Imp

A. Sawatzky 1 and C. E.
Ells 2
Understanding Hydrogen in Zirconium
REFERENCE: Sawatzky, A. and Ells, C. E., "Understanding Hydrogen in Zirconium,"
Zirconium in the Nuclear Industry: Twelfth International Symposium, ASTM STP 1354, G. P.
Sabol and G. D. Moan, Eds., American Society for Testing and Materials, West Conshohocken,
PA, 2000, pp. 32-48.
ABSTRACT: Selected examples of the Canadian contribution to the development of under-
standing of the behavior of hydrogen in zirconium alloys are reviewed. To solve the technical
problems faced by the nuclear industry, it was necessary to have values for the solubility and
diffusivity of hydrogen in the alloys, the diffusivity needed for the three gradients of concen-
tration, temperature, and stress. With these quantities known, sensible models for the accu-
mulation of hydrogen at points in the reactor components and the morphology of the resultant
hydride phase were developed.
KEYWORDS: zirconium, hydrogen, zirconium hydride, Zr-2.5Nb, Zircaloy, diffusion, delayed
hydride cracking
In accepting the Kroll award, we would like to use this medium to thank the committee
for judging us as deserving of this high honor. We are very grateful. The committee noted
that "one of its objectives is to collect first-hand accounts on key aspects of the history of
the application of zirconium and its alloys" and went on to request from us a paper "outlining
the pertinent aspects of the effects of hydrogen and hydrides on the structure and properties
of zirconium-based alloys." In attempting to meet the intent of the request, we have limited
ourselves to some aspects of the Canadian contribution to the understanding of hydrogen
(behavior) in zirconium alloys and, where it seems essential, putting this work in the context
of the worldwide effort on the topic. In the space available in this publication, we can touch
only on a few aspects of the vast literature on hydrogen in zirconium alloys. We deal largely
with topics on which we ourselves worked, and in keeping with the Kroll tradition we are
allowed to greatly over-emphasize our personal contributions. The dependence of the Ca-
nadian-designed CANDU on zirconium alloy tubes to provide the pressure containment
places a particular urgency on avoiding any hydrogen/hydride-induced deterioration of the
component. Thus, in this paper we concentrate on work devoted to the behavior of hydrogen
in the pressure tubes.
The first Canadian power reactor conceptual design, NPD-1, had a steel pressure vessel,
but the concept was changed to zirconium alloy pressure tubes in NPD-2 in 1957. The
Canadian work on hydrogen in zirconium dates from that decision, although one of us, CEE,
had worked previously on the topic. In the sections of text below we will generally describe
the work in the chronological order of its being done; again, because of space constraints
we can list only a few key references. We merely touch on the matter of mechanical prop-
erties and do not cover any aspects of the important topic of hydrogen ingress from water-
P.O. Box 137, Pinawa, Manitoba ROE IL0, Canada.
2 P.O. Box 838, Deep River, Ontario K0J 1P0, Canada. (Tragically, Dr. Ells died in a car accident on
November 30, 1999.)
32
Copyright* 2000 byASTM International www.astm.org

Copyright by ASTM Int'l (all rights reserved); Sun Mar 6 23:27:34 EST 2011
Downloaded/printed by
Arthur Moses Thompson Motta (Pennsylvania State Univ) pursuant to License Agreement. No further reproductions authorized.
SAWATZKY AND ELLS ON HYDROGEN IN ZIRCONIUM 33
side corrosion. We do not need to distinguish between hydrogen (protium) and deuterium
and will list all concentrations of these isotopes in hydrogen equivalent.
Terminal Sol i d Solubility
The phase diagram of the Zr-H system has received an extraordinary amount of attention
in several countries, as detailed in recent reviews [1-3]. From the viewpoint of the CANDU
reactor, the ~/ ~ + hydride phase boundary in the Zr-2. 5Nb alloy (and to a lesser extent in
Zircaloy-2, -4) is an item of overwhel mi ng importance. This boundary is universally termed
the TSS (for terminal solid solubility); when plotted as log hydrogen concentration versus
reciprocal temperature, the boundary is generally linear. At the early stages of the CANDU
development program it was speculated [4] that it might be acceptable to have hydrogen
concentrations in the range 200 to 400 mg/ kg present in the Zr alloy component s at reactor
operating temperatures, compared to the TSS at 300~ of about 70 mg/ kg. However, with
the advent of problems from delayed hydride cracking (DHC), the acceptable level has been
reduced. Due to the large local volume expansion associated with hydride formation, it may
be extremely difficult i f not impossible to measure a true equilibrium value of TSS. In
practice, the solubility line for precipitation during cooling, TSSP, is found at significantly
lower temperatures than the line for dissolution during heating, TSSD. TSSP is dependent
on the thermal history of the specimen prior to the start of cooling. The literature has often
referred to the differences between TSSP and TSSD as a "hysteresis, " and contains various
attempts to explain the source.
In 1957 Sawatzky, studying Zircaloy-2, made the first of the Canadian measurements of
TSS [5]. His diffusion technique resulted in values of TSSP, and these were the reference
values for AECL in the first hal f of the 1960s. Using a metallographic technique, Daniel [6]
in 1965 found TSS(D) values in Zircaloy-2 somewhat lower than those in Ref 5, and at
hydrogen concentrations below about 60 mg/ kg his values were significantly higher than
those predicted from the extrapolation of the higher concentrations, a result now discarded.
In 1967 Kearns [7] at the Bettis Laborat ory reviewed the bulk of the data on TSS(D) in
zirconium, Zircaloy-2, and Zircaloy-4 available from the open literature; his estimate of the
best fit of the data, often termed the "Kearns' line," has long been a standard to follow.
Sawat zky and Wilkins [8] shortly afterwards measured TSSD in both Zircaloy-2 and Zr-
2.5Nb, using a thermal diffusion technique. Their data were indistinguishable from the
Kearns' line. Sawat zky felt that the higher TSSP values were those of a metastable hydride
phase. Comi ng to CRL in the early 1960s, Erickson continued [9] his pressure-temperature-
composi t i on measurements, attempting to find the reason for the hysteresis. While clearly
establishing the existence of the hysteresis, and feeling that much of it was due to a meta-
stable hydride phase, Erickson did start thinking about the effect of strain energy from
hydride formation.
Following the DHC-related leaks in some of the pressure tubes at Pi cket i ng in 1974, there
has been a succession of attempts to refine the measurement of the TSS values in zi rconi um
alloys. In 1975 workers at WL, 3 particularly Ritchie, initiated experiments to make use of
the Poynt i ng effect, the internal friction spectrum, and the dependence of the elastic modulus
on hydride concentration. Of the three techniques, the change in elastic modulus has proven
most useful, clearly demonstrating the difference between TSSD and TSSP, and is still in
use. Pan et al., in the most recent summary [10] of TSS in the Zr-2. 5Nb alloy, have used
3 The laboratory at Pinawa, Manitoba will be listed as WL, the laboratory at Chalk River, Ontario as
CRL, and that of Ontario Hydro in Toronto as OHRD.

