Irradiation behaviour of MOX fuel: Results of an EPMA investigation

Irradiation behaviour of MOX fuel: Results of an EPMA investigation

Nuclear Engineering and Design 131 (1991) 1-16 North-Holland 1 Irradiation behaviour of M O X fuel: Results of an E P M A investigation * C.T. W a l...

3MB Sizes 0 Downloads 26 Views

Nuclear Engineering and Design 131 (1991) 1-16 North-Holland

1

Irradiation behaviour of M O X fuel: Results of an E P M A investigation * C.T. W a l k e r a, M. C o q u e r e l l e a, W. Goll b a n d R. M a n z e l b a Commission of the European Communities, Joint Research Centre, European Institute for Transuranium Elements, Postfach 2340, W-7500 Karlsruhe, Germany b Siemens A G - KWU Group, Postfach 3220, W-8520 Erlangen, Germany

Received 25 June 1990

The redistribution of Pu and the behaviour of Xe and Cs were investigated in three sections of irradiated MOX fuel produced by the AUPuC process. Two of the fuels had been irradihted under steady-state conditions to burn-ups of 23.2 and 38.8 GWd/t; the third had been transient tested to 42 kW m -1 at a burn-up 26.2 GWd/t. The MOX agglomerates become highly porous in the course of irradiation. The porosity is attributed to vacancy supersaturation caused by the locally high fission density. Little U/Pu diffusion had occurred in the fuels irradiated under steady-state conditions. In the central region of the transient-tested fuel as a consequence of fast interdiffusion the MOX agglomerates had dissolved completely in the UO 2 matrix. Release of Xe and Cs from the MOX agglomerates is considered to involve two steps. First, transfer of Xe and Cs atoms to the UO 2 matrix by recoil or thermal diffusion or both, depending on the fuel temperature. Second, diffusion of Xe and Cs atoms to the UO 2 grain boundaries from where release occurs. About 20% of the Xe inventory and 5% of the Cs inventory had been released from the UO2 matrix in the fuels irradiated under steady-state conditions. In the transient-tested fuel, release figures of 50% for Xe and 35% for Cs were obtained.

1. Introduction Recycling of plutonium in Light Water Reactors (LWRs) in the Federal Republic of Germany has almost reached a commercial state. The reprocessing in commercial plants of spent P u O 2 / U O 2 (MOX) fuel from LWRs requires a high solubility of Pu ( > 99%) in pure nitric acid. With the so called AUPuC- and OCOM- MOX fuel manufacturing processes Siemens/ KWU has developed a fuel that is highly soluble even in the as-fabricated (unirradiated) condition [1,2]. The manufacturing process adopted depends on the form in which the Pu is supplied, i.e., whether Pu-nitrate or PuO2. A master-mix concept is applied in both processes so that the fuel enrichment can be varied. Consequently, on a microscopic scale the Pu is distributed heterogeneously within the fuel pellet. S i e m e n s / K W U started to manufacture MOX fuel on a large scale in the early 1970s. Since then its in-pile

* Extended version of the presentation given at the Jahrestagung Kernteehnik '90, Niirnberg, 15-17 May 1990.

performance has been monitored during irradiations spanning 1 to 4 cycles [1]. For highly soluble MOX fuel, irradiation programs have been carried out since 1981, when production on a commercial scale began. P o s t - i r r a d i a t i o n e x a m i n a t i o n s carried out by Siemens/KWU have included both nondestructive techniques in the reactor pool and destructive hot cell techniques [3]. These examinations yielded mainly quantitative data on the macroscopic behaviour of the MOX fuel rods, e.g., rod dimensional changes or integral fission gas release. Data on the behaviour of the fuel at the microscopic level were obtained from metallographic/ceramographic studies of fuel rod cross-sections and various radiochemical analyses. These investigations were later supplemented by electron probe microanalysis (EPMA) and quantitative image analysis at the European Institute for Transuranium Elements (ITU) within the frame of the research project "Oxide Fuel Transients". This paper is concerned with the EPMA examinations at ITU. Nevertheless, results from quantitative image analysis and puncturing are also reported as these provide essential information and improved clar-

0 0 2 9 - 5 4 9 3 / 9 1 / $ 0 3 . 5 0 © 1991 - Elsevier Science P u b l i s h e r s B.V. All rights reserved

2

C.T. Walker et al. / Irradiation behaviour of MOXfuel

ity on some aspects. The objectives of the E P M A investigations were two fold. First, to obtain information on how the MOX agglomerates fare during irradiation with particular emphasis on what happens to the plutonium. Second, to identify the mechanisms by which fission gas is released from the MOX agglomerates and reaches the rod free volume. With these objectives in mind the distributions of Pu and the fission products Xe, Cs and Nd were analysed. As will be seen the local concentration of Nd was used as an indicator of the burn-up in the MOX agglomerates and the UO2 matrix.

