Cyclotron1

E. O. Lawrence and D. Cooksey, Berkeley, USA

I. Introduction

Beginning with the first experiments[^2], carried out six years ago in our laboratory, work has proceeded continuously on the development of methods for obtaining fast charged particles. By this we mean the creation of methods and apparatus that would be useful in applications to research in the field of nuclear physics.

Almost from the very beginning we considered the most promising method to be that of multiple acceleration (of light ions), according to which ions move in a magnetic field in resonance with a rapidly alternating electric field. In accordance with this, the greatest attention in our work was devoted to this method of multiple acceleration. Four years ago an apparatus was described in detail[^3] that made it possible to obtain protons with an energy of 1.2 MeV, and three years ago[^4] a more powerful apparatus of this type, suitable for obtaining deuterons with energies of 5 MeV in quantities of fractions of a microampere. Since then, significant improvements have been made not only in increasing the output to several microamperes at high energies and in the sense of greater reliability and convenience of the apparatus, but also with respect to methods of extracting the ions from their spiral trajectories.

The subject of the present article is a description of these latest improvements, as well as a discussion of certain related questions, since some of the measurements we have made point to the possibility of further improvements.

II. Apparatus

The general appearance of the apparatus, now called[^5] a magnetic resonance accelerator, or more often a “cyclotron,” is shown in Fig. 1. The electromagnet, described in the article published in 1934, has not—

...has undergone no changes. We intended to increase the diameter of the pole tips (at present equal to 687.5 mm), but this change has not yet been carried out. All our attention has been directed toward improving the cyclotron at its present dimensions. The drawing in Fig. 2 gives a general view of the chamber with the cover removed,

Fig. 1. General view of the cyclotron.

in which the following are visible: the ion source, the accelerating electrodes, called dees, the electrode serving to deflect the ions from their spiral paths,

Fig. 2. Perspective view of the cyclotron chamber with the cover removed.

beryllium and tungsten electrodes capable of rotating, and the window through which the ions are emitted from the chamber.

outward. In the following sections we shall consider the action of all these parts of the apparatus.

Ion source. As in the original design, ions are formed in the central part of the chamber as a result of the motion of a beam of electrons flying between the dees in the direction of the magnetic field and ionizing the gas that fills the chamber under the proper pressure. In the original design the spiral filaments, which served as sources of electrons and were located in the upper and lower parts of the central region of the chamber, were mounted on glass tubes enclosed in copper tubes fastened to the walls of the chamber and passing

Fig. 3. View of the cyclotron chamber with the cover removed and one dee removed. In this instrument the ions are not completely removed, but bombard the electrodes on the chamber walls.

between the dees along the radius toward the center. These tubular shields were present only on one side, which led to an undesirable asymmetry of the field between the dees. In the apparatus being described this defect was eliminated1 by introducing the filaments into the central region through one of the dees, as can be seen from Fig. 3. In this figure one can see two filament holders, located one above and the other below one of the dees (in this photograph—the one removed). Flat spiral filaments, consisting of \(5 \tfrac{1}{2}\) turns of tungsten wire of diameter \(0.075\) mm and having a maximum spiral diameter of about \(12\) mm, are mounted by welding to water-cooled copper blocks. The filament holders are placed inside copper protective housings (necessary because high frequency is applied to the dees), and the whole arrangement is made as thin as possible in order to increase as much as possible the gaps between the parts connected with the filaments and the dees. In practice, as experience has shown, in order to avoid difficulties associat-

with gas breakthrough, a clearance of 6 mm in the vertical direction or 3 mm in the horizontal direction is sufficient.

The filament holders are mounted on a special joint that permits the filament to be moved to various points in the central part of the chamber. Theoretically, the ion source should be located at the center of the cyclotron, since the ions begin their motion under the action of the oscillating electric field between the accelerating electrodes along trajectories with small radii of curvature. In practice, however, it is usually found that the most advantageous position of the filament is not central, but somewhat displaced relative to the center in one direction or another. This is a consequence of the fact that, generally speaking, the centers of successive circular paths of the ions are displaced relative to the geometrical center of the apparatus because of the absence of complete central symmetry of the magnetic field. This displacement of the ion paths can be more or less eliminated by making the magnetic field symmetrical, which can be done, for example, by placing pieces of iron of suitable size and shape between the pole pieces of the magnet and the covers of the vacuum chamber. Usually, however, it proves much simpler simply to shift the point at which the ions begin their spiral paths.

