[NCI Logo]

Inadequacy of the IAEA's Air Transport Regulations:
The Case of MOX Fuel

Edwin S. Lyman
Scientific Director
Nuclear Control Institute
1000 Connecticut Avenue, NW Ste 804
Washington, DC 20036 USA

Introduction

It is becoming increasingly likely that in the future, large quantities of commercial plutonium will be shipped internationally not in the form of oxide powder, as was the case during the controversial sea shipments of 1984 and 1992, but in the form of finished mixed- oxide (MOX) fuel assemblies. Some reasons for this are (1) the delay or cancellation of MOX fuel fabrication facilities in the nations which are foreign customers of reprocessing companies, such as Germany and Japan; (2) the expansion of MOX fabrication capacity in the countries that provide reprocessing services; (3) the reprocessing industry's hope that the transport of plutonium in the form of MOX fuel will be perceived by the public as posing less severe environmental and security risks than the transport of plutonium oxide, and will therefore generate less controversy.

Under current regulations, the International Atomic Energy Agency (IAEA) Safety Series No. 6, large quantities of radioactive materials (RAM) may be shipped by air in packages that are no more robust than those used for land-based modes of transport -- the so- called Type B packages. However, because it must be assumed that Type B packages will fail if subjected to the significantly greater mechanical and thermal stresses occurring in typical aircraft accidents, industry has refrained from shipping large quantities of unconsolidated and readily dispersible materials, such as plutonium oxide powder, by air.

On the other hand, plutonium incorporated into MOX fuel assemblies is currently transported by air by British Nuclear Fuels plc (BNFL) from the U.K. to continental Europe in Type B packages. The logic underlying this practice is that even if the package were destroyed in an aircraft accident, release of radioactivity would be limited due to the solid form of the fuel. However, the actual level of protection provided by the MOX fuel itself under aircraft crash conditions has not been quantified. Public controversy over air shipments of plutonium have inhibited the expansion of the practice.

In the IAEA's 1996 revision of its recommendations for the transport of radioactive materials (ST-1), packaging standards specific to air transport are introduced. These include the specification of a Type C package, which must withstand a somewhat more severe accident environment than the Type B package. Under ST-1, large quantities of RAM can be rendered exempt from the Type C requirement if it can be demonstrated that they are "low dispersible material" (LDM), which is given a quantitative definition in the standards.

The ST-1 air transport standards are the result of a decade of deliberations among a number of parties, of which most had a commercial interest in the outcome. As a result, the standards were designed with an emphasis on minimizing their economic burden on shippers of radioactive materials, and fall short of providing adequate assurance that the risks of such transports will be acceptably low. These standards are not intended to provide protection in the event of very severe aircraft accidents, which could result in a widespread dispersal of plutonium from a cargo of MOX fuel. Consequently, adoption of ST-1 will do little to alleviate the serious safety concerns associated with the air shipment of radioactive materials in general, and MOX fuel in particular.

The IAEA argues that although ST-1 is not perfect, it is an improvement over existing regulations and will therefore enhance the safety of air transport of RAM. However, this is not necessarily the case. Adoption of a flawed standard will increase the political acceptability of air transport of RAM and could conceivably lead to a large increase in the number of such transports, thereby increasing the risk of a severe accident that is not covered by the standard. Moreover, it is argued below that the LDM exemption was not derived from a rigorous safety analysis but was carefully constructed to legitimate the existing, risky practice of shipping MOX fuel by air in Type B packages. If this is the case, adoption of ST- 1 will lead to no changes in practice and hence no increase in safety.

Type B Package Standards and the Notion of "Graceful Failure"

Under the 1985 (revised 1990) edition of Safety Series No. 6 (SS6) which will remain in effect until the 1996 version is adopted by individual IAEA member states (a process that can take years), unlimited quantities of plutonium in any form can be transported by land, sea or air in so-called "Type B" casks. Type B casks are designed so that they will lose only a fraction of their contents (a radionuclide-specific quantity known as "A2" per week, an amount deemed "acceptable" by the IAEA and which is tabulated in SS6) if they are exposed to accident conditions equivalent to an impact of 13.2 meters per second (m/s) on an unyielding surface, followed by an fire with a flame temperature of 800C for 30 minutes.

