Artificial Splitting of the Elements1

E. Rutherford.

Since the development by Dalton of atomic theory on an experimental basis, the progress of chemistry rested on the central idea of the constancy and indivisibility of the atoms of the elements. The whole body of experimental material in chemistry, over approximately a century, clearly showed that by ordinary physical and chemical processes it was impossible to destroy the atoms of the elements. However, the rapid growth of our knowledge of the internal structure of atoms during the last twenty years has to some extent modified this conception.

At the present time it is generally accepted that the atoms of different elements all have a structure of one and the same general type. At the center of the atom there is a positively charged nucleus, of negligible dimensions, in which the entire mass of the atom is concentrated. This nucleus is surrounded by a cloud of electrons, which are maintained in equilibrium by forces emanating from the nucleus. The electrons occupy a region whose diameter is of the order of \(2 \cdot 10^{-8}\) cm. The nuclear charge of the atom obeys a very simple rule, first established by Moseley: the resultant charge of the nucleus of an atom is equal to its atomic, or ordinal, number and varies from one “atom” of electricity in the case of hydrogen to 92 in the case of uranium. Thus the ordinal numbers express the number of planetary electrons that surround the nucleus of the atom. According to this conception, the ordinary physical and chemical properties of the atom, with the exception of its mass, are wholly determined by the nuclear charge, for on the latter depends the number and distribution of the external electrons, which chiefly determine these properties. The mass of the atom is a property of its nucleus and affects the distribution of the electrons, and consequently the ordinary properties of the atom, only insignificantly.

From this point of view, isotopes at once receive a simple explanation: they consist of atoms with one and the same nuclear charge, but with different nuclear masses. By acting with light or an electric discharge, we can easily remove from an atom one or several

external planetary electrons; by acting with X-rays or swift β-rays, we can even tear away one or several firmly bound electrons of the system. In this way we can, in a certain sense, bring about a transformation of the atom, but only of a temporary character, since soon a new electron is seized from without and the atom again becomes what it was before. General considerations indicate that even if a certain number of planetary electrons is removed under the action of the appropriate agents, the stability of the nucleus will not be disturbed, and after a short time the atom will again acquire its former structure. In order to bring about a permanent change of the atom, it is necessary to destroy its very nucleus. If one charged unit is removed from the structure of the nucleus, then the nuclear charge undergoes a permanent change, and there are no indications that this process could be reversible under ordinary experimental conditions.

The discovery of the instability of radioactive elements was the first sharp blow to the idea of the constancy of all atoms. However, the property of radioactivity is inherent, apart from the two heaviest elements, uranium and thorium, and their numerous descendants, only in two elements, potassium and rubidium, and even then to an insignificant degree. With these exceptions, the great majority of the remaining atoms are very stable structures and have probably remained unchanged on our earth for thousands or millions of years.

Radioactivity is a property of the nucleus; it manifests itself, generally speaking, in the emission from the nucleus of swift α-particles, or helium nuclei, and sometimes of swift electrons, or β-rays. The number and velocity of emission of these particles do not yield to the influence of even the most powerful physical and chemical agents, and are a property invariably inherent in the above-mentioned highly complex nuclei, owing to their instability.

From what has been set forth it is clearly evident that the nuclei of heavy atoms contain positively charged helium nuclei and negative electrons; and this leads us to the general conception that the complex nuclei of all atoms are built of hydrogen nuclei, helium nuclei, and electrons. It is usually assumed in this connection that the helium nucleus is a secondary unit and is itself built of four hydrogen nuclei and two electrons. If this is so, then we may suppose that the nuclei of all atoms are, ultimately, built of hydrogen nuclei, or “protons,” as they are now called, and negative electrons.

Thus radioactivity provides not only a key to the understanding of the structure of all elements, but also furnishes us with such a powerful method of investigating the internal structure of the atom as swift α- and β-particles. By bombarding the atoms of a substance with α-particles and studying the deflection of their flight from the original direction, we can determine the magnitude of the forces and the law of their action near the nucleus and со-

to form some notion of the dimensions of the latter. General results indicate that the diameter of the nucleus of heavy atoms is of the order of \(4\cdot 10^{-12}\) cm, or, approximately, \(\frac{1}{5000}\) of the diameter of the entire atom. The law of inverse proportionality to the square of the distance for repulsion between electric charges proved valid over a considerable region around the nucleus. Undoubtedly, the nuclei of light atoms are still smaller and, in the case of helium, apparently have dimensions of the order of \(5\cdot 10^{-13}\) cm. Hence it is clear that atomic nuclei, even with a very complex structure, have exceedingly small dimensions.