34 Z I R C O N I U M I N T H E N U C L E A R I N D U S T R Y : T W E L F T H S Y M P O S I U M
the changes in elastic modul us as the primary experimental technique. Workers at WL found
that metallography appeared useful for hydrogen concentrations as low as 8 mg/ kg in pure
zirconium, but was ineffective for Zr-2.5Nb. At CRL, Col eman and Ambl er [11] used the
onset of DHC to give an effective TSSD with data points close to the Keams' line, Differ-
ential scanning calorimetry, measuring the heat absorbed or evolved when hydrides dissolve
or precipitate, was introduced by Tashiro at OHRD in the late 1980s, and has since been
exploited by Khatamian and co-workers [12,13]. An essential component of the work on
TSS is measurement of the hydrogen concentration in the sample. Al t hough commerci al
equipment using hot vacuum extraction has been satisfactory, at CRL a HVEMS technique
(hot vacuum extraction mass spectrometry) developed with many man-years of work [14]
gives better accuracy in addition to distinguishing between the hydrogen isotopes.
Erickson and Hardie [15] seem to have been the first to address the effect of alloy additions
on TSS. They reported on a range of alloys and found little difference in TSS between them.
More recently it has been shown [16] that the alloy Excel, Zr-3. 5Sn-0. 3Mo-0. 8Nb, does have
TSS significantly higher than that of the other alloys studied. From all of the work thus far,
the Kearns line and its linear extrapolation to lower concentrations, i.e., about 5 mg/ kg,
remains the standard for TSSD in zirconium, the Zircaloys, and Zr-2.5Nb. A recent summary
of TSS values for Zr-2.5Nb as obtained by various techniques is presented in Fig. 1.
The various elastic and plastic stresses and strains caused in the hydride and surrounding
lattice on hydride formation have a key role in both the hysteresis in TSS observed and in
effects on TSS from externally applied stress. The Canadian work on this aspect is largely
due to Puls, who started thinking about the problem(s) immediately following the first Pick-
T ( ~
3 5 2 2 8 3 2 2 7 1 8 2 1 4 4
2 0 0 , . ~ I J l I
O} 1 0 0 - ~ " " \ "
E ' " % ,
o
"'"" ' c ~ . " . .
rr" T SSD ~ . " ' - : . ~ '
o
z
0 - T S S P
7 ~ I ' - ' % ,
' - A" 1 % . . .
0
rr
r , , T S S D
>.-
- r 2 i I , I , I i I
1 . 6 1 . 8 2 . 0 2 . 2 2 . 4
1 0 0 0 / T ( K ' I )
FIG. 1--Values of TSS in Zr-2.5Nb obtained by different experimental techniques. Sources of the lines
shown are: TSSD (a) calorimetr3, (b) dynamic elastic modulus. TSSP: (c) calorimeto', (d) Slatterys
data, (e) ~,namic elastic modulus (TSSP2 ), (f) dynamic elastic modulus (TSSP1). The TSSPI results
from cooling from a higher temperature than used f or TSSP2.

ering leaks in 1974 and has continued the work to the present time. (To enable this work to
be done, crystallographic relationships, not known in 1974, were required, and many of the
gaps were filled by Carpenter and co-workers). A measurement of the partial mol ar volume
of hydrogen in the alloys was made [17]. After Puls' preliminary work he elucidated the
effects in a series of four papers [18-21 ], which now constitute the definitive explanation
of hysteresis in metal hydrogen systems. Even a summary of his work on the topic is outside
the scope of this paper.
Diffusion
lsotheJrnal Diffilsion
Because Zircaloy-2 was initially chosen for use as fuel cladding and pressure tube material
in Canadian nuclear reactors, the diffusion of hydrogen in it was investigated over the tem-
perature range 260 to 560~ [5]. Zircaloy-2 rods, with thin hydride layers on the surface,
were heated at the desired temperature long enough to give a suitable hydrogen distribution.
The mid-third of the specimen was then machi ned into seven layers and each anal yzed for
hydrogen by vacuum extraction. The diffusion coefficient was then obtained from the radial
hydrogen distribution by the method described in Crank [22 ]. The results were in fair agree-
ment with those of Mallett and Albrecht for zirconium [23].
The col d-worked Zr-2. 5Nb used for pressure tubes (and some fuel cladding in the WR-1
reactor) as-fabricated has a microstructure consisting of c~-grains elongated in the extrusion
direction separated by a network of 13-filaments. It was expected that the diffusivity of hy-
drogen would be higher in the [3-phase than in the a-phase. The axial hydrogen diffusivity
of this material was investigated over the temperature range 200 to 700~ [24]. In the method
used, a rod of material was hydrided to somewhat less than TSS at the test temperature, butt
welded to a similar unhydrided rod, and heated long enough to give a measurable hydrogen
distribution. The diffusivity was then determined from the axial hydrogen distribution. The
hydrogen diffusivity in Zr-2. 5Nb was somewhat higher than in zi rconi um as observed by
Kearns [25], who used the same experimental method. However, it was about 50 times higher
in as-extruded Zr-20Nb, whi ch is [~-phase, than in zirconium. The experimental diffusivity
in Zr-2.5Nb was much less than predicted using a simple mathematical model based on the
microstructure. The high temperature used in hydriding the specimen and the diffusion anneal
had changed the microstructure of the [3-filaments. Skinner and Dutton [26], using a met hod
in which tritium tracers were implanted over one surface of the specimen, were able to make
diffusivity measurements in times short enough to avoid much change in the microstructure.
Their experimental hydrogen diffusivities for as-extruded Zr-2.5Nb were in good agreement
with predicted values.
Thermal Diffusion
High radial temperature gradients were thought to contribute to the hydride distribution
seen in some defected Zircaloy-2 fuel cladding, so it was decided to investigate the thermal
diffusion of hydrogen in this alloy [27]. The specimens used were hydrided Zircaloy-2
cylinders with constant temperature gradients along their axes. A typical result, obtained
after a 34-day anneal, is shown in Fig. 2. The hydrogen concentration is plotted against
temperature along the specimen and, as shown by the data points, there is a sharp cusp in
hydrogen concentration at about 250~ From the TSS curve, the region to the left of the
discontinuity is a+~- phase, and that to the right is a-phase. (Markowitz [28] also investigated

36 ZI RCONI UM IN THE NUCLEAR INDUSTRY: TWELFTH SYMPOSI UM
130 150
1000 I
TEMPERATURE (~
200 250 300 350 400 450
] I I I I I
n,"
I -
Z
Z
0
0
Z
ill
0
0
Q
>-
I
9 0 C -
800
700
6 0 0 - -
500
400
300
/
100 - -
TERMI NAL SOLI D SOLUBI LI TY
/ /
/
/ - -
/ /
/
/
/
/
/
/ /
. . . . . " I N[ ] ] AL H CONCENTRATI ON
/
/
/
. . . . . q I
0 0 4 0 . 8 12 1 . 6 2 . 0
P OS mONA L ONGS P E CI MA NI c m~
~ G. 2 - - T h e thermal diffusion of hydmgen in Zircaloy-2.
the thermal di ffusi on o f hydrogen in zi rconi um usi ng cylindrical geometry. He t oo observed
a sharp change in hydrogen concentrati on at the a + M a interface, but it remai ned fairly
constant over the c~+5 region. Thi s was s hown to be strictly because o f the geomet ry. )
To expl ai n the s omewhat unexpect ed hydrogen distribution o f Fig. 2, the differential equa-
tion gi vi ng the hydrogen flux as f ound in Denbi gh [29] was used. She wmo n [30] had adapted
the equati on t o the a + 5 regi on usi ng the assumpti ons that: (a) hydrogen is the onl y el ement
di ffusi ng, (b) the hydride (8) is in equi l i bri um wi th the sol i d sol ut i on phase ( a) at every
point, and (c) di ffusi on occurs predomi nantl y in the a- phase. The equati on was s ol ved for
hydrogen flux and concentrati on in the a + 5 regi on for the case o f the linear geomet ry used
in the experi ment. The theoretical hydrogen distribution, usi ng known val ues for TSS, dif-
f usi on coeffi ci ent, and heat o f transport, was cal cul ated and is s hown by the sol i d line in
Fig. 2. The agreement is very good. The s ol i d l i ne in the a- regi on is the steady state distri-
buti on under a temperature gradient and is al so in g o o d agreement wi th experi ment. Sawat zky
and Vogt [31 ] gave a more ext ensi ve mathemati cal treatment o f the probl em, taking account
o f the di f f usi on in the 5-phase, move me nt o f the a + 5 / 5 interface, variable temperature

