Chapter 9b Plasticity
Chapter 9b Plasticity
Chapter 9b Plasticity
Modes of Deformation
The Uniaxial Tension Test
Mechanisms underlying Plastic Deformation
Strengthening mechanisms
MATERIALS SCIENCE
Part of & A Learner’s Guide
ENGINEERING
AN INTRODUCTORY E-BOOK
Anandh Subramaniam & Kantesh Balani
Materials Science and Engineering (MSE)
Yield Criteria Indian Institute of Technology, Kanpur- 208016
Email: [email protected], URL: home.iitk.ac.in/~anandh
http://home.iitk.ac.in/~anandh/E-book.htm
Mechanical Metallurgy
George E Dieter
McGraw-Hill Book Company, London (1988)
Let us start with some observations… Revise: the mechanisms by which materials fail
An Al rod when bent through a large angle does not come back to its original shape.
Steel is more difficult to deform as compared to Al.
A steel piece is easier to deform when heated (as compared to when it is cold).
‘Chinese’ Clay when deformed does not regain its original shape.
‘Silly putty’ deforms like Chinese clay when slowly deformed. However, when one bounces
a ball of silly putty it bounces like a rubber ball.
Plastic deformation of crystalline materials takes place by mechanisms which are very
different from that for amorphous materials (glasses). The current chapter will focus on plastic
deformation of crystalline materials. Glasses deform by shear banding etc. below the glass transition temperature (T g)
and by ‘flow’ above Tg.
Though plasticity by slip is the most important mechanism of plastic deformation, there are
other mechanisms as well. Many of these mechanisms may act in conjunction/parallel to give
rise to the observed plastic deformation. Twinning is perhaps the second most important one.
Phenomenological terms
Shear
Torsion
It is important to note that these are macroscopically defined with respect to a body of given geometry (even in tensile loading inclined
planes will be subjected to shear stress)
Bending
Deformed configuration
Note: modes of deformation in other contexts will be defined in the topic on plasticity
Peak ahead
In addition to the modes of deformation considered before the following modes can be
defined w.r.t fracture.
Fracture can be cause by the propagation of a pre-existing crack (e.g. the notches shown in
the figures below) or by the nucleation of a crack during deformation followed by its
propagation.
In fracture the elastic energy stored in the material is used for the creation of new surfaces
(when the crack nucleates/propagates).
Mode I
Modes This is ‘shearing’ mode
of Mode II
Deformation
Mode III
This is ‘tearing’ mode
Path to understanding plasticity
The following aspects need to be understood to comprehend plasticity*:
External process parameters (Temperature, strain rate, etc.)
Macroscopic and Microscopic aspects of plasticity
Continuum and Discrete views of plasticity
Plasticity in single crystals
Plasticity in polycrystals
Plasticity in multiphase materials
Plasticity in nanomaterials
P L 0 → initial
s e Subscript Note: quantities obtained by performing an
A0 L0 i → instantaneous Uniaxial Tension Test are valid only under
uniaxial state of stress
* Though the hardness test is even simpler, the data is convoluted with various effects.
The Tensile Stress-Strain Curve
The results of a uniaxial tension test can be plotted as in the
Load →
figures to the right.
It is important to note that we prefer stress-strain and not load-
stroke as the axes.
Stroke →
Tensile specimen
s →
elongation
Gauge Length → L0
→
Initial cross sectional area → A0
Important Note
We shall assume cylindrical polycrystalline specimens (unless otherwise stated) →
The experimental setup for the uniaxial tension test
Grip
Cross-head
Sample
Zoomed-in view
N
Zoomed-in view
Sample after the formation of a neck
The grip should be such that the uniaxial loading is maintained (i.e. there is no shear/bending component to the loading)
Problem with engineering Stress (s) and Strain (e)!!
Consider the following sequence of deformations:
1 L0
e1→2 = 1
e2→3 = ½
3 L0
1 L0
1→2 = Ln(2)
2→3 = Ln(2)
3 L0
With true strain things turn out the way they should!
True strains are additive, engineering strains are not.
Schematic s-e and - curves Schematics: not to scale These are simplified schematics which are close to the
Polycrystalline Specimen (e.g. Al) curves obtained for some metallic materials like Al, Cu
etc. (polycrystalline materials at room temperature).
Many materials (e.g. steel) may have curves which are
qualitatively very different from these schematics.
