MSE 323
Materials Characterization Laboratory
Spring 2020
MSE 323 – Materials Characterization Laboratory
Lab 2: SEM Fractorgraphy
Lab Dates: 1/15, 1/22, 1/29
Report Due on 2/7
Aim of project:
•
To obtain an overall understanding and hands-on experience of the operation of FEI
FEGSEM Sirion 200, FEI Quanta 200, or FEI Apreo.
•
To obtain images of the fracture surfaces of different materials and from which, identify the
failure mechanisms.
Experimental Procedures:
•
You will be given 5 fractured samples, they are:
1.
2.
3.
4.
Mild/ Low carbon steel (A36)
Aluminum
YAG polycrystal
Copper
Each sample has been broken in some manner.
•
•
•
Observe and sketch the fracture surfaces of each sample with unaided eye and /or using an
optical microscope.
Obtain representative images (secondary and/or backscattered electron images) of the
fracture surfaces from each sample.
Determine from the observation the type of failure occurred in each sample.
Report:
Please refer to document on Blackboard Learn regarding
• Style of report; and
• Contents expected in the report (grading guidelines)
References:
You may find the following sources of information helpful:
• W.D. Callister, Materials Science and Engineering: An Introduction, 10th ed., Wiley, New
York, 2018. (MSE 201 book)
See chapter 8 Failure (or other similar introductory materials text that discusses fracture).
• ASM Handbook Volume 12 – Fractography, in the reference section in Owen Library.
• http://www.ammrf.org.au/myscope/ -- this link provide a lot of detail information about
SEM. You may also use the virtual SEM to practice your operation skills!
Due to the large size of this file,
it is broken into two parts for easier downloading.
ASM Handbook, Volume 12: Fractography
ASM Handbook Committee, p 12-71
DOI: 10.31399/asm.hb.v12.a0001831
Copyright © 1987 ASM International®
All rights reserved.
www.asminternational.org
Modes of Fracture
Victor Kerlins, McDonnell Douglas Astronautics Company
Austin Phillips, Metallurgical Consultant
METALS FAIL in many different ways and
for different reasons. Determining the cause of
failure is vital in preventing a recurrence. One
of the most important sources of information
relating to the cause of failure is the fracture
surface itself. A fracture surface is a detailed
record of the failure history of the part. It
contains evidence of loading history, environmental effects, and material quality. The principal technique used to analyze this evidence is
electron fractography. Fundamental to the application of this technique is an understanding
of how metals fracture and how the environment affects the fracture process.
This article is divided into three major sections. The section "Fracture Modes" describes
the basic fracture modes as well as some of the
mechanisms involved in the fracture process.
The section "Effect of Environment" discusses
how the environment affects metal behavior and
fracture appearance. The final section, "Discontinuities Leading to Fracture," discusses
material flaws where fracture can initiate.
Fracture Modes
Fracture in engineering alloys can occur by a
transgranular (through the grains) or an intergranular (along the grain boundaries) fracture
path. However, regardless of the fracture path,
there are essentially only four principal fracture
modes: dimple rupture, cleavage, fatigue, and
decohesive rupture. Each of these modes has a
characteristic fracture surface appearance and a
mechanism or mechanisms by which the fracture propagates.
In this section, the fracture surface characteristics and some of the mechanisms associated
with the fracture modes will be presented and
illustrated. Most of the mechanisms proposed
to explain the various fracture modes are often
based on dislocation interactions, involving
complex slip and crystallographic relationships.
The discussion of mechanisms in this section
will not include detailed dislocation models or
complex mathematical treatments, but will
present the mechanisms in more general terms
in order to impart a practical understanding as
well as an ability to identify the basic fracture
modes correctly.
Dimple Rupture
When overload is the principal cause of
fracture, most common structural alloys fail by
a process known as microvoid coalescence. The
microvoids nucleate at regions of localized
strain discontinuity, such as that associated
with second-phase particles, inclusions, grain
boundaries, and dislocation pile-ups. As the
strain in the material increases, the microvoids
grow, coalesce, and eventually form a continuous fracture surface (Fig. 1). This type of
fracture exhibits numerous cuplike depressions
that are the direct result of the microvoid
coalescence. The cuplike depressions are referred to as dimples, and the fracture mode is
known as dimple rupture.
The size of the dimples on a fracture surface
is governed by the number and distribution of
microvoids that are nucleated. When the nucleation sites are few and widely spaced, the
microvoids grow to a large size before coalescing and the result is a fracture surface that
contains large dimples. Small dimples are
formed when numerous nucleating sites are
activated and adjacent microvoids join (coalesce) before they have an opportunity to grow
to a larger size. Extremely small dimples are
often found in oxide dispersion strengthened
materials.
The distribution of the microvoid nucleation
sites can significantly influence the fracture
surface appearance. In some alloys, the
nonuniform distribution of nucleating particles
and the nucleation and growth of isolated microvoids early in the loading cycle produce a
fracture surface that exhibits various dimple
sizes (Fig. 2). When microvoids nucleate at the
grain boundaries (Fig. 3), intergranular dimple
rupture results.
Dimple shape is governed by the state of
stress within the material as the microvoids
form and coalesce. Fracture under conditions
of uniaxial tensile load (Fig. la) results in the
formation of essentially equiaxed dimples
bounded by a lip or rim (Fig. 3 and 4a).
Depending on the microstructure and plasticity
of the material, the dimples can exhibit a very
deep, conical shape (Fig. 4a) or can be quite
shallow (Fig. 4b). The formation of shallow
dimples may involve the joining of microvoids
by shear along slip bands (Ref 1).
Downloaded from https://dl.asminternational.org/handbooks/chapter-pdf/268655/a0001831.pdf
by Washington State University user
Fracture surfaces that result from tear (Mode
I) or shear (Modes II and III) loading conditions
(Fig. 5) exhibit elongated dimples (Ref 2, 3).
The characteristics of an elongated dimple are
that it is, as the name implies, elongated (one
axis of the dimple is longer than the other) and
that one end of the dimple is open; that is, the
dimple is not completely surrounded by a rim.
In the case of a tear fracture (Fig. 6a), the
elongated dimples on both fracture faces are
oriented in the same direction, and the closed
ends point to the fracture origin. This characteristic of the tear dimples can be used to
establish the fracture propagation direction (Ref
4) in thin sheet that ruptures by a full-slant
fracture (by combined Modes I and III), which
consists entirely of a shear lip and exhibits no
macroscopic fracture direction indicators, such
as chevron marks. A shear fracture, however,
exhibits elongated dimples that point in opposite directions on mating fracture faces (Fig.
6b). Examples of typical elongated dimples are
shown in Fig. 7.
It should be noted that the illustrations representing equiaxed and elongated dimple formation and orientation were deliberately kept
simple in order to convey the basic concepts of
the effect on dimple shape and orientation of
loading or plastic-flow directions in the immediate vicinity where the voids form, such as at
the crack tip. In reality, matching dimples on
mating fracture faces are seldom of the same
size or seldom show equivalent angular correspondence. Because actual fractures rarely occur by pure tension or shear, the various combinations of loading Modes I, 1I, and III, as
well as the constant change in orientation of the
local plane of fracture as the crack propagates,
result in asymmetrical straining of the mating
fracture surfaces.
Figure 8 shows the effect of such asymmetry
on dimple size. The surface (B) that is strained
after fracture exhibits longer dimples than its
mating half (A). When fracture occurs by a
combination of Modes I and II, examination of
the dimples on mating fracture surfaces can
reveal the local fracture direction (Ref 5). As
illustrated in Fig. 8, the fracture plane containing the longer dimples faces the region from
which the crack propagated, while the mating
fracture plane containing the shorter dimples
faces away from the region. With the different
M o d e s of F r a c t u r e / 13
, ~
Upper s u r f a c e / ~ ~
(rmax
o~
t~¢
surface
Section A-A
U
(a)
/Oval dimple
Oval dimple
Detail B
~ ' ~ C t "
~ -
Uuppere
Oma x
\ Lower
surface
Ibl
O'ma~
~ma×
Upper
su rface ~
Lower
surface
(c)
Fig.
1 Influenceof direction of maximumstress(Crmox)on the shape of dimples formed by microvoid coalescence.
(a) In tension, equiaxed dimples are formed on both fracture surfaces. (b) In shear, elongated dimples
point in opposite directions on matching fracture surfaces. (c) In tensile tearing, elongated dimples point toward
fracture origin on matching fracture surfaces.
combinations of Modes I, II, and III, there
could be as many as 14 variations of dimple
shape and orientation on mating fracture surfaces (Ref 5).
Metals that undergo considerable plastic deformation and develop large dimples frequently
contain deformation markings on the dimple
walls. These markings Occur when slip-planes
at the surface of the dimples are favorably
oriented to the major stress direction. The
continual straining of the free surfaces of the
dimples as the microvoids enlarge produces
slip-plane displacement at the surface of the
dimple (Ref 6), as shown in Fig. 9. When first
formed, the slip traces are sharp, well defined,
and form an interwoven pattern that is generally
referred to as serpentine glide (Fig. 10). As the
slip process proceeds, the initial sharp slip
traces become smooth, resulting in a surface
structure that is sometimes referred to as rippies.
Oval-shaped dimples are occasionally observed on the walls of large elongated dimples.
An oval dimple is formed when a smaller
subsurface void intersects the wall of a larger
void (dimple). The formation of oval dimples is
shown schematically in Fig. l(b) and 6(b).
Cleavage
Cleavage is a low-energy fracture that propagates along well-defined low-index crystallographic planes known as cleavage planes. The-
Downloaded from https://dl.asminternational.org/handbooks/chapter-pdf/268655/a0001831.pdf
by Washington State University user
oretically, a cleavage fracture should have
perfectly matching faces and should be completely flat and featureless. However, engineering alloys are polycrystalline and contain grain
and subgrain boundaries, inclusions, dislocations, and other imperfections that affect a
propagating cleavage fracture so that true, featureless cleavage is seldom observed. These
imperfections and changes in crystal lattice
orientation, such as possible mismatch of the
low-index planes across grain or subgrain
boundaries, produce distinct cleavage fracture
surface features, such as cleavage steps, river
patterns, feather markings, chevron (herringbone) patterns, and tongues (Ref 7).
As shown schematically in Fig. 11, cleavage
fractures frequently initiate on many parallel
cleavage planes. As the fracture advances,
however, the number of active planes decreases
by a joining process that forms progressively
higher cleavage steps. This network of cleavage
steps is known as a river pattern. Because the
branches of the river pattern join in the direction of crack propagation, these markings can
be used to establish the local fracture direction.
A tilt boundary exists when principal cleavage planes form a small angle with respect to
one another as a result of a slight rotation about
a common axis parallel to the intersection (Fig.
11 a). In the case of a tilt boundary, the cleavage
fracture path is virtually uninterrupted, and the
cleavage planes and steps propagate across the
boundary. However, when the principal cleavage planes are rotated about an axis perpendicular to the boundary, a twist boundary results
(Fig. 1 lb). Because of the significant misalignment of cleavage planes at the boundary, the
propagating fracture reinitiates at the boundary
as a series of parallel cleavage fractures connected by small (low) cleavage steps. As the
fracture propagates away from the boundary,
the numerous cleavage planes join, resulting in
fewer individual cleavage planes and higher
steps. Thus, when viewing a cleavage fracture
that propagates across a twist boundary, the
cleavage steps do not cross but initiate new
steps at the boundary (Fig. 1 lb). Most boundaries, rather than being simple tilt or twist, are
a combination of both types and are referred to
as tilt-twist boundaries. Cleavage fractures exhibiting twist and tilt boundaries are shown in
Fig. 12(a) and 13, respectively.
