Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
CHAPTER 4 ACCELERATOR AND DETECTOR TECHNOLOGIES SUMMARY Radiation can come from radionuclide decay,1 atomic and nuclear reactions (e.g., alpha- n reactions), or from machine sources which are most commonly called accelerators, x-ray generators (including x-ray tubes), and neutron generators or neutron tubes. For the purposes of this report x-rays are electromagnetic radiation from machine sources, and gamma rays are electromagnetic radiation from radionuclide decay. Beta-minus particles from beta decay are electrons no different from the electrons in an e-beam from an accelerator. Similarly, neutrons are the same regardless of their source. Machines and radionuclide sources produce radiation with different energy distributions. Gamma decay emits radiation at discrete energies, although some radionuclides emit gamma rays at many different energies. X rays from accelerators and x-ray tubes are generated with an energy spectrum that extends from the beam energy downward and, with high-energy e-beams, the spectrum can go much higher than the energy of decay gamma rays. Accelerators that produce e-beams and x-rays come in many shapes and sizes; all are more complicated and expensive than x-ray tubes, which operate at lower energies than accelerators. At present, there is a âgray zoneâ in the energy range from 0.5 MeV to 1 MeV for which it is difficult to build x-ray tubes and for which accelerators are not typically constructed, although accelerators could be built to cover this range if there were a market for them. Neutron generators rely on accelerated particles to drive nuclear fusion reactions (typically deuterium-tritium, or D-T, reactions) that release neutrons. These neutrons are highly energetic (14.1 MeV for D-T neutrons) compared to neutrons from radionuclide sources (6 MeV average for americium-beryllium; 2.5 MeV average for californium-252). The intensity of the neutron flux (neutrons emitted per second per unit area) from current neutron generators is lower than is desired, but it has been improving. Radiation detectors are critical components for replacement of radiation sources for some applications of radiation sources, particularly radiography and well logging. This is because more efficient detectors could enable radiographers and well loggers to use lower activity radionuclide radiation sources or substitute machine sources that are unable to generate the flux achieved with high-activity radionuclide radiation sources. Detectors, too, are improving, but not primarily in response to desires to reduce the use of high-activity radiation sources. ACCELERATOR TECHNOLOGIES An atomic particle acceleratorâa category of high-energy equipment, which includes linear and circular machines such as betatrons, synchrotrons, cyclotrons, synchrocyclotrons, microtrons, and radiofrequency linear electron accelerators (rf linacs)âis one of the most important tools that modern science possesses. The very high velocity particles produced by these accelerators can be used to break apart atoms, allowing scientists to probe the fundamental principles of matter and energy, and it was primarily for this purpose that these machines were developed. The industrial and medical applications of the accelerators are 1 This report is concerned mostly with nuclear decay producing alpha, beta, and gamma radiation and neutron emissions from spontaneous fission. 67
68 RADIATION SOURCE USE AND REPLACEMENT classic examples of basic research translated to significant, unforeseen development of commercial markets.2 This chapter describes types of electron accelerators and x-ray and neutron generation techniques, as well as detector technologies that might replace or reduce radionuclide sources in certain applications (see Chapter 1). The accelerators that might be used in these applications are summarized in Table 4-1. The applications and specific replacement technologies are described in greater detail in Chapters 5 through 9. As illustrated in Table 4-1, there are a many choices of accelerator technologies and a range of possible configurations. The majority of the replacement technologies are based on electron accelerators, which can be configured to deliver either electron beams (e-beams) or x- rays. In many applications, x-rays can be advantageous because of their much greater penetration depths. However, because of the poor energy conversion efficiency in generating x- rays, the required beam powers for electron and x-ray beams differ by more than an order of magnitude for the same radiation dose rate. It should be noted that there are many other industrial applications that have been made possible because of the properties of electron accelerators or where the clear advantages of the accelerators have already made them, rather than radionuclide radiation sources, the technology of choice. These include applications in material processing such as cross-linking of polymers or curing of composites where the high energy density in an accelerator beam is required (see, e.g., Masefield, 2004). Another application is cargo inspection where high energy density and high x-ray energy are desired to penetrate dense objects. TABLE 4-1 Summary of Radionuclide Source Applications and Possible Accelerator Replacements Accelerator Energy Power Application Dose (Gy) Type Radiation (MeV) (kW) Radiotherapy Few rf linac E-beam â2â30 â1 or x-ray Self-contained â1â25 X-ray tube X ray â0.1â 0.4 â1 irradiators Panoramic irradiators â100â25,000 dc linac; E-beam â5â10 â10â1,000 rf linac; or x ray Rhodotron Oil well logging Electrostatic Neutrons Accelerator produces â0.001 D-D or D-T â 0.1 deuteron or triton for D-T, 2 deuteron for D-D; 2.45 or 14.1 neutron output Radiography <1 X-ray tube; X-ray â0.1â20 â0.001â1 Betatron; rf linac NOTES: Gy = the dose unit, gray; rf = radio frequency; dc = direct current; D-D = deuterium-deuterium fusion reaction; D-T = deuterium-tritium fusion reaction; MeV = mega-electron volt; kW = kilowatt. SOURCE: Table provided by the committee. 2 For a general description of the commercial uses of industrial accelerators, see Berejka (1995).
