Neutron source with beam shaping apparatus for radiography

Abstract

A neutron generator has a neutron source generating an ion beam bombarding a Titanium target having a first diameter, the target embedded in a pre-moderator, emitting fast neutrons isotropically, a portion of the fast neutrons moderated in passing through the pre-moderator and exiting through a lowermost surface of the pre-moderator, and a plate of moderating material abutting the lowermost surface of the pre-moderator, the plate having an opening therethrough in a shape of a truncated cone with an axis aligning with direction of the ion beam, a depth, a major diameter of at the upper surface of the plate and a minor diameter at the lower surface of the plate, the opening forming a funnel through which neutrons pass. Neutrons enter the funnel and are to exit through the minor diameter of the funnel, providing a neutron beam with a spot size useful for neutron radiography.

Claims

1. A neutron radiography generator, comprising: a modular neutron source generating an ion beam in a first direction, the ion beam bombarding a Titanium target having a surface comprising a first diameter, the target embedded in a pre-moderator having a lowermost surface, emitting fast neutrons isotropically, a portion of the fast neutrons moderated in passing through the pre-moderator and exiting through the lowermost surface of the pre-moderator; and a plate of moderating material abutting the lowermost surface of the pre-moderator, the plate having an upper surface, a lower surface, a thickness, and an opening therethrough in a shape of a truncated cone with an axis aligning with the first direction of the ion beam, a depth, a major diameter of at the upper surface of the plate and a minor diameter at the lower surface of the plate, the opening forming a funnel through which neutrons pass; wherein neutrons exiting the pre-moderator through the lowermost surface thereof enter the funnel and are collimated through the funnel to exit through the minor diameter of the funnel, providing a neutron beam with a spot size useful for neutron radiography.

2. The neutron radiography generator of claim 1 wherein the thickness of the plate of moderating material is greater than the depth of the funnel, further comprising a sleeve of high-density neutron reflecting material having in inside diameter equal to the major diameter of the truncated cone shape of the funnel, the sleeve lining a circular hole in the plate of moderating material, the hole having a first depth from the upper surface to the major diameter of the truncated cone shape of the funnel, such that the depth of the circular hole and the depth of the funnel equals the thickness of the plate, and the inside diameter of the sleeve of reflecting material meets the major diameter of the truncated cone shape of the funnel.

3. The neutron radiography generator of claim 2 further comprising a disk of Bismuth having a diameter equal to the inside diameter of the sleeve and a length less than the first depth of the hole, and a solid cylinder of sapphire crystal having a diameter equal to the inside diameter of the sleeve and a length such that the length of the bismuth disk and the sapphire cylinder equal the length of the hole from the upper surface of the plate of moderating material to the major diameter of the funnel, the bismuth disk and sapphire crystal attenuating gamma radiation and attenuating fast neutrons.

4. The neutron radiography generator of claim 1 further comprising a layer of shielding material cladding the outer surfaces of the generator including the lower surface of the plate of moderating material, with a hole through the shielding material at and equal in diameter to the minor diameter of the funnel.

Description

BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS

(1) FIG. 1A is a transverse cross section of a modular neutron generator in an embodiment of the present invention.

(2) FIG. 1B is a longitudinal cross section of the modular generator of FIG. 1A.

(3) FIG. 2 is a diagram showing geometrical un-sharpness or blur in radiography using neutron sources.

(4) FIG. 3 is a cross section view of the LVFG, its pre-moderator, and a convergent funnel for a small area source in an embodiment of the invention.

(5) FIG. 4A illustrates thermal neutron yield (n/sec-cm.sup.2) as a function of distance x across an axis of the convergent funnel of a simple beam-shaping assembly (BSA) in an embodiment of the invention.

(6) FIG. 4B illustrates measured thermal neutron yield (n/sec-cm.sup.2) as a function of distance x across the axis of the convergent funnel of the simple BSA (Slab with funnel) and for the case of no funnel and the moderator with the thickness L.sub.1+L.sub.2=9.5 cm,

(7) FIG. 5 is a cross sectional view of components required for producing a knife edge image for a computer simulation in an embodiment of the invention.