The entire system of filament holders is fastened, by means of bolts and vacuum mastic, to a plate attached to the side wall of the chamber, and is made easily removable, which is necessary for replacing the filaments. Tubes with flowing water pass along the shield protecting the filament, and also cool the places sealed with mastic. The electrodes supplying current to the filament are insulated from the surrounding electrostatic shield by glass tubes and are led out through a rubber stopper, also sealed with mastic. The insulation of the filament can withstand a voltage of several thousand volts; in operation, however, the negative potential of the filament relative to the shields and the walls of the vacuum chamber usually does not exceed 1000 V.

Although the apparatus has two filaments (one above and one below), only one operates at a time. The other is then moved aside and is switched on when the first burns out. Electrons from the filament, passing between the dees, strike the chamber cover and, since water cooling of it is not provided, this bombardment at a sufficiently large current can cause heating of the metal and gas evolution. Therefore, in the future, when large currents are used, it will be necessary to water-cool this part of the apparatus as well; for the currents used at present (less than 500 μA at 1000 V), this was not necessary.

From a comparison of the magnitude of the current produced by the fast ions in the peripheral part of the apparatus with the number of ions formed at the center, it follows that the focusing action of the electric and magnetic fields is such that a very considerable fraction of the ions produced reaches the collecting electrode. Since the influence of space charge at small currents and high voltages, with which

CYCLOTRON

that we encounter in this device, play no role, it should be expected that the number of fast ions will be proportional to the number of ions formed in the ion source. Therefore it seems important to us to consider ways of increasing the productivity of this source.

The ion current is in general proportional to the electron current, and therefore, in order to increase the number of ions formed, it is sufficient to increase the emission of the filament. Unfortunately, large electron currents are limited by the action of space charge, and in order to increase the ionizing electron stream it is necessary to increase the accelerating voltage. In our present apparatus, increasing the electron stream by raising the accelerating voltage proves disadvantageous, since the probability of ionization (for example, the formation of protons) decreases as the energy of the electrons increases, as was shown by Bleakney[^6]. The optimum value of the energy for the formation of protons or deuterons lies near 200 V. It is therefore desirable to create an electron source that would give large currents at low voltages.

One way of achieving this goal is to increase the size of the filament. This is possible, since in practice we have become convinced that the central region of the apparatus, of considerable dimensions, can serve as the ion source. When the filament is displaced over distances up to 50 mm, the yield of fast ions, relative to the number formed in the source, changes by no more than 50%. A weakening of the action of the space charge can also be achieved by placing the accelerating electrode as close as possible to the filament. A suitable design would therefore be the use of a comparatively thick (of the order of 0.1 mm) straight tungsten filament several inches long, stretched parallel to the edges of the dees midway between them. With such a straight filament, held exactly in the proper position, it would be possible to place, in its immediate vicinity, a water-cooled accelerating electrode with a narrow rectangular slit, which would give a much stronger field at the surface of the filament than in the present device.

It therefore seems to us that it will not be very difficult to achieve a 5- to 10-fold increase in the electron current producing ions and a corresponding increase in the number of fast ions. At present the maximum yield of fast ions is about 25 μA, and the usual stable operation of the apparatus is attained at 20 μA. With an increase in the productivity of the ion source, the yield of fast ions should prove to be greater than 100 μA.

Accelerating electrodes. The semicircular hollow accelerating electrodes, or dees, are clearly visible in Fig. 2 and are shown in more detail in Figs. 3 and 4. These electrodes are made entirely of copper, the copper sheets being brazed to a more massive copper frame carrying tubes with flowing water for cooling, necessary because of the high-frequency charging currents and bombardment by moving ions. The outside diameter of the dees is 606.25 mm. The arrangement of the exit slit, as can be seen from Fig. 4, is such that,

that the diameter of the last circular path of the ions turns out to be equal to 575 mm. The duants at the center have a thickness of 43.75 mm, their thickness decreasing uniformly toward the periphery to 31.25 mm. The greater thickness in the central part helps to increase the number of ions without unnecessarily increasing the capacitance of the system.

