Atomic Energy, Vol. 102, No. 2, 2007

IMPROVEMENT OF THE OPERATING CHARACTERISTICS OF VVÉR OXIDE FUEL

P. N. Alekseev, A. V. D’yakov, A. S. Kolokol, A. A. Proshkin, and A. L. Shimkevich

UDC 621.039.54

Investigating the microstructure and atomic dynamics of solid solutions and the physicochemical processes occurring in them will make it possible to adjust the properties of the solutions in steps, according to prescribed indicators, by using alloying additives. The concept of designing an oxide fuel may be promising for the development of a new generation of nuclear reactors. Alloying uranium dioxide with an additive having a different valence, consisting of the glass-forming elements carbon and silicon, can double its thermal conductivity and can substantially increase the concentration of equilibrium vacancies in the crystal lattice of the solid solution. This will increase the incubation period for swelling above 20 MW·days/kg and, possibly, decrease the emission of gaseous fission products from the fuel.

Reliable operation of, first and foremost, all members of the core (fuel assemblies and fuel elements) during normal operation and during an accident determines the cost-effectiveness and safety of nuclear power plants. In this connection, the state of the oxide fuel in VVÉR reactors merits analysis in order to improve its operating characteristics by, for example, modifying the structure and composition of the fuel. Domestic Results. More than 5·106 fuel elements have operated in nuclear power plants with VVÉR-1000 reactors 6 and 7·10 in plants with VVÉR-440 reactors [1–4]. Fuel for VVÉR-1000 reactors was manufactured in the form of 7.57 mm in diameter, sintered, uranium dioxide pellets with a 2.2 mm in diameter opening at the center. The pellets were placed in a 9.1 mm in diameter casing made of the alloy É110. The fraction of damaged fuel elements for burnup to 50 MW·days/kg did not exceed 1.2·10–5. The operation of fuel assemblies and multifactor tests performed on ~7000 fuel elements with burnup 18–63 MW·days/kg (eight VVÉR-440 and 20 VVÉR-1000 fuel assemblies [2, 3]) showed how the characteristics of oxide fuel can be improved. It is evident from Fig. 1 that the emission of gaseous fission products from VVÉR-440 and -1000 oxide fuel does not exceed 3% of their total amount with burnup B ≤ 50 MW·days/kg, reaches 5% for B ~ 55 MW·days/kg, and rapidly increases as fuel is burned up. At the same time, our estimates show that the pressure of the gaseous fission products in VVÉR-440 fuel elements did not exceed 3 MPa with initial helium pressure 0.6 ± 0.1 MPa in the elements and 6 MPa in VVÉR-1000 fuel elements with helium pressure 2.2 ± 0.2 MPa. Since the initial coolant pressure is higher than the gas pressure in a fuel element, the cladding is compressed until it comes into contact with the fuel pellet. It is evident in Fig. 2 that this happens at burnup B ~ 55 MW·days/kg. Subsequently, only the plasticity of the cladding allows the fuel element to remain operable, since the oxide fuel swells. According to the data in Fig. 3 this swelling is linear: ∆V/V = 0.07B – 1.47 Russian Science Center Kurchatov Institute. Translated from Atomnaya Énergiya, Vol. 102, No. 2, pp. 109–113, February, 2007. Original article submitted January 11, 2006. 1063-4258/07/10202-0129 ©2007 Springer Science+Business Media, Inc.

129

G, % 5 4 3 2 1 0 40

42

44

46

48

50

52

54 56 58 B, MW·days/kg

Fig. 1. Gas emission versus fuel burnup for VVÉR-1000 (b) and -440 (a, A, c) reactors.

