Radiation Source Use and Replacement: Abbreviated Version (2008)

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 3 RADIATION SOURCE RISKS SUMMARY This chapter presents a discussion of hazards and, to some extent, risks associated with uses and misuses of the radiation sources described in Chapter 2. The uses are described in detail in Chapters 5 through 9 with examinations of possible replacements. The committee used a qualitative risk framework in thinking about risks. The use of one or more radiation sources as radiological dispersal devices (RDDs or “dirty bombs”) to cause lasting contamination that prevents regular human access to an area (area denial), and the economic and social consequences that result may be the most important risks of malevolent use of radiation sources. The International Atomic Energy Agency (IAEA) categorization of radioactive sources and the corresponding U.S. Nuclear Regulatory Commission (U.S. NRC) categorization are based almost exclusively on deterministic health effects from exposure to an unshielded source, not on the area-denial RDD potential of the radionuclides. Although these categorization systems provide a basis for regulatory regimes related to safety and some aspects of security, they do not account for other important security-related issues. In this chapter, the committee recommends that the U.S. NRC reexamine its security-based orders and decisions for materials licensees, considering the potential consequences of area-denial RDDs. For the purpose of prioritizing radiation source replacements and considering options for implementing those replacements, the committee concludes that radioactive cesium chloride is the greatest concern among the materials used in radiation sources in the United States because it is used in significant quantities in urban areas in a powdered, dispersible form. Review of previous accidents involving cesium-137 dispersals shows significant consequences. After an incident with radioactive cesium chloride in a wet panoramic irradiator, the U.S. NRC imposed a “qualified ban” on radioactive cesium chloride sources in such applications. In this chapter, the committee finds that radioactive cesium chloride sources should be replaced and that government action is needed to implement such replacements. The committee suggests several steps to implement the options for replacement. RISKS AND PRIORITIES The committee’s charge (Sidebar 1-1) explicitly directs the committee to focus on risk: “The report will contain a review of radiation source use, potential replacements for sources that pose a high risk to public health or safety, and findings and recommendations on options for implementing the identified replacements.” The U.S. NRC helped the committee to interpret the charge by stating at the committee’s first meeting that “options for implementing” include both technical and policy options. The U.S. NRC wrote into its request for this study the description of Category 1 and 2 sources as radiation sources that pose a high risk to public health or safety.1 1 Sources that fall into Category 3 and lower can be assembled into Category 2 or 1 quantities of radioactive material. Further, it may be the case that some radiation sources near the upper threshold for Category 3 pose more serious risks than other sources that fall near the lower threshold of Category 2 in 43 

44 RADIATION SOURCE USE AND REPLACEMENT The committee chose to base its findings and recommendations on the hazards of the radionuclide radiation sources and the risks they pose, and to use these factors in prioritizing and balancing among trade-offs involved in radiation source replacement. The term risk is used imprecisely and often inconsistently in common parlance and sometimes even in technical discussions. The committee uses the risk terminology as defined in a 1983 National Research Council (1983) report and other basic references (e.g., Kaplan and Garrick, 1981). In this framework, a hazard is a potential source of a negative consequence or harm, and risk is defined as the likelihood that such a harm will occur. The risk associated with a particular event involving a hazard (e.g., someone encountering the hazard posed by a knife) is expressed as the product of the probability of the event (someone encountering the hazard by being cut by the knife) and the consequence (a laceration): Risk = Probability x Consequence (3-1). No single result characterizes the risks associated with a radiation source because there are many possible events involving a radiation source that have consequences, so it is more informative to consider an ordered set of risks. The events may be accidents or malevolent uses of the radiation source for which data may be poor and rigorous quantification may be difficult or impossible, particularly in characterizing the probability. In the context of terrorism and other malevolent misuses, it may still be possible to evaluate consequences (in terms of the number of fatalities, economic losses, and social effects) for specific scenarios with relative rigor. Evaluation of probabilities, however, lies beyond the ready reach of traditional analytic techniques because the probability of a successful terrorist attack involves many factors that cannot be objectively quantified. An assessment that examines only the consequences is called a “hazard analysis” or “hazard assessment.” A hazard assessment can be informative, but is not usually used as the sole basis for risk management, because it can lead to inefficient and inappropriate allocations of resources that can increase rather than reduce risk. Stated differently, risks can actually increase if all of society’s risk mitigation efforts are devoted to events that have high consequences but very low probabilities, neglecting lower consequence, but higher probability events. A formalism is emerging for evaluating risks related to terrorism, using heuristics (rules or devices used to narrow the scope of intractable problems) and whatever other data are available for systematic treatment of the problem (see, e.g., Haimes, 2006; Willis et al., 2005; Paté-Cornell and Guikema, 2002). In this approach, the probability term in the risk equation is usually described as the product of two probabilities: one characterizing the threat and the other characterizing the vulnerability. The threat to a target is a measure of the existing intent and capability to cause harm or damage by carrying out an attack, expressed as the probability of a particular kind of attack on a specific target in a given time period. A target’s vulnerability to the threat is a set of conditions or states of a system that can be exploited to harm that system. It is expressed as the probability that damage occurs from a given threat. The probability in the risk equation would then be the product of these two probabilities. The committee was equipped to evaluate hazards or, from the risk formalism, some of the potential consequences, parts of the vulnerabilities, and none of the threats associated with misuse of radiation sources. Further, it is the role of policy makers to decide what levels of risk are acceptable and how to achieve those levels (e.g., through deterrence, detection, denying access, interdiction, mitigation, or reduction of the hazard). At the same time, however, the committee and the readers of this report can use the concept of risk qualitatively and heuristically to help organize thinking and better target society’s response to the threats. Unless scenarios other than those used to create the source categorization system. However, the examination of Categories 3 to 5 is beyond the scope of this report. 

RADIATION SOURCE RISKS 45 otherwise indicated, wherever the term risk is used hereafter in this report, it is meant to connote the portions of risk (consequences and some aspects of vulnerability) that the committee could examine. The risks associated with radiation sources include both the risks from accidents (i.e., equipment failures and unintentional misuse) and the risks from malevolent uses (i.e., sabotage or weaponization). The former can be called safety risks and the latter called security risks. Each of these is discussed below. Radiation Source Safety Risks Safety has historically been the U.S. NRC’s main focus in regulating the possession and use of radiation sources. Most of the agency’s and the Agreement States’ regulations, guidance, and enforcement center on ensuring radiation safety, which they generally do quite well. Examination of the summaries of radiation incidents in the United States2 shows that accidental exposures do occur, most commonly when required radiation safety procedures are not followed, and the doses received are typically well below 10 mSv. Most of these are direct exposures to gamma radiation when a source is not where it should be (such as a radiography source that does not retract into its housing) or when a person is in an area where he or she should not be (someone wanders into close vicinity of an active radiography camera missing warnings or because warnings are not in place). Major accidental radiation exposures—ones that could cause serious injury or death—are rare. Radiation Source Security Risks Security, too, has always been part of the U.S. NRC’s mission, and many of the commission’s regulations serve both safety and security goals. Accidental exposures resulting from inadvertent access have resulted in some of the worst incidents involving radionuclide radiation sources, so many of the commission’s security measures were designed to prevent inadvertent access. Since September 2001, in response to the changing threat environment, the U.S. NRC’s regulatory focus has sharpened to give security risks related to malicious acts a much greater emphasis than they had previously. The agency has taken steps to prevent and mitigate many security risks, including radiological terrorism. There are numerous scenarios for radiological terrorism, but all can be grouped into three categories (see, e.g., Moore, 2003): • exposing people to radiation by using a radiation exposure device (RED) involves the clandestine placement of a large radiation source in an area where large numbers of people are likely to be exposed; • poisoning food or water supplies with radioactive material; or • dispersing radioactive material either through sabotage of a device in place or by fashioning and operating an RDD. Each of these kinds of attacks can result in fatalities. The committee notes, however, that it is not easy to cause a large number of deaths with radiation sources, regardless of the kind of attack, and there are more direct pathways to lethality (such as bullets and bombs) than using radiation sources. However, our society also values more than just health and safety, and so those who seek to harm us can attack in other ways, aiming to harm our sense of well-being 2 These are available from the state regulatory agencies and the U.S. NRC. See, e.g., http://www.dshs.state.tx.us/radiation/pdffiles/is1q03.pdf (accessed May 24, 2007). 