gradi ent , and hydr ogen pi ckup. Mar i no [32] devel oped a comput er met hod based on the
wor k of Sawat zky and Vogt that l ed to very good agr eement wi t h the exper i ment al resul t s
in Fi g. 2 as wel l as wi t h Mar kowi t z' s resul t s [28].
Several results may be obt ai ned f r om Refs 27 and 31. At every poi nt in the ct +8 regi on
the hydr ogen concent rat i on i ncreases cont i nuousl y, and there is a hydr ogen flux t owar d the
col der regi on. I f the hydr ogen di ffuses in the a- phase, the flux is det er mi ned by the t em-
perat ure gradi ent and the TSS gradi ent . The hydr ogen flux at the col d surface l eads to the
f or mat i on of" a hydr i de l ayer at this point. The movement of the a +8/o~ i nt erface is of interest.
For the exper i ment al hydr ogen di st ri but i on in Fi g. 2, the hydr ogen flux t oward the a+S/ ~x
i nt erface is less than away from it and the i nt erface moves t oward the col der end. On the
ot her hand, i f the hydr ogen pi ckup at the hot end l eads to a hydr ogen di st ri but i on in the a-
regi on such that the concent rat i on gradi ent at the i nt erface is the same as that of the TSS
curve, as gi ven by Curve A, the i nt erface posi t i on r emai ns fixed. For hi gher pi ckup rates,
the i nt erface moves t owar d hi gher t emperat ures. In fact, for hi gh enough pi ckup rates, t he
i nt erface may be at the hot surface, in whi ch case a sol i d hydr i de l ayer wi l l f or m at bot h
hot and col d surfaces wi t h an a + ~ regi on in bet ween. Thi s result has been obser ved in
defect ed fuel cl addi ng.
Hydride Orientation
Workers in zi r coni um met al l urgy over the past t wo decades may not real i ze the ext ensi ve
concern fel t in the 1960s on the mat t er of hydri de ori ent at i on and stress ori ent at i on of
hydri des. Al t hough it was wel l known that hydri des coul d have del et eri ous effects on duc-
t i l i t y paramet ers, Mar shal l and Lout han [33] showed that wi t h the normal s to the hydr i de
pl at el et s paral l el to a t ensi l e stress the t ensi l e duct i l i t y coul d be r educed to negl i gi bl e val ues
at r oom t emperat ure from a hydr ogen concent rat i on as smal l as 40 mg/ kg. (Al t hough Ref
33 is dat ed 1963, the i nformat i on was r el eased vi a progress report s in 1961.) Thi s resul t had
wor l dwi de i mpl i cat i ons for the use of zi r coni um al l oy structural component s. Each of the
c ount r i e s - - Uni t e d States, Uni t ed Ki ngdom, Japan, Sweden, and Ge r ma n y - - h a d si gni fi cant
rol es in est abl i shi ng the fabri cat i on procedures that l argel y el i mi nat ed any t ensi l e fai l ures
f r om ori ent ed hydr i des in react or component s. Nevert hel ess, Schemel wrot e [34] "The wor k
at At omi c Ener gy of Canada Lt d. ( AECL) pr ovi ded much of the t echni cal dat a on whi ch
the commer ci al use of zi r coni um was based. "
The first CRL wor k on di r ect i onal preci pi t at i on of hydri des in Zi r cal oy- 2 was compl et ed
in 1962 [35], demonst r at i ng that the hydr i de pl at el et normal s t ended to form in di rect i ons
per pendi cul ar to the di rect i on of pri or strain of the speci men. CRL st aff then wor ked wi t h
Wah Chang, of Al bany, Oregon, to devel op a fabri cat i on route for r ol l ed pl at e up to 50 mm
t hi ck in whi ch the subsequent hydr i de pl at el et s had nor mal s in the t hi ckness di rect i on of the
sheet [36]. Thi s mat eri al gave much mor e sat i sfact ory speci mens than di d ei t her of the thin-
wal l ed tubes, thin sheet, or rod pr evi ousl y used. Evans and Parry [36] st udi ed the i nt errel at i on
bet ween hydr i de ori ent at i on, hydr ogen concent rat i on, and test t emperat ure on t ensi l e duct i l -
ity. When the test t emperat ure reached 300~ the duct i l i t y was i ndependent of hydr i de
ori ent at i on. Parry went on to st udy stress or st ress-i nduced ori ent at i on in both Zr - 2. 5Nb-
0. 5Cu [37] and Zr - 2. 5Nb [38]. Pri ce [39] and t hen Bel l and Duncan [40] were to make
furt her cont ri but i ons to knowl edge on the Zr-2. SNb alloy. Some year s l at er Cheadl e et al.
[41] were to refine the stress ori ent at i on measurement s on tubes, usi ng speci mens that
avoi ded the flattening used in many previ ous studies.
El l s [42] in 1968 revi ewed al l of the open l i t erat ure i nformat i on on ori ent at i on and stress
ori ent at i on of hydri des in the zi r coni um al l oys. By that t i me t here was sufficient i nformat i on
to rel at e hydr i de ori ent at i on to fabri cat i on procedure wi t h consi der abl e confi dence. The basi c

3 8 ZIRCONIUM IN THE NUCLEAR INDUSTRY," TWELFTH SYMPOSIUM
mechani sms of stress reori ent at i on r emai ned unknown, but short l y El l s [43] showed that an
ext ensi on of a t reat ment by Li [44] pr ovi ded at l east a qual i t at i ve expl anat i on of the exper-
i ment al observat i ons. Two decades later, Puls [45], maki ng use of the advances in under-
st andi ng he had devel oped in the i nt eract i on bet ween hydri des and stress, put the expl anat i on
on a mor e sat i sfact ory basis.
Zi rconi um Al l oys i n an Organi c- Cool ed React or
The Feasibility of Using Zirconium
It was deci ded, in the l at e 1950s, that Canada woul d bui l d an exper i ment al or gani c- cool ed
nucl ear react or ( WR- 1) usi ng nat ural urani um oxi de or car bi de as fuel. The use of organi c
cool ant al l owed a hi gher t emperat ure than heavy wat er at equi val ent or l ower pressures, wi t h
the target t emperat ure of 400 versus 300~ for heavy wat er gi vi ng a hi gher t her modynami c
efficiency. The use of natural urani um fuel pr ecl uded the use of st eel for pressure tube or
fuel cl addi ng mat eri al . A si nt ered al umi num mat eri al ( SAP) and ber yl l i um were consi der ed
for this purpose, but bot h were very brittle. The possi bi l i t y of cl addi ng zi r coni um wi t h a
thin met al l ayer was consi dered, but the onl y feasi bl e met al was al umi num and no met hod
of r eadi l y appl yi ng it was found. The Amer i cans had l ooked i nt o usi ng uncl ad zi r coni um
al l oys, but found t hey hydr i ded ver y r api dl y at 400~ in organi cs such as Sant owax OM
[46-48]. However, Shannon [49] had shown that addi ng smal l amount s of wat er vapor to
hydrogen gas r educed the hydr i di ng rate at 400~ by several orders of magni t ude. He at-
t ri but ed t hi s to wat er mol ecul es t aki ng up the si t es on the zi r coni um oxi de ont o whi ch
ot herwi se the hydr ogen mol ecul es woul d be absorbed. It seemed wort hwhi l e to see whet her
wat er woul d have the same effect of reduci ng the hydr i di ng rate of zi r coni um in organi cs
(Sant owax OM and HB-40). In fact, Trout ner [50] had obt ai ned a l ower - t han- expect ed hy-
dri di ng rate in organi cs in at l east one of his tests, whi ch was thought possi bl y due to wat er
in the organi c. An organi c l oop (X-7) in the NRX react or at CRL enabl ed testing bot h in
and out of reactor. Uncl ad Zi r cal oy- 2 in this l oop pi cked up hydr ogen at 400~ at a rate
several orders of magni t ude l ess than in unr epl eni shed aut ocl aves at Canadi an West i nghouse
operat i ng at the same t emperat ure. It seemed that the wat er in the organi c cool ant in the X-
7 l oop was enough to provi de prot ect i on agai nst hydri di ng. When the wat er level in the
aut ocl aves was mai nt ai ned at about 100 mg/ kg, the hydr i di ng rate was al so decr eased by
several orders of magni t ude. The pr el i mi nar y results showed that it mi ght be possi bl e to use
zi r coni um al l oys in or gani c- cool ed react ors. The ear l y tests on zi r coni um in organi cs were
carri ed out mai nl y at CRL in 1961-63, descr i bed by Sawat zky [51 ]. The wor k on hydr ogen
pi ckup by zi r coni um al l oys in organi cs aft er 1963 was done at WL. It resul t ed in Zr-2. 5Nb
and Ohzhenni t e 0.5 becomi ng the al l oys of choi ce for use in or gani c- cool ed reactors. Thi s
wor k is report ed by Boul t on [52].
Hydrogen Sinks in Zirconium Components
The heavy wat er moder at or in the WR- 1 react or was cont ai ned in a tank, the cal andri a,
whi ch had a lattice of vert i cal al umi num cal andr i a t ubes t hrough whi ch the pressure t ubes
passed. CO2 fl owi ng down the annulus bet ween the t wo t ubes served as t hermal i nsul at i on.
The t emperat ure dr op across the gas annul us was about 300~ I f the pressure t ubes were
of zi rconi um, it was t hought that zi r coni um fins, wel ded to the out si de of the pressure tubes,
mi ght be exposed to a sufficient t emperat ure gr adi ent to serve as hydr ogen sinks, t her eby
i ncreasi ng the pressure tube life. Wi l ki ns and Sawat zky [53] demonst r at ed in an out - r eact or