2. Fuel characteristics and irradiation history The design characteristics of the pellets and rods supplied by the German fuel vendor S i e m e n s / K W U are summarized in table 1. The fuel was fabricated by mechanically blending a "master-mix" of 70/30 UO 2PuO 2 produced by the AUPuC process with natural UO 2 powder. The product contained 3.2 wt% Pufiss and consisted of agglomerates of master-mix irregularly dispersed in a matrix of UO2. The master-mix (MOX) agglomerates were less than 100/~m in size and occupied 15 voi% of the fuel. Three sections of irradiated fuel were examined by EPMA. Two of the fuels had been irradiated under steady-state conditions; the third had been transient tested. The irradiation conditions are given in table 2 and the power schedule for the transient test is shown in fig. 1. The power cycling at the end of the test was made in order to check the mechanical integrity of the zircaloy cladding. The base irradiations were carried out in the KWO reactor at Obrigheim (Germany) whereas the transient test was performed in the High Flux Reactor (HFR) at Petten (The Netherlands).

Table 2 KWO irradiation data Rod identification

Average linear power (kW m- l)

Irradiation time (EFPD) a

Burn-up (GWd/t)

PN317 PN225 b 6015

23.2 25.7 25.3

572 572 877

23.2 26.2 38.8

a Effective Full Power Days. b Subsequently transient tested in HFR. The integral fission gas release values obtained from puncturing were 2% for PN317 and 9% for 6015 (both irradiated under steady-state conditions). The transient-tested rod PN225 had released 43% of the total amount of gas generated.

3. Experimental techniques 3.1. Electron probe microanalysis The work was carried out on the shielded electron microprobe at the Institute for Transuranium Elements. The instrument is unique in that it combines the high performance of the Cameca MS46 spectrometers with the electron optical system of the CAMEBAX microprobe [4]. Lead and tungsten shielding makes possible the analysis of samples of irradiated fuel of activity up to 740 GBq (about 20 Ci).

3.1.1. Quantitative analysis of Xe, Cs and Nd The electron acceleration potential was 25 keV and the beam current was 250 nA. Under these conditions the depth of electron penetration in uranium dioxide is about 0.5 g m and the diameter of the region of X-ray excitation is about 2.5 /~m. The conventional microprobe correction procedure was carried out using a

Table 1 Fuel pellet and rod design characteristics Fuel density (% TD) Grain size a (p.m) Enrichment (% 235U) Fissile plutonium b (wt%) MOX agglomerates (vol%) Stoichiometry (O/M) Pellet diameter (ram) He fill gas pressure (MPa) Cladding material a Linear intercept.

b 239pu + 241pu"

94-95 5-6 0.72 3.2 15.0 2.00 9.08 2.25 Zircaloy 4

P, /,.2kWm"1' 52,5h 4

I

/~ 6I0 kWm'l h'l

72.8

I I

I I Time

I

/-,8

h

Fig. 1. Power schedule for the transient test on PN225.

C.T. Walker et al. / Irradiation behaviour of MOX fud modified version of the C O R Z A F program of Tong [5]. The radial distributions of Xe, Cs and Nd in the U O 2 matrix were determined by point analysis at intervals of 50 to 150 /zm. At each location six measurements were made. These were up to 10/xm apart and were placed away from grain boundaries thereby avoiding pores, large gas bubbles and cracks. The concentrations of Xe, Cs and Nd in the M O X agglomerates were determined by point analysis at intervals of about 10 /~m along a line traversing the agglomerate. The specimen current image (absorbed electron current) was used to obtain information about the distribution and morphology of the pores and gas bubbles in the agglomerates selected for point analysis. These features appear white in the photomicrographs (see e.g., fig. 2). Xenon was analysed using the procedure developed at the Institute for Transuranium Elements [6]. For the UO2 matrix this gave the concentration of gas in solution and trapped in intragranular bubbles. Xenon contained in intergranular bubbles did not contribute to the measured X-ray intensity because point analysis was made away from grain boundaries. The confidence interval on the measured Xe concentrations at a signif-

3

icance level of 99% is about 5% relative at 0.5 wt% and 10-20% relative at 0.05 wt%. The values reported for the concentration of Xe in the MOX agglomerates in the outer part of the fuel are certainly imprecise. A large fraction of the gas present was contained in the porosity and therefore was not detected by point analysis. Moreover, the standard Z A F procedure does not correct for the changes in electron penetration and X-ray absorption caused by porosity [7]. This has very likely resulted in an underestimation of the measured Xe concentrations. 3.1.2. Quantitative analysis o f Pu Plutonium was analysed using an acceleration potential of 25 keV and a beam current of 100 nA. Because the Pu M~ (3.701 ~,) and the U M s (3.716 A) X-ray lines overlap, the intensity of the Pu Mr3 line was monitored. A quartz (1011) diffracting crystal was used for this purpose. The measured intensity was corrected for X-ray contributions from the UMiiiNi v line which lies only 0.011 ,~ from the PuM~ line. The correction was made as follows. The spectrometer was positioned on the Pu Mr3 line and the X-ray intensity obtained from a

Fig. 2. Pu X-ray maps and electron absorption micrographs showing the appearance and distribution of the MOX agglomerates at three positions in the as-fabricated fuel.

