Lecture1460085501 2

LECTURE NOTES ON
PHASE TRANSFORMATIONS & HEAT TREATMENT
6th Semester, MME
For Batches 2013-

2013-17, 2014-
2014-18 and 2015
2015--19
Course Coordinator
Avala Lava Kumar

Department of Metallurgical & Materials Engineering (MME)
Veer Surendra Sai University of Technology (VSSUT), Burla-768018
E-mail : lavakumar.vssut@gmail.com Phone: (+91) (7077110110)
Course Objective
The aim of this course is to gain an understanding of the role of phase transformations w.r.t.
heat treatment on the development of microstructure and properties of metallic materials. The
course will highlight a number of commercially-significant applications where phase
transformations due to heat treatment are important.
Course Overview
Thermodynamics of phase transformations; solidification of pure metals and alloys; thermal
supercooling; constitutional supercooling; interface stability; solute redistribution; Solid state
transformations : nucleation and growth of phases; diffusion mechanisms; transformation
kinetics; transformation diagrams. Diffusional and Diffusionless transformations:
decomposition of solid solutions; ordering reactions, spinodal decomposition; eutectoid,
bainitic, martensitic transformations Heat treatment processes, concept of Hardenability and
Cast Irons.
Learning Outcomes
Enhanced critical thinking, analytical and problem solving skills in materials science and
engineering. An understanding of the principles underlying liquid-to solid and solid-state phase
transformations in a range of materials. An understanding of the importance of phase
transformations w.r.t. heat treatment for controlling microstructure and properties in
3
engineering alloys.
REFERENCES
Phase Transformations in Metals and Alloys (This is the major reference for this course)
D.A.Porter, K.E. Easterling, and M.Y. Sharif
CRC Press , Taylor & Francis Group
Diffusion in solids
Prof. Alok Paul , IISC Banglore
NPTEL Web course
Phase Transformations
Prof. Anandh Subramaniam IIT Kanpur
Steel heat treatment : Metallurgy and Technologies
Geroge E. Totten
CRC Press , Taylor & Francis Group
Heat Treatment : Principles and Techniques
T.V.Rajan, C.P.Sharma and Ashok Sharma
PHI Learning Private Limited
Heat Treatment of Materials
Vijendra Singh
Standard Publishers Distributors, Delhi
Phase Transformations & Heat Treatment
Prof. M.P.Gururajan
NPTEL web course
4
CONTENTS
Chapter Chapter Name Page No

1 Introduction 6-16
2 Thermodynamics & Kinetics 17-90
3 Diffusion 91-164
4 Solidification 165-206
5.1 Iron-Cementite Phase Diagram 207-251
5
5.2 TTT&CCT Diagrams 252-306
6 Diffusional Transformations 307-384
7 Diffusionless Transformations 385-406
8 Heat Treatment Processes 407-460
9 Hardenability 461-488
10 Cast irons 489-525
Avala Lava Kumar*
Veer Surendra Sai University of Technology (VSSUT), Burla -768018
*E-mail : lavakumar.vssut@gmail.com
6
Entropic force
UNIVERSE STRONG
WEAK
HYPERBOLIC ELECTROMAGNETIC
EUCLIDEAN SPACE ENERGY GRAVITY
SPHERICAL
nD + t
PARTICLES FIELDS
METAL
ATOMIC NON-ATOMIC
SEMI-METAL BAND STRUCTURE
SEMI-CONDUCTOR
INSULATOR STATE / VISCOSITY
LIQUID
GAS SOLID LIQUID CRYSTALS
STRUCTURE
AMORPHOUS QUASICRYSTALS RATIONAL CRYSTALS

APPROXIMANTS
SIZE
NANO-QUASICRYSTALS NANOCRYSTALS
7
Classification of materials
Materials
Composites: have two (or more)
solid components; usually one is a
Monolithic Hybrids
matrix and other is a reinforcement
Metals (& Alloys) Composite
Sandwich structures: have a
Ceramics & Glasses material on the surface (one
Sandwich or more sides) of a core
material
Polymers (& Elastomers)
Lattice* Structures: typically a
Lattice combination of material and space
(e.g. metallic or ceramic forms,
aerogels etc.).
Segment
Segmented Structures: are Hybrids are

designed to improve
divided in 1D, 2D or 3D
certain properties of
(may consist of one or monolithic materials
more materials).
8
*Note: this use of the word 'lattice' should not be confused with the use of the word in connection with crystallography.
Length scales in metallurgy
• Casting
Thermo-mechanical • Metal Forming
Treatments • Welding
Crystal • Powder Processing
• Machining
Atom Structure Microstructure Component
Electro-
magnetic Phases + Defects + Residual Stress
& their distribution
Structure could imply two types of structure: • Vacancies
Crystal structure
Electromagnetic structure
• Dislocations
Fundamentally these aspects are two sides of the same • Twins
coin • Stacking Faults
Microstructure can be defined as: • Grain Boundaries
(Phases + Defect Structure + Residual Stress) and their
distributions
• Voids
Microstructure can be ‘tailored’ by thermo-mechanical • Cracks
treatments
9
Processing determines shape and microstructure of a component
Length scales in metallurgy
Let us start with a cursory look at the length scales involved in Materials Science
Dislocation Stress fields

Angstroms → Nanometers Microns Centimeters
Unit Cell* Crystalline Defects Microstructure Component
Grain Size
*Simple Unit Cells 10

Transformations in Materials
Phases Defects Residual stress

Phases can transform Defect structures can change Stress state can be altered
Phase Defect structure Stress-state

transformation transformation transformation
Geometrical Physical
Structural Property
Phases Microstructure
Phases Transformations Microstructural Transformations

11
Classification of Transformations
Thermodynamics Order of a phase transformation

Ehrenfest, 1930
Classification of Based on Reconstructive Civilian

Phase Mechanism
Transformations Buerger, 1951 Displacive Military
Non-quenchable Athermal → Rapid
Kinetics
le Chatelier, (Roy 1973) Quenchable Thermal → Sluggish
12
The thermodynamic characteristics
associated with the phase transformations
can be used to classify transformations; in
this classification methodology, if the nth
derivative of free energy (G) with respect to
temperature (T) and pressure (P) is
discontinuous, it is defined as the nth order
transformation.
As shown in Fig., in transformations such as
melting, the first derivative has the
discontinuity; hence, melting is a first order
transformation; on the other hand, in some
of the order/disorder transformations, it is
the second derivative which is
discontinuous, making it the second order
transformation.
Figure: The thermodynamic classification of transformations:
the first derivative of the free energy ‘G’ with respect to
temperature ‘T’ , that is the enthalpy ‘H’ is discontinuous at the
transformation temperature Tc as shown in the first column; the
second derivative of the free energy with respective to
temperature Cp is discontinuous while ‘H’ is not in the second
13
column, making the order of transformation second.
Heterogeneous Transformations Homogeneous Transformations

(nucleation and growth)
Spinodal decomposition
Order-disorder transformation
GROWTH CONTROLLED GROWTH CONTROLLED BY GROWTH CONTROLLED BY A THERMAL

BY HEAT TRANSFER HEAT AND MASS TRANSFER THERMALLY ACTIVATED GROWTH
MOVEMENTS OF ATOMS
Solidification of pure metals Solidification of alloys Martensite
transformations
SHORT RANGE TRANSPORT LONG RANGE TRANSPORT