It has been pointed out by many observers on many occasions that the conditions that a RAM package may encounter in the course of an accident at sea or the crash of an aircraft can be far more severe than those simulated by the Type B test. One of the chief arguments invoked by the IAEA in response is the notion of "graceful failure": namely, the claim that RAM packages are designed and constructed with such a high degree of conservatism that they will be able to withstand accident conditions far more severe than those under which they are tested.

The "graceful failure" principle is central to the philosophy on which the 1996 version of SS6 is based, as will be discussed further below. However, there is hardly any experimental evidence for it, as the IAEA freely admits.1 Package manufacturers have little incentive to carry out testing to a severity beyond what the standards require; such tests are expensive and difficult, and they present the risk of embarrassment should the package fail abruptly rather than gracefully.2 In the absence of experimental verification, "graceful failure" is based entirely on expectation.

The IAEA, in defending the concept of "graceful failure," cites in a recent document the findings of a recent study that "cask-like" structures can survive "forces greatly exceeding those imparted by the regulatory tests" without gross failure --- in particular, that they can survive on-axis impacts (e.g. with maximum protection from the impact limiters at the ends of the package) of 4 times the energy of the regulatory tests (e.g. an impact speed of 26.8 meters per second [m/s], about twice the 13.3 m/s speed associated with the regulatory test). However, it neglects to mention that the same study found that these casks exhibited "significant leakage from both seals" following an impact of 5 times the energy (2.25 times the velocity, or 30 m/s), or an off-axis impact of only 4 times the energy.3

While the graceful failure concept may have limited validity in the case of spent fuel casks, for which the shielding requirements mandate very thick (25 cm) metal walls that also provide mechanical strength and thermal insulation, the Type B packages used for transport of unirradiated MOX fuel require far less shielding. A review of available information on MOX transport cask designs does not provide confidence that they can withstand mechanical and thermal loads far in excess of those for which they were designed.

A typical Type B U-MOX fuel cask design is the Westinghouse Model No. MO-1. Although the 1976 US Nuclear Regulatory Commission (NRC) license for this design was allowed to lapse, it is considered representative of the type of package that will be used in the US for the transport of U-MOX, should the US carry out its plan to dispose of some excess weapon plutonium by fabricating it into MOX and irradiating it in reactors.4 The MO-1 is designed to transport 2 MOX assemblies containing about 45 kg of plutonium. The package weighs 3.9 tonnes and has outer dimensions 1.143 m x 1.194 m x 5.23 m. It consists of two concentric steel shells, each around 3 mm thick. The volume between the shells is filled with an 18-cm thick layer of rigid polyurethane foam to provide shock and thermal insulation. The MOX assemblies are separated by borated stainless steel plates for criticality control. Because of the importance of insulating the assemblies from shocks during routine transport, they are shock-mounted inside the package with clamps.

Less information is publicly available about the Transnucl‚aire FS-69, the Type B package currently used in France to transport U-MOX assemblies by road, and presumably the same one that will be used for international shipments. The FS-69 is a Type B package licensed for the transport of U-MOX containing plutonium with up to 30,000 parts per million (approximately 3 weight-percent) americium-241 (Am-241). This corresponds to plutonium obtained from spent fuel of 45,000 MWD/t burnup and aged for six years, which is the plutonium composition used in the design of the MELOX plant.

Details of the FS-69 are considered proprietary. However, from the general descriptions that can be found in the open literature, the cask can be seen to have many similarities to the MO-1: it carries two assemblies held in place by a borated aluminum basket, weighs 5 tonnes when loaded and consists of two concentric steel shells separated by a neutron-absorbing material.5 On the basis of the weight similarity with the MO-1, the FS-69 must contain a comparable amount of structural metal.

