NEW WORKS ON THE QUESTION OF THE DECOMPOSITION OF ATOMS

E. Khalfin, Leningrad.

§ 1. The decomposition of atoms was first observed by Rutherford in 1919. Studying the scattering of α-particles of RaC in gases, Rutherford noticed that in nitrogen there arise particles with a very large range (up to 40 cm), which, like α-particles, produce flashes (scintillations) when they strike a zinc-sulfide screen. Later investigations showed that these particles are identical with the particles obtained when α-particles pass through hydrogen or hydrogen-containing substances. Since the nitrogen was completely free of hydrogen, and the range of these particles was considerably greater than the range of the nucleus of the hydrogen atom which had acquired its velocity from a collision with an α-particle (< 28 cm), it remained to suppose that these hydrogen nuclei are knocked out by α-particles from the nuclei of nitrogen atoms. This supposition was subsequently confirmed, and it was found that H-particles (with range > 28 cm) are knocked out not only from nitrogen, but also from a number of other elements—B, F, Na, Al, P; moreover, they are knocked out not only in the direction of motion of the α-particles, but also—in approximately the same number—in the opposite direction. On the basis of this latter property of the H-particles (their emission in all directions), Rutherford and Chadwick applied the so-called “rectangular method” for observing them: particles emitted at right angles to the direction of incidence of the α-particle beam are observed. In this case the “natural” H-particles, produced in collisions of α-particles with free hydrogen atoms, cannot be observed. In the “direct” method these particles make it impossible to observe them because of absorption (< 28 cm of air). This method made it possible to study H-particles of small range, emitted in the disintegration of the nucleus. The results obtained by Rutherford and Chadwick reduce to the following: α-particles knock hydrogen nuclei out of atoms of all the light elements up to K inclusive, except Li, Be, C, O, He. In addition, iron was studied, which also did not yield H-particles; α-particles with a range < 3 cm are incapable of splitting Al; the H-particles of aluminium must have a certain minimum range, equal to approximately 9 cm.

Simultaneously with and independently of Rutherford and Chadwick, the “rectangular method” was applied by Kirsch and Pettersson at the Vienna Radium Institute. They themselves, as well as a number of their collaborators, obtained completely opposite results. First of all, they assert that H-particles can be knocked out of beryllium, carbon, oxygen, and iron. Further, their data indicate that for aluminium no “threshold of disintegration” whatever is observed, and that the H-particles knocked out of aluminium can have a very small range (considerably less than 9 cm).

Quite recently the Viennese investigators have published a new series of works,¹ to the exposition of which the present article is devoted.

¹ Zeitschrift f. Physik, vol. 42, pp. 641–758; articles by Kirsch and Pettersson, Pettersson-Holoubek, Schmidt and Stetter. The fullest summary of the earlier works of the Viennese investigators is given in the book: H. Pettersson und G. Kirsch. Atomzertrümmerung. Akademische Verlagsges. Leipzig 1926; see also Handb. der Physik by Geiger and Scheel, vol. XXII.

§ 2. The apparatus in which Schmidt studied the scattering of Al nuclei under the action of impacts of α-particles is shown in Figs. 1 and 2.

The apparatus consists of two halves of a brass box, carefully ground to fit one another. Its total height is 10.5 cm, its diameter 9.7 cm. The source of α-particles is a RaC-coated ring of invar \(P\), with an inside diameter of 14 and an outside diameter of 20 mm (shown in black in the drawing). The source is placed on a lead blank (hatched), which, in turn, is held by a thick brass plate \(M\). Through the lead a channel \(K\), 10 mm in diameter, is drilled; it is lined on the inside with thin brass, projecting a few mm above the upper edge of the channel and forming something like a roof (see the drawing). Along the continuation of the axis of the channel, an opening of 15 mm diameter has been made in the wall of the apparatus; opposite it, on the outside, is fastened a screen for counting particles \(Z\), covered on the side of the apparatus with an aluminum film equivalent to 1 mm of air. Above the screen pass metal plates, one round and one fan-shaped, connected with the pins \(A\) and \(B\). At their periphery holes 10 mm in diameter are cut out, covered with thin plates of mica. Thus, above the screen, by rotating the pins \(AB\), one can set up layers of mica with absorption equivalent to from 0 to 10 cm of air, in steps of 0.5–1.0 cm. Since the measurements are carried out in complete darkness, it proved necessary to attach catches to the pins, which would automatically set the mica disks in the required position. By rotating the pin \(E\), connected with two segments \(Z\), the preparation can be completely closed.