(8) FIG. 6 illustrates a computer simulation of a knife edge made of Gadolinium (Gd) for fast neutrons (E>0.5 eV) and thermal neutrons (E<0.5 eV) in an embodiment of the invention.

(9) FIG. 7A shows image flux as a function of x, the transverse distance across a detector array in an embodiment of the invention.

(10) FIG. 7B shows image flux as a function of x, the transverse distance across the detector array in an embodiment of the invention.

(11) FIG. 8 is image resolution (mm) of the knife edge as a function of L, the distance between the end of BSA and the knife edge in an embodiment of the invention.

(12) FIG. 9 shows estimated time (sec) it takes to measure the knife edge as a function of L, the distance between the end of BSA and the knife edge in an embodiment of the invention.

(13) FIG. 10A is a longitudinal cross-sectional of the modular generator with a BSA composed of Bismuth foil, a funnel, an exit aperture, and a sapphire crystal with a reflecting graphite sleeve in an embodiment of the invention.

(14) FIG. 10B is a transverse cross-section view of the modular generator of FIG. 10A taken along an axis of the acceleration chamber, and along the axis of a turbo vacuum pump in an embodiment of the invention.

DETAILED DESCRIPTION OF THE INVENTION

(15) FIGS. 1A and 1B are cross-sectional views of a modular neutron generator 118 known to the inventors designed to produce a substantial flux of thermal neutrons. Modular generator 118 comprises a pre-moderator 108 that is made of material known to moderate energy of fast neutrons to thermal energies. In most embodiments for thermal neutron production the pre-moderator may be a shaped, solid block of material such as High-Density Polyethylene (HDPE) or Ultra High Molecular Weight (UHMW) polyethylene. Selection of material and its thickness is determined at least in part by the desired neutron spectrum (e.g. thermal or epithermal) and desired neutron yield.

(16) Modular generator 118 has four important elements in this example: (1) a deuterium ion source 102, (2) an acceleration chamber 100, through which deuterium ions 104 are accelerated, and (3) a titanium target 106 that is bombarded by the deuterium ions to produce high-energy neutrons 110. Deuterium ion source 102 has an attached microwave source 160 in the implementation, and microwave slug tuners 172. In operation Deuterium gas is leaked slowly into a plasma ion chamber 174 at an upper end of the acceleration chamber, where microwave energy ionizes the gas, creating deuterium D.sup.+ ions 104. The gas is ionized by microwave energy, and Deuterium ions (D.sup.+) 104 are created and accelerated through an ion extraction iris 138 into acceleration chamber 100, and through an electron suppression shroud 180 which deflects back-streaming electrons from being accelerated back into the plasma source, which could damage the apparatus. Electrons are created by collisions of the D.sup.+ ions in the deuterium gas that are being created in the acceleration chamber.

(17) The deuterium ions are positively charged, and target 106 is negatively charged to a level of from 120 kV to 220 kV, and the D.sup.+ ions are strongly attracted to negatively biased titanium (Ti) target 106. Acceleration chamber 100 is connected to a turbo vacuum pump 124 that provides a modest vacuum in one embodiment of about 10.sup.6 Torr, minimizing scattering of the D.sup.+ ions as they travel from extraction iris 138 to target 106. Titanium target 106 is positioned in a cavity 181 at the bottom of the chamber, the cavity formed in the pre-moderator material. Pre-moderator 108 has a passage for a high voltage cable and fluid cooling channels to and from the target. Pre-moderator 108 acts as a high-voltage insulator and as a mechanical support for the target at a high negative bias. When in operation the D.sup.+ ions in the ion beam are attracted to the titanium target 106, where fast neutrons are produced in a resulting DD fusion reaction.

(18) A major issue for fusion sources using the Deuterium-Deuterium (D-D) reaction to produce fast neutrons that must be moderated to thermal neutron energies is that fast and epithermal neutrons as well as high energy gamma emission are usually part of the moderation of the fast neutrons to thermal energies. These components can accompany the thermal neutrons penetrating the absorbent material of the iris and may effectively increase the aperture size D if the extraneous radiation can penetrate the iris materials, blurring the desired image.

(19) In large reactors, thermal neutrons have been obtained which have mixtures of thermal, epithermal, and fast neutrons along with gamma and x-rays. Applications such as neutron radiography and radiotherapy usually require the neutron energy to be confined to single neutron energy bands without x-ray or gamma components. There need be methods to eliminate the unwanted radiation components.

