Development of an Aryl Amination Catalyst with Broad Scope Guided by Consideration of Catalyst Stability

Abstract

We have developed a new dialkylbiaryl monophosphine ligand, GPhos, that supports a palladium catalyst capable of promoting carbon–nitrogen cross-coupling reactions between a variety of primary amines and aryl halides; in many cases these reactions can be carried out at room temperature. The reaction development was guided by the idea that the productivity of catalysts employing BrettPhos-like ligands is limited by their lack of stability at room temperature. Specifically, it was hypothesized that primary amine and N-heteroaromatic substrates can displace the phosphine ligand, leading to the formation of catalytically dormant palladium complexes that reactivate only upon heating. This notion was supported by the synthesis and kinetic study of a putative off-cycle Pd complex. Consideration of this off-cycle species, together with the identification of substrate classes that are not effectively coupled at room temperature using previous catalysts, led to the design of a new dialkylbiaryl monophosphine ligand. An Ot-Bu substituent was added ortho to the dialkylphosphino group of the ligand framework to increase stability of the most active catalyst conformer. To offset the increased size of this substituent, we also removed the para i-Pr group of the non-phosphorous-containing ring, which allowed the catalyst to accommodate binding of even very large α-tertiary primary amine nucleophiles. Compared to previous catalysts, the GPhos-supported catalyst exhibits better reactivity both under ambient conditions and at elevated temperatures. Its use allows for the coupling of a range of amine nucleophiles, including: (1) unhindered, (2) five-membered-ring N-heterocycle-containing, and (3) α-tertiary primary amines, each of which previously required a different catalyst to achieve optimal results.

Graphical Abstract

Introduction

The coupling of aryl (pseudo)halide electrophiles with amines to form carbon–nitrogen (C–N) bonds is an important transformation with applications in a variety of fields. In particular, transition-metal-catalyzed aryl amination reactions are one of the most used reaction classes in the synthesis of pharmaceutical candidates.^1,2 Palladium-based catalysts are among the most effective for catalytic aryl amination reactions.³ We have a longstanding interest in the development of new ligands for palladium-catalyzed C–N bond-forming reactions.^{4, 5} Specifically, our group has created a variety of dialkylbiaryl monophosphine ligands to support Pd catalysts that are highly active for the coupling of many classes of aryl electrophiles with a broad range of amine nucleophiles.

The mechanism by which palladium catalyzes C–N cross-coupling reactions is well documented (Scheme 1A).^5,6 The elementary steps of these reactions, including oxidative addition (Scheme 1B),⁷ amine binding (Scheme 1C),⁸ and reductive elimination (Scheme 1D),^9,10 occur at or near ambient temperature using Pd complexes ligated by dialkylbiaryl monophosphines. However, most currently used synthetic protocols that exhibit a broad substrate scope are carried out above room temperature. Many early reports of Pd-catalyzed aryl amination reactions included examples of reactions run at room temperature, but the substrate scopes of these protocols were generally limited and included very few primary aliphatic amine nucleophiles.^{5c–e, 11 – 13} By increasing the reaction temperature and developing new ligands, our group has been able to improve both the catalyst reactivity and stability.^4,14–16

Scheme 1.

Mechanistic Hypothesis and Previous Studies of Elementary Steps.

The vast majority of aryl amination reactions that proceed at room temperature use alkoxide bases,^11–13 so we chose to employ NaOt-Bu during our reaction development. Based on previous results from our group indicating that amines can displace dialkylbiaryl monophosphine supporting ligands,¹⁷ we anticipated that a key challenge to facilitating a broader scope of C–N coupling reactions at room temperature would be avoiding the production of off-cycle aryl–Pd species such as V¹⁸ and VI¹⁹ that form through the reaction of on-cycle Pd complexes with excess primary amine or N-heterocycle-containing substrates, respectively (Scheme 1A). The formation of both types of Pd complexes (i.e., V and VI) likely has a negative impact on productive catalytic turnover,^20–22 and minimizing their production could enable more effective catalysis, particularly at room temperature.