test that the hydrogen concentration in the pressure tube could be maintained at well below
TSS for many years.
A practical demonstration that thermal diffusion can be used to protect load-beating reactor
component s against hydrogen embrittlement arose by accident [54 ]. The conventional fuel
elements in WR-1 consisted of UC or UO2 pellets in zi rconi um alloy tubular sheaths with
zirconium wire-wrap spacers spot-welded to the outside. A gas plenum about 1 cm long at
the upper end of each element maintained the fission gas pressure at acceptable levels.
Several fuel elements without the wire wrap were found to have defected over the pl enum
as a result of heavy hydriding, after 200 days in the reactor at a sheath temperature of 450
to 500~ The sheath over the pl enum was 100 to 150~ colder than over the fuel, and most
of the hydrogen picked up by the sheath for a distance of about 7 cm along the fuel pellets
had diffused to the plenum, whereas the rest of the sheath was heavily hydrided. A mathe-
matical model gave a hydrogen distribution along the sheath in excellent agreement with
that observed. The wire wrap in the conventional fuel elements, being colder than the sheath,
was heavily hydrided, leaving the sheath itself hydride-free at the operating temperature,
thereby serving as a hydrogen sink.
Hydr oge n Di s t ri but i on i n Fuel Channel s
Joi nt s i n t he WR- 1 Re a c t o r
The pressure tubes in the WR-1 reactor were fabricated from zi rconi um alloys, preoxi-
dized, and mechanically connect ed to the out-of-core steel piping by sandwich-type steel
joints, Fig. 3a. Since steel is fairly permeable to hydrogen, two pressure tubes were removed
from the reactor in 1971 after about 3000 h of operation at 350~ and the joints checked
for hydriding. The lower joint had picked up about 80 mg/ kg, but the upper joint had picked
up about ten times this amount. An analysis of the hydrogen distribution suggested that in
i i 1 1 1 1 [ 1 1 / / / / 1 1 / 1 1 1 1 , / / / , ' / / 1 1 / 1 1 1 1 1 / 1 1 1 1 1 1 / / / 1 1 / / / I ; / /
5 C 0 2 1
~ ~ I / / / / J . . / J _ I / / ~ / _ Z . L L / J I / / / / / , / / , ' 1
f [ f / / l Y l l l / / / / / / / ~ , " A ~IRCONIUM
, - - v S T E E L " - - - ' / ~ " 2 ~
/
O R G A N I C C O O L A N T
Fig. 3a. The sandwi ch j oi nt i n WR- 1.
+
S T E E L
N2(+Dz )
Zr - 2.5% Nb J
D~OCOOLANT
Fig. 3b. The rolled joint in Pickering.
FIG. 3- - Joi nt s between pressure tube and out-core circuits.

40 ZIRCONIUM IN THE NUCLEAR INDUSTRY: TWELFTH SYMPOSIUM
the lower joint most of the hydrogen had entered by Route 2, Fig. 3a, whereas in the upper
joint a considerable amount of hydrogen had also entered from the CO2 annulus, Route 5.
The hydrogen probabl y entered the annulus through the steel extensions, Route 4. Because
of a 300~ temperature drop across the annulus, the pressure tube may have acted as a gas
separation tube [55], leading to a higher hydrogen concentration over the upper joint than
over the lower one.
A duplicate of the pressure tube annulus system was set up in the out-reactor arm of a
WR-1 organic loop and operated under in-reactor conditions except for the neutron flux. The
annulus had gas sampling positions every 40 cm. The CO2 in the reactor annuli flowed
downward at about 400 mm/ mi n, and at this flow rate the hydrogen concentration over the
upper joint was about 10% by volume and 0.03% over the lower joint. A mathematical
analysis of the situation suggested that downflow of the gas tended to allow the hydrogen
concentration t o build up, whereas reversing the flow would permit the hydrogen to be swept
out. The gas flow direction was reversed and the same flow rate led to hydrogen concentra-
tions of 0.1 and 0.03% over the upper and lower joints, respectively. The hydriding of the
upper joint from the annulus could therefore be largely eliminated by having the CO2 flow
upward rather than downward. Two other approaches to reduce hydriding under the joint
were considered: (a) bare zirconium wire embedded in the joint or (b) zi rconi um fins seam
welded to the outside of the tube inboard of the joint to serve as a hydrogen sink. Either of
the approaches probably woul d have been effective, but work on the organi c-cool ed reactor
was stopped to allow concentration on the heavy water-cooled CANDU.
Joints in the Pickering-3 Reactor
The Zr-2. 5Nb pressure tubes in Pickering-3 had been pre-oxidized to reduce deuteriding.
The DHC cracks in the tubes in whi ch leaks were first observed near the rolled joints in
1974 may have been initiated before reactor startup. Incorrect rolling procedures led to high
tensile residual stresses, and the hydrogen concentration in the as-installed tubes was suffi-
cient to cause delayed hydride cracking during col d shutdowns. However, enhanced deute-
rium concentrations found in the joints [56] raised the possibility of the hydrogen equivalent
becomi ng sufficiently large at later times for DHC to occur at operating temperatures. The
possible routes by which deuterium might enter the zirconium in the joints are shown in
Fig. 3. To get a better understanding of the hydriding mechanism, a number of joints were
rolled ont o short lengths of pre-oxidized Zr-2. 5Nb tubes and placed in an out-reactor hot
water loop at operating conditions. At regular time intervals a joint would be removed and
the hydrogen distribution determined. A mathematical analysis indicated that the hydrogen
entered the tube fairly uni forml y over the joint, at a rate decreasing with time. The outer
surface of the tubes showed regions where there was little oxide. An unexposed joint was
cut up and much of the oxide on the outer surface found to be missing, probabl y damaged
during rolling. The leakage of minute quantities of water along the st eel / zi rconi um alloy
interface apparently slowly repaired the oxide layer. These results suggest that Route 2 is
the main avenue for hydriding. The rolled joint cracki ng problem has been largely solved
by changi ng roiling practices to keep the residual stresses low. In addition, programs have
tested modified joint designs to reduce the deuteriding rate. These include longer joints to
increase the path length for deuterium diffusion, deuterium barriers such as chromi um, the
use of getters such as yttrium, and fins as originally proposed for the pressure tubes in the
organi c-cool ed reactor. The getter concept was developed to a prototype so that it could be
considered for use in future CANDU reactors.