The forces holding together the components of the nucleus are probably very powerful, and therefore large quantities of energy may be obtained when the latter is destroyed. The fast \(\alpha\)-particles of radium or thorium, which represent the most concentrated source of energy of all known to us, are the most suitable instrument for destroying the strongly built nucleus; an \(\alpha\)-particle is emitted from radium with a speed of about ten thousand miles per second and thus possesses a speed two hundred thousand times greater than that of swift rifle bullets.

Although an \(\alpha\)-particle directed straight into a heavy nucleus can undoubtedly penetrate into its structure, its energy may prove too small to bring about destruction. The bombardment of light atoms promises greater results, since the repulsive forces in this case are so much smaller that the \(\alpha\)-particle can retain a large part of its energy after already entering into the very structure of the nucleus.

However, before discussing the experiments devoted to this question, a few words should be said about the collisions of \(\alpha\)-particles with hydrogen nuclei, where there can be no question of the destruction of the atom. When \(\alpha\)-particles pass through gaseous hydrogen, close collisions sometimes occur between the \(\alpha\)-particles and the hydrogen nuclei. As a result of such collisions, \(H\)-nuclei of high velocity arise. These \(H\)-particles traverse a distance four times greater than the bombarding \(\alpha\)-particles, and they can easily be detected by the scintillations which they excite on a zinc-sulfide screen. On the basis of ordinary principles one can calculate that the maximum velocity imparted to an \(H\)-nucleus is 1.6 times greater than the velocity of the incident \(\alpha\)-particles, while the maximum energy is \(0.64\) of the energy of the \(\alpha\)-particle. It was found that the number of these fast \(H\)-atoms far exceeds what could have been expected if one assumes that the \(\alpha\)-particles and the hydrogen nucleus behave like point charges at the minute distances involved in these powerful collisions. Moreover, the variation in the number of \(H\)-particles as a function of the velocity of the \(\alpha\)-particles, as well as the number of \(H\)-particles ejected at various angles to the direction of the \(\alpha\)-particle, differs substantially from the results obtained from the simple theory of point charges.

In all likelihood, what is at work here is not only the structure of the $\alpha$-particle, but also the circumstance that, for very small distances, the law of action of the force is quite different from the inverse-square law. Chadwick and Bieler, on the basis of a careful investigation, have recently drawn the conclusion that the results of collisions can be explained by assuming that the $\alpha$-particle, to which all the complications had been attributed, behaves like a spheroid with axes $8 \cdot 10^{-13}$ and $5 \cdot 10^{-13}$. Outside the surface of this spheroid the inverse-square law holds; but when the $H$-nucleus passes through the surface of the spheroid, the forces increase so rapidly that it is quickly pushed back. Of course, such a model of the helium nucleus is highly artificial, but it gives us some idea of the probable dimensions and extent of that region in which these new powerful forces are played out.

Consequently, we must anticipate that, in a close collision of an $\alpha$-particle with the nucleus of an atom more complex than the nucleus of hydrogen, the ordinary laws of force action are violated when the distances separating the particles and the nucleus become very small. It must be remembered that in such nuclear collisions gigantic forces are developed, and one must expect that only very stable structures can withstand such collisions.

The first observation connected with the principal subject of my lecture was made several years ago. When $\alpha$-rays from a strong radioactive source pass through dry gases, such as oxygen or carbon dioxide, a small number of weak scintillations is observed on the screen beyond the range of the $\alpha$-particles. These “natural” scintillations were attributed to hydrogen atoms coming from the source and arising, probably, as a consequence of a slight “hydrogen contamination” of the source during the action upon it of the radium emanation. If, however, instead of oxygen or carbon dioxide, dry air is taken, the number of scintillations increases three- or fourfold. It was found that this additional effect is caused by the presence of nitrogen, and it was shown that it increases correspondingly if the air is replaced by chemically prepared nitrogen. By a special arrangement of the experiment it was shown that the particles producing these scintillations are deflected by a magnetic field approximately as would be expected if they consisted of rapidly moving charged $H$-atoms. It seemed probable from the very beginning that these additional $H$-atoms, which appear only in dry nitrogen, but not in oxygen or carbon dioxide, owe their origin to the splitting of the nitrogen nucleus in collision with a swift $\alpha$-particle.

