Nickel-Catalyzed Asymmetric Reductive Cross-Coupling to Access 1,1-Diarylalkanes

Abstract

An asymmetric Ni-catalyzed reductive cross-coupling of (hetero)aryl iodides and benzylic chlorides has been developed to prepare enantioenriched 1,1-diarylalkanes. As part of these studies, a new chiral bioxazoline ligand, 4-heptyl-BiOX (L1), was developed in order to obtain products in synthetically useful yield and enantioselectivity. The reaction tolerates a variety of heterocyclic coupling partners, including pyridines, pyrimidines, indoles, and piperidines.

Graphical abstract

Ni-catalyzed cross-coupling reactions have emerged as powerful methods to forge C(sp²)–C(sp³) bonds.¹ Nicatalyzed reductive cross-coupling reactions are one subset of these transformations, which couple two organic electrophiles and use a stoichiometric reductant to turn over the Ni catalyst.^2,3,4 Whereas an array of conventional Ni-catalyzed cross-coupling reactions have been rendered highly enantioselective (e.g. Suzuki,⁵ Negishi,⁶ Kumada⁷ reactions),⁸ less progress has been made in the development of asymmetric reductive cross-coupling reactions.⁹ Given that many of these reactions work well for secondary alkyl substrates and provide chiral products, it would be of value to develop enantioselective variants. In this communication, we report an enantioselective Ni-catalyzed reductive cross-coupling between aryl iodides and secondary benzylic chlorides (Figure 1). This success of this effort hinged on the development of a new chiral bioxazoline (BiOX) ligand, 4-heptyl-BiOX (L1), which provides 1,1-diarylalkanes with both improved yield and enantioselectivity relative to previously disclosed BiOX ligands.

Figure 1.

Ni-catalyzed enantioselective reductive cross-coupling to prepare diarylalkanes.

A number of commercial pharmaceuticals possess stereogenic 1,1-diarylalkane motifs,¹⁰ and as a result, significant effort has been devoted to the enantioselective synthesis of this substructure. As a complementary approach to methods such as asymmetric hydrogenation¹¹ and conjugate addition,¹² Ni-catalyzed stereospecific¹³ and stereoconvergent¹⁴ cross-coupling reactions have been developed.^15,16 For example, in 2013, Fu and coworkers reported the enantioselective Negishi coupling of benzylic mesylates with aryl zinc halides to furnish 1,1-diarylalkane products in good yields and high enantioselectivity.^14a However, a limited scope of heteroaryl substrates was demonstrated.

Based on our previously disclosed research,^9c we hypothesized that Ni-catalyzed reductive cross-coupling could provide improved access to heterocycle containing products. However, a challenge in the development of such enantioselective reactions is that as one changes electrophile class (e.g. from vinyl halides to aryl halides), or as one alters the ligand, the product yield and ee can decrease dramatically. Indeed, efforts to prepare enantioenriched diarylalkanes via asymmetric reductive coupling^14b or Ni/Ir synergistic catalysis^14c have proved challenging. In 2015, Weix reported a reductive cross-electrophile coupling between primary mesylates and aryl bromides; this report contained a single enantioselective coupling of (1-chloroethyl)benzene with 4-bromoacetophenone, which proceeded in both modest yield and ee.^14b Similarly, Molander reported that coupling of (1-phenylethyl) potassium trifluoroborate with 4-t-butylbromobenzene proceeds in 65% ee, and coupling of more electron deficient arenes occurs with lower enantioselectivity.^14c

Consistent with the challenges encountered by others, submission of a mixture of (1-chloropropyl)benzene (1) and 5-iodo-2-methoxypyridine (2a) to the optimal conditions identified for the Ni-catalyzed asymmetric reductive cross-coupling of benzylic chlorides with vinyl bromides provided 3a in only 12% yield and 10% ee (Scheme 1a).¹⁷ Similarly, use of the conditions developed for the reductive cross-coupling of heteroaryl iodides with α-chloronitriles failed to deliver detectable amounts of 3a (Scheme 1b).

Scheme 1.

Application of previously developed cross-coupling conditions to the coupling between 1 and 2a.

Despite these discouraging results, we initiated studies focusing on the cross-coupling between 1 and 2a by screening a variety of chiral bidentate ligands under both sets of conditions shown in Scheme 1. From this study, it was determined that performing the reaction in 1,4-dioxane with ⁱPr-BiOX (L2) as the ligand and TMSCl as an activator produced 3a in 22% yield and 68% ee (Table 1, entry 2). We found that both the yield and the enantioselectivity of the reaction could be improved by increasing the length of the BiOX alkyl chain, with 4-heptyl-BiOX (L1), delivering 3a in 84% yield and 90% ee (entry 1). We note that Bn-BiOX (L5, entry 5) and serine-derived ligand L6 (entry 6), the ligands used by used by Weix/Molander and Fu, respectively, perform poorly under the optimal conditions. Control experiments confirmed that no reaction takes place in the absence of Ni, ligand, Mn⁰, or TMSCl. Zn⁰ and TDAE performed poorly as reductants (entries 7-8).¹⁸ No product was detected when TFA was used as an activator in place of TMSCl (entry 9)¹⁹ and when DMA is used as solvent, the yield and ee both drop substantially (entry 10). Use of aryl bromide 5 instead of 2a delivered 3a in only slightly reduced yield and comparable ee (entry 11), while employing benzylic bromide 6 in place of 1 increased formation of bibenzyl homocoupling product 4 at the expense of 3a, and the ee of 3a was slightly lower (entry 12).

