FORM TI-12

WAPD-T-3173

8/94

THE CORROSION OF ZIRCALOY-CLAD FUEL ASSEMBLIES

IN A GEOLOGIC REPOSITORY ENVIRONMENT

E. Hillner
D. G. Franklin
J. D. Smee

Contract No. DE-AC11-93PN38195

NOTE
This document is an interim memorandum prepared primarily for internal referenc.e and does not
represent a final expression of the opinion of Westinghouse. When this memorandum is distributed
externally, it is with the express understanding that Westinghouse makes no representation as to
completeness, accuracy, or usability of information contained therein.

BETTIS ATOMIC POWER LABORATORY

WEST MIFFLIN, PENNSYLVANIA 15122-0079

Operated for the U.S. Department of Energy

by WESTINGHOUSE ELECTRIC CORPORATION

WAPD-T-3173

THE CORROSION OF ZIRCALOY-CLAD FUEL ASSEMBLIES

IN A GEOLOGIC REPOSITORY ENVIRONMENT

E. Hillner, D.G. Franklin, and J.D. Smee

ABSTRACT
Recently analyzed long-term Zircaloy autoclave corrosion data were used to develop new
Zircaloy corrosion correlations. Calculations were performed to estimate the additional quantity of
corrosion that may be experienced by spent Zircaloy-c1ad fuel assemblies from long-term
exposure to geologic repository conditions. These estimates indicate that the level of general
corrosion is very small and will not be expected to affect the integrity of nonfailed fuel elements
with Zircaloy-4 cladding through the life of a repository.
Twenty-two different autoclave tests were analyzed. The tests included specimens from 46
different heats. of both Zircaloy-2 and Zircaloy-4. The material conditions included different heat
treatments and various prefilms. Maximum exposure time was 10,507 days (-29 years) in a
600F (316C) test and maximum weight gain was 1.665 mg/dm 2 (-4.5 mils of oxide film) for a
640F (338C) experiment. The new kinetic data, with various conservative assumptions. were
used in conjunction with a projected fuel temperature profile during disposal to generate a bestestimate prediction of -0.3 mils of additional oxide film growth after one million years exposure in
the repository. This additional oxide film growth should not pose a threat to the integrity of the
Zircaloy cladding. A review of the fundamental factors affecting corrosion indicates that the
extrapolation from the test database to repository times is reasonable, in part because the
database covers the thickness of oxides expected for repository disposal. It also is shown that
the ionic constituents of typical ground water, in the anticipated concentrations and pH levels, will
have no discernible detrimental effect on the corrosion behavior of Zirc.aloy in the repository
environment.

BACKGROUND
The safe repository disposal of spent fuel elements is important for the nuclear power industry
because it closes the nuclear fuel cycle. Currently, expended fuel assemblies are being stored,
temporarily, in water pools or in above-ground dry facilities at the various power plants. The US
Department of Energy (DOE) has been given the task of establishing long-term disposal
capability for spent fuel, which will allow utilities to continue operations and close plants at the
scheduled end of life.
.
Expended cores from naval nuclear reactors are stored in water pools, with dry storage capability
being planned. Disposal will be in the same national repository being developed for commercial
fuel. Therefore. the Naval Reactors program is performing testing and analyses to demonstrate
that naval expended cores can be disposed of safely in the national geologic repository. The
primary function of the repository is to protect the environment by delaying and controlling the
release of radioactive material into the environment. Although the disposal canisters and the
ground provide the barriers to the release of materials, Zircaloy fuel cladding provides a
significant additional barrier. This report evaluates whether the integrity of the cladding surrounding spent fuel elements to be interred in the repository will be compromised during the prolonged
1

WAPD-T-3173

exposure to the repository surroundings via a corrosion-related mechanism. Cladding integrity is

important since the cladding is the final barrier between the irradiated fuel and the repository
environment, regardless of the final containment design. Of all properties relevant to the entire
wastage package Le., both the spent fuel assemblies and the repository containment materials,
the most important is corrosion. The anticipated cladding corrosion in the repository could also
be factored into the analysis of potential criticality in the repository and the choice of materials for
the outer containment barriers.
To date, the fuel-element cladding of choice for most nuclear power plants has been one of two
Zircaloys, which are alloys of zirconium with small amounts of tin, iron, and chromium as alloying
additions. Zircaloy-4 is used in both pressurized water reactors (PWR) and boiling water reactors
(BWR). Zircaloy-2, which is used extensively in BWRs, also has small amounts of nickel as an
alloying addition. The two properties of zirconium alloys that largely influenced their selection for
cladding are their excellent high-temperature water-corrosion resistance and very low thermal
neutron absorption cross-section. Since these two Zircaloys exhibit similar corrosion behavior at
repository conditions, only the generic term "Zircaloy" will be employed in this report. However,
all discussions and conclusions in this document apply to both Zircaloy-2 and Zircaloy-4.
AQUEOUS CORROSION OF ZIRCALOY
Introduction
The database most appropriate to establish a basis for Zircaloy corrosion behavior in a repository
environment is isothermal aqueous autoclave corrosion data. Under present considerations it is
anticipated that when the final barrier in the repository is breached, and ground water can come
in contact with the Zircaloy-c1ad spent fuel elements, the vast majority of additional repository
exposure would take place at temperatures in the neighborhood of -200F (-93C) and below.
The preponderance of Zircaloy autoclave data have been generated in the temperature range
480F to 680F (250C to 360C) in degassed, deionized water and with relatively short
exposure times for the lower exposure temperatures. The excellent corrosion resistance of
Zircaloy makes it impractical to conduct meaningful testing below -450F (230C) in any
reasonable time period. Thus, it will be necessary to employ corrosion models based on data
generated at the higher temperatures and extrapolations have to be made to the lower
temperatures anticipated in the repository.
Model Development
Zircaloy reacts with water to form a corrosion film of Zr0 2 , by the following reaction:

[1]
Since virtually all of the oxygen generated in Equation [1] reacts with the Zircaloy to form the
corrosion film and the film remains adherent, the weight gain of the corrosion specimens has
been used as a direct gauge of the oxide film thickness1. The Zircaloy corrosion processes are
known to occur in three stages:

1378 mg/dm 2 of weight gain

=1 mil of oxide growth =0.66 mils of metal consumed.

2

WAPD-T-3173

1. The early pretransition regime, characterized by the formation of a thin, black, tightly
adherent corrosion film that grows thicker in accordance with a cubic rate law,
2. The midlife transition, or transitory stage, that lies between the pretransition and
posttransition stages. As initially shown by Bryner [REFERENCE 1], this region appears to
be comprised of a series of successive cubic curves, similar to the initial cubic kinetic
curve, but initiating at shorter and shorter intervals, and
3. The linear posttransition kinetic regime.
The three-stage Zircaloy corrosion behavior is shown schematically in Figure 1. The dashed lines
in Figure 1 indicate that most early corrosion models recognized only the pretransition and
posttransition kinetic regimes. Since some spent naval fuel cladding will be in the posttransition
region upon entry into the repository, Le., oxide film thicknesses in excess of 1 mil, this document
will be concerned only with posttransition corrosion behavior. The use of these kinetics will be
conservative for cladding that may not have attained the posttransition region prior to disposal.
Posttransition corrosion kinetics can be described by an expression of the form:
[2]

where

= specimen weight gain, in units of mg/dm

= exposure time, in units of days,
= empirical constant, usually termed the linear (or posttransition) rate constant, in

llW
t
I<t.
C

units of mg/dm2/day, and

another constant, in the same units as llW (mg/dm 2), which is the intercept of the
linear equation at zero time.

From Equation [2], a plot of the weight gain as a function of the exposure time produces a
straight line with a slope equal to I<t. and an intercept equal to the constant, C. The temperature
dependence of the linear rate constant (I<t.) has been shown to follow an Arrhenius-type behavior
of the form:
[3]

where:

B
QL =
R
T

=
=

an empirical constant, in units of mg/dm2/day,

the activation energy for the posttransition (linear) corrosion region, in units of cal/mol,

universal gas constant, 1.98 cal/mol-oK, and

absolute temperature, in units of oK.

Thus, from Equation [3], a plot of the natural logarithm of the linear rate constant [Ln(I<t.)] against
the reciprocal of the absolute temperature generates a straight line with a slope equal to -QL/R
and an intercept on the Y axis equal to the natural logarithm of the constant, B. Combining
Equations [2] and [3] results in the following general expression for the posttransition corrosion of
Zircaloy as a function of both time and temperature:

WAPD-T-3173

l::.W =(B exp[-QL/RT] x t) + C.

[4]

Many isothermal autoclave corrosion studies of the Zircaloys have been conducted, primarily in
the temperature range of 550F to 750F (-290C to 400C) [REFERENCES 2 - 19]. In 1976
Hillner [REFERENCE 11] compiled the data from REFERENCES 2 - 10 into engineering curves
for each exposure temperature and, generating an activation energy for the posttransition rate
constant, ~, produced the following empirical expression for posttransition behavior:
l::.W = 1.12 X 108 exp[-12,529/T] x t

[5]

where all of the terms have been defined previously. During analyses of the Zircaloy corrosion
data, the constant C was found to be small and not included in Equation [5]. (Note that for ease
of calculation the parameter QL/R has been combined into a single value in the above equation).
Several additional Zircaloy corrosion models have been published in the technical literature
[REFERENCES 12 - 19]. Most models follow the form of Equation [4] for posttransition b.ehavior
and those that do not can be converted easily to this format. For the current application it is of
interest to compare predictions from the various models for prolonged exposure to a lowtemperature environment. Rothman [REFERENCE 20] has proposed an assumption of a
constant temperature of 356F (180C) for 10,000 years as a possible spent fuel disposal
scenario for a compacted-volcanic-ash (tuff) repository. Using Rothman's exposure conditions,
calculations were conducted with eight different corrosion correlations to determine the
anticipated extent of additional Zircaloy corrosion, assuming that the fuel cladding was in the
posttransition kinetic regime upon disposal. Table 1 shows the results of such calculations. The
prediction for the REFERENCE 11 model was based on the use of Equation [5], above.
All of the numbers in Table 1 have been rounded. The predicted extent of corrosion for all of the
models shown in Table 1 reflect the excellent corrosion resistance of Zircaloy, for these assumed
~
conditions for tuff repository disposal.
Comparisons with Equation [5] have been conducted in many investigations [REFERENCES 1518, 20-30]. Almost all of these early models suffer from one major weakness when repository
predictions are concerned; in general, there is a paucity of data for extended exposures in the
posttransition region, especially at low temperatures. The Bettis Atomic Power Laboratory has
generated a considerable quantity of autoclave corrosion data since the publication of REFERENCE 11. The sections below present these new data, discuss the generation of new
posttransition corrosion equations, and assess the impact of these equations on the long-term
repository oxide-film growth predictions.
New Corrosion Test Data
1. Test Description
The tests were conducted in static isothermal autoclaves at temperatures of 520F to 680F
(271C to 360C). Specimen dimensions were typically 1 inch (25 mm) by 1 inch by -0.060 inch
(1.5 mm) to 0.100 inch (2.5 mm) thick. Periodic weight measurements were obtained during
exposure in twenty-two different autoclave tests. These specimens represented a total of 46

WAPD-T-3173

Zircaloy-2 and Zircaloy-4 heats with different heat treatments and with various prefilms. The
number of specimens in each test decreased with increasing exposure since coupons were
removed periodically for destructive analyses. For this repository disposal application, long-term
exposure data are of the most interest. Therefore, this analysis concentrated only on the posttransition kinetic region. It was assumed that weight gains of approximately 45 mg/dm 2 and
above were in posttransition2 and the weight changes were tabulated from this point on.
Maximum exposu~e time for any specimen was 10,507 days (-29 years) in a 600F (316C) test
(test #5) and the maximum weight gain was 1,665 mg/dm 2 (-4.5 mils or - 114 IJm of oxide) in a
640F (338C) experiment (test #12).
2. Test Data
Table 2 provides a summary of the material and test conditions employed in these tests.
For groups of coupons that are nominally identical and tested together in the same autoclave in
each test, specimen weight gains were averaged for each exposure time and the average weight
(mg/dm 2) plotted as a function of total exposure time (days). This averaging was done to prevent
domination of the analysis by times or temperatures at which many data were collected.
Appendix A describes how the SOLO statistical software package was used to handle the raw
data and to generate the algorithms that show the best fit with the weight gain versus time data.
Figure 2A shows a typical data plot at 680F (360C) from test #21. The solid line in this plot is
the least-squares straight line drawn through all of the posttransition data. The curvature of the
data about this line is readily apparent and cannot be ignored. Thus, the posttransition kinetic
region can no longer best be described by a single posttransition rate constant, like that shown in
Equations [2] and [3]. Using the SOLO statistical program, it was found that a series of two
successive linear equations, each active over a different time range, best describe the data.
Figure 28 is a replot of the data of Figure 2A showing the two linear rate constants. For
purposes of discussion the early linear portion has been designated as stage 1 and the later
linear behavior has been called stage 2. The software package automatically calculates both the
time and weight gain at the transition point between the two linear stages. The SOLO program
was then employed to calculate the stage 1 and stage 2 linear rate constants (and associated
data) for all twenty-two tests in this study, for the temperature range 520F to 680F (271C to
360C).
3. Summary of Test Data
Table 3 summarizes all of the corrosion data obtained from the twenty-two tests. The secondstage rate constants [K(2)] for the temperatures 520F and 550F are included in this table for
completeness. However, since corrosion thicknesses at these low temperatures were too thin to
have reached stage-two kinetics, the K(2) data for these two lowest temperatures have not been
included in stage-two analyses. At these low temperatures the constants calculated by the
statistical program for stage-two' posttransition corrosion kinetics were not significantly different
from their stage-one counterparts, which supports the conclusion that stage-two kinetics had not'
been attained.

