522 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS is important and shows the reasons why end-points (b) and (d), at which most of the indicator is in the aqueous phase and nearly all the surfactant in the organic phase, are unsatisfactory with commercial matehals. End- point (a) is rarely used because it gives no warning of its approach. It also has the disadvantage that trace impurities in the indicator, particularly of the oxidation products in methylene blue, may impart a bluish colour to the aqueous layer well before the end-point. The same impurities also interfere in end-point (e) and from our experience we recommend the use of this end-point only with the artionic indicators which are purer and more stable than the cationic ones. This leaves end-points (c) and (f) as the best for cationic indicators such as methylene blue. It might be mentioned at this point that methylene blue seems to be practically the only cationic indicator that is used for the titration and despite the limitations it seems to be preferred to anionic indicators in most commercial analytical laboratories in the U.K. and abroad. A wide choice of artionic indicators of the sulphonphthalein, sulphonic acid, and fluorescein classes is available, but the vast majority of workers appear to follow closely the technique of the early workers, Barr, Oliver, and Stubbins, and to use bromophenol blue. Standardisation At the end of the titration in cases (a) and (e) the whole of the indicator is in combination with surfactant, and a blank correction, which is constant, calcu!ab!e, reproducible and readily determined, must be applied. In cases (c) and (f) only part of the indicator remains combined with surfactant and the necessary correction depends upon the proportion involved, and this depends in turn upon the relative volumes of aqueous and chloroform layers. Experimental determinations of the blank or of a correction factor have been described by several writers, but for routine analysis it is sufficient to standardise the titrant under similar conditions to those of a determination, thus eliminating the correction. The standard substance in this approach must be of a similar composition to that being determined, which means in many cases that it must contain a mixture of isomers or of homologues as do the commercial materials, and its composition can therefore only be established by other analytical methods, principally by extraction or ion- exchange methods. For na!kyl sulphates, it is possible to prepare the pure matehals, and these can be used as standards in this field. Miscellaneous Factors After the choice of indicator, of end-point and of standard, that of the titrant is the most important. Cetyltrimethylammonium chloride and cety!- pyridinium bromide seem to be most frequently used for anionic detergents.

THE ANALYSIS OF SYNTHETIC DETERGENTS The stearyl compounds have the disadvantage of lower solubility, while the myristyl and shorter chain homologues may react incompletely. However, we have found that N-alkyl-N-benzyldimethylammonium chlorides give much sharper end-points than the other two groups of titrant. The stearyl com- pound again is of low solubility, but the lauryl, myristyl and cetyl com- pounds all have similar performances and there is little to choose among them except that the cetyl compound, of lowest solubility, is the most readily purified by crystallisation. Regarding the concentration of the titrant, many workers follow Epton in using titrations of about 10 ml with .004 or .0053{ titrant, but larger titrations with more dilute solutions, e.g. around 20 ml or .0013{ as used by Barr et al normally give more precise results, and are at least as accurate if due regard is paid to the blank. All titration procedures are equally suitable for determining anionic or cationic surfactants, the concentration of the other being known, and it is also immaterial whether the solution of unknown concentration is placed in the titration vessel or in the burette, although the former is usually the most convenient. The single phase titration The earliest work by Hartley and Runnicles used a single-phase titration of aniordc and cationic surfactants, but the end-point with bromophenol blue was not so clear as when the two-phase technique was applied. How- ever, with fluorescent indicators such as eosin •5 and dichlorotetraiodo- fluorescein •6 the end-point may be as sharp as with the two-phase method, and such indicators may repay further study. Their main disadvantage is that large amounts of inorganic salts, and moderate amounts of inactive organic compounds, tend to obscure the end-point. OTHER •/[ETHODS OF ASSAY Methods based on the formation of an amine salt with artionic surfactants, using an ordinary primary amine, have been used for a long time. One procedure •* uses toluidine as the amine, extracts the salt with carbon tetra- chloride, and determines the amine in the extract by addition of ethanol and titration with alkali. A similar procedure •a uses benzidine and separates the salt by filtration. The salt may be weighed before fitration to give an indication of the equivalent weight, which cannot be determined by any volumetric method. For determining cationic surfactants, several precipitants containing large artions may be used, and in a review by Chinnick and Lincoln 19, phosphotungstic acid is recommended. Non-ionic surface active agents with an ethanoxy chain of suitable length for detergency can also be precipitated