(20) Modular DD fusion generator 118 in embodiments of the present invention uses a small titanium target (e.g. a 5-7 cm diameter disk of titanium backed by water-cooled copper fins) to produce neutrons. FIGS. 1A and 1B are cross-sectional drawings of a modular neutron generator to produce maximum thermal neutrons. The target is supported directly on the pre-moderator, which is an integral part of the apparatus in this implementation. The Ti target may be attached with fasteners to the pre-moderator block and may be sealed to the block with an O-ring. Targets in embodiments of the invention can be easily manually removed and replaced. They also have a long lifetime and have been tested for over 4000 hours with no failures.

(21) In the following descriptions reference is made to accompanying drawings that form a part of the disclosure and teaching of the present invention, and which illustrate specific embodiments in which the invention may be practiced. It is to be understood that other embodiments may be utilized, and structural changes may be made without departing from the scope of the present invention.

(22) The accelerator structure in embodiments of the invention is compact and includes a pre-moderator 108 that adds only about 4-5 cm of High-Density Polyethylene (HDPE) or 15-20 cm of polytetrafluoroethylene (PTFE) Teflon to produce a first stage of neutron beam tailoring in embodiments of the invention. The pre-moderator in these embodiments is an integral part of each modular generator, as is taught below with reference to several figures. Other short-length attachments are added to the pre-moderator to further improve the neutron beam in beam purity, size and shape, making the modular neutron generator a highly versatile source of neutrons. A primary application for the unique apparatus taught in this application is thermal neutron radiography, which requires a small source size, high neutron yield (n/cm.sup.2) and high beam purity. High thermal neutron beam purity is achieved in embodiments of the invention by minimizing other neutron and photon components that may be introduced during the DD fusion process and moderation of the 2.5 MeV neutrons to thermal energies. The filtering process is accomplished using neutron filters; both low pass and high pass filters. To maximize the resulting neutron flux and minimize the neutron source size, these filters and collimators are minimalized in length and proximity to the neutron generator. This results in a highly compact and useful neutron source for many applications.

(23) As is known for most radiation sources, a small source size is required for image clarity and sharpness. Geometric un-sharpness or blur refers to loss of image detail caused by the finite size of the source diameter. This is true of sources of radiation such as x-ray tubes, where an electron beam of diameter D strikes the anode (e.g. tungsten target) of the x-ray tube, producing a source size of diameter D. In other sources of radiation used for radiographic imaging (example synchrotron radiators or plasma pinch sources of x-rays), the spot or source size D can be defined by the either a slit or metal aperture that defines where the x rays or neutrons are being emitted. In the present invention, the aperture is defined by an exit aperture D2 of a funnel or conical aperture.

(24) The source or spot size can result in geometric un-sharpness, blur, or the loss of image detail caused by the finite size of the neutron emission size of diameter D.sub.2. In neutron sources the spot size D.sub.2 is defined by an exit aperture after the neutrons have been moderated to thermal neutron energies. The convergent aperture is defined by an exit aperture which can be made of different materials, such as HDPE and graphite, which result in collection and collimation of the thermal neutrons.

(25) As shown in FIG. 2 with simple geometry, minimum blur (B.sub.o, 226) may be achieved if the neutron source-to-detector distance L, 240, and the source-emission diameter, D.sub.2, 228, results in a small D.sub.2/L. This permits a resolution of x(D.sub.2/L) for an object that is x thick 242. For higher resolution (less blur B.sub.o, 226) this relationship requires the source diameter D.sub.2 to be small as possible and as far from the detector distance, L 240, as possible. Placing an aperture over the neutron source that limits source size D.sub.2 will reduce the flux density (N/sec-cm.sup.2) where N is the number of neutrons per sec that pass through the aperture to the detector. The optimum is to reduce D.sub.2 without loss of neutrons. Focusing optics achieve this for photons, but for neutrons in the thermal energy range, this is difficult in a short distance L, without loss of neutrons from an ever-expanding neutron beam.

