See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/236122384

Chloride Ion Activity and Susceptibility of Al Alloys 7075-T6 and 5083-H131 to

Stress Corrosion Cracking

Article in Metallurgical and Materials Transactions A · March 2013

DOI: 10.1007/s11661-012-1500-2

CITATIONS READS

25 874

6 authors, including:

Richard E. Ricker
National Institute of Standards and Technology
163 PUBLICATIONS 1,482 CITATIONS

SEE PROFILE

All content following this page was uploaded by Richard E. Ricker on 09 May 2014.

The user has requested enhancement of the downloaded file.

Chloride Ion Activity and Susceptibility of Al Alloys 7075-T6
and 5083-H131 to Stress Corrosion Cracking
R.E. RICKER, E.U. LEE, R. TAYLOR, C. LEI, B. PREGGER, and E. LIPNICKAS

The influence of chloride ion activity on the susceptibility of aluminum alloys 5083-H131 and
7075-T6 to stress corrosion cracking (SCC) was investigated by conducting slow strain-rate
tensile tests at a strain-rate of 10 7 s 1 in naturally aerated aqueous solutions with varying NaCl
mass fraction (0.001 to 20 pct) and in a 3.5 pct mass fraction NaCl solution with varying strain-
rates (10 8 to 10 4 s 1). This study found that both alloys exhibited reduced strengths and
failure strains (times) in the solutions compared with laboratory air. The extent of these
reductions was greater in alloy 5083 for the conditions examined. The strength and ductility of
both alloys decreased with chloride ion activity in a manner that indicates a chemical reaction is
responsible. The strength and ductility of both alloys decreased with strain-rate in a sigmoidal
manner, but the transition in alloy 7075 occurred at slower strain-rates of approximately two
orders of magnitude. It was deduced that the chloride ion interacts chemically with the pas-
sivated surface in the potential gradient at the crack tip to cause SCC. While no mechanism of
cracking can be eliminated on the basis of these results alone, the results are consistent with the
hypothesis that the absorbed hydrogen causes cracking in alloy 7075 while cracking in 5083 is
the result of a dissolution mechanism.

DOI: 10.1007/s11661-012-1500-2
Ó The Minerals, Metals & Materials Society and ASM International (outside the USA) 2012

I. INTRODUCTION accepting electrons. Because aluminum contains suffi-

cient energy to reduce hydrogen ions from water
ALUMINUM is a relatively active metal that spon- molecules, the oxidizers primarily responsible for caus-
taneously reacts with oxygen or water to form aluminum ing corrosion of aluminum alloys in aqueous environ-
oxides, hydroxides, hydrated oxides, or mixtures of these ments are dissolved oxygen and water.[1,8] If the
phases.[1–3] The normally excellent corrosion resistance of concentration of the chloride ion has no influence on
aluminum alloys is the result of a thin impervious film of (1) the chemical potential of the elements in the alloy, (2)
these products forming on the surface. This film isolates the chemical potential of the elements in the corrosion
the underlying metal from the environment making it products, and (3) the activity of the oxidizers in the
nonreactive, or passive, in environments where the metal solution, then it has no direct influence on the free energy
is thermodynamically unstable.[4,5] As with most passiv- driving corrosion reactions. However, the literature
ated metals, corrosion problems arise if this passivating clearly shows that the chloride ion alters the rate and
film has difficulty protecting the metal because of chem- morphology of attack.[7,9,10] Therefore, these ions inter-
ical or mechanical forces. In particular, localized forms of act with corrosion processes in a manner that enables the
corrosive attack, such as pitting, crevice corrosion, and existing corrosion driving forces to exert their influence
stress corrosion cracking (SCC), may occur when halide in a more detrimental manner. The responsible interac-
ions are present in the environment.[6] tion may be as simple as altering the activated complex,
For aluminum alloys, the chloride ion, which is adsorbing in a manner that interferes with passivation,
commonly found in ambient environments, is one of the or altering hydrolysis equilibria. In all these cases, the
more aggressive of these ions.[4,5,7] The fact that these influence of the chloride should be related to its activity
ions appear to cause corrosion implies that they increase in a manner that indicates the stoichiometry of this ion in
the thermodynamic forces driving corrosion reactions. the responsible process. Quantifying the effect of the
However, these ions are the most reduced state for this chloride ion on SCC over a wide range of concentrations
element and cannot promote oxidation of metals by will be a critical step toward understanding how this ion
accelerates attack and causes SCC. Therefore, SCC
experiments were conducted on two aluminum alloys in
R.E. RICKER, Metallurgist, is with the Materials Science and
naturally aerated solutions with the chloride ion con-
Engineering Division, National Institute of Standards and Technology, centration varying over four orders of magnitude.
Gaithersburg, MD 20899. E.U. LEE, C. LEI, and B. PREGGER An Al-Zn-Mg alloy (AA 7075-T6) was selected for
Materials Engineers, R. TAYLOR, Materials Engineering Technician, this study because the SCC susceptibility of this alloy in
and E. LIPNICKAS, Chemical & Materials Engineer, are with the the high strength T6 temper limits applications.[4,5]
Materials Engineering Division, Naval Air Warfare Center Aircraft
Division, Patuxent River, MD 20670. Contact e-mail: [email protected] These alloys exhibit accelerated cracking not only in
Manuscript submitted March 26, 2012. NaCl solutions, but also in distilled water and
Article published online November 10, 2012 water vapor. In addition, cracking is observed in inert

