|Home | About | Journals | Submit | Contact Us | Français|
Alkenes are attractive starting materials for organic synthesis and the development of new selective functionalization reactions are desired. Previously, our laboratory discovered a unique Pd-catalyzed hydroalkoxylation reaction of styrenes containing a phenol. Based upon deuterium labeling experiments, a mechanism involving an aerobic alcohol oxidation coupled to alkene functionalization was proposed. These results inspired the development of a new Pd-catalyzed reductive coupling reaction of alkenes and organometallic reagents that generates a new carbon-carbon bond. Optimization of the conditions for the coupling of both organostannanes and organoboronic esters is described and the initial scope of the transformation is presented. Additionally, several mechanistic experiments are outlined and support the rationale for the development of the reaction based upon coupling alcohol oxidation to alkene functionalization.
Alkenes are attractive starting materials for organic synthesis because they can be readily converted to a wide range of functional groups.i However, the utility of many alkene functionalization reactions is limited because of poor regioselectivity of the transformation leading to a mixture of products. One strategy to overcome this challenge is the use of metal catalysts to control the selectivity of the functionalization. For example, Pd catalysts have been successfully employed in the Wacker reaction, which selectively oxidizes an alkene to a methyl ketone.ii Previously, our group has developed more effective catalyst systems for the Wacker oxidation.iii As part of this ongoing effort, we were interested in applying these and related catalyst systems to the discovery of new types of selective alkene functionalization reactions. As a result, we discovered a unique PdII-catalyzed hydroalkoxylation reaction of alkenes containing a phenol.iv For example, vinyl phenol (1) is converted to ether 2 in 72% yield using catalytic Pd[(−)-sparteine]Cl2 and CuCl2 (Equation 1).
While there have been numerous reports of metal-catalyzed alkene hydroalkoxylation reactions,v there were no examples that employed PdII salts.vi This is most likely due to facile β-hydride elimination after initial oxypalladation, which would lead to Wacker oxidation products.ii We initially hypothesized that the phenolic substrate allowed for competitive protonation of the Pd-carbon bond formed after initial oxypalladation of the alkene with the alcoholic substrate.iv In order to test this hypothesis and probe the origin of the hydrogen atom incorporated into the alkene framework, we believed that conducting the reaction with CH3CH2OD as the solvent should result in the incorporation of one deuterium atom into the product at the site of Pd-carbon bond protonation.v However, subjecting vinyl phenol 3 to the reaction conditions with CH3CH2OD as the solvent produced the unlabeled ether product 2 (Figure 1, A). This result rules out the plausibility of a Brønsted acid based mechanism. To further investigate the origin of the hydrogen atom incorporated into the product, a similar reaction was performed with CD3CD2OD as the solvent. This led to the formation of isotopomers 4a and 4b in a 2.5 to 1 ratio, which contain one deuterium atom incorporated into the alkene framework. This result suggests that the hydrogen atom originates from the alkyl chain of ethanol rather than the acidic proton.
Based upon these results, we propose a mechanism wherein the PdII catalyst A initially oxidizes the alcoholic solvent to generate the Pd-hydride intermediate B (Figure 1, B). Coordination of the alkene produces the cationic complex C. Based upon the formation of isotopomers 4a and 4b, we propose that the alkene inserts into the Pd-hydride at both the α- and β-position of the styrene yielding D and D′. Since only the benzylic addition product is obtained, we propose that intermediate D′ proceeds to product via formation of an ortho-quinone methide intermediate with concomittant reduction of PdII to Pd0 and generation of E.vii Ethanol subsequently reacts with the quinone methide intermediate to liberate the ether product and the Pd0 species F, which is reoxidized by CuCl2 or O2.ii,viii
While the deuterium labeling experiments suggest a unique alcohol oxidation coupled alkene functionalization process, the scope of the reaction is limited to substrates that contain a phenol. Therefore, our laboratory was interested in taking advantage of this distinctive mechanistic motif in the realm of new reaction discovery. Specifically, we wanted to develop a PdII-catalyzed reductive coupling reaction, which employs alkenes that do not contain a phenol in combination with an organometallic reagent to generate a new carbon-carbon and carbon-hydrogen bond under aerobic conditions. Herein, we present the development of an aerobic reductive coupling reaction of alkenes with both organostannanes and organoboronic esters as well as experiments that showcase the unique mechanism.
We were interested in developing a reductive coupling reaction of alkenes and organometallic reagents that utilizes an alcohol oxidation to initiate the catalysis and generate a Pd-hydride. Mechanistically, this type of transformation differs from Pd0-catalyzed cross-coupling wherein the catalyst initially oxidatively adds an electrophilic organic compound, which generally contains a carbon-halogen bond (R–X), that yields the oxidized species R–PdII–X (Figure 2, A).iia A potential limitation of Pd0-catalyzed cross-coupling reactions is the requirement of a substrate that is capable of oxidizing Pd0 to PdII. Unconventional routes to access the requisite R–PdII–X intermediate from starting materials that circumvent the need to oxidize Pd0 to PdII will increase the versatility and scope of cross-coupling reactions.
