FROM CURRENT LITERATURE

APPLICATION OF SCINTILLATION COUNTERS TO THE STUDY OF COSMIC RADIATION

During the last two or three years a large number of papers have been published devoted to the scintillation method of counting fast charged particles and γ-quanta. This method of detecting radiation has now become common in many studies in the field of radioactivity and in work on accelerators. The principal advantages of the scintillation method of recording radiation are well known: they consist, first, in the absence of inertia in the counting, which makes it possible to measure very short time intervals of the order of \(10^{-8}\)–\(10^{-9}\) sec between the moments at which two charged particles pass through a scintillator; second, in the extremely high counting rate, limited at present only by the properties of the counting and radio-engineering circuits; third, in the high efficiency of scintillators for γ-radiation; and, fourth, in the fact that the scintillation radiation detector is a proportional counter, whose light output is proportional to the energy lost by a fast particle in the scintillator. To this it should be added that modern photomultipliers make it possible, when fast charged particles pass through scintillators of ordinary volumes of a few cubic centimeters, to obtain pulses of the order of tenths of a volt and even several volts at a very low noise level. Meanwhile, the pulses from proportional counters that register relativistic or semi-relativistic particles, under ordinary conditions, have a magnitude of the order of hundredths and thousandths of a volt. Therefore, when using a scintillation counter to measure ionizing power, the radio-engineering amplification can be made quite small, of the order of 10–100. This circumstance may be of great importance in the study of cosmic rays in the stratosphere, when the weight of the lifting apparatus must be as small as possible. Despite all the advantages of the new method, its application to the study of cosmic radiation has taken place on a limited scale. The reason for this, apparently, is that until now there have been no photomultipliers with a sufficiently large photocathode surface, necessary for recording small effects of cosmic radiation. The papers reviewed here \(^{1,2,3}\) were carried out in 1950–1951 and are devoted to the direct detection of cosmic radiation by means of scintillation counters. In all these works scintillation counters are used as proportional counters. Paper \(^{1}\), carried out at sea level, had the aim of checking to what extent a scintillation counter is a proportional device. For this purpose, a cylindrical anthracene crystal (the diameter and height of the cylinder were 30 mm) was placed in a telescope of Geiger counters in such a way that the charged particles of cosmic rays traversed approximately equal path lengths in it. Above the system was placed

40 cm of lead, and, by means of lead filters placed between the counters of the telescope, from the total cosmic radiation three intervals were distinguished, corresponding to the following momentum intervals for \(\mu\)-mesons:

\[ \begin{aligned} &\text{1st interval:}\quad 0.8\cdot 10^8\ \text{eV}/c < P_1 < 1.10\cdot 10^8\ \text{eV}/c;\\ &\text{2nd interval:}\quad 1.10\cdot 10^8\ \text{eV}/c < P_2 < 2.5\cdot 10^8\ \text{eV}/c;\\ &\text{3rd interval:}\quad P_3 > 2.5\cdot 10^8\ \text{eV}/c. \end{aligned} \]

A \(\mu\)-meson with momentum \(2.5\cdot 10^8\ \text{eV}/c\) has the minimum (relativistic) ionizing power; \(\mu\)-mesons whose momenta lie in the second and first intervals have, as can easily be verified from the Bethe–Bloch formula for ionization energy losses\(^4\), ionizing powers respectively 1.2 and 1.6 times greater than the ionizing power of a \(\mu\)-meson with momentum \(2.5\cdot 10^8\ \text{eV}/c\). The pulses from the amplifier were recorded by means of an oscilloscope, the sweep of which was triggered by coincident discharges in the counters of the telescope. The distributions obtained for the pulse magnitudes for all three intervals are shown in Fig. 1. Let us note that the width of the lines obtained is determined mainly by the following causes: 1) statistical fluctuations in the energy losses; L. D. Landau showed\(^5\) that they arise chiefly because of the formation of \(\delta\)-particles, and calculated the shape of the fluctuation curve; 2) fluctuations in the number of photons incident on the photocathode of the multiplier; 3) variations in the length of the particle path in the scintillator; 4) fluctuations in the amplification coefficient of the photomultiplier. By using good scintillators with a high light output, sensitive photomultipliers, and recording only those cases in which the particles traverse equal paths in the scintillator, one can considerably reduce the fluctuations due to causes 2), 3), and 4). However, the fluctuations in the energy losses are unavoidable, and in the reference work it is they that, to a considerable degree, determine the observed width of the lines. This is clearly seen from Fig. 1, in the upper part of which is shown the distribution obtained for the pulse amplitudes

