EXTRACTION LYMAN C . CRAIG The Rockefeller Institute f o r Medical Research, New York, N . Y .
XTRACTION has long been used analytically for preliminary separations of mixtures into groups. With present-day technique it can be further used to separate and quantitatively estimate the components of mixtures of closely related substances such as members of a homologous series. The basis for the first purpose is frequently that of solubility or of widely differing partition ratios. The basis for the latter is the use of a two-phase distribution in conjunction with a countercurrent process of some sort, so that separations may be made in spite of closely related partition ratios. The approach for each of the two objectives may be similar or it may be entirely different, depending on whether or not partition ratios favorable for the puipose are possible. Thus, if a desired solute is present in a mixture and a solvent can be found which will extract only the desired solute, the simplest case is presented. If the particular solute is extracted exclusively but in part only, so that successive extractions are required, a more difficult case is to be treated and a countercurrent process may be desirable. I n either case the only problemis that of performing the extraction so that the desired solute is completely removed from the accompanying substances. Khen such is the case in the literature considered, it is treated in the present review under the heading of “Extraction for Removal Purposes.” On the other hand, when the problem is that of the separation of substances of similar solubilities or of similar partition ratios, a countercurrent extraction process is involved and the process is treated under “Extraction for Fractionation Purposes.” I n order to limit the scope of the reviev, publications appearing before 1943 are not considered; otherwise, a short review would not be advisable. Even with this arbitrary restriction only those articles which appear to the writer to have a special interest are mentioned. As regards previous review articles, the yearly reviews by Elgin (41) should in particular be consulted, although they are presented more from the standpoint of chemical engineering than from that of laboratory analysis. Green (27) has written a general review on extraction, but i t too is written more from the preparative or commercial standpoint. An excellent review dealing mainly with the theory of such extraction has been written by Frey and Scheibel (25). EXTRACTION FOR REMOVAL PURPOSES
Judging from the number of paperson the subject (1,32,43,46), many chemists still have trouble with the continuous extraction of solids when amounts a little larger than ordinary laboratory quantities are encountered. Smaller amounts have apparently caused little difficulty. Wayman and Wright (57)have described efficient apparatus for the convenient extraction of small volumes of either light or heavy liquids. Accounts have been published (7, 2.4) of the use of reduced pressures in order to perform extractions a t lower temperatures and in inert atmospheres. A novel technique in analytical extraction, described by Harrison and Meincke (29),permits the use of pressure so that solvents which are gases a t room temperatures may be employed. Often these solvents of low molecular weight have unusual or specific solvent properties. The apparatus is not elaborate. It involves the use of pop bottles with a synthetic rubber gasket for the cap and hypodermic syringes with a commercially obtainable stopcock fitted a t the base of the needle. Most research workers dealing with natural products sooner or later encounter difficulty with emulsions. Centrifugation is
often time-consuming and ineffective. Emulsions usually result from surface-active solutes, but Davis (19) has published an interesting observation in this connection. When attempting to extract certain solutions which would otherwise cause trouble he found Duponol and Tween useful in preventing emulsions. Pearl ( 4 1 ) and Kieselbach (34) have designed continuous extractors with settling chambers where the emulsions have time to break before being transferred to the concentration chamber, Extraction of fermentation liquors is often not easy. Two somebyhat different continuous laboratory spray extractors (12, 36) have been designed to deal with this problem. I n both the aqueous phase is injected into the organic solvent a t one end of an unpacked column. Often stable emulsions do not form when the organic phase is present greatly in excess. I n the Kolfenbach et al. extractor a single jet was used for injection of the aqueous phase. Air bubbles were continuously passed through the column to help break the emulsion. I n the Bush and Goth extractor the aqueous phase was introduced through a distributor with many holes in i t but the size of the holes was carefully controlled. Stable emulsions did not form when the holes were not smal!er than a determined optimal size which, however, was small enough to give sufficient contact for good extraction. I