Improving Homogeneous Cationic Gold Catalysis through a

2 days ago - ... new approaches to catalysis and green chemistry, and the study of biologically active natural products from Peruvian medicinal plants...
14 downloads 0 Views 5MB Size
Article Cite This: Acc. Chem. Res. XXXX, XXX, XXX−XXX

pubs.acs.org/accounts

Improving Homogeneous Cationic Gold Catalysis through a Mechanism-Based Approach Zhichao Lu,† Gerald B. Hammond,*,† and Bo Xu*,‡ †

Department of Chemistry, University of Louisville, Louisville, Kentucky 40292, United States Key Laboratory of Science and Technology of Eco-Textiles, Ministry of Education, College of Chemistry, Chemical Engineering and Biotechnology, Donghua University, Shanghai 201620, China

Acc. Chem. Res. Downloaded from pubs.acs.org by UNIV AUTONOMA DE COAHUILA on 04/19/19. For personal use only.



CONSPECTUS: Homogeneous gold catalysis is regarded as a landmark addition to the field of organic synthesis. It is the most effective way to activate alkynes for the addition of a diverse host of nucleophiles. However, the literature reveals that a relatively high catalyst loading is needed in many gold-catalyzed applications (1−10 mol %), which is impractical in large-scale synthesis or multistep synthesis because of the high price and recyclization difficulty of the gold. A more thorough understanding of the factors that operate on homogeneous gold catalysis can provide better guidelines for the future design of more efficient gold-catalyzed reactions. In this Account, we will summarize our group’s extensive investigation of factors impacting cationic gold catalysis, namely, the effects of ligands, counterions, additives, and catalyst decay and deactivation, using a mechanism-based approach with the aim of improving the efficiency of homogeneous gold catalysis. Through NMR-assisted kinetic studies, we investigated the above factors. Our systematic ligand effect investigation provided a clearer understanding of how ligands influence each of the three stages in the gold catalytic cycle. On the basis of this study, we synthesized a novel phosphine ligand and achieved parts per million-level gold catalysis by manipulating the electron density of the substituents and the steric strain around phosphorus. Our investigation of counterion effects led to the design of a gold affinity index and hydrogen-bonding basicity index for counterions, which can forecast the reactivity of counterions in cationic gold catalysis. We studied the adverse silver effects in cationic gold catalyst activation and proposed a more efficient practical guide. Our additive effect investigation revealed that additives that are good hydrogen-bond acceptors increase the efficiency of gold-catalyzed reactions in those occurrences where protodeauration is the rate-determining step. The first detailed experimental analysis of gold catalyst decay and the influence of each component in the reaction system (substrate, counterion, solvent) on the decay process was also conducted. We found that high-gold-affinity impurities (halides, bases) in solvents, starting materials, filtration, or drying agents decrease the reactivity of a gold catalyst but that a suitable acid activator can reactivate the gold catalyst and enable the reaction to proceed smoothly at competitively low gold catalyst loadings. The effects of acid additives were also systematically investigated using typical reactions. We are convinced that better mechanistic understandings will offer clearer guidelines for the search for more efficient goldcatalyzed reactions. reacts with an electrophile (E+) to yield the final product via protodeauration with simultaneous regeneration of the cationic gold species (stage 2). Most, if not all, gold-catalyzed reactions show decay or deactivation of the gold catalystlikely through the reduction of cationic gold to gold(0) (stage 3). Most homogeneous gold-catalyzed reactions require a relatively high catalyst loading (1−10 mol %), which is impractical in large-scale synthesis or multistep syntheses.10 In a homogeneous gold catalytic cycle, several factors may contribute to the low turnover number (TON) (Scheme 1). Silver-mediated halogen abstraction is regularly used to generate cationic gold species from a gold catalyst precursor. However, in the case of in situ activation, excess amounts of silver salt precipitate

1. INTRODUCTION Few metals have had such an influential role in human history as gold. Despite its ubiquitous presence in jewelry and currency, not until the end of the last century did homogeneous gold catalysis rouse excitement among synthetic chemists.1−5 This excitement was due to the unique property of gold(I) species to act as π-soft Lewis acids toward unactivated carbon− carbon multiple bonds (alkynes, allenes, alkenes). Upon activation by gold, the addition of a diverse host of nucleophiles, either inter- or intramolecularly, under mild reaction conditions with high functional group tolerance leads to a variety of new chemical entities.6−8 A typical gold-catalyzed reaction cycle contains three key stages (Scheme 1).9 A nucleophile attacks a [AuLn]+-activated alkyne or alkene (a π complex) to produce a trans-alkenylgold complex intermediate (stage 1). The resulting vinyl complex © XXXX American Chemical Society

Received: October 29, 2018

A

DOI: 10.1021/acs.accounts.8b00544 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research Scheme 1. Various Effects in the Cationic Gold Catalytic Cycle

Table 1. Reactions of Cationic Gold with Allenoates 1

Figure 2. Ligand effect on protodeauration.

Figure 1. ORTEP-3 (Farrugia) diagram (50% ellipsoids) illustrating one independent molecule of gold complex 2d.

filtration, or drying agents affect the reactivity of gold catalysts adversely, which may significantly reduce the TONs of cationic gold-catalyzed reactions. In this Account, we discuss our investigation of the major factors that control the efficiency of cationic gold catalysis using a mechanistic approach. Specifically, we have studied the effects of silver, ligand, counterion, additive, and catalyst deactivation with the ultimate aim of improving the efficiency of gold catalysis.