Hydride Blisters
The fuel channel s in the CANDU react ors are hori zont al , with the pressure t ubes kept
f r om t ouchi ng the cal andr i a t ubes wi t h gart er spri ng spacers. The rupture of a Zi r cal oy- 2
pressure t ube in Pi cket i ng NGS- 2 in 1983 was due to movement of the gart er springs,
al l owi ng the pressure t ube to cont act the cool er cal andr i a tube. A row of hydr i de bl i st ers
had f or med at the poi nt s of cont act and had been i nst rument al in i ni t i at i ng the cr ack that
f ol l owed the row of blisters. The bl i st ers f or med at the col d spot s from t hermal di ffusi on.
Usi ng t hermal di ffusi on theory, met hods were devel oped for det er mi ni ng the size of a bl i st er
as a funct i on of t i me [57]. The correct i ve measures t aken for current and future CANDUs
are descr i bed in det ai l in Ref 58.
Hydr i de bl i st ers had been obser ved on a Zr-2. 5Nb pressure t ube r emoved f r om the
WR- 1 r eact or in 1977, but their si gni fi cance had not been real i zed unt i l the 1983 i nci dent
[59]. Four blisters, axi al l y al i gned, up to 1 cm in di amet er and al most t hrough-wal l , were
found in this WR-1 tube, whi ch had been i n-react or for 34 874 operat i ng hours. The t ube
had been moni t or ed for r adi at i on- i nduced el ongat i on and had been el ongat i ng at a const ant
rate when this seemed to stop suddenly. The movi ng end of the pressure tube appar ent l y
j a mme d agai nst rest rai ni ng straps, resul t i ng in bowi ng and event ual cont act wi t h the col der
cal andr i a tube. Aft er a further 6000 h of oper at i on the pressure t ube was r emoved f r om the
reactor. The t i me requi red for growt h of the l argest bl i st er was cal cul at ed to be 6770 h. Thi s
was in reasonabl e agr eement wi t h the 6000 h of est i mat ed cont act bet ween cal andr i a and
pressure t ubes and gave confi dence in the met hod used for cal cul at i ng bl i st er growth. The
OHRD st aff have made a large cont ri but i on to the underst andi ng of bl i st er growth, and then
to its prevent i on.
Delayed Hydride Cracking
The mat t er of DHC in CANDU pressure tubes has i nvol ved a l arge cont i nui ng effort in
Canada in the per i od si nce 1974. Maj or pl ayer s in the st udy have been CRL, WL, and
OHRD, wi t h ot her wor k in Canada bei ng undert aken at the Uni versi t i es of Toronto, Al bert a,
Wi ndsor, and New Brunswi ck. The t hree maj or pl ayer s al so j oi nt l y or i ndi vi dual l y empl oyed
vari ous consul t ant s f r om t i me to t i me, E. Smi t h deservi ng of speci al note. Ot her count ri es
have cont ri but ed to the underst andi ng of DHC, but the maj or proport i on of the wor k has
been done in Canada. It woul d be ri di cul ous to cl ose this paper wi t hout ment i on of DHC,
but at the same t i me i mpossi bl e to do ot her than t ouch on some of the hi ghl i ght s of the
work,
The pri nci pal charact eri st i cs of t he DHC in t he Zr - 2. 5Nb have been wel l est abl i shed.
Hydr ogen di ffuses to regi ons of hi gh t ensi l e stress at a cr ack t i p and forms hydr i de there,
event ual l y resul t i ng in an i ncr ement of crack growth, and the process repeat s at the new
cr ack tip. The dependence of cr ack vel oci t y on the stress i nt ensi t y fact or is shown sche-
mat i cal l y in Fi g. 4. There is a sharp t hreshol d (K~H in St age I) for cr ack i ni t i at i on at a stress
i nt ensi t y factor, t ypi cal l y 5 to 10 MPa~'--mm, f ol l owed by a range in whi ch the cr ack vel oci t y
is fai rl y i ndependent of the stress i nt ensi t y factor val ue (St age II), but event ual l y the fracture
t oughness of the met al mat ri x is exceeded (St age IID. In the earl y exper i ment s the vel oci t y
in St age II was shown to i ncrease exponent i al l y with t emperat ure, t ypi cal l y reachi ng 10 -6
m/ s at 300~ Lat er it was shown that at t emperat ures above some val ue the cr ack growt h
rate was much hi gher on cool down to test t emperat ure compar ed to heat up to the same
t emperat ure, whi ch expl ai ned some of the l arge var i abi l i t y obser ved in the very earl y results.
The current underst andi ng of dependence on t hermal hi st ory is i l l ust rat ed in Fig. 5. Fr om

>-
~ - I II ! I I I
~ _ S
q
' " "
>
O ~
n- :
123 : :
0 K I H K I
FIG. 4--Schematic form of the relationship between stress intensi~ f act or and DHC velocio:
the vi ewpoi nt of i nt egri t y of react or component s, par t i cul ar l y pressure tubes, the val ue of
the stress i nt ensi t y fact or was di sconcer t i ngl y l ow and the crack growt h rates di sconcer t i ngl y
hi gh: it was essent i al to l earn how to mi ni mi ze the hazard. The underst andi ng of DHC fel l
into t wo issues: (a) how the cr ack initiated, and (b) the accumul at i on of hydr ogen at the
cr ack tip that resul t ed in the cr ack growth. Here we are concer ned onl y wi t h the DHC
mechani sm as such and not wi t h the appl i cat i ons of the t heory to repair, operat e, or con-
struction of react or component s.
The first recorded observat i on of DHC in zi r coni um al l oys is pr obabl y that of Wei nst ei n
and Hol t z [60], worki ng on a cont ract sponsored j oi nt l y by AECL and the USAEC, al t hough
t hey acknowl edged some unpubl i shed wor k by Ost berg. They not ed in t hei r summary, " . . .
Zr-2. 5Nb . . . is suscept i bl e to hydr ogen- i nduced del ayed fai l ure. " The observat i ons were
for speci mens cont ai ni ng 500 mg / k g of hydr ogen st ressed to 275 MPa, and at the t i me the
result di d not seem to be of concern to the pressure tubes, whose hydr ogen concent rat i on
woul d be an order of magni t ude less and operat i ng at hal f the stress. Later, in 1971, Ai t chi son
cert ai nl y demonst r at ed DHC in the hi gh-st rengt h zi r coni um al l oy Excel , al t hough it was not
r ecogni zed as such at the t i me and the observat i on was onl y r ecor ded in i nt ernal progress
reports. Al so in 1971, Stevens and Dut t on [61] i ssued an i mport ant document on the mot i on
of hydr ogen in combi ned gradi ent s of t emperat ure and stress. They defi ned a new quantity,
V*, the vol ume of transport, in anal ogy with Q*, the heat of transport. They left the equat i ons
\
\
r
q
IM
O
(.9
q
I
1 / T ( K -~)
FIG. 5- - The dependence of DHC velocity on the temperature of approach. Arrows on the curves
show the d~fference between heating and cooling cycles. A fidl explanation of these effects is provided
in Ref 79.

SAWATZKY AND ELLS ON HYDROGEN IN ZIRCONIUM 4 3
in terms of the flux of hydrogen in the gradients. Using the mathematical technique used by
Sawat zky [27] for the thermal gradient (with stress constant), Ells and Simpson [62] made
the first attempt to integrate the flux equation to obtain the rate of hydrogen (hydride) buildup
at a region of high tensile stress (at constant temperature). Following Ai t chi son' s work, the
next observation of DHC in zi rconi um alloys was in some welded end caps of Zr-2. 5Nb fuel
sheathing, the experimental work being done by Simpson, Aldridge, and Ambler. The work
by Aldridge and Ambl er was not written for the open literature, but the Si mpson work was
published [63] with Ells as a co-author. Aldridge and Ambl er were the first to recogni ze
this failure as DHC and postulated the correct mechanism, essentially as described above.
They solved the technical probl em for the fuel sheathing merely by stress relief of the weld,
and DHC might have merited little further work were it not for the failure of the tubes in
Pickering in August 1974 and concerns for other reactors.
During the autumn of 1974 each of CRL, WL, and OHRD had initiated experimental
programs to define the parameters of DHC in the col d-worked Zr-2. 5Nb alloy (and shortly
the Hitachi laboratory in Japan started similar work on heat-treated Zr-2.5Nb). The early
published work was done using notched tensile-type specimens and then obtaining crack
extension by optical examination [64-66] or measuring time to failure [67]. Better methods
of detecting crack growth were developed, acoustic emission [68,69] at CRL and potential
drop [70] at WL, along with the use of cantilever beam and compact tension specimens.
Apart from the large number of measurements of K~ and crack velocity, various specialized
experiments were undertaken. Col eman [71] showed that the crack velocity increased with
increasing percentages of basal plane normals parallel to the stress direction. At WL, it was
demonstrated that stringers of 13-phase in the direction of crack growth increased the growth
rate [26] attributed to the higher diffusivity of hydrogen on the 13-phase compared to the ~-
phase. Ells et al. [72] showed that a tensile stress aligned parallel with the normals to the
hydride platelets increased the time required for hydride dissolution at a range of tempera-
tures; quantitative information on this topic was desired when thinking of how quickly the
hydride present at crack tips at low temperatures would be eliminated as the pressure tube
went to operating temperatures.
Coincident with the initiation of the experimental programs, in late 1974 Dutton and Puls
started an intensive study on the matter of diffusion of hydrogen to the crack tip and by
mid-1975 had achieved a sophisticated crack growth equation [73], which gave semi-
quantitative agreement with the experimental data for Zr-2.5Nb. The concept of volume of
transport was discarded. Lee-Whiting [74] over the same period developed corresponding
equations, but with more emphasis on the form of the stress field near the crack tip. In this
stage of the theory, transformation from hydride volume to crack growt h required an as-
sumption of hydride thickness. Then workers at WL, particularly Puls, but with contributions
from other colleagues, continued to fine-tune the equation and collect the data to test it. First
the discontinuous nature of the cracking was included [75]. By late 1977 considerably more
data on growth rates were available, and they developed a model [76] to differentiate between
the Stage I and Stage II growth. Late in 1978 they had started to include refinements from
the elastic strain energies at the hydrides, and several years later wrote out a final growth
rate equation [77]. With the elastic strain energies included in the theory, Ambl er [78] was
able to explain the dependence of velocity on direction of approach to test temperature. The
current up-to-date review on the theory of crack velocity is that of Shi and Puls [79]. The
differences in crack velocity between different alloys and the effect of material strength is
not yet completely understood. The extreme example encountered thus far is the factor of
10 hi gher rate in Excel alloy compared to Zr-2. 5Nb [80]. The more important matter is the
effect of irradiation on the rate in Zr-2.5Nb, studied in detail at CRL and WL [81]. In
general the rate is increased by irradiation by factors of up to 5, but the increase is a function

