1960
J. J. BANEWICZ, J. A. MAGUIRE, AND P. S. SHIH
The Electrical Conductance of Tetraethylammonium Perchlorate in Valeronitrile by John J. Banewicz, John A. Maguire, and Pong Su Shih Department of Chemistry, Southern Methodist University, Dallas, Texas 76222 (Received February 6, 1968)
Conductances of solutions of tetraethylammonium perchlorate in valeronitrile have been measured at 25, 30, 40, and 50'. The data have been analyzed using the Fuoss-Onsager 1957 equation. Appreciable ionpair association was found at all temperatures, and values of the ion-pair association constant were calculated. Variations in the association constant and Walden product with temperature are reported and are briefly discussed. Introduction Acetonitrile has been used extensively as a solvent in the investigation of the electrical conductances of salts in nonaqueous solutions.l-5 Acetonitrile is attractive in that it is a polar, nonhydrogen bonding solvent in which many salts are soluble. Several other nitriles have also been used but to a much more limited e~tent.68~ The results of these investigations indicate that salts composed of cations, such as the tetraalkylammonium ions, and symmetric anions, such as perchlorate, are not solvated appreciably in nitrile solvent^.^-^ Owing to this lack of solvation, ion-pair formation in these solvents should be determined predominantly by ion-ion electrostatic interactions. However, in acetonitrile, because of its moderately high dielectric constant (36 at 25"), ion-pair formation takes place to such a limited extent that values of the association constants derived from conductance data are small and difficult to determine with any precision.a Longer chain nitriles, such as valeronitrile, have low enough dielectric constants (about 20) so that moderately high ion-pair association constants would be expected, and these solvents should be interesting ones in which to investigate ion association. As part of an investigation of the solvent properties of valeronitrile, the electrical conductances of solutions of tetraethylammonium perchlorate in this solvent have been determined at temperatures from 25 to 50". Tetraethylammonium perchlorate was chosen as a solute because its ions are small enough to yield fairly high ion-pair association constants, yet sufficiently large so that the possibility of solvation would be small. Experimental Section Eastman reagent grade tetraethylammonium perchlorate (TEAP) was recrystallized twice from water, w* dried at 'Oo0 for hr' and then was further dried w&5 in a vacuum Oven at 6oo for days* prepared by the method of Ferguson.1° The valeroThe Journal of Physicat Chelnietry
nitrile was dried over anhydrous calcium chloride prior to distillahion. The distilled nitrile was allowed t o stand over phosphorus pentoxide and then was fractionated. The fraction boiling at 140-141' was collected. The phosphorus pentoxide treatment, followed by fractionation, was repeated until no coloration was noted on treatment with phosphorus pentoxide. I n later runs, the final distillation was done from an anhydrous potassium carbonate-calcium sulfate mixture. This procedure yielded valeronitrile with specific conductances a t (1.2-3.8) X lo-* mho cm-'. Stock solutions of TEAP in valeronitrile were prepared by weight. Solutions whose conductances were to be measured were made by mixing known weights of the stock solutions with weighed amounts of pure valeronitrile. Molar concentrations were calculated from the weights and densities of the solutions. The densities of the TEAP-valeronitrile solutions, as functions of concentration and temperature, were measured using a Sprengel-Ostwald pycnometer, calibrated with water as a standard. The viscosities of the various solutions were determined a t several temperatures using an Ostwald-Fenske viscosimeter. The viscosimeter was calibrated a t the different temperatures with water as a standard, using the viscosity values of water a t the various temperatures given by Coe and Godfrey.I1 (1) I. Y. Ahmed and C. D. Schmulbach, J . Phys. Chem., 71, 2368 (1967). (2) R. L. Kay, B. J. Hales, and G. P. Cunningham, ibid., 71, 3926 (1967). (3) D.F. Evans, C. Zawoyski,and R. L. Kay, ibid., 69,3878 (1965). (4) J. F. Coetaee and G. P. Cunningham, J. Amer. Chem. SOC.,87, 2534 (1965). (6) A. M. Brown and R. M. Fuoss, J . Phys. Chem., 64, 1341 (1960). (6) P.G.Sears, J. A. Caruso, and A. I. Popov, ibid., 71, 905 (1967). (7) J. F. Coetaee and D. I(. McGuire, ibid., 67, 1810 (1963). (8)J. F. Coetzee and D. K. MoGuire, ibid., 67, 1814 (1963). (9) D. 8. Berm and R. M. Fuoss, J . Amer. Chem. Boo., 82, 5586 (1960). (10) J. W.Ferguson, Proc. Indabna Acad. Sci., 6 3 , 131 (1953). (11) J. R. Coe, Jr., and T. B. Godfrey, J . Appl. Phys., 15, 626 (1944).