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 44A, MARCH 2013—1353

environments following pre-exposure to environments Slow strain-rate (SSR) tensile tests were selected for
containing water.[2,11,12] Therefore, while the chloride this study because each experiment produces a quanti-
ion may promote this form of attack, it is not required tative measurement that can be used as an indicator of
to cause it in this alloy. The second alloy selected for this relative susceptibility.[19–23] Another advantage of the
investigation, AA 5083-H131, is a nonheat-treatable SSR technique is that the continuous application of
Al-Mg alloy known for its excellent resistance to dynamic strain facilitates SCC crack initiation and
corrosion. However, aging of this alloy at slightly propagation. This technique can produce SCC in
elevated temperatures (50 °C to 200 °C) for extended environments where other techniques fail to initiate
periods of time results in the precipitation of the cracking, take a prohibitively long time, or exhibit poor
relatively active b-phase, Al3Mg2, at grain boundaries. reproducibility.[19] The SSR test has been found to be so
The precipitation of this phase at grain boundaries has effective at initiating SCC that it has been used to study
been correlated to increased susceptibility to intergran- cracking in systems where SCC is not observed for other
ular corrosion and SCC.[13] The passive film on the loading conditions.[24–28] The disadvantage of the
b-phase is unstable at the conditions where SCC is smooth specimen SSR technique is that it does not
observed in this alloy, and SCC does not occur when this readily produce crack propagation rate data suitable for
passive film is stable.[13] Therefore, this alloy should only modeling and service life analysis.[29–31] However, a
exhibit SCC when the halide ion concentration is sufficient number of different techniques have been developed to
to cause pitting attack of the grain boundary b-phase. estimate crack propagation rates from these tests.[21,32]
These two alloys were selected because they represent As the aim of these experiments is to evaluate the effects
two extremes of the range of SCC behavior that is of bulk solution chemistry on susceptibility to SCC,
observed in aluminum alloys.[14] In the 5083 alloy, active smooth specimens that emphasize crack initiation in the
dissolution of an anodic phase from the grain boundary bulk solution, and minimize the influence of ion
is almost certainly responsible for SCC. However, there accumulation and hydrolysis reactions in the occluded
should be a critical halide ion activity required to initiate regions, were deemed appropriate.[33]
pitting of the grain boundary b-phase and SCC. In the
7075 alloy, cracking can occur without an aqueous
phase present to solvate ions and in inert environments II. EXPERIMENTAL
following pre-exposure to water. In addition, the pre-
exposure effect diminishes at a rate that correlates to the The nominal composition of the two alloys selected
rate of desorption of hydrogen.[2] The only viable for this study, 5083-H131 and 7075-T6, are shown in
explanation for these observations is that this alloy Table I, and Figure 1 shows the microstructure of these
absorbs hydrogen when exposed to water and that this alloys. The 7075-T6 alloy was tested in the as-received
hydrogen causes cracking.[2,15–18] While it is clear from T6 heat treatment. The 5083-H131 alloy was sensitized
the literature that absorbed hydrogen can cause crack- by artificially aging at 448 K (175 °C) for 240 hour in
ing in 7075, it still remains to be resolved if this is the air.
mechanism responsible for SCC in ambient aqueous The electrochemical behavior of the alloys was
solutions.[14] characterized by conducting polarization experiments

Table I. Chemical Compositions of 7075 and 5083 Aluminum Alloys in Mass Fraction (pct)

Alloy Cu Mn Mg Cr Zn Fe Si Ti Al
5083 0.10 0.4–1.0 4.0–4.9 0.05–0.25 0.25 0.40 0.40 0.15 bal
7075 1.2–2.0 0.30 2.1–2.9 0.18–0.40 5.1–6.1 0.7 0.5 0.2 bal

Fig. 1—Optical micrographs showing the microstructure of alloy AA 5083-H131 (a) and alloy AA 7075-T6 (b).

1354—VOLUME 44A, MARCH 2013 METALLURGICAL AND MATERIALS TRANSACTIONS A

in the solutions on specimens with the heat treatments same quantities for tests conducted in laboratory air. The
used in the SSR tests. For these experiments, specimens fracture surfaces were cleaned, dried, and examined by
approximately 76 mm by 76 mm square and 63 mm optical microscopy and in a scanning electron microscopy
thick were mounted on a flat electrochemical cell that with an accelerating voltage of 20 kV.
would expose a circular 100 mm2 area to the solution. As the chemical activity of ions in aqueous solutions
Approximately 250 ml of solution was added to the cell can deviate from their concentration because of non-
to completely immerse the specimen. The solutions were ideal interactions, the actual chloride ion activity (Cl )
exposed to air during the experiment to represent the was estimated for each solution using the literature data.
ambient conditions and the solutions in the SSR For these estimates, regression analysis was per-
chamber. Before the start of the polarization experi- formed[34] on the values recommended by a survey of
ments, the specimens were allowed to freely corrode in the literature[35] combined with more recent data cover-
the solutions under open-circuit conditions until a ing lower concentrations.[36–38] This analysis included
steady-state potential was reached. For the polarization regression of the literature data on solution density[39] to
experiments, the potential of the specimen with respect enable conversion between measures of concentration.
to a saturated calomel reference electrode (SCE) was The values estimated for the chloride ion activity of each
controlled with a commercial closed-loop, high-gain, solution are given in Table II, and all plots examining
amplifier (potentiostat) and scanned from 100 mV the chemical effects of chloride ion are made using these
below the free corrosion potential, while the current values for the chloride ion activity.
required to maintain the potential of the specimen was Varying the NaCl concentration will also vary a
recorded. Corrosion currents were estimated through number of solution properties that influence corrosion
linear polarization resistance measurements using scans behavior. Therefore, these properties were also esti-
from 15 mV below to 15 mV above the free corrosion mated by regression analysis of the literature data from
potential at 0.167 mV/s. different studies.[39] This analysis included estimation of
Smooth cylindrical tensile specimens, with a gage the viscosity, electrical conductivity, and the solubility
diameter and a length, of 6.35 mm (0.25 in) and of oxygen in the solutions.[34] As altering the chloride
38.1 mm (1.5 in), respectively, were machined out of the ion concentration could also influence the activity of
alloys with the loading axis parallel to the long-transverse water (H2O), this quantity was also estimated by
direction of the rolled sheet. The SSR tests were performed regression analysis of the literature data.[34,35,39] The
in a tensile loading frame at strain-rates varying from 10 8 estimated values for these quantities in each solution are
to 10 4 s 1. The specimens were tested in an environmen- included in Table II.
tal chamber that contained enough solution to completely Unless specified otherwise, each data point represents a
immerse the gage section of the specimen (0.1 L). During single measurement, and the error bars are the estimated
the tests, the solution was continuously circulated between standard deviation for measurements of this type using
the test chamber and a 0.5 L reservoir by a pump at a rate the instrumentation used for these experiments.
of approximately 0.3 L/m. The solution was exposed to
laboratory air in the reservoir to represent ambient
exposure conditions. Neutral pH solutions were mixed
from laboratory grade NaCl and deionized water with III. RESULTS
mass fractions (wNaCl) varying from 0.001 to 20 pct
(1.7 9 10 4 to 3.92 mol/L). The specimens were pulled in A. Electrochemical Experiments
tension to fracture, and the ultimate tensile strength and The results of the electrochemical polarization exper-
time-to-failure were recorded for comparison with the iments are summarized in Figure 2. Examination of this

