Activities from vapor pressure measurements of lithium and of sodium chlorates in water and water-dioxane solvents

Transcription

1 Activities from vapor pressure measurements of lithium and of sodium chlorates in water and water-dioxane solvents A. N. CAMPBELL AND B. G. OLIVER' Department of Chemistry, University of Manitoba, Wirzizipeg, Manitoba Received January 8, 1969 The vapor pressures of solutions of sodium chlorate and of lithium chlorate in water and in solvents consisting of 44.5 % dioxane % water and of 64.5 % dioxane % water, at 25 "C, were determined using a differential manometer. The vapor compositions of the dioxane-water solutions were found by the air-saturation method. The mean molal activity coefficients were calculated from the vapor pressures aed compositions. The activity coefficients of the solute were fitted to the theoretical equations of Stokes and Robinson and of Glueckauf, yielding hydration numbers for the cations of the salts. The minimum dioxanation numbers of sodium and of lithium chlorates in the dioxane-water mixtures were calculated from the experimental activity coefficients by the method of Grunwald. The results show that lithium chlorate is more highly solvated than sodium chlorate and that dioxane plays a major role in the solvation of both electrolytes in the dioxane-water mixtures. Canadian Journal of Chemistry, 47, 2671 (1969) Introduction We had previously determined the conductances of sodium and of lithium chlorates in water and water-dioxane mixtures (I), and it seemed to us appropriate to follow this with an investigation of the activities of the same salts in the same media. The technique chosen was that of vapor pressure and vapor composition. From these data, the partial pressures of the components of the mixed solvents and, by way of the Gibbs-Duhem theorem, the activities of the salts could be obtained. Experimental Materials Salts Fisher certified sodium chlorate (major impurity 0.01 % bromate) was dried at 120" for 48 h or longer, before use. Lithium chlorate was prepared by the method of Campbell and Griffiths (2). The melting point was found to be 127.g0, in agreement with the value of Campbell and Williams (3). Solvents Water, of specific conductance 2-5 x lo-' mhos/cm, was obtained by passing ordinary distilled water through a mixed-resin demineralizing column. Dioxane of spectro quality was further purified by refluxing with sodium, according to the method of Hess and Frahm (4). After purification, the dioxane had a refractive index of ndz5 = 1 A202 and a density of g/cc at 25". This is in close agreement with the refractive index of found by Frey and Gilbert (5) and the density of found by Schott (6). 'Present address: Department of Chemistry, Rensselaer-Polytechnic Institute, Troy, New York. Apparatus A 3 ft high cylinder was used as one of three thermostats, for work with the differential manometer. Vigorous stirring with a 114 hp motor gave a temperature control of ", despite the height of the mass of liquid. The other two thermostats were of conventional type. Temperature was measured with Beckmann thermometers which were calibrated, and frequently re-calibrated, against a platinum resistance thermometer and Mueller bridge. Temperature fluctuations never exceeded k 0.005". Vapor Pressure Measurements Following Gilson and Adams (7) and Shankman and Gordon (8), a differential manometer was used for the determination of total vapor pressure. A modification of Gilson and Adams' apparatus, introducing glass spirals between the manometer and the three-way cocks to the vacuum line, saved us many breakages. As manometer liquid, we used vacuum pump oil for water solutions, as recommended by Shankman and Gordon (8), but we found it necessary to change the oil after about 5 runs, because of contamination of the oil-glass interface with water. From the densities of oil and mercury and the accuracy of the cathetometer, it would seem that vapor pressures could be measured with an accuracy of k0.005 mm of mercury but, as pointed out by McAllan (9), difficulties such as slow drainage time and dissolved gases lead to higher errors than this. We estimate the error of our measurements to be k0.01 mrn of mercury. When investigating solutions in water-dioxane mixtures, we found that pump oil, diethyl phthalate, dibutyl phthalate, and other manometric liquids, all dissolve dioxane, and we were therefore forced to use mercury as manometer liquid. The limbs of the manometer were widened to 22 mm so that the capillary depression became negligible. Using mercury, the accuracy of mezsurement was reduced to k0.05 mm. All measurements were reduced to 0" in the usual way. The solvent was degassed by freezing in dry iceacetone mixture, pumping down to 1.0 p, melting, and