(AgX) disturb the reaction. The ligand and counterion in the gold catalyst also play a vital role in the catalysis. In the protodeauration step, a positively charged intermediate C is reluctant to undergo protodeauration because its positive charge acts as a deterrent for an incoming proton. Another significant factor to consider is the extent of gold catalyst decay and its conversion into nonreactive species (e.g., Au(0)). Impurities (halides, bases) in solvents, starting materials, B

DOI: 10.1021/acs.accounts.8b00544 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Figure 3. Ligand effect on the formation of gold complex 7. Figure 5. Ligand effect on the gold-catalyzed cyclization of propargyl amide 9.

was obtained when gold complex 2a was treated with TsOH (eq 1). In another experiment, 2a was treated with iodine to produce γ-iodolactone 4a (eq 2).11 Since our first report, many other gold intermediates from other gold-catalyzed reactions have been reported by the groups of Hashmi,9,13,14 Gagné,15 Shi,16 Toste,17 and Yu.18

Figure 4. Ligand effect on the decay of cationic gold (R3PAu+OTf−).

2. ISOLATION AND IDENTIFICATION OF THE KEY GOLD INTERMEDIATE: THE VINYLGOLD COMPLEX Although various vinylgold complexes had been proposed as critical intermediates of cationic gold catalysis, these had not been directly identified in homogeneous cationic gold catalysis before our first study was published.11 We found that roomtemperature- and air-stable vinylgold complexes 2 could be isolated in good yields from the stoichiometric reactions of cationic gold species with allenoates 1 (Table 1).11 The structure of 2 was confirmed by single-crystal X-ray diffraction (Figure 1).11 This discovery not only allowed access to organogold compounds on a preparative scale but also provided direct evidence for most of the mechanisms proposed in gold catalysis. The formation of vinylgold complex 2 was considered to arise from the gold-containing oxonium intermediate E, which was identified by in situ 1H and 31P NMR spectroscopy.12 Two controlled reactions (eqs 1 and 2) further underscored the fact that complex 2 was the actual intermediate in the Au(I)-catalyzed cyclization. In one experiment, γ-lactone 3a

3. LIGAND EFFECTS AND LIGAND DESIGN Although ligands play a major role in the reactivity of transition metal catalysts, there are no definite experimental data, despite a computational study,14 that determine the influence of the structure of a ligand on the kinetics of each stage in the gold catalytic cycle. To complicate matters further, the deactivation of the gold catalyst hampers the likelihood of accurate predictions. We carried out a logical investigation of ligand effects in gold catalysis using a two-phase tactic: in phase 1, we studied the structure−activity relationship between a ligand structure and the kinetics in each stage of the cycle; in phase 2, we classified gold-catalyzed reactions on the basis of their ratedetermining steps and their influence on catalyst deactivation.19 Effects of Ligand Structure on the Kinetics (Phase 1)

Ligand Effect on the Protodeauration Step (Stage 2). The kinetic study of the model reaction indicated that C

DOI: 10.1021/acs.accounts.8b00544 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Figure 6. Ligand effect on an intermolecular hydroamination reaction. Figure 7. Ligand effect on the rearrangement of allenyl ester 15.

electron-withdrawing groups decrease the rate of the reaction whereas electron-donating groups increase the reaction rate (Figure 2a). All reactions followed pseudo-first order kinetics. The plot of ln(A/A0) versus time displayed a linear correspondence (Figure 2b). This characteristic permitted computation of kobs for protodeauration of the gold complex with para-substituted phosphine ligands. Our Hammett correlation established the correspondence for the protodeauration process (ρ = −4.03, R2 = 0.98) (Figure 2c). We examined other commonly used ligands such as the N-heterocyclic carbene-type ligand IPr (L1). The sterically more challenging (o-MeC6H4)3P slowed the reaction vis-à-vis (p-MeC6H4)3P, but the ligand (o-PhC6H4)Ph2P (L2) stimulated protodeauration; we accredited this result to the probability that an orthosubstituted phenyl ring stabilized the generated cationic gold complex through an η2 interaction. Ligand Effect on Electronic Activation of the Alkyne/ Allene (Stage 1). The gold-catalyzed triazole−alkyne cyclization developed by Shi16 was selected as a model reaction because no protodeauration could influence the kinetics of vinylgold complex 7 formation. Inded, compound 7 was transformed to 8 over time at a very low reaction rate (Figure 3). Benzotriazole was added to slow the formation of vinylgold complex 7 for kinetic monitoring. This experiment showed that this stage is favored, both

thermodynamically and kinetically, by electron-deficient ligands (e.g., (p-CF3C6H4)3P) rather than electron-rich ligands (e.g., (p-MeOC6H4)3P). Ligand Effect on the Decay of Cationic Gold (Stage 3). We found that at room temperature, the para-substituted Ar3PAu+ complexes that we studied deteriorated significantly with time. The electronic makeup of the ligand did not appear to influence the stability of cationic gold. The biphenyl diaryl phosphine ligand (L3), however, was quite stable under the reaction conditions (Figure 4). Categorizing Gold-Catalyzed Reactions (Phase 2)

When Protodeauration Is the Rate-Determining Step. The cyclization of propargyl amide 9 was selected as a model reaction.20 The ligand effects (Figure 5) demonstrated that electron-poor ligands decrease the reaction rate. The electronrich trialkylphosphine ligand promoted the highest rate of conversion of starting material to product. When Electronic Activation of the Alkyne/Allene/ Alkene Is the Rate-Determining Step. This possibility would result when the nucleophile is relatively weak or the substrate is less responsive, such as an alkene or an allene. An electron-deficient ligand would enable the vinylgold complex formation. This statement was strengthened experimentally by D

DOI: 10.1021/acs.accounts.8b00544 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

performed better but could not prevent the eventual deactivation of gold. We obtained the best results when we added a substituent at the distal position of the biaryl unit (L5).