46 RADIATION SOURCE USE AND REPLACEMENT and our economic prosperity. Psychosocial effects and economic damage, then, are important consequences to consider. REDs Time, distance, and shielding are the key elements of radiation protection and minimizing exposure. REDs use the same elements with an opposing goal: They cause more harm if people are exposed for longer times, at closer proximity, and with less shielding material between the source and the subject. These factors make it difficult to do grave harm to large numbers of people because it is difficult to put many people in close proximity to a source for a long time, and even the human body itself provides some shielding, so that a crowd somewhat shields a radiation source. A single RED, even a large one, might not have major or lasting psychosocial or economic impacts. Multiple REDs that target some critical element of society could have greater consequences than a single attack, but mitigation strategies are readily available, because REDs are inherently easy to detect with radiation detectors. There are a large number of portable radiography cameras at job sites and in transit between licensees. These sources appear to be the most common Category 1 or 2 sources involved in accidental exposures, and RED attacks closely resemble the most common accidental exposures because they involve a person who is unaware that he is in the vicinity of a gamma radiation source. Radiation Sources as Poison Poisoning with radioactive material has garnered some attention since the murder of a former Russian KGB agent in London with polonium-210 (Po-210) in 2006 (see, e.g., Roessler, 2007), but it is difficult to conceive of high-impact physical or economic consequences using this kind of attack. Poisoning large numbers of people to achieve near-term health impacts would be difficult because in food, for example, the radioactive material must be fairly highly concentrated to have a deterministic effect on any individual—bacteria are much more effective at causing harm—so to affect many people requires a very large amount of material. Soluble radioactive material could be introduced into water reservoirs, but almost any plausible number of radiation sources would become too dilute to have much health impact. The material could be introduced closer to the point of consumption, but then the number of people affected would be low. It is possible that a poisoning attack could trigger some mistrust of the food or water supply, but because food-borne and water-borne illness outbreaks occur with some frequency, they are somewhat familiar. Problems with spinach and pet food in 2006 and 2007 have caused temporary economic damage and some concern about food safety, but there is no indication that suppliers of these products will suffer enduring harm. It is also possible that if a reservoir were contaminated with radioactive material, consumers would insist on cleaning up the reservoir, even if the radioactive material had no safety implications for the water in people’s homes. Such cleanup could be costly, but such an attack would no longer be about poisoning; it would be a use-denial radiological dispersal attack. RDDs As with the other modes of attack, it is very difficult to cause serious deterministic health effects for large numbers of people with an RDD, even a very large RDD (Harper and Musolino, 2006). Just as with an RED, time, distance, and shielding are important, and people can 

RADIATION SOURCE RISKS 47 evacuate or shelter in shielded areas, but the concentration or intensity of the source is another important factor. A Category 1 or 2 source is required for an RED to cause any serious deterministic health effects from external exposures. Dispersing radioactive material reduces its concentration, which lowers the likelihood of deterministic health effects from external exposure. People can also get harmful internal exposures from inhaling radioactive particles but, as with poisoning, deterministic effects require fairly high concentrations of the radioactive material, in this case in the form of respirable particles. Dispersing radioactive material increases the likelihood that people will be able to inhale the material, but again, as its concentration goes down, so does the likelihood of deterministic health effects. Dispersal of radioactive material can create persistent area contamination. If dispersed at the right concentrations, the contamination may prevent people from occupying or even using the affected area. An RDD that maximizes this consequence is termed an “area-denial” RDD. RDD risks are decomposed in Figure 3-1. The figure shows a simple block diagram of the major components of RDD risk, dividing it into the two major elements: probability of the event and consequences given a “successful” attack. For an RDD attack to occur, there must be someone or some group sufficiently motivated to undertake it. That group must obtain a quantity of radioactive material, build an RDD, and successfully deliver it to a target. These are the fundamental building blocks or elements that the terrorists require to carry out an RDD attack. The consequences of a successful RDD attack are listed beneath the Consequences box in the figure. They include health effects to the people exposed, economic damage from the area contaminated, and the less tangible psychological and social impacts. These consequences are not totally independent: prompt health effects and the potential for such effects lead to both psychological and economic consequences; specifically, fears and anxieties can cause large adverse economic consequences. However, the quantity of radioactive material needed to cause significant consequences varies according to the type of consequence. As noted earlier, it is very difficult to cause serious deterministic health effects for large numbers of people with even a large RDD. At the other end of the spectrum, even small or ineffective radiation dispersals may stimulate a psychological impact or a government response (Nucleonics Week, 1999). Past accidents have demonstrated that radiation, due to its nature and history, has a unique ability to trigger fear and anxiety in the general population (IAEA, 2006, 1991, 1988). The quantity of radioactive material required to cause economic consequences is somewhere between the large amount needed to cause deterministic health effects and the small amount that might trigger psychological impacts. What is needed to cause economic impacts and the scale of those impacts are discussed below. The federal government has developed plans to deal with natural disasters, accidents, and attacks covering a broad set of incidents, including nuclear and radiological incidents. An annex to this National Response Plan (DHS, 2004) establishes roles and responsibilities for nuclear and radiological incidents. Supporting these plans are proposed guidelines for the levels of contamination that would cause the U.S. government to relocate the inhabitants of a contaminated area and to initiate a cleanup campaign (Federal Register, 2006). The guideline for relocating inhabitants in a contaminated area is 20 mSv (2 rem) in the first year after the incident. That is, if estimates show that an inhabitant who continues to reside in the area would receive 20 mSv in the first year, the inhabitant should be relocated. 

48 RADIATION SOURCE USE AND REPLACEMENT RDD Risk R i Probability Consequences s k Intent Capability Perpetrator Source RDD Delivery & Health Economic & Psychological Motivation Material Development/ Successful Effects Access & Social Acquisition Acquisition Employment Losses Damage Negotiate Physical Intelligence/tips Rad/Nuclear Response Cleanup Socialize plans Appease security on acquisition detection plans plans well Infiltrate Regulatory of material, Points of entry Medical Cleanup Transparency Contain control tools, and Smuggling stockpile technology Reliable Preempt Source expertise for pathways Training of Urban information prioritization dispersal In-country emergency cleanup Realistic vs. sources responders Target engineered and health hardening weapons care Other indicators specialists Prevention Detection/Interdiction Mitigation FIGURE 3-1 RDD Risk Decomposition. SOURCE: Adapted from Connell et al. (2003). Twenty mSv is much lower than the doses required to cause deterministic health effects (immediate or near-term effects, such as those considered in establishing the IAEA source categories). Current best estimates indicate that 20 mSv corresponds to an increase in the average lifetime risk of cancer for an adult from about 42 percent (the current average risk of cancer) to about 42.2 percent (National Research Council, 2006b).3 Under the guidelines mentioned above, the long-term cleanup level or extent of decontamination is to be determined through a case-specific “optimization process.” If history and other contamination incidents are any guide, the cleanup level will be pushed much lower than this: most action levels for cleanup of contaminated sites under Superfund regulations are set at or below the level of 10−4 (1 in 10,000) lifetime risk of excess fatal cancer, which is well below the level of 20 mSv/year based on the linear, no-threshold model of radiation effects. Referring back to Chapter 1, the economic consequences of an area-denial RDD attack are captured by the TRC term: TRC = P (RDD) x Economic Cost (RDD) (3-2). Equation 3-2 states that the terrorist contribution to the cost of using a radiation source is the probability that a terrorist will successfully cause an RDD incident involving this source (P(RDD)) multiplied by the economic consequence of the dispersal of radioactive material (Economic Cost (RDD)). Thus, to actually derive a number for TRC, we need not just the economic cost of the dispersal but also the likelihood that it will occur. Unfortunately, both terms are difficult to estimate. Indeed, we have already discussed that the P(RDD) term is not readily 3 The BEIR-VII lifetime risk model predicts that approximately 1 person in 1,000 would develop cancer from a 10 mSv one-time dose. Some 420 of those same 1,000 people would be expected to develop cancer from other causes (National Research Council, 2006c). Note that mortality from these cancers is perhaps one-half these numbers, depending on age at exposure and several other factors. 