2 2672 CANADIAN JOURNAL OF CHEMISTRY. VOL. 47, 1969 repeating the whole process five times. The solvent was then vacuum distilled on to the salt which had previously been placed in one of the manometer flasks. After the determination of vapor pressure, the concentration of the solution was determined by evaporation of a weighed portion to dryness. Separate experiments showed that the composition of the mixed solvent did not change during the manometric experiment. Vapor Composition The compositions of the vapors in equilibrium with liquid mixtures of given composition were determined by the air saturation method, using what was essentially the apparatus of Bichowsky and Storch (10). Vapor compositions by this method were reproducible to within 0.1 %. The condensed vapor was analyzed refractometrically. Treatment of Results As the mathematical treatment is involved, we confine ourselves, for the most part, to referring to published papers, in which this treatment is described. The activity of the water in the aqueous solutions was calculated from the relation where p, is the vapor pressure of the water in solution andp10 the vapor pressure of pure water. In what follows subscripts 2 and 3 refer throughout to dioxane and electrolyte, respectively. The activity of the solute was then calculated from the activity of the solvent by using the Gibbs- Duhem equation SdT - VdP + z ~,dc, = O which for constant temperature and pressure, simplifies to The details of the method are given by Randall and White (11). Because the data were not sufficiently accurate below 0.1 molal, we took the value of logy at 1 molal as known and integrated to the observed concentration. For sodium chlorate in water at a concentration of 1 molal, we used the figure of Jones (12) for the activity coefficient, viz For lithium chlorate, at 1 molal concentration, we used the value y = from the freezing point data of Scatchard and Prentiss (13). Scatchard's figure refers necessarily to the freezing point of the solution and, in using the same figure at 25", we make an assumption that may not be justified. Such an assumption may lead to an error as great as 3% but, when accurate data become available for the dilute region at 25", our results can be corrected by multiplying by the appropriate factor. The above simple treatment cannot be used when the solvent is a mixture of water and dioxane. It is now necessary to know both the vapor pressure and the vapor composition in order to calculate the activities of the two components of the solvent. Assuming the applicability of Dalton's law and knowing Ni, the mole fraction of the ith component in the vapor, the partial pressure is obtained and.the activity of component i in solution results as The addition of solute to a two-component solvent at constant N, can be represented by the equation where M,, = NIMl + N2M2, is the mean molecular weight of the binary solvent. If the solution is not ideal, the equation shows that a, and a, may alter by different factors. Indeed, where there is strong preferential solvation of the solute by component 1, it is possible for the decrease in a, to be so pronounced that a, actually increases. We used the treatment of Grunwald and Baccarella (14). In this treatment, attention is focussed on d In a,/a, or, what amounts to the same thing, on the experimental quantity, d In u1/u2, where u, = p,/p, * and u, = p,/p,*: p, andp, are the partial pressures of water and of dioxane over the solution containing solute, and p, " andp, * are the partial pressures of water and of dioxane over the solvent of composition N,, in the absence of solute. It is unnecessary to repeat the treatment of Grunwald and Baccarella here. Suffice it that the following equation results for 1-1 electrolytes [I] (1000/~,,)(~,d ln u, + N,d ln u,) = -2dm - 2mdln y, Here m is expressed in formula weights of electrolyte per kg of binary solvent. Following

