Engineering Fracture Mechanics 292 (2023) 109657

Contents lists available at ScienceDirect

Engineering Fracture Mechanics

journal homepage: www.elsevier.com/locate/engfracmech

Growth characteristics of stress corrosion cracking in

high-strength 7075 aluminum alloy in sodium chloride solutions
Tomoyuki Fujii *, Daiki Ito , Yoshinobu Shimamura
Department of Mechanical Engineering, Shizuoka University, 3-5-1, Johoku, Naka-ku, Hamamatsu 432-8561, Japan

A R T I C L E I N F O A B S T R A C T

Keywords: Stress corrosion cracking occurs in alloys subjected to tension in specific corrosive environments.
Stress corrosion cracking Corrosion products that affect crack growth can be formed. This study aims to clarify their in­
Crack growth rate fluence on the crack growth of 7075-T651 aluminum alloy in NaCl solution and proposes a
Aluminum alloy
method to obtain the crack growth property. Crack growth tests reveal that corrosion products
Corrosion product
create crack bridging which hinders crack opening, and as a result, the crack growth is decel­
erated. Hence, to obtain the crack growth property in 7075-T651 in NaCl solution, it is effective to
use additives that prevent the formation of corrosion products.

1. Introduction

Aluminum alloys are lightweight structural metals that exhibit some preferred mechanical properties due to the alloying elements
and heat treatment. High-strength aluminum alloys, such as the 7000-series alloys strengthened by precipitation hardening, are often
used in space equipment, aircraft, and trains. These alloys are unsuitable for use in corrosive environments because they are sus­
ceptible to corrosion, which is due to the alloying elements, such as Cu [1], decreasing corrosion resistance, and the formation of
corrosion-sensitive grain boundary precipitates during heat treatment conducted to improve the mechanical properties [2]. Therefore,
aluminum alloys with high corrosion resistance, such as the 6000-series alloys, are preferentially used [3]. In the 6000-series alloys,
Mg and Si are added instead of Cu, and the alloys are heat treated to achieve precipitation strengthening. Hence, the 6000-series alloys
have a good balance between strength and corrosion resistance. However, the strength of 6000-series alloys is lower than that of 7000-
series alloys, and it is expected that the high-strength 7000-series alloys are used for weight reduction even in a corrosive environment.
The common modes of corrosion occurring in components made of aluminum alloys in corrosive environments are localized pitting
corrosion, intergranular corrosion, and stress corrosion cracking (SCC) [4]. SCC occurs on the surface of a component subjected to
tension whereas pitting and intergranular corrosion occur on the surface of a component, irrespective of the loading condition. Among
these, SCC, which is crack-like corrosion and difficult to detect via nondestructive inspection, remains unclear from the viewpoint of
structural integrity. Many studies on SCC have been conducted. Its behavior is generally categorized into crack initiation and crack
growth. Schwarzenböck et al. [5] conducted the SCC testing of 7085 aluminum alloy in humid air and found that crack initiation was
caused by hydrogen embrittlement, not anodic reaction. Fujii et al. [6] examined the SCC processes of 6061 aluminum alloy in NaCl
solution and found that corrosion pitting followed by crack initiation along the specific plane at the pit root where the stress/strain
concentrated. Euesden et al. [7] conducted an in-situ observation of SCC in 7xxx aluminum alloy in humid air and found that cracks
preferentially initiated on the surface-connected gas pores, sometimes combined with Mg-based intermetallic compounds. To ensure

* Corresponding author.
E-mail address: [email protected] (T. Fujii).

https://doi.org/10.1016/j.engfracmech.2023.109657
Received 6 June 2023; Received in revised form 25 September 2023; Accepted 27 September 2023
Available online 29 September 2023
0013-7944/© 2023 Elsevier Ltd. All rights reserved.
T. Fujii et al. Engineering Fracture Mechanics 292 (2023) 109657

the long-term use of machines and structures, the remaining service life of a component with a crack detected by nondestructive testing
is usually evaluated based on crack growth characteristics. This approach has already been adopted in various industries. In the nuclear
industry, components of the nuclear power plants are periodically inspected using nondestructive testing techniques (NDT) such as
ultrasonic inspection. When SCC is detected, the remaining life to final failure is predicted according to its fitness-for-service as­
sessments [8,9]. If a fatigue crack, not SCC, is found, the life is also predicted according to the assessments related to fatigue crack
growth. Further advanced approaches have been developed in the aerospace industry; a technique considering crack growth from
corrosion pits was developed to assess the structural integrity [10]. Hence, it is necessary to derive the crack growth characteristics of
the structural components of various equipment for defects which can be found by NDT. The characteristics of crack growth can be
expressed in terms of the relationship between the crack growth rate da/dt and stress intensity factor (SIF) K. The characteristics and
mechanisms of crack growth have been reviewed [11,12]. Lee et al. [13] reported that the mechanism governing crack growth in a
high-strength aluminum alloy in an aqueous environment depends on SIF, with anodic dissolution found for small SIF, and hydrogen
embrittlement found for large SIF. Zhang et al. [14] demonstrated that SCC of aluminum alloys in an aqueous solution had three
processes during testing, namely an incubation period, transient growth, and stable growth. The incubation period depends on the load
level and loading history. Ogawa et al. [15–17] conducted SCC testing in NaCl solution and in humid air and obtained the crack growth
rate as a function of SIF for several aluminum alloys. Euesden et al. [7] investigated the growth behavior of small cracks (hundreds of
micrometers long) and found that their growth characteristics were different from those of a long crack and that a small crack might
grow below the threshold SIF KISCC. Although SCC testing has been performed using a variety of techniques, the reported crack growth
characteristics vary even for a similar alloy and corrosive environment. There are two possible reasons for this: (i) crack growth is
affected by microstructures, such as precipitates and grains, which depend on heat treatment and deformation history, and (ii) crack
growth is affected by the testing conditions, such as specimen shape and size, corrosive solution type, and temperature. Therefore, it is
expected that the da/dt-K relation for an aluminum alloy can be obtained as a material property, irrespective of the testing method,
which can be used for the service life assessment of a component. There have been not many efforts to derive the da/dt-K relation as a
material property.
This study proposed a method to obtain the da/dt-K relation of aluminum alloys in NaCl solution as a material property, and the
factors influencing the da/dt-K relation were investigated. SCC testing of the high-strength 7075-T651 aluminum alloy was conducted
in two types of aqueous solutions, namely a 3.5 wt% NaCl solution and a 3.5 wt% NaCl solution with additives which inhibit the
formation of corrosion products. The crack growth rate as a function of the applied SIF was investigated. The length of a crack in each
specimen was estimated based on the change in elastic compliance during testing. The actual crack length was measured based on the
observations of the side surfaces. Crack growth behavior in both solutions was examined by a comparison between the estimated and
actual crack lengths and the factors that affect crack growth were investigated.

