, even at elevated temperatures and in solution, due to their existence as Pd(II) complexes. The rate and mode of activation of these precatalysts are important in determining the outcome of catalytic transformations. Therefore, these unique precatalysts, containing the same ligand, may exhibit different crosscoupling activities and selectivities, depending on their structural profile and reaction conditions.
Currently, there are no comprehensive reviews on this relevant modern topic, and this survey is intended to help researchers both in academia and industry gain a better understanding of this important emerging area.
In the sections that follow, this review highlights the development of various emerging technologies for generating L 1 Pd(0) catalysts, wherein L is a tertiary phosphine or NHC ligand. Each of the approaches for generating L 1 Pd(0) is described and critically evaluated in terms of synthesis, activation mechanism, and unique applications. Overlap with earlier reviews has been avoided; however, relevant reviews from the last 5−10 years are cited to make this review comprehensive. Selected industrial applications focusing on pharmaceuticals are also included. Most importantly, this review should serve as a primer for chemists wanting to become familiar with cross- coupling reactions and as a reference guide for chemists seeking alternative synthetic methods that offer improved reaction efficiency from a process-and atom-economy points of view.

MECHANISTIC STUDIES SUGGESTING THE INVOLVEMENT OF L 1 Pd(0) IN THE CATALYTIC CYCLE
The generally accepted mechanism of cross-coupling reactions involves three principal steps: (i) oxidative addition, (ii) transmetalation, and (iii) reductive elimination (Scheme 1). 37 Miyaura and Suzuki's statement, 37 "palladium complexes that contain fewer than four phosphine ligands or bulky phosphines such as tris (2,4,6-trimethoxyphenyl)phosphine are, in general, highly reactive for the oxidative addition because of the ready formation of coordinate unsaturated palladium species", is based on a kinetic study conducted by Farina and Krishnan on the Stille crosscoupling reaction. 38 Farina's observation that excess phosphine slows down the coupling, coupled with the above statement by Miyaura and Suzuki, may have prompted many modern crosscoupling experts to look into employing bulky ligands to form coordinatively unsaturated palladium complexes to improve the efficacy and efficiency of a given cross-coupling reaction.
In this regard, investigations carried out by Fu on P(t-Bu) 3 , Buchwald on biaryl ligand systems, Hartwig on the QPhos ligand, and Beller on CataCXium (Evonik Degussa GmbH) ligands have clearly indicated that bulky phosphine ligands prefer forming low-coordination Pd species during the catalytic cycle. 11,39 In general, Tolman's cone angle has been useful in measuring the overall sterics of the phosphine ligand 11 while Nolan's 15−20 and Organ's 21,22 work on bulky N-heterocyclic carbenes (NHCs) have invoked the percent buried volume (% V bur ) for ligand parametrization, especially for NHC systems.
One of the fundamental questions in cross-coupling reactions is about the nature of the active species in the catalytic cycle. While it is generally accepted that L n Pd(0) is the active species, the further question is whether or not it is a monocoordinated (n = 1) or biscoordinated (n = 2) palladium or both. In this section, we shall review the available mechanistic studies aimed at answering these questions.

L 1 Pd(0)-Based Catalytic Species in the Oxidative Addition Step
Bulky ligands such as P(t-Bu) 2 (1-Ad) or P(t-Bu) 3 , with a cone angle of ca. 180°, are known to form both monoligated and bisligated complexes. However, based on kinetic and mechanistic studies by Hartwig and co-workers, 40,41 even the isolable L 2 Pd(0) initially loses one of the ligands to form a monomeric oxidative addition complex, L 1 Pd(Ar)X, possessing a T-shaped geometry (Scheme 2). 40−42 This T-shaped intermediate is typically stabilized by a weak agostic interaction between a C−H in the ligand L and the fourth coordination site of the palladium.
In certain cases, the oxidative addition product can be a monoligated dimer, [L 1 Pd(Ar)X] 2 , depending on the nature of the ligand L (e.g., L = P(o-Tol) 3 ) and the nature of the aryl halide, ArX. Kinetic studies by Hartwig and Paul on the oxidative addition of aryl bromides to Pd[P(o-Tol) 3 ] 2 indicated an inverse first-order dependence of the observed rate constants on the concentration of P(o-Tol) 3 . 43 This is in contrast to the classic understanding of the oxidative addition of ArI to Pd(PPh 3 ) 4 , where dissociation of two PPh 3 ligands leads to a 14-electron intermediate, Pd(PPh 3 ) 2 , which then produces a fourcoordinate (Ph 3 P) 2 Pd(Ar)I species. 44,45 By using ion-trap mass spectrometry, McIndoe, Maseras, and co-workers demonstrated that, in the Pd(PPh 3 ) 4 catalyzed oxidative addition of bromobenzene, the reactivity ratio of Step from L 2 Pd(0) and from L 1 Pd-Based Precatalysts with a Bulky Ligand such as P(t-Bu) 3 bromobenzene with L 1 Pd(0) vs L 2 Pd(0) was at least 10 4 :1 by mass spectrometric measurements and 10 5 :1 by theoretical calculations. 46 A computational (DFT) study by Norrby and coworkers on the oxidative addition of aryl chlorides to monoligated Pd complexes revealed that electron deficient aryl chlorides tend to interact strongly with Pd due to back-donation to form stronger prereactive complexes. 47 Further DFT studies, combined with polarized continuum solvation models, carried out by Fu, Liu, and co-workers, revealed that for PhX substrates (X = Cl and Br), the transition state of the oxidative addition to 14-electron-based [Pd(PPh 3 ) 2 ] has a much higher free energy than the transition state of the 12-electron-based [Pd(PPh 3 )] species. 48 Noticeably for the bulky P(t-Bu) 3 , the L 2 Pd transition state does not even exist. Hence, for both bulky P(t-Bu) 3 and less bulky PPh 3 ligands, oxidative additions seem to proceed via a 12electron-based L 1 Pd(0) pathway with PhX (X = Cl and Br).
Systematic studies by Hartwig and co-workers on the oxidative addition of Ar−I to a series of trialkylphosphine− palladium complexes having the general formula L 2 Pd, where L = P(t-Bu) 3 , P(t-Bu) 2 Cy, P(t-Bu)(Cy) 2 , and PCy 3 , indicated that bulky ligands such as P(t-Bu) 3 and P(t-Bu) 2 Cy form the L 1 Pd(Ar)I species by ligand dissociation. In contrast, Pd complexes with the relatively less bulky P(t-Bu)(Cy) 2 and PCy 3 ligands give L 2 Pd(Ar)I by an associative pathway (Scheme 3). 49,50 These findings are in agreement with those of Brown and co-workers. 51 Further computational studies conducted by Harvey, Fey, and co-workers using dispersion-corrected DFT together with solvation models have corroborated Hartwig's findings. 52 The oxidative addition products of PhBr and PhCl are also known to form the stable, four-coordinate trans complexes just as in the case of the iodide. 24,42,53 These studies concluded that less bulky ligands, such as PPh 3 and PCy 3 , tend to form L 2 Pd(Ar)X, while bulkier ones, such as P(t-Bu) 2 (1-Ad) and P(t-Bu) 3 , form L 1 Pd(Ar)X even from L 2 Pd(0). Although Hartwig's pathway in Scheme 3 indicates that ligand dissociation takes place after oxidative addition, for bulkier ligands it could occur prior to oxidative addition, leading to equilibration beteeen L 2 Pd(0) and [L 1 Pd(0)]. Similar observations were made by Shaughnessy with Np-based L 2 Pd(0) complexes, which gave [LPd(Ar)(μ-X)] 2 dimers upon reaction with ArX. 54,55 Very bulky ligands, for example, some biaryl ligands such as t-BuBrettPhos, with very large cone angle may not form L 2 Pd(0); however, these ligands might form L 1 Pd(Ar)X during the oxidative addition even when the Pd complex is generated in situ. Nevertheless, kinetic studies by Colacot and co-workers point to a significant difference in activity between preformed monocoordinated and biscoordinated P(t-Bu) 3 complexes of Pd (Scheme 4). 32 Very recently, Hirschi, Vetticatt, and co-workers carried out a combined study of theoretical and experimental 13 C kinetic isotope effects to gain an understanding of the mechanism of the Pd(PPh 3 ) 4 catalyzed Suzuki−Miyaura cross-coupling of aryl halides with aryl boronic acids, where oxidative addition of the aryl halide takes place onto a 12-electron, monoligated palladium complex, [Ph 3 P−Pd]. The study revealed that the commonly proposed oxidative addition to the 14-electron Pd(PPh 3 ) 2 complex can happen only under stoichiometric conditions or in the presence of excess added ligand. However, after the first turnover and in the absence of excess ligand, the catalytically active species is the 12-electron, monoligated [Ph 3 P−Pd] (Scheme 5). 56

Does Reductive Elimination in the Catalytic Cycle Involve L 1 Pd(Ar)(Nu), Where Nu = Nucleophile?
Several studies have been conducted on the reductive elimination step to form C−C bonds with palladium complexes containing bidentate phosphines and monophosphines with sterics similar to that of PPh 3 . It was found that reductive elimination is faster from three-coordinate than from fourcoordinate complexes in Pd systems with sterically bulky ligands. 57−59 The rates of reductive elimination from palladium complexes containing two monophosphine ligands is inversely dependent on the concentration of the added ligand, suggesting that reductive elimination occurs via a three-coordinate species by losing one of the ligands. 60,61 Hartwig's efforts to develop palladium-catalyzed α-arylations of carbonyl compounds and nitriles based on the reductive eliminations of C(sp 2 )−C(sp 3 ) bonds have been reviewed. 62 Reductive elimination from arylpalladium malonate ligated by a bulky alkylphosphine such as P(t-Bu) 2 Fc (Fc = ferrocenyl) occurred in high yield and at lower temperatures than from complexes of less bulky ligands such as PPh 3 or bisphosphines. The faster reductive elimination from complexes of P(t-Bu) 2 Fc over those ligated by PPh 3 or DPPE is attributed to the increased steric bulk of the alkylphosphine in the tricoordinate species (Scheme 6). 62 Similar trends were observed for the C−N coupling as well. Yamashita and Hartwig reported the first examples of monomeric three-coordinate arylpalladium amido complexes and investigated the reductive elimination of arylamines from these species. 41 The reactions of potassium arylamides with three-coordinate (oxidative addition) arylpalladium halide complexes, L 1 Pd(Ar)(Br) [L = P(t-Bu) 3 , QPhos, and P(t-Bu) 2 Fc], formed the corresponding three-coordinate arylpalladium amido complexes as stable species at room temperature when the aryl and amido ligands bore highly deactivating groups. Upon thermolysis, these complexes underwent reductive elimination to form arylamines. For complexes substituted with the bulkier phosphines, the yield of the coupled product increased in the order: P(t-Bu) 3 > QPhos > P(t-Bu) 2 Fc. The bulky P(t-Bu) 3 based three-coordinate Pd amido complex undergoes reductive elimination with a t 1/2 of 20 min even at −10°C, while the dppf-based four-coordinate amido complex undergoes the same but at a temperature of 75°C and a t 1/2 of 55 min, once again showing the preference of monoligated, tricoordinate Pd for facile elimination in the catalytic cycle (Scheme 7). 41 To our knowledge, reductive elimination from arylpalladium thiolato complexes ligated by bulky monophosphine ligands has not been reported. 63 However, for C−O coupling reactions, Hartwig found that the yields of diaryl ether formed by reductive elimination from isolated arylpalladium aryloxo complexes can be significantly increased by addition of bulky alkylphosphine ligands. The conversion to diaryl ether was highest in the presence of the bulkiest ligand (Scheme 8). 64 In summary, palladium complexes ligated by bulky alkylphosphines are favorable for both the oxidative addition and reductive elimination steps in cross-coupling reactions via monoligated three-coordinate species.

