D. R. STRANKS
906
AND
G. M. HARRIS
Vol. 56
ISOTOPE EFFECT IN SOM,E REACTIONS OF CARBON-14-LABELED CARBONATOTETRAMMINE COBALTIC COMPLEX ION BY
D. R.
STRANKS AND
G. M. HARRIS
Department of Chemistry, University of Melbourne, Melbourne, Australia Received March 6. 1966
The exchange equilibrium in aqueous solution between carbonatotetrammine cobaltic complex ion and free carbonate ion has been investigated by means of radiocarbon tracer. The equilibrium constant of the reaction, defined 8s the ratio of equilibrium specific activity of complexed to that of total free carbonate, was 0.875 d= 0.002 a t OD,increitsing to 0.900 f 0.004 at 30 , The apparent heat ofreaction waa - 140 f 26 cal./mole. The unidirectional acidic decom osition of aqueous solution of labeled carbonato complex, studied by a process of fractional examination of product specilc activity, exhibited no significant isotope effect. The results of both the e uilibrium and decomposition reaction investigations are satisfactorily interpreted by a qualitative application of statistic3 mechanical theory.
Carbonatotetrammine cobaltic 'complex salt (CTC) undergoes a moderately rapid exchange reaction with uncomplexed carbonate in aqueous solution.' A preliminary study2has indicated that the equilibrium distribution of radiocarbon tracer among the exchanging species is appreciably nonstatistical. The present paper describes a more detailed examination of this isotope effect, including its temperature dependence. In addition, the ratio of the rates of acidic decomposition in aqueous solution of labeled and unlabeled CTC were determined. These related researches provide, for a single carbon-14-labeled compound, investigations of two of the types of chemical process in which isotope effects have been previously observed-exchange equilibration3and unidirectional decompositi~n.~ The equilibrium under consideration may be regarded as being made up of either one or both of two possible over-all exchange systems Co(NHa)rCOa+
or CO(NHs)4C03+
+ H C * O a - Z Co(NHs)iC*03+ + HCOs-
+ C*Oa'
IT
Co(NHa)rC*Oa+
+ co3'
The first of these singly is supported by the kinetic investigation mentioned.' However, the equilibrium constant of the system
coi- + HC*Os-
__
of COS- or HCO3- ions. In any case, the practical equilibrium constant
is a straightforward definition of isotopic enrichment effects in this system. Given in terms of specific radioactivities of carbon-14 tracer, K is simply the ratio of the equilibrium specific activity of CTC to that of total uncomplexed carbonate. The unidirectional decomposition of an isotopically labeled compound to give products, only one of which is to contain labeling atom, may be represented k
...+ P + Q + ... k* A* + B + . . . +P* + Q + . . . A+B+
With label present in tracer concentration only, the relative rates of formation of product molecules P* and P is given by
d[P*l
k*[A*] d[P] '= klA]
irrespective of the reaction mechanism.6 Proceeding in a manner analogous to the treatment of the case of fractional examination of reactant,' it is readily shown that
c*os' f HCOa-
has almost certainly a value of approximately unity in the neighborhood of room temperature.5 It is therefore of no great importance whether the complex ion/carbonate exchange is considered in terms (1) G.M.Harris, J . Chem. Phys., 18, 764 (1950); G. M. Harris and D. R. Stranks, Trans. Faraday Soc., 48, 137 (1952). (2) D. R. Stranks and G. M. Harris, J. Chem. Phys., 19, 257 (1951). (3) H. C. Urey, J. Chem. SOC.,562 (1947). gives a full review of stable isotope exchange work, including C'*/C" systems. Nothing other than reference (2) above has yet been published on isotope effects in Cl*/CI( exchange equilibria. (4) Carbon-I4 effects in such reactions have been reported by the following: Albert L. Myerson and F. Daniels, Science, 108, 676 (1948); P. E. Yankwich and M. Calvin, J. Chena. Phya., 17, 109 (1949); J. W.Weigl and M. Calvin, ibid., 17, 210 (1949); W. H. Steven8 and R. W. Attree, Can. J. Res., 827, 807 (1949); J. Chem. Phus., is, 574 (1950); W. 0. Armstrong, Leon Singer, 9. H. Zbarsky and B. Dunsbee, Science, 112, 531 (1960); A. Roe and M. Hellmann, J. Chem. Phya., 19, 660 (1951); E.A. Evans and J. L. Huston, ibid., 19, 1214 (1951). ( 5 ) Experimental data reported i n reference (3) suggest a value of very close to unity for the constant as applied to a Cl*/Cl* system. Statistical mechanical calculations t o be discussed below lead to a figure of 0.982 for the corresponding CI'/CQ constant a t 293*, with a very Blight temperqture coe@cient,
