SPECTROPHOTOMETRIC STUDY OF
THE
213
Nd3+-NO3- ASSOCIATION
A Spectrophotometric Study of the Nd3+-N0,- Association’
by Nathan A. Coward2 and Robert W. Kiser Department of chemistry, Kansaa State Universitv, Manhattan, Kansas
(Received August I,1966)
A differential spectrophotometric method was used to study the association of Nd3+ and NO3- ions in aqueous solutions. The 579-mp band was selected for quantitative determinations. Our results indicate that NdN032+is the only important absorbing species (at 579 mp) other than Nd3+in these solutions and that the NdN032+is formed with an association constant of 0.77 at an ionic strength of 4.2. The differential technique was employed to produce the apparent visible spectrum of the pure NdN032+ complex.
Introduction The influence of various anions on the visible and near-ultraviolet absorption spectra of lanthanide ions has long been known. 3-11 Recently, workers studying the visible and near-infrared spectra of crystals, salt melts, and aqueous solutions, have noted similar effects and in particular have drawn attention to the ~ Jgeneral ~ effect marked effect of the Nos- i ~ n . ~The of the NO3- ion is to shift the spectrum of neodymium to longer wave lengths and to change the shape of the absorption bands. The present work is concerned with a quantitative study of the ion association in the Nd3+-N03- system in fairly concentrated (Nd3+ concentrations to 0.5 M ; NO3- concentrations to 4.0 M ) , nearly neutral, aqueous solutions. Differential spectrophotometric methods in the visible (325-800-m~) region were used. Observations in the ultraviolet were precluded by the strong absorption bands of NOa-. Our results indicate that NdN032+ is the only important absorbing species (at 579 mp) other than Nd3+ in these solutions and that the NdN032+ is formed with an association constant of 0.77 at an ionic strength of 4.2. Experimental Section AU materials used were analytical reagent grade, except for the Nd203. Nd203, 99.9% pure, from Matheson Coleman and Bell was employed to prepare the Nd(ClO& solutions. During the early work, Nd(C104)3solutions were prepared by dissolving the oxide in a stoichiometric amount of HC104, following the method of Moeller and B r a n t l e ~ . ~The neodymium was recovered from used solutions by precipitation of Ndz(C204)3from the perchlorate solutions using Na2C204.
Subsequently, the Nd2(C204)3was ignited to about 1000” to convert the oxalate to the oxide. As our studies progressed, we found it more convenient to convert the used Nd(Cl0a)~solutions to Nd2(C03)8 by addition of NaHC03 and then to dry the Ndz(C03)S overnight at 115”. New Nd(C104)3 solutions were then prepared from the carbonate by addition of the appropriate amounts of HC1O4. The carbonate dissolved in the acid more readily than the oxide. COz was expelled by boiling the slightly acid solutions. NaC104solutions were prepared by direct neutralization of HC1o4 with NaOH. Differential visible (325-800-mp) absorption spectra were recorded using a Cary Model 11 recording spectrophotometer for solutions containing varying amounts of Nd(C104)3, NaCI04, and NaN03 in the sample beam. (1) This work was supported in part by the National Science Foundation under Grant No. GE-350 to Kansas State University and in part by the U. 6 . Atomic Energy Commission under Contract No. AT(11-1)-751 with Kansas State University. (2) N.S.F. College Teachers Research Participant, 1962 and 1963. (3) P. W. Selwood, J . Am. Chem. SOC.,52,4308 (1930). (4) S. Freed, Reu. Mod. Phys., 14, 105 (1942). (5) T. Moeller and J. C. Brantley, Anal. Chem., 22, 433 (1950). (6) A. W. Wylie, J. SOC.Chem. Id.,69, 143 (1950). (7) P. Erumholz, Spedrochim. Acta, 10, 274 (1958). (8) A. Sonnesson, Acta Chem. Scand., 12, 1937 (1958). (9) R. C. Vickery, J. Mol. Spectry., 2, 308 (1958). (10) R. C. Vickery, “Analytical Chemistry of the Rare Earths,” Pergamon Press Inc., New York, N. Y., 1961, pp. 64-80. (11) F. H. Spedding and A. H. Dame, Ed., “The Rare Earths,” John Wiley and Sons, Inc., New York, N. Y., 1961, pp. 16-19, 579, 580, 594-596. (12) G. W. Oetjen, Z. Naturjorsch., 4A, 1 (1949). (13) W. L. Carnall, D. M. Gruen, and R. L. McBeth, J . Phys. Chem., 66, 2159 (1965).
