J . Chem. SOC., Faraday Trans. I , 1982, 78, 371-381 Equimolar Mixtures of Divalent Transition Metal Perchlorates as Constant Ionic Media in Studies on Complex Formation in Non-aqueous Donor Solvents Chloro Complex Formation in Dimethyl Sulphoxide BY WLODZIMIERZ LIBUS,* ROMAN PASTEWSKI A N D TERESA SADOWSKA Department of Physical Chemistry of the Institute of Inorganic Chemistry and Technology, Technical University of Gdansk, 80-952 Gdansk, Poland Received 28th November, 1980 Equimolar mixtures of divalent transition metal perchlorates in dimethyl sulphoxide (DMSO) have been used as constant ionic media for studying the formation of complexes of metal cations with weakly coordinating anions. Isopiestic experiments indicate the same solvent activities and accordingly the same activity coefficients of the solutes in such mixtures.In the present study Co(ClO,), is used as one component of the mixture and optical absorptions due to CoC1,DMSO-, which is formed on addition of Et,NCl, are used to study chloride complexing with several transition metals. Determination of stability constants of ionic complexes is either performed under conditions of constant activity coefficients or variations in the coefficients are taken into account in the calculations. The latter approach, in which dilute ionic solutions are used,'? has the advantage of providing thermodynamic characteristics of complex formation in the absence of a supporting electrolyte and is better for discussing the solvent effect on complex formation.3 However, determination of stability constants in solutions of variable low ionic strength in non-aqueous solvents suffers from our present inadequate knowledge of the concentration dependence of the activity coefficients.On the other hand, the use of a supporting electrolyte such as alkali metal, ammonium or tetra-alkylammonium perchlorates to maintain a constant ionic strength cannot ensure that the activity coefficients remain constant when using weakly complexing ligands, since a wide concentration of the latter is needed to provide sufficient data.4 An alternative possibility for maintaining constant activity coefficients in studies of metal complexes with anions is to use equimolar mixtures of the respective perchlorates or some other metal salts involving non-coordinating anions as the reaction media.In a series of papers Z. LibuS et al. have shown that aqueous equimolar divalent transition metal perchlorates and magnesium perchlorate behave as effective constant ionic This was ascribed to the fact that the respective metal cations all occur as the hexa-aquo complexes, M(OH,)i+, probably preserving constant second layers of hydrogen-bonded water molecules. This suggested that equimolar solutions of the divalent transition metal perchlorates or tetrafluoroborates could be used as constant ionic media in non-aqueous donor solvents like acetonitrile or dimethyl sulphoxide (DMSO). Previously, we have found that the respective solutions contain the metal cations exclusively in the form of octahedral hexasolvo complexes whose mobilities and associating abilities with the non-coordinating anions 377378 COMPLEX FORMATION I N DMSO vary within relatively narrow limits as the nature of the central metal atom changes.12-15 Some results indicated constant activity coefficients of the metal perchlorates in their equimolar mixtures as we11.16 In the present work we have examined the chloro complex formation of divalent transition metal cations with the chloride anion in DMSO solution using equimolar mixtures of cobalt(r1) perchlorate with the other metal perchlorates as effectively constant ionic media. Light absorption due to the CoC1,DMSO- complex anion17 has served as a measured solution property depending on the free chloride anion concentration.EXPERIMENTAL The DMSO-solvated metal perchlorates, Mn(C10,), .6DMSO, Co(ClO,), * 6DMS0, Ni(CIO,), .6DMSO, Cu(ClO,), -4DMSO and Zn(ClO,), * 6DMS0, were obtained from the respective hydrated