3 CAMPBELL AND OLIVER: ACTIVITIES FR :OM VAPOR PRESSURE MEASUREMENTS 2673 Scatchard and Prentiss (1 3) In y. can be expanded in a power series in m1i2 equation for the evaluation of a. Assuming that only the cation is hydrated, the parameter a is In eq. [2], S is the Debye-Hiickel limiting slope and B, C, D, E,. are adjustable constants. The latter are evaluated from the experimental data as follows. Substitution of [2] in [l] and integration leads to [3] (- 1000/Ml,)(Nl ln a, + N, ln a,) = 2 m - 2/3sm3I2 + Bm2 + 6/5Cm5I2 In eq. [3], the terms on the left and the first two terms on the right involve only experimental data and known solvent properties. The adjustable constants in the remainder are evaluated by standard least squares methods,using a computer. Once the constants are obtained, it is a simple matter to insert them in eq. [2] in order to calculate the activity coefficients. In no case were terms higher than Dm2 found to be necessary. Results and Discussion (a) Results in Pure Water (Table I) At comparable concentrations, the activity coefficients of lithium chlorate in water are considerably larger than those of sodium chlorate. This is consistent with the idea that the smaller lithium ion (r = 0.68 A) is more extensively hydrated in aqueous solutions than the larger sodium ion (r = 0.97 A). The extremely high value attained by the activity coefficient of the lithium salt, (y = at m) is probably due to the extensive hydration of the lithium ion. We attempted to treat our results with the theoretical equation of Stokes and Robinson (15) viz. ~41 logy*= - A(Z, z,jz1i2 h - - log a, 1+~Qz112 - log [l W,(V - h)m] This eq. contains two adjustable parameters, h, the hydration number, and a, the distance of closest approach. The first term is the well known Debye-Hiickel equation. In a more recent paper (16) Stokes and Robinson have developed an The partial molal volumes of the sodium and lithium ions which we introduced into the above eq. were obtained from the data of ~ukerjee (17), and that of the chlorate ion from Couture and Laidler (18). These values are FN',,+O = cc/mole, VLi+O = -5.2 cc/mole, and V,,,,- = cc/mole. The values of the hydration number 12, the 'a' values calculated by the above method, and the calculated activity coefficients are presented in Table I. Table I shows good agreement between the experimental activity coefficients of sodium chlorate and the theory over the entire concentration range, viz molal. The results of Jones (12) for the more dilute region when treated in the same way run smoothly into ours. This is to be expected, since sodium chlorate is only slightly hydrated (h = 1.05) and there are plenty of free water molecules even at saturation. On the other hand, the theory begins to break down above 2 molal, for lithium chlorate. This also is not surprising, since lithium chlorate is highly hydrated (h = 6.7) and there are only a limited number of water molecules available. In fact, Stokes and Robinson (15) have stated that in the majority of cases, the equation breaks down when the product of the hydration number, h, and the molality, nz, exceeds 10 or 15. To proceed to higher concentrations it would be necessary to use a hydration number which was a suitably decreasing function of concentration. We then fitted the experimental activity coefficients to an equation developed by Glueckauf (19). This equation has two adjustable parameters, h and a. The number of adjustable parameters was

4 2674 CANADIAN JOURNAL OF CHEMISTRY. VOL. 47, 1969 TABLE I Vapor pressures, water activities, and activity coefficients of sodium chlorate and of lithium chlorate in water at 25 C Molal activity coefficient of salt, y Experimental Interpolated Calculated concentration Vapor Water concentration (moles/kg pressure activity (moles/kg Stokesof solvent) (mm Hg) a, of solvent) Experimental Robinson Glueckauf Sodium chlorate h=1.05 a=3.56 h=0.96 a=3.50 r= Lithium chlorate reduced to one by substitution of eq. [5] for a. The values of the hydration number, h, the a values calculated by eq. 151 and the calculated activity coefficients are shown in Table I. Glueckauf's eq. represents the data as well as that of Stokes and Robinson. In fact, it seems to be valid over the entire concentration range for sodium chlorate, and up to 2.5 molal for lithium chlorate. The hydration numbers found by this method are lower than those obtained by the Stokes and Robinson equation and this agrees with Glueckauf's own findings. Both equations give values for h and a (the distance of closest approach) of the right order of magnitude. Although Glueckauf's derivation takes into account the size differences between the hydrated ions and the water molecules, both Glueckauf (19) and Stokes and Robinson (15) assume that

