Biography of the Alpha Particle1

Sir Ernest Rutherford.

In this lecture I propose to speak about some of the properties of the swift $\alpha$-particle, spontaneously ejected by radioactive substances. This flying atomic nucleus is not only the most powerful of the projectiles known to us, but it is also the mightiest instrument for investigating the structure of atoms. Therefore an acquaintance with the phenomena caused by it is of deep scientific interest.

It is now definitively established that the $\alpha$-particle emitted by radioactive bodies is in every case a helium atom, or, more precisely, the nucleus of a helium atom of mass 4, carrying two positive charges of electricity. And only when the ejected nucleus, in passing through matter, is slowed down and captures two negative electrons—only then does it become a neutral helium atom. It is quite natural to suppose that the helium nucleus, hurled with enormous speed from the heavy nuclei of radioactive atoms, is a constituent part of them. For certain reasons, still not understood, the radioactive nucleus quite accidentally breaks up with an explosion, ejecting in the process the helium nucleus that forms part of it with enormous speed. It is possible that the $\alpha$-particle, on being liberated from the radioactive nucleus, receives part of its enormous kinetic energy while passing through the repelling electric field surrounding the radioactive nucleus, but at present we know neither the nature of the forces binding the separate parts of the nucleus nor whether the $\alpha$-particle in the nucleus is at rest or in orbital motion. We do know, however, that the stability of the nucleus of different radioactive elements varies within very wide limits. In the case of the substance known as radium A, the radioactive atom before ejecting an $\alpha$-particle has a mean lifetime of 4.3 minutes, whereas radium itself has a lifetime of 2,250

years; and in the case of an extremely slowly changing element, say uranium, the average lifetime is over 7,000 million years.

It is known that all \(\alpha\)-particles of a given element are ejected with one and the same velocity, and that this velocity varies from element to element. Obviously, there exists a close connection between the velocity of ejection of the \(\alpha\)-particle and the mean lifetime of the parent element. The shorter the mean lifetime of the element, the greater the velocity of emission. This curious relation between the force of the explosion and the mean lifetime of the element holds in most cases, but at the present time it is difficult to give oneself an account of all the details of this phenomenon. Sir William Bragg showed quite long ago that an \(\alpha\)-particle runs through matter approximately in a straight line and travels in it a perfectly definite distance. This is very well illustrated by the tracks of \(\alpha\)-particles obtained by Wilson’s expansion method. Most of the tracks appear as straight lines; only in some of them are bends observed near the end of the path. Obviously, at the end of the path the photographic and ionizing action of the \(\alpha\)-particle suddenly ceases. Owing to its enormous kinetic energy an individual \(\alpha\)-particle can be detected by the scintillation it produces in crystalline zinc sulphide, or by its action on a photographic plate, or by a special electrical method; and Wilson’s excellent method shows the path of each separate \(\alpha\)-particle as it passes through a gas.

We, in particular, adapted the scintillation method to the counting of individual particles, and thus we had at our disposal a very delicate method for studying the phenomena arising when \(\alpha\)-particles pass through matter. In flying through a gas, an \(\alpha\)-particle passes through the electronic systems of an enormous number of atoms and, liberating electrons, gives rise to strong ionization along its path. The ionization reaches a maximum near the end of the path of the \(\alpha\)-particle and then rapidly falls to zero.

The law of the decrease in the velocity of an \(\alpha\)-particle in passing through matter was carefully investigated by studying the deflection in a magnetic field of a beam of \(\alpha\)-particles before and after their passage through a layer of matter of definite thickness. In most of these experiments we used the \(\alpha\)-particles of radium C, which have a range of about \(7\ \mathrm{cm}\) in air under normal conditions. The initial velocity, \(V_0\), of these particles, as is known, is \(19\,200\ \mathrm{km}\) per second, and it gradually decreases to \(0.41V_0\). In this state the observed range of the \(\alpha\)-particle is less than one centimeter; measurements in this region are very difficult, because the beam of \(\alpha\)-particles becomes non-uniform and contains particles moving with different velocities.

