THE NATURAL
AND ARTIFICIAL DISINTEGRATION
OF THE ELEMENTS




AN ADDRESS BY

Professor Sir ERNEST RUTHERFORD

Kt., D. Sc., LL. D., Ph. D., D. Phys., F. R. S.




ON THE OCCASION OF THE CENTENARY CELEBRATION
OF THE FOUNDING OF

THE FRANKLIN INSTITUTE

AND THE INAUGURATION EXERCISES OF THE
BARTOL RESEARCH FOUNDATION
SEPTEMBER 17, 18, 19, 1924




THE FRANKLIN INSTITUTE

PHILADELPHIA




THE NATURAL
AND ARTIFICIAL DISINTEGRATION
OF THE ELEMENTS


_By_ Professor Sir ERNEST RUTHERFORD, Kt., D. Sc.,

LL. D., Ph. D., D. Phys., F. R. S.


IT is not my intention in this paper to give a detailed account of the
natural disintegration of the radio elements or of the methods employed
to effect the artificial disintegration of certain light elements. I
shall assume that you all have a general knowledge of the results of
these investigations, but I shall confine myself to a consideration of
the bearing of these results on our knowledge of the structure of the
nuclei of atoms.

There is now a general agreement that the atoms of all elements have a
similar electrical structure, consisting of a central positively
charged nucleus surrounded at a distance by the appropriate number of
electrons. From a study of the scattering of _α_ particles by the atoms of
matter and from the classical researches of Moseley on X-ray spectra, we
know that the resultant positive charge on the nucleus of any atom, in
terms of the fundamental unit of electronic charge, is given numerically
by the atomic or ordinal number of the element, due allowance being made
for missing elements. We know that with few exceptions all nuclear
charges, from 1 for the lightest atom, hydrogen, to 92 for the heaviest
element, uranium, are represented by elements found in the earth. The
nuclear charge of an element controls the number and distribution of the
external electrons, so that the properties of an atom are defined by a
whole number, representing its nuclear charge, and are only to a minor
degree influenced by the mass or atomic weight of the atom.

This minute but massive nucleus is, in a sense, a world of its own which
is little, if at all, influenced by the ordinary physical and chemical
forces at our command. In many respects, the problem of nuclear
structure is much more difficult than the corresponding problem of the
arrangement and motions of the planetary electrons, where we have a
wealth of available information, both physical and chemical, to test the
adequacy of our theories. The facts known about the nucleus are few in
number and the methods of attack to throw light on its structure are
limited in scope.

It is convenient to distinguish between the properties assigned to the
nucleus and the planetary electrons. The movements of the outer
electrons are responsible for the X-ray and optical spectra of the
elements and their configuration for the ordinary physical and chemical
properties of the element. On the other hand, the phenomena of
radioactivity and all properties that depend on the mass of the atom are
to be definitely assigned to the nucleus. From a study of the
radioactive transformations, we know that the nucleus of a heavy atom
not only contains positively charged bodies but also negative electrons,
so that the nuclear charge is the excess of positive charge over
negative. In recent years, the general idea has arisen that there are
two definite fundamental units that have to do with the building up of
complex nuclei, viz., the light negative electron and the relatively
massive hydrogen nucleus which is believed to correspond to the positive
electron.

This view has received very strong support from the experiments of Aston
on Isotopes in which he has shown that the masses of the various species
of atoms are represented nearly by whole numbers in terms of O = 16.
From the general electric theory, it is to be anticipated that the mass
of the hydrogen nucleus in the nucleus structure will be somewhat less
than its value 1.0077 in the free state on account of the very close
packing of the charged units in the concentrated nucleus. From Aston's
experiments, it appears that the average mass of the hydrogen nucleus,
or proton as it is now generally called, is very nearly 1.000 under
these conditions. We should anticipate that the whole number rule found
by Aston would hold only to a first approximation, since the mass of the
proton must be to some extent dependent on the detailed structure of the
nucleus. In the case of tin and xenon Aston has already signalized a
definite departure from the whole number rule, and no doubt a still more
accurate determination of the masses of the atoms will disclose other
differences of a similar kind.

While our present evidence indicates that the proton and electron are
the fundamental constituents of the nucleus, it is very probable that
secondary combining units play a prominent part in nuclear constitution.
For example, the expulsion of helium nuclei from the radioactive bodies
indicates that the helium nucleus of mass 4 is probably a secondary unit
of great importance in atom building. On the views outlined, we should
expect the helium nucleus of charge to be built up of four protons and
two electrons. The loss of mass in forming this nucleus indicates that a
large amount of energy must be liberated during its formation. If this
be the case, the helium nucleus must be such a stable structure that the
combined energy of four or five of the swiftest _α_ particles would be
necessary to effect its disruption. Such a deduction is supported by our
failure to observe any evidence of disintegration of the swift particle
itself, whether it is used to bombard matter or whether the _α_ particle
is used to bombard other helium atoms.