Table II. The Chloride Ion Activities and the Properties of the Solutions Estimated by Regression Analysis of the Literature Data
at 25 °C*

Mass Fraction Density** Viscosity** Elect. Dissolved (Cl ) (H2O)

NaCl (wNaCl), Conc.(c) kg/L mPaÆs Cond** S/m Oxygen Conc,**, mol/L Activity§
pct mol/L (0.0002) (0.002) (0.04) lmol/L (0.95) (0.05 pct) (0.0005)
0.0010 0.000171 0.997 0.900 0.03 255 0.000168 1.000
0.0050 0.000853 0.997 0.900 0.03 255 0.000826 1.000
0.010 0.00170 0.997 0.900 0.04 255 0.00163 1.000
0.035 0.00597 0.997 0.901 0.08 255 0.0055 1.000
0.50 0.0856 1.001 0.904 0.88 248 0.0675 0.997
3.5 0.612 1.023 0.936 5.7 211 0.412 0.979
10.0 1.83 1.072 1.071 14.3 144 1.22 0.935
20.0 3.92 1.152 1.455 22.6 74 3.17 0.840
*The number in parentheses indicates the estimated uncertainty in the data.
**Based on aqueous NaCl solution at 25 °C data compiled by Lobo[34,39].

Equilibrium dissolved oxygen concentration when exposed to ambient laboratory air.

Analysis of activity data[35,37,38,40].
§
Based on the vapor pressure data of Partanen[38].

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 44A, MARCH 2013—1355

Fig. 2—Electrochemical polarization curves for alloy AA 5083-H131 (a) and alloy AA 7075-T6 (b) in aerated solutions of varying mass fraction
NaCl. The curves are labeled by the chloride ion activity given in Table II.

found that adding a change in the intercept at a chloride

ion activity of 0.01 mol/L significantly lowered the error
in the model, but that adding a slope change at that
point did not. Table III shows the results of regression
analysis of these models including correlation coeffi-
cients (R2) and the results of a test of the t statistic for
the significance of each parameter. This table shows a
significant probability that the observed slope change of
the four-parameter model is due to statistical scatter,
but that the probability that the intercept change in the
three-parameter model is due to statistical scatter is
essentially zero.
Figure 4 is a plot of the corrosion current density
estimated for the two alloys at different chloride ion
concentrations. In this plot, there is significant scatter,
but there is a clear trend of increasing corrosion current
density with chloride ion activity. Regression analysis of
these data indicates that the scatter is too great to make
Fig. 3—Measured corrosion potential as a function of chloride ion
activity.
a distinction between the slopes and intercepts for the
two alloys. Performing linear regression on the com-
bined data from the two alloys indicates an intercept of
figure shows that the corrosion potential steadily 1.43 (±0.10) V-SCE and a slope of 0.54 (±0.05) V/log
decreases as the activity of NaCl increases for both (mol/L) with a correlation coefficient (R2) of 0.927.
alloys. Plotting the corrosion potential verses chloride Therefore, the positive slopes in these data are signifi-
ion activity in Figure 3 shows that this trend is essen- cant, and the corrosion rate is increasing with the
tially linear with respect to the logarithm of the chloride concentration of the chloride ion.
ion activity [log (Cl )]. However, it also appears that the
corrosion potentials at the higher chloride ion activities
may be on a different line from those at lower
B. Mechanical Properties in Air
concentrations. As the differences between the alloys
were small, the data for the two alloys were combined The mechanical properties of the alloys when tested in
for regression analysis.[41] A commercial statistical laboratory air are given in Table IV. Examination of
analysis package was used to compare three different this table shows that, compared with the 7075 alloy in
models: (1) a single line (two parameters), (2) two lines the T6 temper, the 5083 alloy has a lower ultimate
with the same slope (three parameters), and (3) two lines tensile strength and greater strain-to-failure in air. This
with different slopes and intercepts (four parameters). table also shows that the tensile strength and the failure
The F-statistic was used to evaluate the statistical strain of alloy 7075 were essentially constant with
significance of adding terms to the models.[41] This test respect to strain-rate in air. The tensile strength of the

