QD -NRLF P3 EXCHANGE The Conductance of Aqueous Solutions of lodic Acid and the Limiting Value of the Equivalent Conductance of the Hydrogen Ion BY HENRY C PARKER A DISSERTATION SUBMITTED TO THE FACULTY OF CLARK UNIVERSITY, WORCESTER MASS., IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY, AND ACCEPTED ON THE RECOMMENDATION OF CHARLES A. KRAUS CLARK UNIVERSITY 1922 EASTON, PA. ESCHENBACH PRINTING COMPANY The Conductance of Aqueous Solutions of lodic Acid and the Limiting Value of the Equivalent Conductance of the Hydrogen Ion BY HENRY C PARKER A DISSERTATION SUBMITTED TO THE FACULTY OF CLARK UNIVERSITY, WORCESTER MASS., IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY, AND ACCEPTED ON THE RECOMMENDATION OF CHARLES A. KRAUS CLARK UNIVERSITY J922 EASTON, PA. ESCHENBACH PRINTING COMPANY v>> "2 O THE CONDUCTANCE OF AQUEOUS SOLUTIONS OF IODIC ACID AND THE LIMITING VALUE OF THE EQUIVALENT CONDUCTANCE OF THE HYDROGEN ION Introduction While the conductance of aqueous solutions of the salts has been ex- tensively investigated and we now have fairly reliable data for these sub- stances to concentrations as low as 10 ~ 4 N or even lower, the corresponding data for the strong acids and bases in aqueous solutions remain very un- certain. Obviously, the limiting values of the equivalent conductance for the acids and bases are even less certain. This lack of accurate data for the acids and bases is largely due to the fact that at low concentrations conductance measurements with these substances are attended with difficulties inherent in the nature of the solutions in question. In measur- ing the conductance of a dilute solution it is necessary, on the one hand, that all the electrolyte present shall be in a known state and, on the other, that the specific conductance due to this electrolyte shall be determinable. Except in the case of salts of weak acids and bases, where hydrolysis inter- venes, impurities present in the water have little influence on the state of the salt in solution. At the same time, any impurities present are not mea- surably influenced by the presence of the salt, and accordingly the con- ductance due to these impurities may be corrected for by measuring the conductance of the water from which these solutions are made up. With solutions of the acids and bases, this is not the case. In view of the fact that the present investigation is limited to an acid, it is unnecessary to discuss the case of solutions of bases further than to state that the behavior of these substances is similar to that of the acids. If the impurity present in the water from which an acid solution is made up contains a base, or the salt of a weak acid, then the concentration of the acid will be influ- enced by the presence of this material; furthermore, the correction to be applied for the conductance due to the impurity cannot be determined by measuring the conductance of the original solvent. This difficulty was met with at a very early date. It was found that the equivalent con- ductance of an acid, instead of approaching a limiting value asymptoti- cally as the concentration is decreased, passes through a maximum after which it decreases the more, the lower the concentration. 1 It has been suggested that this result is due to the presence of a base or a salt of a 1 Arrhenius, Bihang Schived. Akad., 8, Nos. 13 and 14 (1884). Kohlrausch, Ann. Physik, 26, 161 (1885). Ostwald, /. prakt. Chem., 31, 433 (1885). Compare also Kohlrausch and Holborn, "Leitvermogen der Elektrolyte," Teubner, Leipzig, 1898, p. 92. 50783i weak acid in the solvent, as a result of which a certain amount of the hydrogen ion disappears on reaction with the base and the equivalent conductance is accordingly decreased. 2 This decrease will, of course, be relatively the greater, the lower the concentration of the acid. Thus far, however, no measurements have been carried out in which this source of error has been eliminated. Indeed, the error in the case of the acids may be traced to two sources: first, impurities present in the water used in making up the solution; and second, impurities dissolved from the containing vessels. The first of these sources of error has been recognized as mentioned above, but the second source of error appears not to have been considered heretofore. It is well known, however, that even the best of glass is appreciably soluble in acids and it is accordingly to be expected that experiments carried out in glass cells will be measurably affected at low concentrations owing to the solubility of the glass. In order to overcome these difficulties, therefore, it is necessary to carry out measurements with water which has been freed from all impurities and in cells which are resistant to the action of acids. In the present investigation, pure water has been pre- pared by a method previously developed in this Laboratory 3 and measure- ments have been carried out in a quartz cell. In selecting an acid for the purpose of this