Conductance of some CobaIt(II1) Complexes in Water at 25°CPart 1 .-Conductance of Salts of trans- and cis-dinitrobis(ethylenediamine)cobalt(rrI)BY ALAN D. PETHYBRIDGE" AND DAVID J. SPIERS?Department of Chemistry, The University, Whiteknights, Reading, Berks. RG6 2ADReceived 4th April, 1975Precise conductivity results are reported for aqueous solutions of seven salts of trans- and cis-[c~(en)~(NO~)~]+. They are analysed in terms of the full Pitts equation and the association con-stants for the salts, although small, are found to increase with increasing radii of the ions, beingparticularly large for ci~-[Co(en),(NO~)~]I. Contrary to popular belief, tran~-[Co(en)~(NO~)~]ClO~is also appreciably associated in dilute aqueous solution. Rate constants for the very slow aqua-tion of these cations are also reported.The existence of outer-sphere complexes has been widely postulated and variousmethods of determining the association constants of these complexes have beendeveloped.A comprehensive survey of earlier work has been published.' Moststudies of outer-sphere complex formation in aqueous solution have been made withlabile inner-sphere complexes or with concentrated solutions and the values obtainedfor the association constants of the outer-sphere complexes are often discrepant anddepend very much on the assumptions made in their calculation. Measurement ofthe electrolytic conductivity of dilute solutions of inert symmetrical complex saltsshould provide a more certain method of measuring the association constants of anyouter-sphere complexes formed, but most work in this direction has been withmultiply-charged salts, where association is likely to be greater for purely electro-static reasons, or with unsymmetrical salts where the resulting outer-sphere complexis still charged.The theoretical conductance equations of Pitts 2* and Fuoss canbe applied to symmetrical multiply-charged electrolytes, although not to unsym-metrical ones, but they are more reliable when used for 1 : 1 electrolytes. The outer-sphere association constants of 1 : 1 electrolytes are likely to be much smaller andboth very precise conductance measurements and a sophisticated analysis of the resultsare necessary to obtain meaningful values of these constants. Only the early workof Densham on the complex ion [ C r ( ~ y ) ~ F ~ l + had been published when we startedthis work, but during its progress the molar conductivities of some halogen substi-tuted acetato- and some oxalato-complexes of cobalt(II1) were reported and arediscussed later.In this work salts of trans- and cis-dinitrobis(ethylenediamine)cobalt(m), [Co(en),-(NO2),]+, were chosen for study because earlier reports in the literature had claimedthat these inner-sphere complexes are inert with respect to isomerization 8* andaquation.*-io By comparing the association constants for salts with a commonanion of the cis-cation with those for the trans-cation, we should be able to investigatethe importance of dipole moments in outer-sphere complex formation.In fact, boththe complex ions were found to aquate very slowly, the conductance increasing byapproximately 0.01 % per hour at 25°C so that the aquation of these ions can reallyt present address : International Nickel Ltd., European Research and Development Centre,Wiggin Street, Birmingham B16 OAJ.6A.D. PETHYBRIDGE A N D D. J. SPIERS 65only bz detected by precise conductimetry. The rate was sufficiently fast to necessitatethe replacement of the usual concentration-run technique of conductance measure-ment by the determination of the electrolytic conductivity of individually-preparedsolutions extrapolated back to the time of preparation. Hence, in addition to themolar conductivities, rate constants for the aquation of these cations are also reported.EXPERIMENTALAPPARATUSThe apparatus and thermostat used for these conductance measurements have beendescribed e1sewhere.l' The conductivity celIs were of the Marsh-Stokes type.12 The elec-trodes were not platinized as there is some evidence that blackened electrodes catalyse thedecomposition of ColI1 complexes.' All conductances