Artificial Fission of Heavy Nuclei

E. V. Shpolsky, Moscow

The last months of 1938 and the first weeks of 1939 were marked by a discovery which, in its significance, may be placed alongside the most important discoveries ever made in the field of physics. Namely, the work of a number of physicists and chemists, especially O. Hahn, L. Meitner, and Strassmann, showed that, alongside the known types of artificial and natural transformations of atomic nuclei, in which particles with one or two elementary charges are emitted, there exists yet another type of artificial transformation, accompanied by the splitting of a heavy nucleus approximately in half and by the release of enormous quantities of energy.

The beginning of these remarkable investigations was laid by the discovery of the so-called “transuranic elements.” In 1934 Fermi, Rasetti, and D’Agostino^1 showed that, when uranium is bombarded with neutrons, an artificially radioactive nucleus arises with a half-life of 13 min., which in its chemical properties is analogous to the elements of Group VII of the periodic system—rhenium and manganese. Since the possibility of the formation from uranium \((Z = 92)\) of a nucleus with atomic number 75 Re or 25 Mn at that time seemed excluded, it was assumed that, upon capture of a neutron by uranium, an isotope of uranium \({}^{239}_{92}\mathrm{U}\) is formed, which by means of \(\beta\)-transformation passes into an element with atomic number 93:

\[ {}^{239}_{92}\mathrm{U} \longrightarrow \bar e + {}^{239}_{93}\mathrm{R}. \]

The nucleus \({}^{239}_{93}\mathrm{R}\), the nearest analogue of which is rhenium, according to established custom may be called “eka-rhenium.” A detailed analysis of the products arising when uranium is bombarded with neutrons, carried out by L. Meitner, Hahn, and Strassmann, showed that in reality a large quantity of ne-

series, which Meitner, Hahn, and Strassmann interpreted in the following way²:

\[ {}^{238}_{92}\mathrm{U}+{}^{1}_{0}\mathrm{n}\to{}^{239}_{92}\mathrm{U}(10\ \text{sec.})\to{}^{239}_{93}\mathrm{Eka\,Re}(2.2\ \text{min.})\to{}^{239}_{94}\mathrm{Eka\,Os}(59\ \text{min.})\to \]

\[ \to{}^{239}_{95}\mathrm{Eka\,Ir}(66\ \text{hr.})\to{}^{239}_{96}\mathrm{Eka\,Pt}(2.5\ \text{hr.})\to{}^{239}_{97}\mathrm{Eka\,Au}(?), \]

\[ {}^{238}_{92}\mathrm{U}+{}^{1}_{0}\mathrm{n}\to{}^{239}_{92}\mathrm{U}(40\ \text{sec.})\to{}^{239}_{93}\mathrm{Eka\,Re}(16\ \text{min.})\to \]

\[ \to{}^{239}_{94}\mathrm{Eka\,Os}(5.7\ \text{hr.})\to{}^{239}_{95}\mathrm{Eka\,Ir}(?), \]

\[ {}^{238}_{92}\mathrm{U}+{}^{1}_{0}\mathrm{n}\to{}^{239}_{92}\mathrm{U}(23\ \text{min.})\to{}^{239}_{93}\mathrm{Eka\,Re}(?). \]

Thus it was assumed that, when uranium is bombarded with neutrons, three isomeric isotopes of uranium³ may arise, which, through a series of β-transformations, give rise to trans-uranium elements with atomic numbers up to 97.

In 1937 Irène Curie and P. Savić⁴ discovered a new substance arising under the irradiation of uranium with neutrons and possessing a previously unknown half-life of 3.5 hours. The identification of this substance, however, proved difficult. Since in all previously known types of transformations particles with one or two elementary charges were emitted, it was usually assumed that the newly found substance was either an isotope of the initial substance, or was situated in the periodic system somewhere near it. Thus, for example, if it turned out that the new substance obtained from uranium behaved in chemical reactions like osmium \((Z=76)\), it was assigned atomic number 94, and not 76; if the substance reacted chemically like platinum \((Z=78)\), it was assigned atomic number 96, and not 78 (see Table 1, in which the last 3 rows of the periodic system are given). However, the results of studying the chemical behavior of the new substance with period \(3.5^{h}\), published by Curie and Savić at the end of 1938, showed that it was impossible to identify it with any of the elements close to uranium. It was possible to establish—and this is extremely important—that the new substance behaves chemically like lanthanum \((Z=57)\), but since, on the other hand, it is excellently separated from actinium \((Z=89)\), it

TABLE 1

could not be identified with the latter (Table 1). In the end, Curie and Savič were unable to establish the chemical nature of the substance \(R^{3.5h}\) which they had discovered.

