79
ANXON EXCHANGE IN W ~ ~ X EORGANIC-AQUEOUS D SOLUTIONS
e
Jensen2and R. M. Diamond”
Luosr crum Radiation Laboralory, Uwiversity of Californ.la, Berkeley, California 9.$7%0
(Received April 6 , 19’7’0)
Publictdhn costs asskied by the C’. S. Atomic Energy Commission
The itlea, developed earlber, that ion-exchange resin selectivity occurs as the result of B competitmn of the exchanging ions for that phase providing the best solvation, a competition won by the ion most in need of solvation (usually that ion with the highest charge density, the most acidic cation, or the most basic anion), has beien applied to the water-dioxane system. Usually, the dilute external aqueous phase provides the better solvating medium, so the smaller ion, or for ions of different structure the more basic anion, goes there, relegating the losing ion to the resin phase. But as dioxane replaces water, the mixed external solution becomes a poorer solvating agent, and since the resin phase preferentially takes up water, the latter becomes relatively a better solvating phase. Thus the ion most in need of solvation decreasingly prefers the external solution; separstion factors should get smaller, and this is what is observed experimentally. I n fact, by properly choosing the mitiire of the macro-electrolyte and of the resin group, the riorrnal aqu&us-phase selectivity order of the halides can be reversed in dioxane-water mixtures.
Introduction Earlier p a ~ e m have - ~ discussed the selectivity shown by (organic) ion-exchange resins with dilute aqueous solutions as a competition between the exchanging ions for solvation in the resin phase and in the external phase. Water molecules usually provide the main means of ionic solval,ion, and so for the usual conditions of a dilute external solution and a (concentrated) resin phase with strongly acidic or basic exchange groups, the external &ohtionfuurnisbes the most favorable medium. That ion whom need for solvation is the more urgent, that is, IviU yield the larger free energy of solvation, goes into that phase which provides the better solvation, and so forces the remaining ion into the other phase even though the latter ion also would obtain a larger free energy of solvation in the better-solvating phase. For cxaniple, %ith 51 family of similar ions such as the halides, the smallest ion goes into the aqueous phase and t b Lrgest ion is pushed into the resin phase, thus achieving a maximum hydration energy and a minimiam free energy for the system as a whole, even though the large? ion is not achieving optimum solvation, For still Zarger monovalent anions such as Reo4and AuCh-, it ha^; also been ~uggested”~Ghat this selectivity order is further enhanced by an entropy eRect; the very Inrge ions tend to tighten up the water structure siirroundiitig them in the sadmeway that (hydrophobic) molecules do. That is, the neighboring water mo1ecules tend to hydrogen-bond out to each other rat,her than in to the large ion, and the zone of the water existing around smaller ions, a disordered region 111 which water molecules can bond either in to the ion (or to the ions’ hydration shell) or out to “ordinary” l ~ u l kwater molecules,6v7disappears. Since this effect would be reduced in the less-structured water of the (oonnwitrated) resin phase, large ions would
tend to be still furt,her squeezed out of the aqueous solution into the resin phase by such an entropy ch-ange, in addit,ion to t,he effect of optimizing the heat, of hydration. Evidence for such behavior is shown by t’be corresponding case of the large t , e ~ ~ ~ ” ~ l ~ ~ l ~ ~ ~ cations iri cation exchange where, in fact,, it, be.8 been shown8that the replacement of Na-t by NPr,:+ or NBu4+ occurs because of a large positive A S and in spite of a positive AH. Thus ithe dilute solution anion-exchange selectivity order expected for the halides and Reo4- would. be F- < CT- < BY- < I--< Reo4--,and this is indeed observed. Furthermore, it has been possible to affect ~ ~nature i~ this order in a predictable way by c h a n ~ the of the exchange group and by s ~ ~ ~ various s t ~ cont ~ ~ t ~ ~ centrated solut’ions for the diluto external phase.? In this paper we want’ to consider t’hes i t u a ~ i where ~ ~ i the water in the system is partially replmed with an organic solvent d i c h does not solvate anioris very well but is miscible with water. Dioxane wag chosen for this first study.