Feather markings are a fan-shaped array of
very fine cleavage steps on a large cleavage
facet (Fig. 14a). The apex of the fan points
back to the fracture origin. Large cleavage steps
are shown in Fig. 14(b).
Tongues are occasionally observed on cleavage fractures (Fig. 12b). They are formed when
a cleavage fracture deviates from the cleavage
plane and propagates a short distance along a
twin orientation (Ref 8).
Wallner lines (Fig. 15) constitute a distinct
cleavage pattern that is sometimes observed on
fractured surfaces of brittle nonmetallic materials or on brittle inclusions or intermetallic
compounds. This structure consists of two sets
14 / Modes
of Fracture
(a)
Fig.
I
5 lira
I
I
(bl
2 i~m
I
2
Examples of the dimple rupture mode of fracture. (a) Large and small dimples on the fracture surface of
a martempered type 234 tool steel saw disk. The extremely small dimples at top left are nucleated by
numerous, closely spaced particles. (D.-W. Huang, Fuxin Mining Institute, and C.R. Brooks, University of Tennessee).
(b) Large and small sulfide inclusions in steel that serve as void-nucleating sites. (R.D. Buchheit, Battelle Columbus
Laboratories)
I
I
5 p.m
Fig. 3
Intergronular dimple rupture in o steel specimen resulting from microvoid coalescence at
grain boundaries.
of parallel cleavage steps that often intersect to
produce a crisscross pattern. Wallner lines result from the interaction of a simultaneously
propagating crack front and an elastic shock
wave in the material (Ref 9).
Fatigue
I
(a)
Fig. 4
I
(b)
I
2 iLm
I
Different types of dimples formed during microvoid coalescence. (a) Conical equioxed dimples in o spring
steel specimen. (b) Shallow dimples in a maraging steel specimen
Mode I
Fig. 5
10 p.m
Mode II
Mode III
Fracture loading modes. Arrows show loading direction and relative motion of mating fracture surfaces.
Downloaded from https://dl.asminternational.org/handbooks/chapter-pdf/268655/a0001831.pdf
by Washington State University user
A fracture that is the result of repetitive or
cyclic loading is known as a fatigue fracture. A
fatigue fracture generally occurs in three stages:
it initiates during Stage I, propagates for most
of its length during Stage II, and proceeds to
catastrophic fracture during Stage Ili.
Fatigue crack initiation and growth during
Stage I occurs principally by slip-plane cracking due to repetitive reversals of the active slip
systems in the metal (Ref 10-14). Crack growth
is strongly influenced by microstructure and
mean stress (Ref 15), and as much as 90% of
the fatigue life may be consumed in initiating a
viable fatigue crack (Ref 16). The crack tends
to follow crystallographic planes, but changes
direction at discontinuities, such as grain
boundaries. At large plastic-strain amplitudes,
fatigue cracks may initiate at grain boundaries
(Ref 14). A typical Stage I fatigue fracture is
shown in Fig. 16. Stage I fatigue fracture
surfaces are faceted, often resemble cleavage,
and do not exhibit fatigue striations. Stage I
fatigue is normally observed on high-cycle
low-stress fractures and is frequently absent in
low-cycle high-stress fatigue.
The largest portion of a fatigue fracture
consists of Stage II crack growth, which generally occurs by transgranular fracture and is
more influenced by the magnitude of the alternating stress than by the mean stress or microstructure (Ref 15, 17, 18). Fatigue fractures
generated during Stage II fatigue usually exhibit crack-arrest marks known as fatigue striations (Fig. 17 to 22), which are a visual record
M o d e s of F r a c t u r e / 1 5
Principal
loading
,
direction 1
]
~
/ /
Lip of
dim le
-~
~
Nucleating
particle
Fracture
Fracture
direction
direction
,
I -c_ c c # c
\ Microvoid
*
d
I Top surface
I
I ~ C , "~- C ~- ~ c- I
.c~
--
ic cC.cCc
I? c.c -c
.......
\
°
"
[ ~ ' ~
%~C"~"Open"en
f - k,._,.--,,,._ I
~ c ~-c- ~~ - Nucleating particle~" II '~..~-"-
' f ~ ~ i : ~ : !
\
"
Bottom
surface
(a)
Oval -----'------
C .c'-c
C-'- C ~- C
cl
,C%_ -r,._. ,~c,.. Top surface
c..--~_ ~ _ ( t _ C
,,_.
(-- ~ - ( "
L'- C~
.,
Principal
loading
direction
~
..Q~ -~_~
~ ~.~)*~.~)'
..t)
' ~ ~ ~)
~) -~'~)
-~-~ Bottom
surface
(b)
Fig. 6
Formation of elongated dimples under tear and shear loading conditions. (a) Tear fracture. (b) Shear
fracture
of the position of the fatigue crack front during
crack propagation through the material.
There are basically two models that have
been proposed to explain Stage II striationforming fatigue propagation. One is based on
plastic blunting at the crack tip (Ref 11). This
model cannot account for the absence of striations when a metal is fatigue tested in vacuum
and does not adequately predict the peak-topeak and valley-to-valley matching of corresponding features on mating halves of the
fracture (Ref 8, 19-23).
The other model, which is based on slip at
the crack tip, accounts for conditions where
slip may not occur precisely at the crack tip
due to the presence of lattice or microstructural
imperfections (Ref 19-21). This model (Fig.
23) is more successful in explaining the
mechanism by which Stage II fatigue cracks
propagate. The concentration of stress at a
fatigue crack results in plastic deformation
(slip) being confined to a small region at the tip
of the crack while the remainder of the material
is subjected to elastic strain. As shown in Fig.
23(a), the crack opens on the rising-tension
portion of the load cycle by slip on alternating
slip planes. As slip proceeds, the crack tip
blunts, but is resharpened by partial slip
reversal during the declining-load portion of
the fatigue cycle. This results in a compressive
stress at the crack tip due to the relaxation of
the residual elastic tensile stresses induced in
the uncracked portion of the material during
the rising load cycle (Fig. 23b). The closing
crack does not reweld, because the new slip
surfaces created during the crack-opening
displacement are instantly oxidized (Ref 24),
which makes complete slip reversal unlikely.
The essential absence of striations on fatigue
fracture surfaces of metals tested in vacuum
tends to support the assumption that oxidation
reduces slip reversal during crack closure,
which results in the formation of striations (Ref
19, 25, 26). The lack of oxidation in hard
vacuum promotes a more complete slip reversal
(Ref 27), which results in a smooth and relatively featureless fatigue fracture surface. Some
fracture surfaces containing widely spaced fatigue striations exhibit slip traces on the leading
edges of the striation and relatively smooth
trailing edges, as predicted by the model (Fig.
23). Not all fatigue striations, however, exhibit
distinct slip traces, as suggested by Fig. 23,
which is a simplified representation of the
fatigue process.
As shown schematically in Fig. 24, the
profile of the fatigue fracture can also vary,
depending on the material and state of stress.
Materials that exhibit fairly well-developed striations display a sawtooth~type profile (Fig.
24a) with valley-to-valley or groove-to-groove
matching (Ref 23, 28). Low compressive
stresses at the crack tip favor the sawtooth
profile; however, high compressive stresses
promote the groove-type fatigue profile, as
shown in Fig. 24(c) (Ref 23, 28). Jagged,
poorly formed, distorted, and unevenly spaced
striations (Fig. 24b), sometimes termed quasistriations (Ref 23), show no symmetrical
matching profiles. Even distinct sawtooth and
groove-type fatigue surfaces may not show
symmetrical matching. The local microscopic
Downloaded from https://dl.asminternational.org/handbooks/chapter-pdf/268655/a0001831.pdf
by Washington State University user
plane of a fatigue crack often deviates from the
normal to the principal stress. Consequently,
one of the fracture surfaces will be deformed
more by repetitive cyclic slip than its matching
counterpart (Ref 29) (for an analogy, see Fig.
8). Thus, one fracture surface may show welldeveloped striations, while its counterpart exhibits shallow, poorly formed striations.
Under normal conditions, each striation is
the result of one load cycle and marks the
position of the fatigue crack front at the time the
striation was formed. However, when there is a
sudden decrease in the applied load, the crack
can temporarily stop propagating, and no striations are formed. The crack resumes propagation only after a certain number of cycles are
applied at the lower stress (Ref 4, 23, 30). This
phenomenon of crack arrest is believed to be
due to the presence of a residual compressivestress field within the crack tip plastic zone
produced after the last high-stress fatigue cycle
(Ref 23, 30).
Fatigue crack propagation and therefore striation spacing can be affected by a number of
variables, such as loading conditions, strength
of the material, microstructure, and the environment, for example, temperature and the
presence of corrosive or embrittling gases and
fluids. Considering only the loading condit i o n s - w h i c h would include the mean stress,
the alternating stress, and the cyclic freq u e n c y - t h e magnitude of the alternating stress
(Crmax -- ~min) has the greatest effect on striation spacing. Increasing the magnitude of the
alternating stress produces an increase in the
striation spacing (Fig. 25a). While rising, the
mean stress can also increase the striation
spacing; this increase is not as great as one for
a numerically equivalent increase in the alternating stress. Within reasonable limits, the
cyclic frequency has the least effect on striation
spacing. In some cases, fatigue striation spacing can change significantly over a very short
distance (Fig. 25b). This is due in part to
changes in local stress conditions as the crack
propagates on an inclined surface.
For a Stage II fatigue crack propagating
under conditions of reasonably constant cyclic
loading frequency and advancing within the
nominal range of 10 -5 to 10 -3 mm/cycle*, the
crack growth rate, da/dN, can be expressed as a
function of the stress intensity factor K (Ref 15,
31, 32):
da
i = C(AK)"
dN
(Eq 1)
where a is the distance of fatigue crack advance, N is the number of cycles applied to
advance the distance a, m and C are constants,
and ZkK = Kmax - Kmi" is the difference
between the maximum and minimum stress
intensity factor for each fatigue load cycle. The
*All fatigue crack growth rates in this article are given in
millimeters per cycle (mm/cycle). To convert to inches per
cycle (in./cycle), multiply by 0.03937. See also the Metric
Conversion Guide in this Volume.
16 / Modes
of Fracture
o- 1
- di'° , es
(a)
I
l mm
I
I
(b)
20 pm
T1
(T 1
Fig, 8
Effectof asymmetry on dimple size
o"
....--Wall of dimple
Original dimple surface
(c)
I
75 p.m
I
(d)
I
20 I-tm
~
I
w
,
slip-created surface
Fig. 7
Elongated dimples formed on shear and torsion specimen fracture surfaces. (a) Shear fracture of a
commercially pure titanium screw. Macrofractograph shows spiral-textured surface of shear-off screw.
Typical deformation lines are fanning out on the thread. (b) Higher-magnification view of (a) shows uniformly
distributed elongated shear dimples. (O.E.M. Pohler, Institut StraumannAG). (c) Elongated dimples on the surface
of a fractured single-strandcopper wire that failed in torsion. (d) Higher-magnificationview of the elongated dimples
shown in (c). (R.D. Lujan, Sandia National Laboratories)
- -
Preferably oriented
slip planes
(r
stress intensity factor, K, describes the stress
condition at a crack and is a function of the
applied stress and a crack shape factor, generally expressed as a ratio of the crack depth to
length.
When a fatigue striation is produced on each
loading cycle, da/dN represents the striation
spacing. Equation 1 does not adequately
describe Stage I or Stage III fatigue crack
growth rates; it tends to overestimate Stage I
and often underestimates Stage III growth rates
(Ref 15).
