3972
K. -1. KRAUS,F.NELSON.\ND C>. E. MOORE [COSTRIBUTIOS FROM THE OAK
Anion-exchange Studies.
lrol. 77
RIDGENATIONAL LABORATORY, CHEMISTRY D I V I S I O N ]
XVII. Molybdenum (VI), Tungsten (VI) and Uranium (VI) in HCl and HCl-HF Solutions1i2
BY KURTA. KRAUS,FREDERICK NELSONAND GEORGEE. MOORE RECEIVED FEBRUARY 23, 1955 The anion exchange behavior of Mo(VI), W(T.1) u i d G(V1) \.vas investigated with a strong base quaternary amine anionexchange resin in HCI solutions and in HC1-€IF mixtures. The elements adsorb strongly a t high HC1 concentrations. .idsorption decreases with decreasing -1.1HCI aiid, except for W( VI) which tends to precipitate, becomes negligible near 0 . 5 31 HCI. The adsorbabilities in HCl solutions containing 1 ;If H F differ widely and separation is readily achieved. A numher of separations of these elements from each other and from other element., p:irticularly Fe(III), are illustrated. The effect of cross-linking on the separatioris is discussed.
In continuation of a systematic survey of the anion-exchange behavior of the metals in HC1 solutions, the adsorbabilities of LIo(VI), W(V1) and U(V1) have been determined. Conditions were not found for the complete separation of these elements in this medium even with resins of different cross-linking. Studies in HC1-HF mixtures were carried out since, on the basis of earlier work with elements of the fourth and fifth groups,3 considerably better separability may be achieved by the use of such niixed systems thrrri with HCl alone. Experimental
acidificatiox. On the basis of a number of column experiments the tracer appeared t o be essentially free of other activities. The L Y J 3tracer was found by a-pulse analysis t o contain less than 1% of non-uranium a-activity.8 The separations experiments were usually carried out with small columns (ca. 1 ml., 1 t o 2 inches high). Small aliquots (ca. 0.5 cc.) of the test solutions were permitted to seep into the columns which had becn pretreated with the same acids its used in the test solutions. The solutions usu:illy contained the appropriate tracers to facilitate analysis. Metal concentrations varied from tracer levels to ca. 10-3 .lL, although higlier concentrations could probably be used provided the columns are not overloaded. Flow rates in the separations were usually between 0.3 and 1 cm./min. The experiments were carried out in a n air-conditioned room a t 2,5 =i= 2 ” . For the experiments involving fluoride solutions, fluoride-resistant containers were used, i.e., Lusteroid test-tubes, columns made of Tygon tubing with retaining plugs of lucite shavings, plastic burets and pipets. Except where otherwise noted, the same batch of quaterna amine polystyrene divinyl benzene resin (Dowex-I, 10‘ Di‘B, 200-230 mesh) was used as in the earlier Further details of the experimental procedure will be given in the rliscussion of the various systeiris.
