I wouldn't even attempt it without fentanyl test strips or some kind of access to analytical laboratory equipment, but there is the fact that fentanyl is only 5 per cent as soluble in water as diacetylmorphine, and considerably more soluble in organic solvents. If this can be put to use in a kitchen chemistry chromatography method would seem to be somewhat likely but could be a very involved process; in theory, a fractional crystallisation method bearing some resemblance to Cold Water Extraction is also possible.
If this is street smack in the States or Canada, in addition to the various fentanyls, there is the potential of Upjohn 44700, Allen & Hambury's 7921, piperidylthiambutene, clonitazene, benzodiazepine RCs and other things in there as well.