Fig. 1.                Fig. 2.

The pin \(C\) makes it possible to place a stopping mica sheet in the path of the α-particles. This mica is shown in Fig. 1 by the thin line above the preparation. To avoid scattered α-particles from it reaching the screen, it must fit closely to the very edge of the channel \(K\). On the disk \(S\), connected with the pin \(D\), is fastened the substance under investigation (aluminum foil of various thicknesses, heated in vacuum).

With such a construction of the apparatus, particles reaching the screen can have as their source either aluminum \((S)\), or the volume of gas enclosed in the solid angle under which the channel \(K\) is seen from the screen \(Z\). Since, during the experiments, the apparatus was either evacuated or filled with helium, then, assuming that helium is not decomposed, all the observed H-particles could have as their source only Al, placed on the disk \(S\). Distinguishing fast scattered α-particles from H-particles presents no particular difficulty, since the scintillations caused by the former are considerably brighter. In those cases where the scintillations were of approximately equal brightness (which occurs when α-particles are observed in the ...

with very low velocities), comparison with “normal” ones—known α- and H-particles—was used. For this purpose, sources of α-particles and H-particles (a Po preparation covered with a layer of paraffin) were fixed on a circular plate above the screen Z. The particles could be given any velocity by placing in their path a retarding mica sheet, placed on the second, fan-shaped plate.

In order to avoid contaminating the apparatus by recoil, the preparation \(P\) was covered with a collodion film equivalent to approximately \(1\) mm of air. By covering the preparation completely with segments, it is always possible to take into account the degree of contamination of the apparatus. The microscope objective used for counting particles had the following data: focal length \(12\) mm, aperture \(0.70\), magnification \(\times 47\), field of view \(9\ \mathrm{mm}^2\).

With this apparatus the results presented in Table 1 were obtained.

TABLE 1.

Number of H-particles knocked out by α-particles of different ranges from aluminum.

Case c) refers to a reconstructed apparatus and therefore cannot be directly compared with cases a and b.

From these data it follows that α-particles with a range \(< 3\) cm can disintegrate Al, and that no minimum range of H-particles is observed. Moreover, the very large number of observed H-particles is striking. Experiments of a qualitative character with Po, carried out in another apparatus, confirmed these results. The authors estimate the number of particles that can be ejected from Al when α-particles are completely absorbed at up to 200 per \(10^6\) α-particles. These results (the large number and the absence of a minimum range of the knocked-out H-particles) were also confirmed by experiments with unretarded α-particles of RaC, in which aluminum foil was bombarded, the retarding action of which was equal to \(1\) cm of air. (In this case, the presence of slow H-particles cannot be ascribed to their having arisen as a result of the slowing down of fast ones in the very substance being disintegrated.)

The results obtained by Shimizu were confirmed by the work of Golubev, who observed the disintegration of Al in a Wilson chamber (see below). According to Golubev’s data, α-particles with a range of \(0.9–1.6\) cm, when incident on Al, knocked out a considerable number of H-particles with a range of approximately \(5\) cm.

§ 3. Another field of the study of the disintegration of atomic nuclei, where the results of the Vienna experimenters diverge from the results of Rutherford’s school, is

question of the possibility of disintegrating the atoms Be, C, O, F. Particularly detailed investigations were carried out by Peterson and Golubek with carbon. For his investigations Peterson used the already described Schmidt apparatus, but on plate \(S\) there was fastened, of course, not Al, but carbon in the form of graphite, diamond splinters in paraffin, or very pure amorphous carbon. RaC served as the source of \(\alpha\)-particles. The results obtained in such investigations are difficult to reproduce, but it may be said with certainty that a considerable number of H-particles with a small range was observed.

