Fracture

 Fracture is related to propagation of cracks, leading to the failure of the

material/component.
 If there are no pre-existing cracks, then a crack needs to nucleate before propagation (to
failure). Crack nucleation$ typically requires higher stress levels than crack propagation.
 A crack is typically a ‘sharp*’ void in a material, which acts like a stress concentrator or
amplifier. Hence, crack is a amplifier of a ‘far field’ mean stress. (Cracks themselves do not
produce stresses!). [A crack is a stress amplifier !].
 Cracks in general may have several geometries. Even a circular hole can be considered as a
very ‘blunt’ crack. A crack may lie fully enclosed by the material or may have ‘crack faces’
connected to the outer surface. Cracks connected with outer surface may be profoundly influenced by the environment.
 Crack propagation leads to the creation of new surface area, which further leads to the
increase in the surface energy of the solid. However, in fracture the surface energy involved
(the fracture surface energy) is typically greater than the intrinsic surface energy as fracture
involves ‘sub-surface’ atoms to some extent. Additionally, the fracture surface energy may
involve terms arising out of energy dissipation due to micro-cracking, phase transformation
and plastic deformation.
A crack in a material
Fracture surface energy (f) > Intrinsic surface energy ()
$ Regions of stress concentrations (arising from various sources) ‘help’ in the process.
* More about this sooner
2a
 Fracture mechanics is the subject of study, wherein the a materials resistance to fracture is
characterized. In other words the ‘tolerance’ of a material to crack propagation is analyzed*.
 Crack propagation can be steady (i.e. slowly increasing crack length with time or load) or
can be catastrophic (unsteady crack propagation, leading to sudden failure of the material)$.
 ‘What dislocation is to slip, crack is to fracture’.
 Under tensile loading if the stress exceeds the yield strength the material, the material
begins to plastically deform. The area under the stress-strain curve is designated as the
toughness in uniaxial tension. Toughness relates to the energy absorbed to fracture.
 Similarly, in the presence of cracks we arrive at a material parameter, which characterizes
the toughness of the material in the presence of cracks→ the fracture toughness.
 In most materials, even if the material is macroscopically brittle (i.e. shows very little
plastic deformation in a uniaxial tension test), there might be some ductility at the
microscopic level. This implies that in most materials the crack tip is not ‘infinitely’ sharp,
but is blunted a little (blunting occurs by plastic deformation). This further avoids the stress
singularity at the crack tip as we shall see later.

$ One of the important goals of material/component design is to avoid Crack after crack tip
Sharp Crack (tip) blunting process
catastrophic failure. If crack propagation is steady, then we can
practice preventive maintenance (i.e. replace the component after
certain hours of service) → this cannot be done in the case of
catastrophic failure.
* Amongst its many other goals!
 The subject of Fracture mechanics has its origins in the failure of WWII Liberty ships. In
one of the cases the ship virtually broke into two with a loud sound, when it was in the
harbour i.e. not in ‘fighting mode’.
 This was caused by lack of fracture toughness at the weld joint, resulting in the propagation
of ‘brittle cracks’ (i.e. crack propagation will little plastic deformation). The full list of factors
contributing to this failure is in the figure below.
 It is seen that welding was done for faster production, but this resulted in micro-cracks and
residual stresses, which led to brittle crack propagation. The problem became ‘global’ as
this provided continuity of crack path across plates (so instead of one plate breaking the
entire ship ‘broke’). High sulphur in steel contributed to the brittleness of the plates.
 Due to the cold sea waters the ships were harboured in, the hull material underwent a
phenomenon known as ‘ductile to brittle transition (DBT)’ (about which we will learn more in this chapter).
 Ironically, this ‘death’ of ships lead to the ‘birth’ of fracture mechanics as a systematic field
of study.

Welding instead of riveting Continuity of the structure

Breaking
of Cold waters Residual stress
Liberty Ships
High sulphur in steel Microcracks
 What is a crack?

