ANALYTICAL CHEMISTRY
250 part of R , which collects during the filling, was determined. The average residue was 4 mg. The weight of solution introduced was obtained by subtracting the weight of the residue in the valve from the weight of the apparatus filled with the unknown solution. No correction for the volume of gas which evaporates into the narrow part of E is necessary, since it is considered within experimental error. The filled apparatus waa next suspended vertically in a glass water bath, and cathetometer readings were taken a t both 25’ and 30” C. By comparing these cathetometer readings with those obtained during the calibration, the volume could be determined. From the mass and the volume the density was calculated.
The densities of the other formulas were measured with a small hydrometer (4j. -411 the densities a t 25’ and 30” C. are given in Table 111.
The density of only aerosol formula 350 a t two temperatures
(4) Smith, C . M.,and Goodhue, L. D., IND.ENQ.CHEY.,. 4 ~ . 4 ~ED., . 16,355-7 (1944).
waa determined in this way, but the apparatus is well adapted for measuring liquefied-gas solutions of any density. The accuracy of this method appears to be good, although no comparisons have been made with other methods.
LITERATURE CITED
(1) Barr, Guy, “Monograph of Viscometry”, p. 116, London, Humphrey Milford, 1931. (2) ~, Bennine. A. F.. and Markwood. W. H... Jr.., J . Assoc. Refria. . E~&;., 37 (4j, 243-7 (1939). (3) Marks, L. S.. “Mechanical Engineers’ Handbook”, 4th edition, p . 250, New York, McGraw-Hill Book Co., 1941.
TEIBwork was conducted under a transfer of funds, recommended by the Committee on Medical Research, from the Office of Scientifio Research and Development to the Bureau of Entomology and Plant Quarantine.
Determination of Beta-Dicarbonyl Compounds WILLIAM SEAMAN, J. T. WOODS, AND E. A. MASSAD’ Calco Chemical Division, American Cyanamid Company, Bound Brook, N. J .
A method is described for determining acetylacetone, ethyl acetyl pyruvate, and sodium ethyl acetyl pyruvate which may be of general applicability to j3-dicarbonyl compounds. The method is based upon the formation of a copper complex by treatment with copper acetate solution; this complex is precipitated and filtered, and the filtrate extracted with chloroform. The excess copper acetate is determined iodometrically as a measure of the copper consumed by the dicarbonyl compound. The precision of the method is expressed by a standard deviation of &0.13% (absolute). There is no reason to suspect any systematic errors. The effects of some possible impurities are discussed.
A
CCORDING to the literature, many pdicarbonyl compounde
react with copper salts, presumably to form chelated compounds, some of which are insoluble, and others soluble in water. Both the water-soluble and the water-insoluble copper compounds in general seem to be soluble in organic solvents such as chloroform. It is the purpose of this paper to show how this reaction with copper may be used for determining a t least some 0-dicarbony1 compounds, using as examples acetylacetone (2,4-pentanedione), ethyl acetyl pyruvate, and sodium ethyl acetyl pyruvate. Further investigation may indicate that t,he method may be of general application for pdicarbonyl compounds. Apparently some of the usual methods for the determination of carbonyl compounds cannot be used for the p-dicarbonyl compounds in question. Thus the reaction with hydroxylamine forms the basis of a well-known method for the determination of monocarbonyl compounds (3, 6 ) . However, acetylacetone reacts with hydroxylamine to give a t least two products, the dioxime and CY, ydimethylisoxazole (I), and with p-nitrophenylhydrazine to form l-(p-nitrophenyl)-3,5-dimethylpyrazole as well m the di-(p-nitrophenylhydrazino) acetylacetone (6jI Meyer (4) reports that acetylacetone can be titrated as a monobasic mid, using Poirrierblau as the indicator. The authors have found that while acetylacetone can be titrated potentiometrically, the value obtained does not mean much because the samples contain other acidic constituents. Beilstein (2) shows that practically all 8-dicarbonyl compounds form copper complexes of some kind. Some of these are soluble in water, and nearly all seem to be soluble in chloroform. The formation of the copper complex is proposed here as the basis of a method for determining p-dicarbonyl compounds. The 1
Present address, 50 Arthur St., Worcester 4, Mass.
