US3746859A - High intensity neutron source - Google Patents

Abstract

Disclosed is a neutron generator in which a gas (such as a heavy isotope of hydrogen) or a mixture of gases, is ionized by any convenient means such as exposure to a magnetically stabilized arc. The ions are directed to an accelerator and the resulting high velocity ion stream is caused to impinge on a target. The interaction of the ion beam with the target material soon builds up a high density of gas in the target surface which, in turn, interacts with the incoming beam to produce an intense isotropic neutron output from the fusion of the isotopes.

Description

United States Patent [191 Hilton et al.

HIGH INTENSITY NEUTRON SOURCE Inventors: John L. Hilton, Walnut Creek;

Gordon W. Hamilton, Livermore, both of Calif.

Assignee: The United States of America as represented by the United States Atomic Energy Commis sion Filed: Apr. 22, 1970 Appl. No.: 30,724

US. Cl ..250/501, 313/61 S Int. Cl G2lg 3/04 Field of Search 250/845; 313/61 S References Cited UNITED STATES PATENTS [451 July 17,1973

Primary Examiner-James W. Lawrence Assistant Examiner-Davis L. Willis Att0rneyRoland A. Anderson [5 7] ABSTRACT 2 Claims, 1 Drawing Figure 3/]964 Reifenschweiler 250/845 X g 6 2 m PUMPING 3 SYSTEM 32 3 l 52 as as wk 64 78 a s m j S 2 J O Patented July 17, 1973 3,746,859

PU MPI N G SYSTEM (/OH/V L. 7 TO/VA 60200 M/ v /A'M/L 70/1/ INVENTORS mam ATTORNEYS 1 HIGH INTENSITY NEUTRON SOURCE BACKGROUND OF THE INVENTION Intense beams of high energy particles such as neutrons have important applications in analysis of materials, therapy (especially cancer therapy), nuclear reactor development, nondestructive testing, etc. and considerable attention has been directed to extending the useful life of the beam neutron output by water cooling the generating target, providing a good vacuum, reducing the sputtering of the target material, and in deuteriumtritium neutron producing systems, by preventing deuterium dilution of the tritium-loaded target.

Satisfactory solutions to most of the above problems have minutes. found in present-day technologies of cooling, high vacuum techniques, mixed isotope beams and the physical chemistry of materials. However, because of target deterioration, there does not yet exist a high neutron yield device which permits a small source size with long life. Typically, assuming a 6 cm stationary target, a neutron output of 10 neutrons/second requires the order of 10 ma at 300 Kv, if a pure deuterium beam is accelerated into a previously loaded tritiated target. This has a deuterium dilution half life of only 36 If the dilution is overcome by using a mixed beam (requiring twice the current) is used, the necessary thin and delicate target surface plating is sputtered or eroded away in a comparatively short time.

BRIEF SUMMARY OF THE INVENTION The present invention provides a structure capable of generating a high intensity neutron yield, for which structure the target life is considerably extended as a result of the incorporation, on a properly cooled surface, of a high beam current density deuterium-tritium beam. This is achieved by using a mixture of both gases in the ion source and target of deuterons and tritons (i.e., a mixed beam); this high beam current density ion beam is accelerated at an energy of the order of 100 Rev or higher onto a mixed gas target formed near the surface of a relatively thick stationary metallic target. The target gas concentration is maintained in dynamic equilibrium by continuous replenishment by the beam and can achieve a useful life limited only by the sputtering of the thick target. Thus, in this generator, it is not necessary to change the target frequently and, consequently, higher neutron yields per source area, economy and operational convenience are achieved.

BRIEF DESCRIPTION OF THE DRAWING The FIGURE shows the neutron source of the present invention in longitudinal cross-section.

DESCRIPTION OF PREFERRED EMBODIMENT Before discussing the shown preferred embodiment which is just one of the several possible configurations, it may be appropriate to outline some of the considerations which the present inventors had in mind prior to and during its development.

When a properly cooled simple Class A metallic target (e.g. Al, Cu, Mo, Ni, etc.) is struck by energetic deuterium and tritium ions, the concentration of hydrogen ions near the target surface, due to implantation by the beam, rapidly exceeds that expected from normal hydrogen solubility in the metal. As the concentration builds up, the neutron production level also rapidly increases. If a uniform current density in the beam is assumed, a one dimensional analysis (infinite extent in the other two dimensions) indicates that the hydrogen will diffuse out of the target, at a rate proportional to the concentration, until the rate of diffusion is equal to the rate of implantation; this presumes that all the diffusion is to the surface of beam impacts, which is a valid presumption since the target thickness is large compared to the l to 3 micron penetration depth of the beam. This limiting steady state condition may be represented by where i is the beam current density e is the electronic charge D is the diffusion coefficient dn/dx is the gradient in the number density of the hydrogen.