The Role of Monoligated Pd in the Transmetalation Step
As discussed above, both the oxidative addition 65−67 and reductive elimination 37,68−71 steps have been studied in detail and are generally applicable to all cross-coupling reactions. However, the transmetalation steps are the least understood, as name reactions such as Murahashi (organo-Li), Kumada− Corriu (organo-Mg), Negishi (organo-Zn), Stille (organo-Sn), Hiyama−Denmark (organo-Si), and Suzuki−Miyaura (organo-B) involve different kinds of organometallic reagents (nucleophiles), where the intricate transfer of the organic moiety to Pd might differ from one metal to another. As early as 1983 and based on a limited study, Labadie and Stille provided some rationale for the transmetalation in the Stille coupling reaction. 72 More recently, Denmark conducted comprehensive investigations of the transmetalation step in the Hiyama−Denmark coupling. 73,74 These studies conclusively demonstrated that, in the cross-coupling with silanolate salts, two mechanistic regimes are operating, both of which involve a discrete Si−O−Pd linked intermediate in which the palladium atom is three-coordinate. 73,74 In one pathway, a neutral, four-coordinate silicon unit   75 undergoes intramolecular transmetalation, whereas in the second pathway, an anionic, five-coordinate silicon unit (10-Si-5) undergoes intramolecular transmetalation as well albeit at a significantly faster rate (Scheme 9). 73,74 The demonstration that silanolate cross-couplings proceed by intramolecular transmetalation of discrete Si−O−Pd intermediates motivated the Denmark group to investigate whether the related Suzuki− Miyaura cross-coupling proceeds by a related mechanism (vide infra). The dominance of the Suzuki−Miyaura reaction in industrial applications provided additional motivation for these studies. 76,77 Denmark's studies sought to provide a fundamental understanding of the transfer of the organic fragment from boron to palladium and were enabled by low-temperature and rapid injection NMR spectroscopic analysis (RI-NMR), combined with a series of structural, kinetic, and computational (DFT) investigations. Prior to Denmark's investigations to identify the "missing link" in the transmetalation step, the research groups of Suzuki and Miyaura, 78−80 Soderquist, 81 Amatore and Jutand, 82, 83 Schmidt, 84 and Hartwig 85 had provided justifications for one of two pathways, path A ("boronate") and path B ("oxopalladium") prior to transmetalation (vide infra). Soderquist's study gave indirect evidence for the Pd−O−B linkage, while Hartwig's kinetic study revealed that path B is favored over path A kinetically by more than 4 orders of magnitude. 85 However, the displacement of bromide with hydroxide from the oxidative addition complex has been found to have a transition-state Scheme 7. Relative Ease of the Reductive Elimination from Three-vs Four-Coordinate Arylpalladium Amido Complexes of the Type LPd(Ar)(NAr 2 ), Where L = P(t-Bu) 3 90 For PPh 3 -ligated complexes, Maseras, Ujaque, and co-workers carried out calculations on the displacement of bromide from the oxidative addition product of trihydroxyphenylboronate and palladium hydroxide complex and concluded that both pathways are capable of forming Pd−O−B linkages; however, this study had its limitations. 91 Harvey and co-workers have also studied computationally the effect of the size of the ligand on the transmetalation step by considering ligands of various electronic and steric properties, such as P(t-Bu) 3 , P(CF 3 ) 3 , PMe 3 , and PPh 3 , and found that the effect of steric bulk was twice as important as that of the electronic parameters. 92 Denmark's group has clearly shown that, in the reaction of trans-[P(i-Pr) 3 ] 2 (4-FC 6 H 4 )Pd(OH) with ArB(OH) 2 vs [P(i-Pr) 3 ] 2 (4-FC 6 H 4 )Pd(I) with [ArB(OH) 3 ] − , path B (oxopalladium) is preferred over path A (boronate) when ArB(OH) 2 is used as the nucleophile (Scheme 10). 93 However, when boronate esters are employed instead of boronic acids, the "boronate" mechanism is also possible as the reaction takes place under anhydrous conditions. 94 The primary conclusion from Denmark's studies is that the Suzuki−Miyaura cross-coupling does indeed proceed via discrete B−O−Pd containing intermediates just as in the Hiyama−Denmark reaction. Here again, two pathways were identified in which both tri-(6-B-3) and tetracoordinate (8-B-4) Pd complexes serve as pretransmetalation intermediates with the latter again undergoing more rapid transmetalation (Scheme 11). 93,95 A key reason for the slower transmetalation of the 6-B-3 intermediate is that it exists as a diligated palladium species (and the kinetic equation for the transmetalation shows an inverse first-order dependence on the phosphine ligand, P(i-Pr) 3 ). Calculated ground-state equilibrium energies for L 1 Pd and L 2 Pd based 8-B-4 complexes provided insight into the instability of the 8-B-4 complex during its synthesis with two ligands on Pd. The loss of water from an initially formed 8-B-4 complex (A), yielding 6-B-3 complex B, was found to be highly exergonic (Scheme 12). 93 Space-filling models clearly show that L 2 Pd(II) intermediate 6-B-3 (B) prevents the binding of the water to access species 8-B-4 as the OH groups on boron penetrate the van der Waals radii of the isopropyl methyl groups on phosphorus, thereby destabilizing the four-coordinate geometry. This was substantiated by the failed attempt to isolate 8-B-4 (A) from 6-B-3 (B) by addition of water or base. 93 This behavior illustrates the point that the steric bulk of the two P(i-Pr) 3 groups on the palladium in A prevents the formation of the requisite 8-B-4 species for rapid transmetalation. Removal of a phosphine ligand from the complex results in a monoligated complex (A′). Calculation of the ground-state energies of monoligated complexes A′ and B′ indicated a reversal of the position of the equilibrium, substantially favoring the 8-B-4 species (A′) with a monoligated Pd. This insight was crucial for focusing research efforts on the study of monoligated arylpalladium(II) complexes in order to form the long-sought-after 8-B-4 activated intermediate.
Denmark also provided experimental evidence for the formation of a dimeric palladium complex, C, from the monoligated precursor [(i-Pr) 3 PPd(4-FC 6 H 4 )(OH)] 2 by mixing the latter with one equivalent of boronic acid (Scheme 13). 93 This intermediate C was stable in the presence of additional boronic acid in THF. However, in THF−methanol solution, dimer C converted into the monomeric adduct in a 1:1 stoichiometry with ArB(OH) 2 to give the long-sought-after monoligated 8-B-4 complex A′ (see Scheme 11) as the pretransmetalation intermediate that is ready to undergo reductive elimination to the final coupled product.
This study also revealed that palladium complexes bearing various phosphine ligands such as PPh 3 , P(i-Pr) 3 , and DPPF also form Pd−O−B species as intermediates in cross-coupling reactions. The rate of the transmetalation is comparable for all ligands but decreases in the order Ph 3 P ≥ P(i-Pr) 3 > DPPF, demonstrating a common mechanism involving a coordinatively unsaturated and electrophilic palladium atom during the transmetalation process. 93 The ability to directly interrogate the transmetalation step kinetically enabled the investigation of the behavior of various boronic esters such as those of catechol and glycol. 94 Interestingly, these boron reagents undergo direct transmetalation via similar transmetalation intermediates but with increased rates as compared to the arylboronic acids under anhydrous conditions. This behavior is distinct from the study by Lloyd-Jones and co-workers conducted in the presence of water with organoboron reagents such as trifluoroborates 90 Suzuki−Miyaura cross-coupling that obviated the need for protodeboronation. 100 In summary, on the basis of comprehensive structural, kinetic, and computational investigations, it is clear that for both the Hiyama−Denmark and Suzuki−Miyaura cross-couplings, the transmetalation intermediate in the catalytic cycle is a monoligated L 1 Pd(II) species as opposed to the corresponding L 2 Pd(II) complex, irrespective of the size and coordination number of the palladium. The very recent study by Hirschi, Vetticatt, and co-workers 56 employing experimental and theoretical 13 C kinetic isotope effects showed that the transmetalation step likely proceeds via a tetracoordinate boronate (8-B-4) intermediate with a Pd−O−B linkage as already observed by Denmark. Having coordinatively unsaturated palladium is crucial for the rapid transfer of the organic group from silicon or boron to palladium.

ROLE OF THE LIGANDS IN FORMING THE L 1 Pd(0) PRECATALYST IN SITU
One of the significant advances in cross-coupling over the last two decades was the development of new ligands that make the technology viable for organic synthesis in both academia and industry. 11,101−104 The facile formation of the low-coordinate, monoligated Pd(0) species is the key factor in determining the efficiency and efficacy of the catalyst system. Until the late 1990s, the most commonly employed ligand for cross-coupling was PPh 3 , which was thought to form a 14-electron Pd(PPh 3 ) 2 complex as the active species during catalysis either in situ from Pd(OAc) 2 or Pd 2 (dba) 3 or from a precatalyst. However, recent computational 105 107 The major breakthrough in this area happened when the known, pyrophoric, and least-explored ligand tri(tertiary-butyl)phosphine, P(t-Bu) 3 , was utilized independently by Fu 5,6,108 and Koie 7,8 in situ for aryl chloride C−C and C−N bond formation. Up until then, it was difficult to effectively couple an aryl chloride under palladium catalysis due to the large C−Cl bond dissociation energy. 109 The unique role of P(t-Bu) 3 in facilitating the cross-coupling was attributed to its steric bulkiness and electron richness, which are related to its cone angle and pK a , respectively. 11 Today, P(t-Bu) 3 is considered a privileged ligand in cross-coupling. Fu also used PCy 3 in certain challenging cross-couplings, where PPh 3 was not effective; 108 however, P(t-Bu) 3 was far superior for challenging coupling reactions, including those of unactivated aryl chlorides. 5 In the same year, Buchwald reported a new type of biaryl ligand containing PCy 2 (DavePhos) for the Suzuki−Miyaura crosscoupling of unactivated aryl chlorides 9 as a modification to his BINAP systems, which had earlier been employed for the C−N coupling of aryl bromides. 110 Interestingly, DavePhos clearly demonstrates the balancing role of the sterics and electronics of ligands in cross-couplings, as PCy 3 alone cannot effect the analogous transformations for certain challenging substrates. If one considers steric factors alone, P(o-Tol) 3 , which has a larger cone angle than that of P(t-Bu) 3 , should facilitate aryl chloride couplings readily but does not. One year later, Buchwald incorporated the P(t-Bu) 2 moiety into the biaryl scaffold to generate P(t-Bu) 2 R (R = biphenyl), known as JohnPhos, which Chemical Reviews pubs.acs.org/CR Review allowed the room temperature amination of aryl chlorides to be carried out effectively. 111,112 These findings were the impetus for developing a new class of biaryl ligands, commonly known as Buchwald ligands, with members such as SPhos, RuPhos, XPhos, tBuXPhos BrettPhos, tBuBrettPhos, AdBrettPhos, GPhos, and others. Today, these are some of the most privileged classes of commercially available ligands for cross-coupling. 102,113−115 In 1998, Hartwig also created a new P(t-Bu) 2 R-type ligand called QPhos, in which R is a pentaphenylferrocene moiety. 10 This is now a commercially available ligand that has proved superior in challenging coupling reactions. Beller also disclosed the bulky electron-rich ligand PAd 2 (n-Bu) in 2000, 116 while Buchwald had reported a year earlier the synthesis of the P(t-Bu) 2 (biphenyl) for challenging cross-coupling reactions. 117 As the bulkiness of the ligand increases, the ligand tends to form coordinatively unsaturated Pd(0) species in the catalytic cycle, thereby increasing the oxidative addition and reductive elimination abilities, although the electronic properties of the ligand also play a key role in the oxidative addition step as we discussed above. The size of the ligand determines the "n" in L n Pd(0). That is why typically PPh 3 forms L 4 Pd(0), while P[3,5-(CF 3 ) 2 C 6 H 3 ] 3 forms L 3 Pd(0). 118 The bulky P(t-Bu) 3 forms L 2 Pd(0). 34 Figure 2 depicts the effect of ligand size on L n Pd(0) formation. As the steric bulk of the ligand increases further, as in BrettPhos, t-BuBrettPhos, GPhos, and others, the ligand tends to form "L 1 Pd(0)", and this has been invoked as an intermediate by desorption electrospray ionization mass spectrometry under C−C and C−N bond-forming reaction conditions 119 and by ESI MS under Suzuki−Miyaura coupling conditions. 120

