We chose to study the SMC under biphasic reaction conditions due to the widespread use of these conditions (\(\sim\)50% of SMCs)10 coupled with the difficulties typically associated with conducting mechanistic studies in such settings. Standard monitoring techniques leveraging infrared (IR) spectroscopy or NMR spectroscopy have difficulties ‘locking’ onto the changing solvent background. The use of 13C kinetic isotope effect was used by Joshia et al. and proved a powerful strategy to study the SMC under biphasic reaction conditions.8i However, this strategy would not lend itself to our ultimate goal of probing the impact of different solvent ratios and additives on the SMC. Manual reaction sampling has also been employed to study the SMC,8c,h,k though reproducibility is typically difficult to achieve in biphasic systems and may explain why these studies tend to limit the proportion of the aqueous phase. Moreover, offline reaction sampling of SMC has been shown to suffer from irreproducibility due to sample aging.8r Recently, we have developed an automated sampling platform which leverages online high-performance liquid chromatography (HPLC) for reaction analysis.11 Notably, this system was shown to provide highly reproducible sampling in heterogeneous settings and thus provides an ideal tool for the current study.11a,b
We chose to target the use of benzylic electrophiles as a model system of study for three reasons; 1) they provide a direct route to the diarylmethane motifs which is broadly present in pharmaceutically relevant compounds,12 supramolecular chemistry,13 as well as being important building blocks in their own right (Fig. 1b);14 2) mechanistic studies on such systems remain exceptionally sparse;15 3) the functional group tolerance of these transformations remain noticeably lacking relative to its C(sp2)-C(sp2) counterpart.16–17 The prevalence of dialkylbiaryl phosphines in SMCs led us to choose XPhos Pd G2 as a representative catalyst.18 The use of a weak inorganic base was targeted due to its widespread use in SMC.10b Thus, the SMC of benzyl bromide (1) with 4-methoxyphenylboronic acid pinacol ester (2) was conducted in a mixture of MeTHF and H2O as the model system.
The reproducibility of the model reaction was probed on our standard automated sampling platform (Fig. 2). Exceptional overlay between two identical reactions was achieved. This control confirms that the system is well stirred and not operating under mass transfer limiting conditions. Moreover, HPLC allowed us to easily monitor the speciation of 2 to the parent boronic acid under the reaction conditions. Notably, the background rate of hydrolysis of 2 was much slower than the rate of product formation, suggesting transmetalation occurs with 2 directly and not the parent boronic acid (see the supporting information Figure S4).8n
With the validity of our reaction monitoring platform confirmed we turned our attention to probing the order of each individual reagent. This will provide an understanding of the turnover limiting step and catalyst resting state, thus providing a foundation upon which to understand the positive or negative impact of solvent changes and additives. We conducted a series of ‘different excess’ experiments for each reagent which can be seen in Fig. 3. A variable time normalization analysis (VTNA) was conducted for each reagent as well (see Supporting Information Figure S5-S12).19 The benzyl bromide electrophile 1 was not observed to have any impact on the rate of reaction (Fig. 3a).The order in 2 was determined to be 0.75 supporting its involvement in the turnover limiting step (Fig. 3b).The palladium catalyst and the base were observed to have orders of 1.0 and 1.8 respectively (Fig. 3c and d). From these data, a turnover limiting oxidative addition or reductive elimination can be ruled out as the former would require a positive order in 1 while the latter would exhibit 0th order behavior in all reagents safe for the catalyst. We believe these data are consistent with a rate determining transmetalation coupled with a base mediated pre-equilibrium. However, these data do not provide insight on whether the pre-equilibrium is between LnPd(aryl)(X) and the arylboronate (Fig. 1, path A) or LnPd(aryl)(X) and LnPd(aryl)(OH) (Fig. 1, Path B).
