3852
17~)i.so
D. D. PERRIN
As another methodologically interesting application, eq. 11 can be made use of for the stochastic charaCteriZatiOn Of relative velocities. For instance, the rate of motion (relative to the ambient ocean) of ships on the high seas might be determined by comparing the limiting current of oxygen obtained in a sample of sea water in a "calibration assembly" aboard, with the limiting current measured using [CONTRIBUTION FROM
THE JOHN
CURTIN S C H O O L
OF
the same indicator electrode imnersed alongside the ship in the ocean. Acknowledgment.-This in\,estigatiorl was SUI,ported in part by Research Grants H-2342 alld H234°C) from the Sational Heart Institute, S a tional Institutes of Health, Public Health Service. ITSIVERSITY
PARR, PA.
MEDICAL RESEARCH, THE-AUSTRALIAN NATIONAL
U N I V B R S I 1 U]
The Ion Fe (CNS)2+. Its Association Constant and Absorption Spectrum BY D. D. PERRIN RECEIVEDDECEMBER 26, 1957 Association constants of t h e complexes FeCNS + + and Fe(CNS)z+ have been calculated from absorbance and oxidationreduction potential measurements. Values at 18" by t h e two methods are K1 = 145 and 133; K 2 = 14 and 10, respectively, at ionic strengths of 0.56 and 0.65. Estimates at p = 0 are Klo = 1090, Kz0 = 40. The absorption spectrum of Fe( CNS)2 shows a maximum near 4850 A., with a molar absorbance index of 9800. Fe(CSS12' is t h e major light absorbing species for thiocyanate concentrations greater t h a n 0.04 M . +
Dilute aqueous solutions of ferric and thiocya- gest that under the present experimental condinate ions contain the FeCNS++, tions about 27, of the ferric ion not present in for which consistent values of the association thiocyanate complexes would exist as ferric per~ o n s t a n t and ~ - ~the molar absorbance indexdp6have chlorate ion pairs. This would not significantly been derived from spectrophotometric measure- affect the results obtained. ments. Although qualitative evidence from a Experimental Methods range of experimental methods indicates that at Reagents.--All rcagents were of A.R. grade and \vcrc higher thiocyanate concentrations complexes of used without further purification. Stock sulutioiis of ferric higher thiocyanate content are p r e ~ e n t , ' . ~few , ~ . ~perchlorate were prepared from ferric ammonium sulfate b y quantitative studies have been reported. Esti- precipitation of ferric hydroxide with amnionin; the premates of the association constants of the complexes, cipitate was washed repeatedly with distilled water a t the centrifuge and then redissolved in excess perchloric acid. FerFe(CNS).(3-x)+, (x = 1 to R), have been made using rous perclilorate was freshly prepared by double deco~nposolvent distribution6 and spectrophotometry,s but sition from ferrous sulfate and barium perchlorate solutions. the results obtained were not in good agreement. Concentrations of freshly prepared sodium thiocyanate soluA later estimate6 of the association constant of tions were checked b y argentometrir titration. Spectrophotometry.