Volume 70, Number 1 January 1966
NATHANA. COWARD AND ROBERT W. KISER
214
The reference beam contained solutions of only Nd(C104)3and NaC104. [Nds+] was varied from 0.05 to 0.5 M . [NOs-] was varied from 0.1 to 4.0 M . The ionic stren.;th was maintained constant (at 4.14.2) by controlling the NaC104 concentrations. All spectra were taken at 20 i 2". The temperature was controlled by water circulation through the cell housing. Quartz and Vycor cells of 1.00-, 5.00-, and 10.00-cm. path length were used. The most intense absorption in the differential spectrum is the peak located at 579.0 f 0.5 mp (17270 f 15 cm.-I). The error figures refer to the precision of duplication of the peak on the chart record. For convenience, we shall subsequently refer to this absorption as the 579-mp peak. All absorption data (AA) reported herein were taken at this wave length. Because of the obvious sensitivity of AA to small changes in the [Nds+], extreme care was employed in the additions of Nd(C104)s to the solutions. Aliquots of N'd(C1Or)s were carefully pipetted into previously weighed volumetric flasks. The flasks were reweighed, and only those aliquots which deviated less than 0.1% were accepted for use. For the other reagents, usual volumetric procedures were employed.
Treatment of Spectrophotometric Data Let us consider the ion-pair association of Nd3+ and NO3-. We ma,ywrite
+
Nd3+ NosNdN0s2+ (1) Let Do be the total concentration of neodymium in the solution. i h y concentration of the association complex formed, C, will cause the concentration of Nd3+ to be D = (Do - C). Similarly, we shall let E be the nitrate ion concentration, E = (EO- C), and Eo be the total nitrate in the solution. Then, the association constant, K , is K =
C [NdNO,"] [Nd3+][NO3-] (Do- C)E
(2)
Rearranging eq. 2, we obtain
C
=
KEDo/(l
+ KE)
(3)
Now the absorbance, AR, at a given wave length (579 mp for this case) in the reference beam, containing no NOS-, is
AR = bedlo (4) The Beer-Lambert law has been assumed, where b is the path length of the cells employed and ED is the molar absorptivity of Nd3+at the chosen wave length. The absorbance, As, at the same wave length in the sample beam, containing NO3- is Ths Jm& of Physical Chemistry
As = b e d
+ be&
(5) Here EC is the molar absorptivity of NdNOs2+at the chosen wave length. If the cell path length is' constant in the reference and sample beams, and if the total quantity of neodymium in the two solutions is constant, the differential absorbance is given by
AA
=
As
- AR = W(EC- ED)
(6)
Combining eq. 3 and 6, and rearranging
(7) Let us define written as
CY
= AA./bDo. Then eq. 7 may be
(&/E)= K ( E C- ED) - KCY (8) A plot of a/E vs. CY should give a straight line with a slope of -K and an intercept of K ( e - ED). Thus K and (EC - ED) may be determined. Further, if ED is known, e may be obtained. Some features of ow approach are similar to (but not identical with) the procedure used by Newton and ArcandI4 in their study of cerous sulfate. We note that the more generalized approach would be to write eq. 1 as Nd3+
+ ?&NOS-I_ Nd(N03),+3-n
(9)
and thus
However, we have found from the plots of our data that n = 1. The use of n = 2 and larger does not fit our data at all. This provides us with evidence that the complex we are studying under the conditions of our experiments is NdNOa2+. If instead of plotting a/E vs. CY, we plot CY/EO vs. a, we expect to find that at the smaller values of E,, the values of a/Eo are in error by being smaller than a/E. This we found to be so for the present study. However, a fit could be made to the curve, and a h t approximation value of K determined. From this K , a first approximation of E could be made. The approximation may be repeated several times, until a convergence to the true value of K is obtained by this reiterative process. Carrying out this method on an IBM 1620 computer with a least-squares program, values of K = 0.86 and (e- CD) = 3.44, which did (14)T.W. Newton and G. M. Arcand, J. Am. Chem. SOC., 75, 2449 (1953).
