IXDUSTRIAL AIVD ENGIKEERING CHEMISTRY
386
Comparison of Results from Different Calorimeters Comparable tests, employing the three types of calorimeters described above, indicated that almost identical results could be obtained from all three calorimeters when due regard was paid to possible sources of error. Results of heat-of-solution tests made on neat specimens cured on the time-temperature curve of corresponding concrete, when converted into temperature rise of concrete, closely represented the time-temperature curve obtained from an adiabatic calorimeter. Likewise, results of heat-of-solution tests on neat specimens cured a t 21.11" C. (70" F.) checked very well with results on similar neat specimens tested in vane Ealorimeters a t 21.11 " C. (70" F.), Results were not in agreement, however, until proper account was taken of (1) immediate heat of hydration, (2) carbonation of heat-of-solution specimens, (3) water-
VOL. 10, KO. i
cement ratio, and (4)variations in specific heat of concrete with temperature. The effects of the other possible sources of error discussed above either were not involved in the comto be subject to proof Of parisons Or were too in the tests that were made.
Literature Cited (1) Carlson, R. W., Proc. Am. Concrete Inst., 34, 89 (1937). (2) Carlson, R. IT., Proc. Am. SOC.Testing Materials, 34, Part 11, 322 (1934). (3) Hornibrook F. B., Kalousek, G. L., and Jumper, C. H., J . Research Na'atl. Bur. Standards, 16, 487 (1936). (4) Lea, F. M., "Comparison of Methods for Measuring the Heat of Hydration of Cements," London Congress Intern. .4ssoc. Testing Materials, 1937. Lerrh. Wm.. Eno. News-Record. 113. 523 (1934). (5) (6) Woods, Hubert,-and Steinour, 'H. H., Proc. Am. Concrete Inst., 3 (31, 195 (1931). Bpril 12, 1938
Separation of the Chrysanthemum Carboxvlic Acids J
Destructive Effect of Steam Distillation on Chrysanthemum Monocarboxylic Acid ATHAN A. PANTSIOS, University of Chicago, Chicago, Ill.
T
HE acid methods of estimating the pyrethrin I and I1
content of pyrethrum flowers and their extracts depend upon the separation of the chrysanthemum acids by steam distillation of the steam-volatile monocarboxylic acid (6, 6, 7). Estimations of pyrethrin I by these methods are usually lower than those obtained by difference after estimation of total pyrethrins by the Gnadinger-Cor1 (1) reduction method and of pyrethrin I1 by the Haller-Acree (4) methoxyl method. TABLEI. Loss OF CHRYSANTHEMLT~ MONOCARBOXYLIC ACID ON REPEATED STEAMD~STILLATION OF THE SAMESAMPLE Operation 1 2 3 4 5 6
7
8 9 10 11
12
I
(Acid expressed as cc. equivalent of 0.0200 N base) Total Acid in Distillate --LossLoss 52.0 4i.3 42.5 38.0 33.5 29.6 25.8 22.2 20.1 18.8 17.4 16.1
4.7
4.8 4.5 4.5 3.9 3.8 3.6 2.1 1.4 1.3 1.3
%
%
9: 1 10.2 10.6 11.8 11.6 12.8 13.9 9.5 7.0 6.9 7.5
9:1 18.3 26.8 35.2 43.0 50.4 56.9 61.4 64.0 66.5 69.0
Acid in Residue 0:i, 0.1 0.05 0.1 0.1 0.1 0.05 0 1 0.1 0 1 0.1
Graham (2) has found that steam distillation of perfumed oil extracts to remove the essential oil (as directed in the Seil method, 6 ) results in a 25 per cent loss of pyrethrin I, and ( 3 ) that there is a lack of uniformity among the results of different analysts. This study was undertaken t o show the effect of steam distillation on chrysanthemum monocarboxylic acid and to develop a different method of separating the acids.
