LIBRARY UNIVERSITY OF CALIFORNIA. RECEIVED BY EXCHANGE Class THE CONDUCTIVITY OF AIR CAUSED BY CERTAIN CHEMICAL CHANGES THESIS Presented to the Faculty of Philosophy of the University of Pennsylvania In partial fulfilment of the requirements for the degree of Doctor of Philosophy S ,irtHA7*7 f UNIVERSITY } BY FAXXY COOK GATES 1909 THE CONDUCTIVITY OF AIR CAUSED BY CERTAIN CHEMICAL CHANGES. THE ionization of gases is the accompaniment of a large number of phenomena which are apparently of totally different character and origin. Many of these phenomena have been studied with consider- able care and have already thrown much light on the nature of elec- trical charges, and have brought forward more satisfactory hypotheses regarding the constitution of matter than had previously existed. Of other phenomena with which seem to be associated the ionization of gases, very little is known; and particularly has this been the case when the ionization occurs during chemical changes in neighboring substances. G. Le Bon first noted that when the sulphate of quinine was heated or cooled, the air about it was made a conductor of electricity and that at the same time a phosphorescent glow was observed over the entire surface of the quinine. He attributed these effects to the hydration and dehydration known to accompany the temperature changes during which the effects were noted and advanced this as an argument for believing that the conductivity of gases in the presence of radio- active substances, is due to chemical processes resulting from tempera- ture changes within the body. For the purpose of refuting or establishing this deduction, the author made a comparative study 1 of the radiations from quinine sulphate under the conditions described, with those from thorium and radium, and found the two to be wholly unlike; the former showing every indication that the discharge of the air condenser used, was due to the ionization of the gas, while in the case of the radio-active substances, the discharge was produced by the projection of charged particles from the substance. The exact effects of cooling quinine through different intervals of temperature were carefully noted with the result that the greatest conductivity of the air in each case was found to take place during the temperature change and rate of change, in which hydration was a maximum. LeBon makes the statement that except for the sulphate of cin- chonine, a substance very similar to quinine, no substance could be found which gave similar effects; but he does not tell what substances were tested and his statement is therefore of little value. Kalahne made an interesting study of the quinine radiations, described in the Annalen for November, 1905, in which he concerned 1 Physical Review, May, 1904. 202475 himself with testing the relative effects during heating and cooling; the relative conductivities produced in air, hydrogen and carbon dioxide by the same amount of quinine; the maximum amount of water which a unit mass of quinine sulphate can hold; and the effect of vapor pressure and temperature on dissociation pressure. He also found, as did the author, that cooling dry heated quinine in dry air gives neither ionization nor light, but that both begin with the intro- duction of damp air. The work described in the present paper was undertaken with the hope of ascertaining more facts regarding the cause as well as the nature of the quinine radiation; whether it is always accompanied by phosphorescence and if so, whether the electrical effect may not be due to it, rather than to the hydration or dehydration alone; and whether other substances can be found which give similar effects. An abstract of some preliminary observations appeared in the Physical Review for January, 1906. Since then, the work has been continued in the Cavendish Laboratory, and the writer takes this opportunity of expressing her appreciation of the courtesy extended by Professor Thomson in placing the facilities of the laboratory at her disposal and of his interest shown during the progress of the work. The question as to what effect, if any, the phosphorescence may have upon the conductivity of the air is a most important one to con- sider. If the phosphorescence is the cause of the conductivity, one reasonable explanation would seem to be that the latter is due to the presence of short ultraviolet light waves accompanying the phosphor- escence, such as Lenard has shown to cause ionization when they strike a negatively charged surface. The only apparent difference between this case and that which Lenard describes is that with quinine, the ionization takes place whether the plate on which the quinine rests, is charged positively or negatively. This seemed to exclude the ultraviolet light theory, until it was found that the intensity of the ionization current was always greater when the quinine was placed on the positive rather than the negative plate of the condenser. This fact would tend to explain the effect by considering that when the lower plate is positive the light strikes the upper negatively charged plate and ionization follows; when, however, the upper plate is posi- tive, the ionization is produced only after the light has been reflected by the upper to the lower negatively charged plate, in which case some of the energy would be necessarily dissipated. The ratio of