Ligand-Promoted Non-Directed C–H Cyanation of Arenes

Abstract

We herein report the first example of a 2-pyridone accelerated non-directed C–H cyanation with an arene as the limiting reagent. This protocol is compatible with a broad scope of arenes, including advanced intermediates, drug molecules, and natural products. A kinetic isotope experiment (k_H/k_D = 4.40) indicates that the C–H bond cleavage is the rate-limiting step. Also, the reaction is readily scalable, further showcasing the synthetic utility of this method.

Graphical Abstract

Cyanoarenes are abundant in natural products, drug molecules, and advanced synthetic intermediates.¹ The cyano functional group is highly versatile in organic synthesis and can be readily converted to carboxyl, carbonyl, amino, and other heterocyclic groups.² Thus, installation of cyano groups in a direct and facile manner is of high interest. However, introduction of cyano groups onto arenes have traditionally relied heavily on the Sandmeyer reaction and other transition-metal-mediated cross-coupling reactions with aryl halides.³ The use of highly reactive diazonium intermediate and the requirement to prefunctionalize arenes render these traditional methods undesirable and inefficient, often with limited scope and functional group tolerance.

On the other hand, direct C–H cyanation of arenes offers a potentially superior alternative. In the past decade, tremendous efforts have been devoted to the development of directed C–H cyanation of arenes. Pyridines, pyrimidines, and oximes have been employed as directing groups for ortho C–H cyanation with Pd, Ru, Rh, and Co catalysts (Fig. 1a).⁴ A meta-selective C–H cyanation has been realized using a U-shaped template.⁵ More recently, two examples of non-directed C–H cyanation of arenes through radical pathways have also been achieved (Fig. 1b).⁶ In a continuing effort to improve the reactivity and selectivity of non-directed C–H functionalizations of arenes, we have successively discovered 2,6-dialkyl pyridine ligands⁷ that can accelerate non-directed C–H olefination of electron-deficient arenes and electron-deficient 2-pyridone ligands⁸ that can enable C–H olefination with arenes as the limiting reagent (Fig. 1c).

Figure 1.

(a) Previous work about transition-metal-catalyzed directed C(sp²)–H cyanation. (b) Non-directed cyanation of arenes via radical mechanism (c) Ligand-accelerated non-directed C(sp²)–H olefination of arenes. (d) Ligand-accelerated non-directed C(sp²)–H cyanation of arenes.

Guided by these ligand developments, we wondered whether these ligands could be used to enable C–H cyanation. We began our study by treating the model substrate mesitylene (1.0 equiv.) with Pd(OAc)₂ (10 mol%), AgCN (1.1 equiv., cyanide source), and AgOAc (3.0 equiv.) in HFIP (solvent) at 100 °C. Encouragingly, the desired C(sp²)–H cyanation product was detected in 18% yield. We first evaluated several mono-dentate pyridine-based ligands (L1-L3), which have been shown to promote C–H activation.⁹ However, only trace amounts of product were observed. Based on our previous studies on non-directed C(sp²)–H olefination promoted by electron-deficient 2-pyridone ligands, we set out to extensively screen 2-pyridone ligands for this non-directed C(sp²)–H cyanation reaction. The addition of a simple 2-pyridone ligand suppressed the reactivity (L4). To our delight, the use of 5-nitro-2-pyridone, which has previously demonstrated efficiency for γ-C(sp³)–H arylation,¹⁰ slightly increased the yield to 25% (L5). Incorporation of a trifluoromethyl or a nitro group at the 3 position of L5 led to lower yields (L6, L7). A significant drop in reactivity was observed with 3-nitro-2-pyridone (L8). Intriguingly, the yield increased to 77% when a trifluoromethyl group was installed at the 5-position of L8 (L9). The replacement of 5-CF₃ in L9 with 5-Cl further improved the yield to 87% (L10). However, the use of 3,5-dichloro-2-pyridone resulted in loss of reactivity (L11). These results suggest that this cyanation reaction is exceptionally sensitive to the steric and electronic environments of the 2-pyridone ligands. Other structural variations of 2-pyridone ligands resulted in low reactivity (L12-L15) (Table 1).

