NOTES
4333
solution was assumptively taken to be equal to that in NaCl solution. It is interesting to note that in NaCl solution the value of { is almost constant in the range of p 5 3 and then decreases with increasing p , while in CaCl2 solution there appears a small maximum around p = 3 in the {-p curve. The surface charge density, u, can be obtained using the data given in ref 8, where the relation between u and is given in tabulated form. Furthermore, the number of effective charges on the micelle, Q, is obtained by using the equation
Q = 4.rrr2u/e where e is the elementary charge. The values of CT and Q of the micelle in solutions of NaC1 and CaClz are given in Table I. The micellar charge can be also expressed as a degree of ionization, a = Q/n, where n is the number of molecules per m i ~ e l l e . ~In Figure 2 the value of a i n KaC1 solution is plotted as a function of p . A similar plot of a us. p in CaC12 solution would be obtained if we have available data of n in this solution. The results shown in Figure 2 indicate that the degree of effective charge on the micelle increases with increasing oxyethylene content, which could promote the repulsion among the charged heads of the polyoxyethylene chains in the micelle and the extension of the chains into the mater owing to this repulsion. This extension will make the oxyethylene units less available for interaction with the dye molecules and, therefore, in this case the degree of solubilization per oxyethylene unit decreases as the oxyethylene content is increased.
Acknowledgment. The author expresses his thanks to Dr. H. Kita, Director of the Research Laboratories, for encouragement and permission to publish this paper. (8) A. L. Loeb, F. H. Wiersema, and J. Th. G. Overbeek, “The Electrical Double Layer Around a Spherical Particle,” The M.I.T. Press, Massachusetts Institute of Technology, Cambridge, Mass.,
Fea+
by T. W. Kewton, Gloria E. RlcCrary, and W. G. Clark2 University of California, Los Alamos Scientific Laboratory, Los Alamos, New lMexico 87644 (Received June 10,1968)
Chloride ion usually increases the rates of aqueous oxidation-reduction reactions which involve the Fe(I1)Fe(II1) couple. The relative slowness of the equilibrium reaction
FeC12+
=
(1)
has made it possible to specify the location of the chloride in the activated complexes for of the more rapid reactions. Thus for the Cr(I1)-Fe(II1) reaction, Dulz and S u b 3 used their stopped-flow technique to distinguish the terms k(Cr2+)(FeC12+)and k’(Cr2+)(Fe3+)(C1-) in the rate law. ;\lore recently, Carlyle and Espenson4 found analogous terms in the rate law for the Eu(I1)-Fe(II1) reaction. These results show the existence of two activated complexes of the same stoichiometric composition, one with chloride coordinated to iron and the other with the chloride elsewhere. I n the present work, the effect of chloride on the slower Kp(II1)-Fe(II1) reaction has been shown to be due primarily to the term k(Np3+)(FeCl2f) with a small contribution from k’(Np3+)(Fe3+)(Cl-). Related, preliminary experiments have shown that the oxidations of Fe(I1) by Ce(IV), V(V), and Np(V1) in the presence of C1- produce FeC12+ in greater than equilibrium concentrations. ExDerimental Section The reagents and procedures used were essentially the same as previously de~cribed.~The spectrophotometric measurements were made in the range from 3350 to 3700 8 using the same stirred 10-cm absorption cells as beforea5 The hydrochloric acid used was prepared by diluting reagent grade acid to 6 ill and redistilling. Results A t wavelengths where both Np3+ and FeC12+absorb appreciably, plots of absorbance us, time showed distinct minima when Fe(II1) was in excess. A typical result is result is shown in Figure 1. The simplest mechanism consistent with such behavior is Fe3+
+ C1-
=
FeC12+
+ C1Np3+ + FeC12+ = Np4+ + Fez+ + C1Np3+ + Fe3+ = Np4+ + Fez+
1961. (9) D. Stigter and K. J. Mysels, J . Phus. Chem., 59, 45 (1955).