the intensity of the ionization current in these two cases when the plates are separated by 2 cm. of air is 1.6, and increases rapidly as the air space is increased, being 2.5 when the plates are 7 cm. apart. This is what would be expected if the above explanation is the correct one. Le Bon declared that the conductivity of the gas could not be due to the phosphorescence since the latter persists longer than the former. The author found this statement not true, but that the ionization con- tinued long after the phosphorescence had ceased to be visible, when tested near the surface of the quinine. This might be equally well claimed as an argument against the ultraviolet light theory unless we can assume the presence of the short ultraviolet waves after the longer visible waves have ceased, both, however, having a common cause. Much stronger would be the evidence against the theory if conditions could be found under which ionization takes place without a trace of phosphorescence, or at least conditions in which a variation in the magnitude of one effect is not accompanied by a corresponding change in the other effect. Tests were first made to determine what effect a reduction of pressure would have on the ionization current. That entire exclusion of air and moisture would cause the effects to cease was to have been expected, and the behavior at merely comparatively low pressures was, therefore, of greatest interest. For these observations two types of testing chambers were used : with one, the quinine was first heated in the outside air and cooled under cover at lowered pressure; while with the other, the heating took place inside the testing vessel itself and at a low pressure. The first apparatus was made as follows: A circular block of ebonite 1 .5 cm. thick was used as a base, and on its upper surface at regular intervals were glued three small blocks to sup- port the metal plate, over which the quinine was spread. A short stiff wire was soldered to the lower side of this metal plate, and was just long enough to dip into a few drops of mercury contained in a small cup- like depression in the centre of the base. In this way, by means of a fine sealed wire piercing through the base, electrical contact was made between the terminal of a battery without, and the plate on which the quinine was spread, w r hen the latter was under cover. The upper por- tion of the apparatus consisted of a cylindrical brass cover 10 cm. in diameter and 5 cm. high. A metal plate the size of the one on which the quinine rested, was screwed on the end of a brass rod piercing an ebonite stopper in the centre of the cover. By means of this rod, this plate could be raised or lowered to any desired distance above the sur- face of the quinine. To the upper outside end of the rod was attached a wire connecting it with a C. T. R. Wilson electroscope. The cover to the above apparatus was made to fit over rubber packing in a circular trough cut in the ebonite base. After heating the plate containing the quinine and placing it in position in this chamber, weights were put on top of the cover and mercury was poured into the outside of the trough to insure an air-tight apartment. (Fig. 1.) One-half of a gram of quinine sulphate was sifted uniformly over the metal plate of the above apparatus and heated to 180 C. in an air bath. It was then quickly placed in position in this apparatus and joined to the positive terminal of a 2CO-volt battery. \Yith the upper plate 2 cm. distant and joined to an electroscope in which a deflection of one division per second was found to be caused by a discharging current of the order of 10~ 13 amperes, when tested directly in con- FIG. 1. To Electroscope Exhaust To Battery nection with a condenser of known capacity, the following readings were obtained from the quinine at atmospheric pressure. Curve A shows the current-time variation for the same. % READINGS FOR CURVE A, ATMOSPHERIC PRESSURE. Time after heating. 1 min. 30 sec. 4 27 52 14 40 7 4 Current in arbitrary unit. Began 37.5 75.0 75.0 43.0 29.0 20.0 12.5 Time after heating. 5 min. 35 sec. 6 7 8 9 10 12 17 9 22 30 46 20 Current in arbitrary unit. 10.0 7.0 5.5 4.5 3.4 3.0 2.6 EFFECT OF PRESSURE. With exactly the same conditions except that the cooling took place under reduced pressures of 10 cm. and 3 cm. respectively, curves B and C were obtained. (Fig. 2.) When lower pressures were used, however, it was found that the ionization current rapidly decreased. This is presumably due to the fact that at lower pressures the remain- ing air in the chamber contained insufficient moisture to produce the normal effect. FIG. 2. 