It is reasonable to assume that with respect to the construction materials and overall design philosophy, the FS-69 is similar to the FS-47, the container that was used to transport cans of plutonium oxide by sea in 1993. The FS-47 structure, like that of the MO-1, is based on two thin concentric steel shells (outer shell a few millimeters thick; inner shell about 1 cm thick) separated by a 5-cm thick layer of heat insulation ("wet", or hydrated, gypsum) and a 15-cm thick layer of a (proprietary) neutron absorber material.6

No information on the type of seal used in the FS-69 could be located. The ability of a cask to maintain containment if exposed to a beyond-design-basis fire depends crucially on the seal material. For instance, the elastomer seals used in the Transnucleaire TN 28 VT cask for vitrified high-level radioactive waste (VHLW) will fail if they are heated to a temperature above about 250C, which could occur if the cask were exposed to an 800C fire for a couple of hours. According to COGEMA, the seal of the FS-47 cask failed after it was exposed to a 1000C fire for 1.5 hours;7 if the FS-69 uses the same type of seal it can be expected to behave similarly.

Even less information is available about the "mystery" package now used by BNFL to transport MOX by air to Switzerland. A recent request for details about this cask was denied by BNFL, on the grounds that "disclosure of package types and capabilities would result in a breach of customer confidentiality."8 However, according to BNFL, "... all packages for MOX transport fulfil the requirements of all the appropriate transport regulations." All this means at present is that the package used to ship MOX by air is Type B, the significance of which will be discussed below. However, one would expect that the package would not be more robust than the FS-69 used by BNFL's commercial competitor, Cogema. BNFL had attempted in the past to design a package specifically for air transport of plutonium oxide, known as the 1680. However, "due to changing commercial priorities ... it has not yet been used."9

It is apparent from what little is publicly known of these package designs that their structural strength and thermal resistance is provided primarily by the filler material between the two steel shells, which are themselves too thin to provide much strength. However, materials such as rigid polyurethane foam typically are not capable of providing resistance to mechanical or thermal stresses well in excess of design stresses. If the energy of an impact is significantly higher than the design energy, the foam will be crushed without causing any deceleration of the package. However, increasing the resistance of the package to high energy impacts by using a denser foam would increase the risk that the package could not withstand lower energy impacts, since the foam will not compress (and therefore act as a rigid surface) if the energy is too low.10 In other words, for any foam density, there is a fairly narrow "window" of impact energies for which the foam is capable of providing protection.

"Wet" (hydrated) gypsum is a material used in boards for building construction. It is not capable of load-bearing and therefore cannot provide the FS-47 cask with significant impact resistance.

With regard to thermal resistance, organic materials such as polyurethane foams function primarily via ablation -- that is, they absorb heat energy by burning and carry heat away from the payload by dispersal of combustion products. Therefore, they can only continue to protect against a fire until they are completely consumed. After this point, the contents of the cask will quickly achieve thermal equilibrium with the fire temperature. In thermal tests of a polyurethane foam shielded by a metal lid, more than 25% had degraded after exposure to a flame of approximately 980C for fifteen minutes.11 Thus the "graceful failure" margin for casks that rely on ablative media for thermal protection is rather slim.

In the FS-47 cask, thermal protection is provided by the layer of hydrated gypsum, which has a low thermal conductivity at temperatures around 40C. However, hydrated gypsum can decompose at temperatures as low as 150C, and it cannot be expected to provide a significant margin of safety against fires of greater severity than the design basis fire.

Behavior of Unirradiated MOX Fuel Under Accident Conditions

Thermal: Although MOX fuel is a refractory ceramic with a very high melting point (around 2700C), if it is heated in the presence of oxygen it readily oxidizes, expands and undergoes comminution (production of fine particles). This process can take place at temperatures as low as 250C, so that significant oxidation and comminution is possible even in thermal conditions of moderate severity.12 In particular, prolonged thermal exposure at relatively low temperature, as is characteristic of smoldering conditions following a severe fire, could result in substantial particulate formation.