(26) Thermal neutron collection can be achieved with a funnel 222 (FIG. 3) to both collect and channel neutrons into a small spot size D.sub.2 with increased thermal flux at the exit D.sub.2 of the funnel. The compact DD fusion source with a short (4-5 cm) moderator slows the fast 2.5 MeV neutrons to thermal energies in a short distance from the fast neutron source (the titanium target). The neutrons are then collected by a relatively short (e.g. L.sub.2=3 to 4 cm) funnel shape formed into a slab of HDPE. Both moderation and scattering continue to occur along the funnel length L. Simulations show that this results in an increase in flux 2 to 3 times and spot sizes of 1 to 3 cm in diameter depending on the geometry of the cone and size of the fast neutron emitter (Ti target).

(27) Compactness of the DD fusion generator, and shortness of the pre-moderator to produce and collect thermal neutrons also allow for the use of other devices in the neutron beam. These devices include short lengths of sapphire crystals and bismuth which can reduce fast neutrons and gamma emission in the neutron beam, thus cleaning up the beam and achieving a relatively pure beam of thermal neutrons.

(28) In fusion devices, such as the LVFG in the present invention, the number of neutrons is limited. The use of a compact fusion generator with relatively small spot sources of neutrons permits neutron filters to also be compact and close together. The modular generator combines multiple functions that were separate functions in the prior art. These integrated functions include both neutron production and neutron energy band selection. This method shortens the overall length of the device and ensures high fluxes.

(29) As was shown in FIG. 2, a critical requirement for image resolution is a small diameter source size of high neutron yield. FIG. 3 is a cross section of the neutron generator. In one embodiment a simple Beam Shaping Apparatus (BSA) in the form of funnel 222 is added, which may produce a small source size for radiography. In one embodiment the convergent funnel has a base diameter D.sub.1 of 6 cm and exit opening D.sub.2 of 1.5 cm. The funnel collects thermal neutrons being emitted by pre-moderator 108. The section of FIG. 3 is taken along an axis of an acceleration chamber 100 for ion beam generation and at a right angle to the axis of a turbo vacuum pump, 124, that is part of the modular generator.

(30) In one embodiment an input aperture D.sub.1 (252) of the funnel 222 is placed approximately at the L.sub.1=5.5 cm away from the titanium target 106 whose diameter is 6 cm. This location is where the thermal neutron flux has been shown to be maximum and where collecting the thermal neutrons maximizes the neutrons at the exit aperture D.sub.2, at least for this particular example.

(31) As shown in FIG. 3, a short Beam Shaping Assembly (BSA) in the shape of funnel 222 is provided after Ti target 106 and the short moderator or pre-moderator 108, where the thermal neutrons may be collected by a large aperture D.sub.1 and directed to a smaller aperture D.sub.2 at the end of funnel 222. In a simple embodiment, funnel 222 is an inverted cone machined into the BSA support structure (or plate) 234 made of HDPE as shown in perspective view FIG. 3 with an entrance aperture D.sub.1 and an exit aperture D.sub.2. Cone-shaped funnel 222 in this embodiment is formed into the plate 234 of HDPE. Other materials such as Teflon, UHMW polyethylene, or graphite can be used for the plate. As shown in FIG. 3, an exit aperture 250 is also machined in the shielding 248 to define the thermal neutrons coming from the exit aperture.

(32) After collection at aperture D.sub.1, the resulting thermal neutron beam exits at aperture D.sub.2, providing an increased flux and smaller source size for the thermal neutrons when compared to a simple pre-moderator.

(33) FIG. 4A is a plot of flux density vs. lateral dimension in cm. at D.sub.2 in one example. In the example of FIG. 4A, the spot size is roughly 1.5 cm at diameter (D.sub.1). Without the funnel 222 the diameter of the source size would be a disc 6 cm in diameter and expanding (FIG. 4). The use of the moderator in direct or near contact with Titanium target 106 is to produce a maximum thermal neutron flux at the top aperture D.sub.1 of the funnel 222. Adding a funnel 222 of modest length (L.sub.2=4 cm in this example) directly on to the pre-moderator permits a maximum number of neutrons and higher thermal neutron flux to be collected and a smaller source size to be obtained. The close stacking of the pre-moderator 108 and the BSA (Funnel 222 in HDPE slab 234) permits a maximum number of thermal neutrons to be obtained with a small source size. FIG. 4A clearly illustrates the beneficial effect of the slab with the funnel.

