'
ANALYTICAL CHEMISTRY
146 ness of the magnesium uranyl acetate method for the determination of this common element and is of value in showing how the glycolate radical is distributed in the aluminum salt. The per cent glycolic acid was determined by a modified colorimetric method as described by Calkins ( 8 ) .
Table IV. Analyses of Aluminum Carboxymethylcellulose Samples Sample A-1 A-2 A-3 A-4 A-5 A-6
Colorimetric 3.18 3.28 3.28 3.64 4.95 5.20
Per Cent .4luminum 8-Hydroxyquinoline 3.17 3.27 3.17 3.58 4.84 5.09
Difference 0.31 0.30 3.35 1.65 2.23 2.11
Sodium,
Glycolic Acid,
0.23 0.15 0.19 1.76 2.98 2.15
17.2 20.0 15.0 16.1 30.0 31.0
%
%
Much work has been done pertaining t o the use of Aluminon, ammonium aurintricarboxylate, in the colorimetric determinat ion of aluminum. Composite aluminum reagents containing stabilizers such as gum arabic and gelatin have been used by Olsen, Gee, and MeLendon (IO) and Craft and Makepeace (4). Both reagents were investigated but were found t o be timeconsuming and the colorimetric curves produced by them were not suitable for the range of aluminum desired. The use of an aqueous solution of Aluminon as described by Sandell (11) proved to be the most suitable, as it was stable and easily prepared and gave the curve desired (Figure 4). The results obtained showed that the method was precise and that the determination was reproducible. LITERATURE CITED
Caley, E. R., and Foulk, C. W., J . Am Chem. Soc., 51, 166474 (1929).
Calkins, IT.P., IND. ENG.CHEM.,AKAL.ED., 15, 762-3 (1943). Cotton, R. H., Ibid., 17, 734-8 (1945). Craft, C. H., and Makepeace, G. R., Ibid., 17, 206-10 (1945). Eyler, R. W., Klug, E. D., and Diephuis. Floyd, ANAL.CHEW, 19, 24-7 (1947).
Kolthoff, I. M., and Sandell, E. B., "Textbook of Quantitative Inorganic Analysis," revised ed., BP. New York. .. 327-8, Macmillan Co., 1943. Ibid., 410.
Lindner, R. C., PEant Physiol., 19, 76-89 (1944). Lindner, R. C., and Harley, C. P., Science, 96, 565-6 (1942). Olsen, A. L., Gee, E. 8 . ,and McLendon, V., IKD.EKQ.CHEM., ANAL.ED.,16, 169-72 (1944). Sandell, E. B., "Colorimetric Determination of Traces of
Per Cent Aluminum
Figure 4.
Determination of Aluminum
Cenco Type B-2 photelometer, with green filter 1.0% aluminrim equals 0.01 m g .
Metals, Chemical Analysis," Vol. 111,pp. 114-18, New York, Interscience Publishers, 1944. (12) Willard, H. H., and Boyle, A. J., IND.ENO. CHEM.,ANAL. ED.,13, 137-9 (1941). (13) Willard, H. H., and Diehl, H., "Advanced Quantitative Analysis," p. 261, New York, D. Van Nostrand Co., 1943. RECEIVED May 31, 1947. Presented before the Diviaion of Analytical and Micro Chemistry at the 111th Meeting of the AYERICAN CHEMICAL SOCIETY, Atlantic City, N. J.
Quantitative Determination of Dicyclopentadiene FELIX BERGMANN' AND HELENE JAPHE The Daniel Siea Research Institute, Rehovoth, Palestine
A method for the quantitative analysis of dicyclopentadiene consists of addition of formic acid to one of the double bonds and determination of the saponification equivalent of the resulting formate ester. The method is also applicable to the analysis of higher polymers of the hydrocarbon.
A
LTHOUGH dicyclopentadiene and its monomer have become of considerable industrial importance in recent years, the few methods that have been developed for the quantitative determination of the dimeric hydrocarbon are subject t o limitations. The method of Schultze ( 9 ) , based upon bromination in the absence of oxygen, involves an,elaborate procedure and is accurate t o no more than about 2%. Uhrig, Lynch, and Becker ( 4 ) have described a method based upon the controlled depoly1
Present address, Polytechnic Institute of Brooklyn, Brooklyn, N. Y.
meriaation of dicyclopentadiene t o the monomer, which is determined colorimetrically in the form of the phenylfulvene, but careful control is required because of the tendency of the monomer to polymerize a t higher temperatures. In a recent investigation of dicyclopentadiene the authors observed that one of the two double bonds of the hydrocarbon possesses an unusual power to add various reagents (1). Although Bruson (2) has stated that organic acids of dissociation constant. higher than about 1.5 X 10-8 react with dicyclopentadiene in the absence of an electrophilic catalyst (sulfuric acid,
147
V O L U M E 20, NO. 2, F E B R U A R Y 1 9 4 8 boron trifluoride, etc.), the authors found that the hydrocarbon (I) reacts with formic acid.@, = 2.1 X a t the reflux temperature to produce dicyclopentenyl formate (11)in very high yield.