ACCELERATOR AND DETECTOR TECHNOLOGIES 69 Electron Accelerators and X-Ray Sources Electron accelerators generate high-energy beams that can be used to directly irradiate an object or to generate x-rays as illustrated in Figure 4-1. Electron accelerators have operated at energies as high as 100 GeV. However, for most applications considered here, the useful energy range is from a few hundred kilo-electron volts (keV) to a few tens of mega-electron volts (MeV). There are many commercial manufacturers of electron accelerators for a wide variety of applications including radiography, materials processing and diagnostics, and medical diagnostics and treatment. An electron accelerator consists of three parts: an electron source or gun, the accelerator, and a target or scanning/focusing system. Figure 4-1 is a diagram of a medical electron accelerator with an x-ray conversion target. Although Figure 4-1 provides a general layout of linear accelerator (linac) components, there are significant variations from one commercial machine to another, depending on final electron-beam kinetic energy and on the particular design used by the manufacturer. The length of the accelerating waveguide depends on the final electron kinetic energy, and ranges from about 30 cm at 4 MeV to about 150 cm at 25 MeV. In many applications such as materials processing, food irradiation, and some cancer therapies, the electron beam is sent directly into the object. The beam characteristics at the object are adjusted with bending or focusing fields which, depending on the beam energy, can be either electrostatic or magnetic. In other cases where greater penetration into the object is desired, the electron beam is directed onto a dense target to generate x-rays and the object is irradiated with the x-ray beam. The energy deposition as a function of depth in water is shown in Figure 4-2 for different energy electrons, x-rays, neutrons, and heavy charged particles. In 2002, it was estimated that there were over 17,000 accelerators in use in industrial or medical applications around the world (Maciszewski and Scharf, 2004). In this chapter, the committee does not discuss the majority of these applications, but concentrates only on those applications where the use of radionuclide sources is common. X-rays are generated through a process known as bremsstrahlung in which the electrons scatter inelastically off heavy atoms in a target; targets typically are made of tungsten or tantalum. This process is relatively inefficient because much of the electron beam energy is deposited in the target itself, although the process becomes more efficient at higher electron energies. For example, the optimized conversion efficiency of a 1-MeV electron beam into x- rays is only about 1â2 percent, whereas this increases to about 8 percent at 5 MeV and to about 12 percent at 7.5 MeV. The x-rays generated through bremsstrahlung have a continuous energy distribution extending down from the electron beam kinetic energy with an average energy that is much lower, typically 20â30 percent of the beam energy; an example of a bremsstrahlung spectrum produced by a 5-MeV electron beam striking a tungsten x-ray target is shown in Figure 4-3. The shape of the bremsstrahlung spectrum as well as the conversion efficiency depend on the kinetic energy of the electron beam striking the target and the thickness and atomic number of the target. The x-ray beam of Figure 4-3 would be referred to as a 5-MV x-ray beam where MV stands for "megavoltage" and the term 5 MV implies that we are dealing with an x-ray bremsstrahlung spectrum that contains photons with energies from 0 to 5 MeV and is produced in the x-ray target by monoenergetic electrons with a kinetic energy of 5 MeV.3 3 By convention, x-ray beams in the megavolt range are abbreviated as MV, but those in the kilovolt range are abbreviated as kVp
70 RADIATION SOURCE USE AND REPLACEMENT FIGURE 4-1 Schematic diagram of a typical medical electron accelerator with x-ray conversion target. SOURCE: Image provided by committee. FIGURE 4-2 Absorbed dose plotted as a function of depth in water for ionizing radiation beams of various types and energies. Parts (a) and (b) are for indirectly ionizing radiation: in (a) for photon beams in the range from 100 kVp to 22 MV and in (b) for neutron beams. Parts (c) and (d) are for directly ionizing radiation: in (c) for megavoltage electron beams in the range from 9 to 32 MeV and in (d) for heavy charged particle beams (187 MeV protons, 190 MeV deuterons, and 308 MeV carbon ions). SOURCE: With kind permission of Springer Science+Business Media.
ACCELERATOR AND DETECTOR TECHNOLOGIES 71 FIGURE 4-3 Bremsstrahlung x-ray (photon) energy spectrum from a 5-MeV electron beam. The ordinate (vertical axis), dn/dE, shows the number of x-rays within a certain energy range per incident 5-MeV electron. SOURCE: Image provided by the committee. The accelerator-based sources described here are grouped into three categories by their acceleration method: high-voltage dc accelerators, rf microwave accelerators, and induction accelerators; x-ray tubes are a special type of dc accelerator. Beam energies between 100 keV and about 30 MeV are potentially useful for most applications that are considered in this report. Typical medical accelerators operate with beam powers of roughly 1 kW; low-power accelerators for radiography may only have beam powers of a few watts; and high-power accelerators for irradiation operate with average beam powers of tens to hundreds of kilowatts. In high-power applications, such as those needed in large irradiators, the beam energy is limited by neutron production and activation of the accelerator and the target materials. The threshold for neutron production is between 8 and 13 MeV in most materials. In the case of food irradiation, the U.S. Food and Drug Administration limits the electron-beam energy to 10 MeV for electron irradiation and 7.5 MeV for x-ray irradiation (Meissner et al., 2000). Radiation hazards and radioactive waste from accelerators are discussed later in this chapter. The energy and beam power requirements for the accelerator system depend on the size and type of radionuclide source that it is replacing. To set the scale, the radiation dose from a radiotherapy source with 370 TBq (10,000 Ci) of cobalt-60 is roughly equivalent to the x-ray dose from a 10-MeV linac with 100 watts of beam power. A portable radiography source with 11 TBq (300 Ci) of iridium-192 could be replaced using the x rays generated by a 1-MeV electron beam with a beam power of a few watts. On a larger scale, the radiation dose in a panoramic irradiator using 110,000 TBq (3 MCi) of cobalt-60 would be roughly equivalent to that from a 20- kW, 10-MeV electron beam or an x-ray beam generated from a 300-kW, 7.5-MeV electron beam. High-Voltage dc Electron Accelerators Direct current (dc) accelerators use a dc voltage to accelerate an essentially continuous wave (cw) beam. The beam is accelerated with an electric field generated with a series of electrodes. To prevent breakdown (direct electrical discharge that short-circuits the voltage gap), the fields are typically limited to much less than 1 MV/m and the energy of the electron beam is usually limited to a few MeV although dc devices have been operated at energies as high as a few tens of MeV. Because of the breakdown limitations, these devices tend to be larger than rf accelerators. Examples of dc accelerators include x-ray tubes as well as devices for materials analysis and processing, semiconductor processing and development, and