The duants are fastened to 25-mm copper rods 550 mm long, which in turn are supported by tubular insulators made of Pyrex glass, clearly visible in Figs. 2, 3, and 4. These insulators are standard flanged tubes with wall thickness of 6 mm and inside diameter of 75 mm, the middle part of which is expanded to an inside diameter of 100 mm (the insulators shown in the photograph of Fig. 3 are not expanded). The glass is joined to the metal parts by means of fused-in wire rims, according to the method proposed by D. L. Webster and P. A. Ross.

Fig. 4. Details of the arrangement and mounting of the accelerating electrodes.

From the mechanical point of view, this method of fastening the duants is quite satisfactory; however, we encounter difficulties connected with breakdown of the insulator under the action of cathode rays and with its overheating in the high-frequency field. These difficulties were eliminated by providing protection from cathode rays (Fig. 4) and by applying air cooling of the insulator. The arrangement shown in the drawing not only makes it possible completely to eliminate the danger due to cathode rays, but also promotes such a distribution of the high-frequency fields that heating is reduced to a minimum. Nevertheless, the heating of the walls still remains so strong that, in order to maintain an admissible temperature, it is necessary to resort to blowing the insulators with air. With air cooling in operation, we had no cases of breakdown over a period of several months, and apparently the insulators can serve under these conditions for an unlimited time.

Since, in the further development of these devices, the tendency will undoubtedly be to strive to increase the number of fast ions with ever greater and greater energies, this will lead to a continuous increase in the high-frequency voltage on the dees. As a result, a situation will inevitably be reached in which avoiding breakdown of the insulators will prove very difficult, and the need will arise for their complete elimination. This can be accomplished by fastening the dees to massive and rigid coils located in vacuum. The application of high frequency to such a contour can be carried out at a comparatively low voltage, and therefore the oscillatory power can be increased many times without encountering difficulties with insulators. Such a device, making use of the insulating properties of vacuum in the manner of Sloan’s high-voltage radio-frequency generator, will undoubtedly be used in our laboratory in the future.

The question of the influence of the thickness of the dee (i.e., the size of the clearance between the covers) on the yield of fast ions is of great interest and practical significance. The fact that fast ions can be obtained in quantities of many microamperes indicates that the focusing action of the curved electric field between the dees is very effective. At the same time, it does not seem possible to predict precisely the influence of the geometry of the accelerating electrodes on the focusing. Without doubt, it can be asserted with confidence that the number of ions formed at the center will be at least a linear function of the thickness of the dees; but how strongly the thickness of the dees at the periphery affects the ion output can be determined only by experiment.

To clarify this question, we recently tested dees of uniform thickness throughout, equal to 50 mm, and established that under the same conditions in which, with narrow dees, the ion output was 20 μA, the wide dees gave a steady output of 50 μA. Thus the ion output increases with the thickness of the dees somewhat faster than according to a linear law. Therefore a cyclotron intended for very large output currents must have not only a powerful ion source, but also very thick dees.

Extraction of ions outward. The device used for extracting ions outward is shown in Figs. 2 and 4. Ions leaving one of the dees are deflected by a transverse electric field and move between the deflecting plate and the dee along an approximately circular path, whose radius is 30% greater than the radius of the last trajectory inside the dee (whose radius is 287.5 mm). The force exerted by the magnetic field on an ion moving along a circle is equal to the centripetal force, which in turn is equal to twice the ion energy divided by the radius of curvature. Therefore, the force with which the magnetic field acts on an ion of energy 5 MeV moving along a circle with a radius of 30 cm corresponds to an electric field of 330,000 V/cm. In view of this, the electric fields used in a modern cyclotron have

gradients of the order of 100,000 V/cm. Since the diameter of the pole tip is 687.5 mm, the ions leaving the deuteron cavity find themselves at a distance of 56 mm from the side wall of the vacuum chamber. By means of deflecting electrostatic fields of the indicated magnitude they are brought to the wall over a path less than a quarter of the circumference from the point of exit from the deuteron. As is seen from Figs. 2 and 4, a water-cooled electrode mounted on a lapped plug can be placed in the path of the ion beam. Usually this system consists of two electrodes—tungsten and beryllium. The first serves chiefly for adjusting the apparatus (measuring the beam intensity with a minimum number of neutrons), the second as the neutron source. The presence of a flattened portion of the chamber wall near the electrodes permits objects to be placed close to and directly in front of the neutron source. Since neutrons acquire momentum from the bombarding deuterons, the intensity of the neutron flux in the forward direction is usually greater, so that the possibility of placing objects near or in immediate proximity to the beryllium electrode is a real advantage.

