"
-
?
-
TOFT
ORNL P
1108
.
.
.
:
-
•
-
f
.
.
&
.
Lim
..
.
}
.
.
|| 1.25 .1.4 16
e
n
MICROCOPY RESOLUTION TEST CHART
NATIONAL BUREAU OF STANDARDS -1963
is
1.
RH
1
.41
A
1
D
CLOd3122
Yell
Vi
AU
.
1
14
.'
'
LEYL
"224
***
11,
**
TH

PI
.
.
'
.
!
.
.
..
te
.
PAYWERSRET
i
O
*..*.
WR
-
!"
"
.
."
.
11,
.
-;
OI
-
NEW
.7
..
'
.
i
.
!
!
!
.
- !
-
-
-
...
.
1
.:
1.1
2
.
.
.
.
,
.'{:;
:
i
'
.-
21
.
.
.
.
-
-
TZ
-
-
'Tu
".
1
-
.
-
.
.
1.
..
.
.
7.
.
":
.
.
LY
..
.
II
...
.
.:
ii
.
..
1
..
I.1
..
-
.
.
.
.
1.
'
.
L
'
.
.
.
.
1..
....
...
.:
YI
.
-
-
--
RE
.-
.
2."
"
.?
.
ti
'my.
.
.
3
.
.
.
.t
.
ii
9.
..
1
1
..
'
11.
.
in
-
.
.
.
.
**
AN
LEGAL NOTICE
This report was prepared as an account of
--Government sponsored work. Neither the
United States, nor the Commission, nor any
person acting on behalf of the Commission:
A. Makes any warranty or representa-
tion, expressed or implied, with respect to
the accuracy, completeness, or usefulness
of the information contained in this report,
or that the use of any information, appa-
ratus, method, or process disclosed in this
report may not infringe privately owned
rights; or
B. Assumes any liabilities with respect
to the use of, or for damages resulting from
the use of any information, apparatus,
| method, or process disclosed in this report.
As used in the above, “person acting on
behalf of the Commission” includes any em-
ployee or contractor of the Commission, or
employee of such Gontractor, to the extent
that such employee or contractor of the
Commission, or employee of such contractor
prepares, disseminates, or provides access
to, any information pursuant to his employ-
ment or contract with the Commission, or
his employment with such contractor.
.
.
.
:-1:-;
I
.
.
.
--
.
.
.
.
....
:
:
i
-
1
.
|
1' 1
BHC
...
!
.
!
!
!,,
E
.14--21.
.1.4
1
-
-
.
.
'
ij:.::::::
,
.
.
*
.
Fr
*..
.
.
'
.
I
INST
AW
AL
.
1-
ni
:.
-
-
..
4
+42
.
.
..
.'
.
.:'
'.
.
.
.
..:
!
:
YAN
.
..
-
t
...
.
-
Fii
..!!
'.
'.
.'.
.
1.1
n
.
.
.-..
.
.
1.
.
'
.
-
E..
2..
-
1
r
.
.
4
.
T
.
.
.
'
*
...
...
.
.
.
*
..PL
i
+
-
:
-
".
-
.
.
;
-
.
.
.
-
-
i
1
.
...
1
.
.
.
i
.
Y
.
Y
.
W
.
>
'
.*
..
.
.
.
A'
.
.
..
-
-
-
-
1
.!
.
.
.
'
.
L
.
:'
..
J
.
i
FORM-P-1108

will
so
R 29 1965
CONF-650904-1
MASTER
THE BEHAVIOR OF FISSION-GAS IN FUHLS*
R. M. Cerroll
Oak Ridge National Laboratory
Oak Ridge, Tennessee
----
- LEGAL NOTICE –

The report wuo prepared una escount of Covenant poured work. Mother the United
Da , nor the couniuko, nor ay para o acts about of the Commodons
A. Maka mrrunty or reprenatation, promod os lagund, Mth rect to the seco-
rhoy, complement, or wantalone of the taformation contebuod ha de reports or that the one
of my bularuation, spranu, methods or proceduncloud ute soport may not latte
petrated oned rights or
D. Am my Habtusor with roupat to the umo, or for denne tenallting from the
w of my takorutta, appart, cathod, or prooww dieclound on this reports
A und an the sboro, "partod act ang ball of the Commission" troludou may on-
Noge or contractor of the Constanton, or neploy of mod contrator, to do that
mol esployee or contractor of the Commission, or uployw at al contractor wropurus,
denuntates, or provide scco to, ay taformation par met do No employee or contract
Wa c ominalan, a dio saplog at the mob contractor.
TOM
INITIVcnR
resowanie Sihat EATIVE OF PATENTIsinan
TIRSINSININO
www
A DAVED FOR PUBLIC RELEASE
*Research sponsored by the U. S. Atomic Energy Commission under
contract with the Union Carbide Corporation.
--- --•••**-*.