2As can be seen in Appendix A, this choice of the initiation of posttransition kinetics had very
little impact on the statistical evaluation of the posttransition data.

WAPD-T-3173

The occurrence of a second posttransition corrosion rate raises the concern that additional
corrosion could result in a subsequent higher rate. Although the extrapolation to low
temperatures and long times results in predictions outside the database for time, the corrosion
rate is not directly determined by exposure time. Rather, the corrosion rate is determined by the
physical state of the Zircaloy metal and oxide. As discussed in a later section on corrosion
mechanisms, this material state is closely correlated with the amount of corrosion that has
occurred. The sole exception is the memory effect that decays quickly and this decay is
beneficial rather than detrimental. As will be seen below, the autoclave database presented
herein covers corrosion thicknesses up to 4.5 mils, which is greater than the corrosion thickness
expected in a repository through one million years.
As in the earlier studies, both stage-one and stage-two linear rate constants appear to follow an
Arrhenius-type temperature dependence. Figure 3 shows the natural logarithm of both linear rate
constants plotted as a function of the inverse of the absolute temperature. From this figure and
the detailed analyses of this plot shown by the data in Appendix A, Table A,2, the following
parameters were derived for the stage-one posttransition region: Q/R -12877 and the preexponential constant (8) 2.46 x 108 . The corresponding values for the stage-two kinetic region
are -11452 and 3.47 x 107 , respectively. Thus, the incremental change in weight obtained by two
successive exposure times entirely within the stage-one linear regime [MW(1)] is given by:

MW(1)

=K(1) x flt = 2.46 x 10

exp[-12877/T] x flt,

[6]

and the corresponding weight change for the second-stage linear corrosion is

MW(2)

=K(2) x flt =3.47 x 107 exp[-11452/T] x flt

[7]

where flt is the change in time (days) between two successive exposures times entirely within
the stage-one or stage-two kinetic regimes and MW is the weight gain (mg/dm 2) generated between these two exposure times. Since K(2) is always equal to or slightly greater than K(1), and
it is assumed that the Zircaloy cladding on the expended fuel elements will be in the
posttransition kinetic region upon disposal, Equation [7] will be used to estimate the extent of
additional corrosion due to geologic repository exposure. Thus, the use of this equation for all
cladding is conservative.
Equations [6] and [7] are known to be valid within the database used for their generation, which
is up to approximately 4.5 mils of oxide film. As shown below, extrapolation beyond the limits of
the high-temperature database is not needed to estimate the corrosion of Zircaloy-clad naval fuel
after disposal in a repository for one million years, which eliminates concerns that there could be
changes in the corrosion kinetics for thicker oxides.
4. Application to Repository Conditions
Prior to applying Equation [7] to the Yucca Mountain Repository environment, it is of interest to compare
Equations [6] and [7] with the eight earlier models used to estimate the extent of corrosion for the tuff
repository conditions. Table 4 is a copy of Table 1 with two additional rows for the results of calculations
with Equations [6] and [7].

WAPD-T.;.3173

For the time and temperature shown in Table 4, Equation [6] predicts an additional oxide film
growth of -1.1 mils (which is in line with the other values in this table) whereas Equation [7]
produces an oxide film thickness growth of -3.5 mils. These calculations demonstrate that the
use of Equation [7] is more conservative than any of the eight models described above or than
the use of Equation [6] for this application.
Figure 4A shows what, at present, is considered to be a conservative temperature profile for
naval fuel disposal in the Yucca Mountain Repository facility. Currently, it is anticipated that the
inner container could be breached, and the fuel cladding exposed to an aqueous environment, in
the period -1000 to 5000 years after depositing the fuel assembly in the Yucca Mountain
Repository. It is assumed that all spent fuel cladding will be in the stage-two linear kinetic regime
upon disposal, which is an upper bound on the corrosion rate. As an additional conservative
factor, the corrosion calculations shown in Figure 4B initiate at 1000 years after disposal and
continue for a total exposure of one million years. An additional oxide film thickness of-0.14 mils
is projected for this history.
There have been several additional algorithms reported inthe literature for the oxidation of
Zicaloy at elevated temperatures. For example, Boase and Vandergraaf [REFERENCE 59] show
an Arrenhius-type plot of rate constants compiled from various documents (some unpublished)
with wet and dry air and steam in the temperature range -298C to 725C (-570F to 1340F).
Similarly, Suzuki and Kawasaki [REFERENCE 60] report on the oxidation of Zircaloy in moist air
in the temperature range 350C to 500C (662F to 932F). Einziger [REFERENCE 61]
generated an equation for Zircaloy oxidation. based upon the work of the previous two references.
The results of these studies were not included in the comparisons of Table 4 because of
differences in test conditions, i.e., temperature or environment. Inclusion of these reported
corrosion rates would not have altered the results of this study, despite the differences in test
conditions. Van Swam and Shann [REFERENCE 62] employed the MATPRO corrosion model
[REFERENCE 63] to compute out-of-reactor comparison values for their in-reactor
measurements. MATPRO calculations conducted for our anticipated repository conditions
generate estimates that fall in the middle of the range of values shown in Table 4. For all
conditions discussed to date, including the information presented above in references 59-63,
Equation [7] of this report produces the most conservative estimate of additional oxide buildup on
spent Zircaloy-clad fuel assemblies due to long-term exposure to expected geologic repository
conditions. These maximum corrosion rates are due to the underlying database that includes the
thickest corrosion films reported to date (up to 4.5 mils).
There is one additional degree of conservatism implicit in the above computations.. All
calculations have been conducted with the assumption that the environment is water. However,
examination of Figure 4A indicates that after breaching (assumed at 1000 years) a considerable
period of time is spent above 100C. Long-term autoclave data have shown that corrosion of
Zircaloy in the steam phase is not as aggressive as corrosion in water at the same temperature
(see Figure 5). Thus, the use of water autoclave data for the entire exposure time adds another
degree of conservatism.
5. Postirradiation Corrosion in Autoclaves
The predictions for the additional corrosion on the Zircaloy cladding during repository disposal
presented above were based on isothermal autoclave testing of nonirradiated specimens. The
corrosion of cladding that has experienced appreciable irradiation exposure prior to disposal in
the repository may be accelerated, especially for the posttransition kinetic regime. At the present
time it is uncertain how long this accelerated in-reactor corrosion rate will persist when
7

WAPD-T-3173

transferred to an ex-reactor corrosion environment, but sufficient data are available to bound the
effect.
Garzarolli et al. [REFERENCE 16] conducted an experiment to determine how rapidly the
enhanced corrosion rate observed in-reactor would decrease to that anticipated from ex-reactor
(autoclave) data. The authors found that samples of cladding from a BWR that were irradiated at
an estimated temperature of 554F (290C) continued to corrode at the irradiated rate for the
first ten days in an autoclave at 536F (280C). In the time frame from 10 to 110 days the rate
continued to decrease until, at 110 days, it became close to that measured in autoclave testing of
nonirradiated specimens. When exposed to higher temperatures in autoclaves (572-662F; 300350C) the specimens were found to attain the ex-reactor rate quickly. At the highest autoclave
temperature there appeared to be no "memory" effect of the prior irradiation exposure, Le., the
measured corrosion rate agreed very favorably with the ex-reactor rate from the onset of the
high-temperature postirradiation autoclave exposure.
Somewhat different results were obtained with specimens from cladding taken from PWRs. In
one batch of irradiated specimens tested in autoclaves at temperatures up to 617F (325C), the
corrosion rate measured after the first cycle of exposure in the autoclave was a factor of 20
greater than the nonirradiated corrosion rate but decreased rapidly to a factor of 2 or less during
subsequent autoclave exposures. The test was terminated after 180 days in the autoclave. It is
not known whether additional long-term autoclave exposure would have resulted in the
convergence of the postirradiated and nonirradiated corrosion rates, as observed with the BWR
specimens for shorter exposure times. At a higher autoclave temperature (662F; 350C) the
corrosion rate for the PWR specimens agreed very well with the expected nonirradiated corrosion
rate. In all cases, the effect of irradiation was reduced to a factor of two or less on or before 110
days of postirradiation exposure at 280C. The technical basis for this memory effect and the
factor of two for postirradiation corrosion are discussed in more detail in a later section.
Cheng et al. [REFERENCE 58] also investigated postirradiation corrosion performance of
Zircaloys. However, they removed the in-reactor oxide film and pickled the surface on almost all
of their samples before postirradiation autoclave exposure. The oxide was removed because
they were concentrating on determining the effect of irradiation on corrosion rates through
irradiation-induced dissolution of second-phase particles in the metal. With the exception of two
of the weld metal samples, all samples postirradiation tested at 316C (600F) corroded at
essentially the same rate as nonirradiated samples, despite the irradiation-induced changes in
microstructure and dissolution of alloying elements into the zirconium matrix, which Cheng et al.
confirmed. Two of the weld metal samples postirradiation tested at 316C corroded at higher
rates than the other samples. This was attributed to the strong susceptibility of the small weldmetal precipitates to an irradiation-induced dissolution. There was no explanation as to why
some weld metal samples corroded at an accelerated rate and some at the corrosion rate of
nonirradiated material. However, all base metal samples corroded at the nonirradiated rate, not
showing the memory effect reported by Garzarolli. This would suggest that the memory effect at
316C (600F) is associated with the effect of irradiation on the oxide films rather than on the
base metal. This interpretation is somewhat different from that given by Cheng et al. The
modified interpretation is necessary to reconcile the posttransition corrosion observations of both
Garzarolli and of Cheng for postirradiation testing with and without the in-reactor oxide removed.
Cheng also postirradiation- tested samples at 400C (752F) and observed accelerated
corrosion. For samples with the oxide removed, the corrosion rate was initially high but
decreased with postirradiation exposure. This decrease in rate was associated with precipitation
of alloying elements during postirradiation testing, which occurred at 400C but not at 316C.
One sample postirradiation tested at 400C with the in-reactor corrosion film left on behaved

WAPD-T-3173

differently from the samples postirradiation tested at 400C with the in-reactor corrosion films
removed. The sample with the in-reactor corrosion film left on initially corroded at a slow rate
that was more like the in-reactor rate than the 400C rate, suggesting that the character of the
oxide initially controlled the rate in this case. After some time, the corrosion rate for the sample
with the in-reactor oxide left on increased to a similar rate as the postirradiation rate of the
samples with the in-reactor oxides removed.
The testing of Garzarolli et al. and Cheng et al. suggest that there are two postirradiation timedependent effects of in-reactor irradiation on postirradiation corrosion rates. The first effect is on
the oxide. This effect is a postirradiation corrosion-rate acceleration that decays relatively
rapidly, at least rapidly compared to repository times, to less than a factor of two of the
nonirradiated posttransition corrosion rate at 316 and lower. The second effect is on the metal.
At relatively high temperatures the alloying elemerits that became supersaturated by the
irradiation-induced dissolution begin to precipitate in a unique microstructure that increases the
corrosion rate at 400C. This unique microstructure and the associated acceleration in corrosion
rate do not occur at 316C. The temperature dependence on postirradiation microstructural
changes at high temperatures has been observed by other investigators. For repository
conditions (Figure 4A) the temperatures are never high enough to induce such microstructural
changes. Therefore, the testing at 316C and below by both Garzarolli et al. and by Cheng et aI.,
despite Cheng's removal of the oxide, support the assumption that the postirradiation repository
corrosion rates will be within a factor cif two of the nonirradiated posttransition corrosion rates
reported and modeled above.
These postirradiation corrosion test results suggest that it is possible for PWR Zircaloy cladding
in a repository to experience an enhanced corrosion rate over that predicted from autoclave test
data. A conservative treatment is to assume that the corrosion rate at the low repository
temperatures will not converge to the preirradiation rate, even after one million years in the
repository. This can be reflected by applying a factor of two to the corrosion rate. Using this
conservative factor of two increase for the entire exposure time, the final additional increase in
the oxide film thickness of Zircaloy cladding interred in the Yucca Mountain Repository is
anticipated to be twice that estimated by Equation [7], for a total of 0.28 (or-0.3) mils. This small
increment in oxide film thickness is insignificant compared to the Zircaloy cladding thickness.
EFFECT OF CHEMISTRY
The autoclave corrosion data discussed above have been generated in water chemistries where
every attempt has been made to eliminate, or at least control to a very low level, all. sources of
contaminants in the test e,nvironment. In contrast, the ground water in the Yucca Mountain
repository will include an unspecified and variable chemistry. To generate a plausible
composition for the ground water anticipated at the repository, a deep well was drilled in the
vicinity of Yucca Mountain (henceforth designated as the J-13 well) and the water analyzed for
ionic species. The analysis results for the J-13 well water are shown in Table 5. The chemical
composition of the J-13 well water is taken as representative of the ground water environment in
the Yucca Mountain Repository.
It appears that, in general, the concentrations of the ionic species in J-13 well water are too low
to have an adverse effect on the Zircaloy corrosion behavior. The following is a brief discussion
of the known effects of the more-aggressive ionic species in the J-13 well water:

WAPD-T-3173

1. Fluoride (F-)
The fluoride ion is probably the most aggressive of the ionic species; in sufficient quantity it will
prevent the formation of the protective oxide and lead to a rapid attack of Zircaloy. Berry
[REFERENCE 31] has shown that at concentrations in excess of 100 ppm the fluoride ion will
degrade the normally excellent corrosion resistance of Zircaloy but concentrations below this
value will not influence corrosion performance. If maintained at this level, the 2 ppm of F' found in
J-13 well water should have no effect on the Zircaloy corrosion behavior.
2. Chloride (CI-)
The combination of Zircaloy and chloride ions results in the possibility of local attack by pitting
and stress corrosion cracking. The effect of the chloride can be magnified by the presence of
ferric (Fe+ 3) or cupric (Cu+2) ions [REFERENCES 32,33]. However, there is no known readilyavailable source of copper whereby cupric ions could contaminate the ground water in the
repository. The exception would be if the repository containers had a high copper concentration.
There is a small amount of copper in the crud on fuel elements, but this has not accelerated
corrosion of fuel cladding during service.
There could be sources of iron in the repository from the disposal containers and repository
construction. It has been shown that room temperature additions of 100 ppm ferric ions to a
solution of concentrated acidic chloride will increase the corrosion rate of a commercial
zirconium-base alloy by a factor of 25 and cause a pitting attack [Reference 34]. However,
Maguire [REFERENCE 64] has shown that pitting of zirconium in acidic ferric chloride solutions
will not occur above a pH of 3 and that this behavior is consistent with the Pourbaix stability
diagram for iron that shows that the ferric ion ( Fe+ 3 ) is not stable above a pH of 3. Thus, if the
pH of ground water remains above 3 there is little chance of generating a ferric-chloride-induced
pitting attack of the cladding. J-13 well water has a pH of 7.4.
3. Lithium (Ln
In neutral or acid solutions lithium has no detrimental effect on the corrosion resistance of
Zircaloy. In basic solution lithium can form lithium hydroxide (LiOH) which, if in sufficiently
concentrated solutions, can have an adverse effect on Zircaloy corrosion performance. However
several investigations have shown that a lithium hydroxide concentration of greater than 1000
ppm is required to cause accelerated corrosion [REFERENCES 35-37]. The concentration of
lithium ion in J-13 water is only 0.05 ppm, which is well below those levels that could cause
concern.

Cox [REFERENCE 38] reports that the corrosion of the Zircaloys at elevated temperatures is
insensitive to the hydrogen ion concentration (pH) in the pH range of 2 to 12. The J-13 water has
a measured pH of 7.4, well within these bounds. As indicated above, lithium hydroxide in
sufficiently high concentrations (>pH 12) can greatly accelerate the Zircaloy corrosion kinetics
and other strong alkalis (e.g., sodium, potassium and calcium) can have similar, but smaller,
effects at these highly basic pH values [REFERENCES 37,39]. These conditions are not credible
because pH levels in the J-13 water would have to concentrate by a factor of 105 (100,000
times), which is not a very likely event.

10

WAPD-T-3173

The Pourbaix diagram calculations of REFERENCE 40 suggest that the base metal, zirconium,
has an excellent corrosion resistance to pure water at temperatures in the region of 25C (77F)
over a wide range of pH and electrode potential values. Figure 2 on page 227 of REFERENCE
40 shows the various stability and corrosion domains generated for zirconium at 25C (77F).
The region marked as "immunity" in this diagram indicates the area wherein the metal does not
react with water; it is completely noble in this domain. In the region titled "passivation" the metal
becomes coated with an oxide (Zr02) that virtually prevents all direct contact between the metal
and the solution. The exposed cladding surfaces will be in this region. This diagram shows the
extensive range of pH and electrochemical potential where the zirconium passivated surface
should not experience a corrosion attack. Due to the fact that the Zircaloys are dilute alloys of
zirconium, and that the passivation product (Zr02) is virtually the same on all dilute zirconium
alloys, this diagram can be employed to represent Zircaloy behavior, as well.
5. Other Ions
By virtue of their low concentrations, the other ions listed in Table 5 for J-13 water also are not
anticipated to degrade the Zircaloy corrosion performance. In fact, at concentrations well above
those in the J-13 water both the nitrate (N03-) and sulfate (804=) ions may be beneficial in that
they can inhibit pitting in the presence of chloride ion.
The discussions in this section suggest that the ionic constituents in the J-13 well water, in the
measured concentrations and pH levels, would have no detrimental effect on the Zircaloy
corrosion behavior in the Yucca Mountain repository, relative to that observed in autoclave
testing. In addition, most of the testing discussed in this section was conducted with. bare (Le.,
unfilmed) zirconium or zirconium-alloy specimens whereas the spent fuel elements should have a
protective passive corrosion film upon disposal. This different surface can have a pronounced
affect on the chemical interactions discussed herein.
MICROBIOLOGICALLY INFLUENCED CORROSION
The term microbiologically influenced corrosion (MIC) is commonly used to designate that type of
corrosion attack associated with microbial activity at metal surfaces. As indicated by Little and
Wagner [REFERENCE 41], MIC is a form of localized corrosion that can result in pitting,
selective leaching of alloys, crevice corrosion, underdeposit corrosion, and enhanced erosion
corrosion. Little and Wagner divided the reactions between biofilms and metal surfaces into the
following general categories: sulfide, acid, hydrogen and ammonia production, and metal
deposition. The two major forms of MIC for materials being considered for containment vessels
in mined geological repositories are sulfide attack through the action of sulfate-reducing bacteria
(8RB) and organic acid production by secretion of certain bacteria.
McNeil and Odom [REFERENCE 42] have indicated, by thermodynamic calculations, that
zirconium would not be affected by 8RB. Considering the tolerance that Zircaloy has for a wide
range of pH values, it is unlikely that production of weak organic acids will have an adverse effect
on the passivation of Zircaloy by the Zr0 2 film.

11

WAPD-T-3173

PERSPECTIVE ON EXTRAPOLATING CORROSION RATES TO REPOSITORY CONDITIONS

Introduction

The analysis of the data presented above uses a standard thermal dependency to extrapolate to
the low temperatures and long times associated with repository corrosion. The uncertainty in this
thermal dependency, that is, the uncertainty in the thermal activation energy, and its impact are
calculated by standard statistical methods. Implicit in this extrapolation is the assumption that
the rate-controlling process remains unchanged between the test-temperature range and the
application-temperature range. To better understand the impact of the extrapolation, physical
aspects of the corrosion process are reviewed.
The corrosion rate of Zircaloy in water is limited by the slowest of the basic corrosion steps:
dissociation of water into oxygen and hydrogen ions, diffusion of oxygen ions through the oxide
film, oxidation of the Zr metal, diffusion of electrons through the oxide, and hydrogen-ion
reduction by the electrons at the water-to-oxide interface. At temperatures for current testing,
that is, above about 450F, diffusion of oxygen ions through the oxide film is rate limiting
[REFERENCE 43]. In extrapolating to lower temperatures, one of the other basic steps could
become rate limiting, which would decrease the corrosion rate to less than predicted by the
extrapolation. In that case, projections based on oxygen-ion diffusion through the oxide would be
conservative. Therefore, analyses herein are based on the assumption that oxygen-ion diffusion
is rate limiting in the repository. That is, if one of the other steps in the currently considered
process had a high enough temperature dependency that it displaced oxygen-ion diffusion as the
rate limiting step, the extrapolation to repository temperatures used herein would be
conservative. The only risk that more corrosion will occur than predicted by the extrapolations
based on a thermal activation energy is that a new more-rapid corrosion mechanism becomes
active.
The corrosion process usually is described and modeled as a function of time. However,
changes in the corrosion rate or the mechanism occur due to physical changes that are occurring
r the environment. Therefore, it is important to understand what changes might influence the
corrosion mechanism. Two aspects of the extrapolation to low temperatures are discussed.
Both are based on considering what physical changes are occurring with corrosion and how
these changes might affect the extrapolation. First, in a repository environment, the active
mechanism controlling the corrosion rate is expected to be as closely correlated with the amount
of additional corrosion that occurs in the repository as with the time in the repository. Therefore,
understanding corrosion for the amount of additional corrosion expected in the repository is just
as important as obtaining data for the amount of time expected in the repository, which can not
be achieved. This important feature of corrosion supports extrapolation of available models for
corrosion to repository conditions as long as the database on which the models were built
extends to the corrosion thicknesses expected in a repository. Second, the temperature
dependencies of the physical changes that may occur are considered to determine if there is any
evidence that a new more-rapid corrosion mechanism could occur at repository conditions. This
review is needed to minimize the risk that extrapolation to repository conditions will not
conservatively represent the expected repository corrosion mechanism.
This reliance on thickness of corrosion rather than time is established by considering three broad
primary categories of parameters to which the corrosion rate is sensitive. As appropriate,
thermal dependencies also are considered. Early studies showed that corrosion is sensitive to
the alloy design. The most important such feature is the alloy chemistry, and the associated type
and size distribution of second-phase particles, which are controlled primarily through heat-

12

WAPD-T-3173

treatment schedules during manufacturing. Subsequent work showed that corrosion of Zircaloys
also is sensitive to the environment, both to water chemistry and to neutron irradiation. Finally,
research has shown that corrosion rate is sensitive to oxide thickness. Each of these three
categories, Alloy Design, Service Environment, and Oxide Thickness, reflects the effects of other
more fundamental parameters. The primary ones are reviewed below.
Alloy Design
Zirconium alloys in world-wide nuclear-core applications, principally the Zircaloys, are alloyed with
tin, iron, chromium, and nickel, and in some cases niobium. Except for tin, these elements are
not very soluble and form second-phase particles. The type and size distribution of these
particles are determined by alloy chemistry and processing, mainly heat treatments. Once
established during manufacture, alloy-design features are only changed by exposure to high
temperatures or to irradiation flux. During exposure in reactor cores, the temperatures are not
high enough to induce significant microstructural changes. However, the irradiation field can
induce significant changes. In particular, at low irradiation temperatures, second-phase particles
are transformed from crystalline form to amorphous form. At intermediate temperatures, iron,
chromium, and nickel are ejected and diffuse out of the otherwise thermally stable second-phase
particles [REFERENCES 44-46]. Reprecipitation may occur for irradiation at higher reactor
temperatures [REFERENCE 46]. These reactor-service changes in microstructure increase the
corrosion rate of Zircaloys, and may be the basis for the short-term postirradiation factor of two in
corrosion rate in the model described elsewhere in this paper.
Alloy design and changes during reactor exposure are fixed on entering repository service. The
primary potential impact on subsequent corrosion is the reprecipitation of the alloying elements
iron, chromium, and nickel, which the irradiation causes to be supersaturated in the zirconium
matrix. In development of zirconium alloys, it was found that corrosion rates decreased with
annealing and precipitation of these alloying elements, with the optimum amount depending on
the in-reactor service conditions. The important observation is that precipitation generally
reduces corrosion to some optimum amount of precipitation. As discussed below, repository
temperatures will be too low to reach this optimum. Therefore, repository exposure is expected
to alter the alloy microstructure in a direction to improve corrosion resistance, although perhaps
by an insignificant amount.
Repository Service Environment
The expected repository environment will not include sufficiently high temperatures nor neutron
irradiation flux to significantly alter alloy-design features, primarily the microstructural character of
the second-phase particles and the associated amounts of iron, chrome, and nickel in solution.
The most commonly used parameter to estimate the effect of thermally induced changes in
microstructure on corrosion is the LA j parameter, as defined by Garzarolli [REFERENCE 47].
The LA; parameter is a measure of the effect of time at temperature on Zircaloy microstructure
and of the associated effect on corrosion. The thermal activation energy in Garzarolli's analysis
is Q/R = 40,OOOK. Based on this value of Q/R and the repository thermal history of Figure 4A,
the integrated IA j parameter for the full repository life is expected to be approximately 1 x 10-30 h.
As shown in Figure 6, essentially all the annealing occurs during the first 50 years, long before
any water might access zirconium-alloy surfaces in a container. This LA; parameter must be
significantly greater than 10-20 h to affect corrosion. Therefore, the amount of annealing that will
occur during repository service is orders of magnitude less than required to significantly affect
thermal corrosion. To the small extent that this effect occurs for irradiated material, it will restore