72 RADIATION SOURCE USE AND REPLACEMENT pharmaceutical research. These devices can have an output power as high as a few hundred kilowatts, and high-efficiency solid-state switching techniques may further improve the power output to a few megawatts. High-voltage dc generators are based on either mechanically transporting charges, as is done in a Van de Graaff or Pelletron accelerator (referred to as electrostatic accelerators), or cascaded rectifier circuits (referred to as electrodynamic accelerators; Norton and Klody, 1997). Both types of dc accelerators are produced by many manufacturers including IBA, Nissin High Voltage, Pelletron, and Vivirad. These accelerators operate at voltages between 100 keV and 10 MeV and are frequently used for electron-beam irradiation in materials processing, for example, to improve the physical properties of plastics, cables, and wires, or for materials analysis; similar sources are also used at a much lower energy for ion implantation in semiconductor development. In general, the electrodynamic accelerators produce greater beam powers and are used for materials processing applications. Low-energy dc accelerators (approximately 100 to 500 keV) have been constructed with a few megawatts of beam power. These high-power4 accelerators are primarily used for materials processing applications. Examples of higher voltage accelerators include the Dynamitron from IBA (Cleland, 1959; see also Figure 4-4) and the ICT from Vivirad, both of which operate at up to 5 MeV with about 200â300 kW of beam power. The dc accelerators are relatively straightforward to control, and the beam voltage and current can be varied over large ranges. For example, a Dynamitron is typically designed to be operated over a range from 1 to 3 MeV at constant beam current, and a much wider range (e.g., 300 keV to 3 MeV) if the beam current is lowered. Such large variation can be more difficult to attain in rf accelerators. The improvement and high power capability of electronics and semiconductor switches, in particular the insulated gate bipolar transistor (IGBT) and the integrated gate commutated thyristor (IGCT), will likely lead to improved performance in the dc accelerators. At present, the typical operating efficiencies are roughly 70 to 80 percent. New multimegawatt power supplies (see, e.g., Bradley et al., 1999; Cassel et al., 1997; and Diversified Technologies, Inc., 2007) with voltages of about 100 kV are being developed for other applications that have efficiencies in the high 90 percents; it is expected that these concepts will also lead to improvements in the higher voltage dc accelerators as well. X-ray Tubes X-ray tubes are low-energy dc accelerators in which the electron gun (cathode), accelerator, and target are contained within a single vacuum enclosure. Tube voltages can range from a few kilovolts (kV) up to about 500 kV and are used in many applications ranging from medical devices to irradiators to industrial materials processing; x-ray tubes are manufactured by a wide range of producers in the United States and abroad. The simplest x-ray tubes are evacuated closed glass tubes with a tungsten filament (cathode) that emits electrons and a tungsten target that generates x-rays. The lifetime of a tube can be limited by the tungsten filament, tungsten plating that forms on the glass and leads to arcing, or the deterioration of the vacuum within the tube. Typical lifetimes of closed glass tubes are about 1000 to 5000 hours of operation. Closed metal and ceramic tubes can achieve significantly longer lifetimes; at least one company manufactures a closed x-ray tube with a 4 Note that there is a difference between energy and power. Energy refers to the energy of the individual particles. Power is the amount of energy that flows through the beam in a given amount of time. An accelerator that produces an intense beam of low-energy particles can be a high-power machine.
ACCELERATOR AND DETECTOR TECHNOLOGIES 73 dispenser cathode5 rather than a tungsten filament and the tube has a rated lifetime of 20,000 hours of operation. Open or dismountable tubes have a separate vacuum pump which allows the tube to be opened and the cathode and anode to be replaced as necessary. With the appropriate preventive maintenance, dismountable tubes can be operated for years without failure. In high-power applications, the anode must be constructed of high-temperature materials because, as noted earlier, the x-ray generation process is inefficient and most of the electron beam energy is deposited into the target. Typical fixed anode tubes are limited to powers of a few kilowatts. A further refinement is a rotating anode tube, where the target is rotated to spread the heat deposition around a larger area. Such tubes can operate at power levels many times higher than the fixed anode tubes. Much of the development of robust x-ray tubes is being driven by the medical imaging and nondestructive testing (NDT) requirements. In both of these applications, high power and long tube lifetime are important. Additional refinements such as cold (field emission) cathodes have increased the energy and power available in compact and miniature x-ray tubes (see Reyes-Mena et al., 2005; Xintek Inc., 2004â2006), which make them an attractive alternative in situations where small radiation sources are required. Low-energy (40 to 50 kVp) miniature x- ray tubes with hot cathodes have also been developed as brachytherapy sources. FIGURE 4-4 A Dynamitron showing the accelerator column. SOURCE: Image courtesy of ION Beam Laboratory, State University of New York at Albany. 5 A dispenser cathode is a tungsten matrix doped or impregnated with a material, such as barium or ruthenium oxide, that lowers the work function or energy required to liberate an electron from the surface of the cathode. Such cathodes can generate the same current at lower temperatures than standard tungsten filaments, which extends the operating lifetime.