A set of same excess experiments were then conducted and confirmed the catalyst suffered insignificant catalyst degradation, though notable halide inhibition was observed (see the supporting information Figure S13). We wanted to probe this halide inhibition behavior more thoroughly through examination of different halide salts: KCl, KBr, and KI (Fig. 4). KI was the strongest inhibitor while KCl was the weakest. Similar behavior was reported previously in C(sp2)-C(sp2) SMC and was attributed to a greater solubility of KI in the organic phase negatively impacting the pre-equilibrium.8c Our results further corroborate this hypothesis as considerable electrophile speciation to benzyl iodide (4) was observed in the presence of KI (Fig. 4A, inset) while such speciation remained undetectable in the case of KCl. Several insights can be drawn from this set of experiments; 1) Substitution based speciation of electrophiles should not be ignored when in benzylic systems which could have implications in more complex settings;16h,t,17b,l,n 2) the low inhibition observed by KCl suggest chloride based electrophiles may be ideal when conducting difficult C(sp3)-C(sp2) couplings. The increased reactivity of benzyl chloride electrophiles was confirmed, achieving reaction completion just over 3x faster than the parent benzyl bromide electrophile (see supplementary information Figure S15).
The significant halide inhibition observed led us to target the development of conditions which would minimize this undesired behavior. Milner and coworkers highlighted that moving from THF to toluene in biphasic systems greatly increased reactivity due to the diminished organic phase solubility of the halide salt.8c However, ethereal solvents are regularly employed in SMC and thus this strategy may not be universal.10b We hypothesized that increasing the proportion of the aqueous phase in these biphasic systems may decrease the concentration of halide in the organic phase thus improving reactivity. Despite the widespread use of biphasic reaction conditions, we are unaware of studies probing the effect of this variable on turnover rate. We began by holding the total volume of the reaction constant while manipulating the ratio of water relative to MeTHF (Fig. 5). A clear trend is observed where decreasing the proportion of the organic phase leads to significant rate accelerations. This behavior is likely attributed to the fact that the reaction primarily occurs in the organic phase of the reaction system.20 Thus, decreasing the volume of the organic phase concentrates the reaction despite holding the total volume constant.
Though the increased concentration of the organic phase stands as a logical explanation for the observed behavior, we wanted to disentangle this effect from other potential impacts. To this end, we conducted a set of experiments where the volume of MeTHF was held constant while changing the amount of water present in the reaction system (Fig. 6). This reveals a significant rate enhancement when increasing the volume of water despite decreasing the concentration of the overall system and holding the organic phase concentration constant. We believe this behavior could stem from decreasing the concentration of the halide salts in the organic phase. Overall, these results highlight that simply manipulating the ratio of the aqueous and organic layer can have stunning impacts on reaction rates. To place these results into perspective, the rate increases observed by simply manipulating the ratio of the aqueous and organic phase were the same (Fig. 6) or larger (Fig. 5) than the impact of manipulating the identity of the organoboron nucleophile (vide infra, Figure S26) a parameter that has been a central focus in the advancement of the SMC.21 These results stand in stark contrast to the typical biphasic SMC where the amount of water present is typically kept low.10
Although manipulating the solvent composition proved a powerful strategy to improve turnover rates, we wondered if similar improvements could be achieved with the introduction of PTCs. In the case of path A being dominant, a PTC may increase the concentration of the required 8-B-4 species thus pushing the equilibrium towards the Pd-O-B complex. Alternatively, if path B is dominant, a PTC would be expected to push the equilibrium towards the desired LnPd(aryl)(OH). The use of PTC in SMCs is well documented however,22 determining its primary role in such systems remains difficult to decipher.20,23 This arises as a result of the competing effects of PTC in SMCs; 1) the ability to influence the concentration of base/boronate in the organic phase, and 2) the ability of PTCs to stabilize palladium nanoparticles thus minimizing the aggregation, and consequently inactivation, of the palladium catalyst. We were surprised to find that, despite the widespread use of biphasic reaction conditions, the use of PTC as additives remains exceptionally rare outside nanoparticle forming systems.