Herein, mechanistic studies and ligand design informed the development of a practically useful catalyst that promotes C–N cross-coupling reactions involving a variety of aryl (pseudo)halides and primary amine nucleophiles. The GPhos-supported catalyst can operate at room temperature in many cases, which allows for a greater tolerance of base-sensitive substrates relative to previous catalyst systems that operate above room temperature, while retaining the desirable qualities of those systems, such as low catalyst loadings and fast reaction kinetics.²³ The catalyst can accommodate sterically hindered aryl halides and amines, which were not successfully coupled by our group’s recently developed Pd catalyst system that employs an amine base.^15d,15e In addition to displaying improved stability and reactivity at room temperature, the GPhos-supported catalyst system shows high activity when heated, enabling the coupling of substrates that do not work well at room temperature. Altogether, the precatalyst based on GPhos can perform the function of catalysts based on three different ligand families: BrettPhos (unhindered primary amines),^14b,c PhCPhos and (t-Bu)PhCPhos (α-tertiary amines),^15a and EPhos (aryl halides or amines containing five-membered-ring N-heterocycles).^15c

Results and Discussion

Development of New Catalysts for Room Temperature Aryl Amination Reactions

Initial testing of catalytic reactions indicated that couplings of ortho-substituted aryl bromide electrophiles with primary amines were especially challenging for catalysts based on BrettPhos (L1) and EPhos (L2) at room temperature^24,25 (see Supporting Information). As noted in the Introduction, we hypothesized that catalyst deactivation is a key factor in the lack of general success for Pd-catalyzed C–N coupling reactions carried out at room temperature. An alternative explanation is that the catalyst is stable, but rate of the productive C–N coupling reaction is slow at room temperature. To differentiate between these two possibilities, reaction calorimetry was used to monitor the progress of the reaction of 2-bromo-1,4-dimethylbenzene and n-propylamine (Figure 1A). Catalysts based on BrettPhos (L1) and EPhos (L2) are among the most active catalysts that our group has developed for the arylation of primary amines, so the oxidative addition complex (OAC) precatalysts bearing these ligands (OA1, OA2) were tested initially.²⁶ In both cases, these catalysts produced small amounts of the C–N coupled product (<10%), but the catalyst activity decreased within the first 10–30 min of the reaction. Additionally, after 1 h of reaction time, free BrettPhos or EPhos was the only detectable phosphorus-containing species in the ³¹P NMR spectrum (Figure 1B, see Supporting Information for analogous data for the reaction using OA2). These results are in agreement with the hypothesis that sequestration of the palladium as a non-phosphine-ligated complex is a cause of catalyst deactivation and the resulting low yields for the reactions carried out at room temperature using OA1 and OA2. However, because the OA1- and OA2-derived catalysts showed activity in the first few minutes of the reactions at room temperature, we anticipated that high yields of the C–N coupled product could be achieved with a catalyst that was more stable toward deactivation under these conditions.

Figure 1.

(A) Comparison of reaction time courses as measured by reaction calorimetry for the reaction shown with catalysts OA1–OA6 (Ar = 4-(2-(trimethylsilyl)ethyl benzoate)). (B) ³¹P NMR spectrum of the reaction employing OA1 as the precatalyst after 1 h. Reaction conditions: 1.0 mmol 2-bromo-1,4-dimethylbenzene, 1.4 mmol n-propylamine, 1.4 mmol NaOt-Bu, 0.1 mmol n-dodecane (internal standard), 2.5 or 5.0 μmol OAn in THF (1.0 M [2-bromo-1,4-dimethylbenzene]) maintained at 26.0 °C in OmniCal calorimeter. Note: OA1 refers to precatalyst with L1, OA2 to that with L2, etc. Reaction GC conversions for each catalyst: OA1 = 9%, OA2 = 9%, OA3 = 60%, OA4 = 65%, OA5 = 90%, OA6 = 96%.

