406 MIDDLETON AND STUCKEY THE DETERMINATION [Vol. 75 The Determination of the Purity of Propylene Glycol BY G. MIDDLETON AND R. E. STUCKEY SYNOPSIS-A description is given of a test using the critical solution temperature of propylene glycol and ether. With this test the presence of 0.1 per cent. of ethylene glycol, 0.1 per cent. of di-propylene glycol or similar amounts of water and ethyl alcohol in propylene glycol can be deter- mined. IN spite of the large number of organic solvents that are at present available, only a few can be administered in ternally, while the number available for parenteral administration is extremely restricted. The addition of propylene glycol to the last group is therefore of special interest, and a high standard of purity is necessary for material required for injection purposes.A number of specifications1 p2 p3 have been published for propylene glycol for pharmaceutical purposes. These specifications include upper and lower limits for physical constants, together with tests for general impurities such as sulphates, arsenic and lead; the American National Formulary VIII has an assay using periodic acid. An examination of these specifications shows them to be, in general, deficient with regard to tests for impurities of a glycol character. Ethylene glycol is an impurity that is not likely to be found to any great extent in commercial propylene glycol, more likely impurities being di- and tri-propylene glycols; at the same time, in view of possible confusion between various glycols, a test that would limit the presence of glycols other than propylene glycol appears to be desirable. The refractive indices of propylene glycol and ethylene glycol are practically the same, and although there is an appreciable difference in density between the two compounds, this can hardly be relied upon to show the presence of small amounts, say a few per cent., of ethylene glycol in propylene glycol.The method given by Warskowsky and Elvig4 for the determina- tion of ethylene glycol and propylene glycol in admixture, being based on periodate oxidation followed by estimation of the relative proportions of acetaldehyde and formaldehyde produced, is equally unsuitable for this purpose. It is, similarly, difficult to determine the presence in propylene glycol of condensed compounds, such as di-propylene glycol, which can be formed from propylene glycol by the elimination of water and the formation of an ether linkage.A search for a chemical test does not appear to be a very promising line of attack. Direct physical measurements, such as refractive index and density, are also, as has been indicated, of little value in solving the problem. Attempts were therefore made, in experiments on mixtures of known amounts of ethylene glycol with propylene glycol, to concentrate the ethylene glycol by extraction with ether. Propylene glycol is about eight times more soluble in ether than is ethylene glycol, but preliminary trial of this method showed considerable difficulties. It did, however, indicate that a more sensitive test could probably be based on the critical solution temperature with a solvent, and ethyl ether was found to be suitable. In using such a test it is necessary to obtain a sample of pure propylene glycol, a matter of some difficulty owing to the ease with which the substance absorbs water.As a suitable criterion of purity, the constancy of the critical. solution temperature of a series of successive fractions obtained by distillation in an efficient fractionating apparatus was used. For the determination of the critical solution temperature, a method based on that of the Institute of Petroleum5 for the determination of the aniline point was used. Owing to the necessity of guarding against er:rors due to the volatility of the ether and to the hygroscopic nature of the components, the test was carried out in a closed vessel and, to avoid the need for stirring, with a falling temperature.A clear mixture of propylene glycol and ether, when cooled slowly, shows at first a faint turbidity, which gradually increases as the temperature falls, so that it is not possible to state a definite point at which turbidity commences. On continuing the cooling there is, at a particular temperature, a sudden increase in turbidity, followed immediately by a “streakiness” due to convection currents and separation into two phases. This point, at which there is a marked increase in the turbidity, was found to be the most easi1.y reproducible, and it was taken as the critical solution temperature, or what we have termeld the “ether-point.”August, 19501 OF THE PURITY OF PROPYLENE GLYCOL 407 TECHNIQUE OF THE TEST- The ether used