 As we have seen crack is an amplifier of ‘far-field’ mean stress. The sharper the crack-tip,
the higher will be the stresses at the crack-tip. It is a region where atoms are ‘debonded’ and
an internal surface exists (this internal surface may be connected to the external surface).
 Cracks can be sharp in brittle materials, while in ductile materials plastic deformation at the
crack-tip blunts the crack (leading to a lowered stress at the crack tip and further alteration
of nature of the stress distribution).
 Even void or a through hole in the material can be considered a crack. Though often a crack is
considered to be a discontinuity in the material with a ‘sharp’ feature (i.e. the stress amplification factor is large).
 A second phase (usually hard brittle phase) in a lens/needle like geometry can lead to stress
amplification and hence be considered a crack. Further, (in some cases) debonding at the
interface between the second phase and matrix can lead to the formation of an interface
cracks.
 As the crack propagates fresh (internal) surface area is created. The fracture surface energy
required for this comes from the strain energy stored in the material (which could further
come from the work done by externally applied loads). In ductile materials energy is also
expended for plastic deformation at the crack tip.
 A crack reduces the stiffness of the structure (though this may often be ignored).
Hard second phase in
Though often in figures the crack is shown to have a large lateral A crack in a material the material
extent, it is usually assumed that the crack does not lead to an
appreciable decrease in the load bearing area [i.e. crack is a local
2a
stress amplifier, rather than a ‘global’ weakener by decreasing
the load bearing area].
Characteristics of Cracks
Cracks can be characterized looking into the following aspects.
 Its connection with the external free surface:
(i) completely internal,
(ii) internal cracks with connections to the outer surfaces,
(iii) Surface cracks.
2a
 Cracks with some contact with external surfaces are exposed to
outer media and hence may be prone to oxidation and corrosion
(cracking).
 Crack length (the deleterious effect of a crack further depends
on the type of crack (i, ii or iii as above).
 Crack tip radius (the sharper the crack, the more deleterious it
is). Crack tip radius is dependent of the type of loading and the
ductility of the material. ~
 Crack orientation with respect to geometry and loading. We will
see modes of loading in this context soon.
a
Modes of deformation of a cracked body (modes of fracture)
How many ways are there to load a cracked body?
 Three ideal cases of loading of a cracked body can be considered, which are called the
modes of deformation:
 Mode I: Opening mode
 Mode II: Sliding mode
 Mode III: Tearing mode
 In the general case (for a crack in an arbitrarily shaped body, under an arbitrary loading), the
mode is not pure (i.e. is mixed mode). The essential aspects of fracture can be understood by considering mode I.

Mode I

Modes of Deformation /
fracture of a cracked body
Mode II

Mode III

Important note: the loading specified and the geometry of the specimen illustrated for Mode II & III above do not give rise
to pure Mode II and II deformation (other constraints or body shapes are required).
 ‘Early Days’ of the Study of Fracture
 C.E. Inglis  Stress based criterion for crack growth (local)
→ C.E. Inglis (seminal paper in 1913)
 Energy based criterion for crack growth (global)
→ A.A. Griffith (seminal paper in 1920)

‘Modern’ Fracture Mechanics

 Stress Intensity Factor (K)
Material Parameter  Fracture Toughness (KC)
 Energy Release Rate (G)
Material Parameter  Critical Energy Release Rate (GC)
 J-integral (J)
Material Parameter: JC
 Crack Tip Opening Displacement (CTOD) ()
Material Parameter: C
Crack growth and failure Brittle Materials

 Initially we try to understand crack propagation in brittle materials (wherein the cracks are
sharp and there is very little crack-tip plasticity). The is the domain of Linear Elastic
Fracture Mechanics (LEFM).
 For crack to propagate the necessary global criterion (due to Griffith) and the sufficient
Global vs. Local
local criterion (due to Inglis) have to be satisfied (as in figure below).
 The kind of loading/stresses also matters. Tensile stresses tend to open up cracks, while
compressive stresses tend to close cracks.

It should be energetically favorable

For crack growth to occur
Sufficient stress concentration should
exist at crack tip to break bonds

Griffith
Energy based Global
Crack growth criteria
Stress based Local
Inglis
Stress based criterion for crack propagation (Inglis criterion)
 In 1913 Inglis observed that the stress concentration around a hole (or a ‘notch’) depended
on the radius of curvature of the notch. I.e. the far field stress (0) is amplified near the hole.
[(max / 0) is the stress concentration factor ()].
 A ‘flattened’ (elliptical) hole (with a sharp tip) can be thought of as a crack. hole crack

 0 → applied “far field” stress σ max

 c  For sharp cracks c 
σ max  σ 0 1  2  σ  2 σ  max → stress at hole/crack tip 0
  
max 0

  → hole/crack tip radius
Sharper the crack, higher the stress concentration.  c → length of the hole/crack

 Sharper the crack (smaller the ) more the stress amplification (higher value of max). A
circular hole has a stress concentration factor of 3 [ = 3].
 From Inglis’s formula it is seen that the ratio of crack length to crack tip radius is important
and not just the length of the crack.