The neutron yield per square centimeter of target area, which is proportional to the product of the beam current and the density of the hydrogen in the target, would therefore vary as the square of the current density if all other factors are constant. However, the diffusion coefficient is a strongly increasing function of temperature, given theoretically by lD(T) D T" exp (b,,/T). An increase in beam current density will generally result in an increased target temperature and a related increase in the diffusion coefficient; Thus because of cooling limitations, the full i dependence of the neutron yield cannot be realized.

The neutron yield using a simple metallic target is therefore expected to follow the proportionality:

Ya Ai /D(T) where y is the neutron yield and A is the area of the target irradiated by the beam.

For mono-atomic single energiedl beams, the concentration of hydrogen in the target increases linearly with depth to the point where the ions stop. This penetration depth for a titanium target, is approximately 2.2 p. for KeV D and -2.8 p. for 170 Kev T and D, while the most probable fusion depth is between I to 1.5 p. for 170 KeV T and D. Therefore, the neutron yield should continue to increase with beam energy, but with a decreasing slope, as the energy is increased. For this type of target, the largest yields will result from the highest practical beam energy, the largest current density consistent with heat transfer requirements for the target, the largest area (consistent with any limits on source size that may be imposed) which implies the largest total beam current, a low diffusion coefficient of the target material at the operating temperature, and the lowest possible target temperature.

Aluminum and copper, have low diffusion coefficients for hydrogen and a relatively high thermal conductivity; they are attractive target materials, but because of its light atomic mass aluminum is expected to sputter at a higher rate than would a heavier target material such as copper.

The target temperature will be proportional to N, where V is the accelerating voltage of the ion source. Since heat transfer from the target :is expected to be the limiting factor in neutron yield optimization, this will limit the values of i and V that can be used for any given size or target.

Metal targets made of Class B metals (e.g. Ti, Er, Zr), if kept relatively cool, can initially contain a large quantity of hydrogen, apparently as a pseudo-hydride. Irradiation of such a pre-loaded target by a D and/or T beam will initially produce a high neutron yield, proportional to the total beam current. Again the yield will be higher for higher beam ion energies and the neutron yield will remain high until the hydride has been sputtered off, or the hydrogen has partially diffused out, or, in the case of a pure D beam impinging on a pure tritiated target, until the tritium in the target has been diluted by deuterium. At this point, the yield from the hydrogen isotopes initially contained in the target will decrease; however, the yield from hydrogen implanted by the beam will remain. Although the diffusion coefficient for hydrogen in the Class B metals is generally very large, the hydride formation'apparently traps the hydrogen in the target structure, giving high hydrogen concentrations if the target temperature is not too high. Thus, the hydrogenconcentration in a Class B metal target is not expected to be a function of the diffusion equations alone; the concentration may be augmented by hydride formation.

At higher beam current densities, the concentration of hydrogen trapped in the hydride becomes less important. For example, at a current density of 1.3 mA/cm and a total current of mA, the yield from a well-cooled target with implanted hydrogen diffusing out of it is about half that of the optimum yield from the hydride. At higher beam densities, if the temperature is held constant, the difference is less important.

With regard to target erosion rates, measurements made with a high beam current at a controlled temperature suggest a large difference between vapordeposited titanium Ti on a copper substrate and cold rolled titanium sheet stock; the sheet target withstands three times the operating period that completely destroys the plating under similar target cooling and beam intensities. Therefore, the use of nonplated sheet metal targets, on which are built up the hydrogen isotope mixture by a drive in mechanism, is preferred. These targets can have a thickness limited only by heat transfer and will not be eroded as quickly.

The mixed beam approach eliminates the regular tritium depletion problem along with another problem that is evident as target beam currents are extended from the presently available 0.1 mA/cm to to 50 mA/cm. This problem arises as the pure deuteron beam loads the target surface with the beam atoms. At a beam current of 0.1 mA/cm, this effect is relatively small, but it becomes significant at the high current density. For a total 75-150 mA beam on a target spot of 3 to 5 cm, near saturation of the target occurs in a hundred seconds or so. While this is an advantage for the drive-in-target on a mixed beam accelerator, it will limit the pure deuteron beam tritiated foil system to a target life of no longer than a few minutes (or perhaps to an hour or so if a large rotating target is used).

During operation, hydrogen concentration in the target will continue to rise until the hydrogen (both D and T) leaks back out of the target surface as fast as it is injected by the beam. At ambient temperature, this concentration will occur with an e folding time of approximately one hundred seconds. Therefore the incoming pure deuteron beam will quickly dilute the needed tritium and both isotopes will outgas from the target, giving the appearance of an enhanced tritium depletion in a matter of a few minutes.

