Simplified Countercurrent Distribution Apparatus SAMUEL RAYMOND College of Physicians and Surgeons, Columbia University, New York, N . Y . OUSTERCURRENT distribution is an experimental C method of great power and versatility in the analysis of mixtures of similar substances. Introduced only a few years ago by L. C. Craig of the Rockefeller Institute (4),it has already become one of the indispensable tools of the chemist. The method consists in distributing a mixture in a certain way through a series of separatory funnels by means of a two-phase solvent system. The identity and quantity of the components of the mixture are determined from an analysis of the contents of each funnel after distribution. One method for carrying out the distribution employs separatory funnels and hand shaking. This apparatus is readily available, and allows a wide choice of solvent volumes and quantities of material, but suffers from the disadvantage that a large number of separate operations is necessary. For 8 24plate separation] 288 shaking operations are necessary. Another method currently in use employs a machine designed by Craig. This machine makes the necessary equilibrations and transfers automatically, but it is limited in solvent volume and is expensive. A semiautomatic apparatus combining the best features of both these methods has now been developed. It is essentially a series of separatory Figure 1. Countercurrent funnels of special design arranged on a Distribution frame for shaking and semiautomatic Plate transfer. APPARATUS
Separatory Funnels. Each unit or “plate” consists of a separatory funnel with a reentrant neck (Figure 1). These units can be made from standard Pyres BO-nil. cylindrical separatory funnels by any competent glass worker. Steps in converting the funnels are: Attach handles to each end of the funnel. Heat and draw down a section beginning about 7.5 cm. from the stopcock and extending 1 cm. toward the neck, leaving a sharp shoulder and a short section 1 cm. in diameter and 2 cm. ong to form the neck. After reheating the shoulder uniformly all around, push the neck into the body of the funnel. Cut off the neck as short as possible and finish in the usual manner. Cut off the tip of the funnel to a length of 2 cm.
The apparatus is driven by an electric motor (Bodine Electric Co., No. NSE33R, l/10 hp. with speed-reducing gears attached in the ratio of 15 to 1) through a cam a t a speed of 330 cycles a minute. The cam is a disk 1.5 inches in diameter set 0.25 inch off center. The total length of thrust is therefore 0.5 inch. The cam actuates an aircraft valve tappet and guide assembly (Aircraft Engine and Parts Co., New York, N. Y . )which drives the shaking frame in vertical reciprocating motion. Power for the downward phase of the motion is derived from a spring between the upper pivot and the guide block. This spring supplies a force of about 40 pounds when compressed 0.5 inch. Shaken together in this apparatus, 10 ml. each of butanol and water are brought into equilibrium within 2 minutes. A solute such as acetic acid, present in either phase, is distributed in the equilibrium concentration in both phases in the same length of time. The funnels are attached to the shaking frame, the tip of each funnel projecting into the neck of the one below. PROCEDURE
At the start of a distribution the desired volume of the less dense solvent, saturated with the heavier solvent, is placed in each plate. The sample, dissolved in the desired volume of the heavier solvent, is then added to the topmost plate of the column, which is shaken until equilibrium is established. (An automatic timer controlling the motor is a great convenience.) The lower layer of the first plate (plate 0) is now drained into plate 1 and a second volume of the heavier solvent is added to plate 0. (The added solvent does not contain an additional quantity of the sample.) This operation constitutes one transfer. After equilibrating, the lower layer of plate 1 is drained into plate 2, and that of plate 0 into plate 1, and fresh solvent is added to plate 1. This operation constitutes the second transfer. The addition of solvent is facilitated by an automatic pipet-e.g., Kimble No. 37075-at the top of the column. The cycle of operations is repeated until the desired number of transfers is reached. If more than a single vertical series of plates is utilized, the transfer from the bottom of one series to the top of the next is made by detaching the funnel and carrying it up by hand. RESULTS
Preliminary testing of the apparatus was carried out using butanol-water as the solvent system and acetic acid as the test ’ substance. The time required to reach equilibrium was deter-
A funnel made in this way will hold a total of 35 ml. A volume of 20 ml. can be shaken vertically in the funnel with no loss of contents. If greater volume capacity is desired, the diameter of the funnel rather than the height should be increased] because efficiency of mixing falls off as height is increased, and the number of plates the machine will hold also decreases. The standard funnel of the given dimensions occupies about 10 em. (4 inches) of vertical height. Figure 2.
Shaking Frame. The funnels are attached to a vertical rod (by means of laboratory clamps) for the shaking and transfer operations. The rod is part of a frame (Figure 2) constructed of 0.5-inch aluminum rods (Fisher Scientific Co., Flexaframe) fastened together with standard connectors. The pivots are special connectors (Emil Greiner Co., No, G23150) whose setscrews are cut off about 0.125 inch to allow for rotation.
Shaking Frame
G. Spring
1292
,
1293
V O L U M E 21, NO. 10, O C T O B E R 1 9 4 9 n
Table I.