2. Experiments

2.1. Material and specimen

Table 1 shows the chemical composition of the utilized alloy, 7075-T651. Fig. 1(a) shows the microstructure of the alloy. The
crystallographic orientation was measured using an electron backscattered diffraction apparatus (TSL Solutions) installed in a field-
emission scanning electron microscope (JSM-7001F, JEOL Ltd.). Following the ASTM E399 standard, the orientation of an as-
received plate of the alloy was referred to as longitudinal direction (L), long transverse direction (T), and short transverse direction
(S). The grain size of the alloy was relatively large (hundreds of micrometers) and each grain was elongated along the longitudinal
direction, which corresponded to the rolling direction. Fig. 1(b) shows the nominal stress - nominal strain curve in the longitudinal
direction, and Table 2 shows the mechanical properties of the alloy used.
Compact tension (CT) specimens were machined from 26 mm thick plates of the alloy. Fig. 2 shows the dimensions of a specimen.
The direction of crack growth was set to be perpendicular to the longitudinal direction, i.e., the L-T specimen as defined in ASTM E399
was used. Also, the specimens were cut from the center of the plates in the thickness direction (short transverse direction) to remove
the work-affected layer near each surface of the plates. A groove on each side of a specimen was machined to achieve a stress state close
to the plane strain condition. The side surfaces including the grooves of each specimen were ground and mirror-polished with
sandpapers and diamond paste, respectively. Then, a pre-crack with a length of approximately 2 mm was introduced from the notch of
each specimen by fatigue testing which was conducted under the sinusoidal cyclic load with frequency of 10 Hz and stress ratio of 0.1
at room temperature in air. The maximum SIF was set to be less than at most 80 % of the applied SIF in SCC testing. Side surface
observation was conducted using a scanning electron microscope (SEM, VE-9800, Keyence Co.).
The value of the applied SIF was calculated [18] as
P (a)
K = √̅̅̅̅̅ F , (1)
BN W W

Table 1
Chemical composition of 7075-T651 aluminum alloy (mass%).
Si Fe Cu Mn Mg Cr Zn Ti Al

0.09 0.19 1.4 0.05 2.2 0.19 5.8 0.03 Bal.

2
T. Fujii et al. Engineering Fracture Mechanics 292 (2023) 109657

Fig. 1. (a) Crystallographic orientation map measured via EBSD and (b) stress–strain curve in the longitudinal direction before necking.

Table 2
Mechanical properties of 7075-T651 aluminum alloy.
√̅̅̅̅
Young modulus, E (GPa) 0.2 % proof stress, σ0.2 (MPa) Tensile strength, σB (MPa) Fracture toughness, KIC (MPa m)

70.5 499 555 29.7

where K, P, BN, a, and W are the stress intensity factor, applied load, net specimen thickness, crack length, and specimen width,
respectively. The function F is a correction factor for the CT specimen. It is given by
)
(2 + α)(0.886 + 4.46α − 13.32α2 + 14.72α3 − 5.6α4
F(α) = , (2)
(1 − α)3/2

where α is the crack length normalized by the width (α = a/W).

2.2. Corrosive environments

SCC testing was conducted in two types of solution consisting of NaCl. The first solution was a 3.5 wt% NaCl solution (hereafter
denoted as 3.5%NaCl), which represents a common corrosive environment (it has the salinity of seawater). As described in the next
section, it was confirmed that many small corrosion products formed on specimen surfaces, including crack faces, in the tests using
3.5%NaCl. Zhang et al. [14] reported the same phenomena and proposed the use of a 3.5 wt% NaCl with the additives, 0.6 M sodium
chloride, 0.02 M sodium dichromate, 0.07 M sodium acetate, and 0.389 M acetic acid to a pH of 4 (hereafter denoted as 3.5%NaCl with
additives), as a corrosive environment for the SCC testing of aluminum alloys. The two types of solution have different pH values. This
should not be a problem since the pH value of the solution near a crack tip becomes approximately 3.3–3.5 during testing, irrespective
of the pH value of the bulk solution [11,12]. It has been reported that the additives inhibit the formation of reaction products without

3
T. Fujii et al. Engineering Fracture Mechanics 292 (2023) 109657

Fig. 2. Dimensions of the CT specimen used (unit: mm).

significantly affecting the corrosion reaction [19].