We believe our current system of study provides a unique opportunity to probe the impact of PTC in SMCs without the confounding effect of nanoparticle stabilization. We are unaware of any reports where Buchwald precatalyst derived from dialkylbiarylphosphine were shown to be in a nanoparticle forming regime. Moreover, the catalyst stability observed from the ‘same excess’ experiment suggests this catalyst system is not prone to deactivation (see supplementary information Figure S13), thus any impact observed is unlikely due to nanoparticle formation/stabilization.
A series of experiments were conducted using tetrabutylammonium (TBA) salt additives. To our delight, the addition of 0.10 M tetrabutylammonium bromide (TBAB) resulted in complete consumption of the starting material in under 15 minutes (see supplementary information Figure S17). Thus, at minimum, the addition of the TBAB salt results in a 12-fold increase in turnover rate! To properly monitor such a rapid reaction, significant modifications to the reaction conditions were made including reducing the catalyst loading, base, and reaction temperature. With these milder conditions, tetrabutylammonium iodide (TBAI), bromide (TBAB), chloride (TBACl), and hydroxide (TBAOH) were tested for their ability to promote the reaction (Fig. 7). Notably, significant amounts of speciation of both the electrophile (Fig. 7A, inset) and nucleophile (Fig. 7C) were observed. Given the significant halide inhibition observed vide supra, this suggests the negative impact of halide ion must be mitigated by the presence of the TBA counterion. The significant amount of arylboronic acid (6) observed in the presence of the TBA salts strongly supports an increased organic phase concentration of hydroxide. Finally, all TBA salts, safe for TBAI, resulted in rate accelerations relative to the standard conditions with no additive. TBAOH provided the highest activity, only slightly outperforming TBACl.
The substantial rate acceleration caused by the addition of TBA salts was believed to arise through increasing the organic phase hydroxide concentration which could significantly impact arylboronate concentration (Fig. 1, path A) and/or catalyst speciation (Fig. 1, path B). Alternatively, one may suspect the hydrolysis of 2 to the more reactive 6 could account for such an increase in rate as TBA salts were observed to promote its formation (Fig. 7c). To probe this possibility, we obtained time course data from multiple different organoboron nucleophiles including the parent boronic acid (6) (Fig. 8). A neopentyl glycol based boronic acid ester (7) had the highest reactivity while 2 was the slowest. Importantly, the drastic rate accelerations observed through the introduction of TBA salts are not explained by the differences in reactivity between 2 and 6. Noteworthy is the comparatively limited impact the nature of the organoboron nucleophile had on turnover rate. While the use of 7 was able to achieve just above a 2-fold rate acceleration, the addition of TBAB was able to achieve at minimum a 12-fold rate acceleration.