Computational Prediction of L 1 Pd vs L 2 Pd Formation in the Catalytic Cycle by Using Ligand Parameters
Ligand parametrization for predicting the outcome of a catalytic reaction with the help of data science is emerging as a new tool in organic synthesis. 121−123 Although the Tolman cone angle 124,125 descriptor has been of great value in understanding and predicting ligand properties in catalysis, it suffers from certain drawbacks, one of which is not taking ligand flexibility into consideration. 126−128 Recently, the groups of Sigman and Doyle reported that the use of a single descriptor, minimum percent buried volume or %V bur (min), could nicely predict the ligation state of the catalytically active complexes in both Pd-and Nicatalyzed reactions. 129 Initially, the authors were inspired by the high efficiency of the DinoPhos ligand family (TyrannoPhos and TriceraPhos) in the Ni-catalyzed coupling reaction of acetal and boroxine (Scheme 14). 130 Even though DinoPhos ligands feature large cone angles, similar to P(t-Bu) 3 , interestingly they have small %V bur , similar to PPh 3 . The authors studied the ligation state, which has been reported to influence the reaction outcome, of various ligands with 4-fluorobenzaldehyde (Scheme 15) 129 and further discovered that the %V bur (min) descriptor would classify the ligands with a clear threshold. Any monodentate phosphine which has a %V bur (min) value below this threshold forms a bisligated complex and otherwise forms a monoligated complex. From a structural perspective, with the help of X-ray diffraction and DFT calculations, several L n Ni(4fluorobenzaldehyde) (n = 1 or 2) structures were revealed. For example, although PteroPhos (see Scheme 15) has a large cone angle compared to P(t-Bu) 3 , it still forms the L 2 Ni(0) complex. As can be seen, cone angle does not reflect the overall topology of the ligand structure; however, %V bur (min) captures the flexibility of the steric bulk after the first ligand coordination.
These studies indicated a possible region of %V bur (min) values in which L 2 M is thermodynamically favored in the resting state, but L 1 M is also found in solution. Within this region, the equilibrium between L 2 M and L 1 M would be influenced by such factors as the temperature, solvent, and concentration of the reaction. The concept of %V bur (min) has also been successfully expanded into the Pd-catalyzed cross-coupling reactions. When plotting the reaction yield against the ligand descriptor % V bur (min), clear thresholds that separate active and inactive ligands can be obtained with the directionality of active and inactive regions determined by whether the active catalyst is L 1 Pd or L 2 Pd. 131,132 In Scheme 16, part (a), L 2 Pd is identified as the catalytically active species, whereas L 1 Pd is identified as the catalytically active species in part (b). When the ligand sterics is not the decisive factor, e.g., Scheme 16, part (c), no threshold can be identified. 133 The authors proposed that this approach should facilitate mechanistic studies of related organometallic reactions and enable reaction development by identifying active and inactive as well as mono-and bis-ligating phosphines before synthesis. Even though the authors acknowledge that %V bur (min) may not capture reactivity trends across all phosphines, its ability to identify outliers (particularly false negatives) can spur the development of new descriptors and targeted mechanistic studies. Besides using a single descriptor to predict the ligation state of palladium, Schoenebeck and co-workers recently reported a compound descriptor method via an unsupervised machine learning process for predicting a rare situation of the palladium ligation state, Pd(I) dimer, rather than L 2 Pd. 134 Various recent review articles and book chapters have listed some of the most sought-after phosphine and heterocyclic carbene ligands for cross-coupling applications. Colacot's book chapter 11 and Nolan's book chapter 16 provide an update on the phosphine and NHC ligands, respectively, up to 2014, while Shaughnessy's 135 and Hazari's 103 reviews highlight recent updates in this area.

Introduction
As mentioned in the Introduction and in the section on mechanisms, it is clear that monoligated Pd(II) "T-shaped" intermediates are favored in all three steps of the catalytic cycle. The size of the ligand and the nature of the precatalyst play key Although various precatalysts (Scheme 17) had been used to generate "L 1 Pd(0)", its physical existence was not wellestablished until very recently. Once the coordinatively unsaturated monoligated "L 1 Pd(0)" is generated, it has to be trapped either as an oxidative addition complex, L 1 Pd(II)(Ar)-(X) (see the section on mechanism), or be coordinated with a neutral ligand such as an olefin to keep it stable as a Pd(0) complex; otherwise, it can disintegrate into palladium black instantaneously. Vilar's mini-review highlights the developments in this area until 2005; 136 however, the majority of the precatalysts featured in Scheme 17 have appeared in the literature only in the past 5 years.
Carrow's recent work has been crucial in establishing the existence of L 1 Pd(0) (L = P(t-Bu) 3 ; P(Ad) 3 ) via the synthetic route outlined in Scheme 18. 137

[L 1 Pd(I)X] 2 Dimers (X = Br, I) as Precatalysts.
Palladium typically forms complexes with 0, +2, or +4 oxidation states. Palladium(I) is relatively rare even in catalysis. The presence of unpaired electrons in Pd(I) potentially favors the formation of a dimeric complex, considering the thermodynamic stability of the dimer (Pd−Pd bond energy is usually worth about 25 kcal/mol). 138 Although over 50 Pd(I) dimers have been reported in the literature over the past decades, only very few are known to be catalytically active, 139,140 and a few good reviews of their structural diversity have been published. 141−144 Based on our knowledge in this area, [(P(t-Bu) 3 )Pd(μ-Br)] 2 is the first example of a well-studied monoligated Pd(I) precatalyst. 145 Although, as early as 1996, Mingos and coworkers had synthesized [(P(t-Bu) 3 )Pd(μ-Br)] 2 as an airsensitive, dark-green 16-electron complex, 39,146,147 its superior catalyst activity was disclosed by Hartwig and co-workers in 2002, 148,149 followed by Prashad and coauthors in 2003. 150 In 2012, Gooßen's group identified it as a highly active isomerization catalyst for the synthesis of enol esters from allylic esters. 151 [(P(t-Bu) 3 Pd(μ-Br))] 2 has been employed as a precatalyst, 145,152,153 which was assumed, and later confirmed, 154 to function as an in situ reservoir of the highly reactive 12-electron complex [P(t-Bu) 3 )Pd(0)], that readily activates aryl halides (Scheme 19). 139 This reactivity difference between Pd[P(t-Bu) 3 ] 2 and [(P(t-Bu) 3 Pd(μ-Br))] 2 was clearly demonstrated in the C−N cross-coupling reactions as well ( Figure 3). 32 The major applications of this unique catalyst in coupling reactions such as C−N cross-couplings, cyanation, thiolation, αarylation, and the Suzuki−Miyaura and other cross-couplings have been reviewed elsewhere. 2,135,139,145 Despite its unique and superior reactivity in challenging crosscoupling reactions, when compared to the in situ system or to Pd[P(t-Bu) 3 ] 2 , making this air-sensitive precatalyst commer- discovered an atom-economical way to make it by reacting Pd(cod)Br 2 or PdBr 2 with 1 equiv of P(t-Bu) 3 , followed by addition of NaOH (1 equiv) in methanol. 156 In 2017, while studying, with the aid of DFT calculations, the mechanism of this atom-economical and very interesting transformation in collaboration with Shoenebeck et al., it was found experimentally that the same transformation is possible with excellent selectivity by using 1.5 equiv of P(t-Bu) 3 as well. 157 Moreover, Goossen et al. published their findings in this regard in 2013. 158 As of today, and although all three processes are patented, these remain the best methods for preparing [P(t-Bu) 3 Pd(μ-Br)] 2 in very high yield (Scheme 20).
The corresponding iodo dimer, [P(t-Bu) 3 Pd(μ-I)] 2 , has been investigated as a cross-coupling catalyst by Schoenebeck and coworkers. 139, 159 Schoenebeck's group also developed a direct comproportionation method to synthesize the iodo-bridged dimer, [P(t-Bu) 3 Pd(μ-I)] 2 , while Colacot et al. developed three convenient routes either directly from PdI 2 or from PdBr 2 / Pd(cod) 2 Br 2 via the corresponding bromo dimer, [P(t-Bu) 3 Pd-(μ-Br)] 2 (Scheme 21). 157 In contrast to the Pd(I) bromo dimer, the Pd(I) iodo dimer was not reactive under comparable conditions. This explains why it had only one known application relating to carbonylations of aryl halides, 160 prior to the detailed investigations by Schoenebeck's group, who showed that the in situ release of the 12-electron-based Pd(0) species [PdP(t-Bu) 3 ] from these Pd(I) dimers is dependent on the adequate choice of additive. 161 Schoenebeck's study (Figure 4) using the N scale 162 identified the minimum nucleophilicity required to effect the activation in each case. Using nucleophiles with N ≥ 16, Pd(I) iodide dimer was successfully activated for the Suzuki−Miyaura crosscoupling. 161 Schoenebeck's group determined that DABCO (Evonik Operations GmbH), a nucleophile with an N-scale value of 18.80, 163 is suitable for the activation of dimer [P(t-Bu) 3 Pd(μ-I)] 2 in a mild and selective method for the direct aromatic C−H activation to form aryl germanes via the tetrafluorothianthrenium salt (Scheme 22). 139,164 The corresponding L 2 Pd(0) based precatalyst, [Pd(P(t-Bu) 3 )] 2 , was ineffective in this transformation.
Based on computational and experimental studies, Schoenebeck's group proposed a dinuclear mechanism during catalysis. 165,166 Transition-state calculations suggested that bond activation occurs primarily at one Pd center of the Pd(I) dimer. After oxidative addition, a Pd(II) dimer was computationally obtained, suggesting that an overall Pd(I)−Pd(I)/ Pd(II)−Pd(II) oxidative addition occurs involving both Pd centers vs a mechanism involving Pd(I)/Pd(III). Details of the application studies have been summarized in the mini-reviews by Schoenebeck 139    earlier, Watson's group developed the Pd(II) dimer [(JessePhos)PdI 2 ] 2 (JessePhos = tert-butyldi(3,5-di-tertbutylphenyl)phosphine) for the coupling of silyl iodides and alkenes with yields comparable to those obtained with in situ catalysis using JessePhos in conjunction with Pd 2 (dba) 3 but with significantly lower catalyst loadings. 168 However, these types of complexes are not completely new; many examples using NHC had been synthesized and characterized by Nolan's group as early as 2002. 169

Palladacycle Precatalysts
The development and application of palladacycles as precatalysts for cross-coupling reactions have been accomplished by several research groups; however, noteworthy original contributions came from the groups of Herrmann 174 and Bedford. 175 A few reviews have been published in this area. 176−178 These precatalysts were popular in academia during that period, mainly because of their stability to air and moisture and high TONs when employed in relatively easy coupling reactions. Shaughnessy's recent mini-review 135 offers an overview of this area; hence, we do not intend to duplicate the effort; rather, we wish to highlight the precatalysts that are utilized in the R&D and process chemistry laboratories.