n connection with the recovery of penicillin from fermentation liquors a number of interesting observations have been made (44) and a continuous countercurrent laboratory extractor has been designed for the purpose (59). Bush and Goth (12) designed their extractor for this purpose. The Podbielniak extractor (@), though too large for a laboratory analytical tool, is ideally suited for the commercial production of penicillin. A common error in the use of an extraction procedure for analytical purposes is the belief that the distribution behavior of a pure substance can be used to predict its extractability from a crude mixture. I n recent years a number of observations bearing on this point have been reported. For example, Dzialoszynski, Mystkowski, and Stewart (20) found that with certain organic solvents carotenes could not be extracted from aqueous solutions containing protein, though carotenes are not soluble in water but are fat-soluble. Extraction occurred readily a t a lower pH. They postulated an association complex in order to explain the effect. Another example is the work of Heymann and Fieser (30). Association of specific solutes is often recognized when the complex is less soluble than the separate components and therefore crystallizes out. Association can also be shown to occur in very dilute solution, by the solubilizing effect of certain solutes. The solubilizing effect of the higher fatty acids, certain proteins, certain sulfonic acids, the bile salts, saponins, etc., is well known. Weil-Malherbe (58) has shown that this effect is also produced by the purines and has studied the influence on the effect of substitution on the ring structure. Unquestionably, the effect is in the same category as the phenomenon known as salting in but occurs a t extreme dilution. I n solution the complex may be largely dissociated, but the tendency to associate is present and may be considered in terms of an equilibrium constant. Often certain added substances in small concentration greatly reduce solubility (salting out). The reason may be that they associate preferentially with some component of the solvent which would otherwise associate with the solute, thus leaving the solute unissociated and less soluble. Such effectsmay be extremely important in extraction. If the complex possesses a different partition ratio from the free solute it may defeat the purpose of the extraction. Conversely, it may form the basis of a highly useful analytical tool. Thus Brodie
86 (10) found that the dye methyl orange forms a salt or complex with antimalarial bases n-hich partitions in a water-chloroform or serum-chloroform system greatly in favor of the chloroform. The free dye remains in the aqueous phase. Colorimetric determination of the dye in the chloroform gives a quantitative measure of the amount of base. Sl’ith the use of other sulfonic acid dyes, the principle was earlier employed by Lehman and Aitken (37) for the estimation of Demerol and by Auerbach (3) for the estimation of certain quaternary bases. Titus and Fried (54) were able to partition streptomycin in a butanol-water system by addition of p-toluenesulfonic acid, whereas without the acid the antibiotic greatly favors the aqueous phase. Obviously such effects offer an enormous field in extraction foi future exploration They alio make soniea hat more complicated the use of simple extraction ar an analytical tool, particularly in biochemistry where substances n ith solubiliiing activities art’ plentiful. When a complication from this source occurs, the ube of systematic multiple extraction is often enlightening Introduction of a knomm solid phase can be helpful (23) in extracting mixtures such as feces EXTRACTION FOR FRACTIONATION PURPOSES
.Ipparently certain advantages of the stepwise or discontinuouk procedure over the continuous column procedures for analytical purposes have been fully realized only recently. In the discontinuous procedure ( 5 ) essential equilibrium can be reached at each stage and the factors required ran be quantitatively studied in a very simple mannrr. In the continuous or column procedure, equilibrium betn eeii the phases is not reached and therefore the rate of exchange of the solute is important. The rate is known to depend on many factors, some purely mechanical. Thc efficiency of the column can therefore be derived only in terms of a height of column equivalent to a theoretical plate. Such a height often changes with the solute mixture, with the different solvents, and with slight changes in operating conditions. The relative partition ratios or separation factors also change with concentration unless dilute solutions are used. On the other hand, the discontinuous procedures can easily be made entirely reproducible. Concentrations a t each stage can be calculatrd and held in a range of fairly constant partition ratios. Bush and Densen (11) have studied systematic multiple extraction with individual units such as separatory funnels and have developed an ingenious systematization which they have called