4

C.T. Walker et al. / Irradiation behaviour of MOXfuel

U O 2 standard at this setting was determined. About 8

C s - 1 were registered which corresponded to 0.76 wt% Pu. The true percentage of Pu at each spot analysed was then obtained by multiplying the reference concentration of 0.76 wt% by the fraction of UO 2 present and substracting the result from the measured Pu concentration. The analysis of MOX agglomerates of known composition in unirradiated fuel has revealed that when E P M A is carded out against a PuO 2 standard the mass concentration of Pu is equal to the measured k value * and a Z A F correction is not required. Consequently, for Pu the uncorrected k data are reported in this paper.

by burn-up determinations on conventional LWR fuel that had been irradiated in the KWO reactor. It has also been assumed that the concentration of Nd increases linearly with burn-up. The presence of fissile 241pu has not been taken into account. In a thermal neutron flux 241pu has a much larger fission cross-section than 239pu (1009 as against 743 barns) and it yields more Nd (0.193 compared with 0.167). However, the concentration of 241pu in the MOX agglomerates changed considerably during the irradiation. It comprised 10% of the fissile Pu at the beginning of life and almost 50% at the end of life. Owing to the omission of 241pu it is judged that the burn-up in the MOX agglomerates has been overestimated by about 5%.

3.2. Optical microscopy and quantitative image analysis The volume fraction of pores in the MOX agglomerates and in the UO2 matrix was measured using a Leitz quantitative image analyser connected to a Reichert Telatom 62 optical microscope. In the UO 2 matrix the percentage of porosity was measured at intervals of 480 p,m along the fuel radius. At each location, three areas approximately 120 × 180 ~ m square were examined at a magnification of 590 × . Quantitative image analysis was performed on selected MOX agglomerates at five radial positions in the outer region of the fuel irradiated to 38.8 G W d / t . At each location three agglomerates 70-100/zm in size were analysed at a magnification of 1033 × . The magnifications of 590 × and 1033 × were chosen because they provided the optimum resolution and field of view. Prior to examination in the optical microscope the specimen surface was ion etched to enhance the visual contrast of the pores.

3.3. Burn-up determination The burn-up in the MOX agglomerates and in the U O 2 matrix was calculated from the local concentra-

tion of fission product Nd. The fission yield for Nd is different for 235U and 239pu in a thermal neutron flux [8]. In this work a burn-up of 10 G W d / t has been assumed to produce 0.11 wt% Nd when the fissile isotope is 235U (UO2 matrix) and 0.09 wt% Nd when the fissile isotope is 239pu (MOX agglomerates). For 235U the figure for Nd production has been confirmed

* k =Co(l/lo), where Co is the concentration of Pu in PuO 2, 88.2 wt%, and I and I o are the intensities of the characteristic Ma X-ray line from the specimen and the standard, respectively.

4. Results

4.1. Fuel restructuring Figure 2 shows the microstructure of the as-fabricated (unirradiated) fuel. The white areas on the Pu X-ray maps are the MOX agglomerates and the white spots on the electron absorption micrographs are pores. It is evident from the figure that the MOX agglomerates were highly dense prior to irradiation. Figure 3 shows the appearance of the MOX agglomerates at three radial positions in the fuel irradiated to 23.2 G W d / t . It can be seen that the agglomerates were extremely porous after irradiation. In the cold outer region of the fuel the agglomerates were characterised by dense clusters of small pores < 3/~m in size. With increase in temperature, however, the number of pores decreased and their size increased until in the central region of the fuel the MOX agglomerates contained just a few coarse pores. The agglomerates in the outer region of the fuel irradiated to 38.8 G W d / t contained between 30 and 50% porosity and had therefore undergone substantial swelling (fig. 4). The level of porosity in the UO2 matrix was dependent on the burn-up and thermal conditions. The fuel irradiated to 23.2 G W d / t showed slight densification with the percentage of porosity varying between 3 and 5% across the section. At 38.8 G W d / t between 5 and 7% porosity was measured in the central region of the section (fig. 4) indicating that the fuel density had changed little during the irradiation. In the outer region of the fuel, however, the porosity increased to almost 10% within 500/~m of the pellet surface (fig. 4). This increase is ascribed to grain loss during specimen

C.T. Walker et al. / Irradiation behaviour of M O X fuel

5 i

22

UO 2 Matrix

60~

i

MOX Agglomerates

20 18

16 Ol

1.0

14

,

,

09

Q8

,

,

Q7

r/to -

-

06

12 "~ 10 13_

8 6

4 2 ;--ar°i, Lo,,--~

0,

I

0'8

0'6

04

!

I

I

!

a2

0

r/ro Fig. 4. Radial distribution of porosity in the UO 2 matrix of the fuel irradiated to 38.8 GWd/t. The inset gives the percentage of porosity in the MOX agglomerates in the outer region of the pellet. Owing to grain loss the porosity in the UO2 matrix is overestimated in the region between the fuel surface and r / r o = 0.8.

20, 18

12 .-_ 1C o

Fig, 3. Electron absorption micrographs showing the appearante of the MOX agglomerates at the pellet surface, mid-radial position and the pellet centre in the fuel irradiated to 23.2 GWd/t.

preparation. Results from pycnometry suggest that in the outer region of the fuel around 3% porosity is to be expected at the most and this is indicated by the broken line in the figure.

w

2 0 10

O8

!

06 0.4 02 0 q Relative Radius r/ro Fig. 5. Radial distribution of porosity in the UO 2 matrix of the fuel transient-tested to 42 kW m-1 at a bum-up of 26.2 GWd/t.

