CLL-The Electrical Conductivity of Potassium Sodiuni, and Barium Chlorides in Mixtures of Pyyidine and Water. By JNANENDRA CHANDRA GHOSH. IN previous papers (T. 1918 113 449 627) it has been shown that in solutions of strong electrolytes the increase in molecular conductivity with dilution is given by the following equations : for univalent binary electrolytes and for uni-bivalent electrolytes where N is Avogadro’s number E the absolute charge on an ion ’Ir the molecular dilution and D the dielectric constant of the solvent. Equations (1) and (2) contain the term pN the molecular conductivity a t infinite dilution which cannot be determined experimentally. They can however be easily put in the following forms which contain no unknown magnitudes whatsoever : an The present investigation was carried out with the object of testing the validity of tho above relation between the dielectric constant of the solvent and variation of molecular conductivity with dilution.It has already been shown that the extensive experimental data of Walden on the molecular conductivity of tetraethylammonium iodide in various solvents can be accounted for (with the exception of the aldehydes) on the assumption that tetraethylammonium iodide a t first undergoes polymerisation and then the double mole-cule dissociates as a uni-bivalent electrolyte. The experimental difficulties in determining molecular conductivity in pure non-aqueous solvents are very great. The data of different observers do not very often agree a t all and it was therefore concluded that the agreement between the observed and calculated value of niole-cular conductivity was within the limits of experimental error.I n the present investigation only the chlorides of potassium, sodium and barium were used which unlike the iodide have very little tendency to complex-formation. Binary mixtures of pyridine and water were selected as they are completely miscible in one another and offer a very wide range of dielectric constants varying from 12 to 80. Preparation of Materials. Chemically pure pyridine was distilled several times over solid potassium hydroxide and the fraction distilling a t 115-116a was collected. This fraction was then allowed to remain overnight over pieces of metallic sodium as recommended by Hopkins (this vol.p. 280). Gas bubbles were evolved showing that moisture was present. The dry substance was distilled directly into a dry flask protected from moisture by a calcium chloride tube. The specific conductivity a t 1 8 O was 0.2 x 10-6 mho and the boiling point 115.6O. Water was purified by distillation from acid and alkaline per-manganates a block-tin condenser being used. It was finally distilled from a quartz distillation apparatus and stored in quartz flasks. The specific conductivity of the water was 1.2 x 10-6 mho a t 1 8 O . The mixtures of pure pyridine and water were also stored in quartz flasks. The salts were crystallised twice from pure samples and dried by heating in a platinum crucible. For experiments a t Oo a bath containing finely powdered ice was used jacketed by a second vessel containing ice.The tempera-ture varied from O0 to 0 . 0 5 O . For experiments a t 18O a thermostat was used the extreme variation of temperature being 0.02O. Fo 1392 GHOSH THE ELECTRICAL CONDUCTIVITY OF POTASSIUM, dilutions below 1000 litres a closed conductivity cell made of Jena resistance glass was used and for higher dilnt'ions a closed quartz cell made by the British Silica Syndicate. The bridge wire was carefully calibrated and the resistances were National Physical Laboratory standards of 1 10 100 and 1000 ohms. Dielectric Constants of Mixtures of Pyridine and Water. The Nernst method was employed as modified by Turner (Zeitsch. physikal. Chem. 1900 35 385). A Wehnelt interrupter was used in the primary circuit of the secondary coil; it consisted of two thin platinum wires dipped in dilute sulphuric acid.The anode could be raised up and down by means of a rack and pinion arrangement and a sharp fine tone obtained a t 16 volts by careful adjustment. The dielectric constant of pure pyridine a t Oo with chemically pure benzene as the liquid for comparison was found to be 12.5 from the equation S-S so -D = (Do - l)-s + 1, where Do is the dielectric constant of the known substance s the number of scale divisions the condenser plate has to be moved when the vessel is empty so the number for the substance of known dielectric constant and S for the unknown substance. For binary mixtures of water and pyridine