(interface controlled)
Polymeric transformations
Continuous Reaction Discontinuous Reaction
Massive transformations
Recrystallization
Grain growth, etc. Precipitation dissolution Eutectoid reactions
14
Discontinuous precipitation
Phase transformations can be classified as homogeneous (transformations which take place
through spinodal mechanism in which transformation takes place throughout the material)
and heterogeneous (transformations which take place through nucleation and growth
mechanism in which transformation takes place heterogeneously at a few places in the
material at the start of the transformation).
Transformations can also be classified as diffusional (or, so called, `civilian') and
diffusionless (or, so called `military') depending on the mechanism. In civilian
transformations, the nucleation and growth take place via diffusion assisted atomic motion.
On the other hand, in the military transformation, the nucleation and growth is by shear and
shuffle of atoms by less than one atomic displacement and the movement of all the
participating atoms is coordinated.
There are transformations which are thermally activated (which typically are based on
diffusion) while there are others which are athermal.
The transformations can also be diffusion controlled or interface controlled.
Transformations can also be differentiated based on whether the interfaces formed are
glissile or nonglissile.
In some transformations there are compositional changes while in some other there are no
composition changes.
Further, transformations which are diffusional can either involve long range diffusion or
short range diffusion. 15
Questions?
1. What is a Phase transformation? How it is depending on Heat treatment?
2. How do the thermodynamic parameters such as G, H, V and Cp vary with temperature for
first and second order transformations
3. Explain the classification of phase transformations.
4. What is meant by Glissile transformations?
5. Can a ferromagnetic to paramagnetic change be described as an order-disorder
transformation? Explain
6. Give two examples of phase transformations where the proper control of the transformation
can result in different combinations of properties.
7. List the possible differences in the nature of short-range diffusion in a polymorphic
transformation versus an order-disorder transformation.
8. What types of phase changes require long-range diffusion?
9. In what type of transformation is a composition change not possible?
10. Consider a transformation in which there is a large volume change during transformation.
What is the order of such transformation?
16
Avala Lava Kumar*
17
Introduction
The fields of Thermodynamics and Kinetics are vast oceans and this chapter will introduce
the bare essentials required to understand the remaining chapters.
Let us start by performing the following (thought) experiment:
Heat a rod of Al from room temperature to 500°C → As expected the rod will expand
(A → B in figure below).
The expansion occurs because of two reasons:
1 Vibration of atoms (leading to an increase in average spacing between atoms→ the
usual reason) (A → M in figure below).
2 Increase in the concentration of vacancies* (a vacancy is created when a Al atom goes
to the surface and for every 4 vacancies created the volume equal to 1 unit cell is added). (M
→ B in figure below). The 2nd reason is a smaller effect in terms of its contribution to the overall
increase in length of the specimen
Metal expands on
heating due to 2 different
physical reasons!
* It costs energy for the system to put vacancies (broken bonds, distortion to the lattice) 18
→ then why does the system tolerate vacancies?
Introduction
Let us next consider the melting of a pure metal at its melting point (MP) (at
constant T and P) → by supplying heat to the sample of metal (so that the metal
sample is only partly molten). At the MP the liquid metal is in equilibrium with the
solid metal.
The liquid has higher potential energy as well as higher kinetic energy than the
solid.
Then why does the liquid co-exist with the solid?
The answer to this question lies in the fact that internal energy is not the measure
of stability of the system (under the circumstances).
We will learn in this chapter that it is the Gibbs Free Energy (G). The molten metal
has higher energy (internal energy and enthalpy), but also higher Entropy. So the
melting is driven by an increase in Entropy of the system. The molten metal and
the crystalline solid metal have the same G → hence they co-exist in equilibrium.
19
Stability and Equilibrium
Equilibrium refers to a state → wherein there is a balance of ‘forces’* (as we shall see
equilibrium points have zero slope in a energy-parameter plot)
Stability relates to perturbations (usually small perturbations** about an equilibrium state)
(as we shall see stable relates to the curvature at the equilibrium points).
Let us start with a simple mechanical system → a rectangular block (Figure in next slide)
(under an uniform gravitational potential).
The potential energy (PE) of the system depends on the height of the centre of gravity (CG).
The system has higher PE when it rests on face-A, than when it rests on face-B.
The PE of the system increases when one tilts it from C1 → C2 configuration.
In configurations such as C1,C2 & C3 the system will be in equilibrium (i.e. will not change
its configuration if there are no perturbations).
In configuration C2 the system has the highest energy (point B) and any small perturbations
to the system will take it downhill in energy → Unstable state.
Configuration C3 has the lowest energy (point C) and the system will return to this state if
there are small perturbations → the Stable state.
* Force has been used here in a generalized sense (as an agent which can cause changes)
20
** Perturbation is usually a small ‘force/displacement’ imposed in a short span of time.
Mechanical Equilibrium of a Rectangular Block
A Ball on a plane
Neutral Equilibrium
Centre B
Of
Gravity
C1 C2 C3
Potential Energy = f(height of CG)
Unstable
Stable
A
Lowest CG of all possible
states
Metastable state C
21
Configuration
Stability and Equilibrium
Configuration C1 also lies in an ‘energy well’ (like point C) and small perturbations will
tend to bring back the system to state C1. However this state is not the ‘global energy
minimum and hence is called a Metastable state.
Additionally, one can visualize a state of neutral equilibrium, like a ball on a plane (wherein
the system is in a constant energy state with respect to configurations).
Points to be noted:
A system can exist in many states (as seen even for a simple mechanical system: block on a
plane)
These states could be stable, metastable or unstable
Using the relevant (thermodynamic) potential the stability of the system can be
characterized (In the case of the block it is the potential energy, measured by the height of the CG for
the case of the block on the plane)
System will ‘evolve’ towards the stable state provided ‘sufficient activation’ is
provided (in the current example the system will go from C1 to C3 by ‘sufficient jolting/shaking’ of
the plane)
Three kinds of equilibrium (with respect to energy)
Global minimum → STABLE STATE
Local minimum → METASTABLE STATE
Maximum → UNSTABLE STATE 22
Law’s of Thermodynamics
Zeroth law of Thermodynamics
If two systems are each in thermal equilibrium with a third, then all three are in
thermal equilibrium with each other. (Similar to the transitive property of equality
in mathematics; i.e. If a = b and b = c, then a = c)
If A B & B Then A C
C
No heat flows No heat flows No heat flows

(A is in equilibrium with B) (B is in equilibrium with C) (A is in equilibrium with C)
First law of Thermodynamics

This is a statement of the conservation of energy i.e. When heat (Q) is added to a
system, it increases the internal energy (ΔU) of the system and the system does
some work (W) on the external world.
Signs of Q and W
ΔU = Q – W
Q Positive System gains heat
Q Negative System loses heat
For infinitesimal change of the state,
W Positive Work done by system
dU = δQ – δW
W Negative Work done on system
23
Law’s of Thermodynamics
Second law of Thermodynamics:
In an isolated system, natural processes are spontaneous when they lead to an
increase in disorder, or entropy i.e. The entropy of a system in an adiabatic
enclosure always increases for spontaneous/irreversible processes and remains
constant during a reversible process but it never decreases.
Entropy S is defined by the equation
δQrev
dS = and is a function of state.
T
Third law of Thermodynamics:
The entropy of a perfect crystal is zero when the temperature of the crystal is equal
to absolute zero (0 K).
lim ∆S = 0
T →0
24
Thermodynamic parameters
In Materials Science we are mainly interested with condensed matter systems (solids and
liquids) (also sometimes with gases)
The state of such a system is determined by ‘Potentials’ analogous to the potential energy of
the block (which is determined by the centre of gravity (CG) of the block).
These potentials are the Thermodynamic Potentials (A thermodynamic potential is a Scalar
Potential to represent the thermodynamic state of the system).
The relevant potential depends on the ‘parameters’ which are being held constant and the
parameters which are allowed to change. More technically these are the
State/Thermodynamic Variables (A state variable is a precisely measurable physical
property which characterizes the state of the system- It does not matter as to how the system
reached that state). Pressure (P), Volume (V), Temperature (T), Entropy (S) are examples of
state variables.
There are 4 important potentials (in some sense of equal stature). These are: Internal Energy,
Enthalpy, Gibbs Free Energy, Helmholtz Free Energy.
Intensive properties are those which are independent of the size of the system
P, T
Extensive Properties are dependent on the quantity of material
V, E, H, S, G
25
Gibb’s free energy, G can be expressed as G = H-TS
H – enthalpy (J/mol) , T – temperature in K, S – entropy (J/mol.K)
Further H= E+PV
E – Internal energy (J/mol), P – Pressure (N/m2) , V – Volume (m3/mol)
In the solid state, the term PV is in general very small (in the temperature and
pressure range we consider) and can be neglected.
On the other hand, internal energy of metals is in the order of kJ/mol. So PV term
is less than even 1%.
Internal energy has two components:
Potential energy , which depends on atomic bonds and
Kinetic energy , which depends on the vibration of atoms at their lattice
position
The relation can be rewritten as H = G + TS
H measures the total energy of the body
TS measures the useless energy that is the energy which can’t be spent for any
work or transformation.
26
So the Gibb’s free energy is defined as the energy that can be “set free” at a
particular pressure to do the work or make a particular transformation possible
Similarly at a particular volume the free energy is called Helmoltz free energy, F,
expressed as
F = E – TS
We shall consider the Gibb’s free energy since we are going to consider the
transformations under constant pressure.
To determine Gibb’s free energy, we need to determine enthalpy and entropy of the
system.
27
Relation between enthalpy and specific heat
Let us first consider a single component system.
If we keep the system isolated and do not allow it to interact with the surroundings, the
internal energy of the system will not change. This is actually a closed system.
However, if the system is allowed to interact with the surroundings (that is if it acts as a
open system) internal energy might change.
Suppose the system does work δW in the surroundings and takes heat δQ from the system,
then according to the first law of thermodynamics, the change in internal energy of the
system is
dE = δQ − δW
So after spending some energy for work, rest of the energy is added to the system to
increase the internal energy.
Here exact differential is used for energy because it does not depend on the path and
function of the state only. On the other hand heat and work depend on the path.
If a system changes its volume by dV at a constant pressure P, the work can be expressed as
dW = PdV
So, dE = δQ-PdV, we know H = E+PV
dH = dE+PdV+VdP
dH = dQ + VdP
28
Relation between enthalpy and specific heat
If we divide by dT on both sides we get dH δQ dP

= +V
dT dT dT
Specific heat capacity, Cp at a constant pressure is defined as the heat required to increase
the temperature of the system by one degree. So it can be expressed as
 δQ   dH 
CP =   = 
 dT P  dT P
After integration, we can write
H T T
∫ dH = ∫ C
H0 0
p dT H T = H 0 + ∫ C p dT
0
T
H T = ∆H 298 + ∫ C p dT
298
HT , H0 , and ∆H298 are the enthalpies at temperature, T, 0, and 298K respectively

Note that we can not measure absolute thermodynamic values. We rather measure a relative
value, H298 is considered as reference value.
For pure metals it considered as zero.
29
Relation between entropy and specific heat
As explained before, the knowledge on entropy (which cannot be spent to do a work) is
required to determine how much energy is available (free energy) in a system to do useful
work from the total heat content or enthalpy.
Second law of thermodynamics states that the entropy of a system will either remain the
same or try to increase.
When system moves from one equilibrium state 1 to another equilibrium state 2 by
changing heat of δQ, the entropy S in a reversible process can be defined as 2
δQ
S 2 − S1 = ∫
1
T
The value at zero Kelvin is zero and considered as the reference state for entropy, so that the
entropy at any temperature T can be expressed as T
δQ
ST = ∫
0
T
Previously we have seen from the relation H = E + PV and using the first law of
thermodynamics
δQ = dH
T
We know dH = Cp dT, So the entropy at T can be expressed as S = C P dT
T ∫0 T
30
Relation between entropy and specific heat
Sometimes value of the entropy at the standard state (298 K) is available and can written as
T
CP
ST = ∆S 298 + ∫298 T dT
So the free energy at temperature, T can be determined
GT = H T − TST
T
 T
CP 