Most ceramics are brittle with very little plastic
deformation.
These diagrams are not to scale as the strain at yield is
~0.001 (eelastic ~10–3)
[E is measured in GPa and y in MPa thus giving
this small strains]
the linear portion is practically vertical and stuck to
the Y-axis (when efracture and eelastic are drawn to the
same scale).
Linear elastic material (obeys Hook’s law) Note the increasing stress required for
continued plastic deformation
(the stress to cause continued plastic
deformation is called flow stress)
Points and regions of the curves are explained in the next slide
* It is better to determine the Young’s modulus from sound propagation experiments,
than from UTT experiments.
Sequence of events during the tension test
O unloaded specimen
OY Elastic Linear Region in the plot (macroscopic linear elastic region)
Y macroscopic yield point (there are many measures of yielding as discussed later)
Occurs due to collective motion of many dislocations finally leaving the grain boundary or crystal surface.
The stress at this point is called yield strength. [i.e. stress strength]
YF Elastic + Plastic regime
If specimen is unloaded from any point in this region, it will unload parallel to OY and the elastic
strain would be recovered. Actually, more strain will be recovered than unloading from Y (and
hence in some sense in the region YF the sample is ‘more elastic’ than in the elastic region OY).
In this region the material strain hardens flow stress increases with strain.
This region can further be split into YN and NF as below.
YN Stable region with uniform deformation along the gauge length
N Plastic Instability in tension Onset of necking
True condition of uniaxiality broken onset of triaxial state of stress (loading remains uniaxial
but the state of stress in the cylindrical specimen is not).
NF most of the deformation is localized at the neck
Specimen in a triaxial state of stress
F Fracture of specimen (many polycrystalline materials like Al show cup and cone fracture)
Notes:
In the - plot there is no distinct point N and there is no drop in load (as instantaneous area has been taken into account in the definition of ) in the elastic
+ plastic regime (YF).
The stress is monotonically increasing in the region YF true indicator of strain hardening.
Information which can be obtained from a uniaxial tension test
(i) Young’s modulus*→ slope of the OY (elastic part of the curve).
(ii) Yield stress (or proof stress)→ stress corresponding to point Y.
(iii) Ultimate Tensile Stress (UTS) → point N (maximum) in s-e plot.
(iv) Fracture stress→ stress corresponding to point F.
(v) Ductility→ measured as: (a) strain at fracture (in %), (b) % reduction in area.
(vi) Resilience (area under the curve elastic portion- OY).
(vii) Toughness (area under the curve total)→ has unit of Energy/volume [J/m3].
(viii) Strain hardening exponent (from - plot).
In a UTT any inclined plane (i.e. except the vertical and horizontal planes) all planes feel
shear. If a slip plane (within a grain) is oriented in such a way as to feel a shear stress in
excess of the Peierls stress (or the PN stress or CRSS) then slip is initiated.
Note how a weak plane in the material shears off due to applied tensile forces
Q&A What is meant by toughness?
Toughness is the energy absorbed by the material upto failure. This could be in a uniaxial
tension test or in an impact test (which is done using a notched specimen Charpy or Izod
specimens).
Toughness is combined parameter involving strength and ductility (i.e. a material with higher
strength for a given ductility or higher ductility for a given strength has higher toughness).
Q&A What are the simple tests we can perform on test specimens to evaluate their
mechanical behaviour?
Uniaxial tension test.
Compression test.
Hardness test.
Bending test (3-point, 4-point bend tests).
Torsion test.
Why does the stress value drop after UTS in a “s-e” plot?
Funda Check
Why does this not happen in the “-” plot?
After necking the load bearing area decreases, however in the computation of ‘s’ we divide
by the original area A0 → leading to a apparent drop in stress (or strength). The word
apparent has been used as in reality the material continues to harden (get stronger and is
capable of ‘bearing’ higher stress).
In the “-” plot we divide by the instantaneous area and hence the hardening effect with
deformation is captured (even beyond the necking).
Hence, in the “-” plot the necking event cannot be ‘captured’.
Funda Check What is the difference between Stiffness and Young’s modulus?
Engg. strain
Careful experimentation has to be done to
capture the full details of the Yield Point
Note the ‘serrated’ yielding
Phenomena region (not done here!)