Experimental Section Reagents. The anion-exchange resins used were Dowex 1-X4, 100-200 mesh, a strong-base resin with a (1) W-ork supported under the auspices of t h e U. 8.Atomic En-ergy Commission (2) Slimmer visitor, 1965, 1967, NSF College Teachers Research Participation Program, Cabrilla College, Apt09, California (3) B. Chu, D. C. Whitney, and R. M . Diamond, J , 1nm-g. Nucl. Chem., 24, 1405 (1962). (4) 6.XI. Jensen and R, M. IXa.mond, J . Phzjs. Chem., 69, 3440 (1965). (5) R. M. Diamond and D. C. Whitney in “Ion Exchange,” Vol. 1, J. Marinsky, Ed., M. Deliker, New York, N. Y., 1966, pp 277-351. (6) N. s. Frank and &!I. w. Evans, J. Ci’LETYl. Phgs,, 13, 507 (1945). (7) R.M. Diitmond, J . Phy8. C!k$fl%.,67, 2513 (1867;. (8) G. E. Boyd and &. V. Larsoxi: 9. Amer Ch,em. &c., (1967). I
The Journal of Physical Chemistry, Val.
80
C. H. JENSEN AND R. M. DIAMOND
polystyrene matrix, and Dowex 3-X4, 100-200 mesh, a weak-base resin having as functional groups a mixture of primary, secondary, and tertiary amino groups. The capacity and water uptake of the Dowex 1 were 4.04 mequiv and 1.73 g, respectively/g of dry C1-form resin, arid the corresponding figures for the Do-vvex 3 were 5.5 mequiv and 0.58 gig of dry C1-- form resin. The solutions of LX1, and N(CH&Cl were prepared by volumetric dilution with conductivity water of analyzed stock solutjons of reagent grade materials. The 18Ftracer was prepared at the Lawrence Radiation Laboratory 88-in. cyclotron from conductivity water by the nuclear reaction 160(a,d)18F.The nzBr-was prepared by neutron irradiation of reagent grade LiBr a t the Valiecitos Rcactor. The l3lI- tracer (carrier-free in n'azS03) was purchased from New England Kuclear Corp. The carrier-free 183-184Re04-was chemically extracted from 8, tungsten deflector plate which had been subject to o! arid d bombardment over an extended period of time at the 88-in. cyclotron. Procedure. Batch measurements were made with all tracer anions by placing weighed samples (0.10000.3500 g) of resin and 10.0 or 20.0 ml of solution of known tracer content into 30-ml polyethylene screwcap bottles and sha king for at least 6 hr. Two 2.00-ml aliquots of solution were removed through filter paper and y-counted using a well-type NaI(T1) scintillation counter with single-channel analysis. Samples of the stock solution were also counted to give the initial tracer activity. After correction for background count, and for the mok~turecontained in the resins at 47% humidity (all resin used in the batch measurements was weighed at 47% humidity), the distribution coefficient was calculated in the usual way
i
[(counts/min)
n
I=
1
- (counts/min),quil J/
g 01 dry C1--form-~ resin
!counts/min)eqUil/mlof solution
All experimental work was done at room temperature, 23 f 2". The ion-invawion into the Dowex 1-X4 by the LiCl \vas determined at various concentrations of dioxane. Three small columna were each filled with 1 g of resin a t 47% humidity (17.3% H20). The columns were washed with the dioxane-water-lithium chloride solution until equilibrium was established and were then placed in a centrifuge for 5 min to remove excess liquid. The columns were then washed with distilled water to remove all trace of LiiCl and the C1- was titrated by the Fajaris method, The quantity of LiCl adhering to the outside surface of the resin was estimated by repeating the above procedure using a volume of glass beads of 17043.0 mesh equal t o that of the resin. The uptake of dioxane-water by the resin at various concentrations was determined in the following manner. A weighed amount of dry resin and a weighed amount of solution of known composition were placed in a 2-ml The Jaumal of Physical Chemiskrg, Vol. :y5,No. 1, 197'1
volumetric flask and shaken for about 8 hr to reach equilibrium. The solution T T ~ Sthen filtered through a small polyethylene filter by centrifugation and weighed. (A correction was made for solutioii caught in the filter and adhering to the walls.) The concentration (mole fraction dioxane) of the solution before and after eyuilibrium was determined by measuring its index o€ refraction with a Bausch and Lomb Abbe refractometer. From these data, the weight and composition of the solution taken up by the resin were calculated.