Table 2 gives some of the results then obtained; source of \(\alpha\)-particles—RaC, substance disintegrated—graphite; the apparatus was filled with helium.

TABLE 2.

Range and number of H-particles ejected from carbon
Source of \(\alpha\)-particles: RaC.

The number of H-particles with a range of \(6\ \mathrm{cm}\) is so small that it cannot be accurately counted.

The data of the first two columns were obtained with an interval of 3 months between measurements. The data of the last column were obtained under especially favorable conditions of observation, when the continuous glow of the screen was very weak, which facilitated the counting of faint scintillations. Characteristic is the presence of a large number of very slow (with a range of approximately \(1.5\ \mathrm{cm}\)) H-particles, which were also observed in other similar measurements.

When using (in an apparatus of another construction) \(\alpha\)-particles of Po, approximately the same results were obtained (see Table 3).

TABLE 3.

Range and number of H-particles from carbon. Source of \(\alpha\)-particles: Po

The figures given in graphs 1–5 were obtained in observations on different days. As can be seen from the table, the fluctuations in the number of particles are comparatively small. Just as in the case of the disintegration of carbon atoms by Ra C α-particles, a large number of particles with a short range is observed. The number of H-particles, their “yield,” is absolutely very difficult to determine, but comparative measurements with aluminum showed that it amounts to approximately \(1/2\) of the quantity that is knocked out of the latter.

Fig. 3.

A collaborator of the same institute, Golubev, attempted to observe the disintegration of various elements in Wilson’s chamber.

Golubev’s apparatus, schematically represented in Fig. 3, was arranged in such a way that only particles knocked out of the substance under investigation entered the chamber itself. This made it possible to use very strong Po preparations as the source of α-particles (Po in the drawing). The substance under investigation \(P\) was fixed at the center of the chamber. Through a narrow slit (see Fig. 3), H-particles could fly out in a direction making an angle of not less than \(80^\circ\) with the direction of flight of the α-particle. The slit could be covered with thin aluminum foil of various thicknesses.

Light entered through the glass ring \(F\). The piston \(Q\) was set in motion by an electric motor. The results obtained with this instrument are given in Table 4. Al was a piece of tinplate 0.5 mm thick, heated in vacuum; iron, hard and the purest, was in the form of foil 20 μ thick; Be was in the form of a metallic plate, O in the form of a 1% solution of agar-agar in water.

TABLE 4.

Thus, here too H-particles were obtained in the case when Po α-particles fell on carbon. In addition, H-particles were also observed under bombardment by α-particles of beryllium, oxygen, and iron. It is noteworthy that the “yield” of H-particles is the same in both sections of the table, although the strength of the Po preparation in section b) exceeds the strength of the preparation in section a) by more than 6 times. The “yield” for iron ne-

doubtless diminished, since, owing to the greater range of the scattered \(\alpha\)-particles, it was necessary to introduce considerable (\(4.1\ \text{cm}\)) absorption in the path of the H-particles.

In order to exclude the possibility that H-particles arise in the air on the path source of the \(\alpha\)-particles—disintegrable substance, the apparatus was modified in such a way that the target was covered not with aluminum foil but with thin mica. Then, from the part in which the disintegration takes place, the air can be pumped out or replaced by helium, which, it is supposed, does not give H-particles. The results obtained with the apparatus modified in this way are the same as with the previous one (the “yield” of H-particles is approximately \(17\cdot 10^6\) with \(1.4\ \text{cm}\) absorption for carbon; \(30\cdot 10^6\), with \(1.7\ \text{cm}\) absorption, for aluminum).

§ 4. As has already been indicated, in the work of the Vienna experimenters the H-particles differed from the scattered \(\alpha\)-particles in the character of the scintillations (Pettersson) or in the appearance of the photograph in Wilson’s chamber (Golubev). Of course, there could be no firm certainty that these were indeed H-particles, and it was necessary to apply some method making it possible to distinguish H-particles accurately from scattered \(\alpha\)-particles. Such a method may be the determination of the ratio \(e/m\). Stetter constructed an apparatus of the type of Aston’s well-known apparatus for the study of isotopes. The arrangement of the apparatus is shown in Fig. 4.