1961
THEELECTRICAL CONDUCTANCE OF TETRAETHYLAMMONIUM PERCHLORATE The dielectric constants of valeronitrile were measured in the temperature range from 25 to 50" using a heterodyne beat circuit operating at 5 Mcps and a glass capacitance cell of the type described by LeFevre.12 The cell was calibrated using benzene as a standard. Conductance measurements were made using an ac bridge similar to that described by Jones and Dike.laJ4 The conductance cell was of the same type used by Daggett, Bair, and Kraus,15 except that in place of the flask reservoir, a flat-bottom cylindrical reservoir with a capacity of approximately 40 ml was used. l6 The electrodes were 15.8-mm platinum disks, spaced about 1.5 mm apart. The cell constant was found to be 0.07911 & 0.00005 cm-' at 25" and was assumed to be constant over the range of temperatures studied. A Sargent viscosimeter bath, Catalog No. S-67528, filled with mineral oil, was used to maintain a temperature constant to within *0.01". The conductance of each solution was measured at 25". The temperature was then increased from 25 to 50", in 5" stepis, and readings were taken at each temperature. Additional conductance measurements were made as the temperature was decreased from 50" back to 25". I n all cases the two readings at the same temperature agreed to within less than 0.1%. Results and Discussion The viscosities, 7, densities, d, and dielectric constants, D, obtained for valeronitrile at the various temperatures are summarized in Table I. Table I : Properties of Valeronitrile Temp,
5
d,
OC
D"
g/ml
IOZtl, P
25 30 40 50
20.03 19.64 18.81 18.06
0.7948 0.7906 0.7816 0.7730
0.6928 0.6485 0.5719 0.5084
2.36
d
J 2.32 2.28 1.68 1.70 IO'/DT, deg-1.
Figure 1.
1.72
The variation of log K Awith l/DT.
Density and viscosity data were also taken on solutions of TEAP in valeronitrile at 25". From these data, a value of F equal to 0.58 was obtained. Since F should be temperature independent in nonaqueous solvents, this value was used in calculations at all temperatures. I n confirmation, Kay, Zawoyski, and Evans20 have found only small variations in F occurring with temperature for a number of symmetrical tetraalkylammonium salts. The x-y methodl9 was used in obtaining the conductance parameters listed in Table 111. The uncertainties in K A are the standard deviations and those in dJ were calculated from the standard deviations in J. Using the best values of K A and J at a particular temperature, a value of A. was calculated at each concentration from eq 1. The uncertainties in A, shown in Table I11 are the average deviations obtained from the calculated A;s. Figure 1 shows a plot of log K A vs. 1/DT for the TEAP in valeronitrile. The best straight-line equation that reproduces these data, as determined by the method of least squares, is log K A = -(1.38
x 104 * 0.15) + (2.2 & D0.9) T
(2)
Dielectric constant.
The values of dielectric constants at 30, 40, and 50", listed in Table I, are in good agreement with those obtained by Krishnaj i and Mansingh at these temperat u r e ~ . ~ 'The density at 30" agrees quite well with the literature value of 0.79058.1s However, the viscosity at 30" is higher than that previously reported.ls The conductances of TEAP solutions at the different temperatures are shown in Table 11. The conductance data were analyzed using the FuossOnsager 1957 equation, which has the forml9 A, =
A0
- S(CY)"* + ECYlog
(CY)
+ JCY - KACY.PA, (1)
in which the symbols have their usual meanings.