the diamond pattern. The system permits ready calculation of the fraction of a single pure substance which would be expected in each unit a t the end of the process, provided the partition ratio remains constant. The importance of employing certain optimum volume ratios is discussed in their work. Stene (49)has made an exhaustive mathematical investigatioii of systematic extraction procedures of many different types. He has shown the value in this field of the mathematics of probability and statistics. Craig (16), Williamson and Craig (60), and Lieberman (38) approached the problem on a similar but much more restricted basis. Craig ( 1 5 ) developed an apparatus by means of which multiple quantitative extractions could be done rapidly in a sequence, so that each step would correspond exactly to a term of the binomial expansion. The name “countercurrent distribution” was given to this particular type of extraction. It thus constitutes the simplest fractionation scheme possible. For exact mathematical interpretation a constant partition ratio is required. Apparently systems wliirh givr constant partition ratios are possible for the largest majority of substances, as the method has been applied with success to fatty acids (8, 47), aromatic acids (31, 45), phenols ( 6 6 ) , penicillins (6, 8, 9, 17), antimalarial bases (63),streptomycins (+55),xanthomycins ( 5 0 ) , purines and pyrimidines (51), polypeptide antibiotics (B),and partial hydrolyzates of a protein or polypeptide material ( 4 , 61). The method is sufficiently quantitative for wide application t o the problem of purity (14,18). Buffers of high salt concentration
(16, 4 ? ) have proved useful. The analogy to the effect later obtained by Tiselius (56) in the use of salt solutions in chromatography is interesting. h recent development of great analytical significance to biochemistry in particular is that called “partition chromatography” (26). The operation is performed as a chromatographic separation but the effect is considered by the originators, Martin and Synge (39), to be due to liquid-liquid eutraction, since first an aqueous phase is adsorbed on some supporting agent such as silica, starch, or cellulose. The supporting agent forms a column and the immiscible organic phase is filtered through it. If the original viciv of the mechanics of operation is correct, then the process is similar to countercurrent distribution, except that it is a continuous process. This has now proved to be too simple an interpretation, as some of the better amino acid fractionations (15, 40) by the method have not required a second immiscible liquidphase. Furthermore, although certain band rates are in surprising agreement with that expected from a liquid-liquid partition ratio, many are in disagreement. It is also difficult t o reconcile the degree of equllibrium apparently reached in the column with the difficulty expected for a purely liquid-liquid system ( 5 ) of this type. I t would thrrefore appear that this development should best he reviewed under chromatography. An excellent review and bibliography on the subject of partition chromatography are given by Consden ( I S ) . A certain amount of study has continued in the attempt to develop an efficient continuous liquid-liquid extraction column (33, %5), but the technical difficulties are considerable. I n spite of the fact that separation factors with liquid-liquid extraction are usually much larger than with distillation and, therefore, such a column should be highly effective, a simple compact column which will furnish up to tn-enty stages was developed ( 2 2 ) only very recently. The column described by Scheibel (48)is of such design that it overcomes many of the mechanical difficulties in liquid-liquid extraction which contribute to the resistance of interchange of solute from one phase to the other and of clear-cut transfer or flow of the phases in the desired diyection. A promising approach where the required volatility is present is the more recent development known as “extractive distillation” ( 2 2 , 35). LITERATURE CITED
dpplesweig, N., IND.E m . CHEM.,ANAL.ED., 16, 472 (1944). Atchley. W. A., J . B i d . Chem., 176, 123 (1948). A4uerbach,M. E., IKD.EXG.CHEM.,ANAL,ED.,15, 492 (1943). Barry, G. T., Gregory, J. D., and Craig, L. C., J . B i d . Chem., 175,485 (1948).
Barry, G. T., Sato, Y.,and Craig, L. C., Zbid., 174, 209 (1948). Ibid., 174, 221, 217 (1948). Barthel, W. F., IKD.ENG.CHEX.,AKAL.ED.,17, 53 (1945). Behrens, 0. K., Corse, J., Huff, D. E., Jones, R. G., Soper, Q.F., and Whitehead, C. W., J . BioZ. Chem., 175, 771 (1948). Boon, W. R., AnaZyst, 73, 202 (1948). Brodie, B. B., and Udenfriend, S., J . Bid. Chem., 158, 705 (1945).
Bush, hl. T., and Densen, P. M., ANAL.CHEX.,20, 121 (1948). Bush, M. T., and Goth, A., IND.ENG.CHEM.,ANAL.ED., 16, 528 (1944).
Consden, R., Nature, 162, 359 (1948). Craig, L. C., Federation Proc., 7, 469 (1948). Craig, L. C., J. Biol. Chem., 155, 519 (1944). Craig, L. C., Golumbic, C., Mighton, H., and Titus, E. O., Ibid., 161, 321 (1945).