Table 2. Relative Rates of Hydroamination for Various Gold Ligands

Ligand Design

The limiting stage in most gold-catalyzed reactions is the regeneration of the cationic gold catalyst. Our investigation of ligand effects highlighted the importance of ortho substitution and electron density matching to gold catalysis. An electron-rich ligand that is able to provide electron density to the gold center is expected to facilitate the redevelopment of the cationic gold catalyst. With this in mind, we designed ligand A1 (Table 2) that features two electron-rich and sterically demanding o-biphenyl groups and one electron-rich cyclohexyl group.22 We prepared A1 in one step from readily available starting materials. The crystallographic structure of A1−AuCl proved that the two o-biphenyl motifs surrounded the gold center. Ligand A1 was superior to other benchmark ligands (Table 2). We evaluated the advantages of A1 using other gold-catalyzed reactions involving C, N, or O nucleophiles. For example, in the gold-catalyzed intermolecular hydroamination of alkynes (Scheme 2a), we managed to reduce the catalyst loading to Scheme 2. Selected Examples Using A1−AuCl in X−H (X = N, O, C) Addition to Alkynes

the report of Toste and co-workers on the intermolecular hydroamination of allenes with hydrazide in the presence of LnAuNTf2.21 When Protodeauration Is the Rate-Determining Step and Catalyst Deactivation Is Significant. We used the intermolecular amination of phenylacetylene as a model reaction (Figure 6). This reaction initially progressed at a practical rate, but it slowed down considerably over time. The Cy3P ligand did not catalyze the hydroamination reaction because it deactivated readily, despite that fact that is considered an ideal ligand for the cyclization of propargyl amides. On the basis of our earlier findings on the protodeauration step, we expected that an electron-rich ligand would accelerate this step. Indeed, we found that L4, an electron-rich ligand with η2 stabilization, produced the best results in the hydroamination. When Electronic Activation of the Alkyne/Allene/ Alkene Is the Rate-Determining Step and Is Accompanied by Significant Catalyst Deactivation. We chose the isomerization of allenyl carbinol ester as a model reaction. We observed substantial decay of the active gold complex and a rate decrease when we used PPh3 as the ligand (Figure 7). A more electron-deficient ligand (p-CF3C6H4)3P (compared with Ph3P) exhibited a higher initial reaction rate but sustained decay. An electron-poor ligand with ο-phenyl substitution (L8)

0.0025% (25 ppm), but the reaction still reached an impressive TON of 31 200. This catalyst performed similarly on a 10 mmol scale. In the gold-catalyzed cyclization of homopropargylic diols (Scheme 2b), we needed only a 20 ppm loading of A1−AuCl to obtain product 7 in 92% yield (Scheme 2b). We obtained similar success in the Conia-ene reaction of a β-keto ester E

DOI: 10.1021/acs.accounts.8b00544 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

(Scheme 2c), where we employed a catalyst loading of only 0.004% (40 ppm) to deliver the desired product in 95% yield.

4. SILVER EFFECTS AND SILVER-FREE PRECATALYSTS The traditional formation of a cationic gold catalyst relies on halide abstraction from a gold catalyst precursor using a silver salt. We discovered that an excessive amount of a silver salt deteriorates the reaction and that a preformed L−Au+X−

Figure 8. Modes of generation of cationic gold catalysts using silver salts.

Scheme 3. Effects of Silver Activators on the Reactivity of Cationic Gold-Catalyzed Reactions

F

DOI: 10.1021/acs.accounts.8b00544 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research complex avoided this problem.23 The two traditional methods of using silver salts are (1) the preformed method and (2) the in situ method (Figure 8). We investigated four manners to generate cationic gold: using a preformed catalyst or using the in situ method and changing the number of AgX equivalents. For the preformed method, we used a sonicated mixture of L−AuCl and AgX; we centrifuged instead of filtering in order to avoid catalyst deactivation caused by components present in filtration aids, such as Celite. We applied the above four modes of forming cationic gold catalysts in six gold-catalyzed reactions (Scheme 3). In all cases, the preformed catalyst was the most reactive. Moreover, we demonstrated that silver caused problems by adding silver salts during the cyclization of propargylic amide 9 (Figure 9).

phthalimide (Pht). The reactivity of the L-Au-Pht/acid system could be fine-tuned with various Brønsted or Lewis acids. The model reaction we employed demonstrated that the reactivity of L-Au-Pht/acid was superior to the traditional silver abstraction method in X−H (X = O, N, C) additions to unsaturated C−C compounds (Scheme 4). Scheme 4. Addition of X−H (X = O, N, C) to an Alkyne/ Alkene

Figure 9. Effect of adding a silver activator during the reaction.