To better understand the mechanism of action of TBA salts, we conducted speciation experiments of 2 and SPhosPd(Ph)(Cl) due to difficulties in the synthesis of the parent C(sp3) oxidative addition complex. Notably, control experiments revealed the impact of TBA salts is general to the C(sp2)-C(sp2) system as well (see supplementary information Figure S29). To probe the Palladium speciation, we set up our typical biphasic reaction system with SPhosPd(Ph)(Cl) in the absence of 1 and 2. The system was equilibrated (see supplementary information Figure S33) following which a sample of the organic phase was taken and analyzed by 31P NMR spectroscopy. We focused on the quantities of monomeric SPhosPd(Ph)(OH), monomeric SPhosPd(Ph)(Cl), and dimeric SPhosPd(Ph)(Cl) as these will dictate reactivity under the reaction conditions (see supporting information for a detailed discussion). Overall, both monomers were present in similar concentrations with the SPhosPd(Ph)(Cl) dimer as the dominant species (Eq. 1). The slight favorability of Pd-Cl complexes is in line with prior reports.8b,c To our surprise, the same experiment conducted in the presence of TBACl resulted in almost the complete loss of the parent SPhosPd(Ph)(OH) complex with significant shift towards the formation of Pd-Cl complexes (Eq. 1). This can be rationalized through the combination of the weak Pd-OH bond strength, coupled with the preference of PTCs to increase the organic phase concentration of the softer chloride anion relative to hydroxide.24
We then turned our attention to the speciation of 2. In this case, we set up our standard biphasic reaction system in the absence of 1 and the palladium catalyst. A sample of the organic layer was taken for analysis by both 19F and 11B NMR spectroscopy. Under standard reaction conditions, 8-B-4 species remained below the detection limit in the organic phase of the reaction (Eq. 2). The parent pinacol ester was the dominant species with small amounts of boronic acid byproduct as expected. In contrast, the presence of TBACl led to the organic phase being dominated by the arylboronate species while some of the parent 6-B-3 species remained (Eq. 2). This result can be rationalized through the ability of PTCs to increase the organic phase concentration of hydroxide, coupled with the low pKa of arylboron species (see supporting information for a detailed discussion).24
Overall, these results suggest that without the TBA additive, the standard conditions are likely dominated by the oxo-palladium pathway (Fig. 1, pathway B). Although both the Pd-OH and Pd-Cl species are present in appreciable amounts, the arylboronate remained below the detection limit in the organic phase. Thus, the positive order in base and negative order in halide salts can be interpreted by manipulating the pre-equilibrium of the catalyst in pathway B while the positive order in palladium and 2 arise from the rate determining transfer of the aryl species onto the catalyst. In contrast, the introduction of TBA salts increases the organic phase concentration of both halides and hydroxides. This results in significant shifts in the speciation of the catalyst towards LnPd(aryl)(Cl), and 2 towards its 8-B-4 complex. These results suggest that the rate enhancements observed from TBA salts result from shifting the dominant mode of transmetalation from the oxo-palladium pathway (Fig. 1, path B) to the boronate pathway (Fig. 1, path A).
To further validate our conclusion, we carried out a series of different excess experiments while in the presence of TBAB (see supplementary information Figure S19-26.). This revealed the system remained in a turnover limiting transmetalation and thus the rate increases observed are not explained by manipulating the kinetic regime. This was followed by a set of experiments set out to study the impact of halide under these new conditions (Fig. 9). A direct comparison of TBAB versus the use of tetrabutylammonium triflate (TBATf) reveals a significant positive impact of the bromide under these conditions. This stands in stark contrast to the comparison of KBr versus potassium triflate (KTf). The later case, as was highlighted earlier, shows a significant negative impact in the halide ion. This fundamental change in the impact of the halide ion further corroborates a change in the dominant mechanism of transmetalation from pathway B to pathway A occurs in the presence of TBA salts.