RADIATION SOURCE RISKS 49 quantified and is itself the product of the probability of an attack and the probability that the attack will result in a particular consequence. Only a few reports have been written on the economic costs of radiation dispersals. Those prior to the September 11, 2001, attacks on the United States (9-11 attacks) focus mainly on plutonium dispersal incidents when nuclear weapons were accidentally dropped from aircraft but did not detonate with a nuclear yield (see, e.g., Chanin and Murfin, 1996). Post 9-11 reports are more directed at RDDs (see, e.g., Reichmuth et al., 2005; Kelly, 2002). More economic studies are likely in development, in view of the current interest in and uncertainties about RDD consequences. The existing reports indicate that the cost of cleanup is an exponentially increasing function of the decontamination factor (DF)4 needed to reduce the existing contamination level down to whatever the public and government mutually agree is clean enough for re-inhabitation of the area. The DF required for a cleanup depends on the level of contamination caused by the radiation dispersal and the cleanup standard. A report by the National Radiation Protection Board in the United Kingdom (NRPB, 1996) states that common “muck and truck” methods of cleanup (i.e., large-scale removal of contaminated material, such as such as sweeping, vacuuming, hosing with water, brushing, and application and removal of strippable coatings) are able to achieve DFs on the order of 10 (i.e., they are able to reduce the contamination level by a factor of 10) although factors of 3 to 5 are more common. The same source states that sandblasting can result in a DF of up to 100 for smooth surfaces, if it is done within 30 days after the dispersal. Over time, the dispersed material diffuses into the surfaces on which they were deposited, making the removal more difficult and reducing the effectiveness of sandblasting dramatically (a factor of 10 or more). According to Reichmuth et al. (2005), achieving a DF of 100 or more is not generally possible except by destructive methods (demolition of the contaminated structures and removal of the debris). The total cost of cleanup includes both direct and indirect costs. Direct costs include the physical cost of the cleanup operation, the cost of disposal of the radioactive debris, and the cost of compensating individuals and business that were forced to relocate outside the contamination zone. Reichmuth et al. (2005) estimate that, for a highly contaminated zone requiring a DF greater than 10 in a high-density urban area, the direct costs would range from $10 billion to $40 billion/km2. This could be consistent with figures in the NRPB report (1996), depending on the extent of contamination and the nature of the environment contaminated. Indirect costs include the overall impact on the nation’s economy resulting from the lost business in the affected zone and how these losses ripple through the economy, causing other losses. Rosoff and von Winterfeldt (2007) carried out an analysis of economic consequences of RDDs on the ports of Los Angeles and Long Beach. Their estimates of cleanup costs range from hundreds of millions to tens of billions of dollars, but these were based on the same work of Reichmuth et al. (2005). Their estimates of the indirect costs resulting from port shutdown and related business losses similarly range from hundreds of millions to tens of billions of dollars, depending on the magnitude of the attack. The committee has not reviewed the studies of direct or indirect costs in detail and so draws no conclusions about the reliability of the results, but notes that these are commonly cited figures. To appreciate the hazards associated with Category 1 and 2 radiation sources, one has to understand the actual devices in which the sources are used. Each of the major applications of Category 1 and 2 radiation sources is described briefly below. Chapters 5 through 9 describe these applications in more detail from the perspective of function and possible replacements. 4 The ratio of the before-and-after contamination levels establishes the decontamination factor. 

50 RADIATION SOURCE USE AND REPLACEMENT Panoramic Irradiators Panoramic irradiators are somewhat self-protecting against attacks that require human proximity because exposure to a 37,000-TBq (1-million-Ci) cobalt-60 source (at 1-m separation) would result in an incapacitating dose in about 10 seconds. Furthermore, the thick concrete structure provides additional security from sabotage attacks, and there are Compensatory Measures (special security requirements) mandated by the U.S. NRC at all of the large U.S. panoramic irradiator sites. The sterilization irradiators do, however, require re-sourcing at least once per year, which involves the transport of large quantities of cobalt-60 throughout the United States and installation of cobalt-60 pencils in the source rack. Requirements for radioactive material quantities of concern (RAMQC) apply to these shipments.5 Self-contained Irradiators As noted previously, the form of the cesium-137 in self-contained irradiators is the same as the source in the Goiânia radiation dispersal accident: radioactive cesium chloride powder. The Goiânia accident showed that the cesium chloride salt pellets are easily dispersed if the source container is breached. Another problem with these and other nondefense high-activity cesium-137 sources is that they currently have no permanent disposal pathway in the United States. As noted in Chapter 2, they are considered “Greater than Class C” low-level radioactive waste, and the United States has not yet established a permanent disposal facility for such waste. The National Nuclear Security Administration’s Offsite Source Recovery Project (OSRP) recovers unwanted and abandoned sources, so that licensees are not stuck with their sources indefinitely. The OSRP is discussed further in Chapters 2 and 10. Teletherapy and Gamma Knife® The U.S. NRC reports that there are over 240 cobalt-60 teletherapy units in the United States, although most are not now used for radiotherapy. The sources are compact and intense gamma emitters, and so would primarily be of interest for an RED, although they contain thousands of small cobalt-60 pellets which could be dispersed (see the discussion of the accident in Juarez, Mexico, later in this chapter). The sources in a Gamma Knife® are similar to those used in teletherapy, but the individual sources in a Gamma Knife® are much lower activity than the teletherapy sources (roughly 10 TBq versus perhaps 550 TBq). Each source is held in the Gamma Knife® container in such a way that retrieval of the 201 sources is a slow process requiring specialized tools. Just like in any other use of high-activity radionuclide sources, the security involved with the transport of fresh cobalt-60 sources from the manufacturer to the customer is of some concern. 5 A recent National Research Council report (2006b) recommends “an independent examination of the security of spent fuel and high-level waste transportation …. provide an integrated evaluation of the threat environment, the response of packages to credible malevolent acts, and operational security requirements for protecting spent fuel and high-level waste while in transport.” The security concerns about shipment of radionuclide radiation sources are similar to those about shipment of radioactive waste. 

