Cu-Catalyzed C–N Coupling with Sterically Hindered Partners

Abstract

Copper, an earth-abundant metal, has reemerged as a viable alternative to the versatile Pd-catalyzed C–N coupling. Coupling sterically hindered reaction partners, however, remains challenging. Herein, we disclose the discovery and development of a pyrrole-ol ligand to facilitate the coupling of ortho-substituted aryl iodides with sterically hindered amines. The ligand was discovered through a library screening approach and highlights the value of mining heteroatom-rich pharmaceutical libraries for useful ligand motifs. Further evaluation revealed that this ligand is uniquely effective in these challenging transformations. The reaction enables the coupling of sterically hindered primary and secondary amines, anilines, and amides with broad functional group tolerance.

Graphical Abstract

Anilines represent a privileged class of amines in nature, as well as pharmaceuticals, agrochemicals, and other valuable organic materials.¹ Consequently, transition metal-catalyzed C–N-bond formation has become critically important to organic synthesis. While a number of d-block metals have been investigated for this reaction, the two most common metals remain palladium (Buchwald–Hartwig amination²) and copper (Ullmann–Goldberg-type³ and Chan−Lam⁴ amination). Exploration of palladium-catalyzed C–N-coupling reactions started almost 25 years ago concurrently by Buchwald and Hartwig.⁵ Concentrated effort across a number of groups has positioned the Buchwald–Hartwig reaction as the dominant amination method with broad substrate scope and high selectivities.⁶ While much progress has been made with the copper-catalyzed variants, the Ullmann-type amination has lagged in terms of both electrophile and amine scope.⁷ Early advances came independently from Ma,⁸ Goodbrand,⁹ Buchwald,¹⁰ and others.^11,12 Recently, the Ma group introduced oxalamide ligands that enable low catalyst loading and the use of aryl chlorides as electrophiles—addressing key challenges for copper-catalyzed amination.¹³ Copper-catalyzed C–N-coupling reactions with hindered partners, however, remain an unsolved problem (Figure 1a).^12b Currently, there exist only a few examples where an ortho methyl group can be tolerated on the electrophile (Figure 1a).^13,14 Furthermore, examples with ortho,ortho′-disubstituted electrophiles with sterically hindered amines have not been reported. Although alternative base metal-catalyzed approaches exist to synthesize sterically encumbered anilines (Figure 1b),¹⁵ they rely on fundamentally different building blocks. Here, we describe a novel ligand that enables very challenging and previously unreported direct aminations with sterically hindered coupling partners using copper catalysis (Figure 1c).

Figure 1.

(a) Existing copper-catalyzed C–N-coupling reactions remain limited in sterically demanding environments. (b) Previous approaches for base metal-catalyzed aniline synthesis. (c) A novel pyrrole-ol ligand enables C–N coupling of sterically hindered partners.

We were interested in identifying a ligand for Cu to couple sterically hindered electrophiles and nucleophiles—a challenge even for Pd-based methods.¹⁶ Despite several new ligands for Cu reported in the recent years, coupling of sterically hindered partners remains a formidable challenge. The discovery of a more effective ligand is critical to the success of this transformation. Given the recent success with evaluating pharmaceutical compound libraries as ligands for metal-catalyzed reactions,¹⁷ AbbVie’s internal compound library was screened as ligands for copper. The coupling of bromobenzene with 1-methyl-3-phenylpiperazine sought to evaluate 54 nonproprietary compounds that were previously examined for Cu-catalyzed C–O coupling (Scheme 1).^17b Interestingly, 2,2,2-trifluoro-1-phenyl-1-(1H-pyrrol-2-yl)ethan-1-ol (L1), a compound that had not been previously applied as a ligand, produced the highest yield of the coupled product (74 HPLC peak area%).^14d,18 Surprisingly, the more sterically hindered substrate combination, 2-bromotoluene and 1-methyl-3-phenylpiperazine, coupled to form product only in the presence of L1.