Table 111. Determination of Various Samples of Dicyclopentadiene
1 2
3
\I/$-
\I/--'
I
4
IIa
5 6
IIb With a view t o the development of an analytical procedure, a study has nom been made of the factors governing the addition reaction and the subsequent saponification of the formate. SAPONIFICATION O F DICYCLOPENTENY I. FORMATE
When a mixture of dicylcopentenyl formate with 1.5 equivalents of 0.5 A' ethanolic potassium hydroxide was allowed t o stand a t room temperature (22" C.), about 93% of the ester was saponified in the first half hour but a very long time was required for completion of the reaction. Further experiments were therefore conducted a t the boiling point. The data reported in Table I were obtained by heating weighed samples of the formate with alcoholic alkali in a boiling water bath for varying periods of time, cooling the solutions, and immediately titrating the excess alkali. The results show that hydrolysis is 99.7 to 99.8y0 complete in 4 to 5 hours; a 6-hour period was decided upon as providing for completion of the reaction. ADDITION O F FORMIC ACID TO DICYCLOPENTADIENE
Pure, crystalline dicyclopentadiene (melting point 32 C.) was used to study the optimum ratio of hydrocarbon to formic acid and the velocity of addition. When mixtures of the reactants in various molar proportions were refluxed for 8-hour periods, it was found that a quantitative conversion could be realized with
Table I. Saponification of Dicyclopentenyl Formate with 1.5 Equivalents of 0.5 NPotassium Hydroxide in Ethanol at Reflux Temperature Time Min. 5 10 20 30 60
90 120 160 180 240 300
Table 11.
Yield b y Titration
% 92.7 92.9 93.3 93.7 94.8 95.7 96.6 97.6 99.2 99.7 99.8
Reaction of Dicyclopentadiene (1 Mole) with Formic Acid (3 Moles) at Boiling Point
(Analysis b y saponification with 0.5 N ethanolic K O H for 6 hours) Time of Reflux Ester Hours % 1 93.1 95.4 98.1 99.5 99.9
Composition of Sample 2.3207 grams of formate 2.4206 grams of formate 1.8586 grams of pure dicyclopentadiene 3.0891 grams of formate 2.0923 g r a m s of technical dicyclopentadiene 2.672 grams of pure dicyclopentadiene 3.272 grams of formic acid 2.3212 grams of technical dicyclopentadiene f 2.7821 grams of formic acid 2.1755 grams of Jechnicalj 2.8104 grams of dicyclopentadiene 0.6672 grams , o f pure formic acid dicyclopentadiene
+ +
+
i
+
Ester Found,
Hydrocarbon Found,
%
%
..
99.8
..
99.7 99.65
..
...
99.9
...
91.8 9 3 . 6 (calcd. 9 3 . 7 )
2 moles of formic acid per mole of hydrocarbon, and a ratio of 1 to 3 was chosen for the further experiments recorded in-Table 11. The results show that a reflux period of 5 hours is satisfactory, DETERMINATION O F DICYCLOPENTADIENE
Procedure. -4n accurately weighed sample of approximately 2.5 grams of dicyclopentadiene or a technical mixture containing the hydrocarbon is mixed with 2.5 grams of 99 to 100% formic acid in a flask with a ground-glass joint and the mixture is heated under reflux for 5 hours in an oil bath kept a t 135" to 140' C. The flask is then removed and cool@d,the condenser is rinsed with about 25 cc. of benzene, the solution is transferled to a separatory funnel, and the reaction vessel is rinsed with two 25-cc. portions of benzene. The organic layer is washed first with 15 cc. of distilled water and then with 15 cc. of saturated sodium bicarbonate solution t o remove excess formic acid, and finally with two fresh portions of distilled water; separation of the layers proceeds very quickly. The benzene layer is returned t o the reaction flask and an accurately measured volume (50.0 cc.) of 0.5 AVpotassium hydroxide in 95% ethanol is run in from a buret. The solution is refluxed for 6 hours and then cooled under the tap, the condenser is rinsed with 25 cc. of alcohol, and the excess alkali is titrated with 0.5 N hydrochloric acid, with phenolphthalein as indicator. Calculation. dicyclopentadiene = 6.6 (a - b)/neight of sample vhere a = cc. of acid required t o neutralize 50.0 cc. of potassium hydroxide solution, and b = cc. of acid used t o titrate sample.
The results reported in Table I11 indicate that the method is reliable to 0.5y0or better and that impurities present in technical preparations of dicyclopentadiene do not interfere with the analysis. Since higher polymers of cyclopentadiene react with formic acid in the same manner as the dimer does (I), mixtures containing higher polymers must be separated prior to analysis by fractional distillation; in view of the large differences in the boiling points, such separation presents no difficulties. Orienting experiments with tricyclopentadiene indicated that interaction with formic acid and saponification of the formate proceed quantitatively also in this series. Since cyclopentadiene dimerizes completely when kept a t room temperature for a few days, the monomer may be determined by conversion to the dimer and application of the above procedure. A mixture of monomer and dimer can be analyzed by distilling off the monomer from one sample and determining the dimer in the residue, and simultaneously keeping a second sample a t 20" to 25" C. for 4 to 5 days and determining the amount of total dimer. LITERATURE CITED
(1) Bergmann, F., and Japhe, H., J. Am. Chcm. Soc., 69, 1826 (1947). (2) Bruson, U.S.Patent 2,395,452(1945). 56, 1552 (1934). (3) Schultze, J.Am. Chem. SOC., (4) Uhrig, Lynch, and Becker, IND.ENG.CHEM.,ANAL. ED., 18, 550 (1946). RECEIVED February 21, 1947.