were measured at six frequenciesand were extrapolated to infinite frequency.The cells, which had constants of 0.211 06,1.406 3 and 5.559 3 cm-l, were calibrated to 0.01 % with solutions prepared from vacuum-dried, twice-recrystallised AnalaR potassium chloride, using the semi-empirical equationsof Lind, Zwolenik and Fuoss l4 and Chiu and Fuoss l5 for the molar conductivity of potas-sium chloride.All solutions were prepared by weight, corrected to vacuum and the concentrationswere calculated using density measurements made on one of the more concentrated solutionsat 25°C using two density bottles calibrated with distilled water.A linear relationshipbetween density and molality was assumed in all cases for our solutions which are allsufficiently dilute for this to be reasonable.All measurements on these complexes were made by the single-point technique describedbelow, to allow for the time-dependence of the conductivity caused by the slow aquationof the cation. Cell constants obtained by this approach agreed with those found by themore usual concentration run method within experimental error. A special routine forwashing, rinsing and drying the cells and preparing the solvent was developed l6 so that alow and reproducible water conductivity could be obtained. This was always in the range2.2k0.3 x lo-' S cin-l, about four times the electrolytic conductivity of pure water, so thatthe maximum solvent correction to be deducted is less than 1 % for the most dilute solution.In a single-point run 0.05-2 g of the purified salt was weighed on an Oertling model 141balance which permits estimation, using extreme care, of the sixth decimal place. Theweighed salt was then placed in the cell reservoir, a weighed quantity of thermostattedconductivity water16 added by nitrogen pressure and the time noted.After stirring magneti-cally for a few minutes the reservoir was connected to the cell and immersed in the thermo-stat.After one hour conductance measurements were started and repeated at hourlyintervals for 6 h. For most salts the conductance increase was of the order of 0.01 % perhour and was linear in time. The conductance value at zero time was obtained by graphicalextrapolation and the rate constant for the aquation was calculated from the slope. Theconductances of solutions of the salt [Co(en),(NO,),][Co(edta)] increased by about 0.1 %per hour and yielded slightly curved conductance against time plots which were fitted to aquadratic expression by a least-squares technique.PREPARATION OF COMPOUNDSConductivity water and materials of the highest quality grade conveniently available(usually AnalaR) were used throughout. Full details of the preparation and purificationof these salts are given e1~ewhere.l~tr~ns-[Co(en)~(NO~)~]NO~.Several different samples of this salt T were prepared bythe method of Holtzclaw, $heetz and McCarthy.s The sample used for conductance measure-ments was purified by recrystallising twice from boiling conductivity water and was thendried in vacuo at 100°C. The solubility at room temperature is approximately0.06 mol dm-3.Analysis : found C 14.4, H 4.9, N 29.7 ; calculated C 14.4, H 4.8, N 29.4 %.trnns-[C~(en),(NO~)~]CIO~ was prepared by mixing equimolar quantities of almostsaturated solutions of T and sodium perchlorate. The yellow precipitate was filtered off,1-66 CONDUCTANCE OF CO"' COMPLEXESwashed with alcohol and ether, recrystallised twice from boiling conductivity water and wasfinally dried in vacuo at 100°C.The solubility at room temperature is estimated to be0.015 mol dm3. Analysis : foundC 12.9, H4.4, N 22.6 ; calculated C 13.0, H 4.35, N 22.7 %.trans-[C~(en)~(NO~)~]Cl was prepared by passing a near-saturated solution of T down anion-exchange column containing analytical grade Amberlite resin IRA400 in the chlorideform. The middle portion of the eluate was reduced to one tenth by evaporation and theproduct precipitated by the addition of a large excess of acetone. The product was purifiedby dissolving in the minimum quantity of water, 10 cm3, and filtering into 90 cm3 of acetone.Air-dried samples of the salt proved