metal perchlorates by dissolving the latter in anhydrous DMSO, followed by boiling the solutions for several minutes and evaporating the solvent under reduced pressure.18 The crystals were filtered off and further purified by repeated crystallizations from DMSO.Reagent grade tetraethylammonium chloride was recrystallized twice from anhydrous acetonitrile and dried in uclcuo at 60 "C. Dimethyl sulphoxide, reagent grade, was dried using calcium hydride and then distilled under reduced pressure. The product was further purified by repeated fractional crystallization carried out under anhydrous conditions. The specific conductivity of the final material was 3 x The stock solutions of the metal perchlorates in DMSO were analysed for the respective metals by standard EDTA titrations.Prior to titration, weighed portions of the stock solutions were dissolved in water. In addition, the stock solutions of Co(ClO,), and Cu(ClO,), were analysed for the respective metals by electrodeposition and the stock solution of Ni(CIO,), was analysed for nickel gravimetrically using dimethylglyoxime. At least five determinations were carried out by each method and the results obtained by different methods agreed to within 0.25 % for cobalt and 0.15% for nickel. Solutions used in the final measurements were obtained by weighing from the respective stock solutions and the solvent. Their concentrations were calculated using the densities determined independently. Preparation of the solutions and other manipulations were carried out in a dry-box.Absorption spectra were measured by means of a VSU 2-P Zeiss spectrophotometer equipped with a thermostated cell compartment. The isopiestic experiments were carried out using the apparatus and techniques described in ref. (19). W1 cm-l at 25 OC. RESULTS ISOPIESTIC MOLALITIES Taking into account the similarity of the solution forms of the divalent metal perchloratesin DMS0,13j l5 as already mentioned, their osmotic and activity coefficients might be expected to display essentially the same concentration dependences. In order to check this expectation, we have attempted t o determine the isopiestic molalities of the metal perchlorates in DMSO solution.It has been found, however, that the isopiestic equilibrium in these systems at 25 "C is only attained after very long times, 1 to 2 months, or even longer for the more dilute solutions. As a result, only a few experiments have been completed and the results are listed in table 1 . Inspection of the data shows that the isopiestic molalities are the same, to a good approximation, for Co(ClO,),, Ni(ClO,), and Zn(C10,),, while they are slightly higher for Cu(CIO,),. This pattern is very similar to that observed for the aqueous perchlorates,lg indicating that in DMSO solution the pattern of the osmotic and activity coefficients is also the same, viz. the coefficients for Cu(ClO,), are slightly lower than those for the otherw. LIBUS, R.PASTEWSKI A N D T. SADOWSKA TABLE 1 .-ISOPIESTIC MOLALITIES (mol kg-l) OF THE DIVALENT TRANSITION METAL PERCHLORATES IN DMSO, AT 25 "c 3 79 Co(ClO,), Ni(C10,), Cu(ClO,), Zn(C10,), 0.1970 0.1964 0.2022 0.1976 0.2017 0.2007 0.2034 0.2003 0.2865 0.2824 -~ ~ _ _ _ _ _ _ _ ~ _ ~ _ _ _ _ _ _ _ - - ~~~~ ~ ~ ~- . ~ _ _ _ divalent metal perchlorates belonging to this group. However, apart from these second-order differences, the present results support the expectation that equimolar mixtures of the divalent transition metal perchlorates in DMSO solution should behave as effectively constant ionic media. SPECTRAL EFFECTS U N D E R L Y I N G THE APPLIED METHOD Fig. 1 shows the visible absorption spectra of cobalt(I1) in a series of solutions containing Co(CIO,), at a constant concentration of 0.175 mol dm-, and Et,NCl at concentrations varying from 0 (curve 1) to 0.0253 mol dm-, (curve 7).The absorption band observed at 535 nm for the pure Co(ClO,), solution (curve 1) is due to the 4TT,,(F) 7 lT,JP) transition within the Co(DMSO):+ octahedral complex.lR