846
J. SILVERMAN AND R. W. DODSON
Vol. 56
THE EXCHANGE REACTION BETWEEN THE TWO 0XIL)ATION STATES OF IRON IN ACID SOLUTION1 BY JOSEPHSILVER MAN^ AND R. W. DODSON Chemistry Departmerd, Columbia University, New York, N . Y . , and Brookhaven National Laboratory, Uplon, N . Y . Reeeiued March 13, I068
The rate of exchange between ferrous and ferric iron in perchloric and hydrochloric acid solutions has been mcasurcd at ionic strength 0.55f: at temperatures from 0 to 20". The reaction is first order in the over-all concentration of each oxidation state. The rate increases with decreasing acid concentration and with increasing chloride concentration. The results suggest four reaction paths involving ferrous ion, namely, with: (1)unhydrolyzed ferric ion, Fe+++,(2) FeOH++, (3) FeCl++ and (4) FeCh+. The specific rate constants at 0" are 0.87, 1.0 X loa,9.7 and about 15f-1 sec.-l, respectively, for these reaction paths. The corresponding experimental activation energies are 9.9, 7.4, 8.8 and about 10 kcal./mole, respectiyely. The rate increases with ionic strength. No evidence was found for a specific perchlorate ion effect.
Introduction The exchange reaction between ferrous and ferric iron in 3 f perchloric acid solution has been the subject of several investigations. N a h i n ~ k yemploy,~ ing a chemical method to separate the exchanging species, reported the exchange to be complete in five seconds. Van Alten and Rice4 used a diffusion technique to separate the reacting species in order t o avoid catalysis which might be induced by a chemical separation. They observed a halftime of 18.5 f 2.5 days. This latter work also yielded the rather unexpected result that ferric iron diffused twice as rapidly as the ferrous species. Linnenbom and Wah15 repeated the Van AltenRice experiment but obtained entirely different results. They found that the half-time was less than two hours and that ferrous iron diffused 40 f 20% faster than the ferric iron. Kierstead6 also performed this same experiment. He reported halftimes of 39 f 2 days and 166 f 12 days, and stated that the ferrous iron diffused three to four times faster than the ferric iron. Betts, Gilmour and Leigh,' using the diffusion separation method, found that the exchange was complete in one-half hour and that the ferrous iron diffused 22 f 4y0 faster than the ferric iron. A chemical separation recently devised8 has yielded half-times of the order of secpnds for the exchange reaction. The present work presents a study of the kinetics of the reaction, made through the use of a modification of this separation. The data obtained confirm the preliminary experimentsJsand are consistent with the results of Linnenbom and Wahl, of Betts, Gilmour and Leigh, and of Nahinsky. However, they are in complete disagreement with the results of Van Alten and Rice, and of Kierstead. In the experiments reported below, the rate of the exchange reaction has been determined as a function of ferrous and ferric concentrations, acid concentration, chloride concentration and tempcruture. The effcct of various neutral salts was also observed.
Materials.-A ferric perchlorate st,ock solut,ion was prepared from a hydrochloric acid solution of high specific activity Fe66, obtained from the U. S. Atomic Energy Conimission, Oak Ridge, Tennessee. This nuclide decays by I< capture with a 2.94 year half-life9 emitting 5.9 kev. manganese X-rays. The radioactive iron was extracted int,o isopropyl ether by a standard method10 to free it from metal contaminants, and was then back-extracted into water. In order to convert this to a ferric perchlorate solution of (lasired concentration and specific activity, additional inaci ivcb iron was required. Iron wire, obtained from the Gencml Chemical Co., was dissolved in nitric acid and oxidized io the ferric state with hydrogen peroxide. After the excvss peroxide had been destroyed by prolonged boiling, the solution was mixed with the purified radioactive iron solutioti and then made alkaline with ammonia. After several cycles of solution with perchloric acid, precipitation with ammonia, and washing with water, the washings gavc negative t,csts for chloride, sulfate and ammonium ions. The purified ferric hydroxide was t,hen dissolved in a calculatcd volumc of standardized perchloric