On these views, we should anticipate that the nucleus of radium of
atomic number 88 and atomic weight 22.6 contains in all 226 protons of
mass 1 and 138 electrons. While this gives us the numerical relation
between the two fundamental units, we have, at present, no definite
information of their arrangement in the minute nuclear volume, nor of
the nature and magnitude of the forces that hold them together. We
should anticipate that many of the protons and electrons unite to form
secondary units, _e. g._ helium nuclei, and that the detailed structure
of the nucleus may be very different from that to be expected if it
consists of a conglomeration of free protons and electrons.

It is thus of great importance to obtain definite evidence of the nature
and arrangement of the components of the nucleus and of the forces that
hold them in equilibrium. We shall now consider some of the lines of
evidence which throw light on the actual dimensions of the nucleus and
the law of force operative in its neighborhood; the structure and modes
of vibration of the nucleus, together with the effects observed when
some light nuclei are disintegrated by bombardment with _α_ particles.




DIMENSIONS OF THE NUCLEI AND THE
LAW OF FORCE


The conception of the nucleus atom had its origin in 1911 in order to
explain the scattering of an _α_ particle through a large angle as the
result of a single collision. The observation that the _α_ particle is in
some cases deflected through more than a right angle as the result of an
encounter with a single atom first brought to light the intense forces
that exist close to the nucleus. Geiger and Marsden showed that the
number of particles scattered through different angles was in close
accord with the simple theory which supposed that, for the distance
involved, the _α_ particle and nucleus behaved like charged points,
repelling each other according to the law of the inverse square. The
accuracy of this law has been independently verified by Chadwick, so
that we are now certain that in a region close to the nucleus the
ordinary laws of force are valid.

These scattering experiments also gave us the first idea as to the
probable dimensions of the nuclei of heavy atoms, for it is to be
anticipated that the law of the inverse square must break down if the _α_
particle approaches closely to or actually enters the nuclear structure.
This variation in the law of force would show itself by a difference
between the observed and calculated numbers of _α_ particles scattered
through large angles. Geiger and Marsden, however, observed no certain
variation even when the _α_ particles of range about 4 cms. were
scattered through 100° by a gold nucleus. In such an encounter, the
closest distance of approach of the _α_ particle to the center of the
nucleus is about 5 x 10^-12 cm., so that it would appear that the radius
of the gold nucleus, assumed spherical, could not be much greater than
this value.

There is another argument, based on radioactive data, which gives a
similar value for the dimensions of the radius of a heavy atom. The _α_
particle escaping from the nucleus increases in energy as it passes
through the repulsive field of the nucleus. To fix a minimum limit,
suppose the _α_ particle from uranium, which is the slowest of all _α_
particles expelled from a nucleus, gains all its energy from the
electrostatic field. It can be calculated on these data that the radius
of the uranium nucleus cannot be less than 6 x 10^-12 cm. This is based
on the assumption that the forces outside the nucleus are repulsive and
purely electrostatic. If, as seems not unlikely, there also exist close
to the nucleus strong attractive forces, varying more rapidly than an
inverse square law, the actual dimensions may be less than the value
calculated above.

At this stage of our knowledge it is of great importance to test whether
the law of force breaks down for the distance of closest approach of an
_α_ particle to a nucleus. This can be done by comparing the observed
with the calculated number of _α_ particles scattered through angles of
nearly 180°. It seems almost certain that the inverse square law must
break down when swift _α_ particles are used. This can be seen from the
following argument. If an _α_ particle, of the same speed as that ejected
during the transformation of uranium, is fired directly at the uranium
nucleus, _it must penetrate into the nuclear structure_. If a still
swifter _α_ particle is used, _e. g._ that from radium C, which has about
twice the energy of the uranium _α_ particle, it is clear that it must
penetrate still more deeply into the nuclear structure. This is based on
the assumption that the field due to a nucleus is approximately
symmetrical in all directions. If this is not true, it may happen that
only a fraction of the head-on collisions may be effective in
penetrating the nucleus. It is hoped soon to attack this difficult
problem experimentally.

