Jonathan Mckinley Holt JNM 2008

Available online at www.sciencedirect.
com
Journal of Nuclear Materials 372 (2008) 182214

www.elsevier.com/locate/jnucmat
In-reactor deformation of cold-worked Zr2.5Nb pressure tubes

R.A. Holt *
Department of Mechanical and Materials Engineering, Queens University at Kingston, Ont., Canada K7L 1W9
Received 20 March 2006; accepted 20 February 2007
Abstract
Over forty years of in-reactor testing and over thirty years of operating experience in power reactors have provided a broad under-
standing of the in-reactor deformation of cold-worked Zr2.5Nb pressure tubes, and an extensive data-base upon which to base models
for managing the life of existing reactors and for designing new ones. The eects of the major operating variables and many of the met-
allurgical variables are broadly understood. The deformation is often considered to comprise three components: thermal creep, irradi-
ation growth and irradiation creep. Of the three, irradiation growth is best understood it is thought to be driven by the diusional
anisotropy dierence (DAD). It is still not clear whether the enhancement of creep by irradiation is due to climb-plus-glide (CPG),
stress-induced preferred absorption (SIPA) or elasto-diusion (ED). The least understood area is the transition between thermal creep
and irradiation where the fast neutron ux may either suppress or enhance the creep rate. The three components are generally treated as
additive in the models, although it is recognized that this is only a crude approximation of reality. There are still signicant gaps in our
knowledge besides the thermal- to irradiation-creep transition, for example, the eect of Mo which is produced from Nb by transmu-
tation in the thermal neutron ux is not known, and on-going work is required in a number of areas. This paper reviews the current
state of knowledge of the in-reactor deformation of cold-worked Zr2.5Nb pressure tubes, and highlights areas for further research.
2007 Elsevier B.V. All rights reserved.
1. Introduction tured and operated to standards published by the Cana-

dian Standards Association [16].
The fuel channel of the CANDU1 nuclear power plant
comprises a Zr2.5Nb pressure tube and two steel end-t-
tings (in contact with the primary coolant), a Zircaloy-2 1.1. Fuel channel design
calandria tube (in contact with the moderator), and four
spacers maintaining the annular gap between pressure tube The core of a CANDU reactor is a cylindrical tank (cal-
and calandria tube. The pressure tubes are the primary andria) containing D2O moderator at 70 C. It is penetrated
containment of the high temperature D2O inside the core by 380 (675 MW units) or 480 (875 MW units) horizontal
of a CANDU reactor. They are subjected to high stresses, fuel channels, Fig. 1.1.1. The cold-worked Zr2.5 wt% Nb
temperatures and fast neutron uxes which cause changes pressure tube is 6 m long, 104 mm in diameter and has a
in the dimensions and material properties. In addition, ser- wall thickness of 4.2 mm. The chemical specication for
vice-induced wear occurs. To ensure the safe, reliable and the alloy is given in Table 1.1. The pressure tube contains
economic performance of the reactor, it is important that the natural UO2 fuel, encased in Zircaloy-4 sheathing, and
these changes are known and that the rate of change can heat transport uid, D2O, operating at temperatures from
be predicted and demonstrated to remain within the design 520 to 540 K at the inlet to 565585 K at the outlet. The
basis. The pressure tubes are designed, specied, manufac- inlet pressure is about 10.5 MPa and the outlet pressure is
about 9.9 MPa, resulting in an initial axial stress in the pres-
*
Fax: +1 613 533 6610.
sure tube wall of about 65 MPa and an initial hoop stress
E-mail address: holt@me.queensu.ca that varies from about 130 MPa at the inlet to about
1
CANDU is the registered trademark of AECL. 122 MPa at the outlet. Superposed on the axial stress from
0022-3115/$ - see front matter 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.jnucmat.2007.02.017
R.A. Holt / Journal of Nuclear Materials 372 (2008) 182214 183
Fig. 1.1.1. Schematic diagram of a CANDU fuel channel.
Table 1.1 damage and wear. In older CANDU units these changes
Chemical specication for Zr2.5Nb pressure tubes for CANDU reactors were not fully understood, and have required monitoring,
Element Limits (by weight) and in some cases maintenance. In newer CANDU stations
adequate allowances for them have been made for the
Niobium 2.42.8%
Oxygen 9001300 ppm design life of the reactor.
Nitrogen <65 ppm To assess their aging characteristics, pressure tubes are
Hydrogen Was <20 ppm, now <5 ppm subject to periodic non-destructive inspection, and material
Iron <1300 ppm surveillance, which, in Canada, requires a pressure tube to
be removed every three years from the unit with the highest
integrated neutron uence (lead unit), for evaluation of
the pressure is a relatively small axial component due to an fracture properties. These requirements are dened in
end-load from the out of core hardware that varies with Ref. [5]. The data that has been obtained from the periodic
time from tensile to compressive. The peak fast ux is up inspection and surveillance programs has been supple-
to 3.5 1017 n m2 s1, E > 1 MeV. The pressure tube and mented by in-service inspection programs, destructive
the annealed Zircaloy-2 calandria tube are separated by examination of tubes removed as a result of in-service
spacers and the annular gap is lled with owing CO2 to damage and irradiation of small specimens in test reactors
insulate the pressure tube from the cold moderator. The cal- [1013].
andria tube has a wall thickness of 1.4 mm and an inside
diameter of 129 mm. The pressure tubes are rolled into
403 stainless steel end-ttings at each end of the channel. 2. Deformation of pressure tubes
The end-ttings have mechanical closures to enable on-
power refueling and are connected to the heat transport sys- During reactor operation, the eects of temperature,
tem (HTS) by carbon steel feeders. stress and neutron ux change the dimensions of the pres-
The annulus gas system has been developed to detect the sure tubes. Irradiation and thermally induced deformation
presence of moisture if a leak develops in the primary pres- of fuel channel components will, in the absence of other
sure boundary inside the reactor core. The ability to detect mechanisms, eventually determine fuel channel life. The
moisture at an early stage of the formation of a sub-critical, dimensional changes in pressure tubes during normal reac-
through-wall crack in the pressure tube is an important tor operation are axial elongation, diametral expansion,
component in assuring leak-before-break of the tubes. sag and wall thinning. One can consider that there are three
mechanisms contributing to the dimensional changes: irra-
1.2. Aging mechanisms diation growth the change in shape at constant volume in
a fast neutron ux, thermal creep the change in shape due
Pressure tubes must accommodate the following aging to the eect of temperature and stress in the absence of a
mechanisms [79]: dimensional changes, material property fast neutron ux, and irradiation creep, the additional
changes, deuterium ingress due to corrosion and in-service change in shape due to stress and fast neutron ux [14].
184 R.A. Holt / Journal of Nuclear Materials 372 (2008) 182214
To a rst approximation the total in-reactor deformation assessing the microstructural and microchemical
rate is often expressed as the sum of these three terms, i.e., changes that occur during service [21];
and most recently:
e_ total
d e_ gd e_ icd e_ tc
d; 2:1
identifying the causes of variability between reactors and
where e_ is strain rate and the subscript d indicates the direc- between tubes in one reactor [22];
tion, and superscripts, g, ic and tc designate growth, irradi- creep at high stresses under multi-axial stress conditions
ation creep and in-reactor thermal creep, respectively. near aws [23,24];
To some extent, the creep and growth phenomena can the transition between thermal- and radiation-induced
be studied independently to determine their important creep near the ends of the tubes [25].
characteristics, however the real behaviour is more com-
plex than this, representing a continuum in a multi-dimen- For general reading on in-reactor deformation of zirco-
sional space of stress, fast neutron ux, temperature, fast nium alloys, see Refs. [26,27].
neutron uence and the microstructure of the material.
The changes in dimensions cause the stress state in the 2.1. Diametral expansion
pressure tube to vary with time, for example, while the
coolant pressure prole along the channel remains essen- Diametral expansion, which corresponds to an increase
tially constant with time, the internal diameter increases, in the circumference of the tube, occurs mainly by
and the wall thickness is reduced, resulting in an increase irradiation creep [14]. There is a small thermal creep com-
in the hoop stress. The change in the stress state must be ponent and the contribution of irradiation growth com-
taken into account to understand and predict the behav- ponent is actually negative, and therefore benecial. The
iour accurately. diametral strain is measured as mandated in Ref. [5].
The main trends of pressure tube deformation at normal For CANDU 6 reactors, approximately 90 diameter pro-
operating conditions were fairly well established by the les have been measured up to 123 000 full power hours
early 1980s based on various types of in-reactor testing in of operation, allowing statistical treatment of the data
the NRU Reactor at Chalk River and early diametral (examples of such data are given in Ref. [14]). A typical
and sag gauging results and elongation measurements from prole is shown in Fig. 2.1.1, and typical fast neutron ux
Pickering A [15]. In the past 20 years the emphasis has been and temperature proles are shown in Fig. 2.1.2. The
on: diametral strain is low near the ends of the tube where
the fast neutron ux is low, and peaks towards the outlet
extrapolating these results to newer reactors that oper- (hot) end.
ate at higher temperatures and with higher fast neutron An important factor aecting the shape of the diametral
uxes [16]; strain prole is the orientation of the coolant ow relative
making end-of-life predictions requiring long term irra- to the manufacturing direction of the pressure tube. Pres-
diations [12,17]; sure tubes are made using hot-extrusion, followed by
understanding sag of the pressure tube and the develop- cold-drawing and heating in a steam autoclave at 673 K
ment of contact between pressure tubes and calandria for 24 h [28]. In the hot-extrusion process the end of the
tubes [1820]; tube emerging from the extrusion press rst is referred to
106.6
106.1
MEAN DIAMETER (MM)
105.6
105.1
104.6
104.1
103.6
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5
Axial Position Referenced to Outlet Rolled - Joint - m
Fig. 2.1.1. Typical prole of diameter from outlet to inlet along the length of a CANDU 6 pressure tube.
300 4 60
Flux - n.m .s x10 , E>1MeV

15.33
290
50
Temperature - C
28.6
3 48.22
280
-17
57.38
40
Strain - 10 -4
Flux 71.13
270 2
-1
30
260
-2
Temperature
1
250 20
240 0 10
0 1 2 3 4 5 6
Distance from the PT Inlet End - m
0
Fig. 2.1.2. Typical fast neutron ux and temperature proles from inlet to 0 10 20 30 40 50 60
outlet along a CANDU pressure tube. Time - kh
Fig. 2.1.4. Strain as a function of time for the fast neutron uxes shown
as the front end and the end emerging last is referred to as (in units of 1014 n m2 s1, E > 1 MeV) showing an initial transient strain
followed by a steady state rate, from Ref. [25].
the back end. The front ends of the tubes exhibit lower
yield and ultimate tensile strengths than the back ends.
In earlier reactors the pressure tubes were always experiments on small specimens [30] that the irradiation
installed in the same geographic orientation relative to creep rate changes with uence, and clear evidence that
reactor core, e.g., the back end of the tube at the west the growth rate does change [12], so reactor operators
end of the reactor. Since the coolant ow is in the reverse should remain vigilant.
direction in alternating channels, half the pressure tubes The design of the fuel channels accounts for stress rup-
then had their back ends at the outlet, and half had their ture/creep ductility and ow bypass. Based on data from
back ends at the inlet. This resulted in two types of diam- in-reactor experiments, 5% is considered to be a very
etral strain proles. Tubes with their back end at the outlet conservative limit with respect to creep rupture and creep
exhibited asymmetrical strain proles like that shown in ductility, Ells et al. [31], see Section 7. Although the mea-
Fig. 2.1.1, and tubes with their back end at the inlet exhib- sured diametral expansions are within the maximum value
ited a more symmetrical strain prole, Fig. 2.1.3. assumed in the analysis, lifetime management strategists
In later reactors tubes, were installed with their back are evaluating the need for increased inspections to pre-
ends at the outlet of all channels, to reduce the likelihood cisely quantify the variability in expansion. This will allow
of a failure mechanism called delayed hydride cracking more accurate assessments of eect of ow bypass on fuel
(DHC) at the rolled joints where the pressure tubes are cooling and the potential approach to fuel dry-out the
joined to the steel end-ttings [29]. In the most recently condition limiting the operating power of the reactor.
built reactors, this concern has been alleviated and the The variability of diametral strain is also important
tubes are installed with the back end at the inlet. because operating margin to fuel dry-out is based on the
After a small positive transient, Fig. 2.1.4, the diametral worst case. A new fuel design has been developed to ame-
strain rate at a given point along the pressure tubes appears liorate the eects of diametral creep on fuel cooling [32],
to reach a steady state, and, unlike the situation for elonga- which, if implemented, would delay any necessity to de-rate
tion, see Section 2.2, there is no evidence from power reac- reactors as a result of degraded operating margins.
tors that this rate changes with time throughout the life of
the reactor. There is some suggestion from test reactor
2.2. Axial elongation
0.012
Axial elongation is due to a combination of irradiation
Diametral strain after 117000
creep and irradiation growth [14,33]. Both of these phe-

0.01
nomena are anisotropic, and result directly from the aniso-
0.008
tropic crystal structure (hexagonal close packed or hcp) of
efph
0.006 zirconium, and the pronounced crystallographic texture

0.004
Back End of PT at Inlet
produced during manufacturing. If Zr2.5Nb were isotro-
0.002 Back End of PT at Outlet pic, there would be no signicant pressure tube elongation.
0
The designers of the earliest CANDU stations assumed
0 1 2 3 4 5 6 7 that zirconium alloys were isotropic and made no allow-
Distance from Inlet m ance in the design for elongation beyond thermal and
Fig. 2.1.3. Comparison of diameter proles along CANDU pressure tubes elastic expansion. Once pressure tube elongation was dis-
with the back end at the inlet and the back end at the outlet, from Ref. covered, it was assumed that it would occur at a steady rate
[14]. (several millimeter per year), after a short initial transient
120 30
100 20
Elongation (mm)
80
Deflection - mm
10
60
0
40
0 150 300 450 600
20 -10
0 -20
-20
-30
0 2 4 6 8 10 12 14
25 2
Fluence (10 n/m ) -40
Distance From Inlet - m
Fig. 2.2.1. Elongation as a function of fast neutron uence (averaged
along the length of each pressure tube) for a CANDU 6 reactor. Fig. 2.3.1. Sag prole for a CANDU fuel channel after several years of
operation.
[15], and much reactor maintenance is still planned based

upon this assumption. However, it is now known that the
tions and the pressure comes in contact with the calandria
elongation rate increases gradually with time [7,14], or
tube. Fuel bundle passage, as proven by tests using pre-
more correctly with fast neutron uence, Fig. 2.2.1. This
dicted end of life curvature, is not impaired during the fuel
increase in rate is attributed to evolution of the microstruc-
channel design life [7].
ture induced by the fast neutron ux [12].
In the design, sag is assessed to address the possibility of
Pressure tube elongation is accommodated by support-
contact between the calandria tube and horizontal mecha-
ing the end-ttings on sliding bearings. One end is initially
nisms that are perpendicular to the fuel channels: the
xed and the other end is free to move outwards. About
liquid-injection-shutdown-system nozzles and horizontal-
halfway through the design life, the channel is recong-
ux-detector-unit guide tubes [7]. Deformation predictions
ured, i.e., the free and xed ends are reversed, allowing
indicate that, although unlikely, such contact could occur
elongation to be accommodated at the other end of the
before 30 years of reactor operation in earlier reactors.
reactor [7].
Measurements are planned on lead units to verify the cal-
Two aspects of elongation are important: the magnitude
culations, and a modication is available to address the
and the dierential between channels. Monitoring is neces-
possibility that contact occurs. The newest units have more
sary to establish the optimum time to recongure and to
sag resistant calandria tubes, and greater clearances
ensure deformation continues at the established rate. The
designed to accommodate the predicted sag with larger
dierence in elongation rates between neighboring channels
margins [7,35].
is also monitored so that interference between feeders or
In early units, the loose-tting spacer design was suscep-
problems with fueling machine access do not occur.
tible to movement during reactor assembly and early oper-
Recently, some progress has been made in understanding
ation [29]. Because of the resulting wide span between some
the sources of the variability in terms of the manufacturing
spacers, pressure tube to calandria tube contact could
variables and their eect on the microstructure [34].
occur. This could lead to the potential for hydride blister
Life management strategies currently involve measure-
formation and growth if the hydrogen concentration at
ment and calculation of the elongation rate of each channel
the contact location exceeds the blister formation thresh-
to determine when maintenance is required. The fueling
old (BFT) [36]. A specialized spacer location and reposi-
machine, in combination with specialized instrumentation,
tioning (SLAR) tool was developed to locate the spacers
can be used to provide this elongation data on-line, see
and relocate them, if necessary [37]. This maintenance
Fig. 2.2.1. The newest CANDU stations have been
operation is being carried out before BFT is reached. The
designed with increased elongation margins to account
operation is controlled by an operator using the computer
for the gradual acceleration of the rate [7].
code SLARON/SLARADE, Badie et al. [20], based on
measured spacer locations and model calculations of the
2.3. Sag sag phenomenon. In 1983, the spacer design was changed
from loose-tting Zr2.5Nb0.5Cu to tight-tting Inconel
Sag occurs mainly by irradiation creep from the weight 600 spacers that do not move [38].
of the fuel and heavy water in the pressure tube, Fig. 2.3.1. The models used to predict sag and control SLAR are
Gross sag deformation is primarily controlled by the rela- still under development. These models are essential because
tively cool calandria tube [18,19]. Sag of the pressure tube the gross sag of the fuel channel is mainly controlled by the
occurs between the spacers, but is only of consequence if calandria tube, and hence sag measurements only give very
spacers are suciently displaced from their design loca- crude information about the sag of the pressure tube per se.
Fig. 3.2(a). The elongated a-grains tend to be somewhat