acid to produce a st,oclc solution of 0.00110 f ferric perchloratc in 0.0507 f perchloric acid. This solution also gavc negative tests for chloride and sulfate ions. A master stock solution of ferrous pcrchlorate was p ~ ! pared by successive recrystallizations of ferrous perchlorate hexahydrate (obtained from the G. Frederick Smith Go.) from 8 f perchloric acid. The resulting solution was about 0.3 f in ferrous perchlorate and about 2 ,f in perchloric acid. I t gave a negative test for sulfate ion but was about 0.004 .f in ferric perchlorate and about 5 X 10-6 f in chloride ioii. An aliquot of the solution was analyzed for ferrous ion by titration with standard ceric sulfate. Another aliquot was passed through a Jones reductor and titrated with ceric sulfate to give the total iron concentration. A third aliquot, was treated with hydrogen peroxide to oxidize all the iron to the ferric state. AFtcr the excess peroxide was boiled OH, the hot sample was titrated to the phcnolphthalein end-point. with standard sodium hydroxide, thus precipitating all thcl iron as ferric hydroxide and neutralizing all the acid.'l The perchloric acid concentration was determined from this titer after t,aking account of the hydroxide used up in precipitating the iron and t,hat produced during the oxidation process. A stock solution containing 0.0011 f ferrous pcrchlorat,e, 1.5 X 10-5,f ferric perchlorate, and less than 10-6f chloride ion was periodically prepared from the master solution. Each time such a solution was prepared, the Inastcr solution was re-analyzed. Both stock solutions of ferrous : a t about -5" in order t>oinininiizc perchlorate ~ e r (stored the oxida1,ion of the fcrrous iron to thc ferric st.atc by (lissolved oxygen and perchlorate ion. Sodiuni perchloratc solutions \cere niadc U ~ Ifroni 1 lie nionohydrat,e, obtained from tlic Einicr and Anielld (h. (1) This work was carried out e t Brookliaven National Lalm%tory These solutions gavc negative tests for chloride and sulrltl o ions. Salt concentrations wcrc determined by- specific under the auspices of the United States Atomic Energy Cominission. (2) Now at Walter Kidde Nuclear Laboratories, Inc., New York, gravity measurcmcnt. N. Y. Pcrchloric acid (70% vacuum distilled), purchased from (3) PI1.D. Thesis. ITniversity of California, 1942. ,J. T. I3akcr C'hi~riiicd Po., \WLS sl.t~~idarclixc~l by 1 ilration (4) L. \'an Altan au(l C . N . ICii:n. J . A N I .C ' I I P ~ So,,.. L. 70, 88.1 (1948) ( 5 ) V. ,J. Liniieiil~oi~ aiid A . C. Wahl, ibid.. 71, 288Y ( I Y - l Q ) . (6) H. A. liierstend, J . Cliem. Phys., 18, 756 (1950). (7) R. H.Betts, H. S. A. Gilmour and K . Leigh, J . A m . Chem. Soc., 72, 4978 (1950). (8) R. W. Dodson, ibid., 72, 3315 (1950).
__
. . (9) Ci. L. Brownall aiid C . .I. hialctakos. l ' h ~ s Ileu., . 80,1103 (1950). (IO) R. W. Dodson, G . J. Forncy and E. I{. Swift, J . Am. Chem. SOC.,68, 2573 (1936). (11) W. C. Sohumb and S. B. Sweetser, i b i d . , 67, 871 (19353,describe this method of titrating acid solutions of ferric iron.
Oct., 1952
847
EXCHANGE RATESOF FERRIC AND FERROUS IRON IN ACID SOLUTION
with standard sodium hydroxide. It gave negative tests for chloride and sulfate. Constant boiling hydrochloric acid, whose concentration was also checked by titration with st8rtndard sodium hydroxide, was the source of the hydrochloric acid solutions employed in the experiments described below. A solution containing 0.05 f perchloric acid and 0.025 f a,a'-dipyridyl was pre ared. The latter was obtained from Eastman Kodak and &mer and Amend. Lanthanum oxide waa obtained from the Varlacoid GO. A s ectrographic analysis detected O . O O l ~ o praseodymium as t i e only metal impurity. The oxide was heated to 1000" for a few hours; after cooling it, was treated with slight'ly less than an equivalent amount of perchloric acid. The filtered solution was 1.91 f in lanthanum perchlorate. The lanthanum concentration was determined by precipitating the hydroxide from a 50% alcohol solution and then filtering, igniting and weighing the precipitate. The solution gave a negative test for chloride ion but proved to have a 9 X 10-4f sulfate ion content as determined by the barium sulfate gravimetric method.'* All other chemicals employed were standard chemically pure reagents.