∆, µm 120

80

40 0 30

35

40

45

50

55 60 65 B, MW·days/kg

Fig. 2. Width of the gap between the fuel and fuel-element cladding versus burnup for VVÉR-1000 (A) and -440 (C) reactors.

with an incubation period ~21 MW·days/kg. On the whole, the domestic oxide fuel with burnup B ≥ 50 MW·days/kg is characterized by the following features [1]: • the diameter of the opening at the center of a fuel pellet remains essentially unchanged; six to eight radial cracks emanate from it and become overgrown at the periphery of the pellet; • the same average grain size as in the initial pellet; • the defective layer at the periphery of the pellet is up to150 µm wide and possesses open porosity and an elevated plutonium content. Foreign Data. In the opinion of foreign specialists, the problems of fuel-assembly operation which occur because of high burnup (B ≥ 65 MW·days/kg) are due to the following: • elevated emission of gaseous fission products from the fuel and a substantial increase of the gas pressure in a fuel element; • appreciable decrease of the thermal conductivity of the fuel; • prolonged fuel–cladding contact resulting in mechanical and corrosion action of the fuel on the cladding; • degradation of the mechanical properties of the fuel-element cladding; • formation of a defective layer and its effect on the properties of the fuel. The thermal conductivity of uranium dioxide fuel decreases with burnup [5]. Figure 4 shows that in the range 900 ≤ T ≤ 1600 K the thermal conductivity of spent nuclear fuel with burnup 60 MW·days/kg h.a. is almost 1.5 times lower 130

S∆V/V, % 4 3 2 1 0 20

25

30

35

40

45

50

55 60 65 B, MW·days/kg

Fig. 3. Fuel swelling versus burnup for VVÉR-1000 (a) and -440 (c) reactors.

λ, W/(m·K) 5 1 4

2

3 2

3

4 5

500

1000

1500

2000

2500

3000 T, K

Fig. 4. Thermal conductivity of oxide fuel versus temperature with burnup 0 (1), 20 (2), 40 (3), and 60 MW·days/kg (4); 5) new data.

than that of the initial uranium dioxide. The dependence of the thermal conductivity λ(T, B) of the fuel on the temperature T and burnup B can be represented in the form λ(T, B) = (7.6 – 5.5·10–3T + 1.4·10–6T 2)/(1 + 7.91B/T) for 800 ≤ T ≤ 2000 K and 0 ≤ B ≤ 60 MW·days/kg. Figure 5 illustrates the gas release of spent fuel as a function of burnup [6]. It is evident that the emission of gaseous fission products does not exceed 7% of the total amount for burnup B ≤ 50 MW·days/kg, ~12% for 70 MW·days/kg, and 30% for ~100 MW·days/kg. According to these data, the emission of gaseous fission products can be represented in the form F = 0.65 – 0.095B + 0.315·10–2B 2 in the range 10 ≤ B ≤ 80 MW·days/kg. This relation can be used to estimate the emission of gaseous fission products from VVÉR fuel elements with fuel burnup above 60 MW·days/kg. The behavior of the gaseous fission products in spent fuel exhibits the following specific characteristics [6–8]: • their emission at the center of a fuel pellet, where the temperature is estimated to be 1273 K, reaches 90% and remains unchanged with increasing average burnup; 131

γ, % 30

20

10

0 20

40

60

80 100 B, MW·days/kg

Fig. 5. Emission of fission products from oxide fuel versus burnup.