(yield point phenomenon)
A0 Li
From volume constancy A 0 L0 =Ai Li
Ai L0
P
Ai
P A0
A0 Ai
L L
s i s 1 i 1 s(1 e)
L0 L0
s( 1 e) Valid till
necking starts
ε ln ( 1 e)
L
dL L L
L ln
L0
ln 1 1 ln(1+e)
L0 L0
In some materials (e.g. pure annealed Cu, gray cast iron etc.) the linear portion of the curve
may be limited and yield strength may arbitrarily determined as the stress at some given
strain (say 0.005).
Funda Check Previous page is too confusing where does yielding really start?
Important Points
Yielding begins when a stress equal to y (yield stress) is applied; however to cause further
plastic deformation increased stress has to be applied i.e. the material hardens with
plastic deformation → known as work hardening/strain hardening.
Beyond necking the state of stress becomes triaxial (in a cylindrical specimen considered).
Technically the yield criterion in uniaxial tension cannot be applied beyond this point.
Plastic deformation (= yielding) and state of stress
We shall keep our focus on plastic deformation by slip.
y is yield stress in an uniaxial tension test (i.e. plastic deformation will start after crossing yield stress
only under uniaxial tensile loading) and should not be used in other states of stress (other criteria of
yield should be used for a generalized state of stress).
I.e in uniaxial tension the yield criterion is very simple:
Yielding starts when: y . Note: is a field variable and y is a material property.
Hydrostatic pressure (leading to hydrostatic stress in a material) does not lead to yielding in
a continuous solid (usually!).
This implies that the stress deviator holds the key to yielding.
For an isotropic material the yield criterion will not be independent of the choice of the axes
(i.e. a invariant function). Hence, for an isotropic material the yield criteria (we will note that there are more than
one) will be a function of the invariants of the stress deviator.
q L A0 1 L L0 A0 1 q
e0
A 1 q
e0 1
1 q
1
1 q
1 q
L0 L0 A
* I.e. does not underestimate ductility.
Comparison between reduction in area versus strain at fracture
We had also seen that strain based measure can be related to area based measure
mathematically as: q
e0
1 q
However, it should be noted that they represent different aspects of material behaviour.
For reasonable gauge lengths, ‘e’ is dominated by uniform elongation prior to necking and
thus is dependent on the strain hardening capacity of the material (more the strain
hardening, more will be the ‘e’). Main contribution to ‘q’ (area based calculation) comes
from the necking process (which is more geometry (& microstructure dependent)).
Hence, reduction in area is not ‘truly’ a material property and has ‘geometry dependence’.
What happens after necking?
Following factors come in to picture due to necking:
Till necking the deformation is ~uniform along the whole gauge length.
Till necking, the points on the - plot lie to the left and higher than the s-e plot (as below).
After the onset of necking deformation is localized around the neck region.
The formulae used for conversion of ‘e’ to ‘’ and ‘s’ to ‘’ cannot be used after the onset of
necking. s( 1 e) ε ln ( 1 e)
Triaxial state of stress develops and uniaxiality condition assumed during the test breaks
down.
Necking can be considered as an plastic instability in tension.
Hence, quantities calculated after the onset of necking (like fracture stress, F) has to be
corrected for: (i) triaxial state of stress, (ii) correct cross sectional area.
Neck
Fractured surfaces
Original sample
Why is the point ‘N’ (the necking point) in the - curve to the left and above
Funda Check
that in the s-e curve? How to compute strains (& stresses) beyond ‘N’?
Let the subscript N refer to ‘at the neck’ for now (after necking has started).
N = FN / AN . sN = FN / A0 . AN < A0 N > sN.
= ln (1+e). Being a logarithm (natural) the value of << e.
Hence, N’ is to left and above N. After necking we cannot use the formulae for conversion of
engineering to true quantities as the state of stress is triaxial.
Beyond ‘N’, the calculation has to be based actual area (eq. (1) below). RHS is for cylindrical specimen (Cy.
Sp.)
(1)
A0 4D
2
D0
ln ln
0
2ln
A D
2
4 D Cy . Sp.
After necking
Eq.(1) has to be used to determine true strain.
The true strains far exceed that calculated
from: e = L/L0.
The flow curve after necking often turns out
Region of validity s( 1 e) be/has: (i) linear or (ii) continuously
of the equations decreasing slope.