One of the objectives of this study was to evaluate crack growth characteristics in a common environment, such as seawater pooled
in gaps of mechanical components. Hence, SCC tests were conducted in a tank open to air and filled with the solution (3.5%NaCl or
3.5%NaCl with additives).

2.3. Testing procedures and conditions

Some of the procedures used in this study are similar to those in previously reported SCC tests on stainless steel [20,21]. In this
section, the testing procedures are briefly described (see the previous papers for details).
SCC tests were performed under constant loading using a custom machine at room temperature. Note that the value of the applied

Fig. 3. Schematic of experimental setup.

4
T. Fujii et al. Engineering Fracture Mechanics 292 (2023) 109657

SIF increases as a crack grows during testing; such testing is thus classified as K-increasing testing. Fig. 3 shows the experimental setup
of crack growth tests in the corrosive solution. First, the specimen was fixed to the testing jigs of the machine and a tank was installed to
immerse the specimen in the solution. The two pins used to fix the CT specimen to the jigs were pre-coated with heat-shrinkable tubing
to insulate the specimen, jigs, and machine. The solution (3.5%NaCl or 3.5%NaCl with additives) of approximately 250 ml was poured
into the tank and the pre-crack was immersed in the solution. Then, tensile loading was applied and maintained at a given load for the
total testing period. After testing, the specimen was removed from the jigs and ultrasonically cleaned in pure water. The side surfaces
were observed with SEM and the crack length was measured. After that, the specimen was forcibly separated via fatigue and both
fracture surfaces of the tested specimen were observed with SEM. If the formation of corrosion products on the fracture surface and/or
side surface during testing was observed, elemental analysis was performed using an electron probe micro-analyzer (EPMA, JXA-
8530F, JEOL Ltd.).
To estimate crack length during testing, the elastic compliance was obtained from the relationship between load and displacement
based on periodic, partial unloading. The crack length can be estimated from the compliance under the assumption that a single crack
grows in a straight line without branching from the pre-crack tip of the specimen [22].
( )
a = W 1.00196 − 4.06319U + 11.242U 2 − 106.043U 3 + 464.335U 4 − 650.677U 5 , (3)

1
U= 1 , (4)
(Be EC)2 + 1

where E is the Young’s modulus, and Be and C are the effective specimen thickness and compliance, respectively, which are given by,

(B − BN )2 (5)
Be = B − ,
B

δ
C= , (6)
P
where B and δ are the thickness of the specimen without side grooves and the opening displacement, respectively. The crack length
based on the side surface observation was also defined as the length projected in a direction perpendicular to loading, which corre­
sponded to the crack length calculated using the change in elastic compliance. The crack lengths on both surfaces of each tested
specimen were measured as a projected distance between the tips of the pre-crack and stress corrosion crack; the average of the crack
lengths on both surfaces was considered to be the crack length of the specimen.
√̅̅̅̅
Table 3 shows the conditions of SCC testing. The initial SIFs, Kini, were set to eight values, from 8.5 to 30 MPa m. A constant load
corresponding to each initial SIF was applied. Note that these values satisfied the small-scale yielding condition required in ASTM
√̅̅̅̅
E399. For an initial SIF of 30 MPa m, rapid fracture occurred before the first measurement of crack length, which approximately
equaled KIC, and thus the crack growth rate could not be evaluated. In all experiments, partial unloading was done up to 25% of the
applied load every 1.8 ks (30 min), and the total testing period was set to be approximately 600 ks (7 days). One specimen was tested
for each condition. The influence of cyclic unloading/loading on crack growth was not investigated, even though periodic unloading
has been reported to slightly affect the da/dt-K relation in SCC testing [23]. Note that the test period was approximately 600 ks,
irrespective of testing conditions, and the number of cycles during the period was the same. The effect of cyclic loading will be dis­
cussed later.

3. Results and discussion

3.1. Crack growth behavior

Fig. 4(a) and (b) show the crack lengths as functions of immersion time obtained from tests conducted with an initial SIF of 25
√̅̅̅̅
MPa m in 3.5%NaCl and 3.5%NaCl with additives, respectively. The crack lengths obtained based on the change in elastic compliance
and side surface observation are plotted. The error bars and plots based on side surface observation denote crack lengths on both sides
and their average, respectively. For the test in 3.5%NaCl, unexpected phenomena were observed in the results based on the change in
elastic compliance. The following describes the investigation of crack growth using crack length estimated based on the change in
compliance. The crack continuously grew from the start of the test until approximately 300 ks (83 h). Then, the crack began to shorten

Table 3
Experimental conditions used for SCC testing.
Pre-crack length Approx. 2 mm from a notch root
√̅̅̅̅
Initial SIF, Kini (MPa m) 8.5*, 9*, 10, 12.5, 15, 20, 25, 30
Loading condition Constant load
Total testing period, tf (ks) 604.8 (7 days)
Unloading period (ks) 1.8 (30 min)
Unload (%) 25 % of applied load
* If da/dt < 10-12 m/s, no test was performed.