We have so far dealt with collisions of an _α_ particle with a heavy
atom. We know, however, from the results of Rutherford, Chadwick and
Bieler that in a collision of an _α_ particle with the lightest atom,
hydrogen, the law of the inverse square breaks down entirely when swift
particles are used. Not only are the numbers of H nuclei set in swift
motion much greater than is to be expected in the simple-point nucleus
theory, but the change of number with the velocity of the _α_ particle
varies in the opposite way from the simple theory. Such wide departures
between theory and experiment are only explicable if we assume either
that the nuclei have sensible dimensions or that the inverse square law
of repulsion entirely breaks down in such close collisions. If we
suppose the complexity in structure and in laws of force is to be
ascribed to the _α_ particle rather than to the hydrogen nucleus,
Chadwick and Bieler, as the result of a careful series of experiments,
concluded that the _α_ particle behaved as if it were a perfectly elastic
body, spheroidal in shape with its minor axis 4 x 10^-13 cm. in the
direction of motion and major axis 8 x 10^-13 cm. Outside this
spheroidal region the forces fell off according to the ordinary inverse
square law, but inside this region the forces increased so rapidly that
a particle was reflected from it as from a perfectly elastic body. No
doubt such a conception is somewhat artificial, but it does serve to
bring out the essential points involved in the collision, viz., that
when the nuclei approach within a certain critical distance of each
other, forces come into play which vary more rapidly than the inverse
square. It is difficult to ascribe this break-down of the law of force
merely to the finite size or complexity of the nuclear structure or to
its distortion, but the results rather point to the presence of new and
unexpected forces which come into play at such small distances. This
view has been confirmed by some recent experiments of Bieler in the
Cavendish Laboratory in which he has made, by scattering methods, a
detailed examination of the law of force in the neighborhood of a light
nucleus like that of aluminum. For this purpose he compared the relative
number of _α_ particles scattered within the same angular limit from
aluminum and from gold. For the range of angles employed, viz., up to
100°, it is assumed that the scattering of gold follows the inverse
square law. He found that the ratio of the scattering in aluminum
compared with that in gold depended on the velocity of the _α_ particle.
For example, for an _α_ particle of 3.4 cms. range, the theoretical ratio
was obtained for angles of deflection below 40° but was about 7 per
cent lower for an average angle of deflection of 80°. On the other
hand, for swifter particles of range 6.6 cms. a departure from the
theoretical ratio was much more marked and amounted to 29 per cent for
an angle of 80°. In order to account for these results he supposes that
close to the aluminum nucleus an attractive force is superimposed on the
ordinary repulsive forces. The results agreed best with the assumption
that the attractive force varies according to the inverse fourth power
of the distance and that the forces of attraction and repulsion balanced
at about 3.4 x 10^-13 cm. from the nuclear center. Inside this critical
radius the forces are entirely attractive; outside they are repulsive.

While we need not lay too much stress on the accuracy of the actual
value obtained or of the law of attractive force, we shall probably not
be far in error in supposing the radius of the aluminum nucleus is not
greater than 4 x 10^-13 cm. It is of interest to note that the forces
between an _α_ particle and a hydrogen nucleus were found to vary rapidly
at about the same distance.

It thus seems clear that the dimensions of the nuclei of light atoms are
small, and almost unexpectedly small in the case of aluminum when we
remember that 27 protons and 14 electrons are concentrated in such a
minute region. The view that the forces between nuclei change from
repulsion to attraction when they are very close together seems very
probable, for otherwise it is exceedingly difficult to understand why a
heavy nucleus with a large excess of positive charge can hold together
in such a confined region. We shall see that the evidence from various
other directions supports such a conception, but it is very unlikely
that the attractive forces close to a complex nucleus can be expressed
by any simple power law.




RADIOACTIVE EVIDENCE


A study of the long series of transformations which occur in uranium and
thorium provides us with a wealth of information on the modes of
disintegration of atoms, but unfortunately our theories of nuclear
structure are not sufficiently advanced to interpret these data with any
detail. The expulsion of high speed _α_ and _β_ particles from the
radioactive nucleus gives us some idea of the powerful forces resident
in the nucleus, for it can be estimated that the energy of emission of
the _α_ particle is in some cases greater than the energy that would be
acquired if the _α_ particle fell freely between two points differing in
potential by about 4 million volts. The energies of the _β_ and _γ_ rays
are on a similar scale of magnitude.

Notwithstanding our detailed knowledge of the successive transformation
of the radio-elements, we have not so far been able to obtain any
definite idea of their nuclear structure, while the cause of the
disintegration is still a complete enigma. In comparing the uranium,
thorium, and actinium series of transformations, one cannot fail to be
struck by the many points of similarity in their modes of
disintegration. Not only are the radiations similar in type and in
energy, but, in all cases, the end product is believed to be an isotope
of lead. This remarkable similarity in the modes of transformation is
especially exemplified in the case of the "C" bodies, each of which is
known to break up in at least two distinct ways, giving rise to branch
products. For example, thorium C emits two types of _α_ rays, 65 percent
of range 8.6 cms. and 35 per cent of range 4.8 cms., and in addition
some _β_ rays.

In order to explain these results, it has been suggested that a fraction
of the atoms of thorium C break up first with the expulsion of an _α_
particle and the resulting product then emits a _β_ particle. The other
fraction breaks up in a reverse way, first expelling a _β_ particle,
while the subsequent product emits an _α_ particle. Similar dual changes
occur in radium C and actinium C, although the relative number of atoms
in each branch varies widely for the different elements.