As is seen, addition of Et,NCl to the solution containing Co(ClO,), in a large excess results in the appearance of a new band with maximum at ca. 682 nm. As the concentration of Et,NCl increases further, the more subtle features of the new band become clearer and the band shape characteristic of the CoC1,DMSO- tetrahedral complex finally develops. For comparison, the spectrum of the latter complex, known from the earlier study,17 has also been indicated in fig. 1 (curve 8, right-hand scale).It may be seen that the intermediate spectra are linear combinations of those of curves 1 and 8. The conclusion might be drawn from these observations that CoC1,DMSO- is the only chloro complex formed. However, formation of an octahedral complex, e.g. CoCl(DMSO)', at a concentration comparable with that of CoC1,DMSO-, having relatively low absorptivity may well remain without a noticeable effect on the measured spectra under conditions of excess cobalt(r1) perchlorate. An indirect demonstration that an octahedral chloro complex of cobalt(r1) in fact is formed is provided by the observation that substituting Co(ClO,), for Ni(ClO,), in the equimolar mixtures containing Et,NCl at a constant concentration results in a marked increase in intensity of the 'blue' band due to the CoC1,DMSO- complex, as illustrated by curve 2 in fig.1. The effect indicates increasing concentration of the 'free' chloride anion which, in turn, may only result from the fact that a part of the total chloride content in the solution was initially bound to cobalt(I1) in the form of octahedral complexes, while nickel(r1) forms weaker chloro complexes. It appears that substitution of Co(ClO,), for Mn(ClO,), has an opposite effect on the intensity of the band due to the CoC1,DMSO- complex (curve 2' in fig. 1). This we take as an indication that the MnCl+ complex has a higher stability than CoCl+. Still more effective in depressing the intensity of the band under consideration appear to be Cu(ClO,), and Zn(C10,),, indicating the high stability of the chloro complexes of the respective cations in DMSO solution.3 80 COMPLEX FORMATION I N DMSO 500 400 309 200 100 0 400 500 600 700 A / nm FIG.1 .--Visible absorption spectrum of cobalt(i1) in the DMSO solutions of: 1, (Co(CIO,), (0.175 rnol drnP3); 2-7, Co(CIO,), (1.175 mol dm-")+ Et,NCl (concentration from 0.00596. curve 2. to 0.0253 mol drnP3, curve 7); 2', Co(CIO,), (0.101 rnol dm-:l) + Mn(CIO,), (0.077 rnol dm-:$) + Et,NCl (0.0062 mol drn-"); 2", Co(CIO,), (0. I46 mol dm-:3)+Ni(CI0,), (0.029 rnol dm-:')+ Et,NCI (0.0062 rnol dm-3); at 25 "C. The broken line 8 represents the spectrum of the CoCl; complex, right-hand scale. O U T L I N E O F T H E METHOD The applied method consists of using a series of solutions of two metal perchlorates, M(ClO,), and M'(ClO,),, of a constant total concentration ct to which a small amount of the complexing anion X is added, so that c, << ct.The equilibrium concentration [XI of the free anion is then determined as a function of the composition of the mixture either directly or from the measured equilibrium concentration of one of the complexes formed. Evaluation of the results makes use of the relation n n C nPTL [MI [XIn + C n& [M'] [XI" +[XI - c, = 0 (1) n=o 71 -0 arising from the material balance for the anion, where are the 'medium' stability constants of the complexes MX, and M'X,, respectively, and the bracketed symbols denote equilibrium concentrations of the respective species. These formulae, as well as those used to denote the chloro complexes studied, neglect possible coordination of the solvent molecules.This is in accord with the known factw. L I B U S , R . P A S T E W S K I A N D T. SADOWSKA 38 1 that the stability constants determined under conditions of a constant ionic medium may, in principle, be sums of the 'true' stability constants of a number of different solution species, all corresponding to the non-committal formula MX,."