There have been literature reports that silver could be involved in the later stages of the gold catalytic cycle because of its high affinity toward gold and the halide atom.24 For example, the groups of Straub,25 Jones,26 and Gagné27 have reported different Au−Ag complexes formed from the halide abstraction process. The detrimental effect caused by excess silver could have been caused by its interaction with a critical gold intermediate such as the vinylgold complex. We investigated the interactions of isolable vinylgold intermediates Au-1 and Au-2 (Figure 10) with silver salts and observed

5. COUNTERION EFFECT Although the groups of Maier,29 Echavarren,30 Hashmi,31 Gevorgyan,32 and Bandini and Macchioni33 have studied the counterion effects in selected gold-catalyzed reactions, a quantitative analysis of counterion effects in gold catalysis is still missing. We have assembled a gold affinity index and Scheme 5. Simplified Representation of Cationic MetalCatalyzed Reactions

Figure 10. Vinylgold intermediates used to study interactions with silver salts.

significant changes in chemical shifts and peak broadening in the 31P NMR spectra. Silver-Free Gold Precatalyst

Because of the adverse silver effect described above, we looked for an alternate method to generate a gold catalyst and found an answer in the Brønsted acid or Lewis acid activation of an imidogold precatalyst, L-Au-Pht. With this silver-free catalyst we obtained higher reactivities and higher turnover numbers.28 We synthesized L-Au-Pht from L−AuCl and potassium G

DOI: 10.1021/acs.accounts.8b00544 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research hydrogen-bonding basicity index for counterions to estimate their reactivity in cationic gold catalysis.34 Although cationic metals are commonly regarded as “free” ions (M+ in Scheme 5a), these ions may complex with reactants (RC1, RC2, etc.) in the transition state (TS1), and in low-dielectric-constant solvents it is possible for a cationic gold complex to exist as an ion pair (M+X− in Scheme 5b) rather than a “free” ion. When the paired cationic metal (M+X−) interacts with RC1 and RC2 to produce TS2, there will be a charge separation between M+ and X− during the formation of TS2. Hence, additional energy will be needed to overcome the Coulombic attraction (i.e., ΔG⧧TS2 > ΔG⧧TS1). Therefore, a catalyst that has a weakly coordinating counterion (low affinity between M+ and X−) will be more reactive. However, if the transition state intermediate contains an active proton (e.g,. O−Hδ+, N−Hδ+), a counterion may hinder the proton transfer. It should be remembered that a proton has the smallest mass and is highly charged. In these cases, we believe that counterions that have high hydrogen-bonding basicity35 will play an important role in the transition state (Scheme 5c). We calculated the H-bonding bonding energies of various anions with phenol as a way to foretell their hydrogen-bonding basicities. Also, we estimated the gold−counterion dissociation energies in order to establish the rudiments of their binding affinities. The hydrogen-bonding basicity index and gold affinity index for counterions that we established allowed us to rationalize the kinetic effects of counterions in gold-catalyzed reactions. For reactions in which no active protons are involved (e.g., cycloisomerization of 1,6-enyne 37; Table 3a), a gold catalyst having a counterion with a low gold affinity (e.g., SbF6−) exhibited fast kinetics. For reactions encompassing an active proton (e.g., cyclization of propargyl amide 9; Table 3b),

Figure 11. Activation of MOTf by KCTf3.

MOTf in Figure 11) with KCTf3.36 CTf3− is a carbon-based soft anion that undergoes ion exchange with transition metal catalysts. When we added KCTf3 to the reaction mixture, we observed that the interaction between M and OTf− was destabilized by the attraction between OTf− and the naked K+. We proved the above supposition by isomerizing an allenyl carbinol ester (Figure 12). We compared KCTf3 with LiNTf2 and

Table 3. Counterion Effects on the Kinetics of GoldCatalyzed Reactions Figure 12. Effect of promoters in the rearrangement of an allenyl ester.

sodium tetrakis[3,5-bis(trifluoromethyl)phenyl]borate (NaBARF), that is, salts whose counterions have demonstrably low coordinating ability, good stability, and highly delocalized negative charge, namely. We found KCTf3 to be a superior promoter.

a counterion with high hydrogen-bonding strength (e.g., OTf−) backed the proton transfer and sped up the reaction. We prepared the catalyst with a CTf3− counterion using the preformed method. However, we found that the same result was accomplished by stirring a commercial catalyst (e.g.,

Figure 13. Effect of promoters in the cycloisomerization of 1,6-enyne 37. H

DOI: 10.1021/acs.accounts.8b00544 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Figure 14. (a) Decay of [Ph3PAu+OTf−] (39) with cyclohexene present. (b) Decay of 39 in the presence of alkene/alkyne/allene. (c) Decay of 39 in various solvents with cyclohexene present. (d) Decay of 39 with cyclohexene present. (e) Decay of 39 in the presence of cyclohexene and additives. (f) Effect of water on the decay of 39.

cause for the low TON. The fact that L−Au+X− is relatively stable alone but decays at a high rate in many reactions suggests that a cationic gold catalyst decays appreciably in the presence of some elements present in a reaction mixture. We conducted a comprehensive study of the deactivation process by analyzing the contribution of each element in the reaction system.37 With a common gold catalyst, Ph3PAu+OTf−, we investigated the impact of various components (e.g., unsaturated substrates, solvent, counterion, nucleophiles, and additives) on

We also compared the reactivities of L-Au-OTf/KCTf3 and L-Au-CTf3 in the cycloisomerization of 1,6-enyne 37 (Figure 13). Not only did KCTf3 increase the reactivity of L-Au-OTf in this reaction, but the reactivity was only slightly below that of L-Au-CTf3 prepared by treating L−AuCl with AgCTf3.