In view of this, a velocity of the α-particle smaller than \(0.38 V_0\) could not be determined with sufficient certainty. We must note that, even at the lowest velocity at which it is still possible to detect an α-particle by means of scintillations or photographically, it nevertheless moves with a velocity much greater than the velocity of a positively charged particle arising in an ordinary discharge tube.

It is known that, at the end of its range, an α-particle slows down so much that it captures electrons and turns into a neutral atom; but until recently we had no proof of the existence of this phenomenon of electron capture. Recently G. Henderson (Proc. Roy. Soc. A, 102, p. 496, 1922) made a considerable advance in this field by studying the deflection of α-rays in a magnetic field at a very good vacuum. For these experiments to be successful it is necessary that the apparatus in which the deflection is observed be pumped down to a very low pressure, corresponding to the pressure of air in a good tube for X-rays. We shall see below the reason for this. When a narrow beam of α-rays was deflected by a magnetic field, two bands were observed on the photographic plate: the main band, corresponding to ordinary α-particles carrying two positive charges, and a second band1, which, according to Henderson’s suggestion, corresponds to particles that have captured one electron, i.e. to helium atoms having one positive charge. Moreover, at small velocities he succeeded in proving the existence of neutral α-particles, produced when a helium nucleus captures two electrons. In these experiments Henderson used platinocyanide plates, in which the film is so thin that particles at small velocities produce the same or even a greater photographic effect than particles with high velocities.

I repeated these experiments by the scintillation method and confirmed Henderson’s conclusions. In observing the deflection of the middle band in electric and magnetic fields, I found that, undoubtedly, the particles producing it have mass 4 and charge 1, i.e. they are singly charged helium atoms, having the same velocity as the particles with two charges that form the main band.

At the same time I carried out several experiments to clarify the conditions under which flying α-particles can capture or lose an electron. The general arrangement of the experiment is shown in Fig. 1 (p. 190). A thin platinum wire was coated with radium \(B + C\) by the method of exposure in emanation; I used it as an almost homogeneous source of α-rays, for the α-particles were emitted in this case only by radium \(C\) atoms, the number of which was so negligible that they could hardly form on the platinum a film with a thickness of one mo—

molecule. The $\alpha$-rays from this source passed through a narrow slit about $0.3\ \mathrm{mm}$ wide and fell on a zinc-sulfide screen. The distribution of $\alpha$-particles on the screen was determined by the scintillation method in a dark room, with the aid of a microscope placed outside the box. The vessel containing the source and the screen was evacuated completely by means of a Gaede diffusion pump; when necessary, the residual pressure was measured with a McLeod manometer. The box was placed between the flat poles of a large electromagnet in such a way that the beam of $\alpha$-rays was bent in the direction shown in Fig. 1. Usually the distance between the source and the screen was $16\ \mathrm{cm}$, the slit being placed midway between them.

Fig. 1.

The rays traversed almost their entire path in a uniform magnetic field, and thus the deflection of the beam of rays was proportional to the strength of the magnetic field. Under normal experimental conditions the beam of $\alpha$-rays from the wire coated with radium C was deflected on the screen by approximately $15\ \mathrm{mm}$ from the zero position (i.e., from the position in the absence of a field). The field of view of the microscope was sufficient to see the base of the entire beam of $\alpha$-rays in the absence of a magnetic field.

Special precautions were taken to eliminate contamination of the screen by active substance released from the wire at low pressures. It should be borne in mind that a source of this type always produces a certain inhomogeneity in the beam of $\alpha$-rays. Such inhomogeneity is produced by $\alpha$-particles emitted from the rear side of the wire, which, passing through the material of the wire, reduce their velocity. This phenomenon is clearly observed in deflection

beam of α-rays by a magnetic field: alongside the principal band of α-rays there are always particles distributed outside the main beam. But the intensity of these inhomogeneities of the beam is usually less than one percent of the intensity of the main beam, and their presence does not seriously affect the accuracy of the conclusions reported in this lecture.