Experimental Procedure The ferric fraction of the reaction mixturc was prepared by thc addition of appropriate volumes of the stock solutions t.o a wide-mouthed amber reagent jar (about 400-ml. capacity), which also served as the reaction vessel. The ferrous fraction was prepared in a 125-ml. erlenmeyer flask. The concentrations of the two fractions in a particular run were nearly identical with respect to all components except for ferrous and ferric iron. This precaution minimized heat and volume effects which might accompany mixing. Both fractions were brought to constant tem erature ( &0.lo) in a thermostatically-controlled bath. eT ! reaction was started by rapidly pouring the ferrous fraction into the reaction vessel. During the mixing process, the contents of the reaction vessel were stirred by a motor-driven glass propeller. Aliquots of the reaction mixture were withdrawn a t measured time intervals. Each aliquot was transferred to a 100-ml. beaker and the reaction was immediately quenched by the rapid successive addition of 10 ml. of t.he dipyridyl stock solution and enough sodium acetate solution to bring the p H to about 5. The function of the dipyridyl is to render the ferrous iron inert to exchange processes by formation, rapid a t this pH, of the highly stable, intensely red fcrrous tris-or,d-dipyridyl ion .13 The two iron species were then separated by precipitation of the ferric iron. I n the absence of lanthanum salts, ferric hydroxide was precipitated with ammonia. When lanthanum was prescnt, the prccipit,ation was performed wit,h 8-hydroxyquinoline. Bccausc of tjhe low reactant concentrations employed, an aliquot of an inactivc ferric: pcrchlorate solut,ion was added following the precipitation to increase the mass of the precipit,ate. Also an aliquot of a fine supension of diatomaceous earth (Johns-Manville SupcrGel) was added to aid the filtration. The amount, of diatomaceous earth added to each sample was such that it took a 15% variation in this amount to produce a 1% change in the self-absorption of the radiat,ion emitted by the precipitate. The precipitates were filtered on Whatman number 42 paper, washed, dried, mounted on cards, and covered with cellophane squares. Most of the samples were counted with end-window Geiger tubes connected to standard scalers. Some runs were counted on a sensitive proportional counter designed for X-ray counting.l4 With this instrument set so as to count only the K lines of manganese, the data confirmed the results obtained with the Geiger count,cbr. 111 one cxperimcnt, pcrfornicd in the study of tho cliloride dependence, the fcrrous fraction was also couritcd. The fillrates froni the fcrric iron tilt,ration were acidified with hydrochloric acid and boiled until they turned colorless. Bromine water was added to oxidize t.hc iron to the ferric state. Counting samples were then prepared from the resulting solutions in a manner identical with that described above for the ferric iron. (12) E. H. Swift, "A Systeiir of Clreinical Aiialy6a." Prentice -Hall Publishers. IIIC.,New York. N. Y., 1939, p. 478. (13) J. €1, Baxendale a n d P. George, Trona. b'aradav SOC.,46, 55, 736 (1950), have made equilibrium a n d kinetic studies of this ion. ( 1 4 ) W. Bernstein. I i . G. Brewer, Jr., a n d W. Rubinson, A'ucleonics, 6, 30 (1950)
400.
?QO
-
2o 100
500
1500
1000
2000
2500
SECONDS.
Fig. 1.-Typical
rate data.
During each run, eight aliquots werc analyzed. The first six were withdrawn before four half-lives had elapsed. The last two were infinite time samples, which were withdrawn after a lapse of a t least ten half-lives. In cases where oxidation of ferrous iron was cxpccted to provide complications, the seventh was removed after the passage of ten half-lives and tho eight.h was removed after a t least fiftccn half-lives. The pairs of infinite time samplos usually agreed within about 3%.