This paper was submitted for publication
in the open üterature at least Months
prior to the issuance date of this Micro-
card. Since the 1.8.A.E.C. has no evi-
dence that it has been published, the pa-
per is being distributed in Microcard
form as a preprint.
they
in
a
-
..
.
-
-
-
-
-
-
1.
.
*
!
THE IWHAVIOR OF FISSION-GAS IN FULLS*
R. M. Carroll
Oak Ridge National Laboratory
Oak Ridge, Tennessee
INTRODUCTION
The behavior of fission-products in nuclear fuel materials 18
important to reactor economics and safety. A very significant group
of fission-products are the noble gases, xenon and krypton, since
about 80% of the fissions produce an atom that exists for some
length of time as one of the fission-gases. If the gases escape
from the fuel structure they will cause an internal pressure on
th? fuel cladding that will increase as burnup progresses. If the
fuel 18 unclad, or if the cladding ruptures, the gases will es-
cape into the primary coolant. Some of the fission-gases are
highly radioactive, and the amount and types of these gases in
the coolant will influence (1) the extent to which the coolant
circuit must be shielded, (2) the amount of contamination from
the radioactive daughters of the gases, and (3) the hazard pre-
sented by rupture of the primary coolant circuit.
The swelling of fuels during irradiation will be influenced
by the behavior of the fission-gas that remains within the fuel
structure (e.g., whether the gas remains widely dispersed through
the fuel or whether it collects into bubbles). This topic is the
subject of another session of this meeting and my discussion will
be limited to the escape of the gases from the fuel structure. I
-
-
-
*Research sponsored by the U. S. Atomic Energy Commission under
contract with the Union Carbide Corporation.
.

'
.
.
Yr
.
SPA
T
will consider only pure, solid, fuel materials. Migration is
more important at high temperatures, and ceramic fuel materials
re
mening
(particularly voz) are the predominate choice for high temperature
reactors. Therefore, my primary subject will be the behavior of
fission-gas in uranium dioxide and the factors which influence
this behavior.
I will try to use the fission-gas behavior 10 Uo, to show,
in a general way, what factors can be expected to influence the
fission product release from any nuclear fuel. The qualitative
effects of irradiation are essentially the same for the various
ceramic fuel materials; principal differences being in the magni-
tude of the irradiation effects and the temperature at which the
irradiation effect occurs. Thus, the factors discussed below
will have implications concerning fission product release from
all types of ceramic fuels and likely metallic fuels as well.
CLASSICAL DIFFUSION RELEASE MODEL
Early investigations of the aigh temperature release of fis-
sion-gas from uo, showed that the gas was emitted as an exponen-.
tial function of temperature, and it was generally accepted that
diffusion was the primary fission-gas release mechanism. It
follows that, if the diffusion coefficient were known, the fission-
gas release in fuel elements could be calculated. A large number
of measurements of diffusion coefficients for uo, have been made;
however, a wide range of values was reported by different inves-
tigators. Diffusion coefficients for sintered vo, are difficult
to calculate because of the semi porous nature of the body, and a
great deal of effort bas been expended in search of a suitable
model upon which to base the calculations.
One widely accepted concept assumes the grains of the sin-
tered vo, body to be close-panked spheres of uniform radius "a".
The fission-gas 18 assumed to diffuse slowly until it emerges
from the surface of the sphere and then to escape rapidly through
the inter-connected channels leading to the specimen surface.
The amount of fission-gas that escapes the specimen 18 considered
to be a function of D/a2 where D is the diffusion coefficient
and (D/a2) 18 called the apparent diffusion coefficient. By
this model, the amount of gas released at a given temperature
18 a function of a constant diffusion coefficient and a paysi-
cal characteristic of the material which might vary widely be-
tween specimens.
A common method of determining the diffusion coefficient
18 by post-irradiation annealing. The specimen 18 irradiated
just enough to produce sufficient gas for measurement, at a
temperature low enough that diffusion does not occur during the
irradiation. After some decay time, the specimen 18 heated and
the amount of libsion-gas evolved at a fixed temperature is
measured as a function of heating time. This method has the
advantage that the specimen may be handled without elaborate
shielding, the temperature can be accurately controlled and
measured, and a large number of specimens may be processed in
a reiatively short time.
DEFECT TRAP CAZCEPTS
More recently the results of work by several experimentors
have led to the conclusion that the temperature-dependent release
of fission-gas is not controlled by diffusion, but is controlled
by a trapping process. The first indication of a trapping pro-
cess was discovered at Chalk River by Stevens and MacEwan when
they investigated the effect of low burnup on the xenon dimin
sion coefficient. (2) Using a post-irradiation heating method,
they found an order of magnitude decrease in the apparent diffusion
coefficient when the irradiation dose was increased from 2 x 2015 to
3.8 x 10+' fissions/cm. Later, they irradiated single-crystal
powäer and sintered UO specimens to burnups ranging from 1014
to 1049 fiosions/cms. The apparent diffusion coefficient of
155xe decreased with increasing burnup for all cases, changing
as much as a factor of 1000 for the single-crystal specimens.
