Tests for Stress-Corrosion

Stress-corrosion cracking is a time-dependent process SCC susceptibility tests

To determine the susceptibility of alloys to SCC,
in which a metallurgically susceptible material several types of tests are available. If the objective
fractures prematurely due to the synergistic is to predict service behavior or to screen alloys for
interaction of a corrosive environment and sustained service in a specific environment, it is often neces-
sary to find SCC information in a relatively short
tensile stress at the metal surface. The tensile stress period of time. This requires acceleration of testing
may be residual stress resulting from heat treatment by increasing the severity of the environment or
or fabrication of the metal, may be developed by the critical test parameters. Testing can be acceler-
ated by raising the temperature or the concentra-
external loading, or may be a combination of these tion of corrosive species, and by electrochemical
conditions. stimulation. Parameters that can be changed to
shorten the testing time include the application of
higher stresses, continuous straining, and pre-
Gerhardus H. Koch* cracking, which allows bypassing of the crack-nu-
CC Technologies Inc. cleation phase of the SCC sequence.
Dublin, Ohio Stress-corrosion specimens can be divided into
two categories: smooth and precracked or notched.
Further distinctions can be made in the loading

S
tress-corrosion cracking (SCC) is a mode, such as constant deflection, constant load,
cracking phenomenon caused by the con- and constant extension or strain rate.
joint action of a tensile stress and the pres- During the production of wrought alloys, the
ence of a specific corrosive environment. metal is forced predominantly in one direction, so
For such cracks to develop on a structure, three that the grains are elongated in the direction of
conditions must be met simultaneously. A spe- flow. Because it is important to relate the appli-
cific crack-promoting environment must be pre- cation of stress and the grain flow direction,
sent, the metallurgy of the material must be sus- two conventions have been chosen to relate the
ceptible to SCC, and the tensile stresses must be parameters.
above some threshold value. In one system, which is primarily for smooth
Stress-corrosion cracking can result in cata- specimens, the three stressing directions are desig-
strophic failure, often without any prior warning. nated by indicating the direction of the stress,
Many different mechanisms for SCC have been pro- namely longitudinal (L), long-transverse (LT), trans-
posed, but in general, these mechanisms can be di- verse (T), and short transverse (ST).
vided into two general groups: anodic dissolution A second system, which is particularly useful for
mechanisms and cathodic mechanisms. The para- precracked specimens, indicates both the cracking
meters that control SCC can be divided into mate- plane and the direction of crack propagation. Three
rials, environmental, and mechanical parameters. letters (L, T, and W) indicate three perpendicular
Stress-corrosion cracking failures resemble brittle directions: L for the longitudinal direction, T for the
fracture. This means that typically little, if any, in- thickness direction, and W for the width direction.
dication of metal ductility is visible at the origin of The crack plane is indicated by the direction normal
fracture. The cracking actually is a form of
subcritical flaw growth, either intergranular of
of l
on g or
o n l Short ina ti na
l
c
or transgranular, depending on the partic- ecti ing ina transverse it ud ire din udi
ir l d D tru t
ular combination of microstructure, envi- D rol itu ng ng
i
Lo ex
Short ong Lo
ronment, and strain rate. However, it is more transverse
L
difficult to design against environmentally Short
assisted cracking than against fracture. Rea- Long transverse Long transverse
transverse
Long transverse
sons for this difficulty include composi- Sheet and plate Extruded and drawn tube
tional, mechanical, and metallurgical syn- Rolled and extruded rod, bar, and thin shapes
ergisms, and the consequent need to
l
consider a range of environmental variables, of na
ion or n al u di l
as well as their variations with time and t di t
c g
ire n gi na
D olli gitu on udi
their interactions with loading and metal- L it
l

r n
na

Lo ng
di

lurgical variables. This article overviews Lo

itu
ng

SCC susceptibility testing, and describes

Lo

constant extension testing and the most suit- Short

Transverse Short transverse Long transverse transverse
able types of specimens .
*Member of ASM International Fig. 1 — Grain orientations in standard wrought forms of alloys.

