Safeguards and Security for High-Burnup TRISO Pebble Bed Spent Fuel and Reactors

Abstract

Several high-temperature thermal neutron–spectrum pebble bed reactors are being commercialized. China has started up two helium-cooled pebble bed high-temperature reactors. In the United States, the X-Energy helium-cooled and the Kairos Power salt-cooled pebble bed high-temperature reactors will produce spent nuclear fuel (SNF) with burnups exceeding 150 000 MWd per tonne. The reactor fuel in each case consists of small spherical graphite pebbles (4 to 6 cm in diameter) containing thousands of small TRISO (microspheric tri-structural isotropic) fuel particles embedded in the fuel of zone these pebbles.

The unique isotopic, chemical, and physical characteristics of this high-burnup SNF create a technical case to eliminate safeguards based on the low risk for use in nuclear weapons, while maintaining safeguards in terms of risk for use in radiological weapons. These safeguards could be reduced to the simple counting and monitoring of pebbles in storage. Alternatively, there is the option to create a special category with reduced requirements for this SNF in storage, transport, and disposal. No safeguards would be required for a repository with only this type of SNF. Reactor safeguards are required for fresh fuel, partly burnt fuel, and to identify unconventional pebbles with depleted uranium or other materials that might be used to create weapons-useable materials.

Keywords:

• Safeguards
• pebble-bed fuels
• TRISO fuels

I. INTRODUCTION

With the worldwide interest in new nuclear technologies in response to climate change, high-temperature reactors have received increased attention, not only for base-load electricity but also by markets that need high-temperature heat. The demand for variable electricity creates large incentives to add heat storage to enable base-load nuclear reactors to provide variable electricity to the grid.^[1] The cost of stored heat systems decreases with increased heat storage temperatures. The second high-temperature heat market is the chemical industry, such as the recent decision of Dow Chemical to buy four high-temperature gas-cooled reactors (HTGRs) for its Seadrift chemical plant site in Texas. The third market is for the production of cellulosic hydrocarbon biofuels to replace all crude oil where the refineries require high-temperature heat.^[2]

The new generation of high-temperature reactors are pebble bed reactors. China started up two helium-cooled pebble bed high-temperature reactors (HTR-PMs) producing a total of 200 MW(electric) in 2022. Dow Chemical, in partnership with X-Energy, has announced plans to build four Xe-100 helium-cooled pebble bed reactors, each with an output of 200 MW(thermal), for its Seadrift Texas site.^[3] The U.S. Nuclear Regulatory Commission (NRC) have granted a construction license^[4] for the Kairos Power 35-MW(thermal) salt-cooled pebble bed HERMES test reactor, the first fluoride salt–cooled high-temperature reactor (FHR).

These modular reactors using pebble fuel bring with them renewed interest in addressing nuclear proliferation concerns for reactors with online refueling capabilities. Today, nonproliferation regimes are primarily focused on the 435 light water reactors (LWRs) deployed in over 40 countries under the auspices of the International Atomic Energy Agency’s (IAEA’s) safeguards program.^[5] Today’s nonproliferation challenge is to address the possible rapid expansion of non-LWR technologies using innovative fuels and processes that are not directly amenable to the inspection and verification techniques that are currently applied by the IAEA. Additionally, the relatively small size [50 to 300 MW(electric)] of these reactors to meet different market requirements will require a large number of reactors to meet the electricity and process heat demands of the world.

One such technology is the use of pebble bed reactors that contain small spherical pebbles (4 to 6 cm in diameter) of graphite containing thousands of small microspheric tri-structural isotropic (TRISO) fuel particles embedded in each graphite pebble. A typical reactor may contain over 200,000 such pebbles, which are continuously circulated in the reactor, discharging only those that have been sufficiently depleted in fissile fuel such that they are no longer useful in sustaining power production.

There are two types of such reactors currently being designed and licensed in the United States. One, a HTR-PM, is now operating in China, producing 200 MW(electric) to their electric grid.^[6] The HTR-PM is a high-temperature, helium gas–cooled pebble bed reactor. U.S. development of this technology^[7] is being done by X-Energy for its Xe-100 [200 MW(thermal)] reactor.^[8] The other type of pebble bed reactor is the FHR, which is cooled by a molten fluoride salt containing over 200,000 4-cm-diameter graphite TRISO particle pebbles. This reactor concept^[9] is being designed and licensed in the United States for deployment by Kairos Power [280 MW(thermal)].

