NEW P H O T O M U L T I P L I E R FOR S C I N T I L L A T I O N C O U N T E R S

A. G. B e r k o v s k i i

The multiplying system in the FEU-33 photomulttplter contains auxiliary accelerating grids (Fig. 1) which reduce the spread in the time of flight of secondary electrons in the gaps between dynodes; this spread arises as a result of the difference in initial electron velocities. These auxiliary electrodes also make it possible to oper- ate the multiplier at higher current loads. The grids in front of the l_th dynodes are connected to the (i + 2)-dy- nodes.

5 t

e

~f00 !,

Fig. 1. Diagram of the multiplication system in the

FEU-33.

The dynodc profile comprises a circular section

and two straight sections; the grid profile is comprised of two straight sections. Calculations show that the time of flight of an electron through the stage along the short- est trajectory is 2.4.10 -9 sec while the maximum time

spread (Fig. 1 , trajectories 1 and 4)is 4.5.10 -~~ sec

(with 100 v applied to the stage).

The area of the seml-transparent ceslum-sulfide

cathode of the FEU-33 is 9 cm 2. The mean integrated intensity for a light temperature of 2850~ is 40 tin/lumen

while the minimum is 30 /Ja/lumen. The quantum yield of the cathode is controlled by the ~blue w sensitivity (a

F8-6 filter 1 mm thick).

The multiplying system, consisting of 13 cesium- sulfide dynodes, provides gain of 10 s at voRages of 2500- 3000 v with auniform voltage distribution along the

divider. The dependence of gain and thermal noise on the supply voltage for three FEU tubes is shown in Fig. 2.

In order to study the current characteristics of the

FEU-33 in pulsed operation a special sysmm was used

in which a pulsed light source was employed. The flash repetition frequency was several hundred cycles per se- cond and the duration of each flash was several millimicroseconds. The signal from the FEU was applied directly to the plates of a 13LO37 tube on the screen of which the spot deflection was calibrated with a 26=I oscillator.

Measurements of the pulsed current in a large number of FEU tubes show that at voitages of 3-4.5 kv the peak currents differ; the average current is 0.5 amp. In individual tubes the pulsed currents were as high as 1 amp; this current corresponds to a loading of apprQximately 0.5 amp/cm 2 at the least dynode. This value is some-

what higher than that reported in [1].

The voltage divider which provided the maximum output current was chosen Indlvidually for each multi- pller. As a rule, the difference in the values of the resistance legs was small: R I = 200-220 kohm; P~ = 20-70

kohm; Rs_ m = 180=200 kohm; Ru = 300-500 kohm; R~ = 800=1000 kohm; Rts = 1.5-2.0 meg.

Measurements of the amplitude discrimination of the FEU-33 with a NaI(TI) crystal and a Cs j'W sample were carried out with a single-channel amplitude analyzer [2] at a supply voltage of 1400-1800 v; this quan= tit), was found to vary from 8% to 11O/o. In most of the tubes the operating condRions for best amplitude dlscri-

604

H

t-6

10 ;' p- ~

I0 s r-~

t0 5 fr

~0 4 -#t

I0~ -9 5uu 7uL/u 7bUU ZUUU ZbUU 3000

Fig. 2. three FEU-83 tubes.

Co ~~

Limiter ~ - ~ N ~ N~-Z ] H Limiter Stilbene

Length of ] ] Fast tranmlissionb---q coincidence I

line 1 [ circmt I I

J Cathode follower I

1 I Amplifier HD.iscrirntrtatolH Counter I

The gain (solid lines) and noise currents (dashed lines) as a function of supply voltage for

Fig. 3. Block diagram of the system for determining the time resolution of the FEU-33.

605

mlnatton were similar to the conditions for maximum pulsed current. The mean noise on the NaI(TI)-Cs I~ scale was 3-4 kev. (The comparison was carried out at a background of 50 noise pulses per second,)

The time resolutatlon of the PEU-33 was studied with the system shown schematically in Fig. 3 (page 605). The ltmiter was a 6ZhlP tube with a 75-ohm load. The coincidence circuit made use of a DG-Ts8 diode. The length of the RK-50 shaping cable was 10 era. The amplifier gain was approximately 7500 at a band width of 4 Me/see. The measurement showed that at a counting efficiency of approximately 50% the width of the coin- cidence curve at half height was less than 4 .10 -9 see. In individual tubes it was possible to achieve 2 r = 2.2.10 -9 see. Measurements of the time resolution of the FEU-33 carried out in a number of physics laboratories indicate similar results. The measurements with the coincidence circuit were carried out with the following divider: R 1 = 500 kohm; R 2 = 2"/0 kohm; R s = 75 kohm; R4-1, = 160 kohm; Rxs = 220 kohm; R~ = 400 kohm; R m = 600 kohm.

Comparative measurements of the time resolution of the FEU-33 and other FEU tubes which make use of circular multiplying systems ( [13],p. 1'35)carried out by us and in the laboratories of V. I. Veksler and G. D. Laty- shev show that the FEU-33 has a somewhat better resolution.

At high values of the gain one usually observes damped oscillations at the end of the anode pulse (after- pulses); these have periods of 4-6 re#see and the amplitude of the first after-pulse is approximately 0.1 of the amplitude of the main pulse. A similar effect has been observed in multipliers such as the N4646 ([3], p. 472), RCA-6810 [4] and in high gain tubes with circular dynode configurations.

In conclusion the author wishes to express his gratitude to A. E. Chudakov and P. V. Vakulov for placing the spark light source at the author's disposal and to Iu. I. Tuzhilkin for constructing the apparatus with which the current and time characteristics of the FEU were studied.

L I T E R A T U R E C I T E D

[1] J. S. Allen and L. R. Megill, IRE Trans. Nucl. Set., NS-8, 4, 112 (1956).

[2] Lettertzen, Glukhovskit and Breido, Crystallography 2, 290 (195'/).*

[3] Chechtk, Fainshtein and Lifshits, Electron Multipliers (State Teeh. Press, 195'/), 2nd ed.**

[4] W. Widmaier and R. W. Engstrom, IRE Trans. Nucl. Sci~ NS-3, 4, 137 (1956).

Received August 2'/, 1957

* See English translation.

* * In Russian.

606