In this, E is the rate constant ratio lc*/lc, and the subscripts 0 and y refer to concentrations initially and at fraction of reaction y, respectively. In terms of specific radioactivities, the expression above is closely approximated up to a high fraction of reaction by SrlSo
=
€(I - y)c-l
(2)
Here, Sy represents specific activity of small prodiicf increment collected over interval of reaction of average fraction y , and So is the initial specific activity of reactant. The form of this function is such that small divergences of E from unity are considerably magnified during the last stages of reaction.a
..
With only ( 8 ) Let the over-all rate law be of the form k[Alm [B]". tracer concentration of labeled reactant A*, the corresponding rate law governing formation of P* Gill be of the form k*[A*][A]"' -1[B]".. , , since the probability of molecules A* participating more than onae in the mechanism is negligible. (7) A. M. Downes and G. M. Harris, J. Chem. Phys., '20, 196 (1952)). ( 8 ) For example, with e = 0.9, the increment of product taken between y = 0.85 and y = 0.90 will have a n average Sy value about 217" greater than its initial value. For the y = 0.90 to y = 0.95 increment, the corresponding figure is 30%.
Oct., 1952
ISOTOPIC
EFFECT IN
007
CI4-LABELED CARBONATOTETRAMMINE COBALTIC ION
It is clear that fractional examination of product specific activity should enable detection of any appreciable isotope effect, notwithstanding the unavoidable experimental unccrtainty in such activity determinations. Referring specifically t o the acid decomposition of CTC in aqueous solution, the over-all reaction is known to be9
+
C O ( N H ~ ) ~ C O ~2H+ +
+ H2O +
Co(NHs),(HzO)z+++
+ Cor
One can make successive additions of small portions of strong mineral acid to tracer-labeled complex salt, and determine the relative activities of the series of carbon dioxide samples evolved. A reasonable approximation to the relevant E factor should then be attainable by application of equation (2).
Experimental A. The Exchange Equilibrium.-CTC nitrate and carbon-14-labeled sodium carbonate were prepared as previously described.’ Aqueous solutions of the reactants were equilibrated at constant temperature (f0.01’) for a minimum of seven half-times of exchange (i.e., to witliin 0.78% of complete equilibrium). The required half-times for various conditions of reactant concentration, pH, ionic strength, and temperature were known from the previous kinetics work.’ After equilibration, uncomplexed carbonate was precipitated as barium carbonate, mounted, and assayed for radioactivity exactly as described in the earlier study. Assay of the filtrate containing the CTC was effected by means of a semi-micro modification of a standard closedsystem technique of carbonate determination.10 A typical experiment proceeded as follows. After evacution of the apparatus to water pressure, 6 N COz-free hydrochloric acid (2 ml.) was run from an attached funnel into a portion of CTC solution (0.2 ml. in 4 ml. Cot-free water) contained in the main reaction vessel. When evolution of gas had ceased, the acidified solution was boiled by immersion in an oil-bath, steam and COZbeing driven over into a cool sidearm containing 0.5 N sodium hydroxide (1 ml.). Ten-minute cooling periods were alternated with the boiling operation. Test experiments showed that carbonate recovery was complete after the third repetition of the boiling-cooling routine. COe-free air was then admitted, and the carbonate solution transferred with washings to a centrifuge tube. Barium carbonate samples were prepared essentially as before, except that in