RADIATION SOURCE RISKS 51 RTGs The former Soviet Union produced approximately 1,000 RTGs for supplying remote power to navigational beacons and lighthouses. Many of these were abandoned after the breakup of the Soviet Union. In addition to concerns about inadvertent exposures, officials in the United States and Russia are concerned that an RTG source might be used in an RDD. The U.S. government and other international partners are currently helping Russia and other former Soviet countries locate and recover abandoned RTGs (see National Research Council, 2007), improve security for RTGs still in use, and replace some with alternative technologies (IAEA, 2005b). The RTGs in the United States are housed on government facilities that are required to have robust security for other reasons. The committee did not explore replacement technologies for RTGs because they are not commercial sources and because the former Soviet sources are not in use in the United States. Well Logging Neutron Sources Well logging sources are used wherever boreholes are found or are being drilled, especially in oil-rich areas of the United States. There are thousands of nuclear logging tools and they are transported by truck from the oil-field service companies to the job sites. The trucks, the devices that use the sources, and the sources themselves are expensive and are provided with some security because of their cost. Industrial Radiography Some (perhaps most) of the thousands of radiography devices are portable (i.e., can be carried by a person) and are used out in the field, making them more vulnerable to theft. Mobile radiography units are heavier devices mounted on wheels or placed on dollies for mobility. Still others are neither portable nor mobile. The U.S. NRC’s Nuclear Material Event Database lists a number of incidents in which radiography devices containing radionuclide radiation sources were lost or stolen. Balancing this is the fact that iridium-192 has a short half-life and is not readily dispersible. Internalizing the Costs of Security Risks As noted in Chapter 1, not all social costs of radiation source use are borne by the users of the radiation sources. One option for implementing replacements that is discussed in Chapter 10 of this report is to make the users bear more of those costs, that is, to “internalize” the costs. With respect to security, the users already bear at least some of these costs. After 2001, the U.S. NRC imposed enhanced security requirements on its materials licensees: Compensatory Measures for panoramic and underwater irradiators, Additional Security Measures for its manufacturers and distributors, and Increased Controls for licensees with Category 1 and 2 devices and sources. The commission (along with the Department of Transportation, with whom the commission shares regulatory authority) also imposed more stringent requirements on transportation of RAMQC. Compensatory Measures, Additional Security Measures, and Increased Controls include varying levels of access controls and alarms with response by security personnel coordinated with local law enforcement and coordination with companies shipping to the facility. The revised shipping requirements are different for Category 1 and Category 2 quantities, and include maintaining constant control and/or 

52 RADIATION SOURCE USE AND REPLACEMENT surveillance during transit and physical controls to serve as barriers to unauthorized removal. The details of the security provisions at any particular site or shipment are not public, and the committee did not review the adequacy of the security. The U.S. NRC indicated that it has carried out inspections for large, panoramic irradiators and manufacturers and distributors and will complete inspections related to the transportation requirements in 2007. Inspections for other licensees who possess Category 1 and 2 sources are scheduled for completion in the summer of 2009. The Agreement States have issued legally binding requirements equivalent to the U.S. NRC’s Orders for Increased Controls. The U.S. NRC told the committee that it plans to promulgate regulations that codify the enhanced security requirements, taking into account lessons learned from their implementation by the U.S. NRC and the Agreement States. Although the committee has not examined the U.S. NRC security requirements in detail, it is aware that their application is based on the IAEA categorization system, specifically applied to licensees possessing IAEA Category 1 and 2 sources/devices. As part of their analysis to identify risk-significant sources and quantities, the U.S. NRC and the Department of Energy (DOE) considered the potential for the material to cause deterministic health effects and contamination of an area greater than 0.5 km2 in excess of the Environmental Protection Agency’s intermediate-phase protective action guide. The latter was a threshold criterion that factored in the potential for contamination (DOE/U.S. NRC, 2003). The U.S. NRC concluded that the results of its own assessment were not significantly different from those found in the IAEA system, and so adopted the IAEA categorization system. The IAEA categorization, as discussed in Chapter 1, is based on defining a “D value” (D for dangerous level) for each radionuclide. Category 1 sources are those that exceed the D value by a factor of 1,000, while Category 2 sources are greater that 10 but less than 1,000 times the D value. Table 3-1 lists the D values and IAEA Category 1 and 2 thresholds for the radionuclides of interest. A low D value indicates a highly hazardous source and a correspondingly low threshold for Additional Security Measures. The key characteristic that yields a low D value is highly energetic decay. Cobalt-60 emits two high-energy gamma rays with each decay, compared with approximately one moderate energy gamma ray for each decay of cesium-137, and so the D value for cobalt-60 is more than a factor of three lower than the D value for cesium-137. Because the IAEA categorization is based on deterministic health effects and safety concerns, it factors in considerations relevant for RED risks and scenarios. It does not, however, account well for RDD risks, which are dominated by area-denial aspects of dispersed radioactive material. Although the DOE/U.S. NRC analysis to identify sources and quantities of concern did include area contamination, it was a single threshold criterion. That is, it could have affected which radionuclide radiation sources are in Category 2, but there would have been no distinction, in terms of the potential for contamination, between Category 2 and Category 1. TABLE 3-1 IAEA D Values and Category 1 and 2 Thresholds IAEA Category 2 IAEA Category 1 IAEA D Value (10 x D) (1,000 x D) Radionuclide (TBq) [Ci] (TBq) [Ci] (TBq) [Ci] Am-241 0.06 [1.6] 0.6 [16] 60 [1600] Co-60 0.03 [0.81] 0.3 [8.1] 30 [810] Cs-137 0.1 [2.7] 1.0 [27] 100 [2700] Ir-192 0.08 [2.2] 0.8 [22] 80 [2200] Pu-238 0.06 [1.6] 0.6 [16] 60 1600] Sr-90 1.0 [27] 10 [270] 1,000 [27,000] SOURCE: Courtesy of IAEA (2005a). 

RADIATION SOURCE RISKS 53 Further, it did not account for differences in cleanup of the contamination or any other factors that might contribute to the economic consequences of an attack. Ultimately, the DOE/U.S. NRC 0.5-km2 criterion may have affected the quantities that define the source categories for a few radionuclides,6 but had no impact on others. To evaluate the adequacy of its security measures, the U.S. NRC carried out security assessments for nuclear materials facilities considering several attractiveness factors in assessing the threat but only deterministic health effects from radiation exposures as the consequences of interest. The U.S. NRC staff characterized this as the commission’s first step in reevaluating security needs for materials licensees. A comprehensive and quantitative examination of the area-denial risks of radiation sources is beyond the scope of this study. However, the key characteristics that make radiation sources hazardous with respect to area-denial RDDs are summarized in Figure 3-2. They are the availability of large radiation sources, the dispersibility of the sources (see Sidebar 3-1), the persistence of the radioactive material once is has been dispersed (this reflects both radioactive decay and the tendency of the material to bind to other materials in the environment), and the potential to cause harm to human health and the environment, which affects the long-term cleanup goal. Several publications (see e.g., U.S. NRC, 2007c; Harper et al., 2006; NCRP, 2006; Argonne, 2005; Van Tuyle et al., 2003; NCRP, 2001) describe features of some or all of the most common high-activity radionuclides with respect to some of these key RDD characteristics. Comparing the radionuclides that emerge from these characteristics, a different ordering of hazards emerges from those shown in Table 3-1 for deterministic health effects. The committee does not advocate a change to the categorization system that is already in place. The IAEA categorization system is being used for multiple purposes by both the international community and the U.S. NRC, spanning regulatory, safety, and security guidance. The system has been used to establish new regulations on the import/export of radioactive material. It is also used both domestically and internationally to help prioritize the recovery of orphaned and unwanted radiation sources. And, as noted earlier, the U.S. NRC and the Agreement States have applied the Increased Controls orders (enhanced security measures) to licensees in possession of IAEA Category 2 and 1 quantities of radioactive material. Radiation Source RDD Hazard Availability of Dispersibility of Persistence Potential to Radiation Source Radiation Source (half-life, reactivity in Cause Harm the environment) (to human health and the environment) FIGURE 3-2 Key characteristics that make radiation sources hazardous with respect to area-denial RDDs. SOURCE: Image provided by the committee. 6 To the committee’s knowledge, there is no documentation that these calculations affected the IAEA categories. The U.S. NRC informed the committee that information the U.S. agencies shared did have effects on the category thresholds for a few radionuclides, and a person involved in the IAEA effort confirmed that conversations with the U.S. agencies led to rethinking of some thresholds. 