Hahn and Strassmann \(^{5}\) also obtained this substance and subjected it to the most detailed study, as a result of which they proposed a hypothetical scheme of the complex transformations arising when uranium is bombarded with neutrons. According to this scheme, upon the capture of a neutron by uranium \({}^{238}_{92}\mathrm{U}\), the uranium isotope \({}^{239}_{92}\mathrm{U}\) arises, which, by two \(\alpha\)-transformations through thorium, gives four isomeric radium isotopes; the latter, in turn, by two successive \(\beta\)-transformations give four isomers of actinium and four isomers of thorium. The whole scheme proposed by Hahn and Strassmann may conveniently be followed from the appended table:

\[ {}^{238}_{92}\mathrm{U}+{}^{1}_{0}n \to {}^{234}_{92}\mathrm{U} \ \xrightarrow{\alpha}\ {}^{235}_{90}\mathrm{Th} \ \xrightarrow{\alpha}\ \begin{cases} \mathrm{Ra}_{\mathrm{I}}\ ? \xrightarrow[\ <1\ \mathrm{min.}\ ]{\beta} \mathrm{Ac}_{\mathrm{I}} \xrightarrow[\ <30\ \mathrm{min.}\ ]{\beta} \mathrm{Th?}\\[6pt] \mathrm{Ra}_{\mathrm{II}} \xrightarrow[\ 14\pm2\ \mathrm{min.}\ ]{\beta} \mathrm{Ac}_{\mathrm{II}} \xrightarrow[\ \text{about }2.5\ \mathrm{hours}\ ]{\beta} \mathrm{Th?}\\[6pt] \mathrm{Ra}_{\mathrm{III}} \xrightarrow[\ 86\pm6\ \mathrm{min.}\ ]{\beta} \mathrm{Ac}_{\mathrm{III}} \xrightarrow[\ \text{several days}\ ]{\beta} \mathrm{Th?}\\[6pt] \mathrm{Ra}_{\mathrm{IV}} \xrightarrow[\ 250\text{--}300\ \mathrm{hours}\ ]{\beta} \mathrm{Ac}_{\mathrm{IV}} \xrightarrow[\ <40\ \mathrm{hours}\ ]{\beta} \mathrm{Th?} \end{cases} \]

As for the substance discovered by Curie and Savič, Hahn and Strassmann proposed that it is in fact a mixture of all these isotopes. Wishing to establish the chemical nature of these substances, Hahn and Strassmann subjected them to an extremely refined chemical investigation. First of all they took up the hypothetical four radium isotopes. Having separated their mixture, they attempted to concentrate these isotopes by separating them from barium salts by fractional crystallization—just as is usually done in concentrating radium from salts containing barium. This attempt, however, did not succeed. Control experiments on the separation in this way from barium of the isotopes of radium, thorium X, and mesothorium I showed that these substances behave as was to be expected, i.e. under fractional crystallization they separate from barium. On the contrary, from a mixture in which mesothorium I and the hypothetical \(\mathrm{Ra}_{\mathrm{IV}}\) were present together with Ba, only \(\mathrm{MsTh}_{\mathrm{I}}\) could be concentrated, whereas \(\mathrm{Ra}_{\mathrm{IV}}\) remained together with the barium. From this Hahn and Strassmann concluded that the substances which they regarded as radium isotopes in fact possess the properties of barium. Further, Hahn and Strassmann showed that their hypothetical \(\mathrm{Ac}_{\mathrm{II}}\) is in reality not an isotope of actinium, since, when a mixture of \(\mathrm{Ac}_{\mathrm{II}}\) with mesothorium II (an isotope of actinium) and lanthanum oxalate is fractionated from nitric acid, only \(\mathrm{MsTh}_{\mathrm{II}}\) is concentrated, whereas \(\mathrm{Ac}_{\mathrm{II}}\) remains together with the lanthanum. This, however, only confirmed the result earlier found by Curie and Savič. From this it was possible to conclude that

