This summer I had a little time to devote to the problem. A considerable help was a copy of the Sigma–Aldrich Handbook of Stains, Dyes and Indicators (2) given to me by an associate. Looking through the UV–vis spectra in this work, I concluded that the unknown dye was an azo dye produced from an aniline and a 2-naphthol derivative. Our experience in lab making water-soluble azo dyes for dyeing multifiber cloth convinced me that the dye must have at least two sulfonic acid groups. NMR of the extract showed a lot of carbohydrate but no aromatic signals. Chromatography of the material separated the dye from the sugars, but did not yield enough material for a spectrum.

TLC showed that the M&M dye was none of the following: Orange I, Orange B, Orange G, Crocein Orange G, Acid Orange 8, Orange GGW.

At this point I decided to synthesize a number of orange dyes. Using ten substituted anilines and six 2-naphthols commonly used in making dyes, 60 orange dyes were produced. Of these nearly half were not water soluble enough for serious consideration. The rest were assayed by TLC leaving four contenders. Careful recording of the visible spectra showed that one was an exact fit. This was the dye formed from sulfanilic acid and Schaeffer’s acid (6-hydroxy-2-naphthalene sulfonic acid). This was confirmed by running this dye and the M&M dye in about 10 different TLC systems.

A trip back to the Color Index in the library brought a great surprise. The dye in question is C.I. 15985 (also known as sunset yellow or FD&C yellow 6), which is not included in the Sigma–Aldrich handbook.

Literature Cited

1. Birdwhistell, K. R.; Spence, T. G. J. Chem. Educ. 2002, 79, 847.
2. Green, F. J. Sigma–Aldrich Handbook of Stains, Dyes and Indicators; Aldrich Chemical Company Library: Milwaukee, WI, 1990.