to be the monohydrate but the non-hygroscopic anhyd-rous salt was formed on heating in vacuo at 100°C and was used for the conductance measure-ments.The solubility at room temperature was estimated to be greater than 2.5 mol dm-3.Analysis for anhydrous salt : found C 15.7, H 5.4, N 27.5 ; calculated C 15.7, H 5.3, N27.4 %.trans-[C~(en)~(NO~),]Br was prepared and purified in a manner similar to that of thechloride. The solubility at room temperature was estimated as approximately 0.5 rnol dm3.Analysis : found C 13.6, H 4.7, N 23.9; calculated C 13.7, H 4.6, N 23.9 %.tram-[C~(en),(NO~)~]I was not prepared directly by ion exchange because of the dangerof oxidation to iodine during the lengthy ion-exchange process. Instead a saturated solutionof the chloride was prepared by ion exchange as described above and an equimolar quantityof saturated potassium iodide solution was added.The precipitate was filtered off andrecrystallised twice from boiling conductivity water, being protected by an atmosphere ofnitrogen at all stages. The product was finally dried in vacuo at 40°C. The solubility atroom temperature was estimated as approximately 0.12 mol d~n-~. Analysis : found C12.2, H 4.1, N 21.1 ; calculated C 12.1, H 4.05, N 21.1 %.tran~-[Co(en)~(NO~),]F was also prepared by ion exchange of a saturated solution of T.The purified product analysed as the trihydrate, but on heating in vacuo at 40°C the non-hygroscopic monohydrate was formed which was used for the conductance measurements.On heating to 100°C in vacuo the anhydrous salt was obtained but this proved to be extremelyhygroscopic.Analysis of monohydrate: found C 15.7, H 5.9, N 27.4, F 6.4; calculatedC 15.6, H 5.9, N27.3, F 6.2 %.from trans-[Co(en),CI,]Cl which was in turn prepared by the method of Bailar and Rollin-son.18 At all stages during the preparation and purification of the salts of ~is-[Co(en)~-(NOz)J+ care was taken to ensure that the temperature did not exceed 60°C since isomeriza-tiong to the trans form becomes significant above 65°C. The final product was purified byrecrystallising twice from water at 55"C, cooling the solution in an ice-salt bath before filter-ing. The product was dried in vacuo at 40°C. The solubility at room temperature isapproximately 0.035 mol dm3. Analysis : found C 14.3, H 4.9, N 29.5 ; calculated C 14.4,H 4.8, N 29.4 %.cik-[C~(en)~(NO~),]r.As with the trans-salt, this was prepared from the chloride (itselfobtained by ion exchange) by precipitation with saturated potassium iodide solution afterconcentration at 40°C on a rotary evaporator in the absence of oxygen. The salt was re-crystallised from water at 40°C and was dried in vacuo at 40°C. The solubility at room tem-perature is approximately 0.07 mol dm3. Analysis : found C 12.1 , H 4.1, N 21.2 ; calculatedby fractionalcrystallisation from a solution containing equimolar proportions of trans-[C~(en)~(NO~),]Cland potassium ethylmediaminetetra-acetatocobaltate(m) (K[Co(edta)J, whose preparationis described in Part 2). The product was recrystallised from boiling water and dried in vacuoat 100°C.The solubility at room temperature is approximately 0.2 mol dm-3. Analysis :found C 27.3, H 4.7, N 18.0; calculated C 27.2, H 4.6, N 18.1 %.cis-[C0(en)~(NO~)~]N0~, C, was prepared by the method of Harbulak and AlbinakgC 12.1, H 4.05, N 21.1 %.trans-[C~(en)~(NO~)~][Co(edta)] was prepared following SchwarzenbachRESULTSThe inolar conductivities of eight salts are given in table 1 together with the rateconstants for the aquation process. The aquation of tran~-[Co(en)~(NO~)~]F proA. D. PETHYBRIDGE AND D. J . SPIERS 67ceeded rapidly so neither conductivities nor rate constants are included here. Thisphenomenon will be discussed later. All conductance readings have been convertedto (absolute ohm)-l or siemens.104 cA106 k104 cA106 k104 cA106 k104 cA106 k104 cA106 k104 cA106 k104 cA106 k104 cA106 kTABLE 1 .