54 RADIATION SOURCE USE AND REPLACEMENT SIDEBAR 3-1 FRAGMENTATION Dispersal of radioactive materials in an explosion has emerged as a possible hazard scenario. Fragmentation during an explosion or as a result of impact has been the subject of extensive studies since the Second World War. A distinction is made between brittle materials, such as ceramics and glasses, and ductile materials, such as most metals. For brittle solids, fracture mechanics models predict that the size, d, of dust fragments is related to the fracture toughness for high-speed fracture, Kcr, of the solid: 2/3 ⎡ K cr ⎤ d = 2.9 ⎢ 3-3, ε ⎥ ⎣ ρ cO έ&0O ⎦ where ρ is the density, c0 is the elastic wave speed and έ0 is the deformation rate (see Grady (1982) and quoted by Freund (1998)). The fracture toughness is an intrinsic property of a material and its processing. Most metals undergo plastic deformation before fracture and have considerably larger fracture toughness than brittle solids. Their ductile behavior makes fragmentation into small dust particles much less likely and instead they break or tear into larger pieces. Iridium is a face-centered cubic metal but, unlike the majority of metals with that structure, such as gold and aluminum (which are ductile and show considerable malleability), iridium exhibits little ductility and fractures in a brittle manner. Americium-beryllium sources, such as those used in oil-well logging, typically are formed by sintering mixtures of americium oxide and beryllium powders to form a pellet which is then either diffusion bonded to a metal strip or sealed in a welded stainless steel container. The committee does, however, conclude that consequences other than deterministic health effects, especially the consequences of area-denial RDDs, should be factored into decisions about security for radiation sources. The committee judges that area-denial RDDs have the greatest potential consequences among the kinds of possible attacks with at least some of the high-activity radiation sources, and may pose the greatest risks, as well. Returning to Figure 3-1, for each element of probability and consequence, there are measures that can be taken to either reduce the probability or mitigate the consequences. On the probability side, for example, the probability of acquiring an RDD-significant quantity of radioactive material can be reduced by increasing the security of such materials, tightening the regulatory controls on their use, or by reducing the overall quantity of radioactive materials in use. Similarly, by understanding the relative difficulty of RDD manufacture and RDD effectiveness based on the radionuclide used, the government can better prioritize those radionuclides based on which is most hazardous in RDD scenarios and which poses the greatest RDD risks. This would enable the government to apply greater security and control over those radionuclides. Preventing an attack is clearly preferable to dealing with the consequences, but addressing the consequence side, which involves measures that mitigate the consequences of an RDD attack should one occur, may be just as important. Another National Research Council study (2002) found that better public awareness and education about the true risks of an RDD attack and a clear, well-planned response would greatly help in mitigating the psychological impact on the public. Taken as a whole, the mix of government countermeasures represents a layered defense against RDD attack. No single layer can be perfect because gaps will always exist and determined terrorists can take the time to test the defenses to find weak points. The defensive layers, therefore, must be examined as a system and improvements in defenses must come from identifying and addressing system weaknesses. Thus, even if one cannot numerically 

RADIATION SOURCE RISKS 55 calculate the risk of an RDD attack, this discussion shows that the concept of risk is helpful in organizing our thinking about the problem. It is also useful in understanding how and where defensive countermeasures can be taken and which part of the RDD risk equation is being affected. Radioactive Cesium Chloride This National Research Council study is devoted to examining the options for replacing high-risk radiation sources with alternatives that pose lower risks of malevolent use. In terms of Figure 3-1, this replacement would impact the Source Material Acquisition box, thus reducing the probability that terrorists could acquire a radiation source that would pose a high risk if used in an attack. When evaluating the potential harm that could be caused by different radionuclides in radiation sources, radioactive cesium chloride sources emerge as a major concern. Cesium chloride is soluble and highly dispersible (other forms of cesium-137 mentioned in Chapter 2 are not as dispersible). It emits penetrating radiation and so it cannot be easily shielded if dispersed. Devices containing sizable quantities of this material are used across the United States, most commonly in facilities that are located in cities, large and small, which are potentially attractive targets, and the number of these sources appears to be increasing. All of these factors contribute to making radioactive cesium chloride such a concern to the committee. This concern is exacerbated by the lack of an avenue for permanent disposal of high- activity cesium radiation sources, which increases the likelihood that unwanted cesium radiation sources will remain in unplanned storage where they are potentially more vulnerable to theft. The alternative available to owners of these sources is to obtain the services of the Conference of Radiation Control Program Directors (CRCPD) source disposition program or the OSRP, which have been subject to budget uncertainties. The IAEA efforts have identified the same characteristics and concerns. In the findings from a 2003 IAEA conference, the group encouraged IAEA (2003a): • to formulate and implement national plans for the management of radioactive sources throughout their life cycle; • to develop, to the extent practical, standards for the design of sealed sources and associated devices that are less suitable for malevolent uses (e.g., alternative technologies, less dispersible forms of radioactive sources); • to establish arrangements for the safe and secure disposal of disused high-risk radioactive sources, including the development of disposal facilities. In the same year, the IAEA guidance on security of radioactive sources (IAEA, 2003b) identified the typical form of cesium-137 as “radioactive material that could be easily dispersed via an explosion or otherwise destroying the source.” The IAEA report on the Categorization of Radiation Sources (IAEA, 2005a) also presents an overview of radiation source applications. Figure 3-3 presents an abridged summary of the IAEA data. That figure displays the various applications on the vertical axis, and the activity range per application on the horizontal axis. The horizontal bars represent the range of activity levels for each application while the black vertical line within each bar delineates the typical activity used. The most common radionuclide used for each application is also listed next to each bar (a number of different radionuclides are used for most applications, but the figure shows only the most common nuclide). Note that these applications cover a very wide range of activities (12 orders of magnitude), from tens-of-kBq (microcurie) smoke detectors to hundred- thousand-TBq (megacurie) sterilization irradiators. The thick line running through the applications represents the radionuclide specific threshold levels for IAEA Categories 1 and 2. 

56 RADIATION SOURCE USE AND REPLACEMENT The interagency Radiation Source Protection and Security Task Force formed at the direction of Congress also highlighted cesium chloride as deserving special attention (U.S. NRC, 2006a): A specific area of concern is the widespread use of cesium chloride (CsCl) in a highly dispersible form in certain devices. The Task Force recommends that high priority be given to conducting a study within 2 years to assess the feasibility of phasing out the use of CsCl in a highly dispersible form. This study should include consideration of the availability of alternative technologies for the scope of current uses, safe and secure disposal of existing material, and international safety and security implications. Any plan to phase out these sources should involve industry and consider not only alternatives for uses of these materials, but also how to compensate owners of these sources so that they do not find their way into environments where less rigorous controls are in place. FIGURE 3-3 Radiation source applications, radionuclides, and activity ranges. NOTE: 1 Ci = 0.037 TBq. SOURCE: Modified from IAEA (2005a). 