GT = ∆H 298 + ∫ C P dT − T  ∆S 298 + ∫ dT 
298  298
T 
T
 T
CP 

GT = ∫ C P dT − T  ∆S 298 + ∫ dT  In pure elements
298  298
T 
Specific heat is expressed as
C
C P = A + BT − A, B and C are constants.
T2
Specific heat is expressed in terms of empirical formula as
expressed above.
In general the constant values are available in the data book.
Specific heat changes with temperature as shown in the figure
31
Single component system
This line slopes upward as at constant
T if we increase the P the gas will
liquefy as liquid has lower volume
(similarly the reader should draw
horizontal lines to understand the effect
of pressure on the stability of various
phases- and rationalize the same).
Phase fields of non-close packed

structures shrink under higher pressure
Phase fields of close packed structures

expand under higher pressure
These lines slope downward as: Under

higher pressure the phase with higher
packing fraction (lower volume) is
preferred
32
Gibb’s free energy change with temperature in a
single component system
An isolated system always tries to maximize the entropy. That means the system is stable
when it has maximum possible entropy.
Instead of considering isolated system, we need to consider the system which interacts with
the surroundings because heat transfer is always with respect to the surroundings only.
Any transformation is possible only when dS+dSSurrou ≥ 0 where dS is the change in entropy
of the system.
In a reversible process, when system can transform back to its previous state it is equal to
zero and in an irreversible process when the system cannot come back to its previous state is
greater than zero.
We need to find the stability with respect to another term, for the sake of convenience,
which can be used without referring to the surroundings. We shall show now that free
energy is rather a suitable property to define stability of the phases.
Let us consider that the system absorbs some amount of heat δQ from the system. Since the
surrounding is giving away the heat, we can write
δQ
dS surrou = −
T
33
Gibb’s free energy change with temperature in a
single component system
We have seen before that in an isobaric system δQ = dH. So we can write
dH
dS − ≥0
T
dH − TdS ≤ 0
We are considering isobaric system. If we consider the transformation at a particular
temperature ( T constant, dT = 0) then
dH − TdS − SdT ≤ 0
d ( H − TS ) ≤ 0
dG ≤ 0
So we have derived a more reasonable relation, which can be used without referring to the
surroundings.
In an reversible process, such as allotropic transformation or the transformation from solid
to liquid or liquid to solid the free energy change is zero. There will be many irreversible
transformations (later we shall see these kinds of diffusion controlled transformations),
where dG is less than zero.
This also indicates that a system will be stable when it has minimum free energy, so that it is
not possible to get anymore dG less than zero by any further transformation.
34
Stability of the phases in a single component system
From previous studies, we understand that system can attain minimum free energy
by decreasing enthalpy and/or increasing entropy.
That means dH ≤ 0 and/or dS ≥ 0 since dH – TdS ≤ 0
One component can have different stable phases at different temperature and
pressure ranges, for example, solid, liquid and the gas phase.
One phase is said to be stable, when it has lower free energy compared to other
phases in a particular range of temperature and pressure.
Let us consider constant pressure.
To compare the stability of different phases in a particular range of temperatures,
we need to calculate the free energy of the phases with respect to temperature.
To determine Gibb’s free energy at a particular temperature, we need to determine
H and S. Similarly it can be calculated at every temperature to gain knowledge on
the change in free energy with temperature.
As mentioned previously the data on specific heat are available in literature and
these parameters can be calculated.
35
From the definition, we know that the slope of

the enthalpy at any temperature is equal to the
specific heat at that temperature.
G = H − TS
dG = dH − TdS − SdT ( H = E + PV )
dG = dE + PdV + Vdp − TdS − SdT (dE = δQ − PdV )

dG = δQ − PdV + PdV + VdP − TdS − SdT δQ = dS
T
dG = TdS + VdP − TdS − SdT
dG = VdP − SdT (Maxwell’s Relation)
So at a constant pressure the slope of the free
energy curve
 ∂G 
  = −S Typical variation of thermodynamic
 ∂T  P parameters are shown in the figure.
If we want to know whether a solid or liquid phase will be stable at a particular

temperature or in a temperature range, we need to find free energy for both the phases.
36
For that we need to know the variation of specific

heat with temperature for both the phases as shown
in the figure.
Specific heat for a liquid phase is always higher than
the solid phase. That means energy required to
increase the temperature by one degree for liquid
phase is higher than the solid phase.
Superscript, S and L denote the solid and liquid
phases.
Free energy for liquid phase changes more
drastically compared to the free energy of the liquid
phase because entropy of the liquid phase is always
higher, which is the slope of the free energy.
At melting point, free energy for both the phases are
the same and the difference between the enthalpy of
these two phases is equal to the latent heat of fusion
L.
37
It must be apparent that one particular phase at certain temperature range will be stable, if
the free energy, G is lower of that phase than the other phase.
At low temperature range, one particular phase will be stable, which has low enthalpy, since
“TS” term will not dominate. That is why solid phase is stable at low temperature range.
On the other hand, at higher temperature range, phase having higher entropy will be stable
since in this range “TS” term will dominate. That is why liquid phase is stable at high
temperature range.
This is the reason that α-Ti with close packed HCP structure is stable at low temperature
range, whereas, at high temperature range β-Ti with relatively more open BCC structure is
stable.
If we fix a temperature, we have seen that one material/element will be stable as one type of
phase with a particular state compared to other.
It can stay in equilibrium if it has minimum free energy and dG=0, it indicates that for small
fluctuation it does not move to another state.
38
Gibbs Free Energy (G) vs T, H, S
Melting transition
39
Pressure Effects
From previous studies
dG = VdP − SdT
G l = G s → ∆G = 0 Tb
∆G
consider : S → L
Ts
dG = V dT − S dT
S S S
dG L = V L dT − S L dT
T Tm
(V − V )dP = ( S − S )dT
S L S L
dT ∆V V S − V L
= = S
dP ∆S S − S L
dT The rate of change of transformation
=
dP temperature as a function of pressure
dT ∆V
=
dP ∆S Is known as Clausius-Clapeyron equation 40
Pressure Effects
dP ∆H
= Is also known as Clausius-Clapeyron equation
dT T∆V
41
Driving force for solidification
In dealing with phase transformations, we are often
concerned with the difference in free energy Solid stable Liquid stable
between two phases at temperatures away from the
equilibrium temperature.
∆Gv
For example if a liquid metal is under cooled by
∆T below Tm before it solidifies, solidification will Solid (GS)
be accompanied by a decrease in free energy ∆G
G →
(J/mol) as shown in figure.
This free energy decreases provides the driving ∆G → −ve
force for solidification. The magnitude of this
change can be obtained as follows. Liquid (GL)
The free energies of the liquid and solid at a ∆T
temperature T are given by Undercooling ∆G → +ve
G = H − TS
L L L
Tm
G S = H S − TS S
Therefore, at a temperature T ∆G = ∆H − T∆S →①
Where ∆H = H L − H S and ∆S = S L − S S 42
At the equilibrium melting temperature Tm the free energies of solid and liquid are equal,
i.e., ∆G = 0. Consequently
∆G = ∆H − Tm ∆S = 0
∆H L
And therefore at Tm ∆S = = →②
Tm Tm
This is known as the entropy of fusion. It is observed experimentally that the entropy of
fusion is a constant ≈ R (8.3 J/mol . K) for most metals (Richard’s rule).
L
For small undercoolings (∆T) the difference in the specific heats of the liquid and solid ( C p
- C pS ) can be ignored.
L
Combining equations 1 and 2 thus gives ∆G ≅ L − T
Tm
i.e., for small ∆T L ∆T
∆G ≅
Tm This is called Turnbull’s approximation
43
Thermodynamic parameters in a binary system
Previously we are considered one element only. Now we consider interaction between
two elements. This is not straightforward since elements can interact differently and
thermodynamic parameters may change accordingly.
Let us consider a Binary system with elements A and B. For our analysis, we consider
XA mole of A and X B mole of B so that XA + X B = 1
That means we consider total one mole of the system That further means we consider
total number of atoms equal to the Avogadro number, N0 (= 6.023 × 1023),
Ni
Where Xi = is the number of atoms of element i.
N0
Unlike single component system, where we

determine the change in free energy with
temperature, in the binary case we shall find the
change in free energy with the change in
composition at different constant temperature at a
time.
Let us consider the free energy for one mole of
element A is GA and one mole of B is GB.
44
Thermodynamic parameters in a binary system
So before mixing when kept separately, XA mole of A and X B mole of B will have the free
energy of XAGA and XBGB respectively
Total free energy before mixing G0 = X AGA + X BGB
After mixing there will be change in free energy

Total free energy after mixing
G = G0 + ∆Gmix
∆Gmix is the free energy change of the alloy because of mixing
∆Gmix = G − G0 = H − TS − ( H 0 − TS 0 ) = ( H − H 0 ) − T ( S − S 0 )
∆Gmix = ∆H mix − T∆S mix
So, once we determine the change in free energy because of mixing, we can determine the
total free energy after mixing.
Let us first determine, the enthalpy change because of mixing (∆Hmix) and the change in
entropy because of mixing (∆Smix )
Note that system always tries to decrease enthalpy and increase entropy for stability.
45
The change in enthalpy because of mixing, ∆Hmix
We take the following assumptions:

The molar volume does not change because of mixing
Bond energies between the pure elements do not change with the change in composition
We neglect the role of other energies.
After mixing, the system can have three different types of bonding, A-A, B-B and A-B
Enthalpy of mixing can be expressed as ∆H mix = N 0 ZX A X B ∆ε
N0 - Avogrado number, Z = coordination number
The change in internal energy

1
∆ε = ε AB − (ε AA + ε BB )
2
εAB is the bond energy between A and B
εAA is the bond energy between A and A
εBB is the bond energy between B and B
It can be written as
∆ H mix = Ω X A X B
Where Ω = N 0 Z∆ε
46
Situation 1: Enthalpy of mixing is zero
∆Hmix = ΩXAXB = 0
1
That means ε AB = (ε AA + ε BB )
2
There will be no preference to choose
neighboring atoms. Atoms can sit randomly at
any lattice points .
Situation 2: Enthalpy of mixing is less than zero
∆Hmix = ΩXAXB < 0