Based upon the deuterium labeling experiments performed on the alkene hydroalkoxylation reaction,iv we hypothesized the catalytic R–PdII–X intermediate could be generated from an alkene (Figure 2, B and C).ix Specifically, oxidation of the alcoholic solvent with PdII catalyst A will lead to the formation of PdII-hydride B. Coordination and insertion of the alkene into the PdII-hydride yields the PdII-alkyl intermediate D (or D′), which resembles the R–PdII–X species formed via oxidative addition of an organic electrophile.iia Transmetalation to form E and subsequent reductive elimination generates the reductive coupling product as well as the reduced catalyst F. Aerobic oxidation of Pd0 to PdII completes the catalytic cycle.ii,viii The sp2-hybridized carbon atoms of the alkene will be transformed to sp3 atoms, which can be difficult to access with traditional cross-coupling because of facile β-hydride elimination.x
Initially, we anticipated the formation of undesired byproducts arising from the oxidative Heck reactionxi and homocoupling of the organometallic reagent,xii which are proposed to originate from initial transmetalation of the PdII catalyst (Figure 2, D). The nature of the catalyst, solvent, and coupling partners were thought to have a significant impact on the formation of these byproducts by modulating the relative rate of transmetalation compared to alcohol oxidation/alkene insertion. Therefore, we initially decided to employ the Pd[(−)-sparteine]Cl2 catalyst because of its ability to readily oxidize alcohols and alkenes.iiib, xiii
Previously, our laboratory has shown that exposing a simple styrene substrate to the alkene hydroalkoxylation reaction conditions with ethanol as the solvent leads to the acetal product 6 (Figure 3, A).xiv Since the proposed mechanism of acetal formation involves initial nucleophilic addition of the alcohol to the metal bound alkene, we thought a less nucleophilic secondary alcohol, such as isopropanol (IPA), would not readily undergo acetal formation. However, performing the reaction with IPA as the solvent unexpectedly produced the hydroalkoxylation product 8xv along with the Wacker oxidation product 7, which most likely originates from an H2O2 mediated pathway (Figure 3, B).xvi Further investigation of these results demonstrates that the hydroalkoxylation product 8 forms via initial hydrochlorination of the alkene followed by nucleophilic substitution with the alcohol.xv Interestingly, decreasing the amount of CuCl2 slows the formation of the hydroalkoxylation product. Thus, we believed that IPA would be an effective solvent and hydride source for the reductive coupling reaction as long as the concentration of CuCl2 was low enough to prevent formation of 8.
In addition to the solvent, the nature of the coupling partners should play an important role in the outcome of the reaction. Therefore, styrene derivatives were selected because of the inability of the olefin to isomerize.xvii Previously, our laboratory has found styrenes to be excellent model substrates for the development of new reactions and catalyst systems. Information garnered from the use of styrene substrates can subsequently be employed to expand the scope of these systems to include a wide variety of alkenes.iii Furthermore, the second coupling partner needed to readily undergo transmetalation and had to be stable to the alcoholic solvent and O2 atmosphere. Many organometallic reagents require a base to facilitate transmetalation with PdII,iia but the presence of exogenous base has been found to affect the rate of alcohol oxidation.xiii Therefore, PhSnBu3 (9a) was initially selected as the organometallic coupling partner because it should be stable under the reaction conditions and does not require base to facilitate transmetalation.xviii
Exposure of 4-methylstyrene (5a) and PhSnBu3 (9a) to a solution of catalytic Pd[(−)-sparteine]Cl2 and CuCl2 in IPA under an O2 atmosphere led to the desired hydroarylation product 10a in 23% GC yield as a >25:1 mixture of regioisomers (Equation 2).ix, xix As expected, a significant amount of the oxidative Heck product 11 was formed (18% GC yield), which is proposed to originate from initial transmetalation of the PdII catalyst.xid Interestingly, the diarylation product 12, which is the addition of two aryl groups across the alkene, was formed in 20% GC yield. The diarylation product is believed to originate from the same PdII-alkyl intermediate that leads to the Heck product. However, instead of undergoing β-hydride elimination, the PdII-alkyl species reacts with another equivalent of the organostannane and subsequent reductive elimination generates the product. While diarylation of alkynes and norbornene have been previously reported, this is the first example of a diarylation reaction of a styrene derivative.xx This unique transformation is currently under further investigation in our laboratory. Additionally, the oxidative homocoupling product 13xii was formed in 9% GC yield and the Wacker oxidation product 714–16 was generated in 3% GC yield.