RADIATION SOURCE RISKS 57 The consequences of dispersal of cesium-137 in the environment can be understood more clearly by examining three accidents involving this radionuclide.7 These events took place in Chornobyl,8 Ukraine; Goiânia, Brazil; and Decatur, Georgia, in the United States. In addition, an accident involving dispersal of cobalt-60 in Juarez, Mexico, illustrates some of the differences between dispersal of discrete radionuclide radiation sources and finely divided radioactive material. Each of these is discussed below. It should be noted that the scale and manner of dispersal of radioactive material by the reactor accident is different from an RDD. Chornobyl dispersed orders of magnitude more cesium-137 than is contained in any device in the United States, but lessons can still be learned from the accident. Relevant Lessons from the Chornobyl Accident The Chornobyl reactor accident of April 1986 (IAEA, 1991) resulted in dispersal of radioactive material worse than any other. A reactor test gone awry led to a sudden tremendous power excursion, causing an explosive breach of the reactor vessel and a subsequent graphite fire that released a plume containing over 3.7 million TBq (100 MCi) of radioactive material. Several days passed before some of the exposed population was made aware of the accident and evacuated. The first responder community of firefighters and other rescue teams suffered the immediate health consequences of Chornobyl—over 30 of them died from acute radiation exposure. A delayed health impact has been seen in increased incidence of cancer. The children exposed to some of the roughly 1.8 million TBq (approximately 48 MCi) of iodine-131 released in the accident (IAEA, 2006) were the most affected members of the population for delayed health effects. It is worth noting that iodine-131 is not available in Category 1 and 2 quantities for commercial use in either the United States or the world market. As significant as these consequences are, the massive economic impact caused by the ground contamination from the cesium-137 released by the Chornobyl accident is its own national-scale disaster. Although many of the radioactive species released from a reactor accident are short-lived or are noble gasses that disperse to harmless levels in the atmosphere, two particular species persist as contamination. They are cesium-137 and strontium-90, both of which are long-lived (approximately 30-year half-lives) and have sufficient dose potential to pose continued risks.9 The Chornobyl accident released roughly 70,000 TBq (2 MCi) of cesium-137 and approximately 11,000 TBq (300,000 Ci) of strontium-90. Thus, the cesium-137 was the dominant radionuclide for ground contamination. The cesium-137 surface contamination was measured, using standard radiation detection equipment. It is presented graphically in terms of the quantity of cesium-137 per square kilometer of surface area, TBq/km2 (Ci/km2) in Figure 3-4, which shows the cesium-137 ground contamination around the Chornobyl site. After much debate and confusion, the Soviet government finally settled on criteria for relocating the populations living on contaminated ground. The upper limit was 1.5 TBq/km2 (40 7 The National Council on Radiation Protection and Measurements recently issued an informative report titled Cesium-137 in the Environment: Radioecology and Approaches to Assessment and Management (2006), which describes cesium-137 contamination in the environment due to releases at sites in the United States, the Ukraine, and Brazil. The report describes near-term countermeasures and long-term cleanup strategies as well. 8 The name of this city and nuclear power plant are commonly spelled Chernobyl in English as a transliteration of the Russian pronunciation. The committee uses here a transliteration of the Ukrainian spelling (Чорнобиль) because the power plant is located in the Ukraine. 9 The cesium-137 deposited as fallout during the accident was not in the form of cesium chloride, but once deposited in a moist environment, the cesium behavior is similar. 

58 RADIATION SOURCE USE AND REPLACEMENT Ci/km2). At this level of contamination, the area was confiscated and the population forcibly removed. No attempt was made to clean up these areas, called confiscated zones (the area illustrated with a mesh of lines in Figure 3-4). The confiscated zones amounted to approximately 3,000 km2, and a population of over 300,000 was relocated. The permanent control zone involved contamination levels from 0.5 to 1.5 TBq/km2 (15 to 40 Ci/km2). In this zone, the inhabitants were given the option to be relocated or to stay and receive financial compensation. FIGURE 3-4 Cesium-137 ground contamination zones from the Chornobyl accident. SOURCE: Image adapted for black-and-white reproduction from De Cort et al. (1998). 

RADIATION SOURCE RISKS 59 From 1986 to 1989, decontamination measures were taken at tens of thousands of residences and public buildings, as well as more than 1,000 farms. Contamination levels were reduced by factors ranging from 10 to 100. In addition to the huge economic costs of cleanup and relocation, the psychological impact on the affected populations of the Ukraine and Belarus were also very significant. According to the 2005 UN review of the Chornobyl incident (WHO, 2005): • Persistent myths and misperceptions about the threat of radiation have resulted in “paralyzing fatalism” among residents of affected areas. • Alongside radiation-induced deaths and diseases, the report labels the mental health impact of Chornobyl as “the largest public health problem created by the accident” and partially attributes this damaging psychological impact to a lack of accurate information. These problems manifest as negative self-assessments of health, belief in a shortened life expectancy, lack of initiative, and dependency on assistance from the state. Relevant Lessons from Goiânia Another serious radiation dispersal accident again involving cesium-137 occurred within two years of the Chornobyl disaster, in Goiânia, Brazil, in September 1987 (IAEA, 1988). An abandoned teletherapy device (see Figure 3-5) was found by scrap metal scavengers. They managed to remove the heavy metal shielding subassembly from the machine (which also contained the sealed source) and, while trying to remove the source holder from the shielding, punctured the 1-mm-thick window of the source capsule with a screwdriver and removed some of the radioactive contents. The scrap was sold to a local junkyard, which then had possession of the source and the remainder of its contents, now exposed to the environment. The cesium-137 source emitted a blue glow and the junkyard owner, believing it to be valuable and possibly supernatural, removed pieces of the source and distributed it to friends and family. The cesium-137 was thus unintentionally dispersed throughout various parts of Goiânia, simply by human handling and normal traffic of humans and animals about the city. The easy dispersal was facilitated by the form of cesium-137 used in these and other high- activity devices in use throughout the world, then and now. As described in Chapter 2, the cesium-137 is in the form of radioactive cesium chloride. Figure 3-5 shows a diagram of the teletherapy device and a picture of the kind of sealed source holder used in the device. The radiation source contained 51 TBq (1375 Ci) of cesium-137. As can be seen from Figure 3-5, the source is small. It was approximately 60 grams of cesium chloride, a quantity that would easily fit inside a typical salt shaker. This quantity of cesium-137, if uniformly distributed, could contaminate an area of approximately 35 km2, using the relocation/confiscation zone criterion of 1.5 TBq/km2 (40 Ci/km2) from the Chornobyl accident. Fortunately, creating a uniform dispersal is difficult, and in Goiânia the dispersal covered a region of only 1 km2. Nevertheless, this small amount of cesium-137 created a huge cleanup problem for the city, resulting in the generation of over 40 tons of radioactive material for disposal. Many RDD relevant lessons can be gleaned from the consequences of the Goiânia accident: First, there were only four initial deaths,10 and primarily among those who actually 10 Over 112,000 people were surveyed for contamination; 249 people were contaminated and of those, 129 had both internal and external contamination. Some 50 people showed signs of whole-body irradiation, and radiation injuries were observed on 28 of them. Fourteen of the 50 had damage to their bone marrow and required intensive care. Four of these people died within 2 months of the accident. 

60 RADIATION SOURCE USE AND REPLACEMENT handled and accidentally ingested cesium-137. Second, psychosocial consequences were considerable; citizens of Goiânia were shunned by the rest of the country and many Goiânians who received no radiation exposure presented with psychosomatic symptoms of radiation sickness. Third, cesium-137 salt is readily dispersible and very active and mobile in the environment. Once on the ground the cesium-137 salt went into solution with the ground moisture. When the ground moisture evaporated in daytime heat, cesium-137 dust particles became airborne, thus enlarging the dispersal area. The large quantities of radioactive waste produced in Goiânia were the result of cesium-137 chemically bonding to standard building materials (such as the tile roofs shown in Figure 3-5). It was not economically feasible to remove the cesium-137 contamination from these surfaces, resulting in the demolition of several contaminated structures. Picture of a sealed source holder, which is approximately three times the size of the sealed source itself: ~ 2.5 cm diameter ~ 50.9 TBq (1375 Ci) cesium chloride powder Schematic view of a teletherapy machine similar to the one in Goiania. FIGURE 3-5 The Goiânia radiation dispersal accident. SOURCE: Image courtesy of IAEA (1988). Many of the people who received high doses (more than 1 Gy or 100 rad) have persistent medical conditions resulting from the exposure. In May 1994, another person who was severely exposed (7.0 Gy) died of chronic liver failure, which is likely a result of the dose received (see Brandao-Mello et al., 2000). 

