Professional Documents
Culture Documents
number of cleavage steps, which may join and form a river pattern. The
convergence of river patterns is always downstream; this gives a
possibility to determine the direction of local crack propagation in a
micrograph. Examples of cleavage steps and river patterns, as observed by
electron fractography. arc presented in figure 2.10. They arc places where
small scale plastic deformation is likely to occur. Plastic deformation
requires energy, and therefore river patterns and steps are observed more
abundantly on cleavage fractures produced at temperatures close to the
transition temperature.
A cleavage tongue is another typical feature of a cleavage fracture. It is
called a tongue because of its apparent shape. Cleavage longues of various
—* [no]
Figure 2.12. Formation of cleavage tongue (BCD) due to
passage of twin. Cul along (110) plane through a coherent twin
in bec lattice
1 36
7 ^Mechanisms of fracture and crack growth
2
2 2.3 Ductile
Mechanisms of fracture /nature
and crack
growth
i
n
el
e
m
e
nt
s.
S
o
m
et
i
m
es
p
ar
ti
cl
es
of
th
is
si
z
e
ar
e
es
se
nt
ia
l
fo
r
th
e
pr
o
p
er
I
ie
i
of
th
e
m
at
er
ia
l,
as
2.3 and
2 Mechanisms of fracture Ductile fracture
crack
growth
4 41
2 Mechanisms oj fracture and crack growth
!IM 1 111........................................................................................F 11! i n i' 11 ■ i p 11' i F11111 ■ir 11 M r 11 r 11' 1! 11 n f pi r 11 p > i' i r' i r ■ i ■ i r ■ I r' r ■ l r ■ r r' r r i r i T ' i r ■ i r i r' i r r i r 11 r i r 11 n I r i r: i n r r l r! I r 11 M r 11 p i r M r i; M i [; r 11 [ 111 ir 11 r 11 M 111 r i' 11 [ 11
........IIIMIIIIIIIIIIIMIMIIIMIIIIIIMIIIIIIII...............................................................
t t t t
M i l
■UUUt
.
i t i __________________
■— ___— - —^
^opposite direction
shear dimples
eguiaxed dimples
l same direction
tear dimples
1 I * by stereoscopic measurements of dimple topography. Apparently, voids
grow mainly in lateral directions, causing them to remain shallow. In oxide-
igurc 2. IK. Occurrence of different types of dimples
dispersion strengthened materials the initiation and growth of voids at the
dispersed particles can be made visible when a thin foil is subjected to a
tensile deformation while under observation in the electron microscope [35.
[16]. Thus it can be asserted from figure 2.20 that the depth-to-width ratio 36, 37]. In normal structural materials voids at the intermediate particles
is low and that dimples are relatively shallow holes. The latter is confirmed can seldom be observed. In a study of 13 different aluminium alloys Broek
5 44
2 Mechanisms oj fracture and crack growth
[4, 16. 29] showed the occurrence of voids at the intermediate particles:
slender particles have a tendency to fracture (figure 2.21) while particles of
other shapes lose coherence with the matrix (figure 2.22). However, in
general only very few voids were found.
Apparently, the cohesive forces between the matrix and particles are
extremely large in conventional materials. They are so large that void
6 44
2.3 growth
2 Mechanisms of fracture and crack Ductile fracture
|
|
5
7
2.3 growth
2 Mechanisms of fracture and crack Ductile fracture
Initiatio of many observable voids leads to the assumption that immediate perpendicular to the tensile stress, as in the ease of a crack This lateral void
n does coalescence must take place at the moment that voids are formed in some growth is confirmed by the shallowness of the dimples A model that is
not take quantity. This implies that voids can only be initiated at such high stresses HI reasonably compatible with the observed facts is the following[16, 29].
place strains that the conditions for coalescence are already fulfilled. This requires During plastic deformation dislocation pile-ups will form at the particles.
until a model for void initiation which predicts immediate and spontaneous These piled-up loops are depicted in figure 2.23a, TllC loops arc repelled by
very late void.growth. (Where void growth in oxide-dispersion strength- i tud the particle through the action of their image forces On the other hand, the
in the materials was reported, it occurred in the direction of the tensile stress. leading loop will be pushed towards the parlicl< by stresses set up by the
fracture This mode of void growth is a normal consequence of longitudinal Straining, pile-up and the applied shear stress. When one or a couple of loops are
process. which would also occur under elastic conditions). From a physical point of pushed to the interface a dccohcsion ol ilu
The luck view, one would expect a cavity to spread preferentially in a direction
4S
5
8
ll l il l l l l l ll ll l l l l l ll ll l l l l l ll ll l l l l l ll ll l l l l l ll ll l l l l l ll ll l l l l l li ...