substances taken to be actinium are in fact isotopes of lanthanum. It is also quite possible that the third group of isotopes, which had tentatively been identified with thorium, is in fact identical with cerium \((Z = 58)\). As the final result of Hahn and Strassmann’s experiments, it was possible to conclude with complete certainty that there exists yet another type of nuclear transformation, in which approximately half of a heavy nucleus is “split off” from it1. If such a splitting of the nucleus into equal parts actually occurs, then these “fragments” must fly apart in the form of particles possessing enormous energy. On the basis of theoretical considerations, Frisch and Meitner[^6] calculated that the total amount of energy released in each elementary act, from the mass defect and from repulsion, should be about \(2 \cdot 10^8\) eV; therefore each of the resulting particles should carry an energy of about 100 MeV \((1\ \mathrm{MeV} = 10^6\ \mathrm{eV})\). The experimenters’ attention was directed toward the search for such particles. Proof of their existence was obtained with astonishing rapidity in various places in Europe and America. It is worth giving several dates in order to characterize this extraordinary pace. The communication of Hahn and Strassmann[^12] was published in Naturwissenschaften on January 6; on January 16 O. Frisch from Copenhagen sent to Nature a communication on experiments[^7] confirming the “division” of the nucleus (published February 18). On January 26 a conference on theoretical physics opened in Washington, with the participation of N. Bohr and E. Fermi, and at the very first session the discovery of Hahn and Strassmann was subjected to animated discussion, but the results of Frisch’s experimental work in America, apparently, were not known. By the end of the conference, on January 28, Hafstad, Roberts, and Meyer[^8] of the Carnegie Institution in Washington were already able to demonstrate to Bohr and Fermi the appearance of particles of enormous energy as a result of the bombardment of uranium by neutrons. It turned out, however, that the same particles had been observed quite independently in America at two other places as well, namely on January 26 by Fowler and Dodson[^9] at Johns Hopkins University in Baltimore (a report of this was sent by cable by Fowler and Dodson to Nature and published on February 11), and by Dunning and co-workers at Columbia University. In the February 15 issue of Physical Review four communications in succession by different authors from different American laboratories were published, confirming the discovery of Hahn—Meitner—Strassmann—Frisch. The method used by all the experimenters was as follows. The walls, or the collector, of a small ionization chamber connected to a linear amplifier were coated with uranium oxide. If a neutron source (beryllium + radon in Frisch’s experiments) is placed at some distance (1–5 cm) from the chamber, then

10–30 times per minute the strongest ionization pulses are observed. The mass of the uranium-nucleus fragments must be equal to 100–150, the effective charge, according to Frisch’s calculation, 40–50, and the energy up to 100 MeV; despite this enormous energy, their range should amount to only a few millimeters of air, since, owing to the large magnitude of their effective charge, these particles must produce extremely intense ionization (according to an approximate calculation, each particle must create about \(3 \cdot 10^6\) ion pairs). In Frisch’s experiments the thyratron “opened” the amplifier circuit only in the case when the ionization pulse was produced by at least \(5 \cdot 10^5\) ion pairs. Thus the amplifier did not respond to ionization produced by \(\alpha\)-particles or by neutrons. Control experiments showed that in fact no pulses are observed when either uranium or the neutron source is absent. If the neutron source is surrounded by a layer of paraffin, the number of ionization pulses doubles, whence it follows that slow neutrons play an essential role in the process of splitting the uranium nucleus. In the experiments of Roberts, Meyer, and Hafstad, neutrons with different values of the maximum velocity, obtained in different neutron reactions (\(\mathrm{Li} + \mathrm{D} — 13.5\) MeV, \(\mathrm{D} + \mathrm{D} — 2.5\) MeV, \(\mathrm{C} + \mathrm{D} — 0.5\) MeV), were used, and slow neutrons were filtered out by cadmium. It turned out that for neutrons with a maximum energy of 0.5 MeV the introduction of cadmium reduced the effect by 90%, whereas for neutrons with a maximum energy of 2.5 MeV and above—only by 30%.