ε ln ( 1 e) Given the triaxial state of stress, it is ‘difficult’
to ascertain the longitudinal tensile stress
accurately.
‘True’ values beyond necking
Note: Further complications arise at strains close to fracture as microvoid nucleation & growth take place
and hence all formulations based on continuum approach (e.g. volume constancy) etc. are not valid
anymore.
Cotd.. ‘True’ values beyond necking
The Correction
exp
uniaxial
2R a
1 ln 1
A 2 R
Cup
Ductility implies the ability of the material to undergo plastic deformation (typically before fracture).
We have to distinguish between macroscopic and microscopic ductility. Often, there might be no/little
visible macroscopic ductility, but considerable amount of plastic deformation at the microscopic level.
Broadly, we can think of three factors which affect the ductility of a material, as shown in the flow
diagram below. Needless to point out, the material properties are to considered at the test conditions and
hence these two aspects are intimately linked.
A material which may be brittle at low T, may become ductile at high temperatures.
At high strain rates (which results from a high loading rate) a ductile material (i.e. under low rates of
loading) may behave in a brittle manner.
Sometimes less importance is paid to the geometry or state of stress. The geometry of the component
plays an important role in deciding the loading conditions on the material. E.g. a spring pulled in tension
leads to torsion at the material level.
We have already seen that a ‘neck’ in the specimen (which is like a blunt crack) leads to a triaxial state
of stress, which suppresses plastic deformation. The worst case scenario is ‘hydrostatic’ state of stress,
wherein there is no plane which experiences shear stresses. Since shear stresses are required for plastic
deformation, this state of stress is not conducive for plastic deformation. This implies that the material
will seem brittle, albeit the fact that the material is still “ductile” (assuming we started off with a ductile material).
The stress corresponding to the blue part of the curve is also called
the ‘flow stress’.
Unloading followed by re-loading
Usually, there is a small hysteresis during unloading and reloading (as in the figure below).
The unloading takes place along the path XM, while the reloading is along the path MR. I.e.
during reloading, the curve joins the flow curve YF (that obtained during uninterrupted
loading) at R. (XR is the missing part of the flow curve YF).
This hysteresis is often neglected.
Variables in plastic deformation , , , T
In the tension experiment just described the temperature (T) is usually kept constant and
the sample is pulled (between two crossheads of a UTM) at a constant velocity. The
crosshead velocity can be converted to strain rate ( ) using the gauge length (L0) of the
specimen. v / L0
At low temperatures (below the recrystallization temperature of the material, T < 0.4Tm)
the material hardens on plastic deformation (YF in the - plot known as work
hardening or strain hardening). The net strain is an important parameter under such
circumstances and the material becomes a partial store of the deformation energy provided.
The energy is essentially stored in the form of dislocations and point defects.
If deformation is carried out at high temperatures (above the recrystallization temperature;
wherein, new strain free grains are continuously forming as the deformation proceeds),
strain rate becomes the important parameter instead of net strain.
ln
n
ln ,T
Deviations from this behaviour often observed (e.g. in Austenitic stainless steel) at low
strains (~10–3) and/or at high strains (~1.0). Other forms of the power law equation are also
considered in literature (e.g. y K plastic n ) (Some of these are more complicated and involve more constants).
An ideal plastic material (without strain hardening) would begin to neck right at the onset of
yielding. At low temperatures (below recrystallization temperature- less than about 0.5Tm)
strain hardening is very important to obtain good ductility. This can be understood from the
analysis of the results of the uniaxial tension test. During tensile deformation instability in
the form of necking localizes deformation to a small region (which now experiences a
triaxial state of stress). In the presence of strain hardening the neck portion (which has been
strained more) hardens and the deformation is spread to other regions, thus increasing the
ductility obtained.
Cu and brass (n ~ 0.5) can be given large plastic strain (before fracture) as compared to
steels with n ~ 0.15.
When true strain is less than 1, the smaller value of ‘n’ dominates over a larger value of ‘n’
Variables/parameters in mechanical testing
There are variables, which we have to ‘worry about’, when we do mechanical testing
(for the test and to interpret the results):
Process parameters (characterized by parameters inside the sample)
Mode of deformation, Sample dimensions, Stress, Strain, Strain Rate,
Temperature etc.
Material parameters
Crystal structure, Composition & segregation, Grain size, dislocation density,
texture, etc.