While different in coolant and design, both discharge the pebbles in large numbers to be placed in onsite storage containers pending disposal in licensed repositories. The current nonproliferation regime requires material accountability of the used fuel coming out of the reactor, which means measuring the amount of special nuclear material remaining and having it monitored to be sure that it is not diverted into weapons-usable material. This is relatively easily done with spent LWR fuel assemblies since only 30 to 50 spent fuel assemblies are discharged every 2 years or so. However, as an example, the X-Energy pebble bed reactor discharges about 157 pebbles per day or about 58,000 pebbles per year, making detailed measurement of every pebble exceedingly difficult, if not impossible, and perhaps unnecessary.

The purpose of this paper is to evaluate the options available to assure proliferation-resistant means to manage the spent nuclear fuel (SNF) from pebble-type reactors so as to allow for their deployment worldwide, specifically high-burnup pebble bed SNF. There were many studies about a decade ago on pebble bed nonproliferation challenges.^[10–17] Since then, there have been major changes. Proposed fuel burnup has doubled. The commercial-scale FHR will achieve 190,000 MWd/tonne burnup in less than 18 months. Fuel testing has fully qualified such fuels to these higher burnups.^[18] These and other changes have significant implications for safeguards. What follows is a brief description of each type of reactor and the characteristics of the spent pebble fuel with suggestions as to how to approach a proliferation-resistant means of accountability based on more recent work^[19–21] and our analysis.

II. SAFEGUARD AND SECURTY GOALS

There are two goals of nuclear safeguards and security: (1) prevent use in the construction of nuclear weapons and (2) prevent radiological weapons. A radiological weapon is where the nuclear material is dispersed with conventional explosives or other methods to expose people to high levels of radiation and contaminate a local area. The threat can be from a nation state or a subnational terrorist organization. In terms of nuclear weapons, nuclear materials are divided into two categories. Direct-use materials, such as plutonium and high-enriched uranium, can be used in nuclear weapons. Indirect-use materials, such as low-enriched uranium, require an enrichment plant to convert the feedstock into a weapons-useable material.

Nuclear materials are divided into different categories based on the difficulty of converting the materials into direct-use nuclear weapons materials. This is the basis for the division between low-enriched uranium and high-enriched uranium at which less than 20% is deemed acceptable for power reactor use. Nuclear weapons safeguards can be terminated on nuclear materials based on the criteria that anyone wanting to build a weapon would not use that feedstock but rather would enrich natural uranium. The classic example is plutonium in high-level waste glass. When reprocessing SNF, uranium and plutonium are removed and the fission product wastes are converted into a waste form for geological disposal. The separation is not perfect, leaving some plutonium in the glass. Nuclear weapons safeguards can be terminated because the proliferator would start with natural uranium as the faster, lower-cost route to weapons materials. There are also requirements to prevent theft because of the potential use in a radiological weapon depending on the physical and chemical characteristics.

Special nuclear material categories are based on traditional nuclear materials in the fuel cycle. High-burnup pebble bed SNF with burnups between 150,000 and 200,000 MWd per tonne of heavy metal is very far from traditional materials coming from LWRs in terms of (1) plutonium and uranium isotopics, (2) fissile material concentrations, and (3) chemical forms. This raises two questions. First, can safeguards in terms of nuclear weapons materials be terminated on these materials because natural uranium is an easier route to direct-use nuclear weapons materials? Second, if nuclear weapons safeguards are not terminated, should these materials be put into a separate category with significantly reduced safeguards requirements? Either case implies that this specific SNF is not a significant proliferation risk. This does not change the safeguards requirements for fresh fuel or partially irradiated fuel.

III. REACTOR AND FUEL DESCRIPTIONS

Table I summarizes the two types of reactors with TRISO fuel. Figure 1 shows the two types of fuel pebbles. In both reactors, the TRISO fuel particles are embedded in a graphite matrix, which is the fuel element, and a neutron moderator. Each TRISO particle contains the fuel, a uranium oxi-carbide particle 0.425 mm in diameter that is surrounded by multiple layers, including a silicon carbide (SiC) cladding. The external diameter of the TRISO fuel with multilayer coatings is 0.855 mm. TRISO fuel failure temperatures are above 1600°C; the graphite can go to higher temperatures. These reactors are designed such that they cannot have core meltdowns. This is due to (1) the physics of the design and (2) the high-temperature capability of the fuel, which enables the decay heat in an accident to be conducted to the environment, thus maintaining temperatures below fuel failure temperatures.