 One way of understanding this formula is that if max For a circular hole
exceeds t (the theoretical fracture stress), then the
 c
material fails (by the extension of the crack). σ max  σ 0 1  2 
 c
 This is in spite of the fact that the applied stress is of
much lower magnitude than the theoretical fracture =c σ max  3σ 0
stress.
 For a crack to propagate the crack-tip stresses have to do work to break the bonds at the
crack-tip. This implies that the ‘cohesive energy’ has to be overcome.
 If there is no plastic deformation or any other mechanism of dissipation of energy, the work
done (energy) appears as the surface energy (of the crack faces).
 The fracture stress (f) (which is the ‘far field’ applied stress) can be computed using this
approach. Note that the fracture stress is of the order of E (i.e. in GPa).

 f → fracture stress (applied “far-field”)

E E
 cohesive  f    → crack tip radius
 4a 0 c  c → length of the crack
 a0 → Interatomic spacing
Griffith’s criterion for brittle crack propagation
 We have noted that the crack length does not appear ‘independently’ (of the crack tip radius)
in Inglis’s formula. Intuitively we can feel that longer crack must be more deleterious.
 Another point noteworthy in Inglis’s approach is the implicit assumption that sufficient
energy is available in the elastic body to do work to propagate the crack.  (‘What if there is
insufficient energy?’)  (‘What if there is no crack in the body?’). Also, intuitively we can
understand that the energy (which is the elastic energy stored in the body) should be
available in the proximity of the crack tip (i.e. energy available far away from the crack tip
is of no use!).
 Keeping some of these factors in view, Griffith proposed conditions for crack propagation:
(i) bonds at the crack tip must be stressed to the point of failure (as in Inglis’s criterion),
(ii) the amount of strain energy released (by the ‘slight’ unloading of the body due to crack
extension) must be greater than or equal to the surface energy of the crack faces created.
 The second condition can be written as:
 Us → strain energy
 U → surface energy
dU s dU 

(Energy per unit area: [J/m2])
 dc → (‘infinitesimal’) increase in the
dc dc length of the crack (‘c’ is the crack length)
Essentially this is like energy balance (with the ‘=‘ sign) → the surface energy for the extended
crack faces comes from the elastically stored energy (in the fixed displacement case)
We look at the formulae for Us and U next.
 The strain energy released on the introduction of a very narrow elliptical double ended
crack of length ‘2c’ in a infinite plate of unit width (depth), under an uniform stress a is
given by the formula as below.

 c2 a2
Reduction in elastic energy  U  U without crack U with crack   Us 
E
 This is because the body with the crack has a lower elastic Should be written with a ve
energy stored in it as compared to the body without the crack sign if U = (Ufinal  Uinitial)

(additionally, the body with the crack is less stiffer). Also, the
assumption is that the introduction of a crack does not alter For now we assume that these stresses
the far-field stresses (or the load bearing area significantly). arise out of ‘applied’ displacements

 Notes:
 The units of Us is [J/m] (Joules per meter depth of the crack→ as
this is a through crack).
 Though Us has a symbol of energy, it is actually a difference
between two energies
(i.e. two states of a body→ one with a crack and one without).
 Half crack length ‘c’ appears in the formula.
 E is assumed constant in the process (the apparent modulus will decrease
slightly).
 a is the ‘far field’ stress (this may result from displacements
rather than from applied forces see note later).
 The formula for Us can be appreciated by considering the energy released from a circular
region of diameter 2c as in the figure below. (The region is cylindrical in 3D).
 The energy released is:


 c2  (1)
2
1
Elastic energy released from a circular region  Uscircular region  a
2 E
 The computation of the actual energy released is more involved and is given by the formula
as noted before:

 c2 a2
Us 
E Energy released from this circular region
is given by the formula (1) as above
Plane stress condition (not a true value, but to get a feel of the
predominant region involved).