The mixed beam accelerator eliminates not only target dilution but also another potential problem of the pure deuteron beam. Within minutes, a pure D accelerator significantly reduces the target tritium by implanting deuterium. For a mixed beam system, where there is an even ratio in the feed gas, the target isotope ratio will, of course, always remain at nearly 50 percent. With too high a D to T ratio the total yield will drop and the unwanted D-D neutrons may become an increasing fraction, producing a serious problem in both diagnostic and therapy applications of a generator. As long as the D to T ratio in the target remains close to 50 percent, it does not produce an objectionable quantity of D-D neutrons, the ratio of the D-D to the D-T fusion cross-section reduces unwanted low energy (i.e., 2.45 MeV) D-D neutrons to about two-thirds of 1 percent of the 14 MeV neutrons (for an accelerator energy of 150-170 KeV).

A useful system could involve a total beam current of -l50 mA at -170 KeV energy with an ion beam current density of 10-30 mA/cm the total power deposited in the target is therefore 13 to 25 Kw over the beam diameter (4 8 Kw/cm perpendicular to the beam for a 2 cm spot diameter).

The useful beam power density is limited by the ability of the target to dissipate this beam power and maintain a fairly low target surface temperature. An elementary heat transfer computation for a thin non-moving target cooled at the rear surface, shows that the limiting beam power density is proportional to KAT f/p where K is the thermal conductivity of the target material,

AT is the temperature difference between the target surfaces f is the allowable tensile stress p is the collant pressure The target is assumed to be as thin as is structurally possible, and therefore the thickness is proportional to f/p- The targets surface temperature must be kept well below its melting point to avoid excessive outgassing of the adsorbed D-T. In selection of a target material for a drive in target, it is desirable to maximize hydrogen content, which is a function of the diffusion rate, as well as K, T and f. Among common target materials, aluminum, copper, nickel and molybdenum are good choices with respect to these figures of merit.

It is practical to build targets at a simple or compound slope to the beam. A fixed target shaped to intercept the beam on a compound sloped surface, such as the surface of a cone could be used. The neutron yield can be observed from the downstream side, since the absorption of the desirable 14 MeV neutrons as they pass through the target will be relatively small; the downstream diameter of the resulting neutron source spot" will still be small, while the heat transfer requirement will be reduced. This will, of course, detrimentally reduce the effective beam current density and shorten the diffusion path length out of the target surface.

It is, of course, necessary to maintain a. stable target temperature by removing the beam energy from the target. It is possible to utilize the improved thermal transfer of a rotating target. However, this type of development indicates a comparatively high cost, and significant reduction of the target loading. Therefore, a non-rotating target is considered preferable. A beam current density of at least 23 Kwlcm or approximately 140 mA/cm, can be tolerated in a water cooled nonrotating cone target, if melt-down heat transfer were the only limitation on current density. The optimum shape for a fixed target will probably be a right cone with the base diameter about one half the depth. For this shape, the surface area capable of cooling is 4.12 times the surface area of a flat plate perpendicular to the beam. Further, the coolant flow rate is highest at the center, where the highest beam density and heat flux are encountered.

A cone-shaped copper target with a depth-todiameter ratio of 2 manifests a surface temperature of 200-225 C for a 100 mA beam at 175Kv, and a spot diameter of approximately 2 centimeters of the water coolant flow rate is 30-35 ft/sec across the back side of the target plate. Also it is possible to cool a thin bonded titanium sheet on a back-up plate of copper or other Class A metal not subject to hydrogen embrittlement. This could possibly maintain a surface temperature below 300 C at these power densities, although Ti has a relatively poor thermal conductivity. However, best yields will be obtained from target temperatures lower than can be obtained from a water coolant, at these power densities. This well-cooled target in combination with the high current density beam will provide startlingly higher yields.

Turning now to the preferred embodiment shown in the FIGURE, the neutron source may be regarded as comprising ion source 10, accelerator 112 and target 14.