investigation, it appeared desirable, if possible, that the acid should be relatively strong in order to reduce the errors due to extrapolation and that it should be readily manip- ulated for the purpose of making up the solutions. For this reason iodic acid was chosen, since it is readily prepared in a pure state as a solid, in which condition it may be weighed and transferred to the cell, while at the same time it is but little weaker than hydrochloric acid. Preparation of Materials Iodic Acid. The iodic acid employed in this investigation was prepared from the best quality of iodic acid obtainable on the market. This acid was purified by repeated recrystallization. Considerable difficulty was at first experienced in the process of recrystallization owing to the fa^ct that, on evaporating an aqueous solution, the acid forms a supersaturated solution which is brought to crystallization with difficulty, even on inoculation. It was found, however, that when a small amount of nitric acid is added to the water solution and the water is extracted slowly in a desiccator, beautiful crystals of the acid are formed. A beaker containing sodium hydroxide was placed in a desiccator with sulfuric acid, together with the solution of iodic acid. The purpose of the sodium hydroxide was to prevent the concentration of the nitric acid from reaching too high a value. After precipitation, the crystals of iodic acid were washed with very dilute nitric acid and then dried over a solution of potassium hydroxide. This was for the purpose of removing any possible remaining traces of nitric acid. The last traces of moisture were removed over sulfuric acid. The final process of desiccation was carried 1 Whetham and Paine, Proc. Roy. Soc., 81A, 58 (1908). Paine and Evans, Proc. Cambridge Phil. Soc., 18, 1 (1914). * Dexter, Dissertation, Clark University, 1917. J. Am. Chem. Soc., 44, 2468 (1922). 5 out shortly before the iodic acid was to be used, since otherwise, as was found, a certain amount of the anhydride, I^Os, is liable to be formed. The process of recrystallization above described was carried out 4 times. The purity of the acid was tested by various methods. First, the acidity was de- termined by comparison with a solution of hydrochloric acid which had been standard- ized against silver by weighing as silver chloride. The average of three determinations gave for the normality of the acid the value 0. 10273 =*= 3 N. The sodiumhy droxide solution used compared with this standard acid, in two titrations, gave the factor 0.099193 =*= 11, weight burets being employed. On titrating the iodic acid against this solution of sodium hydroxide, the results given in Table I were obtained. TABLE I DETERMINATION OF ACIDITY OP IODIC Aero HIO 3 taken NaOH HIO 3 found Error G. G. G. % 0.74285 42.681 0.74307 +0.03 0.51119 39.318 0.51162 +0.08 0.85278 48.9185 0.85367 +0.016 The acidity was likewise determined by a new method which consists in precipitating the iodate with silver as silver iodate and weighing. Since silver iodate is somewhat soluble, the filtrate is saved and analyzed, iodine being liberated with potassium iodide and sulfuric acid and titrated against standard thiosulfate solutions. The results of this method are given in Table II. TABLE II DETERMINATION OF IODIC ACID AS SILVER IODATE HIO 3 taken AgIO 3 HIO 3 in filtrate Total HIO 3 found Error G. G. G. G. % 0.21224 0.33721 0.0027 0.21247 +0.04 0.15159 0.24019 0.00226 0.15168 +0.07 0.16769 0.26617 0.00192 0.16750 -0.11 The nitrate content was shown to be negligible by a colorimetric test, using phenol- sulfonic acid, the iodate being precipitated with silver sulfate. This test showed roughly 6 parts of nitrogen as nitrate per million of iodate. An attempt was also made to deter- mine the iodic acid by dehydrating to anhydride and weighing the difference, as well as by liberating the iodine by means of potassium iodide and estimating by means of standard thiosulfate solution. Neither of these latter tests proved entirely satisfactory, however, the former on account of the difficulty of perfect dehydration, and the latter owing to the large amount of iodine liberated. The former tests, however, in ail cases fell within the limit of experimental error and were thus satisfactory from that point of view. The best proof of the purity of the iodic acid, however, consists in the agreement of the values obtained with samples of iodic acid from the different crystallizations. The precision of the conductance measurements is far greater than that of any analytical method which is here applicable. The four runs in dilute solutions given below repre- sent results with three different crystallizations of the acid ; and, as may be seen, these agree very closely. Water. The water employed in this investigation was purified by the method de- veloped in this Laboratory. Ordinary water was distilled from an alkaline permanganate solution under the usual conditions, the first fraction of the distillation