Results Order of the Reaction.-The ordcr of the rcbaction was determined by independent variatioiis of t)hc conccntrationv of the react,ants. If there are no slow consecutive reactions and the conccntrat'ions of the reactants do not vary wit,h tdmc, thc time tiopendcncc of t>hc fraction of ac:t,ivity cxchanged is givcii by t,he rate law16
whore yo, and represent, thc: spwific activities of the initially activc: iiiatcrial a t xci'o t.inic, time t, and infinite time, respectivcly. R is the constant rate a t which the over-all exchange process occurs, and a and b are the gross concentrations of the two react,ant,s. Representative experimental curves of In (y - y m ) versus t arc shown in Fig. 1. The fact that these plots arc st,raight lines, in agreement with the properties of equation ( l ) ,indicates that the assumptions involved in the derivation of this equation apply to the systems under considcration. The rat,(!law may also bo c!xpr.cwc!cI by R siniilai- (!quation which tlc!sc:ribcs thc: growth of cwtivi1.y with t,iinc i t i the initially inactjive inrtterial
where x aiitl x m wpresctit the specific txtivity of ih:inilially inactive material at timet and infinite time. In the instance where the ferrous fraction was analyzed, a plot of In (1 x/xm ) versus f yieltlctl the anticipated linwr itlationship. Values of' the cischange rate R wcre coniputed from the half-timv of' plots siicli as t.hose in Fig. 1. 11. has been shownll: t 1 1 i ~ t iiiootiil)lel,(~sq)tc.mt.ion 01' partial ( v ~alysis t 01' (15) (a) H. A. C. bIcKay, Nature, 142, 997 (1038); (bj R. B. Duf. field a n d A I . Calvin. J . A m . Chem. Soc., 68,557 (1946). (16) R. Prestwuod a n d A. Wahl, ibid., 71, 3137 (1949).
J. SILVERMAN AND R. W. DODSON
848
Vol. 56
TABLE I DETERMINATION OF THE ORDEROF (HCI),
f
F > CI > Br > I. On the other activated complexes in I and I1 are identical. An X atom crease their effectiveness as “conductors” should decrease in transfer leads to exchange only in the case of mechanism I. hand, the order: I > Br > C1 > OH > F. A study of the relative effectiveness of these ions in facilitating electron transfer ORLO E. MYERS(Oak Ridge National Laboratory).Have you considered the possibility that perchlorate ion should permit a quantitative analysis of the relative immight be a component of the transition state in your Fe++- portance of these opposing factors. Fe+++ exchange work? (Apparently Clod- can enter into R. W. DoDsoN.-Further studies of the effectiveness of 12-10,; and Hg(1)-Hg(I1) cyanide exchange and perhaps anions in facilitating exchanges between different oxidation the ionic strength effects observed in Tl(I)-Tl(III) exchange states will almost certainly give informative and interesting might be approached by considering the possibility of data, which bear on the question of electron transfer C104- participation. This might mean that all four ob- verms atom may transfer as a mechanism. We are inclined to served paths are of the same general type. think that the exceptionally large rate constant for FeOH++ R. W. DoDsoN.-This suggefition has also bcen made by may be understood in terms of a hydrogen atom transfer Sutton (Nature, 169, 71 (1052)). We feel that the present reaction instead of one involving only electron transfer, as evidence does not permit a conclusion on this point. In discussed a t the meeting. In this picture, the exchange is the present work there is no significant evidence for a per- brought about by the transfer of a neutral hydrogen atom chlorate ion effect; and the participation of perchlorate in from the hydration sphere of ferrous ion to the OH group the Hg(1)-Hg(I1) exchange has the status of a speculative of the hydrolyzed ferric ion. suggestion. \Ve have work in progress which may give inC. S. GARNER (U. C. L. A.).-It may be of interest to reformation relevant to these questions. mark that King and I have submitted a letter to the Editor HERBERT C. BROWN(Purdue University).-Study of the of J . Am. Chem. Soc., describing our observations on the gas phase dissociation of “bridged” molecules shows that vanadium(I1)-vanadium( 111) exchange in perchloric acid, the effectiveness of the bridging decreases from nitrogen to a system which formally resembles the iron( 11)-iron(111) oxygen to chlorine. For example, in the dissociation of system. We found complete exchange in the dark at 2” in derivatives of trimethylaluminum [N. Davidson and H. C. exchange times of the order of one minute and for concentraBrown, J . Am. Chem. Soc., 64, 316 (1942)], we observed tions of the vanadium species of ca. 0.06-0.1 f, Separation was effected with a+’-dipyridyl in some runs, and with a that the ease of dissociation increased in the order: cation-exchange resin in others. The possibility exists that (Me2AINMe2)z < (MerAlOCH& < (MetAICI)2 the exchange is comparable in rate to that for the iron sysThe large difference in the rate constants observed by Silver- tem. Incomplete exchange was observed between the trisman and Dodson for FcOH++ and FcCl++ can be under- (a,a’-dipyridy1)-vanadium(I1)ion and vanadium( 111) ion s h o d in tcrms of thc rclstive cffectivcncss of oxygen and in 0.5 f perchloric acid a t 25” for concentratioiis of ca. 0.03 f chlorine in “bridging” two metal ions. In terms of the for the vanadium species. ++++
+ +
+ +