The Chloride Catalysis of the Np(II1)-Fe(II1) Reaction in Aqueous Acid Solutions1
+ C1-
FeC12+ = Fe3+
kl
(1)
k2
(2)
k3
(3)
kd
(4)
This system leads to the simultaneous differential equations dz/dt
=
k3(A - X ) Y
+ ka(A - x ) ( B - x - y)
(5)
and (1) W o r k done under the auspices of the U. S. Atomic Energy Commission; presented in part at the 155th National Meeting of the American Chemical Society, San Francisco, Calif., April 1968. (2) Student summer employee. (3) G. Dull and N. Sutin, J . Amer. Chem. Soc., 86, 829 (1964). (4) D. W.Carlyle and J. H. Espenson, ibid., 90, 2272 (1968). (5) T. W.Newton and N. A. Daugherty, J . Phys. Chem., 71, 3768
(1967).
Volume 7.9, Number 1.9 November 1068
4334
NOTES
dy/dt = ki[Cl-](B
- x - y)
Iczy - k3(A
- X)Y
(6)
where A = [Np(III)],, B = [Fe(III)l0, x = A [IYp(III) 1, and y = [FeClZ+]. A FORTRAN IV program based on the Runge-Kutta methods was written to solve these equations and allow the calculation of the concentrations of the various species us. time for given values of A, B, kl, lcz, k3, and kd. This program was coupled to the Los Alamos nonlinear least-squares program' to determine values for the rate constants which best reproduce the experimental absorbance us. time data. Rather than attempting to find reliable values for four adjustable parameters, kl and ICz were fixed at values determined in separate experiments. Previous work on the equilibration rates for FeC12+8 was done in solutions of unit ionic strength between 16 and 32". Since our work was done a t 0.9' in 2 M solutions, values for kl and kz were redetermined for our experimental conditions. The rate of approach to equilibrium was measured spectrophotometrically after injecting FeC12+into HC10, solutions or injecting Fe3f into HCI-HClOd mixtures. The results were essentially independent of which solution was injected. For almost all of the determinations [Cl-] >> [Fe(III)]; however, reversing the 7 alues of these concentrations was found to be without effect on the rate of approach to equilibrium. An extensive series of determinations was made a t 0.9" in 2 i W (LiC104) solutions in which the acid concentration ranged from 0.5 to 2.0 M , the chloride concentration ranged from 0.01 to 0.11 M , and the Fe(II1) concentrations ranged from 2 X to 2 X M , Although the observed half-times were M
O
.
'
-4
\
0.2c 0.1 I
I
I
.
I
l
I
I
T I M E , sec
Figure 1. Absorbance us. time: curve A, observed; solid line was calculated by least-squares; curve B, contribution due t o FeC12+, ed = 3.45 X 106 M-1; curve concentration of Fe3+ arbitrarily multiplied by loa M-1 since E 0; curve D, contribution due to Npa+, Ed = 2.53 X loa M-I. Conditions: 4.2 X M M Fe(III), 0.04 M HC1, Np(III), 7.5 X 1.96 M HC104, 0.9', 3350 A, d = 10.33 cm path, kt = 0.42 M-l sec-l, kz = 0.248 sec-l, ka = 3.28 X los M-1 sec-l, and k, = 73.6 M-1 sec-1.