(34) To reduce the size of the thermal neutron beam emitted by the HDPE moderator at its maximum thermal neutron flux (n/sec-cm.sup.2), funnel aperture 222 is added along the axis of the generator 118 as defined by direction of the D.sup.+ ion beam 104, and the titanium target 106. As shown in FIGS. 3A and 3B, the dimensions of the apertures in this example are: D.sub.2=1.5 cm, D.sub.1=6.0 cm, L.sub.1=4.0 cm, L.sub.2=3.0 cm., E<0.5 eV. The plot in FIG. 4 is from a Monte Carlo Neutral Particle (MCNP) code simulation for these parameters. For the case with no funnel, the moderator shows a beam spot of FWHM of 10 cm in diameter. With the funnel 222 in place, the FWHM of the emission is roughly 2 cm in diameter, defined by a known technique called Full Width Half Maximum (FWHM) and the neutron flux is 2 times larger than an uncollimated neutron beam. The source size after the funnel has a full width half maximum of 2 cm, whereas for pre-moderator only, it is 10 cm. One can achieve a high resolution if the source is at an enough distance L, and small enough emission size diameter, D.sub.2, to have a small D.sub.2/L. This permits a resolution of x(D.sub.2/L) for an object that is x=1 mm thick. For a D.sub.2 of 1.5 cm and an L=20 cm, the resolution is 0.1 mm.

(35) A prototype of the apparatus has been built and tested at the time of filing the present patent application. The apparatus is shown in FIGS. 3A and 3B with a funnel collimator as shown in the two figures, as described above. The dimension of the prototype funnel collimator are D.sub.1=6 cm and D.sub.2=1.5 cm with L.sub.1=5.5 cm and L.sub.2=4 cm. These dimensions may be different in other embodiments. The method of detection of the thermal neutron is a linear array of small chips of NaCl. These chips were activated by the in-coming thermal neutrons for a measured length of time. The radioactive flux was then determined by neutron spectrographic means. The results are shown in FIG. 4B. The general shape and magnitudes are comparable with the simulated results of FIG. 4A. The peak flux of FIG. 4B is smaller (610.sup.6 n/sec-cm.sup.2) than that of the simulation of FIG. 4A. (910.sup.6 n/sec-cm.sup.2). The differences are from the resolution of the detectors (measured vs simulated). The simulation of FIG. 4A used a shorter pre-moderator, L.sub.1=4.0 cm. vs L.sub.1=5.5 cm. However, in both case L.sub.2=5.5 cm and the comparison between 4A and 4B shows that the HDPE funnel BSA is indeed effective as a method of collimation.

(36) In examples of BSAs, convergent collimators are used. Other geometries can be used such as divergent collimators, which reverse the direction of the truncated cone. These have been used throughout the nuclear reactor source industry. Some collimators have a divergent-convergent shape, which can result in a shorter BSA length and higher thermal neutron flux.

(37) To see how well the conical aperture source performs, an image of a 1.0 mm thick Gadolinium (Gd) knife edge 238 is simulated, placed in front of a detector array 224 made of 5.0 mm of H.sub.2O, 142, as shown in FIG. 5. For this calculation, the desirable image properties are (a) a flux of greater than 10.sup.3 n/(s-cm.sup.2), and (b) a desired resolution is 1 mm, with the contrast of an order of magnitude or greater.

(38) To see how well the conical aperture neutron source works, an image of a 1.0-mm-thick-Gadolinium (Gd) knife edge 238 with a conical BSA 222 is simulated. Water (5-cm, 240), is used to simulate materials in the detector 246 which scatter the thermal neutrons. The Gd knife edge 238 is placed on the upstream side and in front of the H.sub.2O, 240 to determine resolution and contrast. With the arrangement shown in FIG. 5, L is the distance between the BSA 222 and the Gd-knife edge 238. In all the simulations shown in this submission, a diode array detector 246 is assumed to be at a distance of 1.0 mm from the back of the H.sub.2O 240. The BSA 222 is an air cone embedded (machined in HDPE with an entrance aperture of D.sub.1=6 cm and an exit aperture of D.sub.2=1.5 cm). The neutron flux (n/cm.sup.2-sec) is found using a Monte Carlo Neutral Particle (MCNP) simulation shown in FIG. 6, which compares the thermal (E<0.5 eV) and fast (E>0.5 eV) energy neutron flux components.