Results A plot of the resin-phase composition (dioxane mole fraction) for the Dowex 1-X4 resin vs. the equilibrium external-phase dioxane mole fraction is shown in Figure 1. A plot of the LiCl resin invasion in meq of Cl-/g of dry Cl--form resin vs. the equilibrium extcmal-phase dioxane mole fraction €or the Dow-ex 1-X4 rosin with a constant concentration of 0.0104 &I 133in the external solution is shown in Figure 2. Figures 3-3 show the distribution ratio
D
=
mequivig of dry Cl--form resin __________ mequiv/nil of s o h tion
for tracer F-, Br-, I-, and Reo4- as a function of the dioxane mole fraction with LiCl of 0.0104 M (Figure 3), 0.0311 M (Figure 4), and 0.104 M (Figure 5) as the macro-electrolyte in the external phase. Similarly, Figures 6-8 show for the same tracers as a function of the dioxane mole fraction with K(CH,)dCl of 0.0101 M (Figure 6), 0.0303 M (Figure 7 ) , and 0.101 AT (Figure 8) as the macro-electrolyte in the external phase. Finally, Figures 9 and 10 give examples of the results obtained with the weak-base resin Dowex 3-x~.A plot of D for tracer Br-, I-, and Reo4- vs. dioxane mole fraction in the external phase is shown in Figure 9 for 0.0104 M LiCl and in Figure 10 for 0.0101 M N(CH&-
n
c1.
Discussion The partial replacement of water in an aqueous solution by dioxane will certainly have at least three effects. First of all, since the cyclic ether dioxane
0'
\
0 /
CHZ-CHZ does not possess any (acidic) hydrogen capable of hydrogen-bonding to (basic) anions, as does the water molecule, the mixed solvent will become a poorer and poorer solvating agent for anions. That is, if we can think of single-ion activity coefficients, these will rise steeply for anions with an increase in dioxane c ~ n l e n t , ~ (9) E. Grunwald, G. Baughinan, and G. Kohnstsm, J . Amer. %em
see., 82, 5801 (1960).
ANIONEXCEANGE IN ! k X E I ) ORGANIC-AQUEOUS
81.
SOLUTIONS
1.6
2
1.2
,"
0
0)
0.8 1 0.4 0.0
0.4 0.6 0.8 Dioxane male fraction (solution)
I .0
Figure 1. The mole fraction of dioxane in the resin phase and the total weight of solvcbnt (dioxane water) uptake per gram of dry yesin for Dowex 1-X4 us. the mole fraction of dioxane in the equilibrium solutiton.
+
IO
--A-
0.0
0.1
0.2
0.4 0.5 0.6 Dioxane mole lrociion 0.3
0.7
0.8
Figure 3. Plot of D us. dioxane mole fraction in the solution for 0.0104 N LiCl and Dowex I-X4 resin and the tracer anion: F-, V; Br-, 0 ; I-, 0; and Re04-., A.
a
IO'
Dioxane mole froction Figure 2. Nonexchang;e or resin-invasion electrolyte in Dowex I-X4 resin from 0.0104 1%' LiGl solutions of dioxane-water us. the mole frac-lioii of dioxane (disregarding the LiCl) in the equilibrium solution. The left-hand ordinate scale is in mequiv of Cl-/g of dry 61--form resin; the resin capacity is 4.04 meqiiiv/g of dry Cl--txm resin. Also shown is a plot of the dielectric constatit of a dioxane-water solution us. the dioxane mole fraction; the ordinate is to be read off the right-hand scale
and the more so, iche smaller and/or more basic the anion. Secoridly, the dioxane molecule will act as a base toward the wBter molecule, competing with the anions for hydrogen-bonding t o the water, and breaking up the hydrogen bonded water structure. Finally, the addition of dioxane, because of its low dielectric constant and because of the destruction of the water structure, will cause an increasingly marked drop in the dielectric constant of the mixed solvent.10 (This is shown in Figure 2.) For all of these reasons, the dioxane solutiorr provides a much poorer solvating medium for anion53 than the original water solution,
[-
i
LA-/ -1
io0
00
I
0.1
0.2
0.3
0.4
0.5 0.6
Dioxane rnoie f r o c l i o n
Figure 4. Plot of D us. dioxane mole fraction in the solution for 0.0311 M LiCl and Dowex 1-X4 resin and the same tracer anions as in Figure 3.
both chemically (through bonding to the anions) and electrostatically (through the effect of the lower dielectrio constant in raising the electrostatic free energy of the ions), and the effect is more marked the higher the basicity and charge density of the anion. The results of this poorer solvating ability have been remarked on by a number of workers.9s11 But obviously then, to be able to predict the effect of dioxane on the selectivity order of a family of anions, one must first determine the distribution of dioxane between the resin phase and the external solu(10) C. A. Kraus and R . M. Fuoss, J. Amer. Chem. Soc., 5 5 , 2 1 (1933). (11) D. Feakins and D. J. Turner, J. Chem. Soc., 4986 (1965).