5
T. Fujii et al. Engineering Fracture Mechanics 292 (2023) 109657

√̅̅̅̅
Fig. 4. Crack length as a function of loading time obtained from tests conducted with Kini = 25 MPa m. (a) 3.5%NaCl and (b) 3.5%NaCl
with additives.

until 450 ks (125 h). The crack then began to grow again until the end of the test (600 ks). The final crack length was almost the same as
the pre-crack length. In contrast, the results based on the side observation indicate that the crack grew by approximately 170 µm. The
dotted line between the plots of the pre-crack length and the final crack length based on the side observation denotes the average crack
growth during testing. The trend mostly agrees with the early crack growth (0 to 300 ks) estimated via the change in elastic
compliance. This crack growth behavior was also seen at the other load levels in 3.5%NaCl. Hence, the crack length could not be
estimated via the change in elastic compliance, which is commonly used for measuring crack lengths in fracture mechanics testing such
as fatigue testing and fracture toughness testing. For the test in 3.5%NaCl with additives, the results based on the change in elastic
compliance indicate that the crack continuously grew from the start to the end of testing (0 ks to 600 ks). The crack lengths based on
the change in elastic compliance were roughly equal to those based on the side surface observation. This trend agrees with the results
obtained at the other load levels in 3.5%NaCl with additives.
Fig. 5(a) and (b) show the side surface around the crack and the fracture surface after the test conducted with an initial SIF of 25
√̅̅̅̅
MPa m in 3.5%NaCl, respectively. A crack grew from the pre-crack in a straight line. Regarding the surface morphology, small
corrosion pits and corrosion products were observed on the side and fracture surfaces. The fracture surfaces were ambiguous, and it
was difficult to clearly distinguish between fatigue pre-crack, SCC, and forced fracture surfaces in some areas. In particular, the
boundary between SCC and forced fracture was not clear. There were slightly fewer products on the side surface than on the fracture
surface, which may have been caused by the removal of the products on the side surface during ultrasonic cleaning after testing. The
crack grew by 170 µm, which means that it grew through a few grains in view of Fig. 1(a). Although the cracking behavior was unclear
in detail due to corrosion products, significant branching and/or kinking, which are characteristic of intergranular SCC, could not be
observed in Fig. 5. Therefore, it was presumed that transgranular SCC occurred in this experiment. Fig. 6 shows the results of the EPMA
measurement on the fracture surface. Measurement points A and B correspond to the corrosion product and the base metal, respec­
tively. The corrosion product consisted of Al with a small amount of Cl. Traces of Zn and Cu, which are major alloying elements in the
alloy, were also detected. Some reports [24,25] have demonstrated that dissolved Al was re-deposited as oxides/hydroxides, which
was different from the present result. There was the possibility that the elements of the base metal, not the corrosion products, were

6
T. Fujii et al. Engineering Fracture Mechanics 292 (2023) 109657

√̅̅̅̅
Fig. 5. Fracture morphology obtained from test conducted with Kini = 25 MPa m in 3.5%NaCl. (a) side surface around SCC and (b) fracture surface.

detected in this experiment because the measurement depth of the EPMA used was considered to be relatively deep, approximately 1
µm. Hence, the result of this study has been questionable. A higher resolution analyzer is needed to identify corrosion products, and
further experiments should be needed. Since the main objective of this study is to clarify the crack growth characteristics of the Al
alloy, this is a subject for future work.
Fig. 7(a) and (b) show the side surface around the crack and the fracture surface after the test conducted with an initial SIF of 25
√̅̅̅̅
MPa m in 3.5%NaCl with additives, respectively. The crack grew from the pre-crack in a straight line. The corrosion was almost the
same as that in the test in 3.5%NaCl, shown in Fig. 5, except for the lack of corrosion products. The fracture surfaces were also
ambiguous, and it was difficult to distinguish between fatigue pre-crack, SCC, and forced fracture surfaces in some areas. A comparison

7
T. Fujii et al. Engineering Fracture Mechanics 292 (2023) 109657

Fig. 6. Elemental analysis of corrosion product and base metal of the alloy on fracture surface.

of the crack growth behavior in 3.5%NaCl with that in 3.5%NaCl with additives, shown in Fig. 4, suggests that the corrosion products
play a significant role in the change in elastic compliance of a growing crack. To investigate the phenomenon of crack shortening
shown in Fig. 4(a), Fig. 8 shows enlarged images of the fracture surface of the specimen tested in 3.5%NaCl. Fig. 8(a) and (b) show the
two sides of the crack surface. The corrosion products, which are indicated by arrows in the figure, seem to be fractured because their

√̅̅̅̅
Fig. 7. Fracture morphology obtained from test conducted with Kini = 25 MPa m in 3.5%NaCl with additives. (a) Side surface around SCC and (b)
fracture surface.

8
T. Fujii et al. Engineering Fracture Mechanics 292 (2023) 109657

fracture morphology on the two sides matches. On the other hand, the other products near the fractured product were in different
locations on the two sides; these products were thus isolated. That is, the faces of the crack were connected by the growth of corrosion
products. The connected products play a role in crack bridging during testing. Crack bridging inhibited the opening of the growing
crack because the traction acted on its faces to close it, and as a result, the elastic compliance decreased. This mechanism explains the
results shown in Fig. 4(a). When the crack continuously grew at the early stage from 0 ks to 300 ks, corrosion products formed on all
specimen surfaces, but the faces of the pre-crack and the stress corrosion crack were not connected. Hence, the compliance corre­
sponded to the crack length, which could thus be correctly estimated from the change in elastic compliance. At approximately 300 ks,
the crack faces were connected due to the formation and growth of corrosion products and crack bridging occurred, resulting in a
decrease in elastic compliance. The formation and growth of corrosion products continued, and the crack bridging became tighter due
to further corrosion product growth. Hence, the elastic compliance continued to decrease, leading to an underestimation of the crack
length. At approximately 450 ks, the crack bridges may have been partially fractured due to opening loading and the elastic
compliance increased again, resulting in the crack seeming to begin to grow again. Since the crack faces consisted of a mixture of
unfractured and fractured bridges of corrosion products, the crack length was underestimated. Note that it is unknown whether the
separation of the corrosion product shown in Fig. 8 was caused by crack growth during SCC testing or by fatigue after the test.
Crack bridging also occurred at other loading levels. Fig. 9 shows a diagram of the relation between the period at which the
compliance began to decrease and the initial SIF. The period tended to decrease with decreasing applied SIF, indicating that a lower
applied SIF leads to an earlier change in elastic compliance, which corresponds to the earlier formation of crack bridging. This is
explained by the relationship between crack opening displacement (COD) and corrosion product growth; when the applied SIF is low,
the crack faces are easily connected by the formation of small corrosion products because COD is small. On the other hand, when the
applied SIF is high, the value of COD becomes large and the formation of crack bridging requires large corrosion products, which in
turn require a long time to form. Therefore, the crack length estimation based on the change in elastic compliance is not effective after
crack bridging formation. The crack length of the specimen tested in 3.5%NaCl could be estimated before crack bridging formation
based on the change in elastic compliance, whereas that of the specimen tested in 3.5%NaCl with additives could be always estimated
using this technique.