Biphenylamine-Based Palladacycles.
Among the many examples of palladacycle precatalysts, amino-substitutedbiphenyl-based palladacycles have recently become one of the most popular classes of precatalyst. 179 Palladacycles derived from 2-(dimethylamino)biphenyl as a scaffold for use as precatalysts in cross-coupling applications such as the Buchwald−Hartwig, Suzuki−Miyaura, and enolate cross-couplings were originally developed by Nolan and co-workers by incorporating NHC-based ligands. 180−183 Nolan also proposed a mechanism for the formation of the highly active monoligated L 1 Pd(0) from the stable Pd(II)-based palladacycle in the presence of a strong base such as sodium isopropoxide (Scheme 24). 182 Buchwald and co-workers have developed several generations of palladacycles using similar or the same amine scaffold as in Nolan's system. The first two generations of Buchwald palladacycle are based on Pd−Cl systems with κ 2 -N,C phenethylamine (G1) 184 and 2-aminobiphenyl as the chelating N,C-ligand (G2), 185 respectively. Although the G1 and G2 classes of precatalyst are superior to the in situ generated systems that use precursors such as Pd 2 (dba) 3 or Pd(R-allyl)Cl, they nevertheless have many limitations. The multistep synthesis and scale-up of G1 is tedious, mainly because of the synthesis of the thermally unstable (TMEDA)PdMe 2 and its conversion into G1 (Scheme 25). 184 Moreover, the G1 precatalysts are slow to initiate reaction at room temperature in the presence of weak bases such as carbonate or phosphate.
While one important drawback of the G1 precatalysts is their limited ability to accommodate bulkier ligands, they are nevertheless superior in certain coupling reactions to the G2 185 and G3 186 counterparts because they do not produce  187 In addition, 4-aminobiphenyl, present as an impurity in 2-aminobiphenyl, is very toxic; however, one can overcome this by using high-purity 2-aminobiphenyl as the starting material. Consequently, Buchwald's lab developed the G4 and G5 versions of the palladacycle precatalysts. 188 The general synthesis and activation of G2−G5 complexes are summarized in Scheme 26.
Although the Suzuki−Miyaura cross-coupling is well established, the coupling of relatively unstable fluoroboronic acids and heterocyclic boronic acids is challenging because they tend to undergo protodeborylation as they cannot withstand high temperatures and longer reaction times. Therefore, the active catalyst in these reactions, [L 1 Pd(0)], has to be generated in a facile manner and needs to effect the coupling rapidly under milder conditions. The advantages of XPhos-based G4 and G5 precatalysts for a series of challenging cross-couplings of boronic acids with aryl bromides and chlorides are showcased in Scheme 27. 188 It is worth noting that the XPhos-based G2 precatalyst had earlier been effectively utilized by Buchwald for the Suzuki− Miyaura cross-coupling of unstable polyfluorophenyl and fivemembered ring 2-heteroaryl boronic acids with aryl bromides, chlorides, and triflates. 185 There are overlaps in reactivity between generations of G2− G5 precatalysts. G3s, however, appear to be a broader class of precatalysts because of the ability of the framework to accommodate bulky ligands such as tBuXPhos and tBuBrett-Phos. In addition, G3s appear to have wider applications in coupling reactions such as the selective monoarylation of CH 3 -CO-R compounds, the coupling of challenging boronic acids under milder conditions, and the C−N coupling of primary and secondary amines. 186 These G2−G5 complexes are relatively air-stable and commercially available in milligram-to-bulk quantities, although purity and reliability can vary from supplier to supplier.

Acetanilide-Based Palladacycles.
Carrow and coworkers have reported the synthesis and applications of a tri(adamantyl)phosphine-based acetanilide palladacycle ( Figure  5) and demonstrated its superior activity in the Suzuki−Miyaura coupling of aryl bromides and chlorides with the challenging polyfluorophenyl boronic acids including, C 6 F 5 B(OH) 2 . 189 Although the same group did provide a synthetic route for the known ligand PAd 3 , its synthesis in bulk quantity is not trivial. Prior to this report, in 2010, Kantchev and Wing prepared an IPr complex and compared the activity of several IPr-coordinated palladacycle precatalysts in the Suzuki−Miyaura and C−N/P/ S/O cross-couplings; they found that this catalyst system is superior to the corresponding PEPPSI, cinnamyl, and acac systems. 190

L 1 Pd-Based Acyclic Precatalysts
This section describes a class of monomeric precatalysts that represents analogous technological advancements in the synthesis of monoligated Pd complexes and their applications to cross-coupling reactions, with an emphasis on activity, selectivity, and efficiency. However, the precursor to these Pd complexes can be a monomer or dimer. This section also Chemical Reviews pubs.acs.org/CR Review explains how different types of precatalysts could be activated to L 1 Pd(0) while minimizing the "off-cycle species." We hope that using the information provided below, the users from both academia and industry can make a decision in selecting a specific cross-coupling reaction by generating the most appropriate active catalytic species.

Phosphine-Based π-(Allyl, Crotyl, and Cinnamyl) Precatalysts.
A notable report on the synthesis and applications of the tertiary-phosphine-based Pd(π-allyl)(L)Cl, where L = di(tert-butyl)(neopentyl)phosphine (DTBNpP), was published jointly in 2010 by Shaughnessy's and Colacot's groups. 191 Prior to this report, an important publication from Verkade's group described the synthesis and full characterization of moisture-and air-stable monoligated (η 3 -R-allyl)Pd(L)Cl complexes, where R is Me or Ph and L is the bulky phosphinimine ligand (t-Bu) 2 PN = P[N(i-Bu)CH 2 CH 2 −] 3 N (Scheme 28). 192 These complexes were successfully employed by Verkade for the amination of aryl bromides and chlorides, including challenging Buchwald−Hartwig aminations of sterically hindered amine and halide coupling partners.
To clarify the influence of phosphine ligand L and the R-allyl moiety on the reactivity of these complexes, Verkade's team conducted a series of control experiments. These investigations revealed that the ancillary cinnamyl moiety leads to reduction of the ligand loading by half in comparison to the in situ generated catalyst whereby 2 equiv of L are required to achieve high yields, regardless of whether the Pd source is the acetate or the chloride salt. This may be attributed to the need for an extra equivalent of the ligand to reduce Pd(II) to Pd(0) during the in situ catalysis involving both Pd(OAc) 2 and PdCl 2 , whereas preformed complexes of type B (see Scheme 28) are activated to L 1 Pd(0) in the presence of a base. The analogue of complex B in which the ancillary cinnamyl ligand is replaced with crotyl (possessing the smaller Me group) resulted in a slightly lower isolated yield of the Buchwald−Hartwig coupling product C (87%) compared with 98% for the cinnamyl analogue depicted in Scheme 28 192 In spite of the unique and superior reactivity of B and its crotyl analogue, they are, to our knowledge, not commercially available, presumably owing to the timeconsuming and tedious synthesis of the requisite ligand A.
In the aforementioned report, 191 191 Although L 1 Pd(0) has not been isolated or fully characterized, with the exception of Carrow's very recent work, 137 its presence has been well-documented by many (e.g., isolation and characterization of L 1 Pd(Ar)(X) by Buchwald 185 ). Precatalyst D, in the presence of a base such as NaOt-Bu gets activated to L 1 Pd(0), which can further react with the starting precatalyst D to form a "comproportionation dimer" (H) (see Scheme 30). The thermodynamic stability of the dimer, H, determines the  2 Np, (t-Bu) 2 (4-Me 2 NC 6 H 4 )P] to understand the role of the ligand and the substituents on the allyl group (Figure 7). 193 The X-ray crystal structures of the QPhos-based (π-R-allyl)Pd complexes (Figure 8) 193 revealed an increase in dissymmetry of the π-R-allyl moiety in the order Pd(allyl)(QPhos)Cl < Pd(crotyl)(QPhos)Cl < Pd(cinnamyl)(QPhos)Cl, not unlike what Nolan observed for his NHC-based systems. 194 The activation pathway involves either an alkoxide for chloride exchange (path A) or a nucleophilic attack by the tert-butoxide anion onto the π-allyl moiety (path B), leading to the L 1 Pd(0) active species (Scheme 31). 193,194 In agreement with Nolan's observation in his studies on the NHC systems, the dissymmetry could in theory affect the rate of the activation step and hence the rate of the coupling reaction. However, one should not extrapolate from these studies that the superior activity of the cinnamyl-based catalyst will apply to all cross-coupling reactions, because, in various cases, crotyl complexes give similar or better reactivities as exemplified by results from a model Suzuki−Miyaura cross-coupling reaction at room temperature (Scheme 32). 193 It is clear from this study that the crotyl-based catalysts appear to be superior to the other R-allyl catalysts. Surprisingly, the cinnamyl-based precatalyst showed lower activity in this Suzuki−Miyaura coupling, in contrast to Nolan's observation with the NHC systems. 194 In general, the size of L and the substituents on the allyl group are important factors that influence the activity of this class of precatalysts: Pd(crotyl)-(L)Cl, where L = QPhos, (t-Bu) 3 P, or (t-Bu) 2 NpP, stand out as performing the best.
In a related investigation of several catalysts in the Buchwald− Hartwig amination, Pd(crotyl)(QPhos)Cl performed the best, with very low loading (Scheme 33). 193 In contrast, the αarylation of 1-tetralone under analogous conditions (dioxane, 100°C) proceeded best with Pd(allyl)-based catalysts, whereas use of the Pd(crotyl)-based ones resulted in poor yields. 193 4.4.1.3. Role of the L 2 Pd 2 (allyl)X "Comproportionation" Dimer in Catalysis. One of the very interesting findings of Colacot's work is that all of the Pd(π-allyl)(L)Cl precatalysts investigated [L = (t-Bu) 3 P, QPhos, (t-Bu) 2 (Np)P, (t-Bu) 2 (4-Me 2 NC 6 H 4 )P] produce the corresponding Pd(I)(allyl)-based "comproportionation" dimers upon treatment with t-BuONa as well as during the oxidative addition step, whereas no such major comproportionation dimer is observed with Pd(crotyl)(L)X (X = Cl, Br) under identical conditions (Scheme 34). 193 The comproportionation dimer was proven, by isolation and X-ray crystal structure analysis, to be the main intermediate in the Buchwald−Hartwig amination of N-methylaniline with 4bromoanisole.

