[rolnor]

I understand that you can have mixed feelings about using a unfamiliar technique such as RP-silica chromatography but if you submit this paper reviewers will probably question why you have not tested this to purify your material especially if it is intended for testing in some kind of biological assay.

[Babcock_Hall]

The Dowex-501 that I used was old, and it has occurred to me that this might have hurt its effectiveness.  I think we should scale up the synthesis and optimize the yield in the first step (we had to rerun one column, owing to a suboptimal choice of solvent).  Although I have not entirely given up on the mixed-bed idea, I will look forward to trying the RP-silica.

[rolnor]

Yes, I had also had problem with degraded ion-exchange resin. I think you dont need more than 1.2 fold excess TBAF. Have you considered other reagent like HF-pyridine?

[wildfyr]

Why not just take an anion exchange resin, stir in KF or NaF in methanol, wash the heck out of it again with clean methanol, dry it, then stir your compound for several hours over this resin. Filter off resin and rinse, rotovap down and voila pure product. Make sure to use a notable excess of resin eqs (like 2) since it will be slower than liquid phase. No need to mess around with TBAF.

I saw you were doing something with resin, but I didn't really follow and it seemed painful.

OR! Do it biphasic. KF in water, product in organic phase, then stir the christ out of it. Use saturated KF because the high salt concentration should force the organic products to stay in the organic phase since I saw your substrate is rather hydrophilic. I know MeCN and sat KFHF in water are not miscible, that's how I made sulfonyl fluorides out of sulfonyl chlorides. If that solution can keep MeCN out of the water phase, then I bet it will force your product to stay in the organic phase, MeCN is freely miscible with pure water.

[Babcock_Hall]

wildfyr,

Are you suggesting in effect putting the anion-exchange resin into the fluoride form?  That is a very creative idea.  IIUC either there will be some methoxide and it and the methanol will be in the equilibrium with the product, ROH, and RO^-.  If RO^- is formed, it will be the counter-ion to the anion exchange resin.

Rolnor,

I need to look into pyridine•HF more closely; I have been a slacker.

[wildfyr]

Yep that is my suggestion. Then don’t use methanol, use thf Or acetone?  You can find a good solvent. I know you can dissolve a salt in methanol and use a resin and the resin will be the desired anion.

But fluoride is a crappy base, I don’t think there would be much Methoxide present. The autoionization of methanol is very small. KF in methanol plus anion exchange resin will give you 99.9% fluoride in the active sites. Then you can rinse it once with clean methanol follow up by an aprotic solvent to flush all the methanol out.

[rolnor]

Its clever wildfyr, is it maybe possible to have a resin with amino-residues (Sephadex) and make a hydrofluoride-salt of this? It feels as if this will give more free fluoride ions than a quaternary ammonium fluoride resin. It could also be possible  to have the amino-resin loaded with excess hydrogen fluoride like in HF-pyridine.

[wildfyr]

I think the amount of "free" fluoride is based more on the number of active sites than the nature of the cation. This is a bit more of just choosing a resin and less the moiety within.

I don't feel like going back through the post, but perhaps if the substrate is highly sensitive then plain fluoride is preferable to a "HF" source, not to mention a bit safer to handle, since I bet ammonium HF will give off some HF gas in the presence of ambient water.

I think also its worth trying the biphasic method with saturated fluoride salt and MeCN. I used to to convert sulfonyl chlorides to sulfonyl fluorides with KFHF and it worked beautifully. The trick is just to stir like crazy. As with the resin method, workup is incredibly trivial.

[Babcock_Hall]

With four hydroxyl groups, I think that our product will prefer to remain in the water layer.

The Kaburagi/Kishi method worked pretty much as advertised, and if we are able to remove the ~0.1 equivalent of tetrabutylammonium ion, the compound will be nearly pure.  However, my students won't be able to pursue this for a while.

[Babcock_Hall]

We tried the Kaburagi/Kishi work-up method again last week on an aromatic glycoside.  This is the method of using Dowex-50 and calcium carbonate to remove tetrabutylammonium fluoride after removing TBDMS groups from an oxygen.  There are four TBDMS groups in our glycoside intermediate, and the ratio of TBAF to TBDMS group was 2:1.

We increased the amount of Dowex-50 by 50% over our previous attempted clean-up.  By H-1 NMR (using the signal from the methyl groups from the tetrabutylammonium ion) we see about 99.7% removal of the tetrabutylammonium groups.  We saw about 99.2% removal last time. 

[rolnor]

Great! A good end of the week.

[Babcock_Hall]

We tried using TBAF on the para-isomer of our first compound (which had an ortho-relationship between the two functional groups on the aromatic ring).  For some reason, the reaction failed completely.  We stored the TBAF solution in the refrigerator.  When we remove aliquots, we remove using a syringe.  Any thoughts as to how to proceed?

[rolnor]

How do you know it didnt work, NMR? If you have a hydroxy ortho to a carbonyl, deprotection may not show on TLC because you get internal hydrogen bond so Rf does not change

[Babcock_Hall]

Both C-13 and H-1 NMR showed the presence of the tertiary-butyldimethylsilyl group.  The product was not soluble in D₂O.

[rolnor]

Fluorine is so small so steric factors are small, its hard to say. Have you tried to add more reagent? As you know, its almost a fool-proof reaction.