6. CATALYST DECAY As mentioned earlier, the steadfast deterioration of cationic gold through its changeover into nonreactive species (e.g., Au0 or gold mirror gold particles and [L2AuI]+) is a principal I

DOI: 10.1021/acs.accounts.8b00544 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research the kinetics of the decay. This investigation showed that the addition of unsaturated substrates triggered decay (Figure 14a,b). CH3CN stopped the deactivation (Figure 14c). A counterion screening showed that NTf2− was superior in stabilizing the cationic gold catalyst (Figure 14d). The decay of cationic gold was reduced differently depending on the type of additives used (Figure 14e). Water showed little influence, but it helped in the deactivation of gold catalyst 39 when it was mixed with substrates (Figure 14f). Disproportionation is a possible reason for the decay since the direct reduction of AuI with a highly negative potential is difficult. We used X-ray photoelectron spectroscopy (XPS) and electrospray ionization mass spectroscopy (ESI-MS) to evaluate disproportionation. The results showed that a cationic AuI (e.g., AuIOTf) disproportionates readily compared with its noncationic form (e.g., AuICl). In the presence of an appropriate ligand, cationic gold(I) disproportionates, albeit slowly. Complexes of cationic AuI with π donors experience relatively fast deterioration, whereas complexes of cationic AuI and most σ donors (e.g., N-heterocycles and acetonitrile) do not. We proposed the decay mechanism illustrated in Scheme 6. First, cyclohexene complexes with cationic gold to form a

Table 4. Survey of Additive Effects in the Cyclization of 9

Scheme 6. Proposed Mechanism for the Substrate-Induced Cationic Gold Decay

Scheme 7. Poisoning and Reactivation of the Cationic Gold Catalyst

gold−alkene π complex, and then the phosphine is provisionally replaced by 40 and a free molecule of AuOTf because of the trans effect of the alkene. AuOTf then disproportionates into Au0 and AuIII(OTf)3. This disproportionation is improved by trace water in the system. AuIII oxidizes the phosphine to Ph3P(OTf)2, which is hydrolyzed to produce Ph3PO in the presence of water. The decay process depends on the resting state of the gold catalyst. For type I reactions, where the formation of a gold σ complex is the turnover-limiting stage, we expect relatively fast decay. For type II reactions, in which regeneration of cationic gold catalyst (e.g., protodeauration) is the turnover-limiting stage, we expect relatively slow decay and therefore anticipate a high TON.

(Table 4). The initial rate ratio v/v0 (where v0 is the initial reaction rate without additive and v is the initial reaction rate with additive) was calculated to compare the effects of different additives. Basic additives completely repressed the reaction (e.g., B1, B2). Most other neutral additives did not affect the kinetics of the reaction (e.g., B4−B8), but good hydrogenbond acceptors (e.g., B11, B14, B16, B18) showed dramatic acceleration effects, and pyridine N-oxide (B18) increased the reaction rate 168-fold. We found a quantitative correlation between the efficiency of an additive and its hydrogen-bond basicity according to Laurence’s database. The correlation not only offered strong experimental support for the role of hydrogen bonding but also served as a practical guide for the selection of additives. Additives with high hydrogen-bond basicity are helpful only when protodeauration is the rate-determining step because they also compete with an alkyne/alkene starting material in their complexation with cationic gold, acting as reversible inhibitors. If the additive’s acceleration influence outweighs its inhibitory effect, the overall effect will be positive.

7. ADDITIVE EFFECTS Hydrogen-Bond Acceptors as Additives

In stage 2 of the cationic gold catalytic cycle (Scheme 1), a positively charged intermediate C is reluctant to undergo protodeauration because its positive charge acts as a restraint to the arriving proton. This high resistance toward protodeauration may be the reason why charged vinylgold species have been captured (e.g., Au-7 in Table 4). We found a way to enhance the efficacy of gold-catalyzed reactions through hydrogen-bonding-assisted protodeauration, that is, using additives that have high hydrogen-bonding basicity (pKBHX)38−40 rather than high pKa.41 Many additives were screened using a classic gold-catalyzed reaction where stage 2 (protodeauration) is the limiting step

Acid Additive for Catalyst Reactivation

We found that if the loading of the gold catalyst was below a certain minimum, the reaction did not start. This observation J

DOI: 10.1021/acs.accounts.8b00544 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Figure 15. Influence of acid activators in the cycloisomerization of enyne.

Table 5. Influence of Halide and Base on the Hydration of 43

Scheme 8. Combined Acid Catalysis in a Typical Gold Catalytic Cycle

Figure 16. Effect of cocatalyst on the reaction rate.

41 (Figure 15); here the reaction took place smoothly with a 0.6% catalyst loading but became much slower with a 0.2% gold catalyst loading, and its rate decreased to zero when we reduced the catalyst loading to 0.02%. With a Lewis acid activator, In(OTf)3, the reaction was completed in less than 1 h using a very low gold catalyst loading (0.02%). We investigated the impact of halide and bases on the reactivity and the ability of acidic promoters to restore the reactivity (Table 5). Bases and halides like Bu4N+Cl− and Bu4N+OH− actually inhibited the reaction (Table 5, entries 2 and 3), but Ga(OTf)3 restored the reactivity (Table 5, entries 4 and 5).

