LI. On the Ultra-Violet Radiations emitted by Point Discharges. By JOHN THOMSON, M.A., B.Sc., Houlds worth Research Student in the University of Glasgow *. IN Introductory. N recent papers † it has been shown that ionizing electromagnetic radiations similar to those discovered by Wiedemann are emitted by the gas in the vicinity of metallic points charged to a high potential. That such radiations (or radiations of slightly longer wave-length) may be of first importance in determining the mechanism of the spark discharge has been suggested by J. Taylor ‡, and the present writer has described results which appear to support his theory. The experiments described in the present communication were undertaken with a view to obtaining further evidence regarding the nature of these radiations and their relation to the discharge. The experiments are of a preliminary nature, since, as far as the writer is aware, no attempt has so far been made to investigate any part of this region of the spectrum at pressures comparable with atmospheric. The results obtained, however, fully justify further study of the phenomena exhibited. Experimental Arrangements. The final form of the apparatus used in the investigation is represented diagrammatically in fig. 1. The tube containing the points E and F between which the discharge was to pass was cylindrical in form, the platinum electrodes being placed along the axis of the cylinder. The detecting and measuring apparatus was contained in a sidetube placed opposite one of the platinum points E, while the other two side-tubes were fitted with stop-cocks K, H. One tube was branched between the stop-cock K and the discharge-tube. The detecting apparatus consisted essentially of a photoelectric cell, the insulated electrode of which was connected to a tilted electroscope or to a Dolezalek electrometer. C and D were two half-cylinders of brass of length 20 mm., * Communicated by Prof. E. Taylor Jones, D.Sc. + C. E. Wynn-Williams, Phil. Mag. vol. i. p. 353 (1926); J. Thomson, op. cit. vol. v. p. 513 (1928). J. Taylor, Dissertation,' Utrecht (1927); id. Proc. Roy. Soc. A, vol. cxvii. p. 508 (1928). § J. Thomson, loc. cit. sealed to the walls of the tube so that their faces formed the sides of a slot of width 4 mm. A potential difference of 560 volts was maintained between C and D during the course of the experiments. This strong electric field. (X=1400 volts/cm.) was necessary in order to remove any ions which might drift or be projected down the slot towards the photo-electric cell. Theoretically the electric field was sufficiently strong to remove a simple positive molecular nitrogen ion projected into the slot with a speed of 3.10 cm./sec. Experiment showed that at pressures of 5 cm. of mercury and upwards no ions of any kind penetrated to the cell. B was a circular piece of fine-mesh phosphor-bronze wire gauze placed across the tube and insulated from C and D by two half-cylindrical plates of ebonite. This gauze formed one electrode of the cell. Pressed against it was the brass cylinder L which contained the other insulated electrode A. This electrode consisted of a small rectangular copper plate covered on the side facing the slot with a thin film of copper oxide. The latter, though less sensitive to ultra-violet light than other substances, suffers less from "photo-electric fatigue," and consequently could be relied upon to give results which would be comparable over a long period. This plate was placed at a distance of about 2 mm. from the gauze and soldered to a thin copper wire which was imbedded in pure paraffin wax and led down the axis of the cylinder L to the electroscope or electrometer. The latter was placed at a distance of about 2 metres from the tube itself. It will be observed, therefore, that the insulated electrode and the lead (surrounded by earthed tubing) from it to the measuring instrument were well shielded from effects due to induction. This was tested experimentally before any measurements were made. The insulation also was tested, and proved to be of high quality. Finally, the side-tube containing this apparatus was completely sealed at M with a mixture of beeswax and resin. The detecting cell therefore consisted of the gauze B, the cylinder L, and the insulated plate A, while the slot between C and D limited the radiation affecting it to a small pencil from the immediate vicinity of the point E. The connexions to C, D, and B were made by platinum wires sealed into the glass tube. The other side-tubes provided a means of controlling the nature and pressure of the gas in the discharge-tube. One tube was connected through a capillary to a gas container, while the other was connected to an oil-pump. The branch in the latter tube was sealed to a vertical tube dipping into mercury, so that the pressure of the gas could be measured almost directly at any time. Since pressure measurements did not require to be very accurate, such a gauge was sufficient for the purpose. All the connexions were sealed with the beeswax mixture which was found to be very satisfactory for work at pressures greater than 0.1 mm. of mercury. The potentials relative to the earth of the half-cylinders C and D, and of the gauze B and the cylinder L, depended on the particular nature and purpose of each experiment. When observations were being made with the tilted electroscope to measure the photo-electric or ionization current, the half-cylinder C, the gauze B, and the cylinder L were all at earth potential, while the plate A was