Procedure A sample of pure chrysanthemum monocarboxylic acid (b. p., 140-2" a t 9 mm.) was subjected to repeated steam distillationsthe acid after extraction from the distillate and
titration with standardized base was reacidified and stearndistilled again. All steam distillations were carried out as directed by Seil. After eleven steam distillations (Table I) 69 per cent of the original acid present was destroyed. The average loss for each distillation was over 10 per cent. A series of samples containing various amounts of the chrysanthemum monocarboxylic acid was prepared both by direct weighing of the pure acid and by measuring off aliquots from a standardized alkaline solution of the acid. Each sample was steam-distilled and the acid reestimated as in the Seil method. The results of a series of determinations are condensed in Table 11. It is evident that steam distillation destroys a n average of over 10 per cent of the chrysanthemum monocarboxylic acid; consequently, all methods of estimation of the pyrethrins involving steam distillation give low values for pyrethrin I. TABLE11. ON
Sample
-4 B C
D E F
G H I J K
L
LOSS O F CHRYSANTHEMUM MONOCARBOXYLIC ACID STEAMDISTILLATION OF VARIOUSSAMPLES
(bcid expressed as 00. equivalent of 0.0200 N base) Acid Acid in Present Distillate --Loss% 31 . B 37.6 36.0 33.3 36.6 38.3 36.9 43.8 43.9 48.2 33.4 41.0
27.5 34.5 31.8 28.6 32.0 32.8 31.5 39.4 37.6 43.8 28.5 35.0
4.4 3.1 4.2 4.7
;:! 5.4 4.4 6.3 4.4 4.9 6.0
13.8 8.2 11.6 14.1 12.6 14.4 14.6 10.0 14.6 9.1 14.6 14.6
Acid in Reeidue 0 1 0.1 0 0 0 1 0 1 0 1 0 1 0 0 0 1 0 1 0 1 0 1
That any degree of precision is possible in the acid methods of pyrethrin analysis can be explained by the fact that under comparable conditions the loss of the monoacid on steam distillation is approximately proportional to the amount of acid present. The lack of agreement in the results of different analysts is also understandable.
JULI- 15, 1938
.4X.4LUTIC,4L EDTTIOX
Method of Separation I n the course of analyses for the pyrethrins by the Seil method it was observed that low-boiling petroleum ether does not extract the chrysanthemum dicarboxylic acid from an acidified solution. Since the chrysanthemum monocarboxylic acid is soluble in petroleum ether, it should be possible to separate the mono- from the dicarboxylic acid by such a selective extraction. A few preliminary experiments were sufficient to demonstrate that petroleum ether does not extract the dicarboxylic acid; but, in order to prove that i t is possible to separate the two acids, several artificial mixtures of the two acids were prepared and subjected to the following analysis: Aliquots from standardized, freshly prepared, alkaline solutions of the acids were mixed, acidified, and then extracted with lowboiling petroleum ether as in the Seil method, with the exception that care was taken to retain quantitatively the water layer and washings. The petroleum et,her extract was titrated with standardized 0.0200 N base in the usual manner, but the water layer was again retained. The acidified water layer and washings from the petroleum ether extraction w r e then extracted with diethyl ether, and the chrysanthemum dicarboxylic acid was determined by the Seil method. The acids were then recombined and the mixture was subjected to analysis for both acids by steam distillation (the Seil method). TABLE 111. SEPAR.4TION
O F CHRYBAKTHEYUM MOXO- .4SD CARBOXYLIC ACIDS
Sam*le
DI-
(Both acids expressed in cc. equivalent of 0.0200 N base) Acid Found Loss by Selective Solvent Steam DisAcid Present Extraction Steam Distillation tillation, .MonoDiMonoDiMonoDiMono-
% 29.0 34.1 38.7 45.5 38.4 0.0 33.1 35.2
29.2 34.2 38.7 45.6 38.5 0.1 33.1
39.8 41.9 81.4 40.4 41.9 44.0 0.0 49.8
..
25.2 30.3 32.6 41.4 36.0 0.1 30.0 31.6
39.7 41.8 81.2 40.3 41.8 44.0
0.1
..