TABLE I Characteristics of Pebble Bed Reactors

Reactor Type	HTGR	FHR
Vendor design	X-Energy	Kairos Power
Power level [MW(thermal)]	200	280
Temperature in/out (°C)	260/750	550/650
Coolant	Helium	^7LiBeF4 (Flibe salt)
Coolant properties	Gas	Liquid: boiling point >1400°C
Fuel: TRISO (uranium oxi-carbide)	UCO	UCO
Pebble diameter (cm)	6 cm	4 cm
Pebble enrichment (%)	15.5	19.55
Pebble burnup (MWd/tonne)	168,500	193,600
UCO kernel diameter	0.425 mm	0.425 mm
TRISO particles per pebble	~19,000	16,000
Uranium per pebble (g)	7	6
Pebbles in core	220,000	300,000
Pebble geometry	TRISO in carbon matrix/carbon outer layer	Carbon interior TRISO in carbon matrix/carbon outer layer
Pebble residence time (day)	1320 equivalent full power	522
Passes through reactor Number of pebbles discharged day/year	6 157/58,000	8 217/79,200

Fig. 1. Kairos Power and X-Energy fuel pebbles.

The Kairos pebble is 4.0-cm in diameter with fuel particles in an annular zone. It is the size of a ping-pong ball. The X-Energy pebble is larger (6 cm), with fuel particles throughout the sphere. It is the size of a tennis ball. Both have a fuel-free outer matrix shell (Fig. 1).

Liquid-salt coolants are much better coolants than helium; thus, the FHR uses smaller pebbles to increase surface area, with the power density 4 to 10 times that of a gas-cooled reactor.^[19–21] In the FHR, the graphite with embedded TRISO particles is a layer outside a graphite inner sphere. The annular fuel design provides two benefits: (1) reduce peak temperatures to allow for much higher power densities and (2) enable adjusting the inner carbon sphere density to enable the pebbles to float in the liquid salt. The salt provides about half the neutron moderation; thus, the volume of SNF is half to a third of that of a gas-cooled reactor if both fuels have the same burnup. In the HTR-PM, the pebbles are larger. The entire inner pebble is graphite with embedded TRISO fuel particles. Like the FHR pebbles, there is an outer protective layer of carbon. The FHR gamma radiation fields are significantly higher for short-cooled fuel, which can impact some types of diversion scenarios (more shielding required) and radiation measurements (to avoid saturating radiation detectors).

Each reactor uses online refueling, with pebbles going through the reactor core multiple times before discharge. For gas-cooled systems, the pebbles enter at the top and flow by gravity to the bottom of the vessel. In the FHR, the pebbles float in the salt. They are injected under the reactor core and float up to form a packed bed that moves slowly upward to a defueling chute at the top of the core. When each pebble exits the reactor, it is inspected for burnup and wear. If there is insufficient burnup (fuel utilization), it is recycled back to the reactor core. For both types of reactors, the Burn-up Measuring System uses the ¹³⁷Cs activity to determine the extent of uranium utilization.^[10] To perform a measurement typically requires a decay time of hours, after which the measurement uncertainty can be reduced to less than 5%. This measurement can be correlated with detailed burnup and isotopic physics calculations of the core.^[22]

Online refueling results in the nearly uniform burnup of pebbles in contrast to LWR fuel assemblies where there is high burnup at the center of the fuel assembly and low burnup at the top and bottom. It also results in significantly higher burnup than high-temperature reactors with hexagonal fuel blocks that contain TRISO particles in fixed positions, which are then stacked vertically with six to eight blocks in a hexagonal array. While the ability to rotate fuel blocks in the vertical and horizontal directions increases burnup, limits on economically allowable refueling times significantly lower hexagonal fuel block burnups relative to what, in theory, is possible.

Figure 2 shows plutonium quantities and isotopics with burnup for a gas-cooled pebble bed reactor. Plutonium-239 first increases and then decreases. Other plutonium isotopes increase with burnup. The quantity and quality of the fissile uranium decreases with burnup. The intrinsic proliferation risks of the pebble bed fuel decrease rapidly between 100,000 and 200,000 MWd per tonne. There have been detailed studies of nonproliferation risks associated with HTGRs, including the differences between block and pebble bed reactors,^[11,12] but these have been at much lower burnups than the current designs of pebble bed reactors. These changes in burnup require rethinking the safeguards.

Fig. 2. Fissile composition in gas-cooled pebble bed reactor versus burnup.^[18]

Table II shows the detailed average isotopics for a FHR of pebbles on each of the eight passes through the reactor. The original report^[21] assumed an initial heavy metal loading of 3.4 g. The current Kairos Power design has 6 g of heavy metal (uranium) per pebble, with Table II adjusted to an initial uranium loading of 6 g per pebble. The discharge burnup is near 200,000 MWd per ton of initial uranium.