Properties of selected materials
Material n K (MPa)
Annealed Cu 0.54 320
B
A
O
Note the magnitude in x and y axes
At high temperatures (above recrystallization temperature) where strain rate is the important
parameter instead of strain, a power law equation can be written as below between stress and
strain rate.
ln
ln T
Further aspects regarding strain rate sensitivity
1/ m 1/ m
P 1
C ,Tm
C
m P
A C A
From the definition of true strain
1 dL 1 dA
L dt A dt 1/ m 1/ m
1 dA P 1
A dt C A
1/ m
dA P 1 If m < 1→ smaller the cross-sectional area, the
(1 m ) / m
dt C A more rapidly the area is reduced.
If m = 1 → material behaves like a Newtonian
viscous liquid → dA/dt is independent of A.
1/ m
dA P 1
(1 m ) / m
dt C A
Effect of temperature and strain rate on the stress strain curve
Slip is a thermally activated process. So increasing the temperature decreases the flow
stress. This is because dislocations can make a thermally activated jump over the Peierls barrier.
As mentioned elsewhere, increasing the strain rate (by a few orders of magnitude) is like
decreasing the temperature. The strength of the material increases on increasing the strain
rate (with concomitant decrease in ductility). At very high strain rates (as experienced when a bullet hits a
material), dynamic effects (wave propagation in the structure) become important and some of the usual material properties may become
irrelevant.
The effect of these two parameters on the stress-strain curve is slightly different for FCC
and BCC metals.
BCC FCC
Effect of temperature and strain rate on the yield stress
As noted before, slip is a thermally activated process and hence yield stress decreases with
an increasing temperature.
With increasing strain rate (by orders of magnitude) the yield stress increases.
We have seen that below recrystallization temperature ‘n’ is ‘the’ important parameter.
Above recrystallization temperature it is ‘m’ which is important.
We have also noted that it is necking which limits the ductility in uniaxial tension.
Necking implies that there is locally more deformation (strain) and the strain rate is also
higher locally.
Hence, if the ‘locally deformed’ material becomes harder (stronger) then the deformation
will ‘spread’ to other regions along the gauge length and we will obtain more ductility.
Hence having a higher value of ‘n’ or ‘m’ is beneficial for obtaining good ductility.
Plastic deformation by slip Click here to see overview of mechanisms/modes of plastic deformation
C (FCC lattice)
{111} ½<110> 12
Diamond cubic
More examples of slip systems
TiO2
{101} <101>
Rutile
CaF2, UO2, ThO2
{001} <110>
Fluorite
CsCl <001>
{110}
B2
NaCl, LiF, MgO
{110} <110> 6
Rock Salt
C, Ge, Si
{111} <110> 12
Diamond cubic
MgAl2O4
{111} <110>
Spinel
Al2O3
(0001) <1120>
Hexagonal
More examples of slip systems
Cu (FCC)
{111} (a/2)< 1 10> 4 x 3 = 12
Fm 3m
{011} 6 x 2 = 12
W (BCC)
{112} (a/2)<11 1> 12 x 1 = 12
Im 3m
{123} 24 x 1 = 24
{0001} 1x3=3
Mg (HCP)
{10 10} (a/3)<1120> 3x1=3
P63/mmc
{10 11} 6x1=6
CsCl
{110} a<001> 6x1=6
Pm 3m
(100) 1x1=1
Polyethylene
{110} c<001> 2x1=2
Pnam
(010) 1x1=1
Critical Resolved Shear Stress (CRSS) What is the connection between Peierls stress and CRSS?
When one bents a rod of aluminium to a new shape, it involves processes occurring at
various lengthscales and understanding these is an arduous task.
However, at the fundamental level slip is at the heart of the whole process.
To understand ‘how slip can lead to shape change?’; we consider a square crystal
deformed to a rhombus (as Below).
Now visualize dislocations being punched in on successive planes moving and finally
leaving the crystal
This sequence of events finally leads to deformed shape which can be approximated to a
rhombus
We have seen that slip of dislocations weakens the crystal. Hence we have two strategies to
strengthen the crystal/material:
completely remove dislocations difficult, but dislocation free whiskers have been
produced (however, this is not a good strategy as dislocations can nucleate during loading)
increase resistance to the motion of dislocations or put impediments to the motion of
dislocations this can be done in many ways as listed below.