order to obtain precipitates amenable to spreading much slower addition of precipitating reagent was necessary than in the case of less alkaline carbonate solutions. Despite all precautions, occasional samples were obviously badly spread and were discarded. Moreover, any samples the mass of which differed by more than 0.2 mg. from the mean mass (about 3 mg. usually) of a series of identical samplings were also rejected. All counts were done with a conventional thin-end-window G-M tube and acce8sories. They were sufficiently prolonged to reduce the random standard deviation to within 0.5%, and were repeatedly checked. B. The Acidic Decomposition.-Carbon-14-labeled comnlex salt was prepared bv equilibration of inactive CTC in solution with ;adioactiveun;omplexed carbonate. After 13 hours reaction at 25” (about 4 half-times) the solution was cooled to O ” , and ice-cold absolute alcohol was added. The precipitated active CTC nitrate was filtered off on a sinteredglass funnel, thoroughly washed in ice-cold alcohol, and dried in vacuo over anhydrous silica gel. The final product, recovered in 83% yield, contained no detectable free carbonate ion. The step-wise acid decomposition of labeled CTC was accomplished by means of the apparatus illustrated in Fig. 1. COz-free hydrochloric acid (0.2006 N ) was added in accurately measured aliquots from the small funnel to the 8-ml. capacity decomposition flask containing solution of a known (9) Abegg, “Handbuch der anorganisohen Cbemie,” Hirzel, Leipaig, 1935: Bd. IV, Aht. 3, Teil 3, p. 771. ( I O ) H. T. 5. Britton, “Hydrogen Ions.” Vol. 11. Chapman and Hall,
Irondon, 1942, p. 17%.
i I
INCHES Fig. 1.-Acidic
decomposition apparatus.
weight of active CTC nitrate in 2 ml. of CO2-free water. The volume of acid aliquot for each fraction of decomposition was determined by addition of a definite number of drops of known average volume. Tests showed that this procedure was very accurate when the rate of drop addition was carefully standardized. Decompmition of CTC was known from pH measurements to go to near completion very rapidly. For example, a t 25O, the temperature of t>heexperiments, the pH rose from an initial value of ca. 2.2 immediately on addition of the acid to 4.2 in 30 seconds, after which only a slight further rise slowly occurred. Temperature control of the reaction vessel to within f 0.02’ was maintained by immersion of the spring-held decomposition flask in a small thermostat. The COz produced (about 0.015 milliniolc per fraction of decomposition) was entrained in purified oxygen bubbled through the reaction mixture.” Sweeping was continued for five minutes after addition of acid, the COZ-O~mixture bubbling through COz-free sodium hydroxide (0.5 ml. of 0.5 N solution) contained in the 1.5-1111. capacity tapered absorption tubes. The resultant carbonate solutions were transferred to a centrifuge tubc by dropper pipet, together with two washings of the absorption tubos and bubbler capillaries. Prcparation of the barium carbonate platcs and their counting was carried out as already indicated. Preliminary tests showed that quant,itat’ive recovery of carbonate radioactivity was achieved by the procedure outlined. However, despite rigorous purification of reagents and careful attent.ion to details of technique, a small blank of inactive Carbonate could not be eliminated. By standardization of the routine, especially as regards timing, it was found possible to make the blank consistently reproditcible at 0.265 f 0.005 mg. of barium carbonate per sample. With this constant blank, all radioactive assay results were readily expressed on a common standard of comparison.
Results A series of exploratory experiments wag first carried out a t 20.3’ to detect possible major dependences of the ewhmge equilibrium constant on conditions such as PI€, reactant, ratio, ionic strength, light intensity, and surface or cuntainer. The results are recorded in Table I. It is ohviolis (1 1) Boiling could not be employed here to drive off the
g1t9
since
tthermal decomposition of the residual CTC w)uld alw t?tke place.