1
That means ε AB < (ε AA + ε BB )
2
Because of transformation internal energy will
decrease. That means transformation is exothermic.
Atoms will try to maximize A-B bonds.
47
Situation 3: Enthalpy of mixing is greater than zero
∆Hmix = ΩXAXB > 0

1
That means ε AB > (ε AA + ε BB )
2
Because of transformation internal energy will

increase. That means transformation is to be
endothermic. Atoms will try to maximize A-A and
B-B bonds.
48
Slope/maximum/minimum of the enthalpy of mixing curve
∆H mix = ΩX A X B = Ω( X B − X B2 )
d (∆H mix )
= Ω(1 − 2 X B )
dX B
d (∆H mix )
At maximum/minimum =0 This implies XB = 0.5. That means maximum
dX B or minimum will be at XB =0.5
d (∆H mix )
Further at X B → 0 = Ω That means the slope at the beginning
has a finite value of x
dX B
49
The change in entropy because of mixing, ∆Smix
Since we are considering transformation at a particular temperature, the change in entropy

because of the change in temperature can be neglected.
We need to consider only the configurational entropy change. Configurational entropy
change comes from the possibilities of different ways of arrangement of atoms.
Following statistical thermodynamics the configurational entropy can be expressed as
S = l ln ω k is the Boltzmann constant

w is the measure of randomness
∆ S mix = S − S 0 = k ln ω − k ln 1 = k ln ω
since atoms at their pure state before mixing can be arranged in only one way
If we consider the random solid solution, then

(n A + nB )!
ω= nA and nB are the number of atoms of A and B
n A ! nB !
Following Stirling’s approximation ln N ! = N ln N − N
50
The change in entropy because of mixing, ∆Smix
So, ∆Smix can be written as
∆Smix = k lnω = k{[(nA + nB ) ln(nA + nB ) − (nA + nB )] −[nA lnnA − nA ] −[nB lnnB − nB ]}
 nA nA 
∆S mix = − k n A ln + nB ln 
 n A + n B n A + n B 
Number of atoms can be related to the mole fraction, X and the Avogadro number N0 following
nA = X A N0 nB = X B N 0
X A + XB =1 n A + nB = N 0
∆S mix = − kN 0 [ X A ln X A + X B ln X B ]
= R[ X A ln X A + X B ln X B ]
where, R is the gas constant
51
Slope/maximum of the entropy of mixing curve
d (∆S mix )  1 1  XB
= − R − ln(1 − X B ) − (1 − X B ) + ln X B + X B  = − R ln
dX B  (1 − X B ) X B 1− X B
d (∆S mix )
= 0 at maximum, this corresponds to XB = 0.5
dX B
Further, the slope at XB →0 is infinite. That means
the entropy change is very high in a dilute solution
As mentioned earlier the total free energy after mixing
can be written as G = G0 + ∆Gmix
where G0 = X AG A + X B GB
∆Gmix = ∆H mix − T∆S mix
∆H mix = ΩX A X B
∆S mix = − R[ X A ln X A + X B ln X B ]
So ∆Gmix can be written as ∆Gmix = ΩX A X B + RT [ X A ln X A + X B ln X B ]
Following, total free energy of the system after mixing can be written as
G = X AG A + X B G B + ΩX A X B + RT [ X A ln X A + X B ln X B ] 52
Unmixed state
Entropy of mixing - Schematic
In the case of two gases initially separated by a barrier, which
is allowed to mix on the removal of the barrier: the number of
mixed states are very large compared to the unmixed states.
Hence, if all configurational states are accessible, the system
will more likely found in the mixed state. I.e. the system will
go from a unmixed state to a mixed state (worded differently
the system will go from ‘order’ to ‘disorder).
On the other hand it is unlikely (improbable) that the system
will go from mixed state to a unmixed state. (Though this is
not impossible → i.e. a mixed system can spontaneously get
‘unmix’ itself!!)
Mixed states with ‘various’ degrees of mixing
Note: the profoundness of the concept of entropy comes

from its connection to energy (via T)
18 mixed states
2 unmixed states
* We assume that all states have equal probability of

Unmixed state 53
occurring and are all accessible
Free energy of mixing
We need to consider three situations for different kinds of enthalpy of mixing
Situation 1: Enthalpy of mixing is zero
G = G0 + ∆Gmix
= G0 − T∆S mix
= X AGA + X BGB + RT[ X A ln X A + X B ln X B ]
With the increase in temperature, -T∆Smix

will become even more negative.
The values of GA and GB also will decrease.
Following the slope Go might change since

GA and GB will change differently with
temperature.
54
Situation 2: Enthalpy of mixing is negative
G= XAGA + XBGB +ΩXAXB + RT[XA lnXA + XB lnXB]
G0 Negative -T∆Smix
∆Hmix
Here both ∆Hmix and -T∆Smix are negative
With increasing temperature G will become even

more negative.
Note that here also G0 may change the slope

because of change of GA and GB differently with
temperature.
55
Situation 3: Enthalpy of mixing is positive
G = X AG A + X B GB + ΩX A X B + RT [ X A ln X A + X B ln X B ]
G0 Positive ∆Hmix -T∆Smix

∆Hmix is positive but -T∆Smix is negative. At lower
temperature, in a certain mole fraction range the
absolute value of ∆Hmix could be higher than T∆Smix
so that G goes above the G0 line.
However to GA and GB, G will always be lower than
G0 , since ∆Hmix has a finite slope, where as ∆Smix
has infinite slope. The composition range, where G
is higher than G0 will depend on the temperature i.e.,
-T∆Smix
Because of this shape of the G, we see miscibility
gap in certain phase diagrams.
At higher temperature, when the absolute value of -
T∆Smix will be higher than ∆Hmix at all
compositions, G will be always lower than G0 56
Free energy Vs Composition Phase Diagrams
With decreasing temperature, free energy
for the solid, GS and liquid phase GL will
change differently with different rates of
change of GA and GB.
At higher temperature, GL< GS (fig a), so
that the liquid phase is stable. At lower
temperature GS < GL (fig e) so that the solid
phase is stable.
In between, GS and GL intersects to find the
situation as explained in Fig. c.
Common tangent drawn gives the
composition range where both solid and
liquid phases can be found together.
That means, if average composition is XB*
as shown in the phase diagram, it will have
S
the solid phase with composition of X B and
the liquid phase with composition of X BL
Both the liquid and solid phases can stay together Phase fractions can be found with the help
because overall free energy will be less than the free of lever rule. 57
energy when considered separately.
This kind of phase diagram is found

when the system has positive enthalpy
of mixing, especially for the solid
phase.
Because of the shape of the GS at low
temperature as as shown in Fig. e, it is
possible to draw a common tangent,
which indicates that the solid phase
with the average composition between
α1 and α2 will have a phase separation .
This corresponds to the miscibility gap
in the phase diagram.
It may also come to the situation, as it
is shown in Fig. c, GL and GS intersects
twice. In that case these will be two
separate regions where both the solid
and liquid phases coexist.
58
In this system, there are two solid

state phases and one liquid phase.
So three free energy curves
should be considered.
At certain temperature, TE, one
common tangent can be drawn,
which will touch all the free
energy curves.
This indicates that all the three
phases at a particular
composition E, as shown in Fig.
d and f can coexist. This is called
eutectic point.
59
β phase is an intermetallic compound that is an ordered phase with very narrow homogeneity range.
60
Sometimes β phase can be found with wider homogeneity range.
Concept of the chemical potential and the activity of elements
Gibb’s free energy, G is function of temperature, T, pressure, P and amount of elements, nA,
nB.
G = G (T, P, nA, nB ……)
At particular temperature and pressure, partial derivative gives
∂G ∂G
dG = dn A + dnB
∂n A ∂nB
= µ A dn A + µ B dnB
∂G
= µA is the chemical potential of element A. It measures the change in free
∂n A energy because of very minute change of element A.
∂G
= µB is the chemical potential of element B. It measures the change in free
∂nB energy because of very minute change of element B.
It should be noted here that the change should be so minute that there should not be any
change in concentration because the chemical potential is a concentration dependent
parameter.
61
Let us consider an alloy of total X moles where it has XA mole of A and XB mole of B.
Note that x is much higher than 1 mole.
Now suppose we add small amounts of A and B in the system, keeping the ratio of XA : XB
the same, so that there is no change in overall composition.
So if we add four atoms of A, then we need to add six atoms of B to keep the overall
composition fixed. Following this manner, we can keep on adding A and B and will reach
to the situation when XA mole of A and XB mole of B are added and total added amount is
XA + XB = 1
Since previously we have considered that the total free energy of 1 mole of alloy after
mixing is G, then we can write
G = µAX A + µAX B
Previously, we derived
G = X AG A + X B GB + ΩX A X B + RT [ X A ln X A + X B ln X B ]
Further, we can write X A X B = X A2 X B + X A X B2
G = X A (G A + ΩX B2 + RT ln X B ) + X B (GB + ΩX A2 + RT ln X B )
62
Further, comparing the expressions for free energy, we can write
µ A = G A + ΩX B2 + RT ln X A
µ A = G A + Ω(1 − X A ) 2 + RT ln X A
µ B = GB + Ω(1 − X B ) 2 + RT ln X B
In terms of activity
µ A = G A + RT ln a A
µ B = GB + RT ln a B
So the relations between the chemical potential and activity are
RT ln a A = Ω(1 − X A ) 2 + RT ln X A
RT ln a B = Ω(1 − X B ) 2 + RT ln X B
63
Activities and chemical potentials are

determined from a free energy curve after taking
a slope, as explained in the figure.
If we are interested to determine the activities or
chemical potentials in a binary system A-B, let
say at X B* , we need to take a slope on the free
energy curve at the free energy, G * and extended
it to pure element, A (XB = 0) and pure element
B (XB = 1).
The point at which it hits NB = 0, the value corresponds to the chemical potential of element
A ( ). µ *A
From previous slide, we can write − RT ln a A = G A − µ *
A
So, once the chemical potential is known, the activity of the element can be calculated using
the above equation, as explained in the graph.
It can be proved that, by taking slope and then extending to XB = 0 and XB = 1, we can find
the chemical potentials. We can write G * = µ *A + ab
Further we can write ab = cd ⇒ ab = X * cd = X * ( µ * − µ * ) ⇒ G * = µ * X * + µ * X *
X B* 1 B B B A A A B B
64
Further, we can write
 aA  Ω  aB  Ω
ln  = (1 − X A ) 2 ln  = (1 − X B ) 2
 X A  RT  X B  RT
Ω  Ω 
a A = X A exp  (1 − X A ) 2  = γ A X A aB = X B exp  (1 − X B ) 2  = γ B X B
 RT   RT 
Ω  Ω 
γ A = exp  (1 − X A ) 2  γ B = exp  (1 − X B ) 2 
 RT   RT 
γi are the activity coefficient of element i.