3.2. Growth characterization for SCC

As mentioned in Sec. 3.1, the crack length estimated via the change in elastic compliance was incorrect. To calculate the crack
growth rate, the crack lengths based on the side surface observation were used:
da af − ai
= , (7)
dt tf

where af, ai, and tf are the crack length after testing, the initial crack length (pre-crack length), and the total testing time, respectively.
Although the crack growth rate for each test conducted in 3.5%NaCl with additives could be calculated based on the change in elastic
compliance, the rate was also calculated using Eq. (7), as done for the tests in 3.5%NaCl, to compare the rates with those for 3.5%NaCl.
Note that the rate obtained using Eq. (7) denotes the average crack growth rate during the testing period. This is reasonable because the
crack tended to uniformly grow during testing as shown in Fig. 4. To calculate the applied SIF, the average crack length am is used:

Fig. 8. Enlarged images of crack faces in the specimen tested in 3.5%NaCl. (a) fracture surface and (b) fracture surface mirror opposite to that in (a).

9
T. Fujii et al. Engineering Fracture Mechanics 292 (2023) 109657

Fig. 9. Relationship between period at which the compliance began to decrease and initial SIF. The dotted curve was fitted using a polynomial to
discuss its trend.

ai +af
am = . (8)
2
Fig. 10 shows the crack growth rate as a function of the applied SIF for crack growth in 3.5%NaCl and 3.5%NaCl with additives. As
√̅̅̅̅
mentioned in Sec. 2.3, when the applied SIF was 30 MPa m, the specimens fractured before the end of the testing period due to rapid
crack growth, irrespective of the corrosion solution types; the crack growth rates for these specimens are described using plots and
arrows. The crack growth rate increased with increasing applied SIF, irrespective of the solution types. Considering the influence of the
solution, the growth rate was lower in the test in 3.5%NaCl than in 3.5%NaCl with additives within the full range of the applied SIF.
The threshold SIF, KISCC, which was defined as the condition where the rate became slower than 10-12 m/s, was approximately 10
√̅̅̅̅ √̅̅̅̅
MPa m in 3.5%NaCl and approximately 8.5 MPa m in 3.5%NaCl with additives. According to refs. [14] and [15], the additives play a
small role in the corrosion action related to localized corrosion. It is thus reasonable to consider that the difference in crack growth was
due to the formation of corrosion products and crack bridging.
Fig. 11 shows schematic diagrams of the crack growth processes during testing in NaCl solution. Fig. 11(a) shows the initial stage.
Fig. 11(b) shows the early stage of testing, where corrosion pitting occurs on all surfaces, including crack faces, and a stress corrosion

Fig. 10. Crack growth rate as a function of applied SIF for two types of solution.

10
T. Fujii et al. Engineering Fracture Mechanics 292 (2023) 109657

crack grows from a pre-crack. This might result in an increase in the Al concentration of the solution near the surfaces. If the Al
concentration exceeds its saturation level, corrosion products can form on the surfaces, including the faces of the growing crack. Such
localized corrosion continues to occur and the corrosion products continue to form and grow. Fig. 11(c) shows that if a corrosion
product growing on the crack face comes into contact with the opposite crack face or a corrosion product on the opposite crack face, the
crack faces will connect via crack bridging. Although the crack grows, the increase in COD is hindered by the crack bridging, which can
form anywhere on the crack faces. The actual SIF, Kact, at the crack tip is lower than the applied SIF, Kapp, because of the traction acting
on the crack faces to hinder crack opening:
Kact = Kapp − Kbridge (9)

where Kbridge is the SIF considering crack bridging. Fig. 11(d) shows that some crack bridges might fracture due to the tension caused by
the increase in COD associated with the crack growth. Hence, the crack growth is affected by the phenomena occurring on the crack
faces, namely the crack bridging due to the corrosion products and its fracture due to the increase in COD. As a result, the crack growth
in 3.5%NaCl is slower than that in 3.5%NaCl with additives; the da/dt-K curve for 3.5%NaCl shifted to the left by the value of Kbridge
compared with the actual da/dt-K curve, which was derived in the test where no corrosion products formed (i.e., the test conducted in
3.5%NaCl with additives). Note that the formation and/or fracture of crack bridging may change during testing, and quantitatively
determining Kbridge is not very meaningful. The condition under which the crack did not grow was also discussed: Fig. 10 implies that
the value of KISCC increased due to corrosion product formation, which is technically acceptable, but phenomenologically question­
able. This is because the value of KISCC actually represents the threshold for crack growth, but KISCC was defined using the crack growth
rate of 10-12 m/s in this study. If a stress corrosion crack does not grow, the Al concentration in the crack might not be sufficient
because only corrosion pitting occurs. If few corrosion products large enough to create crack bridging form, it is natural to consider
KISCC to be constant, irrespective of the solution type. The change in KISCC should be clarified for engineering applications. To discuss
the influence of corrosion products on KISCC, an experiment with a lower applied SIF should be conducted to examine a slower crack
growth rate; however, this might require an impractical testing period.