-All absorbance iiieasureiiients were Fe(CNS)s+, from kinetic studies, did not support made a t 18 2~ 0.3", using a Beckman inodcl L)U quartz earlier values. photoelectric spectrophotometer. Silica inserts and a range The present investigation was undertaken in an of optical cells provided light paths from 0.14 t o 100 nim. Known concentrations of ferric perchlorate, perchloric endeavor to measure the association constant of sodium perchlorate and sodium thiocyanate were the ion Fe(CNS)s+ and to obtain its absorption acid, mixed in t h a t order and diluted to the desired volume. .4bspectrum. I n this way it was hoped to assess the sorbarice rneasureiueiits were coiii~rieiicccllritliiii t w o uiinsignificance of this species in the ferric thiocyanate utes of mixing, and :i series of readings \vas takvii at each system. Competitive complex formatiou was min- wave length to allow correctioii for ariy observed fading. Fading was slight b u t sigiiificaiit only a t higher tliicicyuriatc. imized by using perchlorate solutions since Kabino- coricentratious. All rentlings were taketi withiii 30 iiiiiiutcs witch and Stockmayerg found no evidence of com- of mixing. Variation of Absorbance with Thiocyanate Concentraplex formation by ferric ion with perchlorate concentrations of up to 3 M. Subsequent datalo sug- tion.-If, as appears t o be generally accepted, thiocyanate (1) h1, Mgller, "Studies on rlqueous Solutions uf I r o n Thiocyanates," Dana Bogtrykkeri, Copenhagen, 1Y37. (2) H. E. Bent and C. L . French, THISJOURNAL, 63, 598 (1941). (3) S. M. Edmonds and S. Birnbaum, ibid., 63, 1471 (19-11). (4) H. S. Frank a n d R. L. Oswalt, ibid., 69, 1321 (1947). (5) J. Y . Macdonald, K. A I . hIitchel1 and A. T. S. Mitchell, J. Chenz. Soc., 1574 (1951). (6) R. H. Betts a n d F. S. Dainton, THISJ O U R K A L , 7 6 , 5721 (1953). (7) (a) H. I. Schlesinger and H. B. Van Valkenburgh, ibid.,53, 1212 (1931); (b) H. 1. Schlesinger, ibid., 63, 1765 (1941); (c) R. K. Gould and W. C. Vosburgh, ibid., 64, 1630 (1942); (d) S. E. Polchlopek and J. H. Smith, ibid., 71, 3280 (1949); (e) S. Baldwin and W. J. Svirbely, ibid., 7 1 , 3326 (1949); (f) A. K. Babko and V. S. Kodenskaya, T~74dy Komissii A n a l . K h i m . , A k a d . A'a7~k S.S.S.R., 3 , 162 (1951); C. A , , 47, 2575 (1953); (9) K . M. hfitchell and J. Y . Macdonald, J. Chern. S o c . , 1310 (1949). (8) A. K . Babko, Compl. ueizd. acad. sii. L7.R.S S I 62, 37 (194G); C. A , , 41, 1913, 4732 (1947). (9) E. Rabinowitch a n d W. H. Stockmayer, T H I S J O U R N A L , 64, 335 (1942). (10) J, Sutton, S a t w e , 169, 71 (1952).
ion forms a series of colured complexes with ferric ion, the variation of t h e absorbance of the system with iiicreasiiig thiocyanate coiiceritration should enable their association constants and absorptioii spectra t o tie calculated. I n a constant enviroiiiiient, froni tlie law of mas5 xctiuii F ~ ( C S S I , ~ ' ~f]- "=: h-,[I~e(CSS),.~ 1 ' 4 ~ - z-'I@ j = h n l K , K , [ F e - + + I O z (1) I =
n
=
0
[F?+- - j J =
[ F e ( C S S ) 7 ' 3r ) i
[Ire-
I-
(1
+ h-,B
f K,K@ $-
,
-1 ,
=
+ KJ
/i,o") ( 2 )
x = n
.1,
CIII.
= x = o
+
+
+
[ F e + + A ] ( ~ l K l~~~ k ' ~ k ' ~ .8. ~. ~,~k'iIiB. .K#) (3) where K, is the association constant for the complex, Fe(CNSj.'3 - z ) +, 0 is [CNS-1, -4is tlie absorbance, t is the molar absorbance itidex, arid [FeT+']o is tlie tutal ferric ion couceiitration. :It tlie conceutratiuns ubed in the present
ASSOCIATIONCONSTANT AND ABSORPTION SPECTRUM OF Fe(CNS)Z+
Aug. 5, 1958
work ferric ion gives no detectable absorption throughout the wave length range studied. Elimination of [ F e + + + ]between equations 2 and 3 gives
+
+ + enKIKz, K,O" + + K1K2. .K,,O"
elKIO ezK1Kz02 . . 1 K1O KlKz02 . .