TABLE II Average-per-Pass Mass in Grams of Isotopes per Pebble and Other Parameters in Kairos Power FHR*

Isotopes and Other	Pass 1	Pass 2	Pass 3	Pass 4	Pass 5	Pass 6	Pass 7	Pass 8
^235U g/pebble	9.60E-1	7.91E-1	6.51E-1	5.37E-1	4.39E-1	3.62E-1	2.98E-1	2.45E-1
^236U g/pebble	3.97E-2	7.13E-2	9.62E-2	1.15E-1	1.30E-1	1.41E-1	1.49E-1	1.55E-1
^238U g/pebble	4.75E-0	4.68E-0	4.62E-0	4.55E-0	4.50E-0	4.45E-0	4.38E-0	4.32E-0
^235U percent enrichment	16.7	14.3	12.1	10.3	8.66	7.31	6.17	5.19
^238Pu g/pebble	2.07E-5	1.77E-4	5,54E-4	1.19E-3	2.14E-3	3.30E-3	4.71E-3	6.30E-3
^239Pu g/pebble	4.08E-2	5.88E-2	6.60E-2	6.92E-2	7.01E-2	7.02E-2	6.99E-2	6.94E-2
^240Pu g/pebble	5.35E-3	1.36E-2	1.98E-2	2.38E-2	2.65E-2	2.77E-2	2.82E-2	2.88E-2
^241Pu g/pebble	1.45E-3	6.90E-3	1.38E-2	2.01E-2	2.49E-2	2.81E-2	3.05E-2	3.16E-2
^242Pu g/pebble	7.11E-5	7.52E-4	2.45E-3	5.10E-3	8.42E-3	1.20E-2	1.58E-2	1.94E-2
Total Pu g/pebble	4.77E-2	8.02E-2	1.03E-1	1.10E-1	1.32E-1	1.42E-1	1.49E-1	1.56E-1

*Initial 6 g uranium per pebble with enrichment of 19.55%.

IV. IMPLICATIONS OF ISOTOPICS

Different reactors use different levels of uranium enrichment. The economics favor using high-assay low-enriched uranium for pebble bed reactors: 19.55% ²³⁵U for the FHR with a burnup near 200,000 MWd/tonne. There are several reasons for the higher uranium enrichment levels. First, they minimize fuel fabrication costs. The burnup limits for coated particle fuel burnup are far beyond 20% ²³⁵U, which is the dividing line between weapons-useable and non-weapons-usable uranium. The coated particle fuel is significantly more expensive to fabricate per kilogram of initial heavy metal. This creates economic incentives for higher enrichments to reduce the total mass of fuel requiring fabrication, transportation, and disposal.

In contrast, LWR fuel burnup has been limited by clad failure. Given the clad burnup limits, there were no incentives for uranium enrichments above 5% until recently with improved cladding, and even then maximum enrichments will be below 8%. Second, the higher enrichments reduce reactor size. The coated particle fuel enables building reactors that can eliminate many accident scenarios, including reactor core melt down. However, this implies large quantities of graphite in the fuel and the use of TRISO fuel particles.

The higher enrichment results in very high power densities within the coated particles but much lower pebble power densities, resulting in pebbles that are easy to cool. Fuel is quickly depleted, minimizing financial charges; that is, the cost of money between the time when the fuel is purchased and when the fuel is converted into energy. Third, in a pebble bed reactor, the uranium fuel is mostly burnt; there is little useable fissile fuel. In a traditional LWR fuel assembly, the top and bottom of the fuel assembly have lower fissile burnup because of the axial thermal flux distributions, so more fissile fuel material is left in the SNF. The flowing pebble bed design enables efficient fuel utilization.

The higher burnups in thermal neutron–spectrum reactors result in very different fuel isotopics compared to other fuels. The plutonium concentration goes up and then levels off at higher burnup. Nuclear weapons are built using fissile ²³⁹Pu produced in reactors with conditions designed to minimize the quantities of other plutonium isotopes. The typical composition of weapons-grade plutonium^[15] is 94.3% ²³⁹Pu, 5.5% ²⁴⁰Pu, and 0.2% ²⁴¹Pu. As shown in Table II, the plutonium composition for this high-burnup pebble bed fuel is 4% ²³⁸Pu, 44% ²³⁹Pu, 18% ²⁴⁰Pu, 20% ²⁴¹Pu, and 12% ²⁴²Pu.

The large buildup of other plutonium isotopes makes their use in nuclear weapons much more difficult. Plutonium-239 becomes less than half the total plutonium. About 90% of the fissile Pu is utilized during operations. At discharge, the high heat–emitting ²³⁸Pu is 4% of the total plutonium mass. The heat generation rate of pure ²³⁸Pu is 0.54 kW/kg. This implies that a kilogram of this plutonium produces 21.6 W. The IAEA defines a significant quantity of plutonium as 8 kg.^[23] A significant quantity of this plutonium generates 173 W.