Since the major byproducts originate from either the oxidative Heck or diarylation reactions, we focused our attention on minimizing the formation of these compounds. Based upon the proposed mechanism, a decrease in the relative rate of initial transmetalation compared to alcohol oxidation/alkene insertion was desired to improve the selectivity for the reductive coupling product. Previously, the addition of Cu salts to Pd0-catalyzed Stille couplings had been shown to facilitate transmetalation.xviiia Therefore, CuCl2 was removed from the catalytic conditions leading to a decrease in conversion, which was attributed to catalyst decomposition (Table 1, entry 2). However, the ratio of 10a to 11 improved to 2.4:1 and, thus, Cu-free conditions were pursued. However, without CuCl2 to promote catalyst regeneration, we needed to identify an additive to help stabilize the catalyst. Previously our laboratory conducted mechanistic studies on the aerobic oxidation of alcohols with Pd[(−)-sparteine]Cl2, which demonstrated a significant rate enhancement and improved catalyst stability when exogenous (−)-sparteine was used.xiii Based upon these data, 40 mol% exogenous (−)-sparteine was evaluated and led to an increase in selectivity for the hydroarylation product compared to Heck with a ~28:1 ratio in low conversion (entry 3). Since the aerobic oxidation of alcohols with Pd[(−)-sparteine]Cl2 is normally conducted at elevated temperatures,xiii the reaction was performed at 60 °C, which led to 63% conversion of 5a and a 45% GC yield of the hydroarylation product (entry 4).
Considering only 63% conversion of the substrate was observed, a time course analysis of the reaction was performed (Figure 4, A). A significant retardation of the rate was observed as the reaction progressed, suggesting either catalyst decomposition or inhibition. Previously, Waymouth and co-workersxxi and Stahl and co-workersxxii have demonstrated that ligands commonly employed in Pd oxidase catalysis can be consumed via oxidation. Therefore, we hypothesized that a possible inhibitor could originate from oxidation of (−)-sparteine. In the past, (−)-sparteine-N-oxide (14a and b) has been synthesized by treating (−)-sparteine with H2O2, which is the product of O2 reduction (Figure 4, B).xxiii To test our hypothesis, two experiments were performed with added 20 mol% (−)-sparteine-N-oxide at two different concentrations of (−)-sparteine. With both sets of conditions, a comparable decrease in the conversion of 5a was observed (Table 1, compare entries 4–6). These results suggest that (−)- disproportionate H2O2 reduce the amount of (−)-sparteine-N-oxide generated during the catalysis. Previously, MnO2 has been shown to disproportionate H2O2.xxiv Excitingly, use of 75 mol% MnO2 led to 10a in 81% GC yield at >99% conversion (entry 7). Interestingly, addition of MnO2 and (−)-sparteine-N-oxide led to similar conversion and GC yield compared to the absence of MnO2 (entry 8). It should noted that the reaction does not effectively proceed without the addition of (−)-sparteine even in the presence of MnO2.
To test our hypothesis that MnO2 was reducing the amount of (−)-sparteine-N-oxide, two reductive coupling reactions were performed using the optimal catalytic conditions except one reaction did not contain MnO2 (Table 1, entries 4 and 7). After completion of the reaction, the crude mixtures were analyzed by ESI-MS. Both (−)-sparteine-N-oxide and (−)-sparteine were observed and the ratio of the respective peak heights were measured. However, the peak heights are dependent on ionization potential and do not directly reflect the absolute quantity of each species in the mixture. The ratio of peak heights of (−)-sparteine-N-oxide to (−)-sparteine was found to be 0.50 for the reaction performed with 75 mol% MnO2 compared to 0.64 for the MnO2 free conditions, which is an approximate 28% difference. These data support our hypothesis that MnO2 disproportionates H2O2 and reduces the amount of (−)-sparteine-N-oxide generated during the catalysis.
Attempts to scale the reaction from 0.2 to 1 mmol led to inconsistent conversion and GC yields of 10a. For example, the reductive coupling reaction performed on 0.2 mmol scale produced 10a in 81% GC yield (Table 2, entry 1). However, conducting the same reaction on 1 mmol scale produced 10a in only 24% GC yield (entry 2). This discrepancy led us to question if there is not enough O2 present in the liquid phase and causes catalyst decomposition. Previously, Stahl and co-workers have found that a large liquid to gas surface area is required to achieve effective mass transport of O2 into solution in the PdII-catalyzed aerobic oxidation of alcohols with DMSO, which could be a problem in this case.xxv Additionally, the solubility of O2 in isopropanol may be decreased at 60 °C.xxvi
In order to test our hypothesis, the pressure of O2 was increased from balloon (~1.1 atm) to 25 psi (1.7 atm), and this led to a significant improvement of the GC yield of 12a on 1 mmol scale (Table 2, entry 3). Even with the increased O2 pressure, MnO2 was still required to achieve effective catalysis (compare entries 3 and 4). The optimization studies demonstrate that ligand oxidation can inhibit the catalysis and efficient mass transport of O2 is required to prevent catalyst decomposition.xxvii Even with the improved catalytic conditions, a large diameter flask is still required to obtain acceptable yields of the reductive coupling product. Since this is a triphasic reaction, vigorous stirring is also essential to ensure a sufficient dispersion of MnO2 and efficient mass transport of O2.
