1. Radiation effects in spent nuclear fuel
Dr. Claire Corkhill
University of Sheffield
IAEA Workshop on radiation effects in nuclear waste forms and their consequences for storage and disposal, Trieste, September 2016

2. Outline Introduction to spent fuel Out-of-reactor radiation damage
Spent fuel microstructure and chemistry
High burn up structure
Out-of-reactor radiation damage
Factors leading to swelling
Defect accumulation
Role of He
Effects of radiation on spent fuel disposal
Spent fuel dissolution
Relationship between microstructure & dissolution
Role of radiolysis on dissolution

3. Introduction
10,000 metric tonnes heavy metal (MTHM) spent fuel (SF) produced per year
Cumulative inventory of 300,000 MTHM stored in pools or dry casks
Destined for long-term storage and geological disposal
Spent fuel pool storage, UK
Spent fuel dry cask storage, USA
Source:
Source:

4. Spent fuel microstructure
Comprised of > 95% UO2
Heterogeneous chemistry
Heterogeneous microstructure
Final composition depends on:
-initial fuel type
Chemical composition
Enrichment of 235U
Neutron energy spectrum
Burn-up
- Spent fuel comprises > 95 % UO2
- The remainder is composed of fission products, transuranium elements and activation products. These are in a number of forms:
fission product gases, such as Xe and Kr, which occur as finely dispersed bubbles in the fuel grains
Metallic fission products (Mo, Tc, Ru, Rh and Pd) that occur as immiscible nm to um sized precipitates (called epsilon particles)
Fission products that occur as oxide precipitates Rb, Cs, Ba and Zr
Fission products that form solid solutions with the UO2 fuel, such as Sr, Zr, Nb and the rare earth elements
Transuranium elements that substitute for U in the UO2.
- Because of a steep thermal gradient within the fuel pellets, (1700C in the centre, decreasing to 400C at the rim), the fuel pellet is not homogeneous; this means that volatile elements (Cs, I) migrate to grain boundaries.
The rim also has a higher burn up than the centre, which lead to enhanced concentrations of 239Pu, and increase in fuel porosity and a reduction in the grain size (more on that in the next slide)
- Final composition of the fuel depends on the initial fuel type, chemical composition and enrichment of 235U, the neutron energy spectrum and the “burn up” (or amount of fission). Typical burn ups are 35 – 45 MWd/kg U, but higher burn ups are also achieved (do a little more looking into this).
Bruno and Ewing, (2006)

5. Fission gas bubbles in spent fuel
23 GWd/t
44 GWd/t
83 GWd/t
Kashibe et al. (1993)
Fission gases coalesce as bubbles (2 – 10 nm) within fuel grains.
Transported to grain boundaries during fission.
Fission gases, such as Xe and Kr, coalesce as bubbles within the fuel grains.
These are typically between 2 and 100 nm in size.
Gas is transported to grain boundaries by simple atomic diffusion and gas mobility at the high temperatures experienced during fission.
When FG achieve a high enough concentration, and the local temperature exceeds 800C, they can form intergranular bubbles ranging in size from 0.1 to 1um.
The density and size depends on the burn-up
At high concentration and > 800˚C, intergranular bubbles (0.1 – 1.0 µm) are formed.
20 µm
2 µm
B = bubbles; T = tunnels
Shoesmith (2007)

6. High burn-up structure (HBS)
Burn-up is 2 – 3 times higher at rim
Results in grain subdivision by a factor of 104 (0.1 – 0.3 µm grains)
Centre structure
Rim structure: HBS
The burn-up at the rim of the UO2 pellet can be 2 – 3 times higher than the average pellet burn-up
Results in what is known as the “high burn up structure”
Usually in the outer few ums of the pellet.
In this region, the grains subdivide by a factor of 10,000 into sub-micron sized grains (0.1 to 0.3um).
As a result, there is a high proportion of defects (which we will come onto next) and a high proportion of micron-sized intergraular pores, which trap fission gases.
High porosity traps fission gases
Rondinella & Wiss (2010)
Cladding
Rondinella & Wiss (2010)

7. Spent fuel activity
Radioactivity of uranium ore
Total for spent fuel
Fission & activation products
Actinides
& daughters
After a typical burn up, the radioactivity has increased by a factor of 1 million (1017 Bq/ MTHM fuel).
With time, the total radioactivity drops quickly, so that after 10,000 years, it is 0.01 percent of the activity one month after the removal from the reactor
After several hundred thousands of years, the total radioactivity equals the radioactivity of the original uranium ore from which the fuel was made
Alpha decay is the dominant form of radiation at long time scales, so we really need to understand how it will affect the fuel integrity over long time scales
Today we’ll focus on processes expected to be important before and after 10,000 years, which is how long the engineered barrier of the geological disposal facility is expected to keep water out (essentially dry and wet conditions).
After decay of β- and γ- products, α-decay from actinides dominates for long time periods