6

C.T. Walker et al. / Irradiation behaviour of MOXfuel

Fig. 6. Electron absorption micrographs and Pu X-ray maps showing the fuel microstructure and the local distribution of Pu in the transient-tested fuel.

C.T. Walker et aL / Irradiation behaviour of M O X fuel

7

seems to decrease steadily to about 0.15 wt% before increasing again to 0,2 wt% at the centre of the fuel. The integral concentration of Nd in the cross-section is 0.19 wt% which indicates a burn-up of 17 G W d / t (see section 3.3). The concentrations of Nd in three M O X agglomerates located at r / r o = 0.84, 0.6 and 0.32 in the same fuel are shown in fig. 8. The highest concentration measured in each case was 1.76, 1.31 and 1.18 wt%. In terms of burn-up, these concentrations correspond to 197, 146 and 131 G W d / t (see section 3.3). The burn-up in the fuel cross-section, br, is given by the sum of the burn-up in the M O X agglomerates and the UO2 matrix as expressed in equation (1):

In the transient-tested fuel the porosity in the U O 2 matrix increased sharply from 3 - 4 % at the pellet surface to about 12% at r / r o = 0.5 before dropping again to about 5% in the central region of the fuel (fig. 5). The increase in porosity at intermediate radial positions was associated with the precipitation of gas bubbles on grain boundaries. From the electron absorption pictures in fig. 6 it can be seen that the precipitation of bubbles on the grain boundaries began between r / r o = 0.75 and 0.8 and that at r / r o --0.7 almost all the boundaries were decorated with bubbles. It is also apparent that in the centre of the fuel grain growth had occurred and that the gas bubbles had coalesced to form pores. The electron absorption image for r / r o ffi 0.4 exhibits numerous pores of the order of 5 t~m in size. A t r / r o ffi 0.2 a marked drop in the pore density is evident and those pores that remain have increased appreciably in size.

bf =faba + fmbm,

(1)

where fa is the volume fraction of the M O X agglomerates (0.15) ba is local burn-up in the agglomerates, fm is the volume fraction of the U O 2 matrix (0.85) and b= is the burn-up in the matrix (17 G W d / t ) . Substituting for b a in eq. (1) gives a fuel burn-up, bf, that ranges from 44 G W d / t at r / r o = 0.84 to 34 G W d / t at r / r o = 0.32. These values are in good agreement with the reported burn-up of 38.8 G W d / t .

4.Z Radial burn-up distribution

The radial distribution of Nd in the U O 2 matrix of the fuel irradiated to 38.8 G W d / t is shown in fig. 7. It is seen that after falling sharply at the fuel surface from 0.4 to about 0.2 wt% the concentration of Nd

0.~

30 0.3

~0.2

ooo •

e oe





O 0



• •



' - -

"0







• O0











"'3

• ::2

0.1

01.0 .,

10

, Q8

, 0.6

, 0.4

, 0.2

0

r / ro

Fig. 7. Radial distribution of Nd in the UO 2 matrix of a MOX fuel irradiated to 38.8 GWd/t. A N d concentration of 0.11 wt% corresponds to a burn-up of 10 GWd/t. The scatter in the outer region of the fuel may result from the presence of Nd transferred by recoil ~rom the MOX agglomerates.

8

C T. Walkeret al. / Irradiation behaviour of MOXfuel

4. 3. Redistribution o f plutonium Five MOX agglomerates were analysed in the asfabricated fuel. These contained between 23.3 and 28.8 wt% Pu. In the UO 2 matrix the concentration of Pu did not exceed the limit of detection which was 0.23 wt%. The concentration of Pu in MOX agglomerates at three radial positions in the fuel irradiated to 38.8

G W d / t can be seen in fig. 8. It is evident that the concentration of Pu had fallen markedly during irradiation to about 11.5 wt%. The latter concentration was measured in the agglomerates at r / r o = 0.84 and 0.32. Figure 9 shows the radial distribution of Pu in the UO 2 matrix of the fuel irradiated to 38.8 G W d / t . Between r / r o ffi 0.85 and the fuel centre the concentration of Pu appears to increase steadily with temperature. As indicated in the figure not all the Pu meas-

Fig. 8, Concentrationsof Pu and Nd in MOX agglomerates at r / r o ffi 0.84, 0.6 and 0.32 in the fuel irradiated to 38.8 GWd/t.

C.T. Walker et al. / Irradiation behaviour o f M O X fuel i

i

!

i

3 F r o m U / P u interdiffusion

\.

"5 .

from n copture in 238U

" -

1

0'.6

I

.0

0.8

•,

I

0./~

I

0.2

0

r/r o

Fig. 9. Radial distribution of Pu in the U P 2 matrix of the fuel irradiated to 38.8 GWd/t. The radial profile for Pu created by neutron capture was deduced from the burn-up profile given by the radial distribution of Nd (fig. 7).

Figure 10 shows the situation in the fuel that had been transient-tested to 42 kW m - 1 at a burn-up of 26.2 G W d / t . In the central region of the fuel the power transient has caused the concentration of Pu in the U P 2 matrix to increase appreciably. It is seen that the concentration of Pu begins to increase at r / r o = 0.5 and reaches an upper limit of 3 wt% at r / r o ffi 0.3.

ured came from the M O X agglomerates; a substantial fraction, about 1 wt%, resulted from neutron capture in 23Su. In the fuel irradiated to 23.2 G W d / t a slight increase in the concentration of Pu could be discerned in the central region of the fuel at r / r o < 0.25. Over most of the fuel radius the matrix contained between 0.8 and 1 wt% Pu.