In work with the original arrangement the scintillations were few in number, weak in intensity, and it was difficult to count them with sufficient accuracy. Further progress depended chiefly on improving the counting microscope

in the sense of increasing the intensity of scintillations and the area of the zinc-sulfide screen accessible to observation. Thanks to the use of an objective with a wide aperture and special ocular lenses of low magnification, the counting of scintillations became considerably easier and much more definite.

We shall now consider the methods developed for a more detailed investigation of the phenomena observed in nitrogen, and for the study of other elements in the same respect. The apparatus adopted (Fig. 1) has the simplest construction and consists chiefly of a brass tube (T), 3 cm in diameter, furnished with stopcocks by means of which a dry gas can circulate through the tube. At one end of the tube there is an opening covered with a thin silver plate. The screen

Fig. 1.

made of zinc sulfide (S) is fastened at a distance of 1.3 mm from the opening and is provided with grooves into which absorbing screens of mica can be inserted. The radioactive source is fixed at the end of a rod, so that its distance from the screen can be varied at will. In order to diminish the illumination of the screen caused by the $\beta$-rays of the source, the entire apparatus was placed in a strong magnetic field. It is not without interest to give a few details to illustrate the magnitude of the effect observed under different conditions. Suppose that the radioactive source, consisting of a brass disk coated on one side with an invisible deposit of radium C, with an activity in $\gamma$-rays corresponding to 40 mg of radium, is placed at a distance of 3.5 cm from the screen, and that a current of dry hydrogen is passed through the apparatus. Let the stopping power of the material between the source and the zinc-sulfide screen correspond to 20 cm of air, so that this stopping power would be sufficient to stop an $\alpha$-particle with a range of 20 cm in air. The passage of $\alpha$-particles (which in the present case had a range of 7 cm) through hydrogen liberates a large number of rapidly moving H-atoms, exciting scintillations on the screen. The number of these scintillations, if the screen is viewed through a special microscope with a field of view of 40 mm², is so great—thousands per minute—that it would be impossible to count them without reducing

activity of the source. If, however, additional absorbing screens of mica are inserted, the number of scintillations rapidly falls, and with an absorption, for example, of 30 cm, not a single \(H\)-scintillation per minute can be observed. The same phenomenon is observed if the hydrogen is replaced by oxygen and the source is covered with a thin film of paraffin or of some other hydrogen-containing compound. The number of \(H\)-scintillations observed at a given absorption depends only on the amount of hydrogen and is quite independent of the chemical combination. This was to be expected, for the forces required to set an \(H\)-nucleus into rapid motion are enormous in comparison with the weak forces manifested in chemical combinations. Thus we draw the conclusion that, for \(\alpha\)-particles with a range of 7 cm, \(H\)-atoms cannot be knocked out of hydrogen, in the free state or in a chemical compound, at an absorption greater than 30 cm of air.

Now the oxygen, which gives no scintillations, is again replaced by dry air. Then, at an absorption of 30 cm, we at once observe more than 100 scintillations per minute, whereas in hydrogen not a single one was observed. By adding new mica screens, we find that the scintillations cease at an absorption of 40 cm. It is clear that these particles, which arise from nitrogen, have a greater range than free \(H\)-atoms bombarded by \(\alpha\)-rays. Thus the effect observed beyond 30 cm cannot be ascribed to hydrogen contamination of the nitrogen.

Next the air is again replaced by neutral oxygen, and the source is successively covered with thin sheets of copper, iron, silver, and gold, with a stopping power corresponding approximately to 3 cm of air. In this case not a single \(H\)-atom is observed for an absorption greater than 30 cm. If now a thin aluminum sheet is placed in front of the source, the number of scintillations at once rises to 100 per minute. Some of these scintillations are very bright, and it turns out that such fast particles are emitted that the absorption must be increased to 90 cm in order for the scintillations to cease. Thus it is clear that aluminum gives rise to a certain number of particles with a very large range.