SPECTROPHOTOMETRIC STUDY OF
THE
Nd8+-NO*- ASSOCIATION
215
L
1.o
1
I
550
0.5 0
6dO
6dO
'
Ti0
800
WAVE LENGTH IN MILLIMICRONS
Figure 2.
0.5
1.0
a
1.5
2.0
2.5
Figure 1. Evaluation of the association constant and ( eC for the Nda+-NOa- system: 0, 0.05 M Nd*+, b = 10 cm., p = 4.2; 0 , 0.1 M Nda+, b = 5 cm., p = 4.2; 0, 0.1 M Nda+, b = 5 cm., p = 4.1; M, 0.5 M Nda+, b = 1 cm., p = 4.2. Nitrate ion concentration was varied in all cases.
Visible spectra of NdNOSZ+ (A) and Nd(C10& (B).
3.0 e=)
not further change after six iterations, were obtained. The final plot made is presented in Figure 1. In another test of t'hese results, our experimental data were treated by a special spectrophotometric data processing program suggested to us by C~nrow.'~The results were K = 0.68 and (e- ED) = 3.82. We believe that an average of these two findings best summarizes the experimental data (K = 0.77 and e - ED = 3.63). The association constant is probably not accurate to much better than 2074, whereas the molar absorptivity difference is believed to be good to 5%.
The Absorption Spectrum of NdNO?+ After the evaluation of the association constant was made, the value of K (K = 0.86 was taken) was used to calculate the preparation of two solutions, one containing Nda+, NOa-, and NdNOs2+,and the other containing only Nd3+. The [Nd3+]was chosen to be identical in these two solutions, and therefore the differential spectrum of these two solutions should provide the absorption spectrum of the NdNOs2+ complex only. Thus, the sample of the beam solution was 0.300 in Nd(C104)3and 2.5 M in NaNO3 (which should provide a solution (0.199 M in NdNOa2+ and 0.101 M in Nd3+). The reference beam solution was 0.101 M Nd(C104)a. The differential spectrum obtained is shown in Figure 2, curve A. This is to be compared with the spectrum of Nd3+ as Nd(C1Od)s shown in
Figure 2, curve B. The red shift observed for many of the absorptions has been noted previously.6JJO~la The fact that the differential peak at 579 mp does not show a significant shift in wave length over the concentration ranges we studied is consistent with the assumption (eq. 5 ) that in these solutions there are only two important species absorbing at this wave length. (See Appendix.) The spectrum of Nd3+in perchlorate solution, shown in Figure 2B, compares very closely with those reported by Carnall, Gruen, and MeBethla for Nd3+ in DC104 solution, and by Krumhold for Nd3+ in HC104. The spectrum of NdN032+ shown in Figure 2A is similar to those reported by Carnall, Gruen, and McBeth,13 Selw~od,~ and Oetjen.12 However, the value of E 3.63 we found for the 579-mp absorption is lower than that reported by Carnall, et uZ.,'~ for 0.2 M Nd3+ in the LiNOrKN03 eutectic. Similarly, OetjenI2 found a large value of E 22 for concentrated solutions of Nd(NO&. This indicates that the species we are observing (NdN032+with E 3.63) is not the same &s the species observed by the other workers. To check this, we prepared a solution 0.1 M in Nd(C10& and 12 M in HNOS and found €679 16.7, more than four times greater than for any other absorption peak in the visible spectrum, and in reasonable agreement with the values reported by Carnall, et uZ.,13 and Oetjen.12 The dependence of the intensity of the 579-mp band upon the environment of the rare earth ion has been discussed by Judd.16 (15) I(. Conrow, G. D. Johnson, and R. E. Bowen, J. Am. Chem. SOC.,86, 1025 (1964). (16) B. R. Judd, Phys. Ear., 127, 750 (1962).