This remarkable similarity between the "C" bodies is still further
emphasized by the recent discovery of Bates and Rogers that both radium
C and thorium C give rise in small numbers to other groups of _α_
particles, some of them moving at very high speeds.

It has often been a matter of remark that the radioactive properties of
the "C" bodies seem to depend more on the atomic number, _i. e._, the
nuclear charge, than on the atomic weight. Confining our attention to
radium C and thorium C, which are best known, both have a nuclear charge
83, but the atomic mass of radium C is 214 and of thorium C 212. The
nucleus of radium C thus contains two protons and two electrons more
than that of thorium C. If it were supposed that the nuclei of these
elements consisted of a large number of charged units in ceaseless and
irregular motion, it is to be anticipated that the addition of the
protons and electrons to the complex structure would entirely alter the
nuclear arrangement and consequently its stability and mode of
transformation. On the other hand, we find that the modes of
transformation of these two nuclei have striking and unexpected points
of resemblance which are in entire disaccord with such a supposition. We
can, however, suggest a possible explanation of this anomaly by
supposing that the _α_ and _β_ particles which are liberated from these
elements are not built deep into the nuclear structure but exist as
_satellites_ of a central core which is common to both elements. These
satellites, if in motion, may be held in equilibrium by the attractive
forces arising from the core, and these forces would be the same for
both elements. On this view the manifestations of radioactivity are to
be ascribed not to the main core, but to the satellite distribution,
which must be somewhat different for the two elements although possibly
showing many points of similarity. It must be admitted that a theory of
this kind is highly speculative, but it does provide a useful working
hypothesis, not only to account for the similarity of the modes of
transformation of the two elements but also immediately suggests a
possible explanation of the liberation of a number of _α_ particles of
different ranges from the same element. There are two ways of regarding
this question. We may in the first place suppose that a certain amount
of surplus energy has to be liberated in the disintegration and that
this energy may be given to any one of a number of satellites. There
will be a certain probability that any particular particle will be given
this energy, and on this will depend the relative number of particles in
the different _α_ ray groups. The ultimate energy of ejection of an _α_
particle will depend on its position in the field of force surrounding
the inner core at the moment of its liberation. On the other hand, we
may suppose that the same _α_ particle is always ejected but that the
particle may occupy in the atom one of a number of "stationary"
positions analogous to the "stationary states" of the electrons in
Bohr's theory of the outer atom. This rests on the assumption that all
the atoms will not be identical in satellite structure but there will be
a number of possible "excited" states of the atom as a consequence of
the previous disintegrations. This satellite theory is useful in another
connection. It has been suggested that possibly the high frequency _γ_
rays from a radioactive atom may arise not from the movement of the
electrons as ordinarily supposed, but from the transfer of _α_ particles
from one level to another. In such a case, the difference in energies
between the various groups of _α_ particles from radium C and thorium C
should be connected by the quantum relation with the frequencies of
prominent _γ_ rays. The evidence at present available is not definite
enough to give a final decision on this problem, but points to the need
of very accurate measurements of the energies of the various groups of
_α_ particles. On account of the relatively small number of particles in
some of the groups, this is difficult of accomplishment.

In considering the satellite theory in connection with the radioactive
bodies, it is at first sight natural to suppose, since the end product
of both the radium and thorium series is an isotope of lead, that one of
the isotopes of lead forms the central core. It may, however, well be
that the radioactive processes cease when there are still a number of
satellites remaining. If this be so, the core may be of smaller nuclear
charge and mass than that of lead. From some considerations, described
later, this core may correspond to an element near platinum of number 77
and mass 192.




FREQUENCY OF VIBRATION OF THE NUCLEUS


One of the most interesting and important methods of throwing light on
nuclear structure is the study of the very penetrating _γ_ rays expelled
by some radioactive bodies. The _γ_ rays are identical in nature with
X-rays, but the most penetrating type of rays consists of waves of much
higher frequency than can be produced in an ordinary X-ray tube. The
work of the last few years has indicated very clearly that the major
part of the _γ_ radiation from bodies like radium B and C originates in
the nucleus. A determination of the frequencies of the _γ_ rays thus
gives us direct information on the modes of vibration of parts of the
nuclear structure. The frequency of some of the softer _γ_ rays excited
by radium B and radium C was measured by the crystal method by
Rutherford and Andrade, but it is difficult, if not impossible, by this
method to determine the frequencies of the very penetrating rays.
Fortunately, due largely to the work of Ellis and Fräulein Meitner, a
new and powerful method has been devised for this purpose. It is well
known that the _β_ rays from radium B and radium C give a veritable
spectrum in a magnetic field, showing the presence of a number of groups
of _β_ rays each expelled with a definite speed. It is clear that each of
the groups of _β_ rays arises from conversion of the energy of a _γ_ ray
of definite frequency into a _β_ ray in one or other of the electronic
levels in the outer atom. The energy ω required to move an electron
from one of these levels to the outside of the atom is known from a
study of X-ray absorption spectra. The frequency ν of the _γ_ ray is
thus given by the quantum relation hν = E + ω, where E is the measured
energy of the _β_ particle.