Pd(π-R-Allyl)(NHC)Cl Complexes.
Nolan's group can be credited with developing and studying the applications of Pd(π-R-allyl)(NHC)Cl complexes as precatalysts that generate the monoligated species, (NHC)Pd(0), in the context of crosscoupling reactions. In their landmark publication in 2006, 194 they screened the Pd(π-R-allyl)(NHC)Cl precatalysts in a model Suzuki−Miyaura coupling. This study demonstrated that precatalysts with an unsubstituted allyl are less active than those with a crotyl, cinnamyl, and prenyl analogues (Scheme 35). 194 Although these authors did not provide a rationale as to why allyl resulted in lower conversions, subsequent studies by Hazari 195,196 and later by Colacot 187 identified Pd(I) allyl dimer formation as a detrimental "off-cycle" species that leads to lower activity (see section 4.4.1.3 above). Additionally, Nolan applied these precatalysts in Suzuki−Miyaura cross-couplings of aryl chlorides and bromides at rt and at 80°C and in the Buchwald− Hartwig amination, including that of sterically hindered coupling partners. 194 In addition to the size of the R substituent on the allyl group, the size of the ligand also plays an important role in activating/ deactivating the catalyst. For example, tetra-ortho-substituted biaryl synthesis was accomplished by using [Pd(anti-(2,7)-SIc-OctNap)(cinnamyl)Cl] precatalyst by Dorta and co-workers (Scheme 36). 197 This catalyst was far superior (90% yield) than the corresponding IPr or SIPr precatalyst (33−35% yields) and Pd(IPent)PEPPSI)Cl (29% yield) for coupling of chloromesitylene with dimethylphenylboronic acid at rt with 2 mol % Pd loading. The superior activity of this catalyst was attributed to the percent buried volume (% V Bur ) of 42.0 vs 36.7 and 37.0 for IPr and SIPr, respectively. 198 Nolan applied this concept by utilizing the IPr* ligand (see Scheme 36 below), with a %V Bur of 44.6%, to couple a variety of challenging ortho-disubstituted aryl and heteroaryl halides with ortho-disubstituted phenylboronic acid at rt using 1 mol % Pd loading. Nolan and Chartoire have published a detailed account of the use of various NHC-based precatalysts, including the synthesis and novel applications of Pd(R-allyl)(NHC)Cl complexes; 16 hence, this chemistry will not be covered in this review and will be touched upon where warranted. 4.4.3. Indenyl NHC Complexes. Balcells, Hazari, and coworkers carried out very detailed experimental and computa- tional studies to understand the role of the Pd(I)(μ-allyl) dimer ("comproportionation" dimer) in catalysis by using NHC-based Pd systems. 195 They were able to prove that the Pd(I)(μ-allyl) dimers are directly observed during catalysis in reactions that utilize Pd(II)-based Pd(allyl)(L)Cl precatalysts and concluded that Pd(I)(μ-allyl) dimer formation is detrimental because it removes the [IPr−Pd(0)] active species from the reaction mixture (Scheme 37). 195 Their studies also clearly indicated that increased steric bulk at the 1 position of the allyl ligand in Pd(IPr)(η 3 -crotyl)Cl and Pd(IPr)(η 3 -cinnamyl)Cl results in a larger kinetic barrier to comproportionation. The slower rate of comproportionations in these two cases permits more of the active [IPr−Pd(0)] species to enter the catalytic cycle. Although Nolan's, Verkade's, and Colacot's groups had noticed the effect of bulky substituents attached to the allyl group, Balcells and Hazari were able to establish the negative role of the "comproportionation" dimer during the catalyst activation process. They found that the increased catalytic activity of the (η 3 -1-R-allyl)Pd complexes was correlated to an increased barrier to dimer formation via comproportionation. In a related study of the effect of the electronic and steric properties of the C-2 substituent in precatalysts of the type Pd(η 3 -2-R-allyl)(IPr)Cl, Balcells and Hazari found that the catalytic efficiency of the precatalysts is inversely related to the thermodynamic stability of the corresponding (μ-2-R-allyl)-bridged Pd(I) dimers. 196 Although (μ-allyl)-bridged Pd(I) dimers do function well as precatalysts in certain catalytic applications, 191,193,195 dimer formation is generally a nonproductive off-cycle pathway, and disproportionation back to [L 1 Pd(0)] and the ligated allylpalladium(II) complex is required for catalytic activity.

Catalyst Design Informed by the Mechanism of Precatalyst Activation.
Guided by an understanding of the off-cycle pathway leading to the Pd(I)-based comproportionation dimer, and aiming to improve catalytic efficiency, Melvin et al. 199 and DeAngelis et al. 187 developed two entirely different approaches to designing precatalysts that would not form the undesirable dimer.
Nova, Hazari, and co-workers' clever approach took advantage of the steric bulk inherent in the tert-butylindenyl motif to design and develop precatalysts of the type (η 3 -1-t-Buindenyl)Pd(L)Cl that do not form the corresponding inactive Pd(I) comproportionation dimers. 199 The precatalysts are either produced in situ or are easily accessed from the precursor (η 3 -1-t-Bu-indenyl) 2 (μ-Cl) 2 Pd 2 by reaction of the latter with a range of NHC or phosphine ligands, L. The higher activity observed for these precatalysts, when compared to the analogous ones generated from (η 3 -cinnamyl) 2 (μ-Cl) 2 Pd 2 is attributed to the bulky tert-butylindenyl hindering construction of the chloride bridge necessary for Pd(I) dimer formation. To this point, when unsubstituted indenyl was used as the auxiliary and IPr as the ligand, the Pd(I) comproportionation dimer was isolated in 85% yield after treating the precatalyst with 2 equiv of K 2 CO 3 in MeOH at rt for 2 h. In contrast, treating (η 3 -1-t-Buindenyl)Pd(IPr)Cl with 2 equiv of K 2 CO 3 under the same conditions led to Pd(0) products such as Pd(IPr) 2 and Pd black (the reactive [IPr−Pd(0)] is unstable in the absence of Ar−X and forms the Pd(0) products).
The authors demonstrated the superior performance of these NHC-supported precatalysts vis-a-vis their cinnamyl-supported analogues in a challenging Suzuki−Miyaura cross-coupling (Scheme 38). 199 In addition to the (η 3 -1-t-Bu-indenyl)Pd(NHC)Cl precatalysts, Nova and Hazari's team also successfully synthesized a series of complexes of the type (η 3 -1-t-Bu-indenyl)Pd(L)Cl, where L is an electron-rich and sterically demanding phosphine [L = SPhos, RuPhos, XPhos, DavePhos, P(t-Bu) 2 (4-Me 2 NC 6 H 4 ), PPh 3 , P(t-Bu) 3 , QPhos, PCy 3 , and P(o-Tol) 3 ]. 199 Similarly, these phosphine-based precatalysts did not form the inactive Pd(I) comproportionation dimer and proved highly active in a number of challenging cross-coupling reactions (Scheme 39). 199 Concurrently, Colacot's alternative approach to minimizing or even preventing Pd(I) dimer formation involved using Buchwald-type biaryl ligands possessing larger cone angles than those of the conventional monophosphines. 187 This research group carried out an extensive and systematic study of neutral Pd(R-allyl)(L)Cl complexes (R = H, Me, Ph; L = relatively less bulky biaryl ligand such as SPhos, RuPhos, XPhos, and BrettPhos). In the case of bulkier ligands, such as tBuXPhos, tBuBrettPhos, and AdBrettphos, precatalysts were designed and synthesized as cationic complexes in which the coordinating chloride anion was replaced with a noncoordinating anion such as TfO − . This substitution of the anion not only freed up more space in the coordination sphere of palladium, thus accommodating the bulkier ligands, but also prevented the formation of the "comproportionation" dimer ( Figure 9). 187 The X-ray crystal structure of one of these precatalysts is given in Figure  10. 187 In this study, Colacot's group demonstrated how to minimize the formation of the off-cycle allylPd(I) dimer by increasing the size of the R group on the allyl (similarly to Hazari's work with

Scheme 39. Representative Examples Showcasing the Superior Performance of Phosphine-Based (η 3 -1-(tertbutyl)indenyl)Pd(L)Cl (L = Phosphine Ligand) Precatalysts in a Number of Challenging Cross-Coupling Reactions
Chemical Reviews pubs.acs.org/CR Review the t-Bu-indenyl system) in conjunction with the size of the ligand, L, and, most importantly, by forming cationic complexes with a noncoordinating anion such as triflate, TfO − instead of a bridging coordinating anion such as Cl − . The importance of the size of the ligand L in disfavoring formation of the Pd(I) dimer is seen by comparing the X-ray crystal structures of Pd 2 (μallyl)(L 2 )(μ-Cl), where L = (t-Bu) 2 (4-NMe 2 C 6 H 4 )P vs L = SPhos ( Figure 11). 187 Application studies confirmed that even allyl complexes with smaller ligands such as SPhos can be made active by this simple approach (Scheme 40). 187 Examples of the applications of Pd(allyl)(XPhos)Cl and Pd(crotyl)(XPhos)Cl precatalysts in challenging ketone arylations, Suzuki−Miyaura cross-couplings, and Buchwald−Hartwig aminations are highlighted in Scheme 41. 187 It is worth mentioning in this context that, although the Buchwald−Hartwig amination is well established, the arylation of primary and secondary cyclopropyl amines had presented a significant challenge. 200−202 Applying this approach, Gildner et al. successfully effected the mono-and diarylation of cyclopropyl amines using the Pd-based technology discussed in this section (Scheme 42). 203 Subsequently, Stradiotto's group developed a related nickel-based cyclopropylamine arylation using an orthophenylene-bridged bisphosphine bidentate ligand. 204 In a related report, a tandem double amination protocol was described by Colacot and co-workers. The three-component, one-pot synthesis is catalyzed by [Pd(allyl)t-BuXPhos]OTf in the presence of RuPhos and provides amino aniline derivatives in high yields and chemoselectivity (Scheme 43). 205 At room temperature, the chloro-substituted (hetero)aryl bromide couples with the benzophenone imine to give an aniline surrogate as an intermediate. Subsequent heating to ca. 80°C, where ligand exchange presumably takes place, followed by hydrolysis provide the aniline derivative. This method is advantageous to keep the anilines protected, as some of them are susceptible to degradation with accompanying black color formation.