-
The Jwrnal of Physical Chemistry
C,
as short a t 1.3 sec, close to the useful limit of our apparatus, the mean deviations for replicate determinations averaged 2.6% (maximum 6.2%). At constant [ H f ] in excess C1- the apparent first-order rate constant for approach to equilibrium is given by lc = ki[Cl-] kz; however, both of these rate constants depend on [Hf]: kd = k,' kt"/[H+] as shown by Connick and CoppeL8 The combined [Hf] and [CI-] dependence is given by
+
k
= kl'[Cl-]
+
+ kl"[Cl-]/
[H+]
+ kz' + k2"/ [H+] (7)
The equilibrium quotient, &, is given by & = kl/lcz for all [H+] in the range studied; so k~' = lcl"lcz'/k2''. The data are in reasonable agreement with eq 7 and were used to find best values for the three independent parameters. The results are: kl" = 0.267 f 0.019 sec-', kz' = 0.169 f 0.002 sec-I, and kz" = 0.158 f: 0.003 M sec-l, where the uncertainties are the standard deviations. These parameters reproduce the experimental values with mean deviation of 1.9% and a maximum deviation of 15%. The derived parameters are k1' = 0.287 ==I 0.019 M-l sec-l and Q = 1.69 0.13 Neptunium rate runs were made a t 0.9 f 0.1" in solutions with p = 2.0 M (LiC104), covering the [H+] range from 0.6 to 2.0 M and up to 0.06 M C1-; initial concentrations of Np(II1) and Fe(II1) were about 3.7 X lo-* and from 4.5 X to 9.0 X loe4 14, respectively. For all of these runs values of le3 and 1c4 were found which reproduced the observed absorbance us. time curves satisfactorily. The least-squares calculation of these quantities also required the known values of [Np(III) lo, [Fe(III) lo, and the appropriate extinction coefficients as well as the values for kl and kz determined as described above. An example of the agreement between observed and calculated values is shown in Figure 1; the experimental points are shown as open circles while the upper curve is the calculated 0ne.l" A more important test of the rate law associated with the proposed mechanism is the consistency of the calculated le3 and k4 values. These are shown in Table I for the various runs. It is seen that k3 is essentially independent of the
*
(6) (a) H. Margenau and G. M. Murphy, "The Mathematics of Physics and Chemistry," D. Van Nostrand Go., Inc., New York, N. Y., 1943, p 469; (b) H.R. Siewert, P. N. Tenney, and T. Vermeulen, University of California Radiation Laboratory Report, UCRL-10575, 1962. (7) This program was written by R. H. Moore and R. K. Zeigler and is described in Los Alamos Scientific Laboratory Report, LA2367, 1959, and addenda. (8) R. E. Connick and C. P. Coppel, J . Amer. Chem. SOC.,81, 6389 (1969). (9) N. Sutin and H. Po, private communication, report IC = 0.269 sec-1 for approach to equilibrium at l o in 1.98 M HC104-0.02 M HCI. This value, determined by stopped-flow, is in good agreement with 0.256 sec-1 from eq 7. (10) The effect of the uncertainties in kl and kz was determined for this run. It was found that b In ka/b In kl = 0.04, b In ka/b In Ica = 1.1, b In kr/b In kl = - 1.1, and b In k4/b In ka = -0.2.
NOTES
4335
concentrations of acid, chloride, and initial iron. The precision with which k4 was determined is far poorer than that for IC3 because reaction 3 is more important than reaction 4. I n spite of this, 124 is seen to be clearly dependent on the concentrations of both H + and C1-. The H + dependence (inverse) is expected on the basis of the previous work in the absence of chloride.6 The increase in the rate of (4) with increasing [Cl-] indicates an additional term in the rate law of the form ~ C [Np3+] I [Fe3+][Cl-] or possibly k[NpC12+][Fe3+]. Chloride complexing of Np(II1) is much less than that for Fe(II1) ; Shiloh and Marcusll have shown that the characteristic absorption peak for the Np(II1)-chloride complex does not begin to appear until the chloride concentration reaches about 5 M . The data in Table I indicate a value of about 500 M - 2 sec-l for ~ C I . -
~~~
Table I: Results for p = 2.0 M (LiC104) a t 0.9 =t0.1""
+ k4[Nps+l [Fea+l
Rate = ka[Np3+][FeCl2+1 IH +I, M
E l -I, M
0.6
0.00 0.01 0.02 0.04 0.06 0.04 0.06 0.06 0.04 0.04 0.00 0.02 0.04 0.04b 0 . 04c 0.04d 0.05 0.06
0.7 1.35 1.46 1.61 2.00
IO-aka, M-1 see-1
10 -%a, M-1 see-1
...
1.05 1.04 1.1 1.19 1.28 1.o2 1.16 0.98 0.95 0.8 0.52 0.66 0.78 0.66 0.68 0.76 0.88 0.93
3.7 3.31 3.21 3.29 3.32 3.51 3.47 3.43 3.35 .
I
.