3o It is of interest to note that except for the time at which the ioniza- tion current is first detected, curves B and C are almost identical with those produced at atmospheric pressure, when the distances between plates are greatly decreased. Thus, it would seem that the strength of the ionization current depends upon the amount of air between the plates, and can be increased by either decreasing the distance between plates or by lowering the pressure. The difference in behavior at the beginning of the effect may be explained by considering either that the lower pressure changes the rate of cooling, or that in the case of the rarer air, a longer time is required for the quinine to receive sufficient moisture to start the effect. To test the effect of pressure on the rate of cooling, one terminal of a thermopile was placed between the plates, while the other terminal was kept without in a constant temperature bath. Differences in temperature within and without were measured during the cooling of quinine at atmospheric pressure and again at a pressure of 3 cm. respectively. Although the galvanometer was sufficiently sensitive to record changes of .05 C., the rate of cooling was found not to be appre- ciably different in the two cases, it requiring ten minutes for the tem- perature within to be reduced to that of the room outside. In order to determine whether the effects on the ionization brought about by changes in pressure were in each case accompanied by a similar change in the phosphorescence, a mica window was inserted in the side of the apparatus and all of the above described tests were repeated in a dark room. These tests showed that with every decrease in ionization there was a corresponding decrease in phosphorescence, and vice versa. In all of these experiments the heating of the quinine took place in a different air chamber from that in which it was finally tested. In order to determine what effect was due to the contact of the hot quinine with the outside air during its removal from the oven to the testing vessel, experiments were made similar to those already described, except that in these, the heating took place within the testing vessel itself. For this purpose, the above apparatus was modi- fied by the introduction of two openings in opposite sides of the cylin- drical cover. These openings were plugged with ebonite stoppers through which stout copper rods were inserted, holding in place a thick sheet of platinum foil to which each was welded. The quinine was sprinkled over the sheet of platinum and a heating current was sent across it. In the case when the pressure was lowered before the heating took place, the conductivity, as shown by the discharge of the electroscope, took place as before; but no effect was obtained upon cooling, in the case when the exhaustion had been in progress during the heating. This again, might have been predicted if we had regarded the ionization as due to hydration, since in the first case, the moisture thrown off during heating was re-absorbed upon cooling, while in the second case, it was pumped out as soon as it was given up by the quinine. EFFECT OF A MAGNETIC FIELD. Tests were made for the purpose of determining whether the pres- ence of a magnetic field would change the ionization current. The testing vessel was placed between the poles of a powerful electro- magnet and the following readings taken during the cooling of quinine under 7 cm. pressure, with and without the presence of the magnetic field. Time. Current. 11 hr. 7 min. 35 sec. 60 8 30 42.2 8 50 37.50 9 50 30 12 30 17 Magnetic field. Off On Off On Off These readings give the following current-time curve, the readings taken while the magnetic field was present being marked o, and those without the magnetic field, x. It is apparent that all lie on a continuous curve and, therefore, no effect can be noticed, as a result of the presence of the magnetic field. (Fig. 3.) FIG. 3. so 7 Time /Z, VARIATION OF CURRENT WITH VOLTAGE AT Low PRESSURE. It has been found that at atmospheric pressure the ratio of current to potential difference between plates remains constant for all values of the potential difference between ICO volts and 1000 volts, thus showing no indication of a saturation current within that range. In order to ascertain whether this remains true at low pressures, the s potential difference was varied from 40 volts to 600 volts during the cooling of quinine at 15 mm. pressure with the following results: Time. 10 min. 40 sec. 11 30 12 12 40 14 20 15 16 55 17 45 19 55 23 Potential Current. Difference. 150 300 70 200 100 180 37 60 100 35 80 600 40 200 80 600 40 80 600 80 Time. 25 min. 10 sec. 26 10 27 .50 29 35 30 55 33 00 33 .50 34 35 38 45 40 Potential Current. Difference. 60 46 46 15 5 25 30 3 20 12 600 200 600 80 40 200 600 40 600 200 By plotting four curves for current-time variations, with values of potential difference of CCO, 200, 80, and 40 volts, respectively, the following are obtained : Voltage. 600 200 80 40 The results on a current voltage diagram, show that after a voltage of about 180, the current ceases to increase with voltage at its original rate and gives indication of a saturation value. (Fig. 4.) FIG. 4. After 12 min. After 17 min. After 23 min. Average Curent . Ratio. Current. Ratio. Current . Ratio. Ratio. 