Table1.

Ligand Evaluation^a,b

^[a]Conditions: mesitylene (0.2 mmol, 1.0 equiv), Pd(OAc)₂ (10 mol%), ligand (20 mol%), AgOAc (3.0 equiv), AgCN (1.1 equiv.), HFIP (0.5 mL), 100 °C, under air, 24 h. See SI for work up procedures.

^[b]The yield was determined by ¹H NMR analysis of the crude product using CH₂Br₂ as the internal standard.

With the optimal ligand L10 identified, we next examineed different cyanide sources and their loadings (Table 2). We first tested several commercially available cyanide salts and found that increasing the loading of AgCN to 2.0 equivalents from 1.1 equivalents decreased yield to 50% from 84%, presumably due to palladium catalyst deactivation by excess cyanide coordination.¹¹ Screening of two other cyanide sources showed that CuCN and Zn(CN)₂ gave lower yields. Other reaction parameters have also been examined, such as oxidants, solvents, and temperature. Further, minimal products were afforded with KCN, K₃Fe(CN)₆, and TMSCN as the cyanation reagents. Oxidants other than AgOAc only provided trace products, while other variations of the reaction parameters resulted in inferior reactivity.

Table 2.

Optimization of Reaction Conditions^a,b,c

Entry	Variations from the ‘Standard Conditions’	Total Yield (%)
1	none	84
2	2.0 equiv. AgCN	50
3	1.1 equiv. CuCN	67
4	2.0 equiv. CuCN	39
5	0.6 equiv. Zn(CN)2	37
6	1.0 equiv. Zn(CN)2	32
7	1.1 equiv. KCN	trace
8	0.2 equiv. K3Fe(CN)6	18
9	0.5 equiv. K33Fe(CN)6	6
10	1.1 equiv. TMSCN	trace
11	no AgOAc	trace
12	1.0 equiv. AgOAc	45
13	2.0 equiv. AgOAc	50
14	2.0 equiv. Cu(OAc)2	trace
15	2.0 equiv. CuBr2	trace
16	CHCl3 as solvent instead of HFIP	trace
17	CHCl3/HFIP = 1/1	30
18	45 mol% ligand	68
19	80 °C	60
20	120 °C	72

^[a]Conditions: o-xylene (0.2 mmol, 1.0 equiv), Pd(OAc)₂ (10 mol%), ligand (20 mol%), AgOAc (3.0 equiv), AgCN (1.1 equiv.), HFIP (0.5 mL), 100 °C, under air, 24 h. See SI for work up procedures.

^[b]The yield was determined by ¹H NMR analysis of the crude product using CH₂Br₂ as the internal standard.

^[c]The product was isolated as a mixture of two isomers: 3-cyano and 4-cyano o-xylene.