As in earlier work,4 adsorbabilities were determined either by the column or by the equilibrium method, depending on the magnitude of the distribution coefficients D (amount per kg. dry resin/amount per liter of solution5). For low values of D the equilibrium method is less accurate than the column method, while the latter becomes relatively cutnbersome when D becomes large. A shift from the m e method t o the other was niade in the neighhorliood of D = c a . 20. Results and Discussion The column method readily yields38 n volume distrihutioii coefficient DV6which is related to D by the relationship D,. 1 . Mo(VI), W(V1) and U(V1) in HC1 Solutions. = p D , where p is tlie bed density. For present purpo~es --The results of the experitnetits are summarized in l / p = 2.2 liters per kg. dry resin nm taken. Fig. 1 as a plot of log D vs. HC1, where D is the Metal concentrations were deterniined radiometrically, using as traccrs MoQQ ( 3 , A/, T J / ?= 67 hr,); W ’ g 9 (a, y , 7-1 i2 = weight distribution coefficient (amount per kg. 73 d.) and 1.T233 ( a , Ti/* = 1.6 X 103yr.). dry resin/amount per liter of solution). The &Iog9tracer was prepared hi- neutron bombardment of The adsorbability of ?rIo(VI) is small in ca. 0.5 reagent grade MoOn in the O R F L Low Intensity Training A 1 HCI (D = ca. rises rapidly to a maximum Reactor ( L I T R ) . The irradiated oxide was dissolved in 0.5 h.1 NaOH and aliquots of this “stock” solution w r c near 5 -11 HC1 ( D = ca. 250) and then slowly deused after acidification. IIalf-life determinations shou.etl creases with increasing 111 HC1. The reliability of the tracer t o be of sntisfactoq- purity. Since tlie tracer the values of D a t the highest and lowest HCl concontained the daughter actix-ity T C “(~ T i / %= 0 lir.) final counting way done after secular equl!ibrium had k e n centrations is somewhat in doubt. At high ill reached. RJ- ertrapolatiori of t h e counting data to “zero HC1, hlo(V1) tends to precipitate while at very time” (time of separation of the phases) it t m s dernoii.;trated low M HCl (dashed line, Fig. 1) slow equilibria that under a11 conditions studied technetium I\-J\ more cause complications. Thus below 1 Jl HC1 310strongly adsorbed than mol!-l~knum. High specific activity TVL8”tracer \vas obtained: iii 0.7 ~b1 (VI) tends to tail and elutes in the form of two KOH and portions of this “stock solution” were used after bands which follow each other closely (e.g., in 0.5
a),
(1) This document is based on work performed for the U. S. Atomic Energy Commission a t t h e Oak Ridge Sational Laboratory. (21 Previous paper: F. Selson and K . A . Kraus, THIS J O U R N A L , 77, 1391 ( I Y S ) , (3) (a) K A . Kraus and G . E. IIoore, t b i d . , 73,9 (19.51); (b) 73, 13 (1951); (c) 73,2H00 (1991); (d) 77, 1353 (1955). (4) See, E . & , R . A. Rraus, F SelTon a n d G . nr. Smith, J. Phys. Che?n., 5 8 , 11 (1954). ( 5 ) Values of D were computed on the basis of weights of resin dried over Anhydrone in a vacuum desiccator a t 603. Metal concentrations were held sufficiently lorv t o permit measurement of D at less than 1Wc loading of t h e resin (6) T h e volume distribution coefficient LA,can be calculated from the relationship D , = ( V l d d ) - i , where V i s the volume a t which the eluted ion appears in maximum concentration in the effluent, where A IS the cross-sectional aiea, d t h e length of the column (;.e., Ad I S one column volume) and !+here i is t h e fractional interstitial space o f t h e column. For present computations i = 0.42 was assumed. ( 7 ) We are indebted t o t h e ORSI, Isotopes Division for the tracer.
31 HCI maxima appear a t 1.4 and 2.4 column volumes). This implies that a t least two species o f Xo(V1) exist under these conditions, which are not in rapid equilibrium with each other, as might be expected from the rather complicated hydrolytic behavior of Mo(V1). When elution is carried out with 1 Jf HC1, Mo(V1) appears as a symmetrical band. Although D (see Fig. 1) is somewhat higher a t this acidity than at lower HC1 concentrations, removal seems to be optimal near 1 M HC1. The adsorbability of TL’(V1) at HGI concentrations larger than 6 ;lL first increases with -21 HC1 (8) We are indebted t o Mrs. Myrlene Davis of t h e O R S I . Instiuments Division for the a-spectiometer analyses (9) For characterization of the resin, see K . A . K r a u s and ( > . 15. LIoore, THISJ O U R S A L , 7 5 , 14J7 (1953).