> 23 Further argument is needed to ascribe the derived stability constants to particular solution species, such as MXL: or MX,L- (L denotes the solvent molecule), or to resolve them into the stability constants of the single species. and pi', the stability constants at zero ionic strength, are related to the above stability constants by where Y , and Y/, are quotients of the activity coefficients for the respective complex-forming equilibria.The assumption is made that under the conditions stated all the activity coefficients remain constant while the relative contents of the two metal perchlorates are varied. Provided that the free anion concentration has been determined for at least 2n solutions of the above type, the best solution of a set of eqn (1) in the stability constants may be found using a computer-based program. In the zeroth approximation initially estimated values of the stability constants must be used and the free cation concentrations [MI and [M'] may be approximated by the total concentrations, provided that association with the perchlorate anion is negligible. The latter seems to be a sufficiently good approximation for the divalent transition metal perchlorates in DMSO solution, as may be seen from recently determined association ~0nstants.l~ EVALUATION OF T H E SPECTROPHOTOMETRIC RESULTS In the present study use was made of the readily determinable equilibrium concentration of the CoCI,DMSO- complex, which served as an indicator of the free chloride anion concentration.Accordingly, equimolar mixtures of Co(ClO,), with the other divalent transition metal perchlorates, containing a small amount of Et,NCl, were studied, the concentration of CoCl; (apart from its solvation) being found as [CoCl,] = c(F- E,,)/E,, where Fdenotes the measured mean molar absorption coefficient of cobalt(I1) at 680 nm and E , = 0.45 and c4 = 524 are the known molar absorption coefficients of Co2+ and CoCl;, respectively, at this wavelength.Taking into account the possible formation of three consecutive chloro complexes of either metal cation, the computer-based iterative procedure of finding the best values of the stability constants from the spectrophotometric data consisted in minimizing the function while the equilibrium concentrations in the ith solution at each step were calculated as and [COCI,], 5 ci - [CoCl,], 1 +pp[Cl], +p:"[cl]; iC11i = (-1 [CO], = ci = 1 +pp[cl], +pp[cl]; +p,"[Cl]; * The necessary values of pfo, the stability constant of the CoC1; complex in the given medium, were found in separate experiments, as described below. The program written for an ODRA 1305 computer permitted elimination of any of the five382 COMPLEX FORMATION I N DMSO complex-forming reactions represented in eqn (1) and (4), apart from that for the CoCl; complex, as well as admission at thejth iteration of any of these complex formations not so far considered, thus providing an insight into the convergence process.Using the above approach, convergent values of the stability constants of the MC1+-type complexes were obtained in few iterations for the equimolar mixtures involving either Mn(C10,), or Ni(C10,), in addition to Co(ClO,), on the assumption that pFo = Dz* = pp = 0. On the other hand, negative values of the constants, lacking physical meaning, were obtained for these systems if the respective terms were admitted into the iterative calculations. These results we take as a confirmation of the expected exclusive formation of the monochloro complexes of the respective metal cations under conditions of excess metal perchlorates.On the other hand, for the systems involving copper(I1) and zinc(II), convergence could only be obtained either by using very well estimated initial values of the stability constants or by performing the calculations in steps. In the first step, formation of only the predominant complex of the metal other than cobalt(r1) was taken into account, viz. of CuCl+ for copper(I1) and of ZnC1; for zinc(II), in addition to the two cobalt(I1) complexes, CoC1+ and CoCI;. After