The Journal of Physical Chemistru, Vol. r5, N o . 1 , 1971
C. H. JENSENAND R. M. DIAMOND
82
-10.0 -I i
I100 0.0
0.1
0.2
0.3
0.4
0.5
0.6
IO
1 -
--.1 0.1
0
0.2
0.3
0.4
0.5
0.4 '
0.5
0.6 L
Figure 7. Plot of D us. dioxane mole fraction in the solution for 0.0303 M (CH3)rNCl and Dowex 1-X4 resin and the same tracer anions as in Figure 3.
0.6
Dioxone mole fraction
Figure 6. Plot of D us. dioxane mole fraction in the solution for 0.0101 M (CE1a)r NC1 and Dowex 1-X4 resin and the same tracer anions as in Figure 3.
tion. Figure 4. shows that with Dowex 1-X4, C1--form, the resin phase preferentially rejects dioxane and takes up water. Up t o an external solution dioxane mole fraction of 0.8, the resin-phase dioxane nzole fraction is that of the equilibrium solution. This only 1/2 to behavior seems to be usual for most anion resins and for most of the mixed solutions studied, but not all.12-1e Why should this be so? We do not know the answer in detail, but surely thiry is a combination of the hydration needs of the ions in the concentrated resin phase, and an electrostatic effect of the high-charge density there. The Journal of Physical Ct5emistry, Vol. 76, No. 1,IS71
0.3
Dioxone mole froction
Dioxane mole fraction Figure 6. Plot of D ZJS. dioxane mole fraction in the solution for 0.104 M LiCl and Dowex 1-X4 resin and the same tracer anions as in Figure 3.
0.1 00.2
0.0
0.1 0.2 0.3 0.4 Dioxone mole fraction
Figure 8. Plot of D us. dioxane mole fraction in the solution for 0.101 M (CHa)4 NC1 and Dowex 1-X4 resin and the same tracer anions as in Figure 3.
The first part of the previous sentence is obvious; the second part perhaps needs further explanation. The (12) H.P. Gregor, D. Nobel, and M . H. Gottlieb, J. Phys. Chem.? 59, 10 (1955). (13) 0.D. Bonner and J. C. Moorefield, &id., 58, 555 (1954). (14) C. W.Davies and B. D. R. Owen, J . Chem. Sac., 1676 (1956). (15) H.Rltckert and 0. Samuelson, Acta Ch,em. Seand., 11, 303 (1957). (le) Y.Marcus and J. Naveh, J. Phys. Chem., 73, 591 (1969).
ANIONEx:cIrArwx
OC
0.1
IN
0.2
MIXEDORGANIC-AQUEOUS SOLUTIONS
0.3
0.4
0.5
0.6
0.7
0.8
Dioxone mal froction
Figure 9. Plot of D us. (dioxanemole fraction in the solution for 0.0104 .iMLiCl and Clowex 3-X4 resin and the tracer anions: Br-, 0 ; I-?(1); and Ke04-, A.
0.0
01
02
03 0.4 0 5 0.6 Dioxone mole froction
1
1
0.7
0.8
Figure 10. Plot CISD us. dioxane mole fraction in the solution for 0.0101 M (CH?)*NC1 and Dowex 3-X4 resin and the same tracer anions as irr Figure 9.
addition of dioxane results in a lowering of the dielectric constant of the mixture. But the resin phase, with its high density of chargca, both fixed resin sites and mobile counter ions, will have a much larger increase in electrostatic free energy with lowering of the dielectric constant than will tho much more dilute external phase. To keep the total sysl,em free energy at a minimum value, water must 1prefert;ntiallygo into that phase with the highest concentratioin of charge to provide electrostatic (but not necessarily chemical) solvation. However, all of this water is not likely chemically bound directly to the quaternary alkyl ammonium ions of the resin, but still bas same freedom to solvate chemically the counterions, Thus, as the concentration of dioxane in the total s,ystan?n increases, the resin phase absorbs a higher proportion of water than exists in the external solution, and the superiority of the external phase over the resin phase for soYvai ing anions decreases. With water done ELS the solvent medium, anions will obtain the best solvation in the dilute external phase and so will go into that phrase in the order of their size or basicity, resulting in the resin selectivity order F- < GX- < Br -