3.34 3.35 3.45 3.55 3.44 3.42 3.63
[Fe(III)lo = 7.4 X 10-4 M unless otherwise indicated. M. [Fe(III)lo = 6.8 X 10-4 M . 4.5 X [Fe(III)]o = 9.0 X l O - 4 M . a
* [Fe(III)]o =
The reaction of Np(II1) with Fe(II1) in chloride solutions thus appears similar to the analogous reactions of Eu(I1) and Cr(I1) in that isomeric activated complexes are involved. It is worthwhile to use our results to extend the comparison of reducing agents given in Table VI1 of the paper by Carlyle and E ~ p e n s o n . ~For the reactions of Np3+with Fe3+,FeOH2+,Fe3+ C1-, and FeCI2+ the relative rate constants are 1.0, 4 X lo3, 16 M-I, and 104, respectively. These values apply at 1' in 2 M (Li, H)C104 solutions; 3.3 X was used for the hydrolysis constant of Fe(II1). When Np3+ is compared with Eu2+ and Cr2+ it is seen that the values for k F e O H z t / l c F e s t are roughly similar: 4 X :lo3, 1.35 X IO3,and 1.4 X lo3,respec-
+
tively. The similarity is much less marked when kFeOHz+/kBeClet is compared; the values are 40, 2.6, and 1.4, respectively. This change in relative efficiency is difficult to explain but seems to be related to size and charge and possibly to the hydration of the activated complexes. It is noteworthy that Fez+ is quite anomalous in this respect; kFeOHzt/Fen+ is 810, not much smaller than the other values, but ~ F ~ o H ~ kFeC1zt is quite large, at least 200. Other Reactions The term k3[Npa+][FeCl2+] in the rate law for the Np(II1)-Fe(II1) reaction and the analogous terms for the Cr(I1) and Eu(I1) reactions suggest that FeC12+ might be the product in some oxidations of Fe(I1) in the presence of chloride. I n this connection, Conocchioli, et U Z . , ' ~ found that FeC12+is formed in the Fe(I1)Co(II1) reaction when chloride is initially present with the Co(II1) but not when it is initially present with the Fe(I1). We have extended this investigation by making preliminary experiments with three additional oxidizing agents: Ce(IV), Np(VI), and V(V). I n all cases the chloride was initially present with the Fe(II), its final concentration was 0.083 M , and FeCI2+ was produced in considerably greater than equilibrium concentrations and dissociated with apparent first-order rate constants in good agreement with the directly determined ones. Details of these experiments are given in Table 11. Although the reactions were too fast to measure, the fraction of the Fe(II1) which was formed as FeC12+was determined by extrapolating the absorbance due to FeC12+to the time of mixing. Parallel paths are involved since the production of FeC12+is not stoichiometric for any of the three oxidants listed in the table. For at least one of the paths, chloride is coordinated to iron in the activated complex, but it is not known whether chloride is in the bridging position, Table 11: Oxidation of Fe(I1) in the Presence of Chloride"
Oxidant concn,
Fe(I1) concn,
MX
Oxidant
M X 106
CeOHS+ Np02 +
6 6
2 2
7
0.5
voz
-b
103
Rateb const. sec - 1
Moles of FeCl** formed/ mole of oxidant addedC
0.29 0.28 0.28
0.59 0.42 0.28
a Conditions: 2 M total acid, 0.9", 0.083 M C1-, present initially with the Fe(I1). Apparent first-order rate constant for the disappearance of FeC12+. Determined by extrapolation of the absorbance due to FeC12+to the time of mixing.
(11) M. Shiloh and Y. Marcus, Israel Atomic Energy Commission Report No. LA-924, April 1964. (12) T. J. Conocohioli, G. H. Nancollas, and N. Sutin, J. Amer. Chem. Soc., 86, 1453 (1964).
Volume 72, Number 12 November 1968
+ /
NOTES
4336
if one exists. This is in contrast to the Co(II1) reaction where it was shown that to be effective the chloride must be in the bridging position. For the Ce(1V) reaction it is reasonable to assume that d[FeC12+]/dt = k'[Ce(IV)][Fe(II)][Cl-] and d [Fe3+]/dt = k" [Ce(IV)] [Fe(II)1; the stoichiometric ratio in Table I1 leads to ,'/,'I = 17. This value is consistent with the observations of Adamson, et C L Z . , ~ ~ who found an insignificant increase in the rate in the presence of 5.9 X M chloride. Our result requires only a 10% incfease in rate for this concentration.
