=2ug. 20, 19.57
ELECTROK-EXCHAXGE
REACTIONS BETWEEN LARGE COMPLEX CATIONS
it is assumed that there is no rotational disordering of the methyl groups in the structure. The same consideration, and others, makes i t seem difficult to interpret n.m.r. or spectral information. The higher homologs of trimethylindium do not appear to be associated. Triethylindium melts a t -32," 120" below trimethylindium. This is not unexpected, since the type oi bridges found in triinethylindium are sterically incompatible with higher homologs. I t is somewhat surprising that freezing point depression measurements of trimethylthallium in benzene indicate that i t is monomeric.18 Since the (17) E. G. Rochow, D. T. Hurd and R . N. Lewis, "The Chemistry (if Organometallic Compounds," John W'iley and Sons, Inc., New 1957, p. 130 York, N. Y . , (18) H. Gilman and R . G. Jones, THISJ O U H X A L , 6 8 , ,517 ( I S l i i ) .
[COSTRIBUTIOS FROM
THE
4145
covalent radii of In and T1 are very similar, one might expect it, too, to be associated, especially since its melting point (38") is relatively high compared to truly monomeric solids such as tetramethyllead. It is, of course, possible that trimethylthallium is tetrameric in the crystal, but t h a t entropy effects lead to dissociation in solution. Acknowledgments.-The authors are much indebted to Drs. Templeton and Senko of the University of California for I. B. M. 650 programs noted in the text and to Mr. D. R. Fitzwater for a three-dimensional block program as well as help with all our computations on the I. B. M. 650. . ~ M E S , IOWA
DEPARTMENT OF CHEMISTRY, WASHINGTON USIVERSITY]
Electron-exchange Reactions between Large Complex Cations' BY EUGENE EICHLER AND ARTHURC. WAHL RECEIVEDMARCH24, 1958 The rates of the electron-exchange reactions (1) between Fe(phen)a++and Fe(phen)a+++zand (2) between Os(dipy),++ and O ~ ( d i p y ) , + + + have ~ been investigated both by optical-active and by isotopic-tracer methods, and the rates have been found to be immeasurably large. Limits on the specific reaction rates, calculated with the assumption of second-order (at 25") and >lo6 .Wsec.-' (at 0 " ) and for reaction 2 are > l o 3 M set.-' (at 4') and kinetics, for reaction 1 are >lo2 Msec >lo6 M sec.-l ( a t O O ) , the smaller numbers coming from the optical-active measurements, which are free from separationinduced exchange uncertainties. Colorimetric observations led to specific rate limits a t 0" of >3 X lo8 X sec. for the net reactions between Os(dipy),++ and Fe(phen)3++', between Os(dipy)a++ and Fe(dipy),+++, and between Fe(phen)a+" and Ru(dipg),+++ and of > 5 X l o 4 M set.-* for the net reaction between Os(dipy),++ and Ru(dipy),++". No reduction in the rate of the Os(dipy),++-Fe(phen)3+++ reaction was observed when resolved rather than racemic reactants were used
Introduction
tween Mn04-- and MnOc-5 and between Fe(CN)$---- and Fe(CN)e---.6 However, since immeasurably large rates were found, we repeated the experiments of Dwyer and Gyarfas3 and applied a similar optical-active method to the F e ( ~ h e n ) ~ + + Fe (phen)3 +++ system. Experimental
The rate of electron exchange between Os(dipy)3++ and Os(dipy)3+++2 was investigated by Dwyer and Gyarfas3 by observing the decrease in optical activity with time after mixing &Os(dipy)3++ with Z-Os(dipy)3+++. They reported that the rate was large but indicated that it was Chemicals.-Mallinckrodt "analytical reagent" grade measurable, 9.5 sec. being required for complete chemicals were used without purification, with these esracemization a t 5' and 5 x 10-5 M reactant con- ceptions: G. F. Smith Co. sodium perchlorate, Eastman centrations. Kodak Co. camphorsulfonic acid, and Eimer and Amend p Eimer and Medalia4 used conventional isotopic- toluenesulfonic acid, while Eastman Kodak Co. Practical sulfate was redistilled a t atmospheric pressure. tracer methods to investigate the exchange reac- dimethyl Eimer and .ilmend C.P.nitromethane was purified b y the tion between the tris-(5,6-dimethyI-l,lO-phenan-method of Thompson, et ~ 2 1 . ~ Mallinckrodt U.S.P. potasthroline) complexes of iron(I1) and iron(II1). sium tartrate was recrystallized from H20. Stock solutions of Fe(phen), were prepared by dissolvThey found complete exchange in 15 sec., indicatequivalent amounts of FeS04.7H?O or Fe(NH4)2(SO& ing either that the rate in 0.5 f H2S04was immeas- ing 6HzO and l,l0-phenanthroline (G. F . Smith). The Feurably large (specific rate > l o 3 M set.-' a t 0') or (phen)s solutions were prepared immediately before each that complete exchange was induced by the sep- run by dilution of appropriate amounts of Fe(phen),++with dilute sulfuric acid and oxidation with PbOl. Excess PbOz aration methods. the product PbSOa were removed by centrifugation. We undertook the measurement of the Os- andTris-(2,2'-dipyridyl)-osmium(II) chloride was prepared (dipy)3 ++-Os(dipy) 3 + + + and Fe(phen13++-Fe by the method of Burstall, Dwyer and Gparfas.* The com(phen)3+++ exchange rates by the isotopic-tracer pound was purified by recrystallization. Stock solutions method modified to include rapid mixing and of Os(dipy), were prepared from weighed amounts of the The Os(dipy),++" solutions were prepared by quenching techniques, which had made possible compound. PbOz oxidation of Os(dipy),++in dilute HzS04. measurements of the large rates of exchange be++
+++
++
(1) This work was supported by the National Science Foundation. The paper was abstracted from the Ph.D. thesis of Eugene Eichler, Washington University, 1955. (2) dipy = 2,Z'-dipyridyl; phen = 1,lO-phenanthroline: en = ethylenediamine. ( 3 ) F. P. Dwyer and E. C Gyarfas, Nature, 166,481 (1950). 141 L.Eimer and A. 1. Medalia, THIS JOURXAL,74, 1592 (1952).