74 RADIATION SOURCE USE AND REPLACEMENT RF Electron Accelerators Radiofrequency accelerators accelerate beams with rf cavities that typically operate at low microwave frequencies (on the order of 1 GHz), although there are accelerators based on lasers (about 100 THz) and relatively low-frequency rf accelerators, which operate in the 100- MHz regime. Usually the rf cavities are limited to accelerating fields of a few tens of megavolts per meter (MV/m). Laser-based accelerators have achieved fields as high as 100 GeV/m, but these are far from being commercial devices. Unlike the dc accelerators, rf accelerators do not have to prevent discharge across the full acceleration voltage along their length, which allows rf accelerators to operate over a large energy range and be relatively compact. There are many different variants of rf accelerators; examples include linear accelerators for medical applications (on the order of 1â10 MeV), synchrotron storage rings for synchrotron radiation generation (on the order of 1 GeV), and high-energy synchrotrons or linacs for high-energy physics (on the order of 1 TeV). Radiofrequency accelerators may be single-pass linacs or multiple-pass accelerators such as microtrons or the Rhodotron, or circular accelerators such as cyclotrons or synchrotrons. These are described below, except for circular rf accelerators, which are used in proton and carbon ion radiotherapy but are not typically used in the energy range of interest as potential radionuclide radiation source replacements. Radiofrequency Linacs Radiofrequency linacs were developed in the 1940s and are used for many applications ranging from the generation of x-rays in a hospital environment to injectors into higher energy synchrotrons at particle physics laboratories. The first rf linacs were built to operate around 3 GHz (similar to the frequency used in household microwave ovens and in some cordless phones), and today most commercial linacs operate at frequencies between 1 and 15 GHz with rf wavelengths that are between 30 and 2 cm, respectively.6 A large number of companies manufacture rf linacs for a variety of applications. The most common use of rf linacs is for medical radiotherapy. A number of companies manufacture such accelerators, including Elekta, Mitsubishi, Siemens, and Varian. The two other major applications that are considered in this report are non-destructive testing (radiography) and irradiation/sterilization. A number of companies, both large and small, produce linacs for these applications; a few examples include AS&E, Hitachi, L3 Communications, Linac Technologies, L&W Research, Tsinghua Tongfang Nuctech, and Varian. A few commercial rf linacs are illustrated in Figure 4-5. An rf linac is constructed from four main elements: (1) a high-voltage power supply (modulator), (2) an rf power source, (3) a microwave cavity, and (4) a charged particle source as illustrated in Figure 4-1. The microwave cavity is the "heart" of the accelerator. It is constructed from a series of cavities with an aperture along the axis for the beam. The size of the cavities is selected on the basis of wavelength (and therefore the rf frequency) of the linac, and is independent of the overall size of the accelerator. For example, the individual 3-GHz cavities in the Stanford Linear Accelerator Center (SLAC) linac are roughly 10 cm (4 in.) in diameter and about 3.8 cm (1.5-in.) thick with a 2.5 cm (1 in.) diameter hole passing through the center through which the electron beam and rf power pass. In SLAC, roughly 90 individual cavities are bonded together for each of its many 3-meter-long accelerator structures, whereas a lower energy linac might have just one accelerator structure with a small number of cavities. Because 6 The ILU-series of linacs operate with rf frequencies around 100 MHz (Auslender, 2005).
ACCELERATOR AND DETECTOR TECHNOLOGIES 75 the cavity dimensions scale with the rf wavelength, a similar structure operating at 12 GHz would be roughly one-quarter the size. Typical acceleration gradients scale with frequency due to electrical breakdown and power limits. Normal conducting 1-GHz linacs typically have maximum gradients that are about 10 MV/m, while at 12 GHz, linacs have operated with gradients greater than 70 MV/m (Adolphsen et al., 2005). However, the rf power required to achieve these high gradients is quite high. High gradients are desirable because they make the linac shorter for a given particle energy and they give less travel time for the beam to spread as a result of the particles electrically repelling each other. In cases where high radiation doses are desired, such as for irradiation facilities, the accelerator efficiency is very important. The overall electrical efficiency from power source to emitted beam power of early rf linacs was only about 20 to 30 percent, compared with efficiencies of 60 to 80 percent for the dc accelerators. The efficiency can be improved by using more efficient rf power sources and maximizing the efficiency of the accelerator cavity by choosing the beam current so that the beam-induced voltage is comparable to the unloaded voltageâthe optimization depends on the detailed cavity design. For normal conducting cavities, this implies beam currents that are comparable to 1 ampere and the resulting rf-to- beam transfer efficiencies can be as high as about 70â80 percent; some examples are listed in Haimson (1975) and Miller et al. (2003). Superconducting cavities can be optimized so that essentially all of the rf energy is transferred to the beam; however, the technology is still relatively novel and there are no commercial superconducting linacs presently operating. The simplest and most common particle source for an rf linac is a thermionic gun where a relatively low dc voltage of about 20 kV is used to accelerate electrons from a heated cathode. This type of gun has the disadvantage that the beam is not appropriately bunched for the rf linac, which results in a broad energy spectrum, particle losses, and a relatively large beam diameter. Variations on the thermionic gun include higher voltage operation (up to 500 kV), the use of high-frequency choppers and subharmonic bunchers to improve the bunching and capture of the electron beam, and voltage grids to clearly define the length of the pulse (from 1 nanosecond to a few microseconds). In most cases, these additional complications are not useful for the applications considered here. (b) (a) FIGURE 4-5 Examples of commercial rf linacs. SOURCE: (a) Courtesy of Varian Medical Systems, Palo Alto, CA; (b) courtesy of L3 Communications.