In fact, experiments indicate that oxidation and pulverization of sintered MOX fuel pellets is most severe at temperatures considerably below 800C, at which the Type B and C regulatory tests are conducted. After 30 minutes at a temperature of 400C, MOX pellets were found to have completely pulverized, with scarcely any coarse pieces and considerable production of fine powder.13 However, at 800C, the pellet shapes were still recognizable. Production of respirable particles (aerodynamic diameter <10 µm) is also greatly enhanced at the lower temperature 1.87% of the initial mass of MOX pellets was converted to such a form after 15 minutes at 430°C in air, as compared to 0.01% at 800°C.

Oxidation can only occur, by definition, if the fuel pellets themselves are exposed to oxygen. Therefore, for oxidation to occur mechanisms must exist for failure of the cladding of fuel rods and for the cask seal (if the casks are filled with inert gas; otherwise, oxygen will be present in the cask even if the seals initially remain intact). Fuel rod cladding can be ruptured either by mechanical impact during the accident or by bursting as a result of thermally induced overpressure. The latter has been observed to occur in spent fuel rods after exposure to temperatures over 725C for a four-hour period.14 Extrapolation of this result to unirradiated fuel rods is not straightforward, because they contain no fission gas and their cladding has not corroded from interaction with fuel and coolant during reactor operation. However, one should note that for PWR fuel rods are internally pressurized with helium gas.

If the cladding has been breached due to mechanical impact, oxidation of the fuel will cause it to expand and exert additional pressure on the cladding, which will cause further ruptures and greater exposure of the fuel pellets to air.

If oxygen access to the fuel pellet surfaces is limited, then there will be little particulate formation. However, the volatility of americium-241 in the fuel (of which there may be a substantial concentration) will be enhanced in reducing conditions, which could result in the release of highly radiotoxic americium vapor from the fuel during a fire, even if the fuel matrix itself is not dispersed. In contrast, volatilization of plutonium (e.g. gaseous release) is not likely to occur to a significant extent for the range of temperatures that would be encountered in a transport accident.

Mechanical: Uranium oxide ceramic pellets are brittle materials, and shatter when exposed to high-energy impacts. The size distribution of particles produced by such impacts is typically log-normal. Experiments on depleted uranium pellets subjected to an impact of energy 0.1 J/g, corresponding to an impact on an unyielding surface of 14.1 m/s (slightly higher than the IAEA Type B impact speed) have found that the pellets will release approximately 0.5% of their mass as particles with diameters less than 100 microns (called the "dispersible fraction"), and 0.01% as particles with diameters less than 10 microns (called the "respirable fraction") [Fig 1].15 Higher impact speeds shift the distribution in the direction of smaller average particle size [Fig 2], and thus in the direction of increasing hazard.16 (The "dispersible fraction" denotes particles that can be transported easily in the form of aerosols; larger ones tend to settle rapidly. The "respirable fraction" denotes particles that tend to remain deep in the lung once inhaled.)

Because there is little or no information available on the behavior of actual MOX pellets following impacts, one must rely on the uranium oxide data given above. This should provide a rough idea of the impact resistance of MOX pellets. However, differences in the microstructure of the two fuel materials, and in particular the inhomogeneity of the MOX pellets, may affect the relative impact behavior. Uncertainties can only be eliminated with tests of the actual material to be transported.

Transport of MOX by Air: The Low Dispersible Material (LDM) Exemption

Air shipment of plutonium and other radioactive materials has been a controversial practice, largely because of the obvious inadequacy of the IAEA transport standards for the air mode. Past versions of SS6 were essentially "mode-independent": the same package standards (Type B) applied for both ground and air transport, despite the fact that the mechanical and thermal stresses encountered in a plane crash would be far greater than those simulated by the Type B test. This deficiency was highlighted most forcefully by the United States, which unilaterally adopted very stringent domestic regulations on air transport of RAM.