+
+
,
(4)
At low thiocyanate concentrations, where the only thiocyanate complex present in significant amounts is FeCSS++,d equation 4 reduces to A I c m / [ F e f + + ] o= elK10/(1 KlB) which 011 rearranging becomes
+
+
[Fe+++]olAlom = 1/e1 l/elK16 (5) Hence el and K1 can be obtained from the plot of [Fe+++]o/ Al against l/e. At higher thiocyanate concentrations other complexes must also be considered. If it is assumed6 t h a t for thiocyanate concentrations less than 0.1 M the concentration of Fe(CNS)2+is very much greater than of Fe(CNSIr and subsequent complexes, i t becomes possible to determine the association constant K Pof Fe(CNS)2+. Under these conditions equation 4 becomes
+
+ KIO + K1K20P)
A I ,,/[Fe+++]o = (elKI0 ezK1KnO2)l(1 which can be rearranged t o give
- f(A)/K1K202 is the molar absorbance index of F e ( C S S ) 2 C . As K I , €1, 0 and F e + + + are known it is possible to obtain e2 and KPfrom the plot of A/[Fe+++]oagainstf(A)/P. For thiocyanate concentrations exceeding 0.1 M,cubic and higher equations can be derived from which to evaluate subsequent association constants. However, uncertainty in the experimental measurements is too great to allow reliable constants t o be obtained. Oxidation-Reduction Potentials.-All measurements were made at 20 f 0.1' with a Tinsley type 4046B potentiometer, easily readable t o 0.1 mv., and a Cambridge Spot galvanometer. Two bright platinum electrodes were immersed in the ferric thiocyanate solution, which also contained a constant, known, concentration of ferrous ion. The reference electrode was saturated calomel, the liquid junction being established across a fine sintered disc. Commercial nitrogen freed from oxygen by passage through Fieser's solution was used t o stir the solution and to maintain an inert atmosphere in the closed electrode vessel. The potential of the cell =
ion. If the concentration of ferrous ion is maintained constant the observed potential change enables the concentration of ferric ion t o be calculated from ( E - Eo')/0.0581 = log ([Fe+++]o/[Fe+++]) where Eo' and [Fe+++]orefer to thiocyanate-free solutions. Values for [Fe+++]o,[ F e + + + ]and [CNS-] can then be substituted in ([Fe+++]o/[Fe+++]- I)/[CNS-] = K1 KiKz[CNS-] K1KzK3[CNS-I2 . . . (7) which is obtained from equation 2. By fitting a curve of the form a bx cx2 . to the plot of the left side of equation 7 against [CNS-] the successive association constants can be evaluated.
+
+
+ +
+
+ ..
Results In some preliminary experiments using the method of continuous variation^,'^ a total ferric perchlorate, sodium thiocyanate concentration of 0.05 M gave a broad maximum which varied with wave length (Fig. 1). The observed scatter of the
1.2 1.0
t
c
€2
where
€2
is given by
E
3853
=
Eo
- 2.3026(127'/F) log
+ E,
(QF~+++/~F~++)
where EOis a constant and Ej is the liquid junction potential. A t 20°, for = 0.69, Eo Ej was found to be -0.5280 v. From the variation of E with concentration of thiocyanate ion the stability constants of the ferric thiocyanate complexes can be calculated if i t is assumed that formation of ferrous complexes is negligible under the experimental conditions. This assumption is reasonable because i t is known t h a t where the formation of a metal complex is attended by a reduction of ionic charge the resulting large entropy change favorable to complex forrnation1l is greater the smaller t h e ion and the higher its charge.'2 Ferric ion would therefore be expected to form much more stable complexes with anions than does ferrous ion. Electronegativity considerations point t o the same conclusion. At constant ionic strength, E , and the activity coefficients of ferric and ferrous ions should be only slightly affected when some of the perchlorate ion is replaced by thiocyanate
+
(11) R. J. P. Williams, J . Chem. Soc., 3770 (19.52). (12) R. E. Powell and W. M. Latimer, J . Chem. Phys., 19, 1139 (1951).
3 0.2
0.4 0.6 0.8 1.0 Mole fraction CNS-. Fig. 1.-Method of continuous variations: [Fe+++]o f [CNS-Io, 0.05 M; [HCIOa], 0.50 M ; 0.14 iiini. light path. Solid lines calculated from associatioil constants and molar absorbance indices for F e C N S + + and Fe(CNS)2+.