In an ideal solution, Ω = 0 , and Ai = Xi.
In a non ideal solution, activity coefficient indicates the deviation from the ideal
line.
65
Ω 
γ A = exp  (1 − X A ) 2 
 RT 
Ω 
γ B = exp  (1 − X B ) 2 
 RT 
In the case of positive enthalpy of mixing, activity deviates positively and in the case of
negative enthalpy of mixing activity deviates negatively from the ideal mixing line.
 Ω 
X A → 0, γ A → exp 
 RT  X A → 1, γ A → 1
 Ω  X B → 1, γ B → 1
X B → 0, γ B → exp 
 RT 
Henry’s law: activity of elements is more Rault’s law: activity is equal to the mole
or less the constant in a very dilute fraction near the mole fraction of 1.
66
solution.
Equilibrium vacancy concentration
Diffusion of elements is possible because of the presence of defects. For example,
substitutional diffusion occurs because of exchange of an atom with vacancies. Further,
impurities are present at the interstitial positions, which could diffuse to any neighboring
vacant interstitial sites. This is called interstitial diffusion.
Certain concentration of these impurities are always present in the pure elements in
equilibrium condition, which will be discussed here.
Further in the intermetallic compounds (ordered phases), antisites defects also could be
present along with vacancies. Moreover, different sublattices could have different
concentration of defects, which makes the diffusion process complicated.
Let us not get confused with the structural defects present in pure elements.
To understand simple atomic mechanism of diffusion at this point, we shall calculate the
equilibrium concentration of point defects that is the equilibrium concentration of
vacancies and interstitial atoms in pure elements
Interesting point to be mentioned here that unlike dislocations, grain boundary or
interfaces, point defects are equilibrium defects. It means that the system will try to have
certain concentration of defects to decrease the free energy.
Let us first calculate the equilibrium concentration of vacancies, because of which
lattice substitutional diffusion is possible
67
Equilibrium vacancy concentration in a pure element
Pure elements A and vacancies (V) can be considered as a mixture of two entities
and follow the similar treatment as we did to calculate the total free energy after
mixing of elements A and B.
It can be easily visualized how the movement of atoms can create vacancies in
the material in few steps.
Equilibrium number of vacancies is so small that we can neglect the interaction
between them.
This means that we are neglecting V-V interactions. This also means that we are,
at this point, neglecting the presence of divacancies, which are actually present in
the material.
68
If we consider that the number of vacancies is very small then the increase of enthalpy
because of formation of vacancies can be written as
∆X V
∆H ≈ X V ∆H V is the mole fraction of vacancy and
∆HV is the increase in enthalpy because of one mole of vacancies
There will be the change in the pattern of vibration of atoms next to vacancies because of
extra free space. The increase in (thermal) entropy because of the extra freedom of
vibration can be written as
∆Sthermal = X V ∆SV ∆SV is the increase in entropy for one mole of vacancies
In both the cases, we are actually assuming linear relationship because of very small
concentration of vacancies.
NOTE: In the case of entropy change of A-B mixture, we did not consider the thermal
entropy change, since we are considering the change at a particular temperature. However,
in this case, although we are considering again at a particular temperature, we still need to
consider the contribution because of the change in vibration pattern of atoms next to
vacancies.
69
Further there will be change in configurational entropy considering the mixing of A and V
and can be expressed as (Note that we are considering XA + XV = 1
∆S config = − R[ X V ln X V + X A ln X A ] = − R[ X V ln X V + (1 − X V ) ln(1 − X V )]
Total entropy of mixing ∆S mix = ∆SV X V − R[ X V ln X V + (1 − X V ) ln(1 − X V )]

(Total contribution from thermal and configurational entropy)
Total free energy in the presence of vacancies
G = G A + ∆G
= G A + ∆H − T∆S
= G A + X V ∆H V − T {∆SV − R[ X V ln X V + (1 − X V ) ln(1 − X V )]}
Note here that G of element A when vacancies are

present decreases. So always there will be vacancies
present in materials. Further G decreases to a minimum
value and then increases with the further increase in
vacancy concentration. So, in equilibrium condition,
certain concentration of vacancies will be present,
which corresponds to Ge.
70
Since the slope is zero at the minima, we can find the equilibrium vacancy concentration
from dG
=0
dX V
 1 1 
∆H V − T∆SV + RT ln X V + X V . − ln(1 − X V ) − (1 − X ). =0
 XV (1 − X V ) 
Since the number of vacancies that can be present in the system is very small
1− XV ≈ 1
∆H V − T∆SV + RT ln X V = 0
The relation for equilibrium concentration of vacancies can be written as

 ∆H V − T∆SV   ∆GV 
X Ve = X V = exp −  = exp− 
 RT   RT 
Although there is increase in configurational entropy because of the presence of vacancies,

it is not included in the activation energy barrier, ∆GV.
Although there is activation energy barrier present because of formation, vacancies will
still be present because of overall decrease in free energy of the materials. 71
From the equation in previous slide, we can also write

 ∆H V 
X Ve = X V0 exp − 
 RT 
∆HV is the activation enthalpy for the formation of vacancies and
 ∆S  Is the pre-exponential factor.
X V0 = exp V 
 R 
The change in concentration of vacancies in pure Cu is shown in below graph,
considering activation energy for vacancy formation to be 83.7 kJ/mole.
The concentration of vacancies in mole

fraction at 1000°C is 3.68 × 10-4. That
means in a cube of 14 atoms in each
direction, only one vacancy will be present
in Cu at this temperature!!!
72
Equilibrium concentration of interstitial atoms
In many metals, especially in transition metals, interstitial atoms such as carbon, nitrogen,
oxygen, hydrogen can be present up to a certain extent depending on the metal.
There are mainly two types of interstitial voids present in the structure, where these atoms
can sit: tetrahedral (surrounded by four solvent atoms) and octahedral (surrounded by six
solvent atoms).
However, since impurities prefer to occupy octahedral interstices, we shall mainly
consider this type of voids only.
LET US FIRST CONSIDER A BCC CRYSTAL
In general the size of the interstitial atoms is much larger

than the interstitial void and lattice surrounding the
interstitial will be strained. That means enthalpy of the
system will be increased.
Consider the increase in enthalpy because of addition of
one mole of interstitial atoms is , ∆HI.
The enthalpy increment of the system because of addition
of XI is expressed by ∆H = X ∆H
I I
nI
whereX I = Number of interstitial atom is nI 73
No
Further, there will be two different types of contribution on entropy
Vibration of atoms A, next to interstitial atoms will change from normal mode of vibration
and will be more random and irregular because of distortion of the lattice
∆S thermal = X I ∆S I
∆SI is the change of the entropy of one mole of atoms because of change in vibration pattern
From the crystal structure, we can say that for 2 solvent atoms there are 6 sites for
interstitial atoms. So if we consider that there are N0 numbers of A atoms then there will be
3N0 numbers of sites available for interstitial atoms.
In other sense, we can say that nI atoms will randomly occupy in 3N0 sites available. So the
configurational entropy can be written as
3N 0!
∆S config = k ln w = k ln Following Stirling’s approximation ln N != N ln N − N
nI !(3N 0 − nI )
∆S config = k [3 N 0 ln 3 N 0 − nI ln nI − (3 N 0 − nI ) ln(3 N 0 − nI )]
 n 3 N 0 − nI 
∆S config = R 3 ln 3 N 0 − I ln nI − ln(3 N 0 − nI ) 
 N0 N0 
74
 3N 0 n nI 
∆S config = R 3 ln − I ln 
 3 N 0 − n I N 0 3 N 0 − n I 
 3 XI 
∆S Config = R 3 ln − X I ln 
 3 − X I 3 − X I 
So the total entropy change
 3 XI 
∆S = X I ∆S I + R 3 ln − X I ln 
 3 − X I 3 − X I 
Free energy in presence of interstitial impurities
G = G A + ∆G = G A + ∆H − T∆S I
 3 XI 
G = G A + X I ∆H I − TX I ∆S I − RT 3 ln − X I ln 
 3 − X I 3 − X I 
75
To find the equilibrium concentration, we need to take
dG
=0
dX I
 3 XI XI 
∆H I − T∆S I − RT − − − ln X I + ln(3 − X I ) − =0
 3− XI XI 3− XI 
XI
∆H I − T∆S I + RT ln =0 since X I ≈ 0
3− XI
 ∆H I − T∆S I   ∆GI 
X I = 3 exp −  = 3 exp− 
 RT   RT 
GI is the activation barrier.
76
LET US NOW CONSIDER A FCC CRYSTAL
If we consider FCC crystal, then the number of sites available for interstitial atoms are 4.
Further in a FCC unit cell, total 4 solvent atoms are accommodated. So we can say that for
N0 solvent atoms there will be N0 sites available for interstitial atoms. Like previous
example, we consider nI interstitial atoms which will occupy randomly.
Then following similar procedure, we can show the equilibrium concentration of interstitial
atoms present in a metal with FCC crystal as
 ∆GV 
X I = exp − 
 RT 
So in general, we can write that the equilibrium concentration of interstitial impurities
present is  ∆GI 
X I = B exp −  B depends on the crystal structure
 RT 
 ∆H I 
X I = X I0 exp − 
 RT 
∆HI is the activation enthalpy for interstitial impurities
 ∆S  is the pre-exponential factor
X I0 = B exp 
 R 
77
Carbon Solubility in Iron
Solubility of carbon in Fe = f (structure, temperature)
Where is carbon located in iron lattice ?
Octahedral
sites
Octahedral
sites
6 faces sharing with two sides (6/2)=3 One interstitial site in center plus
12 edges sharing with four sides (12/4)=3 12 edges sharing with four sides (12/4)=3
Total sites is (3+3), 6 per unit cell Total sites is (1+3), 4 per unit cell
78
Every one Fe atom we have 3 interstitial sites Every one Fe atom we have 1 interstitial site
Why concentration of carbon in α-Fe with BCC structure is less
than γ-Fe with FCC structure?
FIRST LET US CONSIDER FCC STRUCTURE (γ-Fe)
Packing factor of FCC lattice 0.74
This indicates the presence of voids. Let us discuss it more elaborately.
In a FCC crystal there are mainly two types of voids:
Tetrahedral: coordination number (CN) is 4. That means the void is surrounded by 4 atoms.
Octahedral: CN is 6. That means the void is surrounded by 6 atoms.
There are 8 tetrahedral voids in a unit cell. That means 2 voids per atom.
There are 4 octahedral voids in a unit cell. That means 1 void per atom.
However, impurities prefer to occupy octahedral voids.
Because the ratio of the radius of the tetrahedral void to atom is 0.225 and the same for the
octahedral void is 0.414.
The ratio of the radius of the carbon atom (size is 77 pm) to Fe (when it has FCC crystal) is
0.596.
So when a carbon atom occupies any void, lattice will be distorted to increase the enthalpy.
Distortion will be less if it occupies the octahedral voids.
Although it increases the enthalpy, carbon atoms will be present up to a certain extent
because of the gain in entropy, as explained previously, which decreases the free energy.
79
FCC
o
= 1.292 A
Fe Void (Oct) Void (Tet)
Size of Fe atom
CCP crystal rFCC
o FeFCC C
Size of the OV
Fe
xFCC (oct ) = 0.534 A
o
r = 0.77 A
C Relative sizes of voids w.r.t atoms
Size of Carbon atom
Note the difference in size of the atoms

