relative standard deviation was 1.09% for a single determination. As a further evaluation of the method, six 1-mg. uranium-235 metal samples were irradiated at different times in the MTR and analyzed after 1-day cooling both for zirconium-97 and for uranium. For uranium, a colorimetric method ( 4 ) was used with a stated precision of about 1% relative standard deviation. The precision of the ratio of atoms of Zrg7t o milligrams of uranium was 2.8% relative standard deviation for a single determination. Error sources include the uranium determination, the irradiation time, and variations in the reactor neutron flux. From the extracted zirconium, Nb95 is the only niobium activity of any consequence, other than X'ba7, produced during a 1-hour grow-in period. Up to 5 days after a critical incident, its activity is less than 0.5y0 of that of Nb9' in a 1-hour grow-in period. Even
after 2 weeks, with a Nbg5/Nb" activity ratio of 18.6 in a 1-hour grow-in period, the N b g is resolvable by gamma spectral analysis. The distribution of 1-day cooled fission products in the method is shown in Figure 1. The specificity of the separation is indicated by the absence of extraneous activity after a decay time equivalent to six half lives of Nbo7 (as shown by the bottom curve in Figure 1). Based on a weighted regression analysis of activity decay data from six separated Nbg milkings followed for 7.5 hours, the calculated half life of Nb9' is 74.0 i 1.2 minutes. The uncertainty is expressed a t the 95% confidence level. ACKNOWLEDGMENT
The authors thank L. €3. Hansen and Floyd Spraktes for statistical calculations.
LITERATURE CITED
(1) Ginkel, W. L., Bills, C. W., Dodd, A. O., Kennedy, K. K., Tingey, F. H., U. S. At. Energy Comm. Rept. IDO-10035(1960). (2) Katcoff, Seymour, Nucleonics 16, No.
4. 78 (1958). (3) 'Lyon, W. S., Reynolds, S. A., Eldridge, J. S., Ibid., 20, No. 5, 92 (1962). (4)Maeck, W. J., Booman, G. L., Elliott, M. C., Rein, J. E., ANAL.CHEM.31,
1130 (1959). (5) Ibid., 33, 237 (1961). (6) Marsh, S. F., Maeck, W. J., Booman, G. L., Rein, J . E., Ibid., 33, 870 (1961). (7) Ibid., 34, 1406 (1962). (8) Nucleonics 19,No.3,62 (1961). (9) Paulus, R. C., Dodd, A. O., Kennedy, K. K., Tingey, F. H., Warzel, F. M., U. S. At. Energy Conim. Rept. IDO10036 (1961 j. (10) Steinberg, E. P., Ibid., NAS-NS3011 (1960). (11) Wish, Leon, ANAL.CHEY. 34, 625 (1962). RECEIVED for review September 4, 1962. Accepted December 26,1962. Work done under contract AT( 10-1)-205 to Idaho Operations Office, G. S. Atomic Energy Commission.
Solvent Extraction of Certain Transition Metal Ions with 1- (2- Pyridy Ia z 0)- 2 - na pht ho I A Study of Complex Formation and Distribution Equilibria D. BETTERIDGE,' QUINTUS FERNANDO, and HENRY FREISER Department o f Chemistry, Universify of Arizona, Tucson, A r k .
b A study was made of the extraction of the complexes of Mn(ll), Zn(ll), Cu(ll), Ni(ll), and Co(ll) with 1-(2pyridylazo)-2-naphthol (PAN) from water into CClr or CHC13. Mn(ll) and Zn(ll) form extractable complexes containing a 1 to 2 metal-ligand ratio. The over-all formation constant of the Zn(PAN)z complex was found to be 1OZ1J. At pH above 8.5,Zn(ll) formed a different extractable complex, whose composition has not been determined. The formation of an extractable Cu(PAN) (OH) complex has been postulated. The formation constant of the Cu(PAN)+ complex was found to be 1012.6; the hydrolysis constant, 10-6.9. The cobalt and nickel complexes of PAN showed anomalous behavior.