Figs. 2 and 3 show the distribution of singly and doubly charged α-particles on a zinc-sulfide screen. Fig. 2 shows the result obtained when mica is placed in front of the source with a thickness corresponding to the stopping power of a layer of air of 3.5 cm. The main band, corresponding to particles \(He_{++}\), is sharply bounded on the side of the higher velocities, but it also

Fig. 2.

indicates a certain inhomogeneity of the beam produced as it passes through the mica. As we expected, the middle band (particles \(He_+\)) indeed lies exactly between zero and the main band and contains approximately \(1/55\) of the particles of the main beam. Fig. 3 gives the distribution when the thickness of the mica layer is increased to a value corresponding to the stopping power of air of 6 cm. Now both—the main and the middle—bands are no longer outlined as sharply as in the first case, but each of them is formed by particles with very different velocities. The relative number of particles \(He_+\) and \(He_{++}\) is approximately \(1/8\) for the fast particles, but this ratio increases as the velocity decreases. The middle band broadens and merges with the main band, where the latter can no longer be observed. The brightness of the scintillations corresponding to particles \(He_+\) decreases continuously from \(A\) to \(B\). In this case there aris-

also neutral particles. This is indicated by the \(He_0\) band, which is not deflected in the magnetic field, but its intensity is small in comparison with the middle band. Moreover, between the neutral and the middle band there occurs a small number of weak scintillations, in all probability belonging partly to the scattered edges of the beam of \(\alpha\)-particles, partly to oxygen atoms or to other elements entering into the composition of mica. The distribution of charged and neutral helium particles at considerably lower velocities is seen in curves \(A\) and \(B\) (Fig. 4), the analysis of which we shall postpone for the present. It is clear that the ratio of the number of particles \(He_+\) to \(He_{++}\) has here increased, and likewise the relative number of neutral particles is much greater.

Fig. 3.

We can now turn to an explanation of all these observations. It is clear that the particles emerging from mica will be either with two charges, with one charge, or else completely neutral, but the relative number of these three types of particles changes noticeably depending on the stopping power of the mica plates. We may suppose that an \(\alpha\)-particle, in passing through the external electronic structure of an atom, on its way accidentally pushes out an electron and captures it. This electron, falling into a stable orbit around the helium nucleus with double charge, moves together with it.

However, this atom with one charge has a very short lifetime: in passing through other atoms the electron is knocked off, and the \(\alpha\)-particle with one charge again returns to the type of particles with two charges. This process of losing an electron is analogous to the usual phenomenon of ionization, where the electron is knocked out of the atom when it collides with an \(\alpha\)-particle. In exactly the same way, a particle with one charge

He^+ can capture an electron from another atom and, conversely, it can accidentally lose its “extra” electron. Thus, we must distinguish two opposite processes: one consisting in the capture of an electron, and the other leading to its ejection. From the data that we shall present below it will be seen that this process of capture and loss can be repeated more than a thousand times during the flight of an α-particle, so that the mean path traversed by an α-particle before capturing an electron or before losing a captured electron is small in comparison with the entire distance traversed by the α-particle before stopping. From this it is clear that, at a given velocity of the α-particle, there must exist an equilibrium between the number of particles He^+ and He^{++}, i.e., on the average the number of electrons captured over a given flight distance must be equal to the number lost.

Fig. 4.

It is very convenient to suppose that, at a given velocity, each particle He^{++} has a certain mean free path \lambda_1 cm in the substance before it captures some electron, and that the mean free path of a particle He^+ before the loss of an electron is \lambda_2 cm. Without doubt, individual particles travel greater or smaller distances than these mean distances before capturing or losing an electron, but, taking into account a large number of particles, we may assume that there exists a mean distance traveled before the capture or loss of an electron, which we may call the mean free path.

