Control of Selectivity in Palladium-Catalyzed Oxidative Carbocyclization/Borylation of Allenynes

Depending on the reaction conditions a selectivity <80% for either the borylated trienes or the borylated vinylallenes is achieved in the product mixtures.

clization/arylation conditions alkyl-substituted allenynes afforded two different constitutional isomers (arylated trienes and arylated vinylallenes) in a ratio determined by the substitution on the starting allenyne. [13b] The aim of the present study was to develop a carbocyclization/borylation that can be directed towards either a borylated triene or a borylated vinyllallene by control of the reaction conditions (Scheme 1 c). We now report a highly selective oxidative carbocyclization/borylation of allenynes 1 with B 2 pin 2 under Pd II catalysis with p-benzoquinone (BQ) as the oxidant. The use of LiOAc·2 H 2 O in 1,2-dichloroethane (DCE) or BF 3 ·Et 2 O in THF addressed the issue of selectivity, to give either borylated trienes 4 or borylated vinylallenes 5, respectively.
We first studied the reaction of ethyl-substituted allenyne 1 a with B 2 pin 2 under the original carbocyclization/borylation conditions (Scheme 1 b). [13b] The use of a catalytic amount of palladium acetate (2 mol %) and stoichiometric amounts of BQ (1.1 equiv) in THF at 50 8C led to an isomeric mixture of borylated triene 4 a and borylated vinylallene 5 a in 28 % and 14 % yield, respectively (Table 1, entry 1). Analyzing the effect of different solvents showed that a higher selectivity for 4 a was obtained when DCE was used as the solvent (in Table 1, entry 4 vs. entries 1-3). Furthermore, upon the addition of catalytic amounts (20 mol %) of a basic salt, Scheme 1. Palladium-catalyzed borylating carbocyclizations of allenynes: a) under non-oxidative conditions; [11]  The finding that the addition of a basic salt substantially enhanced the selective formation of alkenyl boronate 4 encouraged us to study the effect of acidic reaction conditions. To our surprise, the addition of a Brønsted acid, such as HOAc, generated an approximately 1:1 mixture of 4 a and 5 a in moderate yields (Table 1, entry 9) and the use of ptoluenesulfonic acid (p-TSA) even did not afford any borylation products (Table 1, entry 10). However, the use of a Lewis acid, BF 3 ·Et 2 O, resulted in a high selectivity for 5 and afforded products 4 a and 5 a in 3 % and 78 % yield, respectively (  4 ]-[(BF 4 ) 2 ] was used the same trend in selectivity was seen but a lower yield was obtained ( Table 1, entry 12 vs. entry 11). [16] The structurally similar Lewis acid BEt 3 was also tried and moderate selectivity for 5 a over 4 a was seen with low yields of products (Table 1, entry 13).
With the optimized conditions for the selective formation of borylated triene 4 a established, we applied them to differently substituted allenynes ( Table 2). The allenynes bearing a methyl group on the alkyne moiety (1 b and 1 c) afforded the borylated trienes as the sole products ( Table 2, entries 2 and 3). For substrates with a longer alkyl group (1 d and 1 f) on the alkyne moiety, the competing allene formation took place to a notable extent ( Table 2, entries 4 and 6), but the corresponding triene products 4 d and 4 f/4 f' could be isolated in good to moderate yields. In those cases where the substrates are unsymmetrically substituted at the allene  Entry Allenyne Product 4/5 [b] Yield of 4 [%], [c] ratio [b]  moiety (1 c and 1 e) a comparatively high selectivity for the triene products was observed ( Table 2, entries 3 and 5). However, a mixture of borylated triene products was obtained, with a preference for formation of the products with the more substituted double bond. Products 4 c and 4 e were obtained as a mixture of Z/E isomers in a ratio of 3.3:1 and 3.5:1, respectively. The allenyne 1 f with a cyclohexylidene group on the allene cyclized to give a mixture of isomers 4 f and 4 f', where the formation of 4 f' could be explained by a Pd-catalyzed isomerization of 4 f ( Table 2, entry 6; for