3.3. Discussion

It is important to investigate the crack growth inside the specimen by observing the fracture surfaces, especially the crack front, in
detail. However, it was impossible to clarify the region of SCC on the fracture surface of each tested specimen, as shown in Figs. 5(b)
and 7(b). The boundary between SCC and forced fracture by fatigue was unclear even though the fatigue testing was performed in air
after SCC testing. This problem may be caused by the experimental technique used; the specimen was tried to be thoroughly cleaned
after SCC testing, but the corrosive solution would be left inside the crack. Hence, SCC may grow due to the residual corrosive solution
during fatigue testing in air after SCC testing. Although the crack length is often defined by measuring several different positions of the
crack front on the fracture surface, this study could not define the crack length by such a technique and defined the length based on the
side surface observation. To solve this problem, beach marks can be introduced by varying the stress at the end of SCC testing to enable
detailed observation of the crack front. In addition, this study has another problem; cyclic unloading/loading was applied to measure
the crack length during SCC testing. Such cyclic unloading/loading commonly accelerates crack growth because this promotes the
dislocation emission, breakage of oxide film, and formation of new surfaces at a crack tip. Moreover, the formation of crack bridges
may be also affected because COD was changed with unloading/loading in the tests conducted in 3.5%NaCl. From Figs. 5(b) and 7(b),
no marks formed by cyclic unloading/loading, such as striation formed by fatigue, could be observed. This may be because the effect of
cyclic unloading/loading was not significant on crack growth or marks formed by cyclic unloading/loading were removed by

Fig. 11. Schematic diagrams of phenomena near crack tip in 7075-T651 aluminum alloy in NaCl solution. (a) to (d) show the growth processes of a
stress corrosion crack.

11
T. Fujii et al. Engineering Fracture Mechanics 292 (2023) 109657

corrosion. To investigate crack growth, SCC testing should be performed under constant loading. For such testing, the potential drop
method [26] is effective to measure the crack length during testing in a corrosive solution. Recently, the method has been able to detect
the initiation and growth behavior of multiple cracks [27,28] and crack closure during cyclic loading [29]. Application of the method
to crack growth with crack bridging would enable a detailed investigation of the crack growth characteristics. In the future, this
method will be used to evaluate the crack growth characteristics of Al alloys.
In this study, the SCC tests were conducted in aerated conditions because the corrosive solution was poured into the open tank, as
shown in Fig. 3. The influence of aeration/deaeration on corrosion in Al alloys has been studied [30,31], and dissolved oxygen tended
to promote the corrosion of Al, especially during the initial stage of immersion. Also, Searles et al. [32] and Krishnan et al. [33]
evaluated the SCC susceptibility of Al alloys in the NaCl solution in aerated and deaerated conditions from an electrochemical
viewpoint and demonstrated that the deaeration caused the drop in corrosion potential. Note that the test environment must be
determined according to the objective of SCC testing. For example, in the SCC tests cited in the introduction, the tests were conducted
in aerated conditions [13,15–17] and in deaerated conditions [14]. This study aimed to evaluate the crack growth characteristics in a
common environment, such as seawater pooled in gaps of components, and the SCC tests were conducted in the tank open to air.
Hence, the crack growth characteristics shown in Fig. 10 were valid only in aerated conditions. The crack growth characteristics of Al
alloys tested in aerated conditions may be different from those in deaerated conditions. Further studies of the test environment are also
needed as components made of Al alloys may be used under other conditions such as periodic exposure to seawater and storage of
seawater in sealed containers.
Previous studies have conducted experiments on the growth characteristics of a stress corrosion crack in Al alloys in corrosive
solutions consisting of NaCl, as mentioned in the introduction. This study found that when the pure NaCl solution was used for SCC
testing, corrosion products formed on crack faces, which should mechanically affect the crack growth characteristics; crack bridging
formed due to the connection of crack faces by corrosion products, and as a result, stresses decreased ahead of the crack tip, resulting in
a decreased crack growth rate. Corrosion products formed on all faces of a crack, including the pre-crack, and hence the influence of
corrosion products on crack growth is expected to vary depending on the immersion time, specimen size, pre-crack length, and stress
corrosion crack growth. The service life evaluation of a component must be conservative, and thus crack growth characteristics
without the influence of corrosion products are required even if corrosion products actually occur in a real component. Experiments
have been conducted to evaluate SCC in NaCl solutions to avoid the problem of corrosion products. Landkof and Gal-or [34] obtained
results similar to those in this study; they found that the crack growth in an aluminum alloy in NaCl solution with additives is faster
than that in pure NaCl solution. However, Sprowls et al. [16] reported that additives have little effect on the crack growth rate in an
aluminum alloy, and Connolly et al. [35] reported that the use of the additives led to slightly slower crack growth. These contradictory
results might be due to some differences in the formation of corrosion products and crack bridging. Ogawa et al. [15] proposed a
technique for gripping and immersing a specimen that avoids the influence of corrosion products on crack growth. A specimen is fixed
to the tensile machine with the crack mouth facing down so that the corrosion products that form on the crack faces will fall with the
NaCl solution which is forced into the crack face using a pump. This technique mechanically removes the corrosion products by forcing
the solution to flow and it does not introduce any additives (impurities) into a corrosive environment. However, the solution envi­
ronment in the crack will likely be totally different from the actual situation. This technique may be used for material selection but not
for lifetime evaluation. The crack growth rate reported by Ogawa et al. is much higher than those shown in Fig. 10 even though a
similar alloy was used. To assess the structural integrity of components, it is necessary to evaluate the crack growth rate as a material
property, that is independent of testing conditions. To obtain the relationship between crack growth rate and SIF as a material property
of 7075 in NaCl solution, it is recommended that testing be conducted in a solution consisting of NaCl and the additives that inhibit the
corrosion product formation. Note that it is known that SCC behavior is significantly scattered, depending on various conditions
[21,36–39]. In this study, the crack growth characteristics were evaluated using a single specimen for each condition, and their scatter
could not be evaluated. A number of experiments will be needed to obtain reliable crack growth characteristics which can be used for
structural integrity assessment and life prediction for SCC occurring in actual machines. As explained in the introduction, some re­
searchers have been conducting research on SCC growth characteristics, it is desirable to conduct crack growth tests not only on 7075
but also on various Al alloys and to accumulate crack growth data. Besides, the use of the additives is only valid for evaluating the crack
growth of 7075 in NaCl solution, as shown above. It is well known that the SCC mechanisms differ depending on the combination of
alloy and solution. If the proposed method is applied to other conditions, the obtained crack growth property may be different from
that under the condition to be clarified because the SCC mechanism may change due to the influence of additives. When evaluating a
crack growth property as a material property, it is necessary to clarify how the alloys, solutions, and additives affect the SCC mech­
anisms. Generalization of such influencing factors requires numerous mechanical and electrochemical tests, and further study is
needed.