Experiments with thorium by all the experimenters gave the same effect as experiments with uranium; only slow neutrons play no role in the case of thorium. Experiments with bismuth, lead, thallium, mercury, gold, platinum, tungsten, tin, and silver gave a negative result\(^1\).

Another method for the experimental verification of the splitting of the uranium nucleus, carried out simultaneously by Joliot and Meitner and by Frisch, consists in the following. The nuclei obtained as a result of the splitting possess an abnormal excess in the number of neutrons over the number of protons. The neutrons are successively transformed into protons, as a result of which a chain of radioactive \(\beta\)-transformations is obtained. In fact, the products of the splitting of heavy nuclei always prove to be \(\beta\)-active. Since these “fragments,” moreover, possess a large kinetic energy, it appears possible to collect them on some surface and to study the decay of their activity. Such an experiment was carried out by Joliot in the following way\(^ {10}\). Inside a brass cylinder, coated on the outside with uranium oxide, a neutron source \((\mathrm{Rn} + \mathrm{Be})\) was placed. The brass cylinder, in turn, was arranged coaxially inside a bakelite cylinder, so that the gap between the two cylinders was about 3 mm. After a certain interval of time had elapsed, the bakelite cylinder was removed and inside it

\(^1\) In the case of thallium Joliot observed a weakly positive effect.

a Geiger–Müller counter was placed, which detected the induced activity of the bakelite cylinder, decreasing in a definite manner. A similar experiment was carried out by Meitner and Frisch.

From the theoretical point of view, the splitting of heavy nuclei into two halves at first sight appears enigmatic. Indeed, the emission of charged particles from the nucleus was usually regarded as passage through a potential barrier. However, passage through the barrier by a very heavy particle, or the successive passage within a short interval of time of a large number of particles, is an event of the highest degree of improbability. L. Meitner and O. Frisch, basing themselves on Bohr’s ideas, were the first to point out that the newly discovered phenomenon can be explained classically by analogy with the division of a charged drop. This explanation was later refined by Bohr, and its essential features are as follows. Since the nucleus consists of a large number of very closely

Fig. 1

situated particles bound by enormous forces, it may be likened to a solid body or a liquid. The disintegration of the nucleus under the action of the incident particle should therefore be regarded as a process proceeding in two stages, separated by a relatively considerable interval of time. In the first stage the particle entering the nucleus gives a significant part of its energy to the particles forming the nucleus, and this energy is distributed among many degrees of freedom. As a result a more complex nucleus is obtained, in which the excess energy is distributed among its constituent particles, bringing them into lively motion reminiscent of thermal motions in a solid or in a liquid. This quasi-thermal motion may continue for a very long time—until the excitation energy is given off in the form of electromagnetic radiation (emission of γ-rays). It may happen, however, that, owing to energy fluctuations, some particle situated on the surface of the nucleus receives an excess of energy sufficient to tear itself out of the nucleus. In that case this particle (a neutron, proton, or α-particle) will be ejected from the nucleus, and the remainder of the excitation energy will be emitted in the form of a γ-quantum.

This picture is well illustrated by Bohr’s vivid diagram, which we reproduce here for the convenience of readers (Fig. 1). Before the collision the nucleus is depicted as a symmetric sphere (Fig. 1,1), and its temperature, measured by an imaginary thermometer with scale divisions appropriate to this case ($10^{10}$ degrees Celsius or millions of electron-volts), is equal to zero. After a neutron with an energy of about 10 MeV enters the nucleus, the temperature of the nucleus rises approximately to 18 MeV (Fig. 1,2), and the strongest thermal motions manifest themselves in deformations of the surface of the nucleus. In the next stage a particle (a neutron) flies off from the surface of the nucleus and the temperature falls somewhat (Fig. 1,3); finally, in the last stage (Fig. 1,4), the remaining excitation energy is emitted in the form of a $\gamma$-quantum, the temperature falls to zero, and the nucleus again assumes a symmetric form. Ya. I. Frenkel was the first to point out that this kind of ejection of a particle from a nucleus is very reminiscent of the evaporation of a molecule from the surface of a liquid drop. This analogy, further developed by V. Weisskopf and N. Bohr, proved very useful for interpreting the mechanism of nuclear reactions.