Craig, L. C., Hogeboom, G. H., Carpenter, F. H., and du Vigneaud, V., Ibid., 168, 665 (1947). Craig, L. C., Mighton, H., Titus, E. O., and Golumbic, C., A N A L . CHEM., 20,134 (1948). Davis, B. D., Arch. Biochem., 15, 351 (1947). Dsialossynski, L. M.,Mystkowski. E. M., and Stewart, C. P., Biochem. J . , 39,63 (1945).
Elgin, J. C., 2nd. Eng. Chem., 38, 26 (1946); 39, 23 (1947); 40, 53 (1948).
Fenske, M. R., Carlson, C. S., and Quiggle, D., Ibid., 39, 1322 (1947).
Fowweather, F. S., and Anderson, W. N.. Biochem. J., 40, 350 (1946).
V O L U M E 2 1 , N O . 1, J A N U A R Y 1 9 4 9
(24) Frampton, V. L., and Giles, F. K., IXD.EFG. CHEM.,ANAL.ED., 17,674 (1945). (25) Frey, A. J., and Scheibel, E. G., “Jubilee Volume Emil Barell 1946,” p. 446, Basle, Hoffmann-La Roche & Co., Ltd. (26) Gordon, A. H., Martin, A. J. P., and Synge, R. L. &I.,Biochem. J . , 37,79 (1943). (27) Green, G. C., Chem. Age, 50, 475, 497, 519 (1944). (28) Gregory, J. D., and Craig, L. C., J . B i d . Chem., 172, 839 (1948). (29) Harrison, S.A , and Meincke, E. R., ANAL. CHEW,20,47 (1948). (30) Heymann, H., and Fieser, I,. F., J . Pharm. Etptl. Therap., 94, 97 (194s). (31) Hogeboom, G. H., and Craig, L. C., J . B i d . Chem., 162, 363 (1946). (32) Jonnard, R., IXD.ESG. & E x . , ANAL.ED.,16,61 (1944). (33) Kenyon, R. L., Gloyer, S. W., and Georgian, C. C., Ind. Eng. Chtm., 40, 1162 (1948). (34) Kieselbach, R., IND. ENG.CHEM., AXAL.ED.,15,223 (1943). (35) Knox, 1%’. T., Jr.. Weeks, R. L., Hibshman, H. J., and McAteer, J. H.. Ind. Eno. Chem.. 39. 1573 11947).
(36) Kolfenbach, J. J:, Kooi, E. R., Fulmer, E. I., and Underkofler. L. A., IND. ENG.CHEY.,ANAL.ED., 16,473 (1944). (37) Lehrhan, R. rl., and Aitken, T., J . Lab. Clin. Med., 28, 787 (1943). (38) Lieberman, S. V.,J . B i d . Chem., 173,63 (1948). (39) Martin, -4.J. P., and Synge, R. I,. M . , Biochem. J . , 35, 1358 (1941). (40) Moore, S., and Stein, IT. I€., Ann. S. E’. Acad. Sci., 49, 265 (1948). (41) Pearl, I. -4., IND. ENG.CHEW,ANAL.ED.,16,62 (1944). (42) Podbielniak, Inc., Chicago, Ill., Circ. 13,14,and 15,U. S Patents 2,003,308 and 2,004,011 (June 4, 1935).
(43) Rapp, K. E., Woodmansee, C. W., and McHargue, J. S., IND. ENG.CHEM.,ANAL.ED.,15,351 (1943). (44) Rowley, D., Steiner, H., and Zimkin, E., J . SOC.Chem. Ind., 65, 237 (1946). (45) Rudkin. C.: O., and Nelson, J. M., J . Am. Chem. Soc., 69, 1470 (1947). . I s a ~ ED., . 18, (46) Salkin. 11.. and Kaye, I. A , , 1x0. ENG.CHEM., 215 (1946). (47) Sato. P., Barry, G. T., and Craig, L. C., J . B i d . Chem., 170, 501 (1947). (48) Scheibel, E.G., Chem. Eng. Pmgress, 44,681, 78%(1948). (49) Stene. S., ATkiv Kemi Mineral. GeoZ., 18A,No. 18 (1944). (50) Thorne, C. B., and Peterson, W. H., J . Biol. Chem., 176, 413 (1948). (51) Tinker, J. F., and Brown, G. B., Ibid., 173,585 (1948). (52) Tiselius, A., Arkiv Kemi Mineral. GeoZ., 26B,KO.1 (1948). (53) Titus, E. O., Craig, L. C., Golumbic, C., Mighten, H. R., Weinpen, I. M., and Elderfield, R. C., J . Org. Chem., 13,39 (1948). (54) Titus, E. O., and Fried, J., J . B i d . Chem., 168,393 (1917). (55) Ibid., 174,57 (1948). (56) Warshowskl-, R., and Schanta, E. J., ASAL. C H m f . , 20, 951 (1948). (57) Wayman, X f . , and Fright, G . F., IND. ENG.CHEM., -&N.\L. ED., 17,55 (1945). (58) Weil-Malherbe, H., Biochem. J., 40, 351 (1946). (59) Whitmore, F. C., and Coworkers, Ind. Eng. Chem.. 38, 942 (1946). (60) n?liarnson, B., and Craig, L. C., 6.B i d . Chem., 168,687 (1947). (61) Aoolley. D. W., Federation Proc.. 7, 200 (1948). RECEIVEDSovrrnber 9. 1‘348.