of starting mat eri al , i rradi at i on t emperat ure, i ncrease in yi el d st rengt h due to i rradi at i on, and
changes in the amount and mor phol ogy of the 13-phase. The t heory does not yet gi ve a
sat i sfact ory quant i t at i ve model i ncl udi ng al l of t hese vari abl es.
The cr ack i ni t i at i on process has been much more di ffi cul t to model than the cr ack growth.
Af t er 23 years of st udy i t is felt that t hree si t uat i ons must be consi dered: (a) a sharp cr ack
for whi ch a t hreshol d stress i nt ensi t y fact or KIH can be defined, Fig. 4, (b) a shal l ow smoot h
not ch to be char act er i zed by the maxi mum nor mal stress, and (c) a nomi nal l y smoot h surface
for whi ch the surface t ensi l e stress causes the cr ack initiation. In one of the earl i est studies,
Nut t al l [65] di d not find a val ue of Kin, but Col eman [67] soon found a val ue of about 5
MPa~mm for Zr-2. 5Nb at 170~ usi ng not ched t ensi on speci mens. Col eman and Ambl er
[68] were to cont i nue the wor k usi ng acoust i c emi ssi on wi t h cant i l ever beam speci mens.
Cheadl e was to be a l eader in measur ement s of cr ack i ni t i at i on on nomi nal l y smoot h surfaces
and surfaces wi t h smal l art i fi ci al defects. In part of t hi s work, Cheadl e and El l s [82] st udi ed
i ni t i at i on in bot h cant i l ever beam speci mens and i nt ernal l y pr essur i zed l engt hs of a ful l -si ze
pressure tube. One pr obl em in assembl y of the dat a on cr ack i ni t i at i on arose f r om the di f-
ficulty in deci di ng i f or not t here was a smal l crack in the hydri de at a pr e- machi ned notch.
A further pr obl em came from the effect of t hermal cycl i ng to effect or gener al l y reduce the
val ue of KIH obser ved [83 ], thus gr eat l y i ncreasi ng the number of vari abl es to be consi dered.
In the st udy of nomi nal l y smoot h surfaces there was the di ffi cul t y of char act er i zat i on of the
defect s and the coi nci dent needed to present resul t s in t erms of pr obabi l i t y of fai l ure in a
gi ven t i me for a gi ven cal cul at ed surface stress. A l arge vol ume of dat a was assembl ed on
t hese matters.
The mechani st i c goal s was the same as that for cr ack velocity, viz., obt ai n an underst and-
i ng of cr ack i ni t i at i on in t erms of basi c propert i es of t he mat eri al (s), but t here are far fewer
publ i cat i ons on t hi s t opi c t han on cr ack growth. The l at er fract ure wor k at WL is report ed
in a series of papers havi ng Puls as aut hor or co- aut hor [84-90]. Based on this work, and a
l arge series of tests undert aken by Sagat, a sat i sfact ory model for cr ack i ni t i at i on at sharp
cr ack t i ps and shal l ow not ches has been devel oped by Pul s, Shi, and Sagat. Any descr i pt i on
of the model is beyond the scope of this paper.
Conclusions
In the four decades si nce the deci si on to use zi r coni um al l oy pressure t ubes in the CANDU
reactor, a l arge quant i t y of knowl edge on the behavi or of hydr ogen in the al l oys has been
generat ed. The authors were pr i vi l eged to be abl e to cont ri but e to this knowl edge and regret
that in t he l i mi t ed space avai l abl e for t hi s paper it has not been possi bl e to acknowl edge the
i ndi vi dual cont ri but i ons from al l of t hose i nvol ved.
References
[1 ] Zuzek, E., "On Equilibrium in the Zr-H System," Surface and Coatings Technology, Vol. 28, 1986,
pp. 323-338.
[2] Zuzek, E., Abrita, J. P., San-Martin, A., and Manchester, E D., "The H-Zr (Hydrogen-Zirconium)
System," Bulletin of Alloy Phase Diagrams, Vol. 11, 1990, pp. 385-395.
[3 ] Aladjem, A. in Hydrogen Metal Systems, Ionic and Metallic Hydrides, F. A. Lewis and A. Aladjem,
Eds., Chap. 7, Scitec Publications, Acdermannsdorf, Switzerland, 1995.
[4] Ross, P. A., Evans, W., and Langford, W. J., "Experience with Zirconium-Alloy Pressure Tubes,"
AECL-4262, 1972.
[5] Sawatzky, A., "The Diffusion and Solubility of Hydrogen in the Alpha-Phase Zircaloy-2," Journal
of Nuclear Materials, Vot. 2, 1960, pp. 62-68.