76 RADIATION SOURCE USE AND REPLACEMENT Another type of injector is the rf gun where high-voltage microwave fields accelerate the beam right from the cathode. For low-energy applications of a few MeV, the gun can be integrated onto the accelerator structure to directly deliver the desired beam. Because the beam is rapidly accelerated, many of the effects that limit the dc gun performance due to the self- repulsion of the electrons are mitigated. Radiofrequency guns have been developed using lasers and photocathodes to generate short-pulse electron beams (a few picoseconds) which are matched to the rf wavelength of the accelerator (Sheffield et al., 1988). Photocathode rf guns tend to have extremely high quality beams but require a costly laser system. Radiofrequency laser cathodes are not commercially available. Thermionic rf guns that eliminate the laser system also have been developed (Beczek et al., 2001; Tanabe et al., 1989; Westenskow and Madey, 1984). Thermionic rf guns are becoming more common as the technology becomes more mature and will likely find use for the applications considered in this report. The rf power source that generates the microwaves for the linac is usually either a magnetron oscillator or a klystron amplifier. In both cases, the size of the rf source and the power output capability are roughly proportional to the rf wavelength. Magnetrons were developed during the 1940s and are used in everything from microwave ovens to sophisticated radar systems. Magnetrons are relatively simple and compact but have limited output power and limited control over the rf frequency and phase. Continuous-wave devices can have an output power as high as about 100 kW at 1 GHz with efficiencies of about 75â85 percent while pulsed devices can operate at about 60â75 percent efficiencies. To be an effective power source for an rf linac, the magnetrons usually need a feedback system to stabilize the rf output. Klystrons are high-power amplifiers that tend to have higher power capabilities than the magnetrons but are also larger, heavier, and have lower efficiencies. Continuous-wave klystrons can have output powers of around 1 MW at 1 GHz with efficiencies of 60â70 percent while pulsed devices with efficiencies of about 50 percent have peak powers that range from about 100 megawatts with microsecond pulses to about 10 megawatts with microsecond pulses. New high-efficiency klystrons that are being developed in laboratories and industry use configurations based on planar beams or multiple round beams to reduce the repulsive forces within the beam; efficiencies of about 65 percent are expected for the pulsed klystrons (sheet-beam klystrons; multibeam klystrons; for examples, see Lenci et al., 2004). The rated lifetimes of typical high- power magnetrons or klystrons is usually 5,000 to 20,000 hours. Over the past two decades, the lifetime of typical magnetrons has been extended by more than a factor of three. It is expected that modern design will further improve both the efficiency and reliability of these rf power sources (Vlieks et al., 1998). Finally, the modulator or high-voltage power supply converts the incoming alternating current (ac) voltage into the high-voltage dc power needed for the rf power sources. As noted earlier, the improved high power capability of semiconductor switches, in particular the IGBT and the IGCT, has led to improved power handling, efficiency, and reliability in both modulators and high-voltage power supplies. These systems operate with minimal losses and have expected availability of more than 50,000 operating hours, which is many times greater than that achieved with conventional technology. Like x-ray tubes, the reliability of rf linacs has been driven by the medical and industrial applications. Modern medical linacs have useful lifetimes of 10 to 15 years, after which the computer controls and therapy planning software generally are considered to be obsolete and unable to be upgraded, although the linac may be fully operational. The waveguides and rf cavities tend to be robust against failure and, as described above, the lifetimes, as well as the mean times to repair, of the consumable components, the electron gun, rf power source, and voltage supply have been greatly improved over the past few decades.
ACCELERATOR AND DETECTOR TECHNOLOGIES 77 Multipass rf Accelerators There are two types of multipass rf accelerators that are most useful in the energy range of interest for this study: microtrons (Veksler, 1944) and the Rhodotron (Jongen et al., 1993; Pottier, 1989). A microtron uses a standard accelerator cavity to accelerate the beam. The beam is recirculated with a set of bending magnets in either a circular or racetrack configuration (see Figure 4-6). Typical designs use tens of passes through the same accelerator cavity to accelerate the beam and, depending on the energy gain per turn, the output beam energies can range from a few MeV to 1 GeV. Because the accelerator cavities are used for multiple passes, microtrons have the potential for being less expensive than linacs, although the bending magnets that recirculate the beam can be expensive. In addition, because the microtron can operate in a cw mode, the output powers can be more than 100 kW. There are a number of commercial manufactures of microtrons, including Scanditronix Medical and Sumitomo Heavy Industries. The Rhodotron is also a recirculating accelerator where a very low frequency field is used to accelerate the beam radially through a cylindrical cavity with 5 to 10 passes. The low- frequency operation (about 100 MHz) has low losses and the rf power can be generated with an efficient high-power tetrode (a four-element electron tube). Commercial Rhodotrons are produced by IBA Industrial and operate at energies up to 10 MeV with beam powers as high as 700 kW (IBA Industrial); one such device is shown in Figure 4-7. The Rhodotron operates with roughly 50 percent efficiency at peak output power. Electron path Bending Bending magnet magnet Accelerator FIGURE 4-6 Diagram of a microtron. SOURCE: Image provided by the committee.
