Approach towards the Limit in the Process of Extraction CARROLL W. GRIFFIN AND MARGARETHA
VON
I
N AN EARLIER paper (2) it was shown that for the process of extraction the formula
one reason why the experimental path of Xn/Xo plotted against n might not coincide with the calculated curve. (Another reason would be, as Hill, 6, points out in connection with constancy of K , the slight solubility of one liquid phase in the other. In one case in the present study this factor caused a rather wide divergence between the observed and the calculated curve.)
xn = x0 [ K W y L , % - j ' of solution xO= grams-of solute L = total cc. of extracting solvent K = distribution ratio of the solute for the two solvents X , = grams of solute remaining unextracted after the nth extraction
where W
= CC.
After dissolving the solute in the W cc. of solvent one portion of L cc. of extracting liquid was added and the mixture shaken in a se aratory funnel placed in a water bath kept at a temperature oP25.0 =t0.02' C. In general the mixture remained in the thermostat one day and was then removed, and the layers were separated and analyzed. The first experiment, using L cc. of extractant, was in every case t o check the value of K as found by previous investigators. If the result confirmed the value of K found in the literature, it was assumed that sufficient time had elapsed for attainment of equilibrium, but as an additional
may be employed to calculate the amount of solute which may be expected after a given number of operations (the number of operations being varied by dividing the L cc. of extractant into L / 2 , L / 3 . . . . or L/n cc. portions where 2, 3 . . . . or n extractions are to be made); and the limit of Xfl/X0 as n approaches infinity was found to have the value e-L j K W . The object of the present study has been to learn just how closely the experimental results follow the theoretical path for the approach of Xn/Xo towards its limit, e-LiKw, as n takes on increasing values. In three of the four cases studied little or no disturbance from such influence as dissociation or association occurs. In one case, where acetic acid was used as the distributed substance, the results, uncorrected for such disturbance, are in good agreement with the theoretical. The procedure in the four cases investigated was first to select suitable values for L, W , and Xo. These quantities were chosen so that the concentrations of the solute in the two solvents would approximate initially the conditions used by previous workers in their determinations of K , the distribution constant. Of course the concentration of the solute diminishes rapidly with each extraction; this is only 1
SAAF,l Vassar College, Poughkeepsie, N. Y.
check upon this point the solute was fre uently dissolved initially in the L cc. of the one solvent instead in the W cc. of the other before mixing the two. After the first run, others followed which differed from the first only in that L/n cc. of extracting liquid were used in n successive portions, the second portion replacing the first after the first had reached equilibrium with the W cc. of original solution, and so on. Analysis was finally made to determine X,, the amount of distributed substance remaining unextracted. Furthermore, the total amount of solute-i. e., that in the W cc. and in, for example, the two L/~-cc.portions of extractantwas determined as a check against the weight, X O , originally taken.
01
In the figures the theoretical curve for Xfl/Xo against 12 has been plotted in each case by calculating sufficient points for Xn/Xo from Formula 1. The limiting value for X,/Xo as n + a~ (shown by broken curves in the figures) has been computed from the constant e-L j K W . The points designated in the figures by circles represent the experimental values as obtained by analysis for Xl/Xo, X2/Xo,X,/Xo, etc. The extent by which these fail to fall
+
Present address, 31 Hausaokerweg, Heidelberg, Germany.
FIGURE1. EXTRACTION OF BORICACID FROM ISOAMYL A m o n o L WITH WATER
FIGURE2. EXTRACTION OF IODINEFROM WATER WITH CARBONTETRACHLORIDE
F, 100 cc.
TV 1000 00. Xo, L,'24 0.1cc. gram
L,100 cc. Xo, 1.5 grams
338
SEPTEiMBER 15, 1936
ANALYTICAL EDITION
I
0.3
359
1
2
I
J 6 7 8 9 10 -nFIGURE4. EXTRACTION OF ACETIC ACID FROM ISOAMYL WITH WATER ALCOHOL
4
a5 -n-
FIGURE3. EXTRACTION OF IODINE FROM ETHERWITH ETHYLENE GLYCOL
w,200
CO.
L,200 cc. Xo,0.4 gram
upon the theoretical curve therefore represents the departure of the experimental facts from the theoretical. X o was always weighed to 0.1 mg. and volumes in titrations were read to 0.01 cc. It is therefore believed that the precision of the experiments was limited by the values of K which were employed in the calculations with three significant figures. The chemicals used were c. P. products from the Eastman Kodak Company and Merck and Co. They purposely were not further purified, since a higher purity would be impracticable in large-scale operations. I n every case the standard solutions used in making analyses were standardized against the substance used as the distributed solute.