Volume 70,Number 1 January 1966
NATHAN A. COWARD AND ROBERT W. KISER
216
Effect of Other Ion Associations In the aqueous system of Nd3+, Na+, Clod-, and NO3-, one must consider the possibility of all four different cation-anion interactions. There appears to be no evidence for an association of the perchlorate ion. Previous work,5~17~1s which was rechecked during this study, showed that large concentrations of perchlorate ion have a virtually negligible effect on the spectra of the rare earth ions, but ion pairing has been reported to occur between Na+ and Nos- at very low concentrations, with an association constant of 0.25 given at infinite d i l u t i ~ n . ~ ~However, J~ RaoZ0has reported that NaN03 is completely dissociated, based on a Raman study of nitrates. Because of these discrepancies, we have assumed in the above treatment that the Na+-N03- association is negligible to the first approximation. Of course, it is desirable to remove such an assumption by changing the system to one where the salt is known to be completely dissociated. LiN03 off ers such an advantage, 21 and studies similar to those reported here have recently been concludedz2in which it was shown that the Nd3+, Li+, Clod--,and NO3- system gives K = 0.89 (in good agreement with the present work). Discussion Neither the effects of temperature nor the effects of changes in the ionic strength upon the equilibrium association constant were studied in the present work. Later workz2has shown that temperature has no significant effect. Peppard, et aZ.,z3 using a nonaqueous extractant technique, determined stability constants for MX2+ at p = 1.00, [H+] = 1.0, temperature, 22 i lo, for several different lanthanides and for americium, where X := C1- and NOa-. They did not determine NdN032+; however, their values of K for LaNOS2+ (1.3 ~t0.3), CeN0a2+ (1.3 f 0.3), PrN02+ (1.7 i 3), and EuNO3Z+ (2.0 i 0.3) are in reasonable agreement with our value for NdN02+ in view of the differences in the ionic: strengths employed in the two studies. A study of the influence of ionic strength on this system is contemplated. It may be possible to express such data by the equation of Kraus and Nelson1’ in the Debye-Huckel form log K
=
log KO
+
+
(0.509a~z~‘/~)/(i0 . 3 2 9 a ~ ’ ~ ~(11) )
as done by Newton and Arcand14for the cerous sulfate system. Comparing our data (at p = 4.2) with a value of K = 1.7 interpolated for NdN03z+from Peppard, et aZ.,Z3 (at p = 1.0) by the above equation gives d = The Journal of Physical ChemistTu
8. and KO = 37 (the association constant at p 0). Even though one has no reason to expect this equation to be valid at p = 4.2, the values for a and KO are not completely unreasonable. 3.9 =
Acicnowbdgments. The authors wish to thank John Hassler for writing the computer program which we used. We also wish to acknowledge the advice and help of Professors K. Conrow and G. D. Johnson in the taking of the ultraviolet and visible spectra and their discussions of and comments on our experiments. Special thanks are due Professor Conrow for treating our data with his spectrophotometric data processing program.
Appendix Assume a parent species, F, of initial concentration [F]o of which a fraction @ is transformed into another pure species, G, in the mixed solution. Let AF
=
[F’lf[ll
(AI)
be the equation for the absorbance of pure F at unit cell length as a function of wave length, and similarly let
Ac- = [Glg(X)
(-42)
for pure G. Then a differential spectrum such as we are dealing with would give
AR = b[F]of(X) for the reference beam (b = path length), and
+ b[FlO@g(l)
(A41
- f(U I
(A51
AB = b[FlO(l - @>f(h> for the sample beam. Therefore
AA
= b [E’IOP[9@)
(A3)
Because the wave length dependence of AA would be g(X) - f(X), and this would be constant at any given A, the peak position should remain k e d . That is, the shape of the differential curve should remain fixed. (17) K. A. Kraus and F. Nelson, J. Am. Chem. Soc., 72,3901 (1950). (18) C. W. Davies, “Ion Association,” Butterworth and Go. Ltd., London, 1962. (19) C. W. Davies, T T U ~FSU.T U ~ USOC., Y 23, 361 (1927). (20) N. R. Rao, Indian J . Phus., 15, 185 (1941). (21) W. H. Banks, E. G. Righellato, and C. W. Davies, Trans. FUTCZ~UY SOC.,27, 621 (1931). (22) J. H. Hagfeldt, R. E. Olsen, and N. A. Coward, work done at Wisconsin State University, Superior, Wis., to be published. (23) D. F. Peppard, G. W. Mason, and I. Hucher, J . IT LOT^. Nud. Chem., 24, 881 (1951).