13

WAPD-T-3173

the material condition toward that of nonirradiated material, which will decrease the corrosion
rate, as discussed below.
The factor of two in corrosion rate reported by Garzarolli et al. [REFERENCE 16) and used in the
above model for repository corrosion reflects the impact of irradiation-induced changes in the
material condition on postirradiation corrosion rates. There is no irradiation enhancement during
pretransition corrosion but irradiation during pretransition corrosion enhances corrosion rates
after the pretransition period. That is, the material has a memory of previous exposure through
the changes in the material condition. After a change in exposure conditions, in flux and
temperature, the material condition and the corrosion rate slowly change to reflect the new
exposure conditions. This is important for repository service because it leads to the conclusion
that the postirradiation corrosion rates will slowly decrease toward those measured in autoclaves.
More important, no increase in corrosion rate during repository service is expected.
Experimental evidence of the memory effect and its decay with time are illustrated in the results
of corrosion tests performed in the Advanced Test Reactor (ATR) on Zircaloy-4 corrosion
coupons.
Figure 7a shows the corrosion of a Zircaloy-4coupon exposed in ATR at about 520F for 2500
days, initially at high flux with several decreases in flux to one third of the initial flux at the end of
the irradiation, as shown in the irradiation history provided in Figure 7b. Figure 7a also shows
the corrosion of coupons exposed to the irradiation history of Figure 7c, a nearly constant highflux level that is comparable to the average level of Figure 7b. As the data for the coupon
corresponding the history of Figure 7b show, the irradiation flux level has no effect on the
corrosion rate during the pretransition period of approximately the first 1000 days. However,
after this pretransiton period, the corrosion rate increases beyond what would be expected for
the irradiation flux level during this time, based on isoflux tests. In fact, the rate is higher at this
time than the nearly constant high-flux data of samples irradiated with the history of Figure 7c.
Finally, after another 500 days, the corrosion rate has decreased to a rate that is now less than
the rate of the samples of Figure 7c , even though those samples have experienced about the
same average irradiation flux level. The important memory-related results of this test are that: 1)
the initial high-flux period during the first 500 days was not reflected in the corrosion rate until
about 1200 days, when the sample left the pretransition period, and 2) the decrease in flux to
below that of the flux of the other samples, just as the sample left the pretransition period at
about 1200 days, was not reflected in the corrosion rate until about 500 days later at 1700 days,
after which the corrosion rate became less than that of the other samples. This demonstrates
that decreases in neutron irradiation flux are not initially reflected in corrosion rate; but after
additional corrosion of 1pm to 2,am, corrosion rates decrease.
This memory effect also is illustrated in a coupon initially irradiated for about 100 days at a high
temperature (680F) and then irradiated at a low temperature (520F). The control samples are
the same ones as for the above illustration, being irradiated at low temperature throughout their
exposure. The result is shown in Figure 8a. During the -100 days at 680F approximately ten
microns (0.4 mils) of corrosion occurred. Immediately upon lowering temperature to 520F and
increasing flux, as shown in Figure 8b, the corrosion rate deceases to a lower rate than expected
for the existing corrosion thickness, as shown in Figure 8a, which also shows the corrosion rate
for coupons exposed isothermally at 520F. Approximately 1000 days at 520F are required
before the corrosion rate of the temperature-shifted sample becomes similar to that of
isothermally exposed samples. During exposure at 680F much of the irradiation effect on the
Zircaloy-4 is annealed relatively quickly. As a result, the corrosion rate immediately following the
temperature change to 520F initially is more representative of nonirradiated material than
material aerated at 520F. After about 1000 days at 520F, irradiation damage produces a
material representative of material irradiated isothermally at 520F and the corrosion rate also
14

WAPD-T-3173

becomes the same. A similar memory effect, out-of-pile, has been reported by Beie et al.
[REFERENCE 56].
A possible explanation for this memory effect is as follows: The corrosion rate is determined by
the diffusion of oxygen ions through the thin barrier oxide at the metal-to-oxide interface. Initially
after a change in flux or temperature, oxygen ions diffuse through oxide formed at the previous
conditions. If the existing oxide is more damaged than oxides irradiated at the existing
conditions, as is the case for the flux-shifted sample of Figure 7, a relatively fast diffusion rate
occurs. If the existing oxide is less damaged than oxides irradiated at the existing conditions, as
is the case for the temperature-shifted sample of Figure 8, a relatively slow diffusion rate occurs.
As new oxide forms, its character and resistance to diffusion of oxygen ions is determined by the
new conditions. After a change in conditions, the corrosion rate slowly changes to reflect the
. new conditions. A similar behavior is expected in a repository. Initially, the corrosion rate will be
higher than expected for nonirradiated material, as observed by Garzarolli. Eventually the
corrosion rate will revert to the.corrosion rate for nonirradiated material.
The rapid decay in corrosion rate from the in-reactor rate to that for nonirradiated material is not
due to normal thermal annealing, as estimated by the IA j parameter. Garzarolli's postirradiation
corrosion tests showed that less than 110 days at 280C is required to reduce the corrosion rate
to within a factor of two (or less) of the nonirradiated corrosion rate. One hundred and ten days
at 280C cor(esponds to an IA j parameter of only 10.30 h. Therefore, the rapid decrease in
postirradiation corrosion rate probably is due to a different effect than that associated with
thermal annealing. Although this effect has not been identified, the effect on repository corrosion
is known. Since repository fuel will be at temperatures similar to Garzarolli's 280C
postirradiation testing for on the order of 100 years (see Figure 4A), the postirradiation corrosion
rate enhancement will be well below the factor of two assumed in the analyses herein. Garzarolli
also reported the additional corrosion thickness for postirradiation corrosion rate to decrease to
within a factor of two of the nonirradiated rate. By two micrometers (J.im) of additional oxide, all
samples had a rate within a factor of two of the nonirradiated rate. Two micrometers (-0.08 mils)
is not significant for repository corrosion performance.
In summary, changes to zirconium-alloy microstructures during repository service, other potential
memory effects, and the presence of dissolved chemical species in the water (discussed earlier)
are not expected to be significant enough to result in a different dominant corrosion mechanism,
nor to alter the corrosion rate for the established corrosion mechanisms. Any material changes
that might occur are expected to decrease the corrosion rate due to the slow annealing of
irradiation damage.
Oxide Thickness
Uniform oxide thickness correlates to corrosion directly and indirectly. The direct effect is the
barrier that the oxide provides against the diffusion of oxygen ions through the oxide to the metal.
However, this barrier breaks down after about two microns of oxide thickness, after which only
the indirect effect of oxide thickness is important. The indirect effect is through the correlation
between the oxide thickness and changes in other parameters that influence corrosion. In
posttransition corrosion, the primary indirect factors are the stresses in the oxide and metal and
the concentrations of oxygen and hydrogen in the metal close to the oxide, each of which
correlates to the oxide thickness. In particular, compressive oxide stresses reduce the corrosion
rate [REFERENCE 48], perhaps by stabilizing the tetragonal oxide phase at the oxide-to-metal
interface [REFERENCE 49]. High concentrations of oxygen and hydrogen in the metal substrate,
especially when sufficient hydrogen is present to precipitate hydride phase, increase the
corrosion rate [REFERENCE 50]. The mechanism by which stress or the increased hydrogen
15

WAPD-T-3173

and oxygen concentrations affect the corrosion rate is not known. The important feature is that
the stress and the amount of hydrogen and oxygen increase are correlated to oxide thickness '
increase, notto time. The exception is for creep processes, which in principle could affect
corrosion. The potential for creep to affect corrosion is discussed below.
The increase in hydrogen concentration in the metal due to corrosion has been shown to be
proportional to the amount of oxidation. For in-reactor corrosion, the ratio of increased hydrogen
concentration to the amount of oxidation decreases with decreasing temperature [REFERENCE
51]. Out-of-pile, the ratio of hydrogen concentration increase to amount of oxidation is insensitive
to temperature. Therefore, as long as the expe,cted repository corrosion thicknesses are within
the oxide thicknesses in the current database, the effects of hydrogen concentrations on the
corrosion mechanism are expected to be predicted by the current models.
The oxygen and hydrogen concentrations in the metal close to the oxide also are influenced by
diffusion into the metal. If diffusion is relatively fast, the concentrations close to the oxide will
change little with increasing corrosion, in which case the concentrations will not be a function of
either oxide thickness or time. At autoclave test temperatures, hydrogen diffuses more rapidly
than corrosion occurs. Unless there is a temperature or stress gradient to bias diffusion,
hydrogen will not build up at the oxide interface. The opposite is true for oxygen. Whether there
is a change in going from autoclave test conditions to the repository depends on the relative
activation energies for diffusion of oxygen and hydrogen in the zirconium lattice, and possibly on
the hydrogen supercharging capability of the lattice. Since the activation energy for hydrogen'
diffusion is less than the activation energy for corrosion and the hydrogen diffusion is rapid at
autoclave temperatures [REFERENCE 52], the extrapolation to the lower repository
temperatures will not result in an increase in the concentration of hydrogen in the metal close to
the oxide. The activation energy for oxygen diffusion in the metal (approximately 45 kcal/mol,
REFERENCE 54) is greater than the activation energy for corrosion, but the oxygen diffusion rate
at the autoclave test temperatures of the existing database is slow enough that oxygen
concentrates at the oxide-to-metal interface. Extrapolation to lower temperatures will result in
continued concentration of oxygen in the metal at the oxide-to-metal interface, increasing the
gradient of oxygen concentration into the metal. If this were to result in a new rate-limiting step
at low temperatures, that is, if corrosion were to become limited by the diffusion of oxygen into
the metal, the corrosion rate would be less than predicted by the extrapolation used herein
based on the corrosion activation energy at existing autoclave test temperatures. Therefore, use
of the herein-proposed activation energy for extrapolation to repository temperatures is
conservative.
Stress in the thin protective oxide layer can affect the corrosion rate~ Compressive stresses
decrease the corrosion rate, perhaps by stabilizing the tetragonal oxide phase [REFERENCE 49].
Oxidation in the repository will increase the oxide compressive stresses, as oxidation does in
autoclave tests. These stresses can be relieved by creep in the metal substrate, both in the
repository environment and in autoclave tests. In the absence of a significant irradiation field in a
repository, the relative creep rates in the repository and in autoclave tests depends on the
thermal activation energy for creep, which is on the order of 70 kcal/mol [REFERENCE 53].
Since this is greater than the activation energy for corrosion, extrapolation to low temperatures
will decrease the relative amount of stress relaxation in the oxide compared to the amount of
creep that is reflected in the autoclave test database. This will tend to maintain the oxide
compressive stresses, which reduce the corrosion rate. The activation energy for oxide creep is
not well known. For CaO-stabilized oxide the activation energy is about 94 kcal/mol
[REFERENCE 55]. Extrapolation to lower temperatures will result in a decrease in creep rate
faster than the decrease in corrosion rate, which also will tend to maintain the compressive
stresses in the oxide, which reduce the corrosion rate. .