When we investigate the number of scintillations beyond the range of ordinary \(H\)-atoms, our results are completely independent of any contamination of the substance under test by hydrogen. This is a great advantage, for we need not be concerned about the purity of our material with respect to the absence of hydrogen. In this way Chadwick (Dr. Chadwick) and I examined a large number of elements in order to determine whether these elements emit particles with a range greater than 32 cm. When the element itself could not be used, a compound of that element with an “inactive” element, such as oxygen, was employed. The substance, in the form of a powder, насы-

was applied to a thin gold leaf, fixed with a layer of adhesive substance (the total absorption corresponding on average to 3–4 cm of air), and was exposed to the rays. All elements up to atomic weight 40 were investigated, with the exception of helium, neon, and argon. It turned out that not a single element with an atomic weight greater than that of phosphorus, 31, gives any effect, although it must be said that so far only a small number of elements with large atomic weights have been investigated.

The list of elements from lithium through sulfur inclusive, investigated in this way, is given in the following table. In the third column are given the numbers of scintillations per minute per milligram of activity of the source—radium C—for an absorption of 32 cm of air. These numbers permit only a rough comparison of the effects produced by the various elements, since the experimental conditions—for example, the thickness and distribution of the film of substance—varied from element to element. The fourth column gives approximate values of the ranges of the particles.

In addition to those listed, the following elements with larger atomic weights were also investigated: chlorine—from $\mathrm{MgCl_2}$, potassium—from $\mathrm{KCl}$, calcium—from $\mathrm{CaO}$, titanium—from $\mathrm{Ti_2O_3}$, manganese—from $\mathrm{MnO_2}$, iron, copper, tin, silver, and gold in the form of metallic foils. In no case were particles observed with a range greater than 32 cm of air. The question of whether these elements emit particles with a range smaller than 32 cm has not yet been investigated.

From the table given it is seen that, for absorption in 32 cm of air, scintillations are produced by boron, nitrogen, fluorine, sodium, aluminum, and phosphorus. The number of scintillations for boron and sodium is considerably smaller than for the other elements.

Absorption curves for nitrogen and aluminum.

The change in the number of scintillations with increasing absorption along the path of the rays, starting from 10 cm, is shown in Fig. 2. In all cases the source of the α-rays was radium C. Curve A shows the effect in nitrogen (air), where the maximum range is 40 cm. Curve B is the absorption curve for a mixture of hydrogen and carbon dioxide—approximately 1 volume of hydrogen to 1.5 volumes of carbon dioxide. This mixture has the same stopping power for α-rays as air. The number of scintillations produced by hydrogen is very large, but falls rapidly, and beyond 30 cm not a single scintillation can be detected.

Fig. 2.

scintillation. Curve C gives the “natural” effect when air is replaced by dry oxygen. This effect is small in comparison with the effect observed in nitrogen. Curve D shows the effect observed when an aluminum plate with a stopping power of 3.5 cm of air is placed in front of the source, and the apparatus is filled with dry oxygen instead of air. Thus the particles liberated from aluminum have the capacity to penetrate a much greater thickness than the particles liberated in hydrogen or in nitrogen.

It is very interesting to determine how the absorption curves of these particles with long range vary depending on the change in velocity of the bombarding α-particles. This question was investigated for two typical elements, nitrogen and aluminum, and the results for the latter are shown in Fig. 3. It was found that, to a first approximation, the maximum range of the particles liberated from the element is proportional to the range of the bombarding particles. In all cases the number of scintillations falls rapidly as the velocity of the α-particles decreases.

The influence of velocity is especially noticeable in aluminum, and when the range of the α-particles is reduced to 4.9 cm, only a very small number of particles is observed, if they can be detected at all. The effect shown by curve \(D\) (Fig. 3) is caused almost exclusively by the “natural” scintillations of the source. If we take into account that the decrease in velocity corresponding to a decrease in the range of the α-particles from 7 to 4.9 cm is only 11%, then we shall see how rapidly the number of scintillations falls with decreasing velocity. Apparently, in the case of aluminum, disintegration cannot in general be produced if the velocity of the α-particles is below a certain crit—

Fig. 3.

Labels in the figure:
\( \mathrm{Al_2O_3} \)

\(A\) — range of α-particles 8.6 cm.
\(B\) — same, 7.0 cm.
\(C\) — same, 6.0 cm.
\(D\) — same, 4.9 cm.
\(E\) — curve for oxygen.
\(F\) — ranges of α-particles 8.6–7.6 cm.