partially explained why the TON in gold-catalyzed reactions is generally low, as the catalyst is poisoned by high-gold-affinity impurities such as halides and bases that can be found in solvents, starting materials, and filtration or drying agents. Acid activators (e.g., HOTf and In(OTf)3 that have high affinity toward P (catalyst poison)) reactivated the gold catalyst (Scheme 7). In other words, an acid activator (A+) acted as a sacrificial reagent to bind to possible catalyst poisons (Scheme 7).42 We demonstrated the practicality of this tactic with several examples. One example was the cycloisomerization of enyne K

DOI: 10.1021/acs.accounts.8b00544 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research Scheme 9. Conclusion and Outlook

Acid Additives as Cocatalysts

most crucial factor impacting the efficiency of gold-catalyzed reactions. The precise decay mechanism is still elusive, although in general a cationic gold catalyst itself is relatively stable, so the decay process is likely to occur during the gold catalytic cycle. Usually, the most efficient gold catalyst is often not the most reactive one. When the above mechanistic questions are better understood, we may be able to design more robust and efficient gold catalysts.

Acids can serve as cocatalysts in gold catalysis. The acidity of the cationic gold can be improved using Brønsted acid-assisted Lewis acid catalysis or Lewis acid-assisted Lewis acid catalysis (Scheme 8a). This joint acid system produces a cationic gold species with stronger Lewis acidity, which stimulates bonding of the cationic gold with alkynes/alkenes and the ensuing nucleophilic attack (stage 1) and also helps in the protodeauration stage (stage 2). If the substrate is a terminal alkyne, an acidic cocatalyst should be able to convert the stable acetylide−gold complex A into its active state (Scheme 8b). Lastly, the acidic cocatalyst could influence the decay process (Scheme 8c).43 We reviewed the following typical gold-catalyzed reactions to assess the influence of Lewis or Brønsted acid cocatalysts: hydroamination and hydration of internal alkynes. Through these reactions, we demonstrated that an acidic cocatalyst increased the catalyst turnover significantly (Figure 16).



AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. ORCID

Gerald B. Hammond: 0000-0002-9814-5536 Bo Xu: 0000-0001-8702-1872 Notes

8. CONCLUSION AND OUTLOOK Even though homogeneous gold catalysis is a useful tool in organic synthesis, the relatively high catalyst loadings limit its use in practical larger-scale synthesis. In this Account, we have shown how to improve the efficiency of cationic gold catalysis through a mechanism-based approach. As in other transition-metal-catalyzed reactions, the efficiency of a gold catalyst is improved by utilizing an optimal ligand (usually a bulky and electron-rich ligand) and an optimal counterion (Scheme 9). When a very low loading of catalyst is used (ppm level), the impurities in the starting materials and solvents play a critical role. A suitable Lewis acidic activator improves the efficacy of gold catalysts dramatically. For some gold-catalyzed reactions, a hydrogen-bond acceptor or additive is beneficial. From a mechanistic perspective (Scheme 9), the picture is less clear, despite the considerable effort that we spent elucidating the mechanism of gold-catalyzed reactions. Even though we isolated and identified various gold intermediates, these could be the resting states of gold catalysts and therefore not necessarily kinetically significant. The identification of transitive gold intermediates needs to be further studied. The decay of the reactive cationic gold intermediate is the single

The authors declare no competing financial interest. Biographies Zhichao Lu was born in Hubei, China. He received his B.Sc. and M.Sc. in Medicinal Chemistry from Ocean University of China. After two years working at Chempartner Co. in Shanghai, China, he joined the doctoral program at the University of Louisville in 2011 under the supervision of Professor Gerald B. Hammond and Professor Bo Xu. He obtained a Ph.D. in 2018 and continued a postdoctoral research program in the same lab. His current research interests involve the development of novel gold catalysts and new fluorinating reagents. Gerald B. Hammond was born in Lima, Perú, and received his B.Sc from the Pontifical Catholic University of Perú, his M.Sc from the University of British Columbia in Canada, and his Ph.D. from the University of Birmingham in England. Following an academic career at the University of Massachusetts Dartmouth, he moved to the University of Louisville in 2004, where he currently holds the Endowed Chair in Organic Chemistry. His research interests include the search for new synthetic methodologies in organofluorine chemistry and other halogens, new approaches to catalysis and green chemistry, and the study of biologically active natural products from Peruvian medicinal plants. L

DOI: 10.1021/acs.accounts.8b00544 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

