VELOCITIES OF SLOW NEUTRONS

As is known, neutrons that have first passed through paraffin or water produce very great artificial radioactivity in some substances, and also undergo strong absorption in a number of elements. Fermi explained this by the fact that, when passing through [[unclear: word cut off after “contain-”]]

absorbing hydrogen, neutrons lose part of their energy, colliding with protons, and gradually reduce their speeds to the speeds of thermal equilibrium, which, as is known, depend on the temperature. If one also takes into account the fact that the effective cross section for the creation of a radioactive atom upon capture of a slow neutron by a nucleus increases as the neutron speed decreases, then one may expect that the magnitude of the artificial radioactivity produced by slow neutrons will be affected by the temperature of the water or paraffin through which the neutrons pass. Indeed, Moon and Tillman¹ found that slow neutrons produce effects of different magnitude when they pass through paraffin that is, in one case, at room temperature, and in another case at the temperature of liquid air. In the second case the artificial radioactivity in silver or rhodium was approximately 25% greater than in the first. From this the authors conclude that most of the neutrons producing artificial radioactivity in Ag and Rh have energies comparable with the energies of thermal motion.

Subsequent experiments by Fermi²*, as well as by Lukirsky and Tsareva³, confirmed the existence of the temperature effect discovered by Moon and Tillman. At the same time Lukirsky and Tsareva showed that, owing to neutron absorption in paraffin, an increase in the effect produced by neutrons when the temperature is lowered is observed only for not very thick layers of the cooled paraffin. In layers of considerable thickness, the increase in neutron activity at low temperature can be compensated by a decrease in their number owing to their greater absorption in paraffin at low temperature. Indeed, these authors observed that for a layer of cooled paraffin 10 cm thick there was no temperature effect, whereas for a layer thickness of 0.8 cm this effect reached 45%.

On the other hand, Dunning, Pegram, Fink, and Mitchell⁴, using an ionization chamber with lithium walls as an indicator for slow neutrons, did not find an influence of temperature on the magnitude of the effect. However, in their work it was established that neutrons which had passed through water at the temperature of liquid air are absorbed in cadmium somewhat more strongly (by 5%) than neutrons which had passed through water at ordinary temperature. Unfortunately, the results of these authors cannot be interpreted unambiguously.

More definite results on the presence of neutrons with thermal velocities were obtained by Fermi and collaborators, and also by Frisch and Sørensen⁶. In Fermi’s experiments⁵, at the edge of a large rotating disk there were placed two indicators of slow neutrons, between which a neutron source was placed at equal distances. Parallel to the plane of the disk there was a large paraffin ring. If all the neutrons possess high speeds, then they will fall on both indicators in equal numbers, producing equal activity in them. If, however, there are neutrons with speeds comparable with the speed of rotation of the disk, then the first indicator, situated from the source in the direction opposite to the direction of motion, during the stay of these neutrons in the paraffin will have time to move to the place of their exit from the paraffin ring, while the second indicator will “run away” from these slow neutrons. Therefore the first indicator, with rapid rotation of the wheel, should acquire a greater activity than the second, which was in fact observed.

In the experiments of Frisch and Sørensen⁶, neutrons slowed by paraffin first passed through a rotating wooden disk (speed approximately \(9 \cdot 10^{3}\) cm/sec; disk thickness 2 cm) and a cadmium diaphragm** placed directly behind it, and then entered two ionization

* Let us note that Fermi’s first experiments carried out in this direction gave a negative result.

** Cadmium strongly absorbs slow neutrons and therefore can be used for diaphragming them.

chambers (with boron deposited on the wall), positioned at angles of 45° and 135° with respect to the direction of motion of the disk. As it passes through the rotating disk, the neutrons undergo a series of collisions with protons; in this process the rotational velocity is added to the velocity of the neutrons. If the neutrons have velocities comparable with the rotational velocity of the disk, then on emerging from it they should give an asymmetric angular distribution. For neutrons of higher velocities such an asymmetry should not occur. Experiments carried out by the authors showed that the ionization chamber positioned at an angle of 45° gave 4% more recoils than the other. From a rough calculation of the experimental results it follows that the greater part of the neutrons, strongly absorbed in Cd, have velocities comparable with thermal ones.

Fig. 1. \(A\)—rotating disk with cadmium sectors, \(B\)—stationary disk with cadmium sectors, \(C\)—ionization chamber, \(D\)—paraffin with a neutron source, surrounded by cadmium.

The question of the existence of neutrons with thermal velocities was finally resolved in the recent work of Dunning, Pegram, Fink, Mitchell, and Segre,^7 who showed still more crudely. Their apparatus for determining the velocity of slow neutrons consisted of four duralumin* disks with Cd sectors deposited on them. These sectors had a size of 3.7° and were shifted relative to one another by an angle of 3.5°. Two moving disks were fixed immovably, while the other two had a common axis and could be set in motion with various speeds. On one side of the disks was placed a source of slow neutrons; on the other, an indicator, an ionization chamber with lithium walls (Fig. 1). If all the neutrons emerging from the paraffin had very high velocities, then, after passing through the two first disks, they would all reach the indicator through the corresponding apertures in the second disks, since in the negligibly small time required for them to traverse the distance from the source to the indicator, the moving disks would not have time to turn through an appreciable angle. The situation will be different if the beam contains slow neutrons with thermal velocities. In that case the second moving disk will have time to turn through an appreciable angle while the slow neutrons move from the source to the indicator; therefore these neutrons will not enter the apertures of the second disk, but into its cadmium sector, and will be stopped there, as a result of which the number of recoils in the ionization chamber will decrease. For a given

Fig. 2.

* Duralumin is “transparent” to slow neutrons.

the rotational speed of the disks, all slow neutrons up to a certain limiting velocity will be absorbed in the cadmium sectors; the magnitude of this velocity will increase as the rotational speed of the disks increases. Therefore the decrease in the number of recoils of the ionization chamber, observed when passing from one disk rotational speed to another, is caused by neutrons of a definite spectral interval ceasing to enter the chamber; the width and position of this interval can be determined from the geometrical conditions of the experiment and from the rotational speed of the disks. By comparing the change in the number of recoils of the ionization chamber for neutrons of different spectral intervals, one can establish the velocity distribution of the slow neutrons, as was done in the work of Denning, Pegram, Fink, Mitchell, and Segrè. Fig. 2 shows the curve of the neutron velocity distribution obtained by these authors. The vertical dashes on the curve indicate the probable errors. The curve has a maximum between 2 and 3 km/sec. However, this maximum is not of great significance, since in this curve no account has been taken of the fact that neutrons of different velocities have unequal efficiencies in disrupting lithium and therefore may be recorded by the chamber in different percentage ratios.

Considering the results presented, one may conclude that at present the existence of neutrons with thermal velocities is a firmly established fact.

L. Groshev, Moscow

LITERATURE

1. Moon and Tillman, Nature, 135, 904, 1935.
2. Fermi, La Ricerca Scientifica, VI, 1, No 11—12.
3. Lukirsky and Zarewa, Nature, 136, 681, 1935.
4. Denning, Pegram, Fink, Mitchell, Phys. Rev., 47, 888, 1935; 48, 265, 1935.
5. Amaldi, d’Agostino, Fermi, Pontecorvo, Segrè. La Ricerca Scientifica, VI, 1, No 11—12.
6. Frisch, Sørensen, Nature, 136, 258, 1935.
7. Denning, Pegram, Fink, Mitchell, Segrè, Phys. Rev., 48, 704, 1935.