Scheme 1.

Discovery of Trifluoromethylated Pyrole-ol-Based Ligand

^aArea% of the coupled products determined by HPLC is reported.

Ligand L1 was chosen for further optimization (Table 1). Formation of an almost equal amount of the C–O coupled product 1a was observed as a side-product in the reaction (entry 1). The addition of molecular sieves to the reaction mixture suppressed the C–O coupling product but failed to improve the yield of the desired product (entry 2). We speculated that base-mediated ligand degradation led to catalyst deactivation through the formation of phenyl(1H-pyrrol-2-yl)methanone L1′.¹⁹ Further evaluation of both the target reaction and ligand stability demonstrated rapid, base-mediated decomposition to phenyl(1H-pyrrol-2-yl)methanone L1′, and it functions as an inhibitor for the reaction. Although lowering the reaction temperature suppressed ligand degradation, it also slowed the reaction. In an attempt to increase the lifetime of active catalyst and to block ligand degradation, a variety of mild reductants were evaluated. Interestingly, a stochiometric amount of Hantzsch ester (HE) increased the catalyst lifetime.²⁰ Conversely, ascorbic acid offered no improvement.²¹ In the presence of one equivalent of Hantzsch ester, the desired product was obtained in 33% yield (entry 3). Next, the more reactive electrophile, 2-iodotoluene, was tested in the reaction. The desired product was obtained in a higher 42% yield along with 50% of 1a (entry 4). As observed earlier, addition of molecular sieves suppressed the ether formation (entry 5). The higher yield of 1 (64%) was observed by conducting the reaction in the presence of 0.5 equiv of HE at 80 °C (entry 6). Fortunately, employing additional base (4 equiv) and amine (1.7 equiv) produced the desired product in 86% yield after 24 h (entry 7).

Table 1.

Initial Reaction Optimization

entry	X	additive	1 (%)a	1a (%)a	conversion (%)b
1	Br	-	17	23	40
2	Br	100 mg MS (4 Å)	18	1	20
3	Br	100 mg MS (4 Å), 0.5 equiv HE	32	2	35
4c	I	-	42	50	95
5c	I	100 mg MS (4 Å)	46	8	59
6c	I	100 mg MS (4 Å), 0.5 equiv HE	64	5	70
7	I	24 h, 100 mg MS (4 Å), 1 equiv HE 1.7 equiv amine, 4 equiv K3PO3 90 °C	86	7	98

^aGC yield.

^bBased on unreacted aryl halide.

^cReaction was conducted at 80 °C.

MS = molecular sieves.

As discussed above, this type of sterically demanding coupling is unprecedented with a Cu catalyst. To better understand the catalyst landscape, we also screened 46 established ligands for Cu in this reaction using a high-throughput approach. To our surprise, none of the established ligands resulted in the formation of 1 in any appreciable yield (Table S1).²² These results suggest that the catalyst derived from L1 and CuI may be unique in its reactivity for sterically demanding C–N couplings.

Next, we evaluated the structure–activity relationships for the pyrrole-ol-based N–O type ligands. Systematically replacing the phenyl ring of L1 resulted in significant variations in yield (L2–L8, Table 2). Pentafluorobenzene-containing ligands (L3, L18, L20), for example, showed no reactivity possibly because of decreased ligand stability under the reaction conditions, ejecting C₆F₅⁻ as a leaving group. The pyrrole-ol ligand bearing a benzyl substituent (L7), which is presumably less prone to elimination, provided a minor improvement in yield. The unprotected pyrrole moiety proved critical to afford a ligand which promote the overall reaction in reasonable yields (L9–L13, Table 2). Finally, evaluation of the R³- substituent (L14–L18, Table 2) revealed that the trifluoromethyl group is crucial for the ligand reactivity—presumably due to its steric and electronic properties. On the basis of these results, L7 was chosen for further reaction optimization, providing 90% of the desired product under the optimal conditions (described in Table 1, entry 7).