In response to public criticism and the disparity between SS6 and U.S. domestic regulations, the IAEA made an attempt to increase the credibility of its standards for air transport. IAEA ST-1 defines requirements for packages, called "Type C," which are intended for the air transport of large quantities of RAM. The Type C qualification test includes an impact of 90 m/s on an unyielding surface, and a non-sequential 800C, 1-hour fire. Although the Type C standards are more rigorous than the Type B standards, they fall far short of the US domestic regulations. The Type C standards were set not with a specific safety target in mind, but were taken at the so-called "knee of the curve" point, beyond which significant improvements would come at too great a cost to industry, in the IAEA's judgment. In fact, according to the data which the IAEA used in formulating its standards, 5%-15% of air crashes are associated with impacts more severe than that stipulated by the Type C test.

Moreover, the actual risk of failure of a Type C package in a aircraft accident may be much greater because the Type C impact and fire tests do not have to be performed sequentially. According to ST-1, Type C packages do have to satisfy the Type B test, which involves sequential impact and fire tests of considerably lower severity. However, the ability of a Type C package to withstand an aircraft accident involving both a severe crash and a fire is unknown.

IAEA attempts to justify the decision to make the Type C test non-sequential by claiming that "in severe accidents, high speed impact and long duration fires are not expected to be encountered simultaneously because high velocity accidents cause fuel dispersion." This argument is invalid for two reasons. First, according to data which was used as input for ST- 1, the impact speed above which fuel dispersal becomes a significant factor is around 130 m/s,17 which is far greater than the Type C impact speed. In fact, several accidents have been recorded involving speeds from just below to greater than 90 m/s to that were followed by severe fires, including a crash at an airport in Qatar in 1979 (impact speed 87 m/s, fire duration 3.5 hours). Second, the argument ignores the possibility of combustion from an external source of fuel, such as a severed natural gas line, as was the case in the crash of an El Al jet into an apartment building in Amsterdam.

Moreover, it is illogical to conclude that a Type C package can survive combined impact and thermal loadings characteristic of accidents with impacts in the range in between the Type B and Type C impacts. For example, even if a package remains leak-tight after an impact in that range, the impact may damage the package in a way that greatly degrades its thermal resistance. Conversely, thermal exposure might damage the impact resistance of some packages, for accidents in which a fire precedes the impact.

Having introduced a new but still inadequate standard for air shipment of RAM, IAEA then proceeded to include an exemption from the Type C requirement for so-called "low dispersible material" (LDM). While the actual text makes no mention of specific materials which may fit the LDM criteria, it is widely understood that one of the primary beneficiaries of this exemption will be transporters of MOX assemblies, who would be free to ship unlimited quantities of MOX fuel by air in Type B casks should it be demonstrated that MOX meets the LDM criteria.

One strong indication that plutonium shippers expect that they will be able to transport MOX by air in Type B casks under ST-1 is the noticeable lack of development work in recent years by the nuclear transport industry on a large-capacity air transport package, even though introduction of the Type C standard was anticipated as long as a decade ago. For instance, at the 1995 PATRAM (Packaging and Transport of Radioactive Materials) Conference, there was not a single paper on RAM air transport packaging, in stark contrast to the 1989 and 1992 meetings, at which there were entire panels on the subject. One suspects that attempts to develop Type C packaging for plutonium were abandoned once it became clear that an exemption applicable to MOX would be included in the revised regulations.

This has been confirmed in a recent letter from BNFL, which stated that "BNFL has undertaken some provisional tests to establish whether MOX fuel might be classified as Low Dispersible Material. The results of these test [sic] were encouraging ...".18 BNFL gives no further details or references to support their claim.