Extraction of Boric Acid from Isoamyl Alcohol with Water Abegg, Fox, and Hertz (1) found the value of the distribution constant for this system to range, for varying concentrations, from 3.47 to 3.27 (concentration in water/concentration in alcohol). The results of the present work gave a value of 3.15. The reciprocal, 1/3.15, was employed in the authors' calculations since in Formula 1 K must be defined as concentration in original solution/concentration in extracting phase. Boric acid was determined by titrating the water and the alcohol phases with standard sodium hydroxide solution. Glycerol was employed instead of mannitol, and phenolphthalein was used as the indicator. As shown in Figure 1, six points were secured in this case with a maximum value of 10 for n. The experimental points are in very good agreement with the theoretical curve and with n = 10, over 93 per cent of the boric acid has been extracted from the alcohol.
Extraction of Iodine from Water with Carbon Tetrachloride Jakowkin (4) found the distribution constant for this system to range for a fivefold variation in concentration from 87.9 t o 85.8. The determination of K i n this work yielded a
3
w
100 00.
d, 100 cc.
X o , 5.0 grams
value of 83.9. The reciprocal, 0.0119, was used for calcuIations. Great care must be exercised to effect a complete separation of the tetrachloride with its dissolved iodine from the aqueous solution. Many extremely small droplets of tetrachloride are very slow in settling and the small but significant quantities of the solvent may be repeatedly obtained by whirling the mixture in the separatory funnel. Carbon tetrachloride is slightly soluble in water, but the rate of solution is so much slower than attainment of the distribution equilibrium that its solubility is negligible. The solutions were analyzed, after separation of the two layers, by titrating with standard sodium thiosulfate solution. Potassium iodide was first added and starch was used as the indicator. This case is particularly interesting in that here W was taken some forty times larger than L. Even so the results show a good concordance between the experimental and the theoretical points. Figure 2 gives the curve for this system. Five experimental points were obtained with a maximum value of 6 for n. The coincidence of the observed points and the calculated curve for this case is the closest of all studied. Even with the ratio of W : L being 41:1, the extraction of the iodine when n = 6 is found to be over 82 per cent.
Extraction of Iodine from Ether with Ethylene Glycol Landau (6) found the value of the distribution constant for this system to be 1.78. The present study confirmed this figure. Landau, after separating the two phases, added ethyl alcohol to the ether phase until a homogeneous solution was formed. He then added potassium iodide and titrated nearly to the end point with standard sodium thiosulfate solution, after which starch was added and the titration completed. In this work more consistent results were obtained by omitting the use of ethyl alcohol. When ethylene glycol and ether are mixed the ether layer diminishes in volume to a considerable extent; thus with each glycol extraction W becomes less. This means that the iodine removed by the glycol will become progressively greater than if no such diminution occurred. As a result, as seen in Figure 3, X,, and thus also Xn/X0, are less than theoretically calculated, and the observed points fall below the curve. In fact when n = 4 the extraction has proceeded to 48 per cent, whereas theoretically it should be only 41 per cent.
INDUSTRIAL AND ENGINEERING CHEMISTRY
360
Extraction of Acetic Acid from Isoamyl Alcohol with Water Hertz and Fischer (3) determined the distribution constant for this system a t 20" C., and obtained a value of 0.923. I n this study the constant, uncorrected for dissociation, a t 25" C. was found to be 0.936. Employed in the calculations for the purpose a t hand the value 1/0.936, or 1.07, was used. Analyses of the two layers for acetic acid content were made by titrating with standard sodium hydroxide solution. Figure 4 shows the five points obtained in this case. All experimental points except for n = 10 are very near the theoretical curve. Duplicate runs for n = 10 yielded points, both of which fall slightly below the curve. It should be pointed out that this case is different from that shown in Figure 3 where the experimental points fell progressively further below the curve as n took on increasing values. In
VOL. 8, NO. 5
the present case, although, because of increased dissociation of acetic acid in the aqueous phase with greater dilution, more efficient extraction would be expected as n is increased, nevertheless the dissociation of acetic acid for concentrations prevailing when n = 1 and when n = 10 is of the same order of magnitude; only when n takes on high values would dissociation prove an important factor.