BCC
o
Size of Fe atom
BCC crystal r Fe
BCC = 1.258 A Void (Oct) Void (Tet)
o
(d .tet ) = 0.364 A
Fe FeBCC C
Size of the TV
x BCC
o Relative sizes of voids w.r.t atoms
Size of the OV x Fe
BCC (d .oct ) = 0.195 A
Remember
FCC Size of the largest atom which can fit into the tetrahedral is 0.225 and octahedral void is 0.414
BCC Size of the largest atom which can fit into the d.tetrahedral is 0.29 and d.octahedral void is 0.154
80
NOW LET US CONSIDER THE BCC UNIT CELL (α - Fe)
Packing factor of BCC lattice 0.68. So total void in a BCC unit cell is higher than FCC cell.
However, there are 12 (6 per atom) tetrahedral and 6 (3 per atom) octahedral voids present.
This number is higher than the number of voids present in a FCC unit cell.
Unlike voids in FCC lattice, in a BCC lattice the voids are distorted. That means if an atom
sits in a void, it will not touch all the host atoms.
The ratio of the radius of tetrahedral void to atom is 0.29 and the radius of octahedral void to
atom is 0.155.
The ratio of the radius of the C atom (size is 77 pm) to Fe (when it has BCC crystal) is 0.612.
So it is expected that in a BCC unit cell, impurities should prefer tetrahedral voids.
However, although the octahedral void size is small, planar radius which has 4 atoms on the
same plane is 79.6 pm, which is larger that the C atom size. That means it needs to distort
only other two atoms.
On the other hand if C sits in the tetrahedral void it has to distort all four atoms. So in α – Fe
with BCC unit cell C occupies the octahedral voids
Now the octahedral void size in g-Fe (FCC) is higher than a-Fe (BCC). So naturally the
distortion in a BCC cell will be higher and the activation energy for impurities to occupy a
void in a BCC cell also will be higher.
This is the reason that we find much lower solubility limit of C in a-Fe. 81
Why carbon preferentially sits in the apparently smaller octahedral void in BCC ?
Ignoring the atom sitting at B and assuming the interstitial atom touches the atom at A
o
2a
OA = r + x A = r Fe
BCC = 1.258 A
2
2 6r BCC : 3a = 4 r
r + xA =
3
xA  2 6 
=  − 1  = 0.6329
r  3 
o o o
OX = x A = 0.796 A OY = xB = 0.195 A x Fe
BCC ( d .tet ) = 0.364 A 82
Let us consider the C concentration in α-Fe (BCC)

 77300 J   77300 J 

Wt % = 240 exp − mol  
In terms of mole fraction X 1 = 11.2 exp − mol 
RT   RT 
   
At 727°C, carbon concentration in α-Fe is maximum (0.022 wt%), which is equivalent to

mole fraction of ~10-3. That means there is one C atom present in a cube of 10 atoms in
each direction. That further means that only one carbon atom is present in 3000 octahedral
voids
83
KINETICS
In earlier we are learned about Equilibrium we had seen that the thermodynamic feasibility
of processes is dictated by ‘Thermodynamic Potentials’ (e.g. Gibbs Free Energy, at constant
T, P, Ni).
If (say) the Gibbs Free Energy for a process is negative then the process CAN take place
spontaneously.
However, IF the process WILL actually take place (and if it will take place- how long will it
take to occur?) → will be determined by the ‘Kinetics of the process’.
Deeper the ‘metastable energy well’, higher will be activation energy required to pull the
system out of the ‘well’ and take it to the equilibrium state (or some other metastable state).
To give an example: Fe3C is a metastable phase. it can decompose 3Fe + C after hundreds of
years. ‘Thermodynamics warrants, Kinetics delays’
For a given process to occur heat and mass transfer may have to take place and this would
take time → hence in ‘kinetics’ we deal with time and rates (1/t) .
Thermodynamics dictates that the state should transform to the stable state.
However, how fast this transformation can occur cannot be answered by
thermodynamics alone. Determining the rate of such transformation comes under
the realm of kinetics.
84
KINETICS
For example, consider Figure 1, which shows variation of Gibbs free energy of a system
versus an arbitrary unit (say atomic arrangement).
Thermodynamics tells us that if the system is at some meta-stable state-1, it has to transform
to the stable state-2. The difference of Gibbs energy ( G2 - G1 = ∆GDrive ) is the driving force
for transformation.
However, it should be noted that while moving from

state-1 to state -2, the system has to actually cross
the hump i.e., it has to go through a higher energy
state-A. This state-A is called as activated state and
the difference in Gibbs energy ( GA – Gmeta ) is
called as energy barrier since it actually resists the
system from transforming from metastable state-1
to stable state-2.
The energy ΔGa is also called as Gibbs free energy
for activation of the process.
The energy required to overcome the barrier is usually provided by thermal vibrations of the
atoms and hence these processes are called as “thermally activated processes”. 85
KINETICS
The kinetics tells us that the rate of the process should be proportional to:
 − ∆Ga 
Rate ∝ exp 
 RT 
(or )
 − ∆Ea 
Rate = K o exp  Ko is a constant and Ea is the energy (or
 RT  enthalpy) of activation.
Most of the transformations are brought about by diffusion mechanisms. The diffusion itself
is a thermally activated process. Thus, diffusion coefficient can be expressed as:
 − QD  Do is a constant and Qd is the activation

D = Do exp  energy of diffusion
 RT 
86
Questions?
−3
1. The specific heat of solid copper above 300 K is given by C P = 22.64 + 6.28 ×10 T J mole K .
-1 -1
By how much does the entropy of copper increase on heating from 300 to 1358 K?
2. Draw the free energy temperature diagram for Fe up to 1600oC showing all the allotropic
forms.
3. From the thermodynamic principles show that the melting point of a nanocrystalline metal
would be different from that of bulk metal. Will the melting point be different if the
nanoparticle is embedded in another metal?
4. What is Clausius-Clapeyron equation? Apply it for the solidification of steels and gray cast
iron.
5. If the equilibrium concentration of vacancies in terms of mol fraction at 600 °C is 3x10-6
calculate the vacancy concentration at 800 °C. R = 8.314 J/mol.K
For aluminium ∆HV = 0.8 eV atom-1 and ∆SV/R = 2. Calculate the equilibrium vacancy
concentration at 660oC (Tm) at 25oC
6. Derive ∆SConf = − R[ X A ln X A + X B ln X B ]
mix
7. Show that the free energy of a mixture of two phases in equilibrium in a binary system is
given by the point on the common tangent line that corresponds to that overall composition.
Show that the free energy of a mixture of phases of any other composition or a single phase
is higher.
87
Questions?
8. Draw the G-X diagrams at the temperatures shown in Fig’s. Please note that the crystal
structure of both the pure metals in the phase diagram is the same.
9. For a single-component system, why do the allotropes stable at high temperatures have
higher enthalpies than allotropes stable at low temperatures.
10. Derive following equations G = X G + X G + ΩX X + RT ( X ln X + X ln X )
A A B B A B A A B B
µ A = G A + Ω(1 − X A )2 + RT ln X A
11. Calculate the free energy of Si at 1186 °C from the known values, as given below:
A = 23.698, B = 3.305 × 10 −3 , C = 4.35 × 105 and ∆Sθ298 = 18.81J / mol.k
88
Questions?
12. Identify the errors in the binary phase diagram given in Fig., give the reasons and draw the
corrected phase diagram.
13. Consider an alloy of elements A and B, with the composition of XA=0.25 andXB=0.75. If
the free energy of the pure elements, A (GA) and B (GB) are -45.1 kJ and -48.3 kJ/mole at
950°C, then (a) Calculate the free energy after mixing at this temperature for two cases,
whenenthalpy of mixing ∆Hmix= -3.1 kJ/mole and +2.7 kJ/mole.(b) Further, calculate the
activity and chemical potential of element A at thattemperature for∆Hmix= -3.1 kJ/mole.
14. Derive the Gibb’s phase rule.
89
Questions?
15. 15 g gold and 25 g of silver are mixed to form a single-phase ideal solid solution
a) How many moles of solution are there?
b) What are the mole fractions of gold and silver?
c) What is the molar entropy of mixing?
d) What is the total entropy of mixing?
e) What is the molar free energy change at 500oC?
f) What are the chemical potentials of Au and Ag at 500oC taking the free Au atom is
added? Express your answer in eV atom-1.
g) By how much will the free energy of the solution change at 500oC if one Au atom is
added? Express your answer in eV atom-1
90
Avala Lava Kumar*
*E-mail : lavakumar.vssut@gmail.com 91
Introduction
Diffusion is defined as, random movement of atoms/ molecules in solid, liquid and gas. For
example dissolution of ink in water and smoke in air
It can also defined as, Diffusion is a process of mass transport by atomic movement under
the influence of thermal energy and a concentration gradient.
To comprehend many materials related phenomenon one must understand diffusion.
Role of Diffusion
Oxidation Creep
Sintering Aging
Doping Carburizing
Many More...!
Materials Joining : Diffusion Bonding 92

