Rhodium-catalyzed oxidative amidation of allylic alcohols and aldehydes: effective conversion of amines and anilines into amides† †Electronic supplementary information (ESI) available. See DOI: 10.1039/c5sc03103f

The rhodium-catalyzed oxidative amidation of allylic alcohols and aldehydes is reported.


Introduction
Amides are a common functionality found throughout natural products, pharmaceuticals, agrochemicals, and organic materials. 1 Approaches have been developed for the coupling of carboxylic acids and amines to generate amides in high yields. 2 While these reactions are versatile and widely employed, there are signicant drawbacks associated with them: they generate super-stoichiometric quantities of high molecular weight byproducts and are not functional group tolerant. 3 The transition metal-catalyzed oxidative amidation of alcohols 4 and aldehydes 5 is a promising alternative to traditional coupling methods, 6,7 as it allows for the generation of amides along with molecular hydrogen as the only by-product (Scheme 1). For example, Misltein and Madsen reported an acceptorless oxidative amidation of alcohols using Ru catalysis with pincer 4c and NHC 4e ligands, respectively. Although no hydrogen acceptor is needed, substrate scopes are limited to unhindered alcohols and amines. Secondary cyclic amines require a less-hindered pincer ligand 8b and acyclic amines are still challenging. 8 Moreover, elevated temperatures (>110 C) and reuxing solvents are needed to favor the evolution of H 2 gas 4c-e,i-l or transfer hydrogenation to ketones. 4f Direct amidation of aryl aldehydes with external oxidants have been reported by Rh 5b and Cu 5c catalysis. However, yields are usually low when enolizable aldehydes are used due to the formation of amine byproducts. 5b Recently, Dong reported a Ni-catalyzed C-H activation of aldehydes followed by coupling with alcohols or amines to form esters and amides respectively. 5i A variety of aldehydes and amines are shown to be reactive, but an additional equivalent aldehyde or triuoroacetophenone is required as the hydrogen acceptor. Herein, we report an easily accessible, chemoselective, and general catalytic method, which selectively couples allylic alcohols or aldehydes with aliphatic and aryl amines to form amides.

Research hypothesis
In seeking to develop general conditions for amide bond formation, we proposed that allylic alcohols could serve as an aldehyde precursor. 9,10 Such a system would be advantageous as allylic alcohols have been previously shown to isomerize to an aldehyde in the presence of a [Rh]-catalyst. 10 Then, the in situ generated aldehydes could undergo a Rh-catalyzed oxidative amidation with an amine to afford the amide and a Rh-H.
Scheme 1 Transition metal-catalyzed oxidative amidation of alcohols and aldehydes. Subsequent H 2 formation or transfer hydrogenation would regenerate the active catalyst species. 4g,5b Optimization studies Initial efforts focused on developing a Rh-catalyzed oxidative amidation of cinnamyl alcohol (1a) and N-methylpiperazine (2a). When they are combined with a cationic rhodium catalyst, the enamine (E)-1-methyl-4-(3-phenylprop-1-en-1-yl)piperazine (4a) predominates, 12 suggesting that isomerization/ condensation is signicantly faster than the subsequent oxidation of the hemiaminal (vide infra). Moreover, reduced starting material 1h was also observed, indicating that the cinnamyl alcohol was acting as the hydrogen acceptor. 12 As such, two main challenges in optimization of the desired oxidative amidation reaction were to identify: (1) conditions that promote amide over enamine formation and (2) an oxidant that is selectively reduced over the allylic alcohol.
As summarized in Table 1, and further elaborated in Tables S1-S8, † a variety of reaction conditions were explored: varying solvent systems, bases, and oxidants. As enamine byproducts do not undergo the oxidative amidation, it was postulated that adding water to the reaction might promote reformation of the hemiaminal; indeed, the addition of an equal volume of H 2 O to non-polar solvents like benzene and toluene signicantly suppressed the formation of the enamine 4a ( Interestingly, as seen in Table 2, under the optimized conditions benzyl amine gave only 11% yield of amide 3h and 89% yield of imine 4h ( Table 2, entry 1). Fortunately, by changing the base and hydrogen acceptor to KOH and acetone, respectively, the desired amide 3h is formed in 82% yield along with only 6% yield of 1h (Table 2, entry 6).