16

WAPD-T-3173

In summary, the extent of the current database for autoclave corrosion, in terms of oxide
thickness, provides confidence that key parameters associated with corrosion are adequately
covered. Based on a review of the thermal responses of potential contributors to the corrosion
processes, extrapolation to repository conditions is not expected to result in nonconservative
predictions of corrosion. Therefore, there is an acceptably low risk that the extrapolation to lower
temperatures and longer times will be significantly affected by the repository service
environment.
UNCERTAINTIES IN ESTIMATES OF ZIRCALOY CORROSION
The above estimates of the corrosion of Zircaloy for long periods of time at low temperatures
(FIGURE 4B) do not include conservatism to account for uncertainties. There are several ways
to address this issue and more is involved than just the uncertainty in the corrosion rate.
Uncertainty in the input to the corrosion calculation, for,example, uncertainty in the thermal
history or environmental conditions, also is important. Therefore, the appropriate conservatism
can not be addressed for this limited discussion of Zircaloy corrosion. In the end, it is the
conservatism in the final outcome, for example, release to the environment or risk of criticality,
that is important. Nevertheless, some perspective can be provided.
The uncertainty in the extrapolation from the database temperatures of about 600F and above
to repository temperatures is estimated through calculation of the uncertainty in the thermal
activation process, Q/R, as provided above. The upper 95-percent confidence limit of the
corrosion rate was used in combination with the estimated repository thermal history, which has
its own conservatism, to estimate the amount of Zircaloy corrosion that would occur after one
million years in the repository (see Appendix A for details). This value was found to be 0.63 mils,
compared to 0.14 mils calculated above using the best-estimate corrosion rate and the same
thermal history. This estimate is nota recommended method of handling conservatism but is
provided for perspective. The appropriate method for handling conservatisms should be
addressed in the context of the overall repository analysis procedure.
DISCUSSION AND CONCLUSIONS
It has been shown that, employing a conservative estimate of the Zircaloy corrosion kinetics, the
additional oxide film growth on Zircaloy cladding of spent naval fuel assemblies in the Yucca
Mountain Repository for a million years will be small (-0.3 mils), even when an additional
conservative factor of two enhancement of the rate is used to account for the prior irradiation
exposure of the cladding. This small increment in oxide film thickness should pose no problem to
the integrity of the Zircaloy Cladding. Chemical contaminants in the ground water and
microbiologically influenced corrosion, at currently anticipated levels, should have no adverse
effect on repository disposal of Zircaloy. The current understanding of Zircaloy corrosion supports
extrapolation from the database provided to the relatively low temperatures and very long times
for repository service. Direct measurement of the Zircaloy corrosion behavior at the anticipated
repository temperatures is not amenable to .experimentation due to the long times involved.
However, Zircaloy specimens are being included in the Lawrence Livermore National Laboratory
test program to evaluate the corrosion of various materials in J-13-type water [REFERENCE 57].

17

WAPD-T-3173

ACKNOWLEDGMENTS
The authors wish to acknowledge the direction supplied by, and the many helpful discussions
held with, Mr. Bruce F. Kammenzind. The aid of Dr. W.J. Duffin in compiling the chemistry data
is also appreciated. The assistance of Ms. Cheryl Hall in all phases of data collection, collation
and transmittal is also acknowledged and the use of Figures 4B and 7-9, supplied by Dr. Erik
Mader, is likewise appreciated.

REFERENCES
1.

Bryner, J. S., "The Cyclic Nature of Corrosion of Zircaloy-4 in 633 K Water," Journal of
Nuclear Materials, Vol. 82, 1979, pp. 84-101

2.

Thomas, D. E. in Metallurgy of Zirconium, B. Lustman and F. Kerze, Eds., McGraw Hill,

New York, 1955, Chapter 11, pp. 608-640

3.

Kass, S., Proceedings, U.S: Atomic Energy Commission Symposium on Zirconium Alloy
Development, Castlewood, Pleasanton, CA, 12-14 Nov., 1962, pp. 1-1 to 1-44, WAPD-T1549

4.

Kass, S. in Corrosion of Zirconium Alloys, ASTM STP 368, American Society for Testing
and Materials, 1964, pp. 3-27

5.

Thomas, D. E., Proceedings, International Conference on the Peaceful Uses of Atom

Energy, Vol. 9, United Nations, Geneva, 1955, pp. 407-413

6.

Chirigos, J. N. et ai, Proceedings, Fuel Element Fabrication with Special Emphasis on

Cladding Materials," Vol. 1, Academic Press, New York, 1961, pp. 19-55

1.

Kass, S., Corrosion, Vol. 23, No. 12, Dec. 1967, pp. 374-378

8.

Draley, J. E. et ai, Proceedings, Third International Conference on the Peaceful Uses of

Atomic Energy, United Nations, Geneva, 1964, Vol. 9, pp. 470-478

9.

Kass, S., and Kirk, W. W., Transactions, American Society for Metals, Vol. 55, No.1,
March 1962, pp. 77-100

10.

Kass, S., "The Corrosion and Hydrogen Pickup of Electron Beam Welded Nickel-Free
Zircaloy-2 and Zircaloy-4," WAPD-TM-533, Westinghouse Electric Corporation, March 1966

11.

Hillner, E., "Corrosion of Zirconium-Base Alloys - An Overview," Zirconium in the Nuclear

Industry, ASTM STP 633, A. L. Lowe, Jr. and G. W. Parry, Eds., American Society for
Testing and Materials, 1977, pp. 211-235

12.

van der Linde, A., "Calculation of the Safe Life Time Expectancy of Zirconium Alloy
Canning in the Fuel Elements of the NERO Reactor," RCN Report 41, Reactor Centrum
Nederland, Petten,Ju~1965

13.

Dyce, I. H., Nuclear Engineering, Vol. 9, No. 98, July 1964, pp. 253-255

18

WAPD-T-3173

14.

Dalgaard, S. B., "Long Term Corrosion and Hydriding of Zircaloy-4 Fuel Clad in Commercial
Pressurized Water Reactors with Forced Convective Heat Transfer" in Extended Abstracts
of the Electrochemical Society, Washington, DC, 2-7 May 1976, Vol. 76-1 Abstract #31, pp.
82

15.

Billot, P., Beslu, P., Giordano, A, and Thomazet, J., "Development of a Mechanistic Model
to Assess the External Corrosion of the Zircaloy Claddings in PWRs," Zirconium in the
Nuclear Industry: Eighth International Symposium, ASTM STP 1023, L. F. P. Van Swam
and C. M. Eucken, Eds., American Society for Testing and Materials, Philadelphia, 1989,
pp. 165-184

16.

Garzarolli, F., Jung, W., Schoenfeld, H., Garde, A M., Parry, G. W., Smerd, P. G.,
"Waterside Corrosion of Zircaloy Fuel Rods," EPRI NP-2789, December 1982

17.

Stehle, H., Kaden, W., and Manzel, R, "External Corrosion of Cladding in PWRs," Nuclear
Engineering and Design, Vo1.33, 1975, pp. 155-169

18.

Peters, H. R., "Improved Characterization of Aqueous Corrosion Kinetics of Zircaloy-4,"

Zirconium in the Nuclear Industry: Sixth International Symposium, ASTM STP 824, D. G.
Franklin and R B. Adamson, Eds., American Society for Testing and Materials, 1984, pp.
507-518

19.

Kass, S., "Aqueous Corrosion of the Zircaloys at Low Temperatures," Journal of Nuclear
Materials, Vo1.29, 1969, pp. 315-321

20.

Rothman, A J., "Potential Corrosion and Degradation Mechanisms of Zircaloy Cladding on

Spent Nuclear Fuel in a Tuff Repository," Lawrence Livermore National Laboratory Report
UCID-20172, September 1984

21.

Johnson, A B., Jr., "Behavior of Spent Nuclear Fuel in Water Pool Storage," Pacific
Northwest Laboratories, Richland, WA, Report BNWL-2256, September 1977

22.

Lanning, D. D., Johnson, A B., Jr., Trimble, D. J.; and Boyd, S. M., "Corrosion and
Hydriding of N Reactor Pressure Tubes," Zirconium in the Nuclear Industry: Eighth
Intemational Symposium, ASTM STP 1023, L. F. P. Van Swam and C. M. Eucken, Eds.,
American Society for Testing and Materials, Philadelphia,. 1989, pp. 3-19

23.

Graham, R A, Tosdale, J. P., and Finden, P. 1., "Influence of Chemical Composition and
Manufacturing Varia.bles on Autoclave Corrosion of the Zircaloys," Zirconium in the Nuclear
Industry: Eighth International Symposium, ASTM STP 1023, L. F. P. Van Swam and C. M.
Eucken, Eds., American Society for Testing and Materials, Philadelphia, 1989, pp. 334-345

24.

Garzarolli, F., Jorde, D., Manzel, R, Politano, J. R, and Smerd, P. G., "Waterside
Corrosion of Zircaloy-Clad Fuel Rods in a PWR Environment," Zirconium in the Nuclear
Industry: Fifth Conference, ASTM STP 754, D. G. Franklin, Ed., American Society for
Testing and Materials, 1989, pp. 430-449

25.

Johnson, A B., Jr., "Bases for Extrapolating Materials Durability in Fuel Storage Pools,"
Pacific Northwest Laboratory, Report PNL-SA-24126, December 1994

26.

Graham, R A and Eucken, C. M., "Controlled Composition Zircaloy-2 Uniform Corrosion

Resistance," Zirconium in the Nuclear Industry: Ninth International Symposium, ASTM STP
19

WAPD-T-3173

1132, C. M. Eucken and A M. Garde, Eds., American Society for Testing and Materials,
Philadelphia, 1991, pp. 279-303
27.

Isobe, T. and Matsuo, Y., "Development of Highly Corrosion Resistant Zirconium-Base

Alloys," Zirconium in the Nuclear Industry: Ninth International Symposium, ASTM STP
1132, C. M. Eucken and A M. Garde, Eds., American Society for Testing and Materials,
Philadelphia, 1991, pp. 346-367

28.

Urbanic, V. F., Choubey, R, and Chow, C. K., "Investigation of Variables That Influence
Corrosion of Zirconium Alloys During Irradiation," Zirconium in the Nuclear Industry: Ninth
Intemational Symposium, ASTM STP 1132, C. M. Eucken and A M. Garde, Eds.,
American Society for Testing and Materials, Philadelphia, 1991, pp. 665-682

29.

Pecheur, D., Lefebvre, F., Motta, A T., Lemaignan, C., and Charquet, D., "Oxidation of
Intermetallic Precipitates in Zircaloy-4: Impact of Irradiation," Zirconium in the Nuclear
Industry: Tenth International Symposium, ASTM STP 1245, A M. Garde and E. R Bradley,
Eds., American Society for Testing and Materials, Philadelphia, 1994, pp. 687-708

30.

Garde, AM., Pati, S. R, Krammen, M. A, Smith, G. P., and Endter, R K, "Corrosion

Behavior of Zircaloy-4 Cladding with Varying Tin Content in High-Temperature Pressurized
Water Reactors," Zirconium in the Nuclear Industry: Tenth International Symposium, ASTM
STP 1245, A M. Garde and E. R Bradley, Eds., American Society for Testing and
Materials, Philadelphia, 1994, pp. 760-778

31.

Berry, W. E., "Effect of Fluoride Ions on the Aqueous Corrosion of Zirconium Alloys,"
Corrosion of Zirconium Alloys, ASTM STP 368, 1964, pp 28-40 .

32.

Knittel, D. R and Bronson, A, "Pitting Corrosion on Zirconium - A Review," Corrosion, Vol.

40 #1, January 1984, pp 9-14

33.

Craig, D. B., Ed., Handbook of Corrosion Data, ASM, 1989

34.

Riley, W. D. And Covino, B. S. Jr., "Effect of Ferric Ion on Corrosion Resistance of

Zirconium in HCI-AICI 3 Environment," BuMines RI 8610, February 1982

35.

Manolescu, A V., Mayer, P., and Simpson, C. J., "Effect of Lithium Hydroxide on Corrosion
Rate of Zirconium 2.5 wt % Niobium in 340C Water," Corrosion, Vol. 38 #1, January 1982
pp 23-31

36.

McDonald, S. G., Sabol, G. P., and Sheppard, K. D., "Effect of Lithium Hydroxide on the
Corrosion Behavior of Zircaloy-4," Zirconium in the Nuclear Industry: Sixth International
Symposium, ASTM STP 824, Franklin, D. G. and Adamson, R B. Eds., American Society
for Testing Materials, 1984, pp. 519-530

37.

Hillner, E., and Chirigos, J. N., "The Effect of Lithium Hydroxide and Related Solutions on
the Corrosion Rate of Zircaloy in 680F Water," WAPD-TM-307, Westinghouse Electric
Corporation, August 1962

38.

Cox, B., "Oxidation of Zirconium and its Alloys," Advances in Corrosion Science and
Technology, Vol.5, Fontana, M. G. and Staehle,.R W., Eds., Plenum Press, 1976

20

WAPD-T-3173

39.

Coriou, H., Grail, L., Meunier, J., Pelras, M. and Wilermoz, H., "Corrosion of Zircaloy in
Various Alkaline Solutions at High Temperature," Journal of Nuclear Materials, Vol. 7 #3,
1962, pp 320-327

40.

Pourbaix, M., Atlas of Electrochemical Equilibria in Aqueous Solutions, Pergamon Press,

New York, N.Y., 1966, Section 8.2, p. 227

41.