5 CAMPBELL AND OLIVER: A~IVITIES FROM VAPOR PRESSURE MEASUREMENTS 2675 all the differences between the experimental and the calculated activity coefficients can be ascribed to hydration. It is true that the resulting equations are remarkably useful in describing observed results but this should not obscure the fact that, while hydration is a very important factor, it is not the only one which determines the complicated equilibria in an electrolyte solution. To put it briefly, the secondary hydration has been stretched to cover a number of effects, the quantitative nature of which requires a great deal of investigation. We attempted to get some information about the behavior of the concentrated lithium chlorate solutions, on the basis of the concept of Stokes and Robinson (15) that the ions adsorb layers of water molecules in accordance with the Brunauer, Emmett, and Teller (B.E.T.) equation (20). Modifying the notation of those authors to suit the present case, their equation is r71 maw -- 1 (b - 1) - +- aw 55.51(1 - a,) bg bg where a, is the water activity of the solution; m is the molality, g is the number of molecules of water in the unimolecular hydration layer when it is complete, and b is a constant related to the heat of adsorption E of the molecules in the layer and is given by the approximate relation b = exp (E- EL)/RT, EL being the heat of liquefaction of pure water. Equation [7] may be tested by plotting the left hand side (determined from experimental m and a,) against a,. When we did this, we obtained a straight line for lithium chlorate in water at 25" when the water activity of the solution is less than 0.420, i.e. above 12 molal. The best values of b and g were obtained from the slope and intercept of the line, by the method of least squares. The accuracy of the fit is fairly good over the range molal, (a mean deviation in a, of ). The value ofg, 3.11, is between 3 and 4 in accordance with the values found by Stokes and Robinson (15) for other 1-1 electrolytes. The value of b, 8.75, is also reasonable and corresponds to E - EL = 1.29 kcal/mole of water adsorbed. An unsatisfactory feature of eq. [7] is that it demands a non-integral p value. This can - - its application to this case. The most drastic of these approximations is that of treating all water molecules beyond the first layer as being held by ordinary liquid forces, with a heat of liquefaction of EL. Anderson (21) has introduced a modification of the B.E.T. equation, in which the subsequent layers have a heat of adsorption less than that of water by d. This has the effect of multiplyii~g a,, wherever it occurs in eq. [7] by a factor K = e - and ~ leads ~ to the ~ equation ~ To test the appropriateness of this equation, the value of g is set at 4. Then eq. [8] is written in the form gmaw r91 = (i- aw)/(55.51(1 - Ka,) - a,) and by trial and error, the value of K which leads to a reasonably constant b over the widest possible range of molality is found : this value of Kis The average deviation of eq. [8] is the same as that of eq. [7] but the range fitted is slightly smaller (12-37 molal). The value of the parameter b, 6.9, is plausible and corresponds to E - EL = 1.14 kcal/mole of water in the second and subsequent layers. This relatively small energy might easily correspond to a weak ordering effect on the water molecules concerned. The parameter b has a value of 4. This means that lithium chlorate has 4 sites available for occupation by water molecules in the inner layer, each being held with an energy some 1.14 kcal/mole greater than the latent heat of evaporation of water (10.48 kcal/mole at 25"). The Stokes-Robinson approach used for lower concentrations and the B.E.T. approach used for higher concentrations are not mutually contradictory. As the concentration increases, the hydration number h diminishes and the ions tend to a quasi-crystalline structure with some of the water molecules imbedded in the remnants of crystal lattice (adsorbed water) and some present as "free" solvent. These are the two extremes of a more general theory, which should cover the entire concentration range. scarcely correspond to any physical reality, and (6) Results in Mixed Water-Dioxarze Solvents has probably arisen as a result of approximations (Table 11) in the Brunauer, Emmett, and Teller theory and The properties of solutions in the mixed