D. R. STRANKS AND G. M. HARRIS
908
Vol. 56
TABLE I EFFECTOF CONDITIONS OTHERTHAN TEMPERATURE ON EXCHANGE EQUILIBRIUM CONSTANT Expt.
No.
1 2 3 4 5 6 7
8
Cnncn. CTC (nrole/l.)
Concn. CO: HCOI (niole/l.)
0.03376 .03376 ,05746 .03710 .03710 .03379 .03379 .03389
0.01469 .01469 .01460 .01604 .01604 .01471 .01471 .01470
+
Mean K for Expt.
No. uf detn.
2 2 2 2 3 1
0.886 . 892 .894 .890 .893 .887 .895 .891 0.891
1 1 14
Remarks
pH ca. 10 pH cu. 9 pH 9.47 pH 9.7,p = 0.0799(adjusted by addition of NaNOs) pH 9.7,p = 0.4OOO Exchanged in absence of light. Exchanged under illumination by adjacent 15w.tungsten light Reaction vessel packed with glass granules 10.002weighted mean K and its standard deviation
TABLE I1 TYPICAL SERIESOF K DETERMINATIONS AT 273.2"A. (Applies t o Table I11 also): ma, mg = mass of BaCOa samples from free carbonate and from CTC, respectively; A a , Ag = activities (ct./mg. BaCOa/min.) of free carbonate and CTC samples; y = [CTC]/[free carbonate] = 3.936 Detn No.
1 2 3 4 5 6 7 8
ma, n x .
mg, mg.
2.74 2.68 2.76 2.72 2.73 2.70 2.76 2.71
2.81 2.68 2.67 2.82 2.77 2.72 2.80 2.70
A,
Mean S.D. of Mean Initial activity (check only -mean of 2 detn.)
AI¶
YAB
A,
+ YAP
224 226 224 224 225 226 223 222 224.3 f 0.49
196 194 195 198 196 197 196 196 196.0 1 0.42
772 764 768 779 772 775 772 772 771.6 i 1.65
996 990 992 1003 997 1001 995 994 996.0 1 1.72
996
....
....
996
K
Ag/Aa
0.875 .858 .871 .8&1
*
,871 .872 .879 .883 .874 ,003
TABLE 111 TEMPERATURE DEPENDENCE OF EXCHANGE EQUILIBRIUM CONSTANT Expt. No.
Temp.
1 2 3 4 5 6 7
273.2 273.2 288.2 293.5 298.2 298.2 303.2
(O.4.)
No. of detn.
Mean Aa
Mean A#
8 8 8 6 8 8 8
224.3 224.1 220.0 210.8 211.0 205.7 213.8
196.0 196.1 194.0 187.5 188.2 184.3 192.5
that K is constant to well within 1 1% of 0.89 a t this temperature, regardless of the other variables. A more extensive series of determinations was then undertaken over the temperature range 0-30", other conditions being fixed ~ n 4follows: pH 9.47; concentration of CTC = 0.05746 mole/l.; concentration of free carbonate = 0.01460 inole/l. Higher temperature studies were not attempted because of the observable thermal decomposition of CTC in aqueous solution above 20". (On long standing, a brownish opalescence due to cobaltic oxide precipitation appears.) In order to eliminate this complication in the experiments a t 25 and 30°, the solutions were first equilibrated at 0' (a minimum of seven half-times of exchange), then maintained at the higher temperature for about three additional halftimes. This procedure can be shown to have brought each system finally to within the experimental error in the determination of the true higher temperature equilibrium condition. A total of 54 independent determinations of K was made. The nature of the results obtained is illustrated in detail for one series of measurements at 0" in Table 11. Table 111 Summarizes the data for all experiments, together with the mean K value and its statistical standard deviation, at each temperature. A least squares analysis was performed on the assumption of the linear relationship: log K = - A H / 2.803RT constant. Each of the 54 K values was treated individually and of equal weight. The value of A H 2 tho
+
AB
+ yA#
996.0 996.1 983.7 948.7 951.8 931.1 971.5
Mean K and
Initial Bpecifio activity of free carbonate (check only)
996 998 992 954 962 935 980
its S.D.