RADIATION SOURCE RISKS 61 The cleanup criterion used for exterior surfaces in Goiânia was 0.37 TBq/km2 (10 2 Ci/km ). This corresponds to an annual projected dose to inhabitants (using the standard dose conversion factors described previously) of approximately 5 mSv/yr (500 mrem/yr). This projected dose is roughly comparable to the dose received by the general population from natural background radiation, about 3 to 4 mSv/yr, depending on location. Note that no permanent relocations were performed at Goiânia, all the contaminated zones underwent cleanup, and this threshold for initiating decontamination operations (10 Ci/km2) was lower by a factor of four than the Soviet relocation criterion of 40 Ci/km2. Decatur, Georgia and the Qualified Ban on Cesium Chloride Sources in Panoramic Irradiators In 1986, Radiation Sterilizers, Inc. (RSI), requested permission to use 252 capsules containing a total of 428,000 TBq (12.3 MCi) of radioactive cesium chloride from DOE in a wet- storage panoramic irradiator. The U.S. NRC had announced the previous year that it would accept applications for use of the DOE capsules. The cesium-137 sources provided by DOE were much larger than typical cesium chloride radiation sources sold today. DOE had Sandia National Labortories carry out tests with the capsules in wet-load, dry-storage, dry-irradiator mode. This testing did not demonstrate the capsules’ performance under conditions that cycle the temperature, as would happen in a wet-storage, dry irradiator, so another test campaign was begun at a different RSI irradiator in Westerville, Ohio. The U.S. NRC approved the use in other irradiators before that campaign was completed, and the Decatur facility received permission to use the capsules (DOE, 1990; Setser, 1990). In June 1988, about 300 GBq (8 Ci) of cesium-137 leaked into cooling water from a source containing over 1,850 TBq (50,000 Ci) of cesium-137. The leak apparently was caused by stress from thermal expansion of the bulk cesium chloride, which had melted and relocated to form a block at one end of the source capsule. After the accident, three task forces were formed to investigate the causes and lessons learned: One team was formed by the governor of Georgia, another by DOE, and a third by the CRCPD. The Georgia Task Force, in commenting on the future use of the capsules involved in the incident, quoted a comment from the IAEA report on the Goiânia accident and drew the following conclusion (Setser, 1990): ‘[T]he physical and chemical properties of radioactive sources are very important in relation to radiological accidents. They should be taken into account in the licensing of the manufacture of such sources, in view of the potential influence on these properties on the consequences of accidents with the use or misuse of sources.’ This is not an issue to be taken lightly by DOE or NRC. This issue needs to be fully resolved to the satisfaction of all cognizant regulatory agencies involved. The revised U.S. NRC regulation for panoramic irradiators requires that the sources have a certificate of registration issued under 10 CFR § 32.210, be doubly encapsulated, and use radioactive material that is as nondispersible as practical and that is as insoluble as practical if the source is used in a wet-source-storage or wet-source-change irradiator (10 CFR Part 36). The regulation concerns only panoramic irradiators and underwater irradiators, not the dry-storage self-contained irradiators discussed elsewhere in this report. In February 2001, the U.S. NRC denied a source certificate to the company GrayStar for a GS-42 sealed source containing 1,900 TBq (51,500) Ci of cesium-137 chloride in “caked powder” form. GrayStar designed a dry-storage irradiator for food irradiation using 64 of these 

62 RADIATION SOURCE USE AND REPLACEMENT doubly encapsulated sources (total of approximately 122,000 TBq [3.3 MCi]). The denial was based on the dispersibility and solubility of these sources and the requirement in 10 CFR § 32.210 “to provide reasonable assurance that the radiation safety properties of the source or device are adequate to protect health and minimize danger to life and property” (U.S. NRC, 2001). GrayStar argued that cesium dispersed in glass by vitrification would require a greater amount of material to achieve the same irradiation levels as cesium-137 chloride and that the complexity of producing compounds other than cesium-137 chloride would cause “major difficulties and complexities in hot cell operations for source preparation.” The U.S. NRC Atomic Safety and Licensing Board rejected all of these arguments (U.S. NRC, 2001): The Commission determined that the safety hazards associated with leaks of dispersible cesium chloride, even though the leaks were infrequent, justified restricting its use.…The Staff argues that the longer half-life and decay time of cesium-137, combined with its dispersibility, could actually present an increased risk in comparison with cobalt-60 if a leak or other safety problem occurred, including the possible increased risks associated with a proliferation of smaller (non-cobalt) irradiators "in the vicinity of food processors, whose personnel have no previous training or experience with radiation safety." …The [NRC] staff contends that … cesium-137 chloride powder – even in its “caked” form – is dispersible, not only in water but also in air, by physical forces such as air turbulence, physical contact, fire or explosion, should there be a leak in any of the source capsules…. The Staff asserts that the Commission’s nondispersibility requirement “reflects its general defense-in-depth philosophy, in that it assumes sealed sources will leak, and guards against the consequences caused by the spread of radioactive material after a breach occurs. Relevant Lessons from the Accident in Juarez, Mexico11 In late 1983, two men working for the Centro Médico in Juarez, Mexico, hauled some material and equipment taken from the hospital’s warehouse to a junkyard, Jonke Fénix, across town. Among the equipment dismantled in the warehouse was a Picker 3000 cobalt-60 teletherapy device purchased used from a U.S. company. The 20-year-old device was loaded with a new 107-TBq (2,885-Ci) cobalt-60 source in September 1969. One of the handymen transporting the material from the warehouse took out the radionuclide radiation source capsule and, not knowing what it was, pried it open, and ultimately sold the remains of the capsule to the junkyard. Some of the 6,010 cobalt-60 pellets that the radiation source contained were scattered in the pickup truck and on the road, but most were spilled at a few locations in the junkyard where the loading magnet picked many of them up and intermingled them with the scrap metal.12 The junkyard sent scrap metal carrying the pellets to two steel foundries in Chihuahua, which melted them and produced contaminated steel. Correcting for decay to the time of the accident, the cobalt-60 source contained 16.5 TBq (445 Ci), or about 2.7 GBq (74 mCi) per pellet. No one was aware of the contamination until an investigation was initiated by sheer happenstance: A truck carrying contaminated steel rebar took a wrong turn and passed over a radiation detector in Los Alamos National Laboratory that triggered an alarm in 1984. That alarm led authorities in the United States and Mexico to trace the history of the steel. Hundreds of 11 This discussion is taken from TCPA (1998) and Marshal (1984). 12 Cobalt is a ferromagnetic metal, like iron and nickel. 

RADIATION SOURCE RISKS 63 people received significant doses of radiation, including at least four who received between 3 and 4.5 Sv (300 to 450 rem) as whole-body doses. (Based on statistics, one would expect that half of the people in a group receiving doses around 4 Sv would die from the radiation.) About 4,000 tons of steel was contaminated with about 11 TBq (300 Ci) of the cobalt-60, including 600 tons sent to 23 states in the United States. Some 3,400 tons of steel stayed in Mexico and 109 houses built with contaminated rebar were demolished on orders from Mexican health officials. Using a borrowed helicopter and a radiation detector, officials checked the roads between the foundries and the junkyard, and found 22 contaminated sites, including eight pellets that were embedded in the road. The scale of this accident is larger than one would expect from dispersal of discrete pellets of cobalt-60 because the actions of industrial equipment at the junkyard and the foundries dispersed them more widely and finely. The fact that a survey could be done by helicopter and individual pellets could still be recovered illustrates why scattering discrete pellets of cobalt-60, even in large numbers, imposes less of a cleanup burden than dispersal of radionuclide radiation sources with more finely divided material. Because of this incident and several others, foundries in the United States now have radiation detectors, as do the U.S. land border crossings and major shipping harbors. FINDINGS AND RECOMMENDATIONS Finding: The U.S. NRC ranks the hazards of radiation sources primarily based on the potential for deterministic health effects (especially death and severe bodily harm) from direct exposure to the radiation emitted by the bare (unshielded) sources. The U.S. NRC’s analyses that support the commission’s security requirements for nuclear materials licensees are based only on these potential consequences. The U.S. NRC has ranked radiation sources in terms of hazard using the IAEA system of five source categories, determining that the Category 1 and 2 sources are “high-risk sources.” The IAEA analyses supporting its source categorization system consider only deterministic health effects (such as early fatalities) from direct exposure to ionizing radiation from the unshielded source under different exposure scenarios. The initial DOE/U.S. NRC analysis used the same consequences and added in a contamination criterion that falls short of reflecting the area-denial or economic consequences of a dispersal attack. The U.S. NRC also carried out security analyses of each type of facility licensed to use Category 1 and 2 sources, but these analyses were confined to examining the potential for deterministic health effects caused by attacks involving the Category 1 and 2 sources. The U.S. NRC staff told the committee that this was seen as a first step, and that the commission was considering whether to include other factors. Finding: Factors other than the potential to cause deterministic health effects are important when evaluating hazards from radiation sources, especially the potential to cause contamination of large areas resulting in economic and social disruption (area denial). A radiological incident (an accident or especially an attack) could have its most long- lasting and far-reaching effects as a result of contamination of land, buildings, and infrastructure in densely populated regions, partially or completely disabling those assets for human use for long periods of time. This is illustrated by the radiotherapy source incident in 1987 in Goiânia, 