experimental points probably was due to variation in the length of the light path. A change of less than 3 degrees in the slope of the silica insert in the optical cell altered the cell length by 0.01 mm., that is, about 7yo. Within this rather large experimental uncertainty the maximum a t the lower wave lengths was near a molar ratio (Fe:CNS) of 1: 1; a t higher wave lengths it approached a ratio of 1: 2. This result confirms earlier observationsid and suggests that a t least two complexes, which are probably FeCNS++ and Fe(CNS)2f, are present; also that the latter absorbs more strongly than FeCNS++ a t longer wave lengths. From Fig. 2, elKl was found to be 709000, 522000 and 247500, a t 4600, 5000 and 5-100 A., respectively. Under the experimental conditions, ([CNS-]> [Fe+++]),the intercept, 1 / 6 1 , was excessively sensitive to error, and it was necessary to use published values of el in evaluating K1. The selected values of €4600 = 4946,j E5000 = 357S5 and €5.100 = 17305 (interpolated) are in good agreement with the figures of Betts and Dainton6 but are slightly higher than those of Frank and O ~ w a l t . ~ I n this way K1 was found to be 143.4, 145.9 and 143.0, average 144, a t 15') for p = 0.56. Correction for hydrolysis of ferric ion14 increases K1 by about 1%. (13) P. Job, A n n . c h i m , 9, 113 (1928). (14) W. C. Bray and A. V. Hershey, TUISJOCRNAL, 66, 1889 (1934).
3854
D. D. PERRIN -.
0.02.C
so
Fig. 4, froni which values were found to be: €4600 = 8900, e5000 = 9500, €5400 = 6460. For thiocyanate concentrations exceeding about 0.2 M , the experimental values in Fig. 3 deviate increasingly from linearity, suggesting that higher complexes are present. However, the data are not sufficiently precise to permit the evaluation of higher association cmstants.
1I
P
Vol.
10 ( 1
--i
0 2
4
t;
i / [ c ~ s -x] 10-3,
8
N.
+
Fig. 2.-Estimate of elKl from [Fe"' ' ] o / A = 1 / ~ , ~ / E , K ~ [ C S S - ][Feif+]o, ; 2.0 X lW4 M; [HClO,], 0 57 hf
For b3th Fig. 2 and 3 the thiocyanate concentrations were obtained by successive approximation, beginning with [CKS-] = [CNS-Io. The corrections were small, amounting to about 3y0 a t the lowest thiocyanate concentrations and diminishing to less than lY0 above 0.014 M. Over the thiocyanate range 0.006-0.123 AI, the linearity of the results pl 7tted in Fig. 3 confirms that Fe(CNS)2T and FeCT\;S-+ are tl7e m a i r l coin1'
o'l
I-
I
io00
5000 A,
I'ig
c,Ol'0
A.
4 --Absorption spectra of Fe(CiXS)2+ (solid line) and FeCNS+b (broken line) in tcater
The oxidation-reduction potential measurements provide an independent method of examining complex formation between ferric and thiocyanate ions. The results presented in Fig. 5 confirm that in thiocyanate concentrations a t least up to 0.15 A I only 1: 1 and 1: 2 ferric thiocyanate complexes are present in significant amounts. The good straight line obtained in Fig. 5 over this range leads to __
0
2
4
6
f(n)/e'x
8
10
12
-
__-
_
14
10-6.
Fig. 3.--Evaluation of association constant and molar absorbance indices of Fe(CNS)z+: [Fe+-+]o, 2 0 X 10 - 4 hf; [HCIOo],0 56 ,lT, Linear portions [CSS-1, 0 0151 0 123 -21 Higher concentratmns are [ C S S - ] , 0 31, 0 62, 0 92 M.
ponents present. The slopes of the lines lead to values of Kz = 11.6, 14.5 and 15.7, average 14, a t 18') for p = 0.56. The intercepts are somewhat uncertain owing to the long extrapolations but indicate molar absorbance indices for Fe(CNS):+ of the orders of 9000, 9500 and 6300 a t 4600, 5000 and 5100 A. More accurate values were obtained using a sdution 0.1229 Jf in thiwyanate, 0.56'7 .1/ in perchloric acid and 2.0 x ~ l fin ferric ion. The absorbance due to FeCNSf+ was calculated using the molar absorbance indices of hIacdonald, et aI.,6 and taking K l = 146 and K 2 = 14. Subtraction from the measured absorbances gave the absorption spectrum for Fe(CNS)2+recorded i n
I&
t: 100
I
0
1
0.05
0.10
[CNS-I,
..
.
0.15
0.20
1
0.25
x.