4.5
ner at the back ends of the tubes, Table 3.1 [22]. Occasion-
ally, especially at the front ends of the tubes there are col-
Wall Thickness - mm
4.4
onies of misaligned a-grains, which result from the b- to
4.3 a-phase transformation during cooling after extrusion,
Fig. 3.2(b) [40]. The a-grains contain 0.61.0%Nb, a
4.2 slightly super-saturated solid solution which is thermally
stable because the transformation kinetics are sluggish [42].
4.1
Between many of the a-grains, there is a thin lm of b-
phase, perhaps 2050 nm thick, see Figs. 3.1 and 3.2. After
4
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 extrusion, this phase is metastable body centred cubic (bcc)
Axial Position Referenced to the Outlet End-Fitting - m b-Zr containing approximately 20%Nb, but after cold-
work and autoclaving, this phase has partially trans-
Fig. 2.4.1. Variation in wall thickness (maximum, average and minimum)
of a CANDU pressure tube with distance from the outlet end-tting. formed, through a complicated sequence, towards the equi-
librium phase distribution of a (<0.5%Nb) and bcc b-Nb
(95%Nb) and contains a combination of x (a metastable
The further development of these models and probabilistic hcp phase of intermediate Nb content) and b-Zr enriched
assessment of the likelihood of contact are now probably in Nb to 50% [43]. This partial transformation is recog-
the most important focus of on-going research. nizable by the internal structure of the b-phase observed
after autoclaving, but not immediately after extrusion,
2.4. Wall thinning Fig. 3.3.
Oxygen, which is deliberately controlled to give short
Creep and growth occur at constant volume, therefore term strength [28] is concentrated in the a-phase and
the increase in diameter and length correspond to a reduc- depleted in the b [40], although there is no technique that
tion in the wall thickness. Sag also contributes a small will give a quantitative assessment on such ne microstruc-
amount to the loss of wall thickness on the side opposite tures. The major impurity in Zr2.5Nb is Fe which is a
the centre of curvature. A typical wall thickness prole at b-stabilizer in Zr, and is primarily concentrated in the
mid-life is shown in Fig. 2.4.1. The design of the fuel chan- b-phase. Fe is also found at the aa grain boundaries [44].
nels accounts for the stress increase both due to diametral The dislocation structure in the a-phase arises mainly
expansion and due to wall thinning. Relative to the short from the cold-drawing process (27% reduction in area),
term strength, increased stresses due to diametral expan- and subsequent recovery during autoclaving [45]. It com-
sion and wall thinning are more than oset by the radiation prises both a dislocations, i.e., dislocations with a Burgers
hardening that occurs during service. The recently devel- vector lying in the basal plane of the hcp crystal, and c-
oped capability to measure wall thickness proles accu- component dislocations, i.e., dislocations with a compo-
rately, as shown in Fig. 2.4.1 allows a more accurate nent of their Burgers vector normal to the basal plane.
analysis of the diametral strain proles and elongation with The two types can be observed by transmission electron
the possibility of detecting variations in anisotropy along microscopy, Fig. 3.4 [46]. Using X-ray diraction [47] the
the length of the pressure tube. total dislocation is estimated as 34 1014 m2. and typi-
cally the c-component dislocation density is 1020% of
3. Microstructure and microchemistry of pressure tubes the total [22].
The anisotropic nature of the in-reactor deformation
The manufacturing route of Zr2.5Nb pressure tubes processes in pressure tubes results predominantly from
has varied slightly between the construction of Pickering the anisotropic hcp a-Zr crystal structure and the crystallo-
A (late 1960s) and Qinshan 1 and 2 (early 2000s), but the graphic texture developed during the extrusion process.
microstructure of the pressure tubes has remained reason- The anisotropy is mainly controlled by the orientation of
ably consistent. The basic microstructure and crystallo- the c-axis, or basal plane normal. This is the strong direc-
graphic texture arise during the extrusion process (for tion of the crystal [48]. A typical basal pole gure for a
summaries of the manufacturing route of pressure tubes pressure tube is shown in Fig. 3.5. The basal plane normals
see Refs. [28,39,40]). It comprises mainly hcp a-Zr grains are distributed in the radial/transverse plane and concen-
that are highly elongated in the axial direction of the tube, trated predominantly in the transverse direction [28]. This
as seen in the radial/axial plane, Fig. 3.1(a). In this view concentration tends to be higher at the front ends of the
they are generally between 0.2 and 0.5 lm thick and several tubes. Often there is a small concentration of basal plane
micrometer long. As seen in the radial/transverse plane normals in the axial direction of the tube at the front end
there are three types of grains [41]. Usually they are a associated with the presence of colonies of misaligned a-
few micrometer wide, and are predominantly aligned in grains [49], mentioned above. The basal plane texture is
the transverse direction but often curl around, often quantied by FR, FT and FA, the resolved fractions
Fig. 3.1(b), and a few are aligned in the radial direction, of basal plane normals in the radial, transverse and axial
Fig. 3.1. Transmission electron micrographs from carbon replicas of a pressure tube showing: (a) elongated a-grains in the axial/radial section and
(b) curved attened a-grains in the radial/transverse section, from Ref. [41].
directions, respectively [50]. These quantities represent a tions have been established between the overall diametral
normalized average of a second-order tensor resolution strain [22] and elongation [34] of pressure tubes and micro-
into the tube directions of dilations along the c-axis for structural, chemical and manufacturing information.
all crystallographic orientations [51]. Typical values for The elongation rate of pressure tubes shows a decrease
pressure tubes are given in Table 3.1 [22]. with increasing Fe concentration, Fig. 4.1 [52]. A similar
Unless otherwise specied, the deformation characteris- correlation is seen between the rate of irradiation growth
tics described below pertain to material with the micro- and Fe concentration, Section 5.4. The elongation rate also
structural characteristics given above, and to operating appears to vary from one ingot to the next [53], Fig. 4.2,
conditions that are representative of the behaviour of pres- which could be attributable to ingot chemistry, or could
sure tubes in power reactors. be a manufacturing batch eect [34].
The diametral strain rate decreases with grain thickness,
4. Variability of pressure tube deformation Fig. 4.3, decreases as the distribution of basal plane nor-
mals becomes more pronounced in the transverse direction,
Within one reactor, the deformation rates of pressure Fig. 4.4, and decreases as the oxygen concentration (as
tubes can vary substantially once temperature, fast neutron analysed in the ingot) increases, Fig. 4.5 [22].
ux and tube orientation have been taken into account. In In all cases these correlations may be misleading because
the case of elongation, this variation is of the order of there are many potentially confounding variables.
1520% [33] and in the case of diametral strain the varia- The eect of grain size might be expected from the the-
tion can be of the order of 30% [22]. Further variations are oretical predictions of the eect of grain size on irradiation
seen from reactor to reactor, but these might be attributable growth in the transverse direction [54], see Section 5.6. The
to dierences in operating conditions. A number of correla- eect of texture is expected from its eect on both creep
Fig. 3.2. Transmission electron micrographs from thin foils of a pressure tube showing: (a) transversely oriented a-grains in the radial/transverse section
and (b) a radially oriented a-grain in the radial/transverse section. Both micrographs show some Widmanstatten a-grains (centre in (a) and upper right in
(b)), from Ref. [41].
Table 3.1 placement of atoms from their normal position in the

Front-to-back variation in microstructural parameters for pressure tubes absence of applied stress. There is no volume change a
(1 standard deviation) radiation-induced shape change accompanied by a volume
Parameter Front Back increase is referred to as irradiation-induced swelling, and
FR 0.298 0.025 0.350 0.029 is not signicant in zirconium alloys at the temperature of
FT 0.642 0.031 0.605 0.032 operation of thermal reactor components. Irradiation
FL 0.057 0.017 0.05 0.05 growth is relatively easy to study by measuring the dimen-
Grain thickness (lm) 0.45 0.08 0.34 0.06
sional changes of unstressed specimens as a function of irra-
diation time, or fast neutron uence. The irradiation growth
of Zr2.5Nb pressure tube materials at 525585 K the
and growth [55]. The eect of O concentration could be
operating temperature range of CANDU pressure tubes
related to the solid solution strengthening eect of O in
depends upon their initial microstructure crystallographic
Zr [56], or the eect of O on the phase distribution during
texture, grain size, dislocation density fast neutron ux,
extrusion which could, in turn aect the texture. The eect
temperature, fast neutron uence (which induces changes
of Fe on elongation may be related to its eect on growth.
in the dislocation structure) and chemical composition.
5. Irradiation growth 5.1. Eect of fast neutron uence
Irradiation growth is a shape change induced by any The growth behaviour of Zr2.5Nb pressure tube
energetic radiation that causes displacement damage dis- material in the axial and transverse directions at 550 K is
Fig. 3.3. Transmission electron micrographs from thin foils of a pressure tube showing: (a) undecomposed b-phase before autoclaving and (b) partially
decomposed b-phase after autoclaving for 24 h at 673 K, both in radial/axial section, from Ref. [43].
Fig. 3.4. Transmission electron micrographs from thin foils of a pressure tube showing: (a) a-dislocations b 1=3h1 1 2 0i and (b) c-component
dislocations b 1=3h1 1
2 3i, both in radial/transverse section, from Ref. [46].
A 1.6
Wolsong PT
Relative Strain Rate

1.4
1.2
0.8
11
1 2 3 45 7 9 0.6
R + T
0.4
0.1 0.2 0.3 0.4 0.5 0.6
Mean Grain Width - m
Fig. 4.3. Correlation between peak diametral strain rate of pressure tubes
in operating reactors and grain thickness, from Ref. [22].
Maximum 12
2
Fig. 3.5. Typical basal pole gure of a cold-worked Zr2.5Nb pressure 1.8

tube, from Ref. [17].
1.6
1.4
16 1.2
Elongation Rate - x10 -29 m 2.n
1
14
0.8
12
0.6
10 0.4
8 0.2
0
6
0 0.1 0.2 0.3 0.4 0.5
4 F T - FR
2
0 in operating reactors and the dierence between texture parameters
0 500 1000 1500 2000 2500 3000
FT FR, from Ref. [22].
Iron Content - ppm
Fig. 4.1. Correlation of in the elongation rate of pressure tubes in a

CANDU reactor with the ingot iron analysis, from Ref. [52]. 2
120 1.5
Total elongation (mm)
399149
100
398757 1
80 399012
60 0.5
40
20
0
800 1000 1200 1400 1600
0 Oxygen Concentration - ppm
0 2 4 6 8 10 12 14
Fluence (x10 25 n/m 2 ) in operating reactors and the ingot oxygen analysis for two reactor types,
from Ref. [22].
Fig. 4.2. Elongation as a function of fast neutron uence (averaged along
the length of each pressure tube) for tubes made from three dierent ingots
in a CANDU 6 reactor.
transverse direction of about 0.02%. Subsequently, the
growth is relatively at, or slightly negative in the axial
illustrated for low and high fast neutron uence in Figs. direction for a fast uence of 11.5 1025 n m2,
5.1.1 [57,58] and 5.1.2 [12]. There is an initial rapid increase E > 1 MeV, and establishes a relatively steady negative rate
in length in the axial direction and decrease in length in the in the transverse direction after about 0.5 1025 n m2,
0.15 1.5
Axial
0.1 1 Radial
Growth Strain - %
Growth Strain - %
Transverse
Volume
0.05 0.5
0
0
-0.05
-0.5
-0.1
-1
-0.15
0 1 2 3 4 5 -1.5
-25 -2 0 5 10 15 20
Neutron Fluence x 10 - n.m , E>1MeV
-25 -2
Neutron Fluence x 10 n.m , E>1MeV
Fig. 5.1.1. Growth strain as a function of fast neutron uence for cold-
worked Zr2.5Nb in DIDO at a fast neutron ux of 6 1017 n m2 s1, Fig. 5.1.3. Growth strains in three dimensions as a function of fast
E > 1 MeV, from Ref. [58]. neutron uence for Zr2.5Nb pressure tube material irradiated at 520 K in
Osiris at 2 1018 n m2 s1, E > 1 MeV, from Ref. [59]. Note several
points in the range 913 1025 n m2 are thought to be erroneous.
1
0.8 553 K the strain rates in all three directions evolve with fast u-
0.6
ence, Fig. 5.1.3 [59], and the strain in the thickness direc-
Growth Strain - %
Axial 583 K tion can be inferred to be the negative sum of the strains
0.4 in the axial and transverse directions.
0.2
0 5.2. Eect of fast neutron ux

583 K
-0.2
Above a certain temperature, the irradiation growth
Transverse
-0.4
553 K
strain rate is approximately proportional to the eective
-0.6
displacement damage rate, i.e., the rate at which atoms
0 5 10 15 20 are displaced from their lattice positions by the fast
Neutron Fluence x 10
-25 -2
- n.m , E>1MeV neutrons. Linear ux dependence is demonstrated by the
irradiation of similar materials at dierent ux levels. There
Fig. 5.1.2. Growth strain as a function of fast neutron uence for cold-
is a complication in that dierent irradiation facilities may
worked Zr2.5Nb in Osiris at a fast neutron ux of 2 1018 n m2 s1,
E > 1 MeV, from Ref. [12]. have dierent ux spectra, and therefore dierent degrees
of eectiveness at producing displacement damage. There
are no good data on ux dependance for Zr2.5Nb pres-
E > 1 MeV. After these initial transients, growth is typi- sure tube materials at 525585 K, and there is a signicant
cally positive in the axial direction of the tube, and negative discrepancy between data obtained in the DIDO reactor in
in the transverse direction. The magnitude of growth in the the UK and the Osiris reactor in France [52]. However,
transverse direction is approximately half that in the axial when normalized for ux spectrum, the Osiris data agree
direction. well with data for specimens irradiated in CANDU in a
In the longer term, the growth rate gradually accelerates carrier bundle specically designed for the purpose of
in both directions and after a fast neutron uence about irradiating small specimens in power reactors, Fig. 5.2.1
2 1025 n m2, E > 1 MeV, the behaviour is reasonably [60], so it appears that the ux dependence is indeed
well described by a quadratic equation. The acceleration approximately linear. Also, it has been clearly demon-
continues, but remains gradual until fast uences of at least strated that the ux dependence is linear at 330350 K
2.4 1026 n m2, E > 1 MeV [12], i.e., approaching those for cold-worked Zircaloy-2 pressure tube material irradi-
experienced over their design life by CANDU pressure ated in the ATR Reactor in the US at 24 1018 n m2 s1,
tubes. The rate changes can be related to the changes in dis- E > 1 MeV, and in DIDO in at a ux of about 1/4 of that
location structure with uence [12], in particular an [55]. Theoretically, a linear ux dependence at this temper-
increase in the density of c-component dislocations, see ature always implies a linear ux dependence at higher tem-
Sections 9 and 10. peratures [54], see Section 10.
Often, the rate is assumed to be linear over the later part Linear fast ux dependence appears to be a consequence
of the uence range and growth rates are compared on of the partitioning of the point defects created by the dis-
this basis. Measurements of the growth of compact tough- placement damage amongst various sinks in the micro-
ness specimens in three dimensions at about 520 K, has structure with little recombination, see Section 10. The
conrmed that there is negligible volume change, so that implication of a linear fast ux dependence is that growth
0.12 0.8
Osiris
Bruce A
0.1 DIDO 0.6
NRU
Growth Strain %
Growth Strain - %
0.08
0.4
0.06
0.2
0.04 Medium Fe
0 Low Fe
0.02 High Fe
-0.2
0 520 540 560 580 600
0 0.2 0.4 0.6 0.8 1 1.2 1.4 Temperature - K
-25 -2
Neutron Fluence x10 , n.m , E>1MeV
Fig. 5.3.2. Growth strain in the axial direction of cold-worked Zr2.5Nb
Fig. 5.2.1. Growth strain in the axial direction of cold-worked Zircaloy-2 pressure tube material at a fast neutron uence of 1 1026 n m2,
pressure tube material at 553 K in DIDO at 6 1017 n m2 s1, E > 1 MeV, as a function of temperature, from Ref. [12].
E > 1 MeV; 553 K in Osiris at 2 1018 n m2 s1, E > 1 MeV, and in a
carrier bundle in a CANDU reactor at 2 1017 n m2 s1, E > 1 MeV,
from Ref. [60].
0
strain can be plotted vs fast uence, independent of the fast -0.05