After identifying suitable catalytic conditions, the initial scope of the PdII-catalyzed reductive coupling of arylstannanes and styrenes was explored. This methodology can be utilized to access a variety of diarylmethine containing compounds, which are prevalent in biologically active small molecules.xxviii Many of these compounds would be difficult to selectively synthesize using other types of hydroarylation methodologies.xxix Notably, all of the reactions are highly regioselective (>25:1) for addition of the arene to the α-carbon of the styrene. A variety of arystannanes were competent coupling partners where the steric and electronic nature did not have a significant impact upon the yield of the reaction (Figure 5). In contrast to many previously reported Lewis acid-catalyzed hydroarylation reactions,xxix substrates that contain acid sensitive functional groups can be employed. The scope of the reaction was not limited to arylstannanes, but vinylstannanes were utilized to perform an overall hydrovinylation reaction.xxx Under these conditions, simple alkenes do not lead to appreciable yields of the desired product because of alkene isomerization.xvii Additionally, all of the products were generated with <5% ee even though a chiral ligand was employed.
In addition to the use of organostannes, we were interested in expanding the scope of the reductive coupling reaction to include organoboronic acids or esters. Boronic acids are highly attractive coupling partners because of the low toxicity of the byproducts and excellent functional group compatibility.xxxi However, the mechanism of transmetalation is proposed to be more complex as compared to organostannanes. For example, an exogenous base is typically required to activate the organoboronic acid or ester (A), which yields intermediate B (Figure 6, A).xxxii The activated species then reacts with R-PdII-X via a proposed four-membered transition state C. Integrating an exogenous base to facilitate transmetalation of an organoboronic acid or ester into the reductive coupling reaction was thought to be a significant challenge. The reason for this was based upon the proposed mechanism for the aerobic Pd-catalyzed oxidation of alcohols wherein exogenous base has been shown to accelerate the rate of alcohol oxidation.xiii The proposed mechanism entails initial coordination of the alcohol to the active catalyst D, which leads to species E. Subsequent deprotonation of the bound alcohol by an exogenous base produces Pd-alkoxide F, which undergoes β-hydride elimination to generate Pd-hydride G. Specifically, the rate limiting step of the Pd[(−)-sparteine]Cl2 catalyzed oxidation of alcohols with low concentrations of exogenous base was found to be deprotonation of intermediate E.xiii However, when an exogenous base is employed, an acceleration of the rate of the reaction is observed, which is attributed to a change to β-hydride elimination of F as the rate limiting step.
With this additional requirement in mind, we attempted to identify catalytic conditions for the reductive coupling of styrenes and organoboronic acids. Initially, we evaluated PhB(OH)2 (BA) with the conditions developed for organostannanes.ix This led to the desired product 10a in poor yield.xxxiii During the course of the reaction, a black precipitate was observed, which suggested decomposition of the Pd catalyst. Therefore, we decided to evaluate different ligands, which should better stabilize the catalyst. Previously, our group has shown that the use of N-heterocyclic carbene (NHC) ligands led to highly active catalyst systems for the aerobic oxidation of alcohols.xiiid Inthis case, the [Pd(SiPr)Cl2]2 catalyst was employed to couple 5a and BA in combination with exogenous (−)-sparteine and tBuOK, which led to the desired product 10a in 8% GC yield (Figure 7). Even with the use of an NHC ligand, catalyst decomposition was observed. This led us to question if the acidity of the boronic acid was inhibiting the reaction or was causing catalyst decomposition.
To test our hypothesis, we evaluated the pinacol derived phenyl boronic ester (PE) under the same conditions. This led to a higher yield of the desired product and through further optimization it was found that the use of 2.5 equivalents of PE produced 10a in 64% GC yield. Even though the yield was significantly improved, the reaction required 24 h, which led us to question if the large size of PE was reducing the rate of the reaction. Therefore, both the propanediol (PD) and the ethylene glycol (EG) derived boronic esters were evaluated. Interestingly, PD yielded the desired product in 32% GC yield, but the use of 3 equivalents of EG generated 10a in 91% GC yield.