Table 2.

Screening of Pyrole-ol-Based Ligand^a

^aReaction conditions: iodide 0.1 mmol, amine 1.7 equiv, CuI 10 mol %, ligand 10 mol %, 100 mg 4 Å MS, HE 1 equiv, K₃PO₄ 4 equiv,DMSO 0.2 M, 90 °C, 24 h.

HE = Hantzsch ester

A diverse range of sterically crowded amines with varied N–H pK_a values were explored for the coupling with ortho-substituted aryl iodide. Satisfyingly, the protocol was successful for amines with an N–H pK_a window of ~20 to ~ 40.²³ Next, we explored the substrate scope with regard to various ortho-substituted aryl iodides and 1-methyl-3-phenylpiperazine. Not surprisingly, larger ortho substituents led to lower yields (1–3, Scheme 2). Next, we tested the scope of ortho substituted aryl iodides with the bioactive and sterically demanding 1-phenyl-1,2,3,4-tetrahydroisoquinoline (4–13).²⁴ The catalyst demonstrated excellent reactivity with electron-neutral (4–6) or electron-donating (7) substituents at the ortho position of the aryl iodide. To our delight, a bulky N,N-di-n-butyl-amino group was well tolerated in the reaction (8). Importantly, 2-bromo-iodobenzene promoted the C–N coupling preferentially at the iodide with >20:1 chemoselectivity (9). Surprisingly, aryl iodides with electron-deficient substituents at the ortho position resulted in low-to-moderate yield (10–12) with significant halide-reduction product. Control reactions revealed that this unwanted side product was not due to hydride donation from the Hantzsch ester (10). Excitingly, hindered 2,6-dimethyl-iodobenzene reacted successfully to form 13, a unprecedented reaction for copper catalysis. The ligand even enabled sterically demanding primary amines to couple with ortho-substituted iodides (16–21). As expected, 2,6-dimethyl-substituted aryl iodides provided satisfactory yield of the desired products (18–20, 26). Additionally, this protocol offered excellent reactivity with 2,6-dimethylaniline (23−26, 28). The reactivity with 2,6-diisopropylaniline further established the versatility of this catalyst system (27). Furthermore, this method is also applicable to cyclic amides (29–30), anilide (31), and an amino acid derivative (22). As an investigation into medicinally relevant heterocycles, a series of heteroaryl iodides and amines were tested. Moreover, pyridine, thiophane, and pyrazole moieties were well tolerated by the catalyst (32–37).

Scheme 2.

Substrate Scope for C−N Coupling Reaction

^a10 mol % Cu/L7; ^b2 mol % Cu/L7; ^c5 mol % Cu/L7; ^d8 mol % Cu/L7; ^ewithout HE; ^f48 h; ^g60 h. HE = Hantzsch ester

The origin for the effectiveness of pyrrole-ol-based ligands for enabling the Cu-catalyzed coupling of sterically hindered coupling partners remains worthy of further mechanistic consideration. We expect L7 to be deprotonated under the reaction conditions and act as either mono- or bis-anionic ligand for Cu.²⁵ Likely, L7 forms a highly electron-rich Cu(I) or anionic Cu(I)-complex that promotes oxidative addition even with sterically hindered electrophiles.²⁶

In summary, we have disclosed a new ligand that enables an unprecedented Cu-catalyzed method for coupling ortho- and ortho,ortho′-substituted aryl iodides with sterically hindered primary aliphatic amines, anilines, and amides. The unique reactivity and novel pyrrole-ol structural motif of ligand L7 was discovered by mining AbbVie’s internal compound library—demonstrating the value of screening pharmaceutical compound libraries for ligand discovery in challenging metal-catalyzed reactions. Complementary to the rational ligand design and optimization, this approach has the potential to identify previously unexplored and nonobvious compounds as ligands. Our future studies will focus on detailed mechanistic studies to elucidate the role of pyrrole-ol ligands in Cu-catalyzed reactions.