a number of iterations providing a preliminary convergence of the stability constants, formation of the other chloro complexes was admitted. For copper(II), admission ofeither CuClt or CuCl; resulted in equally good approximations.Of the respective results those corresponding to the formation of CuCli in addition to CuCl+ have been assumed as real, in accordance with the earlier work in which formation of these latter complexes in the dilute solutions of CuC1, in DMSO was inferred from the conductometric data and the visible absorption spectra.,O This does not mean excluding formation of the higher chloro complexes of copper(I1) under conditions of excess chloride anions or in the more concentrated solutions of CuCl, in DMSO. For the system involving Zn(C10,),, self-consistent results have been obtained on the assumption that the ZnC1; neutral complex is formed in addition to ZnCl;, while no significant improvement of the approximations was obtained upon admission of the formation of ZnC1+.Values of /?Fo in the particular media used were needed for the above calculations. One possible way was to calculate them from the previously determined stability constant at zero ionic strength, = (8.4+ 1.5) x 108.17 Magnell and Reynolds have reported (4.21 k0.58) x los assuming the absence of complexes other than CoCl;.l However, in view of the obvious uncertainty involved in selecting a proper equation for the activity coefficients, an attempt has been made to derive together with /?Fo, from the spectrophotometric data for the DMSO solutions of Co(ClO,), containing Et,NCl at smaller and variable concentrations. However, the concentrations of Et,NCl had to cover a fairly wide range in order to obtained significant changes in the mean molar absorption coefficient of cobalt(1r) within the CoCl; band.Accordingly, the Debye-Hiickelequation, log y = - z+z-A %/I/( 1 + Ba'dI), involving the ion-size parameter estimated as 8.0 A,17 was used to make allowance for the possible small variations in the activity coefficients. Values of 1.11 1 and 0.4262 were assumed for the constants A and B, respectively, valid for DMSO at 25 OC. On the assumption that only the CoCl+ and CoCl; complexes are formed under conditions of excess Co(ClO,),, the stability constants pFo and p f o were then derived from the material balances for both ColI and C1- using a separate convergence program written for an ODRA 1305 computer. In this case, the sums of the deviations squared were minimized with respect to the parameters [Co], [CoCl], [Cl], p: and pf, the latter two being constants for all the solutions in which the equilibrium concentration of CoCl; was determined experimentally.The derived values of the constants are listed inw. LIBUS, R . PASTEWSKI A N D T. SADOWSKA 383 TABLE 2.---STABILITY CONSTANTS OF THE MClf-, MC1;- AND MCI, -TYPE COMPLEXES IN EQUIMOLAR MIXTURES OF Co(C10,)2 WITH THE OTHER METAL PERCHLORATES OF TOTAL CONCENTRATION c,(mol dmP3) in DMSO, AT 25 OC Values at ct = 0 are the calculated 'thermodynamic' stability constants (see text). Co(C104)2 + Et4NClu 0.0 620f40 0.086 74.0 k 4.0 0.134 62.0 k 4.0 0.176 56.0 f 4.0 0.0 0.027 1 13.1 k 7.7 0.079 68.9 f 1.4 0.127 57.3k 1.5 0.175 53.7 f 0.9 0.225 48.0 k 0.8 0.285 45.5 & 1.5 - 0.0 0.075 65.8 k 1.5 0.125 56.7 k 3.2 0.175 53.1 2.4 - 0.0 - __ - - - 1720 348.7 f 12.4 201.1 k2.3 175.5k4.5 159.0 2 1.9 146.3 f 1.4 130.4 & 1.8 256 30.4 k 0.7 24.5 f 1 .O 23.5k 1.1 1.02 x lo5 0.1 10 65.0k0.3 (1.25k0.05)~ lo4 0.152 58.0k0.2 (0.81 k0.02) x lo4 0.195 48.0k0.1 (0.89f0.02) x lo4 0.0 0.101 66.0k0.5 - 0.151 58.0k0.5 - 0.185 50.0 k 1 .O - ~ - - (6.3 f0.8) x lo8 - (2.63 k0.35) x lo7 - ( 1 .9 2 k 0 . 3 ) ~ lo7 - (1.68 f 0.3) x lo7 - - (7.82 f 2.2) x 105 - (1.43f0.37)~ lo6 - (2.29k0.29) x lo6 - - __ (1.1+0.14)x lo7 ( 2 . 1 f 0 . 