Design of New Ligands for the Palladium-Catalyzed Arylation of α-Branched Secondary Amines

Abstract

In Pd-catalyzed C–N cross-coupling reactions, α-branched secondary amines are difficult coupling partners and the desired products are often produced in low yields. To provide a robust method for accessing N-aryl α-branched tertiary amines, new catalysts have been designed to suppress undesired side reactions often encountered when these amine nucleophiles are used. These advances enabled the arylation of a wide array of sterically encumbered amines, highlighting the importance of rational ligand design in facilitating challenging Pd-catalyzed cross-coupling reactions.

Tertiary, N-aryl α-branched amines are frequently found as structural components of pharmaceutically relevant compounds and biologically active natural products (Figure 1).^[1] Although Pd-catalyzed carbon–nitrogen (C–N) cross-coupling would provide an efficient means of accessing this valuable class of compounds, the use of α-branched secondary amine nucleophiles has seen only limited success and in many instances low yields of the desired product are obtained.^[2] Other methods for preparing tertiary N-aryl α-branched amines rely on the addition of an amine to an aryne^[3] or nucleophilic aromatic substitution.^[4] While effective, these methods typically have a narrow substrate scope or result in a mixture of regioisomeric products.^[3] Copper-catalyzed electrophilic amination has also been utilized,^[5] with a recent report by Lalic demonstrating its effectiveness for the arylation of sterically hindered secondary O-benzoyl hydroxylamine electrophiles.^[5b] Despite these advances, there remains no general method for the direct arylation of α-branched secondary amines. Therefore, we sought to develop a catalyst system capable of cross-coupling sterically encumbered secondary amines.

Figure 1.

Selected examples of biologically active compounds containing tertiary N-aryl α-branched amines.^[1]

The development of a highly effective catalyst system for the arylation of α-branched secondary amines must address the specific challenges presented by these coupling partners. Their poor nucleophilicity as a consequence steric hindrance can lead to slower rates of amine transmetalation, resulting in the competitive reaction of the alkoxide base and formation of the corresponding aryl tert-butyl ether (ArOtBu) (V, Figure 2). Additionally, β-hydride elimination may occur from the intermediate Pd(II)-amido complex^[6,7] (IV, Figure 2) leading to the formation of the reduced arene (VI, Figure 2). In this regard, the supporting ligand for the palladium catalyst must be carefully designed in order to facilitate the preferential formation of the desired aryl amine while suppressing side reactions.

Figure 2.

Proposed catalytic cycle and potential challenges presented by sterically hindered α-branched secondary amine nucleophiles.

We began our investigation by examining the effect of the supporting ligands on the efficiency of the catalyst system for the reaction shown in Table 1.^[8] RuPhos(L1)-based catalyst systems have been demonstrated to be highly effective for the cross-coupling of secondary amines,^[9] including some cases of reactions between sterically demanding coupling partners.^[2a,2c] However, when RuPhos precatalyst P1 was used in the reaction of 2-bromo-p-xylene (1a) and 2-ethylpiperidine (1b) only a 10% yield of the desired product was obtained (Table 1, entry 1). Other biaryl phosphine ligands such as XPhos (L2) and BrettPhos (L3) have also been used for promoting Pd-catalyzed C–N bond formation.^[9] Nevertheless, these catalyst systems (P2 and P3, respectively) proved to be inefficient in facilitating the desired transformation (Table 1, entries 2–3). In all cases, the major byproduct was the reduced arene, which presumably arises as a result of β-hydride elimination.^[10]

Table 1.

Supporting Ligand Evaluation.^[a]

Entry	Precatalyst	Conversion	Reduction	Yield
1	P1	100%	68%	10%
2	P2	100%	85%	15%
3	P3	37%	15%	2%
4	P4	100%	53%	27%
5	P5	100%	18%	77%
6	P6	94%	Trace	85%
7	P7	100%	Trace	93%[b],[c]

^[a]Reaction conditions: 1a (0.25 mmol), 1b (0.30 mmol), NaOtBu (0.35 mmol), 2 mol % precatalyst, CPME (0.5 mL), 80 °C, 1 h. Conversion, C–N cross-coupling, and reduction product yields were measured by GC analysis of the crude reaction mixture using dodecane as the internal standard.

^[b]The reaction also produced 6% of the corresponding ArOtBu.

^[c]Isolated yield: 89% (1 mmol scale, average of two runs).

CPME = cyclopentyl methyl ether.

Given these results, we turned to CPhos (L4, Table 1), which has been demonstrated to suppress β-hydride elimination in Pd-catalyzed Negishi cross-coupling reactions.^[11] Indeed, CPhos precatalyst P4 produced aryl amine 1c in improved yield, although the reduced arene remained the major product (Table 1, entry 4).