,lug. 5, 19.55
.INION EXCI-IAINCE OF Mo(VI), W(V1)
and reaches a shallow maximum near 8 M HCl. For HC1 concentration less than 6 Ad, D tends to decrease with decreasing M HC1. However, adsorption becomes markedly dependent on equilibration time and elution becomes unsatisfactory. Thus, if W(V1) is adsorbed a t high HC1 concentrations and the column treated with 1 to 3 A 1 HCI, part of the W(V1) is removed within a few column Idumes, though most of it remains on the column. This behavior at low Jf HC1 is probably connected with the formation of hydrolytic polymers of IV(V1) and their tendency to precipitate. =Idsorbability of \Y(.\.I) is generally less than that of Mo(V1) and its adsorption maximum occurs a t higher -1P HC1. The adsorbability of U(VI) rises rapidly with increasing M HC1 from essentially negligible values near 0.5 ;21HC1 to a shallow maximum near 10 A 1 HC1. Up to ca. 4 M HCl the values of D of Mo(VI) and U(VI) are very similar, although U(VI) at higher JP HC1 adsorbs considerably more strongly than Mo(V1). 2. Separations in HC1 Media. Effect of Crosslinking.-The strong adsorbability of Mo(V1) , Ur(VI) and U(V1) a t high AJ HC1 permits separation of these elements from the many elements which under these conditions show little adsorption. Advantage may also be taken of the low adsorbability of Mo(V1) and U(V1) a t low M HC1 to effect separations from elements which adsorb strongly under these conditions. Unfortunately, removal of I17(VI) (and to some extent of Mo(V1)) from columns is complicated by its hydrolytic properties, as discussed in section 1. Furthermore, the similarity of the distribution functions of Mo(VI) and U(V1) a t low JJ HC1 makes their separation from each other difficult. At high M HCl Mo(VI), W(V1) and U(V1) show considerable differences in adsorbability. However, these differences cannot be exploited for separations, since the values of D are too high. Since Herber and Irvine'O recently showed that adsorbabilities of complex ions are greatly dependent on the cross-linking (yoDVB) of the resin," it appeared that a t lower cross-linking than the present 10% DVB, the distribution coefficients a t high M
U(\'I)
AND
3973
IO4 I
5
I
,
I
I
I I
lo3 Q'
LL
:
2
0
l
0
2
4
6
8
1
I
IO
12
MOLARITY OF HCI.
Fig. 1.-Adsorption of Mo(VI), n'(V1) and U(V1) from HC1 solutions.
HCI might become sufficiently low for convenient separations. As shown in Table I, where the volume distribution coefficients D,, are listed, adsorption a t 1% DVB appears to be sufficiently low to permit separation of W(VI) and U(VI), while the values of Dv for Mo(V1) and W(V1) at lY0 DVB, though low, are not sufficiently different for convenient separation with short columns. Typical separations with this resin of low crosslinking are shown in Fig. 2. An aliquot of a solution containing LV(V1) and U(V1) in 12 M HCI was added to a column of a low cross-linked resin (Dowex-1, 1% DVB, 50-100 mesh). Elution was carried out with 12 M HCI. With this relatively short column (5.8 cm.) separation was barely achieved (Fig. Za). After removal of u'(V1) more TABLE I rapid elution, and hence a sharper elution band, of DEPENDENCE OF DISTRIBUTIOS COEFFICIENTS O Y CROSSU(V1) could have been obtained with an HC1 soluLIYKING tion of lower concentration. Vol. distribution DY A separation of U(V1) from Fe(II1) is shown in coefficient (10% DVB) HCI, (DV) D Fig. Zb. This separation is quite difficult with a Element M l o % D V B 1% DVB (l%DVB) highly cross-linked resin in HC1 media since Fe 12 64 6.5 10 Mo(V1) (111) a t high M HC1, where D F e ( I I I ) >> Du(vI), 12 33 4 8 WVI) adsorbs very strongly12while a t low M HC1 where 12 712 19.5 37 U(V1) adsorption is low enough for convenient elution U(V1) 3.05 42 1.8 23 DF~(III) differs only little from D u ~ v I , . An aliquot Fe(II1) 3.05 122 6.4 19 of a solution containing Fe(II1) and U(V1) in 5 M (10) R. Herber and J. W , Irvinc, Jr. fprivate communication). HC1 was added to a column of the low cross-linked We are indebted t o Drs. Heiber and Irvine for making this work availresin. On elution with 3 L$f HC1 U(\TI) appeared able t o us before publication. in the effluent with maximum concentration near (11) These observations are in qualitative agreement with earlier 2.2 column VChrnes. I t was reasonably free Of findings on the effect of cross-linking on adsorbability of simple ions, particularly cations on cation exchange resins. See c g., R. hi. Fe(II1) which appeared in a broad band with maxiWbeaton and W. C . Bnuman, Znd. EW. Chem., 43, 1089 (1951); mum concentration near cu. 7 column volumes. H. P. Gregor, THISJOURNAL. 73, 643 (1951); E. Glueckauf, PYOC. As in the previous experiment a shift to lower M Roy. Sor. ( L o n d o n ) , 214A, 207 (1952); and K. W. Pepper and E i.