6 2676 CANADIAN JOURNAL OF CHEMISTRY. VOL. 47, 1969 solvents will be largely determined by certain physical properties of the mixed solvents, notably the dielectric constant and the viscosity. These, at 25", are as follows for the three solvents Dielectric Viscosity Solvent constant (cp) Water % dioxane- 55.5% water % dioxane % water Figure 1 gives the change of the water and dioxane activities with addition of sodium chlorate to a solvent consisting of 44.5% dioxane, 55.5% water. It is seen that the water activity decreases upon addition of salt whereas the activity of dioxane increases. These results can be explained on the assumption that sodium chlorate is strongly preferentially solvated by MOLALITY FIG. 1. Water and dioxane activity vs. molality for the system: sodium chlorate % dioxane % water, at 25', as determined from vapor pressure and vapor composition; 0, water activity; 0, dioxane activity MOLALITY FIG. 2. Water and dioxane activity vs. molality for the system: lithium chlorate % dioxane % water, at 25", as determined from vapor pressure and vapor composition; 0, water dioxane activity. water. As salt is added to the solution, more water than dioxane is bound by the salt, in forming a solvation sheath. Thus the remaining solvent becomes richer in dioxane and the partial pressure and activity of dioxane in the system rises. Similar changes in the water and dioxane activities are observed when sodium chlorate is added to a solvent consisting of 64.5% dioxane % water and a similar explanation applies. The changes in water and dioxane activity for lithium chlorate in 44.5% dioxane (Fig. 2) and in 64.5% dioxane are quite different from those of sodium chlorate. The water activity in the presence of lithium chlorate decreases to a greater extent than with sodium chlorate, while the dioxane activity at first shows a slight increase and then decreases. The larger decrease in water activity with lithium chlorate is to be expected because of the higher hydration of the lithium ion. The activity coefficients of sodium and of lithium chlorate in these dioxane-water mixtures were calculated by the method of Grunwald and Bacarella (14) as modified by Scatchard and Prentiss (13) (see "Treatment of Results"). The

7 CAMPBELL AND OLIVER: ACTIVITIES FROM VAPOR PRESSURE MEASUREMENTS TABLE I1 Vapor pressures and composition data* Total Mole % Partial vapor Molality vapor dioxane pressure of (moleslkg pressure in dioxane Solution of solvent) (mm Hg) vapor (mm Hg) a,t a23 Sodium chlorate in 44.5 % dioxane Lithium chlorate in 44.5 % dioxane Sodium chlorate in 64.5 % dioxane I Lithium chlorate , in 64.5% dioxane *Total vapor pressures vapor compos~t~ons and partial vapor pressures of solut~ons of sodium chlorate and of l~thium chlorate In 44.5% dioxane and in 64.5% dldxane at 25". togethe; with activlty coeffic~ents according to Grunwald and Bacarella (14).?a, IS defined as (part~al pressure of water over m~xed solvent contalnlng salt)/(partlal pressure of water over the mixed solvent). ia2 is defined as (partial pressure of dloxane over mixed solvent conta~nlng salt)/(partlal pressure of d~oxane over the m~xed solvent). values of the constants, of eq. [2] calculated by higher degree of solvation of the lithium ion. the method of least squares on a computer, are When the activity coefficients of each salt in each given in Table 111. These constants represent the solvent (including water) are plotted against salt experimental quantity (N, In a, + N2 ln a,) of concentration, it is seen that the activity coeq. [3] for all systems in mixed solvent, over the efficients of both sodium chlorate and lithium entire concentration range, with a maximum de- chlorate are greatly reduced on changing the viation of 2%. This means that the activity solvent from water to 44.5% dioxane to 64.5% coefficients are accurate to about the same extent. dioxane. Much of the decrease can be ascribed to Graphs of the activity coefficients of sodium the large change in dielectric constant, with and lithium chlorate in 44.5% dioxane and in consequent increased tendency to ion pair 64.5% dioxane are shown as Figs. 3 and 4. The formation. curve for lithium chlorate lies much above that of Many inconsistencies appear in the literature sodium chlorate, over the entire concentration as to the nature of solvation in dioxane-water range. This behavior is similar to that in aqueous mixtures. Grunwald (22), in a study of a variety solution and can again be explained as due to a of electrolytes in 50% dioxane-water mixtures by