Under the action of cyclic loads cracks can be initiated as a result of cyclic
plastic deformation [38, 39]. Even if the nominal stresses are well below the
I igiire 2 2: IHxvIksion of particles 111 MK>I-\l-«llo>. Top elastic limit, locally the stresses may be above yield due to stress
hull and lioltom half show live same localion. taken al ilillcrenl concentrations at inclusions or mechanical notches. Consequently, plastic
angles of llic election beam [2')\(courtesy I'ergamim) deformation occurs locally on a microscale, but it is insufficient to show in
engineering terms.
interface will ultimately lake place. If this occurs, a void is formed. The
consequence is that the repelling forces on subsequent loops are drastically
reduced "and the greater part of the pile-up can empty itself into the newly
formed void. The dislocation sources behind the loops, which became
inactive because of the constraint of the pile-up ahead, can resume action
and hence the process may lead to unstable lateral void growth and
coalescence as soon as the voids have been initiated (figure 2.23c, d). In
figure 2.24 the model is shown in terms of displacements.
Contrary to cleavage, where the action of a tensile stress is sufficient for
the separation, ductile fracture cannot occur without plastic deformation.
The mechanism of final separation is a direct consequence of dislocation
movements and slip displacements necessary for the growth and
coalescence of voids. Apart from a stress to induce dislocation movement a
certain plastic strain is required for ductile separation to occur. This plastic
deformation may be confined to a small volume of material through which
the fracture passes. Then failure occurs with relatively little plastic-
deformation on a macroscale. requiring only little energy. The fracture is
brittle in an engineering sense. Fractures induced by cracks in high
strength materials are usually of this type.
9
10.4 growth
2 Mechanisms of fracture and crack Fatigue cracking
49
11.4 growth
2 Mechanisms of fracture and crack Fatigue cracking
49
2.4growth
2 Mechanisms of fracture and crack Fatigue cracking
I........IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIM
opening
opening
52 51
2 2,14 Fatigue
Mechanisms of fracture cracking
and crack
growth
54 53
2 Mechanisms o f fracture and
crack growth
m.....I
.........
iiiiiiniiiiiiitmiiiiiiiilHIIIIIIIIIIllllllllllllll
llllllllllllll
56 55
2 Mechanisms o f fracture and
crack growth
56 55
2 2,4 Fatigue
Mechanisms of fracture cracking
and crack
growth
58 57
2 2,4 Fatigue
Mechanisms of fracture cracking
and crack
growth
i.iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii
58 57
2 Mechanisms2.5 Environment
oj fracture assisted cracking
and crack
growth
[58. 69] (in the order of I micron per cycle and above). This is obvious
from figure 2.33. High growth rates are a result of a high stress intensity at
the crack tip (large crack or high loads). Due to the higher stress
concentration, particles in front of the crack tip may cleave or lose
coherence with the matrix, thus initiating a (large) void. The remaining
material between the void and the crack tip now may rupture by ductile
tearing, thus producing a local jump of the crack front. This is obvious
from the areas with dimples in figure 2.33. which are evidence of a
mechanism of void coalescence during ductile rupture.
Al these high propagation rates the effect of inclusions cannot be
neglected. Comparison of the striation spacing with the growth rate
observed in the test [58] reveals a discrepancy as a result of small amounts
of static fracture. At still higher propagation rates striations become very
rare and the fracture surface consists primarily of dimples. One may
conclude that the growth rate would have been much smaller in the absence
of the inclusions. Neglecting inclusions, the "true fatigue" crack
propagation ralc(by striation formation) would have been about 0.5 micron
per cycle in figure 2.33 instead of I micron per cycle actually observed in
the test.
The influence of particles on fatigue crack propagation is limited to high
crack propagation rates. This means that it is limited to the very last and
small part of the crack propagation life. Consequently, it is not very
important technically. This is confirmed by tests on materials with very low
particle content [59].