8. Problem: Volume swelling
Self-irradiation α-damage
He gas accumulation
Volume swelling is the number one enemy, since it can impact on the fuel cladding
In extreme cases, this can lead to clad rupture
This is not good for maintaining containment (particularly of fission gases) during dry storage and geological disposal (wet conditions).
The two main culprits are: i) self-irradiation and alpha damage; and ii) He gas accumulation.
20 µm
30 yr old PuO2; Ferry et al. (2006)

9. Self irradiation of spent fuel
Wiss et al. (2014)
α-irradiation of UO2 results in the formation of Frenkel defect pairs.
Tend to migrate to “sinks” or cluster and form dislocation loops.
Thermal recovery of lattice defects:
Will not occur in storage conditions
α-damage
Fission damage
Stage I: oxygen interstitial migration
Stage II: uranium interstitial migration
Stage III (s.c. UO2 only): He trapped in vacancy sites
Vacancy
Interstitial atom
Frenkel defect pair
Displacement of U and O atoms in spent fuel results in the formation of isolated frenkel defect pairs along the path of the alpha particle.
Extended defect clusters are only seen at very high temperatures or burn-ups so are not generally considered as important in “storage and disposal” conditions.
Defects can be removed from the lattice by recombination, clustering or annihilation.
They can: migrate to grain boundaries, surfaces or fission gas bubbles; or cluster and form dislocation loops.
Removal of defects by thermal annealing can tell us a bit more about the type of defects present
There are 3 isochronal recovery stages in single crystal UO2. The first is interpreted as corresponding to oxygen interstitial migration; the second is thought to be associated with migration of the uranium vacancy. The third in only seen in single crystal UO2 (not spent fuel) – it might relate to He trapped in vacancy sites.
Should note that thermal recovery occurs at temperatures higher than expected in storage and disposal
After Weber (1983)

10. α-damage and swelling ∆𝑎 𝑎 0 =𝐴(1− 𝑒 −𝐵 𝐷 𝛼 )
Annealing rate constant
Defect ingrowth model (Weber, 1981):
𝜌 𝐷 = 𝑁 𝑑 𝐾 𝛼 𝐵 (1− 𝑒 −𝐵 𝐷 𝛼 )
Frenkel defect concentration
Average number of Frenkel defects per α deposited
Local rate of change of Nd (constant)
α-dose
∆𝑎 𝑎 0 =𝐴(1− 𝑒 −𝐵 𝐷 𝛼 )
Relationship between lattice parameter & defect ingrowth:
Lattice parameter change
Saturation value of a/a0
Annealing rate constant
α-dose
Damage ingrowth model postulated by Weber (1981) (at room temperature):
isolated Frenkel defect pairs are produced at a constant rate by the alpha particles
defects are annealed by recombination at a rate proportional to the defect concentration.
Describe equations.
The lattice parameter is proportional to the lattice volume, so an increase in Δa/a0 = an increase in volume

11. α-damage and volume increase
α-recoil damage
fission damage
α-damage and volume increase
Δa/a0
After Weber (1983)
UO2 lattice
a = 5.468(1)Å
Increasing dose
Alpha damage results in a change in the lattice parameter. Since the lattice parameter is proportional to lattice volume, an increase in Δa/a0 results in a lattice volume increase.
As can be seen from this graph, the lattice parameter increases with increasing dose (and damage).
The lower defect concentration in the case of alpha recoil damage may be associated with enhanced defect annealing during irradiation caused by isolated thermal increases associated with the recoil nuclei.
For fission damage, the large thermal spikes associated with the fission fragments greatly enhance defect recombination and defect mobility, effectively reducing the saturation level of the defect concentration.
The relationship between the lattice parameter and the number of vacancies is still under investigation.
In spent fuel, saturation values of between 0.3 and 0.6% have been reported
Corresponds to volume increase of < 4%
Wiss et al. (2014)