ION EXCHANGE ROBERT KUNIN Rohm and Haas Co., Philadelphia, Pa.
I TI1OUGH the recent :ivailability of a host of ion exchange
substances of various properties has stimulated the usage of ion exchange in analytical chemistry, the application of this phenomenon in analytical chemistry is not new. The use of Lloyd’s reagent (14), a hydrated, aluminum silicate cation exchange substance, has found wide application for the removal of ammonia (14) prior to the determination of urea and for the analytical separations of amino acids (S). The concentration of solutions of trace elements on such cation exchangers as aluminum silicates ( 1 ) and filter paper (10) has been a common practice for many years. The consideration of ion exchange principles in the study of the nature of precipitates in gravimetric analysis (Zl),in elucidating the mechanism of the glass electrode (11), and in explaining glass electrode errors in dilute, unbuffered solutions (12) has been of considerable importance in analytical practices. The difficulties in storiiig extremely dilute solutions in glass bottles have also been attributed to ion exchange (41). However, these applications have been but a minor contribution to analytical chemistry in comparison with recent developments. The availability of ion exchange resins containing various functional groups and in several instances having an “analytical grade” purity has been responsible for many new contributions to the field of analytical chemistry. The acceptance of ion exchange as an analytical operation or technique has not been universal, chiefly because the number of true analytical procedures involving ion exchange techniques have been few indeed. However, the possibility of utilizing ionic adsorbents such as ion exchange substances has opened many new vistas in analytical chemistry and i t is probable that in the near future many new procedures founded upon ion exchange principles will be revealed. The chief advantage of an ion exchange technique is that it enables one simply and rapidly to achieve a separation or concentration that would ordinarily be very difficult and time-consuming. In many applications of ion exchange in analytical
chemistry, some accuracy and completeness of separation are sacrificed for simplicity and time. However, for many determinations, this sacrifice is well warranted. The applications of ion evchange in analytical chemistry may tie classified as (1) concentration of dilute solutions, (2) fractionation of ions having similar analytical properties, (3) removal of interfering ions, and (4)miscellaneous analytical applications. A clear-cut distinction between some of the topics is lacking, because all the applications are based upon a common principle, the exchange of an ion in the ionic adsorbent with an ion in solution. BASIC PRINCIPLES OF ION EXCHANGE
The phenomenon of ion exchange may take place, under certain conditions, in all ionic solids (and in some cases, insoluble liquids). If one considers the ionic solid to be completely dissociated-Le., composed of ions and not undissociated molecules-the surface ions may then be considered as being bound to the lattice with a lower binding energy than the internal ions of the same species. When placed in a polar solvent, these surface ions may become solvated and a further lowering of their binding energy results and a marked dissociation from the lattice may also ensue. If a foreign electrolyte is added to the system, it is logical to expect an exchange to take place between these surface lattice ions and ions of the same charge of the foreign electrolyte. The extent of this exchange will depend upon (1) forces binding the ions to the lattice, (2) relative valences of the two ions entering into the exchange, (3) total concentration of ions, (4)sizes of the two ions, (5) accessibility of the lattice ions, and (6) solubility effects. Commercial ion exchange substances (in particular, the ion exchange resins) are ionic solids in nrhich one of the ionic species (either the anion or cation) is a highly cross-linked, polymeric, high molecular weight, nondiffusible ion whose multivalent charge is balanced by relatively small, diffusible ions of the