} 0.875f 0.002 .882i ,003 ,891f .005
} .894f
,002
.goo*
.004
apparent heat of reaction, so obtained was -140 cal./mole. with a statistical standard deviation of i 25 cal./mole. The fractional acidic decomposition data are summarized in Table 1%'. Individual masses of the barium carbonate
TABLE IV FRACTIONAL ACIDIC DECOMPOSITION OF LABELED CTC SALT Mean No. activity T6tal ma88 of samples Specific activities of and its FracRecov., Calod., individual fractions std. Expt. tions mg. mg. No. taken (Counts/rng. BaCOa/min.) deviation 1
2 4 5
3
10
770, 730, 774, 798, 765 802, 803, 826, 779, . . . a 11 796. 801, 801, 785, 792, 796 789, 800, 807, 818, 796 10 781. 819, 781, 796, 795 765, 792, 784, ...,a 786 21 796.770, 789, 796, 812,783, 786, 788, 798, 789, 795, 820, 822,808.814,801,801,796, 804, 785, 783 . b 780, 785, 802, 798, 795, 800
783
f 9.3 798 f 2.7 789 f 4.8 797
i 2.9
32.13
32.44
31.05
32.44
31.36
32.48
64.08
64 88
19.23
19.46
795
i 3.6
Sample spoiled accident,ally. Samples repared by complete decomposition of equal portions of kbeled complex in order to give independent estimate on SOvalue. a
Oct., 1952
ISOTOPIC EFFECTIN C '-LABELED CARBONATOTETRAMMINE COBALTIC ION
samples (all close to 3 mg.) are not recorded. However, the total recovered mass for each series of determinations is seen to agree well with that redicted by calculation from the known quantity of CTC s a i taken and the known standard blank. Inspection of the figures does not suggest any obvious trend in the activity values as reaction proceeds. This observation is confirmed by detailed least squares analysis of all individual activities, treated in term of equation (2) in the form log s y = log 8 0 s ( 6 - 1) log (1 - 7 ) No appreciable deviation of e from unity was indicated," from which it may be concluded that the over-all isotope effect operative in this decomposition process is very slight.
+
Discussion The equilibrium constant of the system, as has been shown above, can be defined in either of two ways. In terms of partition function ratios, the expressions are KI = fCTC/fHCOI Or Kz,= fCTC/fCOl the f's being the ratios of the functions for carbon14-labeled and normal entities, respectively, (f = Ql4/Ql2). The molecular structure data available on such a complicated substance as CTC ion are, of course, inadequate for any direct estimation of the magnitude of its partition function. Nevertheless, by postulating probable structural modifications of a carbonate ion on becoming complexed, a reasonable qualitative interpretation of the observed results is possible. In the CTC ion, two of the carbonate oxygen atoms are covalently attached to the central cobaltic ion.'$ This binding modifies the symmetrical XY3form of the free carbonate ion to what is in effect an XYZa-type of structure. Moreover, the two bonds between the coordinating oxygens and carbon are weakened on complex formation, and the effective mass of these oxygens is increased (say by one-sixth of the cobaltic ion mass). Vibrational modes of the CTC ion not directly associated with its constituent carbon atom will be but slightly affected by the isotopic substitution, so may be neglected in the evaluation of ~ C T C . A reasonable estimate, therefore, of the order of magnitude of ~ C T Cshould be given by evaluating the corresponding function for a molecule approximating to the postulated form of the complexed carbonate ion. Such a molecule is phosgene, C0C12. The partition function ratios required for the determination of K 1 and K z according to the scheme just outlined have been calculated. The procedure was analogous to that of Urey3 for C13/C12systems. Table V records the data relevant to the present discussion, including figures for fco,and a constant K a to be referred to later. The details of the calculations, together with the results obtained for a number of other possible C14/C12 exchange systems, will be published separately." It is seen that the earlier assumption of near equivalence of HC03- and COa- in the exchange system is justified, since both predicted K's are in good qualitative agreement with the experimental find-
-
-
(12) The actual figures obtained were Sor 790.2 += 1.4 and (e - 1) -0.004 f 0.0014, leading to estimates of 0.994 f 0.0048 and 0.996 f O.OOl4, respectively. (13) L. Pauling, "Nature of the Chemical Bond." Cornell University Presn, Ithaca, N. Y.,1940, Chspt. 111. (14) D. R. Stranks and G. M. Harris. to be published.
