104 THE ANALYST. ON THE DETERMINATION OF GLYCEROL IN CRUDE GLYCERINES. BY J. LEWHOWITSCH, Ph.D. (Read at the Meeting, February 4, 1903.) THE two methods recognised as the best for estimating quantitatively glycerol in crude glycerine are the acetin method and the bichromate method, the latter especially in the form given to it by Hehner.‘;’ Whilst, personally, I favoured the former method, I did not dispute the accuracy of the latter. I may be allowed to quote what I wrote several years ago in my ‘‘ Chemical Analysis,” p. 808 : “Hehner has shown by a number of comparative experiments, using both the acetin and the bichromate processes, that the results agree very well. This has been borne out by a large number of experiments made by the writer.” In the course of the last few years, however, a number of crude glycerines came into my hands in which the agreement between the two methods was not of a satisfactory kind throughout. It was soon noticed that discrepancies between the two methods occurred in the case of those glycerines which contained somewhat considerable quantities of organic impurities.I give in the following table a number of analyses in support of my contention : * Joum. SOC. Chem. Ind., 1889, 5.THE ANALYST. 105 Acetin Method. TABLE I. Comparison of the Acetin and Bichrornate Methods for Determining the Percentages of Glycerol in CrzLde Glycerines. 1896-1897. Bichromate Method (Copper Sulphate). NO. Description. 1 2 3 4 5 6 7 8 9 10 Soap lye crude Chemically pure Soap lye crude Candlemaker’s crude Soap lye crude 7 7 7 7 77 7 7 7 9 7 7 7 ’ 7 7 ‘ 7 7 7 ’ 7 7 ) 9 ) 7 7 -77 86.66 86.15; 86.06 75-02 99.04; 99.17 80.51 ; 80.66 78-93 85.94 67-76 79-84 83.05 ; 83-07 86.45 86.01 ; 86.34 75-38; 75.10 101.00 ; 101.90 80.37; 80.04 79.41 85.43 67.67 79.23 83.99; 84-65 1900.No. I Description. Bichromate Method (Copper Sulphate). Acetin Method. Candlemaker’s crude Soap lye crude 89.91 ; 90.75 ; 90.43 78.48; 78.65; 78.89 80.46 ; 80.82 ; 80.28 ; 80.61 76.56; 77-45 90.69; 90.74 86.26; 86-16; 86.22 8 1 ; 44; 81.19 94.02; 94-12 93.68 ; 94.21 ; 94.71 ; 94.90 81-42; 81-60; 81.23 82.56 ; 83-03 ; 83.23 ; 83.89 79.36; 79.01 92.70; 93-08 89-70; 89-22 82.51; 82.61 97-21; 97.94 9 9 7’ 7 ’ 7 7 ’ 9 Y P Candlemaker’s crude 7 9 7 7 I Soap lye crude 8 Candlemaker’s crude 1901 - 1902. Bichromate Method (Copper Sulphate).Bichroina te Method (Silver Carbonate and Lead Subacetate). NO. Description. Acetin Method. 1 2 3 4 5 6 9 7 7 8 9 10 Candlemaker’s crude Soap lye crude 7 7 9 ’ 2 7 7 7 ? 7 7 ’ 7 9 ’ 7 7’ 7 7 ’ ? 7 7 Same crude, concen- Soap lye crude Candlemaker’s crude Soap lye crude Candlemaker’s crude trated further 83.51 ; 83.68 81 -42 79-13 75.20 72.98 72.01 74.73 77-31 83.1 ; 82-98 76.53; 76.61 85.23; 85-42 89-44; 89.73 82-21 8152; 81.94 7 8.79 76.04 77.44 78.35 77.96; 78-60 83.9; 83.56; 83.52 78.17 ; 78-24 86.72; 86.41 78.30; 78-33; 78.69 79.25 ; 79-53 -106 THE ANALYST. Where agreement occurs the glycerine was fairly pure ; in the other instances this was not the case. The fact that I did obtain concordant results for a number of years is seen by a glance at the first series of figures, and this fact disposes at once of a possible objection that the bichromate method has not been carried out as directed.I may say that I found it more convenient to use copper sulphate and caustic potash for purifying the crude glycerine than silver oxide and lead subacetate, or, preferably (acting on a private communication from Mr. Hehner), silver carbonate and lead subacetate. In order to invalidate any objection on that score, some of the samples were treated side by side with copper sulphate, and with silver carbonate and lead subacetate. Curiously enough, the results obtained with the help of the latter reagents are a little higher, a fact which Richardson and Jaff6 also have noticed (Journ. Xoc. Chem. Ind., 1898, 332).I may point out that this is not due to an error caused by the larger volume of the precipitate obtained with silver carbonate and lead subacetate. Of course, if the volume of the precipitate should be somewhat considerable, it is imperative to filter off and fill up the filtrate to a known volume. The duplicate tests, I may distinctly mention, were all done with separately weighed quantities of the samples. I think that the numbers given in the table warrant me in uttering a note of warning against the exclusive use of the bichromate method. It is true Hehner advises using both methods in the examination of a sample, and taking the mean of the two results. As to this suggestion, I remarked some years ago that it entails too much work in commercial analysis, and that it is therefore preferable to make two tests by the acetin method.If Hehner's suggestion be adopted, such large discrepancies as I have shown to occur will no doubt lead the analyst to entertain serious misgivings as to the accuracy of one of the two methods, or possibly of both. I for one adhere to the acetin method as the more accurate one, and if attention is paid to all the details, very concordant results indeed are obtained in duplicate