being discarded. The product obtained was then again distilled from a dil. alkaline permanganate solution in a special still of the type mentioned above in which carbon dioxide was removed by fractional condensation. In order to avoid contamination due to the surrounding atmosphere, distillation was carried out under a slight pressure of air which had previously been purified, as will be described below. The water employed in these measurements had a specific conductance varying between 0.09 X 10 ~ 9 and 0.5 X 10 ~ 6 . As will be seen below, with water having the higher value of the specific conductance, the influence of the impurities could be detected in the conductance values. Purification of Air. Pure air was necessary for the purpose of stirring the solution in the cell as well as for the preparation of the water as stated above. The air used for this purpose was purified by means of a special set of continuously acting towers, a sketch of which is shown in Fig. 1. The air enters through the tube A, carrying before it a column of the purifying solu- tion which is raised to the top of the tower and there projected over the beads with which the tower is filled. The air then passes down among the beads over which the solution is con- tinuously flowing. The chamber containing the beads has a length of approximately 50 cm. and a diameter of 3 cm. so that the rate of flow of the air through the purifying apparatus is relatively low. This method of purification was found very successful. In order to remove both ammonia and carbon dioxide, 3 towers were employed, the first of which was filled with a solution of sulfuric acid and the remaining 2 with solu- tions of sodium hydroxide. The necessary pressure for the air was obtained by means of an aspirator. Fig. 1 . Construction of air purifying towers. Measuring Apparatus Bridge. A drum-wound, slide-wire bridge with extension coils was used in measur- ing the resistance of the solutions. The bridge wire was calibrated by the Kelvin method, and by means of the corrections obtained it was possible to check resistance readings to better than 0.01%. Air condensers were employed to balance out the effects of induc- tance and capacity in the circuit. The entire apparatus was carefully shielded, all con- nections being lead covered and grounded and all measuring instruments being kept with- in a lead-lined case. Source of E.M.F. A Vreeland oscillator was employed as a source of alternating current. The oscillator was fitted out with a variable air transformer by means of which it was possible to regulate the voltage, while the frequency of the oscillator was adjusted by means of suitable capacities introduced in the oscillator circuit. In general, 4 combina- tions of capacities were employed and the fre- quencies corresponding to these were deter- mined by means of an oscillograph. Prints of the records thus obtained are shown in Fig. 2. The horizontal dashes appearing in the figure are due to the timing device. It is seen that the oscillator gave a perfect sine wave. Since Fig. 2. Oscillogram of Vreeland oscillator. the oscillograph had a frequency considerably above 10,000, any overtones which might have been present would unquestionably have appeared on the plate. 4 The four fre- 4 This result would appear to contradict the statement of Eyster [/. Am. Inst. Elec. Eng., 39, 889 (1920)] to the effect that the Vreeland oscillator does not give a sine quencies thus determined were found to be 1747, 1217, 986.5 and 489.5 per second. It was found that in most cases a frequency of 1217 gave the most satisfactory result, both as regards the accuracy of the readings and the distinctness of the setting. This fre- quency was, therefore, used practically throughout this work. Resistances. The resistance boxes consisted of Curtis-wound coils, ranging from 10 to 10,000 ohms capacity. These were calibrated against resistance standards which had recently been calibrated at the Bureau of Standards. In the older measurements, which were carried out with solutions contained in glass cells, the resistances had not been calibrated, for which reason the measurements are not as accurate as those later obtained. These earlier results, however, have not been used for the purpose of deter- mining the final conductance values. In the later measurements with quartz and Pyrex glass cells, which were begun in about September, 1919, and completed in June, 1920, calibration corrections were made. Two series of calibrations of the resistances were carried out, one in September, 1919, and another in May, 1921. All the coils below 10,000 ohms were found to be constant over this interval to better than 0.01%, but the 10,000 ohm coils showed a variation of 0.88%. Accordingly, in correcting the measurements, it was assumed that the resistance varied as a linear function of the time, and a correction applied which amounted to 0.044% per month. Four standard resistances of 10, 100, 1,000 and 10,000 