In the proposed catalytic cycle, the β-hydride elimination pathway competes with reductive elimination from the Pd(II)-amido intermediate (IV, Figure 2). We thus envisoned that using a less electron-rich biaryl phosphine ligand would increase the rate of C–N reductive elimination.^[12] A less electron-rich biaryl phosphine ligand could also increase the rate of transmetalation (amine binding and deprotonation, Figure 2) by rendering the Pd(II) intermediates II and III more electrophilic (Figure 2).^[13] Based on this hypothesis, we examined a catalyst system utilizing the ligand L5 (P5, Table 1).^[14,15] The use of precatalyst P5 dramatically increased the yield of 1c along while decreasing the amount of reduced arene formed (Table 1, entry 5). Following these results, we changed the phosphorus substituents from phenyl to 3,5-bis(trifluoromethyl)phenyl groups to provide ligand L6 (P6, Table 1); this led to additional improvement in the yield and further diminished the formation of the reduced arene (Table 1, entry 6). To achieve additional improvements in catalyst performance, we incorporated methoxy groups in the 3 and 6 positions of the biaryl framework (Table 1) as these groups are known to increase the rate of reductive elimination from Pd(II) complexes.^[16] This modification produced L7 (P7), which provided the most efficient catalyst system for the transformation (Table 1, entry 7).^[17]

Precatalyst P7 was found to enable a wide variety of C–N cross-coupling reactions with α-branched secondary amines (Scheme 1). Hindered cyclic secondary amines were found to be well-tolerated, including in reactions with aryl halides containing ortho-substituents (2a, 2c, 2e, 2g, and 2i, Scheme 1). Lower yields were obtained in the more sterically encumbered cases,^[18] where formation of the reduced arene byproduct was observed. Acyclic α-branched amines could also be efficiently arylated (2b and 2h, Scheme 1). Previously, the arylation of diisopropylamine via Pd-catalyzed C–N cross-coupling has resulted in very low yields,^[2f,20] presumably due to its steric hindrance. By using P7, however, diisopropylamine was successfully arylated in 65% yield (2h, Scheme 1), although additional equivalents of amine and base were necessary to favor formation of the desired product.^[21,22]

Scheme 1.

Scope of C–N cross-coupling reactions using P7. Reaction conditions: aryl halide (1.0 mmol), amine (1.2 mmol), NaOtBu (1.4 mmol), 2 mol % P7, 0–2 mol % L7, CPME (2 mL), 60–80 °C, 6–16 h. Yields are of isolated products, average of two runs. [a] 1:49 cis:trans isomers of the arylated amine. Determined by GC analysis of the crude reaction mixture. 2% reduction, 4% ArOtBu. [b] 9% ArOtBu. [c] 27% reduction, 6% ArOtBu [d] 22:1 cis:trans isomers of the arylated amine. Determined by GC analysis of the crude reaction mixture. [e] 28% reduction. [f] K₃PO₄ (6.0 mmol) used as base. [g] 34% reduction. [h] Amine (9.6 mmol), NaOtBu (10.8 mmol), 7% reduction, 9% ArOtBu. [i] 37% reduction.

We were interested in applying the developed conditions to the amination of heteroaryl halides due to their presence in many pharmaceutically relevant compounds.^[2] However, our initial attempts to utilize activated heteroaryl electrophiles (3a, 3b, and 3c, Scheme 2) resulted in low yields and the formation of significant amounts of the corresponding ArOtBu.^[23,24] Through systematic ligand modification^[25] we found that ligand L8 (P8, Scheme 2) provided higher yields in these cases. With all other substrates, P7 was again very effective in producing high yields of the desired product. In certain instances, the use of additional equivalents of the amine was necessary to further deter the formation of the ArOtBu (3a, 3g, and 3i, Scheme 2). Additionally, a trace of the epimerized product was observed in cases where cis-2,6-dimethylpiperidine (3g, Scheme 2) or an enantiomerically enriched amine was used (3h and 3i, Scheme 2). Despite these considerations, the combined substrate scope using precatalysts P7 and P8 allows for efficient cross-coupling of a wide variety of challenging α-branched secondary amines with different heteroaryl halides (Scheme 2).

Scheme 2.

The scope of C–N cross-coupling reactions with heteroaryl halides and hindered secondary amines. Reaction conditions: aryl halide (1.0 mmol), amine (1.2 mmol), NaOtBu (1.4 mmol), 2–3 mol % P7 or P8, 0–2 mol % L7 (used only with P7), CPME (2 mL), 60–80 °C, 16 h. Yields are of isolated products, average of two runs. [a] Amine (2.4 mmol), NaOtBu (2.8 mmol). [b] 9% reduction, 8% ArOtBu. [c] 2% reduction, 3% ArOtBu. [d] 13% reduction. [e] Amine (3.6 mmol), NaOtBu (4.2 mmol); 20:1 cis:trans isomers of the arylated amine product. Determined by GC analysis of the crude reaction mixture. [f] Starting amine ee: 99% ee; Product ee: 98% ee. [g] Amine (2.4 mmol), NaOtBu (2.8 mmol), dioxane (2 mL); 24% ArOtBu, 6% reduction; Starting amine ee: ≥97% ee; Product ee: 83% ee.

In summary, we have developed two new catalyst systems for the arylation of sterically demanding α-branched secondary amines. Notably, the unprecedented levels of reactivity in C–N cross-coupling reactions with these amines are achieved due to the ability of the new precatalysts to suppress both the β-hydride elimination pathway and arylation of the alkoxide base. Overall, this work highlights the potential of rational ligand design to modulate catalyst behavior and ultimately facilitate the cross-coupling of sterically demanding amine coupling partners.