If one wants to build a weapon, the plutonium in the middle of the weapon is generating heat like an incandescent light bulb surrounded by high explosives, which act as insulators. Separate from the weapons design, there are major challenges in fabricating plutonium components with tight dimensional tolerances with thermally hot metal. The completed weapon would require assured cooling to prevent damaging the weapon or causing a nonnuclear burn or explosion of the conventional explosives in storage and transport.

Second, the ²⁴⁰Pu is 18% of the plutonium and emits spontaneous neutrons that cause pre-ignition in nuclear weapons. Third, the required quantities of plutonium to build a weapon are significantly larger with this high-burnup plutonium. While the critical mass of ²³⁹Pu and ²⁴¹Pu are similar, the critical mass of ²⁴⁰Pu is about four times larger, and the critical mass of ²⁴²Pu is 7 to 10 times larger; much more plutonium is required to make a nuclear weapon from this plutonium than other sources of plutonium.

Last, the plutonium quantity and quality in the SNF decreases with time. The half-life of ²⁴¹Pu is 14 years, so the quantities of this isotope decrease by about 5% per year as it decays into ²⁴¹Am. As the SNF ages in storage, a larger fraction of the remaining plutonium becomes ²⁴⁰Pu and ²⁴²Pu, which are highly undesirable in a weapon, with the quantity of plutonium required to make a weapon going up with time.

The recovery of this plutonium from the pebble SNF is challenging, partly because the concentration of plutonium in the SNF is very low relative to any other SNF and partly because of the complications of chemically separating the carbon and SiC from the fuel. Because the fuel is designed to withstand more extreme conditions than any other fuel (part of its safety case), the recovery of fissile material is more difficult than with other SNF. In a FHR, one needs to reprocess over 51 000 pebbles to acquire a significant quantity of plutonium, assuming no losses in the process. This assumes the IAEA definition of 8 kg as a significant quantity of plutonium, which is based on the plutonium being ²³⁹Pu. If the IAEA definition of a significant quantity of plutonium is adjusted upward because much more of this plutonium is required to make a weapon than other sources of plutonium, significantly more pebbles must be reprocessed. It is concluded that if a weapons designer had a choice, plutonium in high-burnup pebble bed TRISO SNF would be at the bottom of the list.

While the plutonium isotopic composition may make it very undesirable for nuclear weapons, it would be a valuable fuel for fast neutron–spectrum reactors, where it could be blended with other plutonium to produce acceptable fuels. The ²³⁸Pu and ²⁴⁰Pu would be converted to fissile ²³⁹Pu and ²⁴¹Pu.

In parallel, the uranium isotopic composition for use in nuclear weapons degrades with high burnup and the buildup of ²³⁶U and smaller quantities by secondary reactions of ²³²U and ²³³U. In a FHR, the ²³⁶U is 62% of the ²³⁵U content in the discharge SNF. This high-burnup uranium would not be recycled today because the buildup of minor uranium isotopes results in higher radiation exposures in the enrichment plant and limited fuel value—and in any weapon built of such materials by the enrichment process. This includes ²³²U that has a decay product that builds up over time producing 2.6-MeV gamma rays, air inhalation risks,^[15,24,25] and degradation of equipment. Uranium-232 has a short half-life (T_1/2 = 69 years); thus, there is significant alpha radionuclide decay from the ²³²U and its decay products. The ²³³U half-life (T_1/2 = 159,000 years) is longer, but it and its decay products contribute to the alpha activity. This decay will degrade the centrifuge plant through multiple mechanisms, including radiolysis of UF₆ with the generation of highly reactive fluoride.

If a proliferator builds an enrichment plant for a uranium weapon, the technology choice today would be centrifuge.^[26,27] In a centrifuge, essentially all of the light isotopes (^232,233,234U) will concentrate with the ²³⁵U if there is simple separation of ²³⁵U from ²³⁸U. If it is decided to separate the lighter and heavier isotopes from the ²³⁵U, a complicated three-component separation would be required. Alternatively, such uranium can be blended with other uranium in the enrichment process to dilute the effects of these minor isotopes. Recent work^[24,25] has better defined what is required.

The separation is significantly more difficult than separating ²³⁵U from ²³⁸U starting with natural uranium because the separation factor that measures the difficulty of isotopic separation in a centrifuge is^[27]

$\mathrm{Separation}\mathrm{Factor}=\exp \left[\left({\mathrm{M}}_2-{\mathrm{M}}_1\right){\mathrm{V}}^2\left(1-{\mathrm{r}}^2/{\mathrm{a}}^2\right)\right]/2\mathrm{RT},$

where

${\mathrm{M}}_{2,}{\mathrm{M}}_1$ =	=	molecular weights of the two uranium hexafluoride isotopes that are being separated
$\mathrm{V}$ =	=	velocity of the centrifuge with “a” the external radius of the centrifuge and “r” the radius where the separation factor applies
$\mathrm{R}$ =	=	gas constant
$\mathrm{T}$ =	=	absolute temperature.