12. Accumulation of He(g) He solubility in UO2 = ~0.02 at%
Wiss et al. (2014)
He is created in spent fuel grains by alpha decay of actinides
The rate at which alpha dose, and hence He production, changes as a function of time is shown in the graph (for different types of fuels – higher BU = more He)
The solubility of He in UO2 is very low, in the range of 0.01 to 0.02 at%. This concentration can be exceeded within a few decades
It also has very slow diffusion rates at room temperature (~10-25 m2 s-1) – after 10,000 years one He atom in spent fuel will have a diffusion path of 0.15um!
These factors mean that He accumulates within grains, forming bubbles. They can also become trapped in FG bubbles.
What does this mean?
At very high alpha doses, He bubbles may lead to volume swelling, which may compromise the integrity of the fuel cladding (bad for storage and disposal)
At the temperatures of geological disposal, dissolution and migration of bubbles will be very slow
Next slide on thermal recovery?
He solubility in UO2 = ~0.02 at%
He diffusion in UO2 (at 25˚C) = ~10-25 m2 s-1
→ in 10,000 yrs, He diffusion path is 0.15 μm
He accumulates within grains as bubbles or trapped in fission gas bubbles
Wiss et al. (2014)

13. Mechanical effects of He(g) accumulation on spent fuel
He bubbles = volume increase
Increasing He concentration linked to enhanced stress and strain (interstitial defects and He clusters)
He bubble accumulation may result in microcracks
Increase in hardness and lowering of toughness may correspond to brittle fracture
What does this mean? Well, this doesn’t seem to be very well constrained at present. Some ideas are:
At very high alpha doses, He bubbles may lead to volume swelling, which may compromise the integrity of the fuel cladding (bad for storage and disposal)
Strain and stress have been shown to increase as a function of He concentration, it is thought that the formation interstitial defects and He clusters may cause this.
It has been postulated that, thanks to the low diffusion coefficient, a build up of He bubbles could lead to microcrack formation
The vicker’s hardness of spent fuel has been shown to increase with cumulative dose. This graph compares the hardness of fuel just taken from the reactor, with 30 yr old fuel. The maximum hardness is reached at 0.1dpa.
If the increase in hardness can be correlated to lowering of toughness, it could be considered that brittle fracture could occur in the denser parts of the fuel
Wiss et al. (2014)

14. Mechanical changes in the HBS
Distance from pellet edge
Centre
Rim
Lattice parameter
Porosity
Local burn-up
Lattice parameter (pm)
548.2
547.5
Porosity increases in HBS
Lattice parameter decreases due to grain subdivision
HBS is softer and tougher than pellet centre → beneficial to reduce stress on cladding
Softened zone
Toughened zone
Recall from earlier that FG tend to migrate to the rim into the high burn up structure – this leads to quite significant swelling in this region.
However, the grain subdivision process is accompanied by a decrease in the lattice parameter, so these effects are quite complicated.
The figure shows the change in lattice parameter as a function of distance from the pellet centre. A sharp increase in porosity is observed corresponding to the HBS region. The lattice parameter increases slightly in a postulated “intermediate region” prior to a sharp decrease in lattice parameter in the rim.
Micro-hardness measurements show that the HBS is softer and tougher than the pellet centre. This is beneficial to ensure tolerable mechanical stresses on the cladding.
Rondinella & Wiss (2010)
Spino et al. (2003)

15. Summary: radiation-induced changes during fuel storage
α-damage is the dominant effect modifying the structure of spent fuel during long-term storage
The lattice parameter can show a substantial increase, leading to volume swelling, but the fluorite structure is preserved
Intragranular He bubbles are formed, which may lead to volume swelling and loss of mechanical stability
The temperature of fuel storage and disposal is not sufficient for annealing of defects, nor is it sufficient to lead to He diffusion
As a result of these processes, the fuel is likely to experience high stress levels
The fluorite structure of UO2 is preserved during a-irradiation, but the lattice parameter can show a substantial increase.
Konings et al. (2015)

16. Spent fuel disposal scenario
Swedish KBS-3V concept
Both Sweden and Finland have advanced plans for design, construction and operation of geological repository for spent fuel.
Both are planning to use the KBS-3V concept:
After a cooling period of at least 30 years, the fuel elements are placed in cast iron canisters with an outer layer of copper.
The canisters are sealed and placed in vertical holes drilled in tunnels around 500 m below ground, and surrounded by compacted bentonite clay.
These so-called engineered barriers are applied to prevent or delay the contact of groundwater with the fuel.
This is not expected to take place before 10,000 years of storage – which if we recall, is when alpha radiation is expected to dominate.