1356—VOLUME 44A, MARCH 2013 METALLURGICAL AND MATERIALS TRANSACTIONS A

 Table III. The Results of the Regression Analysis Models

Model Parameter Estimate Uncertainty t Statistic* Prob>|t|** Corr. Coef. R2

4 Parameter intercept, V-SCE 0.8646 0.0275 31.48 <0.0001 0.9950
slope, V/log (mol/L) 0.0719 0.0045 16.08 <0.0001
D intercept, V 0.0621 0.0232 2.67 0.0284
D slope, V/log (mol/L) 0.0102 0.0113 0.90 0.3931
3 Parameter intercept, V-SCE 0.8861 0.0136 65.21 <0.0001 0.9945
slope, V/log (mol/L) 0.0740 0.0038 19.36 <0.0001
D intercept, V 0.0794 0.0130 6.13 0.0002
2 Parameter intercept, V-SCE 0.8044 0.0058 138.60 <0.0001 0.9715
slope, V/log (mol/L) 0.0519 0.0028 18.46 <0.0001
*The value for the t statistic for a test of the hypothesis that the actual value is zero.
**The probability based on these data that the actual value of the parameter is zero.

both alloys in all solutions, but the extent of this

reduction was greater for the 5083 alloy.
Figure 6 shows the variation of the tensile properties
of the alloys with strain-rate in a one of the solutions.
Figure 6(a) shows that the tensile strength for both
alloys was constant with respect to strain-rate in air, but
decreased in the solution as the strain-rate was reduced.
Figure 6(b) shows that the strain-to-failure follows a
similar trend for both alloys in the solution. Both
quantities appear to be following a sigmoidal curve with
a high strain-rate plateau that approaches the values
observed in air and a low strain-rate plateau that is
determined by the environment. If this is the case, then
the 7075 alloy has yet not reached the low strain-rate
plateau at the lowest strain-rate and the 5083 alloy has
not reached the air value at the highest strain-rate.
Assuming a low strain-rate plateau that is above zero
and a high strain-rate plateau that equals the value
Fig. 4—Corrosion current density estimated from the polarization
curves as a function of the chloride ion activity.
observed in air, the tensile strength and failure strain
data for both alloys can be fitted to sigmoidal curves.
When this was done with commercial curve fitting
5083 alloy exhibited similar behavior, but the observed software, the midpoints for the transition in tensile
failure strains increased as the strain-rate decreased. strength and failure strain for alloy 7075 were both 2.25
orders of magnitude below those determined for alloy
5083. This software also estimates the uncertainty in the
C. Slow Strain-Rate Tensile Tests midpoint estimates, and the sum of the uncertainties in
Figure 5 shows the variation of the tensile properties the midpoint estimates for the alloys was 0.24 orders of
of the alloys with chloride ion activity. By examination magnitude for both tensile strength and failure strain.
of Figure 5(a), it can be seen that the observed tensile Figure 7 compares the fracture surfaces generated in
strength of both alloys increased as the chloride ion alloy 5083 when tested in solutions of low and high
concentration decreased until a maximum or plateau chloride ion activity. Examination of this figure shows
was reached at a chloride ion activity of 0.0055 mol/L that crack propagation was predominately intergranular
(wNaCl = 0.035 pct). At chloride ion activities below this in both solutions. In the lower concentration solution,
point, the ultimate tensile strength decreased slightly for there was less secondary intergranular cracking, and
both alloys, but this may be due to the normal scatter in there is evidence of ductile processes between regions of
these tests. For the 7075 alloy, this peak or plateau in intergranular fracture. For the higher concentration
tensile strength at low chloride ion concentrations was solution, there was no evidence of ductile processes, and
close to that observed in air, but in the 5083 alloy it was the frequency of secondary cracking increased. This
well below the air value. The time-to-failure data in secondary intergranular cracking is shown more clearly
Figure 5(b) follow a similar trend with a maximum in the metallographic cross sections of Figure 8. This
observed at the same chloride ion activity. The estimated figure shows two metallographic cross sections:
failure times for these alloys in air at this strain-rate are (a) through-the-fracture surface, and (b) through-the-
well beyond the upper limit of this plot: (205 ± 5) hour side surface of a specimen. In the cross section through
for 7075 and (365 ± 20) hour for 5083. Therefore, the the fracture surface, Figure 8(a), multiple secondary
failure times and strains were significantly reduced for cracks can be seen. In the second cross section,

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 44A, MARCH 2013—1357

 Table IV. Results of Slow Strain-Rate Tensile Tests in Air

Ultimate Tensile Strength

Strain-rate Time-to-Failure Strain-to-failure
Alloy s 1 (3 Pct)* MPa (7) ksi (1) h (0.02) (3 Pct)
5083 1.0E-04 306 44 0.32 0.115
7075 512 74 0.20 0.072
5083 1.0E-05 315 46 3.70 0.133
7075 519 75 2.14 0.077
5083 1.0E-06 318 46 40.57 0.146
7075 511 74 20.10 0.072
5083 1.0E-07 311 45 + +
7075 501 73 + +
7075 4.0E-08 511 74 + +
*The number in parentheses is the estimated uncertainty.
+ not recorded.

Fig. 5—The ultimate tensile strengths (a) and failure times (b) plotted as a function of the chloride ion activity. The uncertainty in failure time is
less that the size of the points.

Figure 8(b), a deep secondary crack is shown along with in the concentrated NaCl solutions were macroscopi-
evidence of crack branching, intergranular penetrations, cally smoother. In addition, the fracture surfaces appear
and pitting. to have a different luster near the periphery of the
Figure 9 compares the fracture surfaces generated in fracture surface from that found in the center. Exam-
alloy 7075 when tested in air and in a concentrated NaCl ination of these two regions at higher magnifications
solution. At lower magnification, Figure 9(a), the frac- found MVC in the central region similar to that found in
ture surface generated in air has a macroscopically air, Figure 9(d). The morphology of the peripheral
granular appearance indicative of intergranular fracture, regions differed from that as shown in Figure 9(e).
while at higher magnification, Figure 9(d), it is clear that Unfortunately, the corrosion deposits on these regions
the surface is covered with microvoids indicative of prohibited more detailed analysis. However, the metal-
ductile microvoid coalescence (MVC). Precipitation- lographic cross sections shown in Figure 10 clearly
hardened alloys with a precipitate-free zone (PFZ) reveal the intergranular nature of crack propagation and
along the grain boundaries will exhibit preferential secondary cracking. In addition, this figure shows
deformation in the PFZs, particularly at SSRs, creating secondary cracks that initiated from the sides of the
macroscopically granular, but microscopically ductile, specimen and from the bottom of pits.
fractures of this type. This type of fracture mode, where
ductile deformation occurs predominately in the PFZ
adjacent to the grain boundaries, is not uncommon in IV. DISCUSSION
precipitation-hardened 7000 series alloys during SSR
testing.[42–45] A typical fracture surface generated in the Chloride Ion Activity and Corrosion
NaCl solutions is shown at low magnification in As pointed out in the introduction, the chloride
Figure 9(c). Comparison of this figure to that in ion cannot promote oxidation by accepting electrons
Figure 9(a) indicates that the fracture surface generated from metallic elements. It was also pointed out that a