Vertical axis: Number of particles per minute.
Horizontal axis: Absorption in cm of air.

—ical value. It is not easy to prove this sufficiently convincingly, but if this is so, it indicates that the α-particle must possess some critical energy in order to liberate an \(H\)-atom from the nucleus.

A very remarkable result was found in the case of aluminum. It was to be expected that the greater part of the liberated particles would move in the direction of the bombarding α-particles. In reality, however, it was found that almost as many particles are emitted backward as are emitted forward. In nitrogen no analogous phenomenon was observed. The remaining elements have not yet been investigated from this point of view, but we must expect that an element such as phosphorus, which gives particles with a large range, will also exhibit a similar effect1.

Trans.

The Nature of the Ejected Particles.

It is easy to show that particles with a long range, liberated from elements, are deflected in a strong magnetic field. With the aid of special methods it proved possible to compare the magnitude of the deflection of these particles with the deflection of fast \(H\)-atoms produced when \(\alpha\)-particles pass through ordinary hydrogen. It was found that particles from nitrogen are deflected almost to the same degree as \(H\)-particles from hydrogen, and in all respects behave as fast \(H\)-atoms carrying one positive charge. At first sight it seems probable that the corresponding particles from fluorine, phosphorus, and aluminum are likewise \(H\)-atoms, liberated from the nucleus with velocities depending on the nature of the element and on the velocity of the incident particles. This has been confirmed by recent experiments by Chadwick and myself, using a method analogous to that which was employed in the investigation of nitrogen1. The deflection of the particles in the magnetic field was studied for an absorption greater than 32 cm, in order not to complicate the experiment by the possible presence of hydrogen “contamination” in the substance under investigation. The experiments proved difficult, since under the conditions of their arrangement only a small number of particles was available for observation. For this investigation it was necessary to construct a special microscope with a large field of view. The results of all the experiments showed complete agreement with the view that the particles from fluorine, phosphorus, and aluminum are fast atoms of hydrogen, and we may conclude that in all cases \(H\)-atoms are liberated from the nuclei of these elements.

The maximum velocity of ejection of an \(H\)-atom from various elements may be estimated approximately by assuming that the law connecting the velocity and the range of an \(\alpha\)-particle is also valid for an \(H\)-atom, i.e., by assuming that the velocity is proportional to the cube root of the range. It was calculated—and experiment confirmed this calculation—that the maximum velocity imparted to an \(H\)-atom in a central collision with an \(\alpha\)-particle of velocity \(V\) is \(1.6V\), and that its range in air is approximately 28 cm.2 Consequently, the maximum velocity of an \(H\)-atom from nitrogen with a range of 40 cm will be \(1.8V\), and the maximum velocity of an \(H\)-atom from aluminum, with a range of 90 cm, will be \(2.37V\). In a direct collision the \(\alpha\)-particle gives a free \(H\)-atom 0.64 of its energy, and it may be calculated that all \(H\)-particles which have a range greater than 56 cm are ejected with an energy greater than the energy of the bombarding \(\alpha\)-particles. In the case of aluminum the maximum energy of the \(H\)-atom is 1.4 times greater than the energy of the incident \(\alpha\)-particles. This very interesting result shows,

that in some cases, in the disintegration of the aluminum nucleus, there is indeed a gain in energy. We may therefore conclude that in every case, in all collisions in which the liberated \(H\)-atom has a range greater than 56 cm of air, part of the energy of the \(H\)-atom is borrowed from the disintegrated nucleus. In some respects there is here an analogy with the well-known gain in energy when an \(\alpha\)-particle is liberated from a radioactive nucleus.

It should be remembered that the quantity of disintegrated substance is extremely small. When an \(\alpha\)-particle of radium \(C\) passes through aluminum, it probably penetrates through the electronic structures of about 100,000 atoms, and only, approximately, two \(\alpha\)-particles in a million pass sufficiently close to the nucleus to cause the liberation of an \(H\)-atom. We know that the \(\alpha\)-particles of 1 gram of radium, if collected, yield 163 cubic millimeters of helium per year. If we assume that all the \(\alpha\)-particles from 1 gram of radium were directed into aluminum, then the quantity of hydrogen liberated as a result of the disintegration of aluminum nuclei should not exceed 0.001 cubic millimeter per year. Thus the quantity of hydrogen liberated under the actual conditions of the experiment lies far beyond the possibility of detection by ordinary chemical methods. The study of the process of disintegration became possible only through the use of so delicate a method, which permits each liberated \(H\)-atom to be observed by the scintillations on a zinc-sulphide screen.