(5) J. C. Sheppard and A. C. Wahl, ibid, 79, 1020 (1957); 711, 5133 (1953). (6) A. C. Wahl and C . P. Deck, i b i d . , 76,4054 (1954). (7) C. J. Thompson, H. J. Coleman and R. V. Holm, ibid., 76, 3445 (1954). (8) F.H.Burstall, F. P. Dwyer and E. C Gyarfas. J . Chem. Soc., 953 (1950).
E. EICHLER AND A. C. WAHL
414G
Tris-(2,2'-dipyridyl)-ruthenium(II) chloride was synthesized by the method of BurstalP and purified by slow recrystallization a t room temperature. Tris-(ethylenediamine)-cobalt(II1) chloride and tris-(ethylenediamine)platinum(1V) chloride were prepared by the methods of IVorklo and Werner,'I respcrtively. Si(phen)a++ solutions were prepared by dissolution of equivalent amounts of NiSO46H20 and 1,lo-phenanthroline. Radioactivity.-The iron tracer, Feu, and the osmium tracer, a mixture of Os191and Os's, were obtained from the Oak Ridge Kational Laboratory on allocation from the U. S. Atomic Energy Commission. The Fes6 tracer had been purified by Hudis,I2and it wasconvertedtoFe(phen)a+i in much the same way as the inactive Fe(phen)a++ was made. The osmium tracer was purified by distillation of 0~01, and converted to Os(dipy),++ by the method employed for the inactive material.* Radioactive Fe(phen)8++ samples were mounted for counting as the slightly soluble perchlorate salt on filter paper and were counted with an argon-methane proportional counter. All samples were diluted before precipitation to 0.04 mmole with inactive Fe(phen)S++ to ensure equal counting efficiencies for the 5.9 kev. X-rays. Radioactive Os(dipy),++ samples were counted in solution, the ?-rays from both and 0s'" being detected by a stilbene-crystal scintillation counter Optical Activity.-Dextroand levorotatory forms of Fe(phen)s++ were prepared according to the procedure of Dwyer and Gyarfas.13 The active forms of Os(dipy),++ were resolved by the method of Burstall, Dwyer and Gyarfas.5 All rotations were observed with a Rudolph polarimeter. A General Electric Ka-1 sodium light served as the light source for the F e ( ~ h e n+) +~ measurements, and a Hanovia mercury arc with a Baird Associates 3770-5480A interference filter was used for the Os(dipy)a++work. Separation Methods.-The +3 reactant was tagged, and appearance of radioactivity in the $2 form was observed. Reduction of the +3 form was negligible during the course o f an experiment. All separations involved removal of the +2 ion from solution. Both precipitation and solventextraction separation methods were used for each reaction. Table I indicates the degree of separation for each method. TABLEI
SEPARATION METHODS Clod- precipitation method
Solvent extraction method
Reactant: Carrier: yo Removal of +2 form: % + 3 form in precipitate :
Fe(phen)r Ni(phen)a++
Reactant: Solvent:
Fe(phen)r++ Dimethyl sulfate and nitromethane IO-Camphorsulfonic acid
+ +
-40 0.95 - < o . ~ o-- >0.5 - Fo 1 - lo3M-' set.-'. solved in C104- solution, and the + 2 reactant was Fe(phen)3++-Fe (phen)3+++.-Application of the made up as usual in dilute H2S04. The two solu- above method to this system was not feasible betions were mixed rapidly by ejection of the +2 cause of the rapid racemization of resolved Fesolution into the +3 solution being stirred. Inhen)^+++ (h,:I1 = 44 =t 9 sec. a t 25' in 3 j complete exchange, -85%, was observed for both HzS04). Resolved Fe(phen)3++ also racemizes, the iron and the osmium systems. The conditions but more slowly ( t l / ; I = 25 5 min. a t 25" in 3f were the same as for the eighth entry of Table I1 H,S04). These values were the same a t various and the last two entries of Table 111. reactant concentrations, confirming the first-order Since the conditions were similar to those exist- rate laws, and they are consistent with other reing during the separation process in a homogeneous (15) 0. E. Myers and R. J. Prestwood, "Radioactivity Applied to exchange reaction, i t might reasonably be assumed Chemistry," A C. Wahl and N. A. Bonner (Editors), John Wiley and that the competitive exchange is a measure of the Sons, Inc , h'ew York, N. Y.,1951, p. 16.