the detailed mechanism, please see the Supporting Information). Moreover, the reaction of allenyne 1 g, having two benzyl ether groups on the linker part X, also gave borylated triene product 4 g selectively in 57 % yield ( Table 2, entry 7), thus proving that the malonate group of linker X is not necessary for a successful transformation. The optimized reaction conditions for the selective formation of borylated vinylallene 5 (20 mol % of BF 3 ·Et 2 O, Method B) were applied to various allenynes (Table 3). Allenynes 1 a-1 f were transformed into vinylallenic boronates 5 a-5 f; for most cases the yield was between 70 % and 80 % and the formation of the corresponding triene isomers 4 a-4 f was efficiently suppressed. Even the methyl-substituted substrate 1 b, which intrinsically favors formation of triene 4 b, [13b] displayed opposite selectivity under these reaction conditions, that is, favoring vinylallene formation ( Table 3, entry 2). The reaction of allenyne 1 g under the standard conditions of Table 3 was sluggish and did not give the desired product 5 g, probably because of the incompatibility between the benzyl ether group and BF 3 ·Et 2 O. However, by switching the palladium catalyst to [Pd(CH 3 CN) 4 ]-[(BF 4 ) 2 ] and in the absence of BF 3 ·Et 2 O, product 5 g was obtained in 37 % yield (entry 7).
To gain further insights into the mechanism of the oxidative carbocyclization/borylation, kinetic deuterium isotope effects were studied (Scheme 2). An intermolecular competition experiment using 1 d and its hexadeuterated derivative [D 6 ]-1 d under the conditions for selective triene formation for 1 h provided a large intermolecular KIE value of 6.7 [17] (Scheme 2 a). This result indicates that the allylic C À H bond cleavage involved has to occur prior to any irreversible step of the reaction, for example, the carbocyclization step. [18] On the other hand, when a 1:1 mixture of 1 d and [D 2 ]-1 d was subjected to the conditions for selective vinylallene formation for 1 h the ratio between 5 d and [D 1 ]-5 d was 2.4, from which the KIE was determined to 2.7 [19] (Scheme 2 b). [17] The intrinsic KIE from intramolecular competition for vinylallene formation was determined to 5.3 [17] [20] (Scheme 3 c).
The results in Scheme 2 and Scheme 3 indicate that competing allylic and propargylic C À H bond cleavage occurs in 1, and this determines the ratio of boronates 4 and 5 (Scheme 4). The allene attack on Pd II complex A through allylic CÀH bond cleavage [12, 13a-d] would give B and subsequent alkyne insertion would generate intermediate C.
Transmetalation of C with B 2 pin 2 and reductive elimination would form product 4. The competing alkyne attack through propargylic C À H bond cleavage in A would produce allenylpalladium intermediate D. Intramolecular vinylpalladation of the allene moiety would generate (p-allyl)palladium intermediate E. Transmetalation with B 2 pin 2 and subsequent reductive elimination would give 5. The mechanism in Scheme 4 is supported by the kinetic isotope effects and the experiments with deuterium-labeled compounds (Scheme 2 and Scheme 3). The lower kinetic isotope effect observed for the competitive experiment in Scheme 2 b compared to the intramolecular experiment in Scheme 2 c may reflect that 1 and A are not in full equilibrium under the conditions for formation of 5. In the path for formation of 5 it is likely that BF 3 ·Et 2 O creates a cationic palladium species, which interacts better with the acetylene compared to the allene in A. [21] In summary, we have developed an unprecedented selective Pd II -catalyzed carbocyclization/borylation of allenynes under oxidative conditions. By controlling the reaction conditions the reaction can be directed to either the triene 4 or the vinylallene 5. On the basis of the results of deuteriumlabeling experiments, we propose that the reactions of allenynes proceed through competing allylic and propargylic C À H bond cleavage pathways to give borylated trienes and borylated vinylallenes, respectively.