Reichenberg, 2. Eleklrochem., 57, 183 (19531.
(12) G , E . Moore and K. A. Kraus, THISJ O U R N A L , '72, 5792 (1950).
K. A. KIZAUS, F. NELSONAXD G. E.MOORE
3!04
Vol. 77
1G4
A
5
a' $
w3
103
5
-
LL
g
2
3
102
m 3
E
5
G 2 10
2
2
L U V B E R OF COLIIVN VOLUMES.
Fig. 2.-Separations
with low cross-linked (1% DVB) resin.
HCl after removal of U(V1) would have eluted Fe(II1) quicker and in a sharper band. To summarize, by the use of resins of low crosslinkage distribution coefficients may be sufficiently lowered to permit some separations which cannot readily be achieved with a highly cross-linked resin. This effect is expected to be general. 3. Mo(VI), W(V1) and U(VI) in HC1-HF Mixtures.-A study of the adsorption behavior in HC1-HF mixtures appeared promising since, on the basis of earlier work with the elements of the fourth and fifth g r o ~ p s addition ,~ of fluoride was expected to affect greatly the adsorbabilities of these elements of the sixth group, magnify their differences, and avoid difficulties resulting from hydrolytic precipitation. A series of equilibration and column experiments was carried out in 1 Af H F solutions as a function of M HC1. The results are summarized in Fig. 3. The adsorbability of Mo(V1) is considerably decreased by the addition of l M H F to a minimum D = 1s near 2 M HC1. At lower M HC1 the distributiw coefficients of Mo(V1) rise rapidly with decreasing HC1. A similar effect of HF is found for W(V1). An adsorption minimum for n'(V1) occurs near 8 111 HC1 (D = cn. 7 ) where elution of W(V1) is practical, though not necessarily optimal. A search for optimum conditions (D 6 2 ) would involve a more detailed study of the effect of H F coilcentration on adsorbabilities. The effect of H F on the adsorbability of U(V1) is relatively small a t high Jf HC1 but increases with decreasing ill HC1. At HC1 concentrations less than 1 ill,D of U(V1) rises rapidly with decreasing M HCl. In this respect U(V1) is similar to MO(VI) and W(VI), although the adsorbability of
Fig. 3.-Adsorption of Mo(VI), iL'(V1) and U(V1) from HCl-HF solutio~is(1 III €IF)
U(V1) a t low llf HCl i n the presence of H F appears to be considerably less than that of the other two elements. The marked adsorption differences of these elements of the sixth group in HC1-HF mixtures permit satisfactory separation of these elements. After this study had been completed, a publication by Hague, Brown and Brighti3 on the anionexchange separation (Dowex-1) of Ti(IV), h'b(V), Mo(IrI) and \V(lrI) in HCl-HF mixtures appeared. Our data in fluoride solutions are in good agrrement with theirs in the HC1 concentration range in which our studies overlap. 4. Separations in HC1-HF Mixtures. (a) Separation of Mo(VI), W(V1) and U(VI).-Inspection of Figs. 1 and 3 reveals that :I varietyof conditions can be selected for separatioris involving these elements. Two typical separations are demonstrated in Fig. 4. An aliquot of a solution containing \Y(VI) and I;(VI) in 9 ,If HC1-1 JI H F w:is tldded to a small column of the highly cross-linked rcsin. On elution with 0 J1 HC1-1 11 H F (see Fig. 4a'i \\-(VI) appeared in the effluent in a relatively sharp band within the first few column volume^ U(V1) remained strongly adsorbed mi the coluniii and mas eluted with dilute HCl (0.1 -11). In Fig. 4b the separation of l l o ( V I ) , \V(VI) and U(V1) from each other is illustrated. Conditions were selected to reverse the elution order of \ . ( V I ) and U(V1). An aliquot of a wlution containing these elements in 0.5 itf HCl-1 AI H F was added to a column. On elution with 0.5 JI HCl-1 X HF, U(T.'I) appeared in the effluent in a narrow band (13) J. L Hdgue. E D Brown a n J 11 .I Bur Sfanrlorrls, 53, 261 (1954).
Bri:ht.
.f
Kesrorch S ? t 1
.hg. 5, 1955
-ANION
EXCHANGE O F MO(VI), W(v1)
AND
U(v1)
3975
The separation of Fe(II1) from Mo(V1) can similarly be achieved. For example, if a mixture of these elements in 9 AI HC1 is passed into :I column, both elements are adsorbed while non-adsorbable elements pass through the column. Fe(111) may then be eluted with HC1-HF mixtures of low HCl concentration (e.g., 0.01 AI HCl-1 A I HF) under conditions where Xo(V1) remains adsorbed. Mo(V1) may then be removed with cu. 1 M HCl in the absence of HF. By combination and modification of these techniques Fe(II1) can be removed from both Mo(V1) and W(V1) as illustrated in Figs. 3a and b. For N' MEER OF COLUMN VOLUMES. f the experiment described in Fig. 5a a small aliquot of a solution containing Fe(III), \['(VI) and Mo6- 5 1 1 , $ OCIMHC:, I M H F COLUMN SIZE (VI) in 0.5 M HC1-1 AI H F was added to a column. 0!7crnZXS6cm t , On elution with 0.5 ;If HC1-1 JI HF, Fe(II1) practically immediately appeared in the effluent while Mo(V1) and W(V1) were retained by the column. W(V1) was removed with 9 ill HC1-1 JI H F and then Mo(V1) with 1 A I HC1. 4-
I
1-
~
0
5
Fig. 4 -Separations
10 15 NUMBER OF COLUMN VOLUMES.
20
involving iUo(\rI), UI'(V1) and U(V1).