0 6 ) ~ 10" (2.9 k 0.2) x lo7 (2.1 & 0.07) x lo1' (1.8k0.8) x lo7 (3.8k0.3) x 10" a Concentration up to 15 that of CO(CIO,)~. table 2 along with the 'medium' stability constants determined in the equimolar mixtures of Co(CIO,), with the other divalent metal perchlorates.We note that /l: = (6.3 0.8) x lox is intermediate between the two values cited above, at the same time being lower than the value derived from the study of the dilute solutions of CoCl,. The agreement is not particularly important for the present purposes, since the value of a," derived here should only be considered as an empirical constant providing the best approximation of the variation in the 'medium' stability constant /lFo within the range of moderate concentrations of the ionic medium, as arising from the Debye- Hiickel equation. DISCUSSION The stability constant of the CoCl+ complex has in the present work been determined in a number of mixtures of Co(ClO,), with the other divalent transition metal perchlorates playing the role of constant ionic media.Fig. 2 shows a plot of384 COMPLEX FORMATION I N DMSO the respective values against total concentration of the mixtures. As is seen, there is a good agreement between the values obtained from independent experiments involving the different metal perchlorates. This we take as evidence of the real formation of the CoCP complex in DMSO solution, contrary to the earlier claims by some other At the same time the agreement confirms the expected constancy of the activity coefficients of the species involved in the respective complex-forming equilibrium in equimolar mixtures of the divalent transition metal perchlorates, apart from the nature of the metal cation. It also appears that the variation of the activity coefficients with the total concentration of the metal perchlorates is the same for both CoC1+ and MnCI+, as shown by the fact that the ratio of their derived stability constants remains constant when the total concentration of the metal perchlorates forming the ionic medium is varied (see table 2).( 3.0 I ( 2.5 L n - % 2.0 1.5 I I I 1 0.0 0.1 0.2 0.3 0.4 clmol dm-3 FIG. 2.-Dependence of the stability constant /I:'() of the CoCI' complex in DMSO solution on the totai concentration of the metal perchlorates, Co(ClO,),+M(CIO,),, forming the ionic medium: 0, M = Mn; 0, M = Ni; A, M = Cu; +, M = Zn; at 25 "C. The upper curve represents the Debye-Huckel plot on the assumption that ii = 8.0 A. These two observations confirm an earlier suggestion that, for some systems, the activity coefficients of solution species may be uniquely determined by the solution coordination state, being the same function of the latter for analogous complexes of different metals,24> 25 There is little doubt that the MCl+-type complexes are inner-sphere octahedral, MCl(DMSO),f, at least for Co" and Nil1, as indicated by the visible spectral effects accompanying their formation.17.2o Outer-sphere association producing the M(DMSO)g+ - Cl- ion-pairs, which are the other possible forms of occurrence of the MCI+ 'empirical' complexes, seems to be of minor importance in these systems, as also found for the weakest NiCl+ complex.2o It follows that the complex-forming reactions should be formulated as (8) and the above inferences concerning the activity coefficients relate to the quotient of the activity coefficients for these type of equilibria.Using equimolar mixtures of the divalent transition metal perchlorates as effectively constant ionic media in studies on complex formation of the respective metal cations M(DMSO);+ +C1- e MCl(DMS0); + DMSOw. LIBUS, R. PASTEWSKI AND T. SADOWSKA 385 provides a method of devising conditions where the activity coefficients remain constant. This approach to controlling activity coefficients seems to be superior to the more traditional one in which constant formal ionic strength is adjusted while the composition and the coordination state of the ionic medium are varied considerably. However, changing the concentration of the reactants, necessary in any equilibrium study, only in rare cases may be accomplished without changing the solution coordination state, as was achieved in the present work.Another example is provided by the isoconducting mixtures of ZnC1, and CoC1, in acetonitrile, containing complex electrolytes of the type M(AN)i+. 