Stage III is the terminal propagation phase of
a fatigue crack in which the striation-forming
mode is progressively displaced by the static
fracture modes, such as dimple rupture or
cleavage. The rate of crack growth increases
during Stage III until the fatigue crack becomes
unstable and the part fails. Because the crack
propagation is increasingly dominated by the
static fracture modes, Stage III fatigue is sen-
sitive to both microstructure and mean stress
(Ref 17, 18).
Characteristics of Fractures With Fatigue Striations. During Stage II fatigue, the
crack often propagates on multiple plateaus that
are at different elevations with respect to one
another (Fig. 26). A plateau that has a concave
surface curvature exhibits a convex contour on
the mating fracture face (Ref 29). The plateaus
are joined either by tear ridges or walls that
contain fatigue striations (Fig. 19 and 20a).
Fatigue striations often bow out in the direction
of crack propagation and generally tend to align
perpendicular to the principal (macroscopic)
crack propagation direction. However, variations in local stresses and microstructure can
change the orientation of the plane of fracture
and alter the direction of striation alignment
(Fig. 27).
Large second-phase particles and inclusions
in a metal can change the local crack growth
Downloaded from https://dl.asminternational.org/handbooks/chapter-pdf/268655/a0001831.pdf
by Washington State University user
Fig. 9
Slip step formation resulting in serpentine
glide and ripples on a dimple wall
rate and resulting fatigue striation spacing.
When a fatigue crack approaches such a particle, it is briefly retarded if the particle remains
intact or is accelerated if the particle cleaves
(Fig. 18). In both cases, however, the crack
growth rate is changed only in the immediate
vicinity of the particle and therefore does not
significantly affect the total crack growth rate.
However, for low-cycle (high-stress) fatigue,
the relatively large plastic zone at the crack tip
can cause cleavage and matrix separation at the
particles at a significant distance ahead of the
advancing fatigue crack. The cleaved or matrixseparated particles, in effect, behave as cracks
or voids that promote a tear or shear fracture
between themselves and the fatigue crack, thus
significantly advancing the crack front (Ref 33,
Modes of Fracture
I \
Cleavage steps
Cleavage
planes
~
~
/ 17
Fracture
direction
subgrain boundary
(a)
/" \
Twist ~
---L
I
Cleavage f e a t h e r s ~
~
~
River pattern
I
/ ~ . .
J
Fracture
d,rect,on
I
5 i~m
Fig.
10
Serpentine glide formation (arrow) in
oxygen-free high-conductivity copper
specimen
Cleavage step
subgrain boundary
(b)
Fig. 1 1
(b)
(a)
Fig. 1 2
Schematicof cleavage fracture formation showing the effect of subgrain and grain boundaries. (a) Tilt
boundary. (b) Twist boundary
I
20 p m
I
Examples of cleavage fractures. (a) Twist boundary, cleavage steps, and river patterns in an Fe-O.O1C-O.24Mn-O.02Sialloy that was fractured by impact.
(b) Tongues (arrows) on the surface of a 30% Cr steel weld metal that fractured by cleavage
Downloaded from https://dl.asminternational.org/handbooks/chapter-pdf/268655/a0001831.pdf
by Washington State University user
1 8 / Modes
Fig.
of F r a c t u r e
Cleavage fracture in Armco iron showing
a tilt boundary, cleavage steps, and river
patterns. TEM p-c replica
13
1
34). Relatively small, individual particles have
no significant effect on striation spacing (Fig.
17b).
The distinct, periodic markings sometimes
observed on fatigue fracture surfaces are known
as tire tracks, because they often resemble the
tracks left by the tread pattern of a tire (Fig.
28). These rows of parallel markings are the
result of a particle or a protrusion on one fatigue
fracture surface being successively impressed
into the surface of the mating half of the
fracture during the closing portion of the fatigue
cycle (Ref 23, 29, 34). Tire tracks are more
common for the tension-compression than the
tension-tension type of fatigue loading (Ref
23). The direction of the tire tracks and the
change in spacing of the indentations within the
track can indicate the type of displacement that
occurred during the fracturing process, such as
lateral movement from shear or torsional loading. The presence of tire tracks on a fracture
surface that exhibits no fatigue striations may
indicate that the fracture occurred by low-cycle
(high-stress) fatigue (Ref 35).
Decohesive
Rupture
A fracture is referred to as decohesive rupture when it exhibits little or no bulk plastic
deformation and does not occur by dimple
rupture, cleavage, or fatigue. This type of
fracture is generally the result of a reactive
environment or a unique microstructure and is
associated almost exclusively with rupture
along grain boundaries. Grain boundaries contain the lowest melting point constituents of an
alloy system. They are also easy paths for
diffusion and sites for the segregation of such
elements as hydrogen, sulfur, phosphorus, antimony, arsenic, and carbon; the halide ions,
such as chlorides; as well as the routes of
penetration by the low melting point metals,
such as gallium, mercury, cadmium, and tin.
The presence of these constituents at the boundaries can significantly reduce the cohesive
(a)
J,
20 izm
(b)
Fig. 14
Examples of cleavage fractures. (a) Feather pattern on a single grain of a chromium steel weld metal
that failed by cleavage. (b) Cleavage steps in a Cu-25 at.% Au alloy that failed by transgranular
stress-corrosion cracking. (B.D. Lichter, Vanderbilt University)
strength of the material at the boundaries and
promote decohesive rupture (Fig. 29).
Decohesive rupture is not the result of one
unique fracture process, but can be caused by
several different mechanisms. The decohesive
processes involving the weakening of the
atomic bonds (Ref 36), the reduction in surface
energy required for localized deformation (Ref
37-39), molecular gas pressure (Ref 40), the
rupture of protective films (Ref 41, 42), and
anodic dissolution at active sites (Ref 43) are
associated with hydrogen embrittlement and
stress-corrosion cracking (SCC). Decohesive
rupture resulting from creep fracture mechanisms is discussed at the end of this section.
The fracture of weak grain-boundary films
(such as those resulting from grain-boundary
penetration by low melting point metals), the
rupture of melted and resolidified grainboundary constituents (as in overheated aluminum alloys), or the separation of melted material in the boundaries (Ref 44) before it
solidifies (as in the cracking at the heat-affected
zones, HAZs, of welds, a condition known as
hot cracking) can produce a decohesive rupture.
Figures 30 to 32 show examples of decohesive
rupture. A decohesive rupture resulting from
hydrogen embrittlement is shown in Fig. 30.
Figure 31 shows a decohesive rupture in a
precipitation-hardenable stainless steel due to
SCC. A fracture along a low-strength grainboundary film resulting from the diffusion of
liquid mercury is shown in Fig. 32. More
detailed information on hydrogen embrittlement, SCC, and liquid-metal embrittlement can
be found later in this article in the section
"Effect of Environment." When a decohesive
rupture occurs along flattened, elongated grains
that form nearly uninterrupted planes through
Downloaded from https://dl.asminternational.org/handbooks/chapter-pdf/268655/a0001831.pdf
by Washington State University user
I
I
1 ixm
Fig.
1 5 Wallner lines (arrow) on the surface of a
fractured WC-Co specimen. TEM formvar
replica. Etched with 5% HCI. (S.B. Luyckx, University of
the Witwatersrand)
the material, as in severely extruded alloys and
along the parting planes of some forgings, a
relatively smooth, featureless fracture results
(Fig. 33).
Creep rupture is a time-dependent failure
that results when a metal is subjected to stress
for extended periods at elevated temperatures
that are usually in the range of 40 to 70% of the
absolute melting temperature of the metal. With
few exceptions (Ref 45-49), creep ruptures
Modes
(a)
I
Fig.
!6
Fig.
! 7
20 i~m
I
of Fracture/19
I
Ib)
20 p.m
I
Stage I fatigue fracture appearance. (a) Cleovagelike, crystallographically oriented Stage I fatigue fracture in a cast Ni-14Cr-4.5Mo-] Ti-6AI-1 .SFe-2.0(Nb + To)
alloy. (b) Stair-step fracture surface indicative of Stage I fatigue fracture in a cast ASTM F75 cobalt-base alloy. SEM. (R. Abrams, Howmedica, Div. Pfizer Hospital
Products Group Inc.)
Uniformly distributed fatigue striations in an aluminum 2024-T3 alloy. (a) Tear ridge and inclusion
(outlined by rectangle). (b) Higher-magnification view of the region outlined by the rectangle in (a)
showing the continuity of the fracture path through and around the inclusion. Compare with Fig. 18.
Fig.
exhibit an intergranular fracture surface. Transgranular creep ruptures, which generally result
from high applied stresses (high strain rates),
fail by a void-forming process similar to that of
microvoid coalescence in dimple rupture (Ref
45-47). Because transgranular creep ruptures
show no decohesive character, they will not be
considered for further discussion. Intergranular
creep ruptures, which occur when metal is
subjected to low stresses (often well below the
yield point) and to low strain rates, exhibit
decohesive rupture and will be discussed in
more detail.
Creep can be divided into three general
stages: primary, secondary, and tertiary creep.
The fracture initiates during primary creep,
propagates during secondary or steady-state
creep, and becomes unstable, resulting in failure, during tertiary or terminal creep. From a
practical standpoint of the service life of a
structure, the initiation and steady-state propagation of creep ruptures are of primary importance, and most efforts have been directed
toward understanding the fracture mechanisms
involved in these two stages of creep.
As shown schematically in Fig. 34, intergranular creep ruptures occur by either of two
fracture processes: triple-point cracking or
grain-boundary cavitation (Ref 50-63). The
strain rate and temperature determine which
fracture process dominates. Relatively high
strain rates and intermediate temperatures promote the formation of wedge cracks (Fig. 34a).
Grain-boundary sliding as a result of an applied
Downloaded from https://dl.asminternational.org/handbooks/chapter-pdf/268655/a0001831.pdf
by Washington State University user
! 8
Local variations in striation spacing in a
Ni-O.O4C-21Cr-O.6Mn-2.5Ti-O.7AI alloy
that was tested under rotating-bending conditions. Compare with Fig. 17(b).
tensile stress can produce sufficient stress concentration at grain-boundary triple points to
initiate and propagate wedge cracks (Ref 50-52,
55, 56, 58-61). Cracks can also nucleate in the
grain boundary at locations other than the triple
point by the interaction of primary and secondary slip steps with a sliding grain boundary (Ref
61). Any environment that lowers grainboundary cohesion also promotes cracking (Ref
59). As sliding proceeds, grain-boundary
20 / Modes
Fig. 19
of Fracture
Fatigue striations in a 2024-T3 aluminum
alloy joined by tear ridges
as deformation continues, the cavities join to
form an intergranular fracture. Even though the
fracture resulting from cavitation creep exhibits
less sharply defined intergranular facets (Fig.
35c), it would be considered a decohesive
rupture.
Instead of propagating by a cracking or a
cavity-forming process, a creep rupture could
occur by a combination of both. There may be
no clear distinction between wedge cracks and
cavities (Ref 70-72). The wedge cracks could be
the result of the linkage of cavities at triple
points.
The various models proposed to describe the
creep process are mathematically complex and
were not discussed in detail. Comprehensive
reviews of the models are available in Ref 59,
63, 73, and 74.
Unique
cracks propagate and join to form intergranular
decohesive fracture (Fig. 35a and b).
At high temperatures and low strain rates,
grain-boundary sliding favors cavity formation
(Fig. 34b). The grain-boundary cavities resulting from creep should not be confused with
microvoids formed in dimple rupture. The two
are fundamentally different; the cavities are
principally the result of a diffusion-controlled
process, while microvoids are the result of
complex slip. Even at low strain rates, a sliding
grain boundary can nucleate cavities at irregularities, such as second-phase inclusion particles (Ref 54, 57, 63, 64). The nucleation is
believed to be a strain-controlled process (Ref
63, 64), while the growth of the cavities can be
described by a diffusion growth model (Ref
65-67) and by a power-law growth relationship
(Ref 68, 69). Irrespective of the growth model,
76-83). In steels, the cleavage facets of quasicleavage fracture occur on the {100}, {110}, and
possibly the {112} planes. The term quasicleavage can be used to describe the distinct
fracture appearance if one is aware that quasicleavage does not represent a separate fracture
mode.