(16) Chen, Y.; Wang, D.; Petersen, J. L.; Akhmedov, N. G.; Shi, X. Synthesis and characterization of organogold complexes containing an acid stable Au−C bond through triazole-yne 5-endo-dig cyclization. Chem. Commun. 2010, 46, 6147−6149. (17) LaLonde, R. L.; Brenzovich, W. E., Jr.; Benitez, D.; Tkatchouk, E.; Kelley, K.; Goddard, W. A., III; Toste, F. D. Alkylgold complexes by the intramolecular aminoauration of unactivated alkenes. Chem. Sci. 2010, 1, 226−233. (18) Zhu, Y.; Yu, B. Characterization of the Isochromen-4-yl-gold(I) Intermediate in the Gold(I)-Catalyzed Glycosidation of Glycosyl ortho-Alkynylbenzoates and Enhancement of the Catalytic Efficiency Thereof. Angew. Chem., Int. Ed. 2011, 50, 8329−8332. (19) Wang, W.; Hammond, G. B.; Xu, B. Ligand Effects and Ligand Design in Homogeneous Gold(I) Catalysis. J. Am. Chem. Soc. 2012, 134, 5697−5705. (20) Hashmi, A. S. K.; Weyrauch, J. P.; Frey, W.; Bats, J. W. Gold Catalysis: Mild Conditions for the Synthesis of Oxazoles from NPropargylcarboxamides and Mechanistic Aspects. Org. Lett. 2004, 6, 4391−4394. (21) Wang, Z. J.; Benitez, D.; Tkatchouk, E.; Goddard, W. A., III; Toste, F. D. Mechanistic Study of Gold(I)-Catalyzed Intermolecular Hydroamination of Allenes. J. Am. Chem. Soc. 2010, 132, 13064− 13071. (22) Malhotra, D.; Mashuta, M. S.; Hammond, G. B.; Xu, B. A Highly Efficient and Broadly Applicable Cationic Gold Catalyst. Angew. Chem., Int. Ed. 2014, 53, 4456−4459. (23) Lu, Z.; Han, J.; Hammond, G. B.; Xu, B. Revisiting the Influence of Silver in Cationic Gold Catalysis: A Practical Guide. Org. Lett. 2015, 17, 4534−4537. (24) Wang, D.; Cai, R.; Sharma, S.; Jirak, J.; Thummanapelli, S. K.; Akhmedov, N. G.; Zhang, H.; Liu, X.; Petersen, J. L.; Shi, X. Silver Effect” in Gold(I) Catalysis: An Overlooked Important Factor. J. Am. Chem. Soc. 2012, 134, 9012−9019. (25) Weber, S. G.; Rominger, F.; Straub, B. F. Isolated Silver Intermediate of Gold Precatalyst Activation. Eur. J. Inorg. Chem. 2012, 2012, 2863−2867. (26) Zhu, Y.; Day, C. S.; Zhang, L.; Hauser, K. J.; Jones, A. C. A Unique Au−Ag−Au Triangular Motif in a Trimetallic Halonium Dication: Silver Incorporation in a Gold(I) Catalyst. Chem. - Eur. J. 2013, 19, 12264−12271. (27) Weber, D.; Gagné, M. R. Dinuclear Gold−Silver Resting States May Explain Silver Effects in Gold(I)-Catalysis. Org. Lett. 2009, 11, 4962−4965. (28) Han, J.; Shimizu, N.; Lu, Z.; Amii, H.; Hammond, G. B.; Xu, B. Efficient Generation and Increased Reactivity in Cationic Gold via Brønsted Acid or Lewis Acid Assisted Activation of an Imidogold Precatalyst. Org. Lett. 2014, 16, 3500−3503. (29) Zhdanko, A.; Maier, M. E. Explanation of Counterion Effects in Gold(I)-Catalyzed Hydroalkoxylation of Alkynes. ACS Catal. 2014, 4, 2770−2775. (30) Homs, A.; Obradors, C.; Leboeuf, D.; Echavarren, A. M. Dissecting Anion Effects in Gold(I)-Catalyzed Intermolecular Cycloadditions. Adv. Synth. Catal. 2014, 356, 221−228. (31) Schießl, J.; Schulmeister, J.; Doppiu, A.; Wörner, E.; Rudolph, M.; Karch, R.; Hashmi, A. S. K. An Industrial Perspective on Counter Anions in Gold Catalysis: On Alternative Counter Anions. Adv. Synth. Catal. 2018, 360, 3949−3959. (32) Dudnik, A. S.; Xia, Y.; Li, Y.; Gevorgyan, V. ComputationGuided Development of Au-Catalyzed Cycloisomerizations Proceeding via 1,2-Si or 1,2-H Migrations: Regiodivergent Synthesis of Silylfurans. J. Am. Chem. Soc. 2010, 132, 7645−7655. (33) Rocchigiani, L.; Jia, M.; Bandini, M.; Macchioni, A. Assessing the Role of Counterion in Gold-Catalyzed Dearomatization of Indoles with Allenamides by NMR Studies. ACS Catal. 2015, 5, 3911−3915. (34) Lu, Z.; Han, J.; Okoromoba, O. E.; Shimizu, N.; Amii, H.; Tormena, C. F.; Hammond, G. B.; Xu, B. Predicting Counterion Effects Using a Gold Affinity Index and a Hydrogen Bonding Basicity Index. Org. Lett. 2017, 19, 5848−5851.

Bo Xu was born in Hubei, China, and received his B.Sc. and M.Sc. from East China University of Science and Technology in Shanghai, China. After receiving his Ph.D. from the University of Louisville in 2008 with Professor Gerald B. Hammond as his research advisor, he worked as a research assistant professor at the University of Louisville until 2014. He is currently a professor in the College of Chemistry, Chemical Engineering and Biotechnology at Donghua University in Shanghai, China. His research interests include the development of novel and environmentally friendly synthetic methodologies, especially the synthesis of novel fluorinated building blocks for drugs and advanced materials, and studies of organic reaction mechanisms.



ACKNOWLEDGMENTS We are grateful to the National Science Foundation (CHE1401700) for financial support. B.X. is thankful to the National Natural Science Foundation of China for financial aid (NSFC21672035).