Since each _γ_ ray may be converted in any one of the known electronic
levels in the outer atom, a single _γ_ ray is responsible for the
appearance of a number of groups of _β_ rays, corresponding to conversion
in the K, L, M, etc., levels. In this way, an analysis of the _β_ ray
spectrum allows us to fix the frequency of the more intense _γ_ rays
which are emitted from the nucleus. The energy of the shortest wave
measured in this way by Ellis corresponds to more than two million
volts, while other evidence shows that probably still shorter waves are
emitted in small quantity from radium C.

Ellis and Skinner have shown that the energies of these rays show
certain combination differences, such as are so characteristic of the
energies of the X-rays arising from the outer electrons. A series of
energy levels may thus be postulated in the nucleus similar in character
to the electron levels of the outer atom, and the _γ_ rays have their
origin in the fall either of an electron or of an _α_ particle between
these levels. This is a significant and important result, indicating
that the quantum dynamics can be applied to the nucleus as well as to
the outer electronic structure.

The probability of levels in the nuclear structure is most clearly seen
on the satellite hypothesis, but in our ignorance of the laws of force
near the core we are at the moment unable to apply the quantum dynamics
directly to the problem. The outlook for further advances in this
direction is hopeful, but is intimately connected with a further
development of our knowledge of the laws of force that come into play
close to the nucleus in the region occupied by the satellites.




ARTIFICIAL DISINTEGRATION OF ELEMENTS


We have seen that it is believed that the nuclei of all atoms are
composed of protons and electrons and that the number of each of these
units in any nucleus can be deduced from its mass and nuclear charge. It
is, however, at first sight rather surprising that no evidence of the
individual existence of protons in a nucleus is obtained from a study of
the transformations of the radioactive elements, where the processes
occurring must be supposed to be of a very fundamental character. As far
as our observations have gone, electrons and helium nuclei, but no
protons, are ejected during the long series of transformations of
uranium, thorium and actinium. One of the most obvious methods for
determining the structure of a nucleus is to find a method of
disintegrating it into its component parts. This is done spontaneously
for us by nature to a limited extent in the case of the heavy
radioactive elements, but evidence of this character is not available in
the case of the ordinary elements.

As the swift _α_ particle from the radioactive bodies is, by far, the
most energetic projectile known to us, it seemed from the first possible
that occasionally the nucleus of a light atom might be disintegrated as
the result of a close collision with an _α_ particle. On account of the
minute size of the nucleus, it is to be anticipated that the chance of
a direct hit would be very small and that consequently the
disintegration effects, if any, would be observed only on a very minute
scale. During the last few years Dr. Chadwick and I have obtained
definite evidence that hydrogen nuclei or protons can be removed by
bombardment of _α_ particles from the elements boron, nitrogen, fluorine,
sodium, aluminum and phosphorus. In these experiments the presence of H
nuclei is detected by the scintillation method, and their maximum
velocity of ejection can be estimated from the thickness of matter which
can be penetrated by these particles. The number of H nuclei ejected
even in the most favorable case is relatively very small compared with
the number of bombarding _α_ particles, viz., about one in a million.

In these experiments the material subject to bombardment was placed
immediately in front of the source of _α_ particles and observations on
the ejected particles were made on a zinc sulphide screen placed in a
direct line a few centimetres away. Using radium C as a source of _α_
rays, the ranges of penetration, expressed in terms of centimetres of
air, were all in these cases greater than the range of free nuclei (30
cms. in air) set in motion in hydrogen by the _α_ particles. By inserting
absorbing screens of 30 cms. air equivalent in front of the zinc
sulphide screen the results were quite independent of the presence of
either free or combined hydrogen as an impurity in the bombarded
materials. Some of the lighter elements were examined for absorptions
less than this, but, in general, the number of H particles due to
hydrogen contamination of the source and the materials was so large that
no confidence could be placed in the results.

In such experiments many scintillations can be observed, but it is very
difficult to decide whether these can be ascribed in part to an actual
disintegration of the material under examination. The presence of
long-range particles of the _α_ ray type from the source of radium C
still further complicates the question, since in general the number of
such particles is large compared with the disintegration effect we
usually observe.