If N_1 is the number of particles He^{++} traversing a small distance dx in the substance, then the number of them capturing electrons will be N_1 dx/\lambda_1. If likewise N_2 is the number of particles He^+, then the number of them losing one electron is equal to N_2 dx/\lambda_2. But we have already seen that when equilibrium sets in, the number of electron captures over a given distance must be ...

be equal to the number of losses. From comparison of these two expressions it is clear that \(N_2/N_1=\lambda_2/\lambda_1\), or, in other words, the relative number of particles \(He_+\) and \(He_{++}\) is proportional to the ratio of the mean free paths of losses and captures. Since, by the scintillation method, the ratio \(N_2/N_1\) can be measured for any velocities, using absorbers of various thicknesses for this purpose, we can in this way determine, for a given velocity, the ratio of the mean free paths before capture and loss of an electron.

The true value of the mean free path \(\lambda_2\) of the particle \(He_+\) before loss of its electron can be directly determined experimentally. Suppose that the microscope is focused on the middle strip (see Fig. 2), and that we count the number of scintillations in one minute at a good vacuum. If the pump is suddenly stopped and air, or another gas, is introduced into the vessel in small portions, then, as was found, the number of scintillations decreases with increasing pressure until the strip disappears completely. This occurs already at a sufficiently low pressure; for air, for example, at a pressure of about \(1/4\) mm.

The explanation of this result is entirely clear. A particle \(He_+\), flying out of the mica, accidentally collides with a gas atom encountered on its path, and in doing so gives up the electron which it had captured while flying through the mica. In this way the particle \(He_+\) again turns into a particle \(He_{++}\), and the latter is deflected in the magnetic field twice as much as the particle \(He_+\). Suppose that the collision took place at point \(P\) (Fig. 1). Then, after giving up the electron, the particle will follow the new path shown in Fig. 1 and will no longer strike the part of the screen visible in the microscope. It was found that the number of scintillations visible in the microscope decreases exponentially with increasing pressure. This result was not unexpected, and from the data of the experiment it was possible, quite simply, to calculate the mean distance traversed by the particle \(He_+\) before it loses the electron captured by it. Of course, some necessary corrections had to be introduced in calculating the final width of the scintillation strip visible in the microscope, but we shall not go into the details of this experiment. It is very convenient to express the mean free path \(\lambda_2\) in air of the particle \(He_+\) not by the mean length of the path traversed in the rarefied gas before loss of the electron, but by the distance traversed by it in the same gas at normal pressure and temperature. For example, in one of the experiments the mean free path in air for the particle was found to be 12 cm at a pressure of 0.040 mm; this corresponds to a mean free path of 0.0063 mm at normal pressure and temperature.

Thus, the mean free path in air before loss of an electron was measured at different velocities, and it was found that the mean free path changes with change in the velocity of the \(\alpha\)-parti-

BIOGRAPHY OF THE ALPHA PARTICLE

particles, so that as the velocity of the $\alpha$-particle decreases, its mean free path becomes shorter. Since the loss of an electron by a particle carrying one charge may be regarded as the result of ionization, there is no doubt that, if we take into account the strong bond of the individual electron with the nucleus $He_{++}$, the mean free path before the loss of an electron will be of the same order as that calculated from the number of ions produced per $1\ \mathrm{cm}$ during the flight of an $\alpha$-particle in air or in some other gas. In this connection the mean free paths in air were compared with the paths in hydrogen and helium. This quantity in hydrogen is 4 to 5 times longer than in air, and it is 5 times longer than in helium.