Grinding studies indicated that increasing the irradiation ex-
posure caused the proportion of xenon trapped in small pockets
A
:
to increase relative to the amount in large pockets.12) This
result implies that the smaller traps are formed by irradiation.
Frigerio and Gerevini also have found that increasing the
burnup resulted in lowering the escape rate of fission-gas."
Several laboratories have used autoradiographic techniques to
show that fission products are concentrated at defects and
grain boundaries in Irradiated vo(4) Here again the conclu-
Bion 18 drawn that defects in the UO structure behave as traps
for Pission products and that fission product traps can be
created by irradiation.
The general observation that irradiation could produce a do-
fect capable of trapping fission-gas has been made for several
different materials. Zuwvalt observed that "single crystal" UC
particles had xonon trapped or attached to detect sites and inter-
preted the data in terms of a combined trapping-diffusion mechan-
imm.5Iajina has observed that slaston-gas la trapped in gra-
phite at defect sites created by tission fragents.(o) Logerwall
and Schweling studied the diffusion rate of Ar in car as a func-
tion of neutron dose and teasperature and concluded that a trap-
ping mechanism lovered the dimrusion rate.)
In viow of these similar observations on a vide variety of
Irradiated material, one should suspect a defect-trap mechanism
to have a dominating effect on the fission product escape from
almost any solid fuel material. From studies of the babavior of
xenon in Vo fuel, Lovis has stated,'°) northe general conclusion
--- 18 that the mean free path of an isolated Xepon atau before
It encounters & vacancy or an irregularity that serves as a do-
laying trap 18 quite short. The migration of these traps as a
whole, however, does not necessarily require such a high acti-
vation energy, as the escape of the xenon from the trap. These
gas-f111ed vacancy traps agglomerate and grow into the observ-
able bubbles which albo move under thermal gradients or other
stressos."
6.
MEASURDENTS DURIG FISSION LG
II
All of the above evidence for defect trapping vas obtained by
annealing or by electron microscope studies of the fuel material
arter irradiation. For an operating tual material, all the nigra.
tion of the fission-gas viu occur during the irradiation. For
this reason it seens essential that fission-gas release experi-
ments should be partorzed on ceramic ruel materials at high tem
peratures and waile the fuel 18 producing nuclear pover, both to
better understand the gas release mechanism and to eatablish the
critical parameters controlling the fission-gas release.
We have conducted a series of experiments at oral in vaich
fission-gas release was measured during irradiation of the to
under conditions were the fission dansity and temperatura coula
be careruly controlled.191 The Sission rate was regulated by
moving the specimen into or out of the neutron flux of the Onk
Ridge Research Reactor. The specimen was beated by its own pu-
clear power and the temperature was controlled by air-cooling the
capsule which contained the fuel specimen. Pission-gas was antrained
in a continuously tloving stream of sweep-gus and the amount of
fission-gas was measured by gamma-ray spectrometry. (see Fig. 1)
The neutron flux at the specimen was measured by using a
sweep gas with a known argon concentration and measuring the
noutron activation of the argon. (The argon peak may be recorded
along with the fission-gas spectrum and 16 therefore a very
simple measurement.) (10)

CLASSIFIED
ORAL-DWG 64-7270
HYDRAULIC
POSITIONER
PROCESS WATER
(REVERSIBLE FLOW)
HIGH-LEVEL
COUNTER
PURIFICATION
SYSTEM
CHARCOAL
TRAP
TO
STACK
ARGON
HELIUM
GAMMA
SPECTROMETER
AIR
SUPPLY
SMR
COUNTER
COOLING AIR
A:R OUT
ΤΟ
STACK
UO2 SPECIMEN-
Fig. 1. Flow System of In-Pile Experiment
.
NON
.
T2
The UO specimens were selected with great care, since it
has been well established that the physical properties of UO
(e.8., density, surface area, and oxygen-to-uranium ratio) bave
a very strong effect upon fission-gas release. In order to re-
duce the number of parameters affecting the gas release, all
the specimens were of nearly theoretical density (10.96 g/cm3),
stoichiometric composition, contained no detectable porosity,
and the total impurities were less than 200 rpm.
The specimens
were machined into thin discs 0.10 cm thick by 1.27 cm in dia-
meter. By sandwiching two slabs together, the highest tempera-
ture was between the slabs and this was readily measured by
thermocouples. Because the slabs were thin the temperature drop
across the specimens was small and a definite temperature could
be assigned to a measured fission-gas release. Because of the
small temperature drop the specimens could be operated at power
densities in excess of 130 watts/cms without thermal stress rup-
ture. The experiment capsule 18 shown in Fig. 2. The tempera-
ture was controlled in order to determine the change in gas re-
lease rate with change in temperature. The neutron flux was
varied independent of temperature in order to determine the ef-
fect of fissioning rate.