Strengthening mechanisms
1 2 3 4
Q
Ae kT
Q
d
Strain rate
dt
Al2O3
Covalent
300 Si
18-8 SS
Ni-FCC
150 Ni
Cu-FCC
Cu
RT is like HT and P-N 0.0 0.2 0.4 0.6
stress is easily overcome
T/Tm →
Strengthening Mechanism 1 Solid Solution Strengthening
Solid solutions offer greater resistance to dislocation motion than pure crystals (even solute
with a lower strength in bulk form gives strengthening!).
The stress fields around solute atoms interact with the stress fields around the dislocation to
leading to an increase in the stress required for the motion of a dislocation.
The actual interaction between a dislocation and a solute is much more complex.
The factors playing an important role are:
► Size of the solute more the size difference, more the hardening effect.
► Elastic modulus of the solute
(higher the elastic modulus of the solute greater the strengthening effect).
► e/a ratio of the solute.
A curved dislocation line configuration is required for the solute atoms to provide hindrance
to dislocation motion.
As shown in the plots in the next slide, increased solute concentration leads to an increased
hardening. However, this fact has to be weighed in the backdrop of solubility of the solute.
Carbon in BCC Fe has very little solubility (~0.08 wt.%) and hence the approach of pure
solid solution strengthening to harden a material can have a limited scope.
Size effect
The magnitude of solid solution strengthening depends on the size difference between the
‘matrix’ atoms & the ‘additive’ atom and the concentration of the additive atoms.
For the same size difference the smaller
Size difference atom gives a greater strengthening effect
Size effect depends on: ~c
Concentration of the solute (c) y
200 Sn (1.51)
Matrix: Cu (r = 1.28 Å)
Be (1.12)
150
y (MPa) →
Zn (1.31)
50
(Values in parenthesis are atomic radius values in Å)
By ↑ i (lattice friction) X
↑ y
→
Solute atoms ↑ level of - curve
→
Relative strengthening effect of Interstitial and Substitutional atoms
Interstitial solute atoms have a non-spherical distortion field and can elastically interact
with both edge and screw dislocations. Hence they give a higher hardening effect (per unit
concentration) as compared to substitutional atoms which have (approximately) a
spherical distortion field.
However, it should be noted that interstitial atoms usually have very limited solubility
(e.g. 0.01 wt.% C is soluble in BCC Fe at RT). On the other hand, substitutional atoms can
have complete solubility (e.g. isomorphous Au-Ag system→ satisfies Hume-Rothery
rules).
Solute atoms
Substitutional Spherical
distortion
field Gsolute / 10
Elastic Interstitial → Edge and screw dl.
Substitutional → edge
Long range
Modulus
(T insensitive)
Long range order
Mechanisms of interaction of
Solute-dislocation interaction dislocations with solute atoms
Stacking fault
Short range
Electrical
(T sensitive)
Schematic diagram
[Cold work in a uniaxial tension test] What causes Strain hardening? → multiplication of dislocations
The increase in flow stress with strain is called strain hardening (or work hardening).
Plastic deformation is caused by dislocations moving and leaving the crystal this implies
that dislocation density should decrease with plastic deformation. However, it is observed
that dislocation density increases with strain (Eq.1 below)! Further, if dislocations are the agent
weakening the crystal then with increased dislocation density the material should get
weaker! But, we observe on the contrary an increase in strength!
→
creation should exist. [&]
More dislocations is somehow causing a ‘traffic jam’ kind
of scenario and leading to strengthening.
→
How can dislocation density increase with plastic deformation? Dislocation sources
It is difficult to obtain crystals without dislocations (under special conditions whiskers have
been grown without dislocations).
Dislocation can arise by/from the following processes.
Solidification. Plastic deformation (nucleation and multiplication).
Irradiation. Etc.
Some (details) regarding the specific sources/methods of formation/multiplication of dislocations include
the below.
During solidification errors in the formation of a perfect crystal lattice can lead to
dislocations.
Grain boundary ledges and steps emit dislocations during deformation.
Aggregation and collapse of vacancies to form a disc or prismatic loop (FCC Frank partials).
These ‘excess’ vacancies could arise due to quenching from high temperature or irradiation.
High stresses arising due to ► heterogeneous nucleation at second phase particles and
► during phase transformations can lead to the nucleation of dislocations.