A key difference between EPhos and BrettPhos is the Oi-Pr substituent at the 3-position in EPhos (vs. OMe in BrettPhos), which was designed to greatly favor the C-bound conformation of the OAC (Figure 2A).^15c Because the O-bound isomer exhibits slower reductive elimination^10a and can thus behave as an off-cycle Pd reservoir,^15c we hypothesized that adding a larger substituent at the 3-position of the ligand framework could impart additional stability onto the resulting catalyst by further favoring the C-bound isomer relative to the O-bound isomer. In accord with this hypothesis, changing the C3-substituent from Oi-Pr (OA2) to Ot-Bu (OA3, OA4) significantly decreased the rate of catalyst deactivation relative to the productive reaction rate, although the reaction still failed to reach full conversion within 1 h (Figure 1A). When the 6-OMe group that is present in BrettPhos, but not EPhos, was added to the ligand framework containing the Ot-Bu substituent (OA5, OA6), the amination process was fast enough relative to catalyst deactivation to nearly reach completion within 1 h. The progression from OA1 and OA2 to the most active catalyst, OA6, shows the benefit of improving the ratio of the rate of productive reaction to that of catalyst deactivation. Consideration of catalyst stability is less often an explicit focus of aryl amination catalyst development efforts, but it appears to be an important metric in C–N cross-coupling reactions.

Figure 2.

(A) C,O-isomerism observed in some dialkylbiaryl monophosphine-based OACs. A bulkier R group decreases the relative population of the O-bound isomer. (B) Amine binding mode previously proposed for XPhos-supported OAC.²⁷ (C) Comparison of the performance of precatalysts (OA5, OA6; Ar = 4-(2-(trimethylsilyl)ethyl benzoate)) for the coupling of α-branched primary amines. Reaction conditions: 0.4 mmol 2-bromo-1,4-dimethylbenzene or 1-(tert-butoxy)-4-chlorobenzene, 0.56 mmol cyclohexylamine or tert-octylamine, 0.56 mmol NaOt-Bu, 0.04 mmol n-dodecane (internal standard), 0.4 or 2.0 μmol OAn in 0.2 mL THF at RT.

We next sought to examine each catalyst’s reactivity with different aryl halides and amines, with a particular emphasis on bulkier α-branched primary amines. It has previously been suggested that amine binding and/or deprotonation may occur when the Pd is positioned away from the sterically hindered triisopropyl aryl fragment of the ligand (Figure 2B).²⁷ Such an amine binding mechanism is unlikely with catalysts supported by ligands L2–L6, which force their corresponding OACs (OA2–OA6) into the C-bound conformation. However, we hypothesized that the catalyst’s activity might be increased in coupling reactions involving more hindered α-branched amines if the 4’-i-Pr group were removed to reduce the steric hindrance associated with the transition states for amine binding and/or deprotonation. This modification proved critical for enabling the coupling of some α-branched primary amines. For example, OA6 is significantly more effective than OA5 for coupling reactions involving cyclohexylamine or tert-octyl amine nucleophiles (Figure 2C). Overall, employing OA6 provided the best combination of catalyst stability and substrate scope of the catalysts tested,²⁸ likely because it merges the most important features of ligands used in previous catalytic systems (Figure 3): a large substituent ortho to the dialkylphosphino group (cf. EPhos) to stabilize the catalyst, an electron-donating methoxy group in the 6-position (cf. BrettPhos) to improve the reaction rate, and a hydrogen as the 4’-substituent (cf. PhCPhos, (t-Bu)PhCPhos) to enable the binding of sterically demanding amine nucleophiles.

Figure 3.

Common dialkylbiaryl monophosphine ligands used to support Pd catalysts for the arylation of different types of primary amine nucleophiles. Key ligand features are highlighted.

