Oligosaccharide analogues of polysaccharides .5. Studies on the cross-coupling of alkynes and haloalkynes

The cross-coupling of the homopropargylic ether 1 and the halopropargylic ethers 2a and 2b was optimized, and aspects of the coupling mechanism were studied. Coupling promoted by Pd-0 and Cu-I in the presence of an amine yielded a mixture of the heterodimer 3 and the homodimers 4 and 5 (Scheme I). Optimizations were first directed at suppressing home-coupling. Home-coupling is partially due to a H/I exchange (1+2a --> 6+7) promoted by CuI and an amine. The exchange, but not the formation of homodimers, was largely suppressed in DMSO. The influence of phosphine ligands was also evaluated. Weaker a-donors (with the exception of PPh(3)) lead to a faster coupling and to a higher ratio of hetero- to homodimers, with P(fur),leading to the cleanest reaction. Homodimers are also formed (together with I-2 .(i-Pr)(2)NH) by reductive dimerization of the iodoalkyne 2a in the presence of [Pd-2(dba)(3)], CuI, and (i-Pr)(2)NH. Bulky and acceptor-substituted amines reduced the extent of the dimerization of 2a, but the bulkiest amines did not promote coupling. Better results were obtained by using the bromoalkyne 2b. Neither dimerization of 2b, nor H/Br exchange between 1 and 2b were observed. Coupling of 1 and 2b was slower than the one of 1 and 2a, but gave higher yields of the heterodimer 3. The yield of 3 and the ratio of hetero- to homodimers was greatly improved by addition of LiI; no phosphine ligand is then required. While the oxidative addition of the iodoalkyne 2a to [Pd(PPh(3))(4)] (2a --> 8a) was rapid, the one of the bromoalkyne 2b was much slower and proceeded via the eta(2)-complex 9 as evidenced by H-1-NMR spectroscopy. The rearrangement of 9 to the bromopalladium sigma-complex 8b follows first-order kinetics (k = 0.014 min(-1)). CuBr greatly increased the rate of this rearrangement. LiI caused rapid substitution of Br by I in the Pd sigma-complex (8b --> 8a), but not in 9, nor in 2b. The sigma-complex 8a did not react with the alkyne 1 in the presence of (i-Pr)(2)NH, unless Cu-I was added. The alkynes 10 or 1 did not react with CuI and either TMEDA or (i-Pr)(2)NH to yield detectable amounts of the Cu-acetylides 11 or 12. These observations are rationalized by the mechanism shown in Scheme 3, postulating the intermediacy of the binuclear alkyne-Pd-Cu complexes C and J, and some or all of E-H, and highlighting the role of Cu-I in this coupling.