It is often desirable to subject objects to the direct action of a deuteron flux, for which purpose it is necessary to let the ion flux out. This is accomplished by means of a thin platinum window, consisting of platinum foil 0.0025 mm thick, soldered to a brass grid cooled by water and capable of dissipating the heat released by the bombarding deuterons and by the deuteron flux passing through the window. In its absorption the platinum foil used is equivalent to a layer of air 1.5 cm thick. The apertures of the grid are such as to allow a flux somewhat greater than half of the incident flux to pass. A deuteron flux penetrating into air, with an intensity of several microamperes, produces an intense lavender-colored glow. Fig. 5 shows a photograph of this glow, formed in air by a beam of deuterons with energies of 5.8 MeV and an intensity of 5 μA; the length of the glow is 25 cm. Although the device described here is quite applicable and convenient for a whole range of investigations, in some cases it would be desirable to lead the ion flux through the vacuum to a considerable distance from the accelerating electrodes, in order to eliminate the background—radiation produced by ions moving between the deuterons and bombarding various parts of the apparatus other than the special electrodes. This improvement can easily be carried out by installing, in place of the platinum window, one (open) end of a tube of the required length, the other end of which is equipped with a platinum window. If measures are not taken to focus the beam inside the tube, it will diverge, partly owing to the inhomogeneity of the magnetic field beyond the vacuum chamber of the cyclotron. This inhomogeneity, however, can easily be eliminated by placing around the tube pieces of iron of suitable dimensions and shape. In this way focusing of the ion beam at the new exit window can be achieved. These measures have not yet been carried out.

conducted in our laboratory, but we intend to carry them out in the near future in connection with some leading investigations here.

Generators. The source of high frequency is a self-excited generator (with tuned grid and anode circuits), which has the advantage of simplicity and of a minimal number of auxiliary devices. The generator operates on two tubes of our design, recently described,^8 connected in a push-pull circuit. The tubes are rated for 30 kW, but in practice much less power is required. The anode circuits of the tubes are supplied from a three-phase full-wave rectifier, giving a rectified current at a voltage of approximately up to 10,000 V. The rectifier consists of three 220/6600 V transformers, connected for three-phase full-wave

Fig. 5. Photograph of a 5 μA beam of deuterons with energies of 5.8 MeV (range in air equal to 25 cm), emerging from the platinum window of the cyclotron.

rectification and of Kenotrons of type 872 A. In order to draw currents of any required strength from the rectifier, 12 such Kenotrons operate simultaneously, connected in groups of two with stabilizing resistances for each group. During operation of the cyclotron with an output of 20 μA of deuterons with energies of 5.5 MeV, the anode voltage of the generator is 8000 V and the anode current is about 3 A. Of the supplied power of 25 kW, about 12 kW of oscillatory power is delivered to the circuit accelerating the ions.

Despite the fact that the simplicity of the self-excited generator recommends it for many practical purposes, such a generator has disadvantages with respect to the constancy of frequency and a known instability in operation. The point is that a change in the frequency of the generator means a change in the yield of fast ions, whereas in many cases it is desirable to maintain the greatest possible constancy of ion energy and beam intensity. We are at present engaged in constructing a new generator, consisting of stabilizing-

of the driving oscillator and several stages of power amplification. This device must supply to the accelerating electrodes a high-frequency voltage of constant frequency and very nearly constant value. With the existing oscillator, constancy of beam intensity can be maintained to an accuracy of up to 10% by occasionally adjusting the magnetic field. The new oscillator will make it possible to obtain from the cyclotron a constant and uniform beam of fast ions for considerable intervals of time without supervision.