Assessment of C–N Coupling Catalysis at Higher Temperatures

Although catalysts based on BrettPhos (L1) often do not produce C–N coupled product in high yield at room temperature, they are effective catalysts at higher temperatures.^14b,c To reconcile the difference in catalyst performance between reactions carried out at room temperature and those that are heated, several mechanistic experiments were performed using L1-based catalysts. The studies were initiated by collecting reaction time course data for a model amination reaction similar to the one used for the ligand development described above (cf. Figure 1). Two identical series of reactions were allowed to proceed for 1 h at room temperature, during which time they each produced approximately 20% yield of coupled product (Figure 4). Subsequently, one series of reactions was allowed to continue at room temperature for up to 24 hours. During this extended reaction period, minimal additional product was formed, consistent with the result shown in Figure 1A. The other series of reactions was heated to 90 °C after the first hour of reaction time at room temperature. In this case, a quantitative yield of product was formed after ~7 h (~6 h at 90 °C). These results, taken together with the results in Figure 1, indicate that C–N coupling promoted by the L1-supported OAC can occur readily at room temperature, but when the L–Pd complex deactivates and only free L1 is observed in solution (Figure 1B), the reaction mixture must be heated to facilitate productive C–N coupling. The need for heating after dissociation of the phosphine ligand suggests that the re-entry of off-cycle species (e.g., V/VI, Scheme 1A) is an elementary step that necessitates higher reaction temperature in many catalytic protocols.

Figure 4.

Assessment of unheated and heated OA1’-catalyzed aryl amination. Reaction conditions: 0.5 mmol 2-bromo-1,4-dimethylbenzene, 0.7 mmol n-hexylamine, 0.7 mmol NaOt-Bu, 0.05 mmol n-dodecane (internal standard), 2.5 μmol OA1’, 2.5 μmol L1 in 0.5 mL 1,4-dioxane at RT (1 h time point) followed by RT (gray box) or 90 °C (red box). Calibrated GC yields. See Supporting Information for full details.

To probe whether putative off-cycle Pd complexes similar to V (Scheme 1A), formed via displacement of the supporting ligand, can serve as competent catalyst precursors, complex A (Figure 5) was prepared.²⁹ When the reaction mixture containing the model coupling partners was heated to 90 °C in the presence of 0.5 mol% A as the Pd source and 1.0 mol% L1 (to match the amount of catalyst and ligand used in Figure 4), a high yield of product was observed after 24 hours (Figure 5). When our new ligand, L6, was used in place of L1 (in combination with A), the reaction was complete within 1 h at 90 °C. This result indicates that L6 promotes a higher population of active catalyst (cf. I–IV, Scheme 1A) relative to A (cf. V, Scheme 1A) than L1, and/or the population of Pd that enters the productive cycle is significantly more active when supported by L6 than with L1. At room temperature, the L6-based catalyst showed the highest ratio for the rate of the productive reaction relative to the rate of catalyst deactivation. We suspect that the same structural features of L6 that led to this high ratio at room temperature are also responsible for the higher reactivity of the L6-based catalyst relative to the L1-based catalyst observed at 90 °C using A as the catalyst precursor. Reactions with A and L1 (or L6) that were performed at room temperature did not yield any desired product. The notion that non-phosphine-ligated off-cycle Pd species, such as A, may recombine with free ligand to form on-cycle catalysts when heated is consistent with the beneficial effect of added equivalents of dialkylbiaryl monophosphine ligand in many Pd-catalyzed C–N cross-coupling reactions.⁴

Figure 5.

Reaction time course using A as a precatalyst. Reaction conditions: 0.5 mmol 2-bromo-1,4-dimethylbenzene, 0.7 mmol n-hexylamine, 0.7 mmol NaOt-Bu, 0.05 mmol n-dodecane (internal standard), 2.5 μmol A, 5.0 μmol L1 or L6 in 0.5 mL 1,4-dioxane at 90 °C. Calibrated GC yields. See Supporting Information for full details. Dashed lines are intended to guide the eye and do not reflect a kinetic fit.