Mechanism of Disintegration.

The study of radioactivity had already suggested the idea that the \(\alpha\)-particle, or helium nucleus of mass 4, is one of the units in the structure of the atom. The experiments described in the present lecture gave the first definite proof that the hydrogen nucleus is also one of the units in the structure of certain light elements. It is interesting to note that \(H\)-atoms are liberated only from those elements whose atomic masses are expressed by the formulae \(4n+2\) or \(4n+3\), where \(n\) is an integer. Elements such as carbon or oxygen, whose atomic masses are expressed by the formula \(4n\), do not yield \(H\)-atoms. This is shown in the following table:

Such a result was to be expected if the nuclei of the listed elements are built up from helium nuclei of mass 4 and hydrogen nuclei. In order to

explain the liberation of \(H\)-atoms from these elements, it is natural to suppose that the \(H\)-nuclei are satellites of the main mass of their nucleus. If the satellite is not too close to the latter, then an \(\alpha\)-particle, in a close collision, can impart to the satellite sufficient energy to free it from the system. It must be anticipated that, in the case of aluminum, the \(H\)-satellites are situated closer to the nucleus than in the case of nitrogen, and consequently that, in the case of aluminum, a greater energy is required to liberate the satellites. It is interesting to note that the probability of liberating a fast \(H\)-atom from nitrogen amounts to no more than \(1/20\) of the probability that a free \(H\)-atom will be set into the corresponding motion. This indicates that there must exist certain limits to the velocity of the satellite and certain positions of it relative to the central nucleus for which liberation of this satellite is possible.

We have already pointed out the fact that \(H\)-atoms from aluminum seem to be liberated in all directions. In reality, however, the velocity of \(H\)-atoms ejected in the direction opposite to the motion of the \(\alpha\)-particle is considerably less than the velocity of those ejected in the same direction. At first sight this result seems to suggest that the \(\alpha\)-particle plays the role of a detonator for the aluminum nucleus and that the energy of the fragments liberated is borrowed chiefly from the nucleus. I think, however, that the following explanation is more probable and is in better agreement with experiment. If we suppose that the \(H\)-satellite describes an orbit around the central nucleus, then the direction of ejection will depend on the relative position of the \(\alpha\)-particle and of the nucleus at the moment of a close collision with the satellite. For example, in the collision shown in Fig. 4 A, the \(H\)-atom is liberated in the forward direction; in the collision of Fig. 4 B the \(H\)-atom describes an orbit around the nucleus and is liberated in the reverse direction. The velocity imparted to the residual nucleus in the forward direction is considerably greater in the latter case than in the former. Such a representation assumes that the forces between the positively charged satellite and the nucleus in the immediate vicinity of the latter are attractive, not repulsive. This change of sign of the forces at very small distances from the nucleus seems quite probable, for otherwise it is difficult to understand how a positively charged compound nucleus maintains its stability.

Fig. 4.

From the possibility of liberating \(H\)-atoms by light elements there follows still another rather interesting consequence. It is generally assumed—although it is very difficult to give direct proof of this assumption—that the helium nucleus is built of four hydrogen nuclei and two electrons. Such a combination is accompanied by a loss of mass1, and this circumstance is attributed to a very close combination of the structural units. From the modern conception of the relation between mass and energy it follows that the energy liberated in the formation of the helium nucleus is more than three times greater than the energy of the fastest \(\alpha\)-particle of radium2. Consequently, we should not expect the possibility of destroying the helium nucleus by an ordinary \(\alpha\)-particle, which is in complete agreement with experiments, since so far they have remained in vain. Indeed, the helium nucleus appears to be the most stable of all nuclei.

However, in view of the fact that in the case of nitrogen, for example, we can liberate an \(H\)-atom by means of slow \(\alpha\)-particles, it is clear that in the nitrogen nucleus the \(H\)-satellite is bound not so tightly as in the helium nucleus. Consequently, the change of mass caused by the radiation of the energy that bound the \(H\)-satellite must be considerably smaller in nitrogen than in helium. The mass of the satellite should not differ greatly from the mass of a free \(H\) nucleus \((1.0077,\ \text{if } O=16)\).