909
ing (see Table 111). The AH values, calculated from the K1 and K a data according to the conventional isochore treatment, are -115 and -104 cal./mole, respectively, again reasonably consistent with the observed AH(-140 & 25 cal./ mole). TABLEV PARTITION FUNCTION RATIOSFOR VARIOUSMOLECULES Temp., "A. frrcor/COS/COI 273.18 1.5085 1.4984 1.4444 293.16 1.4507 1.4239 1.3973 313.16 1.4025 1.3784 1.3580
-
fcoclr
fCTC Ki K: KI 1.2948 0.858 0.878 1.118 1.2834 .871 .887 1.106 1.2370 .882 ,897 1.098
In order to consider the acidic decomposition reaction, the first requirement is to assign it a definite mechanism. That suggested by the previously published exchange kinetics study' is ki + Ha0 + CO(NHa)rHCOs.HpO kz CO(NHs)dHCOa.HzO + H30+ + CO(NHa)rCOa+
++
+
++
+
Co(~Ha)4(&0)z+++ HzO
+ GO1
with the second reaction rate-determining.16 On the basis of the statistical mechanical theory of isotope effects in unidirectional reactions16 the rate constant ratio required for the present system is given by (3)
This assumes equality of transmission coefficients for labeled and unlabeled reactant. The symbols f~ and f~ are the partition function ratios as defined above for activated complex and bicarbonatoaquotetrammine cobaltic ion, respectively, The quantities m are defined as the reduced mass of the atoms forming the bond broken in the reaction," in each instance here a cobalt-oxygen bond. The ratio m/m* will thus have a value unity. The activated complex may be visualized as H
H
0 '
.0 , .
I
I c=o \
H
Clearly, the vibrational characteristics of the coordinated bicarbonate ion should not be greatly modified by the activation process. It is therefore reasonable to expect that f~ and f~ will not differ greatly in magnitude, if at all. Consequently, k2*/k2 should have a value close to unity, as is found experimentally. A similar argument to that just given can be applied with a similar result if kl is assumed to be ratedetermining rather than kz. With other possible (15) In view of the rapidity with which the carbon dioxide is formed and removed from the system (aee Experimental) it ia deemed unnecessary in the present instance to consider posaible isotopic exchange 00111plirations involving the reverse reactions. (16) J. Bigeleiaen. J . Chcm. Phus., 17, 675 (1949). (17) N. B. Sleter, Proc. ROY.Soc. (London), 184, 112 (1948).
D. R.STRANKS AND G. M. HAIWS
910
mechanisms of the reaction, however, the expectation is otherwise. For example, if the activated complex relevant to kz is conceived as a simple dissociated entity Co(NHa)JI20+++, HCO-a, would be equivalent to ~ H C O ~ This . would predict for k2*/k.2 a value of the order of magnitude of l/&, since f~ should not differ greatly from ~ C T C . Alternatively, it might be assumed that, on addition of acid, complete equilibrium is established prior to removal of the carbon dioxide fraction, as CO(NHa)rCOs+
+ 2H + HzO Jc +
Co(NHa)r(HzO)z+++
+ COz
The portion of gas removed a t fraction of reaction y would then have its isotopic composition defined by
Ka is the ratio of the equilibrium constants of the reaction above with reference to labeled and unlabeled reactants, respectively. It is readily seen to have the magnitude KS = ~ C O J ~ C T C ,the predicted values for which are given in Table V. It is clear from the form of equation (4) that these figures lead to a result contrary to observation. One can safely conclude that neither of the mechanisms last discussed can be operative in this reaction.