ux an almost universal practice. The deformation rate is
Growth Strain%
then expressed in units of m2 n1, E > 1 MeV.
-0.1
5.3. Eect of temperature -0.15
Medium Fe
Typical irradiation growth curves for pressure tube -0.2
Low Fe
materials at 553 and 583 K are shown in Fig. 5.1.2, in High Fe
which the magnitudes of the positive axial and negative -0.25
transverse growth strains both decrease with increasing 540 560 580 600
temperature. Over a temperature range of 330623 K, the Temperature - K
growth rate of Zr2.5Nb pressure tube material in the axial
Fig. 5.3.3. Growth strain in the transverse direction of cold-worked Zr
direction decreases with temperature, Fig. 5.3.1 [58,61,62]. 2.5Nb pressure tube material at a fast neutron uence of 1 1026 n m2,
Over the temperature range of pressure tube operation, the E > 1 MeV, as a function of temperature, from Ref. [12].
same behaviour is seen for a variety of dierent pressure
tube materials, Fig. 5.3.3, the curves being displaced for
materials with dierent Fe concentrations (see Section 0.15
5.4). The (negative) magnitude of the transverse growth
553 K
rate also decreases with temperature, for a range of tubes, 0.1
Growth Strain - %
623 K
0.05
10
Strain Rate x 10 29 m 2n-1 (E>1 MeV)
5 -0.05
-0.1
0 1 2 3 4 5
-25 -2
0 Neutron Fluence x 10 , n.m , E>1MeV
Fig. 5.3.4. Growth strain as a function of neutron uence showing the

eects of temperature changes between 553 K and 623 K for cold-worked
Zr2.5Nb pressure tube material irradiated in DIDO at
-5
300 350 400 450 500 550 600 650 6 1017 n m2 s1, E > 1 MeV, from Ref. [58].
Temperature (K)
Fig. 5.3.1. Eect of temperature on the growth rate of cold-worked Zr Fig. 5.3.4. When the parameters of quadratic equations t-
2.5Nb pressure tube material, from Ref. [61]. ted individual growth curves are plotted as a function of
temperature, it appears that the temperature aects primar- provide recombination sites for vacancies and interstitials.
ily the initial growth rate, but has less eect on the rate of Also, its presence at aa grain boundaries [44] could inu-
acceleration. ence their role as sinks. Finally, Fe changes the volume
When the temperature is changed during a growth test, fractions of a and b during extrusion, and could thus have
there are signicant transients, Fig. 5.3.4, and the growth an indirect inuence through the texture.
rate does not immediately re-establish itself at the expected
rate for the new temperature [58]. Analysis of the anisot- 5.5. Eect of crytallographic texture
ropy of these transients suggested that there may be a long
lasting eect of the dislocation microstructure developed at Based on the growth behaviour of a typical pressure
the earlier irradiation temperature. This could be impor- tube material, Section 5.1, it can easily be inferred that,
tant in future reactors that are designed to load follow with on the crystalline level, growth comprises, to some degree,
signicant changes in temperature over the power range. shrinkage along the c-axis and expansion along the a-axis
of the hcp crystal. Growth tests on small diameter tubes
with two dierent crystallographic textures irradiated to
5.4. Eect of Fe concentration
high fast neutron uence in the Osiris reactor in France
clearly illustrate this [17]. Fig. 5.5.1 shows the crystallo-
Based on the fact that Fe has a signicant eect on the
graphic textures of the two types of tubes referred to as
diusional characteristics of vacancies in Zr [63,64], and
FS (fuel sheathing) and MPT (micro-pressure tube), and
the observation that the elongation rates of pressure tubes
Table 5.5.1 gives the microstructural parameters.
correlate with Fe concentration (see Section 4), pressure
Fig. 5.5.2 illustrates the growth behaviour in two directions
tubes with dierent Fe concentrations were chosen for a
for both types of tubes at 553 K. Both FS and MPT mate-
long term growth experiment in the Osiris reactor in
rial grow a similar amount in the axial direction, in which
France [12,52]. These showed a clear trend, with the mag-
there are very few basal plane normals. The MPT material
nitude of the irradiation growth rate in either direction
exhibits a much greater shrinkage in the transverse direc-
decreasing with increasing Fe concentration. Fig. 5.4.1
tion than FS material, corresponding to the much higher
shows growth curves for pressure tube materials at a nom-
concentration of basal plane normals in the transverse
inal temperature of 553 K with two dierent Fe concentra-
direction in MPT material than in FS material.
tions. Fig. 5.3.3 shows the eect of Fe on the axial growth
At the single crystal level, this behaviour corresponds to
rate as a function of temperature, and Fig. 5.3.4 the eect
a net accumulation of vacancies at defects (dislocations,
of Fe on the transverse growth rate as a function of temper-
dislocation loops) lying on the basal plane and an accumu-
ature. As in the case of temperature, the initial rate is more
lation of self-interstitial atoms (SIAs) at defects perpendic-
sensitive to Fe concentration than the acceleration in the
ular to the basal plane, as originally proposed by Buckley
growth rate with increased uence, Fig. 5.3.4. The eect
[65]. As seen in the next section, however, there are other
of Fe has not yet been satisfactorily explained. During irra-
components to the behaviour which is clearly modied by
diation it is dispersed into the a-phase [44], but whether it
very ne elongated grain structures. In addition, we now
remains in solution or exists as sub-microscopic precipi-
know that the dislocation structures in dierent crystal ori-
tates is not known. As a solute, it may aect the vacancy
entations is dierent [12] and this will no doubt aect the
diusion characteristics [63], while as a precipitate, it might
growth anisotropy [66].
1.2
5.6. Eect of dislocation density and grain size
1 Low Fe (<500ppm)
High Fe (~1000ppm)
There are no growth data for Zr2.5Nb pressure tube
0.8 material in which the initial dislocation density2 has been
Longitudinal
Growth Strain - %
0.6 varied independently of other parameters notably the

0.4
grain size and the condition of the b-phase. Similarly, there
is no growth data for which the grain size has been varied
0.2
independently of other parameters. The work on Zircaloy-
0 2 and Zircaloy-4 [55] has clearly demonstrated that the
-0.2 growth rate increases approximately linearly with increas-
-0.4 Transverse
ing dislocation density, and also probably depends upon
the types of dislocations present (a-type vs c-component),
-0.6
0 5 10 15 20
and whether or not the material has been stress relieved.
-25 -2 Growth rates of Zircaloy-2 and Zircaloy-4 pressure tubes
Neutron Flux x 10 , n.m , E>1MeV
2
Fig. 5.4.1. Growth strain as a function of fast neutron uence for cold- Here, we are referring to the dislocation density of the material prior to
worked Zr2.5Nb with dierent Fe contents in Osiris at a fast neutron ux irradiation, e.g., due to cold-work, and not the dislocation structure
of 2 1018 n m2 s1, E > 1 MeV. induced by irradiation, as discussed in Section 9.
A A
Micro PT Fuel Sheathing
1 1
3 2
3 4
2 3 4 5 6 7 8 R + T
R + T
+ Maximum 9 + Maximum 5.3

Fig. 5.5.1. Basal pole gures for small tubes irradiated in Osiris, from Ref. [17].
Table 5.5.1 A deliberate attempt to take advantage of these charac-

Microstructural parameters for Zr2.5Nb fuel sheathing and micro- teristics was made with the development, in the early 1980s
pressure tubes irradiated in Osiris
of modied pressure tube manufacturing routes, referred to
Parameter Fuel sheathing Micro-pressure tube as the Task Group 3 (TG3) program [39,57,58]. In this pro-
2
a-type dislocation density 14
2.3 10 m 3.4 1014 m2 gram, pilot orders of pressure tubes were made with the
c-component dislocation 0.6 1014 m2 0.7 1014 m2 target of modifying the grain shape, grain size and disloca-
density
tion density. In addition standard pressure tube materials
FR 0.57 0.34
FT 0.37 0.57 were modied by cold-working and heat treatment to
FL 0.07 0.09 change the dislocation density and grain size. Irradiation
growth tests were carried out to uences of 6
8 1025 n m2, E > 1 MeV [58], Table 5.6.1. Of seven
materials tested, seven had a grain size between 0.31 and
0.37 lm within the typical range for pressure tubes, with
0.8
dislocation densities varying between 1.7 and
0.6
4.8 1014 m2. These show an approximate linear depen-
0.4 dence upon the dislocation density, Fig. 5.6.1. It should
Growth Strain - %
0.2 be kept in mind that other variables (e.g., the degree of

transformation of the b-phase) may be at play.
0
0 5 10 15 20 In one case, referred to as TG3 R1, the dislocation den-
-0.2 sity was reduced to 2.4 1014 m2 by a high temperature
-0.4 Axial MPT
stress-relief, and the corresponding reduction in tensile
-0.6
Transverse MPT strength compensated for by reducing the grain size to
Axial FS
Transverse FS 0.27 lm. This resulted in an irradiation growth rate that
-0.8
was actually negative in the axial direction while still
-1 remaining negative in the transverse direction, Fig. 5.6.2
-25 - -2
Neutron Fluence x 10 n.m [58,62]. If the growth rate data are corrected to a dislocation
Fig. 5.5.2. Growth strains for small tubes with two dierent crystallo- density of 4 1014 m2 using the slope of Fig. 5.6.1, and
graphic textures irradiated at 553 K in Osiris at a neutron ux of adjusted to a grain aspect ratio of 15:1 and an FA of 0.03
2 1018 n m2 s1, E > 1 MeV, from Ref. [17]. using the growth anisotropy factor described in [54], the
grain size sensitivity can be plotted, Fig. 5.6.3. This shows
a very rapid increase from negative to positive growth rates
derived from the total strain in power reactors also varied between 0.27 and 0.31 lm, followed by a saturation between
approximately linear with dislocation density [67]. Grain 0.37 and 1.5 lm. Again there may be other factors at play,
size, per se (typically in the range >5 lm for the Zircaloys), but both Figs. 5.6.1 and 5.6.3 are remarkably similar to
appeared to have little eect, but the anisotropy of growth the theoretical curves for the sensitivity of axial growth to
was also inferred to depend upon the grain shape. dislocation density and grain size [54], see Section 10.
Table 5.6.1
Microstructural details and long term irradiation growth rates at 553 K of Zr2.5 Nb pressure tube materials modied to vary grain size, grain shape and
dislocation density [61]
Material Dislocation density FR FT FA Grain aspect Grain Axial growth Transverse
designation ( 1014 m2) ratioa thicknessa (lm) rateb growth rateb
AR 4.2 0.36 0.61 0.03 1520 0.32 3.2
SR 1.8 0.36 0.61 0.03 1520 0.32 0.8
CW 4.7 0.28 0.64 0.08 2.3 1.5 3.7
R1 2.4 0.40 0.56 0.03 1520 0.25 2.1 0.7
R2 4.8 0.37 0.59 0.03 510 0.34 3.2 1.4
R3 2.7 0.36 0.61 0.03 510 0.37 2.3 1.5
D1 4.5 0.35 0.62 0.04 1520 0.31 2.7 2.2
a
Grain length in the axial direction/grain thickness in the radial direction.
b
Averaged in units of 1029 m2 n1 between 2 and 8 1025 n m2, E > 1 MeV.
4 6
Normalized Growth Rate x10 - m .n , E>1MeV

- m .n , E>1MeV
-1
3
-1
2
2
29
3
29
2
Growth Rate x10
1
1
0
0 0.5 1 1.5 2
-1
0
0 1 2 3 4 5 6
-2
-14 -2
Dislocation Density x 10 -m Grain Thickness - m
Fig. 5.6.1. Eect of dislocation density on growth rate for cold-worked Fig. 5.6.3. Eect of grain thickness on the growth rate for cold-worked
Zr2.5Nb with a grain thickness of 0.310.37 lm. Zr2.5Nb corrected to a dislocation density of 4 1014 m2, to a grain
aspect ratio of 15:1 and an FA of 0.03.
10 in the range 0.30.5 lm, where the eect of grain size is

small and other factors dominated.
5
Axial
4
Growth Strain x 10
5.7. Eect of residual stress

0
-5 Intergranular residual stresses are developed in Zr alloys

Transverse during heating and cooling (as a result of the anisotropic
-10 thermal expansion of the hcp crystal), and during deforma-
tion as a result of plastic anisotropy [68,69]. In Zircaloy-2
-15 these have been shown both theoretically and experimen-
tally to result in long term transient irradiation growth
-20
0 2 4 6 8
response whose sign and magnitude depend in a sensible
Neutron Fluence x 10
-25
- n.m
-2 way upon the prior thermal or mechanical treatment
[6870]. The transient irradiation growth strain observed
Fig. 5.6.2. Growth strain as a function of neutron uence for modied in the rst 2 1025 n m2, E > 1 MeV, Fig. 5.1.1, may well
pressure tube material with a low dislocation density and ne grain size.
be attributable to the relaxation of such stresses [71].
Irradiated in DIDO at 553 K and 6 1017 n m2 s1, E > 1 MeV, from
Ref. [61].
6. Irradiation creep
Curiously, there is no discernable eect of pre-extrusion Irradiation creep is more dicult to study than irradia-
b-quenching of the material [52], an operation expected to tion growth because any shape change that occurs under an
rene the grain size. No microstructural data were applied stress incorporates some components of the shape
reported, and possibly the tubes studied had grain sizes change that occurs without stress or due to thermal creep.
For practical purposes, the irradiation growth strain is sim- ponent of growth which also has a linear dependence on
ply subtracted from the total strain, the dierence being fast ux (Section 5.2) and also includes thermal creep. Note
dened as creep. However, on a microscopic level, the that if thermal creep were a signicant contributor to the
processes associated with the growth strain may be dier- total strain rate (e.g., as the temperature increases) the total
ent with and without stress depending upon the mechanism strain rate would exhibit a ux exponent, n < 1 even when
(see Section 10). The issue of the thermal creep is even more the ux dependence of irradiation creep was linear.
complicated because the creep rate without a fast neutron
ux is not relevant to the in-ux behaviour, since it is sup- 6.2. Eect of fast neutron uence
pressed by radiation damage [14,15,25] as will be discussed
in Section 7.2. Frequently, it is found that the latter is small Typically, if diametral strain is plotted against time
relative to the radiation-induced component of creep and or fast neutron uence for a pressure tube irradiated at fast
in many cases, to a rst approximation, can be ignored. neutron uxes typical of power reactors (2
1017 n m2 s1, E > 1 MeV) with an internal pressure, there
6.1. Eect of fast ux is a small positive intercept at the abscissa, Fig. 6.2.1
[15,25], suggesting a transient or primary creep strain.
The dependence of irradiation creep of Zr alloys on fast The transverse irradiation growth strain is negative and
neutron ux (displacement damage rate) is often expressed exhibits only a small transient over the rst 1 1025 n/m2,
as a power law, i.e., E > 1 MeV [56], Fig. 5.1.1, so this transient must be attrib-
uted mainly to creep. A similar transient is also observed at
e_ ic / /n ; 6:1:1
the edge of the reactor core where the irradiation creep is
where / is the fast neutron ux and n is the ux exponent. negligible [25]. Pressurized micro-pressure tube specimens
For pressure tubes at reactor operating temperatures the irradiated at very high ux (2 1018 n m2 s1, E >
fast ux dependence is generally accepted to be approxi- 1 MeV) do not exhibit this intercept [17], presumably
mately linear, i.e., n 1 [15]. because the irradiation creep strain is relatively much lar-
Although no similar experiments exist for Zr2.5Nb, the ger, and hence this initial transient is likely attributable
linear sensitivity of irradiation creep to displacement dam- to thermal creep.
age has been clearly demonstrated by proton irradiation in We have recently shown, however, that over the long
hot-rolled and annealed Zircaloy-2 plate at 553 K [72], term (16 1025 n/m2, E > 1 MeV) the creep rate can vary
Fig. 6.1.1. slowly [30], as shown for example in Fig. 6.2.2. The change
The eect of fast ux is most graphically demonstrated in strain rate may be positive or negative and its sign and
for Zr2.5Nb pressure tubes by the diametral strain proles magnitude depend upon temperature, stress and direction
of pressure tubes in power or test reactors, and examples of and probably relates to long term evolution of the micro-
which are shown in Fig. 2.1.3. Here the strain follows the structure, see Section 9.
prole of the fast ux through the core (although these
may be skewed because of the temperature gradient and/ 6.3. Eect of temperature
or end-to-end eect). The diametral strain includes a com-
From uniaxial in-reactor creep tests at high temperature
10 (>600 K), stress relaxation tests at low and intermediate
5
Axial MPT
4 Transverse MP
-
Creep Rate x10 h
Axial FS
6
Creep Strain %
3 Axial MPT
1
2
0
0 5 10 15 20
0.1 -1
0.1 1 10 -25 2
Neutron Fluence x 10 - n/m , E>1MeV
7 -1
Dose Rate x10 dpa.s
Fig. 6.2.1. Creep strain (total strain minus growth strain) as a function of
Fig. 6.1.1. Creep rate as a function of damage rate for cold-worked neutron uence for small pressurized tubes (Section 5.5) irradiated at
Zircaloy-2 pressure tube material irradiated with 4 MeV protons, from 553 K in Osiris with a nominal stress of 160 MPa at a neutron ux of
Ref. [72]. 2 1018 n m2 s1, E > 1 MeV, from Ref. [17].
0.16 e_ ic / expT act =T 5:5 106 ; 6:3:2