If one assumes, for example, that the nitrogen nucleus is built from three helium nuclei and two hydrogen nuclei1, then the mass of the nitrogen atom should not be exactly 14.00, but should be closer to 14.01. It seems probable that, in the case of the light elements, the effective mass of the protons composing the nucleus must vary in different atoms approximately from 1.007 to 1.000, depending on the greater or lesser tightness of the combination. Consequently, we should expect that Aston’s “whole-number rule,” which is valid for atomic masses with an accuracy of about 1 per mille, would have to be abandoned if measurements could be made with considerably greater precision.

Next there arises the question whether, under bombardment by \(\alpha\)-rays, particles other than hydrogen may be liberated. Some time ago I found that when radium \(C\) is used as the source, one can observe a small number of bright scintillations corresponding to a maximum range in air of 9 cm. Naturally, from the very beginning it was to be supposed that these scintillations were caused by a new type of \(\alpha\)-rays issuing from the radioactive source. However, the influence of aluminum screens in reducing the range of these particles at first led me to think that they arise in the volume of gas filling the apparatus, namely in nitrogen and oxygen. Comparing the deflection of these rays in a magnetic field with the deflection of \(H\)-particles from hydrogen, I concluded that these rays consist of atoms having a mass of about 3 and carrying two positive charges. Further experiments directly indicated to me the defect in the method used for mounting the radiation source, which caused a noticeable nonuniformity in the thickness of the metallic films. Using a more direct and simple method, I have recently become convinced that, in the case of oxygen, the particles undoubtedly arose in the radioactive source, and not in the volume of the surrounding gas. Under such conditions the comparative method of estimating the mass of the particles is already unreliable. In order to establish precisely the nature of this radiation, a large number of experiments is required. However, the general indications point to its consisting of particles of mass 4, which are emitted from the source and characterize a new mode of transformation of radium \(C\).

By the methods described one can discover particles that travel a distance greater than the original \(\alpha\)-particles. However, if a splitting of an element is possible in which a massive particle is liberated, it is very probable that the latter will have a smaller range than an \(\alpha\)-particle. For the investigation of such cases we can make use of the excellent method developed by Wilson (C. T. R. Wilson) for demonstrating the paths of ionizing particles. Similar experiments, by a modified method, were carried out by Shimizu

(Shimizu) in the Cavendish Laboratory. A number of photographs were obtained showing sharply defined tracks of particles near the end of their range. However, despite the fact that these photographs were carefully measured and compared with one another, it is difficult to say with certainty whether the branched tracks can be explained by the collision of an $\alpha$-particle with nitrogen or oxygen nuclei. Apparently, however, the nuclei mentioned can travel considerable distances in the gas before being absorbed. If a large number of photographs are taken, it will be possible to decide definitely whether collisions accompanied by the destruction of an atom occur, and to find the probability of such collisions. Such a direct method of approach to the problem, although it seems laborious, should yield very valuable information on the question that interests us.

It seems probable that an $\alpha$-particle may sometimes be capable of knocking a helium atom out of a complex nucleus, such as the nuclei of carbon or oxygen, which may be regarded as built up respectively of three and four helium nuclei. The very fact that the masses of these atoms are approximately equal to whole multiples of the mass of the helium atom suggests that the helium nuclei in them are bound to one another by forces considerably weaker than the $H$-components of the helium nucleus itself. If the structure of a complex nucleus—for example, the nucleus of oxygen—is such that an $\alpha$-particle can impart a considerable part of its momentum and its energy to the individual components of this nucleus, then we must expect that such a splitting may occur. It is also possible that, as secondary structural units of the complex nuclei of certain elements, there are also particles with masses of about 2 or 3, but up to now no definite proof of their liberation has been obtained.