III. Adjustment of the Cyclotron

Dees. A special joint in the system for mounting the dees makes it possible to change their position during operation of the apparatus. As might have been expected, the most advantageous position of the dees is that in which they are exactly parallel to one another and symmetrical with respect to the magnetic field. Deviations from this symmetrical arrangement, for example by 1 mm, are sometimes found as a ten-percent decrease in the ion output. This circumstance emphasizes, on the one hand, the importance of a design of the dee mounting which would not allow them to shift from the proper position, and, on the other hand, the desirability of an arrangement permitting adjustment of the apparatus. With regard to setting the dees, it was found that small departures from parallelism can, to a certain extent, be compensated by changing their relative position. The system for mounting the dees permits changing the distance between them, which, when they are parallel, is about 18 mm. A change in the distance between the dees does not lead to significant changes in the output, but in general, as the distance increases, the yield of fast ions decreases. It is not known to us whether this decrease occurs because of a change in the number of ions formed at the center of the apparatus, or as a result of a change in the focusing action of the fields.

In light of the preceding remarks it may seem that, with sufficient rigidity of the dees and of the rods supporting them, one can dispense with movable joints, since the dees can be placed in the optimal symmetrical position during assembly of the apparatus, and this position will not be disturbed during operation. It is difficult, however, to decide whether in practice this simplification of the design of the dee holders will more than compensate for the absence of the possibility of adjusting the position of the dees during operation.

Magnetic field. Neglecting relativistic corrections, in view of their smallness, the simple theory of cyclotron operation assumes that the magnetic field is uniform throughout the entire ion-accelerating system. In reality these ideal conditions are not satisfied. The reason for this is, on the one hand, the imperfection of the geometry of the magnetic circuit and the inhomogeneity of the iron composing it, and, on the other hand, the fact that saturation

iron pole shoes leads to a weakening of the field in the direction from the center toward the periphery of the dees. This latter inhomogeneity, strange as it may seem, is desirable. The actual distribution of the magnetic-field intensity from the center to the periphery (for the cyclotron now in existence) is shown in Fig. 6; from it one sees that at the distance from the center of the pole shoe to its circumference the field strength falls by 25%, and that even at the distance from the center to the place where the ions leave the accelerating electrodes (radius 387.5 mm) the field decreases by 7%. At first glance, such a considerable radial inhomogeneity of the magnetic field would make acceleration of ions at the periphery impossible. However, on closer examination it turns out that if the accelerating high-frequency voltage, when obtaining ions with an energy of 5 MeV, is 100,000 V, then the ions make only about 25 complete revolutions, and under these conditions the field inhomogeneity, amounting during motion toward the periphery to only 7%, cannot substantially affect the synchronization between the high-frequency field and the ions moving around the circumference. The high-frequency voltage used in the present cyclotron has not been reliably measured, but by approximate estimate it lies between 50,000 and 100,000 V.

Fig. 6. Radial distribution of the magnetic field in the cyclotron; the hatched regions depict the pole shoes of the electromagnet and the iron covers of the cyclotron chamber.

This radial gradient of the magnetic field, which becomes very considerable near the periphery, produces an “opening” of the spiral trajectories of the ions. Instead of moving along circles between the dees, the weakening of the field in the direction toward the edge compels those ions whose trajectory centers do not exactly coincide with the centers of the magnetic field to describe spirals with gradually increasing radii of curvature. The degree of influence of this factor on the operation of the currently existing cyclotron is illustrated by a simple approximate calculation of the trajectory of an ion, under the assumption that the center of the spiral described by the ion at the last moment before escape is displaced by 3 mm relative to the center of the field and that the accelerating high-frequency voltage is equal to 50,000 V. The ion trajectory, calculated by this simplified method, proves to extend beyond the limits of the dee at a distance—

less than 12 mm from the place where the preceding one emerges. Thus we see that this radial gradient of the magnetic field has a very noticeable effect. This is evidently a very valuable property of it, since it facilitates the removal of ions from the accelerating system.