S
Cheng and Bray (4) pointed out the possibility of using PAX as an extractant, it has been used by several workers to effect analytical separations. The extraction of uranium(V1) has been studied by Cheng (3) and Shibata (IO), vanadium(Vi) by Staten and Huffman ( 1 4 , cobalt(II1) by Goldstein, Manning, and Menis (8), palladium by Busev and Kiseleva (a), INCE
294
ANALYTICAL CHEMISTRY
and zinc and cadmium by Berger and Elvers (1). Shibata (11, 12), has surveyed several of these systems as well as Ga, In, Cu, Bi, Xi, Fe(III), Hg(TI), Mn, Pb, Ce(III), Y, and La. The present study was undertaken to determine the types of extractable metal complexes formed with PAX and to evaluate the equilibrium constants that govern the distribution of the metal-PAN complexes between an aqueous and an organic phase. This work paralleled a potentiometric study of the same complexes carried out simultaneously in this laboratory by Corsini et al. (6). EXPERIMENTAL
Apparatus. The solutions were equilibrated in 30-ml. vials fitted with polyethylene caps and liners. Twenty to 40 vials were shaken at a time in a box shaker. A Suclear-Chicago DS5 scintillation counter and 183 B scaler were used to detect and count the radioactive disintegrations. A Beckman Model G p H meter with a glass and saturated calomel electrode pair, standardized with Beckman buffers, was used for all p H measurements. A Cary Model 11 recording spectro-
photometer was used for recording spectra and a Beckman LModel D U spectrophotometer was used for measurements at a fixed wavelength. Electron spin resonance spectra were recorded on a Varian 100-kc. modulstion spectrometer. Reagents. The C058, Ct164, and Zn66 isotopes were obtained from Union Carbide Chemicals Co., Oak Ridge, Tenn. 4 sample of P A N was obtained commercially and subsequently recrystallized from aqueous ethanol to givc a sample melting a t 140-lo C. The following standard buffers were made to cover the p H range 1 to 13: p H 1 to 2, perchloric acid; p H 2 to 5.5, phthalate; pH 6 to 8, phosphate; pH 8 to 10, borate; pII 10.5 to 13, sodium hydroxide. A11 buffer solutions -ere adjusted, where necessary, to a n ionic strength of 0.1 i 0.01 with sodium perchlorate. .4nalytical grade chemicals, or the purest available, Kere used throughout. All water was deionized. Procedures. D E T E R M I S B T I O N O F S C I D DISSOCIATION
COSsTA4NTs S N D
DISTRIBUTION COEFFICIESTOF REAGEXT, K D ~ .Sixty milliliters of a 1 Present address, United Kingdom Atomic Energy Research Establishment, Harwell, England.
buffer solution was shaken with 5 ml. of a 10-2iM PAN solution in CCL or CHClr for 4 hours. An aliquot of the aqueous phase was taken, its p H measured and adjusted to approximately 0.6, and the absorbance measured a t 430 mp in a 10-cm. cell. A calibration curve was used to calculate the reagent concentration, in the aqueous phase, from which the reagent distribution was calculated a t various pH values. Let the distribution coefficient of the reagent = KDR,the distribution ratio = D, the neutral species of PAN be represented by HR, the protonated species of PAN, by H2R+, and the anionic species of PAN, by R-.
aliquot of the organic layer. first case,
Then the acid dissociation constants of PAN, K,,, and K,,, as well as RDR and D, are given by the following equations, where all terms in square brackets represent molar concentrations and subscripts and uI refer to the organic and aqueous phases, respectively.