complexes formed are in general well-defined, stable compounds having definite formulas and melting points. The method involves treating with copper acetate to form the copper complex which has a limited solubility in water, filtering off the precipitate and eltracting the remaining complex in solution with chloroform. The excess copper is determined iodometrically. For compounds which form water-soluble copper complexes, a method involving extraction with chloroform out of the aqueous solution can be used; the procedure for such compounds will be reported later. The reaction with copper salts may be of interest for dicarbonyl compounds in general, but it will have to be modified in some of its details for the determination of any particular compolllid. REAGENTS
Cupric acetate solution. Fifty grams of cupric acetate monohydrate and 200 grams of sodium acetate trihydrate are dissolved in 2 liters of water. The solution is filtered before use, and standardized iodometrically by titrating a 100-ml. aliquot with 0.1 dVsodium thiosulfate solution in the manner described below. Let this volume equal A . This solution was designed to give the correct pH for the determination of acetylacetone. For other B-dicarbonyl compounds it is necessary to modify the composition of the solution, so that the optimum pH is obtained after precipitation or extraction. This is done by adding either more acid or more sodium acetate as may be needed. Sodium thiosulfate, 0.1 N. Hydrochloric or sulfuric acid, 1 volume of concentrated acid diluted with 1 volume of water. Potassium iodide, C.P. Starch solution. PROCEDURE’
If the sample is soluble in water, a 1.5- to 2.0-gram sample is weighed accurately into a 258-m1. glass-stoppered flask containing
V O L U M E 19, NO. 4, A P R I L 1 9 4 7 Table I. Sample No.
9
251
Analyses Using Copper Method
Type of Sample Acetylacetone Acetylacetone Acetylacetone Acetylacetone Acetylacetone Acetylacetone Ethyl acet 1 pyruvate Sodium e d y l acetyl pyruvate Sodium ethyl acetyl pyruvate
Values Found, % Single values Average 98.4,98.3, 98.5, 98.4 98.4 72.6, 72.4, 72.5 72.5 8 7 . 7 , 8 7 . 3 ,8 7 . 5 , 8 7 . 0 87.4 87.2,87.5 83.3, 83.4 83 4 98.5, 9 8 . 6 98 6 98.8, 98.8 98.8 91.8, 91.5, 91.8, 91.7, 92.0 91.8 90.7, 90.7 90.7 9 8 . 4 , 9 8 . 4 , 9 8 . 3 , 9 8 . 4 , 98 3 98.0,98.3,98.6,98.4,98.7
98.4
Table 11. Copper Content of Complex Per Cent Copper Theory Found 24.29 24 2 2 , 2 4 27 Av. 2 4 . 2 5 16.82 16 8 3 , 1 6 . 8 3 Av. 16.83
Complex Acetylacetone Sodium ethyl acetyl pyruvate
a 100-ml. aliquot of cupric acetate solution. If the sample is not soluble in water, it is dissolved in a few milliliters of ethanol and then the standard copper solution is added. The flask is stoppered and shaken frequently over a period of 5 minutes. The mixture is filtered on paper through a 6-cm. Buchner funnel into a clean suction flask, and the flask is rinsed with three or four 5-ml. portions of distilled water, each wash being poured separately into the funnel. The contents of the funnel are washed with three or four 5-ml. portions of distilled water. The filtrate is transferred quantitatively by means of a funnel into a 250-ml. separatory funnel and the suction flask is rinsed with three or four 5-ml. portions of distilled ryater. To the separatory funnel are added 20 ml. of chloroform. The funnel is shaken vigorously for 1 minute. The chloroform layer is then run into another seporatory funnel, containing about 40 ml. of water, and shaken about 0.5 minute, and the chloroform layer is discarded. This procedure is repeated with three or four more 20-ml. portions of chloroform until the last chloroform extract is colorless. Each chloroform layer is washed with the same 40-ml. portion of water. After the final wash the water layers from both funnels, containing the cupric acetate, are transferred to a single 500-ml. iodine flask, and each funnel is rinsed with two or three 5-ml. portions of distilled water. To this solution are added 15 ml. of hydrochloric acid and 8 grams of potassium iodide. The flask (tightly stoppered) is shaken several times over a period of 10 minutes and then the solution is titrated with 0.1 N sodium thiosulfate. When sufficient thiosulfate has been added to render the mixture light yellow in color, 4 ml. of starch solution are added and the titration is continued to the disappearance of the blue color. Let this volume equal B. The equivalent weight of 4dicarbonyl compounds is equal to twice the molecular weight.