1358—VOLUME 44A, MARCH 2013 METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 6—The influence of strain rate on the tensile strength (a) and strain-to-failure (b) in solution with a chloride ion activity of 0.41 mol/L
(wNaCl = 3.5 %).

7 1 3
Fig. 7—SEM fractographs of alloy 5083 tested at a strain rate of 10 s in solutions with a chloride ion activity of 5.5 x 10 mol/L
(wNaCl = 0.035 %) and 1.22 mol/L (wNaCl = 10 %).

first-order approximation indicates that the free energy mation is valid for dilute solutions, but that in concen-
change (DG) driving the corrosion reactions should not trated solutions the chloride ion decreases the solubility
change with chloride concentration. Examination of the of oxygen and the activity of water. Lowering the
activity of water in Table II shows that this approxi- activity of water will reduce the magnitude of DG for the

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 44A, MARCH 2013—1359

 7 1
Fig. 8—Optical micrographs of a section normal to the fracture surface of alloy 5083 tested at a strain rate of 10 s in a solution with a chlo-
ride ion activity of 3.17 mol/L (wNaCl = 20 %).

8 1 7 1
Fig. 9—Fracture surfaces for alloy 7075 tested at a strain rate = 4 9 10 s in air and at a strain rate of 1 9 10 s in a solution with a
chloride ion activity of 3.17 mol/L (wNaCl = 20 %).

reaction between water and aluminum. The activity of the metal. Therefore, Table II shows that, at high
dissolved oxygen is fixed by the assumption of equilib- concentrations (>0.5 mol/L), the chloride ion actually
rium with air, but reducing its concentration and lowers the thermodynamic forces driving corrosion
increasing the viscosity of the solution will reduce that reactions and reduces the limiting rates of mass trans-
rate that mass transport allows it to reach the surface of port. However, when the driving force stays larger than

1360—VOLUME 44A, MARCH 2013 METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 10—Optical micrograph of a plane normal to the fracture surface of alloy 7075-T6 tested at a chloride ion activity of 3.17 mol/L
(wNaCl = 20 %) and a strain rate of 10 7 s 1.

that required to grow the passive film, the corrosion rate but these plots are not identical. Relating these measures
of passivated metals is determined by the stability of of susceptibility to actual, or relative, crack propagation
their passivating films, not the driving force.[1,46] thresholds or rates requires additional research beyond
Figures 2 and 3 show that the corrosion potentials the scope of the current study. Accordingly, interpreta-
steadily decreased as the chloride ion concentration tion of SSR results in terms of the relative rates of crack
increased, but there was an abrupt change in the initiation and propagation without additional measures
intercept of the trend line at a chloride activity of about to detect or quantify these stages is speculative, but it is
0.01 mol/L. When the environment has sufficient oxi- clear that these stages will influence the measured tensile
dizing potential to cause pitting, the corrosion potential strengths and failure times differently.
will only be slightly above the critical potential for the Figure 5 shows the influence of chloride ion activity
nucleation of pits. This is because anodic currents on the SCC susceptibility of the alloys as measured in
increase very rapidly at potentials greater than this. As it these tests. In Figure 5(b), both alloys exhibited a
is well known that pitting potentials vary linearly with maximum failure time at a chloride activity of
log (Cl ),[7,47,48] the linear trend for corrosion potentials 0.0055 mol/L (wNaCl = 0.035 pct), and reduced failure
at high chloride ion concentrations is attributed to the times at higher and lower activities. As even the maxima
dependence of the pitting potentials on chloride ion in these curves are well below the air values, the
concentration. However, as the chloride concentration is environment is assisting crack propagation in all cases.
reduced, the pitting potential will increase, until it The tensile strengths also exhibited maxima at this
exceeds the oxidizing potential of the environment. At activity, but the decrease at lower concentrations is
concentrations less than this critical concentration, the small compared with the uncertainty in these quantities.
corrosion potential will be the potential where the These low chloride concentration plateaus in the tensile
passive current is equal to the cathodic current. There- strength were close to the yield strength for both alloys.
fore, the break in the corrosion potential line is Reducing failure times, while the tensile strength is
attributed to this transition from the pitting current constant, implies that crack initiation is not changing as
and the passive current determining the corrosion propagation rates are increasing. As discussed above,
potential. As the data on solution properties in Table II lowering the chloride ion concentration in naturally
indicate that cathodic currents should be constant in this aerated solutions results in a transition from pitting
concentration range, the trend at low chloride ion corrosion to passive corrosion. As this transition occurs
concentrations is attributed to a change in the passive at the same chloride ion concentration as these maxima,
current density with chloride ion concentration. the peaks in these curves is attributed to a change in the
SCC crack initiation mechanism. At high chloride
concentrations, pits nucleate before cracks can initiate.
A. Chloride Ion Activity and Susceptibility to SCC
Then, SCC cracks nucleate from the acidified solution
While the basic concept of susceptibility to SCC is produced by hydrolysis reactions inside the pits. As the
understood, measurements for quantifying it are not well chloride activity increases, the rates of pit nucleation
developed. The in situ SSR tensile test was developed and growth increase, thereby reducing the observed
because continuously rupturing the passive film should tensile strengths. In the absence of pitting, crack
promote crack initiation and propagation. The aims were initiation has to occur on the surface of the specimen.
to reduce testing times, improve detection of susceptibil- It appears that this mechanism of crack initiation
ity, and identify critical crack tip strain-rates.[19] In the requires longer time, greater stress, or greater strain.
SSR test, susceptibility is usually quantified by the extent As the tensile strengths observed for both alloys in the
of reduction in ductility, strength, or both.[49,50] In this low chloride solutions are close to their yield strengths,
context, the plots in Figure 5 are graphical presentations it appears that the nonpitting crack initiation mecha-
of the influence of chloride ion activity on susceptibility, nism requires general yielding in these SSR tests.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 44A, MARCH 2013—1361