To overcome these difficulties, inherent in the direct method of
observation, Dr. Chadwick and I have devised a simple method by which we
can observe with certainty the disintegration of an element when the
ejected particles have a range of only 7 cms. in air. This method is
based on the assumption, verified in our previous experiments, that the
disintegration particles are emitted in all directions relative to the
incident rays. A powerful beam of _α_ rays falls on the material to be
examined and the liberated particles are observed at an average angle of
90° to the direction of the incident _α_ particles. By means of screens
it is arranged that no _α_ particles can fall directly on the zinc
sulphide screen.

This method has many advantages. We can now detect particles of range
more than 7 cms. with the same certainty as particles of range above 30
cms. in our previous experiments, for the presence of hydrogen in the
bombarded material has no effect. This can be shown at once by
bombarding a screen of paraffin wax, when no particles are observed on
the zinc sulphide screen. On account of the very great reduction in
number of H nuclei or _α_ particles by scattering through 90°, the
results are quite independent of H nuclei from the source or of the
long-range _α_ particles. The latter are just detectable under our
experimental conditions when a heavy element like gold is used as
scattering material, but are inappreciable for the lighter elements.

A slight modification of the arrangement enables us to examine gases as
well as solids.

Working in this way we have found that in addition to the elements
boron, nitrogen, fluorine, sodium, aluminum, and phosphorus, which give
H particles of maximum range in the forward direction between 40 and 90
cms., the following give particles of range above 7 cms.: neon,
magnesium, silicon, sulphur, chlorine, argon, and potassium. The numbers
of the particles emitted from these elements are small compared with the
number from aluminum under the same conditions, varying between ⅓ and
¹⁄₂₀. The ranges of the particles have not been determined with
accuracy. Neon appears to give the shortest range, about 16 cms., under
our conditions, the ranges of the others lying between 18 cms. and 30
cms. By the kindness of Dr. Rosenhain we were able to make experiments
with a sheet of metallic beryllium. This gave a small effect, about
¹⁄₃₀ of that of aluminum, but we are not yet certain that it may
not be due to the presence of a small quantity of fluorine as an
impurity. The other light elements, hydrogen, helium, lithium, carbon,
and oxygen, give no detectable effect beyond 7 cms. It is of interest to
note that while carbon and oxygen give no effect, sulphur, also probably
a "pure" element of mass 4n, gives an effect of nearly one-third that of
aluminum. This shows clearly that the sulphur nucleus is not built up
solely of helium nuclei, a conclusion also suggested by its atomic
weight of 32.07.

We have made a preliminary examination of the elements from calcium to
iron, but with no definite results, owing to the difficulty of obtaining
these elements free from any of the "active" elements, in particular,
nitrogen. For example, while a piece of electrolytic iron gave no
particles beyond 7 cms., a piece of Swedish iron gave a large effect,
which was undoubtedly due to the presence of nitrogen, for after
prolonged heating _in vacuo_ the greater part disappeared. Similar
results were experienced with the other elements in this region.

We have observed no effects from the following elements: nickel, copper,
zinc, selenium, krypton, molybdenum, palladium, silver, tin, xenon, gold
and uranium. The krypton and xenon were kindly lent by Dr. Aston.




EXAMINATION OF LIGHT ELEMENTS FOR PARTICLES
OF RANGE LESS THAN 3 CMS. OF AIR


When _α_ particles are scattered from light elements, the simple theory
shows that the velocity of the scattered particles depends on the angle
of scattering. For example, using bombarding _α_ particles of range 7
cms., the range of the _α_ particles scattered through more than 90°
cannot be greater than 1.0 cm. for lithium (7), 2.0 cms. for beryllium
(10), 2.5 cms. for carbon, 3.2 cms. for oxygen, 4.3 cms. for aluminum,
and 6.8 cms. for gold.

Provided we introduce sufficient thickness of absorber to stop the _α_
particles scattered through 90°, we can examine for disintegrated
particles from carbon, for example, whose range exceeds 2.5 cms. Certain
difficulties arise in this type of experiment which are absent when the
thickness of absorber is greater than 7 cms.; any heavy element present
as an impurity will give scattered _α_ particles of range greater than
those from carbon and thus complicate the observations. In addition,
serious troubles may arise due to the volatilization or escape of active
matter from the source. This is especially marked if the vessel
containing the radioactive source is exhausted. To overcome this
difficulty, we have found it desirable to cover the source with a thin
layer of celluloid of 2 or 3 mm. stopping power for _α_ rays. By this
procedure we have been able to avoid serious contamination and to
examine the lighter elements by this method. We have been unable to
detect any appreciable number of particles from lithium or carbon for
ranges greater than 3 cms. If carbon shows any effect at all, it is
certainly less than one tenth of the number from aluminum under the same
conditions. This is in entire disagreement with the work of Kirsch and
Patterson (Nature, April 26, 1924), who found evidence of a large number
of particles from carbon of range 6 cms. A slight effect was observed in
beryllium in accordance with our other experiments. No effect was noted
in oxygen gas. Apart from beryllium, no certain effect has been noted
for elements lighter than boron.