[6] Daniel, A. R., "Terminal Solid Solubility of Hydrogen in Zircaloy-2," Canadian General Electric
Report DF65CAP31, June 1965.
[7] Kearns, J. J., "'Terminal Solubility and Partitioning of Hydrogen in the Alpha Phase of Zirconium,
Zircaloy-2 and Zircaloy-4," Journal of Nuclear Materials, Vol. 22. 1967.
[8] Sawatzky, A. and Wilkins, B. J. S., "Hydrogen Solubility in Zirconium Alloys Determined by
Thermal Diffusion," Journal of Nuclear Materials, Vol. 22, 1967, pp. 304-310.
[9] Erickson, W. H., "Hydrogen Solubility in Zirconium Alloys," Journal of Electrochemical Tech-
nology, Vol. 4, 1966, pp. 205-211.
[10] Pan, Z. L., Ritchie, I. G., and Puls, M. P., "'The Terminal Solid Solubility of Hydrogen and
Deuterium in Zr-2.5Nb Alloys," Journal of Nuclear Materials, Vol. 228, 1996, pp. 227-237.
[11] Coleman, C. E. and Ambler, J. F. R., "Measurement of the Effective Solvus Temperature of Hy-
drogen in Zr-2.5wt% Nb Using Acoustic Emission." Canadian Metallurgical Quarterly, Vol. 17,
1978, pp. 81-84.
[12] Khatamian, D., Pan, Z-L., Puls, M. P., and Cann, C. D., "'Hydrogen Solubility Limits in Excel, an
Experimental Zirconium-based Alloy," Journal of Alloys and Compounds, Vol. 231, 1995, pp. 488-
493.
[13] Khatamian, D. and Ling V. C., "Hydrogen Solubility Limits in ct- and B-Zirconium," Journal of
Alloys and Compounds, Vols. 253-254, 1997, pp. 162-166.
[14] Green, L. W., Bickel, G. A., Lesson, P. K., James, M. W. D., Lamarche, T. G., and Michel, H.,
"A Hot Vacuum Extraction Mass Spectrometric System for Determination of H and D in Zirco-
nium," AECL 11342, 1996.
[15] Erickson, W. H. and Hardie, D., "The Influence of Alloying Elements on the Terminal Solubility
of Hydrogen in a-Zirconium, " Journal of Nuclear Materials, Vol. 13, 1964, pp. 254-256.
[16] Khatamian, D., Pan, Z. L., Puls, M. P., and Cann, C. D., "Hydrogen Solubility Limits in Excel,
an Experimental Zirconium-Based Alloy," Journal of Alloys and Compounds, Vol. 231, 1995, pp.
488-493.
[17] MacEwen, S. R., Coleman, C. E., Ellis, C. E., and Faber, J., "Dilation of h. c. p. Zirconium by
Interstitial Deuterium," Acta Metallurgica, Vol. 33, 1985, pp. 753-757.
[18] Puls, M. P., "Comments on the Interaction of Hydrogen with Precipitates," Scripta Metallurgica,
Vol. 12, 1978, p. 595.
[19] Puls, M. P., "The Effect of Misfit and External Stresses on Terminal Solid Solubility in Hydride-
Forming Metals," Acta Metallurgica, Vol. 29, 1981, p. 1961.
[20] Puls, M. P., "Elastic and Plastic Accommodation Effects on Metal-Hydride Solubility," Acta Me-
tallurgica, Vol. 32, 1984, p. 1259.
[21 ] Puls, M. P., "On the Consequences of Hydrogen Supersaturation Effects in Zr Alloys to Hydrogen
Ingress and Delayed Hydride Cracking," Journal of Nuclear Materials, Vol. 165, 1989, p. 128.
[22] Crank, J., Mathematics of Diffusion, Oxford, 1956.
[23] Mallett, M. W. and Albrecht, W. M., "' Low Pressure Solubility and Diffusion of Hydrogen in
Zirconium," Journal of the Electrochemical Societ); Vol. 104, 1957, pp. 142-146.
[24] Sawatzky, A., Ledoux, G. A., Tough, R. L., and Cann, C. D., "'Hydrogen Diffusion in Zirconium-
Niobium Alloys," Metal-Hydrogen Systems, Proceedings of the Miami International Symposium
1981, Pergamon Press, Oxford, 1982, pp. 109-120.
[25] Kearns, J. J., "Diffusion Coefficient of Hydrogen in Alpha Zirconium, Zircaloy-2 and Zircaloy-4,"
Journal of Nuclear Materials, Vol. 43, 1972, pp. 330-338.
[26] Skinner, B. C. and Dutton, R., "Hydrogen Diffusivity in Zirconium Alloys and its Role in Delayed
Hydride Cracking," Hydrogen Effects on Material Behaviour, N. R. Moody and A. W. Thompson,
Eds., The Minerals, Metals, and Materials Society, 1990.
[27] Sawatzky, A., "Hydrogen in Zircaloy-2: Its Distribution and Heat of Transport," Journal of Nuclear
Materials, Vol. 2, 1960, pp. 321-328.
[28] Markowitz, J. M., "Hydrogen Redistribution in Thin Plates of Zirconium Under Large Thermal
Gradients," WAPD-TM-104, 1958.
[29] Denbigh, K. B., The Thermodynamics of the Steady State, Methuen & Co., Ltd., London, 1951.
[30] Shewmon, P. G., "The Redistribution of a Second Phase During Annealing in a Temperature
Gradient," Transactions of the Metallurgical Society AIME, Vol. 227, 1958, pp. 642-647.
[31 ] Sawatzky, A. and Vogt, E., "Mathematics of Thermal Diffusion of Hydrogen in Zircaloy-2," Trans-
actions of the Metallurgical Society AIME, Vol. 227, 1963, pp. 917-928.
[32 ] Marino, G. P., "A Numerical Calculation of the Redistribution of an Interstitial Solute in a Thermal
Gradient," Nuclear Science and Engineering, Vol. 49, 1972, pp. 93-98.
[33] Marshall, R. P. and Louthan, M. R., "Tensile Properties of Zircaloy with Oriented Hydrides,"
Transactions ofASM, Vol. 56, 1963, pp. 693-700.
[34] Schemel, J. H., Zirconium Alloy Fuel Clad Tubing, Publication of Sandvik Special Metals, 1989.

[351 Parry, G. W. and Evans, W., "'The Effect of Strain on the Directional Precipitation of Zirconium
Hydride in Zircaloy-2," AECL-1707, 1962.
[361 Evans, W. and Parry, G. W., "The Deformation Behaviour of Zircaloy-2 Containing Directionally
Oriented Zirconium Hydride Precipitates," Electrochemical Technolog 3, Vol. 4, 1966, pp. 225-231.
[37 Parry, G. W., "Strain Induced Directionality of Zirconium Hydride Precipitates in Zirconium-
2.5wt% Niobium-0.5wt% Copper Alloy," AECL-1888, 1963.
[38 Parry, G. W., "Stress Reorientation of Hydrides in Cold Worked Zirconium-2.5% Niobium Pressure
Tubes," AECL-2624, 1966.
[391 Price, E. G., "Hydride Orientation and Tensile Properties of Z5-2.5%Nb Pressure Tubing Hydrided
While Internally Pressurized," Canadian Metallurgical Qttarterl3; Vol. 11, 1972, pp. 129-138.
[401 Bell, L. G. and Duncan, R. G., "Hydride Orientation in Zr-2.5%Nb; How it is Affected by Stress,
Temperature and Heat Treatment," AECL-5110, 1975.
[411 Cheadle. B. A., Coleman, C. E., and Ipohorski, M., "Orientation of Hydrides in Zirconium Alloy
Tubes," Zirconium in the Nuclear Induso T (6th blternational Symposium), ASTM STP 824, 1984,
pp. 210-221.
[421 Ells, C. E.. "'Hydride Precipitates in Zirconium Alloys," Journal of Nuclear Materials, Vol. 28,
1968, pp. 129-151.
[43 ] Ells, C. E., "Stress Orientation of Hydride in Zirconium Alloys," Journal of Nuclear Materials,
Vol. 35, 1970, pp. 306-315.
[44] Li, J. C. M., "'Nucleation of Fe~6N 2 Under Applied Stress," Transactions ofASM, Vol. 60, 1967,
pp. 226-227.
[45] Puls, M. P., "Effect of Stress on Hydride Reorientation in Zirconium Alloys," Proceedings of the
International Symposium on Solid-Defect Interaction, Kingston, Ontario, August 1985, pp. 426-
433.
[46] Boyd, W. K., Etickson, G. L., Stewart, O. M., Miller, P. D., and Peoples, O. S., "'Corrosion Studies
in Organic Heat-Exchanging Media," BMI-1160, 1957.
[47] Kline, H. E., Gioseffi, N. J., and Bley, W. N., "Dynamic Corrosion in Polyphenyls Under Irradi-
ation," NAA SR-2046, 1957.
[48] Davies, H. E_, "The Compatibility of Some Materials with Santowax," AERE M/ M153, 1957,
[49] Shannon, D. W., "' The Role of the Oxidation Rate on the Hydriding of Zircaloy-2 by Gaseous
Hydriding," HW-67811, 1960.
[50] Troutner. V. H., "Hydrogen Corrosion in Organic Reactor Coolants," Corrosion, Vol. 16, 1960,
pp. 281t-285t.
[51 ] Sawatzky, A., "The Behaviour of Zirconium Alloys in Santowax OM Organic Coolant at High
Temperatures," AECL-2118, 1964.
[52] Boulton. J., "Behaviour of Zirconium Alloys in Organic Coolants," Canadian Metallurgical Quar-
terl3, Vol. 11, 1972, pp. 1-5.
[53 ] Wilkins, B. J. S. and Sawatzky, A., "Thermal Diffusion of Hydrogen in Finned Zircaloy-2 Pressure
Tubes," AECL-2634, 1966.
[54 ] Sawatzky, A., "Hydrogen Distribution in Zirconium Sheathing of Nuclear Reactor Fuels," Hydro-
gen in Metals, No. 2 in the ASM Materials/Metal Working Technical Series, I. M. Bernstein and
A. W. Thompson, Eds., 1973.
[55] Grew, K. E. and Gibbs, T. L., Thermal Diffusion and Gases, Cambridge University Press, 1952.
~56] Perryman, E. C. W., "' Picketing Pressure Tube Cracking Experience," Nuclear Energ3, Vol. 17,
1978, pp. 95-105.
[57] Sawatzky, A., "Formation of Hydride Blisters in Zirconium Alloy Pressure Tubes," Canadian
Metallurgical Quarterl3; Vol. 24, 1985, pp. 227-234.
[58] Price, E. G., "'Highlights of the Metallurgical Behaviour of CANDU Pressure Tubes," AECL 8338,
1984.
[59] Sawatzky, A. and Brady, G. R., "'The Growth of Hydride Blisters in the Zr-2.5wt% Nb Pressure
Tube WRI-12(ZN),' " WNRE-651, 1983.
[60] Weinstein, D. and Holtz, E C., "'Susceptibility of Zirconium and Zirconium Alloys to Delayed
Failure Hydrogen Embrittlement," Transactions ofASM, Vol. 57, 1964, pp. 284-293.
[61 ] Stevens, R. N. and Dutton, R., "'The Effect of Temperature, Pressure and Concentration Gradients
on Diffusion," AECL-3793, 1971.
[62] Ells, C. E. and Simpson, C. J., "'Stress Induced Movement of Hydrogen in Zirconium Alloys,"
Proceedings of the International Conference on Effects of Hydrogen in Metals, AIME, [973, pp.
345-360.