PEPPSI-Type Catalysts.
In 2006, Organ and coworkers reported a unique class of NHC (N-heterocyclic carbene) based air-and moisture-stable Pd catalysts called pyridine enhanced precatalyst preparation stabilization and initiation (PEPPSI; Total Synthesis Ltd) that can generate monoligated LPd(0) upon activation by a suitable base. 206 The PEPPSI precatalysts (commercially available from Sigma-Aldrich, now a part of Merck KGaA, Darmstadt, Germany) 207,208 proved to have broad applications in crosscoupling reactions such as the Kumada, Negishi, Suzuki− Miyaura, Heck, Sonogashira, and Buchwald−Hartwig crosscouplings. 21,22 The precatalyst, which has the general structure shown in Scheme 44, is synthesized by reacting an imidazolium salt with palladium chloride in the presence of K 2 CO 3 and a pyridine ligand. 206,209 These complexes are usually designated in short form based on the nature of the R group on the NHC unit; for example, PEPPSI-IMes, PEPPSI-IEt, PEPPSI-IPr, PEPPSI-IPent, and PEPPSI-IHept. However, Pd is sometimes used in front of the name. 209 3-Chloropyridine is a "throwaway" ligand that helps to stabilize the monoligated Pd by recoordination.   The two chlorides cause the Pd to be in the +2 oxidation state, thereby making the complex stable to air and moisture. It is worth noting that the PEPPSI precatalyst requires an external reductant such as a strong nucleophilic coupling partner or a base to activate it to the [(NHC)Pd(0)] active state. The reduction mechanism is also shown in Scheme 44. Early applications of the PEPPSI precatalysts were primarily in the cross-coupling of aryl halides and organometallic reagents such as organozincs. In these reactions, the organometallic reagent undergoes transmetalation with the precatalyst species to afford (NHC)PdR 2 (pyridine). Reductive elimination of R−R and dissociation of the pyridine ligand provide the active [(NHC)Pd(0)] species. 209 In C−N coupling reactions of alkylamines, the amine can serve as the reductant through βhydride elimination followed by deprotonation (see Scheme 44). 210,211 Although Organ's continuing contributions in this area are significant, newer versions of the PEPPSI precatalysts have also been developed by other groups by changing the sterics and electronics of the NHC ligands for specific substrates. For example, Nolan's precatalyst, based on PEPPSI-IPr* (IPr is modified to IPr* by changing the isopropyl group to a diphenylmethyl group), exhibited superior activity in C−N coupling reactions in comparison to IPr-and SIPr-based PEPPSI precatalysts. 210 Some of the important advances in the area of C−C and C−N couplings are only referenced here due to space limitations. 211−217 4.4.6. PEPPSI-Related NHC Catalysts. In addition to the conventional C−C and C−N coupling reactions, Szostak and co-workers reported PEPPSI-IPr as a highly active precatalyst in the direct Suzuki−Miyaura cross-coupling of a wide range of amides as substrates with various arylboronic acids to produce ketones in very good yields (Scheme 45). 218 The same group employed IPr(cinnamyl)PdCl as a precatalyst for the same transformation just prior to this work and claimed that the NHC systems are superior to the PR 3 systems. 219 Subsequently, Szostak's group expanded the work to esters as substrates, where C−O bond cleavage takes place to form ketones under the Suzuki−Miyaura cross-coupling conditions (Scheme 46). 220,221 The same year, Szostak's group reported a Pd-PEPPSI-IPr catalyzed Buchwald−Hartwig coupling of both common esters and amides via a highly selective C(acyl)−X (X = O, N) bond cleavage to rapidly access a variety of aryl amides. 222 Very recently, the same research group disclosed a Pd-PEPPSI-IPr catalyzed Suzuki−Miyaura cross-coupling of N-acylcarbazoles and N-acylindoles with arylboronic acids by a highly selective N−C(O) bond cleavage to produce aryl ketones in moderateto-excellent yields. 223 To improve the activities of the Pd-PEPPSI precatalyst in coupling reactions, Nolan and Cazin, 224 Navarro, 225 Shao, 226,227 and Organ 228 have reported on efforts to replace 3-chloropyridine in Pd-PEPPSI with other throwaway ligands such as P(OPh) 3 , Et 3 N, methylimidazole (MeIm), and morpholine ( Figure 12).
For example, Navarro's precatalyst is more active at 25 or 50°C than the corresponding PEPPSI-IPr in both the Suzuki− Miyaura and Buchwald−Hartwig cross-couplings of aryl chlorides. 225 This higher activity might be due to the facile dissociation of triethylamine (TEA) to generate the [L 1 Pd(0)] species or the facile recoordination of TEA to the Pd(0) species, thereby imparting more stability to the Pd catalyst during the catalytic cycle. 225 Organ's morpholine adducts of IPent and IPent Cl precatalysts provided a similarly efficient reduction pathway for catalyst activation through a β-hydride elimination. 228 The IPent NHC, with morpholine as the throwaway ligand, gave a 96% yield in the C−S coupling of 1-chloro-2,6dimethylbenzene with thiophenol at room temperature, while the highly active PEPPSI-IPent catalyst gave no conversion under these conditions. 228 Nolan and Cazin's precatalyst with P(OPh) 3 as the throwaway ligand was also effective at rt in the Suzuki−Miyaura coupling of aryl chlorides, although an alcohol was required to reduce Pd(II) to [(NHC)Pd(0)] with the formation of acetone. 224 These studies clearly reveal that a weak ligand is required to break up the dimer, [(NHC)PdCl 2 ] 2 , in order to form a monomeric tetracoordinate Pd as the precatalyst. It is important to note that [(NHC)PdCl 2 ] 2 can also act as a precatalyst to generate [(NHC)Pd(0)], but it is less active than the PEPPSI-type complexes; although there have not been enough control experiments. Because of space limitation, the published work in this area has not been covered in detail; the interested reader should consult the relevant reviews and book chapters for further insight into this area. 15 235−246 In this section, we shall focus only on the applications of L 1 Pd(Ar)X as efficient isolable or in situ generated precatalysts that are valuable in a number of organic transformations. As discussed in the section on mechanisms (section 2), transition-metal-catalyzed cross-coupling reactions involve three elementary steps, with the oxidative addition being the first step (see Scheme 1). 37 Although oxidative addition of a d 10 metal center and an aryl halide or pseudohalide can occur through three possible pathways (radical, S N Ar, and 3-center 2electron), Pd(0)/Pd(II) catalytic cycles generally involve a 3-center 2-electron transition state to yield oxidative addition complexes of the general formula LPd(Ar)X as intermediates.
As discussed in previous sections, studies have shown that the generation of catalytically active and coordinatively unsaturated monoligated Pd(0) species is crucial for the success of modern cross-coupling reactions. From mechanistic studies, it is well understood that L 1 Pd(II) species are involved in all three catalytic steps, namely oxidative addition, transmetalation, and reductive elimination (see section 2). While the [L 1 Pd(0)] species have not been isolated (with the exception of Carrow's work in detecting them), 137 the air-stable OACs such as L 1 Pd(II)(Ar)X, have been isolated and used as precatalysts that can lead to [L 1 Pd(0)]. In one of the earlier applications of a Pd(II)OAC, (t-Bu 3 P)Pd(Ph)Br was used as a Pd(0) precatalyst to carry out a chain-growth polymerization via the Suzuki− Miyaura coupling in the synthesis of polyfluorene. 247 Pioneering work by Buchwald's group uncovered the potential of oxidative addition complexes (SPhos)Pd(Ph)Cl and (XPhos)Pd(Ph)Cl as precatalysts for a rapid Suzuki− Miyaura coupling of unstable polyfluorophenylboronic acids (Scheme 47). 185 After demonstrating the stoichiometric reaction of 2,4-difluoroboronic acid with 4-chloroanisole using SPhos G6, the catalytic reaction was performed in a mixture of THF and 0.5 M aqueous K 3 PO 4 (1:2 ratio) to obtain 93% of the coupled product within 30 min at room temperature. Increasing the reaction time did not improve the yield but resulted in protodeboronation, while increasing the temperature also gave a lower conversion. Complete conversion of 4-chloroanisole at room temperature was achieved by using XPhos G6 instead of SPhos G6. Although this was a significant result for the challenging Suzuki−Miyaura coupling involving time-and temperature-sensitive boronic acids, the authors commented at that time "the preparation and isolation of an individual oxidative-addition complex for each substrate is clearly impractical and often impossible". 185 Related work from Buchwald's group has described the reductive elimination of L 1 Pd(Ar)F (L = bulky Buchwald The same year, Stradiotto's group prepared OAC, L 1 Pd(Ar) Cl by using L = Mor-DalPhos and characterized it by singlecrystal assays to be a square planar complex via N-coordination of the morpholine. However, the basic structure resembles the T-shaped geometry of oxidative addition complexes. The group also tested this as a precatalyst for the direct arylation of ammonia with deactivated aryl chlorides (Scheme 48). 249 Later, Buchwald and co-workers demonstrated the broad applicability of L 1 Pd(Ar)X complexes as precatalysts, 250 in which L 1 is a Buchwald ligand, for C−C, 185  In addition, the Pd G6 precatalysts offer several comparative advantages over the previous generations of Buchwald precatalysts. First, catalyst activation does not require a base and generates innocuous byproducts (as opposed to carbazole inhibitors as byproducts). Second, the Buchwald Pd G6 precatalysts are OACs, which means they are "on-cycle" intermediates, and typically provide higher reactivity and selectivity. Third, Buchwald Pd G6 precatalysts are prepared in a single step at room temperature. Fourth, the synthesis of Buchwald Pd G6 precatalysts allows for a versatile and tunable precatalyst design: (i) Each of the three ligands (L, Ar, X) can be independently fine-tuned. (ii) Improved solubility, greater stability, increased reactivity, and/or easier purification can be achieved by judicious selection of X, L, and Ar. (iii) Bulky ligands (e.g., L = t-BuBrettPhos, AdBrettPhos, and AlPhos) are easily accommodated in the structure of the precatalyst. While these incremental improvements in precatalyst technologies can be likened to the incremental improvements in smart phone technologies, 101 the syntheses of G6 catalysts are relatively more difficult to scale up for commercial use and hence process chemists need to carefully weigh the selection of one catalyst over another. Ingoglia and Buchwald have described the synthesis of the oxidative addition complexes of very bulky biaryl ligands such as t-BuBrettPhos and AlPhos and demonstrated their applications as effective precatalysts for C−N, C−O, and C−F crosscoupling reactions. 251 This technology is a convenient alternative to the previously developed technologies utilizing G1−G5 precatalysts, particularly in the case of the bulkiest biarylphosphine ligands, for which palladacycle-based precatalysts are difficult to isolate. The advantages of this technology are exemplified by the unique applications of AlPhos G6 to the effective fluorination of aryl halides and triflates (Scheme 49). 251 Encouraged by these results, Cernak, Buchwald, and coworkers devised an alternative approach for carrying out Pdcatalyzed cross-couplings of densely functionalized molecules by using stoichiometric quantities of palladium OACs as substrates. These OACs were formed from drugs or drug-like aryl halides. In most cases, these stoichiometric cross-couplings gave better results under milder conditions than their catalytic counterparts. The OACs are remarkably stable under ambient conditions, maintaining their reactivity after months of storage on the benchtop. These workers validated the utility of OACs in various reactions, including automated nanomolar scale couplings between a rivaroxaban-derived OAC and hundreds of diverse nucleophiles and in the late-stage derivatization of the natural product k252a. 256 Carrow and co-workers took a similar approach by utilizing (PAd 3 )Pd(4-C 6 H 4 F)Br as a highly efficient precatalyst for the room-temperature Suzuki−Miyaura coupling of aryl bromides and base-sensitive polyfluorinated arylboron nucleophiles that are very prone to protodeboronation. 257,258 The study claims that this unique catalyst system is superior in terms of efficiency to the in situ generated catalysts formed from PAd 3 and Pd 2 (dba) 3 and to other precatalysts such as (PAd 3 )Pd(η 3cinnamyl)Cl, [(P(t-Bu) 3 )PdBr] 2 , SPhos G2, XPhos G3, and IPr PEPPSI, thereby demonstrating the unique role of PAd 3 as ligand in the G6 technology.
Typically, the Buchwald−Hartwig coupling requires harsh inorganic bases; however, by choosing the appropriate ligand of the G6 system or (cod)-coordinated (LPd) 2 , milder organic bases such as DBU can be used (Scheme 50). 252 The same group expanded the scope of the work further to C−S cross-couplings at room temperature in the presence of soluble bases by using a G6-based biaryl ligand system. 255 Subsequently, a collaboration between the Buchwald and Jensen groups at MIT's chemistry and chemical engineering departments involved the utilization of an automated microfluidic optimization platform to determine the optimal reaction conditions for the cross-coupling of an aryl triflate with four types of commonly employed amine nucleophiles: anilines, amides, and primary and secondary aliphatic amines. 253 By analyzing trends in catalyst reactivity across different reaction parameters, such as temperature and base concentration, they were able to develop a set of general protocols for C−N crosscouplings that rely on organic bases. The optimization algorithm revealed that AlPhos G6 was the most active system in the coupling of each amine nucleophile. Furthermore, their automated optimization showed that the phosphazene base BTTP [tert-butylimino-tri(pyrrolidino)phosphorane] could be employed to assist the coupling of secondary alkylamines with aryl triflates. Very recently, Buchwald's group modified BrettPhos by introducing minor alterations in the biaryl group (replacing the MeO group ortho to PCy 2 with t-BuO and even replacing the i-Pr group at C′-4 with hydrogen), resulting in a new, monophosphine ligand, GPhos. The G6 complex of GPhos was utilized for the primary amination of aryl halides with loadings as low as 0.25 mol % (NaOt-Bu, THF, rt), and its activity was compared to those of Brettphos G6 and other new versions of G6 complexes (Scheme 51). 254 This work clearly demonstrates how one can improve the efficiency of the Buchwald−Hartwig amination, even when using inorganic butoxide base at rt, by the judicial choice of ligand and precatalyst.
In the same vein, Carrow and co-workers utilized an Ad 3 P G6 system to effect Buchwald−Hartwig aminations under mild conditions in which the beneficial roles that H 2 O plays were also highlighted. Thus, the cross-coupling of aryl amines, amides, and secondary amines with aryl bromides and chlorides was achieved in the presence of the weak, soluble base Et 3 N (Scheme 52). 259 The advantage of this technology is that the scope of the C−N coupling can be expanded to substrates with base-sensitive functional groups and that it can be more suited for flow chemistry where the handling of solids is a major impediment.
Using SPhos G6, research groups at Cornell and BASF jointly demonstrated that the halide salt, formed as a byproduct in the cross-coupling reaction, causes the transmetalation step to be reversible, and leads to strong reaction inhibition in the case of (hetero)aryl iodides. Kinetic and stoichiometric studies showed that halide inhibition likely results from the formation of the Scheme 50. Mild-Base-Assisted C−N Coupling Using L G6, Demonstrating the Dramatic Effect of the Ligand on the Outcome of the Cross-Coupling Scheme 51. Reactivity of G6 Precatalysts Supported by BrettPhos-Derived Ligands, in Particular GPhos G6, in the Room-Temperature Amination of Aryl Halides Using NaOt-Bu highly reactive Pd−OH intermediate being disfavored. By changing the solvent in the biphasic reaction system from THF to toluene, this inhibition was effectively minimized. The study also revealed that inhibition by halide is likely a more general problem in metal-catalyzed cross-coupling reactions, in particular the ones that involve a reversible transmetalation step. 260 In a joint effort by the Pentelute and Buchwald groups, OACs were employed for the modification of complex biomolecules via cysteine bioconjugation. Key features of the Pd(II)OACs employed are their ease of preparation, storage, and handling. In particular, these reagents enabled the synthesis of new classes of stapled peptides and antibody−drug conjugates. 261−263 The synthesis of disulfide-containing polypeptides has been a long-standing challenge in peptide chemistry, and versatile methods for the construction of disulfides are always in demand. Furthermore, a limited number of strategies are known for onresin formation of disulfides directly from their protected counterparts. Recently, Stockdill and co-workers disclosed a novel peptide modification method, whereby Pd-mediated onresin disulfide formation proceeds directly from the protected peptide without loss of any acid-labile side chain protecting group. 264−266 Pd(II)OACs could prove highly valuable in drug discovery efforts as demonstrated by Cernak, Buchwald, and co-workers, who developed a practical cross-coupling method that applies to densely functionalized targets, as would be required in late-stage diversification of pharmaceuticals and other biologically active compounds. 256 These workers generated a library of stable, isolable, and easy-to-handle Pd(II)OACs from complex, druglike aryl halides that can lead to the formation of C−C (alkylation and alkynylation), C−N, and C−S bonds, and to cyanation as well as carbonylative amination. While the substoichiometric reaction generally showed low-to-no conversions, the stoichiometric reaction provided moderate-to-high conversions.