The uncertainties reported in eq 2 are the standard deviations. Several equations have been proposed describing the (12) R. J. W. LeFevre, "Dipole Moments," 2nd ed, Methuen and Co.,London, 1948,p 36. (13) G . Jones and R. C. Josephs, J . Amer. Chem. SOC.,50, 1049 (1928). (14) P. H.Dike, Rev. Sci. Instrum., 2 , 379 (1931). (15) H.M. Daggett, Jr., E. J. Bair, and C. A. Kraus, J . Amer. Chem. SOC.,7 3 , 799 (1951). (16) For details of the cell construction, see P. S. Shih, M.S.Thesis, Southern Methodist University, Dallas, Texas, 1967. (17) Krishnaji and A. Mansingh, J . Chem. Phya., 41, 827 (1964). (18) A. Weissberger, "Organic Solvents," Vol. VII, 2nd ed, Interscience Publishers, New York, N. Y . , 1955,p 227. (19) R. M. Fuoss and F. Accascina, "Electrolytic Conductance," Interscience Publishers, New York, N . Y., 1959,p 234. (20) R. L. Kay, C. Zawoyski, and D. F. Evans, J . Phya. Chem., 69, 4208 (1965). Volume 70,Number 6 June 186%
1962
J. J. BANEWICZ, J. A. MAGUIRE, AND P. S. SHIH
Table 11: Conductance of TEAP in Valeronitrile a t Various Temperaturesa Temp, 25"-
-Temp,
-Temp, l0'C
30"-
104c
A
104c
A
63.37 51.50 44.29 35.43 31.54 25.74 19.209 13,173 12.187 9.948 7,953 7.347 5.946 4.975 3.979 3.720 3.307 1.736 1.587
48.69 51.46 53.28 55.76 57.38 59.64 62.89 67.19 67.59 69.56 71.45 72.58 74.10 75.55 76.97 77.29 78.05 81.51 81.87
64.03 51.23 44.06 35.24 25.60 19,108 13.103 12,123 9.895 7.911 7.308 5.914 4 * 949 3.958 3.700 3 290
51.67 54.59 56.51 59.12 63.30 66.75 71.37 71.82 74.00 75.96 77.09 78.84 80.34 81.91 82.27 83.04
I
63.30 50.65 43.55 34.84 25.31 18.890 12.954 11.985 7.821 7.225 5.847 3.913 3.658 3.252 1.561
40°-
A
57.68 61.05 63.25 66.17 71.01 74.91 80.24 80.74 85.52 86.83 88.80 92.30 92.73 93.68 98.33
-Temp, lO'C
62.60 50.01 43.07 34.46 25.03 18.682 12.812 11.853 9.678 7.735 7.146 5.783 4.839 3.870 3.618 3.217 1.543
60°-
A
63.65 67.43 69.87 73.14 78.76 83.16 89.19 89.75 92.40 95.20 96.66 98-92 100.94 103.05 103.40 104.42 109.96
Units: C, M ; A, cma/ohm equiv.
Table I11 : Conductance Parameters and Constants J
25 30 40 50
88.3 93.93 106.3 119.1
rtO.l *0.09 f0.1 i0.2
5154 5722 6790 7716
3.8 i 0.4 4 . 0 i0 . 2 4 . 2 f 0.2 4 . 1 i 0.2
variation of the ion-pair association constant, K A ,with temperature. For salts with both ions electrically symmetrical, Fuoss derived the equation2I (3)
194 rt 4 202 rt 4 221 i 4 237 rt 5
0.6117 0.6094 0.6079 0.6055
Additional evidence for the lack of solvation can be obtained by comparing the values of the Walden product found in this investigation with those for TEAP in other solvents. Values of for this salt, calculated from literature data, in a number of organic solvents were found t o be between 0.60 and 0.65.118p25*a6The values reported in Table I11 are also in this range. This indicates fairly constant Stokes radii for the ions of TEAP in a variety of solvents. The small values of dJ shown in Table I11 are also consistent with the assumption of unsolvated ions. From the slope of the plot in Figure 1, a value of d of 3.3 f 1.0 A is obtained, and from the intercept a value of 2.6 f 0.3 A is obtained. These values of a are lower than those obtained from the values of J shown in Table 111. However, considering the indeter-
where D is the dielectric constant, d is the ionic size parameter, k is the Boltzmann constant, N is Avogadro's number, E is the electronic charge, and T is the absolute temperature. Equation 3 was derived assuming the solvent t o be a continuum, and the equation does not allow for solute-solvent interaction. Specific solutesolvent interactions can be accounted for by the introduction of additional terms to the exponential of eq 3.22 Over the temperature range studied in this work, eq 3 seems quite sufficient to account for the variation in K A with temperature. It should be pointed out, how(21) R. M.Fuoas, J. Amer. Chem. SOC.,80, 5069 (1968). ever, that over this limited temperature range, any (22) A. D'Aprano and R. Triolo, J . Phye. Chem., 71, 3474 (1967). nonlinearity due to specific solvation effects such as (23) W.R. Gilkerson, J. Chem. Phya., 2 5 , 1199 (1966). those proposed by G i l k e r ~ o nor ~ ~Beard and P l e ~ c h ~ ~ (24) J. H. Beard and P. H. Plesoh, J. Chem. SOC.,4879 (1964). could be detected only if these effects were quite large. (26) F. H.Healey and A. E. Martell, J. Amer. Chem. Soc., 7 3 , 3296 For a salt such as TEAP in a solvent such as valero(1951). nitrile, this does not seem likely. (26) J. F. Coetzee and G. P, Cunningham, ibid., 87, 2629 (1966). The J O U T T of M~ Physical ~ Chemistry
