Determining Glycerol in Crude Glycerin and in Soap Lyes WILLIAM J. GOVAN, JR., Pacific Soap Company, Ltd., San Diego, Calif.
A
again noticeable. The bottles are then brought to room temperature in a desiccator over concentrated sulfuric acid, s t o p pered, and weighed to the fourth decimal place. The drying with the addition of 5 ml. of ether is repeated until the loss in weight is less than 0.002 gram. Usually, only one extra drying is necessary. Thq;,final weight, minus the tare, may be termed "residue at 100' C.
though simple and rapid, is lacking in accuracy. Distillation methods are usually lengthy and require special apparatus. Experimental data (6) showed that the loss of glycerol from boiling aqueous glycerol solutions was negligible from a quantitative standpoint] and that the loss of glycerol from anhydrous 0.5-gram samples in small glass beakers at 100" C. was on the order of 2 mg. per hour. This latter fact indicated that a margin of safety could be expected for a quantitative method of removing water from aqueous glycerol solutions by evaporation. An accurate, quick method (4) had already been found for removing both water and glycerol from an aqueous glycerol sample. It was reasoned that, if water alone could be evaporated from an equal aliquot portion of the same sample, the percentage of glycerol could be calculated from the difference in weight of the two residues. The evaporation of water from a n aqueous glycerol solution at 100" C. would be slow work a t best. Therefore, t o speed u p the evaporation, many combinations of low-boiling solvents such as ethyl alcohol, ethyl ether, methyl alcohol, and benzene were tried. Methyl alcohol and ethyl ether were finally adopted as the most satisfactory.
Removal of Both Water and Glycerol from Aliquot Portion
SIMPLE, rapid method for determining glycerol in crude glycerin and soap lyes has been needed for a long time, The international acetin method (S), though accurate, is lengthy and is applicable only to samples containing a maximum of 60 per cent water. The dichromate method (1) requires a preliminary purification and filtration, and its results tend to be high. The Bertram-Rutgers method (2)]
For this determination an infrared drying oven is used, either as described in a previous pa r (4) or in the following manner: The top and bottom of an orgnary tin can, 15 cm. (6 inches) in diameter and 17.5 cm. (7 inches) high, are cleanly cut out, and four inverted V notches, about 2.5 cm. (1 inch) in height, are cut from the base. The can is mounted on its notched end upon a porcelain or asbestos surface. A 250-watt infrared reflectordrying lamp (General Electric R-40) is supported directly over the top opening of the tin can. A thermometer is inserted into one of the notches, so that its bulb rests directly beneath the center of the drying lamp. Two 5-ml. aliquot portions of the sample are pipetted into tared, shallow, flat-bottomed eva rating dishes of about 70-mm. diameter. The tin can is set a s i g a n d the evaporating dishes are placed adjacent and on each side of the thermometer bulb. The tin can is replaced and current to the lamp is switched on. Preliminary evaporation is done with the lamp a t a distance of about 20 cm. (8 inches) from the dishes. When the residue is almost dry, the lamp is raised slightly to prevent spattering of salt crystals. The temperature up to this point is disregarded. When the fumes of glycerol are scarcely to be seen coming off, the reading of the thermometer ia brought to 160' C. by adjusting the height of the lamp. This temperature is maintained for 30 minutes, after which the evaporating dishes are removed, cooled to room temperature in a desiccator over concentrated sulfuric acid, and weighed rapidly to the fourth decimal place. This weight minus the tare may be termed "residue a t 160" C." The over-all drying time rarely exceeds 90 minutes. CALCULATION.
Preparation of Sample SOAP LYE CRUDE GLYCERINAND SAPONIFICATION CRUDE GLYCERIN, Eight to 10 =t0.001 grams are weighed into a tared 100-ml. measuring flask and about 50 ml. of water are added. The contents are well mixed, adjusted to the phenolphthalein end point with dilute hydrochloric acid or sodium hydroxide, made up to the 100-ml. mark with more water, and thoroughly mixed. SOAPLYE. Thirty-five to 50 grams of lye are weighed into a tared 100-ml. measuring flask, the smaller sample for lyes of high salt content. The sample is adjusted t o the phenolphthalein end point with hydrochloric acid or sodium hydroxide. Oneto 2 ml. of a 10 er cent aqueous solution of a wetting agent (Aerosol OT) are ad8ed, and the sample is made up to the 100-ml. mark with water and thoroughly mixed. The use of a wetting agent tends to reduce spattering.