With the optimal reaction conditions established, we next explored the scope of arenes for this non-directed cyanation reaction. Extensive screening of 2-pyridone ligands proved that L9 gave similar even higher yields than L10 in some cases, thus both of the two ligands were used to explore the substrate scope of this reaction. Electron neutral to rich arenes such as simple benzene, alkyl substituted benzenes, and alkoxy benzenes consistently afforded the mono-cyanated products in good yields (2a–2q, 2t–2u). The mono selectivity could be explained by the electron-deficient nature of cyanated arenes. Mono-alkylated arenes were cyanated at the less sterically hindered meta and para positions, providing meta- and para-isomers as the main products (2b–2d). Arenes with bulkier substituents such as tert-butyl group (2d) afforded considerably better para site selectivity. Di-alkylated and tri-alkylated arenes all provided the mono-cyanated products in good yields with various selectivities depending on the substitution patterns of the alkylated arenes (2e–2h). Strong electron-donating substituents such as alkoxy groups, only provide ortho- and para-cyanated isomers, owing to its electronic effect. The bulkier silyl-protected phenol gave para-cyanated product as the main isomer because of steric effect (2p and 2q). Naphthalene also provided the cyanated pruducts in 73% yield, with a 2.0/1.0 selectivity ratio (α/β) (2l). This non-directed cyanation occurred selectively on the relative electron-rich arene for diaryl substrates (2m). For 2-chloroanisole and 3-chloroanisole (2n and 2o), the yields dropped to only 20% under the same conditions due to the electron-withdrawing effect of the chloride on the benzene ring. However, heating up the reaction to 120 °C improved yields to 64% and 50% respectively. Electron-deficient arenes such as chlorobenzene and methyl benzoate, are unreactive to this reaction condition, thus 8.0 equiv. of arenes and 1.0 equiv. of cyanide were used to secure 45% and 50% yields respectively (2s and 2t).

To further examine the practicality of this reaction, we also examined cyanation on advanced intermediates and biologically relevant substrates (2v–2x, 2ab–2ac). Gladly, O, S, N- heterocycles could be cyanated in good yields. Interestingly, for both dibenzofuran and dibenzothiophene, cyanation proceeded sleectively at the 4 position, while olefination occurred at the 2 position in our previous studies for non-directed olefination.^ref This suggests that electronic effect is more important to selectivity compared to steric effect in this cyanation reaction. Some natural products and drug molecules can also be cyanated under the optimal conditions in moderate to good yields (2z, 2aa, 2ad). 2-Alkyl/aryl anisoles were observed to give better selectivity (2ae–2aj), which could be rationalized by a combination of electronic effect and steric effect (Table 3).

Table 3.

Substrate Scope^a,b

^[a]Conditions: 1 (0.2 mmol, 1.0 equiv), Pd(OAc)₂ (10 mol%), L9 (20 mol%), AgOAc (3.0 equiv), AgCN (1.1 equiv.), HFIP (0.5 mL), 100 °C, under air, 24 h. See SI for work up procedures.

^[b]The product was isolated as mixture of regio-isomers.

^[c]3.0 equiv. of benzene were used.

^[d]8.0 equiv. of substrates were used.

^[e]The reaction was conducted at 120 °C.

^[f]15 mol% Pd(OAc)₂ were used instead of 10 mol%.

^[g]20 mol% Pd(OAc)₂ were used instead of 10 mol%.

^[h]L10 was used instead of L9.

^[i]The reaction was conducted at 90 °C for 12 h.

To further showcase the synthetic utility of this method, non-directed cyanation of mesitylene (1h) was carried out on 10.0 mmol scale, yielding the desired product in 62% yield (Fig. 2a).

Figure 2.

(a) Scale-up reaction. (b) Primary kinetic isotope effect.

To gain insight into the reaction mechanism, we carried out kinetic isotope effect (KIE) experiments. Intermolecular one-pot and parallel experiments provided a P_H/P_D value of 4.29 and a k_H/k_D value of 4.40, respectively, which suggest that the C–H activation step is rate-limiting step (Fig. 2b).

A plausible catalytic cycle is outlined in Figure 4. 2-pyridone ligand functions as an analogue of OAc and accelerates C–H cleavage to form intermediate II. Ligand Exchange of CN⁻ between AgCN and Pd complex intermediate II forms III. Then reductive elimination from III provides the desired product (Fig. 3).

Figure 3.

Proposed mechanism

In summary, we have developed the first example of a Pd-catalyzed non-directed cyanation of arenes, enabled by an electron-deficient 2-pyridone ligands. The reaction features broad substrate scope and high functional group compatibility, exemplified by successful C–H cyanation of a range of complex molecules. Currently, we are attempting to achieve better site selectivity for non-directed C–H activation through ligand design.