Fig. 1. Distribution of pulse amplitudes in three intervals.

on the oscilloscope screen for \(\mu\)-mesons whose momenta are greater than \(2.5 \cdot 10^8\) eV/\(c\) (the third interval). The solid curve drawn in this figure represents the fluctuation Landau curve, normalized to the experimental distribution at its maximum. On the side of small amplitudes the experimentally obtained distribution agrees well with the Landau curve, whereas on the side of large amplitudes a systematic excess of the experimental distribution over the theoretical one is observed. The resulting discrepancy lies outside the limits of statistical errors and is explained by the fact that the spectrum of \(\mu\)-mesons at sea level extends far into the region of momenta exceeding \(2.5 \cdot 10^8\) eV/\(c\). It follows from the Bethe–Bloch formula that for such mesons there is a slow logarithmic increase of ionization losses with increasing momentum. If the Landau curve is constructed taking this increase of losses into account, using the known form of the meson spectrum at sea level, then the theoretical distribution thus obtained will be in good agreement with the experimental one. Let us now consider the distribution corresponding to the second interval of momenta. For comparison, the distribution for the third interval, normalized to its maximum, is plotted on the same graph with a dotted line. As should be expected, the second distribution is shifted relative to the third toward larger amplitudes. If the mean amplitude obtained from distribution 3 is taken as 1, then the mean amplitude from distribution 2 is \(1.20 \pm 0.02\). Above we saw that such an increase of ionization losses should be expected on the basis of the Bethe–Bloch formula. The distribution for the first interval turns out to be shifted still further toward larger amplitudes and gives for the mean energy loss the value \(1.54 \pm 0.08\), which is also in good agreement with the Bethe–Bloch formula. Let us note that for particles whose energy does not exceed 10 Mev, the proportionality between the energy loss in the scintillator and its light output is not a thoroughly established fact\(^6\). From the work considered it follows that the scintillator retains the properties of a proportional counter also for cosmic-ray \(\mu\)-mesons possessing energies of hundreds of megaelectron-volts and more.

The second work\(^2\), carried out at an altitude of 3500 m above sea level, consisted in measuring the distribution in magnitude of pulses produced by cosmic rays in a single scintillation counter. In such use the counter is similar to an ionization chamber. To create a large scintillating surface, a layer of scintillating substance about 0.5 cm thick was poured into a glass cell 11 cm in diameter. The photomultiplier was placed at a distance of 8 cm from the lower base of the glass cell and was cooled to reduce thermal noise with solid carbon dioxide. As the scintillator there were used 1) flakes of commercial naphthalene, 2) flakes of thoroughly purified anthracene, and 3) naphthalene recrystallized from a solution of commercial naphthalene in \(\mathrm{CS}_2\); it was a compact mass of small crystals. After amplification, the pulses from the photomultiplier were fed to a Dumont-248 oscilloscope, whose screen was photographed. In order to measure the background caused by noise in the photomultiplier, a light-tight filter was placed between the cathode of the latter and the scintillator. This background was subtracted from the measured effect. The result of the work being reviewed is the establishment of a power-law dependence of the form \(N(>A)=C\cdot A^{-3.2}\) for the number of pulses \(N(>A)\) with amplitude greater than \(A\). Apparently, large pulses in a scintillation counter are produced mainly as a result of nuclear disintegrations occurring in the scintillator itself or in the material surrounding it. Such nuclear disintegrations have repeatedly been recorded with the aid of unshielded ionization chambers\(^7\). In this case, for the distribution of the magnitudes