A quasi-cleavage fracture initiates at the
central cleavage facets; as the crack radiates,
the cleavage facets blend into areas of dimple
rupture, and the cleavage steps become tear
ridges. Quasi-cleavage has been observed in
steels, including quench-and-temper hardenable, precipitation-hardenable, and austenitic
stainless steels; titanium alloys; nickel alloys;
and even aluminum alloys. Conditions that
impede plastic deformation promote quasicleavage fracture--for example, the presence
of a triaxial state of stress (as adjacent to the
root of a notch), material embrittlement (as by
hydrogen or stress corrosion), or when a steel is
subjected to high strain rates (such as impact
loading) within the ductile-to-brittle transition
range.
Flutes. Fractography has acquired a number
of colorful and descriptive terms, such as dimple
rupture, serpentine glide, ripples, tongues, tire
tracks, and factory roof, which describes a
ridge-to-valley fatigue fracture topography resulting from Mode III antiplane shear loading (Ref 84). The term flutes should also be
included in this collection. Flutes exhibit elongated grooves or voids (Fig. 38 and 39) that
connect widely spaced cleavage planes (Ref
85-90). The fracture process is known as fluting.
The term flutes was apparently chosen because
the fractures often resemble the long, parallel
grooves on architectural columns or the pleats in
drapes.
Fractures
Some fractures, such as quasi-cleavage and
flutes, exhibit a unique appearance but cannot
be readily placed within any of the principal
fracture modes. Because they can occur in
common engineering alloys under certain failure conditions, these fractures will be briefly
discussed.
Q u a s i - c l e a v a g e f r a c t u r e is a localized,
often isolated feature on a fracture surface that
exhibits characteristics of both cleavage and
plastic deformation (Fig. 36 and 37). The term
quasi-cleavage does not accurately describe the
fracture, because it implies that the fracture resembles, but is not, cleavage. The term was
coined because, although the central facets of a
quasi-cleavage fracture strongly resembled
cleavage (Ref 75), their identity as cleavage
planes was not established until well after the
term had gained widespread acceptance (Ref
I
I
I
10 izm
Fig. 20
Fatigue striations on adjoining walls on the fracture surface of a
commercially pure titanium specimen. (O.E.M. Pohler, Institut
Straumann AG)
Downloaded from https://dl.asminternational.org/handbooks/chapter-pdf/268655/a0001831.pdf
by Washington State University user
I
lO izm
Fig. 21
Fatigue striations on the fracture surface of a tantalum heat-exchanger
tube. The rough surface appearance is due to secondary cracking caused
by high-cycle low-amplitude fatigue. (M.E. Blum, FMC Corporation)
M o d e s of F r a c t u r e / 21
I
I
I
B p~m
Fig.
22
10 p.m
High-magnlflcatlon views of fatigue striations, (a) Striations (arrow) on the fracture surface of an oustenitic stainless steel. (C.R. Brooks and A. Choudhury,
University of Tennessee). (b) Fatigue striations on the facets of tantalum grains in the heat-affected zone of a weldment. (M.E. Blum, FMC Corporation)
o-
l
Slip
Orientation of
active slip planes
$ /" \\
'•
~//
\\
\\
/
\\
~
\',
c~ (Principal
tensile
tensile stress)
(a)
arrangement that resembles cleavage river patterns (Ref 89). Although fluting has been observed primarily in hexagonal close-packed
(hcp) metal systems, such as titanium and
zirconium alloys, evidence of fluting has also
been reported on a hydrogen-embrittled type
316 austenitic stainless steel (Ref 90). Titanium
alloys having a relatively high oxygen or aluminum content (o~-stabilizers) that are fractured
at cryogenic temperatures or fail by SCC may
exhibit fluting (Ref 89).
Tearing Topography
f
(Compressive
closure stress)
Relatively smooth trailing
edge of striation
S l i p
First
Second
Crack advance
during one
load cycle
traces on leading
edge of striation
Third cycle
(b)
Mechanism of fatigue crack propagation by alternate sllp at the crack tip. Sketches are simplified to
F l g o 2 3 clarify the basic concepts. (a) Crack opening and crack tip blunting by slip on alternate slip planes with
increasing tensile stress. (b) Crack closure and crack tip resharpening by partial slip reversal on alternate slip planes
with increasing compressive stress
Although flutes are not elongated dimples,
they are the result of a plastic deformation
process. Flutes are the ruptured halves of tubular voids believed to be formed by a planar
intersecting slip mechanism (Ref 85, 88, 89)
and have matching tear ridges on opposite
fracture faces. The tear ridges join in the
direction of fracture propagation, forming an
Downloaded from https://dl.asminternational.org/handbooks/chapter-pdf/268655/a0001831.pdf
by Washington State University user
Surface
A tentative fracture mode called tearing topography surface (TTS) has been identified and
described (Ref 91). The TTS fracture occurs in
a variety of alloy systems, including steels,
aluminum, titanium, and nickel alloys, and
under a variety of fracture conditions, such as
overload, hydrogen embrittlement (Ref 92),
and fatigue. Examples of TTS fractures are
shown in Fig. 40 to 42.
Although the precise nucleation and propagation mechanism for TTS fracture has not
been identified, the fracture appears to be the
result of a rnicroplastic tearing process that
operates on a very small (submicron) scale (Ref
91). The TTS fractures do not exhibit as much
plastic deformation as dimple rupture, although
they are often observed in combination with
dimples (Fig. 40 and 41). The fractures are
generally characterized by relatively smooth,
often flat, areas or facets that usually contain
thin tear ridges. Tearing topography surface
fractures may be due to closely spaced microvoid nucleation and limited growth before
coalescence, resulting in extemely shallow
dimples. However, this hypothesis does not
appear to be probable, because TTS is often
observed along with well-developed dimples in
alloys having relatively uniform carbide dispersions, such as HY-130 steel, and because TTS
22 / Modes
of Fracture
(a)
(b)
(c)
Fig. 24
Sowtooth and groove-type fatigue fracture profiles. Arrows show crack propagation direction.
(a) Distinct sawtooth profile (aluminum alloy). (b) Poorly formed sawtooth profile (steel). (c)
Groove-type profile (aluminum alloy). Source: Ref 23
is observed under varying stress states. A detailed discussion of the TTS fracture mode is
available in Ref 91.
centered cubic (fcc) metals and alloys are generally considered to have good resistance to
hydrogen embrittlement, it has been shown that
the 300 series austenitic stainless steels (Ref
95-98) and certain 2000 and 7000 series highstrength aluminum alloys are also embrittled by
hydrogen (Ref 99-107). Although the result of
hydrogen embrittlement is generally perceived
to be a catastrophic fracture that occurs well
below the ultimate strength of the material and
exhibits no ductility, the effects of hydrogen
can be quite varied. They can range from a
slight decrease in the percent reduction of area
at fracture to premature rupture that exhibits no
ductility (plastic deformation) and occurs at a
relatively low applied stress.
The source of hydrogen may be a processing
operation, such as plating (Fig. 30) or acid
cleaning, or the hydrogen may be acquired
from the environment in which the part operates. If hydrogen absorption is suspected.
prompt heating at an elevated temperature (usually about 200 °C, or 400 °F) will often restore
the original properties of the material.
The effect of hydrogen is strongly influenced
by such variables as the strength level of the
Effect of E n v i r o n m e n t
The environment, which refers to all external
conditions acting on the material before or
during fracture, can significantly affect the
fracture propagation rate and the fracture appearance. This section will present some of the
principal effects of such environments as hydrogen, corrosive media, low-melting metals.
state of stress, strain rate, and temperature.
Where applicable, the effect of the environment
on the fracture appearance will be illustrated.
Effect of Environment
on Dimple
Rupture
The Effect of H y d r o g e n . When certain
body-centered cubic (bcc) and hcp metals or
alloys of such elements as iron. nickel, titanium, vanadium, tantalum, niobium, zirconium. and hafnium are exposed to hydrogen,
they are susceptible to a type of failure known
as hydrogen embrittlement. Although the face-
(a)
Fig.
J
25
2 i~m
I
(b)
alloy, the microstructure, the amount of hydrogen absorbed (or adsorbed), the magnitude of
the applied stress, the presence of a triaxial
state of stress, the amount of prior cold work.
and the degree of segregation of such contaminant elements as phosphorus, sulfur, nitrogen.
tin. or antimony at the grain boundaries. In
general, an increase in strength, higher absorption of hydrogen, an increase in the applied
stress, the presence of a triaxial stress state.
extensive prior cold working, and an increase in
the concentration of contaminant elements at
the grain boundaries all serve to intensify the
embrittling effect of hydrogen. However. for an
alloy exhibiting a specific strength level and
microstructure, there is a stress intensity. K I.
below which, for all practical purposes, hydrogen embrittlement cracking does not occur.
This threshold crack tip stress intensity factor is
determined experimentally and is designated as
Kth.
A number of theories have been advanced to
explain the phenomenon of hydrogen embrittlement. These include the exertion of an internal
gas pressure at inclusions, grain boundaries.
surfaces of cracks, dislocations, or internal
voids (Ref 40. 108. 109): the reduction in
atomic and flee-surface cohesive strength (Ref
110-116): the attachment of hydrogen to dislocations, resulting in easier dislocation
breakaway from the pinning effects of carbon
and nitrogen (Ref 38. 112. 117-122): enhanced
nucleation of dislocations (Ref 112. 123): enhanced nucleation and grow:h of microvoids
(Ref 109. 110. 113. 116. 122. 124-126): enhanced shear and decrease of strain for the
onset of shear instability (Ref 112. 127, 128):
I
10 i~m
I
Variations in fatigue striation spacing. (a) Spectrum-loaded fatigue fracture in a 7475-T7651 aluminum alloy test coupon showing an increase in striation spacing
due to higher alternating stress. (b) Local variation in fatigue striation spacing in a spectrum-loaded 7050-T7651 aluminum alloy extrusion. (D. Brown, Douglas
Aircraft Company)
Downloaded from https://dl.asminternational.org/handbooks/chapter-pdf/268655/a0001831.pdf
by Washington State University user
M o d e s of F r a c t u r e / 2 3
Convex
Tear r
i
d
g
//~-~..]~/
Concave~
Fig. 26
e
~
Crack propagation
direction
Schematic illustrating fatigue striations on
plateaus
the formation of methane gas bubbles at grain
boundaries (Ref 129, 130); and, especially for
titanium alloys, the repeated formation and
rupture of the brittle hydride phase at the crack
tip (Ref 131-137). Probably no one mechanism
is applicable to all metals, and several mechanisms may operate simultaneously to embrittle
a material. Whatever the mechanism, the end
result is an adverse effect on the mechanical
properties of the material.
If the effect of hydrogen is subtle, such as
when there is a slight decrease in the reduction of area at fracture as a result of a tensile
test, there is no perceivable change in the
dimple rupture fracture appearance. However,
the dimples become more numerous but are
more shallow at a greater loss in ductility (Fig.
43).
Hydrogen Embrittlement of Steels. At low
strain rates or when embrittlement is more
severe, the fracture mode in steels can change
from dimple rupture to quasi-cleavage, cleavage, or intergranular decohesion. These
changes in fracture mode or appearance may
not occur over the entire fracture surface and
are usually more evident in the region of the
fracture origin. Figure 44 shows an example of
a hydrogen-embrittled AISI 4340 steel that
exhibits quasi-cleavage.