must be anhydrous and free from alcohol; it can be prepared by distilling pure Analytical Reagent grade ether over sodium; the distillate may, if required, be kept for a short time over sodium.The ether and propylene glycol were mixed in a 10-ml. stoppered measuring cylinder, the stopper of which was specially ground with fine corundum. Propylene 10 PERCENTAGE OF GLYCOL IN THE PROPYLENE GLYCOL-ETHER MIXTURE Fig. 1. Temperatures of separation of propylene glycol -ether mixtures. Curve A, pure propylene glycol and ether; curve B, propylene glycol containing 1 per cent. of water and ether. glycol was first added quickly from a wide-bore bulb tube (an inverted pipette is satisfactory), care being taken to see that none of the glycol was allowed to come into contact with the ground portion of the neck of the cylinder. When the propylene glycol had run down in the cylinder, the volume was noted and the requisite quantity of ether was added.The ether - propylene glycol mixture was then stoppered, warrned and shaken to obtain a clear solution. The cylinder, suitably weighted, was placed in a rectangular glass trough forming a water-bath and containing a stirrer, and cold water was allowed to drip in from a tap- funnel to give the desired rate of fall of temperature. The test was carried out with top lighting, and a dead black background with a horizontal white line 0-5cm. in width was placed so that the cylinder cut across the line. The "ether-point" was taken as that point at which the white line was indistinguishable through the cylinder.This point can be repro- duced very accurately in successive experiments. The results of experiments with pure propylene glycol and ether in different proportions are shown in Fig. 1 (curve A). It will be seen that the two liquids are completely miscible above 21.6" C., but that at lower temperatures two solutions of different concentrations are in equilibrium with each other. The summit of the curve is flattened, so that at concentra- tions of propylene glycol between 30 and 35 per cent. the temperature of separation is only slightly altered by variation in concentration. For this reason a concentration of propylene glycol of 33-3 per cent. (one volume of propylene glycol plus two volumes of ether), practically a t the peak of the curve, was selected as a convenient standard concentration.408 MIDDLETON AND STUCKEY : THE DETERMINATION [Vol.75 Pzcri$cation of the propylene &cod-After preliminary work had shown the value of the ether-point in classifying samples of the glycol, a quantity of 1 litre of purified glycol was fractionated through an electrically heated column, 4 feet long, containing glass helices, with a reflux ratio head. Fractions 1D to 9, showing no significant difference in ether-point, were mixed together, the resulting product being taken as pure propylene glycol of ether-point 21-62" C. TABLE I Details of the fractionation are given in Table I. Fraction 1A 1B 1c 1D 2 3 4 5 6 7 8 9 10 Residue FRACTIONATION OF PF.OPYLENE GLYCOL Boiling-point, O c.23-26 90-95 to 96 to 97 97 97 97-8 97 97 97 97 97 97 - Pressure of distillation, mm. of Hg. 24 24 22 24 24 24 24 23 23 23 23 23 23 - Approx. vol. of fraction, ml. 2 25 25 40 80 100 140 100 100 100 140 100 40 20 "Ether-point" of fraction, O c. - 30 21.95 21-60 21.60 21-67 21-62 21.65 21.60 2 1-60 21-63 21-60 21-76 I The physical constants of pure propylene glycol were determined on the sample prepared as described above and the following values were obtained- Weight per ml. at 20" C, = 1.0374 g. Refractive index at 20" (3. = 1.4330 The effect of 1 per cent. of water added to the carefully purified propylene glycol is shown Fig. 2 shows the alteration in the ether-point resulting from the' addition of water and Within the range of concent.rations examined, the graphs are practically The alteration in ether-point produced by the presence of 0.1 per cent.of in Fig, 1, curve B, and it will be noted that this curve remains parallel to the original one. other substances. straight lines. various impurities is shown in Table 11. TABLE: Ir EFFECT OF 1 MPURITIES Impurity Alteration of "ether-point" per 0.1 per cent., Water . . .. .. .. .. + 0-35 Ethylene glycol . . .. .. .. + 0.15 Di-propylene glycol . . .. .. - 0.055 Ethanol .. . . .. .. .. - 0.25 APPLICATION OF THE TEST-DEHYDRATION OF THE SAMPLE- The ether-point cannot be used directly a s a measure of purity as it is greatly affected by the moisture that is usually present in samples of propylene glycol. In passing, however, it may be mentioned that, if the absence of other substances can be assumed, the ether-point can be used accurately to determine the percentage of moisture present