Under the conditions of our experiment, it seems clear that neither H
nuclei nor other particles of range greater than 3 cms. can be liberated
in appreciable numbers from these elements in a direction at right
angles to the bombarding _α_ rays. This is, in a sense, a disappointing
result, for, unless these elements are very firmly bound structures, it
was to be anticipated that an _α_ particle bombardment would resolve them
into their constituent particles.

We hope to examine this whole question still more thoroughly, as it is a
matter of great importance to the theory of nuclear constitution to be
certain whether or not the light elements can be disintegrated by swift
_α_ particles.

In considering the results of our new and old observations, some points
of striking interest emerge. In the first place, all the elements from
fluorine to potassium inclusive suffer disintegration under _α_ ray
bombardment. As far as our observations have gone, there seems little
doubt that the particles ejected from all these elements are H nuclei.
The odd elements, B, N, F, Na, Al, P, all give long-range particles
varying in range from 40 cms. to 90 cms. in the forward direction, the
even elements, C, O, Ne, Mg, Si, S, either give few particles or none at
all as in the case of C and O, or give particles of much less range than
the adjacent odd numbered elements. The differences between the ranges
of even-odd elements become much less marked for elements heavier than
phosphorus.

This obvious difference in velocity of expulsion of the H nuclei from
even and odd elements is a matter of great interest. Such a distinction
can be paralleled by other observations of an entirely different
character. Harkins has shown that elements of even atomic number are
much more abundant in the earth's crust than elements of odd atomic
number. In his study of Isotopes, Aston has shown that in general odd
numbered elements have only two isotopes differing in mass by two units,
while even numbered elements in some cases contain a large number of
isotopes. This remarkable distinction between even and odd elements
cannot but excite a lively curiosity, but we can at present only
speculate on its underlying cause.




VELOCITY OF ESCAPE OF HYDROGEN NUCLEI


We have seen that the experiments of Bieler on the scattering of _α_ rays
by aluminum and magnesium indicate that a powerful attractive force
comes into play very close to the nuclei of these atoms. If this be the
case, the forces of attraction and repulsion must balance at a certain
distance from the nucleus. Outside this critical point the forces on a
positively charged body are entirely repulsive. Certain important
consequences follow from this general view of nuclear forces. Suppose,
for example, that, due to a collision with a swift _α_ particle, a
hydrogen nucleus is liberated from the nuclear structure. After passing
across the critical surface, it will acquire energy in passing through
the repulsive field. It is clear, on this view, that the energy of a
charged particle after escape from the atom cannot be less than the
energy acquired in the repulsive field; consequently we should expect to
find evidence that there is a minimum velocity of escape of a
disintegration particle. We have obtained definite evidence of such an
effect both in aluminum and sulphur by examining the absorption of H
nuclei from these elements. The number of scintillations for a thin film
was found to be nearly constant for absorption between 7 and 12 cms.,
but falls off rapidly for greater thicknesses. This is exactly what is
to be expected on the views outlined. No doubt the limiting velocity
varies somewhat for the different elements, but a large amount of
experiment will be required to fix this limit with accuracy. From these
results it is possible to form a rough estimate of the potential of the
field at the critical surface, and this comes out to be about 3 million
volts for aluminum. The value for sulphur is somewhat greater. This
brings out in a striking way the extraordinary smallness of the nuclei
of these elements, for it can be calculated that the critical surface
cannot be distant more than 6 x 10^-13 cm. from the centre of the
nucleus. These deductions of the critical distance are in excellent
accord with those made by Bieler from observations of the scattering of
_α_ particles.

Another important consequence follows. It is clear that an _α_ particle
fired at the nucleus will not be able to cross this critical surface and
thus be in a position to produce disintegration, unless its velocity
exceeds that corresponding to the critical potential. In an experiment
made a few years ago, we found that the number of H nuclei liberated
from aluminum fell off rapidly with diminution of the velocity of the _α_
particle and was too small in number to detect when the range of the _α_
particle was less than 4.9 cms. This corresponds to the energy of an _α_
particle falling between about 3 million volts--a value in good accord
with that calculated from the escape of H nuclei.

Further experiments are required with other elements to test if this
relation between the minimum velocity of H nuclei and the minimum
velocity of the _α_ particle to produce disintegration holds generally;
but the results as far as they go are certainly very suggestive.

It is of interest to note that these results afford a definite proof of
the nuclear conception of the atom and give us some hope that we may
determine the magnitude of the critical potential for a number of the
light elements.