Frank-Read source and Orowan bowing mechanism and lead to an increase in the dislocation
density during plastic deformation.
Intersection of dislocations can lead to an increase in the dislocation length (i.e. effectively an increase
in the dislocation density).
* Current understanding is that Frank-Read sources are rare and grain boundary ledge sources are more
common.
Strain hardening
Strain hardening is caused due to three reasons:
1 Dislocations moving in non-parallel slip planes can intersect with each other
(“intersection with forest of dislocations”). This can lead to jogs on the ‘forest
dislocations’→ resulting in an increase in stress required to cause further plastic deformation
Strain Hardening / work hardening.
2 Dislocations can also be rendered immobile as in the formation of the Lomer-Cottrell
barrier.
3 Interaction of the stress fields of dislocations can make it difficult for dislocations to
move.
Dynamic recovery
1
[0 11]
2 (100)
1
[1 0 1]
2 ( 1 11)
1 1 1
[ 1 1 0] [1 0 1]
2 (111) + 2 ( 1 11)
[0 11]
2 (100)
Lomer-Cottrell barrier →
The red and green dislocations attract each other and move towards
their line of intersection
[0 1 1] They react as above to reduce their energy and produce the blue
dislocation
The blue dislocation lies on the (100) plane which is not a close
packed plane → hence immobile → acts like a barrier to the motion
of other dislocations
Other dislocations present in the material which intersect (or be present in) the slip plane
can impede the motion of dislocations (forest of intersecting dislocations). This implies that
increase in dislocation density can lead to an increase in the strength of the material (as given by
the empirical relation as below).
Dislocation can intersect with each other and ‘react’ to give jogs and kinks (apart from
sessile dislocations).
Cold work will increase the strength of the material via an increase in dislocation density,
but will lead to a decrease in the ductility (as expected!).
Empirically, the strength (in shear) depends on the root of the dislocation density (eq. below).
Empirical relation 0 → base stress to move a dislocation in the crystal in the absence of other
dislocations (i.e. if dislocation density is ‘effectively’ zero)
Studies across many materials has shown that on decreasing the grain size the yield
strength of the material increases. The formula making this connection is the famous Hall-
Petch relation.
At vary small grain sizes this formula is expected to break down as we will approach the
theoretical shear strength of the material. In some nanostructured materials even inverse
Hall-Petch relation has been reported (i.e. decrease in strength with decrease in grain size).
The usual mechanism suggested for this behaviour is the dislocation ‘pile-up’ mechanism*.
Grain size is related to the ASTM grain size number Hall-Petch constants
by the following formula Material i [MPa] k [MPa m1/2]
645 d → Grain diameter in meters Cu 25 0.11
d
(2 n 1 )1010 n → ASTM grain size number Ti 80 0.40
N 2( n 1) Mild steel 70 0.74
* other mechanisms have also been suggested. Ni3Al 300 1.70
At the heart of the ‘pile up’ mechanism is a pile up of dislocations at a barrier (typically the
grain boundary).
Dislocations get piled up at a barrier (say grain boundary) and produce a back stress. This
back stress impedes the motion of other dislocations on the same slip plane (towards the
boundary).
Photo courtesy: Dr. Eswar Prasad (Johns Hopkins University, 2013). Mg alloy (AZ31, 3% Al , 1% Zn)
1
Glide through the precipitate If the precipitate is coherent* with the matrix
* Or is semi-coherent with some continuous planes
Dislocation
Get pinned by the precipitate
2
A complete list of factors giving rise to hardening due to precipitates/inclusions will be considered later
1 Dislocation Glide through the precipitate
Only if slip plane is continuous from the matrix through the precipitate precipitate is
coherent with the matrix.
Stress to move the dislocation through the precipitate is ~ that to move it in the matrix
(though it is usually higher as precipitates can be intermetallic compounds).
Usually during precipitation the precipitate is coherent (if the lattice misfit is small) only
when it is small and becomes incoherent on growth.
Small
Growth
Growth
Large
Coherent Partially coherent Incoherent
Glide of the dislocation causes a displacement of the upper part of the precipitate w.r.t the
lower part by b → this is usually called cutting of the precipitate by the dislocation.
Schematic views
edge dislocation glide through a coherent precipitate (shearing the
precipitate in the process). This is also called ‘cutting’ of the precipitate.
b Precipitate particle
b
If the particle is sheared, then how does the hardening effect come about?