78 RADIATION SOURCE USE AND REPLACEMENT FIGURE 4-7 This 200-kW Rhodotron is used to irradiate U.S. mail at a facility in New Jersey. SOURCE: Image courtesy of IBA-RDI. Induction Electron Accelerators Induction accelerators operate using magnetic induction where the accelerating voltage is generated by the changing magnetic field. The two types of induction accelerators are betatrons (Kerst, 1940) and induction linacs (Christofilos et al., 1964). The beam current in a betatron tends to be limited by space charge effects (self-repulsion of the electrons) and they operate as pulsed devices; these are relatively low power accelerators. Betatrons were used for radiotherapy (see, e.g., Kapetanakos et al., 1993); however, these have largely been replaced with rf linacs, which provide greater beam power in a more flexible package. Presently, betatrons are primarily manufactured as portable devices for nondestructive testing (see, e.g., Inspecta, 2007; JME Ltd, 2007; Kaplin et al., 2002) and have energies of a few MeV with a radiation output of a few roentgen per minute. Figure 4-8 is an illustration of such a device. In contrast, induction linacs can accelerate high-current beams. They are inherently low- impedance devices and thus are used to produce low-energy (a few MeV), high-current (kA) pulses at a relatively low repetition rate. They are primarily used for high-intensity flash radiography, inertially confined fusion drivers, and directed energy weapons. Examples include the AIRIX accelerator at Centre dâEtudes Scientifiques et Techniques dâAquitane (Eyharts et al.,1995) and the DARHT accelerator at Los Alamos National Laboratory (Burns et al., 1996). Because the high peak power is not necessary when considering radionuclide source replacement and induction linacs are relatively expensive for a given average power, they are not considered as potential radiation source replacements in this report, although additional induction linac research and development is mentioned below.
ACCELERATOR AND DETECTOR TECHNOLOGIES 79 FIGURE 4-8 The JME Betatron Data Pack for industrial radiography. SOURCE: Image Courtesy of JME, LTD. New Developments in Electron Accelerators A number of cutting-edge research demonstrations that could improve accelerator capabilities have been carried out or are underway. These are described below. Although none of these is close to commercial deployment, they illustrate the potential for higher efficiency, higher power, more compact, and specialized radiation generators. High-gradient superconducting rf accelerators have been developed over the past few decades and can provide a very efficient method of accelerating an electron beam. A number of accelerators are using superconducting rf cavities, including the recirculating linac used for nuclear physics experiments at the Jefferson Laboratory in Virginia and the FLASH UV laser facility at the DESY laboratory in Hamburg, Germany. These are research and development programs that indicate possible future directions. Recently, pulsed rf cavities, operating at about 1 GHz and 1 m long, have demonstrated gradients in excess of 35 MV/m, and gradients in excess of 15 MV/m have been maintained in continuous wave operation. Such cavities are being industrially produced by companies such as Advanced Energy Systems in the United States and ACCEL in Europe, which was recently purchased by Varian Medical Systems. However, turnkey superconducting linacs are not yet available. High-power electron guns using superconducting rf cavities are being developed for various purposes. These guns are designed to produce very high quality continuous wave beams with beam energies of a few MeV and beam powers of 100 to 1,000 kW. For example, one gun presently being fabricated by a collaboration between Jefferson Laboratory and Advanced Energy Systems is designed to produce a 7-MeV beam with a current of 100 milliamperes; this device is presently being commissioned at lower power levels. Similar devices are being designed and constructed at many laboratories around the world as injectors for energy recovery linacs. Such accelerators may provide efficient high-power beams for irradiation and sterilization facilities in the future.
80 RADIATION SOURCE USE AND REPLACEMENT The primary disadvantage of the superconducting rf technology is that additional cryogenic cooling is required, which increases the capital cost of a new facility. For best performance, the superconducting rf cavities typically are operated at about 2 K (â271Â°C). For a large installation, the incremental cost and additional complexity in the cryosystem is relatively small, but it may not be cost-effective for a smaller stand-alone facility. Additional analysis is needed to understand these cost trade-offs. Other accelerator developments include: 1. Work on more cost-efficient high-gradient induction linacs. The advanced solid-state switches are leading to improved efficiency and lower cost per module. There is active research and development on the use of dielectric-loaded structures which could permit about 10 times higher gradients, approaching 20 MeV/m (Sampayan et al., 2005). This would make the high-power characteristics of the induction technology much more attractive. 2. Development of compact portable mega-electron volt accelerators that could replace hand-portable iridium-192 sources for radiography (Breidenback, M., Stanford Linear Accelerator Center, private communication with Raubenheimer, T., 2007; Yamamoto et al., 2006). 3. Demonstration of high-gradient laser accelerators that have the potential to make a very compact source of medium-energy (about 20â100 MeV) electrons (Faure et al., 2006). 4. Development of narrow-band x-ray sources based on Compton backscattering or crystal diffraction. Some of these sources are available commercially (Vlieks et al., 2006; Dobashi et al., 2005; Mondelaers et al., 2000; Jolie et al., 1998). ACCELERATOR-DRIVEN NEUTRON SOURCES Particle accelerators can also be used to generate neutrons. Setting aside spallation sources, which require much higher energies (GeV) and therefore large facilities, accelerator- driven neutron sources direct a beam of deuterium nuclei at a target loaded with deuterium or tritium,7 causing fusion reactions. The deuterium-deuterium reaction (D-D reaction), the deuterium-tritium (D-T reaction), and a third reaction that is not typically used, the tritium-tritium (T-T reaction), are illustrated below in Equation 4-1. The reactions produce helium nuclei and neutrons: 3 D + Dâ He + n, D + Tâ4He + n, T + Tâ4He + 2n (4-1). The D-D reaction generates monoenergetic neutrons at 2.45 MeV, but has a competing reaction that generates no neutrons. The neutrons produced by the D-T reaction are emitted monoenergetically at 14.1 MeV. Laboratory-scale accelerator-driven fusion sources based on the D-T reaction, such as the Rotating Target Neutron Source, have been in operation for several decades (see, e.g., Booth, 1967) and generate moderate neutron flux (e.g., 5 x 1011 neutrons per square centimeter per second). Kaman Nuclear, formerly in Colorado Springs, produced neutron generators from at least 1963 with advertised flux ratings ranging from 107 to 1011 neutrons per square centimeter per second. The T-T reaction generates neutrons with energies between about 1 MeV and 10 MeV, and averaging about 5 MeV. Figure 4-9 shows the cross section, which scales directly with the reaction rate, for several fusion reactions. The D-T cross section peaks below 100 keV in center of mass coordinates, which translates into around 7 Deuterium and tritium are isotopes of hydrogen: hydrogen-2 and hydrogen-3, respectively.