Only a small amount of free L1 was observed when excess n-hexylamine was stirred with OA1’ at room temperature for 1 h, suggesting that displacement of L1 occurs from an intermediate other than an OAC (cf. II/III, Scheme 1). This contrasts with previous studies in which it was observed that the addition of excess primary amine to P(o-tol)₃- or Pt-Bu₃-ligated Pd OACs (cf. II, Scheme 1) resulted in the formation of phosphine-free compounds analogous to A.¹⁸ Additionally, Hartwig observed a similar bis(amine) Ni complex when a (BINAP)Ni(Ar)Cl species was treated with an excess of primary amine.³⁰ Although A catalyzed C–N bond formation in the presence of L1 when heated (Figure 5), related bis(amine)Pd(Ar)Br and bis(amine)Ni(Ar)Cl complexes do not always catalyze aryl amination reactions. For example, the combination of P(o-tol)₃ and bis(amine)Pd(Ar)Br complexes did not form an active catalyst.^18a Further, the aforementioned Ni-based bis(amine) complex could not promote stoichiometric C–N coupling when heated in the presence of BINAP supporting ligand.³⁰ These collective observations suggest that complexes such as A are relevant in many primary amine arylation reactions, and the facility with which they re-enter the productive catalytic cycle depends on both the metal (e.g., Pd or Ni) and the supporting ligand.

Scope of Room Temperature C–N Coupling Reactions using OA6 as the Precatalyst

The use of precatalyst OA6^31,32 enabled the room temperature coupling of aryl (pseudo)halides with a variety of primary aliphatic amine and aniline coupling partners with low catalyst loadings and short reaction times (typically 1 h).²⁴ Ortho-substituted aryl chlorides (3k, 3p) and aryl bromides (3a, 3l) were coupled efficiently, even though these are difficult classes of electrophiles for catalysts ligated with BrettPhos (L1) and EPhos (L2) when the reactions are run at room temperature (see Supporting Information). An unhindered aryl iodide (3b) was coupled in high yield, which is noteworthy because aryl iodide electrophiles often show reduced reaction rates relative to aryl bromides.³³ Finally, an unhindered aryl triflate was readily coupled with an aniline (3c). In some cases, heat release was noted in the first several minutes of the reactions,^24,25 but in only one instance was the reaction negatively affected by this exotherm: on 1.0 mmol scale, a decreased product yield was observed for 3c when the reaction vial was not submerged in a room temperature water bath, a modification that was not needed when the coupling was carried out on 0.2 mmol scale.³⁴

Aryl halides or amines containing a free primary alcohol (3d, 3t),³⁵ secondary amine (3n, 3r), or amide (3c, 3o) functional group gave high yields, and several N-heterocycle-containing aryl halides and amines are featured as substrates. Despite the high reactivity of the OA6 catalyst, several chemoselective reactions were achieved. For example, methyl 4-bromobenzoate was selectively coupled in the presence of an aryl chloride (3e). Additionally, 4-aminopiperidine was coupled predominantly at the primary amino group (3n). For the reaction of aryl bromide 1e, NaOMe was used as the base to avoid competitive transesterification.³⁶ Reduction of the aryl halide to the arene was not observed,³⁷ although a small amount of aryl methyl ether was observed when the crude reaction mixture was analyzed using ¹H NMR (~5%).