0.14
where Tact has a value of 6500 K.
160 MPa
0.12 For the temperature range 555 K to 588 K this expres-
0.1 sion predicts an increase in strain rate of 56%, very close
122 MPa
0.08
to the variation of creep compliances later observed in
the axial and transverse directions of pressurized capsules
0.06 82 MPa
of cold-worked Zr2.5Nb irradiated in a fast neutron ux
0.04 of 2 1018 n m2 s1, E > 1 MeV, i.e., 53% and 49%,
43 MPa
0.02 respectively [75], Table 6.3.1 and Fig. 6.3.2. As the temper-
0 ature is reduced, Eq. (6.3.2) becomes at at about 450 K.
0 5 10 15 20
-25 -2
Neutron Fluence x 10 - n.m , E>1MeV 6.4. Eect of stress
Fig. 6.2.2. Variation in steady state creep rate with neutron uence
for small pressurized tubes irradiated at 553 K in Osiris at 2 At pressure tube operating temperatures, the stress
1018 n m2 s1, E > 1 MeV, from Ref. [30]. dependence of irradiation creep of pressure tube materials
is approximately linear, for stresses up to about 200 MPa,
i.e.,
temperature (310570 K) and from the asymmetry of the
diametral strain proles caused by the coolant temperature e_ ic / rn ; 6:4:1
changes in pressure tubes, we have previously inferred a
where r is stress and n 1. The early evidence for this came
general temperature dependence of the in-reactor creep
from full size pressure tubes irradiated in the NRU Reactor
rate. This is at at low temperatures with a transition in
at Chalk River Laboratories with zones of dierent stress
the operating temperature range of pressure tubes in power
created by machining to dierent wall thicknesses [76],
reactors (520580 K) to a relatively steep temperature
Fig. 6.4.1. These tests typically ran for a relatively short
dependence at higher temperatures (P600 K) [15],
time, typically 10 000 h at uxes of about 2 1017 n
Fig. 6.3.1. In the high temperature range the temperature
m2 s1, E > 1 MeV, or a fast uence of 1 1025 n m2,
dependence can be expressed approximately by [14]
E > 1 MeV. Hence the contribution of growth was small
e_ total / expT act =T ; 6:3:1 compared to the total strain (see Fig. 5.1.1), and was not
detected. Subsequently, stress relaxation of bent beams
where e_ is the strain rate, T the absolute temperature and [77,78] supported the observation of linear stress depen-
Tact is the activation temperature and has a value of
around 17 000 K. This is about the value for thermal creep Table 6.3.1
in this temperature range, variously reported as 15 000 Creep compliances as a function of temperature for micro-pressure tube
specimens irradiated in Osiris at 2 1018 n m2 s1, E > 1 MeV
25 000 K [73]. Possibly in-reactor creep is entirely thermal
at these temperatures. Temperature Axial creep Transverse creep
(K) compliance compliance
In the low to intermediate temperature range (310 (m2 n1 MPa1) (m2 n1 MPa1)
580 K) we have inferred a temperature dependence of irra-
555 0.55 1030 0.88 1030
diation creep given by [74] 588 0.84 1030 1.16 1030
400 350 300 250

-5.00
m .n , E>1MeV m /n
Temperature - C 20
2
583K Transverse
--1
15 583K Axial
Log Strain Rate - h
553K Transverse
-6.00
553K Axial
10
-1
2
5
-29
60C
-7.00
Strain Rate x 10
-5
-8.00 -10
1.4 1.6 1.8 2 2.2 0 20 40 60 80 100 120 140 160 180
3 -1
Inverse Temperature x 10 K Hoop Stress - MPa
Fig. 6.3.1. Temperature dependence of the transverse strain rate of cold- Fig. 6.3.2. Variation of strain rate in the transverse and axial directions
worked Zr2.5Nb pressure tubes and pressure tube material, from Ref. with stress for Zr2.5Nb micro-pressure tube material irradiated at 553 K
[15]. and 583 K in Osiris at 2 1018 n m2 s1, E > 1 MeV, from Ref. [75].
1000
100
Stress Exponent - n
10
Fig. 6.4.1. Variation in accumulated strain with elevation for a Zr2.5Nb 0.1
pressure tube in the core of NRU. The wall thickness is varied to give the 0 200 400 600 800
stress indicated and the peak neutron ux is 2 1017 n m2 s1, Applied Stress - MPa
E > 1 MeV, from Ref. [76].
Fig. 6.4.3. Stress exponent of in-reactor creep of cold-worked Zr2.5Nb
pressure tube material as a function of applied stress, from Ref. [79].
dence. After a small transient the logarithm of stress de-
cayed linearly with fast uence or time indicating linear
stress dependence of both the axial and diametral creep 3
rates, Fig. 6.4.2.
Tests on pressurized capsules irradiated to very high fast 2.5
uences (1.6 1026 n m2, E > 1 MeV) continue to show out of flux
Stress Exponent, n
that irradiation creep varies approximately linearly with 2 in flux

stress to at least 160 MPa [75], Fig. 6.3.2. In this case
growth in the unloaded condition is obvious at the origin 1.5
of the strain rate vs stress plot.
At stresses above 200 MPa, the stress sensitivity 1
increases as shown graphically in Fig. 6.4.3 which is based

0.5
on stress relaxation and uniaxial creep tests [79]. The latter
include growth, but its contribution is small (typically
0
<2 108 h1 at the uxes used to obtain the data) relative 530 580 630 680
to the creep rates at the higher stresses (107 h1).
Temperature - K
At temperatures above about 600 K, the stress sensitiv-
ity also increases gradually, based on tests on full sized Fig. 6.4.4. Stress exponent as a function of temperature for cold-worked
pressure tubes in the WR1 reactor at Whiteshell Laborato- Zr2.5Nb pressure tubes irradiated in the WR1 reactor. Data for sections
of tubes both in and out of ux, from Ref. [15].
ries [14,15,80], Fig. 6.4.4. In these tests the ratio of trans-
verse to axial stress (roughly 2:1 for a thin walled, closed
end pressurized tube) was often less than 2:1 because the
pressurizing medium was an organic coolant, exhibiting a
1 relatively high pressure drop relative to the average pres-
sure as it owed through the tube. In WR1, similar stress
sensitivities were observed in both the fueled sections of
the tubes (in-ux) and the pressure tubes extensions that
Stress Ratio
Transverse
passed through the shielding of the reactor (thermal creep
only).
6.5. Eect of crystallographic texture

Axial
Before any observations were available, it was expected,
0.1 based on the hcp crystal structure of zirconium, that irradi-
0 5 10 15 20 25 30 ation creep would be anisotropic and that this anisotropy
Time - h would be controlled by the crystallographic texture [81].
Fig. 6.4.2. Stress relaxation of bent beams of cold-worked Zr2.5NB Uniaxial in-reactor creep tests and stress relaxation tests
pressure tube material at 2 1017 n m2 s1, E > 1 MeV, in NRU at subsequently showed the expected higher creep rates in
573 K, from Ref. [77]. the axial than in the transverse direction of pressure tube
materials, Fig. 6.4.2 [76,77]. Holt and Ibrahim [51,67,74] 10
-1
- m .n .MPa
developed a methodology for separating irradiation creep
and growth in pressure tube deformation based on calcula-
-1
2
tions of the expected anisotropy from the texture and on
the elongation and diametral strain rates of pressure tubes
30
Creep coefficient x 10
in power reactors. This showed that creep made a substan-
tial contribution to the elongation of pressure tubes and
0.20
that growth reduced the diametral strain rate. We now
know that about 33% of the elongation rate is due to irra-
diation growth and 67% due to irradiation creep while the
contribution to diametral strain rate of creep is reduced
1
about 25% by growth [61].
1 10
The eect of texture on the biaxial deformation behav- -14 -2
Dislocation Density x 10 -m
iour has since been clearly demonstrated in tests on inter-
nally pressurized tubes of two dierent textures Fig. 6.6.1. Irradiation creep rate measured by stress relaxation as a
[17,52,75]. Some results are shown in Fig. 6.2.1. Here the function of pre-irradiation dislocation density. Irradiation in NRU at
irradiation growth has been subtracted from the total 2 1017 n m2 s1, E > 1 MeV and 573 K, from Ref. [83].
strain rate and only the irradiation creep is shown. The cor-
responding pole gures are shown in Fig. 5.5.1 and the tex- 50% over a grain size range of 0.370 lm, but there are
ture parameters are given in Table 5.5.1. many potentially compounding factors. There are no stud-
Here it is clear that a material with a strong transverse ies in which the creep rate has been measured as a function
texture, similar to that of a pressure tube, exhibits a rela- of grain size alone. The creep rate has been shown, how-
tively low transverse creep rate and a high axial creep rate, ever, to increase from front to back of pressure tubes, coin-
whereas a material with a more radial texture exhibits a rel- ciding with a reduction of grain size from about 0.4 to
atively high transverse creep rate and a small negative axial 0.3 lm and also with an increase in dislocation density
creep rate. Note that the axial component of texture is and less transverse texture [22], see Section 3. Also the
small in both cases, and the axial creep rate depends upon diametral strain rate of pressure tubes increases with
the distribution of basal plane normals in the radial/trans- decreasing grain size, Fig. 4.3, an observation which might
verse plane. Factors other than texture are believed to also be attributed to a decrease in the absolute irradiation
inuence the anisotropy however, in particular a correla- growth rate in that direction [54].
tion of the dislocation substructure with crystal orientation
[12,66] and possibly the interaction of the hcp a-phase with
the bcc b-phase [82]. 7. Thermal creep
6.6. Eect of dislocation density Thermal creep contributes little to the peak diametral
strain of a pressure tube or to elongation [14]. However
Causey et al. [77,83] showed, using stress relaxation thermal creep contributes the majority of the deformation
tests, that the irradiation creep rate is weakly sensitive to at the ends of the channel where the fast neutron ux is
the initial (as-manufactured) dislocation density, i.e., low, but the bending moment in the pressure tube is high.
e_ ic / qp , where q is the dislocation density and p has a value It may thus contribute signicantly to controlling the sag
of 0.23, Fig. 6.6.1. Within the experimental variations prole of the pressure tube and the potential for pressure
this is comparable to the value (q 0.16) found for Zirca- tube-to-calandria tube contact. Also thermal creep may
loy-2 pressure tubes by extraction of creep and growth contribute signicantly to the strain at high stresses, for
components from elongation and diametral strain data example in the vicinity of a crack tip, and hence assist
[67]. Like the growth, the creep rate and anisotropy prob- the relaxation of stresses that can potentially cause failure
ably depend on the mixture of dislocation types and their by the process of delayed hydride cracking [84,85].
distribution amongst the dierent crystal orientations as The thermal creep of Zr alloys at about 570 K is compli-
well as the overall density [66]. cated by the existence of strain aging [26,27], which can
produce transient increases in strain rate during steady
6.7. Eect of grain size state creep and anomalous stress sensitivity in certain tem-
perature and strain rate ranges. Strain aging in Zr is
The eect of grain size (the thickness of the attened, believed to be associated with solute eects.
elongated a-grains in pressure tubes) on irradiation creep
has not been clearly established. Causey et al. [77] showed 7.1. Thermal creep at aws
a very weak grain size dependence of the in-reactor creep
rate for a variety of Zr alloys over a very wide grain size The stress concentrations at aw tips in pressure tubes
range, the creep rate increasing with grain size by about can initiate delayed hydride cracking under certain
conditions of stress, temperature history and hydrogen iso- 25
-1
Transverse Strain Rate x 10 - h
tope concentration. Often the aws will be present for many
8
operating hours before the conditions exist which could ini- 20
tiate cracking. In this situation the peak stress at the aw tip
can relax by creep. The creep rate is fastest at the tip of the 15
aw where the stress is highest because of the strong sensi-

10
tivity of creep rate to stress [79]. In the early stages of relax-
ation where the stress is still high, thermal creep may
5
contribute to this relaxation.
The thermal creep behaviour of pressure tube and pres- 0
sure-tube like material at temperatures of 373596 K and 3 4 5 6 7 8 9
stresses of 150500 MPa with a variety of loading condi- Elevation - m
tions is reported by Christodoulou et al. [73]. They noted Fig. 7.2.1. Diametral strain rate as a function of elevation in NRU at 550
substantial specimen to specimen variability which they 580 K with a peak neutron ux of 3 1017 n m2 s1, E > 1 MeV
attributed to variations in microstructure in their materials. showing suppression of thermal creep at the edges of the core (4.4 m and
They found that the steady state creep rate over most of the 7.7 m), from Ref. [25].
temperature range was approximately proportional to the
4th power of stress and the rate exhibited an activation locations were inaccurate, the quantitative interpretation
temperature of 10 600 K above about 470 K and 5000 K of the shape of this prole in relation to the fast ux was
below about 400 K. Analysis of the interaction of the crys- not possible. Christodoulou et al. [14] interpreted this min-
tallographic texture with the stress state showed that the imum as corresponding to in-reactor thermal creep, e_ tc , of
creep was anisotropic with the creep compliance being Eq. (2.1). A method of retrospective estimation of fast neu-
higher in the basal plane than in the c direction of the tron uence using 93Nb was reported in 1990 [86], and
hcp crystal lattice. Pre-irradiation reduces the creep rate, results from such estimates were recently used to make a
but the form of the creep law remains the same and the quantitative assessment of these strain proles [25].
information derived above has been used to develop a The analysis shows that there is hardening that sup-
model for local relaxation of stresses by thermal creep of presses the thermal creep at uxes lower than about
pre-irradiated material at aws [23,24], (see Fig. 7.1.1). 5 1015 n m2, E > MeV. This is below the threshold fast
ux for the formation of dislocation loop, 2 1016 n
m2, E > MeV, and is attributed to immobilization by
7.2. Thermal creep at the pressure tube ends
climb of dislocations that enable thermal creep. There is
a further stage of hardening, in this case suppression of
Early irradiations of Zr2.5Nb pressure tubes in the
the irradiation creep above 2 1016 n m2, E > MeV,
NRU Reactor at Chalk River Laboratories showed a min-
by the formation of dislocation loops that either inhibit
imum in the diametral strain rate at the edge of the fuelled
glide or alter the distribution of self-interstitial atoms and
zone where the fast neutron ux is low [15], Fig. 7.2.1, for
vacancies at various sinks or features in the microstruc-
example [25]. This minimum was interpreted to represent
ture. This distribution of point defects amongst various
the suppression of thermal creep by radiation damage
sinks is believed to be the basic mechanism of irradiation
accumulated at a neutron ux too low to induce signicant
creep, see Section 10.
irradiation creep. Because the estimates of ux in these
Making use of this information to improve the calcula-
tion of sag and potential contact is still some time o, how-
900 ever, because there appears to be a complicated interaction
Time (hours)
800
0.0 between temperature, ux, uence and microstructure, and
700 60.0
600 the stress dependence and anisotropy in this region are not
well dened. This is one of the important remaining
Stress (MPa)
500
400 unknowns in the area of pressure tube deformation.
300
200
100
0 8. Creep ductility
-100
-200 The statement by Ells et al. [31] that an operating limit
-300
-400
of 5% circumferential expansion seems to be very conserva-
0 2 4 6 8 10 tive is based on the argument, rst put forward by Wood-
Distance (mm) ford [87,88] that the total elongation of materials that fail
Fig. 7.1.1. Finite element calculation showing the relaxation of peak stress in a ductile manner under tension is inversely related to
at a aw in cold-worked Zr2.5Nb pressure tube material after 60 h at the stress sensitivity of the strain rate (the inverse of the
573 K, from Ref. [23]. exponent n in Eq. (6.4.1)), Fig. 8.1. Superplastic materials
10000 quoted above is based on the fact that these tests are likely
to be very conservative. One might debate this conclusion:
The assumed failure mechanism is ductile necking. The