provided that the amount is small.When 3 per cent. of water is present, the critical solution temperature approaches the boiling-point of the ether. A method was, therefore, required for the dehydration of samples of propylene glycol. This was first attempted by the direct addition of suitable substances to the propylene glycol and to the propylene glycol - ether mixture. None of the agents tried, viz., calcium oxide, barium oxide, anhydrone, silica gel, sodium sulphate, magnesium sulphate, sodium sulphite, copper sulphate and sodium metal, was of any value; either they were ineffective or they dissolved in the liquid and altered the equilibrium.August, 19501 OF THE PURITY OF PROPYLENE GLYCOL 409 Both ethylene glycol and di-propylene glycol have a boiling-point higher than that of propylene glycol, and although the difference in boiling-point between ethylene and propylene glycols is only about 10” C., this should be sufficient to ensure that the concentration of the forrner would increase as part of the mixture is removed by distillation.This was confirmed in experiments in which samples of propylene glycol, to which were added small quantities 0-5 1-5 2.0 PERCENTAGE OF IMPURITIES IN THE PROPYLENE GLYCOL Fig. 2. The effect of impurities on the temperature of separation of propylene glycol -ether mixtures. of the other glycols and of water, were distilled in a simple distillation apparatus.The results given in Table I showed that the water present in an ordinary commercial sample was removed in the first 5 per cent. of distillate. The results shown in Table I1 were obtained by distillation at atmospheric pressure in an ordinary distillation apparatus with a splash head. TABLE I11 DETECTION OF IMPURITIES IN PROPYLENE GLYCOL Impurity in residue, Residual “Ether-point” calculated from Added impurity Added water, quantity, of residue, “ether-point,” YO O c. % nil . . .. .. 1.0 90 21.6 0 Y O Ethylene glycol 1.0 per cent. . . .. 1.0 0.2 99 . . .. 1.0 0.1 9 ) . . .. 1.0 Di-propylene glycol 0-5 per cent. . . .. 1.0 0.1 w . . .. 1.0 90 90 10 23.3 21.9 22.05 1.2 0.22 0.32 90 21.2 0.66 10 21-15 0.75 The results given in Table 111 show that it is possible by this method to detect less than The percentage 0.1 per cent.of ethylene glycol or of di-propylene glycol in propylene glycol.410 MITCHELL: THE DETERMINATION OF THE MARC CONTENT [Vol. 75 of the latter was found gradually to increase in the residue during distillation, although the increase was slight. The sensitivity of the test can be increased by taking the ether-point of the residual 10 per cent. rather than of the residual 90 per cent. (Le., with only 10 per cent. removed by distillation) in which only slight concentration of the impurity had been achieved. In view of the ease with which propylene glycol polymerises in the presence of a trace of acid or alkali,B it was thought advisable to check the possibility of such a reaction occurring on prolonged boiling. Pure propylene glycol was boiled under reflux for 4 hours, 10 per cent. was distilled off in order to remove any water, and the ether-point was determined on the residue.It was found to be 21.58”, which indicates at the most such a slight degree of polymerisation as to be negligible. In order to apply the method to a sample of propylene glycol about which little is known, the glycol should be distilled and the ether-points of fractions of the distillate determined. Water will be removed in the first fractions and the trend of the ether-point of subsequent fractions and of the residue will give an indication of the purity of the glycol under test. Although the only impurities studied were ethylene glycol, di-propylene glycol, ethyl alcohol and water, the test can obviously be extended to other miscible impurities. It is possible that suitable amounts of impurities, e.g., ethyl alcohol and ethylene glycol, would mutually compensate each other and would produce an ether-point of pure propylene glycol; distillation, however, followed by a determination of the ether-point of the fractions, would reveal such impurities. The authors wish to thank the Directors of the British Drug Houses Ltd., for permission to publish this paper. REFERENCES 1. 2. 3. 4. 5. 6. THE BRITISH DRUG HOUSES LTD. GRAHAM STREET, CITY ROAD “National Formulary VIII,” 1947, p. 420. “New and Non-Official Remedies,” 1945, p. 488. “British Pharmaceutical Codex 1949,” p. 375. Warskowsky, B., and Elving, P. T., Ind. Eng. Chem., Anal. Ed., 1946, 18, 253. Institute of Petroleum, “Standard Methods for Testing Petroleum and its Products,” Aniline Levene, P. A., and Walti, A., J . Biol. Chem., 1927, 75, 325. Point (1.P.-2/47), Method C. LONDON, N.l March, 1950