Although I began to discuss this reaction on another thread, I think it is better to start a new topic, to focus on the particulars of this chemistry. We ran a reaction using TBSOTf/lutidine on a β-aromatic glucoside with the intention of adding four TBDMS groups. There were two similar protocols in the literature, but we did not strictly follow either one (it ended up being hybrid between the two). We did not know which spot corresponded to our product by TLC, and because we were unsure whether or not the reaction had gone to completion, we added more of our silylation reagents than one protocol indicated, moving from about 1.5 TBDMSOTf per -OH group to about 2.0. We washed with aqueous HCl and other aqueous solutions, and we ran a silica gel column using 7% ether in hexanes.
Our product is not pure; in our three pools of fractions, there was only modest separation between what appears to be our product and a colored impurity. The impurity has strong signals near 0.85 ppm and 0.00 ppm, and the ratio of the two peaks is about 3:2. Its identity is not yet clear to us, although it would seem to be a derivative from unreacted TBDMSOTf. We think we can get better separation with a less aggressive mobile phase, and we can also examine other systems. If we had the reaction to do over again, we would strongly consider adding a copper sulfate wash step. I am not sure whether HCl or copper sulfate is better at removing lutidine. Does anyone have any comments on the probable identity of the impurity or on other things we should try?