almost immediately, while LIo(V1) and W(V1) remained on the column. Removal of "(VI) was eflecied with 0 31 HC1-1 AI H F and removal of hlo(V1) with 1 JI HCl. The Mo(V1) band tailed slightly, though not seriously. The addition of fluoride in the beginning of a separation is often undesirable since some elements may be present which are precipitated by it (e.g., alkaline earths or rare earths). However, adsorption of l I ? ( V I ) , W ( V I ) and U(V1) can be carried +MHF,O.O1MHCI i MHCi out from HCl sulutions (e.g., 0 111 HC1) in the ab7 A sence of fluoride and interfering ions washed out. 3 W(V1) may then be eluted with 9 AI HCl-1 N HF, U(V1) with 1 JI HCl-1 114 H F and hIo(V1) with 1 JI HC1. 2 (b) Separation of Fe(II1) from Mo(VI), W(V1) and U(VI).-In a series of equilibration experiments 4 with Fe(II1) (Fe59)in 1 M H F solutions containing varying amounts of HC1, it was established that 1 2 3 4 5 6 7 8 9 H F decreases DF~(III) though only slightly in concentrated HCl. The effect increases with decreasKuniber of coluniii volunies. ing AI HCl in a manner similar to the effect of H F Fig. 5.-Separations involving Fe(lII), &Io(\.I), ll'( \'I j a i i t l on DTJ(VT). However, in contrast to the behavior U(V1). of U(V1) (as well as Mo(V1) and bY(V1)) D F ~ ( I I ~ ) in the presence of 1 A I H F reniains low, a t least Conditions may also be chosen to reverse the A 1 HCl. down to order of elution of Fe(II1) and W(V1) (Fig. 5b). For the separation of Fe(II1) from W(V1) ad- An aliquot of a solution containing Mo(VI), Wvantage may be taken of the high adsorbability of (VI), Fe(II1) and Cs13' in 9 111 HCI-1 M H F was Fe(II1) and the low adsorbability of W(V1) in 9 &I added to a column. In this experiment C S ' was ~~ HC1-1 111 HF. If a 9 111 HCl-1 41 H F solution added to demonstrate the behavior of a typical containing these elements is passed into a column non-adsorbable element. On elution of the column and elution carried out with 0 M HCl-1 1cI HF, with 9 A1 HCI-1 ill HF, CsI3' and W(V1) appeared W(V1) can De renioved while Fe(II1) is adsorbed in separate bands with maximum concentrations strongly. Removal of Fe(1II) can be effected with a t 0.4 and 3.7 column volumes, respectively. MoHCl of low concentration, e.g., 0.5 M HCl. (VI) and Fe(II1) were retained by the column.
3976
K. A. KRAUS,F. NELSONAND G. E.~ I O O K E