36 ADVANCED MATERIALS & PROCESSES/AUGUST 2001

Cracking
to the crack, and the crack propagation is indicated
by one of the directions L, T, or W. Figure 2 shows
the various fracture plane orientations.
Other parameters that play an important role
in SCC testing are surface condition and residual
stress. The nucleation of stress-corrosion cracks
strongly depends on initial surface reactions. The
surface condition of the test specimens, particu-
larly smooth specimens, has a significant effect on
the test results. Smooth test specimens are often
tested with a mechanically (machined or abraded)
or (electro)chemically treated surface. It is very
important to avoid or to remove machining
marks or scratches perpendicular to the loading
direction.
Constant extension testing
Smooth SCC specimens allow for the evaluation
of the total SCC life, which includes crack nucle-
ation and propagation. Specimens may be tested
under constant extension or strain, constant load,
and constant extension or strain rate. The selection
of a specific test method for SCC strongly depends
on the particular service application and the time
allowed for testing.
Constant-extension or constant-strain tests on
smooth specimens do not require elaborate fixtures.
Depending on the specific configuration of the test
articles, different types of constant-extension tests
are available. The most common types are bent-
beam, U-bend, C-ring, and tensile type specimens.
• Bent-beam specimens: The different types of bent-
beam specimens
L are illustrated in
H Fig. 3. These speci-
t
mens are for sheet,
plate, and flat-ex-
(a) truded materials, or
wires and extru-
H sions with a circular
y t
cross section. Spec-
imens may be bent
(b)
in several ways, de-
H pending on dimen-
A t
y sions. They may be
stressed by bending
in a stressing device
(c)
while restraining
H the ends. During
h Weld stress-corrosion
s testing, both spec-
(d) t imen and stressing
devices are exposed
Fig. 3 — Schematic specimen and
holder configurations for bent-beam to the test environ-
specimens. (a) Two-point loaded spec- ment. The most
imen. (b) Three-point loaded spec- simple loading ar-
imen. (c) Four-point loaded specimen. rangement is the
(d) Welded double-beam specimen. two-point loaded
ADVANCED MATERIALS & PROCESSES/AUGUST 2001
C
L-T S R-C
T C-R R
T-L L R-L LT-S
L-R
L-TS
L-C
S-T C-L
S-L
T-S L-S TS-L
L
Rectangular section Cylindrical sections Non-primary
Fig. 2 — Fracture plane identification. L, direction of grain flow; T, transverse
grain direction; S, short transverse grain direction; C, chord of cylindrical cross section;
R, radius of cylindrical cross section; first letter, normal to the fracture plane; second
letter, direction of crack propagation in fracture plane.
bent-beam, which is suitable only for relatively thin
sheet or wire material. The elastic stress at the mid-
point of the specimen can be estimated:
L = (ktE/σ) sin–1 (H/ktE)
where L is the specimen length, σ is the maximum
stress, E is the elastic modulus, H is the length of
holder, t is the specimen thickness, and k is the em-
pirical constant, 1.280.
Three-point bend specimen tests are commonly
chosen because of the ease of load application and
the ability to load the same rigs for different stresses.
The load is applied by turning a bolt in the rig, de-
flecting the specimen. The elastic stress at the mid-
point of the specimen is then calculated:
σ = 6Ety/H 2
where σ is the maximum tensile stress, E is the
elastic modulus, t is the specimen thickness, y is the
maximum deflection, and H is the length of the
holder.
However, this test has a number of disadvan-
tages. For example, dissimilar metal corrosion
and/or crevice corrosion can develop under the
bolt. Once the crack has formed, the stress condi-
tions change such that the outer layer of the spec-
imen is not subject to a tensile stress alone, but to a
complex combination of tensile and bending
stresses. The propagating crack will then deviate
from the centerline. Thus, the three-point bend test
should be considered only as a qualitative test to
determine the susceptibility to stress-corrosion
cracking.
• Four-point bend specimens provide a uniform
tensile stress over a relatively large area of the spec-
imen. With the four-point bend test, tensile stresses
can be maintained during the growth of the crack.
The elastic stress in the outer layer of the specimen
between the two inner supports can be calculated:
σ = 12Ety/(3H 2 - 4A2)
where s is the maximum tensile stress, E is the
elastic modulus, t is the specimen thickness, y is the
maximum deflection, H is the distance between
outer supports, and A is the distance between outer
and inner supports.
• U-bend specimens are prepared by bending a
strip 180 degrees around a mandrel with a pre-
determined radius. Figure 4 shows that speci-
mens may also be bent less than 180 degrees.
37
 (Standardized test
methods are de-
scribed in ASTM
G 30.) Because of
the ease of fabrica-
or tion, a large num-
ber of specimens
may be fabricated,
Insulator Insulator Nut and this test is
Bolt
therefore prefer-