Literature Cited Abegg, Fox, and Hertz, 2. anorg. Chem., 35, 129 (1903). Griffin, IND.ENG.CHEM., Anal. Ed., 6,40 (1934). Hertz and Fischer, Ber., 37,4746 (1904). Jakowkin, 2. physik. Chem., 18, 585 (1895). (5) Landau, Ibid., 73,200 (1910). ( 6 ) Taylor, "Treatise on Physical Chemistry," p. 478, New York, D. Van Nostrand Co., 1931.
(1) (2) (3) (4)
RECEIVBID June 5, 1936.
Solubility of Naringin in Water GEORGE N. PULLEY, Bureau of Chemistry and Soils, U. S. Citrus Products Station, Winter Haven, Fla.
T
H E possibility of utilizing grapefruit residue for the production of naringin and its hydrolytic products, together with the fact that there is a limited commerical production of naringin at the present time, has shown the desirability of determining the solubility of naringin in water at various temperatures. Naringin (C27HS2014.2H20) was discovered by DeVry (2) in the flowers of grapefruit trees growing in Java. Will (6, 7 , 8), Zoller (Q),and Asahina and Inubuse (I) have conducted studies to determine its properties. Its increased solubility in hot water or juice has been noted by Fellers (S),and Segal and de Kiewiet (5) in technological studies on grapefruit products. The content of naringin in both peel and juice appears to diminish as the fruit matures. It is soluble in alcohol, acetone, and water. When crystallized from these solvents and dried a t 110" C. it melts a t 171" C. When crystallized from water it has an additional 6 molecules of water and melts a t 83" C. The bitter taste of naringin is pronounced: a water solution containing one part in ten thousand has a distinctly bitter taste. The naringin used in these experiments was made from grapefruit peel, purified by t h e method outlined by Poore (4) and dried at 110°C.; themeltingpoint was 171°C. (uncor.).
TEMPERATURC ' C
FIGURE1
The solubility of narin in was determine$ by adding an excess of the purified material t o 150 CC. of distilled water containedin aflask which was closed with a rubber stopper and immersed in a constant-temperature water b a t h . T h e flask was left in the bath 2 hours and was shaken every 15 minutes. At the end of 2 hours the solution in the flask was rapidly filtered, using a
water-jacketed funnel. A measured volume of the clear filtrate was transferred to a weighed evaporating dish and evaporated to dryness over a steam bath, then dried at 110" C., cooled, and weighed. The amount of naringin dissolved per 1000 cc. was calculated from the average of two determinations. The solubility at 6' C. was determined by placing the flasks in an electric refrigerator, while the solubility at 20" C. was determined in an ice-cooled box. The variation in the temperature at these two points was greater than at the higher temperatures, but, since the increase in solubility of naringin between 6" and 35' is so small, fluctuations in temperatures at 6" and 20" C. would have no significant effect upon the solubility value. Solubilities at other temperatures were carried out in a water bath, the temerature of which was controlled by means of a gas thermoregufator. The water in the bath was kept in constant motion by means of compressed air. The data given in Table I and Figure 1 show that up to 45" C. the increase in solubility with increase in temperature is not pronounced. From 45" C. to the melting point (83" C.) the solubility increases rapidly with increase in temperature. TABLEI. SOLUBILITY OF NARINQIN IN WATER Solubility in Water U . / l O O O cc. 6 0.17 20 0.50 35 0.79 1.96 45 7.16 55 42.21 65 76 108.24 The decreased solubility of naringin a t low temperatures may a t times cause the precipitation of this substance in canned grapefruit juice and sections as has been pointed out by Fellers and by Segal and de Kiewiet (6), This 'is especially true if the juice or sections have been prepared from immature or frozen fruit. I n the case of canned juice the glucoside generally settles t o the bottom of the container as a yellow sludge. Sections may show light yellow spots, which macroscopically have the appearance of mold. At times the juice has a milky appearance due t o minute crystals of naringin. Temperature of Water
c.
Literature Cited (1) Asahina, Y., and Inubuse, M., J . P h a r m . Sac. J a p a n , 49, 1928-35 (1929). (2) DeVry, Jahresber. Ph,armacog., 1886, 132. (3) Fellers, C. R., Canner, 69, No. 18, 11-12 (1929). (4) Poore, H. D., IWD. ENG.CHEM., 26, 637-9 (1934). (5) Segal, B., and de Kiewiet, T., J . South African Chem. I n s t . , 14, 43 (1931). (6) Will, W., Ber., 18, 1311-25 (1885). (7) Ibid., 20, 294-304 (1887). (8) Ibid., 20, 1186 (1887). ENQ.CHEM.,17, 1065 (1925). (9) Zoller, H. F., IWD. RECEIVED June 13, 1936.
Food Research Division Contribution No. 288