8 CANADIAN JOURNAL OF CHEMISTRY. VOL TABLE I11 Constants of the equation of Scatchard and Prentiss (13) In y, = -Stnl/' + Bnz + Cm3l2 + DmZ System S B C D Sodium chlorate in 44.5 % dioxane Deviations Lithium chlorate in 44.5 % dioxane Deviations Sodium chlorate in 64.5 % dioxane Deviations Lithium chlorate in 64.5 % dioxane Deviations FIG. 3. Mean molal activity coefficients vs. molality for sodium chlorate and for litliium chlorate in the solvent 44.5% dioxane % water, at 25"; 0, sodium chlorate; 9, lithium chlorate. a vapor pressure technique, observed partial solvation of the electrolytes by dioxane. Fratiello and Douglass (23) used nuclear magnetic resonance (n.m.r.) to show that the chlorides of Li, Na, K, Mg, Ca, Zn, and Cd, along with the sodium salts of I, NO,, ClO,, and SO, are preferentially solvated by water in 50% dioxane - FIG. 4. Mean molal activity coefficients vs. molality for sodium chlorate and for lithium chlorate, in the solvent 64.5 dioxane % water at 25"; 0, sodium chlorate; 9, lithium chlorate. water mixtures: they found no evidence of solvation by dioxane. Other n.m.r. evidence by Hinton et al. (24) indicates solvation of aluminium perchlorate by dioxane. Analysis of conductance data by Hyne (25) also indicated interaction between dioxane and electrolyte in similar mixtures. We attempted to determine the nature of the

9 CAMPBELL AND OLIVER: ACTIVITIES FROM VAPOR PRESSURE MEASUREMENTS 2679 solvation in the mixed solvents by the method of Grunwald (14, 22, 26). His final equation is Here G,' is the mean value of Go for cation and anion. The coefficient of the m3i2 term was calculated from the Debye-Hiickel limiting slope S, along with the slope of the curve of S plotted against N,. The values of the other coefficients on the right in eq. [lo] were obtained by fitting the data by the method of least squares on a computer. The constants that reproduced the experimental quantity ln a,/a, over the entire concentration range, with a maximum deviation of 2%, are presented in Table IV, together with other useful quantities. The values of dg,o were calculated from the values of 2/RT.dG,O/dN1 of Table IV. The values of the quantities d In &/dn,, d In al/dn, and d In a2/dnl were calculated by plotting In&, In a,, and In a, against N,. The required data for the dielectric constants (E) of various dioxane-water mixtures were obtained from Critchfield et al. (27), while the data required for the activities of water and dioxane in water - dioxane mixtures were obtained from Bacarella et al. (28). If it is assumed that the solute is present at extreme dilution and that it exists as-an equilibrium mixture of molecular complexes of the type S.iH,O-jdioxane, it is possible to obtain the average solvation numbers h0 and do. The treatment is complex and it is unnecessary to go into it here but suffice it that the following equations are obtained + ~ORT- dln a, dn1 where b + is a harmonic average, defined by the eq. +.?.xi ~j anions 2) bij where $ijo is the fraction of total formula weight which exists in the form S-iH,O.jdioxane. In order to apply eqs. [ll] and [12] to our data, it is convenient to represent b, as a function of r,, ho, and do. It is assumed that the harmonic average of the bij values may be approximated by the radius calculated for a hypothetical "average solvate", defined as ion with unsolvated radius r, complexed with h0/2 water molecules and d0/2 dioxane molecules. Assuming spherical molecules and closest packing of spheres, the following eq. results where r, and r, were taken as 1.6 and 2.7 A, respectively. (These radii are computed from the molar volumes of the pure liquids, assuming random packing of spheres, i.e. the actual volume occupied by the molecules is 58% of the total