All of this is in the exponent of the separation factor. There is a mass difference of 3 separating ²³⁸U from ²³⁵U, but only a mass difference of 1 separating ²³⁶U from ²³⁵U. Separating the ²³⁵U from the ²³⁶U is much more difficult than separating ²³⁵U from ²³⁸U with a mass difference of 3. The difficulty of these isotopic separations implies that if one wants useable weapons-grade uranium, natural uranium likely becomes the preferred starting material.

This uranium can be blended with other uranium to fuel low-enriched reactors, such as LWRs and CANDU reactors.^[15] Recent work has further explored how to enrich uranium^[24] with unusual isotopic compositions into commercial reactor fuels where the required enrichments are relatively low. This includes blending with other uranium, which would be viable in a large enrichment plant but not in a small enrichment plant as might be used by a proliferator.

The United States uses high-enriched uranium in Navy nuclear-powered vessels and high-performance test reactors, such as the Advanced Test Reactor (ATR). These reactors have high-burnup fuels. The ATR average burnup is about 250,000 MWd/tonne,^[28] with a starting fuel that is 93% ²³⁵U. The result is that the ²³⁶U to ²³⁵U ratio is several times less than high-burnup pebble bed fuel with a starting enrichment of 19.55%. This is an indicator of how unusual the uranium isotopics of high-burnup pebble bed SNF are, partly enabled by the uniform burnup of the discharged pebbles versus the average burnup of a normal fuel assembly.

V. ARE NUCLEAR WEAPONS SAFEGUARDS REQURED FOR HIGH-BURNUP PEBBLE BED SNF?

There is the assumption that the IAEA requires safeguards for all SNF because of the potential direct use of the plutonium in nuclear weapons. Earlier work by Moses and Ehinger^[12] and Moses^[15] showed that this is not necessarily true for some types of high-burnup pebble bed fuel. The basis of this finding was not to terminate safeguards, but instead to limit the accountancy of plutonium where the risks of diversion are very low. The IAEA^[29] has “provisional” guidelines for the termination of safeguards on “measured discards.” This applies to wastes from reprocessing plants, but could also be applied to high-burnup pebbles. These limits vary from 1200 g/m³ of plutonium for externally contaminated clad hulls stabilized in concrete to 2500 g/m³ for fixed clarification sludges and plutonium in high-level waste glass.

The high-burnup FHR pebble bed SNF with a burnup of ~200 000 MWd per tonne has ~4460 g/m³ of pebbles. With a packing density of 50%, the plutonium content is 2230 g/m³ of pebbles. Most plutonium from reprocessing plant wastes contains high concentrations of ²³⁹Pu. If one calculates plutonium based on the fissile content, then these pebbles are significantly below the plutonium safeguards termination limits. This would limit the safeguards requirements to simply accounting for the number of pebbles in storage to prevent diversion. This rationale is based on existing IAEA regulations and does not account for the other characteristics of this plutonium, as discussed previously, which create major barriers for use in nuclear weapons.

The uranium in the SNF is not sufficiently diluted to meet the IAEA provisional guidelines for the termination of safeguards, although the quantities of ²³²U and ²³⁶U above ~91 GWd/tonne exceed commercial limits for recycling reprocessed uranium. However, with burnups of 150,000 to 200,000 MWd/tonne, natural uranium becomes more attractive as a feedstock for the production of highly enriched uranium, as discussed previously. A proliferator will also look at the economics to choose the lowest risk and lowest cost route. Thus, applying safeguards for pebble fuel containing uranium at high burnups is also not necessary, as discussed previously.

The other difficulty for diversion is that pebble bed TRISO fuels are much more difficult to reprocess than LWR fuels.^[11,12] In a LWR fuel assembly, about two thirds of the weight is uranium. In HTGR fuel, the fuel is 2% of the total weight and the SiC is 3% of the total weight; most of the remainder is carbon in different forms. For the FHR, the uranium and SiC contents are several times larger. Separately, the SiC clad around each microsphere creates added difficulties for reprocessing. The challenges are first to separate the graphite from the TRISO particles and then to remove the hard SiC coatings of 10,000 or so TRISO particles in each pebble to obtain, at most, a few grams of useful material. All of these factors make reprocessing, particularly on a small scale, more difficult because of the high volumes where uranium and plutonium are effectively impurities in the SNF, a very low-grade fissile ore.