On the magnesia plate M lies a very thin (wall thickness \(<10\ \mu\)) capillary with emanation, Q. The \(\alpha\)-particles emitted from it strike the disintegrable substance \(S\). Some of the H-particles knocked out in all directions, passing through the complex target \(C\), enter as a parallel beam into a transverse electric field of about \(100\,000\ \text{V}/\text{cm}\). Beyond it, just as in Aston’s apparatus, a “focusing” magnetic field acts on them.

Fig. 4.

In these investigations, instead of a photographic plate, a screen of zinc sulphide was used, and scintillations at different points were counted. Stetter studied the disintegration of Al, C, B, Fe. The results of his investigations are given in Fig. 5.

As is seen, on all the curves there are maxima at those points where the H-particles should collect. In addition, maxima are observed, of course, also at those places where the scattered \(\alpha\)-particles should be. The case of iron is especially interesting. The large number and high velocity of the scattered \(\alpha\)-particles cause such a considerable broadening of the \(\alpha\)-line that it absorbs the H-line. If, however, the angle which the direction of the target makes with the direction of flight of the \(\alpha\)-particles is increased (position \(S_1\), angles \(\alpha\), in Fig. 4), then, owing to the decrease in the number of scattered \(\alpha\)-particles, the maximum for H-particles begins to stand out. This last is also presented for the different positions of the iron \(S_1\) in Fig. 4. Consequently, these results too confirm the data that carbon and iron are “disintegrable” elements.

On considering the curves in Fig. 5 it is striking that the ratio of the number of scattered \(\alpha\)-particles to the number of H-particles for carbon is considerably greater than for aluminum, whereas Rutherford’s theory of scattering requires precisely the opposite result. This is explained (according to Stetter) by the fact that in the case of aluminum a diaphragm is used which lets through only relatively fast (i.e. little-scattered) \(\alpha\)-particles. If such a diaphragm is placed in the path of the particles when working with carbon, then the maximum for \(\alpha\)-particles, while remaining sharply defined, has an approximately half as large numerical value.

G. KHALFIN

§ 5. The discrepancy between the results of the Vienna and Cambridge investigators is so great that it does not seem possible to explain it by observational errors. It is necessary to suppose that one of the observers made some fundamental error. The Vienna observers believe that such an error in Cambridge was the use of a low-aperture microscope. A comparison carried out in Vienna of the “counting ability” of microscopes of different construction showed that in the microscope used for counting scintillations in Vienna (aperture 0.70), a considerably larger number of scintillations is visible than in a microscope of the Cambridge type (aperture 0.43). This investigation was conducted in the following way: a very strong Po preparation was covered with a layer of paraffin about 40 μ thick. Above the paraffin there was also placed a sheet of mica with an air equivalent of 4 cm, whose purpose was to absorb α-particles that might pass through possible holes in the paraffin. Counting the H-particles of this preparation with the aid of both of the indicated microscopes showed the following advantages of the Vienna microscope in comparison with the Cambridge one: a larger

Fig. 5.

number of scintillations, especially from H-particles of low velocities, observed with it, the small effect of the general glow of the screen under the action of γ-rays, and, finally, the ease of working with it, owing to the smallness of the “subjective field of view” (the field of view of the objective multiplied by the square of the linear magnification of the microscope). The Vienna authors consider these advantages so serious, especially for such weak flashes as are given by H-particles with a small range, that they ascribe to them the difference between their own and the Cambridge results.

A factor which was apparently not taken into account by the Vienna authors is the introduction into the instruments of mica as a “braking” substance for α- and H-particles. Mica is not a definite chemical compound, but a mineral, different types of which may possess different properties. Moreover, what is most important, mica not only contains a number of undoubtedly “splittable” elements, such as Mg, Al, S, but also free hydrogen, apparently in the form of acid and basic salts. From the drawings and dimensions of the instruments given in the papers, it is difficult to judge what role is played by protons knocked out of the mica in the experiments of the Vienna experimenters; in any case, they cannot be called entirely irreproachable, especially in view of the presence of a large number of H-particles with a very small range.