of pulses in the chamber was also obtained as a power-law dependence on the range of the wall, close to the three-halves law. A direct result, obtained with all three scintillators, shows that the scintillation counter, like the ionization chamber, detects nuclear disintegrations and is suitable for their detection. In the second part of the work[^2], the relative efficiency of various scintillators with respect to cosmic radiation was measured, i.e., the relative magnitude of the pulses produced at the output of the photomultiplier when different scintillators were used. For these measurements, as in work[^2], the scintillator was placed in a telescope of Geiger counters. The relative efficiency of various crystals and liquids, measured in this way, is given in the table, where the efficiency of a naphthalene crystal is taken as 100. It is placed at the right. The data presented characterize the counter as a whole, i.e., the scintillator and the photomultiplier.

Work[^3], representing the greatest interest in terms of its results, is devoted to the study of primary cosmic radiation. The apparatus, consisting of a scintillation counter, a telescope of Geiger counters, an amplifier, and an oscillograph with a camera, was raised on pilot balloons to an altitude of 30 km, where it remained for 5 hours. A diagram showing the relative arrangement and dimensions of the telescope and the scintillation counter is given in Fig. 2. During the entire time of measurements at an altitude of 30 km, 30,000 pulses were recorded, and Fig. 3 gives the distribution of pulses by amplitude. Let us recall that the pulse produced by a relativistic particle with charge $Z$ is proportional to $Z^2$. Therefore pulses from fast particles with different charge $Z$ must fall into different places in the distribution shown in Fig. 3: the greater the particle charge, the larger the pulse it produces. In Fig. 3

Fig. 2. 1, 2, 3 — Geiger counters.
$T$ — T-shaped tube into which the photomultiplier and scintillator are mounted.

Fig. 3.

it is also shown where in the spectrum the maxima of the distribution for particles with different \(Z\) should be located. The same Fig. 3 gives the distribution obtained during a flight of the setup between altitudes of 5 and 16 km (dashed line). A comparison of the two spectra clearly shows the presence of multiply charged particles at an altitude of 30 km. The authors point out that they observed a noticeable increase in the number of large pulses with time. They attribute this change to a change in the intensity of the multiply charged particles. Stein[^8] was the first to report such an effect. He observed the intensity of primary cosmic radiation with the aid of photographic plates and found a noticeable diurnal effect, consisting in the fact that nuclei with \(Z > 10\) arrive 2–3 times more often during the day than at night. For particles with \(Z < 10\) Stein did not find a diurnal effect. The authors of paper[^3] indicate that, simultaneously with the increase in the number of large pulses (see Fig. 4, “flux of heavy particles”), they observed a noticeable increase in the number of \(\alpha\)-particles in the spectrum. This is seen from Fig. 4, where three distributions obtained at different times are presented. Obviously, the question of the diurnal effect of primary cosmic radiation is of great importance for the whole theory of the origin of cosmic rays: its existence would definitively indicate a solar origin for at least part of the radiation. The scintillation-counter method makes it possible to obtain statistically reliable data on this problem in a short time, whereas the corresponding investigations carried out by the photographic-plate method require very prolonged measurements.

Fig. 4.

A. Weisenberg

CITED LITERATURE

1. F. X. Roser and T. Bowen, Phys. Rev. 82, 284 (1951).
2. D. D. and C. G. Montgomery, Phys. Rev. 80, 757 (1950).
3. E. P. Ney and D. M. Thon, Phys. Rev. 81, 1068 (1950).
4. Rossi and Greisen, Interaction of Cosmic Rays with Matter, GIIIL, 1948, p. 15.
5. L. D. Landau, J. Physics USSR 8, 201 (1944).
6. R. W. Pringle, Nature 166, 11 (1950).
7. C. G. and D. D. Montgomery, Phys. Rev. 76, 1482 (1949).
8. J. J. Lord and M. Shein, Phys. Rev. 80, 304 (1950).