Procedures using OA6 were able to efficiently couple sterically hindered primary amines under room temperature conditions, an advantage of this method relative to our group’s previous work using soluble amine bases, which could not accommodate α-tertiary amines or ortho-substituted anilines.^15d For example, the reaction of 2-chloropyrazine and tert-octylamine occurred under much milder conditions than those formerly required (3f).^15a,38 In addition to hindered aliphatic amines, ortho-substituted anilines (3g, 3h, 3i) could be coupled in high yield, though in the cases of an extremely hindered aniline (3h) or a hindered electron-deficient aniline (3i), a longer reaction time was required (24 h). Electron-deficient anilines (3e, 3i, 3j, 3k, 3o, 3p, 3t) were efficiently converted to product under the coupling conditions, including fluorinated anilines (3i, 3j), which have been described as challenging nucleophiles in Pd-catalyzed coupling reactions.³⁹ Some of the electron-deficient anilines (3i, 3o, 3p) performed best when NaOPh was used in place of NaOt-Bu as the base, perhaps because these anilines are sufficiently acidic to be deprotonated prior to binding to the Pd catalyst (cf. III, Scheme 1A) or because of their instability in the presence of strong base.⁴⁰ Several potentially base-sensitive functional groups were also accommodated,⁴¹ including an N-trifluoroethylaniline (3l),⁴² nitrile (3p), and several substrates containing acidic C–H bonds (3m, 3n, 3o).

In several cases, we compared our conditions to those used for similar or identical coupling reactions that were previously reported. For example, reactions involving fluorinated anilines 3i and 3j were formed using less catalyst (6–15-fold) and, in the case of 3j, shorter reaction time (1 h vs. 24 h) while still operating at room temperature,³⁹ even though the anilines we employed are either more electron-deficient or more sterically hindered than those in the previous study. Compound 3p was previously prepared by our group using a BrettPhos-based catalyst.^14c Under our new conditions, we were able to simultaneously reduce the amount of catalyst (by 3-fold), temperature (RT vs. 110 °C), and reaction time (1 h vs. 14 h), highlighting the improved reactivity of OA6. Finally, the room temperature conditions allowed for the use of NaOt-Bu to prepare 3l. Previously published conditions heated the reaction mixture to 100 °C in the presence of a weaker base, KOPh, necessitating longer reaction times (6 h), but our room temperature conditions using NaOt-Bu resulted in full conversion to product within 1 h, while maintaining a catalyst loading similar to that employed in the previous report.⁴²

In some instances, poisoning or slowing of reactions has been observed with N-heterocycle-containing aryl halide or amine substrates. For example, 2-aminopyridine can function as a ligand for Pd(II). Still, 3k was formed efficiently at room temperature in 1 h, even though similar coupling processes previously required heating (80–100 °C) with more catalyst (1.7–13-fold) for longer reaction times (24–30 h).⁴³ Finally, 3f had been previously prepared using our PhCPhos-based catalyst.^15a Now we are able to use a shorter reaction time (1 h vs. 24 h) at a lower temperature (RT vs. 120 °C), while still using less catalyst. The faster rate of C–N bond formation (i.e., shorter reaction time) and lower temperature avoid the formation of the ArOt-Bu side product, which was competitively produced under the previous reaction conditions, resulting in a lower yield than that observed here (90% vs. 50%). It is possible that such a significant improvement is observed for this reaction because the large Ot-Bu group on the ligand “protects” the catalyst from degradation by the pyrazine, while the removal of the i-Pr group in the ligand’s 4’-position still allows for binding of the sterically demanding tert-octylamine nucleophile.

We next examined OA6 in the reactions of more complex substrates under our room temperature conditions. C–N cross-coupling reactions involving pharmaceutical derivatives possessing multiple functional groups have been shown to exhibit a high failure rate.⁴⁴ The OA6-based catalyst system enabled the coupling of several high-complexity molecules while generally allowing for low catalyst loadings and short reaction times. These included the arylation of a pyridine- and pyrimidine-containing aniline (2q), which is a fragment of the anti-Leukemia drug Gleevec, to form 3q. Additionally, several aryl halide-containing pharmaceuticals bearing multiple functional groups, such as amoxapine (1r), loratadine (1s), perphenazine (1t), and etoricoxib (1u), were efficiently transformed to the C–N coupled product.