· The.object of the in-pile experiment was to study the release
of flesion gas during libsioning of the fuel. In addition, there
are several distinct advantages to making in-pile measurements
of fission-gas release.

UNCLASSIFIED
ORNL-LR-DWG 40238R3
SWEEP GAS
IN
THERMOCOUPLE LEADS
AUS

15
IN
.
1111
AIR OUT
III
TI
III
MINIMIT
III
1
IND
IIIIIIIIIII
TI
.
UT
IIIIIIIII
IIIIIIIII
MI
-
WWW
-SWEEP GAS OUT
.
w
-
- POSITIONING TUBE'...
WIMMINIMIZ
.
.
per
IN
IN
ten
I
.
US
h
.
.
LI
WO
.
I
!
ON
SWEEP GAS OUT ..
MO!
Al2O3 HOLDER-
1
.
VIDIA
-SPECIMEN CONTAINER
TITUT
- INSULATED THERMOCOUPLES
TITUINTILI
-
=
w
-
REACTOR
CENTER LINE
N
UO, SPECIMENS
II
N
2
00
ZA
WIKIUIIIIIIIII
Fig. 2.
In-Pile Specimen Container
.
.
.
10.
.
1. The measurements may be made under equilibrium conditions
of fission-gas production and release.
2. The simultaneous emission of several isotopes of xenon
and krypton from the same specimen may be studied under identical
conditions. By comparing the isotopic ratios of the gas for equi-
librium conditions and transient conditions one gains valuable in-
formation about the mechanism o? release. .
3. The same specimen may be used for a series of measure-
ments, thus eliminating the differences caused by composition,
... ---,.
.
.
.
closed porosity, grain boundaries, surface-to-"volume ratio, and
impurities, which must be taken into account when a series of post-
irradiation measurements are made. . .
4. The effect of short term burnup is somewijat easier to
follow in an in-pile experiment than in a series of postirradia-
tion annealing experiments.
. .
. .
.
.
.
.
.. .
.
.
.
...
..
.
..
.
.
...
...
..
.
DEFECT-TRAP MODEL
From the results of the in-pile experiments, we have formed
a tentative defect-trap model for fission-gas release during ir-
radiation. Ichty Lag 1 This model assumes that the rate-controlling
process for fission-gas release in the temperature dependent re-
gion (above 600°C for U02) is the probability of being trapped and
the probability of being released from traps. The traps are
divided into three catagories: (1) intrinsic traps, which are
voids, grain boundaries, or other flaws in the material (2) point
defects which are formed in the wake of a fission fragment, and
(3) clusters of point defects. The second and third types of de-
fects are formed by irradiation, while the first is an inherent
.
.
.
.
.
.
..
.
.-.-.
-.-
mminile
-
,
-
-property of the material.
A
.
11.
· For the defect-trap theory, we propose that when a diffusing
fission-product atom encounters a point defect it is trapped in
the defect, but that it has a probability of escape from the trap
which increases with temperature. Also, an energetic fission
fragment pass:ing nearby will free the fission product from the
point defect trap. If the point defect does not either trap a
fission product or combine with a cluster of point defects, it
will soon anneal and vanish. The time required for annealing 18
also a function of temperature. Consequently, the population of
- point defects will reach a production-decay equilibrium. .'.
By the defect-trap theory, the fission-gas would diffuse
rapidly through the Uoa matrix and the escape rate from the speci-
men would be controlled by the probability of not being trapped
in a defect. The proportion of fission-gas which escapes from
the specimens 18 small (less than 10-3 at temperatures below
1300°c), indicating that the probability of fission products
being trapped is large and that only gas near the surface of the
specimen will escape.
The trapping probability decreases as the temperature in-
creases because the annealing rate of the point defects increases
with temperature. Also the probability of escape for a fission
product trapped in a point defect (or a cluster of point defects)
would increase with temperature. The probability of a trapped
Pission product escaping from the defect would vary with the size
of the defect; the intrinsic flaw presents a deep trap from which
:
12.
there is small probability of escape. The intrinsic flaws are
large enough so the self-diffusion of the flaw (at temperatures
not exceeding 1500°C) would be negligible.
Once having established a production-decay equilibrium ol
point defects, it follows that the production rate of defects will
increase proportionally as the fission rate is increased (at least
until the fission rate is so great that the fission tracks are elimi-
nating a significant portion of the existing point defects).
Consequently, the trapping probability would increase with fission
rate and this would produce a lower fractional escape rate (release
rate/production rate).
EXPERIMENTAL EVIDENCE FOR THE DEFECT-TRAP MODEL
During irradiation, there are two basic processes by which
fission-gas is released from uo. The first is a recoil activated,
knock-out process, which is important at temperatures below 700°C.