Ion source has been described in the publication Plasma Physics, Volume 10, pages 687 to 697, Multimomentum 650 ma Ion Source" by G. W. Hamilton, J. L. Hilton and J. S. Luce as a further development of a ion source considered in the U. S. Pat. application, Ser. No. 712,197, filed Mar. 11, 1968. Briefly, a mixture of elemental gases, typically deuterium and tritium is supplied from a reservoir (not shown) at inlet to the area separating cathode 22 and anode 24 between which power supply 26 provides an arc discharge, thereby resulting in an ion plasma consisting of the nine Species 2( 3( A z T(3), T (6) and T (9), where the numbers in parenthesis are the ion mass numbers. A pair of electromagnets 30, 32 constrain the resulting ion plasma in area and density; electromagnet 32 is operated with a polarity opposing that of electromagnet 30 so that the resulting flux line can be controlled in density and area. An ion beam is extracted and electrostatically focused by a three level arrangement consisting of reflector electrode 34, extractor electrode 36 and decelerating electrode 38, fed from power supplies 40, 42. Typical operating voltages are, between electrodes 34 and 36 +10 to +20 Kv and between electrodes 36 and 38 -2 to 4 Kv. The beam is fed through apertures 44 in electrodes 34, 36, 38, and through halo baffle 46 into vacuum vessel 48. [on source 10 is cooled as appropriate through coolant fittings 50, 52. Halo baffle 46 operates to protect accelerator 12 from the diffuse, nonfocussed halo surrounding the ion beam as well as to protect ion source 10 from high energy, backstreaming electrons.

The active elements in accelerator 12, which is entered by the beam through entry 56in case 58, are electromagnet 54, of about six Kgause capacity, which collimates the beam, and case 58 is energized by voltage on the order of l l0l Kv to accelerate the beam to ward target 14. Electromagnet 54, as shown, is oriented coaxially. The voltage is supplied to case 58 through insulator 60, and insulator 60 also serves as partial support for accelerator 12. Flow of coolant is provided through fittings 64, 66. With regard to structural materials, case 58 is of mild steel to provide a magnetic flux return path, entry 56 is of copper brazed to case 58 and copper separator 57 such that the electric field abode entry 56 and the fringing magnetic: lines are separated.

The beam emitted from accelerator 12 impinges upon target 14, thereby releasing neutrons isotropically by the fusion reaction.

Target 141 comprises sheet 78 of a metal such as copper, aluminum, titanium, nickel or molybdenium, the rear surface of which is cooled by coolant entering through fitting 76. The beam of ions (the hydrogen isotope mixture) impinges on the front surface of sheet 78 and builds up thereon a film of mixed hydrogen isotopes which, as it diffuses out and generates neutrons, is continuously replenished by the beam. With regard to target cooling, water may be used to give fair yields and life. However, results are surprisingly good if the cooling system is capable of very high power density at relatively low temperatures; an admirable example has been found to be a non-boiling liquid metal heat transfer loop such as the eutectic NaK or mercury to maintain a target temperature at the range 75 to C. Other high heat flux systems are just as valuable if the front surface of sheet 78 can be maintained at the low temperature while being heated by the high current density beam.

In the present system, the accelerator and ion source operate in a relatively high vacuum (3 X 10 torr), which reduces background ionization, backstreaming electrons and x-ray radiation, and the feed gas (T and D) is continuously recirculated through the system by a fail-safe closed loop vacuum system, thereby minimizing danger due to operating with several curies of tritium in a free gaseous form. Thus, gas is emitted from vessel 48 at fitting 47, piped to pumping system 62 for compression and transfer back to inlet 20. Obviously, an additional storage inventory of the gas may be sealed in this recirculation loop and used as needed.

What is claimed is:

1. A neutron generator, comprising:

a source of ions;

beam-forming means for the ions from said source,

the beam including a plurality of isotopic ions;

an accelerator for the beam from said beam-forming means;

a target to receive the beam from said accelerator having a surface capable of occluding some of the beam atoms so as to emit neutrons as a result of impact by the beam and capable of being replenished by the ions of the beam, said target comprising a homogeneous metallic plate;

a vacuum vessel in which said accelerator and said target are enclosed, and

means to cool said target at its back side surface, said cooling means comprising a flow of liquid metal capable of maintaining said target at a temperature in the range of 75 to 200 C.

2. A neutron generator, comprising:

a source of ions;

beam-forming means for the ions from said source,

the beam including a plurality of isotopic ions;

an accelerator for the beam from said beam-forming by the ions of the beam, and

means; a pumping system connected between said target and a target to receive the beam from said accelerator said ion source operative to receive gas in the vicinhaving a surface capable of occluding some of the ity of said target, compress the gas and transmit the beam atoms so as to emit neutrons as a result of imcompressed gas to said ion source. pact by the beam and capable of being replenished

Claims (1)

1. 2. A neutron generator, comprising: a source of ions; beam-forming means for the ions from said source, the beam including a plurality of isotopic ions; an accelerator for the beam from said beam-forming means; a target to receive the beam from said accelerator having a surface capable of occluding some of the beam atoms so as to emit neutrons as a result of impact by the beam and capable of being replenished by the ions of the beam, and a pumping system connected between said target and said ion source operative to receive gas in the vicinity of said target, compress the gas and transmit the compressed gas to said ion source.

Images