516 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS It deteriorates after a few days, possibly because of reaction between the ester and ammonia. Upward development of the chromatogram is the preferred technique because of its simplicity. Pinacryptol yellow is a very useful reagent for detecting the surface- active agents in the chromatogram, giving under ultra violet illumination an orange fluorescence with most anionic surfactants, and blue with catiordcs. Amine oxides and non-ionics surfactants, of higher R F values, also give blue colours with pinacryptol yellow, and toluene and xylene sulphonates, of lower R F values, give orange colours. Characteristic changes occur in the spots on heating, and a subsequent spraying with Rhodamine B also modifies their appearance and makes many spots more definite. Exposure of the paper to iodine vapour gives dark brown spots with ethylene oxide derivatives, amine oxides and cationic surfactants. It also gives spots with alkanolamides, alkanolamines, and metals, but these dis- appear in a few minutes. Spraying the paper next with bromocresol green solution shows all alkaline regions due to the amine oxides and cationic surfactants already detected, to the various alkanolamines which can then be identified from their various R F values, and to alkali metals which remain at the origin. The metals can subsequently be separated on another chromatogram with a more mobile solvent of methanol and ammonia solution (9-51). A spray with cobalt thiocyanate is useful for confirming the presence of ethylene oxide derivatives and amines that were revealed by the iodine treatment, and a spray with dimethylaminobenzaldehyde is used to test for urea. l)uring the last few months the ninhydrin tests for amines has been found useful. Purple spots are developed after heating to 80øC with primary amines or 120øC for secondary and tertiary amines. These confirm the presence of the alkanolamines detected with bromocresol green, but also indicate the presence of mono- and diethanolamides in spots near the solvent front. It is suspected that these spots are due not to the substituted amides as such but to amine-ester impurities. The ninhydrin reagent is also sufficiently sensitive to detect the traces of free alkanolamine present in the alkanolamides added in the manufacture of the detergent. The ninhydrin spray may be followed by acetaldehyde and nitroprusside which gives a useful distinct/on between primary amines (khaki spots on clear paper, not readily visible after ninhydrin) and secondary amines (deep blue spots). I)rewry described the successive application of all spray solutions to the same chromatogram, but the incorporation of nin- hydrin into the list of treatments tends to overwhelm the paper and it is best to develop two chromatograms and apply the first five sprays to one series and the last two to a duplicated series of spots. Another small change

THE ANALYSIS OF SYNTHETIC DETERGENTS 517 made from the published procedure is the rather obvious simplification of using one marker solution containing all the reference compounds. Another method of identifying amides, applicable both to the nonionic alkanolamides used as additives and to amide sulphonates and carboxylates used as the main anionic surfactants, is to apply paper chromatography after acid hydrolysis. The aidehyde nitroprusside spray is the most useful as it distinguishes primary amines such as monoethanolamine and taurine (khaki spots) from secondary amines such as diethanolamine, N-methyltaur- ine and sarcosine (blue spots), all liberated by hydrolysis of the corresponding fatty acid amide. •)UANTITATIVE ANALYSIS BY SOLVENT EXTRACTIONS The most reliable technique of analysing mixtures of surfactants is a series of extractions, with solvents and ion-exchange resins, before and after hydrolysis, to separate the individual fractions which are then weighed. Each fraction can, if desired, be characterized by further analysis including physical methods such as gas chromatography. A series of separations can be assembled in a variety of ways and therefore it seems best here to discuss the subject under sub-headings of the solvents. The general method for liquid-liquid extractions is to use stoppered separating funnels, and for liquid-solid extractions is simple stirring in a beaker or centrifuge tube, followed by filtration or centrifuging. Light petroleum Typical procedures for petroleum extraction are described in the B.P. and U.S.P. monographs for sodium lauryl sulphate. The sample is dissolved in 50% ethanol, for the unsulphated alcohol is less soluble in this than in an aqueous solution of the sample (owing to reduced miceilar effects) and emulsification difficulties are fewer. Three extractions with petroleum are made. The combined extracts are dried, the solvent is distilled off and the residue is weighed. A certain proportion of free lauryl alcohol remains solubilized in the surfactant solution, as is shown by tests on synthetic mixtures, even after 5 or 10 extractions. The loss depends on the con- centration of surfactants and, for reproducible results, the latter is arbitrarily fixed at about 5% w/v. Besides the question of completeness of extraction there are several other difficulties in arriving at a standard method of high reproducibility. Washing the petroleum extracts is always liable to cause losses, and these must be balanced against the errors due to contaminants. Drying the extracts with sodium sulphate or another desiccant may lead to a loss, but omission of the drying step may cause losses by volatilisation in steam. Removal of the solvent without loss of volatile alcohol is also a problem and, finally, errors associated with weighing the residue in a large