Acknozoledgments. We gratefully acknowledge helpful discussions with Professor L. 0. Morgan and Dr. Norman Sutin. We also thank Dr. C. E. Holley, Jr., under whose general direction this work was done. (13) M. G. Adamson, F. S. Dainton, and P. Glentworth, Trans. Faraday Soc., 61, 689 (1965). We have confirmed these authors' results for 1 M HC104. Twelve spectrophotometric determinations using -60 pM Ce(1V) gave an average rate constant of 928 f 22 M-1 sec-1 at 0.9O. The original authors report 960 4 60 M-1 aec-1 at 0.3O.
Electrochemical and Electron Paramagnetic Resonance Investigation of Nitrotriphenylamine Reductions by Robert F. Selson Sacramento State College, Sacramento, California 96819
and Ralph N. Adams University of Kansas, Lawrence, Kansas (Received J u n e 17, 1968)
66044
During the studies of the electrochemical behavior of various triphenylamine derivatives in acetonitrile, several nitrotriphenylamine (NTPA) compounds were synthesized and their electrochemical reductions at platinum were investigated. The epr spectra of the anion and, in some cases, cation radicals were recorded. Several interesting steric effects were observed, both in the electrochemical and spectroscopic data, and are reported herein. Table I summarizes the compounds studied and their reduction potentials. The reduction currents of the NTPA's were compared with test molecules of similar size known t o undergo reversible one-electron processes. These comparisons showed the initial reductions of the mononitrotriphenylamines (compounds I-IV) were all reversible, one-electron reactions leading to relatively stable monoanion radicals. The reduction potentials of the mononitrotriphenylamines are all very similar; i.e., substitution on the rings The Journal of Physical Chemistry
Table I: Reduction Potentials of p-Nitrotriphenylamine Compounds ----Substituted
I 11 111 Iv V irI
TPA,a (RiRzRsNA)----
RI
Compd
Ra
(V us. sce)
CeH6 C6H5 CeH40hk CEH4;\/Ie CsH, CeH6
-1.18 -1.21 - 1.24 -1.22 -1.02 -1.05
R2
CeH4N02 C~HG CsH4NO2 CEH40Me CsH4N02 C6H40hfe CeHaNOz C&P\'le CsH4NOz CsH4N02 CeH.jNO~OMeCCsH4N02
a All substituents para to the triphenylamine nitrogen unless otherwise noted. Actually the half-peak potential, since these measurements were made a t a stationary platinum electrode. These potentials were measured a t voltage scan rates of 2-16 V/min and are within a few millivolts of the corresponding El/, a t the dropping mercury electrode. For purposes of the present qualitative comparisons they are listed as Ei/,'s. ' The methoxy group is in the 2 position.
not occupied by the nitro group has little effect on the ease of reduction. Moreover, compounds 11-IV have El/,'s much closer to that of nitrobenzene (-1.13 V) than to p-nitroaniline (-1.33 V), all measured in the same medium.l (The Ell2 of the dimethoxy derivative, compound 111, tends toward that of p-nitroaniline, which is to be expected with two strongly electron-donating substituents.) One can see that the inductive influence of the triphenylamine nitrogen and the substituted phenyl rings on the nitro reduction is significantly diminished relative to the -NHZ group of p-nitroaniline. This implies a high degree of ttvisting of the phenyl rings, destroying extensive conjugation throughout the entire T system. The epr spectra of these anions are perfectly consistent with these steric effects. The spectral interpretations of the anion radicals of compounds I-IV are given on Table 11. It is apparent all have almost Table I1 : Coupling Constants of Nitrotriphenylamine Anion Radicals" Compd
I I1 I11 IV
v
Nitrobenzene p-Nitroaniline
CZN~
-------axortho
meta
para
10.6 10.3 10.5 10.5 10.5 10.32 12.18
3.35 3.34 3.35 3.35 3.38 3.39 3.36
1.10 1.08 1.13 1.11 1.12 1-09 1.12
3.97
7
aN (amine)
Unresolved Unresolved Unresolved Unresolved 0.4-0.6'
1. 12d
a All coupling constants in gauss and have an accuracy of f 3 % . *Nitro group nitrogen. Poorly resolved, only an approximate value as indicated. Coupling constants for amine N and its protons.
(1) A. H. Maki and D. H. Geske, J . Amer. Chem. Soc., 83, 1862 (1961).