2MC1,AN- at constant concentrations, in addition to the neutral coordination forms MCl,(AN), believed to be of minor importance in determining the activity coefficients (AN denotes the acetonitrile There are reasons to believe that equimolar mixtures of some trivalent metal perchlorates, or some other trivalent metal salts involving non-coordinating anions, in aprotic donor solvents may also behave as effectively constant ionic media owing to the correspondence of their coordination statesz7 The expected agreement of the quotient of the activity coefficients for all the complex-forming reactions analogous to reaction (8) may be made use of for estimation of the respective stability constants at zero ionic strength from the ‘medium’ ones.In order to obtain a standard of the variation of the quotient of the activity coefficients for the MC1+-type complexes, the data for CoCl+ are used. The upper part of fig. 2 shows that the stability constant of the CoCl+ complex shows a variation with the concentration of the metal perchlorates forming the ionic medium which is approximately consistent with that arising from the Debye-Huckel equation with the ion-size parameter of 8.0 A (corresponding to Bao = 3.41). The latter value has been estimated from the molecular model of the complex cations involved in reaction (8) and the ionic radius of the unsolvated anion.The indicated values of the stability constant at zero ionic strength of this complex are those found previously from spectrophotometric and conductometric studies, respectively, of very dilute solutions of CoCl, in DMS0.I7 The straight line is that of the theoretical Debye-Huckel slope providing the best fit to the presently obtained ‘medium’ stability constants. The observed reasonable agreement of its intercept with the independently determined values of the stability constant at zero ionic strength indicates the reasonable validity of the assumed equation for the activity coefficients. Assuming its validity for the other complex-forming reactions similar to reaction (8), the respective stability constants at zero ionic strength have been calculated from the ‘ medium ’ ones determined in the ionic media of the metal perchlorates.The stability constants derived in the present work relate mainly to the MCP-type complexes (apart from their solvation) preferentially formed under conditions of excess metal perchlorates. Exceptional in this respect are the systems involving zinc(I1). The ZnC1: and ZnCl; complexes formed in them, most probably both tetrahedral, are so much stronger that they depress formation of the ZnC1+ complex, making derivation of its stability constant impossible. Qualitatively, this is in accord with the earlier findings of Ahrland and Bjork, who studied chloro complex formation of zinc(rr) using the NH,ClO, + NH,Cl solutions as reaction media of constant formal ionic strength.lqj For the other systems, the MCl+ complexes have been well- characterized and their real formation seems to be beyond doubt.It seems most probable that they all are of the MCl(DMS0): inner-sphere type. For M = Mn, confirmation of this solution form of the complex is provided by the agreement, discussed above, of the quotients of the activity coefficients of the complex-forming reactions for MnCl’ and CoCl+ Note that it is formation of the MC1+-type complexes that has the largest effect on the properties of the most dilute solutions of the respective386 COMPLEX FORMATION I N DMSO metal chlorides in DMSO, as well as in the other donor solvents. As a result, the stability order of these complexes is important for an understanding of the observed differentiation in behaviour of the divalent transition metal chlorides in their dilute solutions in these solvents.Fig. 3 shows plots of log PI for the MCl+-type complexes in DMSO solution against position of the metal within the Mn-Zn series. The plotted values relate to the same total concentration of 0.1 mol dmP3 of the metal perchlorates forming the ionic media and have been found by interpolation from the data listed in table 2. Also indicated are the thermodynamic stability constants calculated from the 'medium' ones as discussed above. The indicated values for ZnC1+ have been calculated from the data of Ahrland and Bjork4 using the Debye-Hiickel aquation for the activity coefficients. For obvious reasons they are not strictly comparable with our data determined in different media. 