Fig. 5.-Estimate of association constants of F e C S S and Fe(CNS)2+ from potentiometric measurements: [ F e t + ] , 4.88 X ill; [Fe+++]a, 1.67 X lo-' AT; [HClOJ, 0.569 Ai. Addition of 0.500 A f N a C S S : 0 , [CNS-] set equal t o [CSS-l0 - ([Fe++']a - [Fe'++l); g , limiting values correcting for C N S - in FeCNS++ a n d Fe(CSS)Z+. +
+
values of K 1and K 2 ,at 20' and p = 0.65, of 133 and 10.6, respectively. A subsequent experimelit gave K 1 = 130, l i z = S.9, under similar conditions. The small correction for combined thiocyanate ion was made by successive approximation, commencing with [CNS-] = [CNS-]o - ([Fe++f]"- [Fe+++]). The temperature coefficient of K Z should not be namely, AH = very different from that of K1,9
Aug. 5 , 1958
ASSOCIATION CONSTANT AND ABSORPTION SPECTRUM O F Fe (CNS)?'
-1.6 kcal./mole,6 so that K 1 and Kz, corrected to lSo, are approximately 133 and 10. The slight curvature shown by the results in Fig. 5 with increasing thiocyanate concentration is in the direction opposite t o that expected for higher complex formation. This deviation from linearity, which was found to amount to 30% in a solution 0.8 M in thiocyanate ion and of ionic strength 1.08, is probably due to the increasing error in assuming constancy of liquid junction potential and ferric ion activity coefficients with change in thiocyanate concentration and ionic strength.
Discussion The values for K , obtained spectrophotometrically and by potentiometric measurement are in good agreement and confirm previous data for perchlorate media, as listed in Table I.
0,144 mm.; measurement by micrometer gave 0.14 mm. The scatter of the experimental values about these lines is within the estimated experimental uncertainty. Similarly for a solution 0.33 M in ferric nitrate and 0.67 Af in potassium thiocyanate a value of Ago$& mm. = 1.5 was predicted. This is in satisfactory agreement with the reported7dvalue of 1.7 (temperature not stated). The fractions present as Fe+++, FeCNS++ and Fe(CNS)z+for varying thiocyanate concentrations have been calculated using K1 = 146, Ks = 14 and are plotted in Fig. 6. As the absorption maxima for FeCNS++ and Fe(CNS)Z+ occur near 45305 and 4850 8.,respectively, the variation of the ratio of these complexes with increasing thiocyanate concentration leads t o a shift of the absorption maximum to longer wave lengths. 100
TABLE I ASSOCIATION CONSTANTS FOR FeCNS + + SOLUTIONS AT 18'
3855
rT----------
1
PERCHLORATE
IN
&Oat
s 0 65
KI M Od Present work" 133 1040 Present workb .56 145 1140 Frank and Oswalt4 .5 138c 1050 Edmonds and Birnbaum3 1,O 127c 1020 Macdonald, et ~ 1 . ~ 1.0 128 1030 Betts and Dainton6 1.28 121 920 Macdonald, et aL5 1.8 123 775 a Potentiometric. Spectrophotometric. C At room Extrapolation to ,u = 0, using equation of temperature. Rabinowitch and S t o ~ k m a y e r . ~
Both methods also give values for K z reasonably close to the figure deduced by Betts and Dainton6 (Table 11). TABLE I1 ASSOCIATION CONSTANTS FOR Fe( CNS)2 Method
+
s
AT
18O Ka
Potentiometry 0 . 6 5 10" Spectrophotometry 0 . 5 6 14" Betts and Daintons Kinetics 1 . 2 8 20 zk 5" 57" Spectrophotometry 1.8 Macdonald, et aL5 Spectrophotometry 1.8 105' Babkos Spectrophotometry 87c Condi0 Perchlorate solutions. Nitrate solutions. tions not specified. Present work
The conclusion that in solutions less than 0.15 M in thiocyanate ion the only complexes present are FeCNS++ and Fe(CNS) 2 + accounted quantitatively for the magnitudes of the absorbances that are used in Fig. 2 and 3. Over a range of thiocyanate concentration from 0.00014 to 0.123 M , values of the association constants and molar absorbance indices led to absorbances agreeing in almost all cases within 3y0 of the appropriate experimental value. Similarly a comparison was made with the experimental results in Fig. 1 by calculating from the initial ferric and thiocyanate ion concentrations the amounts of FeCNS++ and Fe(CNS)2+ present, and hence their contributions to the absorbances. Results obtained in this way are shown by the solid lines in Fig. 1. Best fit for the data required that the cell length be taken as
0.05 0.10 0.15 [CNS-1, Af. Fig. 6.-Variation in fractions of ferric ion present as complexes in solutions containing thiocyanate ion. Vertical distances between lines give percentage as F e + + + , F e C N S + + and Fe(CNS)?+.