SAWATZKY AND ELLS ON HYDROGEN IN ZIRCONIUM 4 7
[63] Simpson, C. J. and Ells, C. E., "Delayed Hydrogen Embrittlement in Zr-2.5wt%Nb." Jotwnal of
Nuclear Materials, Vol. 52, 1974, pp. 289-295.
[64] Nuttall, K., McCooeye, D. E, Rogowski, A. J., and Havelock, E, "Metallographic Observations of
the Interaction of Hydride, Stress and Crack Growth at 600K in a Zr-2.5Nb Alloy," Scripta Me-
tallurgica, Vol. 10, 1976, pp. 979-982.
[65] Nuttall, K., "Some Aspects of Slow Crack Growth in Hydrided Zr-2.5Nb Alloys," Effect of Hy-
drogen on Behaviour of Materials, AIME, New York, 1976, pp. 441- 45l .
[66] Nuttall, K., "The Effect of Stress and Thermal Cycling on Hydride Orientation and Slow Crack
Growth in Zirconium Alloys," Paper 3C7, Proceedings of the Second bnernational Congress on
Hydrogen in Metals, 1977.
[67] Coleman, C. E., "Susceptibility of Cold-Worked Zirconium-2.5wt% Niobium Alloy to Delayed
Hydrogen Cracking," AECL-5260, 1977.
[68] Coleman, C. E. and Ambler, J. E R., "Susceptibility of Zirconium Alloys to Delayed Hydrogen
Cracking," Zirconium in the Nuclear Industo' (3rd Conference), ASTM STP 633, 1977, pp. 589-
607.
[69] Cox, B. and Coleman, C. E., "Acoustic Emission During the Tensile Testing of Zirconium Alloys,"
AECL-5831, 1977.
[70] Simpson, L. A. and Clarke, C. E, "Application of the Potential Drop Method to Measurements of
Hydrogen-Induced Sub-Critical Crack Growth in Zirconium-2.5wt%Nb," AECL-5851, 1977.
[71 ] Coleman, C. E., "Effect of Texture on Hydride Reorientation and Delayed Hydride Cracking in
Cold-worked Zr-2.5Nb," Zirconium in the Nuclear Industry (5th Symposium), ASTM STP 754,
1982, pp. 293-411.
[72] Ells, C. E., Cheadle, B. A., Coleman, C. E., and van der Kuur, J. H., "'The Decomposition of
Hydrides in Zirconium Alloys," Proceedings of the Third International Congress on Hydrogen and
Materials, 1982, pp. 137-142.
[73] Dutton, R. and Puls, M. E, "'A Theoretical Model for Hydrogen Induced Sub-Critical Crack
Growth," Effects of Hydrogen on Behaviour of Materials, AIME, 1976, pp. 345-360.
[74] Lee-Whiting, G. E., "'Calculation of Hydrogen Diffusion toward a Crack in a Stressed Solid,"
AECL-5612, 1976.
[75] Dutton, R.. Nuttall, K., Puls, M. P., and Simpson, L. A. in Proceedings of the Second International
Congress on Hydrogen In Metals, Paper 3C6, 1977.
[76] Simpson, L. A. and Puls, M. E, "The Effects of Stress, Temperature and Hydrogen Content on
Hydride-Induced Crack Growth in Zr-2.5Pct.Nb,'" Metallurgical Transactions, Vol. 10A, 1979, pp.
1093-1105.
[77] Puls, M. E, "Effects of Crack Tip Stress and Hydride-Matrix Interaction Stresses on Delayed
Hydride Cracking," Metallurgical Transactions, Vol. 21A, 1990, pp. 2905-2917.
[78] Ambler, J. E R., "'Effect of Direction of Approach to Temperature on the Delayed Hydrogen
Cracking Behaviour of Cold-Worked Zr-2.5Nb," Zirconium in the Nuclear Industry (6th Interna-
tional Symposium), ASTM STP 824, 1984, pp. 653-674.
[79] Shi, S.-Q. and Puls, M. E, "Advances in the Theory of Delayed Hydride Cracking," Hydrogen
Effects in Materials, TMS, 1944, pp. 611-621.
[80] Cheadle, B. A. and Coleman, C. E., "Del ayed Hydride Cracking in a Strong, Creep Resistant
Zirconium Alloy," Proceedings, JIMIS-2, 1987, pp. 517-520.
[81 ] Sagat, S., Coleman, C. E., Griffiths, M., and Wilkins, B. J. S., "' The Effect of Fluence and Irra-
diation Temperature on Delayed Hydride Cracking in Zr-2.5Nb," ASTM STP 1245, 1955, pp. 35-
61.
[82] Cheadle, B. A. and Ells. C. E., "Crack Initiation in Cold-Worked Zr-2.5 wt%Nb by Delayed
Hydrogen Cracking," Proceedings, 2nd International Congress on Hydrogen in Metals, Paper 3C8,
Paris, 1977.
[83] Simpson, C. J., "Nucleation of Pressure Tube Rolled Joint Cracks," Ontario Hydro Report 76-480-
H, 1976.
[84] Puls, M. R, "'Fracture Initiation at Hydrides in Zirconium," Metallurgical Transactions, Vol. 22A,
1991, pp. 2327-2337.
[85] Shi, S.-Q. and Puls, M. P., "Dependence of the Threshold Stress Intensity Factor on Hydrogen
Concentration During Hydride Cracking in Zirconium Alloys," Journal of Nuclear Materials, Vol.
218, 1994, pp. 30-36.
[86] Choubey, R. and Puls, M. P., "Crack Initiation at Long Radial Hydrides in Zr-2.5Nb Pressure
Tubes at Elevated Temperatures," Metallurgical Transactions, Vol. 25A, 1994, pp. 993-1004.

[87] Shi, S.-Q. and Puls, M. E, "Crack Initiation Criterion at Notches in Zr-2.5Nb Alloys," Materials
Science Engineering, Vol. A176, 1994, pp. 237-247.
[88] Shi, S.-Q., Shek, G. K., and Puls, M. E, "Hydrogen Concentration Limit and Critical Temperatures
for Delayed Hydride Cracking in Zirconium Alloys," Journal of Nuclear Materials, Vol. 208, 1995,
pp. 189-201.
[89] Shi, S.-Q. and Puls, M. P., "Criteria for Fracture Initiation at Hydrides in Zirconium Alloys I.
Sharp Crack Tip," Journal of Nuclear Materials, Vol. 208, 1994, pp. 232-242.
[90] Shi, S.-Q., Puls, M. E. and Sagat, S., "Criteria of Fracture Initiation at Hydrides in Zirconium
Alloys II. Shallow Notches," Journal of Nuclear Materials, Vol. 208, 1994, pp. 243-250.


Imp

Diunggah oleh

Informasi Dokumen

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

Imp

Diunggah oleh

Hak Cipta:

Format Tersedia

A. Sawatzky 1 and C. E.

Anda mungkin juga menyukai