Little, B., and Wagner, P., "An Overview of Microbiologically Influenced Corrosion of Metals
and Alloys Used in the Storage of Nuclear Wastes," Canadian Journal of Microbiology, Vol.
42, 1996, pp.367-374

42.

McNeil, M., and Odom, A.,"Thermodynamic Prediction of Microbiologically Influenced

.Corrosion (MIC) by Sulfate-Reducing Bacteria (SRB)," Mic.robio/ogically Influenced
Corrosion Testing. ASTM STP 1232, Kearns, J. R and Little, B. J., Eds., American Society
for Testing and Materials, Philadelphia, 1994, pp.173-179

43.

Cox, B., "Rate Controlling Processes During The Pre-Transition Oxidation of Zirconium
Alloys," Journal of Nuclear Materials, Vol. 31,1969, pp 48-66

44.

"Precipitate Stability in Zircaloy-4," EPRI Report NP-5591, January 1988

45.

Yang, W.J.S., Tucker, RP., Cheng, B., and Adamson RB., "Precipitates in Zircaloy:
Identification and The Effects of Irradiation and Thermal Treatment," JOl.lrnal of Nuclear
Materials, Vol. 138, 1986, pp 185-195

46.

Griffiths, M., "A Review of Microstructure Evolution in Zirconium Alloys During Irradiation,"
Journal of Nuclear Materials, Vol. 159, 1988, pp 190-218

47.

Garzarolli, F., Steinberg, E., and Weidinger H.G., "Microstructure and Corrosion Studies for
Optimized PWR and BWR Zircaloy Cladding," Zirconium in the Nuclear Indusfry: Eighth
Intemational Symposium, ASTM STP 1023, Van Swam, L.F.P. and Eucken, C.M., Eds.,
American Society for Testing and Materials, Philadelphia, 1989, pp 202-212

48.

"Corrosion of Zirconium Alloys in Nuclear Power Plants," IAEA-TECDOC-684, IAEA Vienna,

January, 1993

49.

Godlewski, J., "How the Tetragonal Zirconia is Stabilized in the Oxide Scale that is Formed
on a Zirconium Alloy Corroded at 400C in Steam," Zirconium in the Nuclear Industry:
Tenth International Symposium, ASTM STP 1245, Garde, A.M., and Bradley, E.R, Eds.,
American Society for Testing and Materials, Philadelphia, 1994, pp. 663-686

50.

Kido, T., "A Study on Enhanced Uniform Corrosion of Zircaloy-4 Cladding During High
Burnup Operation in PWRs," Proceedings, Sixth Symposium on Environmental
Degradation of Materials in Nuclear Power Systems-Water Reactors, Gold, RE., and
Simonen, E.P., Eds.,National Association of Corrosion Engineers, Houston, TX, 1993, pp
449-453

51.

Kammenzind, B.F., Franklin, D.G., Peters, H.R, and Duffin, W.J., "Hydrogen Pickup and
Redistribution in Alpha-Annealed Zircaloy-4," Zirconium in the Nuclear Industry: Eleventh
International Symposium, ASTM STP 1295, Bradley, E.R, and Sabol, G.P., Eds., American
Society for Testing and Materials, 1996, PP 338-369

21

WAPD-T-3173

52.

Keams, J.J., "Diffusion Coefficient of Hydrogen in Alpha Zirconium, Zircaloy-2 and Zircaloy4," Journal of Nuclear Materials, Vol.43, 1972, pp 330-338

53.

Ardell, AJ., and Sherby, 0.0., "The Steady-State Creep of Polycrystalline Alpha Zirconium
at Elevated Temperatures," Transactions of the Metallurgical Society of AIME, Vol.239,
October 1967, p. 1547

54

Rosa, C. J., "Oxidation of Zirconium - A Critical Review of the Literature," Journal of Less
Common Metals, Vol. 16, 1968, pp. 173-201

55.

St. Jacques, R G. and Angers, R, "Creep of CaO-Stabilized Zr02 ," Journal of the
American Ceramic Society, Vol. 55, No. 11, 1972, pp. 571-574

56.

Beie, H.-J., Mitwalsky, A, Garazarolli, F., Ruhmann, H., and Sell, H.-J., "Examinations of
the Corrosion Mechanism of Zirconium Alloys," Zirconium in the Nuclear Industry: Tenth
Intemational Symposium, ASTM STP 1245, Garde, AM., and Bradley, E.R, Eds.,
American
Society for Testing and Materials, Philadelphia, 1994, pp 615-643

57.

Clarke, C. W. L., "The Safe Disposal of Nuclear Waste," Science and Technology Review,
a publication of the Lawrence Livermore National Laboratory, March 1996, pp 6-16

58. Cheng, B.-C., Kruger, RM., and Adamson, RB., "Corrosion Behavior of Irradiated
Zircaloy," Zirconium in the Nuclear Industry: Tenth International Symposium, ASTM STP 1245,
Garde, AM. and Bradley, E.R, Eds. American Society for Testing and Materials, Philadelphia,
1994, pp 400- 418
59.

Boase, D.G., and Vandergraaf, T.T., "The Canadian Spent Fuel Storage Canister: Some
Materials Aspects," Nuclear Technology, Vol. 32, January 1977, pp 60-71

60.

Suzuki, M., and Kawasaki, S., "Oxidation of Zircaloy in Air," Journal of Nuclear Materials,
Vol. 140, 1986, pp 32-43

61.

Einziger, RE., "Preliminary Spent LWR Fuel Oxidation Source Term Model," High Level
Radioactive Waste Management, Proceedings of the Fifth Annual International Conference,
Las Vegas, Nevada, May 22-26, 1994, American Nuclear Society, Inc., La Grange Park,
Illinois, Vol.2, 1994, pp 554-559

62.

Van Swam, L.P.F., and Shann, S.H., liThe Corrosion of Zircaloy-4 Cladding in Pressurized
Water Reactors,"Zirconium in the Nuclear Industry: Ninth Intemational Symposium, ASTM
STP 1132, Eucken, C.M., and Garde, AM., Eds. American Society for Testing and
Materials, Philadelphia, 1991, pp 758-781

63.

"MATPRO: Version 10, A Handbook of Materials Properties for Use in the Analysis of Light
Water Reactor Fuel Rod Behavior," TREE-NUREG-1180, 1978

64.

Maguire, M., "The Pitting Susceptibility of Zirconium in Aqueous CI" , Br, and I" Solutions,"
Industrial Applications of Titanium and Zirconium: Third Conference, ASTM STP 830, R.T.
Webster and C.S. Young, Eds., American Society for Testing and Materials, Philadelphia,
1984, pp.177-189

22

WAPD-T-3173

Table 1: Comparison of Predictions From Eight Models for the Corrosion of

Zircaloy after 10,000 Years at 180C (356F)
6.W

Thickness

Q,/R

mQ/dm 2

mils

Hillner [11]

1.12E+08

-12529

1.10E-04

402

1.1

van der Linde [12]

2.30E+09

-14451

3.25E-05

119

0.3

Dvce [131

6.53E+09

-15109

2.16E-05

79

0.2

Dalgaard [14]

1.84E+07

-11222

3.23E-04

1181

3.1

Billot [151

1.13E+08

-12567

4.95E-05

181

0.5

Garzarolli [16]

1.18E+09

-13815

1.02E-04

374

1.0

Stehle [17]

2.21E+09

-14242

6.80E-05

248

0.7

Peters [18]

8.12E+08

-13512

9.12E-05

333

0.9

Investigator [REF1

23 .

WAPD-T-3173

Table 2: Summary of Material and Test Conditions

Test
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22

Test
ID

Spec.
No.'s*

Temp,

' 5201
'5501
'5502
'5503
\ 6001
' 6002
'6003
'6004
'6005
'6006
'6007
'6401
' 6801
' 6802
' 6803
' 6804
'\ 6805
'\ 6806
'6809
' 68010
' 68011
68012

4/4
10/9
6/5
6/5
5/5
5/5
3/3
15/3
20/9
15/3
21/9
6/3
7/6
2/2
7/7
4/4
3/3
17/6
71/57
12/12
29/27
10/9

520
550
550
550
600
600
600
600
600
600
600
640
680
680
680
680
680
680
680
680
680
680

of

Material
Zr-2
Zr-2
Zr-2
Zr-4
Zr-2
Zr-2
Zr-2
Zr-2
Zr-4
Zr-4
Zr-4
Zr-2
Zr-4
Zr-4
Zr-4
Zr-4
Zr-2
Zr-2
Zr-4
Zr-4
Zr-4
Zr-4

Heat

Treat

a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
~
a
a
a
a
a
a&~

The number to the left of the / shows the number of specimens at the start of the
posttransition period and the number to the right of the / shows the number of
specimens at
the completion of the test.
*

24

WAPD-T-3173

Table 3: Summary of Dual Linear Posttransition Analyses

Test
No.

Temp.
Deg.F 11T(K)

520

2
3
4

time*
tran.

W**
tran.

time*
max

W**

max

In[K(1)] In[K(2)]

1.84E-03 0.012 0.014 19.4

3000

60

-4.425

550
550
550

1.78E-03 0.021 0.031 -3.4

1.78E-03 0.028 0.024 -9.4
1.78E-03 0.032 0.030 -2.5

9347
6381
6381

229
163
203

-3.863

-3.583

-3.442

5
6
7
8
9
10
11

600
600
600
600
600
600
600

1.70E-03
1.70E-03
1.70E-03
1.70E-03
1.70E-03
1.70E-03
1.70E-03

-22.7
-23.9
-28.3
-26.5
-13.6
-17.9
-15.4

6874
6461
5855
4251
4320
4522
4498

419
473
472
366
310
342
334

10507
10059
8441
7039
7039
7039
7039

801
912
839
943
674
600
725

-2.745
-2.565
-2.461
-2.382
-2.592
-2.531
-2.555

-2.274

-2.210

-2.235

-1.608

-2.102

-2.269

-2.004

12

640

1.64E-03 0.200 0.271 -41.8

3134

584

6392

1665

-1.611

-1.307

13
14
15
16
17
18
19
20
21
22

680
680
680
680
680
680
680
680
680
680

1.58E-03
1.58E-03
1.58E-03
1.58E-03
1.58E-03
1.58E-03
1.58E-03
1.58E-03
1.58E-03
1.58E-03

679
720
672
493
887
869
1184
1292
1256
1009

225
237
225
154
351
214
421
438
407
406

1568
1554
1560
1568
1568
1848
2854
2854
2854
2730

688
651
722
663
941
647
1265
1169
1246
1485

-1.011
-1.034
-1.017
-1.113
-0.825
-1.365
-0.977
-1.037
-1.082
-0.899

-0.719

-0.717

-0.702

-0.819

-0.336

-0.928

-0.813

-0.887

-0.821

-0.587

K(1)

0.064
0.077
0.085
0.092
0.075
0.080
0.078

0.364
0.356
0.362
0.329
0.438
0.255
0.376
0.354
0.339
0.407

K(2)

0.103
0.110
0.107
0.200
0.122
0.103
0.135

0.487
0.488
0.496
0.441
0.715
0.395
0.444
0.412
0.440
0.556

C(1)

-22.4
-18.8
-18.2
-7.7
-38.0
-7.8
-24.2
-20.4
-18.4
-4.1

* Time, in days
**Weight Gain. in mg/dm 2
Transition time and transition weight gain, in this context, are between stage-1 and stage-2 linear
kinetics

25

WAPD-T-3173

Table 4: Comparison of Predictions From Nine Models for the Corrosion of

Zircaloy after 10,000 Years at 180C (356F)

fiW

Thickness

Q,/R

I<t

mg/dm 2

mils

Hillner [11J

1.12E+08

-12529

1.10E-04

402

1.1

van der Linde [12]

2.30E+09

-14451

3.25E-05

119

0.3

Dyce [13]

6.53E+09

-15109

2.16E-05

79

0.2

Dalgaard [14]

1.84E+07

-11222

3.23E-04

1181

3.1

Billot [151

1.13E+08

-12567

4.95E-05

181

0.5

Garzarolli [16]

1.18E+09

-13815

1.02E-04

374

1.0

StehleJ171

2.21E+09

-14242

6.80E-05

248

0.7

Peters r181

8.12E+08

-13512

9.12E-05

333

0.9

This Work, Equ. [6]

2.46E+08

-12877

1.12E-04

410

1.1

This Work, Equ. [7]

3.47E+07

-11452

3.67E-04

1341

3.5

Investigator [REF]

As with Table 1, all of the numbers in Table 4 have been rounded.