 Above the critical chloride activity for pitting, the presence of a strain-rate minimum is considered as
time-to-failure data for both alloys in Fig- evidence that a slip-dissolution mechanism is responsible
ure 5(b) decreased in a linear manner with log (Cl ). for crack propagation.[19,49,50,58] According to this mech-
This behavior indicates that the time-to-failure is deter- anism, a strain-rate minimum results from the competition
mined by a chemical reaction with the chloride ion. As between repassivation and blunting at the crack
these failure times include the influence of the chloride tip.[19,33,58,59,60] As hydrogen absorption and diffusion
ion on pitting, crack initiation, and propagation, it is should increase with reducing strain-rate, no minimum
difficult to conclude which of these defines this trend, should occur for hydrogen embrittlement mechanisms of
but if any one of these deviated significantly from linear SCC. While no strain-rate minima were observed in these
behavior, this linear trend would not be observed. Below experiments, the solution used contained sufficient chlo-
the critical chloride activity for pitting, a linear trend in ride and oxygen concentrations to cause pitting. As pit
time-to-failure is also observed, but with a positive propagation should not have a strain-rate dependence,
slope. While only a few points define this trend, it one cannot conclude anything about the fundamental
appears that failure times increase with chloride con- mechanism of cracking from the absence of a minimum in
centrations in this range. As pitting is not occurring in the strain-rate curve.
these environments, and the tensile strength data indi- While one cannot rule out a mechanism based on the
cate that initiation has not changed significantly, the sigmoidal shape of the strain-rate curves, it is interesting
results indicate that this trend is due to crack propaga- to note that the transition with strain-rate in alloy 7075
tion rates decreasing with increasing chloride ion con- appears to occur at much slower strain-rates. This
centration in this concentration range. There are two indicates that the rate-determining step in the cracking
potential explanations. First, increasing corrosion rates mechanism is proportionately slower for alloy 7075 than
may increase crack tip blunting in this concentration alloy 5083. If dissolution is responsible for cracking in
range. Second, increasing solution conductivity may alloy 5083, then these results indicate that the mecha-
increase anodic activity at the crack tip and reduce nisms for cracking in alloy 7075 has a rate-determining
hydrogen absorption. In any case, additional research step that is approximately two orders of magnitude
will be required to confirm this trend, determine the slower. Additional study will be required to identify the
stage of crack propagation responsible, and evaluate the rate-determining steps on each alloy, but these results
mechanism. are consistent with the hypothesis that a dissolution
As discussed above, if an ion reduces the effectiveness mechanism is responsible for cracking in low strength Al
of the passivating layer, then it can accelerate corrosion alloys and that hydrogen embrittlement causes cracking
without being an oxidizer. It can accomplish this in the high strength 7075 alloy.[14]
reduction in a variety of ways. First, the chloride ion
may form intermediate complexes that alter reaction
paths and activation energies in a manner similar to that
suggested by Foley.[51,52] However, recent density func- V. SUMMARY AND CONCLUSIONS
tional theory studies of postulated aluminum-chloride
complexes have determined that the chloride ion cannot The effect of chloride ion activity on the susceptibility
penetrate the shell of water molecules that surround of aluminum alloys AA 7075-T6 and AA 5083-H131 to
aluminum ions in aqueous solutions.[53,54] Another SCC in ambient solutions was investigated by conduct-
possibility is that the chloride ion is absorbed onto the ing SSR tests at strain-rates from 10 8 to 10 4 s 1 in
passive film and displaces oxygen atoms, thus increasing aqueous solutions of varying NaCl concentrations from
the solubility of the passive film. However, the literature 0 to 20 pct mass fraction. This study found that both
on the influence of chlorides on the dissolution of alloys exhibited reduced strengths and failure strains in
aluminum oxide minerals indicates that the solubility of the solutions compared with laboratory air. The extent
these phases in the presence of chlorides is much lower of these reductions was much greater for the 5083 alloy
than that in the presence of other species that do not over the conditions examined. At low chloride activities,
cause SCC.[55,56] Therefore, it is concluded that the tensile strengths varied little, but at concentrations
interaction responsible for SCC is unique to the condi- greater than a critical value, tensile strengths decreased
tions present on the surface of passivated metals where a with increasing chloride concentration. The failure time
thin insulating, or semiconducting, film isolates a results exhibited a maximum with respect to the chloride
conductor from an electrolyte with steep potential ion concentration that is attributed to the transition
gradients through the film. from pitting corrosion to passive corrosion as the
chloride ion concentration changes. The strength and
strain to failure for both alloys declined with decreasing
B. Strain-Rate and SCC Susceptibility strain-rates in a sigmoidal manner. No evidence of a
Figure 6 illustrates the influence of strain-rate on the minimum indicative of a critical crack tip strain-rate was
ultimate tensile strength and strain-to-failure of specimens found. As these experiments were conducted under
tested in air and in a NaCl solution. This figure shows that ambient conditions with active pitting corrosion, the
these quantities were observed to continuously decrease absence of a minimum does not support, or refute, any
with decreasing strain-rate. A number of investigators mechanistic hypothesis. However, the strain-rate results
have reported finding minimums in these quantities at indicated that the rate-determining step for cracking in
strain-rates varying from 10 5 to 10 7 s 1.[19,33,57,58] The alloy 7075 is two orders of magnitude slower than that