ohms were employed. These were found to be mutually consistent. All calibrations were carried out with a direct current and galvanometer. With all corrections applied, it was found possible to obtain check measurements within 0.005%. The Measuring Cells. In all, 5 cells were used in measuring the conductance of the acid, in addition to an auxiliary cell of the pipet form which was employed for the Fig. 3. ' Construction of various types of cells employed. purpose of calibrating the measuring cells. Two of the measuring cells were constructed of soda-lime glass, and two of quartz, and a fifth cell was constructed of Pyrex glass. The cells are shown in outline hi Fig. 3. In this figure, V A and V B represent front and side elevations, respectively, of the glass cell which was employed in measuring the con- ductance of the dilute solutions. The electrodes of this cell had an area of approximately 1.5 sq. cm. and were placed 4 cm. apart. The body of the cell was approximately 60 cm. long and 5 cm. in diameter and had a volume of approximately 1 liter. The glass cell used for measurements at higher concentrations had electrodes of the same cross section as Cell V which were placed at a distance of 17 cm. from each other. This cell is shown as VI A and VI B in the figure. The height of this cell was approximately 30 cm. and its volume approximately 200 cc. The electrodes in all cells were platinized accord- ing to the method of Kohlrausch, and were then heated in a blowpipe flame until gray. In the same figure, II represents the large quartz cell employed in measurements with the more dilute solutions. This cell consisted of a Vitreosil flask having a volume a little over 3 liters. The electrodes had an area of 1.5 sq. cm. and were placed at a dis- tance of 2.5 cm. from each other. The electrodes were provided with a brace under- neath in order to guard against displacement. The electrode stems were constructed of glass, but the tube F, through which air was blown for the purpose of stirring the solu- tion, was constructed of quartz. A rubber stopper D served to hold the electrodes in position in the cell. The cell. was calibrated in order to determine the effect of the posi- tion of the electrodes, especially with reference to the level of the liquid. A curve of resistance readings for different heights of the liquid was made and the height corre- sponding to a minimum distance was chosen for the position of the electrodes. In this position a difference of 1.25 cm. in the level of the liquid was required to make a difference of 0.01% in the resistance. The quartz cell used for carrying out the measurements at higher concentrations is shown as IV A and IV B . Circular electrodes, having an area of approximately 1.5 sq. cm., were placed in a horizontal quartz tube of 30 cm. length and of 2 cm. diameter. The leads were introduced after the manner described by Kraus. 6 The form of the cell as here shown was so designed that its contents could readily be mixed by shaking without removing it from the thermostat. The capacity of this cell was approximately 1.5 liters. The construction of the Pyrex glass cell was similar to that of Cell IV. The Thermostat. The thermostat employed in these measurements was kept at a constant temperature within 0.002 by means of a steel-contained mercury thermo- regulator. In carrying out the measurements with the glass cells, the thermostat was filled with water, but in all the later measurements the thermostat was filled with kero- sene. At very low concentrations, where the resistance of the cell is large, capacity effects apparently are introduced when a water-filled thermostat is employed, and it was for this reason that kerosene was later employed in place of water. Balance. The balance used for weighing out the acid was a standard analytical balance especially adjusted for a sensitivity of 0.02 mg. It was found possible to obtain check weighings to 0.01 mg. The weights used were calibrated against a set of assay weights which had been standardized by the Bureau of Standards. Preliminary Measurements Cell Constant. The values of the cell constant throughout this investigation are based upon the measurements of Kohlrausch and Maltby 6 with solutions of potassium chloride at 18. The constant of the auxiliary cell, from which the constants of the other cells were determined by inter- comparison, was determined at 18, according to the method described in the preceding paper by the present authors. The cell constants as determined with a series of independent solutions are given in Tables III and IV. In Table III are given values of the con- * Kraus, U. S. pat., No. 1,093,997, 1914. 8 Kohlrausch and Maltby, Wiss. Abh. Phys.-Tech. ReichsansL, 3, 157 (1900). 