THE

LONDON, EDINBURGH, AND DUBLIN

PHILOSOPHICAL MAGAZINE

AND

JOURNAL OF SCIENCE.

[SEVENTH SERIES.]

JULY 1928.

I. The Complete Photo-electric Emission from Potassium. By Miss JESSIE BUTTERWORTH, B.Sc.*

THIS paper is concerned mainly with the negative

emission from potassium, but some account of experiments designed to investigate the possibility of a positive photo-electric emission from the metal will be given at the outset. Since Dember (Ann. d. Phys. xxx. p. 142, 1910) found that a cylinder surrounding a metal plate acquired a positive charge, when the plate was illuminated with ultra-violet light, a suitable field being applied, many experimenters have endeavoured to justify Dember's assumption that the phenomenon is due to the emission of positive photo-electric particles. Most of the later work seems to show that the effect is the result of the emission of electrons by metal parts of the apparatus which are not adequately screened from stray ultra-violet light. Du Bridge (Phys. Rev. p. 201, Feb. 1925) and E. J. Lorentz (Phil. Mag. vol. i. p. 499, 1926) in particular were unable to obtain a positive charging up when precautions against defective screening were taken. No observer, however, appears to have worked with the alkali metals, though, if the positive photo-electric effect does exist, it seems, a priori, probable that these metals would exhibit the effect to a greater degree than other metals known to be much less active as regards the normal effect.

* Communicated by Prof. William Wilson, F.R.S. Phil. Mag. S. 7. Vol. 6. No. 34. July 1928.

B

Experiments on the Positive Photo-electric Effect.

A glass bulb (fig. 1) 10 cm. in diameter, with two inlet tubes which terminated in a bottle-neck, was used. Sealed through the inverted end of the neck, as shown, was a filament of platinum wire, so that the loop of the wire was at the centre of the spherical bulb. Round the filament, but not touching it, was a spiral of platinum wire. Thus the filament was surrounded at the sides and bottom by a grid, and both grid and filament could be electrically heated. A fine

Fig. 1.

To Charcoal Tube

platinum wire was sealed through the bottom of the bulb to make contact with the potassium, which was to act as the third electrode.

A glass bottle containing clean potassium was allowed to slide down one of the side-tubes until stopped by a constriction. By means of the other side-tube the bulb was connected to liquid-air traps and a mercury-diffusion pump, backed by a "Hyvac" pump. After baking and pumping out, with the grid and filament glowing and the charcoal tube heated, the potassium was run over the constriction, which was then sealed off. This was done while the pumps were

working and the liquid-air traps were being cooled. After further pumping, the tube was sealed off and left overnight. Then the potassium was heated slightly once more, and run down to the bottom of the bulb to make contact with the platinum wire.

The glowing filament acted as a source of light, and in so doing gave off thermions, both negative and positive. The latter emission would, of course, almost disappear with constant glowing and a field helping the emission. If a positive photo-electric effect did exist, there would be positive and negative photo-electric currents from both the platinum grid and the potassium. When the connexions are made, as shown in fig. 2, the negative current from the grid and the positive one from the potassium alone can affect the quadrant electrometer.

The grid is negative with respect to the potassium, so that any electrons from the potassium are turned back, while any positive particles are attracted across to the grid, leaving a negative charge on the potassium, which, when the key K is open, causes a deflexion of the needle of the electrometer. The field between the filament and grid will turn back any electrons emitted by the filament, while any positive particles from the same source will be turned back once they have passed through the grid. Hence the emission from the filament cannot affect the electrometer. There may be a negative photo-electric current from the grid, however, which will reach the potassium and charge it in the same manner as the positive current leaving it would do. Any positive photo-electric current from the grid will be turned back. The highest threshold frequency of potassium is known to be well in the infra-red portion of the spectrum, while that of platinum is in the ultra-violet. Consequently,

if no effect is produced upou the electrometer, using light whose wave-length is between the two threshold frequencies, it will mean that no positive photo-electric current is emitted from the potassium-at any rate, when those particular wave-lengths are employed.

The sensitivity of the electrometer was such that with 120 volts on the needle and a potential difference of one volt between the quadrants the deflexion produced would be 140 cm. on the lamp scale. The capacity of the system, the electrometer, and the bulb in series was very nearly 100 cm.

Owing to a slight charging up of the sealing-wax key and also to a small leak across the inside of the glass bulb, even when the filament was unheated, the image moved in the direction indicating a positive charging up of the potassium. Fortunately it was not too fast to be timed, and was constant over a short range of time-long enough for measurements to be taken. The effect looked for would mean slower rate of charging.

a still No change in the rate was noticed until the filament current rose to a value of 18 amps., corresponding to a temperature of over 800° C., and it seemed reasonable to suppose that the negative platinum current was causing the change at this temperature. However, at this temperature the negative current was quite large (nearly 10-6 amp.). A stop-clock reading to 2 sec. was used, and it would have been easy to detect a change in the rate of 1 cm. in 10 secs. The difference of potential between the potassium and the grid was 96 volts: sufficient to saturate the negative currents at the temperatures employed.

For

A change in the rate of charging up of 1 cm. in 10 secs. would mean a photo-electric current of 10-13 amp. the potential difference

Hence a current of little over 1x 10-14 amp. could have

been detected.

Dember found the ratio of his negative to positive photo

electric currents to be

10-9

10-13, i. e. 104, whereas these experi

ments give the ratio as more than

10-6 10-13,

i. e. 107. Hence, if the positive photo-electric effect does occur, it is at least a thousand times smaller than the current which Dember attributed to the emission of positive ions by photo-electric action.

The Threshold Frequencies of Potassium.

Richardson and Young (Proc. Roy. Soc. ser. A, vol. cvii. p. 377, 1925) found for a normal potassium surface a threshold frequency corresponding to a wave-length of 7000 A.U. After sensitizing the potassium surface by passing a glow-discharge through an atmosphere of hydrogen, the critical wave-length appeared to be somewhere between 9000 A.U. and 10,000 A.U. A usual thermionic threshold at a temperature of 200° C. is 10,000 A.U., and another occurs at about 30,000 A.U. No trace of a photo-electric threshold of this magnitude, however, was to be found, even for the sensitized surface. The following account gives details of the production, by vaporizing the metal, of a sensitized surface, for which two threshold frequencies were found.

Experiments on the Threshold Frequencies of Potassium. The apparatus used to investigate the existence of the positive photo-electric effect was also used to measure the threshold frequency or frequencies of potassium, by measuring the complete photo-electric effect when in equilibrium with black-body radiation of known temperature. By heating the potassium at the bottom of the bulb with a small electric heater, it was vaporized and deposited in a thin film over the cool part of the spherical surface. To prevent the vapour condensing on or near the filament and grid, both