tests. Unfortunately, the acetin method lends itself only to fairly strong crude glycerines, containing, say, above 60 per cent. glycerol, and is therefore not applicable to the examination of soap lyes in their original state. I n the latter case the bichromate method seems to be the more convenient one, but here also I find that the bichromate method leads to too high results if the lyes are very impure.A correct method of testing the glycerine is, as I have stated elsewhere, to take 1,000 C.C. or 1,000 grammes, purify the lye, and concentrate down so as to prepare a crude glycerine, in which the glycerol may be estimated by the acetin method. In my own practice, based on a number of checked experiments, I rapidly heat the ether-alcoholic solution of the crude glycerine, after evaporating off on the water- bath, to 150" C., weigh, burn off the glycerine, and weigh again. The difference represents the contents of glycerol. The following table (No. 1I.j gives a number of lyes tested by both methods. I t brings out the same point which I have illustrated by the first table-namely, that agreement is only obtained in the case of pure lyes, but that the bichromate method yields too high results if the lyes are impure.THE ANALYST.107 No. TABLE 11. Comparison of the Acetin and Bichromate Methods for Determining the Percentages of Glycerol in Soap Lyes. Specific Gravity. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 1.120 1.114 1.216 1-1025 1.0975 1.1050 1.0925 1.1025 1 *09 5 1.2025 1.0925 1.09 1.085 1-22 - - Free Alkali, a.s Na,O, per Cent. 2-39 0.58 0.927 0.452 0.678 0.809 1.6 0.863 1.933 0.51 0.768 0-657 0.49 - - - Glycerol, per Cent. Bichroniate Method (Copper Sulphate). 1,000 C.C. Method. 6.62 ; 6.80 8.48; 8.43 5.89; 5.81 7.45; 7.33 5.90; 6.09 5.34 ; 5.43 5.92 ; 5-88 5.65 6.92; 6.93 9.68; 9.67 6-95; 7.07 6.78; 6.42 5.42; 5.45 4.59 7.66; 7-69 12.5 ; 12.6 5.86 7.36 5-70 6.69 5.90 5.25 6-10 5.64 6.95 9.75 6.96 6.62 5.70 11-57 3.57 7.55 The bichromate method suffers from the inherent fault that everything which is oxidizable to carbonic acid is reckoned as glycerine, whereas the acetin process accounts only for those substances which contain hydroxyl groups.The chief impurities are, of course, lower fatty acids, which in the bichromate method are oxidized. In the acetin method, impurities of a, non-alcoholic nature do not interfere with the accuracy of the results. But even if an impure crude glycerine did contain trimethyleneglycol, the acetin method would rather indicate lower results than the true one, whereas the bichromate method would yield too high a number. Chau*meil (Bull. Soc.Chim., xxvii., 629) prefers iodic acid as an oxidizing agent to bichromate, and proposes a method of determining glycerol, which is based on the reaction expressed by the equation : 5C,H,O, + 7I,O, = 15C0, + 20H20 + 71,. I have not examined this method, as it has the same inherent fault as the bichromate test-namely, that everything oxidizable counts as glycerine. It appeared, therefore, unnecessary to carry out any experiments, all the less so as Bernard (Pharm. Central-Halle, xliii., 541) has shown that the method is unreliable even in the case of distilled glycerine. Gailhat’s oxidation method, using permanganate in presence of manganese sulphate, need only be noticed in passing. Verley and Bolsing proposed (Berichte, 1901, 3354) as a general method for the quantitative estimation oi alcohols, and amongst them of glycerol, a modification of108 THE ANALYST.the acetin method, consisting in the addition of pyridine to the acetic anhydride, their contention being that the pyridine accelerates the reaction so considerably that the determination can be carried out by warming on the water-bath for a quarter of an hour without even requiring a reflux condenser. I have tested this method on crude glycerine and also on chemically pure glycerine, but the results obtained were utterly unsatisfactory. The ideal method for the determination of glycerol would, of course, be one which would enable us to isolate the glycerol either as such or in the form of a derivative; therefore the method proposed by Zeisel and Fanto (Zeit. f.d. Zand- wirthsch. Versuchswesen in Oest., 1902) naturally attracted my attention, The method is based on the reaction expressed by the folIowing equation : The glycerol is converted by means of hydriodic acid of 1-7 and even 1.9 into isopropyl iodide, and the latter distilled over into an alcoholic solution of silver nitrate. The iodine, which is formed s t the same time, is retained by red phosphorus. Silver iodide is precipitated and weighed as such, and each molecule of AgI calculated to 1 molecule of glycerol. I have made two series of experiments on crudeglycerine, but the results have been so much below the truth that I do not feel encouraged to continue my experiments. Besides, the method is so cumbersome when compared with those in vogue that there is little inducement for adopting it in the estimation of crude glycerine.Messrs. C. D. Robertshaw and G. Warburton have assisted me in the experiments