3



o 1

2.

•.

t ," •

0 to

"

••





I

I

IT o l •

o'.s



OQQ

0:6

OO

rlro



o'.~

0'.2

o

Fig. 10. Radial distribution of Pu in the UP2 matrix of the fuel transient-tested to 42 kW m - 1 at 26.2 GWd/t. The concentration of Pu increases in the central region of the fuel due to the dissolution of the MOX agglomerates.

10

C.T. Walker et al. / Irradiation behaviour of MOX fuel

Fig. 11. Concentrations of Pu, Xe and Cs in MOX agglomerates at r / r o = 0.85, 0.5 and 0.06 in the fuel irradiated to 23.2 G W d / t . Note that apparently the concentration of Xe increases with fuel temperature whereas the concentration of Cs decreases.

C.T. Walker et al. / Irradiation behaviour of M O X fuel

11

Fig. 12. Xenon line scan across a MOX agglomerate at r / r o = 0.85 in the fuel irradiated to 23.2 GWd/t. At closed pores the intensity of the Xe signal increases dramatically.

This concentration corresponds to the average Pu content of the fuel. 4.4. Behaviour o f xenon a n d caesium

Figure 11 shows the results of point analysis on MOX agglomerates located at r / r o = 0.85, 0.5 and 0.06 in the fuel irradiated to 23.2 G W d / t . It can be seen that the agglomerate at r / r o = 0.85 apparently contained half as much Xe as the surrounding matrix (0.08 wt%), whereas at r / r o = 0.06 about three times more Xe was measured in the agglomerate (0.15 wt%) as in the UO 2. Thus, the EPMA results indicate that the concentration of Xe retained in the MOX agglomerates increased with temperature. The true trend is masked because a large fraction of the gas present in the MOX agglomerate at r / r o = 0.85 was not detected by EPMA because it was contained in the gas bubbles and pores that formed in the course of irradiation. A Xe line scan across the MOX agglomerate at r / r o --0.85 is shown in fig. 12. At closed pores the intensity of

the Xe L~ X-rays increases dramatically revealing that fission gas was contained in the porosity. Moreover, the data in table 3 indicate that up to 60% of the gas created may be contained in the agglomerate porosity at r / r o = 0.85. More than half the Cs created was measured in the MOX agglomerates in the outer region of the fuel. With increase in fuel temperature, the concentration of Cs in the agglomerates decreased. As seen from fig. 11, in the fuel irradiated to 23.2 G W d / t the concentration of Cs dropped from 0.8 wt% at r / r o = 0.85 to 0.25 wt% at r / r o = 0.06. Figure 13 shows the radial distributions of Xe and Cs in the UO2 matrix of the fuel irradiated to 38.8 G W d / t . Both curves show clear evidence of release. The Xe profile exhibits two release steps and is not significantly different from profiles that have been measured in conventional UO 2 fuel irradiated in the KWO reactor. From the profiles it would appear that about 20% of the Xe inventory and 5% of the Cs inventory have been released from the UO2 grains.

C.T. Walker et aL / Irradiation behaviour of MOX fuel

12 0.8-

o Xenon

0.6

~ 0.t,-

• Caesium

o~,~ o~O ~ , ~ . ~ , , o . ~ .

o . 0 0'~'~, 0 0 0 ~O~C



--

0

0.2

."

.

~'-" ° - ° - ° - -

~a

0'.6

'

r/ro

0'.~.

:. o- ~ 2

o'.2

o

Fig. 13. Radial distribution of Xe and Cs in the UO 2 matrix of the fuel irradiated to 38.8 GWd/t.

Similar release figures were o b t a i n e d for the fuel irrad i a t e d to 23.2 G W d / t . T h e radial distributions of r e t a i n e d X e a n d Cs in the U O 2 matrix of t h e fuel t r a n s i e n t - t e s t e d to 42 k W - 1

are shown in fig. 14. Again, similar profiles can be f o u n d in c o n v e n t i o n a l U O 2 fuel t h a t h a s b e e n t r a n sient-tested to powers of 40 k W m - 1 a n d more. It can b e seen t h a t the U O 2 matrix has b e e n almost e m p t i e d

0.5

0.~'

~o 0.3.

~ °0° °"~:)'~~ o

0

O~~ % 0

~o

0.2-





0

~ ~

o Xenon • Caesium

0.1.

0"9"0.....o .~ .a o_Q.o_O_ "P - _ _ q" .~ o _.o_o _ o _ ~_ o -I~--

°tO

.n

018

0.6

r/ro

O.t,

0.2

Fig. 14. Radial distribution of Xe and Cs in the UO 2 matrix of the fuel transient-tested to 42 kW m - 1 at a burn-up of 26.2 G W d / t .

13

C.T. Walker et at / Irradiation behaviour of MOX fuel

of Xe and Cs out to r / r o = 0.7. The level of release from the UO 2 grains has increased accordingly to 50% for Xe and 35% for Cs.

