MONO-ESTER CONTENT OF PROPYLENE GLYCOL MONOSTEARATE will give the entire range from 0--100% ester content, with a mixture of pure mono and pure distearate. Essentially, any mixture of the two esters will give a quantity of propylene glycol between these limits from which the original monoester content is calculated. Thus, estimations within the range of less than 9•o correspond to a range of 0--100% of calculated percentage of monoester content. If, however, instead of dealing with the propylene glycol content, an attempt is made to base the calculation on saponification value, then there is a little improvement. The saponification value of the pure diester is 193.4 and of the pure monoester 171. The complete range of values that can be assumed by monoester content is represented by a range of 22 in saponification value. Consequently, an error in assessment of saponification value is multiplied by a factor of approximately 5 on the calculated monoester content. When we were first confronted with the necessity to determine the monoester content of propylene glycol monostearate, the T.G.A. method was tried, and we found startling variations from one determination to another. Variations were often of the order of 30•o. In fact, in quite a number of cases, we found the result could be calculated to a negative value. Naturally should the combined propylene glycol estimation be calculated to less than 14'39•o then the value of propylene glycol monostearate must be negative. Similarly, should the tota! propylene glycol content be esti- mated in excess of 23% then the calculated propylene glycol monostearate content will be over 100•o. We noted that the titration end-point in this determination was unsatis- factory due to the slow liberation of iodine. Following the T.G.A. method, acetic acid is used and when potash has been used to saponify in the early stages, you have a buffered system in which the free acetic acid occurs in the presence of a considerable quantity of potassium acetate. Consequently, the pH will be rather on the high side for the optimum condition for the liberation of iodine. Adding hydrochloric acid immediately before titration produced a substantial improvement but even so, the method still gave unsatisfactory results. In particular, we noted variations in the blanks. Obviously, small variations in the blanks are a serious matter as the difference between the sample titration and that of the blank is usually about 3.3 ml, so that an error of 0.1 cc in the blank immediately gives a 3% error before multiplying by a factor of approximately 12. It seemed possible that there was some reaction between the periodic acid and the alcohol used for the alcoholic potash solution. We compared blanks obtained with alcoholic potash with those obtained with aqueous alkali. Reproducible results were obtained in the aqueous phase but there was considerable variation amount- ing sometimes to 1.5 cc when titrating with blanks in the alcoholic medium.

20 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS We modified the T.G.A. method by using aqueous caustic soda instead of alcoholic potash for the hydrolysis of the ester, and found some improve- ment but the results were still nothing like as reproducible as we would have liked them to be. At that stage, we considered investigating the problem with a view to obtaining a more reliable method of estimation. A moment's consideration of the arithmetic of the problem suggests that the estimation of hydroxyl value is attractive. The saponification value of pure propylene glycol monostearate is 171 but on acetylation the saponi- fication value of the product increased to 303. In the case of the diester, the saponification value is 193.4. The saponification values may, therefore, vary within a much greater range than before. That is to say, 109.6. Small errors in this determination do not seriously influence the calculated result of the monoester content. In order to carry out the estimation, it is first necessary to remove the free glycol from the product. We have tried several different methods and have found that the most satisfactory is that described above, i.e., the sample is dissolved in diethyl ether and the free glycol is removed by extraction with water. Having thus prepared the dry sample free from propylene glycol, we believe the most accurate method of determining the hydroxyl value is to acetylate using a large excess of acetic anhydride, employing the sodium acetate dehydrar•t. This is followed by the sub- sequent hydrolysis of excess acetic anhydride and then washed free of impurities and dried. The saponification value is determined in the usual manner, using alcoholic potash under reflux for 2 hours. As previously stated, we believe that this is the most accurate assessment of monoester content which is at present available. We have examined numerous other esters and esterfication methods but in one respect or another, they have proved less reliable. The method is obviously a rather lengthy procedure and we considered that it might be possible to reduce the time required for analysis by examining the liquid/solid chromatography method described above. We would, however, draw your attention to some points which require careful attention. The moisture content of the silica gel is critical. We find satisfactory results are obtained if the gel is dried in an oven at 105 ø for several hours. On removing from the oven for a short time, it should not cloud a test tube filled with cold water held over the surface. When no clouding appears, virtually no water vapour is being given off from the gel. This does not mean that the silica gel is anhydrous. If you heat it to redness a loss of 2-3% of water will occur. Alternatively, heating the gel at 105 ø to constant weight is equally satisfactory. The dried silica gel is introduced into benzene and made into thick slurry. If the silica gel is still hot then there is very little trouble due to occlusion of air. The