enumerated above. C3H5(OH)3 + 5HI = C,H$ + 3H,O + 21,. The test is carried out in the apparatus shown on the lecture table. DISCUSSION. Mr. HEHNER said that it had long since been pointed out that, while the results of the bichromate method had a tendency to be too high, the tendency of those of the acetin process was in the reverse direction. The hydrolysis of triacetin was very readily brought about, and if, in the neutralization of the free fatty acid in the acetin process, any alkaline reaction at a11 were allowed to occur, the result would be a partial hydrolysis of the triacetin, and as a consequence the final result would be too low.Evidently, however, some change (which no doubt Dr. Lewkowitsch would be able to point out) had taken place in the actual conditions of manufacture since the time when he (Mr. Hehner) had compared the two methods and had found them to give substantially agreeing results with both soap and candle glycerines. Possibly, as a result of more rapid boiling at higher temperatures, there was now a tendency towards the formation of condensed glycerines or polyglycerines, which, while being oxidized by bichromate almost in the same way as glycerine, would certainly not be acted upon to a corresponding extent by acetic anhydride. He had investigated the question of the possible oxidation by bichromate of those lower fatty acids which were not precipitated by lead acetate or silver carbonate, and had come to the con- clusion that, with proper precautions, no such oxidation occurred.Certainly there was no oxidation in the case of butyric acid, provided there was not too much sulphuric acid present.THE ANALYST. 109 Dr. LEWKOWITSCH said that undoubtedly the acetin process was a somewhat difficult one to carry out accurately, but his own work included a great number of glycerine analyses, so that there was full opportunity in his laboratory for attention to even the smallest details, and consequently, in such a matter as the avoidance of an alkaline reaction in neutralization, the proper limit was not overstepped. The fact was that since 1897 the prices of raw materials had risen so greatly that both Roap-makers and candle-makers had been obliged to fall back upon the cheapest materials they could get.Soap-makers, especially the smaller ones, could not afford to buy high-priced tallow, oils, or fats, and had to use bone-grease, skin-grease, and like materials, containing, as a matter of course, large amounts of free fatty acids, and, what was especially important in the present case, of lower fatty acids, which were brought down into the crude glycerine. Theoretically, the lower fatty acids ought to boil away, but they did not, and were to some extent retained in the impure crude glycerine. They were then, in the course of the analysis, oxidized by bichromate, especially on heating, although the pure fatty acids might not be oxidized so readily. In candle-making, in consequence of this forced utilization of materials of low quality, acid saponification methods had to be resorted to, with the result that, starting with impure materials, the crude glycerine was likewise impure.I n all the cases in which large discrepancies were shown between the two methods, the glycerines contained large proportions of ash. In crude candle glycerine, for instance, the ash ought not to exceed about 0.5 or 0.6 per cent., whereas in these samples it amounted to nearly 2 per cent., consisting mainly of sulphate of lime or some other sulphate. Polyglycerines would, of course, not be accounted for in the acetin process ; but, as a matter of fact, polyglycerines were only to be found in distilled glycerines, or in crude glycerines which had been subjected to distillation, and not in such ordinary crude glycerines to which the present paper was confined.The tempera- ture at which such glycerines were concentrated rarely exceeded 90" C., which was insufficiently high for polymerization to take place. Mr. HEHNER suggested that the relative accuracy of the two methods might be to some extent tested by ascertaining the amount of specific gravity due to glycerine, after deducting that due to ash. If there were present a sufficient proportion of, say, caproic acid to appreciably affect the result yielded by the bichromate method, the specific gravity of the sample would be lower than that indicated by the glycerine and ash percentages. Dr. LEWKOWITSCH said that, as a matter of fact, the influence of the glycerine on the specific gravity was slight as compared with the influence exercised by the ash, and both were liable to variation. In the concentration of glycerine counter- acting influences were exercised, on the one hand, by the salt dropping out of solution, and on the other by increase in the proportion of glycerine. At every stage of the process there existed an equilibrium between the water, glycerine, and salt ; and if that equilibrium were disturbed by the dropping out of some of the salt, the specific gravity could only be raised by increasing the amount of glycerine and decreasing the amount of water.