Substrate scope
The optimization experiments suggest that the allylic alcohol rapidly converts into the aldehyde under the reaction conditions, which could then go on to form the amide. This suggested that under the optimized conditions, aldehydes should be effective substrates for the oxidative coupling reaction. Indeed, when 3-phenylpropanal (5a) is subjected to the optimized conditions amide 3a is formed in a 76% yield, which is nearly identical to the 81% yield from the cinnamyl alcohol (Table 3).  The scope of amines was then explored with both 1a and 5a (Table 3). Secondary cyclic amines, including 1-methylpiperazine, morpholine, piperidine, pyrrolidine, tetrahydroisoquinoline (2a-2e) and more sterically hindered acyclic amines, such as dimethyl amine (2f) and N-benzyl methyl amine (2g) are incorporated in very good, and nearly equivalent, yields from either the 1a or 5a. cis-Cinnamyl alcohol also undergoes the amidation reaction with morpholine to afford 3b, although at a signicantly reduced rate and in moderately reduced yields. 12 Likewise, primary amines afforded 3h-3l in good to excellent yields. The reaction is sensitive to steric hindrance of the amine, as cyclohexylamine affords 54%/48% yield of amide 3j while n-butylamine affords amide 3i in 72%/63% yield. Unlike other oxidative amidation processes, 4a,d-l electron rich and electron poor aniline derivatives (2m-2r) were all effective nucleophiles in the coupling reaction without requiring higher temperatures or specialized reaction conditions. 11 It is important to note, that for all substrates, the allylic alcohol and the aldehyde gave similar yields.
The allylic alcohols that undergo the oxidative amidation reaction were explored, as seen in Table 4. Functional groups such as ketones, esters, ethers, aryl bromides, and chlorides are tolerant under the standard conditions based on a chemical robustness screen. 12 Products bearing active aryl bromides could be synthesized in good yields (3ba, 3ha), allowing for facile subsequent coupling reactions to increase molecule complexity. 1,1-Di, 1,2-di-, tri-, and tetra-substituted allylic alcohols give the corresponding aor/and b-branched secondary (6i-6p) and tertiary (6a-6h) amides in moderate to good yields with primary and secondary amines, respectively.
Finally, as seen in Table 5, the scope of aldehydes that undergo the oxidative amidation reaction was investigated. The reactions are tolerant of a variety of functionalities, including ethers (7b), acetals (7c), aryl bromides (7d), aryl uorides (7g), triuromethyls (7i), nitriles (7h), and heteroaromatics (7j). Benzaldehyde derivatives with electron donating groups, such as p-MeO, afford the desire amide in excellent yield; electron poor benzaldehydes, such as p-CN or p-F 3 C, undergo the oxidative amidation to afford 7h and 7i, albeit in reduced yields. Steric hindrance of the aldehyde did not affect its reactivity, as 2,6-dimethylbenzaldehyde afforded amide 7f in 88% yield. Aliphatic aldehydes, which have proven challenging for other oxidative amidation reactions, 5a-h also afford the desired amides in good to very good yields (7k and 6g).

Competition experiments
The synthetic utility of this oxidative amidation reaction would be signicantly increased if the reaction proved to be chemoselective for allylic alcohols and aldehydes over other oxidizable functionalities, i.e., simple primary alcohols. To explore the  In competition experiment 1, cinnamyl alcohol (1a) competes against hexanal (5k) to compare the relative rates of the two coupling partners. Unsurprisingly, given the rapid rate of 1,3-hydride shi, 10c the reaction is unselective, affording a 33% yield of 3a and a 25% yield of 8 aer four hours. This lack of selectivity supports the rapid Rh-catalyzed isomerization of the cinnamyl alcohol to the corresponding aldehyde; the oxidative amidation of the two aliphatic aldehydes then occurs at similar rates.
Next, in competition experiment 2, the comparative reactivity of two different allylic alcohols was explored. Under the standard reaction conditions, cinnamyl alcohol (1a) reacts selectively over cyclohex-1-en-1-ylmethanol (1g). This excellent chemoselectivity is consistent with the known rates of the Rhcatalyzed isomerization of allylic alcohols. 10c Importantly, 1g was observed, unisomerized, at the end of the reaction.
Competition experiments 3-5 investigate the chemo-selectivity of the oxidative amidation reaction with respect to benzylic alcohols, homoallylic alcohols, and aliphatic alcohols. When equimolar amounts of cinnamyl alcohol (1a) and benzyl alcohol were treated with 3.0 equivalents of 2a under the standard reaction conditions, product 3a was formed in 93% yield while 10 was not observed. This indicates that our conditions are highly selective for allylic alcohols over easily oxidizable benzylic alcohols. 4h,m Moreover, in competition experiment 4, less than 5% amide product 11 was observed when cinnamyl alcohol and 3-buten-1-ol are subjected to the reaction conditions, which afforded an 82% yield of 3a. Notably, 3-buten-1-ol affords <5% yield of 11 under the standard reaction conditions, in the absence of cinnamyl alcohol. 12 Finally, competition experiment 5 demonstrates that cinnamyl alcohol (1a) reacts selectively over hexan-1-ol (12), to afford an 88% yield of 3a; 8 was not observed. These experiments exhibit the excellent chemoselectivity of this Rh-catalyzed oxidative amidation reaction for coupling allylic alcohols and aldehydes selectively.