When an annealed type 301 austenitic stainless steel is embrittled by hydrogen, the fracture
Fig.
27
Striations on two joining, independent
fatigue crack fronts on a fracture surface
of aluminum alloy 6061-T6. The two arrows indicate
direction of local crack propagation. TEM p-c replica
occurs by cleavage (Fig. 45a). An example in
which the mode of fracture changed to intergranular decohesion in a hydrogen-embrittled
AISI 4130 steel is shown in Fig. 45b.
When a hydrogen embrittlement fracture
propagates along grain boundaries, the presence of such contaminant elements as sulfur,
phosphorus, nickel, tin, and antimony at the
boundaries can greatly enhance the effect of
hydrogen (Ref 111, 139). For example, the
segregation of contaminant elements at the
grain boundaries enhances the hydrogen embrittlement of high-strength low-alloy steels
tempered above 500 °C (930 °F) (Ref 92). The
presence of sulfur at grain boundaries promotes
hydrogen embrittlement of nickel, and for
equivalent concentrations, the effect of sulfur is
nearly 15 times greater than that of phosphorus
(Ref 140).
I4ydrogen Embrittlement of Titanium. Although titanium and its alloys have a far greater
tolerance for hydrogen than high-strength
steels, titanium alloys are embrittled by hydrogen. The degree and the nature of the embritdement is strongly influenced by the alloy, the
microstructure, and whether the hydrogen is
::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
....
iii:
i
(a)
(b)
Fig. 29
(c)
Schematic illustrating decohesive rupture along grain boundaries. (a) Decohesion along grain
boundaries of equiaxed grains. (b) Decohesionthrough a weak grain-boundary phase. (c) Decohesion
along grain boundaries of elongated grains
Downloaded from https://dl.asminternational.org/handbooks/chapter-pdf/268655/a0001831.pdf
by Washington State University user
I
I
5 I*m
Fig. 28
Tire tracks on the fatigue fracture surface
of a quenched-and-tempered AISI 4140
steel. TEM replica. (I. Le May, Metallurgical Consulting
Services Ltd.)
present in the lattice before testing or is introduced during the test. For example, a Ti-8A11Mo-IV alloy that was annealed at 1050 °C
(1920 °F), cooled to 850 °C (1560 °F). and
water quenched to produce a coarse Widmanstatten structure exhibited cracking along the
et-[3 interfaces when tested in 1-arm hydrogen
gas at room temperature (Ref 137). The fracture
surface, which exhibited crack-arrest markings.
is shown in Fig. 46(a). The arrest markings are
believed to be due to the discontinuous crack
propagation as a result of the repeated rupture
of titanium hydride phase at the crack tip (Ref
137). Also, Fig. 46(b) shows a hydrogen embrittlement fracture in a Ti-5AI-2.5Sn alloy
containing 90 ppm H that was 13 processed at
1065 °C (1950 °F) and aged for 8 h at 950 °C
(1740 °F). The fracture occurred by cleavage.
Cleavage was also the mode of fracture for a
Ti-6A1-4V alloy having a microstructure consisting of a continuous, equiaxed et phase with
a fine, dispersed 13 phase at the (x grain boundaries embrittled by exposure to hydrogen gas at
a pressure of 1 atm (Fig. 47a). However. when
the same Ti-6A1-4V alloy having a microstructure consisting of a medium, equiaxed et phase
with a continuous 13 network was embrittled by
1-atm hydrogen gas, the fracture occurred by
intergranular decohesion along the ct-[3 boundaries (Fig. 47b and c).
Hydrogen
Embrittlement
of Aluminum.
There is conclusive evidence (Ref 99-107) that
some aluminum alloys, such as 2124, 7050,
7075, and even 5083 (Ref 143), are embrittled
by hydrogen and that the embrittlement is
apparently due to some of the mechanisms
already discussed, namely enhanced slip and
trapping of hydrogen at precipitates within
grain boundaries. The embrittlement in alumi-
24 / Modes
of Fracture
(a)
I
1 pm
I
I
I
2.5 i~m
(b)
Fig. 30
Decohesive rupture in an AISI 8740 steel nut due to hydrogen embrittlement. Failure was due to inadequate baking following cadmium plating; thus, hydrogen,
which was picked up during the plating process, was not released. (a) Macrograph of fracture surface. (b) Higher-magnification view of the boxed area in (a)
showing typical intergranular fracture. (W.L. Jensen, Lockheed Georgia Company)
num alloys depends on such variables as the
microstructure, strain rate, and temperature. In
general, underaged microstructures are more
susceptible to hydrogen embrittlement than the
peak or averaged structures. For the 7050 aluminum alloy, a low (0.01%) copper content renders all microstructures more susceptible to
embrittlement than those of normal (2.1%) copper content (Ref 106). Also, hydrogen embrittlement in aluminum alloys is more likely to
occur at lower strain rates and at lower temperatures.
The effect of hydrogen on the fracture appearance in aluminum alloys can vary from no
significant change in an embrittled 2124 alloy
(Ref 99) to a dramatic change from the normal
dimple rupture to a combination of cleavagelike
transgranular fracture and intergranular decohesion in the high-strength 7050 (Ref 106) and
7075 (Ref 105) aluminum alloys. Figure 48
shows an example of a fracture in a hydrogenembrittled (as measured by a 21% decrease in
the reduction of area at fracture) 2124-UT
(underaged temper: aged 4 h at 190 °C,
or 375 °F) aluminum alloy. It can be seen
that there is little difference in fracture appearance between the nonembrittled and embrittled
specimens. However, when a low-copper
(0.01%) 7050 in the peak-aged condition (aged
24 h at 120 °C, or 245 °F) is hydrogen
embrittled, a cleavagelike transgranular fracture results (Fig. 49a). This same alloy in the
underaged condition (aged 10 h at 100 °C,
or 212 °F) fails by a combination of inter-
lal
I
1 mm
I
granular decohesion and cleavagelike fracture
(Fig. 49b).
The Effect of a Corrosive E n v i r o n m e n t .
When a metal is exposed to a corrosive environment while under stress, SCC, which is a
form of delayed failure, can occur. Corrosive
environments include moist air; distilled and
tap water; seawater; gaseous ammonia and
ammonia in solutions; solutions containing
chlorides or nitrides; basic, acidic, and organic
solutions; and molten salts. The susceptibility
of a material to SCC depends on such variables
as strength, microstructure, magnitude of the
applied stress, grain orientation (longitudinal or
short transverse) with respect to the principal
applied stress, and the nature of the corrosive
environment. Similar to the Kth in hydrogen
(b)
I
100 p.m
I
F l g o 3 1 17-4 PH stainlesssteel main landing-gear deflection yoke that failed because of intergranular SCC. (a) Macrograph of fracture surface. (b) Higher-magnification
view of the boxed area in (a) showing area of intergranular attack. (W.L. Jensen, Lockheed Georgia Company)
Downloaded from https://dl.asminternational.org/handbooks/chapter-pdf/268655/a0001831.pdf
by Washington State University user
Modes
embrittlement, there is also a threshold crack
tip stress intensity factor, K=scc, below which a
normally susceptible material at a certain
strength, microstructure, and testing environment does not initiate or propagate stresscorrosion cracks. Stress-corrosion cracks normally initiate and propagate by tensile stress;
however, compressive-stress SCC has been observed in a 7075-T6 aluminum alloy and a type
304 austenitic stainless steel (Ref 144).
Stress-corrosion cracking is a complex phenomenon, and the basic fracture mechanisms
are still not completely understood. Although
such processes as dealloying (Ref 145-148) in
brass and anodic dissolution (Ref 149, 150,
151) in other alloy systems are important SCC
mechanisms, it is apparent that the principal
SCC mechanism in steels, titanium, and aluminum alloys is hydrogen embrittlement (Ref 38,
100, 107, 137, 143, 152-166). In these alloys,
SCC occurs when the hydrogen generated as a
result of corrosion diffuses into and embrittles
the material. In these cases, SCC is used to
describe the test or failure environment, rather
than a unique fracture mechanism.
Mechanisms of SCC. The basic processes
that lead to SCC, especially in environments
containing water, involve a series of events that
begin with the rupture of a passive surface film
(usually an oxide), followed by metal dissolution, which results in the formation of a pit or
crevice where a crack eventually initiates and
propagates. When the passive film formed during exposure to the environment is ruptured by
chemical attack or mechanical action (creepstrain), a clean, unoxidized metal surface is
exposed. As a result of an electrochemical
potential difference between the new exposed
metal surface and the passive film, a small
electrical current is generated between the anodic metal and the cathodic film. The relatively
small area of the new metal surface compared
to the large surface area of the surrounding
passive film results in an unfavorable anode-tocathode ratio. This causes a high local current
density and induces high metal dissolution (anodic dissolution) at the anode as the new metal
protects the adjacent film from corrosion; that
is, the metal surface acts as a sacrificial anode
in a galvanic couple.
If the exposed metal surface can form a
new passive film (repassivate) faster than the
new metal surface is created by film rupture,
the corrosion attack will stop. However, if the
repassivation process is suppressed, as in the
presence of chlorides, or if the repassivated
film is continuously ruptured by strain, as when
the material creeps under stress, the localized
corrosion attack proceeds (Ref 167-172). The
result is the formation and progressive enlargement of a pit or crevice and an increase in the
concentration of hydrogen ions and an accompanying decrease in the pH of the solution
within the pit.
The hydrogen ions result from a chemical
reaction between the exposed metal and the
water within the cavity. The subsequent
I
Fig.
50 p.m
of F r a c t u r e
/ 25
I
32
Fracture surface of a Monel specimen that
failed in liquid mercury. The fracture is
predominantly intergranular with some transgranular
contribution. (C.E. Price, Oklahoma State University)
Stress-corrosion fracture that occurred by
decohesion along the parting plane of an
aluminum alloy forging
reduction of the hydrogen ions by the acquisition of electrons from the environment results
in the formation of hydrogen gas and the
diffusion of hydrogen into the metal. This
absorption of hydrogen produces localized
cracking due to a hydrogen embrittlement
mechanism (Ref 173, 174). Because the metal
exposed at the crack tip as the crack propagates
by virtue of hydrogen embrittlement and the
applied stress is anodic to the oxidized sides of
the crack and the adjacent surface of the
material, the electrochemical attack continues,
as does the evolution and absorption of
hydrogen. The triaxial state of stress and the
stress concentration at the crack tip enhance
hydrogen embrittlement and provide a driving
force for crack propagation.
In materials that are insensitive to hydrogen
embrittlement, SCC can proceed by the anodic
dissolution process with no assistance from
hydrogen (Ref 149, 155, 161). Alloys are not
homogeneous, and when differences in chemical composition or variations in internal strain
occur, electrochemical potential differences
arise between various areas within the microstructure. For example, the grain boundaries
are usually anodic to the material within the
grains and are therefore subject to preferential
Fig. 33
o-
(T
T
/
(T
(a)
Fig. 34
Downloaded from https://dl.asminternational.org/handbooks/chapter-pdf/268655/a0001831.pdf
by Washington State University user
=
\
ff
(b)
Triple-point cracking (a) and cavitation (b) in intergranular creep rupture. Small arrows indicate
grain-boundary sliding.
26 / Modes
(a)
Fig. 35
of Fracture
I
I
18 I~m
I
!