§ 6. Soon after the appearance in print of the works described above, a paper by Bothe and Fränz1 was published; they studied the action of Po α-particles on the nuclei of atoms of various—

of elements. The counting of H-particles was carried out with the aid of a Geiger counter. They used the “direct” method of observation and compared the number of H-particles entering the counter when the α-particles of the substance under study were placed in their path with the number obtained when it was replaced by copper. It turned out that an appreciable difference in the number of H-particles in the two cases was obtained when boron, nitrogen, magnesium, and aluminum were introduced into the path of the α-particles. A negative result (absence of H-particles with a range \(> 7\) cm) was given by graphite, \(\mathrm{CaF_2}\), \(\mathrm{Na_2CO_3}\), Si, P (red), S, and KCl. Since the report is preliminary in character, the authors do not give the exact “yields” and ranges of the particles obtained, which might have decided the question of which results—those obtained in Vienna or those obtained in Cambridge—are closer to the truth.

It must be emphasized that the results of Bothe and Fränz do not contradict the work of Rutherford and Chadwick. The fact that in their apparatus they did not succeed in splitting a number of elements that had been split by Rutherford can be explained by the fact that they used Po as the source of α-particles (range of the α-particles 4 cm), whereas Rutherford and Chadwick used RaC (range of the α-particles 7 cm). Nor is there a direct contradiction with the work of the Vienna school, although there are data that are difficult to reconcile with these results. Thus, for example, carbon does not give H-particles with a range \(> 7\) cm. Peterson (see above) likewise did not observe particles with a range \(> 5\) cm (approximately), but that was the case for particles emitted in the direction opposite to the direction of fall of the α-particles, whereas Bothe and Fränz observed H-particles emitted in the direction of fall of the α-particles. As Rutherford showed, the range of H-particles emitted in the direction of motion of the α-particles is considerably greater than that of those emitted in the opposite direction. Therefore the absence of H-particles in the observations of Bothe and Fränz is in any case not a confirmation of Peterson’s results. The same may be said of oxygen (in \(\mathrm{Na_2CO_3}\)).

§ 7. Another result obtained by the Vienna experimenters can with difficulty be reconciled with previously known facts—namely, their observation that the “yield” of H-particles for aluminum is extremely large (200:100), and that the energy which an α-particle must possess in order to split the Al nucleus is comparatively small (corresponding to the energy of an α-particle with a range of about 1 cm). If, from these two data, using the formulae of Rutherford’s theory, one calculates the dimensions of the nucleus of the aluminum atom (for the method of calculation see, for example, Frenkel, Zh. R. F. O., 1917), then for its radius (i.e. the minimum distance to which an α-particle with a range of 1 cm can approach the center of the nucleus) one obtains the value \(2 \cdot 10^{-12}\) cm. Since the dimensions of the α-particle are considerably smaller (several units in \(10^{-13}\)), in order to split the nucleus it is necessary that, simultaneously with the α-particle being at a distance of \(2 \cdot 10^{-12}\) cm from the center of the nucleus, some one of the protons composing the nucleus should be near it. The probability of this event is less than unity. If this correction is introduced, then the value \(2 \cdot 10^{-12}\) cm will be the lower limit for the size of the Al nucleus.

One may suppose that if a proton forming part of the nucleus can move away to a distance of \(2 \cdot 10^{-12}\) cm from the center of the nucleus, then at a distance \(< 2 \cdot 10^{-12}\) cm there will be some electric field acting on a charge that has flown into it (for example, an α-particle) according to a law different from Coulomb’s law. Experiments on the scattering of α-particles, however, have shown that Rutherford’s formulae remain valid (and hence also the interaction of the α-particle with the nucleus according to Coulomb’s law) down to distances of \(1.2 \cdot 10^{-12}\) cm from the center of the nucleus.

All that has been said about the work of the Vienna experimenters, and the comparison of it with other work in the same and neighboring fields, compels one to conclude that the data obtained by them, despite all the care of the work, cannot be accepted as final. The question of which school is right—the Vienna or the Cambridge—still awaits its solution.

1. Bothe und Fränz, Zeitsch. f. Phys., 43, 456, 1927. ↩