After identifying the optimal boronic ester derivative for the reductive coupling reaction, the nature of the exogenous base was evaluated. Initially, a significant number of nitrogen containing bases were tested, but all of the bases, except (−)-sparteine, did not lead to effective catalysis. However, without (−)-sparteine, only 2% conversion of 5a was measured and suggests that (−)-sparteine is playing an important role in the reaction (Table 3, entry 1). Therefore, we optimized the concentration of (−)-sparteine and found 6 mol% produced the highest yield of the desired product. Interestingly, removal of tBuOK led to >99% conversion of 5a and 78% GC yield of 10a (entry 3). The lower selectivity for the reductive coupling product suggests that both bases are required to achieve higher yields. Based upon these results, we evaluated higher concentrations of tBuOK, which decreased the yield of 10a (entries 4 and 5). After identifying the optimal concentration of tBuOK, we questioned if other inorganic bases would be more effective. Therefore, were evaluated five different inorganic bases, which led to lower yields of 10a compared to tBuOK (entries 6–10). The optimization studies showcase the importance of the nature of the boronic ester derivative and that both (−)-sparteine and tBuOK are required to achieve optimal yields of the reductive coupling product.
Using the optimized catalytic conditions, the scope of the reductive coupling of boronic ester derivatives was investigated. All of the diarylmethine containing compounds synthesized with this method had >25:1 selectivity. The electronic nature of the styrene did not significantly impact the yield of the transformation (Figure 8). However, electron donating groups on the arylboronic ester reduced the rate of the reaction leading to lower yields. Acid sensitive functional groups, such as an acetal, are stable under the conditions and are readily removed during workup. Unfortunately, under these conditions vinylboronic ester derivatives did not undergo reductive coupling. Additionally, straight chain alkenes rapidly isomerizes.xvii Finally, considering that the chiral substance (−)-sparteine is employed, the enantioselectivity was measured for several reductive coupling products revealing less than 5% ee using either boronic ester derivatives or organostannanes.
After exploring the initial scope of the reductive coupling reaction of styrenes with both organostannanes and organoboronic esters, we had several mechanistic questions. The first question we addressed was where does the hydrogen atom incorporated into the alkene framework originate from? This was explored utilizing two isotopic labeling experiments similar to those used for the alkene hydroalkoxylation reaction (Figure 1, A).iv The reductive coupling of 5a and 9a was performed with (CH3)2CH(OD) as the solvent (Figure 9).ix As expected, no deuterium was incorporated into the product 10a, which rules out the involvement of acidic protons in the catalysis. In contrast, employment of (CH3)2CD(OH) as the solvent led to three hydroarylation products, which were quantified by 1H NMR and GC/MS. Isotopomers 16a and 16b account for 70% and 22% of the isolated products respectively and the remaining 8% is the unlabeled product 10a.
The labeling studies support our original mechanistic proposal wherein oxidation of the alcoholic solvent initiates the catalysis and yields the PdII-hydride intermediate B (Figure 2, C). The formation of two isotopomers suggests that the alkene inserts into the PdII-hydride species at both the α- and β-position. This leads to intermediates D and D′, which equilibrate via β-hydride elimination. However, when (CH3)2CD(OH) is employed as the solvent, deuterium is incorporated into the styrene during the course of the reaction, which suggests that the alkene can dissociate from C. This process implies that the PdII-deuteride can exchange with the alkene substrate via an insertion/β-hydride elimination pathway to generate a PdII-hydride intermediate. This may account for the formation of 8% of the hydroarylation product, which does not contain any deuterium.
Since the deuterium labeling studies support the formation of both intermediates D and D′, this led to the next mechanistic question of why is the regioselectivity >25:1 for substitution of the arene at the benzylic position of the styrene? The high regioselectivity suggests that only D leads to product even though both isomers are proposed to form. To account for this observation, we thought intermediate D, which is a PdII-η1-alkyl species, can be stabilized by converting to a PdII-η3-π-benzyl intermediate when using a styrenyl substrate.xxxiv To test our hypothesis, we believed a conjugated diene substrate would lead to a similar PdII-η1-alkyl intermediate that can be stabilized via a possible PdII-η3-π-allyl species (Figure 10). Reaction of diene 17 with vinylstannane 9f yields the reductive coupling product in 40% yield as a single regioisomer and 1:1 mixture of diastereomers.
The deuterium labeling experiments suggest the possible formation of a well defined PdII[(−)-sparteine](H)X intermediate B (Figure 2). Since the bidentate ligand (−)-sparteine is employed, this causes the hydride and anionic ligand to be cis to each other, which is the required geometry for the complex to undergo reductive elimination to generate Pd0[(−)-sparteine]. To the best of our knowledge, there are no known isolated cis-PdII-(H)X complexes, which can be attributed to facile reductive elimination of HX.xxxv This led to the question does the ligand, (−)-sparteine, in combination with the polar alcoholic solvent facilitate chloride dissociation to generate a cationic PdII-hydride intermediate, which does not readily undergo reductive elimination?xiii To investigate this question, we attempted to synthesize and characterize a cationic PdII[(−)-sparteine](H)(X) complex.