10 ~m
(c)
I
1 iLm
I
Examples of intergranular creep fractures. (a) Wedge cracking in Inconel 625. (b) Wedge cracking in Incoloy 800. (c) Intergranular creep fracture resulting from
grain-boundary cavitation in PE-16. Source: Ref 59
anodic dissolution when exposed to a corrosive
environment. Inclusions and precipitates can
exhibit potential differences with respect to the
surrounding matrix, as can plastically deformed
(strained) and undeformed regions within a
material. These anode-cathode couplings can
initiate and propagate dissolution cracks or
fissures without regard to hydrogen.
(al
(bl
Although other mechanisms may operate
(Ref 175-178), including the adsorption of
unspecified damaging species (Ref 177) and the
occurrence of a strain-induced martensitic
transformation (Ref 178), dezincification or
dealloying (Ref 145-148) appears to be the
principal SCC mechanism in brass (copper-zinc
and copper-zinc-tin alloys). Dezincification is
the preferential dissolution or loss of zinc at the
fracture interface during SCC, which can result
in the corrosion products having a higher concentration of zinc than the adjacent alloy. This
dynamic loss of zinc near the crack aids in
propagating the stress-corrosion fracture.
Some controversy remains regarding the precise mechanics of dezincification. One mechanism assumed that both zinc and copper are
dissolved and that the copper is subsequently
redeposited, while the other process involves
the diffusion of zinc from the alloy, resulting in
a higher concentration of copper in the depleted
zone (Ref 179). However, there is evidence that
both processes may operate (Ref 180).
Like hydrogen embrittlement, SCC can
change the mode of fracture from dimple rupture to intergranular decohesion or cleavage,
(b)
Fig. 36
Examples of quasi-cleavage. (a) Fracture surface of an austenitized Fe-O.3C-O.6Mn-5.0Mo specimen
exhibiting large quasi-cleavage facets, such as at A; elsewhere, the surface contains rather large
dimples. (b) Charpy impact fracture in an Fe-O.18C-3.85Mo steel. Many quasi-cleavage facets are visible. The
rectangle outlines a tear ridge.
Downloaded from https://dl.asminternational.org/handbooks/chapter-pdf/268655/a0001831.pdf
by Washington State University user
Fig. 37
Small and poorly defined quasi-cleavage
facets connected by shallow dimples on
the surface of a type 234 tool steel. (D.-E. Huang, Fuxin
Mining Institute, and C.R. Brooks, University of
Tennessee)
Modes
(a)
I
I
(b)
of Fracture
I
60 l~m
to)
}
17 i~m
/ 27
I
15 izm
I
(dl
I
10 i~m
I
Examples of fluting. (a) Flutes and cleavage resulting from a mechanical overload of a Ti-0.350 alloy. (b) Flutes and cleavage resulting from SCC at 13-annealed
F i g . 3 8 Ti-8AI-1Mo-IV alloy in methanol. (c) Flutes and cleavage in 13-annealed Ti-8AI-1Mo-IV resulting from sustained-load cracking in vacuum. (d) Flutes occurring near
the notch on the fracture surface of mill-annealed Ti-SAI-1Mo-IV resulting from corrosion fatigue in saltwater. Source: Ref 89
although quasi-cleavage has also been observed. The change in fracture mode is generally confined to that portion of the fracture that
propagated by SCC, but it may extend to
portions of the rapid fracture if a hydrogen
embrittlement mechanism is involved.
Stress-corrosion fractures that result from
hydrogen embrittlement closely resemble those
fractures; however, stress-corrosion cracks usually exhibit more secondary cracking, pitting,
and corrosion products. Of course, pitting and
corrosion products could be present on a clean
hydrogen embrittlement fracture exposed to a
corrosive environment.
SCC of Steels. Examples of known stresscorrosion fractures are shown in Fig. 50 to 56.
Steels, including the stainless grades, stress
corrode in such environments as water, seawater, chloride- and nitrate-containing solu-
tions, and acidic as well as basic solutions,
such as those containing sodium hydroxide or
hydrogen sulfide. Stress-corrosion fractures
in high-strength quench-and-temper hardenable or precipitation-hardenable steels occur primarily by intergranular decohesion, although some transgranular fracture may also
be present.
Figure 50 shows a stress-corrosion fracture in
an HY- 180 quench-and-temper hardenable steel
tested in aqueous 3.5% sodium chloride. The
stress-corrosion fracture was believed to have
occurred predominantly by hydrogen embrittlement (Ref 154). Increasing the stress intensity
coefficient, K 1, resulted in a decreased tendency for intergranular decohesion; however,
the oppposite was true for a cold-worked type
316 austenitic stainless steel tested in boiling
aqueous magnesium chloride (Ref 181). It was
Downloaded from https://dl.asminternational.org/handbooks/chapter-pdf/268655/a0001831.pdf
by Washington State University user
shown that increasing K~ or increasing the
negative electrochemical potential resulted in
an increased tendency toward intergranular decohesion (Fig. 51). When the 300 type stainless
steels are sensitized----a condition that results in
the precipitation of chromium carbides at the
grain boundaries, causing depletion of chromium in the adjacent material in the grains-the steel becomes susceptible to SCC, which
occurs principally along grain boundaries.
Figure 52 shows the effect of the electrochemical potential, E, on the fracture path in a
cold-worked AISI C-1018 low-carbon steel that
stress corroded in a hot sodium hydroxide
solution. At an electrochemical potential of
E = - 0 . 7 6 VsH E, the fracture path is predominantly intergranular; at a freely corroding potential o f E = - 1.00 VsH E, the fracture path is
transgranular (Ref 182).
2 8 / M o d e s of F r a c t u r e
I
Fig.
20 ~m
I
39
Flutes and cleavage resulting from SCC of
13-annealed Ti-8AI-1Mo-IV in methanol.
Source: Ref 89
SCC of Aluminum. Aluminum alloys, especially the 2000 and 7000 series, that have been
aged to the high-strength T6 temper or are in an
underaged condition are susceptible to SCC in
such environments as moist air, water, and solutions containing chlorides. The sensitivity to
SCC depends strongly on the grain orientation
with respect to the principal stress, the shorttransverse direction being the most susceptible
to cracking. Figure 53 shows examples of stresscorrosion fractures in a 7075-T6 (maximum tensile strength: 586 MPa, or 85 ksi) aluminum
alloy that was tested in water. The fracture occurred primarily by intergranular decohesion.
SCC of brass in the presence of ammonia and
moist air has long been recognized. The term
season cracking was used to describe the SCC
of brass that appeared to coincide with the
moist weather in the spring and fall. Environments containing nitrates, sulfates, chlorides,
ammonia gas and solutions, and alkaline solutions are known to stress corrode brass. Even
distilled water and water containing as little as
5 x 10 3% sulfur dioxide have been shown to
attack brass (Ref 176, 178). Depending on the
arsenic content of the Cu-30Zn brass, SCC in
distilled water occurs either by intergranular
decohesion or by a combination of cleavage and
intergranular decohesion (Fig. 54). When brass
containing 0.032% As is stress corroded in
water containing minute amounts of sulfur dioxide, it exhibits a unique transgranular fracture containing relatively uniformly spaced,
parallel markings (Fig. 55). These distinct periodic marks apparently represent the stepwise
propagation of the stress-corrosion fracture.
SCC of titanium alloys has been observed in
such environments as distilled water, seawater,
aqueous 3.5% sodium chloride, chlorinated
organic solvents, methanol, red fuming nitric
Fig.
40
Appearance of TTS fracture in bainitic HY-130 steel. (a) Areas of complex tearing (T) and dimple
rupture (DR). (b) Detail from upper left corner of (a) showing particle-nucleated dimples (DR) and
region of TTS. SEM fractographs in (c) and (d) show additional examples of TTS fractures. Source: Ref 91, 93
Fig.
41
Appearance of TTS fracture. (a) An essentially 100% pearlitic eutectoid steel (similar to AISI 1080)
where fracture propagates across pearlite colonies. (b) Fractograph showing dimple rupture (DR) and
TTS fracture in a quenched-and-tempered (martensitic) HY-130 steel. Source: Ref 91, 93
acid, and molten salts. Susceptiblity depends
on such variables as the microstructure (Ref
183-185), the amount of internal hydrogen
(Ref 186), the state of stress (Ref 187, 188),
Downloaded from https://dl.asminternational.org/handbooks/chapter-pdf/268655/a0001831.pdf
by Washington State University user
and strength level (Ref 188). In general,
microstructures consisting of large-grain et
phase or containing substantial amounts of
phase in relation to [3 phase, high levels of
Modes
Fig.
42
Examples of TTS fracture in Ti-6AI-4V a-J3 alloys. (a) Solution treated and aged microstructure
consisting of about lO-~xm diam primary a particles in a matrix of about 70 vol% of fine
Widmanst~itten a and 13. The microstructural constituents are not evident on the fracture surface as verified by the
plateau-etching technique (Ref 91, 94). (b) Fractograph of a 13-quenched Ti-6AI-4V alloy consisting of a fine
Widmanst6tten martensitic microstructure. The tearing portions of the fracture surface exhibit TTS.
internal hydrogen, the presence of a triaxial
state of stress, and high yield strengths all
promote the susceptibility of an alloy to SCC.
If hydrogen is present in the corrosive
environment, SCC will probably occur by a
hydrogen embrittlement mechanism. Depending on the environment, alloy, and heat
treatment (microstructure), mild stresscorrosion attack can exhibit a fracture that
cannot be readily distinguished from normal
overload, while more severe attack results in
cleavage or quasi-cleavage fracture.
Figure 56 shows a stress-corrosion fracture in
an annealed Ti-8AI-IMo-IV alloy that was
tested in aqueous 3.5% sodium chloride. The
stress-corrosion fractures in titanium alloys exhibit both cleavage (along with fluting) and
quasi-cleavage.
Corrosion products are a natural by-product
of corrosion, particularly on most steels and
aluminum alloys. They not only obscure fracture detail but also cause permanent damage,
because a portion of the fracture surface is
chemically attacked in forming the corrosion
products. Therefore, removing the corrosion
products will not restore a fracture to its original condition. However, if the corrosion damage is moderate, enough surface detail remains
to identify the mode of fracture.
Depending on the alloy and the environment,
corrosion products can appear as powdery residue, amorphous films, or crystalline deposits.
Corrosion products may exhibit cleavage fracture and secondary cracking. Care must be
exercised in determining whether these fractures are part of the corrosion product or the
base alloy. Some of the corrosion products
observed on an austenitic stainless steel and a
niobium alloy are shown in Fig. 57 and 58,
respectively. Detailed information on the cleaning of fracture surfaces is available in the article
"Preparation and Preservation of Fracture
Specimens" in this Volume.
Effect of E x p o s u r e to L o w - M e l t i n g M e t als. When metals such as certain steels, tita-
nium alloys, nickel-copper alloys, and aluminum alloys are stressed while in contact with
low-melting metals, including lead, tin, cadmium, lithium, indium, gallium, and mercury,
they may be embrittled and fracture at a stress
below the yield strength of the alloy. If the
embrittling metal is in a liquid state during
exposure, the failure is referred to as liquidmetal embrittlement (LME); when the metal is
solid, it is known as solid-metal embrittlement
(SME). Both failure processes are sometimes
called stress alloying.
Temperature has a significant effect on the
rate of embrittlement. For a specific embrittling
metal species, the higher the temperature, the
more rapid the attack. In addition, LME is a
faster process than SME. In fact, under certain
conditions, LME can occur with dramatic
speed. For liquid indium embrittlement of steel,
the time to failure appears to be limited primarily by the diffusion-controlled period required to form a small propagating crack (Ref
189). Once the crack begins to propagate,
failure can occur in a fraction of a second. For
example, when an AIS14140 steel that was heat
treated to an ultimate tensile strength of 1500
MPa (218 ksi) was tested at an applied stress of
1109 MPa (161 ksi) (the approximate proportional limit of the material) while in contact
with liquid indium at a temperature of 158 °C
(316 °F) (indium melts at 156 °C, or 313 °F),
crack formation required about 511 s. The
crack then propagated and fractured the 5.84mm (0.23-in.) diam electropolished round bar
specimen in only 0.1 s (Ref 189). In contrast, at
154 °C (309 °F), when the steel was in contact
with solid indium, crack nucleation required
4.07 x 103 s (1.13 h), and failure required an
additional 2.41 x 103 S (0.67 h) (Ref 189).