We have seen that as the dislocation glides through the precipitate it is sheared.
If the precipitate is sheared, then how does it offer any resistance to the motion of the dislocation? I.e.
how can this lead to a hardening effect?
The hardening effect due to a precipitate comes about due to many factors (many of which are system
specific). The important ones are listed in the tree below.
Dislocations can bow around widely separated inclusions. In this process they leave
dislocation loops around the inclusions, thus leading to an increase in dislocation density.
This is known as the Orowan bowing mechanism as shown in the figure below. (This is in ‘some
sense’ similar to the Frank-Read mechanism).
The maximum stress required to cause the bending between the pinning points is given by the formula below. More about this can be found
here: double ended Frank-Read source.
The next dislocation arriving (similar to the first one), feels a repulsion from the dislocation
loop and hence the stress required to drive further dislocations increases. Additionally, the
effective separation distance (through which the dislocation has to bow) reduces from ‘d0’ to
‘d1’.
Gb
max ~
d0
d0
The dislocation moves on
leaving behind loops around
precipitates
Precipitate Hardening effect (Complete List)
When there are two (or more) competing mechanisms are available to respond to a stimulus
(say applied load which results in a stress state in the material), then the mechanism which operates at a
lower magnitude of the stimulus (stress in the current situation) operates (in preference to other competing
mechanisms).
When stress is increased beyond the yield stress the mechanism of deformation changes.
Till ‘Y’ in the s-e plot, bond elongation (elastic deformation) gives rise to the strain.
After ‘Y’, the shear stress resulting from the applied tensile force, tends to move dislocations
(and cause slip) rather than stretch bonds as this will happen at lower stresses as
compared to bond stretching (beyond ‘Y’).
If there are not dislocations (e.g. in a whisker) (and for now we ignore other mechanism of
deformation), the material will continue to load along the straight line OY till dislocations
nucleate in the crystal.
What are the internal mechanisms (with respect to plastic deformation by slip),
Funda Check
which oppose an externally applied force (and hence give rise to strengthening)?
Note: some of the below mechanisms may not be operative in all materials/may become
important only under some special circumstances (e.g. high strain rate).
These effects give rise to resistance to the motion of dislocations and are responsible for
strengthening.
Basic internal resistance → Peierls stress.
Hardening due to increase in dislocation density.
Grain size strengthening.
Solid solution hardening.
Precipitation hardening.
Phonon and Electron drag.
Yield Criteria
We have seen that the most common method of yielding is by slip, which is mediated by
dislocations. At the ‘crystal structure level’ slip is initiated when > PN.
Slip at the macroscopic level in a UTT is supposed to begin when > y.
Under a general state of stress*, we would like to know when plastic deformation is initiated
and further would like to use the material parameter obtained from a UTT (i.e. y).
Multiple theories of yield have been developed at various times. The most commonly used
ones are by Tresca (Maximum Shear Stress Condition) and von Mises (Distortion Energy
Density Condition). It is clear that shear stresses cause plastic deformation and hence the
acceptance of these ones.
Some of these criteria are better suited for other modes of failure and not for yielding.
Failure is supposed to take place if the maximum normal stress in the member reaches a
critical value.
Since this criterion is expected to be true for any state of stress, it will be true for an uniaxial
tension test. Hence, this critical value is the yield stress obtained from an uniaxial tension
test (i.e. failure in this context implies the onset of plastic deformation).
The maximum normal stress is one of the principal stresses.
The yield stress in compression and tension is assumed to be equal and failure implies that
either 1 or 2 should exceed Y.
This applicability of this criterion to yielding is rather limited and may be more valid for
fracture (wherein normal stresses play a key role).
Yield Criterion Maximum Shear Stress Tresca
Failure is supposed to take place if the maximum value of the shear stress in the member
reaches a critical value. Further, we make the assumption that if this criterion is valid for any
stress state, it must be valid for that in a UTT and hence this critical value is obtained from
an uniaxial tension test (i.e. the value of shear stress at yield in a UTT).
The shear stress at yield in a UTT is 1/2 (=max).
This criterion is expected to have reasonable validity as shear stresses are responsible for
slip.
1
UTT max Cos 45 Cos 45 max 1 Y at yield
Cos Cos 2 2
TWINNING