Elongation to Failure - %
relationship between ductility and stress sensitivity is

1000
not valid if some other failure mechanism intervenes.
Irradiation creep in the low stress range involves the
redistribution of material by the formation and migra-
tion to sinks of SIAs and vacancies. Over the course
100 of a reactor lifetime each atom in a pressure tube is dis-
placed form its position and rearranged 3050 times to
sinks such as dislocations on boundaries. The combina-
tion of elongation and diametral creep at ends of life
represents an atomic rearrangement by radiation-
10
induced mass transport approaching 10%. Hence micro-
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
structural and microchemical changes occur which are
Strain Rate Sensitivity - m
not experienced over the time scale of the accelerated
Fig. 8.1. Total elongation to failure in a tensile test as a function of strain tests and there is a possibility of an as-yet unidentied
rate sensitivity, m = 1/n, showing superplastic behaviour for m > 0.5, from failure mechanism.
Refs. [87,88].
The biaxial tests cited were done on fuel cladding, which
has a predominantly radial texture [89]. This resists wall
generally have values of n in the range 12. For zirconium thinning under biaxial conditions and, as in the case of
alloys in-reactor, n 1 for stresses up to about 200 MPa, short term burst tests [90], and thus would be expected
then n gradually increases, see Section 5.4. Under normal to impart increased ductility compared with the predom-
reactor operating conditions, at stresses of 100150 MPa, inantly transverse texture of pressure tubes.
it takes tens of years to reach strains approaching 5%, so
in order to investigate that the limits of creep ductility in It should be noted that the safety analysis of a CANDU
any practical time scale, it is necessary to do some form reactor assumes the possibility of a pressure tube rupture,
of accelerated test. The simplest way of doing this is to and the system is able to tolerate this possibility without
increase the stress, and hence the stress exponent, and thus risk to the workers or the public.
such tests are thought to be naturally conservative. Ells
et al. [31] summarized the data for cold-worked Zircaloy- 9. Changes in microstructure and microchemistry during
2 and Zr2.5Nb, in biaxial and uniaxial tests at stresses service
>300 MPa, where n has a value >5, Fig 8.2. The minimum
failure ductility observed was 5.9%. Their conclusion The changes that occur in the microstructure of pressure
tube materials due to exposure to fast neutron ux and
temperature during service have been documented by Grif-
ths and co-workers [12,22,9193].
1000 The most readily observable change is the appearance of
a-type dislocation loops Burgers vector 1=31 1 2 0 a
few tens of nanometers in diameter in the a-grains,
Hoop Stress - MPa
Fig. 9.1. The number density of these increases rapidly in

the rst 1 1025 n m2, E > 1 MeV (roughly one year of
operation), from the level present initially, then tends
towards saturation, Fig. 9.2. The saturation level is a func-
tion of temperature and fast ux, increasing with fast ux
and decreasing with temperature, so the proles are
expected to vary from channel to channel and reactor to
reactor [93].
Coincident with this change is a slow increase in the c-
100 component dislocation density over longer periods which
1.00E-07 1.00E-06 1.00E-05 1.00E-04 depends upon crystal orientation [22]. Crystals with their
Creep Rate - h -1 c-axis in the radial direction of the tube initially contain
mostly 1=31 1 2 3 screw dislocations (approximately nor-
Fig. 8.2. Stress sensitivity of cold-worked Zr2.5Nb pressure tube
material, in-reactor varies from 1 (at <200 MPa) to 10 (at 300 mal to the basal plane), and their dislocation density
500 MPa). The tests shown exhibited creep ductility of at least 5.9% and increases continuously with fast uence, Fig. 9.2, by helical
up to 17.2%, from Ref. [31]. climb, as illustrated in Fig. 9.3. Crystals with their c-axis in
Fig. 9.1. Electron micrographs showing the a dislocation structure of Zr2.5Nb pressure tube material, (a) before and (b) after irradiation at 543 K to a
neutron uence of 1.1 1026 n m2, E > 1 MeV. Note the high density of loops in (b), from Ref. [12].
12 the transverse direction contain mainly 1=31 1 2 3 edge

dislocations (approximately parallel to the basal plane).
-1
Initially these tend to split into two partials 1= 62 0

2 3,
Integral Breadth x 10 - m
10
causing an initial apparent reduction (based on X-ray line
7
8 broadening) but also actually reecting some increase in

L(3030) PRISM the c-component dislocation density.
6 R(0004) BASAL
After a uence of few 1025 n m2, E > 1 MeV, the Nb
4
in the a-phase, which is slightly super-saturated before irra-
diation, precipitates as ne Nb rich particles 25 nm in
2 diameter, Fig. 9.4, which are uniformly dispersed, except
near the grain boundaries where a narrow region is
0 denuded [92]. There is an indication that the concentration
0 5 10 15 20
of Nb in the matrix is reduced to an estimated 0.3%.
-25 -2
Neutron Fluence x 10 , n.m The state of b-phase, which before irradiation has
Fig. 9.2. Change X-ray line broadening due to a- and c-type dislocations partially transformed towards equilibrium and contains
density with neutron uence at 543 K and 2 1018 n m2 s1, 50%Nb, tends towards a steady state after irradiation
E > 1 MeV, from Ref. [12]. to about 2 1025 n m2, E > 1 MeV, Fig. 9.5. The steady
Fig. 9.3. Electron micrographs showing the c-type dislocation structure of Zr2.5Nb pressure tube material, (a) before and (b) after irradiation at 543 K to
a neutron uence of 1.1 1026 n m2, E > 1 MeV. Note loop nucleated on a screw dislocation at arrow in (b), from Ref. [12].
state condition depends upon temperature and fast ux rium is reversed and the b has about 40%Nb. At the inlet
[93]. At the high ux region towards the inlet end and in end there is a very slow progression towards the equilib-
the middle of the core, the transformation towards equilibrium state and at the outlet end this progression is more
rapid. Fig. 9.6 shows a compilation of data for uences
of >2 1025 n m2, E > 1 MeV.
The Fe which is initially concentrated in the b-phase is
dispersed by irradiation and can no longer be detected
within the b or in the a-grains because its average concen-
tration is below the limit of detection (0.15%) [44]. Fe is
still found at the aa sub-boundaries however at approxi-
mately the same level as before irradiation.
10. Mechanisms of irradiation-induced deformation
This section will strictly deal with the deformation

mechanisms associated with irradiation. Thermal creep is
a vast subject about which many textbooks have been writ-
ten, however, its contribution to in-reactor deformation of
pressure tubes is small once irradiation hardening has
Fig. 9.4. Electron micrograph showing Nb precipitates (arrows) in the a- occurred. For an overview of thermal creep in Zr alloys,
phase of Zr2.5Nb pressure tube after irradiation at 573 K to a neutron
uence of 0.9 1025 n m2, E > 1 MeV, from Ref. [92].
see the article by Harbottle [26] and for a more detailed
account see the book by Franklin et al. [27].
The eects of irradiation on dimensional stability are
80
generally acknowledged to be due to displacement dam-
age that is the knocking of atoms o their lattice sites by
70
momentum transfer from the radiation. In pressure tubes
Nb Concentration - %
60 the vast majority of displacement damage comes from the

50 fast neutrons. The damage occurs in displacement cas-
cades. Once an energetic neutron has displaced one Zr
40
atom, sucient energy is usually transferred that the zirco-
30 nium atom (which interacts very much more strongly with
20 the other Zr atoms that the initial neutron) displaces many
more atoms. The residual damage after a very short (ns)
10
period during which the kinetic energy of the displaced
0
atoms is dissipated, comprises vacant lattice sites (vacan-
0 2 4 6 8 10 12
-25 -2 cies), Zr atoms in interstitial sites (self-interstitial atoms
Neutron Fluence x 10 , n.m
or SIAs) and clusters of two or more of each of these
Fig. 9.5. Variation in the Nb concentration in the b-phase with neutron defects. The point defects (vacancies and SIAs) which
uence in a CANDU pressure tube operating at 560570 K, with a peak may represent only a few per cent of the total number of
ux of 3 1017 n m2 s1, E > 1 MeV, from Ref. [92].
atoms displaced, are mobile at reactor operating tempera-
tures, and the dimensional changes caused by creep and
70
growth are related to the disposition of these defects [94].
At higher temperatures (>700 K) where signicant num-
60 bers of vacancies are formed thermally and at lower tem-
Nb Concentration - %
50 peratures (<400 K) where the vacancies are immobile, the

clusters are believed to play an important role [95,96].
40
30 10.1. Irradiation growth

ex-service
20 average of offcuts
In the temperature range at which power reactors oper-
10 ate, the SIAs and vacancies can be annihilated in two ways:
0 by reaction with a defect of the opposite sign either singly or
0 1 2 3 4 5 6 7 in a cluster, or reaction with an extended lattice defect or
Distance from Inlet - m sink (dislocation, loop, phase- or grain-boundary). In the
Fig. 9.6. Variation in the Nb concentration in the b-phase along a former case no strain occurs. In the latter case a strain
CANDU pressure tube operating at 520570 K (inlet to outlet) with a occurs locally in a direction characteristic of the defect. In
peak neutron ux of 3 1017 n m2 s1, E > 1 MeV, from Ref. [92]. the case of a dislocation or loop the strain is a contraction
(for vacancies) and a dilation (for SIAs) in the direction 10

of the Burgers vector. In the case of a boundary it is a 6 microns
8 2 microns
-1
contraction or dilation in the direction of the boundary
-s
1 microns
normal. 6 0.3 microns
11
Axial Growth Rate x 10
0.15 microns
Two conditions are necessary for a macroscopic strain
4
to occur from the disposition of point defects at sinks.
First, the sinks must experience net uxes of point defects, 2
i.e., the ux of one species of point defect must, over time
0
exceed the ux of the species of the opposite sign. Obvi- 0 1 2 3 4 5 6
ously, if a particular sink experiences at net ux of one spe- -2
cies, some other sink must experience a net ux of the other
-4
species. Second, the orientation distribution of the sinks
that experience net uxes of the dierent species of point -6
defect must be dierent. Network Dislocation Densityx10
-14
-m
-2
The factor which determines the net ux of a particular

Fig. 10.1.1. Calculated eect of dislocation density and grain thickness of
point defect to a particular sink is referred to as a bias and
the growth rate of Zr2.5Nb in the axial direction, from Ref. [54]. In the
is by denition negative for vacancies and positive for region between the heavy line and the x-axis both the axial and transverse
interstitials. Note that it is only necessary for one type of growth rates are negative.
sink to have a bias. By default unbiased or neutral sinks
will experience a net ux of defects of the opposite sign.
A number of sources of bias have been proposed for Zr 2
alloys [54], but the commonly accepted one is diusional 1.5 0.15 microns
-1 0.3 microns
anisotropy dierence (DAD) rst proposed by Woo and -s 1 1 microns
11
Goesele [97,98]. They proposed that vacancies and SIAs 2 microns

Transverse Growth Rate x 10
0.5 6 microns
exhibit anisotropic diusion in Zr a natural consequence 0
of the hcp crystal structure, and that the diusional aniso-
-0.50 1 2 3 4 5 6
tropies of the two species of defects are dierent, a natural
-1
consequence of the dierent relationship of the two defects
to the hcp lattice. This provides a very powerful driving -1.5
force to segregate the defects amongst sinks of dierent ori- -2