We will discuss this process later. The second gas release process
increases exponentially with temperature and becomes the primary
release process at temperatures above about 800°C (see Fig. 3).
Fission Rate Effect
At first, we assumed that the temperature dependent process was
diffusion, but then we observed that a change of neutron flux had almost
no effect on the fission gas release rate. We should recall that the
fission rate in the specimens was directly proportional to the neutron
flux. Since t?ese measurements were taken under equilibrium conditions,
UNCLASSIFIED
ORNL-DWG 64-7276
- O FLUX = 2.5x103 neutrons/cm2. sec –
• FLUX= 5.0x1013
RELEASE RATE (atoms/sec)
الا
5x105
L
t
500 600 700 800 900 1000 4100 4200
TEMPERATURE (°C)
Release Rate of 88 kr from Fine Grain Specimen (C1-12) of UO,

UNCLASSIFIED
ORNL-DWG 64-7279

THERMAL FLUX (neutrons/cm2. sec)
.
v 1.9
• 2.4
• 3.1
A 3.6” x 1013
0 5.0
o 6.0
A 8.4
88kr RELEASE RATE (atoms/sec)
51,000 cal/mole ACTIVATION
ENERGY
12
5 6 7 8 9 10 11
10,000/= 10K)
Temperature Dependent Release Rate of 50 Kr from
Fine Grain VO2
-
.
,
•
15.
the amount of Kr in the vo, specimen was directly proportional to
the neutron flux. The data shown in Pigs. 3 and 4 indicate that the
amount of Kr in the specimen can be changed by more than a factor
:
of 4 without changing the escape rate significently. This result is
directly contrary to diffusion theory which postulates that the release
rate should be directly proportional to the concentration. On the
other hand, the defect-trap model predicts that the trapping probability
will increase as fission-rate increases. The increased concentration
of "Kr 18 thus opposed by the higher trapping probability, resulting
in only a slight change in fission-gas release.
Activation Energy
•
The temperature dependence of the fission-gas release determines
the activation energy of the release process. We have found that
the activation energy is the same for both xenon and krypton in Uo
This result was unexpected because these elements have different
atomic radii and therefore should have different activation energies
for diffusion. However, the defect-trap model implies that the
observed activation energy was for the migration or annealing of
point defects and therefore all. noble gases would be released with
the same temperature dependence.
Burst Effect
A burst of fission-gas is observed when irradiated Uo under-
goes an increase of temperature. We have measured the amount of
fission-gas released in a burst when the temperature of a UO
specimen was suddenly increased (see Fig. 5). By measuring the
equilibrium release rates before and after the burst, we could
UNCLASSIFIED
ORNL-LR-DWG 79721
RELEASE RATE (atoms/sec)
Kr88
Kr 85m
Kp88
Kr87
85m
Kr87
0
20 40 60 80 100 420 140 160 180 200 220 240 260 280. 300
TIME (min)
Krypton Release when Single Crystal UO, Specimen C1-9, Temperature was
Increased from 865°C to 1040°C at Constant Flux of 3.3 x 1013 nv.

17.
make a simultaneous solution of the steady state diffusion equations .
during the burst shown in fig. 5 as could be predicted by diffusion
theory. The relative release rates of the krypton isotopes before,
during, and after the burst indicated the burst was composed of gas
that was liberated from traps.
Grain Size Effect
The influence of the grain boundaries on fission-gas release
(when the microstructure of vo, is not changing) has not been well
understood. Generally, when a diffusion model is used, the grain
boundaries are ignored, although in some cases they are assumed to
provide an avenue of rapid diffusion. On the other hand, the defect-
trap model postulates that grain boundaries will anchor migrating
fission-gas. Therefore a good test of one aspect of the defect-
trap theory is to compare fission-gas releases for specimens
with different grain sizes. To do this we have made in-pile tests
ns
comparing the release of fission-gas from single crystal and line
grain uo 2) Both type specimens were cut and ground to the same
size and surface finish. The density, chemical composition, and
irradiation conditions of both type specimens were almost identi-
cal. The essential difference between the specimens was that one
had many grain boundaries and the other had none.
The kr release from the fine grain specimens (~ 3u) was found
to be 17 times lower (at 1000°c) than for the single crystal (see
Fig. 6). Metallographic examination showed a broadening of the
UNCLASSIFIED
ORNL-DWG 64-7284

SINGLE CRYSTAL
88 Kr RELEASE RATE (atoms/sec)
EFINE GRAIN
- 5x10² L
. 600 700 800 900 1000 1100 1200 1300
SPECIMEN TEMPERATURE (°C)
Temperature Dependent 8° Kr Release Rate Normalized
to Single Crystal Surface Area.
grain boundaries for the irradiated fine grain specimen (2.5 x 1019
fission/cms burnup) indicating a trapping of fission products at
the grain boundaries. Grain boundary broadening in uo has also
been observed by Roberts, (14) Golyanov and Pravdyuk, (25) and
Daniel, (103 among others. These observations support the defect.
trap theory and indicate that grain boundaries in vo, act as traps
rather than avenues of rapid escape.