Applications of OACs as Tools for Mechanistic and Kinetic Studies.
There are many examples in which arylpalladium(II) halide complexes are employed as tools for mechanistic and kinetic studies. Seminal work by Hartwig has provided a better understanding of the factors that affect the reversibility of the oxidative addition of Pd(0) to ArX. 267−269 In particular, Hartwig showed that Pd(II)OACs of the type Pd(Ar)(L)X (Ar = o-Tol, L = P(t-Bu) 3 , and X = I, Br, and Cl) undergo reductive elimination to regenerate the starting ArX under certain reaction conditions induced by the addition of a ligand and reaction temperature. 269 This study led to the following noteworthy conclusions: (i) While monomeric, threecoordinate arylpalladium(II) halide complexes had previously been proposed for cross-coupling reactions utilizing sterically demanding phosphine ligands, 108,270 this was the first time these species were isolated for study. 269 (ii) Reductive elimination of ArCl is more thermodynamically favorable than reductive elimination of ArBr and ArI, reflecting the Ar−X bond strength.
(iii) The observed kinetics show that ArBr eliminates faster than ArCl due to the stability of the palladium haloarene intermediate. 269 (iv) A ligand screen highlighted the positive impact of steric bulk on the efficiency of the reductive elimination.
In 2017, Shaughnessy and co-workers studied the mechanism of the Buchwald−Hartwig amination of aryl halides with anilines utilizing [(PNp 3 )Pd(Ar)(μ-X)] 2 as the Pd G6 precatalyst (Scheme 53). 54,55 The reaction of sterically hindered aryl bromides with derivatives of anilines occurred within 5 min, in contrast to that utilizing the L 2 Pd(0) based Pd(PNp 3 ) 2 system which required ca. 1 h at 80°C for the same conversion with 1 mol % loading. These results offered a direct comparison of the turnovers (specifically TOFs) associated with the facile generation of [L 1 Pd(0)] species from the G6 complexes versus that from the L 2 Pd(0) complexes. They also indicated that the type of aryl halide, the steric demand of the aryl halide and aniline, and the choice of ligand affect reaction efficiency.

AQUEOUS-PHASE CATALYSIS USING L 1 Pd(0) SPECIES
The use of water as a solvent in organic synthesis for the largescale application of green chemistry technology has been known for a few decades and is well documented. 271 In the early days, water was introduced as part of a biphasic system, and the organometallic catalysts (preformed or generated in situ) employed were engineered to be water-soluble to interact with the reactants, while the products formed migrated to the organic layer and got separated ( Figure 13). The catalyst in the aqueous phase was typically recycled, and the number of recycles depended on the life of the catalytic species. Commonly, these catalyst systems were simple metal salts with or without phosphine or related ligands and in which the ligands were typically made water-soluble by attachment of polar functional groups such as sulfonate, carboxylate, ammonium, phosphonium, or hydroxyl. 272,273 Kobayashi's recent review highlights the importance of creating a sustainable society by carrying out organic reactions in water, even though many reactants and catalysts are incompatible due to their immiscibility and/or degradation in water. 274 Kobayashi states that, "After the "watershed" in organic synthesis revealed the importance of water, the development of water-compatible catalysts has flourished, triggering a quantum leap in water-centered organic synthesis". He goes on to say, "Given that organic compounds are typically Chemical Reviews pubs.acs.org/CR Review practically insoluble in water, simple extractive workup can readily separate a water-soluble homogeneous catalyst as an aqueous solution from a product that is soluble in organic solvents".

Water-Soluble Ligands and Catalysts
Although numerous examples exist of water-soluble ligands and catalysts that are suitable for cross-coupling applications, this review will focus only on precatalysts that give rise to L 1 Pd(0) catalytic systems. This is important, as L 1 Pd(0) systems are expected to be extremely reactive to air and presumably to moisture under normal conditions. The following examples are intended to highlight the practicality of these catalytic systems in cross-coupling reactions.

PEPPSI-Type Soluble Catalysts.
Examples of Nheterocyclic carbene (NHC) based precatalysts made watersoluble by incorporation of an SO 3 Na functional group were disclosed by Poẗhig, Kuḧn, and co-workers in 2014 (Scheme 54). 275 These PEPPSI-type precatalysts were utilized in the Suzuki−Miyaura cross-coupling at room temperature in water and in air. Palladium complex III exhibited the best catalytic activity in the cross-coupling of aryl bromides with boronic acids at a low catalyst loading of 0.1 mol %. Complex III could be recycled at least four consecutive times without significant loss of activity, thereby reducing the effective loading to as low as 0.025 mol %. Its higher catalytic activity was attributed to the bulky and electron-rich isopropyl groups on the benzene ring attached to the NHC. Based on TEM analysis and kinetic and mercury poisoning experiments, the authors posited that Pd nanoparticles formed during the reaction, presumably from L 1 Pd(0), are responsible for the observed catalytic activity.
A new air-and moisture-stable PEPPSI-type complex, [Pd(L)Br 2 (Py)] [L: 3-(2-fluorobenzyl)-1-(4-methoxyphenyl)-1H-imidazoline-2-ylidene] was utilized to catalyze the Mizoroki−Heck cross-coupling reaction of aryl bromides and iodides with styrene in water. According to the authors, this is the first report of a Pd-PEPPSI-type catalyst successfully employed in the aqueous-phase Mizoroki−Heck reaction. Good-to-excellent yields of the coupled products were obtained for a range of aryl bromides and iodides at 100°C and with 1 mol % catalyst loading. 276

Micellar Technology
As enumerated in Kobayashi's review, there exist today several technologies for carrying out catalysis in water. 274 Of these, surfactants have been an important part of a simple way to solubilize hydrophobic substrates in water by forming emulsions. The dissolution of the catalyst and reagents takes place in nanosized apolar aggregates formed by the surfactant in the aqueous medium via intermolecular interactions such as ion pairing and hydrophobic effects. These interactions mimic somewhat the biosynthesis that takes place in nature through enzymatic action in the aqueous medium. Consequently, the design and development of novel surfactants have been taking place in earnest for the purpose of engineering micelles that can promote organic synthesis in a way that competes favorably with traditional catalysis in organic solvents ( Figure 14). 277 Chemical Reviews pubs.acs.org/CR Review often as four times; however, unlike in the case of the aforementioned water-soluble catalyst systems, the catalyst here could not be recycled. In an aqueous system containing the proposed "micellar nanoreactors", 282,283 solubilization of the precatalyst is a crucial parameter to consider. In this regard, the underpinning of efficient Suzuki−Miyaura couplings in water is the binding constant of a reagent to the micellar inner core: the greater the incentive to enter the site of reaction, the more catalytic activity is to be expected and the lower the catalyst loading. Thus, the two isopropyl groups on the biphenyl moiety of the palladium complex make it more lipophilic, which leads to better activity and efficiency in the Suzuki−Miyaura cross-coupling (see Scheme 55). The low loading and the recycling of the aqueous reaction mixture, involving the same reactants or different coupling partners, were demonstrated several times. However, because the authors had to add catalyst in each recycling run, the recycling was done to reuse the surfactant and water only. The products were easily isolated by a very simple workup, and because no organic workup was needed, an E factor 284,285 of zero was assumed based on the reaction conditions. 281 Even when aqueous waste is factored in, the E factor remained considerably low at 1.7. Low-temperature microscopy (cryo-TEM) established the nature and size of the micellar particles acting as nanoreactors. ICP-MS analyses of residual palladium in the coupled products indicated very low levels of Pd that are within the U.S. FDA's allowable exposure limits (<100 μg/day). 286−288 Lipshutz's lab further modified the palladacycle precatalyst, whereby one N−H of the biphenyl amino group was replaced with i-Pr, while one of the i-Pr groups on the biphenyl moiety was removed, unlike in the previous system. Interestingly, the Ni-Pr substituted version of the precatalyst (A), with the new EvanPhos ligand that is relatively easy to make in comparison with the HandaPhos ligand, 289,290 gave 97% conversion vs 1% conversion for the conventional unsubstituted palladacycle (B) (Scheme 56). 281,291 However, the Pd loadings were a few orders of magnitude higher than those of the HandaPhos system described earlier. 281 Surprisingly, the analogous EvanPhos− palladacycle complex, with an i-Pr group on each ring of the biphenyl moiety, gave much inferior results, presumably due to the mismatch of sterics to be able to fit into the "nanoreactors". 291 In addition, changing the isopropyl to a t-Bu group 291 also gave inferior results.

