NMR is the most versatile tool for the analysis of organic compounds and, in combination with Diffusion-Ordered Spectroscopy ('DOSY'), can give information on compounds in complex mixtures without the need for physical separation. In mixtures where the components' diffusion coefficients are nearly identical, for example because of similar sizes, Matrix-Assisted DOSY ('MAD') can help separate the signals of different constituents, resolving their spectra. Unfortunately, DOSY (including MAD) typically fails where signals overlap, as is common in H NMR. Using F NMR avoids such problems, because the great sensitivity of the F chemical shift to local environment leads to very well-dispersed spectra. Another advantage is the absence of any F background signals from the matrices typically used, avoiding interference with the analyte signals. In this study, differentiation among fluorophenol and fluoroaniline isomers was evaluated using normal and reverse micelles-of sodium dodecyl sulfate (SDS), cetyltrimethylammonium bromide (CTAB) and dioctyl sodium sulfosuccinate (AOT)-as matrices. These surfactants provide useful diffusion separation in these difficult mixtures, with all the solutes interacting with the matrices to different extents, in some cases leading to differences in diffusion coefficient of more than 30%. The best matrices for separating the signals of both acid and basic species were shown to be AOT and CTAB, which are useful over a wide range of surfactant concentration. Copyright © 2016 John Wiley & Sons, Ltd.