• there is no appreciable emission of gaseous fission products in the outer ~100 µm thick defective layer even with average fuel burnup ~60 MW·days/kg. The outer defective layer of a fuel pellet is characterized by burnup ~120 MW·days/kg, which is 2.5 times higher than the average burnup in a pellet, by small grain size (0.1–0.2 µm), which 10 times smaller than the initial size, and by a high concentration of pores with diameter 0.5–1 µm, fission products, and plutonium. Apparently, this peripheral (low-temperature) zone of a fuel pellet is, under irradiation, a sink for impurities, including vacancies produced by the fission products. The formation of the external defective layer and the fuel swelling are characterized by a prolonged incubation period (>20 MW·days/kg) [9], which is seen in Figs. 3 and 6. Improvement of Oxide Fuel. To increase the technical and economic performance of nuclear power plants, foreign engineers are striving to improve the operating characteristics of uranium dioxide and mixed oxide fuel [7]. For example, the General Electric Company (USA) is developing plastic uranium dioxide pellets with an aluminosilicate additive to decrease the fuel–cladding interaction. The possibility of increasing the thermal conductivity of oxide fuel is being studied. For example, before sintering the fuel can be permeated with a heat-conducting suspension consisting of a mixture of silicon carbide and polymer filler [7]. The French Atomic Energy Commission is financing a search for a mixed oxide fuel with burnup to 120 MW·days/kg and operating time of the fuel elements up to 10 yr. Reactor tests are being conducted on fuel elements with chromium oxide or other metals added and larger oxide-fuel grain sizes. This makes it possible to decrease the emission of gaseous fission products and the swelling rate of the fuel. The possibility of developing an oxide ceramic with high plasticity at working temperature up to deep burnup of the fuel is being examined. Vacancy Doping. The dynamics of radiation point defects must be controlled in order to stabilize the fluorite structure of oxide fuel under reactor conditions and to suppress the segregation of impurities from it under irradiation [10], since the fission products generate in the crystal lattice vacancies whose local concentration, as a rule, is higher than their solubility as impurity particles. Consequently, the vacancies and together with them fission products flow onto the grain boundaries and into micropores and dislocation loops. It is easy to increase the solubility of vacancies in oxide fuel [11]. For this, it is sufficient to dope uranium dioxide with the oxides of metals, such as Ca, Cr, or Y, whose valence is lower than that of uranium. The choice of dopant must be based on the formation of a solid solution with a prescribed region of homogeneity and its thermodynamic stability in the working temperature range. Then, the oxygen sublattice of the mixed oxide (UO2)1–x(Me2O3)x with the structure of fluorite CaF2 acquires equilibrium vacancies whose atomic fraction is ~0.5x, where the molar fraction x of the oxide additive in the solid solution can reach 0.15. Ultimately, approximately 7% of the equilibrium vacant sites in the oxygen sublattice of the fuel with density ~1020 cm–3 exceed many-fold the possible concentration of radiation-induced vacancies. Such vacancy-doped 132

h, µm·102 15

10

5

0 20

40

50

60

70

80

90 100 B, MW·days/kg

Fig. 6. Thickness of the outer defective layer of the oxide fuel versus burnup.

–logPO2 UO2 Energy of free electron 70 0

UO

ηc

30

εc

20

εF

Conduction band

Impurity conduction band

χe

εg

U4O9 UO2.01 0

U3O8

y–U3O 600

1000

a

1400

εv Valence band

T, K

b

Fig. 7. Temperature dependence of –log PO2 for uranium oxides (a) and the arrangement of the electron energy levels in the oxide fuel (b); ηe – electron affinity; χe – electronic work function; εF – Fermi level, which is the boundary between the free and occupied electron energy levels.

oxide fuel becomes susceptible to the accumulation of fission products, and the high density of centers of recombination of radiation defects, actually “frozen” as a result of their weak mobility, can hinder vacancy swelling and segregation of impurities in the entire defective layer. Increasing the Thermal Conductivity. The fluorite phase of nonstoichiometric uranium dioxide (UO2±x) is thermodynamically stable in a wide range of oxygen pressures (Fig. 7a). This corresponds to a band gap of width εg = εc – εv > 2 eV 133

