An efficient protocol for Sharpless-style racemic dihydroxylation

It is usually easier to carry out a reaction to produce racemic rather than optically active products. There is one notable exception: the Sharpless asymmetric hydroxylation 1 is so straightforward and practical that it is easier to carry out than previous racemic versions. In using this reaction on a variety of olefins we needed to make the racemic compounds to provide reference samples of diols so that we could measure the enantiomeric excess in the asymmetric dihydroxylation. We present a simple and practical racemic version (RD) of the Sharpless dihydroxylation using quinuclidine (which is, of course, achiral) as the ligand and osmium() chloride as the catalytic reagent and demonstrate its use on a variety of functionalised olefins. Our initial experiments were carried out on substituted stilbenes. The optically active diols were used to make dibenzophosphepines 2,3 for use as chiral auxiliaries. We were unable to detect the other enantiomer of these diols in their NMR spectra in the presence of the Pirkle shift reagent. We needed racemic material to check that the enantiomers did indeed give distinct signals in the NMR spectra with Pirkle’s reagent. As we were already familiar with the excellence of the Sharpless AD procedure, we felt that a similar racemic procedure might be devised which would be more convenient and efficient than the older Upjohn method. Sharpless mentions in passing doing a racemic dihydroxylation in work on kinetic resolutions 6 and on double diastereoselections. Although he has not described a general racemic protocol, Narasaka’s modification has been used to improve diastereoselectivity in the racemic dihydroxylations of trienes. Although Sharpless reported 9 that quinuclidine retarded dihydroxylation under single phase conditions using Nmethylmorpholine N-oxide (NMO) in aqueous acetone, Minato made the observation 10 that quinuclidine accelerates dihydroxylation under the biphasic conditions of tert-butyl alcohol and water. This coupled with a footnote 9,11 by Sharpless that solid OsCl3 gave exactly the same results as osmium tetroxide led us to use a mixture of catalytic OsCl3 with K3Fe(CN)6 as cooxidant and quinuclidine as ligand in biphasic aqueous tert-butyl alcohol. A brief study (Table 1) with 2,29dibromostilbene 1 revealed that quinuclidine suppressed the formation of the bright yellow dione 3 while methanesulfonamide was necessary to accelerate the reaction, by increasing catalytic turnover (Scheme 1). Using this protocol we dihydroxylated a number of stilbenes in good yields (Table 2). The stilbenes were prepared from the