The previous considerations raise the question: Is high-burnup pebble bed SNF a credible source of weapons-usable materials? A technical case can be made for the termination of weapons safeguards for these high-burnup fuels based on a proliferator choosing natural uranium as a preferred starting material. If weapons safeguards are not terminated, then SNF storage, transport, and disposal should be in in a different safeguards category relative to other types of SNF with relaxed safeguards requirements. This leads to several recommendations for future work to better quantify safeguards requirements relative to other materials.

V.A. Indirect Proliferation from Recycled Uranium

A study should be undertaken comparing the production of high-enriched uranium as would be used in a nuclear weapon from natural and depleted uranium versus pebble bed recycled uranium from SNF with burnups up to 200,000 MWd/tonne. The IAEA requires safeguards of more than 10 tons of natural uranium and more than 20 tons of depleted uranium.^[23] Such a study could provide more definitive answers on when natural or depleted uranium is a preferred route to weapons materials versus this recycled uranium.

Such a study should include the separative work units required for different feedstocks using a small enrichment plant like a proliferator would use. With a large-capacity commercial enrichment plant, one can dilute down the challenge by co-enriching with massive quantities of natural uranium. However, any large commercial plant would be under safeguards, and it would be easy to detect any significant high-burnup uranium in an enrichment cascade from the radiation signature and in any commercial enriched uranium from such a plant. All of the uranium isotopes must be included: ²³²U, ²³³U, ²³⁴U, ²³⁵U, ²³⁶U, and ²³⁸U.

The analysis should evaluate the buildup of ²³²U decay products in the uranium enrichment plant with time and the associated radiation fields from the resultant 2.6 MeV, as well as the high alpha levels that may degrade equipment^[25] and cause radiolysis of UF₆. The radiation dose assessment should be extended to include the increasing gamma radiation dose with time from any nuclear weapon in storage containing this uranium.

V.B. Difficulty of Plutonium Weapons from High-Burnup Plutonium Pebble Bed Fuel

A study should be undertaken to determine whether it is credible for a proliferator to build nuclear weapons from high-burnup plutonium (150,000 to 200,000 MWd/tonne) from pebble bed SNF. The historical basis for defining weapons-useable uranium and plutonium is based on physics^[30]: Can you make a weapon. There are major weapons challenges with pebble bed SNF plutonium and its high decay heat loads, high alpha radiation levels, spontaneous neutron fission, and higher critical masses from design to fabrication to storage.

The question is whether the challenges are so large that any proliferator would choose to build a small enrichment plant or a weapons production reactor to obtain materials for nuclear weapons rather than a reprocessing plant to process massive quantities of pebble bed SNF to recover such low-grade plutonium. This analysis would be separate from the very low plutonium concentration in the SNF that may allow for the termination of safeguards in the context of existing IAEA safeguards.

Recent studies have evaluated safeguards for repositories^[31] and the number of waste canisters that must be recovered to obtain a significant quantity of fissile material. In these studies, the TRISO SNF is the least attractive by large margins; large numbers of waste packages must be recovered to obtain a significant quantity of fissile materials. These studies assumed hexagonal fuel assemblies and significantly lower SNF burnup than discussed herein. The IAEA does require safeguards for repositories.^[32] A strong case exists for no safeguards for these SNFs in geological repositories or boreholes because the risks are equal to or less than those for mining natural uranium.

VI. PEBBLE BED REACTOR SAFEGUARDS

There are safeguards concerns associated with the reactor. The first is the diversion of fresh fuel to obtain high-assay low-enriched uranium to further enrich for nuclear weapons. The IAEA defines a significant quantity of uranium enriched to less than 20% as 75 kg of ²³⁵U. The X-Energy and Kairos designs have, respectively, 7 g and 6 g of heavy metal per pebble with corresponding ²³⁵U enrichments of 15.5% and 19.55%. That is, respectively, 1.085 g and 1.173 g of ²³⁵U per pebble. To obtain a significant quantity of uranium, one must respectively divert 69 100 pebbles and 63 900 pebbles. This is diversion by the truck load.

Second, a nation state can fabricate pebbles made of natural or depleted uranium and irradiate them for short times in a pebble bed reactor for the purpose of producing plutonium—not the SNF in storage. Alternatively, a nation state could irradiate pebbles containing thorium for the production of ²³³U. However, the volumes of fuel that must be irradiated and processed to recover a significant quantity of plutonium or ²³³U are one to two orders of magnitude larger than other types of reactors.

In terms of the production of weapons-usable materials, the safeguards priority is detecting pebbles with depleted uranium or thorium (easily identified by the 2.6-MeV gamma ray from the ²³²U decay chain), not fully burnt fuel pebbles. This is simplified by one other characteristic of pebble bed reactors. With continuous refueling, there is very little excess nuclear reactivity in the core. If a significant number of depleted uranium pebbles are sent through the reactor core, there is the need for much more fuel to maintain reactor power.^[10,14] By monitoring reactor power and fueling loading, the insertion of depleted uranium or other breeder pebbles can be detected.