ACCELERATOR AND DETECTOR TECHNOLOGIES 81 110 keV for a deuteron from an accelerator striking a stationary tritiated target. The peak D-D cross section is more than an order of magnitude lower and peaks closer to 1 MeV in center of mass coordinates. The first commercial pulsed neutron well logging tool was introduced in 1963 by Dresser Atlas, but its capabilities were very limited because the neutron flux was quite low. Within the past 10 to 15 years, higher flux compact neutron generators have become technically feasible. Although several companies manufacture compact neutron generators for use in well logging and other applications, the units remain quite costly. They are manufactured by the All-Russia Research Institute of Automatics (VNIIA); Baker Hughes, Inc.; China Petroleum Technology and Development Corporation; Eads Sodem; Halliburton Company; Schlumberger, Ltd.; and Thermo Electron Corporation. Figure 4-10 shows a schematic of the unit designed by Sandia National Laboratories and manufactured by Thermo Electron Corporation. Typical neutron tubes generate 106 neutrons/pulse and pulse 100 times per second to yield 108 neutrons/s. Higher neutron output rates (1011 neutrons/s) can be achieved in some larger units. Several neutron generators with neutron outputs ranging from 107 to 1014 neutrons/s have been developed (Lawrence Berkeley National Laboratory). One of these approaches uses a low-temperature tritium plasma formed inside a tubular deuterium-loaded target and the ions are accelerated toward the walls where the reactions take place. Another promising idea that has been suggested is essentially the inverse of this approach, using plasma immersion ion implantation techniques (I. Brown, Lawrence Berkeley National Laboratory, personal communication with M. Lowenthal, May 8, 2007). In this idea, a low-temperature tritium plasma would be formed around a deuterium-loaded target. The target would then be pulsed to a voltage of â100 kV, and the ions would strike the target. Like the preceding approach, this technique is not limited by the intensity achievable in an ion beam and could lead to higher intensity neutron sources with microsecond pulse widths. To the committeeâs knowledge, this approach has not been examined in detail. Reaction rates, and therefore the neutron output rates, could in principle also be raised by developing targets that hold higher concentrations of hydrogen isotopes. RADIATION HAZARD AND RADIOACTIVE WASTE FROM ACCELERATORS It must be noted that switching from radionuclide radiation sources to machine sources of radiation does not obviate the need for radiation safety and radioactive waste management. Particle accelerators emit radiation primarily along their beamlines, but also to much lesser extent they emit x-rays in other directions. Where and how much radiation is emitted depends on the design of the accelerator. Operators of accelerators and equipment that uses x-ray tubes are required to undergo radiation safety training. The ability to turn off a radiation generator is an obvious advantage for worker safety. Particle accelerators also can, and do over time, induce low but measurable concentrations of radioactive material in their bremsstrahlung targets and other objects subjected to extensive irradiation by electrons or photons in the several-MeV range and higher through photonuclear reactions (see NCRP, 2005, 1984, 1977). Accelerators that operate at energies below 10 MeV pose fairly insignificant radiation hazards when they are not operating. Neutron generators induce radioactivity in the material surrounding the fusion reaction (the target, the device, the housing, and any shielding) through neutron capture and other nuclear reactions. The fusion target or source plenum contains tritium, which is itself radioactive, and the neutron generator becomes contaminated and must be treated as radioactive waste. The radioactive material imposes only a minor waste management burden, as the concentrations of radionuclides are typically very small compared to Category 1 and 2 radionuclide radiation sources.
82 RADIATION SOURCE USE AND REPLACEMENT 1.00E+01 1.00E+00 1.00E-01 Cross Section (barns) 1.00E-02 1.00E-03 D (d,p) D 1.00E-04 D (d,p) D 1.00E-05 T (d,n) He-4 T (t,2n) He-4 1.00E-06 1.00E-07 1.00E-08 1 10 100 1000 10000 Energy of Incident Particle (keV) FIGURE 4-9 Fusion reaction cross sections as functions of kinetic energy. SOURCE: Image provided by the committee. Rear Ion Source Anode Cathode Accelerator Electrode Exit Cathode Ion Source Target Magnet Ion Source Anode Vtarget Gas Reservoir Element Vacuum Envelope Vsource Vaccelerator FIGURE 4-10 Schematic of a neutron tube. SOURCE: Image courtesy of Burkhart, B. (2006). DETECTOR TECHNOLOGIES Each radiation source application discussed in this report uses the radiation for one of two purposes: to deposit energy within the irradiated material or to gather information about the irradiated material. Radiation sterilization and radiotherapy use the deposited energy to kill bacteria, viruses, or cancer cells; blood irradiators use it to kill white blood cells. This can be accomplished with a radionuclide radiation source or radiation generator working alone. In
ACCELERATOR AND DETECTOR TECHNOLOGIES 83 contrast, radiography and nuclear well logging aim to learn more about the structure and composition of the irradiated material. These latter applications use a radiation source or radiation generator in conjunction with a detector. In gamma radiography the detector is commonly a sheet of film. In x-ray radiography, charge coupled device (CCD, like those used in digital cameras) detectors are increasingly replacing sheets of film. In well logging, the gamma rays triggered by radiation interactions in the rock layers adjacent to the well hole are usually detected with thallium-doped sodium iodide crystal scintillators. How useful an apparatus is depends on the performance of the overall system, not just the source, and so, improving the detector can relax the demands on the source. That is, a more sensitive detector may enable users to accomplish the same tasks with a lower intensity radiation source, including a small radiation generator. The underlying physical principles of all radiation detectors are the same. Radiation passing through the detector interacts with some material within the detector. Neutral particles, such as gamma rays, x-rays, and neutrons have relatively large ranges in materials, depositing their energy over tens of centimeters of solid material. Charged particles, such as alpha particles and energetic electrons, deposit their energy in microns or