Now, knowing the mean path length $\lambda_2$, we can calculate $\lambda_1$ (before electron capture), if the ratio $N_2/N_1$ is also known to us. However, a difficulty arises here, for in order to measure the ratio $N_2/N_1$ it is necessary that the active source be covered with mica or some other solid body. Gas is not convenient for this purpose. It was found, however, that the ratio $N_2/N_1$ remains the same, within the limits of observational error, regardless of whether we reduce the velocity of the $\alpha$-particles by celluloid, mica, aluminum, or silver. In doing this we used one and the same mica plate, each time covering it with a very thin layer of the substance under investigation. The thickness of the layer was sufficient to establish a new equilibrium between particles carrying one and two charges, but insufficient to produce a substantial change in the velocity of the ionizing rays.

• Since the value of the ratio $N_2/N_1$ gives no appreciable changes for absorbers with such different atomic weights, we may quite safely conclude that this ratio for a hypothetical layer of solid air will be the same as for the given layer of mica.

Thus, we now have all the data necessary for determining the quantities $\lambda_1$ and $\lambda_2$ relating to $\alpha$-particles at different velocities. The results are given in the following table for three different velocities. The mean free path is expressed in millimeters of air at normal pressure and temperature. The maximum velocity $V_0$ of the $\alpha$-particles of radium $C$ is equal to $1.9 \times 10^9\ \mathrm{cm/sec}$.

From the table it is evident that the mean free path before loss of an electron changes parallel to the velocity, and moreover it changes approximately in the ratio 1 to 2 within the limits of change of the velocities given in the table. On the other hand, the ratio \(\lambda_2/\lambda_1\) increases very rapidly with decreasing velocity, varying approximately proportionally to \(V^{-5}\). Hence it follows that \(\lambda_1\) changes proportionally to \(V^6\), decreasing by 60 or more times while the velocity is reduced by half.

From these data and relations it may easily be calculated that the mean free path before capture of an electron must be equal to the path before its loss at a velocity equal to \(0.3 V_0\), and that at this velocity the numbers of particles \(He_+\) and \(He_{++}\) must be equal. A special experiment showed that the actual value of the velocity at which the quantities of particles of both kinds are identical is \(0.29 V_0\), i.e. it agrees very well with the calculated value. It is exceedingly difficult to determine the quantities \(\lambda_1\) and \(\lambda_2\) for velocities smaller than \(0.3 V_0\), not only because the scintillations are weak in intensity and are difficult to subject to exact counting, but also because the rays produced are very inhomogeneous and do not give a sharply outlined edge on the side of the greater velocity. It was observed, however, that at velocities smaller than \(0.3 V_0\) the ratio \(N_2/N_1\) increases rapidly.

Up to now we have studied the equilibrium between the particles \(He_+\) and \(He_{++}\). It is clear that, at small velocities of the \(\alpha\)-particles, considerations similar to those set forth above are applicable to the equilibrium between neutral helium particles and particles carrying a single charge. It was found that neutral particles become particularly conspicuous when the rays pass through a layer of mica with a retarding power of \(6\ \mathrm{cm}\), but, without doubt, they can be detected at a much smaller retarding power. These neutral particles cause scintillations, in intensity corresponding to \(\alpha\)-particles of small velocity. In all probability, before striking the zinc sulphide or any other absorbing substance, these neutral particles many times lose and capture an electron. This phenomenon was shown by admitting gas at low pressure into the apparatus, whereby the scintillations caused by the neutral particles gradually diminished in number and finally disappeared completely. This phenomenon is explained in the same way as the disappearance of the \(He_+\) streak: the neutral particles accidentally lose an electron as they pass through the gas, and, as a result, the magnetic field deflects them from their zero position.

Calculation shows that the mean free path in air between successive neutral states of the helium particle is, in order of magnitude, approximately \(1/600\ \mathrm{mm}\). But, of course, this quantity is an average for particles with various velocities, entering the neutral streak at a given moment.

For high velocities we have a moving equilibrium

$He_{++} \rightleftarrows He_+$. For velocities less than $0.5 V_0$, $He_+ \rightleftarrows He_0$; in all probability, the same holds for velocities less than $0.3 V_0$. As Henderson showed, at extremely low velocities most $He_{++}$ particles are destroyed, and the particles $He_+$ and $He_0$ predominate.