volume.) Using the values quoted in Table IV, eqs. [l 1 ] and [13] can be simplified to the form for 44.5% dioxane, while for 64.5% dioxane the following simplified equation can be written Equations [14], [15], and [16] can now be used to estimate solvation numbers. The first step is to show that do is greater than zero. When dg,o/dn, for sodium chlorate in 44.5% dioxane is plotted against ho, the calculated value of dg,o/dn1 is always more negative than the experimental value. Similar behavior is observed with sodium chlorate in 64.5% dioxane and with lithium chlorate in the same solvents. Agreement between the experimental and calculated values is possible only if do is greater than zero. This means that there are two unknown parameters, h0 and do, but since there is only one experimental quantity, dg,o/dn,, it is evidently not

10 CANADIAN JOURNAL OF CHEMISTRY. VOL. 47, 1969 TABLE IV Constants for eq. [lo] and derived quantities for eq. [Ill (g (z*; - -, DI) - - ~ r ) - ~ r ) 4 din& a d In a2 System 1 3 dn dn1 dn1 dn1 NaC103 in 44.5 % d~oxane LiCI03 in 44.5 % dioxane NaC103 in % d~oxane LiCIOB in 64.5 % dioxane possible to obtain a unique solution. It is possible, 1. A. N. CAMPBELL> E. M. KARTZMARK~ and B. G. OLIVER. Can. J. Chem. 44, 925, (1966). substituting the of h0 and 2. A. N. CAMPBELL and J. E. GRIFFITHS. Can. J. Chem. do that are consistent with the ex~erimental (1956). results, to find the smallest value of' do which 3. A.' N. CAMPBELL and D. F. WILLIAMS. Can. J. Chem. 42, 1778 (1964). gives the experimental value for dg*oldn,. 4. K. HESS and H. FRAHM. Ber. B 71,2627 (1938). Minimum dioxanation numbers, calculated in 5. P. R. FREY, and E. C. GILBERT. J. Am. Chem. Soc. 59, 1344 (1937). this way, along with the of 6. H. SCHOTT. J. Chem. Eng. Data, 6, 19 (1961). ho, are given in Table V. 7. R. E. GILSON and L. H. ADAMS. J. Am. Chem. SOC. 55, 2679 (1933). TA,BLE V 8. S. SHANKMAN and A. R. GORDON. J. Am. Chem. Minimum values of the average dioxanation SOC. 61, 2370 (1939). number at 25" 9. J. V. MCALLAN. J. Sci. Instr. 42, 290 (1965). 10. F. R. V. BICHOWSKY and H. V. STORCH. J. Am. Solution d0m n h0 Chem. SOC. 37, 2696 (1915) M. RANDALL and A. M. WHITE. J. Am. Chem. Soc. NaC103 in 44.5 % dioxane , 2514 (1926). LiC103 in 44.5 % dioxane J. H. JONES. J. Am. Chem. Soc. 65, 1353 (1943). NaC103 in 64.5 % dioxane ' G. SCATCHARD and S. S. PRENTISS. J. Am. Chem. LiC103 in 64.5 % dioxane SOC. 56, 1486 (1934) E. GRUNWALD and A. L. BACARELLA. J. Am. Chem. SOC. 80, 3840 (1958). Thus dioxane plays amajor role in the solvation 15. R. H. STOKES and R. A. ROBINSON. J. Am. Chem. of electrolytes in dioxane-water mixtures. This at 16. SOC. R. H. 70, 1870 (1948). R. A. Trans. Faraday first sight may appear unlikely since pure dioxane SOC. 53, 301 (1957). has such a low dielectric constant (2.21 at 25") 17. P. MUKERJEE. J. Phys. Chem. 65,740 (1961). 18. A. M. COUTURE and K. J. LAIDLER. Can. J. Chem. and, therefore, a low dipole moment. Hyne (25), 35, 207 (1957). however, has shown that addition of electrolyte 19. E. GLUECKAUF. Trans. Faraday Soc. 51, 1235 to dioxane-water mixtures greatly increases the (1955). 20. S. BRUNAUER, P. H. EMMETT, and E. TELLER. J. Am. microdielectric constant above the bulk value. Chem. Soc. 60, 309 (1938). Such an increase is in keeping with the displace- 21. R. B. ANDERSON. J. Am. Chem. Soc. 68, 686 (1946). merit of the boat-chair equilibrium for dioxane 22. E. GRUNWALD. Electrolytes. The Pergamon Press, Ltd., New York p. 62. in favor of the boat form, under the influence of 23. A. FRATIELLO and D. C. DOUGLASS. J. Chem. ~ hy~. the Coulombic field of the ionic species. Due to 39, 2017 (1963). 24. J. F. HINTON, L. S. McDowELL, and E. S. Ahl1s. the opposition of the bond moments in the chair Chem. Commun. 776 (1966). form, the preponderance - - of this form in Dure 25. J. B. HYNE. J. Am. Chem. Soc (1963). dioxane results in a low dielectric constant,abut 26. E. GRUNWALD, G. BAUGHMAN, and G. KOHNSTAM. J. Am. Chem. Soc. 82, 5801 (1960). the preponderance of the boat form, with its 27. F. E. CRITCHFIELD, J. A. GIBSON, and J. L. HALL. J. bond moments acting together, when electrolyte Am. Chem. Soc. 75, 199 (1953). is added to the system, may explain the strong 28. A. L. BACARELLA, A. FINCH, and E. GRUNWALD. J. Phys. Chem. 60, 573 (1956). interaction between dioxane and electrolytes.