EVOLUTION OF NUCLEI


In concluding, I would like to make a few remarks of a more speculative
character dealing with the fundamental problem of the origin and
evolution of the elements from the two fundamental building units, the
positive and negative electrons. It must be confessed that there is
little information to guide us with the exception of our knowledge of
the nuclear charges and masses of the various species of elements which
survive to-day. It has always been a matter of great difficulty to
imagine how the more complex nuclei can be built up by the successive
additions of protons and electrons, since the proton must be endowed
with a very high speed to approach closely to the charged nucleus. I
have already discussed in this paper the evidence that powerful
attractive forces varying very rapidly with the distance are present
close to the nuclear structure and it seems probable that these forces
must ultimately be ascribed to the constituent proton. In such a case it
may be possible for an electron and proton to form a very close
combination, or neutron, as I have termed it. The probable distance
between the centre of this doublet is of the order of 3 x 10^-13 cm. The
forces between two neutrons would be very small except for distance of
approach of this order of magnitude, and it is probable that the
neutrons would collect together in much the same fashion as a number of
small movable magnets would tend to form a coherent group held together
by their mutual forces.

In considering the origin of the elements, we may for simplicity suppose
a large diffused mass of hydrogen which is gradually heated by its
gravitational condensation. At high temperatures the gas would consist
mainly of free hydrogen nuclei and electrons, and some of these would in
course of time combine to form neutrons, emitting energy in the process.
These neutrons would collect together in nuclear masses of all kinds of
complexity. Now the tendency of the groups of neutrons would be to form
more stable nuclear combinations, such as helium nuclei of mass four,
and possibly intermediate stages of masses two and three. Energy would
be emitted in these processes probably in the form of swift surplus
electrons which were not necessary for the stability of the system. In a
sense, all these nuclear masses would be radioactive, but some of them
in their transformation may reach a stable configuration which would
represent the nucleus of one of our surviving elements. If we suppose
that nuclear masses over a wide range of mass can be formed before
serious transformation occurs, it is easy to see how every possible type
of stable element will gradually emerge. If we take the helium nucleus
as a combining unit which emits in its formation the greatest amount of
energy, we should ultimately expect many of the neutrons in a heavy
nucleus to form helium nuclei. These helium nuclei would tend to collect
together and form definite systems and it seems not unlikely that they
will group themselves into orderly structures, analogous in some
respects to the regular arrangement of atoms to form crystals, but with
much smaller distances between the structural units. In such a case,
some of the elements may consist of a central crystal type of structure
of helium nuclei surrounded by positive and negatively charged
satellites in motion round this central core. Assuming that such orderly
arrangements of helium nuclei are possible, it is of interest to note
that the observed relations between atomic charge and atomic mass for
the elements can be approximately obtained on a very simple assumption.
Suppose that helium nuclei form a point centred cubic lattice with an
electron at the centre of a crystal unit of eight helium nuclei. A few
of the possible types of grouping are given in the following table, with
corresponding masses and nuclear charges. The structure 4. 3. 2. means
a rectangular arrangement with sides containing 4. 3. 2. nuclei
respectively. It will thus contain 24 helium nuclei, have a mass 96, and
will contain 6 intranuclear electrons. Its nuclear charge will therefore
be 48 - 6 = 41.


Structural arrangement of    Calculated     Calculated  Known element of
       helium nuclei       nuclear charge      Mass       equal charge

3. 2. 2.                        22              48      Ti 48

3. 3. 2.                        32              72      Ge 74, 72, 70

3. 3. 3.                        46             108      Pd 106.7

4. 2. 2.                        29              64      Cu 63.35

4. 3. 2.                        42              96      Mo 96

4. 3. 3.                        60             144      Nd 144

4. 4. 3.                        78             192      Pt 195


While the agreement is far from perfect for all these structures, there
is a general accord with observation. If we take the view that some of
these structures can grow by the addition of satellites, there is room
for adjustment of masses and to include the intervening elements. This
point of view is admittedly very speculative and there may well be other
types of structure involved. At the same time, the general evidence
suggests that there are some basal structures on which the heavier atoms
are progressively built up. The failure of the whole number rule for the
mass of isotopes, observed in some cases by Aston, _e.g._, between tin
and xenon, certainly supports such a conception. From a study of the
artificial disintegration of the elements we have seen that carbon and
oxygen represent very stable structures probably composed of helium
nuclei. It is possible that oxygen nuclei, for example, may be the
structural basis of some of the elements following oxygen, but our
information is at present too meagre to be at all certain on this point.

I think, however, it will be clear from this lecture what a difficult
but fascinating problem is involved in the structure of nuclei. Before
we can hope to make much advance, it is essential to know more of the
nature of the forces operative close to protons and electrons, and we
may hope to acquire much information by a detailed study of the
scattering of swift _α_ rays and _β_ rays by nuclei. Fortunately, there is
now a number of distinct lines of attack on this problem, and from a
combination of the results obtained we may hope to make steady, if not
rapid, progress in the solution of this, the greatest problem in
Physics.