Application of L 1 Pd[π-(R-allyl)](X) Systems in Cross-Couplings in Aqueous Media.
Surfactant-based micellar technology has also been very effective for conducting various cross-couplings in water using L 1 Pd[(π-(R-allyl)](X) precatalysts (R = H (allyl), Me (crotyl), Ph (cinnamyl); L = biaryl ligand; X= Cl, OTf) developed by Colacot and coworkers. 187 Lipshutz and co-workers screened a variety of precatalysts, including mono-and biscoordinated Pd complexes, for the Buchwald−Hartwig N-arylation of indoline, a challenging model substrate, in aqueous medium. Among the various ligands tested, t-BuXPhos stood out as the best ligand, and the cationic (t-BuXphos)Pd(π-cinnamyl)(OTf) precatalyst led to a quantitative (NMR) yield of the C−N coupling product. Moreover, the cationic crotyl and allyl counterparts gave slightly lower (NMR) yields, 93% and 95%, respectively (Scheme 57). 292,293 It is worth noting that the neutral Buchwald (t-BuXPhos)Pd G1 and G3 complexes also gave high yields. Handa and co-workers have also achieved the sp 2 −sp 3 coupling of nitroalkanes with aryl bromides in water by utilizing L(π-allyl)PdOTf 187 in conjunction with surfactant PS-750-M (FI-750-M) to mimic polar solvents such as DMF and 1,4dioxane (Scheme 59). [277][278][279]294 This method proved superior to the conventional technology that employs the Pd 2 dba 3 / XPhos system with 10 mol % Pd loading under glovebox conditions in anhydrous 1,4-dioxane at 80°C. 277−279 Following screening of various types of Pd sources while keeping t-BuXPhos as the ligand, the in situ catalysis employing Pd 2 dba 3 or Pd(OAc) 2 as the Pd source resulted in inferior yields even at higher Pd loadings in comparison to the Pd(π-allyl)(t-BuXphos)OTf system. In contrast to Lipshutz's observations that substituents on the allyl group improve the activity and efficiency of the catalyst, 292,293 the presence of an unsubstituted allyl group seems to be important in Handa's results. However, Handa's team did not screen the corresponding crotyl or cinnamyl Pd complexes with OTf as the counterion. Compared to chloride, the OTf counterion typically imparts a cationic character to the Pd complex. It is worth noting that when the same workers added propylene and allyl bromide to the in situ system containing Pd(OAc) 2 and t-BuXPhos, comparable conversion was observed, albeit with a higher Pd loading. In the proposed mechanism (Scheme 60), 294 it is postulated that the allyl group remains attached to the Pd throughout the catalytic cycle. However, detailed mechanistic studies are still needed to support this hypothesis, because, in such transformations, the typical oxidative addition takes place on L n Pd(0). Slow conversions were observed when these workers tried full recycling of the catalyst. However, the greenness of the process was demonstrated when they carried out full reaction medium recycling together with partial Pd recycling, resulting in a low E factor of 5.3 (when the solvent used in the chromatographic separation was recovered) or of 18.4 (chromatography solvent not recovered in the last cycle). Based on the results from Handa's and Lipshutz's studies, it would appear that the catalyst is getting inactivated after each reaction cycle. This is not surprising as the L 1 Pd(0) species is highly reactive and hence susceptible to degradation.
Despite the significant advances achieved in understanding and applying cross-coupling chemistry, the sustainable crosscoupling of quinoline and isoquinoline had remained under-  298 In this type of chemistry, the size of the ligand seems to be important. The best result (100% conversion) was obtained with the medium sized PCy 3 ligand in Cy 3 PPd(crotyl)Cl, while a good conversion (84%) was observed with the bulkier (t-Bu) 3 P based catalyst, (t-Bu) 3 PPd-(crotyl)Cl. Among the Buchwald ligands, SPhos, XPhos, and RuPhos gave 90−94% conversions, while the bulky t-BuBrettPhos gave only a trace amount of the cross-coupling biaryl product. 298 One of the unprecedented findings in this area is the formation of ultrasmall palladium nanoparticles (Pd NPs) under micellar conditions from the precatalyst XPhosPd(crotyl)Cl (Scheme 61). 299 The authors noted that only π-allyl complexes 187 formed Pd NPs, whereas other phosphine-PdCl 2 or phosphine-Pd(OAc) 2 complexes proved ineffective in this regard. The presence of the crotyl group in the precatalyst is key to the fast reductive elimination of crotyl chloride, resulting in formation of the Pd NPs both in the small (1 g) and large (20 g) scale runs.
The composition, ligation, morphology, and size distribution of the Pd NPs were determined by employing analytical techniques such as 31 P NMR spectroscopy, high-resolution transmission electron microscopy (HRTEM), scanning transmission electron-microscopy-based high-angle annular darkfield imaging (STEM-HAADF), and energy-dispersive X-ray spectroscopy (EDX) mapping. Very interestingly, the XPhoscomplexed ultrasmall Pd NPs showed a single peak at 43.1 ppm (confirming the binding of the phosphine ligand) vs −12.9 ppm for the free XPhos; Ph 3 P in a sealed capillary served as the internal standard (−6 ppm).
The Pd NPs formed from XPhosPd(crotyl)Cl catalyzed the αarylation of nitriles in aqueous micellar medium using PS-750-M as surfactant. There is evidence that Pd-bound carbanions or keteniminates are formed as intermediates, which are stabilized inside the hydrophobic core of the micellar environment and thus protected from quenching (protonation) by water. The generality of the reaction was established with about 35 examples, including one on a 50 g scale (Scheme 62). 299 The details of the cross-coupling reaction pathway, such as oxidative addition to Pd NPs, transmetalation, and reductive elimination, were established with control 31 P NMR experiments on a stoichiometric variant.
The broader reactivity and applicability of the Pd NPs catalyst system and micellar conditions were demonstrated in the synthesis of biaryl ketones in 65−78% yields by a one-pot αarylation of nitriles with heteroaryl bromides followed by oxidation with elemental oxygen. More importantly, in a proofof-concept experiment, the protocol was shown to be effective for the Buchwald−Hartwig amination of indoles with aryl bromides (Scheme 63). 299

SELECTED INDUSTRIAL APPLICATIONS OF L 1 Pd(0) PRECATALYSTS
Several industrial applications of the methods and protocols described in the preceding sections have been reported. In Table  1   acid that is available only in a 90% purity from commercial sources with varying amounts of anhydride impurity. 301 In the Heck coupling step en route to Letermovir (Table 1, entry 12), instead of the Pd(OAc) 2 /P(o-Tol) 3 catalyst system 302 with a possible loading of 9 mol %, only 0.2 mol % of (t-Bu) 3 P G2 was needed with significant improvement in reaction time, i.e., 5 h vs 48 h (Scheme 65). 303 Both preformed and in situ formed 12-electron catalysts have been effective in the indole ring formation for the synthesis of the API Lirametostat (Table 1, entry 13) (Scheme 66). 304,305 This step employed RuPhos G3, which is an air-stable precatalyst in comparison to the one from the pyrophoric P(t-Bu) 3 . 306 A one-pot tandem Suzuki/oxetane ring-opening/cyclization has been employed to form the eight-membered ring in the synthesis of Inavolisib (Table 1, entry 14) (Scheme 67). 307 In the synthesis of Paxalisib, replacing (dppf)PdCl 2 ·CH 2 Cl 2 with XPhos G2 facilitated the Suzuki−Miyaura coupling with a lower loading of 0.5 mol % compared with the 2 mol % loading

SUMMARY AND OUTLOOK
Although about 4−5 Nobel Prizes in chemistry have been awarded to homogeneous catalysis from 2001 to 2021, the 2010 Nobel Prize winning technology, namely the Pd-catalyzed crosscoupling for carbon−carbon bond-forming reactions is the most practiced reaction in both academia and industry. This technology is also an example of how incremental innovation, with contributions from many pioneers, has led to a Nobel Prize. Moreover, the Murahashi reaction has recently been resurrected as the Murahashi−Feringa coupling with the significant contributions of Nobel Laureate Feringa. 311,323 The C−N cross-coupling has also become very popular and industrially relevant; hence, we are anticipating another Nobel Prize in cross-coupling in the near future. These innovations in all the areas of cross-coupling are due to the continued development of new ligands and catalysts. 101 Over the past decade, several new precatalyst technologies, such as Buchwald's palladacycle technology, Colacot's work on allyl/crotyl/cinnamyl based cationic complexes, and Buchwald's OACs (G6), have been developed to create precursors for L 1 Pd(0) catalysts containing phosphine ligands as outlined in this review. Additionally, technologies employing NHCs for generating L 1 Pd(0) catalysts have been developed by Organ, Nolan, Szostak, and Hazari. The choice of ligand for the construction of the desired precatalyst is crucial for achieving better selectivity, reduced catalyst loading, and milder reaction conditions, all important considerations for successfully and sustainably carrying out a challenging reaction for real world applications. The facile activation of the precatalyst to L 1 Pd(0) varies from one system to another and, therefore, the ease of synthesis of the ligands and precatalysts is also important.
With the aid of mechanistic and kinetic studies, our understanding of the cross-coupling mechanism has evolved somewhat, whereby L 1 Pd(0) seems to be the active species in the catalytic cycle for all three major steps (oxidative addition, transmetalation, and reductive elimination), even with smaller ligands such as Ph 3 P. However, sterically demanding ligands help to form L 1 Pd(0) more easily than the less bulky ligands such as Ph 3 P. Having said that, the stability of the L n Pd(0) species is important for higher TONs, as L 1 Pd(0) may get deactivated depending on the reaction conditions and the type of ligand used. Based on Hirschi and Vetticatt's recent studies combining DFT calculations and experimental 13 C kinetic isotope effects, the proposed catalytic cycle even for Ph 3 P-based systems is shown in Scheme 5 (cf. section 2.1), where L 1 Pd(0) is the active catalytic species after the first cycle in the absence of excess Ph 3 P. 56 AI based computational and machine learning technologies have emerged as powerful tools in predicting the active ligand/ catalyst system by making use of ligand parametrization. In this context, Gensch, Sigman, Aspuru-Guzik, and co-workers very recently presented "kraken", a discovery platform for monodentate organophosphorus(III) ligands that provides comprehensive physicochemical descriptors on the basis of representative conformer sets. By utilizing quantum-mechanical methods, they were able to calculate descriptors for over 1500 ligands, including some representative commercially available ones, and used machine learning simulations to predict properties of over 300 000 new ligands. 121,123 Our company (MilliporeSigma, business of Merck KGaA, Darmstadt, Germany) is expanding the scope of this AI-based digital technology for commercialization to help industrial and academic customers to accelerate their research. Related predictive studies by Sigman, Doyle, and Schoenebeck, already discussed in section 3, clearly show that newer parameters beyond Toleman's cone angle can be utilized to predict the formation of monoligated Pd complexes generating 12-electron-based catalytic systems. However, to be able to make AI-based, foolproof predictions, there is still a need to incorporate substrate parametrization, reaction conditions, and clean and reliable data, in addition to ligand and catalyst parametrization. Nevertheless, these methods have the potential to be employed for industrial applications with a focus on identifying suitable precatalysts, thereby minimizing Pd loadings, reducing harsh reaction conditions, and increasing reaction output in terms of yield and selectivity. MilliporeSigma (Merck KGaA, Darmstadt, Germany) is currently working on commercializing a digital tool based on AI technology which will support both academic and industrial customers to optimize With the increase in palladium metal prices in recent years, the cost-effective high yield synthesis of new-generation monoligated precatalysts is crucial for the cross-coupling technology to grow further. One way this could be achieved is by minimizing the number of steps in manufacturing commercial catalysts, which in turn minimizes Pd losses during the process. Also developing technologies for low loading Pd catalytic processes as well as efficient and cost-effective Pd capture and recovery methods need to be developed to efficiently recycle the expensive palladium metal. In addition, alternate technologies involving earth abundant metals, photocatalytic and electrochemical methodologies and enzymatic methodologies need to be developed from a sustainability point of view.  He has also carried out postdoctoral work in the U.S. and is a fellow of the Royal Society of Chemistry. He has given more than 400 presentations globally.