20

Plutonium //~~2fiWcll

tU

155. Discussion

5.1. Redistribution o f plutonium

During irradiation Pu diffuses out of the MOX agglomerates at a rate that is dependent on the local fuel temperature. Little U / P u interdiffusion had occurred in the fuel sections that had been irradiated under steady-state conditions. In the fuel irradiated to 38.8 G W d / t interdiffusion had increased the concentration of Pu in the U O 2 matrix from 1 wt% at r / r o = 0.85 to 1.5 wt% at the pellet centre (fig. 9). From the latter concentration it appears that the MOX agglomerates at the centre of the fuel had lost about 12% of their initial Pu concentration to the U O 2. Even less U / P u interdiffusion had occurred in the fuel irradiated to 23.2 G W d / t . In this case a small rise in the Pu content of the U O 2 matrix was measured in the central region of the fuel. Thus, it would appear that interdiffusion increases with burn-up. It is pointed out, however, that it is not burn-up as such that is relevant, but the burn-up related parameters, irradiation time and linear power. In the transient-tested fuel substantial redistribution of Pu had occurred in the central region of the fuel. In this fuel in the region between r / r o = 0.3 and the pellet centre the M O X agglomerates had evidently dissolved completely in the U O e matrix (see figs. 6 and 10). During the transient test the fuel centre temperature was probably of the order of 2000°C. As so little interdiffusion had occurred in the fuels irradiated under normal power reactor conditions, the large decrease observed in the Pu content of the MOX agglomerates must be the result of fission. The effect of fission can be seen in fig. 15 where the concentration of Pu in the MOX agglomerates in the outer regions of the fuels irradiated to 23.2 and 38.8 G W d / t are compared. It is evident that much more Pu was measured in the M O X agglomerates at the lower burnup (15.5 as against 11.2 wt%). Apparently, fission is also responsible for the formation of pores in the M O X agglomerates during irradiation. It appears that this porosity is a consequence of vacancy supersaturation and can be explained by the high fission density in the M O X agglomerates. The vacancy flux that accompanies interdiffusion can also lead to the formation of pores (Kirken-

5-

0

0

2'0

i0

~0 ~ 160 pm Fig. 15. Concentration of Pu in M O X agglomerates at r / r o = 0.85 and 0.84 in the fuel irradiated to 23.2 and 38.80Wd/t. Appreciably more Pu is present at the lower burn-up.

dall Effect [9]). However, since little interdiffusion had occurred in the fuel irradiated under steady-state conditions, this mechanism is considered to be of secondary importance. 5.2. Release o f xenon and caesium

During irradiation the M O X agglomerates lose Xe and Cs to the U P 2 matrix. In the fuel irradiated to 38.8 G W d / t , the burn-up in the U P 2 matrix is estimated from the concentration of Nd present to have been about 17 G W d / t which is calculated to give 0.21 wt% Xe and 0.13 wt% Cs. In the outer region of the fuel, however, where there is negligible release much higher concentrations of 0.37 wt% Xe and 0.24 wt% Cs were measured. The excess Xe and Cs could only have come from the M O X agglomerates. Two things are particularly remarkable about this observation. First, and most importantly, it suggests that significant amounts of Xe and Cs were lost from the M O X agglomerates in the cold outer regions of the fuel at temperatures below 1000°C. Second, it is seen that Xe and Cs were present in the U P 2 matrix in concentrations that accorded with their fission yields even though about 40% of the Xe and 45% of the Cs measured was not created there. Thus the transfer of Xe and Cs from the MOX agglomerates to the U P 2 matrix evidently involves a mechanism that is temperature independent and which is directly related to the fission process. A mechanism that possesses both these characteristics is recoil. Gen-

14

C T. Walker et al. / Irradiation behaviour of MOX fuel

erally this term describes the loss of fission gas atoms from the pellet surface as energetic fission fragments. In the present context it is assumed that Xe and Cs atoms are rejected from the MOX agglomerates as fission fragments and implanted in the surrounding UO2 matrix. Since the range of fission fragments in UO2and PuO 2 is of the order of 10 t~m [10], only Xe and Cs atoms that are created within this distance of the agglomerate surface have a chance of ending up in the UO 2 matrix. At the agglomerate surface there is a 50% chance that a Cs or Xe atom will be propelled by recoil into the UO2 matrix and this chance decreases continuously with depth [11]. Its worth mentioning at this point that the scatter in the Nd concentrations in the outer region of the fuel irradiated to 38.8 G W d / t (see fig. 7) can also be explained if it is recognised that part of the Nd measured in the UO 2 matrix had been transferred there by recoil. Fission product atoms can also be rejected from the MOX agglomerates as a result of knock-on. This involves the interaction of a fission fragment, collision cascade or fission spike with a stationary atom [11]. Since the atom that is struck; travels just a few hundred angstroms only Xe and Cs atoms in the surface layers of the MOX agglomerates can be transferred to the UO 2 matrix by this process. For this reason the role of knock-on in the transfer of Xe and Cs to the UO 2 matrix is considered to be far less important than that of recoil. To test whether the additional Xe and Cs in the UO 2 matrix in the outer region of the fuel irradiated to 38.8 G W d / t can be attributed to recoil from the MOX agglomerates a mass balance was made. The situation at r / r o --0.84 was considered where complete gas retention is assumed. It will be noted that all fission fragments created in the agglomerates are taken to reach the UO2 matrix. This simplification can be