entation. Growth results when the strain direction of a sink -2.5
correlates with its orientation. -3
For boundaries this is self-evident, since the strain direc- -3.5
tion is normal to the boundary. In the case of a-type and c- Network Dislocation Density x 10 -14 - m-2
component network dislocations the orientation correlates
with dislocation character so that c-component edge dislo- Fig. 10.1.2. Calculated eect of dislocation density and grain thickness of
the growth rate of Zr2.5Nb in the transverse direction, from Ref. [54]. In
cations are eectively on the basal plane and a-type edge
the region between the heavy line and the x-axis both the axial and
dislocations are eectively perpendicular to it. Since the transverse growth rates are negative.
point defects interact more strongly with and edge than a
screw dislocation [99], this provides an orientation bias
[54]. In the case of a-type loops the loop plane is normal 10.2. Irradiation creep
to the basal plane as is some part of the line direction. Both
the plane and line direction of c-component loops are on Unlike growth, there is no consensus regarding the
the basal plane. In a material containing only dislocation mechanisms of irradiation creep in Zr alloys, other than
sinks (or coarse grained material), preferential diusion that it is related in some way to net uxes of point defects
of SIAs in the basal plane or vacancies normal to it will to sinks, probably dislocations and loops.
lead to the observed shrinkage along then caxis and Models for irradiation creep that have been considered
expansion in the basal plane. If a high density of bound- for zirconium include yielding-creep, climb-plus-glide
aries are introduced (ne grained material) the growth (CPG) including I-creep and various forms of stress-
can be a complex function of the dislocation densities induced preferred absorption (SIPA) including SIPA-
and grain shapes. Qualitatively, the growth behaviour of DAD or elasto-diusion (SIPA induced by the eect of
the TG3 R1 material, i.e., simultaneous shrinkage in the stress on diusional anisotropy). These models, amongst
axial and transverse direction of the pressure tube (see Sec- others, have been reviewed by Matthews and Finnis [100].
tion 5.6) is predicted for very ne grains [54,98], Figs. The yielding creep model [101,102] is applicable to mate-
10.1.1 and 10.1.2. rials exhibiting irradiation growth. According to this model
Detailed mathematical treatments of models for irradia- the internal stresses generated by growth of individual
tion growth are given in Refs. [54,9499]. grains of dierent orientations approach the yield stress,
and the applied stress elevates the stress in crystals of a cer- Detailed mathematical treatments of the models for
tain orientation above the yield stress causing them to creep are given in the Refs. [98110].
yield. Continued growth causes continued yielding in a
direction that is biased by the applied stress. This model 11. Modelling pressure tube deformation
has generally been ignored in recent times, but has not been
discounted by experimental facts. 11.1. The need for models
The climb-plus-glide model [103] is an extension of the
dislocation creep model [104] of thermal creep in which dis- Models are required to predict the deformation of mod-
locations are able to bypass obstacles by climb due to the ied pressure tubes operating in new reactor designs under
absorption of thermally generated vacancies. The strain conditions dierent from those that have been experienced
producing step is glide. Under irradiation, a net ux of previously or to predict the future deformation of pressure
either vacancies or interstitials to a given dislocation will tubes in existing power reactors for the purposes of plant
allow it to bypass obstacles and glide. In the late 1970s this life management (PLiM).
model was supported by an experiment in which the neu- In the case of new reactor designs the operating condi-
tron ux to a creeping specimen was turned on and o tions, ux, temperature, coolant pressure, may be more
by moving the creep rig in and out of reactor [105]. A posi- severe so that the thickness of the tube must be adjusted.
tive strain transient was seen in both cases, even though on Also, one may want to take advantage of the knowledge
initiating irradiation, the dislocations in the specimen of the eects on deformation of microstructural, chemical
would experience a net ux of SIAs corresponding to the and manufacturing variables to improve the deformation
build up of a vacancy super-saturation, while on terminat- behaviour in order to achieve a satisfactory pressure tube
ing the irradiation a net ux of vacancies would be life in the new design. An example of such a new design
expected as the super-saturation decayed. This result has is the 1000 MWe class Advanced CANDU Reactor
never been explained by another mechanism. A modica- (ACR 1000 [16]) which has higher coolant temperatures
tion of this model is Gittis I-creep model [106], in which and pressures for improved thermodynamic eciency,
there is coordinated movement of arrays of dislocations and higher channel powers for increased average power
by glide after climb. density of the core.
More recently Semenov and Woo [107] have developed In the case of PLiM, a general algorithm for managing
an analysis of the stochastic nature of displacement cascade the life of the pressure tubes in a CANDU reactor is as fol-
damage in both time and space, show that a net ux of lows [7]:
point defects to dislocations may not even be necessary
for them to surpass obstacles and glide, given that the uc- plan maintenance based on predictions using models
tuations in the defect uxes of the two species are large based on R&D and surveillance data;
enough. monitor on-going changes to verify predictions/modify
SIPA [106,108110] assumes a bias of the ux of point models;
defects to dislocations as a function of the orientation of disposition out-of-design conditions, e.g., aws and
the Burgers vector with respect to the applied stress. Dislo- contact of PT with CT (this could involve removal of
cations whose Burgers vectors are aligned with the tensile tubes);
stress experience a net ux of SIAs. By default, those continued R&D including end-of-life testing to improve
aligned with a compressive stress experience a net ux of models;
vacancies. Any deviatoric stress tensor will induce a bias. modify maintenance plans accordingly.
This bias was initially attributed to the elastic interac-
tions of the point defects with and the strain elds of the In many cases, it may be sucient simply to extrapolate
dislocations in a stress eld [100]. The point defects are data from previous inspection campaigns, for example, the
considered as inclusions with elastic properties dierent elongation of each pressure tube in a CANDU 6 reactor
from the matrix. The key feature for SIPA to be viable is can be measured every time that the fuelling machine refu-
that the SIA must to exhibit an eective shear modulus less els that channel. Here a linear extrapolation of the data to
than that of the matrix while the vacancy exhibits a shear the next maintenance outage may be sucient, and for
modulus similar to that of the matrix. This might not longer term planning it may be simply a matter of know-
always be the case. Another possibility is that the disloca- ing how to extrapolate the data (e.g., a quadratic
tions are not perfect sinks (as is often assumed in calcula- equation).
tions of creep and growth) but that there is a barrier to In other cases, there is no alternative but to use a model.
absorption. This barrier may then be a function of stress This is the case for spacer location and repositioning
and dierent for interstitials and vacancies [100]. Finally (SLAR)[37], the objective is to prevent contact of the hot
there is the concept that the anisotropy of point defect pressure tube with the cold calandria tube late in life when
migration is stress dependent [98], so the applied stress the deuterium concentration of the pressure tubes has
modies the DAD eect. This is referred to as SIPA_DAD increased to reach a critical level. In earlier CANDU reac-
or elasto-diusion (ED). tors the spacers, helical springs circling the pressure tube to
centre it in the calandria tube, Fig. 1.1.1, were loose and aspect of such a model is the dependence of the anisotropic
were found to move during service [38]. This caused the deformation behaviour on the multi-axial stress tensor.
rupture of a Zircaloy-2 pressure tube in the Pickering Unit
2 reactor in 1983 as a result of pressure tube sagging into 11.2. Models for anisotropic creep and growth
contact with the calandria tube and the formation of zirco-
nium hydride blisters on the outside of the pressure tube at To describe the deformation behaviour of pressure tubes
the line of contact [29]. under a multi-axial stress state, an anisotropic deforma-
At that time, there were 20 reactors constructed or in- tion equation is required for anisotropic creep and growth
service with the loose spacer design. Fortunately 18 of these of the material. Combined with an equivalent equation for
units had Zr2.5Nb pressure tubes which exhibit a much the calandria tube, such an equation can then be used as
lower deuterium pick-up rate than Zircaloy-2. The two input to a nite element model to predict the deformation
reactors with Zircaloy-2 tubes were immediately shut down of the fuel channel, and in particular the sag and the poten-
for retubing, the spacers for new reactors were changed to tial for pressure tube-to-calandria tube contact. The sim-
a tight-tting design [38], and a maintenance procedure plest anisotropic formulation for creep is that derived by
called SLAR was developed for the 18 reactors with loose Ross-Ross et al. [76] from Hill [111]. Such an equation con-
spacers and Zr2.5Nb tubes. tains three independent anisotropy parameters, and thus
The response of the horizontal pressure tube-calandria requires the measurement of three independent deforma-
tube assembly to the weight of the fuel and coolant is an tion rates under at least two dierent stress states. The
elastic deection several millimeter as a result of the bend- parameters for such a model cannot, therefore, be derived
ing moment. During operation a permanent sag is devel- solely from the elongation and diametral strain data from a
oped due to irradiation creep of both the calandria tube pressure tube. Because the stress exponent n 1 in Eq.
and the pressure tube a process referred to as creep-sag. (6.4.1), the expressions derived by Ross-Ross et al. are con-
The annealed Zircaloy-2 calandria tube exhibits a much sistent with the polycrystalline approaches described
higher resistance to creep-sag than the pressure tube so in below. However, the polycrystalline models allow assump-
the rst few years of operation the pressure tube sags to tions about the deformation mechanisms to be substituted
transfer most of the load of fuel and coolant to the calan- for measurements.
dria tube through the spacers [18]. Thereafter the gross sag The rst anisotropic deformation model for pressure
of the assembly is controlled by the calandria tube and the tubes [75] did not separate irradiation growth, which had
pressure tube properties determine the rate at which the not yet been observed, from the total deformation rate
pressure tube approaches contact with the calandria tube which was treated entirely as creep. The rst deformation
between the spacers [19]. If the spacers are displaced, this equations separating creep and growth were developed in
may happen before the end of the design life and there is the late 1970s [51,67,74]. The anisotropic eects of texture
a potential for blister formation. were taken into account by averaging the behaviour of
In the SLAR process, an instrumented tool is inserted individual crystals, assuming certain strain-producing
into each pressure tube which can detect the spacers, apply mechanisms at the single crystal level in what is referred
an upward bending moment to the inside of the tube to to as a polycrystalline model. This was a lower bound
lift it so that the spacers are free to move and apply an polycrystalline model, which ignores the fact that dierent
electromagnetic pulse to move the spacers with a linear orientations of crystals have dierent strain tensors. The
induction motor [37]. This process has proven very success- anisotropy of the polycrystal (the pressure tube) depends
ful at all the reactors with the loose spacer design. The upon the relative contributions of the strain mechanisms
question arises, Where should the SLAR operator leave at the single crystal level. This was derived from bent beam
the spacers? Since the pressure tube and calandria tube uniaxial stress relaxation tests in which the growth
have already undergone plastic deformation, each assembly component does not contribute to the measured strain
is now unique. A deformation model is used to answer the [7779,83,112] and allowed the prediction of the creep
question [19,20]. anisotropy of a pressure tube under biaxial loading. The
The model comprises three elements, a two dimensional growth anisotropy was derived again using a lower bound
nite element representation of the fuel channel, CDEPTH polycrystalline model, by assuming, that at the single crys-
[19], a calandria tube deformation equation and a pressure tal level, dislocations were the sinks for SIAs, and grain
tube deformation equation. The deformation equation for boundaries were the sinks for vacancies. Once the anisot-
the pressure tube must predict its response to bending ropy of creep and growth were dened, their contributions
moments essentially a uniaxial loading in the axial direc- to elongation and diametral strain in a given power reactor
tion of the tube. Because the calandria tube controls the could be derived from the elongation and diametral strain
gross sag (which can be easily measured) there is no reactor data for that reactor, and the response of a pressure tube to
measurement of this response. A model is required to pre- a multi-axial stress state, including sag [113], could be
dict this response, based upon the measurable deformation predicted.
responses of the pressure tube to a biaxial stress state, i.e., In the mid 1980s, an upper bound approach was used
elongation and diametral strain. The most important (in which all crystals are assumed to exhibit the same strain
tensor) to study the eects of interactions between the crys- the ends of the pressure tube and for high stresses relevant
tals [68,69]. The approach was developed mainly to explain to the relaxation of stresses at aws [23,24].
the irradiation growth response of calandria tube materials
[35,69]. The parameters of the model were derived by the 11.3. Model for pressure tube deformation
same methodology as those for the earlier lower bound
model, but including data for creep in shear as well as stress Although the means to predict polycrystalline behaviour
relaxation [83]. At the same time nite element methods from the behaviour of individual crystals now exist as
were used to calculate the sag more accurately [19,114]. described above, the basic physical parameters that would
Both the upper and lower bound approaches can lead to be needed to construct reliable mechanistic models to pre-
systematic errors in predictions for stress tensors dierent dict the deformation of even a pure Zr single crystal are not
from those for which the data are obtained to normalize known (e.g., mobility and anisotropy, conguration, elastic
the model. As concerns about the accuracy of sag predic- properties of point defects, structures of dislocation
tions for the SLAR application developed in the early cores, and the inuence of impurity atoms on these charac-
1990s (see previous section), a self-consistent polycrystal- teristics). We therefore still rely a phenomenological
line model, SELFPOLY, was developed which allowed for approach.
individual crystals to deform dierently, but under con- The complex interaction between the eects of tempera-
straints imposed by the surrounding matrix of crystals with ture and fast neutron ux on the deformation has led to the
dierent orientations [115]. The single crystal creep and development of analyses that assume that long term steady
growth are deconvoluted from in-reactor data for polycrys- state deformation consists of separable, additive compo-
talline pressure-tube and pressure-tube like materials tested nents from thermal creep, irradiation creep and irradiation
with a range of stress tensors including uniaxial tension, growth as given in Eq. (2.1) [14]. All three components are
biaxial tension and shear [116,117]. The single crystal anisotropic and contribute to length as well as diameter
growth parameters could for the rst time be based on changes. They are given by
growth data coming from tests in high ux reactors [12,61].
This model, currently in use, represents the behaviour of e_ tc d d 2
d K 1 C 1 r1 K 2 C 2 r2 expQ1 =T
the tubes quite well over a narrow range of textures close to
those of pressure tubes, but there are inconsistencies in the K 3 C d1 r1 expQ3 =T ; 11:3:1a
predicted creep anisotropy when the texture is changed sig- e_ icd K c K 4 xC d4 xrx/expQ4 =T K 5 ; 11:3:1b
nicantly [118], Fig. 11.2.1. This is likely due to secondary
contributing factors to the anisotropy, including aniso- e_ ig
d K g K 6 x; /tC d6 x/ expQ6 =T : 11:3:1c
tropic grain shape, anisotropic dislocation distributions
introduced during manufacturing, and the presence of an The in-reactor thermal creep component has two terms [14]
isotropic second phase. A variation of the self-consistent that dominate at temperatures above and below 570 K,
polycrystalline model, applicable for the case when the respectively. The last two terms describe ux-dependent
stress exponent n 5 1 in Eq. (6.4.1), SELFPOLY-N has creep and irradiation growth, respectively. The stress expo-
also been developed [119] for the case of thermal creep near nent for thermal creep varies with stress: for stresses below
120 MPa, the stress exponent is 1, while for stresses be-
tween 120 and 200 MPa, the stress exponent increases to
2 [78]. The parameters in Eq. (11.3.1) are dened as
1.5 follows:
Anisotropy ratio - CBL/CBT
1
e_ d strain rate in a direction d (i.e., radial, transverse,
SELFPOLY axial), h1;
0.5
e_ K 1 ; K 2 constants for high temperature in-reactor thermal
creep;
K3 constant for low temperature in-reactor thermal
0
creep;
K4(x) function describing the variation of irradiation
-0.5
creep due to variations of microstructure along
the length of the tube;
-1
-0.4 -0.2 0 0.2 0.4 x distance from the back end of the tube;
Texture Parameter - FT - FR K6(x, /t) function describing the variation of irradiation
growth due to variations of microstructure along
Fig. 11.2.1. Comparison of measured and calculated (from texture)
the length of the tube as a function of uence;
anisotropy ratios (axial strain rate:transverse strain rate) for internally
pressurized tubes under irradiation creep conditions, from Ref. [118]. The C d1 ; C d2 anisotropy factors due to texture for in-reactor
open square represents CANDU pressure tubes. FT and FR are dened in thermal creep in a given direction d and for stress
Section 3. exponents, n, of 1 and 2, respectively;
C d4 x; C d6 x anisotropy factors due to texture for irradia- F x F b F f F b x=6;

tion creep and growth, respectively, in a given Gx Gb Gf Gb x=6; 11:3:3
direction d along the length of the tube;
Kc, Kg constants for irradiation creep and growth, respec- H x 1:5 F x Gx:
tively; Here Fb, Ff, Gb and Gf are the values of Hills anisot-
Q1, Q3, Q4, Q6, K5 activation temperatures and constant, ropy constants F and G at the back and front ends of the
respectively; tube.
r1, r2 eective stresses in MPa for thermal creep and Using the terminology employed in Eq. (11.3.2), the
stress exponents, n, of 1 and 2; anisotropy factors due to texture for in-reactor thermal
r(x) eective stress for irradiation creep, MPa; or irradiation creep are given by
T temperature, K;
/ fast ux, n/m2/s (E > 1 MeV); C ri H i rr ra Gi rt rr ;
t irradiation time, s. C ti Gi rt rr F i ra rt ; 11:3:4
C ai F i ra rt H i rr ra :
The equivalent stresses r1, r2 and r(x) are related to the
radial, axial, and transverse stress rr, ra and rt, respec- Here i = 1, 2, 4 or 6, and C is a function = nction of (x)
tively, by means of the Hills anisotropy coecients when i = 4 or 6.
[74,111], namely, The coecient describing the end-to-end eect of the
rate of irradiation creep along the length of the tube is
ri F i ra rt 2 Gi rt rr 2 H i rr ra 2 1=2 : given by
11:3:2 K 4 x K 41 K 42 x: 11:3:5
The subscript i stands for 1 (i.e., n = 1), 2(n = 2), or in the Here K4 1 and K4 2, are constants derived from in-reac-
case of irradiation creep ri = r(x). tor stress relaxation tests from specimens taken from dier-
The values of the Hills anisotropy coecients for ther- ent positions relative to the back end of a tube.
mal creep, Fi, Gi and Hi (i = 1, 2) are average values cal- The growth coecient describing the end-to-end eect
culated from the crystallographic textures of pressure and the dependence of growth on uence is given by
tubes based on the non-linear polycrystalline model,
K 6 x; /t K 61 K 62 x1 C=B/t: 11:3:6
SELFPOLY-N [119], assuming a combination of slip on
prism, basal and pyramidal systems that is consistent with Here K6 1 and K6 2 are constants derived from the aver-
out-reactor biaxial, uniaxial and torsion creep tests on age end-to-end variation in grain thickness and from a the-
small tubes with textures and microstructure similar to oretical model for the eect of grain thickness on growth
pressure tubes. The dependencies of Hills anisotropy [54]. The growth anisotropy factors are given by
constants for thermal creep on x were neglected because
of the relatively small magnitude of the thermal C a6 x Gba Gfa Gba x=6;
component. C t6 x Gbt Gft Gbt x=6; 11:3:7
Once the anisotropy parameters for thermal creep C r6 x C a6 x Gt6 x:
have been established, the values of K1, K2 and Q1
(17 000 K) for high temperature thermal creep are Here Gba , Gfa , Gbt and Gft are anisotropy parameters derived
derived from high temperature biaxial creep data from from polycrystalline irradiation growth data from OSIRIS
the WR1 Reactor and uniaxial creep data from NRU [12,52]. The single crystal growth strain rate tensors are
[15]. Similarly, the values of K3 and Q3 (1000 K) are deconvoluted from the polycrystalline data using SELF-
obtained from diametral strain data at the edge of the POLY and the single crystal creep compliance coecients
core of a pressure tube irradiated in NRU [14,25]. In mentioned above. SELFPOLY is then used to calculate
both cases, the data reect thermal creep of irradiation Kg, Gba , Gfa , Gbt and Gft . The temperature dependence of
hardened materials. growth represented by Q (3000 K) is derived from the
The values of the Hills anisotropy coecients for growth data from OSIRIS.
irradiation creep, Fi, Gi and Hi, i = 4, are average values Once the growth and thermal creep parameters are
derived from the crystallographic textures of power reactor established, the irradiation creep rates can be extracted
pressure tubes using the self-consistent polycrystalline from the power reactor data (in this case data from the
model, SELFPOLY [115], based on single crystal creep pressure tubes in Pickering A were used to obtain biaxial
compliances. The single crystal creep compliances are data comprising only irradiation creep). Values of Kc, Q4
deconvoluted from in-reactor data as described in Section (9900 K) and K5 (1.1 107) could then be derived.
10.2 [117]. Fig. 11.3.1 shows the model prediction of the diametral
The Hills anisotropy constants for irradiation creep strain proles of pressure tubes in Pickering A. Note that
depend on the distance, x, from the back end of the tubes, the model is tted to the average strain rate of the
and for a 6-m tube this dependence is given by Pickering A data at the 3.0 m position, but the shapes of
10 10
-1
-
Diametral StrainRate x 10 - h
Diametral Strain Rate x 10 - h

8
8
8 8
6 6
4 4
2 2
0 0
0 1 2 3 4 5 6 0 1 2 3 4 5 6
Distance from Inlet - m Distance from Inlet - m
10 10
-1
-1

8
8
8 8
6 6
4 4
2 2
0 0
0 1 2 3 4 5 6 0 1 2 3 4 5 6
Distance from Inlet - m Distance from Inlet - m
Fig. 11.3.1. Comparison of predicted and measured strain rate proles for Zr2.5Nb pressure tubes in a Pickering A reactor: (a,b) back end at outlet and
(c,d) back end at inlet, from Ref. [14].
20 50
-1
Measured Diametral Strain Rate x 10 - h
CANDU 6
8
Bruce A 40
Elongation - mm
15 Pickering A
30
10
20
10
5
0
0 1 2 3 4 5 6
0 -25 -2
Neutron Fluence x 10 - n.m , E>1MeV
0 5 10 15 20
8 -1
Predicted Diametral Strain Rate x 10 - h Fig. 11.3.3. Comparison of the model predictions with elongation data for
pressure tubes in a Bruce A reactor, from Ref. [14].
Fig. 11.3.2. Comparison of measured and predicted diametral strain rates
of Zr2.5Nb pressure tube in CANDU reactors, from Ref. [14].
tends to under-predict peak diametral strain rates for
the proles are derived from other information. The model CANDU 6 pressure tubes with outlet temperatures above
is a reasonably good predictor of the diametral strain rates, 570 K, Fig. 11.3.2. This tendency, which at the time was
Fig. 11.3.2, and elongation rates, Fig. 11.3.3 of Zr2.5Nb thought to be within the predictive uncertainty of the
pressure tubes in Bruce A and Pickering A, operating tem- model, has since proven to be even greater [120].
peratures up to 570 K. It also predicts well the axial and
diametral strain rates for small tubes with textures similar 12. Outstanding issues and future work
to pressure tubes irradiated at high fast neutron ux at
553 and 583 K [12], Fig. 11.3.4, and diametral strain rates Over 40 years of in-reactor testing and over 30 years of
for pressure tubes operating at temperatures of up to operating experience in power reactors have provided a
670 K at low stress (<80 MPa), Fig. 11.3.5. However, it broad understanding of the in-reactor deformation of
200 structural factors are clearly important. Although the tex-

tures of pressure tubes are not likely to change much in
the foreseeable future, this deciency may also indicate a
-1
-h
150 deciency in the capability for making predictions for

-8
stress states for which no experimental information exists.