(Fig. 7b). Consequently, phonon modes of the metal sublattice dominate in the thermal conductivity of uranium dioxide and cooperative motion of the oxygen anions dominates at high temperatures. In (UO2)1–x(Me2O3)x, the transition metal ions Me+q, being phonon scatterers, lower the thermal conductivity of the solution, and doping uranium dioxide with a metalloid with different valence (C–4, Si–4), conversely, increases it because the mobility of the anions increases. Thus, when uranium dioxide is doped with glass-forming impurities such as carbon and silicon, it is possible at least to double its thermal conductivity and greatly increase the concentration of equilibrium vacancies in the crystal lattice of the solid solution. This can increase the incubation period of mixed uranium dioxide above 20 MW·days/kg, decrease the rate of swelling of the fuel and, possibly, decrease the emission of gaseous fission products from the fuel. Conclusions. The emission of gaseous fission products from oxide fuel and the thermal conductivity of uranium dioxide were determined as functions of the burnup and temperature of the fuel on the basis of an analysis of domestic and foreign experimental data. For initial density 10.4–10.5 g/cm3, the rate of swelling of the fuel is 0.6–0.7% per 10 MW·days/kg up to burnup ~60 MW·days/kg and 1–1.1% per 10 MW·days/kg for burnup ≥65 MW·days/kg. Doping of uranium dioxide with impurities having a different valence with formation of a solid solution can increase the thermal conductivity of the fuel, decrease the swelling of the fuel, and decrease the emission of gaseous fission products from the fuel. This was performed with the financial support from the Russian Fund for Basic Research (Grant No. 05-08-18066).

REFERENCES 1. 2.

3.

4.

5. 6.

7. 8.

9. 10.

134

V. V. Goncharov, K. P. Dubrovin, P. A. Platonov, et al., “Testing of experimental VVÉR-1000 fuel elements in the MR reactor,” At. Énerg., 63, No. 1, 40–44 (1987). A. V. Smirnov, A. K. Panyushkin, V. V. Rozhkov, et al., “Characteristics of VVÉR fuel after operation in stationary and transient regimes. Prediction of superhigh burnups,” in: International Conference on Atomic Energy at the Threshold of the 21st Century, Élektrostal’, June 8–10, 2000, pp. 45–47. A. Smirnov, V. Smirnov, B. Kanashov, et al., “VVER-1000 and -440 fuel operation experience in ANS meeting light water reactor fuel,” in: Proceedings of the 1994 International Topical Meeting on Light Water Reactor, Fuel Performance, West Palm Beach, Florida, April 17–21, 1994, p. 23. A. Smirnov, V. Smirnov, B. Kanashov, et al., “Behaviour of VVER-440 and VVER-1000 fuel in a burnup range of 20–48 MWd/kg U,” in: Proceedings of the 2nd International Seminar on VVER Reactor Fuel Performance, Modeling, and Experimental Support, Sandanski, Bulgaria, April 21–25, 1997, p. 40. Yu. Bibilashvili, V. Velyukhanov, A. Ioltukhovski, et al., “Operation experience of VVER fuel, including analysis of abnormal conditions,” ibid., p. 164. F. Southeimer, H. Landsktron, and M. Billaux, “A fuel thermal conductivity correlation based on the latest experimental results,” in: Proceedings on Thermal Performance of High Burnup LWR Fuel, Cadarache, France, March 3–6, 1998, p.119. A. Seibold, F. Garzarolli, and R. Manzel, “Verification of high burn up materials behavior,” in: International Conference on Top Fuel-2001, Stockholm, May 27–30, 2001, Rep. No. 4. J. Barner, M. Cunningham, M. Fresley, et al., “Relationship between microstructure and fission gas release in high burn-up UO2 fuel with emphasis on the rim region,” in: International Topical Meeting on LWR Fuel Performance, Avingnon, France, April 21–24, 1991, p. 538. C. Walker, “Assessment of the radical extent and completion of crystallization in high burnup UO2 nuclear fuel by EPMA,” J. Nucl. Mater., 275, 56–62 (1999). P. N. Alekseev, N. N. Ponomarev-Stepnoi, A. A. Proshkin, et al., “On the design of nuclear fuel with a low swelling ratio,” Preprint IAÉ-6259/11 (2003).