During irradiation, grain-growth (recrystallization) will occur
in vo, at temperatures above about 1600°C. Essentially all the
libsion-gas trapped in defects will then be released because the de-
fects will migrate to a free surface. Measurements of fission-gas
pressure in an operating fuel element indicate that the pressure can
be related to the volume of uo, at temperatures above 1600°c. (17)
During post-irradiation measurements of fission-gas released from
stressed UOz, Fitts has found that the gas release was proportional
to the amount of creep. (10) It appears that nearly all the fission-
gas will escape during any process involving gross reordering of the
Uo, microstructure.
RECOIL ESCAPE
During irradiation, Pission-gas will escape from the fuel struc-
ture by a recoil process, since when fission occurs within about 104
of the surface of yo, it is possible for a fission fragment to recoil
free of the fuel. The amount of recoil escape can be easily.calculated
but such calculations are often misleading for the following reasons.
(1) The fission fragment usually leaves the vo surface with sufficient
20.
rea
res
energy to embed in solid surfaces near the specimen and thus will not
be released. (2) When the fission fragment leaves the vo, surface an
average of about 2000 UO molecules are ejected, (19) and fission
products in this knock-out zone are ejected along with the Uog
We can determine if the fission-gas 18 liberated by direct re-
coil or knock-out by comparing the relative amounts of different iso-
topes in the fission-gas. This comparison shows that at vo, tempera-
tures below 600°C most of the fission-gas is emitted by the knock-
out process.() The amount of gas liberated by knock-out is directly
proportional to the total (B.E.T.) surface area of the fuel whereas
the amount of gas by direct recoil is proportional to the geometric
surface area. We have measured the knock-out rate and find that it
is directly proportional to the fission rate (see Fig. 7). The specimens
used for lig 7 have different amounts of knock-out because of the
different total surfacetareas of the specimens.
Amother effect of knock-out is that vaporized Uo will plate
out upon nearby surfaces: (29, 20) The amount of plated Uo, will
reach an equilibrium when a layer about 10ả thick has been de-
posited. This rapid equilibrium was unexpected, since the amount
of Pissions in the deposited layer is only 10-4 those in the adja-
cent vo, surface within recoil range. Taus, the amount of knock-
out by fission-fragments from the vo surface must be much greater
than from the deposited layer. Most likely, a fission fragment
knocks out vo, as it emérges and splashes out uo, from the adjacent
layer as it enters. The fact that the splash-out from the thin de-
posited layer equals the knock-out from the vo, surface leads me to
believe that only a thin surface layer (in the order of loÅ) is in-
nas
(x105)
UNCLASSIFIED
ORNL-DWG 64-7280
FINE GRAIN
O RELEASE RATE (atoms/sec)
ESINGLE CRYSTAL -
1 2 3 4 . 5 (x1013)
. THERMAL NEUTRON FLUX (neutrons/cm2. sec)
Comparison of 88 Kr Recoil Release Between Single Crystal
and Fine Grain Uo, Specimens of the Same Size and Density.

.
...
wi
.
es
r
e
AY
t imativna tema
22.
22.
.
volved in the knock-out, splash-out mechanism.
On a microscopic scale, even polished UO, has a convoluted sur-
face. Many direct recoils leaving the surface will impinge upon an-
other part of the UO,, producing both knock-out and splash-out en-
tirely on the fuel surface. These ejected molecules will redeposit
upon the to, surface, producing an effect somewhat like a heavy
snowfall (since small surface irregularities will be smoothed away
while large irregularities will not be changed). This smoothing
causes a decrease in the total surface area of the specimen and can
affect the Pission product release rate at all temperatures.
We have found that the steady state release rate of xenon and
krypton from single crystal vo, decreased 8.8 burup progressed. (d)
By the defect-trap model, we could attribute this to the accumula-
tion of traps as irradiation progresses; however, the knock-out
release rate declined in exactly the same way as the high temperature
***YMPHP WWWWW
W
release and this could be explained by neither the diffusion or the
defect-trap model. Therefore, we postulated that surface smoothing
by knock-out was causing the decrease. Post-irradiation examination
showed that the specimens were indeed smoother than before irradiation. Inde)
In addition, a single crystal specimen which was highly pol.ished be-
fore irradiation did not show a decrease in gas release with burn-
prodhor
up.(12)
DISCUSSION
Considerable effort has been expended in obtaining suitable
diffusion coefficients for f'ission products in fuel materials.
1.
--
z
23.
re.
Uranium dioxide has likely received as much attention as any other
fuel material and the literature abounds in values for diffusion
coefficients of fission-gas in UO . However, recent studies have
shown that the fundamental assumptions used in obtaining these
values were incorrect. Therefore, a simple diffusion model cannot
be used to estimate the amount of fission products that will be re-
leased from ceramic fuels during irradiation and I would suspect
such a model for metallic fuels.