Although gallium and mercury rapidly erabrittle aluminum alloys, all cases of LME and,
Downloaded from https://dl.asminternational.org/handbooks/chapter-pdf/268655/a0001831.pdf
by Washington State University user
of Fracture
/ 29
especially, SME do not occur in such short time
spans. The embrittlement of steels and titanium
alloys by solid cadmium can occur over months
of exposure; however, when long time spans
are involved, the generation of hydrogen by the
anodic dissolution of cadmium in a service
environment can result in a hydrogen embrittlement assisted fracture. The magnitude of the
applied stress, the strain rate, the amount of
prior cold work, the grain size, and the grainboundary composition can also influence the
rate of embrittlement. In general, higher applied stresses and lower strain rates promote
embrittlement (Ref 112), while an increase in
the amount of cold work reduces embrittlement
(Ref 189). The reduction in embrittlement from
cold work is believed to be due to the increase
in the dislocation density within grains providing a large number of additional diffusion paths
to dilute the concentration of embrittling atoms
at grain boundaries.
Smaller grain size should reduce embrittlement because of reduced stress concentration at
grain-boundary dislocation pile-ups (Ref 189);
however, in the embrittlement of Monel 400 by
mercury, maximum embrittlement is observed
at an approximate grain size of 250 Ixm (average grain diameter), and the embrittlement
decreases for both the smaller and the larger
grain sizes (Ref 112). The decrease in embrittlement at the smaller grain sizes was attributed
to a difficulty in crack initiation, and for the
larger grain sizes, the effect was due to enhanced plasticity (Ref 112). An example of a
Monel specimen embrittled by liquid mercury
is shown in Fig. 32.
When fracture occurs by intergranular decohesion, the presence of such elements as lead,
tin, phosphorus, and arsenic at grain boundaries
can affect the embrittlement mechanism. The
segregation of tin and lead at grain boundaries
of steel can make it more susceptible to embrittlement by liquid lead, while a similar grainboundary enrichment by phosphorus and arsenic reduces it (Ref 190). Grain-boundary
segregation of phosphorus has also been shown
to reduce the embrittlement of nickel-copper
alloys, such as Monel 400, by mercury (Ref
191, 192). It has been suggested that the
beneficial effects of phosphorus are due to a
modification in the grain-boundary composition
that results in improved atomic packing at the
boundary (Ref 192).
The mechanisms proposed to explain the
low-melting metal embrittlement process are
often similar to those suggested for hydrogen
embrittlement. Some of the mechanisms assume a reduction in the cohesive strength and
enhancement of shear as a result of adsorption
of the embrittling metal atoms (Ref 112, 114,
189, 193). It has also been suggested that the
diffusion of a low melting point metal into the
alloy results in enhanced dislocation nucleation
at the crack tip (Ref 123,127, 194). A modified
theory for crack initiation is based on stress and
dislocation-assisted diffusion of the embrittling
metal along dislocation networks and grain
30 / Modes
Fig.
43
of Fracture
Effect of hydrogen on fracture appearance in 13-8 PH stainlesssteel with a tensile strength of 1634 MPo (237 ksi). Top row: SEM fractographs of a specimennot
embrittled by hydrogen. Bottom row: SEM fractographs of a specimen charged with hydrogen by plating without subsequentbaking.
boundaries (Ref 189). The diffused atoms
lower the crack resistance and make slip more
difficult; when a critical concentration of the
embrittling species has accumulated in the penetration zone, a crack initiates. The mechanism
for the extremely rapid crack propagation for
LME is not well understood.
Diffusion processes are far too slow to transport the embrittling liquid metal to the rapidly
advancing crack front. For embrittlement by
liquid indium, it has been proposed that the
transport occurs by a bulk liquid flow mechanism (Ref 189, 195); for the SME mode, the
crack propagation is sustained by a much
slower surface self-diffusion of the embrittling
metal to the crack tip (Ref 189).
Examples of low-melting metal embrittlement fractures are shown in Fig. 32 and 59 to
61. Figure 59 shows fractures in AISI 4140
steel resulting from testing in argon and in
liquid lead. Figure 60 shows the embrittlement
of a 7075-T6 aluminum alloy by mercury, and
Fig. 61 shows the embrittlement of AISI 4140
steel by liquid cadmium. The articles "LiquidMetal Embrittlement" and "Embrittlement by
Solid-Metal Environments" in Volume 11 of
the 9th edition of Metals Handbook provide
additional information on the effect of exposure
to low melting point metals.
Effect of State of Stress. This section will
briefly discuss some effects of the direction of
the principal stress as well as the state of stress,
that is, uniaxial or triaxial, on the fracture modes
of various metal systems. This section will not,
however, present any mathematical fracture mechanics relationships describing the state of
stress or strain in a material. The effects of stress
will be discussed in general terms only.
Downloaded from https://dl.asminternational.org/handbooks/chapter-pdf/268655/a0001831.pdf
by Washington State University user
The effect of the direction of the applied
stress has been presented in the section "Dimple Rupture" in this article. Briefly, the direction of the principal stress affects the
dimple shape. Stresses acting parallel to the
plane of fracture (shear stresses) result in
elongated dimples, while a principal stress acting normal to the plane of fracture results
in primarily equiaxed dimples. Because the
local fracture planes often deviate from
the macroscopic plane and because the fracture is usually the result of the combined
effects of tensile and shear stresses, it generally exhibits a variety of dimple shapes and
orientations.
The state of stress affects the ability of a
material to deform. A change from a uniaxial to
biaxial to triaxial state of stress decreases the
ability of a material to deform in response to the
Modes
I
5 ~m
I
,,~=:", 4 4
Quasi-cleavage fracture in a hydrogenembrittled AISI 4340 steel heat treated to
an ultimate tensile strength of 2082 MPa (302 ksi).
Source: Ref 138
applied stresses. As a result, metals sensitive to
such changes in the state of stress exhibit a
decrease in elongation or reduction of area at
fracture and in extreme cases may exhibit a
change in the fracture mode.
The fcc metals, such as the aluminum alloys
and austenitic stainless steels, and the hcp
metals, such as the titanium and zirconium
alloys, are generally unaffected by the state of
stress. Although there can be a change in the
nature of the dimples under biaxial or triaxial
stresses, namely a reduction in dimple size and
depth (Ref 196, 197), fcc and hcp metal systems usually do not exhibit a change in the
mode of fracture. However, the bcc metals,
such as most iron-base alloys and refractory
metals, can exhibit not only smaller and shallower dimples but also a change in the fracture
mode in response to the restriction on plastic
deformation. This response depends on such
variables as the strength level, microstructure,
and the intensity of the triaxial stress. When a
change in the fracture mode does occur as a
result of a triaxial state of stress, such as that
present near the root of a sharp notch, the mode
of rupture can change from the normal dimple
rupture to quasi-cleavage or intergranular decohesion (Ref 198). These changes in fracture
mode are most evident in the general region of
the fracture origin and may not be present over
the entire fracture surface.
Figure 62 shows the effect of a biaxial state
of stress on dimples in a basal-textured Ti-6AI4V alloy. Under a biaxial state of stress, the
size and the depth of the dimples decreased. For
a pearlitic AISI 4130 steel (Ref 198) and a PH
13-8 precipitation-hardenable stainless steel, a
triaxial state of stress resulting from the presence of a notch with a stress concentration
factor of at least K t = 2.5 can change the
fracture mode from dimple rupture to quasicleavage (Fig. 63). When a high-strength AISI
4340 steel is subjected to a triaxial stress, the
mode of fracture can change from dimple rupture to intergranular decohesion.
Effect o f S t r a i n R a t e . The strain rate is a
variable that can range from the very low rates
observed in creep to the extremely high strain
rates recorded during impact or shock loading
by explosive or electromagnetic impulse.
Very low strain rates (about 10 - 9 to 10 - 7
s - l ) can result in creep rupture, with the
(a)
Fig°
I
45
10 ~m
I
(b)
of Fracture
/ 31
accompanying changes in fracture mode that
have been presented in the section "Creep
Rupture" in this article.
At moderately high strain rates (about 102
s - l ) , such as experienced during Charpy impact testing, the effect of strain rate is generally
similar to the effect of the state of stress,
namely that the bcc metals are more affected by
the strain rate than the fcc or the hcp metals.
Because essentially all strain rate tests at these
moderate strain rates are Charpy impact tests
that use a notched specimen, the effect of strain
rate is enhanced by the presence of the notch,
especially in steels when they are tested below
the transition temperature.
A moderately high strain rate either alters the
size and depth of the dimples or changes the
mode of fracture from dimple rupture to quasicleavage or intergranular decohesion. For example, when an AISI 5140 H steel that was
tempered at 500 °C (930 °F) was tested at
Charpy impact rates, it exhibited a decrease in
the width of the stretched zone adjacent to the
precrack and an increase in the amount of
intergranular decohesion facets (Fig. 64). The
same steel tempered at 600 °C (1110 °F)
showed no significant effect of the Charpy
impact test (Ref 199).
At very high strain rates, such as those
observed during certain metal-shearing operations, high-velocity (100 to 3600 m/s, or 330 to
11 800 ft/s) projectile impacts or explosive
rupture, materials exhibit a highly localized
deformation known as adiabatic* shear (Ref
*Adiabatic process is a thermodynamic concept where no
heat is gained or lost to the environment.
I
10 itm
I
Examples of hydrogen-embrittled steels. (a) Cleavage fracture in a hydrogen-embrittled annealed type 301 austenitic stainless steel. Source: Ref 98.
(b) Intergranular decohesive fracture in an AISI 4130 steel heat treated to an ultimate tensile strength of 1281 MPa (186 ksi) and stressed at 980 MPa (142 ksi)
while being charged with hydrogen. Source: Ref 111
Downloaded from https://dl.asminternational.org/handbooks/chapter-pdf/268655/a0001831.pdf
by Washington State University user
32 / Modes
of Fracture
la)
F |g./I.6
(al
I
5 ixm
I
(bl
Examples of hydrogen-embrittled titanium alloys. (a) Hydrogen embrittlement fracture in a Ti-8AI-1Mo-IV alloy in gaseous hydrogen. Note crack-arrest marks.
Source: Ref 137. (b) Cleavage fracture in hydrogen-embrittled Ti-5AI-2.5Sn alloy containing 90 ppm H. Source: Ref 141
(b)
I
10 ~m
I
(c)
Fig. 47
Influence of heat treatment and resulting microstructure on the fracture appearance of a hydrogenembrittled Ti-6A-4V alloy. Specimens tested in gaseous hydrogen at a pressure of 1 atm.
(a) Transgranular fracture in a specimen heat treated at 705 °C (1300 °F) for 2 h, then air cooled. (b) Intergranular
decohesion along ~-I~ boundaries in a specimen heat treated at 955 °C (1750 °F) for 40 mln, then stabilized.
(c) Coarse acicular structure resulting from heating specimen at 1040 °C (1900 °F) for 40 min, followed by stabilizing.
The relatively flat areas of the terraced structure are the prior-~ grain boundaries. See text for a discussion of the
microstructures of these specimens. Source: Ref 142
200-208). In adiabatic shear, the bulk of the
plastic deformation of the material is concentrated in narrow bands within the relatively
undeformed matrix (Fig. 65 to 67). Adiabatic
shear has been observed in a variety of materials, including steels, aluminum and titanium
alloys, and brass.
