[Babcock_Hall]

Previously we oxidized a sulfide to a sulfone in the presence of a primary alcohol and a phosphonate diester using Oxone^R (this reaction is mentioned in a different thread).  The product was sufficiently pure that it did not require a silica column to purify.  We ran a second oxidation and were suprised to observe two products are apparent in the H-1 spectrum as well as by TLC.  My first thought was that we had incomplete oxidation, producing a mixture of sulfoxide and sulfone.  However, the P-31 NMR spectrum gave two peaks in the vicinity of 11 ppm, suggesting that both are sulfonyl phosphonates (a sulfoxide should be roughly 5 ppm downfield of a sulfone, and the starting material should be roughly 24-25 ppm).  One possibility is that we converted the alcohol to an aldehyde, but that seems unlikely.  We will attempt to capture and characterize both products next week (they separated reasonably well by TLC), and one of them should match what we made before.  Any guesses about the identity of the other product?   

[Corribus]

FTIR would be really helpful here.

[Babcock_Hall]

Good idea; we will look into it assuming that we are able to perform the separation successfully.  I think that it is unlikely that the phosphonate lost an ethyl group, because if it had, it might have gotten pulled into the aqueous layer during the extraction.  There are only three functional groups on the molecule, which is why I am scratching my head.

[rolnor]

Can the alcohol form a ring with the phosphonyl group? Trans-esterification/ringclosure?

[Babcock_Hall]

That is an interesting thought.  It would be a 7-membered ring.

[rolnor]

Thats less likely, a 5 or 6-membered ring would be a real possibility, I have done this type of chemistry on phophonatesters.

[Babcock_Hall]

Interim update:  We isolated two products by chromatography from the second synthesis.  Both are probably sulfonyl phosphonates on the basis of their P-31 chemical shifts.  One is close to the product of the first synthesis, but there is an extra signal in the H-1 NMR that has a chemical shift (2.07) and integral consistent with an acetyl methyl group, CH₃C(O)-.  The other product has two very broad, non-Lorentzian peaks (near 2.5 and 1.8 ppm), each of which integrate to 1H.  I don't recall seeing anything like this.  On the other hand, there is an indication that the (EtO)₂2P(O)CH₂- portion of the molecule is intact. 

Time permitting, we plan to look into these questions further in the next few days with the following limitation.  If a compound is not our desired compound, its identity may help us learn how not to make it the next time we perform this synthesis; we will not pursue learning about its identity further than achieving this goal.

[Babcock_Hall]

After reexamining the H-1 data and looking at the carbon-13 data for pool A, I am convinced that pool A became acetylated on the hydroxyl group.  The only donor I can think of is ethyl acetate, which was used during extraction and the purification.  I do not have a detailed mechanistic rationale.  Another puzzling piece of information is that the reason for running a column on this batch was that we believed that two compounds were present in the crude product after oxidation.  Perhaps some of the crude product was activated in some way by Oxone^R, and somehow trans-esterification occurred.  The acetyl group could function as a protecting group in a subsequent synthesis.

I think that pool B, the later eluting fraction, is the product without an acetyl group, but there is a minor difference in the H-1 signal of the -CH₂OH.

[Babcock_Hall]

Just to follow up, the second synthesis sat longer than the first synthesis after the extraction.  I reexamined the first synthetic batch for signs of transesterification, which we made a few months ago.  I did see a small singlet at a similar chemical shift (2.04 ppm vs. 2.07 ppm in the second batch).  However, the signal at 2.04 ppm might arise from ethyl acetate that was not completely removed after the extraction.  Ideally, we would like to control whether or not transesterification occurs, but I am satisfied with simply having identified the two pools.