At such low velocities the counting of scintillations is extremely difficult and unreliable, and in this case the photographic method used by Henderson is preferable. It would be very interesting to see whether the relative number of particles of the three types changes when $\alpha$-particles move slowly through various substances. This part of the work was carried out by Henderson at the University of Saskatchewan.

Here it is necessary to note one very interesting circumstance. It had already been shown that singly and doubly charged $\alpha$-particles always appear after the passage of $\alpha$-rays through mica or other absorbing substances; but do singly charged particles appear when $\alpha$-particles fly out of a wire covered with an infinitely thin layer of active substance? Such a phenomenon was in fact discovered for a platinum wire covered with a deposit of radium $B+C$, obtained by emanating radium, and it was established that singly charged helium atoms are obtained in approximately the equilibrium ratio for the given velocity. This observation was quite unexpected. Initially we explained it by the fact that particles of radium $B$, arising from radium $A$ as a result of recoil, penetrate to some depth into the material of the wire. Under these conditions many of the $\alpha$-particles emitted by radium $C$ pass through small but appreciable thicknesses of the layer of substance before flying out of the wire, and on the way may capture an electron. However, this explanation seems improbable because the mean distance penetrated by the atom produced in recoil constitutes only a small part of the mean free path before electron capture for an $\alpha$-particle at such high velocities. The experiment was repeated with a nickel wire, on the surface of which a deposit of radium $C$ was obtained by the well-known method: immersing the wire in a hot solution of radium $C$. In this case there are no recoil phenomena, but the number of singly charged particles was the same as in the first experiment.

It is very important that the relative number of singly and doubly charged particles is approximately equal to the equilibrium ratio obtained when, after activation, the wire is covered with an appreciable thickness of a layer of copper or of some other substance. It can hardly be thought that singly charged particles, along with doubly charged ones, are actually released from the radioactive nucleus; even if it is supposed that an $\alpha$-particle is thrown out with a companion electron, the latter may be lost upon passing through the ...

through a very strong electric field near the nucleus. In all probability, an $\alpha$-particle with a double charge, in passing through the densely distributed electrons surrounding the radioactive nucleus, accidentally captures an electron, and thus the process of capture and loss of an electron goes on throughout the entire time of the flight of the $\alpha$-particle out of the radioactive atom. At first glance such an explanation seems less plausible, if we take into account the relatively large number of atoms usually traversed by an $\alpha$-particle before equilibrium is established between captures and losses of electrons; but, on the other hand, it is well known that the chances of electronic collisions are in general much greater for charged particles emitted from a central nucleus than for particles from outside passing through the electronic atmosphere of an atom. It is possible that those electrons whose orbital motion around the nucleus is comparable with the velocity of the $\alpha$-particle play a special role in cases of capture or loss.

Until now we have been dealing chiefly with the distribution of particles in a magnetic field in vacuum after their release from the surface of mica. Several very interesting facts emerged in the study of the distribution in the presence of a sufficient quantity of gas, which caused a rapid alternation of captures and losses of electrons along the path of the $\alpha$-particle. This is very well illustrated by the diagram of Fig. 4, on which are plotted the results for $\alpha$-particles emerging from mica with a maximum residual range of about 4 or 5 mm in air. Curves $A$ and $B$ give the approximate distribution of $He_+$ and $He_{++}$ particles in vacuum, whereas $C$ gives the relative number of neutral particles under the given experimental conditions. Suppose now that air is introduced into the vessel in an amount sufficient to cause numerous captures of electrons during the flight of the $\alpha$-particles in the gas, but still insufficient to produce a noticeable decrease in the velocity of the $\alpha$-particles. Then what first of all catches the eye is the remarkable fact that the distribution according to $A$, $B$, $C$ disappears and there remains only a distribution of particles approximately midway between $A$ and $B$ (curve $D$). This band, in comparison with $A$ and $C$, is narrower, and at its maximum higher. This indicates that the particles are collected into a band narrower than occurs in the normal distribution in curve $B$.