4. Conclusions

This study evaluated the growth rate of a stress corrosion crack in the 7075-T651 aluminum alloy in corrosive solutions consisting
of NaCl. Constant load testing was conducted using CT specimens in the solutions. The factors that affected crack growth were
examined. The following results were obtained in this study:

1) The crack length could not be measured based on the change in elastic compliance in the tests conducted in the pure NaCl solution
but could be measured using this technique in the NaCl solution with additives that hinder the formation of corrosion products. This

12
T. Fujii et al. Engineering Fracture Mechanics 292 (2023) 109657

difference was caused by the corrosion products and crack bridging that formed in former solution. The crack bridging hindered
crack opening, resulting in a decrease in the elastic compliance of the specimen.
2) The crack growth rate increased with increasing applied SIF, irrespective of the solution type. The growth rate of a crack in the pure
NaCl solution was lower than that in the NaCl solution with additives. This was due to the formation of crack bridges via corrosion
products and the actual stress intensity being lower than the applied stress intensity. Hence, to characterize the stress corrosion
crack growth of 7075 in NaCl solution, it is recommended that testing be conducted in a solution consisting of NaCl and the ad­
ditives used in this study. It should be noted that this combination is only valid because the additives do not affect the SCC
mechanisms.

CRediT authorship contribution statement

Tomoyuki Fujii: Writing – original draft, Visualization, Validation, Supervision, Resources, Project administration, Investigation,
Funding acquisition, Formal analysis, Conceptualization. Daiki Ito: Investigation, Formal analysis. Yoshinobu Shimamura: Vali­
dation, Formal analysis.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to
influence the work reported in this paper.

Data availability

Data will be made available on request.

Acknowledgment

The present work was supported by the Suzuki Foundation, Japan.