red for qualitative
evaluation of the
susceptibility of al-
loys and heat treat-
ments to stress-
corrosion cracking.
Fig. 4 — Schematic two-stage stressing of a U-bend specimen A good approx-
imation of the
strain at the apex of the U-bend is:
ε = t/2R, when t < R
where t is the specimen thickness and R is the ra-
dius of the bend. Then, an appropriate value for
the maximum stress can be determined from the
stress-strain curve of the test material.
• C-ring specimens are preferred for determining
the susceptibility to stress-corrosion cracking of al-
loys in different product forms. This test is partic-
ularly useful for tubing, rod, and bar in the short-
transverse direction. The specimens are typically
bolt-loaded to a constant strain or constant load per
ASTM G 338. If the stresses in the outer layers of
the apex of the C-rings are in the elastic region, the
stresses can be accurately calculated:
Df = D - ∆
∆ = σ d2/4EtZ
where D is the outer diameter of the C-ring before
stressing, Df is the outer diameter of the stressed C-
ring, σ is the elastic stress, ∆ is the change of D at
the required stress, d is the mean diameter (D - t),
t is the wall thickness, E is the elastic modulus, and
Z is the correction factor for the curved beam.
The stress on C-ring specimens can be more ac-
curately determined by attaching circumference
and transverse strain gages to the stressed surface.
The circumferential elastic stresses (σc) and trans-
verse elastic stresses (σt) can be calculated by:
σc = E (1 –µ 2) × (εc +µ e t)
σ t = E (1 –µ 2 ) × (εt +µ e c)
where E is the elastic modulus, µ is Poisson’s ratio,
e c is the circumferential strain, and et is the trans-
verse strain.
Tensile specimens
For specific purposes such as alloy development,
a large number of stress-corrosion specimens need
to be evaluated. Tensile specimens may be suitable
for this purpose if those for measuring tensile prop-
erties in air are adapted to SCC, as discussed in
38 ADVANCED MATERIALS & PROCESSES/AUGUST 2001
ASTM G 49. When uniaxially loaded in tension, the
stress pattern is simple and uniform, and the mag-
nitude of the applied stress can be accurately mea-
sured. Specimens can be quantitatively stressed by
equipment that applies either a constant load, a con-
stant strain, or an increasing load or strain.
This type of test is one of the most versatile
methods of SCC testing, because of the flexibility
permitted in the type and size of the test specimen,
stressing procedures, and the range of stress levels.
It allows the simultaneous exposure of unstressed
specimens (no applied load) with stressed speci-
mens, and permits subsequent tension testing to
distinguish between the effects of true SCC and me-
chanical overload.
A wide range of specimen sizes would be ap-
propriate, depending primarily on the dimensions
of the product. Note that stress-corrosion test re-
sults can be significantly influenced by the cross
section of the test specimen. Although large spec-
imens may be more representative of most struc-
tures, they often cannot be prepared from the avail-
able product forms being evaluated. They also
present more difficulties in stressing and handling
in laboratory testing.
Therefore, smaller cross-sectional specimens are
also frequently selected. They have a greater sen-
sitivity to SCC initiation, usually yield test results
rapidly, and permit greater convenience in testing.
However, the smaller specimens are more difficult
to machine, and results are more likely to be influ-
enced by extraneous stress concentrations as a con-
sequence of nonaxial loading, corrosion pits, and
so on. Consequently, specimens less than about 10
mm (0.4 in.) in gage length and 3 mm (0.12 in.) in
diameter are not recommended, except when
testing wire specimens.
Tension specimens including machined notches
are suitable for studying SCC and hydrogen em-
brittlement. The presence of a notch induces a tri-
axial stress state at the root of the notch, in which
the actual stress will be greater by a concentration
factor that is dependent on the notch geometry. The
advantages of such specimens include the local-
ization of cracking to the notch region and acceler-
ation of failure. However, unless directly related to
practical service conditions, the results may not be
relevant.
Tension specimens can be subjected to a wide
range of stress levels associated with either elastic or
plastic strain. Because the stress system is intended
to be essentially uniaxial (except in the cast of
notched specimens), great care must be exercised
in the construction of stressing frames to prevent
or minimize bending or torsional stresses. ■
For more information: The information in this article was
extracted from the “Corrosion Testing” section of the ASM
Metals Handbook, Desk Edition. It is available in both book
and CD ROM. See the table of contents on the ASM Web site,
www. asminternationl.org. Click on “Shop ASM,” then click
on the picture of the book, then click on the blue underlined title
to view the table of contents. You can also order online.