CHEMICAL ANALYSIS IN THE COSMETIC INDUSTRY 263 Ethylene diamine tetra-acetic acid, and its salts, may be determined by means of the colour produced with cobalt nitrate and hydrogen peroxide • 8, and this method is very satisfactory for small amounts of EDTA. It is quite sensitive, even in the presence of a large excess of alkyl sulphates. The details are as follows: Prepare a solution or dilution of the material, to contain approx. 0.2 per cent of ethylene diamine tetra-acetic acid, or its disodium salt. Pipette 5-10 ml of this solution into a test tube, add 0.5 ml glacial acetic acid and 1 ml of a 2 per cent cobalt nitrate solution and heat the tube in boiling water for exactly 15 minutes. Remove from the bath, cool to about 40 ø C and add 2 ml of 30 per cent (100 volumes) hydrogen peroxide. Heat for a further 5 minutes in boiling water, and cool to room temperature. Transfer to a 25 ml volumetric flask, make up to the mark with water and mix. Measure the optical density of the pink solution in a cm cell in a photoelectric absorptiometer, using Ilford Green No. 604 filters, taking care that no bubbles of oxygen adhere to the sides of the cell. A standard solution of disodium ethylene diamine tetra-acetate is used to prepare a calibration curve. Many cosmetic raw materials these days contain ethylene or propylene oxide condensed on to a reactive group, and most of them are very resistant to chemical attack, particularly if the linkage is of the ether type. Many methods have been proposed for their assay, but most of them are empirical and depend on a knowledge of the particular compound used. The identifica- tion of the type of condensate and the functional group, on which the alkylene oxide has been condensed, has been the subject of several papers and Rosen •6 has developed a method of detection and differentiation between polyoxy- propylene and polyoxyethylene, which is based on pyrolysis with syrupy phosphoric acid and a detection of the decomposition products by collecting them in a dilute solution of sodium nitroprusside containing diethanolamine. Several papers have been published which deal with the actual determina- tion of alkyloxy groups, and they all depend on a modification of the ordinary alkoxyl determination in which the compound is refluxed with hydriodic acid, and the alkyl iodide so formed is absorbed and titrated. In the case of polyglycol ethers, however, the reaction does not proceed in this simple way. Instead, the alkoxyl group is split into alkyl iodide and alkylene, and as there seems to be no strict proportionation between the amounts of each produced, each has to be absorbed and determined separately. The alkyl iodide is normally absorbed in alcoholic silver nitrate, and the excess silver deter- mined by titration with thiocyanate, while the alkylene is collected in a second trap containing a solution of bromine in acetic acid. A stream of inert gas is passed through the apparatus during the reflux period to sweep the products of the reaction out of the flask.

264 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Morgan •7 postulates the following equation as a summation of the reactions occurring in this process, where ethylene oxide is concerned: R.O-CH•.CH•.OH + (3 + x)HI = RI + x CH•.CH• I + (1 -- x) CH•:CH• + I• + 2H•O The ratio of ethyl iodide to ethylene varies with a number of reaction conditions---hence the need to determine both. A simple silver nitrate trap serves to absorb all the alkyl iodide, but in order to absorb all the ethylene a special absorber, consisting of a long tube containing a glass spiral to increase the time of contact between the gas and the absorbing solution, is necessary. A third trap containing potassium iodide solution is necessary to collect any bromine swept out by the gas stream. Carbon dioxide is generally used as carrier, but nitrogen works equally well. Figure I shows a slightly modified version of Morgan's apparatus. Trap E contains 10 ml 10 per cent potassium iodide solution, the spiral absorption tube D contains 15 ml of bromine solution (1 ml bromine, 300 ml glacial acetic acid and 5 g dry potassium bromide) and trap C contains 10 ml alcoholic silver nitrate solution (15 g silver nitrate are dissolved in 50 ml water and added to 400 ml absolute alcohol. Several drops of nitric acid are then added and the solution is standardized against 0.05N potassium or ammonium thiocyanate using Volhards method). Trap B contains a sus- pension of red phosphorus in water or, if sulphur is present, in 5 per cent cadmium sulphate. A weighed sample of the ethylene oxide derivative (0.05-0.12 g) is placed in flask A and 10 ml of constant boiling hydriodic acid (S.G. 1.70), freshly distilled over red phosphorus is added. The flask is connected to the apparatus and a slow stream of carbon dioxide is passed while the flask is gradually brought to a temperature of 140ø-145 ø C by means of an oil or glycerin bath. The flask is kept at this temperature for a minimum of one hour. Five minutes before the end of the heating, the silver nitrate trap is heated to 500-60 ø C, by means of a hot water bath, to drive out any dissolved ethylene. Tubes D & C are then disconnected in that order, the carbon dioxide is disconnected and the heat removed from flask A. The tube D is then connected by its lower adaptor to a 500 ml flask, containing 10 ml 10 per cent potassium iodide, and 150 ml water. The potassium iodide tube E is removed and the contents of the side arm rinsed into it. The bromine solution is run into the titration flask through the stop-cock and the tube rinsed out with water. The contents of the tube E are washed into the titration flask, which is then stoppered and allowed to stand 5 minutes. 5 ml of 10 per cent sulphuric acid are