Countercurrent Analysis of Lower Aliphatic Acids
Tube No. 0 1 2
c9
10 Extra transfers Totals
%
Total Titration 0.28 1.41 2.79 3.77 4.15 4.44 4.36 4.17 3.60 2.71 1.25 1.08 34.01 100
Acetic Acid Calcd. 0.00 0.04 0.20 0.63 1.53 2.75 3.73 3.92 3.19 2.11 0.99 0.49 19.58 57.6
Propionic Acid Calcd. 0.33 1.38 2.65 3.14 2.58 1.56 0.71 0.25 0.07 0.01 0.00 0.00 12.68 37.2
Total Calcd. 0.33 1.42 2.85 3.77 4.11 4.31 4.44 4.17 3.26 2.12 0.99 0.49 32.26 94.8
Difference, Col. 2-Col. 5 0.05 0.01 0.06 0.00 0.04 0.13 0.08 0.00 0.34 0.59 0.26 0.59
mined by a method similar to that of Barry ( 2 ) and was found to be 2 minutes for the given system. The K found for acetic acid was 1.23, in good agreement with the accepted value, 1.23 (1).
The efficiency of the distribution can be greatly improved for a given number of plates by carrying out a greater number of transfers than there are plates. The procedure is equivalent to procedure 2 of Craig (3, p. 523) except that most of the sample is retained in the apparatus, allowing calculation of distribution coefficients for all components. The standard operations are carried through until all the plates of the apparatus are utilized. Additional transfers are then made. The sample from the last plate is not transferred to a new plate, but is immediately analyzed. The procedure is continued until the desired substance begins to show up in the last sample. Providing the K for the component is greater than 1, a distribution can be obtained equivalent to half again as many plates as are actually used. In the analysis of a mixture of volatile aliphatic acids, obtained from another department of the college, an eleven-plate butanolwater system was used. After the fifteenth transfer the aqueous layer in the last plate had a titration value of 0.73 ml. of base, indicating the appearance of acid in this plate. The total contents of each plate were thereupon titrated, giving the values shown in column 2 of Table I. Calculations. In the mathematical analysis of the system the following symbols are used: r
T,
= serial number of a plate (numbering from 0) = quantity of any component in the combined phases of
23.
= number of transfers completed
K = distribution coefficient, equal to 1.23 for acetic acid and 3.65 for propionic acid
The fundamental equation which governs the behavior of normal substances in the countercurrent system is
A derivation of this equation has been presented by Williamson (6). Application of the fundamental equation shows that a maximum of acetic acid would appear in plate 7 and of propionic acid in plate 3. Preliminary values of T , for propionic acid in the other plates were calculated from the total titration in plate 3, assuming this to represent pure propionic acid. The propionic acid value calculated for plate 7, subtracted from the total acid in that plate, gave a corrected value for acetic acid, from which column 3 of the table was calculated. New calculations from the corrected propionic acid in plate 3 now gave column 4. The sum of the calculated values for each plate is given in column a. The difference (column 6 ) between this sum and the values found for each plate was within experimental error of titration over nearly all of the range. Of the total acidity of the sample taken, 94.8y0 was accounted for as acetic and propionic acids. It is easily seen from the difference column that acetic and propionic acids quantitatively account for the results over tubes 0 to 7 . Beginning with tube 8 there are positive deviations which are small but definitely larger than experimental error. These deviations indicate the presence of a t least one additional acidic component, more soluble in water than in butanol. ACKNOWLEDGMENT
The author's thanks are due to Beatrice Carrier Seegal for advice and encouragement, and to Herbert Wohl for technical assistance. This work was supported in part by a grant from the National Institute of Health. LITERATURE CITED
(1) Archibald, R. C., J . A m . Chem. SOC.,54, 3180 (1932). (2) Barry, G. T., Sato, Y., snd Craig, L. C., J . Biol. Chem., 174, 209 (1948). ( 3 ) Craig, L. C., Ibid., 155,519 (1944). ( 4 ) Ibid., 161, 321 (1945). (5) Williamson, B., and Craig, L. C., Ibid., 168, 687 (1947).
plate r
RECEIVED November 10, 1948.
Sulfasuxidine (p-2-Thiazolylsulfamylsuccinanilic Acid)
a t temperatures above 35 C. Good crystals of either form can be obtained by slow crystallization in the proper temperature range. Methyl Cellosolve was used to prepare form I1 and 25y0 ethanol to prepare form I in this study. The commercial samples of sulfasuxidine are form I; form I1 in all cases transforms through the solid phase over a period of a few weeks to give I pseudomorphs of 11.
HC-N 0 0 0 H ~ ~ -H~ I-1 ~ ~ - n -H C I/r H - C H 2 -OH -
e
0 Structural Formula of Sulfasuxidine Sulfasuxidine exists in a t least two polymorphic forms. Form I is stable below about 35' C. and melts a t 125" C.; form I1 is stable above about 35" C. and melts a t about 180" C. Crystals of either form may be prepared from methyl Cellosolve, 25% ethanol, or dioxane. Form I crystallizes if the solution is supercooled below 35' C. before crystallization; form I1 crystallizes
SULFASUXIDINE I
CRYSTAL MORPHOLOQY (determined by W. C. McCrone). Crystal S stem. Monoclinic. Form an$ Habit. Rods elongated parallel to b from 25% alcohol. The most common forms and usual1 the only forms are the orthopinacoid { loo}, clinopinacoid (%lo),and hemiorthodome ( 101 j. Axial Ratio. a : b : c = 5.098 :1 : 3.793.