charged to 12 voits. In the later experiments in which the Dolezalek electrometer was employed, C, B, and L were maintained at +250 volts, while the plate A was initially at earth potential. The current across the discharge-tube was supplied by a large induction coil (10-inch spark) used in conjunction with a motor mercury jet interrupter. Such an arrangement, when used to produce peak potential differences corresponding to spark lengths of less than half an inch in air at atmospheric pressure, gives a remarkably steady mean current through the secondary of the coil. This current was measured by a Gaiffe milliamperemeter, and could be controlled to 0.1 milliampere. In order to eliminate as fat as possible any inverse current through the coil, the battery E.M.F. in the primary circuit was kept as small as possible. Experiment indicated that the current at "make" was negligible. Four gases were employed in the experiments to be described-oxygen, carbon dioxide, nitrogen, and hydrogen,— and of these, only the two latter were used in quantitative investigations. The hydrogen and oxygen were prepared by electrolysis; the nitrogen was prepared by the action of ammonium chloride on potassium nitrite in concentrated solution, and was purified by passing it through concentrated sulphuric acid and over red-hot copper filings; the carbon dioxide was prepared by the action of dilute hydrochloric acid on marble chips. No attempt was made to dry any of these gases. The The instruments used to measure the ionization and photoelectric currents have already been mentioned. The tilted electroscope, being easy to adjust, was used in the preliminary investigations, but all the quantitative experiments were performed with the Dolezalek electrometer. suspension used in the latter was a fine quartz fibre which was thoroughly platinized by placing it near the platinumwire cathode in a low-pressure discharge. The conductivity of the suspension was thus permanently ensured, and it was verified that the sensitivity of the instrument remained constant. During the work to be described, the deflexion of the needle due to one volt potential difference between the quadrants corresponded to 960 mm. on the scale, which was 3 m. from the instrument. Ionization and Photo-Electric Currents in Different (1) Oxygen.-No results were obtained in oxygen on account of a peculiar phenomenon exhibited by the gas. As this phenomenon is one which might invalidate any ionization experiments carried out in this gas, it may be well at this point to describe it briefly. With the apparatus described in the previous section and the tilted electroscope as the detector of ionization, the normal conductivity of the gas at A was observed. This was very small. Then a spark discharge at atmospheric pressure was passed across the gas between E and F, a current of 1 milliampere flowing for about one minute. The discharge was then discontinued, and again the leakage of the electroscope was observed. It was found that the gas in the Phil. Mag. S. 7. Vol. 6. No. 36. Sept. 1928. 2 M vicinity of the insulated electrode A had become conducting, the rate of leakage of the electroscope being about one hundred times as great as before. The apparatus was now allowed to stand, the same gas was being left in the tube, and the rate of leak of the electroscope was read at halfhour intervals. The leakage gradually and consistently decreased, until, twenty-four hours after the passing of the discharge, the rate of leak was the same as that observed at the beginning of the experiment. On account of the strong electric field in the slot between the half-cylinders C, D, it was impossible for any charged particles to penetrate from the discharge between E and F to the plate A. Hence it must be concluded that the conductivity of the gas at A was caused by ionization which took place in the vicinity of A. This conclusion was verified by another experiment performed in a specially constructed tube, where all the ions directly produced by the discharge were immediately removed by a strong electric field. Spontaneous ionization of the gas continued to take place for some hours after the discharge had passed. It is suggested that this ionization accompanies the gradual change of the O, molecule (and perhaps others) formed during the spark discharge to the normal 0, molecule. It is immediately obvious that no experiments on the ionization produced by electromagnetic radiation could be performed in this gas with the apparatus described. (2) Carbon Dioxide.-The ionization and photo-electric currents in this gas were exceedingly small. At atmospheric pressure the current at A was comparable with the normal leakage of the apparatus, and consequently any observations taken would be of little value. Even at the comparatively low pressure of 30 cm. of mercury the currents were still small, and investigation of the effects in carbon dioxide was therefore postponed until a later date. (3) Nitrogen. This gas, when purified, showed no signs of the phenomenon just described as occurring in oxygen, but a trace of oxygen in the gas was sufficient to give an observable effect. The nitrogen was therefore purified and repurified until all trace of the effect had vanished. The first experiment was made for the purpose of obtaining a table of ionization and photo-electric effects from the vicinities of the anode and cathode in the discharge-tube. It was hoped that this table might be comparable with that given in a former paper* * J. Thomson, loc. cit. |