pure pyridine was used as the comparison liquid and the dielectric vessel consisted of a platinum crucible and a platinum disk attached to the end of a platinum rod passing through a hole in the ebonite cover of the crucible.Although pure pyridine and pure water have themselves very little specific conductivity the conductivity of the binary mixture increases with increasing proportion of water and for a mixture containing 90 per cent of water it is as high as 10 x 10-6 mho a t 18O. The data for the dielectric constants of the mixtures contain-ing a large proportion of water are therefore not' very accurate as it is very difficult to get a sharp sound mininium in the telephone. Percentage by weight of pyridine in water. 96.0 93.8 so-0 67.0 60.0 40.0 30-0 TABLE I. Dielectric constant at 0". 13.0 15.3 22.9 27-7 40.0 56.4 68.3 Dielectric cons tmt at 18".12-5 16.0 21.0 25.1 37.0 52-3 64-SODIUM AND BARIUM CHLORIDES ETC. 1393 Jlolecidar C'onductiuity of Salts in Mixtures of Pyridine and 'CVater. Mixtures containing 80 or 60 per cent. by weight of pyridine do not dissolve the chlorides of barium potassium and sodium easily. A definite amount of salt was therefore first dissolved in a given quantity of pure water and pyridine then added. Barium chloride is not very readily soluble in these mixed solvents. Even in the case of N/40-solutions crystals generally separate on keep-ing overnight. The state of supersaturation can however be easily maintained for the few hours necessary for taking a complete set of readings a t various dilutions.The molecular conductivity of the solutions is rather small and hence the correction for the conductivity of the solvent becomes quite appreciable when the dilution is 80. The accuracy of the data for higher dilution depends on this solvent correction and it was therefore necessary to determine the specific conductivity of the solvents every few hours. In tables 11 111 IV V VI and VII are given the observed values of equivalent conductivities a t Oo and 18O. They are compared with those calculated from the observed value of A,, from equation (3) in the case of uni-TABLE 11. t=O0. Solvent containing 40 per cent by weight of pyridine. Salt. T7 = 10. 40. 80. 640. NaCl X obs. .................. 22.1 24.6 25.5 27.6 cdc. from A,, obs. = 56.3 22.5 24-8 25.6 27.4 'KCI X obs................... 25-6 28-3 29.3 31.4 X calc. from A,, obs. = 30.0 25.9 28.2 29-2 31-1 BaCl Xobs. .................. 20.0 23.6 - 27-7 X calc. from X,,,obs. = 36.1 19.6 23.4 - 28.0 TABLE 111. t = O 0 . Solvent containing 60 per cent. by weight of pyridine. 2560. 28-6 28.1 32-2 31-8 29.0 29.3 Salt. V = 10. 20. 40. 80. 640. 2560. NaCl X O ~ S . .................... 12.7 - 14.8 - 17.3 18.0 Xcalc. from XIGo obs. = 16.1 13-9 - 14.8 - 17.0 17.5 RCl Xobs.. . . . . . . . . . . . . . . . . . . . - 14.55 15.7 16.5 16-3 19.1 Xcalc. from h16 obs. = 17.2 - 14.80 13.7 16-5 18-2 18.8 13~C1 Xobs.. . . . . . . . . . . . . . . . . . . . 10.3 - 13-9 - 16.1 17-0 Xcalc. from hlGO obs. = 14-8 9.9 - 12.7 - 16.3 17.3 3 E 1394 GHOSH THE ELECTRICAL CONDUCTlVITY OP POTASSIUM, TABLE IV.t =oo. Solvent containing 80 per cent. by weight of pyridine. Salt,. V == 20. 40. SO. 640. 5170. NaCl Xobs . . . . . . . . . . . . . . . . . . . . . . . . . . . b.9 10.1 - 13.5 15.2 hCalC. from A,, obd. Z= 12.1 .... 9.4 10.4 - 13.2 14.4 X calc. from Xle0 obs. = 11.9 .... - 10.3 11.1 13.1 14.3 KCl Xobs ........................... - 20.1 10.8 13.8 15.5 TABLE V. t = 18'. Solvent containing 40 per cent. by weight of pyridine. Salt. V = 10. 40. 80. 640. 2560 NaCl Xobs.. ......................... 4G.4 44.6 46.5 50-1 51-S hcalc. from A160 obs. = 47.9 . . . . 40.9 45.1 46.7 49.8 51.1 KCI Xobs ........................... 46.8 51.8 53.8 57.8 59.4 Aca!c. from A,, obs. = 55.6 . . . . 47.4 52.3 54.1 57.8 59.2 BaCl hobs........................... 38.2 44.2 - 53.5 55.6 hcalc. from A,, obs. = 50-1 .... 37.5 44.8 - 53.8 56.2 TABLE VI. t = 18O. Solvent containing GO per cent. by weight of pyridine. Salt. V = 10. 20. 40. 80. 640. 2560. NaCl Xobs.. . . . . . . . . . . . . . . . . . . . 25.1 - 29.4 - 34.1 35.2 X cnlc. from XI, o h . = 32.0 25.6 - 29.5 - 33.6 34.7 KCl Xohs ..................... - 28.6 31.0 32.3 55.8 37-6 hcalc. from h,,o obs. = 34.1 - 29-4 31.3 32.5 35.9 37.3 BaC1 hobs.. ................... 19.6 - 24.9 - 31.4 33-2 hcalc. from A,, 011s. = 28.7 19.1 - 24.5 - 31.7 33.8 TABLE VII. t = 180. Solvent containing 80 per cent. by weight of pyridine. Salt. 17 == 20. 40. 80. 640. 6120 N&l lobs.. ......................... 