REFERENCES

(1) Bond, G. C. The catalytic properties of gold. Gold Bull. 1972, 5, 11−13. (2) Fukuda, Y.; Utimoto, K. Effective transformation of unactivated alkynes into ketones or acetals with a gold(III) catalyst. J. Org. Chem. 1991, 56, 3729−3731. (3) Teles, J. H.; Brode, S.; Chabanas, M. Cationic Gold(I) Complexes: Highly Efficient Catalysts for the Addition of Alcohols to Alkynes. Angew. Chem., Int. Ed. 1998, 37, 1415−1418. (4) Hashmi, A. S. K.; Frost, T. M.; Bats, J. W. Highly Selective GoldCatalyzed Arene Synthesis. J. Am. Chem. Soc. 2000, 122, 11553− 11554. (5) Hashmi, A. S. K.; Schwarz, L.; Choi, J.-H.; Frost, T. M. A New Gold-Catalyzed C−C Bond Formation. Angew. Chem., Int. Ed. 2000, 39, 2285−2288. (6) Hashmi, A. S. K. Gold-Catalyzed Organic Reactions. Chem. Rev. 2007, 107, 3180−3211. (7) Li, Z.; Brouwer, C.; He, C. Gold-Catalyzed Organic Transformations. Chem. Rev. 2008, 108, 3239−3265. (8) Arcadi, A. Alternative Synthetic Methods through New Developments in Catalysis by Gold. Chem. Rev. 2008, 108, 3266− 3325. (9) Hashmi, A. S. K. Homogeneous Gold Catalysis Beyond Assumptions and ProposalsCharacterized Intermediates. Angew. Chem., Int. Ed. 2010, 49, 5232−5241. (10) Blanco Jaimes, M. C.; Böhling, C. R. N.; Serrano-Becerra, J. M.; Hashmi, A. S. K. Highly Active Mononuclear NAC−Gold(I) Catalysts. Angew. Chem., Int. Ed. 2013, 52, 7963−7966. (11) Liu, L.-P.; Xu, B.; Mashuta, M. S.; Hammond, G. B. Synthesis and Structural Characterization of Stable Organogold(I) Compounds. Evidence for the Mechanism of Gold-Catalyzed Cyclizations. J. Am. Chem. Soc. 2008, 130, 17642−17643. (12) Liu, L.-P.; Hammond, G. B. Reactions of Cationic Gold(I) with Allenoates: Synthesis of Stable Organogold(I) Complexes and Mechanistic Investigations on Gold-Catalyzed Cyclizations. Chem. Asian J. 2009, 4, 1230−1236. (13) Hashmi, A. S. K.; Schuster, A. M.; Rominger, F. Gold Catalysis: Isolation of Vinylgold Complexes Derived from Alkynes. Angew. Chem., Int. Ed. 2009, 48, 8247−8249. (14) Döpp, R.; Lothschütz, C.; Wurm, T.; Pernpointner, M.; Keller, S.; Rominger, F.; Hashmi, A. S. K. Gold Catalysis: Hydrolysis of Di(alkoxy)carbenium Ion Intermediates as a Sensor for the Electronic Properties of Gold(I) Complexes. Organometallics 2011, 30, 5894− 5903. (15) Weber, D.; Tarselli, M. A.; Gagné, M. R. Mechanistic Surprises in the Gold(I)-Catalyzed Intramolecular Hydroarylation of Allenes. Angew. Chem., Int. Ed. 2009, 48, 5733−5736. M

DOI: 10.1021/acs.accounts.8b00544 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research (35) Pike, S. J.; Hutchinson, J. J.; Hunter, C. A. H-Bond Acceptor Parameters for Anions. J. Am. Chem. Soc. 2017, 139, 6700−6706. (36) Han, J.; Lu, Z.; Wang, W.; Hammond, G. B.; Xu, B. Potassium tris(triflyl)methide (KCTf3): a broadly applicable promoter for cationic metal catalysis. Chem. Commun. 2015, 51, 13740−13743. (37) Kumar, M.; Jasinski, J.; Hammond, G. B.; Xu, B. Alkyne/ Alkene/Allene-Induced Disproportionation of Cationic Gold(I) Catalyst. Chem. - Eur. J. 2014, 20, 3113−3119. (38) Laurence, C.; Brameld, K. A.; Graton, J. r. m.; Le Questel, J.-Y.; Renault, E. The pKBHX Database: Toward a Better Understanding of Hydrogen-Bond Basicity for Medicinal Chemists. J. Med. Chem. 2009, 52, 4073−4086. (39) Okoromoba, O. E.; Han, J.; Hammond, G. B.; Xu, B. Designer HF-Based Fluorination Reagent: Highly Regioselective Synthesis of Fluoroalkenes and gem-Difluoromethylene Compounds from Alkynes. J. Am. Chem. Soc. 2014, 136, 14381−14384. (40) Han, J.; Lu, Z.; Flach, A. L.; Paton, R. S.; Hammond, G. B.; Xu, B. Role of Hydrogen-Bonding Acceptors in Organo-Enamine Catalysis. Chem. - Eur. J. 2015, 21, 11687−11691. (41) Wang, W.; Kumar, M.; Hammond, G. B.; Xu, B. Enhanced Reactivity in Homogeneous Gold Catalysis through Hydrogen Bonding. Org. Lett. 2014, 16, 636−639. (42) Kumar, M.; Hammond, G. B.; Xu, B. Cationic Gold Catalyst Poisoning and Reactivation. Org. Lett. 2014, 16, 3452−3455. (43) Barrio, P.; Kumar, M.; Lu, Z.; Han, J.; Xu, B.; Hammond, G. B. Acidic Co-Catalysts in Cationic Gold Catalysis. Chem. - Eur. J. 2016, 22, 16410−16414.

N

DOI: 10.1021/acs.accounts.8b00544 Acc. Chem. Res. XXXX, XXX, XXX−XXX