A variety of strategies have been previously reported to access trans-PdII-hydrides and we believed the use of a non-coordinating counterion, such as triflate, would help prevent reductive elimination of HX.xxxv However, attempts to synthesize PdII[(−)-sparteine](H)(OTf) using a variety of different reported methods to access trans-PdII-hydrides led to the formation of a black precipitate.xxxvi These results suggest that the PdII-hydride may have been generated, but underwent reductive elimination to yield a Pd0 complex, which decomposed. Since the initial attempts did not employ a polar alcoholic solvent, we decided to attempt to synthesize the desired complex via an alcohol oxidation.xxxv Therefore, the PdII[(−)-sparteine](OTf)2 complex (20) was synthesized by treatment of PdII[(−)-sparteine]Cl2 with two equivalents of AgOTf. The PdII[(−)-sparteine](OTf)2 complex was subsequently dissolved in isopropanol, which unfortunately led to the formation of a black precipitate (Figure 11, A).
The formation of a black precipitate in all of our attempts suggests that the PdII[(−)-sparteine](H)(X) complex is not stable and undergoes facile reductive elimination to generate Pd0[(−)-sparteine] and HX. This led to the question how does the reductive coupling reaction progress effectively if the proposed catalytic intermediate B readily undergoes reductive elimination (Figure 2, C)? One of the key differences between the stoichiometric studies we performed and the catalytic conditions is the presence of an alkene. Previously, it has been shown that cationic PdII-hydride species, which are not stable, can be trapped by reacting with conjugated dienes.xxxvii Based upon this report, we dissolved the PdII[(−)-sparteine](OTf)2 complex (20) in an isopropanol solution that contained 10 equivalents of styrene (Figure 11, B). The desired [(−)-sparteine]PdII-η3-π-benzyl(OTf) complex (21) was generated and characterized by ESI-HRMS. Unfortunately, attempts to isolate the complex led to the formation of a black precipitate and 1H and 13C NMR analysis of the complex generated in situ reveals the presence of multiple unidentifiable species.
The generation of complex 21 supports our mechanistic hypothesis and suggests that the unstable PdII-hydride intermediate B, which is generated during the catalysis, undergoes competitive alkene insertion instead of reductive elimination of HX. This proposal led to the question of how efficient is the alcohol oxidation compared to alkene functionalization? To investigate this question, a time course analysis of the reductive coupling of styrene 5a and organoboronic ester 15a was conducted with sec-butanol as the solvent (Table 4).xxxiii The higher molecular weight alcohol allows for quantification of the amount of ketone product being formed by GC. The GC yields of the oxidized alcohol product butan-2-one (22) and the reductive coupling product 10a as well as the percent of 5a remaining were measured over time. At 6h or about 50% conversion of 5a, the GC yield of 10a and 22 are similar, which indicates that oxidation of one alcohol yields one product. These data suggest that the PdII-hydride species undergoes faster alkene insertion than reductive elimination.xxxviii However, as more of the alkene substrate is consumed, reductive elimination or other processes become competitive.
During the optimization studies conducted on the reductive coupling of organoboronic esters, it was found that three equivalents of the arylboronic ester derivative were required to achieve acceptable yields. This observation led to another question we wanted to investigate, which is why are multiple equivalents of the transmetalating agent required for organoboronic esters, but not for organostannanes? In both cases, the transmetalation agent can undergo oxidative homocoupling. However, arylboronic esters can be oxidized to the respective phenol derivatives by H2O2 generated in situ.xiid We initially hypothesized that the excess boronic ester may be oxidized under the catalytic conditions. Therefore, as the reaction progressed, the fate of the boronic ester was investigated. At 6 h, approximately two equivalents of 15a were consumed (Table 4). Interestingly, at 14 h, a 139% GC yield of phenol (23) was measured, which supports our hypothesis that the excess 15a required is oxidized by H2O2. Additionally, oxidative homocoupling of 15a converts approximately 0.28 equivalents of 15a to biphenyl (13). The time course highlights the undesired side reactions that lead to the requirement of multiple equivalents of organoboronic esters. Typically organostannanes are more oxidatively stable to H2O2, which allows for lower quantities to be employed. However, the excess boronic ester scavenges H2O2, which potentially prevents oxidation of the exogenous (−)-sparteine that would inhibit the catalysis as observed with the use of organostannanes.