The certification process for LDM is as follows: the test material will be subjected to the Type C impact and fire tests (also non-sequentially), without the protection of any packaging. The material qualifies as LDM if it then does not release an amount of activity greater than 100 A2 in gaseous and particulate forms of up to 100 microns in diameter. As is the case with package tests, compliance can be shown by "direct physical tests, analytical methods or a proper combination of these." One should note that the Type B sequential impact and fire tests which Type C packages must undergo are not required for LDM qualification.

This precision of the LDM definition, which was originally introduced by Germany, falsely gives the impression that it was derived from detailed technical analysis. In fact, a review of the supporting documents shows that the technical basis of the LDM definition is obscure. The radiological arguments advanced in the German working papers on the subject seem crude and arbitrary.19 There is evidence, however, that the definition may have been constructed from German data on the oxidation behavior of MOX fuel pellets so that MOX fuel would automatically qualify.

The first odd aspect of the LDM definition is that it appears to be inconsistent with the radiation protection standards upon which the remainder of the SS6 is based. For instance, following an accident, a Type B or Type C package is supposed to release an amount of activity no greater than A2 in one week (with no restriction on particle size). However, material qualified as LDM can release more than one hundred times that amount (100 A2 in particles with diameters less than 100 microns, plus an unlimited amount of activity in particles greater than 100 microns). Therefore, in order for the transport of LDM to be consistent with this standard, the Type B package must be in sufficiently good shape following a plane crash that it can prevent more than 99% of the activity released from the LDM from escaping into the environment. This invocation of the "graceful failure" hypothesis requires that one believe that a Type B cask will remain largely intact if it is subjected to an impact energy 50-100 times greater than that which it was designed to withstand.

The German working group papers on LDM are vague on exactly why they believe their criterion provides an acceptable standard. At one point, they make reference to "graceful failure"; at another, they cite the low rate of aircraft accidents; at another, they cite the effect of atmospheric dilution of the release. None of these explanations provides a convincing rationale for adopting release rates for LDM that are inconsistent with Type B and Type C release rates.

The definition of the LDM test proposed by Germany also underwent a significant change. Originally, the German proposal suggested that the impact and fire tests be conducted sequentially, because it was realized that the 1-hour fire test alone conducted on a sample such as a MOX fuel rod would probably cause little or no dispersal, whereas a fire following an impact which ruptured the fuel rod (and permitted the ingress of oxygen) could cause a much greater release. Therefore, use of non-sequential impact and fire tests would give a misleading impression of the non-dispersibility of the material. However, they proposed that the fire test in the sequence be limited to only 10 minutes at 800C.20 In later versions of their proposal, the recommendation that the impact and fire tests be performed sequentially was omitted.

A 1982 paper from the Fraunhofer Institute in Germany on the oxidation behavior of MOX fuel pellets in a kerosene fire may shed some light on the mysterious origins of the LDM standard.21 In this paper, it was found that heating MOX pellets to 800C for 15 minutes in an air atmosphere led to a release of 0.01% of the initial material in the form of fine particles. Therefore, for a Type B cask carrying about 50 kg of plutonium in two MOX assemblies, this release fraction would correspond to a release of 5 grams of plutonium, which is identical to 100 A2 ~ 5 grams (for a typical reactor-grade plutonium composition). Thus it is possible that a release of 100 A2 was proposed by German scientists based on their expectation of how a MOX fuel rod would perform with respect to their initial LDM test proposal (e.g. an impact test causing fuel pellets to be exposed to air, followed by a 10- minute fire test).

Another important result from the Fraunhofer Institute is that oxidation of MOX pellets and fine particle formation increases as the oxidation temperature decreases below 800C, and is greatest at 430C (see above). This indicates that hazardous releases from MOX fuel may actually be greater for less severe thermal conditions than the Type C thermal test. A material that meets the LDM criterion could fail if exposed to a fire test at a lower temperature. This is a further indication of the inadequacy of the LDM concept as articulated in ST-1.