(Residue at 100' C. - residue at 160' C.) X 100 = % of glycerol 1/20 original sample Samples 1 and 2 are American Oil Chemists Society standard sample crude glycerin (February 1, 1929). The figure 83.33 per cent is the "true glycerol" content and represents the averaged findings of the laboratories of several soap companies. Samples 3 to 9, inclusive, are crude glycerins, both saponification and soap-lye, obtained from five different soap companies. The figures in the first column represent
Removal of Water from Aliquot Portions Two 5-ml. aliquot portions of the sample are pipetted into tared 60-ml. Erlenmeyer weighing bottles (the bottles have a diameter of 50 mm. a t the base, are 75 mm. hi h, and take a 24/12 F stopper), and 5 ml. of methanol are afded. The unstoppered bottles are placed in a natural convection drying oven maintained at 100" * 2" C. The oven must have ample top and bottom vents for quick heat transfer, and its make and break contact points must be located below the samples for safety. Narrow strips of 0.6-cm. (0.25-inch) asbestos millboard placed on the metal shelf of the oven are effective in preventing spattering. The Sargent electric drying oven 5-63995 has been found satisfactory. When the residue in the bottles is reduced to a thin sirupy consistency by evaporation and the characteristic sweet odor obtained when glycerol is heated under these conditions is faintly perceptible, the bottles are removed from the oven. (The time varies between 65 and 80 minutes.) Five milliliters of ethyl ether are then added and swirled in the bottom of the bottles for 2 minutes. The bottles are put back into the oven and dried a t 100' C. for about 25 minutes, until the acrid odor of ether is no longer perceptible and the odor associated with glycerol vapor is
TABLE I. EXPERIMENTAL RESULTS Sample No. 1
2
3
4 5
6 7 8 9
10 11
260
Net Glycerol Acetin method
Evaporation method
%
%
Difference %
83.33 83.33 85.04 79.62 78.23 88.20 88.07 89.56 90.37
83.66 83.51 85.20 79.80 78.39 88.45 88.26 89.85 90.38
0.33 0.1s 0.16 0.18 0.16 0.25 0.19 0.29 0.01
Glycerol expected Gram
Glycerol found Gram
Difference Gram
0.1655 0.3106
0.1646 0.3111
0.0009 0.0005
April 15, 1943
26 1
ANALYTICAL EDITION
independent analyses b y the acetin method. Samples 10 and 11 are “synthetic soap lyes”, made from weighed amounts of the standard A. 0. C. S. crude glycerin, salt, and water, t o simulate the average soap lyes found in practice.
t h a t the author would like, he feels t h a t i t is essentially useful as it stands and t h a t refinements may be added when i t is generally tried out.
Summary
Literature Cited
This method offers many advantages. The over-all time is about 4 hours, and the applied time is less than 1 hour. The technique is simple and t h e apparatus is readily available. The deviation from accepted or independent analyses of crudes is less than 5 parts per thousand, a fact that recommends it as a cost-accounting tool. Although the method has not been perfected t o the degree
(1) Am. Oil Chem. Soc., “Official and Tentative hlethods”, p. D-5 (1941). (2) Andrews, J. T. R., et al., Oil & Soap, 18, 14 (1941). (3) Committee on Fats, Soaps, and Glycerine, Division of Industrial Chemists and Chemical Engineers, AM. CHEM.SOC.,J. IND. ENC.CHEX,3, 682 (1911). (4) Govan, W.J., Jr., Oil & Soap, 19,27 (1942). (5) Lawrie, J. W., “Glycerol and the Glycols”, pp. 234-5, New York, Chemical Catalog Co., 1928.
Determination of Tin in Babbitts, White Metal Alloys, and Bronze EDWARD T. SAXER AND ROBERT E. MINTO, Otis Works, Jones and Laughlin Steel Corp., Cleveland, Ohio
WHILE
seeking a more satisfactory method for the determination of small amounts (