For the case of the splitting of a heavy nucleus in two, Meitner and Frisch indicated another classical analogy. Since the nucleus may be likened to a liquid drop, the quasithermal motions arising when a neutron enters the nucleus lead to deformations of the nuclear surface (Fig. 1,2). These deformations may ultimately lead to the nucleus dividing into parts, in the same way that a drop divides. In the case of the nucleus, division is promoted by the enormous repulsive forces arising between its like-charged parts. Since the effective cross sections for such nuclear-division processes, for neutrons of different velocities, are of the same order of magnitude as for ordinary nuclear reactions, it may be concluded from this that the energy of these deformations leading to nuclear division is of the same order of magnitude as the energy required to eject one particle from it. That purely classical methods may be used for considering this nuclear problem, according to Bohr, follows from the following consideration. It is evident that the deformations leading to the splitting of a heavy nucleus must be large in comparison with the amplitudes of the quantum-mechanical “zero-point” oscillations of the unexcited nucleus, since otherwise it would be impossible to understand the stability of heavy nuclei in the normal state.

In conclusion, one should return to the question of the existence of “trans-uranium elements.” After the experiments of Hahn and Strassmann, doubt naturally arose as to the reality of these elements. In fact, at present it may be considered proven that all periods which had been attributed to elements with atomic number greater than 92 actually belong to “fragments” of heavy nuclei. L. Meitner and O. Frisch carried out a special experiment[^11] in which the bombarded layer of uranium was placed near the surface of water, and the products of the division of the uranium nucleus entered the water, where they were subjected to further investigation. It turned out that,

that in this way the activity curve corresponding to the “trans-uranium” elements can be reproduced, which also confirms the conclusion indicated above.

References

1. E. Fermi, E. Amaldi, O. D’Agostino, F. Rasetti, and E. Segrè, Uspekhi fizicheskikh nauk, 14, 933, 1934.
2. L. Meitner, O. Hahn, and F. Strassmann, Z. Physik, 106, 249, 1937; Ann. d. Physik, 28, 246, 1937; Ber., 69, 905, 1936; 70, 1374, 1937.
3. N. Dmitriev, Uspekhi fizich. nauk, 19, 535, 1938; 21, 60, 1939.
4. Irène Curie and P. Savitch, J. Phys. et Radium, 7, 385, 1937; 9, 355, 1938.
5. O. Hahn and F. Strassmann, Naturwiss., 27, 11, 1939.
6. L. Meitner and O. Frisch, Nature, 143, 239, 1939.
7. O. Frisch, Nature, 143, 276, 1939.
8. Hafstadt, Roberts, and Meyer, Phys. Rev., 1939.
9. Fowler and Dodson, Phys. Rev., 1939.
10. F. Jollot, C. R., 208, 341, 1939; J. Physique, 10, 159, 1939.
11. L. Meitner and O. Frisch, Nature, 143, 470, 1939.
12. O. Hahn and F. Strassmann, Naturwiss., 27, 89, 163, 1939.
13. F. A. Heyn, A. H. W. Atten jun., C. I. Bakker, Nature, 143, 616, 1939.
14. P. Abelson, Phys. Rev., 55, 418, 1939.
15. N. Feather and E. Bretscher, Nature, 143, 516, 1939.
16. N. Bohr, Nature, 143, 330, 1939.
17. N. Bohr, Uspekhi fizich. nauk, 18, 337, 1937.

1. Subsequently it turned out that the products of this division may be very diverse[^13–^15]. Thus it was shown that, besides barium and lanthanum, strontium, yttrium, cesium, iodine, and also krypton and xenon arise. ↩