These shear bands are believed to occur
along slip planes (Ref 201, 202), and it has
been estimated that under certain conditions,
such as from the explosive-driven projectile
impact of a steel target, the local strain rate
within the adiabatic shear bands in the steel can
reach 9 × 105 s -~ and the total strain in
the band can be as high as 532% (Ref 204). An
estimated 3 × 106-s - j strain rate has been
reported for shear bands in a 2014-T6 aluminum alloy block impacted by a gun-fired (up
to 900 m/s, or 2950 ft/s) steel projectile (Ref
205).
The extremely high strain rates within the
adiabatic shear bands result in a rapid increase
in temperature as a large portion of the energy
of deformation is converted to heat. It has been
estimated that the temperature can go high
enough to melt the material within the bands
Downloaded from https://dl.asminternational.org/handbooks/chapter-pdf/268655/a0001831.pdf
by Washington State University user
(Ref 205,206). The heated material also cools
very rapidly by being quenched by the large
mass of the cool, surrounding matrix material;
therefore, in quench-and-temper hardenable
steels, the material within the bands can contain
transformed untempered martensite. This transformed zone is shown schematically in Fig.
65(b).
The hardness in the transformed bands is
sometimes higher than can be obtained by
conventional heat treating of the steel. This
increase in hardness has been attributed to the
additive effects of lattice hardening due to
supersaturation by carbon on quenching and the
extremely fine grain size within the band (Ref
203). However, for an A1SI 1060 carbon steel,
the hardness of the untempered martensite
bands was no higher than that which could be
obtained by conventional heat treating (Ref
206). In both cases, the hardness of the adiabatic shear bands was independent of the initial
hardness of the steel. For a 7039 aluminum
alloy, however, the hardness of the shear bands
was dependent on the hardness of the base
material. The adiabatic shear bands in an 80HV material exhibited an average peak hardness of about 100 HV, while those in a 150-HV
material had an average peak hardness of about
215 HV (Ref 208). For the Ti-6AI-4V STA alloy
shown in Fig. 66, there was no significant difference in hardness between the shear bands and
the matrix. In materials that do not exhibit a
phase transformation, or if the temperature generated during deformation is not high enough for
the transformation to occur, the final hardness of
Modes
Fig. 48
Hydrogen-embrittled 2124-UT aluminum alloy that shows no significant change in the fracture
appearance. (a) Not embrittled. (b) Hydrogen embrittled. Source: Ref 99
lal
Fig.
(b)
49
Effect of heat treatment on the fracture appearance of a hydrogen-embrittled low-copper 7050
aluminum alloy. (a) Transgranular cleavagelike fracture in a peak-aged specimen. (b) Combined
intergranular decohesion and transgranular cleavagelike fracture in an underaged specimen. Source: Ref 106
the adiabatic shear band is the net result of the
competing effects of the increase in hardness due
to the large deformation and the softening due to
the increase in temperature.
The width of the adiabatic shear bands depends on the hardness (strength) of the material
(Ref 206, 208). Generally, the harder the material, the narrower the shear bands. In a 7039
aluminum alloy aged to a hardness of 80 HV,
the average band width resulting from projectile
impact was 90 Ixm, while in a 150-HV material, the band width was only 20 Ixm (Ref 208).
The average width of the shear bands observed
in a Ti-6A1-4V STA alloy (average hardness,
375 HV~kg) was 3 to 6 txm.
When an adiabatic shear band cracks or
separates during deformation, the fractured
surfaces often exhibit a distinct topography
referred to as knobbly structure (Ref 205-208).
The name is derived from the surface appearance, which resembles a mass of knoblike
structures. The knobbly structure, which has
been observed in 2014-T6 and 7039-T6
aluminum alloys, as well as in an AISI 4340
steel (Fig. 67) and AISI 1060 carbon steel, is
believed to be the result of melting within the
shear bands (Ref 205, 206). Although the
cracked surfaces of adiabatic shear bands can
exhibit a unique appearance, adiabatic shear
failure is easiest to identify by metallographic,
rather than fractographic, examination.
Effect of T e m p e r a t u r e . Depending on the
material, the test temperature can have a significant effect on the fracture appearance and in
Downloaded from https://dl.asminternational.org/handbooks/chapter-pdf/268655/a0001831.pdf
by Washington State University user
of Fracture
/ 33
many cases can result in a change in the
fracture mode. However, for materials that
exhibit a phase change or are subject to a
precipitation reaction at a specific temperature,
it is often difficult to separate the effect on the
fracture due to the change in temperature from
that due to the solid-state reactions. In general,
slip, and thus plastic deformation, is more
difficult at low temperatures, and materials
show reduced ductility and an increased tendency for more brittle behavior than at high
temperatures.
A convenient means of displaying the fracture behavior of a specific material is a fracture
map. When sufficient fracture mode data are
available for an alloy, areas of known fracture
mode can be outlined on a phase diagram or can
be plotted as a function of such variables as the
test temperature and strain rate (Fig. 68). Similar maps can also be constructed for lowtemperature fracture behavior.
Effect of Low Temperature. Similar to the
effect of the state of stress, low temperatures
affect the bcc metals far more than the fcc or
hcp metal systems (see the section "Effect of
the State of Stress" in this article). Although
lower temperatures can result in a decrease in
the size and depth of dimples in fcc and hcp
metals, bcc metals often exhibit a change in the
fracture mode, which generally occurs as a
change from dimple rupture or intergranular
fracture to cleavage. For example, a fully
pearlitic AISI 1080 carbon steel tested at 125 °C
(255 °F) showed a fracture that consisted entirely of dimple rupture; at room temperature,
only 30% of the fracture was dimple rupture,
with 70% exhibiting cleavage. At - 1 2 5 °C
( - 1 9 5 °F), the amount of cleavage fracture
increased to 99% (Ref 210). This transition in
fracture mode is illustrated in Fig. 69.
Charpy impact testing of an AISI 1042 carbon steel whose microstructure consisted of
slightly tempered martensite (660 HV) as well
as one containing a tempered martensite (335
HV) microstructure at 100 °C (212 °F) and at
- 196 °C ( - 3 2 0 °F) produced results essentially
identical to those observed for the AISI 1080
steel. In both conditions, the fracture mode
changed from dimple rupture at 100 °C (212 °F)
to cleavage at - 1 9 6 °C ( - 3 2 0 °F), as shown in
Fig. 70. Similar changes in the fracture mode,
including a change to quasi-cleavage, can be
observed for other quench-and-temper and
precipitation-hardenable steels.
A unique effect of temperature was observed
in a 0.39C-2.05Si-0.005P-0.005S low-carbon
steel that was tempered 1 h at 550 °C (1020 °F)
to a hardness of 30 HRC and Charpy impact
tested at room temperature and at - 8 5 °C
( - 1 2 0 °F) (Fig. 71). In this case, the fracture
exhibited intergranular decohesion at room
temperature and changed to a combination of
intergranular decohesion and cleavage at - 8 5
°C ( - 1 2 0 °F). This behavior was attributed to
the intrinsic reduction in matrix toughness by
the silicon in the alloy, because when nickel is
substituted for the silicon the matrix toughness
34 / Modes of Fracture
(a)
I
25 ~m
I
(bl
Fig. 50
Stress-corrosion fractures in HY-180 steel with an ultimate strength of 1450 MPa (210 ksi). The steel
was tested in aqueous 3.5% sodium chloride at an electrochemical potential of E = - 0 . 3 6 to - 0 . 8 2
VsHE (SHE, standard hyd,r.ggen electrode). Intergranular decohesion is more pronounced at lower values of stress
intensity, K~ = 57 MPaVm (52 ksi~n~n.) (a), than at higher values, KI = 66 MPaN/'-mm(60 ksiN/~n.) (b). Source: Ref
154
room temperature to 600 °C (1110 °F) (Ref
213).
Figure 73 shows the effect of temperature on
the fracture mode of an ultralow-carbon steel.
The steel, which normally fractures by dimple
rupture at room temperature, fractured by intergranular decohesion when tensile tested
at a strain rate of 2.3 × 10 -2 s -L at 9 5 0 ° C
(1740 °F). The change in fracture mode was
due to the precipitation of critical submicronsize MnS precipitates at the grain boundaries.
This embrittlement can be eliminated by aging
at 1200 °C (2190 °F), which coarsens the MnS
precipitates (Ref 209).
A similar effect was observed for Inconel
X-750 nickel-base alloy that was heat treated by
a standard double-aging process and tested at a
nominal strain rate of 3 × 10 5 s ~ at room
temperature and at 816 °C (1500 °F). The
fracture path was intergranular at room temperature and at 816 °C (1500 °F), except that the
room-temperature fracture exhibited dimples on
the intergranular facets and those resulting from
fracture at 816 °C (1500 °F) did not (Fig. 74).
The fracture at room temperature exhibited
intergranular dimple rupture because the material adjacent to the grain boundaries is weaker
due to the depletion of coarse 3" precipitates.
The absence of dimples at 816 °C (1500 °F) was
the result of intense dislocation activity along
the grain boundaries, producing decohesion at
M 2 3 C 6 carbide/matrix interfaces within the
boundaries (Ref 214).
A distinct change in fracture appearance was
also noted during elevated-temperature tensile
testing of Haynes 556, which had the following
composition:
Element
(al
(b)
I
50 ~m
I
Fig. 51
Stress-corrosion fractures in a 25% cold-worked type 316 austenitic stainless steel tested in a boiling
(154 °C, or 309 °F) aqueous 44.7% magnesium chloride solution. At low (14 MPa~'-mm, or 12.5
ksi ~ - ~ ) KI values, t~:~ fracture exhibits a combination of cleavage and intergranular decohesion (a). At higher (33
MPaV m, or 30 ksiVin.) values of KI the principal mode of fracture is intergranular decohesion (b). Source: Ref 181
is increased and no cleavage is observed (Ref
211).
The temperature at which a sudden decrease
in the Charpy impact energy occurs is known
as the ductile-to-brittle transition temperature
for that specific alloy and strength level.
Charpy impact is a severe test because the
stress concentration effect of the notch, the
triaxial state of stress adjacent to the notch, and
the high strain rate due to the impact loading
combine to add to the reduction in ductility
resulting from the decrease in the testing
temperature. Although temperature has a strong
effect on the fracture process, a Charpy impact
test actually measures the response of a
material to the combined effect of temperature
and strain rate.
The effects of high temperature on fracture
are more complex because solid-state reactions, such as phase changes and precipitation,
are more likely to occur, and these changes
affect bcc as well as fcc and hcp alloys. As
shown in Fig. 72, the size of the dimples
generally increases with temperature (Ref 209,
212, 213). The dimples on transgranular
fractures and those on intergranular facets in a
0.3C- 1Cr- 1.25Mo-0.25V-0.7Mn-0.04P steel
that was heat treated to an ultimate strength
of 880 MPa (128 ksi) show an increase in
size when tested at temperatures ranging from
Downloaded from https://dl.asminternational.org/handbooks/chapter-pdf/268655/a0001831.pdf
by Washington State University user
Composition, %
Iron .................................
28.2
Chromium ............................
Nickel ...............................
Cobalt ...............................
Tungsten .............................
Molybdenum
..........................
21.5
22.2
19.0
2.9
2.9
Tantalum .............................
Manganese ............................
0.8
1.4
Silicon ...............................
Copper ..............................
Nitrogen .............................
0.5
0.1
0.1
Three specimens were tested at a strain rate of
approximately 1 S - 1 at increasin...
Purchase answer to see full
attachment