Following the completion of our mechanistic study, we wanted to probe the impact of the acquired knowledge on both the catalyst loading as well as the substrate scope of this transformation. The former was probed by testing our model substrate at increasingly diminished catalyst loading (Eq. 3). TBAOH was chosen as the additive as it had resulted in the highest catalyst turnover rate (Fig. 7). Notably, loadings as low as 0.001 mol% were possible while retaining exceptionally high yields. A control reaction with no added catalyst confirmed that such reactivity arises from the catalyst and not the presence of trace metals in the starting materials. This result is one of the lowest catalyst loadings reported to date involving benzylic electrophiles.16–17
We next turned our attention to the substrate scope of this reaction. The use of exceptionally low catalyst loadings (0.1 mol% ≥) for the SMCs with benzylic electrophiles have been reported however the substrate scope of these reports remains exceptionally limited. To our knowledge, reports where at least one heterocyclic substrate is included range from 1.0 mol% to 10 mol% Pd loading16 with Botella et al.16b and Chahen et al.16p having a single example at 0.05 mol% and 0.004 mol% respectively. Thus, we chose to probe the substrate scope using 0.1 mol% catalyst loading. Our model system produced the desired product (3aa) in 93% yield upon scale up and isolation (Scheme 1). We then proceeded to probe the substrate scope of the organoboron coupling partner. Both electron rich (3ba-da) and electron poor (3ea-3fa) aryl nucleophiles proved competent under these conditions. Heterocyclic pyridyl (3ga-ha) nucleophiles produced the product in good yields. The use of thiophene nucleophiles were also well tolerated under the reaction conditions (3ia-la) however, in some cases (3ia-ja) the competitive formation of benzyl alcohol was observed. This issue was largely resolved by replacing TBAOH with TBAB. Pyrrole (3ma) and pyrazole (3na) derivatives were also well tolerated however the latter required the use of slightly higher catalyst loading as well as TBAB to reduce competitive hydrolysis. These conditions are also compatible with the use of a styryl nucleophile with no issues. Finally, we wanted to probe the applicability of our conditions on base labile substrates due to the increased organic phase hydroxide concentration in the presence of TBA salts. An organoboron species bearing a methyl ester (3qa) or an enolizable ketone (3ra) were well tolerated, providing the desired product in an hour with no signs of competitive degradation.
We then turned our attention to the substrate scope of the electrophile (Scheme 2). The use of both electron poor (3ab-ad) and electron rich (3ae) coupling partners were well tolerated under the reaction conditions. Thiophene derivatives provided the desired diarylmethanes in high yield (3af-aag). Notably, these conditions were observed to produce the monocoupled product at the sp3 center selectively over the sp2 (3ag) when both are present on the electrophilic motif in line with prior reports.16h Increasing the loading of the organoboron coupling partner drives the reaction to the dicoupled product (3aag). Lewis basic pyridyl motifs were observed to be competent coupling partners when present within the electrophile (3ah-3aj). Finally, we wanted to probe the applicability of further substitutions at the benzylic center. The introduction of a methyl group resulted in modest yield (3ak). It Is well known that β-hydride elimination can be competitive under such conditions producing styrene as a dominant byproduct.16f,17j Our optimized conditions provided the desired product in acceptable yield though styrene was observed as a byproduct. Bromodiphenylmethane derivatives have also been reported to be difficult coupling partners in SMCs.16f We were pleased to obtain the desired product in modest yield again highlighting the significant beneficial effect of the incorporation of PTCs.
In conclusion, we have conducted a thorough kinetic study of the SMC with benzylic electrophiles under biphasic reaction conditions. Our automated sampling system provided a robust platform to achieve reproducible reaction sampling despite the biphasic nature of the reaction mixture. A summary of findings in this study is as follows:
-
The addition of PTCs resulted in remarkable rate accelerations when used as additives. While these additives are commonly used in Pd nanoparticle catalyzed systems for their stabilizing effects, our results indicate that the observed rate enhancements in our system primarily arise from their phase transfer behavior. Thus, these additives should not be overlooked when optimizing biphasic SMC with molecular Pd catalysts which has to date been scarcely employed.
-
The rate accelerations of PTCs were found to be the result of a shift in the dominant mode of transmetalation from the oxo-palladium pathway (without additives), to the boronate based pathway (with additives).
-
Optimizing the solvent ratio when applying biphasic reaction conditions should not be overlooked. Rate accelerations were similar or larger than those observed when changing the nature of the boron species, simply by reducing the volume, or the proportion of the organic layer.
-
The significant inhibitory effect of KX salts under standard conditions can be explained by an unfavorable effect on the oxo-palladium pre-equilibrium.
The findings in this study were leveraged to develop a SMC of benzylic electrophiles with an exceptionally broad scope while holding the catalyst loading low. Conceptually, the introduction of TBA salts turns the presence of the halide byproduct from a liability to a beneficial species. We believe these findings will be particularly impactful in the development of telescoped processes or cascades, where significant buildup of halide byproducts is inevitable.