40 2 42.1 81.7 40.6 42.0 44.1 0.1 49 9
13.1 11.3 16 0 9 0 6 3 9' 4 10 3
Several such determinations were carried out, and the results are summarized for comparison in Table 111. It is apparent that tihe separation of the two acids by the selective solvent extraction method is quantitative, well
387
within the experimental errors involved. The destructive effect of steam distillation on the monocarboxylic acid is again demonstrated; for, while the results of the two methods for the dicarboxylic acid are comparable, those for the monocarboxylic acid differ by over 10 per cent. The use of this method for the estimation of the pyrethrin content of pyrethrum flowers and its extracts was suggested. But, when it was applied to samples prepared in accordance with the Seil method, the results were somewhat high for both acids, perhaps because of the presence of other acidic constituents. Further work is being pursued, and there a r e indications of a successful solution. Selective extraction of the monoacid by other solvents is also being investigated.
Conclusion Because of the destructive effect of steam distillation on the chrysanthemum monocarboxylic acid, the acid methods of pyrethrin I analysis are inaccurate and unreliable (8). It is possible to separate the two chrysanthemum acids by the selective extraction of the monocarboxylic acid with lowboiling petroleum ether, but this method requires further study before it can be applied to the estimation of the pyrethrin content of pyrethrum flowers and their commercial extracts.
Acknowledgments The writer is greatly indebted to F. Y. Klock and A. A. Srebren, formerly of the Murray and Nickell Manufacturing Company, who made this investigation possible by granting the use of their laboratory facilities and supplies.
Literature Cited (1) Gnadinger, C., and Corl, C., S. Am. Chem. SOC.,51, 3054 (1929). (2) Graham, J. T., IND,ENG.CHEM.,Anal. Ed., 8, 222 (1936). (3) Graham, J. T., J . &soc. Oficial A g r . Chem., 20, 388 (1937). (4) Haller, H. L., and .4rree, F., Jr.. IND.ESG. CHEX.,Anal. Ed., 7. 343 (1935). (5) RipertlJ.:-Ann. jals , 28, 27-38 (1935). (6) Seil, H. A . , Soap. 10, 89 (1934); Chem. Trade J., 95, 168 (1934). (7) Tattersfield, F., and Hobson, R., J . A g r . Sci., 19, 433-7 (1929). ( 8 ) Wilcoxon, F., Contrh. Boyce Thompson Inst., 8 , 175 (1936).
RECEIVED March 1 1 , 1938.
Evaluation of Commercial Arsenious Oxide by Titration with Iodine P4UL LUNDXIAS, Bolidcns Grul .4/BB>produhtslaboratorium, Ronnskar. Skclleftehamm, Sweden
T
HE following procedure has been used in the analysis of
commercial arsenic trioxide: Weigh out samples of 5.000 grams, cover with 70 ml. of 6 N sodium carbonate solution, and boil until completely dissolved. Cool, make up to 1liter, and take a 50-ml. aliquot for the analysis. Make the solution barely acid with 6 iV sulfuric acid, add 3 grams of sodium bicarbonate, and titrate with 0.1 N iodine to a starch end point. TABLE I. ANALYSISOF ARSENICTRIOXIDE (95 to 97 per cent pure Xs203)
% Probable purity Direct titration value Artificial standard 100 - (nonvolatile Nonvolatile residue Ass01
+
so3
SbnOs Se
%
%
%
96.44 97.01 96.50 96.34 3.01 0.20 0.45 0.46 0 21
95.79 96.58 96.00 95.89 3.27 0.34 0,50 0.55 0.17
95.06 95.48 95.07 95,2S 3.96 0.31 0.43 0.78 0.21
If the iodine solution is standardized in exactly the same way against a sample of arsenic trioxide known to be pure, the results are very exact, provided nothing other than arsenic is present that is oxidized by iodine. It has been known for some time, however, that the values thus obtained are usually too high. The experiments described in this paper were undertaken to ascertain the reason for the error and t o see if the difficulty could not be overcome in some simple manner. I n Table I, the results obtained in the analysis of four samples containing 95 to 97 per cent of arsenic trioxide are given. The first horizontal column gives the actual purity of the sample obtained by determining all the impurities known to be present and subtracting the sum of these values from 100. The second column gives the values obtained by direct titration as outlined above; these values are about 0.5 per cent too high. The third column gives values which were