-MOLAR CONDUCTIVITIES AND RATE CONSTANTS(c/mol dm-3, A/S an2 rnol-', klmin-')7.913 6102.932.36.667 0104.061.910.942101.382.85.486 297.4261.610.86291.8404.324.888100.132.114.48695.3343.017.50758.75220trunr-[Co(en)~(NO~)zlCl29.370 35.786 59.201 73.554100.59 100.13 98.561 97.7843.0 2.6 2.4 2.3trans-[Co(en)z(NOz)zIBr9.491 3 11.602 25.898 50.082103.525.218.268100.463.217.78295.7330.859.73087.7230.637.33 198.6913.417.51594.9891.732.21957.08418103.27 101.69 99.8431.7 2.0 1.2trms-[Co(en)z(NOd~JI28.654 57.610 72.35899.390 97.270 96.3542.0 1.6 1.2f runs-[Co(en)~(NOz)zlNOj26.193 46.20 1 80.47594.942 93.403 91.4731.1 0.7 0.5trons-[Co(en)~(NO2)2ClO472.076 101.45 125.1787.121 85.605 84.6311.1 1.9 1.8cis-[Co(en)Z(NOz)&41.865 53.22498.154 97.1752.2 1.5cis-[Co(en)z(NOz)dNO~35.803 58.907 79.71693.209 91.681 90.4541.2 1.3 5.4trms-ICo(en)z(NOz) zl[Co(edta)l67.821 110.145 154.7954.563 52.276 50.54618 19 1799.76396.6202.171.53498.6083.1102.9794.8221.691.86590.9471.1107.4889.1380.9123.9695.6441.993.21397.5292.7127.0893.7421.3137.9289.0452.0150.2187.4710.6114.6396.6232.5DISCUSSIONA" AND KA VALUESOur general method of fitting conductance data in terms of A", KA and d, thedistance of closest approach of free ions, is based on that of Duer, Robinson andBates,20 and has been described elsewhere.We used the full conductivity equationsof Pitts 2* and Fuoss and Hsia in our analysis rather than the various expandedversions of the two equations as we have found that for water, just as for hexamethyl-phosphorotriamide,' the actual parameters producing a statistical best fit of the datavary significantly according to which expanded version is chosen. Accordingly weconsider that any detailed interpretation of precise conductance data should be interms of the full equations even if this means an added complexity of computation.All calculations were performed on the CDC7600 computer at U.L.C.C. using theprogram SEEKER written in FORTRANIV. Throughout this work we reportparameter values obtained using the Pitts (P) equation.In general the Fuoss-Hsia(FH) equation gives a comparable fit but with slight systematic changes in the para-meters which will be noted68 CONDUCTANCE OF COMPLEXESThe simplest method of fitting the data is to assume that the salts are completelydissociated in aqueous solution, in which case the association distance d and contactdistance a are identical. Values of A" and a which fit the data for each salt are givenin table 2. The FH equation gives a similar fit but at a values larger by approximately0.3 A. The values of the deviation function cA( %), defined asOA( %) = {c[loo(l -A~~lc/Aobs>12/(n-2>): (1)TABLE 2.-ANALYSIS BY FULL PITTS EQUATION. BEST-FIT PARAMETERS ; ELECTROLYTES TREATEDAS FULLY DISSOCIATEDsalt 4 Am/S cm2 mol-1 a-4 %Itran~-[Co(en)~(NO~) 2]C1 0.52 105.33 0.04trans-[ C0(en)~(N0~)~]Br 0.28 106.24 0.05tran~-[Co(en)~(NO~)~]I 0.11 104.24 0.05trans-[ Co(en) (NO2)2]C104 0.10 94.61 0.07trans-[Co(en)z (NO2)2]N03 0.26 99.38 0.04ci~-[Co(en)~(NO~)~]N0~ 0.12 98.56 0.09where n is the number of experimental points, are quite reasonable but the values ofa are all absurdly small for such large ions.In fact ~is-[Co(en),(NO,)~]I requiresa negative value of a (which cannot be determined by SEEKER) for the best statisticalfit, so it is omitted from table 2. Consequently, it is logical to introduce an additionalvariable, the association constant KA, and the best-fit parameters Am, KA and d areshown in table 3 for the P equation.Not surprisingly, smaller values of a,( %) arenow obtained and these results, which are obtained by the single-point technique,always show a unique position of best fit with each equation. In general the fit withthe FH equation is equally good but occurs at values of d, A" and KA which are largerby approximately 0.9 A, 0.010 S cm2 mol-1 and 0.20 dm3 mol-1 respectively.TABLE 3 .