Mechanistic investigations
Next, the requirement of excess amine and super-stoichiometric base was investigated. When a equimolar amounts of amine and allyl alcohol/aldehyde are added to the reaction the desired amide is formed in only 20% yield along with 13% enamine along with both unreacted aldehyde and aldol condensation products (Table S6, † entry 1). However, in the presence of excess amine excellent yields of the amide are observed. It is proposed that the excess amine forces the equilibrium, between aldehyde and enamine/imine, to the enamine/imine thereby decreasing  the concentration of aldehyde and preventing the undesired aldol condensation from occurring. The base additive, CsOAc or KOH, could serve two possible roles, either acting as a proton shuttle or as a ligand on the catalyst. The pH of the water layer in the oxidative amidation reactions is 9.2; when the CsOAc solution is replaced with various buffers (pH ¼ 5.2, 7.5, or 9.0) the reaction are equally efficient as to those in the absence of base (Table S9 †). This suggests that the base is acting as a ligand to the catalyst and that the requirement of excess base is due to the solubility differential of the anion in water versus benzene.
To gain further mechanistic insight into this reaction, we conducted an isotope incorporation experiment, replacing H 2 O with D 2 O. Aer the reaction had gone to completion, 83% deuterium was incorporated at the a-position of the product (Scheme 3). Importantly, in the absence of catalyst and N-methylpiperazine, no deuterium incorporation is observed into amide 3a or 3-phenylpropanal. This suggests that water reacts with the enamine/imine to reform the reactive hemiaminal, 13 which is different from many Ru-catalyzed dehydrogenative coupling of alcohols with amines, where in situ formed aldehydes stay bounded to metal center to avoid the imine formation. 4c-f,i-k Comparison of the initial rate of the reaction in H 2 O and D 2 O revealed a k H /k D ¼ 2.2, indicating the enamine-hemiaminal equilibrium occurs at or before the turnover-limiting step. Alternatively, the primary k.i.e. could be attributed to cleavage of N-H(D) bond in hemiaminal formation, due to the exchange of amine proton with D 2 O.
The proposed dual catalytic cycles for this oxidative amidation reaction are shown in Scheme 4. First, a Rh-mediated 1,3-hydride shi 10 occurs to form an aldehyde from the allylic alcohol. The aldehyde then condenses with the amine to generate the enamine/imine, which is in equilibrium with the hemiaminal. The oxidation of the hemiaminal could occur through either a Rh(I)/Rh(III) or a redox neutral Rh(I) catalytic cycle. In the rst, oxidative addition of the hemiaminal OH into the Rh(I) generates a Rh(III) (H)OCHNR 2 (III). Subsequent b-hydride elimination generates the amide and a Rh(III)-(H) 2 (IV) complex which is reduced to Rh(I) through hydrogenation of the styrene or acetone. Alternatively, in a redox neutral catalytic cycle, ligand exchange of the [Rh] (I) with the hemiaminal affords the Rh-alkoxide (III) then undergoes b-hydride elimination to generate the amide and the Rh-H (IV). Insertion of the hydride into styrene/acetone followed by protolytic cleavage or ligand exchange affords the ethyl benzene/isopropanol, respectively, and either [Rh] complex I or III. Currently, we can not distinguish between the two possible mechanistic pathways and further mechanistic investigations are ongoing. 14

Conclusions
In conclusion, conditions have been developed for the chemoselective oxidative amidation of allylic alcohols or aldehydes, using styrene or acetone as hydrogen acceptors. This methodology presents a general protocol for the synthesis of amides, which is effective for both primary and secondary alkyl/aryl amines. Current efforts are focusing on expanding the scope of nucleophiles and developing asymmetric conditions 15 for the transformation.