Additionally when sent to storage, separate storage containers will be needed for these special pebbles, which can also be monitored at the plant site. Should a proliferator choose to use single-pass pebbles of ²³⁸U to obtain a significant quantity (8 kg) of ²³⁹Pu, the proliferator would have to collect 20,000 pebbles from a Xe-100 pebble bed reactor. A safeguards system that monitors reactor power and fuel loading during operations would be sufficient to detect such activities. In addition, monitoring the discharged pebbles in the spent fuel storage room to see if they are diverting certain pebbles to special casks would be relatively easy to do. Clearly, it would be a detectable and observable amount. Separately, the added fresh fuel consumption would be easily detectable.

Separate from weapons proliferation is the potential to use pebbles as a dirty bomb; that is, use conventional explosives to disperse the radioactivity in the pebbles. Limited work^[13] suggests that the consequences would be low. The TRISO fuel particles are embedded in a graphite matrix where the explosive shock waves would break up the graphite, but not break up the TRISO particle. Instead the fuel particles would be dispersed over a short distance. We are not aware of any full-scale tests to confirm this assessment. The risk of use as a radiological weapon may set the minimum safeguards requirements.

VII. ACCOUNTABILITY SAFEGUARDS REQUIREMENTS

Statistical sampling of pebbles to determine fissile quantities may provide better safeguards measurements than attempting precision measurement of the fissile content of each pebble. The detection of pebbles with depleted uranium (rather than enriched uranium) would likely be the highest safeguards priority, not high-burnup pebbles. The best safeguards system may be counting the number of pebbles going to storage, satisfying the material accountability objective monitored by a physical security system and random sampling of pebbles for isotopic content. This approach, with monitoring core power levels looking for the use of depleted pebbles and tracking offsite shipments, is a practical and realizable safeguards system for this type of fuel. There is ongoing work on the development of such methods.^[33]

The assay of pebble SNF in storage is relatively easy to measure versus other types of SNF^[11,12] because of two specific characteristics of this fuel. The burnup of each pebble is about the same, resulting in nearly uniform radiation fields in a storage tank filled with pebbles. Second, the low-density carbon with highly dispersed heavy metal content does not shield gamma rays as effectively as metal clad fuel. This combination of characteristics makes it much more difficult for a nation state to substitute dummy pebbles into a SNF storage container while diverting pebbles for fissile fuel recovery or to dump low-burnup pebbles into SNF storage for later recovery (see Sec. VI). Channels can be left in the tank for calibrated gamma probes to measure radiation versus height. Alternatively, with two channel systems, one channel can have a neutron source while the other channel a sensor to measure burnup. Nonuniformity of the signal would raise a flag.

VIII. SECURITY

The security requirements for SNF are to assure it is not used as a source for the dispersal of radioactivity to the general environment. The limited studies indicate significantly lower risks relative to other types of SNF.^[13] This follows from two fuel characteristics. First, the graphite provides a high-volume protective matrix. Second, the very small TRISO particles provide sealed containers for the fission products even if highly energetic forces destroy the individual pebbles. The same characteristics that enable this fuel to withstand severe accidents and make it very difficult to reprocess provide major barriers for the dispersal of radioactivity beyond a few tens of meters.

IX. CONCLUSIONS

High-burnup TRISO pebble bed SNF using enriched uranium has low safeguards and security risks relative to other types of reactor SNF. Historically studies have only examined lower-burnup pebble bed fuels under 100 000 MWd/tonne, not thermal neutron–spectrum SNF with burnups between 150 000 to 200 000 MWd/tonne. The isotopic and physical characteristics of these high-burnup SNFs may enable the termination or relaxation of weapons safeguards, or the creation of a separate category for these SNFs with reduced weapons safeguards requirements for storage, transport, and disposal.

Safeguards can be terminated on this SNF in a closed repository. This is the only fuel today with such characteristics, and is potentially a major advantage in terms of the global deployment of nuclear energy. The accountability requirements for high-burnup pebble bed fuels can be reduced to simple monitoring of the number of pebbles discharged into SNF canisters and the relative activity of the radiation levels in the SNF canisters using assay technologies.

Safeguards are required for fresh fuel and partly burned low-burnup fuel. Safeguards are also required to detect the misuse of the reactor to produce weapons-useable materials, such as pebbles with depleted uranium. Safeguards methods should be modified to address the specific differences in pebble bed reactors as suggested in this paper.

Disclosure Statement

No potential conflict of interest was reported by the authors.