|Home | About | Journals | Submit | Contact Us | Français|
A cationic (CAAC)gold(I) complex promotes the addition of all types of non-tertiary amines to a variety of allenes, affording allylic amines in good to excellent yields; the amino fragment always adds to the less substituted terminus of the CCC skeleton.
The most widely used chemical reaction for the functionalization of alkenes, alkynes and allenes is the addition of an A-B bond across a carbon-carbon π-bond.1 Since nitrogen-carbon bonds are ubiquitous in molecules and of significant importance in human life, as well as for the chemical industry;2 the so-called hydroamination reaction has attracted considerable interest.3 However, a close look at the literature reveals that although the intramolecular hydroamination of allenes4 has been widely studied,5 the intermolecular version has received limited attention. Depending on the regiochemistry, this reaction can lead to imines, enamines or allylamines. The allylamines, (Markovnikov adducts) are especially interesting since they are among the most versatile intermediates in synthesis, components of many naturally occurring and biologically active molecules, and of industrial importance.6 In 1992, Bergman reported that zirconium bis(amides) were efficient catalysts to promote the addition of primary arylamines to 1,2-propadiene giving the corresponding imines,7 and more recent papers confirm that early transition metal catalysts lead to the anti-Markovnikov adducts (Scheme 1, top).8 The Markovnikov addition has first been achieved in the presence of equimolar amounts of mercury(II), palladium(II) and platinum(II) salts,9 and then used a palladium(0) catalyst in the presence of triethylammonium iodide10 or acetic acid11 as cocatalysts. Gold catalysts12 have more recently been developed.13 Yamamoto et al., have first demonstrated that 10 mol% of AuBr3 allowed the room temperature addition of primary arylamines to allene to afford allylic amines,14a and then described the addition of morpholine to mono- and disubstituted allenes at 80 °C, using 10 mol % of a 1/1 mixture of (Ar3P)AuCl and AgOTf (Scheme 1, middle).14b In 2008, Widenhoefer reported the Markovnikov hydroamination with N-unsubstituted carbamates using catalytic amounts of an equimolar mixture of LAuCl and AgOTf (L = NHC, phosphine) (Scheme 1, bottom).15 Interestingly, an opposite regioselectivity was observed, when compared with Yamamoto’s results.14b Recently, we have shown that cationic gold(I) complexes (Cat) (Figure 1),16 featuring a bulky cyclic (alkyl)(amino)carbene (CAAC),17 as ancillary ligand, efficiently promoted the addition of NH3 to non-activated alkynes and allenes.18 These first examplehs of hydroamination reaction with ammonia prompted us to investigate the scope of application of our catalytic system. Here we show that Cat1 [a 1/1 mixture of (CAAC)AuCl and KB(C6F5)4] is efficient to promote the intermolecular Markovnikov hydroamination of allenes A-C (scheme 2) with a variety of primary and, more importantly, secondary amines.
Since, as mentioned above, several catalytic systems are known to promote the hydroamination of allenes with primary amines, we briefly investigated if Cat1 was efficient as well (Table 1, Entries 1–3). In the presence of 5 mol % of Cat1, we were pleased to observe that after 12 h at 70 °C, aniline reacted with 1,1-dimethyl allene A to afford a 60/40 mixture of N-allyl- and N,N′-diallylaniline in 95% yield (Entry 1). Clearly, the N,N′-diallylaniline results from the addition of the primarily formed N-allylaniline on allene A, which gives a clear indication that secondary arylamines are suitable substrates for our catalytic system. Mesitylamine, and even the strongly basic and bulky tert-butylamine react with A, although under more drastic conditions, giving the secondary allylamines in 74 and 54% yield, respectively (Entries 2 and 3).
Then, we turned our attention to secondary (aryl)(alkyl)amines using again 5 mol % of Cat1 (Table 1, Entries 4–9). In an initial experiment, 1,1-dimethylallene A was treated with N-methylaniline, and after 12 h at only 70 °C, the desired hydroamination product was obtained in 98% yield (Entry 4). Para-substitution of the aniline by a chlorine or methoxy group are well tolerated, the corresponding adducts being formed in 97% yields (Entries 5 and 6). 1,2,3,4-Tetrahydroquinoline and indoline have been recognized as important synthetic intermediates,19 and exhibit extensive biological activities and potential pharmaceutical applications. Using allene A, their allylic amine derivatives have been prepared using our catalytic procedure with up to 99% yield (Entries 7 and 8). The hydroamination reaction with secondary (aryl)(alkyl)amines is not limited to 1,1-disubstituted allenes. The parent allene B can also be used as shown in Entry 9.
Based on these positive results, we investigated the catalytic activity of Cat1 for the hydroamination of allenes with more basic secondary amines (Table 2). By analogy with the work of Yamamoto, we first reacted morpholine with allenes A-C, and in all cases the Markovnikov adduct was obtained in good to excellent yields (Entries 1–3). Benzylic and benzocyclic amines also smoothly react with allenes A and B at temperatures between 70 and 90°C affording the hydroamination products in 93 to 99% yields after only 8 to 12 h (Entries 4–7). Lastly, we used diethylamine as the prototypical secondary alkylamine. Although drastic conditions are required (130–165°C, 24 to 36 h), the addition occurred to yield the Markovnikov adduct in 61 to 98% yields (Entries 8–10). These results emphasize the robustness of the catalyst.
To gain insight into the mechanism of the gold(I)-catalyzed hydroamination of allenes, the following experiments were performed (Scheme 3). Addition of excess allene A to complex Cat2 did not give the desired allene gold complex at room temperature and even upon heating at 50 °C. This is in marked contrast with the case of alkynes, for which an η2-alkyne complex was readily formed and isolated.18,20 On the other hand, upon exposure at room temperature of a solution of Cat2 in deuterated benzene to one equivalent of diethylamine, the amine complex I was spontaneously and quantitatively obtained. Then, addition of excess allene A to a C6D6 solution of I gives after 16 h at 70°C, complex II, which was isolated in 95% yield. These observations call into questions an outer sphere mechanism involving the attack of the amine to an initially formed gold π-allene complex,15 and rather suggest an inner sphere mechanism with transient formation of a tricoordinated gold complex, as proposed by Yamamoto14 and Tanaka21 for gold-catalyzed hydroamination of allenes and alkynes, respectively, with arylamines. Moreover, it seems quite likely that the amine complex I is the resting state of the catalyst, just as Cat3 is in the case of the hydroamination with ammonia.18
Since we have already shown that Cat1 promotes the addition of ammonia to allenes,18 it appears that the catalytic system described here has a broad scope of application; many types of non-tertiary amines can be used, as well as a variety allenes. Importantly, the regioselectivity observed (Markovnikov addition) is opposite to that obtained with early transition metal catalysts (anti-Markovnikov addition).7,8 The regioselectivity is also different from that reported by Windehoefer,15 using unsubstituted carbamates and a cationic gold(I) complex bearing an NHC or a phosphine ligand; the Markovnikov adduct was formed but at the more substituted terminus of the allene (Scheme 1). This striking difference has not yet been rationalized, and mechanistic studies are under active investigation. Moreover the hydroamination of 1,3-disubstituted allenes allows for the formation of a chiral center, and we are investigating the possibility of using gold(I) complexes bearing an optically active CAAC ligand for the preparation of enantiomerically enriched allylic amines.22
Financial support from the NIH (R01 GM 68825) is gratefully acknowledged. Thanks are also due to the China Scholarship Council for a Graduate Fellowship (X.Z.).