I have so far confined myself to considering the effect of disintegration by fast $\alpha$-rays. It is important to discuss whether fast $\beta$-rays or the energetic $\gamma$-rays of radium might not cause analogous phenomena. We have found that neither $\beta$- nor $\gamma$-rays are capable of imparting sufficient energy to a free $H$-atom in ordinary hydrogen for this $H$-atom to be detected by the scintillation method. Consequently, still less should we expect these rays to liberate a fast $H$-atom from a complex nucleus. It is possible, however, that these agents, and especially $\gamma$-rays of very short wavelength, will be capable of liberating an electron and thus causing a change in the atomic number. It would be difficult, however, to ascertain that such a transformation had occurred, except in the case where the resulting atom, in turn, proves unstable and disintegrates with the emission of fast particles of the type of $\alpha$-rays. It should be noted that Slater showed that $\alpha$-rays, when passing through ordinary matter, can excite certain highly penetrating rays. This observation may serve as some evidence that similar radiation of high frequency can arise only from the nucleus of the atom. If this should prove to be the case, it is possible

it would be that $\alpha$-rays in some cases cause the ejection of $\beta$-particles from the nucleus and a subsequent transformation. However, this effect must be extremely insignificant.

Already in the past several attempts had been made to decompose the ordinary atom by means of special agents. The late Sir William Ramsay, with his characteristic instinct for choosing the best “line of attack,” performed a series of experiments on the effect of the $\alpha$-rays of radium on matter and concluded that he had succeeded in producing neon from water and liberating lithium from copper. These conclusions were not confirmed by later investigators, and in the light of the experiments described in this lecture it appears very doubtful that the scale of the transformation—even if it did occur—was sufficient for it to be detected by the ordinary chemical methods then employed.

Several cases were also noted of the appearance of helium in discharge tubes. It was assumed that helium was a product of the transformation of the electrodes under the action of intense electric discharges. The most remarkable experiment in this direction was carried out by Prof. Collie, but subsequent detailed investigations by Strutt did not confirm Collie’s conclusions. It is extremely difficult to prove that the formation of helium was not caused by its liberation from an occluded state in the electrodes as a result of the intense heating during the discharge. Similarly, several observations were made of the continual liberation of hydrogen from electrodes. Winchester, who investigated this phenomenon in detail, using thin aluminum electrodes, found that hydrogen is liberated until the electrodes are completely sputtered away. It is very difficult to suppose that in this case hydrogen is a product of the transformation of aluminum; one need only recall the great energy of the $\alpha$-particle required to cause such a transformation. As in the case of helium, it seems more probable that the hydrogen had originally been absorbed by the electrodes.

Although it is dangerous in the present case to be dogmatic, the general indications point to the fact that atoms, as a rule, possess so stable a structure, and their nuclei are held together by such powerful forces, that only $\alpha$-particles, as the most concentrated sources of energy, are suitable for attacking these well-protected structures. Even then, when splitting occurs, it takes place on so insignificant a scale that only a few $\alpha$-particles out of a million are effective. If we had at our disposal charged atoms with an energy ten times greater than the energy of the $\alpha$-particle of radium, we could probably penetrate the nuclear structure of all atoms and sometimes even bring about their destruction.

Translated by E. Shtolsky.

1. And, obviously, in addition, one electron, since the nuclear charge of nitrogen is equal to seven. — Translator. ↩↩↩↩↩

2. Indeed, the relation between mass \((m)\) and energy \((E)\) has, as is known, the following form:

   \[ E = mc^{2}, \]

   where \(c\) is the velocity of light. Thus in our case the energy released in the formation of the helium nucleus will be

   \[ \Delta E = (\Delta m)c^{2} = 0.029\,c^{2}. \]

   Let us now calculate in gram-atoms the kinetic energy of a radium \(\alpha\)-particle \(C\) (of velocity \(\sim 2 \cdot 10^{9}\ \frac{\text{cm}}{\text{sec.}}\)):

   \[ E=\frac{1}{2}mv^{2}=\frac{4}{2}\left(\frac{2}{30}\right)^{2}c^{2}=0.009c^{2}. \]

   Thus \(\dfrac{\Delta E}{E} \approx 3.2\). The thermochemical equation for the formation of helium will have the form:

   \[ 4H+2El=He+Q. \]

   If the “heat of formation” of a gram-atom is converted into large calories, then, obviously, one obtains

   \[ Q=\frac{0.029c^{2}}{4.19\cdot 10^{10}}=6.25\cdot 10^{9}\ \text{Cal}. \]

   The magnitude is enormous, if one recalls that the thermochemical effects of ordinary reactions are of the order of \(100\ \text{Cal}\). (A. Sommerfeld, Atombau, p.).
   Translator. ↩↩