ON A METHOD FOR REDUCING DARK CURRENTS IN PHOTOMULTIPLIERS

It is known that dark currents in photomultipliers, which limit the sensitivity of experimental measuring photoelectric apparatus to weak light fluxes, are determined primarily by the thermionic emission of the photocathode. Numerous attempts to reduce the thermionic emission of the photocathode by choosing a favorable structure of its surface without lowering the sensitivity to light have produced only a limited effect. Such, in particular, are diffuse silver–cesium \(^{1,2}\) and antimony–cesium \(^{3,4}\) photocathodes.

Considerably greater possibilities for reducing the thermionic emission of the photocathode are provided by a favorable choice of its geometry. Conversely, an unsuccessful choice of the photocathode geometry may considerably complicate the experimenter’s work. Let us show this by an example. In photoelectric multipliers for highly sensitive photometers (spectrophotometers, stellar photometers, etc.), hollow photocathodes were formerly readily used \(^{5,6,7}\), on the assumption that an increase in the light absorption coefficient due to multiple reflection is accompanied by a lowering of the sensitivity threshold of the photomultiplier. However, the relatively small increase in the coefficient of useful utilization of the light flux as a result of using a hollow photocathode is accompanied by a considerably greater increase in thermionic emission, which nullifies this gain in sensitivity for very weak light fluxes.

The density of the current \(j\) of thermionic emission is determined only by the work function and the temperature of the photocathode, whereas the photoelectric emission is directly proportional to the flux \(F\) of light (of constant spectral composition), regardless of its distribution over the photocathode surface of identical integral sensitivity \(b\).

It follows from this that the sensitivity threshold of a photoelectric device with a photomultiplier for small light fluxes is determined by the ratio of the intensities of the photoelectric and thermionic currents from the photocathode surface, equal to

\[ K=\frac{F\cdot b}{jS}, \]

where \(S\) is the active surface of the photocathode.

The value of this ratio proves, for constant \(F\), \(b\), and \(j\), to be the greater the smaller the active surface of the photocathode. Consequently, when using a photomultiplier for weak light fluxes approaching threshold intensities, one should choose a device with the minimum photocathode surface on which the photometric image can be focused. Precisely this circumstance should make it necessary to refrain from using hollow photocathodes in photomultipliers employed for work with weak light fluxes.

For the same reason, in modern photomultipliers the active surface of the photocathode has been substantially reduced in comparison with a number of models manufactured earlier \(^{9,10}\). The size of the active surface of the photocathode in some photomultipliers has been reduced to \(1\ \mathrm{cm}^2\) or less. Further \(^{11}\) reduction of thermionic emission by still greater diminution of the active surface of the photocathode is possible only in photomultipliers for instruments in which narrow beams of light are used, readily concentrated on a small photocathode. Such instruments include spectrophotometers, stellar photometers, and others; similar devices make it possible to reduce the dimensions of the active surface of the photocathode to an area of \(0.1 \times 0.1\ \mathrm{mm}\).

Comparing the thermionic emission of such a photocathode with the thermionic emission of a hollow photocathode having an active surface of the order of \(10\ \text{cm}^2\) (at the same temperature and with the same work function of the photoactive surface), we obtain a reduction in thermionic emission by approximately \(10^5\) times.

Reducing the active surface of the photocathode proves expedient also in photoelements of the external photoelectric effect and in photoelectric counters, making it possible in the latter to use photoactive surfaces with a small work function.

Reducing the active surface of the photocathode gives a good result only if the geometry of the electric field at its surface does not lead to the creation of increased electric-field intensities that promote considerable cold emission. The latter is easily prevented by two known methods of reducing the active surface of the photocathode that are not accompanied by an increase in the electric-field gradient at the latter. One of them is based on using, as the photocathode, an electrode of considerable dimensions whose surface has a high work function, with the exception of a small activated “eye,” which is a small-sized photocathode.

Fig. 1.

The second method consists in using a photocathode with a considerable active surface, whose thermionic emission is not limited by the cathode itself. From the stream of electrons emitted by the entire surface of the photocathode, an electron-optical system cuts out the stream emitted only by a small region \(^{12, 13, 14}\).

A diagram of a photoelement in which the second method is used is shown in the figure. Electrons emitted by the photocathode \(1\) are focused by the cylindrical coil \(2\) onto the plane of the anode \(3\), which has at its center a small aperture \(4\), through which the selected electron beam passes into the electron multiplier \(5\). The coils \(6\) create a transverse field that shifts the electron image over the surface of the anode \(3\), in order to select the emitting point of the photocathode whose electron stream is directed into the multiplier.

Thus, for example, by projecting the analyzed spectrum onto the surface of the photocathode, one can, by sweeping the entire electron stream over the surface of the anode \(3\), successively direct into the photomultiplier the streams of electrons emitted by different elements of the photocathode under the action of light from individual spectral lines. In a similar way, the illumination of two neighboring elementary regions \(7\) and \(8\) of the photocathode can be compared; that is, comparative photometry of two illuminated surfaces can be carried out without optomechanical switching of the light streams.

L. G.

References Cited

1. S. Asao, Proc. physic. math. Soc. Japan 22, 448 (1940).
2. P. Görlich, Zeits. f. Physik 116, 704 (1940).
3. P. Görlich, Zeits. f. Physik 101, 335 (1936).
4. N. S. Zaitsev, ZhTF 9, 661 (1939).
5. S. F. Rodionov, ZhTF 9, 1180 (1939).
6. N. N. Pavlov, Proceedings of the Main Astronomical Observatory 59, series II (1946).
7. V. B. Nikonov and E. K. Nikonova, Izvestiya of the Crimean Astrophysical Observatory 9 (1952).
8. L. A. Kubetskii, Avtomatika i telemekhanika 1, 17 (1936).
9. W. E. Turk, Photoelectric. Spectr. Gr. Bull. No. 5, 100 (1952).
10. N. I. Chistyakov (ed.), Applied Electronics, L. A. Goncharskii, Photoelements, VINITI and P, Moscow, 1953.
11. A. N. Argue, Proc. Roy. Irish. Acad. A 55, 117 (1953).
12. P. Farnsworth, Trans. Frankl. Inst. 218, 411 (1934).
13. W. Heimann, Postarchiv 68, 120 (1940).
14. N. Schaefti, Helv. Phys. Acta 22, 225 (1949).