It is not surprising that the gas release from UO was considered
to be by diffusion process. The fission-gas release is temperature dependent
and a plot of the gas release in an Arrhenius plot yields a straight line,
as would be expected for diffusion (see Fig. 4). Actually, all processes
governed by an Arrhenius factor will have the same general form. The
first evidence that unknown factors were influencing the data was the
widely scattered values of D obtained by different investigators.
Mostly these variations were considered to be caused by differences in
the UO specimens.
Almost all the published values of D for Uo have been obtained
by postirradiation annealing. In the past two years MacEwan and
others have demonstrated that the values of D obtained by this method
are changed greatly by the amount of irradiation. I 2, 4) One of the
fundamental assumptions in obtaining values of D has been that the
amount of irradiation did not change the diffusion coefficient. Each
experimenter usually gave about the same irradiation dose to his
specimens but not the same dose an other experimenters. Usually each
experimenter produced a set of consistent data which did not agree well
with other sets of data. Tais dose factor explains much of the scatter
24.
.
.
.
..
in reported values of D..
....
. .
Another fundamental assumption has been that the fissioning
-
process does not influence the diffusion of the fission-gas and
-
therefore postirradiation annealing experiments can be used to pre-
..
- - - ...
-
--
- ...
-
dict the amount of l'ission-gas that will be released from an operat-
-
-
.
.
.
ing fuel element. Accary attempted such a correlation and reported:
"--there is disagreement between the apparent diffusion coefficient
.-.. .
-
.
.
.
deduced from an activation analysis and those that would have to be
.
. .
.
.
assumed if the same mechanisms were responsible for the release of
.
. .
.
stable Kr and Xe isotopes, as found by puncturing irradiated fuel
.
.
.
elements. It is therefore unwise to rely on apparent diffusion co-
.
. .
.
efficients as measured in postirradiation experiments."(21) Our in-
. .
.
pile work shows that, over a wide range of temperatures and fission
.
-
rate, the Pission gas release is not changed by a change of fission
rate (see Fig. 4). This result indicates that the fissioning process
retards the fission-gas migration. (If there were no effect of
fission rate, the amount of gas release would be directly proportional
to the production rate of the isotopes.)
As yet we do not have a mathematical model to predict the fission
product release from a fuel material during irradiation. We are work
ing on one that shows promise. (+5) The defect-trap model offers a quali-
tative explanation for our data and on this basis we present our con-
clusions below (formed from literature studies as well as our own work).
CONCLUSIONS
The Lission product release from UO is small for temperatures
well below the recrystallization temperature ( 1600°C). The escape of
Pission-gas is controlled by a knock-out process at temperatures be-
-
25.
en dienen een medisc
.
.
low 700°C, and, at temperatures of 700-1500°C, by the probability of
the gas being trapped in lattice defects. At higher temperatures,
12
movement of pores and grain boundaries control the fission-gas release.
Nearly total release of Pission gas should be assumed for irradiation
at temperatures above 1600°C.
Moderate burnup (<4 x 102° fissions/cm3) decreases the fission-
gas release rate by forming defects in the to, structure. These se-
fects along with the ones already present (such as grain boundaries
or closed pores) will act as traps for fission-gas. The trapping
probability is quite large, and at temperatures 2500°c only gas
near the vo, surface will ever escape.
The knock-out process will smooth microscopic irregularities in
the fuel surface thus reducing the total surface area of the fuel.
It should be noted that the relative effect of irradiation-produced
defects depends on the density of defects already present in the UO .
Also, the amount of smoothing will depend on the condition of the
original surface of the fuel.
For higher burnups, the retention of Pission-gas within the UO,
fuel depends on temperature and fuel structure. Lewis has concluded
"--dense sintered UO, will retain effectively all the fission-gas
produced at least up to 100,000 Mwd/tonne (2.5 x 102 fissions/cm3)
1f kept below perhaps 800°C. More definitely it can be stated that
retention is effectively complete to 15,000 Mwa/tonne up to 1200°c"10)
The temperature dependence of fission-gas release during irradia-
tion is caused by the escape probability of the trapped gas increas-
ing with temperature. Also, because irradiation-created defects are
. 26.
annealed, the trapping probability decreases as the temperature in-
creases.
A range of activation energies are observed because the es-
cape probability depends on the type and size trap as well as the tem-
perature. Since the activation energy is related to the nature of the
trap, the same activation energy applies to all inert gases. The gas
which is released at temperatures below 1500°C originates near the
surface of the fuel and therefore the activation energy is influenced
by the surface conditions of the fuel.