introduced for the purposes of calculation because the MOX agglomerates were generally of the order of 25 /~m in size (see the electron absorption micrograph for r / r o = 1.0 in fig. 6). The results are shown in table 3. It is seen that in the UO 2 matrix apparently 0.16 wt% Xe and 0.11 wt% Cs (column 8) emanate from the MOX agglomerates. These levels of increase require that the concentrations of Xe and Cs in the MOX agglomerates fall by 0.9 and 0.6 wt%, respectively (column 5). Thus, the excess Xe and Cs in the UO 2 matrix correspond to 33 and 38% of their concentrations created in the MOX agglomerates, respectively (column 2). A 30-40% decrease in the concentrations of Xe and Cs is considered to be compatible with recoil. It is also evident from table 3 that the concentrations of Xe and Cs measured in the MOX agglomerates (column 3) and their concentrations lost to the UO2 matrix (column 5) do not sum to the concentrations generated (column 1). The discrepency is large for Xe, but is negligible in the case of Cs. All the missing material is assumed to have been contained in the agglomerate porosity. For Xe this corresponds to about 60% of the created amount. Only 0.1 wt% Cs, however, is unaccounted for. Such a small concentration is clearly within the limits of uncertainty on the quantities calculated and, therefore, should be interpreted only as indicating that relatively little Cs was contained in the agglomerate porosity. Thus, it would appear that under steady-state irradiation conditions release from the MOX agglomerates to the rod free volume involves two distinct steps. First, the transfer of Xe and Cs atoms to the UO 2 matrix by recoil or thermal diffusion. Second, diffusion of Xe and Cs atoms to the grain boundaries from where release occurs. With increase in temperature the percentages of gas and Cs that leave the MOX agglomer-

Table 3 Mass balance in weight per cent for Xe and Cs in the outer region of the fuel irradiated to 38.8 GWd/t 1

2

Fission

Agglomerates (bum-up 197 GWd/t) a

3

4

5

6

Product

Generated b

Measured

in Pores

to UO z matrix c

Generated d

Measured

from Agglomerates

Xe Cs

2.7 1.6

-- 0.2 = 0.9

1.6 0.1

0.9 0.6

0.21 0.13

0.37 0.24

0.16 0.11

UO 2

7

8

matrix (burn-up 17 GWd/t)

a agglomerate at r / r o = 0.84. b Theoretical production assumed to be 0.26 at% Xe and 0.15 at% Cs per 1 at% burn-up of 239pu. c Concentrations in col. 8 multiplied by the vol. fraction ratio, matrix/agglomerates (5.7). d Theoretical production assumed to he 0.24 at% Xe and 0.14 at% Cs per 1 at% burn-up of 235U.

C.T. Walker et al. / Irradiation behaviour of MOX fuel

ates by thermal diffusion increase and recoil as a mechanism of transfer declines in importance. Under transient conditions thermal diffusion is without doubt the principal mechanism by which Xe and Cs atoms are transfered from the M O X agglomerates to the U P 2 matrix. As in conventional L W R fuel, release from the grain boundaries involves the interlinkage of gas bubbles and the subsequent formation of escape tunnels (see electron absorption micrographs in fig. 6). At high temperatures direct release from the MOX agglomerates to the rod free volume is possible in situations where open grain boundaries or cracks in the U P 2 matrix connect with the agglomerates. It appears that in the fuels irradiated under steadystate conditions only part of the Xe and Cs that had reached the grain boundaries was released to the pin free volume; an appreciable amount remained trapped on the grain boundaries. This is evident when the release data derived from puncturing and E P M A are compared. Puncturing gave a gas release value of 2% for the fuel irradiated to 23.2 G W d / t and 9% for the fuel irradiated to 38.8 G W d / t , although E P M A revealed that in both cases about 20% of the gas had been removed from the U P 2 grains. For the transienttested fuel the puncturing result of 43% was in reasonable good agreement with the E P M A result of 50% release. Thus it appears that almost all the gas that had reached the grain boundaries in this fuel had been released to the rod free volume. The release levels quoted in section 4.4 refer to release from the U P 2 matrix. The values do not, however, appear to increase greatly when release from the M O X agglomerates is taken into account. For example, the Xe release figure quoted for the fuel irradiated to 38.8 G W d / t is 20%. Assuming that the volume fraction of the M O X agglomerates is 0.15 and taking the release level to be 85% for agglomerates that lie in the region between r / r o = 0.7 and the fuel centre, and 0% for all others, an integral release value for the fuel of 26% is obtained. It is pointed out that in the transient-tested fuel the agglomerates between r / r o -- 0.7 and the pellet centre had probably released all their Xe and Cs. This would suggest that the percentage of Xe released from the fuel was also somewhat higher than the figure of 50% given in section 4.4.