272 100 144 100 76 100 100 220 80.8 120 83.3 63 82.8 82.3 130 47.7 69 47.9 35.0 46.1 47.2 70 25.7 37 25.7 20.0 26.3 25.9 Coo DIRECTION OF ELECTRIC FIELD. No ionization current could be detected without the presence of an electric field, and it was further found that it made no difference whether the quinine plate was attached to the battery and the upper 9 plate to the electroscope, or vice versa, providing the direction^ electric field remained constant. As regards the direction of the field, it has already been shown that a larger current is produced when the lower plate on which the quinine rests, is the positive one. This would seem to indicate that the phosphorescence may have a direct influence on the current. Observations were therefore made in which the upper plate was of different materials in the different tests. Lenard has shown that in such cases when the ionization is due to ultraviolet light, the current is greatest when the light falls on a negatively charged zinc plate; but in the case when the charging is due to the quinine radiations, the intensity of the ionization was found to remain the same, whether the upper plate was of zinc, brass, aluminum, or of brass covered with lampblack. The accompanying table shows these values obtained when the distance between the plates was 1.8 cm. and at a potential difference of 200 volts. An upper plate with its surface covered with a soap film was also found not to alter the effect. In each of these cases the ratio of the current with the lower plate positive to that with it negative remained constant, being 1.4 for a P. D. of 2CO volts. Brass. Zinc. Lampblack. Time. Current. Time Current. Time. Current. 1 min. 30 sec. Began 1 min, 35 sec. Began 1 min. 30 sec. Began 1 50 60 2 00 62 2 00 60 2 30 80 2 52 86 2 20 75 3 10 75 3 22 75 3 20 75 3 50 65 3 47 67 3 45 60 4 20 45 4 30 43 4 45 40 5 00 35 5 45 30 o 40 30 5 50 29 6 30 25 6 15 27 7 00 23 8 45 16 7 15 22 8 . 10 19 10 22 14.3 8 00 17.50 00 14 13 15 9.5 10 10 12.50 1 30 11 12 00 10 That there is no ground for belief that the ionization is directly due to ultraviolet light waves was further proved by the fact that pho- tographic plates exposed to the phosphorescent light after the latter had passed through a quartz plate, showed no difference from those obtained when the quartz was replaced by glass of the same thickness. Observations of the effect of cooling in damp and in dry air at atmos- spheric pressure, showed that both the ionization and the phosphor- escence were hastened and intensified by the presence of damp air. Unheated quinine which had been kept for some time in an air-tight chamber containing phosphorus pentoxide, caused ionization and phosphorescence as soon as it was exposed to the outside air. The results so far obtained seem to show conclusively that since ionization cannot be produced or altered without producing or altering 10 the phosphorescence and since the latter does not give evidence of ultraviolet light waves, the two phenomena are taking place inde- pendently of one another, but are produced simultaneously by the same cause or series of causes; and that each is produced during FIG. 5. /oo 9o It 7o It. O fo o 3o /c t s (> 7 2 <} ID // /z Time in minutes O Brass X Zinc Lampblack hydration and dehydration in the case of quinine sulphate. In order to ascertain whether either effect could be detected during the forma- tion of quinine sulphate from its alkaloid, the latter was sprinkled over a sheet of asbestos soaked with sulphuric acid and examined in the testing chamber and also in a dark room, but neither ionization nor 11 phosphorescence could be detected. Observations were made with quinine mixed with quicklime, with salammoniac and with the sulphate of sodium, all of which affect the solubility of the sulphate of quinine in water. The effects in each case were not different from those which one would expect if mixed with any inactive matter. TESTS WITH OTHER SUBSTANCES. Tests similar to those made with the sulphate of quinine, were made with many of the more common substances known to hydrate easily, but none of them was found to give similar effects. Among the sub- FIG. 6. Quinine Sulphate Anthracene Grape Sugar Time 12 stances tested were calcium chloride, calcium sulphate, zinc sulphate and copper sulphate. Other substances tested for other reasons were potassium hydroxide, sodium hydroxide, barium oxide, copper nitrate, lead nitrate, grape sugar, aceto-acetic ether, fluorescine, eosine, anthra- cene, hydrocollidin-dicarboxylic ether and sesculin. Although none of these gave effects comparable in magnitude to that resulting from the sulphate of quinine; sesculin, grape sugar, and anthracene were found to produce a slight ionization of the surrounding air when cooled be- tween the plates of a condenser. In order to show the relative effects produced by quinine sulphate, anthracene, grape sugar, and sesculin, equal quantities of each were heated to the same temperature (120 C.) and cooled between the plates of a small condenser with an air space of 1 cm. and a P. D. of 600 volts between the plates. Ionization curves for each plotted according to the same scale are given side by side in the accompanying diagram. (Fig. 6.) The effect from grape sugar like that from the sulphate of quinine might be attributed to hydration, but there seems to be no reason for attributing the anthracene effect to that cause. Both substances, however, are known to be relatively unstable and capable of presenting themselves in slightly different molecular forms, and it is the author's belief that the effect from them as well as from the quinine, is due to some change in the linkage of the molecule in each case. This varia- tion in linkage may itself persist for only a short time, after which the molecule may return to its former stable condition, and thus the effect may be incapable of chemical analysis; but it may nevertheless be real while it lasts and be accompanied by a temporary ionization of the surrounding gas. UNIVERSITY OF CALIFORNIA LIBRARY, BERKELEY THIS BOOK IS DUE ON THE LAST DATE STAMPED BELOW expiration of loan period. - 20m-ll,'20