Despite the apparent effort to work backward from the known properties of MOX fuel to devise a definition of LDM, it remains highly uncertain whether U-MOX will be able to meet the LDM standard, especially with regard to its impact resistance. An impact of 90 m/s corresponds to an energy input of around 4 J/g, which, based on depleted uranium oxide impact tests, would cause the release of (substantially) more than 0.5% of the material as particulates with diameters less than 100 microns. For the 2-assembly package, this would be equivalent to a release of more than 250 grams, well in excess of 100 A2. Only if the assembly structural materials and fuel rod cladding provide a great deal of impact resistance, or remain largely intact with only very small ruptures, will a MOX assembly be able to meet the LDM impact test.

Another large source of uncertainty comes from the LDM qualification test itself. The feasibility and potential environmental consequences of carrying out these tests on actual MOX fuel have not been seriously considered. It is possible that the industry will resort to demonstrating LDM with simulant materials or relying entirely on computer modeling. Neither of these methods will provide assurance that the behavior of actual MOX pellets is being accurately represented. France's Institute of Nuclear Safety and Protection (IPSN) recently voiced its concern about the feasibility and reproducibility of LDM qualification tests.22

Nonetheless, BNFL has indicated that it is moving ahead on qualification of U-MOX as LDM, and that "together with other European nuclear companies, [BNFL] has embarked on a full development of an LDM test facility in line with the requirements of the published LDM regulations."23

For the above reasons, the LDM definition, as currently formulated, is fatally flawed. To improve it, at a minimum, the following three changes should be carried out:

Potential Consequences of the Crash of a MOX Cargo

In the event that a MOX Type B package experiences an accident of greater severity than it is designed to withstand, the amount of material released will be determined by the response of the contents to the accident and by the "graceful failure" behavior of the package. As discussed above, MOX packages that use ablative materials for fire protection will not be able to withstand a prolonged fire (greater than a few hours). Also, the elastomeric seal materials used in many RAM packages will fail after prolonged exposures to relatively low temperatures (~ 250C).Of special concern is an accident which first causes the rupture of many fuel rods and is then followed by a long-duration fire. Even if the fire smolders at a low temperature, substantial oxidation of the fuel rods can take place if the package is ruptured or the seals fail. The amount of fuel oxidized would be limited only by the duration of the fire and the availability of oxygen.

If industry is able to qualify MOX as LDM using the current, inadequate standard, permitting large-scale MOX air shipments in Type B package for the indefinite future, the consequences could be disastrous. For a crash at the not inconceivable speed of 140 m/s, one must assume that a Type B MOX package would fail completely. Such an impact, corresponding to an energy input of around 10 J/g. By extrapolating between the data in Figs. 1 and 2, one sees that such an energy input can cause a release of fuel on the order of 1% for particles smaller than 10 microns (corresponding to 500 grams of plutonium per cask) and 10% for particles smaller than 100 microns (corresponding to 5 kg of plutonium per cask). A subsequent fire would cause further fine particle formation and dispersion. A typical cargo might consist of several casks, all of which must be presumed to fail under such severe conditions. For a 10-cask shipment, a release of 5 kg of plutonium in respirable form is therefore not out of the question.

These releases are quite significant from a radiological perspective. The consequence code MACCS2, developed for the U.S. NRC, was used to assess the consequences of the release of 5 kg of reactor-grade plutonium as a result of an air crash in an area with a population density of 250 persons/km2. For a buoyant release as a result of a hot fire, neutral atmospheric conditions and a light wind, committed effective doses resulting from the passage of the radioactive plume were as high as 52 Sievert (Sv) at 200 meters from the crash site, and remained above 50 mSv for more than 40 km (64 mi) from the crash site. There were more than 4300 cancers committed from the initial passage of the plume. The total number of cancers, including those resulting from resuspension of the ground contamination, exceeded 16,000 over a 100-year period.

A crash in a densely populated urban area could result in total casualties several times greater. The crash and prolonged fire of an El Al jet at an apartment complex in Amsterdam in 1992 serves as a stark reminder that this type of accident is very much in the realm of possibility.