U s 2 c a2
Hence 
c E

 For a body in plane strain condition (i.e. ~ thick in the z-direction, into the plane of the
page), E is replaced with E/(12):

 c2 a2 As plane strain is more severe on the

Us 
E (1  2 ) Plane strain condition material it is better to do experiments in
plane strain condition.
 The surface energy of the crack of length 2c & unit width/depth is:
U
Fracture surface energy  U   4 γ f c [J/m]  4 f
c

 This is the difference in the energy between a body with a crack and one without a crack.
 As pointed out before, the surface energy is the fracture surface energy and not just the
surface free energy. The origin of this energy is contributions from dissipative mechanisms
like plastic deformation, micro-cracking & phase transformation, in addition to the energy
of the ‘broken bonds’.
 The units are Joules per meter depth of the body: [J/m].

Important note

 The “Griffith experiment” is easily understood in displacement control mode (i.e. apply a
constant displacement and ‘see’ what happens to the crack) and is more difficult to
comprehend it in the force control mode (by applying constant ‘far-field’ forces).
 In force control mode, the forces do work on the system and hence the ‘energy accounting’
process is more involved.
 Hence, it is better to visualize  as arising from ‘far field’ applied displacements.
Understanding Griffith’s equation
dU s dU  2 c  2

 According to Griffith’s criterion:   a

 4 f
dc dc E
*
At criticality (written with ‘*’) crack propagation just starts:  2 c a2 
  4 f 
 E 
Putting the ‘*’ around the variables: 2 c 2 *  4 E
 a f
  c  *
 * 2
a  2E f
2 E 2 E f
Since, * is f :  c   2E f
* 2
Which, can be written in two ways:  f  c* 
f  c*  2f

 This can be understood as follows: keep displacement imposed on the ends of the specimen
constant (& hence far field) and keep increasing ‘c’ till the crack beings to propagate (& hence
far field = f).
 Else, one can keep ‘c’ constant and increase displacements (leading to an increase in far field)
till crack propagation starts (i.e. c  c*,   f).

2 E 2 E
Griffith  f 
At constant c (= c*) f 
when  exceeds f then specimen fails  c* (1  2 )
 c*
Plane strain conditions
 ‘Modern’ Fracture Mechanics

 Linear Elastic Fracture Mechanics (LEFM)

 Stress Intensity Factor (K)
Material Parameter  Fracture Toughness (KC)
 Energy Release Rate (G)
Material Parameter  Critical Energy Release Rate (GC)

 Elastic-Plastic Fracture Mechanics (EPFM)

 J-integral (J)
Material Parameter: JC
 Crack Tip Opening Displacement (CTOD) ()
Material Parameter: C
 Linear Elastic Fracture Mechanics (LEFM)
 A fundamental question in linear elastic fracture mechanics (LEFM) is: what is the state
of stress around the tip of a crack in a linear elastic solid?
 This question has been answered by a number of investigators including Westergaard,
Sneddon and Williams. Westergaard (1939) introduced a stress function, now known as
Westergaard stress function, to solve linear elastic crack problems.
 Williams (1952) showed that the stress singularity around a crack, regardless of the
actual configuration, is of the inverse square root type.

Mode I Mode II
Stress fields at crack tips
 For a body subjected far field biaxial stress 0, with a double ended crack of length 2c, the
stress state is given by (this is mode-I loading):

KI       3  
 xx  Cos   1  Sin   Sin    (1)
2 r  2  2  2 
KI       3  
 yy  Cos   1  Sin   Sin    (2)
2 r  2  2  2  Fig.1

KI        3 
 xy  Cos  Sin   Sin   (3)  ij 
1
2r 2 2  2  r

 Note the inverse square root (of r) singularity at the crack tip. The intensity of the
singularity is captured by KI (the Stress Intensity Factor). I.e. KI is the scaling factor for the singularity.
 As no material can withstand infinite stresses (in ductile materials plasticity will intervene),
clearly the solutions are not valid exactly at (& ‘very near’) the crack tip.
 At  = 0 and r → the stresses (xx & yy) should tend to 0. This is not the case, as seen
from the equations ((1) & (2)). This implies that the equations should be used only close to
crack tip (with little errors) or additional terms must be added to the equations.
Understanding the stress field equation

KI        3   f ( )
 xx  Cos 1  Sin  Sin    →  xx  K I (1)
2r  2   2   2  2r
 xx  K I f ( r, )

K I  Y  0 c  Half the crack length (for a fully internal crack)

Indicates mode I ‘loading’

‘Shape factor’ related to ‘Geometry’

 “KI (the Stress Intensity Factor) quantifies the magnitude of the effect of stress singularity at
the crack tip”[1]. K has units of [N/m2.m] or [Pam] or typically [MPam].
 Quadrupling the crack length is equivalent to doubling the stress ‘applied’. Hence, K
captures the combined effect of crack length and loading. The remaining part in equation(1)
is purely the location of a point in (r, ) coordinates (where the stress has to be computed).
 Note that there is no crack tip radius () in the equation! The assumptions used in the
derivation of equations (1-3) are:   = 0,  infinite body,  biaxial loading.
 The factor ‘Y’ is considered in the next page.