17. Dissolution of spent fuel
Instant release fraction
Johnson et al. (2012)
Fission gas bubble release
Oxidation of the fuel matrix
UO2 → UO2+x
U4+ → U6+
Gap release (volatile FPs & fission gas)
U4+
U6+ e-
H2O, HCO3-
UO22+(aq)
U6+ secondary phase precipitation
Cladding
Two stages of spent fuel dissolution:
1. Instant release fraction
- fission gases
- volatile fission products
2. Slow matrix dissolution
- UO2 and actinides
Spent fuel dissolution occurs in two general stages: i) instant release fraction and ii) UO2 dissolution
The instant release fraction is the inventory of the fuel that is released rapidly when the metal waste package and fuel cladding are first breached.
The radionuclides released in this fraction are the fission gases, such as Xe and Kr, and volatile elements such as I, Cs and Cl – that have migrated to the HBS and the gap between the cladding and the fuel.
Because the concentration of fission products increases with increasing burn-up, so too does the instant release fraction, as shown here for UO2 fuel.
The dissolution of UO2 is a much slower process, which involves:
i) oxidation of U4+ to U6+ and the formation of higher oxide structures on the fuel surface and at grain boundaries;
ii) bulk dissolution of UO2 and release of actinides substituted into the UO2 structure;
iii) formation of secondary alteration products, such as coffinite (USiO4) under reducing conditions, or uranyl (UO22+) complexes under oxidising conditions.

18. Dissolution of spent fuel: redox conditions
Redox conditions play a very important role in the dissolution of spent fuel.
Oxidising conditions result in fuel dissolution and production of mobile species
Reducing conditions prevent dissolution
It is clear that redox conditions play an important role in the dissolution of spent fuel. I’d just like to expand on that a little as it is important later on.
This pourbaix diagram shows how redox influences uranium chemistry.
Under oxic conditions, UO2 , with an average oxidation state of U4+, will oxidise to U6+. When this occurs, U6+ compounds such as UO22+ will be formed, which is highly soluble and mobile. At high pH, uranyl carbonate complexes can also form, which are also highly mobile.
So, in the repository, we would prefer reducing conditions, which will keep UO2 as a stable, solid phase.
Anything that could act as potential source of oxidants is NOT good.
Oxidants are bad for UO2 fuel durability!
Ewing (2015)

19. e-(aq), H˙, OH˙, OH-, H3O+, H2, H2O2, HO2
Radiolysis of water
Ionising radiation
H2O
excitation
ionisation
H2O*
H2O + e-
H˙ + OH˙
2OH˙ + H2
OH˙ + H3O+
OH˙ + OH- + H2
e-(aq)
e-(aq), H˙, OH˙, OH-, H3O+, H2, H2O2, HO2
H2+O(1D)
H2O+
H- + OH˙
Time (s)
10-15 s
10-12 s
10-6 s
0 s
Some notes here about this…

20. Radiolysis effects on dissolution
α particles travel ~ 40 μm in water
Absorption of radiation in liquid results in radiolysis, forming:
Fission gas bubble release
Oxidation of the fuel matrix
UO2 → UO2+x
U4+ → U6+
Gap release (volatile FPs & fission gas)
U4+
U6+ e-
H2O, HCO3-
UO22+(aq)
U6+ secondary phase precipitation α
H2O
H2O2, OH˙, HO2˙, H2, O2
Fuel oxidation / reduction by radiolytic oxidants
Oxidants
Reductants
HO˙
e-(aq)
HO2˙ H˙
H2O2 H2
The redox conditions at the surface of the fuel will be affected by these species
Very little is known about radiation-induced processes in spent fuel
Alpha particles travel a distance of around 40 um in water
Absorption of radiation in liquid water leads to the formation of free radical and ionic species.
Radiolysis of water produces reductants and oxidants. These are the hydroxyl radical (HO˙), the hydrogen atom (H˙), the hydroperoxyl atom (HO2˙), the hydrated electron (e-(aq)), hydrogen (H2) and hydrogen peroxide (H2O2).
The redox conditions at the surface of the fuel will be affected by these products, therefore, so too will the dissolution reaction.
Very little is really known about radiation-induced processes at the interface between spent nuclear fuel and water.