-ANALYSIS BY FULL PITTS EQUATION. BEST-FIT PARAMETERS ; ELECTROLYTES TREATEDAS ASSOCIATEDsalt d / A Awls em2 mol-1 KA/drn3 mol-1 aKA 0.4 %Itrans-[Co (en) (NO 2 ) ,]C1trans-[Co(en),(N02),]Brtrans-[Co(en),(NO,),]Itrans-[ Co (en) (NO ,) 2]N0tran~-[Co(en)~(NO~),]ClO~~is-[Co(en)~(NO,)~]I~is-[Co(en),(NO,)~]N03tran~-[Co(en)~(NO~),][Co(edta)]3.023.053.053.123.163.103.064.28105.36106.27104.3399.4394.69105.5098.6762.361.591.952.522.082.606.322.558.050.130.410.130.160.120.180.230.930.030.040.030.020.050.060.060.21However, we have shown that when comparing results of different workers forthe same system, or indeed different runs by the same worker on the same system,it is necessary to compare Am and KA values at a common d to obtain results that aremutually consistent.The actual value of d is not critical within the range 2-8 A buta convenient value is the Bjerrum critical distance q (= z2e2/2DkT) which is 3.58 Ain water at 25°C. Table 4 shows the Am, KA and cA( %) values obtained for the fitat d = q using the P equation.The FH equation gives an equally good fit underthese conditions with A" values about 0.005 S cm2 mol-1 larger and KA values about0.20 dm3 mol-l smaller than those in table 4. We accept that for these salts q is lessthan the sum of the crystallographic radii. However, all the points of statisticaA . D . PETHYBRIDGE AND D. J . SPIERS 69best-fit obtained in table 3 occur at even smaller values of d, and we believe that dcannot be interpreted as a physically meaningful distance, but that it is an artefactof the conductivity equations primarily determined by the charge on the ions andthe dielectric constant of the solvent. Similar calculations at d = 2q simply raise allthe KA values by about 1 .O dm3 mol-l. Because of the arbitrary nature of the valuechosen for din table 4, the exact values of the association constants are of no absolutephysical significance and should only be quoted in conjunction with the value of dselected.On the other hand, association constants calculated for a fixed value ofd should be capable of a valid comparison, since they are free of the random elementthat is inherent in the best-fitting procedure when A*, KA and d are all adjusted.Although the values of crA( %) are slightly larger than those for our conductance dataobtained by the concentration-run technique (see Part 2), they probably represent amore realistic estimate of the errors involved, as each measurement is quite independentof the others in all but the purity of the solute.TABLE 4.-ANALYSIS BY FULL PITTS EQUATION : d = 4 (3.58 A) : ELECTROLYTES TREATED ASASSOCIATEDsalt Am/S cm2 mol-1 &/dm3 mol-1 OKA GA(%)trans-[Co(en),(NO,),]Cl 105.36 1.82 0.13 0.03trans-[Co (en) (NO ,) ,]Br 106.27 2.16 0.41 0.04rrans-[C~(en)~(NO~)~]I 104.33 2.73 0.13 0.03trans-[C~(en)~(NO~)~]NO~ 99.43 2.27 0.16 0.02tuans-[C~(en)~(NO~)~]ClO~ 94.69 2.78 0.12 0.05cis-[Co (en), (NO 2 ) ,]I 105.50 6.52 0.18 0.06cis-[C~(en),(NO~)~lNO 98.67 2.77 0.23 0.06The results in table 4 show that the association constants of the outer-spherecomplexes increase in the order Cl < Br < NO3 < I M C104.The high associationconstant for outer-sphere complex formation for trarzs-[C~(en)~(NO,),]ClO~ is ofparticular interest and is rather disturbing in view of the fact that perchlorates areoften added as swamping electrolytes in equilibrium and kinetic studies because oftheir low tendency to form inner-sphere complexes (and nitrates are usually the secondchoice).This suggests that in such a medium many of the ions of interest are notfree but are in fact paired with the perchlorate ions of the medium. Burnett 2 2 hasshown from kinetic studies that such outer-sphere complex formation with perchlorateis significant and must be allowed for in studies on the rates of exchange of, forexample,[CO(NH~),(H~~)]~+ + C1- + [Co(NH3)=,C1I2+ + H20.We have shown 21 that there is a strong correlation between the value of KA at anyfixed d value such as the Bjerrum critical distance and the sum of the crystallographicradii for a wide range of salts. The results for most of the salts studied in this workagree with this pattern, shown in fig.1, where we have also plotted results for thosesalts where two independent sets of precise results are in good agreement. With oneexception, our complexes all lie close to the band containing the KA values of the othersalts. The exception is cis-[C~(en),(NO,)~]I where association is likely to be enhancedboth by the dipole moment of the cation (the cis-nitrate is also slightly more associatedthan the trans-nitrate) and by the polarizability of the iodide. The association constantof Bu,NI is also much larger 21 than can be predicted on account of its size from fig. 1.The ionic radii used in plotting fig. 1 were taken from standard corn pi la ti on^.