Removal of Fe(II1) was achieved with 0.01 JT HC1-1 M HF, and removal of Mo(V1) with 1 M HC1. Separation of Fe(II1) and U(VI), which is quite difficult a t moderate HC1 concentrations where these elements adsorb similarly even in the presence of HF, can be achieved readily a t low HC1 concentrations where U(V1) adsorbs strongly and Fe(II1) shows only negligible adsorption. For the experiment described in Fig. 5c a small aliquot of a solution containing Fe(II1) and U(V1) in 0.01 ~11HC1-1 -If H F was passed into a column. Fe(II1) was removed with 0.01 JI HC1-1 31 HF and then U(V1) with 1 111 HC1. The separation of Fe(II1) from Mo(YI), \V(\-t) :tiid U(V1) is also feasible a t low HCl concentrations in 1 AI H F where Fe(II1) does not adsorb and these elements of the sixth group adsorb strongly. Unfortunately, however, the separation of these four elements from each other is difficult to achieve in one pass. I t was hoped that one could adsorb these elements from concentrated HCl, remove W(V1) as in the experiment Fig. 3b with 9 III HC1-1 *If H F , remove Fe(II1) with 0.01 AI HCl-1 H F , but retain C(V1) oii the column under these conditions. -1lthough U(VI) adsorbs strongly from this last medium, desorption of U(VI) occurs while the column adjusts itself from 0 11 HC1-1 J1 HF to 0.01 -11HC1-1 A1 €-IFsince between these extrenies is a regioii of HC1 cmceritrations where U(V1) shows negligible adsorption. 5 . Some Considerations Regarding Species in Solution.-The great similarity- of the distribution functions of Mo(V1) arid U(V1) and particularly the similarities of the slopes d log D, d Y Y I H C Ia t low :lI HCl suggest that the initial coniplexing reactions of these elements are similar and that the equilibrium constants for these chloride complex equilibria are approximately the same. This i n plies that the uncomplexed 11o(V1) species in the more acidic soliltions (-11HC1 3 1) is lIoO2++ It has in analogy with the uranyl ion C02t'. been known for soiiie time that a t acidities larger than ca. 0.1 1Io(VI) exists in the fciriii o f positively charged ions, probably polymeric. ' ' Kecently it has been suggested from solubility measurementsIdb that positively charged polymeric llo(V1) species occur a t least up to ciz. 1.2 JI acid (HC104j. ITith respect to formation of such polymeric cations 1Io(irI) thus appears to be similar to U(V1) l5 although the region of pqlyiner forniation extends t o considerably higher acidities follTo(V1) than for U(V1). 'The slow anion exchange c,cjuilibria of Mo(V1) a t 1!w acidities arc consistent with the postulated existeiicc of polyiiieric. species. It is tempting to concliide froin the anioii-exclialige results that these polymers depolymerize to lIoOz++ (and its chloride complexes) a t higher acidities, a t least a t the low 1IojYI) curiceritratiurls used in the equilibration experiments ( c ~ . AI i .
