I was going over some old H-1 NMR data this week, and I noticed something strange. I had synthesized the di-tert-butylester of the attached compound (R = C(CH3)3), and there were two signals in CDCl3. One was a doublet near 4.40 ppm, presumably split by P-31. The other was a quartet near 4.17 ppm, presumably split by F-19. I then removed the tert-butyl groups in acid (R = H). In d6-DMSO I saw a doublet near 4.36 ppm, presumably P-CH2O, and a quartet near 5.04 ppm, presumably CF3CH2S. The splitting is about 9.6 Hz. I am puzzled about the apparent downfield shift of the methylene group next to the trifluoromethyl group. Thoughts?.