The circumstance that a small displacement of the center of the spiral trajectory (3 mm) leads to such considerable changes in the trajectory itself imposes very stringent conditions on the magnetic field with respect to constancy in the azimuthal direction. If the mean value of the magnetic-field strength in one quadrant differed appreciably from the values in the other quadrants, the result would be a displacement of the centers of the spiral trajectories and, as a consequence of this, considerable distortions of them, greatly increased by the action of the radial field gradient. Therefore, in adjusting the magnetic field of the cyclotron it proves necessary to place pieces of iron between the pole tips and the covers of the chamber. For this purpose, as can be seen from Fig. 6, 6-mm gaps are provided between the chamber and the pole tips. The introduction of iron is carried out entirely by trial and error—the criterion is the elimination of azimuthal inhomogeneity. In general it may be stated that an increase of the magnetic field, say, in the southern part of the interpole space, results in a displacement of the centers of the spiral trajectories to the east. Therefore, if it turns out that the stream of ions leaving the exit slit of the duant has an energy less than it should have, this means that the centers of the spiral trajectories have shifted in the direction of the exit slit, so that the radii of curvature of the last trajectories prove to be smaller than the distance between the center and the exit slit. In this case the energy of the particles can be increased by placing pieces of iron in such a way as to produce an increase of the field in a definite region, so that the centers of the spiral paths are displaced back toward the center of the apparatus and the radii of the last paths increase. It should be emphasized once more that in practice the adjustment of the magnetic field by introducing iron is to a considerable extent an empirical operation. Although the action of the iron cannot always be precisely explained, this method of adjustment, fortunately, is not difficult to carry out, since the ion current can be obtained even without the introduction of additional iron. Thus the homogeneity, intensity, and energy of the ion beam can be continuously observed throughout the entire adjustment of the magnetic field.

IV. Operation of the Cyclotron

Hydrogen ions. The cyclotron now in existence gives a deuteron current of several microamperes with energies from 5 to 6 MeV without any adjustment, immediately after the vacuum chamber is placed between the poles of the electromagnet. At the beginning of operation the bombardment of various parts of the appli-

boron by moving ions leads to the liberation of gas, so that during the first several hours of operation the ion yield continuously increases as the gas evolution decreases. It was established that, for satisfactory operation of the apparatus, the pressure of extraneous gases must be below \(10^{-5}\) mm Hg. The most advantageous pressure of hydrogen, at which the largest currents are obtained, is a pressure of the order of \(10^{-4}\) mm Hg. By adjusting the apparatus (i.e., by selecting the deflecting voltage, the distribution of the magnetic field, the position of the dees, the deflecting electrode, and the filament, which is usually located at the center), the ion currents can usually be brought to an optimum value of 20–25 \(\mu\)A of deuterons with an energy of 5.8 MeV (when using dees with an internal peripheral thickness of 31.25 mm) and to 50 \(\mu\)A of ions with an energy of 4.3 MeV (with dees 50 mm thick and of smaller diameter). We found that the intensity of the ion beam obtained does not depend appreciably on the energy of the ions. Thus, for example, with narrow dees we obtained identical currents both for energies greater than 5 MeV and for energies smaller than this value. The greatest deuteron energy that we obtained with the modern cyclotron was 6.3 MeV.

Beams of fast protons are usually the same in current strength as beams of deuterons, but these particles have a smaller energy, since to impart the same energy to them a higher frequency of the alternating potential is required.

Helium ions. A cyclotron adjusted for obtaining deuterons is also adjusted for obtaining doubly charged helium ions. Since helium ions have twice the charge and mass as compared with deuterons, at each passage between the dees they acquire twice the amount of energy and reach the periphery of the dee with twice the deuteron energy under the same conditions. Therefore—apart from the difficulties associated with obtaining doubly charged helium ions at the center of the apparatus—the cyclotron is just as suitable for generating \(\alpha\)-particles with energies of 10 MeV as for generating deuterons of 5 MeV.

In one case we replaced the deuterium in the cyclotron chamber with helium and obtained 0.1 \(\mu\)A of \(\alpha\)-particles with energies of 11 MeV (moreover, with improved adjustment there was undoubtedly the possibility of obtaining currents several times larger). That the ion beam obtained consisted, in fact, of helium ions with so great an energy, and not of slower deuterons, was established with certainty in several ways. Thus, for example, it was found that the \(\alpha\)-particle beam had its greatest intensity at a magnetic field somewhat smaller than that required for a deuteron beam; its range was half that for a deuteron beam under the same conditions; and the scintillation it produced in air was equivalent to the scintillation produced by a deuteron beam four times more intense. Up to the present time we have not yet carried out experiments with these \(\alpha\)-rays, with the exception

...by the observation that bombardment by them of beryllium does not give as large a neutron emission as bombardment by deuterons of half the energy.