Table 1.

Optimization of reaction conditions.^a

Entrya	Ligand	Deviation from Standard Conditions	Yield 3a (%)b	Yield 4 (%)b	ee 3a (%)c
1	L1	None	84	8	90
2	L2	--	22	30	68
3	L3	--	64	21	75
4	L4	--	74	20	80
5	L5	--	4	0	60
6	L6	--	31	9	86
7	L1	Zn0 instead of Mn0	0	26	--
8	L1	TDAE instead of Mn0	3	0	66
9	L1	TFA instead of TMSCl	0	25	--
10	L1	DMA instead of dioxane	14	13	67
11	L1	5 instead of 2a	72	24	89
12	L1	6 instead of 1	8	37	81

^aReactions conducted under N₂ on 0.05 mmol scale for 18 h.

^bDetermined by ¹H NMR versus an internal standard.

^cDetermined by SFC using chiral stationary phase.

With optimized reaction conditions in hand, we evaluated the substrate scope of the aryl iodide coupling partner (Table 2). Pyridyl iodides bearing substitution at the 2-position couple smoothly (3a–3d, 3f), as do pyrimidines (3e, 3g) and indoles (3h). Non-heteroaryl iodides bearing either electronrich (3j, 3k) or electron-poor (3i, 3m–o) functional groups couple smoothly, although slightly lower ee is observed with more electron-rich arenes. It is notable that acidic protons are tolerated (3k); no protodehalogenated byproducts observed. The cross-coupling is orthogonal to aryl triflates and boronates, affording 3o and 3l in excellent yields and providing handles for further derivatization. When the reaction was conducted on 2.0 mmol scale, pyridine 3a was produced in 63% yield and 91% ee.

Table 2.

Scope of (hetero)aryl iodides.^a

^aReactions conducted on 0.2 mmol scale. Isolated yields; ee is determined by SFC using chiral stationary phase.

^b2.4 equiv of 1 is used.

Next, we turned our attention to the scope of the benzylic chloride coupling partners (Table 3). Substrates with either electron-donating or -withdrawing substituents at the para position couple in comparable yields and enantioselectivity (8a–8d). In addition, o-substitution with either methoxy (8e) or fluorine (8f) is tolerated, although the products are formed in decreased yields. Substituents of varying steric encumbrance can be incorporated at the α-position of the benzylic halide (8g–8m). Of particular interest, good chemoselectivity is observed for coupling of the benzylic chloride in preference to the primary chloride (8h). N-Bocpiperidine (8l) and dibenzofuran (8m) groups are also tolerated, providing the products in serviceable yields and excellent ee.

Table 3.

Scope of benzylic chlorides.^a

^aReactions conducted on 0.2 mmol scale. Isolated yields; ee is determined by SFC using chiral stationary phase.

To demonstrate the synthetic utility of our method, we synthesized diarylalkane 11, an intermediate in the synthesis of the commercial anti-depressant sertraline (Scheme 2a).²⁰ Cross-coupling of 1-chloro-1,2,3,4-tetrahydronaphthalene (9) with commercially available iodobenzene 10 provides chiral tetrahydronaphthalene (11) in 70% yield, and 84% ee. Benzylic oxidation of 11 using 3 equiv of CrO₃ in AcOH/H₂O²¹ afforded tetralone 12 in 51% yield (unoptimized) with no erosion of ee.²² Tetralone 12 is a known intermediate in the synthesis of sertraline.²³

Scheme 2.

Synthetic applications and mechanistic experiments.

To probe for the intermediacy of radical species, the reaction was conducted in the presence of TEMPO.²⁴ No significant decrease in yield was observed, and no TEMPO trapping adducts were detected. When cyclopropyl chloride 14 was subjected to the standard cross-coupling conditions, alkene 15 was obtained in 57% yield (Scheme 2b).^25,26 These findings are consistent with a mechanism that proceeds through a non-persistent alkyl radical, however we cannot rule out the possibility of a Ni-mediated cyclopropane opening pathway. Further studies of the mechanism are ongoing; it is unclear at this time whether the absolute stereochemistry is set during the oxidative addition or reductive elimination steps.^14c

In conclusion, we have developed a Ni-catalyzed asymmetric reductive cross-coupling between (hetero)aryl iodides and benzylic chlorides. This transformation enables the synthesis of enantioenriched 1,1-diarylalkanes from simple organic halide starting materials. These efforts resulted in the discovery of a new chiral BiOX ligand, 4-heptyl-BiOX, which we expect will find application in other transition metalcatalyzed cross-coupling processes.²⁷