Measured Strain Rate x 10
At best, the uncertainties of such predictions are larger

100 than those for predictions for stress states for which exper-
imental data exist. The recent development of accurate
techniques to measure the thickness of pressure tubes in-
50
service [121] allows the variation in anisotropy along the
length of a pressure tube to be assessed, an important step
in verifying or improving the axial variations in deforma-
0
tion represented by Eq. (11.3.7).
-50
12.2. More severe reactor operating conditions
-50 0 50 100 150 200
8
Calculated Strain Rate x 10 , h
-1 The model described in Section 11.3 was normalized to
data from the Pickering A Nuclear Generating Station and
Fig. 11.3.4. Comparison of measured and predicted axial and transverse tested against data from the Bruce A Station and early data
strain rates for micro-pressure tubes irradiated at 2 1018 n m2 s1,
from CANDU 6 units which operate at higher tempera-
E > 1 MeV, in Osiris at 553 and 583 K, from Ref. [75].
tures and fast uxes than exist in the Pickering A units.
Already one could see a tendency to under-predict the
CANDU 6 data, Fig. 11.3.2. Although, in 1996, this dis-
70
crepancy was deemed to be within the predictive error of
60 the model, we now know that there is a systematic under-
-1
prediction of peak diametral strain rates for later reactors

Calculated Strain Rate x 10 - h
50 with more severe operating conditions. This requires at