The fission rate will effect the behavior of the fission-gas. First,
the knock-out release will increase as lission rate increases. The migrat-
ing fission-gas will be effected also because an increase of fission
rate will increase the number of defect-traps per unit volume of fuel
(for a given temperature). An increase of fission rate will therefore
increase the probability of migrating fission-gas being trapped. However,
the release rate will not change much because the increased production of
fission-gas will compensate for the increased trapping (Fig. 4).
Bursts of fission-gas, emitted when to 18 heated, are caused
by annealing of defect-traps. Bursts of gas can also be caused by
stresses great enough to cause creep or by changes in oxygen-uranium
ratio. Specifically, any process involving rapid rearrangement of
the uo, structure will allow the migration or elimination of defect-
traps resulting in a burst of fission-gas.
27.
:
REFERENCES
1. W. H. Stevens and J. R. MacEwan, discussion following D. F. Toner
and J. L. Scott paper on Study of Factors Controlling the Release
of Xenon 133 from Bulk UO2, Symposium on Radiation Effects in
Refractory Fuel Compounds, ASIM Special Technical Publication
No. 306, p. 97, 1961.
2. J. R. MacEwan and W. H. Stevens, "Xenon Diffusion in UQ2: Some
Complicating Factors," J. of Nuclear Materials, Vol. 11, No. 1, :
1964.
3. C. Frigerio and T. Gerevini, On the Dependence upon the Irradiation
Exposure of the Apparent Diffusion Rates of Xenon in Uranium
Dioxide, FIAT, Sen, Italy, R, 69 bis (April, 1963).
4. R. M. Carroll, "Fission-Product Release from UO2," Nucl. Safety,
3(4): 356-360 (Summer, 1964).
5. L. R. Zumwalt et al., "Fission-Product Release From Monogranular
UC2 Particles," Nucl. Sci. Eng. 21(1): 1-12 (1965).
6. S..Yajima et al., Bulletin of the Chem. Soc. of Japan, 35(8): 1263
(1963).
7. T. Logerwall, P. Schmeling, "Interpretation and Evaluation of Non-
Noble Gas Release in Post-Activation Measurement," HMI-B27, Oct., 1963.
8. L. B. Lewis et al., "Fission-Gas Behavior in UO2 Fuel," Proc. 3rd UN
Int. Conf. Peaceful Uses of At. Eng., A/Conf. 28/P/19, Geneva (1964).
9. R. M. Carroll, P. E. Reagan, "Techniques for In-Pile Meas
Fission-Gas Release," Nucl. Sci. Eng., 21(2): 141-146 (1965).
10. R. M. Carroll, "Argon Activation Measures Irradiation Flux Continuously,"
Nucleonics, 20(2): 42 (Feb., 1962).
ments of
.."
28...?
REFERENCES (Continued)
12. R. M. Carroll, o Sisman, "In-Pile Fission-Gas Release from
Single-Crystal NO2," Nucl. Sci. Eng. 21(2): 147-158 (1965).
12. R. M. Carroll et al., "Fission-Gas Release During Fissioning
of UO2," Reactor Chem. Div. Ann. Progr. Rept. Jan. 31, 1965,
ORNL-3789 (in press).
13. R. M. Carroll et al., "Release of Fission-Gas During Fissioning
of U02," J. of Amer. Ceram. Soc., 48(2): 55-59, 1965.
L. E. J. Roberts, "The Behavior of UO2 and (U,PU)02 Fuel Materials
Under Irradiation," Proc. 3rd UN Int. Conf. Peaceful Uses of At.
Eng. A/conf. 28/P/155, Geneva (1964).
15. V. M. Golyanor, N. F. Pravdyuk, "Behavior of Nuclear Fuel Under
Irradiation," Proc. 3rd UN Int. Conf. Peaceful Uses of At. Eng.,
A/Conf. 28/P/338, Geneva (1964).
16. R. C. Daniel et al., "Effects of High Burnup on Zircaloy-Clad
Bulk UO2 Plate Fuel Element Samples," NAPD-236, Sept., 1962.
17. M. V. Reynolds, "Measurement of Free Fission-Gas Pressure in
Operating UO2 Filled Fuel Rods," Nucl. Sci. Eng. 20(4): 386-391
(1964).
18. R. B. Fitts et al., "Stress-Induced Release of Fission Gases
from UO2 and ThO2-U02," ORNL-IM-740, Jan., 1964.
19. M. D. Rogers and J. Adam. "Ejection of Atoms from Uranium by
Fission Fragments," J. of Nucl. Mat., 6(2): 182-189 (1962).
20. J. L. Bates et al., "Irradiation Effects in Uranium Dioxide
Single Crystal," HW-73959 (Spet., 1962).
21. A. Accary, R. Delmas, "Recent Developments in Refractory Fuels,"
· Proc. 3rd UN Int. Conf. Peaceful Uses of At. Eng., A/Conf. 28/F/59,
Geneva ( 1964).
L
.
ES
Liisa
407
END

DATE FILMED
9/9/65
...
..
.
::
.
i..
.
.
--
-
"
S
__
no
.
,