26

WAPD-T-3173

Table 5: Composition of J-13 Well Water

Chemical Species

Concentration (mg/L or ppm)

Na+

46

Si4+

29

Ca 2+

13

K+

Mg2+

8 3+

0.13

Li+

0.05

HCO::\-

130

S04=

18

N03-

CI-

7.1

F-

2.2
pH = 7.4

27

WAPD-T3173

POSTTRANSmON
(UNEAR)

en
en
w

~
(J

:cto-

TRANSITORY

)(

o
a:
o
z
Ci
c,:)

to-

iii

PRETRANsmON
(CUBIC)

,,'

"

_ _ _ J

CORROSION TIME

Figure 1:

Schematic Drawing Showing the Three Zircaloy Corrosion Regions:

Pretransition, Transitory, and Posttransition. The Dashed Unes Indicate that Early
Models Recognized Only the Pretransition and Posttransition Kinetic Regimes.

28

WAPD-T-3173

1,200

Test No. 21

r-----...,....---.....,..------,----.......---...,...----,

C)

800 1-----_+-----+-----+--~----:-_+---r.:I_---+-------I

C
.-

co

600

I-------+-----t-----t----,,'A-L....--+-----+----t

400

I-------+-----t-...".a::;ri=:l---t----,---+-----+----t

200

1------+-...,-::i:3="----+---~-+----_t_----+------t

CJ
+-'

-,

.0)

OL...........I..---I..-.......----l.--"---....I..---I..-.......----l.-----.......- ....

Figure 2A:

500

1,000

1,500

Time, Days

2,000

2,500

Corrosion of Zircaloy-4 in 680F (360C) Water - Test No. 21.

Data Show Curvature About a Single Least-Squares Straight Line.

29

3,000

WAPD-T-3173

Test No. 21

1 ,200 r------r------r----...,....----.,.------,~----,

1,000
C'\I

-....
en

"'C

800

E
...

.-ro

600

....,
..c
.-en
m

400

CJ

K(1 ) - 0.339

200

500

1,000

1,500

Time, Days

Figure 2B:

2,000

2,500

Same Data as in Figure 2A, Replotted to Show that a Dual Linear Relationship
is a Better Representation of the Weight Gain/Time Measurements.

30

3,000

WAPD-T-3173

0.0 .....- -....- -......---.,.....--..,..--....,.---.,....---,

--

......

-1.0

C\I

~
......

c:

...J

-2.0

---- -

-3.0

- - - - -

- - - - - - - - - - - -

-4.0

- - - - -

- - - - - - -- - - - -

L-

--

......
~

~
......

c:

...J

- - -- -

-5.0 L-_ _.J-_ _....L.._ _....L._ _.....L.

1.55

1.6

1.65

1.7

1IT
KU)

1.75

I.o..-

1.8

..

1.85

10" 3 (K)

K(2)

-~-'~

Figure 3:

Stage-One and Stage-Two Linear Rate Constants as a Function of the Inverse

Absolute Temperature.

31

1.9

WAPD-T-3173

Temperature Profile

250.,....-..,.....,~.,.,.".r--~""l""n"'m'T"---,.....,....n"T'l'TTT"'"--,-""T""T'''T''T''I''m--r-'''I'"'T'"T"TTTrr--r-!'""T'T''T'T''TTr--r-r~''T'!'T1
I
I

I I III
11111

I I I
III

I I III
11111

I
I

11111
IIIII

III
III

11111
11111

IIIII

11.1

IIIII

III1I

III

11111

IIIII

III

11111

III

11111

III

IIIH

III

I
I

I I I II
I I III

I I II
I I I

200 t--.L...l.~~~~-+-...l:;;j,.~~+--+-++-;+!~l.:--;-+-H+;++J--;-7::-. H:-7:-7:7:;t:1-~~-;-;-;-:-;:-;:7:t-I-;-' ~:-;-:

IIIII
11111
IIIII

11111
1IIII
IIIII

"

11111

II'

IIIII
11111

,,
,

IIIII
I 1111

I1II
, ,I I11111
, , IIII1
, I
, I I11111
IIII
, I 11111

,
,

11111
t 1111

11111
11111

11111
I I "I
I I I II

I III
11111
I I "I

I I "I

I 11111
I

1111\
I 11111

IIIII

IIIII

IIIII

1IIII

!'' I

I
11111

100

10

,
I
,

,
, '1111
I

I III
I

III

I "I
I II I

111111

111111

"

I I III

11111

III

'"

I1II

11111

';-;j

T:;-:

I III

I I III
11111
I till

,I
,

I II1II

11111
11111

I I I II

I 11111
I 111111
I
I

111111

111111

~ ~ I ~ ~I
10.000

1.000

1OO~000

1.000.000

Time (yr)

Figure 4A:

Anticipated Decay Heat Curve for Expended Naval Fuel in Yucca Mountain
Repository.

Oxide Thickness

0.7

0.6 f - -

-e
-

0.5 ~

--

95% Upper Confidence

-Best-Estimate

.!!!

0.4

Gl

.2

.-

I/)
I/)

.:.::

",i

.- .-I
.- .!I

0.63

0.3

..c

0.2

v
I

0.1

II

0
0

10

100

1.000

;...---------;0.14

"

10.000

100.000

1.000.000

Time (yr)

Figure 4B:

Estimated Additional Corrosion on Zircaloy Cladding During Yucca Mountain

Disposal. Based on Temperatures from Figure 4A and Use of Equation [7].

32

WAPD-T-3173

1,600

1,400

-C
......

0)

1,200

1,000

.-c:

800

J:

.-

600

400

..

CJ

....

0)

200
0

1,000

2,000

3,000

4,000

5,000

TIME, days
Steoam Water
,0

Figure 5:

Corrosion of Alpha-Annealed Zircaloy-2 at 640F (338C)

Steam vs. Water Exposure.

33

6,000

7,000

WAPD-T-3173

1E-29

, , , ""I
, , 111111
, ,
, ,,
,,
, ,,
11111
IIIII

,
,
,
,
,
,

, ,,
,

11111

I 11111

1E-30

t 1111

IIIII

1E-31

I-

CI)
CI)

E
ns

l-

ns

a.

C)

.-iiiC

1E-32

, ,
,
, ,

11111

t I I II

11111

,
, , , ,,
, , I
, , , '"

1E-33

IIIII
"
"

, ,,

1E-34

, ,

1E-35

'"

III

I III
I

'"

I
,

II:

1E-36

, ,
, ,
,
, ,,
I

1E-37

I1III
I 1.1111
11.111

,
, ,
, ,
,
, ,,
, ,,
I

1E-39
1E-40

I
I

1E-41

I !Ill
I I !I II
I I Ill!

11111

I 11111

11111
t 1111
11111
I II III

, ,
,

, ,

II I 1'1

11111'
, , I, 11111

1E-42
1E-43

1II1I
I 11111

, ,
,,

1E-38

,
,
I

,
,

, I1II1
, , t 1111
, IIIII

,.,

, ,,

"

, III
, , It I1111
, , II1II
11111

,
,
,
,

II'Jrj

till'

IIII

111111
: 11111

II'
I III

I III

I III

I III

,
,

I 11111
11111

"

III

I 11111
I I 1111

I
I

' Itlll

10

1II1I
IIIII
I 11111

"I

I
I

III

11111
I 11111

"

"

I 11111

I
I

I 11111
I 11111

I I II II

I 11111

,,

I I I III

I 11 III

,
,
, ,
I
I

I
I
I
I

I
I

; I I III
I I I III
: 11111!

I IIIII

I IIIII

II1I

11111

I 11111

11111

I IIIII
I 11111

11111

I
I

,
I

11111

, 1111

11111

I 11111

,,

,
,
,

.11111

, .,
,
I

I 11111

I 11111

I I I III

11111

I 11111

, 11111
III1I
,

II II
11 III
I 11111

, 11111

I1III
11111

,
,

,
I
I

11111

I 11111
II 111

,
I
I

tlill
11111

I 11111

I
I

11111

I "

11111

11111

I 11111

I
I
I

I
I

I III

I III

I 11111

I III
I III

I I I III

IIIII
I 11111
I

I 11111
I 11111
I I I III
11111
I 11111
I 11111

I Illtl

! 1111
11111

I
I

I
I

I I II II

11111
11111
I II II

I I II II
I I I III

I 11111

II III

I 11111

" III
: 'III

I
;

I 11111
I I II II

1,000

I I II J

I IIIII

I1I1I

! 1111

,,

I I III

I II I"
I 11111

I 11111
I 11111

11111

I IIIII

11111

,,

, ,
,
,,
I

I 1111

,,
,

,
,

I 11111

II I II
I 11111
I 1111/

100

I 1111

I 1111

, ,

I 1111
I 1111

'1

, IIII

,
, ,,
,
, ,
, ,

11111

I
I

11111

,,
I

III

I 11111

' ,

, , ,I

I I
11111
I I

, ,

,
,,
, ,
,

11111

I 11111

I IIIII

I I II II

, I I 11111
, , , 11111
, ,
, II I 11111
11111
, 11111
, , ,I 11111
, , ,
, , , 11111
11111
, I , 11111
, I I Illl!
I Jill

I IIII

,
,,
,
,

I I I III

, , IIIII

I 11111

I II'
III

11111

11111
I 1111

,
, , ,

11111

I J 1111

, 11111

I Hili

,
,
,

III

I
I

IIIII
I1III

,
,

, ,
, ,
, ,
,

I III

,,

IIIII

I 1111

'1

I.

,,

, ,

,
,
I

, '"
, ,
, ,,
, ,
, , ,"
, ,
'"
I

,,

, ,

~'''I

I II'"

111111

11111
I II II
I II II

I "II

,
I

" '" I

I 11111

IIIIII

,7::'
,
,

:/'"''

I J 1111

, , 11111
I I III

I
I

, ,I II I I I ,
, , I , ., ,
, , I ," ,

CI)

C
C

Annealing Parameter LA j

, ,

11111

I 11111

I II II

11111

I
I

10,000

,
, ,

,
, ,
I
I

II

,I

, ,

,
,

' .,.,

,,
I

till

11111
11111

I
I

, ,
,,

11111
11111

..

11111

1111.

, ,

11111

, ,
, ,
,
,

I
I

,
I

t 1:11
11 1 11
, 1111

llill

, , llill
11111

,,
,

11III

11111
Iii"

,,
, , ,,

II1II

"

IIIII
I II II
I 11111
I I I III
I 11111

,I
I

, ,I 11111

I IIII
I III

, ,,
,
,,
I

III1

I
I 11111
I 11111
I 11111
"I

I II III
I 11111

I 11111
I I II II
I I II II
Illli

11';1

I1I1I

"

11111
I II II

,
,

, , r 1111
, , 11111

I III

i'l:

,
I

11'11
I !!II

I II I I

, IIIII

I I1III

,
, ,,
, ,,

11 III
11111

, , 1111!
, , I II III
,
,; , , I111:1
lfli
, , , 111.1
, , illli
I

100,000

ll,ll

1,000.000

Time (yr)

Figure 6:

Cumulative Annealing Parameter (LA) Based Upon Anticipated Repository

Temperature Profile of Figure 4A.

34

WAPD-T-3173

(a)

High Flux vs. High-Then-Low Flux Corrosion Data

3

.--

2.5

.:.::

1.5

High Flux Data

High-Then-Low Flux Data

....
I

:.

.!!

en
en
Q)
c

.2

I-

High-Then-Low Flux
Coupon

Q)

"C
x

0.5

~~.

500

1000

Ii

;AIi

.c

I' iI

lJIi

. I

I
I

I
I
I

1500

2000

3000

2500

3500

Time (days)
(b)

High-Then-Low Flux Exposure History

,..--680

-iii

-~

It

640

.a~

600

'

- -

-:-

--

----- - - ----I

:
- - - - - - - -,- -.- - - - - - - - - - - - - ,
~

680

-. - 1 - I

1.6

."

C
)C
1.2 )(

It

;:

~
CD

- -.. ,.
'

...o

0.8 ~
-------------------------3II>
.!!.
0.4

CP

~ 560

{!!.

520
480

+:
- - - - ., - - - ,

(c)
,. 2

.I--

Figure 7:

500

1500

2000

.,-----,.------,.-----r-----r 2
-

640

Time (days)

-1- -

: ...:+

-1- -

-I -

-: -

I '

'

- - - - - - "':" i' - - -

.. -

600

-,-

'

.'
I

- j .... -

-:

500

-;'

.. ~

1.6

C
)C
1.2 )(

...o

I - - -

."

;:

0.8 ~
3
III

520

,:.

--;---7""---i---

0.4 -

0
1000 1500 2000 2500 3000 3500
Time (days)

Effect of Shift in Flux on Subsequent In-Pile Corrosion Rate.