Scope of C–N Coupling Reactions using OA6 Catalyst with Heating

As noted in the Introduction, most broad-scope protocols for Pd-catalyzed aryl amination are carried out above room temperature. We endeavored to compare the effectiveness of OA6 to that of previous catalysts under such conditions, and to examine whether reactions that were unsuccessful at room temperature using OA6 would work with heating.⁴⁵ First, we examined several coupling reactions that were successful at room temperature, and that did not contain functional groups that would be problematic at 90 °C, to probe the general performance of the OA6 catalyst when heated (Figure 8). Although these exact products (3a, 3d, 3k, 3o) have not been previously prepared using catalysts supported by dialkylbiaryl monophosphine ligands, the amount of OA6 used for these reactions is at or below the levels previously reported by our group for the simplest coupling reactions involving primary aliphatic amines.^14b,c

Figure 8. Scope of the aryl amination with heating^a,b.

^aIsolated yields are reported as the average of two runs. Standard reaction conditions: aryl halide (1.0 mmol), amine (1.4 mmol), NaOt-Bu (1.4 mmol), [x mol%] OA6, THF (0.5 mL), 90 °C, 1 h. ^bPrevious conditions refer to previously published conditions for the same or similar coupling reactions. Pd = Pd loading, L = total ligand loading. ^cRT results from Figure 6. ^d1.4 equiv NaOPh. ^e2.4 mmol NaOt-Bu, 2.5 mL THF. ^f75 °C. ^gReaction conditions: aryl halide (1.0 mmol), amine (1.2 mmol), NaOPh (1.2 mmol), [x mol%] OA6, 2-MeTHF (4 mL), 100 °C, 3 h.

In addition to the coupling reactions repeated from Figure 6, we evaluated reactions that were previously reported by our group with other ligands. For example, under the conditions employed for 3aa, the amount of catalyst (decreased 10-fold) and reaction time (1 h vs. 20 h) were both substantially improved relative to those with a BrettPhos-based catalyst.^14c Additionally, OA6 performed much better than PhCPhos- or (t-Bu)PhCPhos-based catalysts for coupling reactions involving α-tertiary primary amines,^15a consistent with our observation at room temperature (cf. Figure 6, 3f). Compounds 3bb, 3cc, and 3dd were prepared using OA6 with less catalyst (4–5-fold) and shorter reaction times (1 h vs. 6–24 h) than the previous report.^15a

Figure 6. Substrate scope of the room temperature aryl amination protocol.^a,b.

^aIsolated yields are reported as the average of two runs. Standard reaction conditions: aryl halide (1.0 mmol), amine (1.4 mmol), NaOt-Bu (1.4 mmol), [x mol%] OA6, THF (0.5 mL), RT, 1 h. ^bPrevious conditions refer to previously published conditions for the same or similar coupling reactions. Pd = Pd loading, L = total ligand loading. ^c1.4 equiv NaOMe, 45 min reaction time. ^d24 h. ^e1.4 equiv NaOPh.

Certain reactions involving five-membered-ring N-heterocyclic substrates were not effective at room temperature. For example, 3ee and 3ff gave no yield at room temperature.⁴⁶ Despite these difficulties under room temperature conditions, at higher temperatures the GPhos-based OA6 precatalyst enabled the coupling reactions that formed 3ee and 3ff with less catalyst (3–4-fold) than our group’s previously reported EPhos (L2)-based catalyst, under otherwise identical conditions.^{15c, 47} Additionally, imidazole-containing amines gave low product yields at room temperature, which could be improved in some cases upon using heated reaction conditions (see Supporting Information for details). Although these coupling reactions are quite different than the model reaction in Figure 1, we suspect that the improved reactivity of OA6 relative to the EPhos-based catalyst for these reactions is due to the improved stability of the catalyst toward deactivation by N-heterocyclic substrates.