Diffusion Phenomena
Mass flow process by which species change their position relative to their neighbors.
Driven by thermal energy and a gradient
Thermal energy → thermal vibrations → Atomic jumps
Atoms move from higher to lower concentration region. If this movement is from one
element to another e.g. Cu to Ni, then it is termed as Inter-diffusion. If the movement is
within similar atoms as in pure metals, it is termed self-diffusion.
Gradient
Electric Stress
Magnetic Chemical potential
93
Diffusion under thermodynamic driving force
As explained before and in the figure, the

chemical potential of elements at any
composition can be determined by taking a slope
on the Gibb’s free energy curve and then by
extending it to XB = 0 (chemical potential of
element A, µ A ) and XB = 1 (chemical potential
of element B, µ B)
From the chemical potentials, we can determine
the activities, aA and aB from the knowledge of
free energies of pure elements.
− RT ln a A = G A − µ A − RT ln aB = GB − µ B
So it must be clear that µ B decreases from GB at XB = 1 to a infinitely small value close to
XB = 0.
Following, aB decreases from 1 at XB = 1 to 0 at XB = 0. It may vary ideally, that is aB = XB
or deviate positively/negatively, as explained previously.
Similarly, µ A decreases from GA at XB = 0 to a infinitely small value close to XB = 1. aB
decreases from 1 at XB = 0 to 0 at XB = 1
So, µ A and aA, and μB and aB follow the same trend of increasing or decreasing with XB.
94
Now let us consider, two different AB alloys P (A rich) and Q (B rich).

If these alloys are kept separately, these alloys will have free energy of GP and GQ,
respectively.
However, if we bond these two blocks, they will not be anymore in equilibrium condition.
If the amount of material P and Q are taken such that overall mole fraction is R, then the
equilibrium free energy will be GR.
So the system will try to move to its new equilibrium free energy.
Now if we don’t melt the alloy, it can only move to the equilibrium composition by solid
state diffusion.
95
If it is held at elevated temperature at which diffusion rate is reasonable, the system will
reach to the equilibrium after certain time depending on the diffusion rate.
It can be seen that at equilibrium alloy R, the chemical potential of A is µ A , which is lower
R
than the chemical potential of the same element, µ A , in alloy P but higher than in alloy Q,
P
that is µ A .
Q
On the other hand, µ B is less than µ B but higher than µ B .

R Q P
So to reach the equilibrium, the alloy P should decrease the chemical potential of A and
increase the chemical potential of B. On the other hand, the alloy Q should increase the
chemical potential of A and decrease the chemical potential of B.
Since the chemical potential of A decreases from the A rich to the B rich alloy and the
chemical potential of B decreases from B rich to A rich alloy, P should decrease the content
of A and Q should decrease the content of B.
In this condition it is possible only by the diffusion of element A out of P and diffusion of
element B out of alloy Q, as shown in the figure in the previous slide.
If we allow the system to go to equilibrium at temperature T, there will be no composition in
the blocks P and Q at time zero that is t0. Then with the increase in time in the order of t3 >
t2 > t1 > t0 , as shown in the figure (next slide), interaction zone increases. Following, after
infinite time of annealing, tα it will reach to equilibrium composition.
96
Note here that infinite time is rather notional. It means that long time enough to reach it to
equilibrium. So this time will depend on the thickness of the blocks and the diffusion rate at
the temperature of annealing.
97
Now let us consider a system with miscibility gap.
If we anneal two block with any compositions at
temperature, T2, the explanation of the diffusion process
will be similar as explained before.
However, if we couple two blocks with the composition
of P and Q, which are inside the miscibility gap, different
situation will happen.
From the slopes taken at P and Q, we find
µ AP < µ AQ and µ BP < µ BQ
That means, the chemical potential of A is less in A rich
alloy, P compared to the A lean alloy, Q.
Similarly, the chemical potential of B is higher in B lean
alloy, P compared to the B rich alloy, Q.
If we couple blocks of P and Q then the average free Gα1 Gα 2
energy of the systems, let say, R depending on the relative
amounts of P and Q.
However, since the system always tries to decrease free
energy, it will try to reach to the lowest free energy Gα1
and Gα 2
98
That means A rich alloy P should increase the content of A and the B rich alloy Q should
increase the content of B.
From the chemical potential point of view also it must be clear that B will diffuse out of the
B lean alloy P towards Q and A will diffuse out of the B lean alloy Q towards P.
The direction of elements is just opposite compared to the previous example explained.
Since elements diffuse up the concentration gradient instead of down the concentration
gradient, it is called uphill diffusion.
In terms of chemical potential gradient
µ AP − µ AQ µ AP − µ AQ dµ A
J Aα ⇒ J Aα ⇒ J A = − DA Since A diffuse from Q to P and
∆x ∆x dx and B diffuse from P to Q
µ BQ − µ BP µ BQ − µ BP dµ
J Bα ⇒ J Aα ⇒ J B = − DB B
∆x ∆x dx
In terms of concentration gradient
C AP − C AQ dC A
J A = DA ⇒ J A = DA
dx dx
C BQ − C BP dC B
J B = DB ⇒ J B = DB
dx dx
It can be seen that negative sign is not present in the Fick’s first law for this case.
99
In previous slides we have shown diffusion of elements because of chemical potential driving
forces. However, diffusion occurs even without the presence of thermo dynamical driving force
or any other driving forces.
For example, in pure material, where there are no forces present but atoms still to jump to another
position.
In a low melting point material, like in Sn or Pb, jump of atoms possible even at room
temperature. However, jump of atoms in this case is completely random. Presence of driving
forces rather make net flux of atoms to a particular direction.
To test the possibility of diffusion without any driving forces tracer diffusion experiments are
conducted. Radiotracer elements which has one or two neutron mass difference are deposited on a
material of the same element and then annealed at elevated temperature.
Concentration of these tracer elements can be measured at different depths even in low
concentration because of radiation of different rays
.As shown in a schematic figure, these tracer elements diffuse
inside the material. Since both are pure elements there is no
chemical potential difference.
These tracer elements are used to trace the diffusion of elements.
There will be very small gain (increase in entropy) because of
spreading of these tracer elements.
If we do not use the tracer elements we will not be able to detect
the jump of atoms. 100
Diffusion Mechanism
Diffusion of atoms involves movement in steps from one lattice site to the another. An
empty adjacent site and breaking of bonds with the neighbor atoms are the two necessary
conditions for this.
Vacancy Mechanism
This mechanism involves movement of atoms (we are interested in substitutional atoms)
from a regular lattice site to an adjacent vacancy. Since vacancy and atoms exchange
position, the vacancy flux is in the opposite direction.
101
Diffusion Mechanism
Interstitial Mechanism
This mechanism Involves migration of atoms from one interstitial site to a neighboring
empty interstitial site.
Usually the solubility of interstitial atoms (e.g. carbon in steel) is small. This implies that
most of the interstitial sites are vacant. Hence, if an interstitial species wants to jump, ‘most
likely’ the neighboring site will be vacant and jump of the atomic species can take place.
This mechanism is more prevalent for impurity such a hydrogen, carbon, nitrogen, oxygen
which are small enough to fit into an interstitial position.
Interstitial
Interstitial
atom
atom
102
Diffusion Mechanism
Atom Interchange Mechanism
It is possible for movement to take place by a direct interchange between two adjacent
atoms or by a four –atom ring interchange.
However, these would probably occur only under special conditions, since the physical
problem of squeezing between closely packed neighboring atoms would increase the barrier
for diffusion.
Note: The rate of diffusion is much greater in a rapidly cooled alloy than in the same alloy
slow cooled. The difference is due to the larger number of vacancies retained in the alloy by
fast cooling.
103
Diffusion Mechanism
Pipe Diffusion
When diffusion occurs via edge dislocation, it is called pipe diffusion.
Since it feels like movement of atoms through a pipe.