Vol. 77
The uncomplexed species of Li'(V1) may also become monomeric iV02++, although only a t much higher acidities than for Mo(V1). At acidities less than G -11the anion-exchange results are consistent with the assumption that positively charged polymeric ions (which may slowly precipitate) predominate, even a t the low W(V1) concentrations used here ( ;ITin the equilibration experiments). Xt the present time relatively little can be stated definitely regarding the negatively charged chloride complexes which are formed by these elements. The complexes are probably of the type X02C1,,-" + and for the negatively charged complexes fz = 3 or Iis likely. If the position of the adsorption maxima is taken as an indication of the relative strength of the chloride complexes the order becomes SIo(L'1) > iT(V1) > U(V1). Comparison of the distribution coeficients iii HC1 (Del) antl HCl-HF solutions (Up) gives soinc iiiforinntion regarding the fraction of the metal in the form of fluoride complexes. Let mt and m; be the stoichiometric concentrations of the metal in the aqueous arid resin phases, then the distrihution coeficients can he expressed as
where subscripts C1 and F refer to the nieasureinetits in HC1 and HC1-HF solutions, respectively. If we desigtiate the concentration of species by parentheses, consider only miiiionieric species and suin over a11 species, the stoichiometric concentrations can be expressed as I1
(illtIC1
(sIc1i)Cl;
= i=ii
+ c ( M F ~ ) F(2) n'
?A
(ft2t)p
(SIC~,)E.
= i=O
i=1
Combination of equations 1 and 2 yields
If D F antl D C Iare obtained a t the same chloritlc conceiitration, if ~ n :is small conipared with the capacity (low loading, linear isotherm) and if the assumptions are made that only chloride complexes are adsorbed and that the activity coefficient ratios for the complex equilibria and resin equilibria are not affected by the presence of HF. then Z(hlC1;)F Z(1ICli)cl -_ _~~(I/l;iF
im:lc,
(4)
Siiice furthermore S(MF ' I k- = Z(51Fi),(ilct1,: -_2 (?I?;JF
(nzt)F(mt;r
1- q-- N F i ) Y [IF
(,?,I
(7l7th
equation ( 3j simplifies to
which gives the fraction of the nietal iu the form of fluoride complexes ( 2 ( h E ? l ) ~ , / ( ~ ~ast ) a~ )function of the ratio D F ~ D c I . As mentioned, these computations are of course applicable only under conditions where adsorbable fluoride coniplexes are either riot formed or show only negligible adsorption. Thus t h e cornputations prohnhly apply
(14) See e g. (a) G Jander and K . F. J a h r , Kolloid B e d . , 41, 27 (1935); (b) hl 31. Jones, TIIISJ O I X N A L , 76, 4233 ( I O j l i . (1;) See e s , , (a) L . G. I.ungsw.orth and D. A . XIcInnes. Report A:i!',U, Kat. Defense Res Committee, Ofice of s c i , Re, a i d l i e v , h-~\,., !!149; (b) J Sritton, J . S a t i o n . i l R e s Caiincil o f Canada S o . ICil2. port .&I