The molar absorbance index a t the maximum increases with thiocyanate content in the ferric complexes, from €4530 = 50006 for FeCNS++ to €4850 = 9800 for Fe(CNS)z+. For the complex, believed to be Fe(CNS)S, obtained by extraction of an aqueous ferric solution, 1.5 ilf in thiocyanate, ion, with an equal volume of amyl alcohol, €4950 was about 13,800. Association constants and molar absorbance indices point to Fe(CNS)z+ as the main light absorbing species for thiocyanate concentrations greater than 0.04 M . By assuming that the activity coefficients of the ferric thiocyanate system are similar to those for ferric chloride comnlexes, estimates can be obtained of K1° and K20. Extrapolation to p = 0, using the same function as Rabinowitch and Stockm a ~ e r ,gives ~ K10 = 1090. Similarly from the function of Bray and Hershey,I5 Kzo = 40. The ratio of Kl0/Kz0= 27 lies between thevaluesof 6.7, for Fe+++, Cl-9, and 175 for Fe+++, OH-.lS A '.'statistical" factor of 6 would be anticipated for a mechanism involving simple substitution ; the difference may indicate less ionic binding in the complex. From simple electrostatic considerat i o n ~it~ is predicted that K30/Kz0is appreciably smaller than Kz0/Kl0. This has been confirmed for (15) A. B. Lamb and A G. Jacques, Tars JOURNAL, 60, 1215 (1938).
3S5G
LEONED. COCKERELL ,\ND PATRICK H. SVOODS
the similar molybdenum(V) thiocyanate system.I6 The potentiometric measurements reported above also suggest that K3 is numerically small. Estimation of Iron as Thiocyanate.-The present results indicate the need to maintain a constant, high, thiocyanate concentration in any spectrophotometric method for the estimation of ferric ion by formation of thiocyanate complexes. The association constants of these complexes are not large and, as is apparent from Fig. G, almost complete conversion to any one complex does not occur, even a t high thiocyanate concentrations. Also, (16) D D. Perrin, THE J O I J R K A I . , 80,
1-01.
so
although ferric hydroxyl complex formation can be reduced t o negligible proportions by using sufficiently acid solutions, ferric ion forms with many other anions complexes of stabilities comparable with, or exceeding, those with thiocyanate. Except under carefully controlled conditions such a method is, therefore, not likely to be capable of high precision. When greater accuracy is desired reduction to ferrous ion and the use of much more strongly complex-forming reagents such as a,a'-clipyridyl ion or o-l'henanthroliiie appears to be preferable. C A N R B R R A , .-\liSTRAI.IA
(Iq58)
-
[COYTRInIJTION FROM THE CHEMISTRY D E P A R T M E S T O F B A Y L O R UXIVERSITT]
Equilibrium Studies of the Copper(I1) Oxalate Complex Between an Aqueous Solution and an Anion-exchange Resin1 RY
LEOh'E
n.COCICERE1,L AND PATRICK R R C E I V Jr:r,r. E~
H.
b700DS2
20, 1937
T h e equilibrium of t h e copper(11)oxalate complex existing Iictweeti an aqueous solution and a n ion-exchange resin 1x1s been studied. T h e copper(I1) bioxalatc conlplex is stable on t h e resin, althougl~it tloes not exist in a dilute aqueous solution. Evidence for a coordination number of sis for cnplm-jIIj is given. '1 higli p11 and low owlate in solution favor the uptake of copper(I1) on a n anion-exchange resin
Introduction The copper(I1) oxalate complex has been investigated and its formula and stability constant in solutions of a i oxalate salt determined. In 1936 Britton and Jarrett3 in an electronietric investigation of the copper(I1) oxalate complex in solutions of sodium oxalate showed the complex to have the formula C U ( C ~ O ~ ) ~This = . work has been confirmed several times, most recently by hleites4 in 1950 by polarographic means. Thc purpose of the present work was to study the equilibrium of the copper(I1) oxalate complex existing between an aqueous solution and an ion-exchange resin. The general approach used was that of Stokes and Walton5 and 'or Salmon.6 The use of an ion-exchange resin previously saturated with the ligand is suggestive of Fronaeus' but his treatment was not applicable in this case. The metal ion A1f) concentration was much too high (7 X and the existence of the uncharged complex could not be assumed. Instead, the data were plotted after the manner of Bjerru~ii.~ Experimental Materials.-Reagent grade chemicals wcrc used tlirouglinut t h e investigation except for primary standard sodium oxalate and potassium dichromate which were used for standardization purposes. ( 1 ) Presented a t t h e Eleventh Southwest Regional Meeting of t h e .4merican Chemical Society a t Houston, Texas, December 1, 1B.55. ( 2 ) T a k e n f r o m a thesis presented by Patrick 13. Woods i n pnrtial fulfillment of requirements for t h e Doctor of Philosophy degree. (3) H. 7'. S. Britton a n d M . E. D . J a r r e t t , J . Cizem. .TOG., llSfl (193G).