Neutrons. Among the many elements that emit neutrons under the action of deuteron bombardment, beryllium is especially remarkable. The emission of neutrons from a beryllium electrode bombarded by a 7 μA beam of 5 MeV deuterons is illustrated by the Wilson photographs in Fig. 7. With a chamber filled with hydrogen, and at a distance of 1.8 m from the beryllium electrode, the number of protons proves to be too large for a count to be made.

a                    b

Fig. 7. Photograph of recoil protons (in a Wilson chamber filled with hydrogen at \(p = 100\) m of Hg) produced by neutrons from a beryllium electrode bombarded by a 7 μA beam of deuterons with an energy of 5.5 MeV; a—distance from the chamber to the electrode 1.8 m, b—distance 12 m.

More suitable conditions are obtained by placing the Wilson chamber at a distance of 12 m. Although comparison of greatly differing numbers of neutrons is difficult, approximate determinations show that the emission of neutrons from the beryllium electrode of the cyclotron is \(10^5\) times greater than that given by a mixture of one curie of radon and beryllium.

This high intensity of the neutron flux facilitates the study of many problems of nuclear physics, but, on the other hand, also makes certain investigations more difficult. Among the practical difficulties connected with work with neutrons is the biological action of neutron rays, which makes protection of personnel necessary. Some biological investigations already carried out\(^{9-11}\) have shown that, with respect to biological action, the usual neutron emission of the beryllium electrode of a modern cyclotron is equivalent to the \(\gamma\)-radiation of 100 g of radium. Accordingly, at present, in order to protect the operator, the control...

the cyclotron is conducted from a distance of 12 m from the beryllium electrode, and suitable absorbing screens are placed in the path of the neutrons.

Artificial radioactive elements. With 5 MeV deuterons it proved possible[^12] to obtain radioactive isotopes of many elements from all groups of the periodic system. In many cases the yield of radioactive elements proved very considerable. Thus, for example, as a result of bombarding metallic sodium for one day with 5 MeV deuterons, more than 200 mg equiv. of radiosodium[^13] was obtained, i.e. an amount of radiosodium whose $\gamma$-activity is equal to the $\gamma$-activity of 200 mg of radium. The production of such large quantities of radioactive isotopes under laboratory conditions is of great significance, since it opens up new possibilities for physical and biological research.

LITERATURE

1. E. O. Lawrence and N. E. Edlefson, Science, 72, 376, 1930.
2. E. O. Lawrence and M. L. Livingston, Phys. Rev., 40, 19, 1932.
3. E. O. Lawrence and M. S. Livingston, Phys. Rev., 45, 608, 1934.
4. E. O. Lawrence, E. McMillan and R. L. Thornton, Phys. Rev., 48, 495, 1935.
5. M. S. Livingston, Rev. Sci. Inst., 7, 55, 1936.
6. W. Bleakney, Phys. Rev., 35, 1180, 1930.
7. D. H. Sloan, Phys. Rev., 47, 62, 1934.
8. D. H. Sloan, R. L. Thornton and F. A. Jenkins, Rev. Sci. Inst., 5, 675, 1935.
9. J. H. Lawrence and E. O. Lawrence, Proc. Nat. Acad., 22, (2), 124, 1936.
10. R. E. Zirkle and P. A. Aebersold, Nat. Acad. Sci., 22 (2), 1934, 1936.
11. J. H. Lawrence, P. A. Aebersold and E. O. Lawrence, Proc. Nat. Acad., 22, 543, 1936.
12. J. M. Cork and E. O. Lawrence, Phys. Rev., 49, 788, 1936; J. J. Livingood, Phys. Rev., 50, 425, 1936.
13. E. O. Lawrence, Phys. Rev., 47, 17, 1935.

1. In the Cornell cyclotron this defect was eliminated by the symmetrical arrangement of four filaments.[^5] ↩↩