35

-:r ~ -~ ----: ------- ~ ---

480

2500

,
I;
.: _
- - -.--- - - - - - .... - - - , - - - ~ ... - - :- - - -

,~ 560

------_....l. 0

1000

Characteristic High Flux Exposure History

---------~ ~ ~---~~--

WAPD-T-3173

(a)

Low Temperature vSo High-Then-Low Temperature

Corrosion Data

-Os

2.5

Low Temperature Data

Temperature Downshift Data

.!!

f I)

fI)

Temperature Downshift
Coupon

Q)

.~

1.5

I---

.c

IQ)

"'D

0;c

0.5

I~

_______

Ii

i .:

i..

I'1500 11

1000

:
... t

I.

~/I
I

500

I
i

I
i
I

I
!

I
0

I
2000

3000

2500

3500

Time (days)
(b)

680 ;:

- 640
If>-

---,- 2

T - - - - - - - -:- - - - -, - - - - ~ - - - -

- - - -.- - - - -:- cjl - - - - - - - - - - - -

.:.

'

____

1.

!!!I __

~L

., ,

.!-

Figure 8:

500

1000
1500
Time (days)

Characteristic Low Temp. Exposure History

r-----,.---:-------~-,
,
,
o
o

."

- 640
If>-

C
loC
1.2 )(

....o

.a
QI

5
n

480

r - -

II
~

t l '

I"
, 0

'I'

""

--

1.6

- -

T'"

"

-t- .... - -:- - - I

. . ., '

'

500

36

0.8

5n

w
0.4 -

0
1000 1500 2000 2500 3000 3500
Time (days)

Effect of Shifts in Temperature and Flux on Subsequent In-Pile Corrosion Rate.

."

C
loC
1.2 )(
....
o

520

2500

-I

--------------------------

~ 560
~

.
,

I.

-j- -

- - -1--- - - - - - - - - - - - -

600

QI

.--

---l----;-----------r--.-~---

1.6

0.4

_.,..--~

2000

680

0.8

_______~r-----.~--------

520
480

(c)

Temperature Downshift Exposure History

.....-

Appendix A to

WAPD-T-3173

STATISTICAL ANALYSIS OF AUTOCLAVE

CORROSION DATA

WALTER A. MEYER

37

Appendix A to

WAPD-T-3173

Statistical Analysis
This section documents the analysis methodologies used to arrive at the conclusions presented
in the report. The outline is in four sections:
1.
2.
3.
4.

Handling the raw data and identifying outliers.

Fitting the curves with a multiphase model.
The database of results and Arrhenius behavior.
Uncertainty of the best fit line.

1. Handling the raw data and identifying outliers

The data from twenty-two separate autoclave tests were recorded in a spreadsheet.
Most runs were comprised of only a single heat of material. Since the data spanned many years
and some of it was recorded and entered into a computer manually, the first procedure used to
analyze the data was to check for outliers. Each run consisted of three to fifty-seven individual
specimens tested over a few thousand days to up to ten thousand days. To detect changes in
the variability, the standard deviation of the specimens was computed and plotted for each time it
was recorded. A sample pioUs shown in Figure A.1 for one run at 600 of. As the figure shows,
the variability among the specimens tends to increase as the weight gain increases. This is
typical behavior of a gage that has error proportional to range or scale. The figure also indicates
two possible anomalies in that run. Each anomaly was investigated. For the circled point in the
figure, it was found that one of the specimens had an unusually low value recorded at that
temperature which caused the standard deviation of the specimens to roughly triple. It was also
observed that the recorded weight gain was inconsistent with both the value immediately prior to,
and immediately following, this exposure time. Since that data point was obviously bad it was
removed. Without that point, the standard deviation came back into normal range. The other
point in the figure that seemed too high was investigated and no highly unusual values were
found. Thus, this point was retained in the database. No further data were removed from this
run.
For the rest of the runs, this charting technique identified potentially bad data, which were
removed for various reasons, e.g., digits reversed, a digit left out, data for two temperatures
reversed, etc. When anomalies were found the only two actions taken were to leave the data
point in the database or to remove the data point from the database. Thirteen bad data points,
outof a total of 11,062 data, were found and removed.
2. Fitting the curves with a multiphase model
With the identifiable outliers removed, the average specimen weight was computed for
each of the twenty-two runs. There were from two to twenty-one specimens in this average. For
most runs, more specimens started than finished the test. Although all of the data were used to
compute the average weight gain for the run, the recorded number of specimens for a test
represented only those specimens that lasted through the entire duration of the test.
The average weight gain versus time data was then put into the SOLO statistical software
package. This package found that a two-phase linear model produced the simplest fit to the
data. The multiphase linear-linear model is made up of two linear equations, each active over a
different time range. The software automatically calculates the transition point from one phase to
another, and no smoothing was used in calculating the transition point. A sample of the output

38

Appendix A to

WAPD-T-3173

from the multiphase curve fit routine (for the same run at 600 of used to make Figure A.1) is
shown in Table A.1 and Figure A.2. All of these data were obtained from test #5 in the text.
Of interest in Table A.1 is the summary of parameters in the common model at the
bottom of the figure. It shows the two linear models with intercepts C(1) and C(2) and slopes
K(1) and K(2) and the transition point o. The transition point 0 is shown in days. Figure A.2
graphically shows the data along with the two-phased linear model. The eye immediately
focuses on the lack of fit in the low weight gain portion of the curve. This is due to a different set
of kinetics early in the corrosion process. The net effect of including this early weight gain data in
the model is to bias the slope for the first linear phase and a smaller bias in the transition point.
Th~ effect is quite small, however. For example, when the data is restricted to weight gains of
greater than 100, the slope changes from K(1) = 0.064 (shown in Table A.1 as 6.4224E-02) to
K(1) = 0.068 and 0 changes from 6,874 days to 7,468 days.
3. The database of results and Arrhenius behavior
For each of the twenty-two runs, the five parameters (C(1), K(1), 0, C(2), K(2)) for the
two-phase linear models were recorded into a summary database. The values for K(1) and K(2)
are reproduced in Table A.2. Also included in the table are the temperatures, the natural logs of
the slopes and the inverse of the absolute temperature. From this data two linear models were
fit. The fit of interest was the natural log of the slope versus the inverse of the absolute
temperature. Materials theory suggests that these fits should be straight lines so simple linear
regression was used to fit the data. (Note: A weighted least squares model, where the weights
were the number of specimens completing each run, was also analyzed. The results were less
conservative than those obtained with simple linear regression. The weighted model was
subsequently abandoned in favor of the simple dual linear-regression model.)
A separate line was fit for each phase of the weight gain curves. Since no transition to a
higher slope was observed for temperatures below 600F, only the 600,640, and 680F data
were used to fit the line through the natural logarithms of the phase-two slopes (K(2)'s). A plot of
the data and the best fit lines is shown in Figure A.3 (and Figure 3 in the text).
4. Uncertainties of the best line fit
Since simple linear regression was used to fit the data, a standard 95% confidence
interval for the mean In(Q/R) at a chosen point, x*, (where x* is in units of 1fT) is:
a + bx* tn_2 Se [(1/n) + (x* - x)2/S xx]Y%
The values for the parameters for line 2 can be found in Table A.2 and used with any x*:
a = 17.36226
b=-11,451.6
n 18
x = 0.00163
tn-2 = 2.12
Se = 0.1954
Sxx = 5.87E-8
This uncertainty for the mean In(Q/R) is applicable to temperatures within the 600F to 680F
band.

39

Appendix A to
WAPD-T-3173
Table A.1
Estimation Summary Report
Model: Y = A+BX+(X-D) * C * SGN(X-D)
Y: Weight Gain X: Time
T-Value Prob

( It I >T)

R-Squared

6.60

-23.5

0.0000

0.9984

8.3537531 D-02

7.94E-04

105.1

0.0000

1.931366D-02

7.94E-04

24.3

0.0000

6874.11

91.45

75.2

0.0000

Coefficient Estimate

Std. Error

-155.4187

Term

Source

df

Sum-Sqr

Mean Square

SQR(M.S.)

F-Ratio

Prob (f>F)

Model

3739838

1246613

1116.518

22253.7

0.0000

Error

109

6105.982

56.01819

7.48453

Total

112

3745944

33445.93

182.8823

=C(1) + K(1) (X)

Y =C(2) + K(2) (X)
Y

=-22.65
K(1) =6.422E-02

Parameters in Common Model

X<=a
X>a

C(1)

=6874

where a

C(2) = -288.1
K(2)

=.1028

=transition time, days

40

Table A.2

.,.,

-a.

Test No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22

Temp (F)
520
550
550
550
600
600
600
600
600
600
600
640
680
680
680 .
680
680
680
680
680
680
680

Regression Output:
Constant
Std Err of Y Est
R Squared
No. of Observations
Degrees of Freedom
X Coefficient(s)
Std Err of Coef.
Sxx
95_U
95_L

K(1)
0.0120
0.0211
0.0276
0.0317
0.0642
0.0769
0.0854
0.0923
0.0749
0.0796
0.0777
0.1997
0.3640
0.3557
0.3616
0.3285
0.4382
0.2553
0.3763
0.3544
0.3391
0.4071

Slope Parameters and Regression Fits

K(2)
0.0142
0.0313
0.0236
0.0296
0.1029
0.1097
0.1070
0.2002
0.1222
0.1034
0.1348
0.2706
0.4875
0.4881
0.4955
0.4407
0.7148
0.3953.
0.4436
0.4120
0.4399
0.5558

Line 1 (Stage-one Linear)

19.3224
0.1377
0.9852
22
20
-12876.96
352.36
1.53E-07
-12140.5
-13613.4

InK(1)
-4.4251
-3.8590
-3.5887
-3.4502
-2.7454
-2.5652
-2.4609
-2.3822
-2.5917
-2.5306
-2.5553
-1.6110
-1.0106
-1.0336
-1.0173
-1.1131
-0.8251
-1.3651"
-0.9774
-1.0373
-1.0815
-0.8988

InK(2)
-4.2556
-3.4645
-3.7456
-3.5189
-2.2745
-2.2098
-2.2347
-1.6083
-2.1017
-2.2696
-2.0043
-1.3073
-0.7185
-0.7172
-0.7022
-0.8195
-0.3357
-0.9280
-0.8128
-0.8867
-0.8213
-0.5873

11T( K)

1.84E-03

1.78E-03

1.78E-03

1.78E-03

1.70E~03

1.70E-03

1.70E-03

1.70E-03

1.70E-03

1.70E-03

1.70E-03

1.64E-03

1.58E-03

1.58E-03

1.58E-03

1.58E-03

1.58E-03

1.58E-03

1.58E-03

1.58E-03
1.58E-03

1.58E-03

Regression Output:
Constant
Std Err of Y Est
R Squared
No. of Observations
Degrees of Freedom
X Coefficient(s)
. Std Err of Coef.
Sxx
95_U
95_L

Line 2 (Stage-two Linear)

17.3623

0.1954

0.9264

18

16

-11451.63

806.74
5.87E-06
-9741.36
-13161.9

~~

"tJi

~&
SC"
~>
c:t o

Appendix A to

WAPD-T-3173

10 . . . . . - - - - - " " " " ' - - - - - - - - - - - - - - - - - - - - : - - - - - - - - - - ,

Outlier found and removed (one data

point was abnormally low, inflating sigma.)

r-

I-

I~

II

4 -

r-

111111 \

III'

""

, I I ! , I I I ' I I I I ' , , ' 'i

Ii""""

I I

"'! I I I

I I I '-l

' I I , , ' , , I ' , I , I , , , , , ' , I I I I I I I '_I I "

I I I I I I I ' I I

Time

Figure A.1:

Plot of Standard Deviation as a Function of Test Time, which was Used to Find
Outlier Data in the Database.

42

Appendix A to

WAPD-T-3173

800r-------:-------~--------....- - N

<

-0

"

C>

600

H
<I:

. .

400.

. . . .. .

t.!)

r-

::r:
t!)

H
W

. . . . .

200 . . .

. ... .

. . . .

. .

.,
OL..-c;.......

..:.....-

....;.-.

-=--

3000

6000

9000

-l

12000

CORROSION TIME, DAYS

Figure A.2:

Weight Gain as a Function of Test Time Compared to Statistically Fit Curve for a
Typical Set of Data.

43

Appendix A to

WAPD-T-3173

Or-------------------------------,

. -1

-2

Phase 1

Phase 2

!oS

-3

Best 1
8 Best 2

-4

-5 '--1.5

----'----'-'-1.6

-"-

-L-

1.7

1.8

---'

1.9

Thousandths

11T (Kelvin)

Figure A.3:

Corrosion Weight-Gain Rates Versus Reciprocal Temperature, Which Reflects the

Statistical Analysis to Determine the Thermal Activation Energy, Q.

44