Another way in which cracks can be initiated and grow at low stress is
environmental cracking, which is considered here in its broadest sense, A
liquid metal environment may cause cracking even under zero stress. A
corrosive environment which would not normally attack the metal may
cause cracking under the assistance of mechanical stresses. Several theories
have been put forward to explain this stress corrosion cracking, yet its
mechanism is far from well understood. In particular, the role of the
mechanical stresses is difficult to comprehend. It seems inconceivable that
any single theory is likely to explain all observations and it seems
reasonable that different mechanisms operate under different conditions and
in different materials. In view of this, the discussion of stress corrosion
cracking will be limited to it being mentioned as a mechanism for cracking.
In many materials stress corrosion cracks are intergranular, which may
be due to a potential difference between the grain boundary and the interior
of the grains as a result of a segregation of solute. Alternatively, it may be
attributed to the presence of second phase particles at the grain boundaries.
The typical appearance of intergranular stress-corrosion fracture surfaces is
shown in figure 2.34. Fracture surfaces resulting from
Figure 2.34. Intergranular stress corrosion fracture surface in 707') Al-alloy (courtesy
Van Leeuwen)
60 59
2.5 Environment assisted cracking 2 Mechanisms of fracture and crack growth
The present interest in fracture mechanics partly stems from the fact that
service failures still take place. Proper engineering applications of fracture
mechanics may improve the situation, but service failures will continue to
occur. A thorough investigation of service failures provides experience for
other cases and it can give information on the shortcomings of the applied
fracture mechanics approach.
Part of service failure analysis is the electron fraclography. As has been
shown in the previous chapters there are a number of very distinctive
fractographic features that allow distinction between the various fracture
mechanisms. Inventories [9-13] of the primary fracture characteristics and
of many secondary fracture features are very useful to service failure
62
22 Chapter 1 Introduction to Fracture Mechanics
- U
result in an equation of the following form for the yield locus in the fourth quadrant:
(1.5d)
By a similar argument the yield locus equation for the second quadrant can be constructed. The
resulting yield locus for plane stress is shown in Figure 1.5c. As with the previously described
failure theories, failure is assumed to occur when the normalized stresses are outside of the shaded
region.
An alternative graphical description of the Coulomb-Mohr theory can be constructed by
considering the limiting uniaxial cases on a Mohr circle plot. According to the Mohr theory, all
possible states of allowable stress can be represented by Mohr's circles bounded by the limiting
circles of uniaxial tension and uniaxial compression and the straight lines corresponding to points
connecting the circles' common tangent. The Mohr's circle representation of the Coulomb-Mohr
failure theory is shown in Figure 1.6.
fracture as a single event. However, from theoretical considerations, the forces required to
accomplish this feat are very large—much larger than the measured tensile strength of brittle
materials. More realistically, the measured forces are more closely related to the force required to
separate a few lines of atoms at a time. This perspective of fracture by progressive separation of
the fracture plane is fundamental to the development of a modern fracture theory.
A series of experiments conducted at the Naval Research Laboratory in Washington, DC, in
the late 1940s on crack propagation behavior in zinc foils has provided important insight into the
mechanism of progressive fracture in metals. In these studies, a razor blade was used to introduce
an initial crack into a thin sheet of zinc foil loaded in tension. The load was slowly increased until
slow extension of the crack occurred. By using back lighting, it was observed that small holes
formed ahead of the crack tip, and the fracture proceeded by successive joining of these advanced
voids. This process, now called "void coalescence," is characteristic of crack propagation in most
metallic materials. Because of the discontinuous nature of this type of crack propagation process
—that is, the joining of discrete holes by isolated events—the fracture surface has a rough
appearance when viewed on a fine scale. This type of fracture is also called "fibrous" because of
the stretching of the ligaments between holes prior to their linking up. These "fibers" are readily
observed under an electron microscope, as illustrated in Figure 1.7. This fracture mechanism has
associated with it a relatively large amount of dissipated energy related to the plastic work done as
the crack progresses.
23 Chapter 1 Introduction to Fracture Mechanics
■
r-3
906 In contrast to the fibrous appearance associated with ductile fractures are the smooth, flat
�22 fracture surfaces characteristic of crack propagation associated with brittle or semibrittle
4655 materials. This mechanism, called "cleavage" fracture, is characterized by the progressive
�06
separation of bonds between atoms or molecules as a result of the high strain levels that exist at
8
Figure 1.8
Cleavage fracture
surface in tungsten
[courtesy C. D.
Beachem, RE.].
the crack tip. The fracture surface associated with this type of fracture is often mirrorlike in
brittle, nonmetallic materials. Although characteristic of fractures in nonmetals, cleavage fractures
are also observed in otherwise ductile metals in the presence of lower temperatures and weak
initiation sites, as shown in Figure 1.8. In marked contrast to the large energy requirements of
fibrous fracture, cleavage fractures proceed with only small amounts of energy required for each
increment of crack extension.