References

[1] Hardwick D, Thompson A, Bernstein I. The effect of copper content and microstructure on the hydrogen embrittlement of Al-6Zn-2Mg alloys. Metall Mater Trans
A: Phys Metall Mater Sci 1983;14:2517–26.
[2] Puiggali M, Zielinski A, Olive JM, Renauld E, Desjardins D, Cid M. Effect of microstructure on stress corrosion cracking of an Al-Zn-Mg-Cu alloy. Corros Sci 1998;
40:805–19.
[3] Ahluwalia RK, Hua TQ, Peng J-K, Lasher S, McKenney K, Sinha J, et al. Technical assessment of cryo-compressed hydrogen storage tank systems for automotive
applications. Int J Hydrogen Energy 2010;35:4171–84.
[4] Foley RT. Localized corrosion of aluminum alloys - a review. Corrosion 1986;42:277–88.
[5] Schwarzenböck E, Wiehler L, Heidenblut T, Hack T, Engel C, Maier HJ. Crack initiation of an industrial 7XXX aluminum alloy in humid air analyzed via slow
strain rate testing and constant displacement testing. Mater Sci Engng A 2021;804:140776.
[6] Fujii T, Sawada T, Shimamura Y. Nucleation of stress corrosion cracking in aluminum alloy 6061 in sodium chloride solution: Mechanical and microstructural
aspects. J Alloy Compd 2023;938:168583.
[7] Euesden RT, Aboura Y, Garner AJ, Jailin T, Grant C, Barrett Z, et al. In-situ observation of environmentally assisted crack initiation and short crack growth
behaviour of new-generation 7xxx series alloys in humid air. Corros Sci 2023;216:111051.
[8] ASME. Boiler and Pressure Vessel Code Section XI 2003 Addenda. New York: ASME; 2003.
[9] JSME. Fitness-For-Service Code S NA1-2004, JSME, Tokyo; 2004.
[10] Liao M, Bellinger NC, Forsyth DS, Komorowski JP. A new probabilistic damage tolerance analysis tool and its application for corrosion risk assessment. In:
Proceedings of the 23rd ICAF Symposium; 2005. p. 241–52.
[11] Speidel MO. Stress corrosion cracking of aluminum alloys. Metall Trans A 1975;6:631–51.
[12] Holroyd NJH, Scamans GM. Stress Corrosion Cracking in Al-Zn-Mg-Cu Aluminum Alloys in Saline Environments. Metall Mater Trans A 2013;44:1230–53.
[13] Lee SM, Pyun SI, Chun YG. A critical evaluation of the stress-corrosion cracking mechanism in high-strength aluminum alloys. Metall Trans A 1991;22:2407–14.
[14] Zhang J, Kalnaus S, Behrooz M, Jiang Y. Effect of Loading History on Stress Corrosion Cracking of 7075–T651 Aluminum Alloy in Saline Aqueous Environment.
Metall Mater Trans A 2011;42:448–60.
[15] Ogawa T, Hasunuma S, Watanabe S, Sogawa N, Kanezaki T, Mano S, et al. Study on evaluation methods for stress corrosion cracking and fatigue crack growth of
aluminum alloys for hydrogen containers. J High Pressure Inst Japan 2016;54:13–24.
[16] Ogawa T, Hasunuma S, Kato S, Suzuki S, Nakamura Y, Mano S, et al. Crack growth characteristics of aluminum alloys dominated by the mechanisms of fatigue
and stress corrosion cracking. Mater Trans 2019;60:2346–52.
[17] Ogawa T, Hasunuma S, Shirawachi T, Fukada N. Effect of chemical composition and relative humidity on the humid gas stress corrosion cracking of aluminum
alloys. J High Pressure Inst Japan 2019;57:24–33.
[18] Murakami Y. Stress intensity factors handbook. Pergamon Press 1987:18–20.
[19] D.O. Sprowls, M.B. Shumaker, J.D. Walsh, J.W. Coursen, Evaluation of stress corrosion cracking susceptibility using fracture mechanics techniques, Pt. 1. Final
Report, Contract NAS 8-21487, Contract Report NASA CR-124469, NASA, Washington, DC, 1973.
[20] Fujii T, Tohgo K, Ishizuka N, Shimamura Y, Takanashi M, Itabashi Y, et al. Fracture mechanics study on stress corrosion cracking behavior under corrosive
environment. J Solid Mech Mater Eng 2013;7:341–56.
[21] Fujii T, Tohgo K, Kawamori M, Shimamura Y. Characterization of stress corrosion crack growth in austenitic stainless steel under variable loading in small- and
large-scale yielding conditions. Eng Fract Mech 2019;205:94–107.
[22] Saxena A, Hudak Jr SJ. Review and extension of compliance information for common crack growth specimens. Int J Fract 1978;14:453–68.
[23] Seifert HP, Ritter S. Stress corrosion cracking of low-alloy reactor pressure vessel steels under boiling water reactor conditions. J Nucl Mater 2008;372:114–31.
[24] Zhu Y, Sun K, Frankel GS. Intermetallic phases in aluminum alloys and their roles in localized corrosion. J Electrochem Soc 2018;165:C807–20.
[25] Kharitonov DS, Dobryden I, Sefer B, Ryl J, Wrzesińska A, Makarova IV, et al. Surface and corrosion properties of AA6063-T5 aluminum alloy in molybdate-
containing sodium chloride solutions. Corros Sci 2020;171:108658.

13
T. Fujii et al. Engineering Fracture Mechanics 292 (2023) 109657

[26] Bar J. Crack detection and crack length measurement with the DC potential drop method - Possibilities, challenges and new developments. Appl Sci 2020;10:
8559.
[27] Ghajar R. An alternative method for crack interaction in NDE of multiple cracks by means of potential drop technique. NDT&E Int 2004;37:539–44.
[28] Saton Y, Kawaguchi N, Ogura N, Kitayama T. Automated visualization of surface morphology of cracks by means of induced current potential drop technique.
NDT&E Int 2012;49:83–9.
[29] Andersson M, Persson C, Melin S. Experimental and numerical investigation of crack closure measurements with electrical potential drop technique. Int J
Fatigue 2006;28:1059–68.
[30] Katoh M. The influences of pretreatment and oxygen on corrosion of aluminium. J Soc Chem Industr Japan 1965;68:1636–42.
[31] Anwar UH. Electrochemical behavior of Al alloys in aerated and deaerated NaCl solution. Corros Prev Control 2001;48:151–9.
[32] Searles JL, Gouma PI, Buchheit RG. Stress corrosion cracking of sensitized AA5083(Al-4.5Mg-1.0Mn). Metall Mater Trans A 2001;32A:2259–867.
[33] Krishnan MA, Raja VS, Shukla S, Vaidya SM. Mitigating intergranular stress corrosion cracking in age-hardenable Al-Zn-Mg-Cu alloys. Metall Mater Trans A
2018;49A:2487–98.
[34] Landkof M, Gal-or L. Stress corrosion cracking of Al-Zn-Mg Alloy AA-7039. Corrosion 1980;36:241–6.
[35] Connolly BJ, Koul MG, Moran AL. A comparison study of stress corrosion crack growth rates for AA7XXX alloys as a function of bulk aqueous chloride
concentrations. Corrosion 2023;61:976–86.
[36] Gertsman VY, Bruemmer SM. Study of grain boundary character along intergranular stress corrosion crack paths in austenitic alloys. Acta Mater 2001;49:
1589–98.
[37] Andresen PL, Morra MM. Effects of positive and negative dK/da on SCC growth rates. In: Proceedings of the 12th International Conference on Environmental
Degradation of Materials in Nuclear Power System; 2005. p. 167–83.
[38] Vasudevan AK, Sadananda K. Role of internal stresses on the incubation times during stress corrosion cracking. Metall Mater Trans A 2011;42:396–404.
[39] Chen K, Wang J, Du D, Andresen PL, Zhang L. dK/da effects on the SCC growth rates of nickel base alloys in high temperature water. J Nucl Mater 2018;503:
13–21.

14