if the same volumes of organic and aqueous phases are used for counting. Counting errors were kept below 2% reliable error (QOyoconfidence level), as far as practicable. Wherever there was the possibility of adsorption or precipitation of the metal ion or complex, both the aqueous and the organic layers were counted. Unless more than 90% of the isotope could be accounted for, the results were discarded. -4 spectrophotometric method was used to determine the distribution of the manganese and nickel complexes of PAN. -4fter equilibration, the absorbance of the organic layer was measured a t a suitable wavelength (570 mp for manganese and 538 mp for nickel). These absorbance values (for solutions containing the same amount of metal ion and reagent) increased with pH and finally reached a maximum. The distribution of the metal complex was calculated from the following equation :
Values of K D R , K,,,, and K,, were obtained from a plot of log D us. pH. Table I shows a typical set of data. DETERMINATION OF METALDISTRIBUTION CURVES.In each case the distribution of the PAS-metal complex was measured, with fixed reagent concentration and varied pH (ionic strength = 0.1 *0.01), nith fixed pH and varied reagent concentration (ionic strength = 0.1 31 001), and sometimes with fixed pH and ieagent concentration and varied salt concentration. ('arbon tetrachloride n-as the organic phaw used throughout, with one exception. In the eytraction of cobalt, chloroform was used, when it was found that the PAN complex was virtually insoluble in CCI,. In all cases the buffer concentration was kept as low as possible to avoid auxiliary complexing, and it was confirmed that there n-as no detectable difference in the distribution of the metal complex if a different buffer was used a t the same pH-e.g., acetate instead of phthalate or phosphate. Equilibrium was normally established during the 4-hour shaking period, although kinetic effects were clearly noticeable in the case of cobalt, The cobalt-PAS compleves were extracted into chloroform by shaking overnight. The distribution of the metal complexes of C O ~Cu", ~ , and Zna5was measured by counting an aliquot of the aqueous phase and, if necessary, an
D =
Ctotsl
-
C.qUCoUI
vu X E
C.PU.OUD
In the Table 1.
Distribution of PAN in CC14/H20 Systems
(2)
where = total number of counts in a standard aliquot C.qu.ou.= number of counts in a standard aliquot of aqueous phase V D= volume of organic phase Vu = volume of aqueous phase. Ctatal
I n the latter case, -
Corpanio
(24
C*qu.ou.
DI =
Aobad
Amax
-
Aobsd
vw E
(3)
D' will be indistinguishable from D, a t low values, if the manganese is almost completely extracted when Aobsd = A,,,; but since no independent measurements of the concentration of the metal ion in the aqueous phase were made, high values of D' could not be calculated accurately. The temperature could not be controlled in the box shaker, but during any single experiment the temperature variation was less than l oC. For all determinations the temperature was 31' + 2' C. Tables I, 111, and V list values that were selected as typical of many others. Tables I1 and IV list all the values obtained.
(Initial PAN concentration in CHClt = 0.01M) pH Log D pH Log D 1.10 1.54 1.75 2.62 2.95 3.37 3.97 4.78 5.42 Table II.
2.30 2.74 2.92 3.35 3.55 3.87 3.95 3.89 3.91
pH dependence 7.5
x
If chelation takes place in the aqueous phase and a neutral, extractable species is formed, in the absence of side reactions, the distribution ratio, D, of the metal is given by:
Reagent dependence
10-6M manganese, 5.0 X 10-aM PAN
Log D '
pH
-2.17 -2.26 -1.28 -0.37 -0.09 0.35 i.52 1.42 -0.27
6.16 6.56 6.85 7.31 7.61 7.78 8.5I 9.63 9.91
Table 111.
4.15 4.07 4.07 4.00 3.88 3.36 3.22 2.76 2.75
Extraction of Manganese into CCld with PAN
pH = 9 . 1
7.5 X 10-6M
manganese L o g D ' Log[HR]o -1.34 -1.34 -1.34 -1.34 -1.07 0.22 0.66 1.03
-6.00 -5.30 -5.00 -4.30 -4.00 -3.52 -3.30 -3.00
Extraction of Copper into CCI, with PAN
Reagent pH"dependence dependence