1362—VOLUME 44A, MARCH 2013 METALLURGICAL AND MATERIALS TRANSACTIONS A

for alloy 5083. This is consistent with the hypothesis that 24. T.B. Cassagne, J. Kruger, and E.N. Pugh: in Environmentally
absorbed hydrogen causes cracking in alloy 7075, while Assisted Cracking: Science and Engineering, W.B. Lisagor, T.W.
Crooker, and B.N. Leis, eds., Vol. STP 1049, ASTM, Phila., PA,
cracking in 5083 is the result of a dissolution mechanism. 1989, p. 59–75.
25. T.B. Cassagne: PhD Dissertation, The Johns Hopkins University, 1988.
26. U. Bertocci, E.N. Pugh, and R.E. Ricker: in Environment-Induced
Cracking of Metals, R.P. Gangloff and M. B. Ives, eds., NACE
ACKNOWLEDGMENTS Intl., Houston, TX, 1990, pp. 273–86.
27. G.L. Bianchi and J.R. Galvele: Corro. Sci., 1993, vol. 34,
The authors would like to thank the Office of pp. 1411–22.
Naval Research for supporting this investigation at the 28. S.B. Farina, G.S. Duffo, and J.R. Galvele: Corro. Sci., 2005,
vol. 47, pp. 239–45.
Naval Air Systems Command (Reference No. 29. R.E. Ricker: J. Met., 1995, vol. 47, pp. 32–35.
N0001412AF00002). Special thanks are due to Drs. 30. R.E. Ricker: in Hydrogen Effects on Materials Behavior and Cor-
R. Williams and A. K. Vasudevan, and Mr. W. C. rosion Deformation Interactions, N.R. Moody, A.W. Thompson,
Nickerson for monitoring and providing technical R.E. Ricker, G.S. Was, and R.H. Jones, eds., TMS, Warrendale,
guidance to this study. PA, 2003, pp. 629–38.
31. R.E. Ricker: J. ASTM Int., 2008, vol. 5, p. 10.
32. P.W. Slattery, J. Smit, and E.N. Pugh; in Environmental Sensitive
Fracture: Evaluation and Comparison of Test Methods, S.W. Dean,
E.N. Pugh, and G.M. Ugiansky, eds., Vol. STP-821, ASTM,
REFERENCES Phila., PA, 1984, p. 399–411.
1. E. Deltombe, C. Vanleugenhaghe, and M. Pourbaix: in Atlas of 33. C.D. Kim and B.E. Wilde: in Stress Corrosion Cracking: The
Electrochemical Equilibria in Aqueous Solutions, M. Pourbaix, ed., Slow Strain-Rate Technique, G.M. Ugiansky and J.H. Payer, eds.,
NACE Intl., Houston, TX, 1974, pp. 168–76. Vol. STP 665, ASTM, Phila. PA, 1979, pp. 97–112.
2. R.E. Ricker and D.J. Duquette: Metall. Trans. A, 1988, vol. 19A, 34. R.E. Ricker: Properties of NaCl Solutions at 25 °C that Influence
pp. 1775–83. Corrosion Reactions and SCC, National Institute of Standards and
3. M.R. Stoudt, A.K. Vasudevan, and R.E. Ricker: in Corrosion Technology, 2012, p. 10.
Testing of Aluminum Alloys. V.S. Agarwala and G. Ugiansky, eds., 35. W.J. Hamer and Y.-C. Wu: J. Chem. Ref. Data, 1972, vol. 1,
Vol. STP-1134, ASTM, Phila. PA., San Francisco, CA, 1992, pp. 1047–99.
pp. 196–213. 36. C.-F. Pan: J. Chem. Eng. Data, 1981, vol. 26, pp. 183–84.
4. E. Ghali: Corrosion Resistance of Aluminum and Magnesium Al- 37. J.I. Partanen and P.O. Minkkinen: Acta Chem. Scand., 1993,
loys: Understanding, Performance, and Testing, John Wiley & vol. 47, pp. 768–76.
Sons, Inc., New York, 2010. 38. J.I. Partanen and A.K. Covington: J. Chem. Eng. Data, 2009,
5. H.P. Goddard, W.B. Jepson, M.R. Bothwell, and R.L. Kane: The vol. 54, pp. 208–19.
Corrosion of Light Metals, John Wiley & Sons, Inc., New York, 39. V.M.M. Lobo and J.L. Quaresma: Electrolyte Solutions: Litera-
1967. ture Data on Thermodynamic and Transport Properties, Published
6. J. Kruger: Int. Mater. Rev., 1988, vol. 33, pp. 113–30. by the author, Coimbra, Portugal, 1981, vol. 2.
7. R.T. Foley: Corrosion, 1986, vol. 42, pp. 277–88. 40. J.I. Partanen: J. Chem. Eng. Data, 2009, vol. 54, pp. 882–89.
8. D.D. Wagman, W.H. Evans, V.B. Parker, R.H. Schumm 41. W. Mendenhall and T. Sincich, Statistics for Engineering and the
I. Halow, S.M. Bailey, K.L. Churney, and R.L. Nuttall: J. Phys. Sciences, Dellen Publ. Co., San Francisco, 1992.
Chem. Ref. Data, 1982, vol. 11, pp. 1–392. 42. S.W. Ciraldi, Ph.D. Dissertation, Univ. of Illinois, 1980.
9. Z.A. Foroulis and M.J. Thubrikar: J. Electrochem. Soc., 1975, 43. S.W. Ciraldi, J.L. Nelson, R.A. Yeske, and E.N. Pugh: in
vol. 122, pp. 1296–1301. Hydrogen Effects in Metals, I.M.B.A.W. Thompson ed., TMS/