End Notes

1. International Atomic Energy Agency, "The Air Transport of Radioactive Material in Large Quantities or with High Activity," IAEA-TECDOC-702, April 1993. Back to document

2. Even attempts to design casks to withstand more severe accident conditions than the Type B test have not been successful; for example, when Japan in 1987 tested a cask specifically developed to meet the stricter US domestic air transport standards, it failed to survive the impact test [P. Leventhal, M. Hoenig and A. Kuperman, “Air Transport of the Japanese from Nuclear Fuel States,” March 3, 1987, p.5]. Back to document

3. D. Ammerman and J. Bobbe, “Testing of the Structural Evaluation Test Unit,” Proceedings of PATRAM '95, Volume III, p.1123. Back to document

4. Westinghouse Electric Corporation, “Plutonium Disposition in Existing Pressurized Water Reactors,” report prepared for the U.S. Department of Energy, DOE/SF/19683-6, June 1, 1994, p. 2.7-2. Back to document

5. J. Charles and O. Konirsch, “Transportation by Road of Plutonium as a Reusable Product,” Proceedings of the 11th International Conference on the Packaging and Transportation of Radioactive Materials (PATRAM ‘95) (December 3-8, 1995, Las Vegas, Nevada), p. 777. Back to document

6. Science and Technology Agency of Japan, reapplication for Design Change Approval of Nuclear Fuel Transport Cask, FS-47 (in Japanese). Back to document

7. COGEMA, “Le Retour Au Japon Du Plutonium,” 1992. Back to document

8. Alan Hughes, BNFL Public Affairs Division, letter to Fred Barker, 24 January 1997. Back to document

9. Ibid. Back to document

10. F. Henry and C. Williamson, “Rigid Polyurethane Foam for Impact and Thermal Protection,” Proceedings of the 11th International PATRAM Conference, December 3-8, 1995, Las Vegas, Nevada, p. 1161. Back to document

11. Ibid. Back to document

12. T. Sanders et al., "A Method for Determining the Spent-Fuel Contribution to Transport Cask Containment Requirements,” Sandia Report SAND90-2406, Sandia National Laboratories, November 1992, p. IV-23. Back to document

13. H. Seehars and D. Hochrainer, "Durchfuhrung Experimenten zur Unterstutzung der Annahmen zur Freisetzung von Plutonium bei einem Flugzeugabsturz,” (in German), Fraunhofer-Institute, SR 0205A, March 1982. Back to document

14. Ibid, p. III-6. Back to document

15. Ibid. Back to document

16. Ibid, p. IV-12 - IV-16. Back to document

17. L. Fischer, J. VanSant and C. Chou, “Draft Criteria for Controlled Tests for Air Transport Packages,” Lawrence Livermore National Laboratory, UCRL-ID-103684, August 1990. Back to document

18. Alan Hughes, BNFL Public Affairs Division, op cit. Back to document

19. F. Lange, F. Nitsche, F-W. Collin and M. Cosack, “Requirements for Very Low Dispersible Material (VLDM)," TC-946, Working Paper No. 11, IAEA Technical Committee Meeting, Vienna, 15-19 May 1995. Back to document

20. F. Lange and F. Nitsche, "Contribution to Technical Committee Meeting,” Working Paper 8, IAEA, Vienna, 29 August - 2 September 1994. Back to document

21. H. Seehars and D. Hochrainer, "Durchfuhrung Experimenten zur Unterstutzung der Annahmen zur Freisetzung von Plutonium bei einem Flugzeugabsturz,” (in German), Fraunhofer-Institute, SR 0205A, March 1982. Back to document

22. A. MacLachlan and G. Seneviratne, “France’s ISPN Raises Concerns About New IAEA Transport Standards,” Inside N.R.C., September 16, 1996, p. 12. Back to document

23. Alan Hughes, BNFL Public Affairs Division, op cit . Back to document



[Air Transport Page]] Air Transport Page[What's New Page]] What's New Page [Home Page]Home Page

mail@nci.org