DETECTION OF INDIVIDUAL IONS, ELECTRONS, AND PHOTONS USING AN ELECTRON MULTIPLIER

The electron multipliers developed by L. A. Kubetskii, P. T. Farnsworth, and V. K. Zworykin are used chiefly as indicators of very weak fluxes of primary electrons. In the USSR these devices are being successfully applied in spectrophotometric investigations, in stellar photometry, for monitoring the concentration of dissolved substances in colored solutions in certain industries, etc. In other words, the devices that have come into practice here are exclusively those in which the primary electron flux is produced by the action of light, i.e., photoelectron multipliers. Meanwhile, the possibilities of using electron multipliers are far from exhausted by this. In the USA, for example, an interesting and successful attempt was recently made to use this device as an indicator of individual elementary particles—positive ions, electrons, and quanta of γ-rays[^1].

The possibility of detecting individual elementary particles is determined by the sensitivity threshold of the electron multiplier. This parameter of the multiplier depends on the intrinsic noise generated by the device, and has not been defined in a sufficiently universal measure. For photoelectron multipliers, the sensitivity threshold is usually taken to be that luminous flux which gives at the output of the device a signal equal to the dark-current signal, i.e., the signal in the absence of illumination.

The dark currents of multipliers are composed of several components, namely:

1) ohmic leaks in the output circuit of the multiplier (in the collector circuit);

2) cold electron emission of the electrodes;

3) the so-called “ion feedback,” a phenomenon consisting in the fact that (when the vacuum is insufficiently high, in particular due to the presence of alkali-metal vapor) the electron flux ionizes the gas in the region of the last cascades, and the positive ions thus formed knock new electrons out of the preceding cascades;

4) the so-called “optical feedback,” when bombardment of the last cascades by the electron flux produces luminous phenomena on them and this light releases electrons at the photocathode or on the first cascades;

5) thermoelectron emission of the photocathode or emitters.

Whereas factors 1), 2), 3), and 4) can comparatively easily be eliminated by appropriate design changes (improving the insulation of the collector, proper arrangement of the electrodes to prevent the formation of large electric-field gradients that would impede the motion of positive ions and light quanta from the last cascades to the first), factor 5) is of a more fundamental character, since it is determined by the thermionic constants of the emitter materials and especially of the photocathode. In photoelectron multipliers with oxygen–silver–cesium cathodes at room temperature (~20° C), the sensitivity threshold apparently cannot be made lower than \(10^{-9}\) lumens. When the cathode temperature was lowered to −30° C, it was possible to reach a threshold of \(10^{-11}\) lumens.

In accordance with this, in the multiplier intended for the detection of individual elementary particles, beryllium exposed to the action of oxygen (air) was chosen as the emitting material. This material has a high work function (~4 V) and at the same time a large secondary-emission coefficient (up to 3 at 300 V)[^2].

Another advantage of beryllium emitters is the very high yield of secondary electrons per single positive ion (proton), which can reach eight[^3].

In its design, the multiplier (with 11 cascades) belonged to the type of electrostatic multipliers. A diagram of its arrangement is shown in Fig. 1, where the layout of the entire installation for counting elementary particles is also depicted. A special feature of the design was a very

careful insulation of the collector [elimination of factor 1]). As a result, the dark current could not be detected with a galvanometer having a sensitivity of \(5\cdot10^{-10}\) A per scale division. The vacuum in the multiplier was maintained at a level higher than \(10^{-6}\) mm Hg. Power was supplied through a voltage divider connected to a stabilized rectifier of the Ivans system.^4

Under operating conditions (330 V per stage), with this multiplier (for a proton beam) an overall amplification of \(10^5\) times was obtained \([2\cdot10^{-9}\) A, 100 keV protons at the first stage and 200 \(\mu\)A (electrons) at the collector], which, for an output of 8 electrons per 1 proton at the first stage, gives for the following stages \(\delta\), equal to from 2 to 3.

The output of the multiplier was connected to the input stage of a linear tube amplifier. The input resistance of the latter was \(10^3\ \Omega\). An oscilloscope was connected to the output of the amplifier and recorded pulses in those cases when the magnetic field had been adjusted so that ions formed in the ion source could enter the multiplier. Since the number of these pulses, by reducing the electron current in the ion source, could be brought down to several per minute, it is evident that they were produced by individual ions.

Fig. 1.

This instrument was calibrated with a Wilson chamber using protons and \(\alpha\)-particles. Subsequently it was used to count ions with masses from 1 (\(\mathrm{H}^{-}\)) to 32 (\(\mathrm{O_2}^{+}\)), at energies from 50 to 20,000 eV; electrons with energies from several hundred to 6,000 eV; and also individual \(\gamma\)-quanta. The author especially emphasizes one valuable property of this new elementary-particle counter: the very small number of pulses due to background.

This successful attempt to create a counter for individual elementary particles based on the principle of an electron multiplier is not the first. In 1938 Bay^5 published his experiments in this direction, which did not yield satisfactory results. The cause of the failure in this case was that a suitable material for the emitters had not been chosen. Bay worked with a multiplier whose electrodes had been treated with cesium. This led to the presence of large dark currents and the necessity of immersing the instrument in liquid air. Even then it was possible to reduce the background only to 40 electrons per minute. Bay also made the first attempt to use other materials for the emitters: he tried to use nickel electrodes coated with barium oxide.

N. Khlebnikov, Moscow

LITERATURE

1. J. S. Allen, Phys. Rev., 55, 966, 1939.
2. H. Bruining and J. H. De Boer, Physica, 6, 473, 1937; H. S. Khlebnikov, Journal of Technical Physics, 8, 994, 1938.
3. J. S. Allen, Phys. Rev., 55, 336, 1939.
4. R. D. Evans, Rev. Sci. Instr., 5, 371, 1934.
5. Z. Bay, Nature, 141, 284, 1011, 1938.