This is precisely what we may expect. Indeed, the fast particles, as compared with the slow ones, undergo a smaller number of captures; consequently, the mean charge of the fast $\alpha$-particles in the gas will be less than $2e$, and their deflection less than the deflection of the faster particles shown on curve $B$. On the other hand, the mean charge of the slow $\alpha$-particles is closer to $1e$ than to $2e$, and therefore their relative deflection will be less than that of the fast particles. Thus it is clear that the total distribution of the particles in air inside the vessel will be concentrated in a narrower band than the main

of the band of \(He_{++}\) particles. From calculations based on the laws of capture and loss, the width of the band under the experimental conditions can be computed and, in fact, has been found to be in excellent agreement with experiment. Let us note that similar results have also been obtained for hydrogen under the corresponding conditions.

GENERAL DISCUSSION OF THE RESULTS.

Our attention must now be turned to an analysis of the results obtained and to a possible explanation of them from the modern point of view. First of all, there is the strikingly large number of captures and losses of an electron occurring during the flight of a radium \(C\) \(\alpha\)-particle. Whereas the mean free path of a radium \(C\) \(\alpha\)-particle with a range of \(7\ \mathrm{cm}\) in air is approximately \(3\ \mathrm{mm}\), with decreasing velocity of the \(\alpha\)-particle this quantity rapidly falls, and, in all probability, at a velocity \(0.3 V_0\) it is equal to \(0.0015\ \mathrm{mm}\). It is not difficult to calculate that the charge of the particle during its flight in air changes about a thousand times while its velocity changes from \(V_0\) to \(0.3V_0\). Although the data obtained so far do not allow us to calculate the number of changes of charge occurring as the velocity decreases from \(0.3V_0\) to 0, it nevertheless seems probable to us that this number is considerably greater than a thousand. We have already noted that at small velocities the changes \(He_+ \rightleftarrows He_0\) predominate. If we take into account the rapidity of the changes of charge of \(\alpha\)-particles at intermediate velocities, it is clear that we cannot hope to observe any appreciable difference in penetrating power between beams of rays of a given velocity, depending on whether they originally consisted of singly charged or doubly charged particles. It is quite clear that a singly charged particle, after penetrating a small distance into a substance, becomes doubly charged, and conversely; and it is clear that the effects produced by the two beams cannot be distinguishable. Henderson, using the photographic method, is attempting to perform such experiments on the absorption of particles, but so far he has not obtained definite results.

When an \(\alpha\)-particle captures an electron, the latter probably falls into the very orbit around the helium nucleus which characterizes an ionized helium atom, i.e. an atom that has lost one electron. When an \(\alpha\)-particle with its companion electron rapidly traverses atoms of a gas on its path, it not only ionizes the gas, but may itself also be ionized, i.e. lose its companion electron. If we take into account the strong binding of the first electron with the helium nucleus—the ionization potential is about 54 volts—then the mean free path for losses of the captured electron in air, in magnitude, will be of the same order as that calculated from the assumption of ionization by \(\alpha\)-particles per unit path in air. Whereas

is deflected through a very large angle as a result of a close collision with the nucleus. I have also omitted an even rarer case, when the disintegration of the atomic nucleus occurs, as happens with nitrogen or aluminium. Thus we see that the $\alpha$-particle has an interesting history. Usually it is retained in a radioactive nucleus as a constituent part of it for intervals of time of many thousands of millions of years. But suddenly a catastrophe occurs: the $\alpha$-particle gains its freedom and lives an independent life for about one hundred-millionth of a second, during which everything described in this lecture happens to it.

