We have performed this reaction to make vinyl sulfones or acrylamides about 25 times , with 0.0-0.2 mole excess phosphonate over aldehyde. In a few cases we see the starting phosphonate by H-1 and C-13 NMR (it is easily identified by the signals which couple to P-31) in the purified product, suggesting co-elution during chromatography. We made a new vinyl sulfone a few weeks ago, and the percentage of phosphonate was much higher than I had seen in other reactions. We are performing the resynthesis (the present reaction) with no excess phosphonate. This is the first time that I tried P-31 to check for completeness. In our hands mixtures of EtOAc/hexanes, acetone/hexanes, DCM/ethanol, and DCM/iPrOH have all failed to separate the phosphonate from our impure product standard from the first synthesis.
Because the starting phosphonate (P-31 chemical shift of 11.7 ppm in CDCl3) cannot yet be separated from the product by TLC or by silica chromatography, I decided to use P-31 NMR to gauge the completion of the reaction. I used d6-DMSO, and I expected that the product, diethyl phosphate, should be near 0-5 ppm. After the the reaction had run for three hours, which is longer than we sometimes use, I took the P-31 spectrum. I saw a peak near 22.6 ppm and one near -0.77 ppm; their integrals were nearly 1:1. I have not check my reference books, but -0.77 ppm sounds reasonable for diethyl phosphate, but identity of the compound giving rise to the peak at 22.6 ppm has me stumped (the difference between the two peaks is over 2700 Hz). I took another time point using chloroform as the solvent, but I did not see any signal (I did see a hazy precipitate) I do not know how to interpret the DMSO spectrum in terms of completion of our reaction. The spectra were of the crude reaction mixture in acetonitrile. I attempted another P-31 experiment in CDCl3, but a hazy precipitate developed, and I did not observe any P-31 signal (perhaps the phosphorus-containing compound(s) precipitated).