[1] Anthony C. Fischer-Cripps, “Introduction to Contact Mechanics”, Springer, 2007.

The Shape factor (Y)
 It is obvious that the geometry of the crack and its relation to the body will play an
important role on its effect on fracture.
 The factor Y depends on the geometry of the specimen with the crack.
 Y=1 for the body considered in Fig.1 (double ended crack in a infinite body).
 Y=1.12 for a surface crack. The value of Y is larger (by 12%) for a surface crack as
additional strain energy is released (in the region marked dark grey shade in the figure
below), due to the presence of the free surface.
 Y=2/ for a embedded penny shaped crack.
 Y=0.713 for a surface half-penny crack.
Fracture Toughness (Irwins’s K- Based)
 One of the important goals of fracture mechanics is to derive a material parameter, which
characterizes cracks in a material. This will be similar to yield stress (y) in a uniaxial
tension test (i.e. y is the critical value of stress, which if exceeded (  y) then yielding
occurs).
 The criterion (critical condition) for fracture in mode-I can be written as:

K I  K IC Where, KIC is the critical value of stress intensity factor (K) and is known
as Fracture Toughness
 KIC is a material property (like yield stress) and can be determined for different materials
using standard testing methods. KIC is a microstructure sensitive property.
 The focus here is the ‘local’ crack tip region and not ‘global’, as in the case of Griffith’s
approach.
 All the restrictions/assumptions on K will apply to KIC:
(i) material has a liner elastic behaviour (i.e. no plastic deformation or other non-linear
behaviour),
(ii) inverse square root singularity exists at crack tip (eq. (1)),
(iii) the K-dominance region characterizes the crack tip.

f ( )
 xx  K I (1)
2r
Fracture Toughness (KIC) for some typical materials

Material KIC [MPam]**

Cast Iron 33
Low carbon steel 77
Stainless steel 220
Al alloy 2024-T3 33
Al alloy 7075-T6 28
Ti-6Al-4V 55
Inconel 600 (Ni based alloy) 110
 Is KIC really a material property like y?

 Ideally, we would like KIC (in mode-I loading) (KIIC & KIIIC will be the corresponding
material properties under other modes of loading$) to be a material property, independent of
the geometry of the specimen*. In reality, KIC depends on the specimen geometry and
loading conditions.
 The value KIC is especially sensitive to the thickness of the specimen. A thick specimen
represents a state that is closer to plane strain condition, which tends to suppress plastic
deformation and hence promotes crack growth (i.e. the experimentally determined value of
KIC will be lower for a body in plane strain condition). On the other hand, if the specimen is
thin (small value ‘t’ in the figure), plastic deformation can take place and hence the
measured KIC will be higher (in this case if the extent of plastic deformation is large then KI
will no longer be a parameter which characterizes the crack tip accurately).

 To use KIC as a design parameter, we have to use its ‘conservative

value’. Hence, a minimum thickness is prescribed in the standard
sample for the determination of fracture toughness.
 This implies that KIC is the value determined from ‘plane strain tests’.

$ Without reference to mode we can call it KC.

* E.g. Young’s modulus is a material property independent of the geometry of the specimen, while stiffness is the equivalent ‘specimen geometry
dependent’ property..
Summary of Fracture Criteria
 The table below summarizes many criteria found in standard literature. Some details can be
found in other pages of the chapter.
Fracture
Criterion Comments Relevant formulae
occurs if
Inglis (f) Involves crack tip radius (& crack E
  f f 
1913 length) 4a 0 c

Griffith (f) 2 E
Involves crack length   f f 
1920  c*
Irwin [K]
Concept of stress intensity factor.
KI  KIC
(in mode I)

K I  Y  0 c 
Energy release rate based. Same as
Irvin [G] G  GC
K based criterion for elastic bodies.
Wells (, CTOD) Involves crack (tip) opening  K I2 
  C     * * 
 E y 
1961 displacement

Rice [J-integral] Generalized energy release rate J  JC HRR fields.

concept. Applicable to non-linear (crack initiation)
1968 elastic and elastic-plastic materials.