9 stant for the standard cell I as determined at different times during the course of the investigation. TABUS III CONSTANT OF STANDARD CELL I Before Run I After Run V After Run VII 3.67125 3.67039 3.67041 3.67113 3.67098 3.67037 3.67108 3.67123 Av. 3.67117 3.67068 3.67039 It will be seen that there is a slow drift of these values with the time. A sufficient check, however, was kept upon this variation so that the cell constant as determined for the measuring cells may be relied upon. In Table IV are given the constants of the large quartz cell IT. the Pyrex glass cell III, the small quartz cell IV, the large lime glass cell V, and the small lime glass cell VI. TABLE IV CONSTANTS OF DIFFERENT CELLS USED IN MEASURING THE CONDUCTANCE OP TODIC Acm Cell II Cell III Before Run I Before Run II Before Runs III, IV, V After RunV 0.28616 0.285436 0.284748 0.284697 8.26770 0.28644 0.285379 0.284711 0.284796 8.26665 0.285391 0.284741 Av. 0.28630 0.285402 0.284733 0.284747 8.26717 Cell IV Cell V Cell VI Run VI Run VII 6.98116 6.99346 0.715824 6.35052 6.98186 6.99317 0.715403 6,35013 6.99278 Av. 6.98151 6.99313 0.715613 6.35033 The constants of Cells III, IV and VI were determined at 25, using the specific conductance given in the preceding article. The values of the constant of Cell II, which was used for the most accurate determina- tions in dilute solutions, are given as determined at different times. These determinations were made each time with a series of independent solutions made up in the cell and then inter compared with the standard Cell I. The resistance of the solution in Cell II was usually in the neighborhood of 400 ohms, although at one time the resistance of the solution was varied between 250 and 1300 ohms in order to determine, if possible, whether the cell constant varies as a function of the concentration. The measure- ments in the more dilute solutions, however, did not check well because 10 of various sources of error. 7 The cell constants finally employed in this investigation are the mean values given at the bottom of Tables III and IV. The actual measurements on iodic acid, and most of the measurements on the cell constants, were carried out at a temperature of 24.958 instead of 25. All data have been converted to the corrected temperature by factors which were determined by measurements made at the two tem- peratures with a series of solutions. Thsee factors are as follows: to convert the resistance of the potassium chloride solution from 24.958 to 25, multiply by 0.999202; to convert the resistance of iodic acid solu- tions from 24.958 to 25, multiply by 0.999434. In determining these factors, the solutions were carried back and forth between the tempera- tures in question several times. The readings agreed to better than 0.01 %. The values given are the averages of all the readings taken. Temperature. The temperatures of 18 and 25 were established by means of 2 Beckmann thermometers which had been calibrated against a platinum resistance thermometer. 8 A series -of comparisons were made over a range of temperatures in the neighborhood of the temperatures in question and the results were plotted in order to determine the readings at 18 and 25. The temperature interval from 18 to 25, which had been used previous to this calibration, was found to be correct to 0.01. The final calibration was found to agree exactly with the reading of a mercury thermometer which had been calibrated by the Reichsanstalt in 1910. Density Measurements. In order to reduce the concentrations to a volume normal basis, it was necessary to determine the density of iodic acid solutions. The results of these measurements are given in Table V and are shown graphically in Fig. 4. The results of Heydweiller and Groschuff, made at 18 and 0, respectively, and reduced to 25, are also shown on the figure. In correcting these results to 25, it was assumed that the density change of the solution between the temperatures in ques- tion is the same as that of pure water between the same temperatures. As may be seen from the figure, the three series of measurements check closely up to a concentration of about 0.5 N. It is probable that above 7 A further investigation, in order to determine whether or not the cell constant varies as a function of the concentration, has since been carried out by the author using a very accurate method of intercomparison. This investigation showed that the constant of this cell and, in fact, the constant of every cell investigated, varies appre- ciably with the concentration. The results will appear shortly in a publication from this Laboratory. 8 The thermometer in question is one which was in use in the Research Laboratory of Physical Chemistry of the Massachusetts Institute of Technology and which had been carefully calibrated by Dr. James A. Beattie. The author wishes to express his in- debtedness to Dr. Beattie for his assistance in connection with the calibration and also to Dr. Frederick G. Keyes, Director of the Research Laboratory. 11 this concentration the assumption made for the reduction of the results of Heydweiller and Groschuff to 25 does not hold. For the measure- ; x 1.16 y .12 / x ^x^ a x j] X ,O4 A DENSITY AT i5' = O