21. Radiolysis effects on dissolution
Surface processes in radiation-induced dissolution
UO2(CO3)n(2-2n)(aq) + nH+
nHCO3˙
OH˙ + HCO3 → H2O + HCO3˙
Competing reaction for OH˙ ε
2H+ H2
2e-
2OH˙
H2O2
H2O + (1/2)O2
U6+
U4+
OH˙ + H2 → H2O + H˙
H˙ + H2O2 → H2O + OH˙
Consumption of H2O2 by H2
This diagram summaries the most important surface processes in radiation-induced dissolution.
Some of the important equations are also given.
I’ll describe the processes shown
Firstly, the presence of noble metals, or epsilon particles (e.g. Pd, Mo, Ru fission products), catalyses the reduction of uranium 6+ to uranium 4+. This reaction can completely suppress the oxidative dissolution of UO2.
The second set of reactions are those governed by the oxidant, hydrogen peroxide. H2O2 can react with the surface in two ways: i) it can either directly oxidise the surface, or undergo catalytic decomposition, whereby further oxidising species are produced, which then go on to oxidise the surface.
Under conditions of a geological disposal facility, radiolytically produced hydrogen peroxide is predicted to have the greatest oxidising potential of all the radiolysis products.
BUT, the reaction of hydroxyl radicals with hydrogen produces a hydrogen atom that, in turn, reacts with hydrogen peroxide, driving a chain reaction that consumes the oxidant hydrogen peroxide.
Finally, other groundwater components, such as S, Fe and carbonate can influence the process of radiation-induced oxidative dissolution of UO2. For example, bicarbonate and carbonate ions can be converted to a CO3-˙ radical upon reaction with the hydroxyl radical. Because the hydroxyl radicals are used up in this reaction, they are not consuming hydrogen peroxide so much – so there will be more oxidant (H2O2) in the system.
H2O2 either: directly oxidises surface or produces other oxidising species upon catalytic decomposition. Both enhance oxidative dissolution
Reaction of OH˙ with H2 begins a chain reaction that consumes H2O2, reducing UO2 dissolution
ε-particles catalyse the reduction of U6+ to U4+, supressing oxidative dissolution
Other groundwater components compete with H2 for OH˙ (e.g. CO3-) supressing decomposition of H2O2.

22. Radiolysis effects on dissolution
Dissolution in the presence of H2O2
Doping decreases redox reactivity
Make sure to mention difference between Westinhouse fuel and Simfuel.
Pehrman et al. (2012)
ε-particle containing pellets have the lowest U release
Dissolution in the presence of H2O2 also has a strong dependence on dopant type → due to influence of the redox reactivity of the pellet upon doping.

23. Influence of microstructure on dissolution
Phase
Lattice volume (Å3)
Lattice strain
CeO1.82
43.76
0.092 ± 0.01
CeO1.98
39.61
0.034 ± 0.01
Previously, we discussed that the lattice parameter increases with increasing alpha dose.
The effect of this on the dissolution of spent fuel is not precisely known, but investigations with a CeO2 analogue for UO2 suggest that it might be significant.
It was found that change in lattice volume and strain when you take reduced CeO2-x and oxidise it to CeO2 – in this case, CeO1.75 to CeO1.82, also has a significant influence on the microstructure. These images show that the change in lattice volume and strain caused particles of CeO1.75 to break apart into individual grains upon dissolution.
Breaking the material apart in this way caused the dissolution rate in increase by a factor of 15.
It is currently not known how changes in lattice parameter will affect the dissolution of UO2 or spent fuel.
Change in lattice volume may have significant consequences for dissolution of UO2
Corkhill et al. (2016)

24. Influence of grain boundaries (GB) on dissolution
GBs are preferentially dissolved
Eventually, grain decohesion can occur
Crystallographic orientation of GBs influences dissolution rate
Effect on spent fuel unknown
CeO2
60˚
(025)/(001)
(001)/(356)
Misorientation angle (± 0.01)
36.01°
59.84°
GB retreat rate (μm d-1) (± 0.001)
0.014
0.017
36˚
Corkhill et al. (2014)

25. Summary: radiation-induced changes during spent fuel disposal
Effect of α-damage related changes to lattice on dissolution not well understood in spent fuel, but seems to enhance dissolution rate in fluorite analogues
Radiolysis influences the redox conditions of the spent fuel surface
Radiolytic reducing species expected to supress oxidative dissolution
Grain boundaries are sites of preferential dissolution → related to instant release fraction
Ewing (2015)

26. References Bruno & Ewing, Elements, 2, 343 (2006)
Corkhill et al. Applied Materials and Interfaces, 6, (2014)
Corkhill et al. Applied Materials and Interfaces, 8, (2016)
Ewing, Nature Materials, 14, 252 (2015)
Johnson et al. Journal of Nuclear Materials, 420, 54 (2012)
Kashibe et al. Journal Nuclear Materials, 206, 22 (1993)
Konings et al. Nature Materials, 14, 247 (2015)
Pehrman et al. Journal of Nuclear Materials, 430, 6 (2012)
Rondinella & Wiss. Materials Today, 13, 24 (2010)
Shoesmith, NWMO Report, TR (2007)
Spino et al. Journal of Nuclear Materials, 322, 204 (2003)
Weber, Journal of Nuclear Materials, 98, 206 (1981)
Weber, Journal of Nuclear Materials, 114, 213 (1983)
Wiss et al. Journal of Nuclear Materials, 451, 198 (2014)