(39) To attempt to achieve these properties, various modulator and BSA arrangements are considered. The object is a 1.0-mm-thick Gd knife edge, 238, backed by 5-mm of H.sub.2O, 240, and is placed on the upstream side of the H.sub.2O to determine resolution and contrast. In the generator, the ion beam 104 strikes a 5-cm diameter Ti target 106 and 2.5-MeV neutrons are emitted into the pre-moderator of thickness L.sub.1=4 cm. FIG. 6 shows the simulated Monto Carlo in Transport (MCNP) thermal flux 244 and fast 242 neutron fluxes as a function of x across the detector array 224. For this arrangement, the resolution is calculated to be 8.5 mm. Other simulations with somewhat different parameters gave resolutions of 3.3 mm.

(40) The fast neutrons created from the moderation process are shown in the top curve in FIG. 6. Note, the fast neutrons 242 in the simulation are not being contrasted across the Gd knife edge. Indeed, most neutron detectors cannot easily distinguish between fast and thermal neutrons. However, separation between the fast and the thermal neutrons can be achieved by pulsing the neutron beam. The generator can be modulated by pulsing the ion beam, interrupting the microwave power that is creating the D.sup.+ ions. Pulses as short as 10 secs have been produced by the inventors using this method. The fast neutrons are created at the titanium target and then pass through 4 cm of the HDPE pre-moderator where approximately 50% are reduced to thermal energies. For a 10 sec pulse of neutrons, the difference in speed (2.2 km/sec for thermals and 1.410.sup.4 km/sec for fast neutrons) results in thermal neutrons lagging by 20 sec when they reach the detector at 20 cm from the aperture. Timing of the reading of the fast and thermal neutron pulses detected by a charge-coupled device (CCD) camera allows to distinguish between the two images; one caused by the thermal neutrons and the other by the fast neutrons. The neutron source properties are sufficient neutron flux (e.g. 10.sup.3 n/(s-cm.sup.2) or greater coming from a small spot size. Millimeter resolution is desired having an image contrast of an order of magnitude or greater.

(41) With the arrangement in FIG. 5 and L=2.0 cm, FIG. 6 shows thermal and higher energy neutron flux. For these parameters, the flux is good, the contrast is good, the transverse dimension is good, but the resolution is 8.5 mm.

(42) To improve the resolution, different parameters for D.sub.1 and D.sub.2 are selected, and different distances from the BSA aperture D.sub.2 228 to the knife edge are tried, L. All other parameters for the generator, knife edge and detector array are the same. Larger apertures D.sub.2=5 cm, and D.sub.1=8 cm. The distance to the knife edge L=2 cm. Flux as a function of x(cm) is shown in FIG. 7A. The maximum thermal flux is a healthy 3.410.sup.6 n/cm.sup.2-sec, but the resolution is 6-mm. Increasing the distance L to 100 cm, as shown in FIG. 7B, sub-mm resolution is achieved. Throughout these simulations, HDPE is used in both the moderator 108 and in the BSA 222, the thickness of the Moderator is L.sub.1=4 cm, and the thickness of the BSA is L.sub.2=4 cm. Plotting the resolution as a function of L in FIG. 8, the resolution continues to improve. We can achieve the desired resolutions of 1 mm for L=20 to 50 cm.

(43) However, with increasing distance L, the available neutron flux for imaging decreases resulting in an increase in measurement time for the collection of neutrons. This may be estimated with a simple assumption that each diode of the array needs around 250 neutrons for a measurement. Plotting the measurement time in FIG. 9 as a function of L, we see that we achieve measurement times in the order of seconds. With L=20 to 50 cm, the detection of the knife edge image with 1-mm resolution can take place within 1 to 3 seconds, a time more than adequate for achieving a high-quality image.

