[Babcock_Hall]

We have made CH₃SO₂CH₂P(O)(OCH₂CH₃)₂ multiple times by reacting Oxone with the corresponding sulfide in methanol/water.  We extract from DCM and use the product without further purification (Blumenkopf 1986 and others).  We removed the solvent using rotary evaporation and high vacuum.  Our most recent synthesis had more than 10% more mass than theoretical and showed a broad peak in the H-1 NMR spectrum near the doublet arising from the bridging CH₂ group.  The appearance is more amorphous than crystalline.  These observations are entirely unexpected based upon our past experiences.  The only difference of which I am aware is that some of the starting sulfide came from a new bottle from the same manufacturer.  This week we tried dissolving in chloroform and removing the solvent, in the hope of stripping away what might be methanol that was present.  The mass is pretty close to what it was before, and I have not checked the NMR yet.  Does anyone have a suggestion about what to try next?

[rolnor]

Can it be simply water? Can you run chromatography, TLC?

[Babcock_Hall]

TLC has not yet shown us a second peak.  We were thinking about running a column using one of our TLC solvent conditions when a colleague suggested dissolving the product in chloroform.  That will be what we try next.

[rolnor]

Its a very polar compound. Yes, that's a good idea.

[Babcock_Hall]

I just reread my previous comment, and it was worded poorly.  We tried the chloroform trick about a week ago, and that was unsuccessful.  We have not yet tried a column.

[rolnor]

To have a purification system is of course very important and this can be time-consuming to develop, but it's a necessity.

[Babcock_Hall]

We worked out solvent conditions a week or two ago, and PMA and anisaldehyde give adequate staining in TLC, although neither one is great.  We made this compound at least five times previously, and it was always so pure that no column or recrystallization was necessary.  A big mystery is what was different on this occasion, and I don't even have a good hypothesis.  If we are able to purify it, I won't let the lack of a hypothesis trouble me.

[rolnor]

Can you make another batch? It does not look expensive to make?

[Babcock_Hall]

Yes the sulfone is easy to make, and a new synthesis also occurred to me as possibly being the better choice.  Of course, the same problem may occur twice, whatever it is.  BTW we obtained a H-1 NMR of the starting sulfide, and it looks fine.  We have not yet tried recrystallization, but my bet is that hexanes or heptane would work.

[MOTOBALL]

Two suggestions to determine if the additional mass is due to solvent of crystallization,

1) Send a sample for micro-analysis for C, H and P

2) Heat a weighed amount of sample in a vacuum oven overnight at 50-60 deg; any weight loss will indicate a volatile component (solv. of xtlln or a byproduct/impurity).

[rolnor]

Yes the sulfone is easy to make, and a new synthesis also occurred to me as possibly being the better choice.  Of course, the same problem may occur twice, whatever it is.  BTW we obtained a H-1 NMR of the starting sulfide, and it looks fine.  We have not yet tried recrystallization, but my bet is that hexanes or heptane would work.

That sounds like very non-polar solvents? I think it will not dissolve in those? I don't know, toluene maybe? If you boil it and let it cool?

[Babcock_Hall]

After silica gel chromatography the compound has less mass (now the mass is below theoretical) and looks more crystalline.  The broad peak that overlaps with the PCH₂S(O)₂ doublet in the H-1 spectrum is smaller, but it is not gone.  There are other low, very broad peaks in several places in the NMR spectrum: 3.3, 1.6 and 0.85 ppm being the most prominent.  I saw something similar before, but I did not give it much thought.  My current thinking is to try the synthesis over again.  If the second batch is impure, then we will develop recrystallization conditions.  I will also give the two NMR spectra (before and after chromatography) a more careful look.
EDT
I added two pdf files.  Each one shows a portion of the H-1 NMR.  One was taken after we stripped with chloroform but before the silica gel column.  The other was taken after the column.

[rolnor]

Good! Isn't 3.6 typical for water? I am very rusty on NMR You could try DMSO-d6 as solvent It looks like poor resolution? You can also heat the sample in the probe. Or use MeO-D4, then the OH, if OH should go away or move. It's strange with a broad signal like that if not OH, it must be a polymer or something that does not dissolve properly? It can not be PEG? It's to up-field?

[Babcock_Hall]

I would call our purification a partial success.  When there is a small amount of water present in CDCl₃, it usually shows up around 1.6 ppm.  I looked over the H-1 NMR of the starting material, and there is a low, broad peak that might be a polymer.  I am not sure whether I should ignore this signal in the starting material; I plan to buy some more of this starting material soon.

[rolnor]

OK, I don't remember the shift of water in CDCl3 correctly. The next step is a condensation or? I don't think this polymer will cause much trouble, your molecule is very acidic, and the base will pick the protons from this, not from the polymer. Yes, of course, you almost got rid of it, that's good.