17.2 19.5 - 26.3 29.9 Xcalc. from X16 obs.= 23.4 .... 18.1 20.1 - 25.5 27.9 KCl Aobs ........................... - 19-1 21.0 250 28.9 Xcalc. from obs. = 22.7 .... - 19.5 21-2 25.2 27.3 univalent salts and from equation (4) in the case of barium chloride. I n the tables V is the equivalent dilution in litres and A the equivalent conductivity SODIIJiVl AND BARIUM CHLORIDES BTC'. 1395 It will be noticed froni the above tables that the agreement between the observed and calculated values of equivalent con-ductivities are within the limits of experimental error except in the case of the solvent .containing 80 per cent. by weight of pyridine. Here the extreme deviations between the observed and calculated values are as much as 7 per cent. Perhaps in the mixed solvent rich in pyridine there is a slight complex-formation.Hartley Thomas and Applebey (T. 1908 93 538) have deter-mined the molecular conductivity of lithium nitrate in mixtures of varying proportions of pyridine and water. It appears interest-ing to exaniine whether their experimental data agree with those calculated from equation (3). The observed values of molecular conductivity in table V I I I are taken from their paper. TABLE VIII. t = 2 5 O . Salt Lithium Nitrate. Mol. per cent. of ppridine in solvent.. V = 16. 33. 64. 128. 256. 512. 1024. 133.98 XObS.. ............. 20.1 24.4 27.5 32.1 35.4 38.3 -15.8 ............ 19.3 23.3 27.5 31.9 36.1 40.1 - Xcak from obr;. = 73.76 hobs.. ............. 23.1 26.6 30.3 33.2 35.9 38.1 40.5 19.1 ............ 22.9 26.5 30.3 33.8 37.0 39.6 41.6 Xcdc.from obs. = 46.67 Xobs.. ............. 24.9 27.4 29.5 30.7 32.2 33.2 -21.5 ............ 24.4 26.6 28.4 30.0 31.2 31.9 -hcelc. from A obs. = 31-28 Xobs.. ............. 27.5 29.4 31.0 32.4 33.5 34.4 -25.2 ............ 27.4 29.1 30.4 31.4 32.2 33.5 -Xcalc. from A obs. = 5-71 Xobs.. ............. 60.8 63.1 64.9 66.3 67.5 68-5 69.5 58.1 ............ 61.0 63.3 65.1 66.7 67.6 68.2 68.5 X calc. from X obs. = It will be a t once seen that' the agreement between the observed and calculated values of molecular conductivity is quite satis-factory. It is peculiar that in the solvent containing a 46.67 molecular percentage of pyridine which is equivalent t o 79 per cent. by weight of pyridine the observed values of equivalent conductivity for lithium nitrate agree with the calculated values within the limits of experimental error.It is strange thai; tho simpler salts like potassium and sodium chlorides should show greater deviations. 3 E* 1396 THE ELECTRICAL COND UCTlVITY 04' PO'I'ASSIUAS ETC. Viscosity of Xixtures of Yyridinc U I L ~ Itrater. Nartley Thomas and Applebey (Zoc. cit.) and Dunstan and Thole (T. 1907 91 1728) have determined the viscosity of mixtures of pyridine and water a t Oo and 2 5 O . No simple relation like that discovered by WaIden between the molecular conductivity and viscosity of the solvent in solutions of tetraethylammonium iodide exists here. Following the lead of Bousfield Hartley, Thomas and Applebey conclude from the comparison of viscosity and conductivity data of lithium nitrate in mixtures of pyridine and water that there is a change in the size of the solvent atmo-sphere attached to the ions as the composition of the solvent changes.Although speculations in this field do not lead to tangible results it was thought advisable t o complete the investi-gation by determining the viscosities of the mixed solvents used in this invesstigation a t 18". For this purpose a Dunstan and Thole type of viscosimeter was used. The results are given in table IX. TABLE IX. Percentage of pyridine in the solvent mixture. relative to water at 18". 80 2.375 60 2.572 40 2.153 Viscosity of the solvent mixture It will be observed that the conductivity of the salts continually increases as the percentage of pyridine in the solvent diminishes from 80 to 40. The viscosity of the mixture of pyridine and water, however passes through a maximum when the percentage by weight of pyridine is approximately 65. Conclusion, The molecular conductivity of the chlorides of potassium sodium, and barium in mixtures of pyridine and water was studied. The dielectric constants of the solvents varied from 12 to 68; the experimental data confirm the hypothesis of complete ionisation of strong electrolytes as developed by the author. My best thanks are due t o Prof. F. G. Donnsn F.R.S. for his kind interest and help in providing the apparatus necessary for this investigation and to my friend Mr. J. N. Mukherjee. CHEMICAL LAB ORATORY, UNIVERSITY COLLEGE, LONDON