(44) Convergent collimators are used throughout this submission, but divergent collimators or combinations of both convergent and divergent collimators may also be used. In the divergent conical collimator, the cone may be lined with grazing angle reflective materials such as Cadmium, Indium, B.sub.4C or Boron. The cone is made of a machinable material and lined with Cd, In, or B. The use of HDPE without a lining (Cd, In, or B) in the convergent collimator, as taught in this disclosure and demonstrated by MCNP simulation, ensures that both collection and further moderation of the neutrons to thermal energies can be achieved.

(45) The resolution may further be improved by attenuating the fast neutrons by means of a low pass filter, in which thermal neutrons are transmitted, while fast neutrons are attenuated. Fast neutrons need to be attenuated or the detector's sensitivity to the fast neutrons needs to be suppressed. To eliminate fast neutrons, a 9-cm long sapphire crystal may be added to the BSA. Sapphire (Al.sub.2O.sub.3) is an effective fast-neutron filter because its transmission for neutrons of wavelengths less than 0.04 nm (500 meV) is less than 3% for a 100 mm thickness.

(46) Current technology enables large diameter, single-crystal sapphire ingots to be grown using what is known as the Kyropoulos technique. Diameters of sapphire ingots may be 5 to 12 cm with thickness of 5 to 20 cm long. It is estimated that 7 cm of sapphire may reduce fast neutron yield by an order of magnitude while transmitting roughly 80% of the thermal neutron flux.

(47) FIGS. 10A and 10B show a modular generator with many of the same elements as shown in FIG. 3, but with additional detail of HDPE rectangular plate 234 and elements in a passage through plate 234 for filtering a neutron beam before the collimating funnel 222. Pre-moderator 108 using 5.5 cm of HDPE produces in this example a neutron beam that is roughly equal in fast and thermal neutrons. As shown in FIGS. 10A and 10B a sapphire crystal 220 and a Bismuth filter 238 may be added that reduce the number of high energy gamma rays produced by thermal neutron capture of hydrogen in the HDPE. Sapphire crystal 220 in this embodiment may be a cylinder with an outer diameter fitting into sleeve 236, between 5 to 15 cm long with a high enough average atomic number, Z, to attenuate the gamma background. Bismuth filter 238 is a disk of the diameter of the sapphire crystal fitting into sleeve 236 just above the sapphire crystal. Sapphire filter 220 is added just below the pre-moderator 108 and into a sleeve 236 which encloses the sapphire filter 220, acts as a reflector and attenuates the neutrons that are outside the sapphire crystal 220. A funnel 222 which is L.sub.2=4 cm long and a reducing aperture D.sub.1 which directs the thermal neutrons to the desired minimal aperture D.sub.2 for a desired small beam diameter D.sub.2. It is important to minimize the distance (L.sub.1+L.sub.2+L.sub.3+L.sub.4+L.sub.5) from the Ti target 106, to the aperture D.sub.2, 250. This maximizes the thermal neutron yield delivered to the small aperture D.sub.2, while minimizing spurious radiation of gammas and fast neutrons. This maximizes the number of thermal neutrons required for good image resolution and contrast.

(48) The D.sup.+ ion beam 104 strikes the titanium target 106, where D.sup.+ ions are embedded and creates the DD fusion reaction, resulting in the isotropic emission of fast (2.5 MeV) neutrons. To maximize the flux being transmitted through the sapphire filter, the crystal 220 is aligned with its axis in line with the ion beam 104 direction and the maximum incoming thermal neutron beam. The sapphire crystal length and orientation is selected to maximize the thermal neutron transmission preferably in a wavelength range of 1.2 to 2.5 A, while minimizing fast neutron wavelengths of less than 1 Angstrom. Fast neutron transmission, T, decreases exponentially with crystal length, L: or T=I/I.sub.o=exp(L/L.sub.o). In this embodiment a sapphire crystal length of 70 mm is selected, which roughly gives an order of magnitude decrease in the fast neutrons relative to the thermal.