8
least a rescaling of the parameters of the model, and per-

40 haps a re-examination of the interaction of temperature
and fast neutron ux in the range of operating conditions
30
of these reactors.
20
12.3. Sources of variability
10
Diametral strain rates of nominally similar pressure
0 tubes, under similar operating conditions in a given power
reactor can vary by 30%, while elongation rates can
-10
vary by 20%. There also appear to be reactor to reactor
-10 0 10 20 30 40 50 60 70
8 -1 variations [122]. Some of these variations appear to be
Measured Strain Rate x 10 - h
due to dierences in microstructure, texture and possibly
Fig. 11.3.5. Comparison of measured and predicted strain rates of cold- chemical composition, Figs. 4.14.5, and are also clearly
worked Zr2.5Nb pressure tubes irradiated in the WR1 reactor at related in some way to manufacturing variables. This
temperatures between 573 and 673 K, from Ref. [14].
appears to be a protable area for further study [34]. Sim-
ilar studies on the variability of mechanical properties of
cold-worked Zr2.5Nb pressure tubes and an extensive early pressure tubes led to a substantial improvement in
data-base upon which to base models for maintaining exist- fracture toughness by the introduction of quadruple- vs
ing reactors, and for designing new ones. However there double-arc melting during primary ingot production
are still signicant gaps in our knowledge and on-going [123,124].
work is required in a number of areas.
12.4. Creep at aws
12.1. Polycrystalline modelling The principle that creep can relax stresses at aws has
been clearly demonstrated experimentally. The susceptibil-
As stated in Section 11.2, models for anisotropy of creep ity of notched specimens of pressure tube material is signif-
and growth that are based solely on crystallographic tex- icantly less if the notched specimen is held at temperature,
ture suer from some deciency when one attempts to under load, to allow creep to relax the notch-tip stress prior
extrapolate over a wide range of textures. Other micro- to testing for DHC susceptibility [125].
The model for localized creep at the high stress regions 12.8. Creep at the pressure tube ends
near aws is based on a very limited data set and this cer-
tainly needs to be augmented. The least understood aspect of in-reactor deformation
of pressure tubes is the transition region between thermal
12.5. Temperature cycling eects creep and irradiation creep, i.e., the fast ux range of
0.16 1016 n m2 s1, E > 1 MeV. This corresponds to a
Because they have a high capital cost and a low oper- region about 0.3 m long, adjacent to the end-ttings at each
ating cost, most existing nuclear generating stations end of the reactor, where the bending moment is high. The
operate as base-load electricity suppliers. Future nuclear current representation of the creep behaviour in this region
plants may operate in a load-following mode, particular is crude, and is not a good match to the data. It is very
if, as seems likely, nuclear power occupies a larger share likely that a more accurate model will change the results
of the electricity supply in some markets. Operating reac- of sag and contact predictions, and it is therefore most
tors in load-following mode may involve signicant tem- important that eorts to address the deformation rates in
perature changes. The large transients in growth strain this fast ux range [25] should continue.
observed after temperature changes, see Section 5, suggest
that temperature cycling may contribute a component to Acknowledgements
the strain, dierent from that predicted by the models
presented in Section 11. This area requires further I would like to acknowledge John Slade of New Bruns-
investigation. wick Power for permission to use previously unpublished
data from the Point Lepreau Nuclear Generating Station,
12.6. Creep rupture and Grant Bickel, Pam Simons and Eric Nadeau of AECL
for providing the gures showing that data. I would also
The case for operating pressure tubes to creep strains of like to thank Navid Badie, Grant Bickel and Stephen
the order of 5% is based on the supposition that pressure Douglas of AECL for permission to refer to their unpub-
tubes will exhibit quasi-superplastic behaviour under irra- lished data, Malcolm Griths of AECL for providing
diation creep conditions because of the low stress sensitiv- micrographs and Nick Christodoulou of AECL for provid-
ity of the strain rate, see Section 8. As pressure tubes in ing digital versions of some of the gures. The preparation
current reactors are approaching this level of peak diame- of this review was sponsored by the Natural Sciences and
tral strain (some tubes in the older CANDU 6 reactors Engineering Research Council of Canada, Ontario Power
have strains >4%), it is important to assess the microstruc- Generation Inc., the CANDU Owners Group Inc., and
ture at the peak-strain location to verify that there is no Nu-Tech Precision Metals Inc., under the Industrial Re-
evidence of impending failure due to an as-yet unidentied search Chair program in Nuclear Materials at Queens
mechanism associated with the high levels of radiation- University.
induced mass transport. This is especially important in Figs. 2.1.3, 2.1.4, 3.2, 4.1, 4.34.5, 5.1.1, 5.1.2, 5.3.1
view of the microchemical changes induced by the thermal 5.3.4, 5.4.1, 5.6.1, 5.6.2, 6.2.1, 6.6.1, 7.2.1, 9.19.6, 11.2.1,
ux, see Section 12.7. 11.3.111.3.5 were adapted, with permission, from pro-
ceedings of the 6th, 8th, 9th, 10th, 11th, 12th, 13th, 14th
12.7. Transmutation eects and 15th International Symposia on Zirconium in the
Nuclear Industry, ASTM STPs 824, 1023, 1132, 1245,
A little recognized fact is that Nb transmutes to Mo as 1295, 1354, 1423, 1457 and the Journal of ASTM Interna-
a result of irradiation by thermal neutrons. It is estimated tional Vol. 2 copyright ASTM International, 100 Barr Har-
that about 25% of the Nb is transmuted to Mo at the bor Drive, West Conshohocken, PA 19428.
mid-core location after 30 years of operation [126]. The Figs. 3.4, 3.5, 5.1.3, 5.5.1, 5.5.2, 6.2.1, 6.3.2, 6.4.1, 6.4.2,
alloy content will be changed from 2.52.8 wt%Nb to 6.4.6, Figs. 8.2, 10.1.1, 10.1.2 and 11.3.4 were adapted with
1.92.1 wt%Nb and 0.60.7%Mo. There is very little permission from the Journal of Nuclear Materials, Vol-
information on irradiation behaviour of alloys containing umes 60, 98, 149, 159, 317 and 335, copyright Elsevier B.V.
Mo except for Excel (Zr3.5 wt%Sn0.8 wt%Nb
0.8 wt%Mo) which behaves similarly to Zr2.5Nb, but
References
this information only applies to neutron doses corre-
sponding to only 12 dpa, well before any signicant long [1] Canadian Standards Association Standard CSA N286.6 Series 88,
term microstructural evolution begins. Therefore, we have Materials Standards for Reactor Components for CANDU
no information on the potential consequences in terms of Nuclear Power Plants, General Instruction No. 1, 1st Ed., Toronto,
microstructural or microchemical evolution, and subse- 1994.
[2] Canadian Standards Association Standard CSA Z299 Series 85-
quent impact on deformation. Clearly, if such eects are
CAN, Quality Assurance Program Categories 1, 2, 3 & 4, General
signicant they must occur gradually during the rst Instruction No. 1, 2nd Ed., Toronto, 1997.
80% of life, but there still may be some consequences near [3] Canadian Standards Association Standard CSA N285.2-99-CAN/
the end-of-life. CSA, Requirements for Class 1C, 2C and 3C Pressure Retaining
Components and Supports in CANDU Nuclear Power Plants, [26] J.E. Harbottle, in: C.E. Coleman (Ed.), The Technology of
General Instruction No. 1, 2nd Ed., Toronto, 1999. Zirconium, International Atomic Energy Agency, submitted for
[4] Canadian Standards Association Standard CSA N289.1-80-CAN3, publication.
General Requirements for Seismic Qualication of CANDU [27] D.G. Franklin, G.E. Lucas, A.L. Bement, in: Creep of Zirconium
Nuclear Power Plants, General Instruction No. 1, 2nd Ed., Toronto, Alloys in Nuclear Reactors, ASTM STP 815, 1983.
1992. [28] B.A. Cheadle, in: Proceedings of the Third International Symposium
[5] Canadian Standards Association Standard CSA N285-4-94-CAN/ on Zirconium in the Nuclear Industry, ASTM STP 633, 1977, p. 457.
CSA, Periodic Inspection of CANDU Nuclear Power Plant Com- [29] G.J. Field, J.T. Dunn, B.A. Cheadle, Can. Metall. Quart. 24 (1985)
ponents, General Instruction No. 1, 2nd Ed., Toronto, 1994. 181.
[6] Canadian Standards Association Standard CSA N285.8-05, Tech- [30] R.A. Holt, N. Christodoulou, G. Bickel, J. Nucl. Mater., in press.
nical Requirements for In-service Evaluation of Zirconium Alloy [31] C.E. Ells, E.F. Ibrahim, A.R. Causey, in: Proceedings of the
Pressure Tubes in CANDU Reactors, Toronto, 2005. International Conference on Creep, JSME/IMechE/ASME/ASTM,
[7] R.A. Holt, H. Wong, Nucl. Energy 41 (2002) 69. Tokyo, 1986, p. 59.
[8] J.R. Hopkins, E.G. Price, R.A. Holt, H.W. Wong, in: ASME/JSME [32] L.K.H. Leung, J.S. Jun, G.R. Dimmick, D.E. Bullock, W.W.R.
Conference on Pressure Vessels and Piping, San Diego, July 1998. Lynch, H.C. Suk, in: Proceedings of the Seventh International
[9] E.G. Price, R.A. Holt, H.W. Wong, P.J. Ellis, J. Slade, in: Conference on CANDU Fuel, The Canadian Nuclear Society,
Proceedings of the IAEA/COG Technical Meeting on Operating Toronto, Ont., 2001.
Experience in Heavy Water Reactors, Mangala, Romania, Septem- [33] A.R. Causey, V. Fidleris, S.R. McEwen, C.W. Schulte, in: Proceed-
ber 1998. ings of the 13th International Symposium on Inuence of Radiation
[10] G. Van Drunen, R.J. Dunn, W.R. Mayo, D.A. Scott, Atomic on Material Properties, ASTM STP 956, 1987, p. 54.
Energy of Canada Ltd., Report AECL-12026, June 1999. [34] G. Bickel, AECL, unpublished data, 2005.
[11] R.R. Hosbons, P.H. Davies, M. Griths, S. Sagat, C.E. Coleman, [35] V. Fidleris, A.R. Causey, R.A. Holt, in: F.A. Garner, D.S. Gelles,
in: Proceedings of the 12th International Symposium on Zirconium F.W. Witten (Eds.), Optimizing Materials for Nuclear Applications,
in the Nuclear Industry, ASTM STP 1354, 2000, p. 122. The Metallurgical Society, Inc., Warrendale, PA, 1985, p. 35.
[12] R.A. Holt, A.R. Causey, M. Griths, E.T.C. Ho, in: Proceedings of [36] T.P. Byrne, D.R. Metzger, M. Leger, in: Proceedings of the 11th
the 12th International Symposium on Zirconium in the Nuclear Conference on Structural Mechanics in reactor Technology (SMiRT
Industry, ASTM STP 1354, 2000, p. 86. 11), Tokyo, paper B13/1, 1991.
[13] B.A. Cheadle, C.E. Coleman, D.K. Rodgers, P.H. Davies, C.K. [37] J.R. Hopkins, in: Proceedings of the Second International Confer-
Chow, M. Griths, in: Proceedings of the International Conference ence on CANDU Maintenance, Canadian Nuclear Society, Tor-
on CANDU Maintenance, Canadian Nuclear Society, November onto, paper 4B, 1987.
1988. [38] M. Liska, B. Churchill, I. Laughlin, in: Proceedings of the Sixth
[14] N. Christodoulou, A.R. Causey, R.A. Holt, C.N. Tome, N. Badie, Annual Conference of the Canadian Nuclear Society, CNS,
R.J. Klassen, R. Sauve, C.H. Woo, in: Proceedings of the 11th Toronto, 1985, p. 12.9.
International Symposium on Zirconium in the Nuclear Industry, [39] R.G. Fleck, E.G. Price, B.A. Cheadle, in: Proceedings of the Sixth
ASTM STP 1295, 1996, p. 518. International Symposium on Zirconium in the Nuclear Industry,
[15] R.A. Holt, A.R. Causey, V. FidlerisProceedings of the International ASTM STP 824, 1984, p. 88.
Conference on Dimensional Stability of Irradiated Metals and [40] R. Choubey, S.A. Aldridge, J.R. Theaker, C.D. Cann, C.E.
Alloys, vol. 1, British Nuclear Energy Society, 1983, p. 175. Coleman, in: Proceedings of the 11th International Symposium
[16] J. Hopwood, in: Proceedings of the 26th Annual Conference of the on Zirconium in the Nuclear Industry, ASTM STP 1295, 1996, p.
Canadian Nuclear Society, paper P23, 2005. 657.
[17] R.A. Holt, N. Christodoulou, A.R. Causey, J. Nucl. Mater. 317 [41] M. Griths, R.A. Holt, J. Li, S. SaimotoMicrostructural Science,
(2003) 256. vol. 26, ASM International, Materials Park, 1999, p. 293.
[18] N. Badie, R.A. Holt, C.W. Schulte, R.G. Sauve, in: Proceedings of [42] C.E. Coleman, R.W. Gilbert, G.J.C. Carpenter, G.C. Weatherly, in:
the Ninth Annual Conference of the Canadian Nuclear Society, June Phase Stability During Irradiation, The Metallurgical Society of
1988. AIME, Pittsburg, 1980, p. 857.
[19] R.G. Sauve, N. Badie, R.A. Holt, in: Proceedings of the 10th [43] M. Griths, J.E. Winegar, Atomic Energy of Canada Ltd., Report
International Conference on Structural Materials in Reactor Tech- AECL-10835, December 1994.
nology (SMiRT 10), vol. D, 1989, vol. L, p. 219. [44] V. Perovic, A. Perovic, G.C. Weatherly, L.M. Brown, G.R. Purdy,
[20] N. Badie, G.D. Morandin, R. Sauve, in: Proceedings of the ASME R.G. Fleck, R.A. Holt, J. Nucl. Mater. 205 (1993) 251.
Pressure Vessel and Piping Conference, Boston, 1999, vol. 385, p. [45] R.A. Holt, J. Nucl. Mater. 59 (1976) 234.
247. [46] R.A. Holt, M. Griths, R.W. Gilbert, J. Nucl. Mater. 149 (1987) 51.
[21] M. Griths, C.K. Chow, C.E. Coleman, R.A. Holt, S. Sagat, V.F. [47] M. Griths, D. Sage, R.A. Holt, C.N. Tome, Metall. Mater. Trans.
Urbanic, in: Proceedings of the Eects of Radiation on Materials: 33A (2002) 859.
16th International Symposium, ASTM STP 1175, 1993, p. 1077. [48] A. Akhtar, J. Nucl. Mater. 47 (1973) 79.
[22] M. Griths, W.G. Davies, G.D. Moan, R.A. Holt, A.R. Causey, [49] R.A. Holt, P. Zhao, J. Nucl. Mater. 335 (2004) 520.
S.A. Aldridge, in: Proceedings of the 13th International Symposium [50] J.J. Kearns, U.S. Atomic Energy Commission Research and
on Zirconium in the Nuclear Industry, ASTM STP 1423, 2002, p. Development Report (Contract AT-11-1-GEN-14), WAPD-TM-
796. 472, Bettis Atomic Power Laboratory, November 1965.
[23] N. Christodoulou, P.A. Turner, B.W. Leitch, C.N. Tome, N. Badie, [51] R.A. Holt, E.F. Ibrahim, Acta Metall. 27 (1979) 1319.
R.G. Sauve, in: Proceedings of the THERMEC 2000, International [52] R.G. Fleck, J.E. Elder, A.R. Causey, R.A. Holt, in: Proceedings of
Conference on Processing and Manufacturing of Advanced Mate- the 10th International Symposium on Zirconium in the Nuclear
rials, Las Vegas, Nevada, 2000. Industry, ASTM-STP 1245, 1994, p. 168.
[24] B.W. Leitch, N. Christodoulou, C.N. Tome, P.A. Turner, in: Sixth [53] A.R. Causey, AECL, private communication, 2000.
International Conference on Residual Stress, Oxford, UK, July [54] R.A. Holt, J. Nucl. Mater. 159 (1988) 310.
2000. [55] V. Fidleris, J. Nucl. Mater. 159 (1988) 22.
[25] M. Griths, N. Christodoulou, S.A. Donohue, in: Proceedings of [56] J. Winton, R.A. Murgatroyd, in: J.P. Pemsler, E.C.W. Perryman,
the 14th International Symposium on Zirconium in the Nuclear W.W. Smelzer (Eds.), Zirconium and its Alloys, The Electrochemical
Industry, J. ASTM Int. 2 (2005) 686. Society, Inc., New York, 1966, p. 322.
[57] J.D. Parker, V. Perovic, M. Leger, R.G. Fleck, in: Proceedings of the [91] M. Griths, J.F. Mecke, J.E. Winegar, in: Proceedings of the 11th
Ninth International Symposium on Zirconium in the Nuclear International Symposium on Zirconium in the Nuclear Industry,
Industry, ASTM STP 939, 1987, p. 86. ASTM-STP 1295, 1996, p. 580.
[58] R.A. Holt, R.G. Fleck, in: Proceedings of the Eighth International [92] V.F. Urbanic, M. Griths, in: Proceedings of the 12th International
Symposium on Zirconium in the Nuclear Industry, ASTM STP Symposium on Zirconium in the Nuclear Industry, ASTM-STP
1023, 1989, p. 705. 1354, 2000, p. 641.
[59] R.A. Holt, A.R. Causey, J. Nucl. Mater. (2004) 529. [93] M. Griths, P.H. Davies, W.G. Davies, S. Sagat, in: Proceedings of
[60] T.S. Gendron, R.A. Holt, R.G. Fleck, in: Proceedings of the the 13th International Symposium on Zirconium in the Nuclear
International Conference on Irradiation Technology, Commissariat Industry, ASTM-STP 1423, 2002, p. 507.
a lEnergie Atomique and Commmission of the European Commu- [94] R. Bullough, M.H. Wood, J. Nucl. Mater. 90 (1980) 1.
nities, Paris, paper V4, 1992. [95] R.A. Holt, C.H. Woo, C.K. Chow, J. Nucl. Mater. 205 (1993) 293.
[61] R.A. Holt, R.G. Fleck, in: Proceedings of the Ninth International [96] C.H. Woo, Radiat. E. Defects Solids 144 (1998) 145.
Symposium on Zirconium in the Nuclear Industry, ASTM STP [97] C.H. Woo, U.M. Gosele, J. Nucl. Mater. 119 (1983) 219.
1132, 1991, p. 218. [98] C.H. Woo, J. Nucl. Mater. 159 (1988) 237.
[62] R.G. Fleck, R.A. Holt, V. Perovic, J. Tadros, J. Nucl. Mater. 159 [99] H. Weidersich, Radiat. E. 12 (1972) 111.
(1988) 75. [100] J.R. Matthews, M.W. Finnis, J. Nucl. Mater. 159 (1988) 257.
[63] G.M. Hood, H. Zou, D. Gupta, J. Schultz, J. Nucl. Mater. 223 [101] A.H. Cottrel, Atomic Energy Research Establishment Harwell
(1995) 122. report AERE M/M1 02, 1955.
[64] A.D. King, G.M. Hood, R.A. Holt, J. Nucl. Mater. 185 (1991) [102] P.K. Madden, R. Hesketh, Central Electricity Generating Board
174. Report, CEGB RD/B/N3113, 1975.
[65] S.N. Buckley, in: D.J. Littler (Ed.), Properties of Reactor Materi- [103] G.R. Piercy, J. Nucl. Mater. 26 (1968) 18.
als and Eects of Radiation Damage, Butterworth, London, 1962, p. [104] J. Weertman, Trans. Am. Soc. Met. 61 (1968) 681.
413. [105] S.R. MacEwen, V. Fidleris, Philos. Mag. 31 (1975) 1149.
[66] C.N. Tome, N. Christodoulou, Philos. Mag. A 80 (2000) 1407. [106] J. Gittis, Philos. Mag. 25 (1972) 345.
[67] R.A. Holt, J. Nucl. Mater. 82 (1979) 419. [107] A.A. Semenov, C.H. Woo, J. Nucl. Mater. 233237 (1996) 1045.
[68] R.A. Holt, A.R. Causey, J. Nucl. Mater. 150 (1997) 306. [108] P.T. Heald, M.V. Speight, Acta Metall. 23 (1975) 1389.
[69] R.A. Holt, T.M. Holden, A.R. Causey, V. Fidleris, in: Proceedings [109] R. Bullough, M.R. Haynes, J. Nucl. Mater. 65 (1977) 184.
of the 10th Ris International Symposium on Metallurgy and [110] W.G. Wolfer, J. Nucl. Mater. 90 (1980) 175.
Materials Science: Materials Architecture, Ris National Labora- [111] R. Hill, The Mathematical Theory of Plasticity, Oxford University,
tory, Roskilde, Denmark, 1989, p. 383. London, 1954.
[70] C.N. Tome, N. Christodoulou, P.A. Turner, M.A. Miller, C.H. [112] A.R. Causey, J. Nucl. Mater. 54 (1974) 64.
Woo, J. Root, T.M. Holden, J. Nucl. Mater. 227 (1996) 237. [113] M.J. Pettigrew, S.B. Lambert, in: Proceedings of the Fifth Interna-
[71] A.R. Causey, C.H. Woo, R.A. Holt, J. Nucl. Mater. (1988) 225. tional Conference of Structural Mechanics in Reactor Technology
[72] O.J.V. Chapman, R.J. McElroy, B.E. Sheldon, in: Proceedings of (SMiRT-5), Berlin, paper L8-2, 1979.
the Sixth International Symposium on Zirconium in the Nuclear [114] R.G. Sauve, in: Proceedings of the 13th Annual Reactor Simulation
Industry, ASTM STP 824, 1984, p. 343. Symposium, Atomic Energy of Canada Ltd., Report CRNL-4139,
[73] N. Christodoulou, P.A. Turner, C.N. Tome, C.K. Chow, R.J. 1987.
Klassen, Metall. Mater. Trans. A 33A (April) (2002) 1103. [115] C.N. Tome, C.B. So, C.H. Woo, Philos. Mag. A67 (1993) 917.
[74] E.F. Ibrahim, R.A. Holt, J. Nucl. Mater. 91 (1980) 311. [116] C.N. Tome, N. Christodoulou, R.A. Holt, C.H. Woo, R.A.
[75] A.R. Causey, R.A. Holt, N. Christodoulou, E.T.C. Ho, in: Lebensohn, P.A. Turner, in: Proceedings of the 15th Ris Interna-
Proceedings of the 12th International Symposium on Zirconium in tional Symposium on Materials ScienceNumerical Predictions of
the Nuclear Industry, ASTM STP 1354, 2000, p. 74. deformation Processes and the Behaviour of Real Materials, Ris
[76] P.A. Ross-Ross, V. Fidleris, D.E. Fraser, Can. Metall. Quart. 11 National Laboratory, Roskilde, Denmark, 1994.
(1972) 101. [117] N. Christodoulou, A.R. Causey, C.H. Woo, C.N. Tome, R.J.
[77] D.E. Fraser, P.A. Ross-Ross, A.R. Causey, J. Nucl. Mater. 46 Klassen, R.A. Holt, in: 16th International Symposium on the Eects
(1973) 281. of Radiation on Materials, ASTM STP 1175, 1993, p. 1111.
[78] A.R. Causey, J. Nucl. Mater. 98 (1981) 313. [118] A.R. Causey, J.E. Elder, R.A. Holt, R.G. Fleck, in: Proceedings of
[79] C.E. Coleman, A.R. Causey, V. Fidleris, J. Nucl. Mater. 60 (1976) the 10th International Symposium on Zirconium in the Nuclear
185. Industry, ASTM-STP 1245, 1994, p. 202.
[80] R.J. Klassen, V. Fidleris, AECL unpublished data, 1996. [119] P.A. Turner, C.N. Tome, N. Christodoulou, C.H. woo, Philos. Mag.
[81] J. Schroeder, M. Holicky, J. Nucl. Mater. 33 (1969) 52. A 79 (1999) 2505.
[82] B.A. Cheadle, S.A. Aldridge, J. Nucl. Mater. 47 (1973) 255. [120] R.A. Holt, A.R. Causey, W.G. Davies, unpublished data, 1998.
[83] A.R. Causey, R.A. Holt, S.R. MacEwen, in: Proceedings of the [121] W.R. Mayo, G. Van Drunen, D. Kalenchuk, R. Gunn, in:
Eighth International Symposium on Zirconium in the Nuclear Proceedings of the Sixth CNS International Conference on CANDU
Industry, ASTM-STP 824, 1984, p. 269. Maintenance, Canadian Nuclear Society, Toronto, Ont., paper 3a20,
[84] C.E. Coleman, J.F.R. Amber, Rev. Coat. Corros. 3 (1979) 105. 2003.
[85] M. Resta-Levi, M.P. Puls, in: Proceedings of the 18th International [122] N. Badie, D. Leemans, private communication, 2003.
Conference on structural Mechanics in reactor Technology (SMiRT [123] I. Aitchison, P.H. Davies, J. Nucl. Mater. 203 (1993) 206.
18), Beijing, paper G10-3, 2005. [124] J.R. Theaker, R. Choubey, G.D. Moan, S.A. Aldridge, L. Davis,
[86] M.F. Banham, A.J. Fudge, J.A. Tibbles, B.E. Sheldon, R.G. Fleck, R.A. Graham, C.E. Coleman, in: Proceedings of the 10th Interna-
R.A. Holt, in: Proceedings of the Seventh ASTM Euratom Sympo- tional Symposium on Zirconium in the Nuclear Industry, ASTM-
sium on Reactor Dosimetry, Kluwer Academic Publishers, Norwell, STP 1245, 1994, p. 221.
MA, 1990, p. 749. [125] S. Sagat, G.W. Newman, D.A. Scarth, in: Proceedings of the
[87] D.A. Woodford, Mater. Sci. Eng. 4 (1969) 146. International Conference for Hydrogen Eects on Material Behav-
[88] D.A. Woodford, Trans. Am. Soc. Met. 62 (1969) 291. iour and Corrosion Deformation Interactions, Moran, Wyoming,
[89] E.F. Ibrahim, J. Nucl. Mater. 96 (1981) 297. 2002.
[90] K.P. Steward, B.A. Cheadle, Trans. AIME 239 (1967) 504. [126] S. Douglas, AECL, unpublished data, 2005.

Jonathan Mckinley Holt JNM 2008

Uploaded by

Copyright:

Available Formats

Jonathan Mckinley Holt JNM 2008

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Jonathan Mckinley Holt JNM 2008

Uploaded by

Copyright:

Available Formats

Available online at www.sciencedirect.

Journal of Nuclear Materials 372 (2008) 182214

In-reactor deformation of cold-worked Zr2.5Nb pressure tubes

Received 20 March 2006; accepted 20 February 2007

1. Introduction tured and operated to standards published by the Cana-

Fig. 1.1.1. Schematic diagram of a CANDU fuel channel.

Flux - n.m .s x10 , E>1MeV

creep and irradiation growth [14,33]. Both of these phe-

0.006 zirconium, and the pronounced crystallographic texture

[15], and much reactor maintenance is still planned based

Fig. 3.2(a). The elongated a-grains tend to be somewhat

Table 3.1 placement of atoms from their normal position in the

5. Irradiation growth 5.1. Eect of fast neutron uence

Relative Strain Rate

Relative Strain Rate

Fig. 4.1. Correlation of in the elongation rate of pressure tubes in a

0 5.2. Eect of fast neutron ux

strain can be plotted vs fast uence, independent of the fast -0.05

5.3. Eect of temperature -0.15

Fig. 5.3.4. Growth strain as a function of neutron uence showing the

0.6 varied independently of other parameters notably the

+ Maximum 9 + Maximum 5.3

Table 5.5.1 A deliberate attempt to take advantage of these charac-

0.2 be kept in mind that other variables (e.g., the degree of

Normalized Growth Rate x10 - m .n , E>1MeV

10 in the range 0.30.5 lm, where the eect of grain size is

5.7. Eect of residual stress

-5 Intergranular residual stresses are developed in Zr alloys

0.16 e_ ic / expT act =T 5:5 106 ; 6:3:2

400 350 300 250

that irradiation creep varies approximately linearly with 2 in flux

increases as shown graphically in Fig. 6.4.3 which is based

6.5. Eect of crystallographic texture

materials, Fig. 6.4.2 [76,77]. Holt and Ibrahim [51,67,74] 10

conditions of stress, temperature history and hydrogen iso- 25

aw where the stress is highest because of the strong sensi-

The assumed failure mechanism is ductile necking. The

relationship between ductility and stress sensitivity is

Fig. 9.1. The number density of these increases rapidly in

12 the transverse direction contain mainly 1=31 1 2 3 edge

Initially these tend to split into two partials 1= 62 0

8 broadening) but also actually reecting some increase in

10. Mechanisms of irradiation-induced deformation

This section will strictly deal with the deformation

60 the vast majority of displacement damage comes from the

50 peratures (<400 K) where the vacancies are immobile, the

30 10.1. Irradiation growth

(for vacancies) and a dilation (for SIAs) in the direction 10

The factor which determines the net ux of a particular

Goesele [97,98]. They proposed that vacancies and SIAs 2 microns

force to segregate the defects amongst sinks of dierent ori- -2

C d4 x; C d6 x anisotropy factors due to texture for irradia- F x F b F f F b x=6;

Diametral Strain Rate x 10 - h

Diametral Strain Rate x 10 - h

200 structural factors are clearly important. Although the tex-

150 deciency in the capability for making predictions for

stress states for which no experimental information exists.

At best, the uncertainties of such predictions are larger

prediction of peak diametral strain rates for later reactors

50 with more severe operating conditions. This requires at