5 4 - c2 w 2 3 2 1 0- 0- 0 I \ I I I 1 I 1 I Mn Fe Co N i Cu Zn FIG.3.-Variation in the stability of the MCF-type complexes in DMSO solution within the Mn to Zn transition metal series: lower curve, in 0.1 mol dm-3 metal perchlorates; upper curve, in pure DMSO; points for Zn2+ recalculated from the data of Ahrland and Bjork.4 Inspection of fig. 3 shows that the stabilities of the MC1+-type complexes in DMSO solution do not follow the regularity known as the Irving-Williams series,28 but rather an 'inverted' Irving-Williams series. This is in accord with our earlier observations concerning some others systemsg+ 29+ 3o and seems to be typical of complexes formed by metal cations with weakly coordinating anions in strongly coordinating donor solvents like DMSO.To our knowledge, regularities of the present type are not generally known and the Irving-Williams series is believed to be obeyed, as it is in aqueous systems.:51 We postpone a more detailed discussion of this question to a later paper in which the thermodynamic characteristics of the complexes will be completed by the enthalpy data. ' K. R. Magnell and W. L. Reynolds, Znorg. Chim. Acta, 1972, 6, 571. W. LibuS, L. Frqczyk and B. Chachulski, Pol. J . Chem., 1978, 52, 493. W. LibuS, Muter. Sci., 1979, 5, 85. S. Ahrland and N. 0. Bjork, Acta Chem. Scand., Ser. A , 1976, 30, 257; 265. L. Sestilli and C. Furlani, Inorg. Nucl. Chem., 1970, 32, 1997. Z. LibuS, J . Phj,s. Chem., 1970, 74, 947.Z. LibuS, Znorg. Chem., 1973, 12, 2972. Z. LibuS and H . Tialowska, J . Solution Chem., 1975, 4, 101 1. j S. Ahrland, N. 0. Bjork and R. Portanova, Acta Chem. Scand., Ser. A . 1976, 30, 270. l o Z. LibuS and G. Kowalewska. Pol. J. Chem., 1978, 52. 709. ' I Z. LibuS and W. Maciejewski, Roc;. Chem., 1976, 50, 166. '' W. LibuS and H. Strzelecki, Electrochim. Actu, 1970, 15, 703; 1971, 16, 1749.w. LIBUS, R . PASTEWSKI A N D T. SADOWSKA 387 ' : I W. LibuS and M. Pilarczyk, Bull. Acad. Pol. Sci., Skr, Sci. Chim., 1972, 20, 539; 1973, 21, 773. I J W. LibuS and B. Chachulski, J . Solution Chem., 1980, 5, 355. W. LibuS, W. Grzybkowski and R. Pastewski, J . Chem. Soc., Furaday Trans. I , 1981, 77, 147. Ifi W. LibuS and H. Strzelecki, Electrochim. Actu, 1972, 17, 577. W. LibuS, M. Pilarczyk, R. Pastewski and T. Szuchnicka, Electrochim. Acta, in press. H. L. Schlifer and W. Schaffernicht, Angew. Chem., 1960, 72, 618. '!I Z. LibuS and T. Sadowska, J . Phys. Chem., 1969, 73, 3229. ?'I W. LibuS, M. Pilarczyk and T. Szuchnicka, Electrochim. Actu, 1975, 20, 831; 1980, 25, 1033. .'I V. Gutmann, Coord. Chem. Rer., 1967, 2, 239. ?? F. J. Rossotti and H. Rossotti, The Determinution of Stability Constants (McGraw-Hill, New York, 1961), p. 3. W. LibuS, Rocz. Chem., 1976, 50, 1813. ?-I W. LibuS, XIIIrh Int. Conf: Coordination Chemistry (Zakopane, Krakow, 1970), Section Lectures, p. 241. 2,i W. Libus, T. Sadowska and Z. LibuS, J . Solution Chem., 1980, 9, 341. ''Ii W. LibuS. D. Puchalska and T. Szuchnicka, J . Phys. Chem., 1968, 72, 2075. " i W. LibuS and M. Pilarczyk, Bull. Acud. Pol. Sci., Ser. Sci. Chim., 1974, 22, 717. :'" H. Irving and R. J. P. Williams, J. Chem. Soc., 1953, 3192. 2B W. Libus, B. Chachulski, L. Frgczyk and H. Strzelecki, Rocz. Chem., 1975, 49, 19. .I1' W. LibuS, Mater. Sci., 1979, 5, 135. .'I S. Ahrland, The Chemistry of'Non-Aqueous Solvents, ed. J. J. Lagowski (Academic Press, New York, 1978), vol. VA, p. 45. (PAPER O/ 1836)