(49) Assuming a mixed neutron beam is being transmitted thru the Sapphire filter, it is desired to maximize the thermal neutrons while suppressing the fast and epithermal neutron components and the gamma rays produced in the HDPE pre-moderator material. It is desired in this example to maximize transmission of a 2.5 cm beam, defined by the definition of Full Width Half Maximum (FWHM), of thermal neutrons down a cylinder 220 composed of Sapphire crystal. Thermal neutrons are being scattered during transmission and some are lost outside the sapphire crystal. However, a sleeve 236 of high density (or high Z) reflecting material just outside the crystal surface may reflect the thermal neutrons back into the crystal and thereby increase the total neutron yield at the exit to the BSA. In this example the sleeve 236 is Bismuth surrounding the sapphire crystal. The high Z sleeve critical angle reflects any grazing-angle thermal neutrons but scatters and absorbs the higher energy neutrons that pass from the sapphire to the Bismuth. Ideally, the thermal neutrons travel down the sapphire cylinder and the fast neutrons get absorbed or scattered. The conical aperture at the end of the Sapphire crystal acts to transmit the thermal flux out a small aperture (D.sub.2). The conical aperture 228 diameter tapers from D.sub.1=6 cm to D.sub.2=1.5 cm. in this implementation There are other parameters and materials, such as graphite, that can be used to form the conical aperture 222 and the rectangular plate 234.

(50) Because tungsten target 106 is on the plastic (HDPE or Teflon) pre-moderator 108, fast neutrons coming from the target immediately enter the pre-moderator and can be moderated to thermal or epithermal energies. A short Beam Shaping Assembly (BSA) is provided below the Ti target and the pre-moderator, where some of the thermal neutrons may be collected and directed to a small aperture at the end of the BSA. A short, L.sub.5, iris 250 is placed just below the BSA. The material of the iris 250 may be made of lead and B.sub.4C. In its simplest embodiment, the BSA is an inverted cone 222 as shown in FIGS. 3A and 3B. The HDPE of the BSA acts as a reflector and collimator of the thermal neutrons. As stated elsewhere, the collimator can be made of other materials such as graphite. The resulting thermal neutron beam at the aperture of the BSA gives a higher flux and smaller source size for the thermal neutrons when compared to a simple moderator. In the example shown, the spot size is roughly 1.5 cm in Diameter (D.sub.2). Without the BSA the diameter of the source size at the pre-moderator would be a disc 6 cm in diameter and expanding. The idea behind the use of the pre-moderator 108 in direct or near contact with Titanium target 106 is to produce a maximum thermal neutron flux. Adding a BSA of modest thickness (L.sub.2=4 cm in this example) directly on to the pre-moderator permits a maximum number of neutrons and higher thermal neutron flux to be collected and a smaller source size to be obtained. The close stacking of the pre-moderator and the BSA permits a maximum number of thermal neutrons to be obtained with a small source size D.sub.2.

(51) In embodiments of the invention, thermal neutron collection can be achieved with a conical funnel to both collect and channel neutrons into a small spot size with increased thermal flux at the exit of the cone of the funnel. The compact DD fusion source with a short thermal moderator (such as HDPE, or UHMW plastics with a high concentration of hydrogen atoms) quickly scatters the fast 2.5 MeV neutrons to thermal energies in a short distance (L.sub.1+L.sub.2+L.sub.3+L.sub.4+L.sub.5) from the fast neutron source (the titanium target 106). As shown in FIGS. 3A and 3B, the neutrons are then collected by a relatively short (e.g. L.sub.2=3 to 4 cm) funnel 222 in the slab of HDPE. Both moderation and scattering continue to occur along the cone length L. Simulations show that this results in an increase in flux 2 to 3 times and spot sizes of 1 to 3 cm in diameter depending on the geometry of the cone and size of the fast neutron emitter (Ti target). Additions of short spatial and energy filters improve the image by improving the brightness of the neutron source and limiting the effects of spurious radiation of fast neutrons, gamma emission.

(52) In embodiments of the invention, the shortness and compactness of the DD fusion generator and the moderation process to produce and collect thermal neutrons also allows for the use of other devices in the beam including short lengths of sapphire 220 and bismuth crystals 240, which can reduce the fast neutrons and gamma emission in the neutron beam, thus cleaning up the beam and achieving a relatively pure beam of thermal neutrons. The use of a compact fusion generator with relatively small spot sources of neutrons permits these neutron filters to also be compact and close together. This results in a useful source of neutrons that can be used in many laboratories and field locations, unlike the fixed, large and expensive reactor sources.