^^^ 24aThe value of 4.3 A for the radius of trans- or ci~-[Co(en),(NO,)~]+ was found from asimple geometrical calculation using bond lengths in similar complexe~.~ We hav70 CONDUCTANCE OF CO"' COMPLEXEScommented 21 on the fact that the trend in KA values with increasing Y is the reverse ofthat predicted by the Bjerrum theory of association.The consequences of the trendobserved in fig. 1 will be discussed further in a later paper.From the A" values shown in table 2 for the various salts it is possible to show,using published d2 that the limiting molar conductivities of traits- and cis-[Co(en),(NO,),]+ are 27.7 * 0.4 and 27.4f 0.3 S cm2 mol-1 respectively. The largescatter in these values is undoubtedly due to varying amounts of hydrolysisproducts produced during the preparation and purification of the compounds andnot entirely removed during recrystallisation procedures.This was particularly true++00 2 4 6 8&lystlAFIG. 1 .-Plot of KA values at d = q using the P equation against the sum of the crystallographic radiiof the ions. x alkali metal halides, 0 alkali metal oxosalts, + tetra-alkylammonium halides,A salts studied in this work.of tran~-[Co(en),(NO,)~]Cl. The values of KA obtained are not seriously affected bythese small changes in A". The point for tran~-[Co(en),(NO,)~][Co(edta)] is notshown in fig. 1 (its coordinates are 7.68 dm3 mol-1 and 8.6tf) because its A@' value,see table 3, is approximately 9 conductance units higher than would be expected fromthe individual ionic conductances.Therefore the association constant of the pure1 : 1 double complex salt would be rather different but probably still large. The trans-[ Co(en),(NO,),JF, which aquated much more rapidly, also had A" approximately9 conductance units too high. However, the scatter on the individual results wassuch that a meaningful value of KA could not be obtained, the value with d = q being4.8k4.5 dm3 mol-l.We have also reanalysed the precise conductivity data for similar salts in theliterature ' 9 and the values of A" and KA which fit at d = q for the Pitts equationare given in table 5. These parameters show general agreement with the trends of K,with r reported in table 4 and fig. 1, especially the high association constant for[Co(ox)(en),]I.The results for the acetatocobalt(rI1) complexes ' are not includeA . D. PETHYBRIDGE AND D . J . SPIERS 71because these salts were all aquated relatively quickly and only 3 or 4 experimentalpoints are reported for each salt.RATES OF AQUATIONFor most of these complex salts the aquation of the cation leads to an increase inthe conductance of the order of 0.01 % per hour and could only be observed bysuch a sensitive technique. Even this change could be masked by a change in thetemperature of only 0.005"C so the values obtained for the rate constants for theaquation are bound to be imprecise. However, the values for the rate constantsreported in table 1 were obtained from each experimental run and show quite goodagreement with no obvious trend with increasing concentration or dependence uponthe anion.These values of the rate constant, k, were obtained from the time-dependence of the electrolytic conductivity, dK/dt, from the equationc dt 100- 83.65 x C'k = !!!fTABLE 5.-ANALYSIS OF LITERATURE RESULTS BY THE FULL PITTS EQUATION, d = 4salt n m / S cm2 mol-1 KA/dm3 mol-1 UKA UA( "/u,[Cr(PY),F2lC104 a 86.71 4.88 0.58 0.02[Co(ox)(en),]Cl 98.52 -0.19 0.27 0.02[Co(ox)(en)2lI 98.69 7.33 0.23 0.03[~o(ox)(en)~l~10~ 90.26 0.95 0.18 0.01a ref. (5) ; b ref. (6).[Cr(PY)4F21SCNba 84.97 6.29 3.34 0.04[Co(ox)(en)zlBf; 100.31 2.28 0.08 0.01in which the Onsager limiting conductance law has been used and Am(NO;) is takenas 72.1 S cm2 r n ~ l - ' . ~ ~The average first-order rate constant for the aquation of trans- or cis-[Co(en),-(NO,),]+ is 2 x min-l in the presence of C1-, Br-, I-, NO: and C102.Thereappears to be no previous determination of k for these ions which had previouslybeen reported as inert to aquation.