6. Summary and conclusions Under steady-state reactor operating conditions negligible U / P u interdiffusion occurs between the

15

M O X agglomerates and the U P 2 matrix. In the fuel irradiated to 23.2 G W d / t a small amount of Pu may have diffused into the U P 2 matrix at the centre of the pellet. Moreover, in the fuel irradiated to 38.8 G W d / t , although the concentration of Pu in the U P 2 matrix apparently began to increase at r / r o = 0.85 it did not exceed 1.5 wt% at the pellet centre. Under transient conditions appreciable U / P u interdiffusion may occur in the central region of the fuel and at high temperatures the M O X agglomerates can dissolve completely in the U P 2 matrix. In the fuel transient-tested to 42 kW m -1, for example, the MOX agglomerates had disappeared completely in the region between r / r o --0.3 and the pellet centre producing an homogeneous solid solution containing about 3% Pu. The burn-up in the M O X agglomerates is much higher than that encountered in conventional U P 2 fuel. In the fuel irradiated to 38.8 G W d / t the burn-up in the agglomerates was in the range 130 to 200 G W d / t in the interval between r / r o = 0.84 and 0.32. These values contrast with the low burn-up of 17 G W d / t or less in the surrounding U P 2 matrix. As a result of the high burn-up, the Pu content of the agglomerates had decreased considerably during irradiation from about 27 to around 11 wt%. The MOX agglomerates become highly porous in the course of irradiation. The porosity is mainly a consequence of vacancy supersaturation arising from the high fission density in the agglomerates. Since little U / P u interdiffusion occurs under steady-state irradiated conditions it cannot be attributed to the Kirkendall Effect. The appearance of the porosity changes with radial position (i.e., with increase in fuel temperature). The MOX agglomerates in the outer region of the fuel contain a high concentration of small pores whereas those in the centre of the fuel exhibit just a few large pores. In the outer region of the fuel the pores contain a large fraction of the fission gas present in the M O X agglomerates. For instance, in the outer region of the fuel irradiated to 38.8 G W d / t more than 50% of the Xe created was contained in the agglomerate porosity. During irradiation the M O X agglomerates lose Xe and Cs to the U P 2 matrix. Depending on the fuel temperature the principal mechanism of transfer is either recoil or thermal diffusion. Under steady-state irradiation conditions recoil dominates in the outer region of the fuel at low temperatures whereas thermal diffusion is prevalent in the central region of the fuel. As in conventional LWR fuel the release of Xe and Cs to the rod free volume occurs via the U P 2 grain boundaries and involves the interlinkage of gas bubbles

16

C.T. Walker et al. / Irradiation behaviour of MOX fuel

and the formation of escape tunnels. Thus, the release process consists of two steps. First, the transfer of Xe and Cs atoms to the UO 2 matrix by recoil or thermal diffusion or both. Second, diffusion of Xe and Cs atoms to the UO 2 grain boundaries from where release occurs. Under transient conditions thermal diffusion is doubtless the principal mechanism by which Xe and Cs are transferred from the MOX agglomerates to the UO~ matrix. The radial distributions of Xe and Cs in the UO 2 matrix of both the steady-state irradiated and transient-tested fuels were not significantly different from those found in conventional UO2 fuels. In the fuels irradiated under steady-state conditions a large fraction of the gas that had been released from the UO 2 grains was retained on the grain boundaries. In the case of the transient-tested fuel almost all the gas that reached the grain boundaries was released to the rod free volume.

References [1] H. Roepenack, F.U. Schlemmer and G.J. Schlosser, Development of thermal plutonium recycling, Nucl. Tech. 77 (1987) 175-186. [2] R. Wiirtz, L6slichkeit bestrahlter MOX-Brennelemente bei der Wiederaufarbeitung, Atomwirtschaft 32 (1987) 190-192. [3] F.U. Schlemmer, H.P. Fuchs and R. Manzel, Status of

irradiation experience with recycled fuel materials in the FRG for Siemens/KWU type fuel assemblies, IAEA Tech. Committee Mtg. on Recycling of Plutonium and Uranium in Water Reactor or Fuels, Cadarache, France, 13-16 Nov. 1989, to appear. [4] G. Giacchetti and C.T. Walker, A shielded microprobe facility for the analysis of irradiated nuclear fuel, in Proc. 9th Int. Congress X-ray Optics and Microanalysis, Vol III, ed. P. Brederro and V.E. Cosslett (The Hague, 1980) pp. 38-39. [5] M. Tong, Methode de correction en microanalyse, J. Microsc. (Paris) 8 (1969) 276-306. [6] C.T. Walker, Measurement of retained xenon in advanced fuels by microprobe analysis, J. Nucl. Mater. 80 (1979) 190-193. [7] C. Ronchi and C.T. Walker, Determination of Xe concentrations in nuclear fuels by electron microprobe analysis, J. Phys. D: Appi. Phys. 13 (1980) 2175-2184. [8] M.E. Meek and B.F. Rider, Compilation of fission product yields, Vallecitos Nuclear Centre, report NEDO12154-1, 1974. [9] A.D. Smigelskasand E.O. Kirkendall, Zinc diffusion in a brass, Am. Inst. Mining Met. Engrs., Inst. Metals Div., Metals Technol. 13, No. 7, Tech. Pub. No. 2071 (1946). [10] Hi. Matzke, Radiation damage in crystalline insulators, oxides and ceramic nuclear fuels, Radiation Effects 64 (1982) 3-33. [11] D.R. Olander, Fundamental Aspects of Nuclear Reactor Fuel Elements (Tech. Info. Centre Energy Research and Development Administration, Springfield, Virginia, 1978) pp. 287-293.