64 RADIATION SOURCE USE AND REPLACEMENT Brazil, and the Chornobyl nuclear reactor accident in the Ukraine. Although an event like the Chornobyl reactor fire is not possible with radiation sources and the scale of the contamination from an incident with radiation sources would inherently be smaller, that 1986 accident showed that radioactive contamination can create sizeable areas that are deemed uninhabitable. The economic and social disruptions caused by such incidents can be difficult to quantify, but they are critical to understanding the scope of the impact beyond the fatalities and severe bodily injuries caused by these events. Recommendation: For prioritizing efforts to reduce risks from malicious use of radiation sources, the U.S. NRC should consider radiation sources’ potential to cause contamination of large areas resulting in economic and social disruption (area denial) to determine what, if any, additional security measures are needed. Having taken an essential first step in considering deterministic health effects from possible radiation exposure from an incident involving radiation sources, the U.S. NRC should now include economic and social disruption in its risk analyses of radiation sources. These impacts can vary significantly depending on the scenarios considered, but that variability does not make them less important. Further, even with such variability, certain factors emerge as important in other analyses of these issues (e.g., Van Tuyle et al., 2003). In carrying out its analyses, the U.S. NRC should not confine itself to the numeric source-activity cutoffs defining the lower limits for Category 1 and 2 sources because the source categorization system itself is based on deterministic health effects. For example, many self-contained irradiators are Category 2 devices, but are near the Category 1 threshold and most americium-beryllium well logging sources have activities near but below the Category 2 limit. Review may show that each set of devices should be regulated similarly. After 2001, the U.S. NRC imposed enhanced security requirements on its materials licensees: Compensatory Measures for panoramic irradiators, Additional Security Measures for its manufacturers and distributors, and Increased Controls for licensees with Category 1 and 2 devices and sources. Compensatory Measures include fairly robust access controls and alarms with response by armed security personnel, along with other measures. Increased Controls include access controls and alarms with response by security personnel, and other measures. After review of the risks associated with some sources and devices, considering more fully the potential for contamination from an attack, the U.S. NRC might conclude that more stringent measures are needed at facilities licensed for some Category 1 and 2 sources and devices. The committee did not examine these security matters in detail and so cannot prejudge the outcome of such analyses. The committee does note, however, that such measures could improve the security of the devices and create a disincentive for owning them. Finding: Because of its dispersibility, solubility, penetrating radiation, source activity, and presence across the United States in facilities such as hospitals, blood banks, and universities, many of which are located in large population centers, radioactive cesium chloride is a greater concern than other Category 1 and 2 sources for some attack scenarios. This concern is exacerbated by the lack of an avenue for permanent disposal of high-activity cesium radiation sources, which can result in disused cesium sources sitting in licensees’ storage facilities. As such, these sources pose unique risks. Radioactive cesium chloride sources are in the form of a steel-encapsulated, compressed powder. The salt is highly dispersible and water soluble. There are approximately 1,300 high-activity cesium chloride devices (each with an activity of tens to hundreds of 

RADIATION SOURCE RISKS 65 terabecquerels [hundreds to thousands of curies]) across the United States, nearly all of which are self-contained irradiators. The number of these devices and sources appears to be increasing. Because it emits energetic gamma rays and its half-life is long enough that an irradiator does not need to be reloaded over the device’s expected lifetime, cesium-137 has been the key component of self-contained irradiators for blood irradiation and research for many years. Cesium chloride is the least expensive and highest-specific-activity form of cesium-137 available today. Because of the nature of the applications that employ these irradiators, they are most commonly located in hospitals, blood banks, and universities, many of which are located in cities, large and small. The presence of these sizable sources in areas that are potentially attractive targets is a major factor making radioactive cesium chloride such a concern to the committee. Finding: In view of the overall liabilities of radioactive cesium chloride, the committee judges that these sources should be replaced in the United States and, to the extent possible, elsewhere. Finding: Discussions in Chapter 2 and 5 show that in most (and perhaps all) applications, radioactive cesium chloride can be replaced by (1) less hazardous forms of radioactive cesium, (2) radioactive cobalt, or (3) nonradionuclide alternatives. However, not all of these alternatives are available now, and all are currently more expensive than radioactive cesium chloride for the users. Finding: Government action is required to implement replacement of radioactive cesium chloride sources because the alternatives cost more and the liabilities or social costs of the sources are not borne by the end users. There is no indication that replacement of devices containing Category 1 and 2 radioactive cesium chloride sources with lower hazard alternatives will improve or worsen the performance of the devices in their standard and proper uses. The act of replacement incurs monetary costs and the replacements themselves currently cost more in most cases than the radioactive cesium chloride devices. All of these costs would be borne by the end users (paying more for the alternatives) and the current device manufacturers (depending on the price elasticity of demand and potential loss of sales). The benefits of replacement are in reducing the liabilities and social costs (including the costs associated with the risk of terrorist attacks and, in some cases, the full costs of disposal, discussed in Chapters 2 and 10), which are shared by the public rather than borne by the end users. Except in cases where the replacements prove to be cheaper, end users have little incentive to shift away from their current devices; and unless there is a demand for the alternatives, manufacturers are unlikely to invest in making the alternatives available. Government action can, however, provide the requirements or incentives to implement replacement. Recommendation: In view of the overall liabilities of radioactive cesium chloride, the U.S. government should implement options for eliminating Category 1 and 2 cesium chloride sources from use in the United States and, to the extent possible, elsewhere. The committee suggests these options as the steps for implementation: 

66 RADIATION SOURCE USE AND REPLACEMENT i. Discontinue licensing of new cesium chloride irradiator sources. ii. Put in place incentives for decommissioning existing sources. iii. Prohibit the export of cesium chloride sources to other countries, except for purposes of disposal in an appropriately licensed facility. In Chapter 10, the committee offers several suggestions as its lead candidates for how to implement the replacement, and they are summarized here. First, to stop the addition of new Category 1 and 2 cesium chloride sources to the nation’s inventory, the U.S. NRC should discontinue all new licensing and importation of these sources and devices. This includes import of new sources from other countries and recycling of sources from decommissioned devices. Second, many licensees may need incentives to decommission13 their existing sources or devices because the devices still have use value. Indeed, there are now also disincentives to decommissioning beyond the loss of use, including the costs of decommissioning. Third, if the sources recovered from decommissioned devices (or the devices themselves) are simply sold outside the United States, then the sources are still potentially available for use in an attack on another country or even the United States. Therefore, disposition options are needed in the United States. These are discussed in more detail in Chapter 10. The overall policy could make exceptions based on unique needs that cannot be met with alternative technologies, but the threshold for creating exceptions should be set high, similar to what the U.S. NRC has done for panoramic irradiators. 13 U.S. NRC’s technical definition of the term decommissioning applies mainly to facilities and involves removing licensed radioactive material to an extent that allows public release of the facility and termination of the license. The committee’s usage here and throughout the report is slightly different: A decommissioned device is retired from service and sent to whatever disposition option is available (disposal or storage pending dismantlement and disposal), and the license for the device is terminated.