% 6-dicarbonyl compounds =
In regard to the possibility of low values, it would appear that the extraction is reasonably complete. The values in Table I11 show how much of the copper complex is extracted by each chloroform extraction; after about the fourth, the extraction is practically complete. Important Variables. Probably the most important variable that must be controlled is the pH. If the p H is too high, copper hydroxide is precipitated with the complex. If too lorn, the keto form of the carbonyl is produced, which apparently does not form the copper complex. The optimum range seems to he in general between 5 and 7: for acetylacetone it is 5.2 to 6.0; for sodium ethyl acetyl pyruvate and ethyl acetyl pyruvate, 4.8 to 6.5. In order t o be sure that the solution has this pH after precipitation, it is best to add sodium acetate or a suitable mixture of sodium acetate and acetic acid to the copper acetate used as a precipitant. The values in Table IV show the extent of the error obtained if the pH is not controlled closely. In the determination of acetylacetone, if acetoacetic ester is present, the pH must he controlled more carefully than between 5.2 t o 6.0. At 5.2, acetoacetic ester is not precipitated with copper, but a t 6.0 a considerable amount is precipitated. Thus, at a pH of 5.2 the addition of 14YGacetoacetic ester produces an increase of only about 0.15y0 in the acetylacetone values; but a t pH 5.6 the addition of 14y0 causes an increase of 1.1%; and a t 6.3 the addition of 14yGcauses an increase of about 47, in the acetylacetone value. If this impurity is known to be present, the pH must be controlled a t about 5.2. The other most likely impurity in this substance is acetic acid, which does not interfere because the pH is controlled. Another variable that must be controlled is the extraction of the last of the copper complex. While the copper complexes of acetylacetone and ethyl acetyl pyruvate are relatively insoluble, there still is enough of the complex left in solution t o give erratic values if the last of it is not extracted. The values shown in Table 111 give an indication of the extent of this error if the extraction is neglected, and indicate how nearly complete the extraction is. Table 111. Completeness of Extraction No. of Extraction 1 2
3 4 5 6
Table IV.
2 ( A - B ) X molecular weight X 100 1000 X weight of sample
DISCUSSION
Analytical Values. Using this method, the values in Table I were obtained for acetylacetone, sodium ethyl acetyl pyruvate, and ethyl acetyl pyruvate. Precision and Accuracy. The standard deviation of a single value from its mean is *0.1370 (absolute). There is no reason to believe that the values for dicarbonyl include any systematic error. However, no 100% pure samples were available to test t,his point. As having some bearing on this question, it may be noted that the combined values for moisture (by the Karl Fischer method) and for dicarbonyl content were 99.5% for a sample of sodium ethyl acetyl pyruvate and 99.1% for a sample of acetylacetone. Furthermore, as some indication of the possible absence or presence of other substances in these samples, xhich might precipitate copper, the copper in the copper complexes has been determined and found to agree closely with the theoretical values (see Table 11). The close agreement of these values with the theoretical values indicates a high probability of the absence of substances which form interfering copper complexes.
Copper Complex Found in Each Extraction (Calcd. as % Diketone of Total Sample) Sodium ethyl Acetylacetone acetyl pyruvate 2.49 3 :40 0.60 1.73 0.24 0.36 0.08 0.14 0.04