*"*The value of k for trans-[Co(en),(NO2),][Co(edta)] is some 10 times greater thanthat obtained for the above salts. The pH of a nitrogen-saturated 0.015 mol dm-3solution was 6.6 and the pH of a similar solution of K[Co(edta)J was 6.2 suggestingthat the anion is responsible for the low pH and that the free hydrogen ions appearto catalyse the aquation of the trans- complex ion. This is presumably due to theequilibriafor which Schwarzenbach l9 has quoted some evidence.The conductance-time curves of tran~-[Co(en),(NO~>~]F were much steeper andmarkedly curved. From the initial slope of these curves, obtained by a curve-fittingprocedure, a first-order rate constant some 2 x lo6 times larger than that for the othertrans- complexes was obtained.The pH of a 0.01 mol dm-3 solution was 5.1 threehours after preparation, but increased to 5.9 when the solution was left overnight inthe cell. This may be due to the reactionSiO,(s) + 4H+ + 6F- + SiFz-+2H20which would also explain why the conductance of the solution, although initially[Co(edta)J- + H20 + [Co(edta)(H,O)]- + [Co(edta)(OH)I2- + H72 C 0 N D U C T A N C E 0 F CO"' C 0 M P L EXESincreasing rapidly, was found to be decreasing by approximately 0.1 % per hour thefollowing day. The free hydrogen ions were probably due to the presence of a smallamount of HF; which was selectively absorbed from the unbuffered fluoride solutionused to convert the IRA-400 resin to the fluoride form, well-known to be an un-favourable exchange, and which was then eluted during the preparation cycle.It is difficult to estimate from these results the quantitative effect of hydrogen ionsupon the aquation, but the evidence is that they appear in the rate equation to apower higher than one.One of us (D.J. S.) is grateful to the S.R.C. and the University of Reading forfinancial support.Note added in proof. We have reanalysed our data using the latest Fuoss equation (J. Phys. Chem.,1975,79, 525). The best-fit values of d occur at around 3.9 A, rather higher than those recorded intable 1, but lower than those reported by Fuoss for alkali metal halides.The new equation withd = (I yields ha values 0.02 S cm2 mol-l larger than those from the P equation in table 4 but KAvalues 0.7 dm3 mol-' lower.M. T. Beck, Co-ordination Chem. Rev., 1968,3, 91.E. Pitts, Proc. Roy. SOC. A, 1953, 217, 43.E. Pitts, B. E. Tabor and J. Daly, Trans. Faraday Soc., 1969, 65, 849.R. M. Fuoss and K.-L. Hsia, Proc. Nat. Acad. Sci. U.S.A., 1967, 57, 1550; 1968, 58, 1818.A. B. Densham, Trans. Faraday SOC., 1937, 33, 1513.M. Yokoi and K. Kuroda, Bull. Chem. SOC. Japan, 1971,44,3293. ' E. Kubota, Nippon Kagaku Zasshi, 1971,92, 1112. * H. F. Holtzclaw, Jr., D. P. Sheetz and B. D. McCarthy, Inorg. Synth., 1953, 4, 176.E. P. Harbulak and M. J. Albinak, Inorg. Synth., 1966, 8, 196.chap. 3.l°F. Basolo and R. Pearson, Mechanisms of Inorganic Reactions (Wiley, New York, 1958),l1 E. M. Hanna, A. D. Pethybridge, J. E. Prue and D. J. Spiers, J. Solution Chem., 1974, 3, 563.l2 K. N. Marsh and R. H. Stokes, Austral. J. Chem., 1964, 17,740.l3 W. A. Millen and D. W. Watts, J. Amer. Chem. SOC., 1967,249,6858.l4 J. E. Lind, J. J. Zwolenik and R. M. Fuoss, J. Amer. Chem. SOC., 1959, 81, 1559.l 5 Y. C. Chiu and R. M. FUOSS, J. Phys. Chem., 1968,72,4123.l6 A. D. Pethybridge and D. J. Spiers, Electroanalyt. Chem., in press.l7 D. J. Spiers, Ph.D. Thesis (Reading University, 1974).Is J. C. Bailar, Jr., and C. L. Rollinson, Inorg. Synth., 1946, 2, 222.l 9 G. Schwarzenbach, Helv. Chim. Acta, 1949, 32, 839.2o W. C. Duer, R. A. Robinson and R. G. Bates, J.C.S. Furaday I, 1972, 68, 716.21 A. D. Pethybridge and D. J. Spiers, Chem. Comm., 1974,423.22 M. G. Burnett, J. Chern. SOC. A , 1970, 2480, 2486,2490.23 Interatomic Distances (Chemical Society Special Publications, London, No. 11, 1958 ; No. 18,24R. A. Robinson and R. H. Stokes, Electrolyte Solutions (Butterworth, London, 2nd edn.25 P. J. Protzenko, 0. N. Shokina and N. P. Chekhunova, Russ. J. Phys. Chem., 1964, 38, 1013.1965).rev., 1959), (a) pp. 125 and 461 ; (b) p. 463.(PAPER 5/641