DIFFUSION OF SOLUTIONS IN THE SYSTEM H~P04-Ca(HzP04)z-H20
A similar treatment, assuming a parent and an additional subspecies H of concentration y [I310 with
AH
=
[Hlh[Xl
gives the equation
217
+
AA = b[FIo[P(g(A) - !(A>> r(h(A) - !(A)>] (A6) It can be seen that if p, g(A), y, and h(A) are all appreciable, then only extremely fortuitous circumstances would give a constant shape for the differential spectrum.
Diffusion at 25” of Solutions in the System Phosphoric Acid-Monocalcium Phosphatewater’
by 0. W. Edwards, R. L. Dum, J. D. Hatfield, E. 0. Huffman, and K. L. Elmore DCiswn of Chemid Dmelopment, Tennessee Vdley Authority, Wa7sson Dam, Alabama (Received August $3, 1966)
Isothermal diffusion measurements at 25’ were made with a Gouy Musiometer on 18 solutions in the ternary system H3P04-Ca(H904)2-Hz0. The solutions were 0.95 to 4.65 M in &PO4 and 0.045 to 1.0 M in Ca(H2P04)2. The four volume-fixed diffusion coefficients were calculated on a digital computer by a least-squares method based on published theoretical equations, and the results were in good agreement with those calculated by the area method of Fujita and Gosting. The ranges of values of the Musion coefficients (cm.2/sec., X106) were [subscript 1 refers to HaP04; 2 refers to Ca(HJ?O&]: Dll, 12.3 to 6.3; 0 1 2 , -7.1 to f5.2; D22, 7.1 to 2.4; and 0 2 1 , -1.8 to -0.05. The large values of the cross-term coefficient DIZat certain concentrations showed strong interactions between the flows of the solutes.
Diffusion coefficients for 0.036 to 16 M phosphoric acid solutions, determined by the Gouy method, have been reported.2 I n a continuing study of transport in phosphate systems, measurements were made of the four relevant diffusion coefficients for each of 18 solutions in the system HsP04-Ca(HzP04)z-H20at 25’. Several methods, which differ in accuracy and general applicability, for calculating diffusion coefficients for three-component systems from Gouy data were described by Gosting, et aLa4 A new general method for calculating diffusion coefficients from Gouy data is described in a companion paperlSand this “least-squares” method was used in the present study. The results are compared with those calculated by the area method.8
Measurements Apparatus. The two-lens Gouy diffusiometer, water bath, cell holder, and diffusion cells were those used in (1) Presented in part before the Division of Physical Chemistry, 145th National Meeting of the American Chemical Society, New York, N. Y., Sept. 8-13, 1963; Abstracts, p. 34-T. (2) 0.W.Edwards and E. 0. Huffman, J . Phys. Chem., 63, 1830 (1959). (3) D. F. Akeley and L. J. Gosting, J . Am. Chem. SOC.,75, 5685 (1953). (4)R.L. Baldwin, P . J. Dunlop, and L. J. Gosting, ibid., 77, 5235 (1956). (5) P. J. Dunlop and L. J. Gosting, ibid., 77, 5238 (1965). (6) H. Fujita and L . J. Gosting, ibid., 78, 1099 (1956). (7) P. J. Dunlop, J . Phys. Chem., 61, 994 (1957). (8) H.Fujita and L. J. Gosting, ibid., 64, 1256 (1960).
Volume 70, Number 1 January 1966