*
*
E. EICHLER AND A. C. WAHL
4148
ported valuesl6,l7 determined for different acid media. A mixture of the same optical-active forms of Fe(phen)a++ and Fe(phen)3L++will racemize a t two independent rates, one characteristic of each species, if there is no electron exchange between the species. If electron exchange is rapid, the species, being in equilibrium, will racemize a t the same rate, the half-time t l ; being an average of the racemization half-times of the individual species weighted for their concentrations, (11)and (111). [.A similar expression is derived in Eichler's thesis.') Experimentally, a portion of Z-Fe(phen)3++ in solution was oxidized rapidly18 by CeIV, the mixture was allowed to stand a t 2.5" for the desired time, FeI1 was added to reduce rapidly18 the Fe(phen)3+++,and the optical rotation of the resulting solution was measured. Racemization times were varied, and the results were compared with curves calculated for fast and slow electron exchange as illustrated in Fig. 1. -1.0
L
2
L
-0.5
---------- -------
- 0.2 -0.1 ~
- 0.5
W
w -01
a a
1
------------
t
(e'
Vol. 80
Since the data agree best with the curves for rapid electron exchange, we take the rate of electron exchange to be greater than the largest raceAs. mization rate, i.e., that for F e ( ~ h e n ) ~ + + + suming a second-order rate law for the electronexchange reaction, a lower limit may be calculated for the specific rate constant k. k [Fe(phen)3+ + I [Fe(phai)s +I > k , , , . [Fe(phen)S + k > Ill 2 x = 160 -11-1 sec.-l 44 sec. (lo-( dl) ++
++I
Net Oxidation-Reduction Reactions-Net oxidation-reduction reactions between Os(dipy)S++ and Fe(phen)3 + Fe(dipy)a++, or Ru(dipy)3 + + + , and between Fe(phenIa++and Ku(dipy)3+++ were performed. The intense colors of the +Z forms of the reactants made visual observation of the progress of the reactions possible. The reactants were mixed in the rapid-mixing device described by Gordon and Wahl,Igand colors were observed in the reaction tube close to the mising chamber. Comparison tubes containing standards whose intensity of color represented 50 and 100yo reaction were used to aid in estimating when a reaction was half, or more, complete. The results, given in Table IV, show that the rate of electron-exchange between the complex cations is large. This is consistent with our conclusions concerning the Fe(phen)3++-Fe(phen)3+++ and O ~ ( d i p y ) ~ ~ + - O s ( d i p exchange y ) ~ ~ + ~ reactions and with the observations of George and IrvinelY that the reactions between Fe(dipy)3++and Fe(phen)3+-++and between Fe(dipy)a++ and Ku(dipy)3+++are very fast > lo5 J.lr-l sec.-l). 90 reduction in the rate was observed when resolved rather than racemic reactants were used.
+-'->
TABLE I\-
-005
S S T ( ~ x I u . ~ T I O N - K E r ) v C T I u NREACTIOX b
i -0.02
,TI; 0')
(Reactant concn.,
2 -05 0
(H*SOd),
a
f
Reactants
- 0.2
O s ( d i p y ) ~-'
+ Fe(phen)a+'+
l-Os(dipp)a" f d-Fe(phen)s'I-Os(dipy)a++f l - F e ( p h e n ) ~ + + Os(dipy)s f Fe(dipy)a Os(dipy)P+ Ku(dip~)3~+* Fe(phen)s+' Ru(dipy)s+++ +
0
2
4
6
8
IO
+ - +
< < < <
3 X >3 >3 >3 >i' 6 >5 >3
106 106
X l0eh
X 106' X 106' X 10'
x
x
1044 io0
Second-order rate law assumed. 23". 8 X JI reactant concentrations. Colors of + 2 ions appeared very similar in 1-mm. capillary of rapid mixer, so reaction was observed in 40-nil. centrifnge tube. a
-
. .
+ +
i
3.0 0.1 3.0 3.0 3.0 3.0 0.1 0.1
Halftime, sec.