The last mechanistic question we wanted to address originated from the optimization studies. Specifically, why is (−)-sparteine required to achieve catalysis considering other amines such as 2,2′-bibyridine, TMEDA, and TEA are not viable? One of our initial hypotheses was that (−)-sparteine was responsible for breaking up the dimeric [PdII(SiPr)Cl2]2 catalyst. To investigate the plausibility of our proposal, we dissolved [PdII(SiPr)Cl2]2 and two equivalents of (−)-sparteine in 1,2-dichloroethane (DCE) and heated the mixture to reflux (Equation 3). After 2 h, the solution was analyzed by ESI-MS and one major Pd species was observed [m/z (M+H)+ = 801.3]. The molecular weight corresponds to the PdII[(−)-sparteine](SiPr)Cl2 complex (24). We propose that the NHC ligand is trans to the monodentate (−)-sparteine ligand based upon the trans geometry of previously reported PdII(NHC)(pyridine)Cl2 complexes.xxxix However, attempts to isolate complex 24 were unsuccessful, but instead [PdII(SiPr)Cl2]2 and free (−)-sparteine were obtained. This suggests that complex 24 is in equilibrium with the dimer and (−)-sparteine. Under these conditions, PdII[(−)-sparteine]Cl2 was not detected by ESI-MS. These data suggest the possibility of the role of (−)-sparteine is to break up the dimer or the free nitrogen of complex 24 may act as an intramolecular base.xiii
Based upon mechanistic insight garnered from the study of the hydroalkoxylation reaction of vinyl phenols, we were able to design and development a fundamentally new reductive cross-coupling reaction of styrenes and organometallic reagents. This transformation is proposed to proceed by initial oxidation of the alcoholic solvent followed by alkene insertion to yield a PdII-alkyl species, which subsequently undergoes transmetalation with either an organostannane or an organoboronic ester to ultimately generate a new carbon-carbon bond. Overall, this process rapidly yields a valuable diarylmethine core structure via a sp3-sp2 cross-coupling in high regioselectivity, which is proposed to arise from stabilization via a PdII-η3-π-benzyl complex. One interesting aspect of this transformation is the use of O2 as a terminal oxidant in what is an overall formal reduction of the alkene. This unique mode of catalysis is currently being investigated in our laboratories in the development of new hydrofunctionalization reactions as well as applications to asymmetric catalysis.
Isopropanol was dried by refluxing over calcium oxide for 12 h followed by fractional distillation. Pd[(−)-sparteine]Cl2 was synthesized according to a previously reported procedure.xl (−)-sparteine was prepared from (−)-sparteine sulfate pentahydrate according to a previously reported procedure.xli HRMS were obtained with either an ESI or APCI source with a Waters TOF. GC separations were performed with an HP6890 GC with a flame ionization detector equipped with a DB-5 column using a 50:1 split. While our laboratory has not encountered any problems, great caution should be taken when heating flammable solvents and adding solid metals to a flammable solution in the presence of O2.
The procedure and characterization data for the synthesis of compounds 10a–j has been previously reported.ix The following general procedure was used to synthesize compounds 10a–j: A three-way joint was fitted to the side arm of an oven dried 100 mL thick-wall glass pressure vessel equipped with a stir bar. An O2 tank was connected to the three-way joint and O2 was flowed through the vessel. To the vessel, was added 65.2 mg of MnO2 (0.750 mmol, 0.750 equiv.), 4.30 mL of isopropanol, 200 μL of a 2.00 M solution of (−)-sparteine (0.400 mmol, 0.400 equiv.) in isopropanol, 1.00 mmol of the alkene (1.00 equiv.), and 1.50 mmol of the organostannane (1.50 equiv.). The vessel was sealed, pressurized to 25 psi, evacuated via water aspiration, and re-pressurized to 25 psi O2. This procedure was repeated three times. The vessel was sealed and the mixture was stirred vigorously for ca. 20 min at room temperature at 25 psi O2. The vessel was opened and O2 was flowed through the vessel. To the stirred mixture, was added 10.3 mg of Pd[(−)-sparteine]Cl2 (0.0250 mmol, 0.0250 equiv.). The vessel was immediately pressurized to 25 psi O2 and was sealed. The three-way joint was removed and the reaction mixture was stirred vigorously for 5 min at room temperature. The mixture was then heated to 60 °C in an oil bath and was stirred vigorously for 18 h. The vessel was removed from the oil bath, cooled to room temperature, and the pressure released. To the reaction mixture, was added 5.00 mL of a 1.00 M solution of aqueous NaOH and was stirred for 1 h.xlii The mixture was filtered through Whatman filter paper, rinsed with ca. 10.0 mL of Et2O, and was transferred to a separatory funnel. The aqueous layer was extracted three times with 20.0 mL of Et2O, all of the organic extracts were combined, washed with 40.0 mL of brine, and dried over MgSO4. The mixture was filtered and the solvent was removed in vacuo. The product was purified via flash chromatography.