Finally, we investigated the use of several common alternative Pd sources (with free L6) as catalyst precursors, to compare their performance to OA6 (Figure 9). Of these, only [Pd(cinnamyl)Cl]₂/L6 formed an active catalyst at room temperature (3a, 3j). This combination performed as well as OA6 for the reaction of 2,6-difluoroaniline to provide 3j, but gave a lower yield for the coupling of an ortho-substituted bromoarene with a primary aliphatic amine (3a). Using Pd₂dba₃/L6 at room temperature resulted in no yield of 3j. The Pd(OAc)₂/L6 catalyst system required heating in the presence of water to form an active catalyst,⁴⁸ which could then catalyze the formation of 3j at room temperature, albeit with a lower yield than reactions with OA6 or [Pd(cinnamyl)Cl]₂/L6. At 90 °C, all of the Pd sources tested were capable of producing significant amounts of 3bb, though the Pd(OAc)₂-based catalyst system performed significantly better with the water activation protocol.⁴⁸ While in some cases the reaction yields using these alternative Pd sources were comparable to those obtained using OA6, none equaled the overall effectiveness of OA6 as a precatalyst. From the perspective of convenience, the use of a one-component precatalyst (containing both ligand and Pd) has advantages on small scale. For larger scale reactions, a variety of Pd precursors can be used with L6.

Figure 9. A comparison of reactions employing OA6 and other, commonly employed, Pd sources.^a.

^aYields determined by ¹H NMR. Standard reaction conditions: aryl halide (1.0 mmol), amine (1.4 mmol), NaOt-Bu (1.4 mmol), [x mol%] Pd, [x mol%] L6 (Pd:L6 = 1:1), THF (0.5 mL), RT or 90 °C, 1 h. ^bResults from Figure 6 (3a, 3j) and Figure 8 (3bb). ^cReaction conditions: aryl halide (0.4 mmol), amine (0.56 mmol), NaOt-Bu (0.56 mmol), [x mol%] Pd(OAc)₂, [2x mol%] L6 (Pd:L6 = 1:2), THF (0.2 mL), RT or 90 °C, 1 h. ^dPd:L6 = 1:2. Using water preactivation protocol.⁴⁸

Conclusions

Guided by a combination of mechanistic analysis and ligand design, we developed a new dialkylbiaryl monophosphine ligand, GPhos (L6), that supports a palladium catalyst capable of promoting highly efficient coupling between a variety of aryl halide and primary amine coupling partners. The OA6 catalyst system derived from GPhos enabled room temperature C–N coupling reactions with substantially more complex substrates than had previously been reported, with high levels of efficiency, both in terms of catalyst required and reaction time. Certain coupling reactions involving five-membered-ring N-heterocycle-containing substrates required heating, but when heated these reactions proceeded in excellent yield. Overall, the new catalyst system promotes the coupling of a wider range of amines than our group’s previously described biarylphosphine-supported systems with equal or greater efficiency. We identified and synthesized a bis(amine)Pd–aryl complex (A), a putative off-cycle catalyst species.²⁹ This complex was not capable of entering the catalytic cycle at room temperature but was found to be a competent catalyst precursor at 90 °C, which is a temperature typical of many Pd-catalyzed C–N coupling protocols. When heated with A as the Pd source, the GPhos-based catalyst exhibited a much faster reaction rate than the corresponding BrettPhos-based catalyst. We believe the greater efficiency of the new catalyst at room temperature compared to previously developed dialkylbiaryl monophosphine-based catalysts arises because the new catalyst exhibits an improved ratio of the rate of productive on-cycle catalytic steps relative to that of detrimental catalyst deactivation. At elevated temperatures, the increased reactivity arises from a combination of this improved ratio with accessible activation of off-cycle species back into the productive cycle.

Figure 7. Scope of the room temperature aryl amination of drug-like substrates.^a.

^aIsolated yields are reported as the average of two runs. Standard reaction conditions: aryl halide (1.0 mmol), amine (1.4 mmol), NaOt-Bu (1.4 mmol), [x mol%] OA6, THF (0.5 mL), RT, 1 h. ^b1.2 mmol aryl halide, 1.0 mmol amine. ^cReaction conditions: aryl halide (0.5 mmol), amine (0.7 mmol), NaOt-Bu (0.7 mmol), 0.75 mol% OA6, THF (0.25 mL), RT, 1 h.