If we are dealing with dense and solid minerals of uranium or thorium, then the $\alpha$-particle, after acquiring two electrons and becoming a neutral helium atom, is retained in the mineral for as long as that mineral exists. Occluded helium can be obtained from the mineral by the action of high temperature and after the removal of all other gases; its presence can be detected by the characteristic bright glow under the influence of an electric discharge. In such experiments a small quantity of helium is liberated. However, enormous quantities of helium, sufficient for filling large airships, are obtained from natural gases freely issuing from the earth in various parts of Canada and the United States of America. And it is remarkable that, in all probability, each individual atom of this substance has had the life history described here.

APPENDIX1

It seems of interest to give here a brief survey of a few more facts concerning the $\alpha$-particle. It has long been known that $\alpha$-particles, although emitted from a source with one and the same speed, nevertheless travel different distances in a gas. For example, the greatest distance traversed by the $\alpha$-particles of radium C in air at a pressure of 760 mm and $15^\circ$C is $7.04$ cm, the smallest about $6.4$ cm; hence the mean will be about $6.8$ cm. A certain “longitudinal scattering” of $\alpha$-particles can be foreseen on general grounds. Indeed, an $\alpha$-particle loses its energy chiefly by liberating electrons from the atoms of the substance encountered on its path. Meanwhile, according to the laws of probability, one $\alpha$-particle may encounter a larger number of atoms and liberate more electrons than some other particle, and in this way the former loses energy more rapidly than the latter. However, the observed longitudinal scattering is much greater than can be obtained by a calculation made on this assumption, and it is necessary

It should be noted that large deflections of $\alpha$-particles, corresponding to collisions with nuclei, are, except at the end of the range, so rare that they cannot seriously affect the final distribution of $\alpha$-particles.

Henderson suggested that the property of an $\alpha$-particle to capture and lose electrons should be regarded as a new factor in longitudinal scattering. This phenomenon undoubtedly occurs, but the observed ratio of electron capture to loss seems too large to allow all the discrepancy between theory and experiment to be attributed to it.

Another interesting suggestion concerning the magnitude of this scattering was made by P. L. Kapitsa. From Chadwick and Bieler’s experiments on collisions between $\alpha$-particles and hydrogen nuclei it was clear that the $\alpha$-particle, or the helium nucleus, possesses a nonsymmetrical field of forces. This asymmetry of the electric field may already be small at distances of the electron orbits in a neutral atom, but it may be sufficient to fix the plane of the electron’s orbit relative to the axis of the helium nucleus.

Let us suppose that the $\alpha$-particles emitted by a radioactive source have axes oriented in space, and that the direction of the axes of each individual particle remains unchanged during its motion. For example, in some cases the captured electron describes an orbit whose plane is close to the direction of motion; in others it is almost perpendicular to it. However, one must expect that the chances of losing the captured electron in a collision will in one case be greater than in the other; or, in other words, the mean free path of an $\alpha$-particle carrying a single charge before losing an electron will not be the same in the two cases.

From this point of view one may suppose that one part of the $\alpha$-particles will lose energy more rapidly than another, and that their ranges will likewise be different. To test whether such a difference between $\alpha$-particles, predicted by the theory, really exists, Kapitsa, in the Cavendish Laboratory, photographed the paths of a large number of $\alpha$-particles, using Wilson’s method and a strong magnetic field, approximately 70,000 gauss, produced by an instantaneous current of great strength. The magnetic field was strong enough to produce a noticeable curvature of the path of the $\alpha$-particle. It was found that the curvature of the paths at equal distances from their end shows a noticeable difference. Before drawing any definite conclusion, it is necessary to obtain a large number of tracks, to measure them carefully, and to separate out the sudden deflections caused by collisions of the nucleus with nitrogen or oxygen atoms. The frequency of such deflections near the end of the range complicates the interpretation of the curvature being measured. The experiments, which are still in progress, are extremely difficult and require colossal technical dexterity; they will be of great interest if, by one method or another, they establish a definite asymmetry for the orbits in singly charged $\alpha$-particles.

1. This appendix is not part of the lecture delivered at the Royal Institution, but it may usefully supplement certain points in this lecture. ↩↩↩