10. G.S. Frankel: J. Electrochem. Soc., 1998, vol. 145, pp. 2186–98. AIME, New York, 1981, p. 437.
11. N.J.H. Holroyd and D. Hardie: Corros. Sci., 1981, vol. 21, p. 129. 44. N.J.H. Holroyd and G.M. Scamans: in Environment-Sensitive
12. D. Hardie, N.J.H. Holroyd, and R.N. Parkins: Met. Sci., 1979, Fracture: Evaluation and Comparison of Test Methods, S.W. Dean,
vol. 13, pp. 603–10. E.N. Pugh, and G.M. Ugiansky, eds., Vol. STP-821, ASTM, Phil.
13. J.L. Searles, P.I. Gouma, and R.G. Buchheit: Metall. Mater. PA, 1984, pp. 202–41.
Trans. A, 2001, vol. 32A, pp. 2859–67. 45. M. Khobaib and C.T. Lynch: in Environment-Sensitive Fracture:
14. T.D. Burleigh: Corrosion, 1991, vol. 47, pp. 89–98. Evaluation and Comparison of Test Methods, S.W. Dean, E.N.
15. F.J. Bradshaw and C. Wheeler: Int. J. Fract. Mech., 1969, vol. 5, Pugh, and G.M. Ugiansky, eds., Vol. STP 821, ASTM, Phil. PA,
p. 255. 1984, pp. 242–55.
16. G.M. Scamans, R. Alani, and P.R. Swann: Corros. Sci., 1976, 46. H.H. Uhlig and R.W. Revie: Corrosion and Corrosion Control, 4th
vol. 16, pp. 443–59. ed., John Wiley & Sons, New York, 2008.
17. G.A. Young and J.R. Scully: Metall. Mater. Trans. A, 2002, 47. H. Bohni and H.H. Uhlig: J. Electrochem. Soc., 1969, vol. 116, p. 906.
vol. 33, pp. 101–15. 48. R.E. Ricker and D.J. Duquette: in Aluminum Lithium Alloys II,
18. N.J.H. Holroyd and G.M. Scamans: Metall. Mater. Trans. A, T.H. Sanders and E.A. Starke, eds., The Metallurgy Society of
2011, vol. 42A, pp. 3979–98. AIME, Warrendale, PA, 1984, pp. 581–96.
19. R.N. Parkins: in Stress Corrosion Cracking: The Slow Strain-Rate 49. R.H. Jones and R.E. Ricker: in Metals Handbook Ninth Edition
Technique, G.M. Ugiansky and J.H. Payer, eds., Vol. STP 665, Vol. 13: Corrosion, J.R. Davis, ed., ASM Intl., Materials Park,
ASTM, Phila. PA, 1979, pp. 5–25. OH, 1987, pp. 145–63.
20. J.H. Payer, W.E. Berry, and W.K. Boyd: in Stress Corrosion 50. R.H. Jones and R.E. Ricker: in Stress-Corrosion Cracking:
Cracking: The Slow Strain-Rate Technique, G.M. Ugiansky and Materials Performance and Evaluation, R.H. Jones, ed., ASM Intl.,
J.H. Payer, eds., Vol. STP 665, ASTM, Phila. PA, 1979, pp. 61–77. Materials Park, OH, 1992, p. 1-40.
21. R.N. Parkins: in Slow Strain-rate Testing for the Evaluation of 51. R.T. Foley: J. Electrochem. Soc., 1975, vol. 122, pp. 1493–94.
Environmentally Induced Cracking: Research and Engineering 52. R.T. Foley: I&EC product R&D, 1978, vol. 17, p. 14.
Applications, R.D. Kane, ed., Vol. ASTM STP 1210, American 53. S.A. Matwiyoff and W.E. Wageman: Inorg. Chem., 1970, vol. 9,
Society for Testing and Materials, Phil., PA, 1993, pp. 7–21. pp. 1031–36.
22. J.A. Beavers and G.H. Koch: in Slow Strain-rate Testing for the 54. X. Jin, W. Yang, Z. Qian, Y. Wang, and S. Bi: Dalton Trans.,
Evaluation of Environmentally Induced Cracking, R.D. Kane, ed., 2011, vol. 40, pp. 5052–58.
ASTM STP, Vol. 1210, American Society for Testing and Mate- 55. M. Dietzel and G. Bohme: Geochim. Cosmochim. Acta, 2005,
rials, Phil., PA, 1993, pp. 22–39. vol. 69, pp. 1199–211.
23. ASTM: in Annual Book of ASTM Standards, Vol. 03.02, ASTM 56. D.J. Wesolowski, D.A. Palmer, and P. Benezeth: Geochim. Cos-
Intl., West Conshohocken, PA, 2006. mochim. Acta, 2006, vol. 70, pp. 2140–42.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 44A, MARCH 2013—1363

57. H. Buhl, in Stress Corrosion Cracking-The Slow Strain-Rate Technique, G.M. Ugiansky and J.H. Payer, eds., Vol. STP 665,
Technique, G.M. Ugiansky and J.H. Payer, eds., Vol. STP-665, ASTM, Phila. PA, 1979, pp. 254–65.
ASTM, Phil. PA, 1979, pp. 333–46. 59. F.P. Ford and P.L. Andresen, in Corrosion Mechanisms in Theory
58. G.M. Ugiansky, C.E. Johnson, D.S. Thompson, and E.H. and Practice, P. Marcus and J. Oudar, eds., Marcel Dekker, Inc.,
Gillespie: in Stress Corrosion Cracking: The Slow Strain-Rate New York, 1995, pp. 501–46.

1364—VOLUME 44A, MARCH 2013 METALLURGICAL AND MATERIALS TRANSACTIONS A

View publication stats