The procedure and characterization data for the synthesis of compounds 10k-r has been previously reported.xxxiii The following general procedure was used to synthesize compounds 10k-r: To an oven dried 100 mL Schlenck flask equipped with a stir bar, was added 4.3 mg of [Pd(SiPr)Cl2]2 (0.0038 mmol, 0.0076 equiv.), 300 μL of a 100 mM solution of (−)-sparteine (0.030 mmol, 0.060 equiv.) in isopropanol, and 7.4 mL of isopropanol. A dried water condenser and a three-way joint fitted with a balloon of O2 were installed on the flask. The flask was evacuated via water aspiration and refilled with oxygen three times and the mixture was stirred vigorously for ca. 20 minutes at room temperature under an O2 atmosphere. To the mixture, was added 1.0 mL of a 500 mM solution of the alkene (0.50 mmol, 1.0 equiv.) in isopropanol, 1.0 mL of a 1.5 M solution of the organoboronic ester (1.5 mmol, 3.0 equiv.) in isopropanol, and 300 μL of a 100 mM solution of potassium tert-butoxide (0.030 mmol, 0.060 equiv.) in isopropanol. The mixture was then heated to 55 °C and was stirred vigorously for 24 hours. The reaction mixture was cooled to room temperature. The mixture was concentrated under reduced pressure. To the residue, was added 10 mL of water and was extracted two times with 10 mL of hexanes. To the combined organic extracts, was added 1.0 g of magnesium sulfate and 1.00 g of silica. The mixture was stirred at room temperature for 10 minutes, filtered, washed with hexanes, and concentrated under reduced pressure to yield a colorless oil. The product was purified via flash chromatography.
To a flame dried 100 mL round bottom flask equipped with a stir bar under a nitrogen atmosphere, was added 618 mg of Pd[(−)-sparteine]Cl2 (1.50 mmol, 1.00 equiv.) and 60.0 mL of CH2Cl2. The orange solution was stirred for 5 min and 771 mg of silver triflate (3.00 mmol, 2.00 equiv.) was added. The cloudy orange solution was stirred for 1 h and the precipitate was removed by filtering the solution through Celite rinsing with CH2Cl2. The solution was concentrated in vacuo to approximately 10 mL and ca. 10 mL of hexanes were added until the solution turned cloudy. The solvent was removed in vacuo to yield an orange solid, which was rinsed with hexanes and the residual solvent was removed by applying a vacuum to the solid. Yield: 76% (730 mg); orange solid mp: 120–123 °C (decomposition); IR 3140, 2941, 2870, 1453, 1271, 1224, 1161, 1025, 982, 625, 576; HRMS (ESI/APCI) m/z (M-OTf)+ calcd.: 489.0646 obsd.: 489.1531.
To a flame dried 5 mL round bottom flask equipped with a stir bar under a nitrogen atmosphere, was added, 3.2 mg of Pd[(−)-sparteine](OTf)2 (0.0050 mmol, 1.0 equiv.) and 500 μL of isopropanol. The orange mixture was stirred for 10 min and a black precipitate began to form. The mixture was diluted by adding 1 μL to 1 mL of isopropanol. The solution was infused into an ESI-MS instrument and no identifiable Pd complexes were observed.
To a flame dried 5 mL round bottom flask equipped with a stir bar under a nitrogen atmosphere, was added, 5.2 mg of styrene (0.050 mmol, 10 equiv.) and 500 μL of isopropanol. Next, 3.2 mg of Pd[(−)-sparteine](OTf)2 (0.0050 mmol, 1.0 equiv.) was added and the orange mixture was stirred for 10 min. The mixture was diluted by adding 0.5 μL to 1 mL of isopropanol. The solution was infused into the (APCI/ESI)-HRMS instrument. HRMS (ESI/APCI) m/z (M)+ for 106Pd calcd.: 445.1835 obsd.: 445.1841. Unfortunately, attempts to isolate the complex result in the formation of black precipitate. Additional attempts to employ (CD3)2CD(OD) as the solvent to characterize the complex by 1H and 13C NMR indicated the presence of multiple species, which could not be identified.
This work was supported by the National Institutes of Health (Grant NIGMS RO1 GM3540). M.S.S. thanks the Dreyfus Foundation (Teacher-Scholar) and Pfizer for their support. K.M.G. thanks Sanofi-Aventis for a graduate research fellowship. Y. I. thanks Kyorin Co., Ltd. for financial support. We would like to thank Dr. Jim Muller for assistance with the ESI-MS experiments. We are grateful to Johnson Matthey for the gift of various Pd salts.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.