MEMBRANE POTENTIALS OF FUSED SILICA IN MOLTENSALTS minations in the values of d and the reported dependence of dJ on dielectric constant,27the values of d are in reasonable agreement. Although the Walden product is in the range reported for other solvents, from Table I11 it can be seen that there is a definite variation in this quantity with temperature. The variation of the Walden product for a particular ion in the absence of specific solvation has been considered by a number of investigators. Taking into account the additional frictional force produced as a result of the dielectric relaxation induced by ionic motion in a polar medium, an equation can be written relating the Walden product to such parameters as the static and optical dielectric constants and the dielectric relaxation time of the solvent. These solvent parameters have been measured for valeronitrile a t 30, 40, and 50” by Krishnaji and Mansingh.” Their variations with temperature predict a change in the Walden product, in the same direction and of about the same
1963
magnitude observed in this work. However, a quantitative application of this equation is not possible, owing to the lack of single ion conductivities in valeronitrile. Although the temperature range covered in this work is small, solvation complications seem to be minimal, and the variations in both K A and the Walden product can be accounted for with the simpler quantitative theories used in interpreting results from mixed-solvent systems. Therefore, valeronitrile, with its moderate dielectric constant and wide liquid temperature range (-97 to 142’) should be quite well suited as a solvent in which to investigate the variations in ion-ion interactions with temperature.
Acknowledgement. This research was supported by The Robert A. Welch Foundation of Houston, Texas, through Grants No. N-056 and N-142. (27) R. M.Fuoss, L. Onsager, and J. F.Skinner, J . Phys. Chem., 69, 2681 (1966). (28) R. Zwanzig, J . Chem. Phys., 38, 1603, 1605 (1963).
Membrane Potentials and Ion Selectivity of Fused Silica in Molten Salts by Kurt H.Stern Institute for Basic Standards, National Bureau of Standards, Washington, D . C . $0884
(Received June 26, 1967)
The effect of alkali metal cations on membrane potentials at the molten salt-fused silica interface has been studied using the concentration cell AglAgC1, M1C1, MzCl. . . (glasslMIC1,MzCl. . ,, AgCIIAg, where MI, Mp. . . are alkali metal cations. Vycor and several types of fused silica glass were used. The liquid junction (LJ), phase boundary (PB), and ion-exchange (IE) theories are examined for their applicability to the interpretation of the data. The IE theory is found to provide the most appropriate model, but the quantitative application of the theory requires data on ion exchange and ionic mobility in fused silica which are not yet available. The effect of anions on the membrane potential was studied with the cell MIMCl, NaCllfused silica(MBr,NaBrIM, where M = Ag, T1, and Pb. The results for the three metals are quite analogous. The emf of the cells could be calculated by straightforward thermodynamic methods, without the assumption of specific anion-glass interactions. The emf measurement of these cells at low sodium halide concentration can also serve as an analytical tool for sodium determinations down to mol %, but not lower,
Introduction The well-known utility of the glass electrode for pH measurements in aqueous solutions has recently led to attempts to account for the ion selectivity of various glasses in terms of ionic mobility in the glass, ionexchange equilibria a t the solution-glass interface, and glass composition. 1-8 If the factors responsible for the selectivity of the glass could be completely elucidated and controlled, it should be possible to construct glass electrodes responsive to a given ion irrespective of the presence of other ions in solution,
I n molten salts glasselectrodes would be equallyuseful.
If the melts in contact with the glass contain at least a few per cent sodium, the emf of galvanic cells with glass membranes can be accounted for by a liquid junction model in which the non-Nernst part of the emf (1) B. P. Nicolskii, Acta Physkochim., 7 , 697 (1937). (2) (a) G. Eisenman, Biophys. J., 2 , 269 (1962); (b) G. Eisenman, “The Electrochemistry of Cation-Sensitive Glass Electrodes,” in “Advances in Analytical Chemistry and Instrumentation,” C. N. Reilley, Ed., Vol. 4, John Wiley and Sons, Inc., New York, N. Y., 1965,pp 213-369. (3) F. Conti and G. Eisenman, Biophys. J . , 5 , 247 (1965). Volume 72, Number 6 June 1068