(1) 1,. Rfeites, THISJ O V R N A L , 72, 184 (1950). ( 5 ) R u t h H.Stokes a n d H . F. Walton, {bid , 76, 3327 ( I O N ) . (G) J. E. Salmon, Rizls. P w e and A p f d . C h c n z . , 6, 2 1 (lO56). ( 7 ) S. Fronaeus, Sveitsk K e i n Tidsiiv., 65, 1 (1953). (8) J. Rjerrum, "Rletal-Ammine Formntion in Arjiieoiis Solution P. Haase and Son, Coprnhaqrn. 1911.
.Iinberlite IRA-401, a strongly basic quateriiary ammoilium type resin was used, l~ecauseit has a low degree of crosslinkage which permits tlie exchange of large anions. T h e chloride form of the resin was converted t o the oxalate forni with concentrated solutions of sodium oxalate. It was washed until the eiTluent was free of t h e oxalate ion, Ijackivashed, column dried and finally air dried before use. T h e moisture coiitcnt, tleterinincd by drying at 110" for 3 lir., was 35.9%. T h e swollen volume (in water) of the air-dried resin was detcrinined pycnoinetrically. T h e dry volume of tlie air-dried resin was determined by tlisp1:iceiiient of IICS:tne.B Tlie values are: swollen vnlunie, 1 .G82 cc./g. of airdried resin; d r y vciluine, 0.888 ce.,/g. ( i f air-dried resiii; internal solution ruluiiie, 0.794 cc./g. of air-tlrictl rcsin. To tibtain tlie ion capacity, a weighed quantity of t l ~ cos:ilate form irf the resin was placed in sulfuric acid anti the clutcti oxalate ion was titrated with Imtassiutii permanganate. This value was checked b y convcrting t o tlie hytlrciside form, adding cseess nitric acid and back titrating with stantlard sotliuni liydroside. T h e exclinnge enpacity per gram of air dried resin was 3.150meq./g. Preparation of Solution.-The copper(I1) oxalate C ~ I I I plex solution used for equilibration was prepared by bringing solid copper oxalate (J. T. Baker ,Analyzed >99.5%#)into solution with sodium oxalate. St:ilitlardization was dotic b y potassium permanganate for t h e oua1:Lte ant1 iotlometrically for t h e copper. Equilibration Technique.-One gram (1.150 nlcq.) of the oxalate resin mas placed in a ~30-11iI.polyetliylene bottle with n measured volume of the solution containing the cornplex and free sodium oxalatc. Water mas added to make a total o f 100 i d . of solution. This was agitated at 30" for 18 hr. Preliminary runs had slioivn t h a t equilibrium was a t tained shortly after 12 hr. Analytical.-Copper content, total oxalate and PH were determined on the solution after equilibration. Tlie @€Iwas measured with a Beeknian Model G meter. Copper u i d oxalate were determined as in tlie original analysis o f the s(11ution containing the complex. The anall-sis of t h e rcsin TWS done b y difference since all ionic species were known. No attempt was made t o correct for activities as available colistants are not tlierinodyiiamic ones. KOattempt was niade t o maintain constant ionic strength because this noultl have rcquired addition of another salt whose anions noulti have cntereti tlie resin n n d caused unnecessary c(iiiip1icatioiiq.'" ~
''
~
-~ (0) I€. 1:. w i l l l o l l , .7. P i f j S . c / l e ~ n 47, . , ::71 (I!4M) (10) S lironacus. : 1 r / n Chrm. . S r n ; i l , 8 , 1171 (I!d:t) ~~