millimeters of solid material. In some detection schemes, the interaction is converted into an electrical signal, which can trigger an alarm, be processed for display, or simply be recorded. In some radiation detectors, the electron or the ion directly creates the electrical signal. For example, in a gas ionization chamber or a solid-state detector, a voltage difference is placed across the detector media so that when electrons are liberated and ions (or holes, in the case of a solid-state photoconductor or photodiode) are formed, an electrical current flows. Other radiation detectors use phosphors or scintillators to convert the incoming radiation to visible light or some other relatively long wavelength radiation. The detectors can then use a photomultiplier tube or a CCD to convert light into electrical signals. Detectors have improved over the past several decades, primarily as semiconductors have been applied and improved and as more efficient phosphors and scintillators have been discovered. These improvements enable detectors today to map the radiation spatially, forming an image, and to discriminate energies. Although it has been documented that x-rays can cause faint visual effects in the human eye, called radiation phosphenes, x-rays and gamma rays are, for practical purposes, invisible. Historically, the first radiation detection devices were photographic plates, modern versions of which, particularly large-format, are still used today in radiography. The sensitivity of the early photographs was improved substantially by the advent of âintensifying screensâ placed in front of the emulsion. These screens fluoresced under x-ray irradiation and the light from the screen supplemented the x-ray interaction with the emulsion (Frame, 2004). Faster films replaced the early plates and screens, but the general approach of converting the x-ray âsignalâ into an intermediate, more readily sensed or recorded form underlies many detection techniques today. The majority of gamma-ray detectors are scintillator devices; they convert a gamma ray into visible light which in turn is then converted into an electrical signal with a semiconductor detector, as described above. In solid state, the gamma ray is directly absorbed in a p-n junction semiconductor to create an electrical output. (This is analogous to the operation of a solar cell.) The main thrust for the development of gamma-ray detectors until very recently has been for high-energy particle physics experiments where the energies are much higher than those used in radiography and a very wide range in energies must be detected simultaneously. Recently, though, there has been renewed interest in developing detectors for 511-keV photons used in positron emission tomography and 662-keV gamma rays. At the present stage of detector development, the scintillator approach is generally preferred because it enables both the scintillator detector and the light detector to be selected for optimum performance. The development of detectors for medical computed tomography (CT) x-ray scanners provides an example of how a combination of technological advances has enabled more imaging information to be obtained with the same or reduced source intensity. For CT, a series
84 RADIATION SOURCE USE AND REPLACEMENT of x-ray exposures have to be recorded rapidly to obtain sufficient information for the computer reconstruction of a three-dimensional image while minimizing the x-ray exposure of the patient. To minimize exposure, this has necessitated the development of more efficient scintillator materials that can convert each x-ray photon to light while minimizing afterglow in order to capture images as quickly as possible. Different manufacturers have developed different scintillator materials specifically for this purpose, such as Gd2O2S:Pr,Ce,F and (Y,Gd)2O3:Eu,Pr. The common feature of these materials is that they have been designed and doping has been manipulated to optimize the efficient conversion of high-energy x-rays to light that can be most efficiently detected by CCD. Concurrently, there have been dramatic improvements in CCD technology, computer software and compact x-ray sources dedicated to the development of the state-of-the-art CT scans commercially available today. Recognizing that improved detectors for well logging applications would be beneficial in decreasing the activity of the current radiation sources, several alternative materials, such as bismuth germanate (BGO), have been investigated, but none has yet demonstrated the reliability, reproducibility, and long-term stability that the current alkali-halide (NaI;Tl) exhibits. However, the development of these alternative materials has not been directed specifically for the detection of gamma-ray energies for well logging applications but rather for the detection of gamma rays over a broad energy range. The detection of neutrons, such as used in well logging, poses particular challenges because their interaction with detectors is so weak. Neutron detectors typically utilize nuclear reactions in helium-3 (n + 3He â 3H + 1H + 0.764 MeV), boron-10 (n + 10B â 7Li + 4He + 2.31 MeV), or lithium-6 (n + 6Li â 3H + 4He + 4.78 MeV). The probability of these reactions occurring with an incoming neutron is high at low neutron energies (fractions of an electron volt), but it diminishes significantly with increasing neutron energy to be rather smaller (by a factor of 1,000 or more) for neutrons in the MeV range. To take advantage of the higher probability of interaction at low energy some detectors lower the energy of the neutrons by making them pass through a material with low mass nuclei that slow the neutrons down through collisions, a process called moderation. For neutrons in the MeV range, tens of centimeters of a moderator material such as polyethylene surrounding the detector are required. Some other detectors simply rely on the fast neutrons to induce the reactions, and still others rely on the protons (simple hydrogen nuclei) that recoil from collisions between energetic neutrons and hydrogen atoms in the detector. The relatively low efficiency of all of these detectors at high neutron energies make fast neutrons less attractive for signal transmission (see Knoll, 2000). FINDING Finding: A variety of accelerator systems is available which can generate electron beams and x-ray radiation at different energies and with different output powers. Accelerators are widely used in industry and are theoretically able to replace radionuclide sources in almost any application, although practical and economic factors limit their value as replacements for some applications.