From the preceding discussion it would appear natural (in hindsight) to describe fracture in
terms of the energy dissipated per unit extension of the crack. This description was the view of
the mechanics of fracture first described by Irwin [1948] and developed in simplistic form by
Irwin and Kies [1952,1954]. These papers developed the concept of energy release during
progressive fracture and provided a means to measure the material's resistance to fracture using
the compliance approach. We can gain an understanding of this concept by considering a
hypothetical experiment. Imagine a body containing an initial crack of area, A, that is subjected to
a slowly increasing load. If the body is linearly elastic, the load-deflection behavior is given by
the line O B shown in Figure 1.9.
After the cracked body has been loaded to point B, the body is fixed so that no deflection of
the load points is possible. This condition, called "system isolated," is
Section 1.3 Crack Extension Behavior 25
a.
a.
CO
necessary in the present discussion to ensure that no energy is lost due to potential energy changes
through the load points. (We will relax this requirement in the more general discussion of energy
concepts in Chapter 7.) Imagine now that some process results in a change in crack area to A + A A.
Since the body is now less stiff, the load must drop to point C in Figure 1.9. If we now release the
constraint at the load point and remove the applied load, the load-deflection diagram must return to
zero along path O C . The triangle O B C represents the energy lost during the load-unload process;
however, because the "system isolated" condition prevented any loss of-energy" through load points,
all of this energy must be related to the crack extension area, A A .
Following Irwin, we will define a quantity called the strain energy release rate, Gj, as the spatial
rate of change of stored strain energy under system isolated conditions, i.e.,
(1.6)
where U is the strain energy of the system, and the negative sign is introduced to make (g a positive
quantity.
For the linear, elastic conditions of our experiment,
U
S (1.7)
but
S c p (1.
8)
where C, called the compliance, is the reciprocal of the slope of the load-deflection line. Since C is
a measure of the stiffness of the body, it is a function of the crack length. Substituting Eq. (1.8) into
Eq. (1.7) and differentiating with respect to A, we obtain the result
2 3A
In deriving equation (1.9), no restrictions were placed on the form of the specimen, and, as a result,
the equation can be applied to any geometrical shape. Accordingly, the compliance calibration
method can be used to develop a fracture mechanics measure for any specimen geometry. This was
the approach taken by Irwin and Kies to develop suitable specimens and to compute the fracture
mechanics parameter, (gc, that formed the basis for the early confirmation of linear elastic fracture
mechanics.
1.4 SUMMARY
26 Chapter 1 Introduction to Fracture Mechanics
Spurred on by several major failure events and by an increasing awareness that classical failure
theories did not predict observed phenomena (such as the effect of size on failure strength), the
development of a new theory of fracture was inevitable. Building on ideas developed by G. R.
Irwin and his coworkers at the Naval Research Laboratory, the new engineering discipline of linear
elastic fracture mechanics quickly spread throughout the 1950s to interested researchers,
particularly in the aircraft, aerospace, and electric utility industries, in both the United States and
England.
Irwin's generalization of the Griffith argument of an energy-related fracture criterion provided
a convenient starting point for the development of a much more comprehensive theory of fracture.
However, calculation of energy changes is not reiwa^^practical and provides no insight into the
behavior of the material near the ^crack*iip>Further, although phenomenologically sound, the
energy approach has an intangible. Quality that makes it hard to visualize. For these and other
reasons, the focus of attention in the theory of fracture mechanics in the 1960s was on the stress
field approach to LEFM. From this viewpoint, the theory of fracture is treated as a subset of the
engineering discipline of solid mechanics. In particular, fracture mechanics is a study of the
mechanics of deformable bodies containing crack-like singularities. In order to develop the concept
of fracture mechanics from this perspective, it is first necessary to master a working knowledge of
linear elasticity. Accordingly, in the next few chapters, we will first treat those aspects of solid
mechanics necessary to develop a theory of fracture based on the stress analysis approach and then
go on to formulate a suitable theory and study its consequences.
REFERENCES
ASTM, 1960, "Fracture Testing of High-Strength Sheet Materials (A Report of a Special ASTM
Committee)," ASTM Bulletin, No. 243, January 1960, pp. 29^0, and No. 244, February 1960,
pp. 18-28.