DISCUSSION J. P. HUNT(University of Chicago).-The experiments of Hunt, Rutenberg and Taube ( J . Am. Chem. SOC.,74, 268 (1952)) on aquation of Co(NHa)b(COa)+ show that here the bond. It is sugC-0 bond is broken rather than the (20-0
gested that such would also be the case for the aquation of CO(NH~)~(H~O)(HCO~)++. One would ask whether or not this affects the argument involving the isotope effect. If the exchange H2018 C O ( H ~ O ) ~ ~ + C O ( H ~ O ) ~ ( H ~ O ~HzO *) \vas rapid, one might conclude that the Co-0 bonds are broken relatively easily in this system. The experiments at present do not distinguish between rapid water exchange and catalysis by Co+2 aq. due to electron transfer. J. BIGELEISEN.-h the paper which 1 prcscnted on 1Vednesdny morning, I called attciition to some work by Dr.
+
+
VOl. 56
Lewis Friedman and myself ( J . Chern. Phys., 18, 1325, (1950)) on the isotope effects in the decomposition of ammonium nitrate. It was established in this work that there are some bonds in the transition state which are stronger than in the normal molecule. A similar situation exists in the thermal deammonation of phthalamide (Canadian Journal of Chemistry, 30, 443 (1952)). If a similar situation exists in the acid decomposition of the carbonatotetrammine cobaltic ion, there would be no conflict with the absence of an isotope effect measured by Stranks and Harris and the possibility that the C-0 bond breaks. G. M. HARRrs.-one can also get around the difficulty by accepting the alternative mechanism, in which kl is assumed to be rate-determining, though this would require re-interpretation of the exchange kinetics data (Harris and Stranks, Trans. Faraduy Soc., 48, 137 (1952)). PETERE. YANKWICH (University of Illinois).-Dr. John McNamara and I have measured the equilibrium constant for Cia and CI4 exchange and the corresponding relative rates in the reaction, involving the carbonato-bisethylenediamine cobalt(II1) ion, a system similar to that studied by Stranks and Harris. We found for K, as defined by their equation ( l ) , 0.990 8.015 for C14 and 0.991 f 0.013 for CIS; CIS exchanged 1.033 f 0.028 times as rapidly as 0 4 . A double check of K for Cia, in which a technique independent of blank corrections was employed, gave the. value 1.OOOq f 0.0004. One could predict from the “mssing” isoto e effects in this system that exchange would be more rapiB than in the carbonato-tetrammine system. Our kinetic data could be compared with that reported recently by Stranks and Harris, in the Transactions of the Faraday Society, only at two points; at both these points the exchange in the bis-ethylene system is somewhat the more rapid. J. BIGELEISEN(Brookhaven Nat. Lab.).-The assumptions made by Stranks and Harris in the calculation of the partition function ratios, ‘if” values, are such that it is improper to list their deviation from unity to four significant figures. I believe that three significant figures are all that are justified and in some instances just two. It is rather interesting that in the equilibrium experiments the C 1 concentrates in the CO,‘. This means that the bonding of the carbon to oxygen in carbon is appreciably weakend in the formation of the complex, possibly as a result of dcstruction of resonance in the carbonate ion. G. M. HARRIs.-This is the implication of our assumption that the complexed carbonate ion has propertie8 more similar to COC12 th?n to C03. In regard to the system studied by McNamara, it would appear that in the bis-ethylcnediamine complex the carbonate ligand is much less modificd than in the tetrammine. Such a difference is not inconceivable, in view of the known considerablc differences in thc chemistry of these two substanccs.