Note that both interstitial and substitutional diffusion can occur via dislocations.
Even impurities are attracted by the dislocations because of availability of more space and
chance to release stress.
This is also the reason (more open space) that it has lower activation barrier and diffusion
rate through dislocation is higher than the diffusion through lattice.
104
Diffusion Mechanism
Grain boundary diffusion
Diffusion occurs via grain boundaries even easily and it is called grain boundary diffusion.
Since grain boundaries are relatively more open structure compared to atomic structure
inside the grains, the barrier for diffusion through grain boundary is much less and the
diffusion rate is higher.
Rate of diffusion increases with the increase in misorientations.
105
Diffusion Mechanism
Surface diffusion
When diffusion occurs over a surface, it is called surface diffusion.
Here activation energy for diffusion is the lowest since there are no atoms above the atom
of interest, which exchanges position with the vacancy. So diffusion rate is the highest
compared to other diffusion mechanisms.
Note that both interstitial and substitutional diffusion can happen through, lattice,
dislocations, grain boundaries and surface.
Slope of the diffusion coefficient vs. 1/T gives the activation barrier for diffusion.
Activation barrier is the lowest and diffusion rate is the highest for the surface diffusion.
Activation barrier is the highest and diffusion rate is the lowest for lattice diffusion. 106
Diffusion Mechanism
Diffusion Couple
When two blocks with different compositions are joined (coupled) together and annealed to
allow diffusion between them, it is called diffusion couple.
Interdiffusion (chemical diffusion)
Since elements in the diffusion couple interdiffuse to each other, it is called interdiffusion.
The diffusion coefficient is in general called as interdiffusion coefficient and if sometimes
any new compound forms during the diffusion at the interface, occasionally, it is called
chemical diffusion coefficient.
Note that actually elements A and B diffuse. Diffusion of these elements are called intrinsic
diffusion of elements.
107
Diffusion Mechanism
Self diffusion
When diffusion occurs without any presence of driving force, it is called self diffusion.
Atoms vibrate at their positions and because of some violent jumps, it can cross the
activation barrier to make the jump successful.
Since there is no driving force to direct the jump of atoms to a particular direction, self
diffusion is truly random in nature.
This indicates that when a pure metal is kept at elevated temperature jump of atoms is
always happening.
In low melting point metals, like In or Sn, even at room temperature, atoms exchange their
position.
However, since these are very small in size, we cannot follow a particular atom jump.
We shall see that the jump can be many orders of magnitude in one second. This makes it
even difficult to follow the jump of atoms.
To obviate this problem, concept of tracer diffusion is introduced.
108
Kirkendall Effect
If the diffusion rates of two metals A and B into each other are different, the boundary
between them shifts and moves towards the faster diffusing metal.
This is known as kirkendall effect. Named after the inventor Ernest Kirkendall (1914-
2005). It can be demonstrated experimentally by placing an inert marker at the interface
A B
Materials A and B welded together with inert Marker

marker and given a diffusion anneal motion
Usually the lower melting component diffuses
faster (say B)
Inert Marker – thin rod of a high melting material which is

109
basically insoluble in A & B
Kirkendall Effect
Zn diffuses faster into Cu than Cu in Zn. A diffusion couple of Cu and Zn will lead to
formation of a growing layer of Cu-Zn alloy (Brass).
Same will happen in a Cu-Ni couple as copper diffuses faster in nickel than nickel in
copper.
Since this takes place by vacancy mechanism, pores will form in cu (of the Cu-Ni couple
as the vacancy flux in the opposite direction (towards Cu) will condense to form pores.
110
steady and non-steady state diffusion
Diffusional processes can be either steady-state or non-steady-state. These two types of
diffusion processes are distinguished by use of a parameter called flux.
It is defined as net number of atoms crossing a unit area perpendicular to a given direction
per unit time. For steady-state diffusion, flux is constant with time, whereas for
non-steady-state diffusion, flux varies with time.
A schematic view of concentration gradient with distance for both steady-state and non-
steady-state diffusion processes are shown below.
Flux (J) (restricted definition) → Flow / area / time [Atoms / m2 / s]

Steady state diffusion Non-Steady state diffusion
C1 C1
Concentration →
Concentration →
D ≠ f(c)
Steady state
D = f(c)
J ≠ f(x,t) t1
t2
C2 t3
Non-steady state C2
J = f(x,t) 111
Distance, x → Distance, x →
steady and non-steady state diffusion
Steady State : Concentration profile not changing with time.
Concentration ‘C’ in the box

J (left) J (right) doesn’t change with time
Steady State
X J(x) Left = J(x) Right
Non--Steady State : Concentration profile changes with time.

Non
dx
Concentration accumulating
in the box
J (left) J (right)
Non-Steady State
J(x) Left ≠J(x) Right
112
Fick’s I Law
Steady-state diffusion is described by Fick’s first law which states that flux, J, is
proportional to the concentration gradient.
The constant of proportionality is called diffusion coefficient (diffusivity), D (cm2/sec).
diffusivity is characteristic of the system and depends on the nature of the diffusing
species, the matrix in which it is diffusing, and the temperature at which diffusion occurs.
Thus under steady-state flow, the flux is independent of time and remains the same at any
cross-sectional plane along the diffusion direction. for the one-dimensional case, Fick’s
first law is given by
J ≡ atoms / area / time ∝ concentration gradient

dc dc 1 dn dc
J∝ J = −D J= = −D
Steady state
dx dx A dt dx
J ≠ f(x,t)
No. of atoms crossing Concentration gradient
dn dc
= − DA
area A per unit time
Diffusion coefficient/ Cross-sectional area
diffusivity dt dx The minus sign in the equation means that
113
diffusion occurs down the concentration gradient
Fick’s I Law
In Ficks I law, minus sign comes from the fact that matter flows down the concentration
gradient. It is not necessarily true in all the cases.
Matter may also diffuse up the concentration gradient, which is called uphill diffusion.
Fick’s first law can directly be applied in the case of
steady state, as shown in the example below.
Steady state means that there will not be any change in
the composition profile with time.
If we pass carburizing gas through the pipe as shown in
the figure and decarburizing gas outside the pipe, a
steady composition profile may develop.
Concentration gradient can be calculated following:
dc C − C0 C − Ci
=− i =− 0
dx d d
From this, one can calculate the flux from the known
diffusion coefficients of carbon or the diffusion
coefficient from the flux determined.
114
The steady-state diffusion is found in the purification of hydrogen gas.
Solved Compute the number of kilograms of hydrogen that pass per hour through a
Example -1 6-mm-thick sheet of palladium having an area of 0.25 m2 at 600°C. Assume a
diffusion coefficient of 1.7 × 10-8 m2 / s, that the concentrations at the high-
and low-pressure sides of the plate are 2.0 and 0.4 kg of hydrogen per cubic
meter of palladium, and that steady-state conditions have been attained.
This Problem calls for the mass of hydrogen, per hour, that diffuses through a pd sheet.
From the Fick’s I law:
∆c
M = JAt = − DAt
∆x
−8  0.4 − 2.0kg / m 3 
= −(1.7 × 10 m / s )(0.25m )(3600 s / h) 
2 2
−3 
 6 × 10 s 
= 4.1×10 −3 kg / h
115
A sheet of BCC iron 2 mm thick was exposed to a carburizing gas atmosphere on one
Solved side and a decarburizing atmosphere on the other side at 675ºC. After having reached
Example -2 steady state, the iron was quickly cooled to room temperature. The carbon
concentrations at the two surfaces of the sheet were determined to be 0.015 and 0.0068
wt%. Compute the diffusion coefficient if the diffusion flux is 7.36 × 10-9 kg/m2-s
(Hint : convert the concentrations from weight percent to kilograms of carbon per
cubic meter of iron.
This problem calls for computation of the diffusion coefficient for a steady-state diffusion
situation. Let us first convert the carbon concentrations from weight percent to kilograms
carbon per meter cubed using below equation.
For 0.015 wt% C    
 C   0.015 
C 'C =  C
× 10 
3 C 'C =  × 10 3  = 1.18kgC / m 3
C C 0.015 99.985
 C+ C   + 
 ρ C ρ Fe   2.25 7.87 
Similarly, for 0.0068 wt% C  

 0.0068 
C '' C = ×103  = 0.535kgC / m3
0.0068 99.9932
 + 
 2 . 25 7 .87 
X − XB  − 2 × 10 −3 m 
D = −J A −9
= −(7.36 × 10 Kg / m − s ) 2
3
C A − CB 1.18 Kg / m − 0.535 Kg / m 
3
116
= 2.3 ×10−11 m 2 / s
Fick’s II Law
However, just the Fick’s first law may not be useful to relate the flux and the concentration
profile, since, in most of the cases, systems develop non steady state concentration profile.
It means that concentration at a particular position changes with time.
For example, if we keep concentration of carbon on one of the Fe surfaces anneal,
composition profile will change with time , as shown in the figure.
We can’t apply Fick’s first law directly to

evaluate the concentration profile that develops
during diffusion in this case, since, as shown in
the figure, composition changes with annealing
time at a particular position.
So we need to have a relation, which can relate
time along with the concentration and the
position.
For that Fick’s second law is derived. It is
derived using conservation of mass and Fick’s
first law.
117
Fick’s II Law
Let us consider a very thin slab in the material.
J1 is the incoming flux, J2 is the outgoing flux
So the amount of carbon coming in short time δt is J1 δt
(mole/m2) and going out is J2δt (here J1>J2)
If the slab thickness is ∆x, then
( J1 − J 2 )δt
δC = Equation 1
∆x
Further, flux change in the thin slab can be considered,
linear and we can write
∂J J 2 − J1 J1 − J 2
= =- Equation 2
∂x ∆x ∆x
From Eq. 1 and Eq. 2 ∂C ∂J
=−
∂t ∂x
Using Fick’s first law
∂C ∂  ∂J  ∂C ∂ 2
J
=  D  If D is Constant → =D 2
∂t ∂x  ∂x  ∂t ∂x
118
Solutions to Fick’s second Law
Solution of the Fick’s second law depends on initial and boundary conditions.
Further, D, in some cases, may be considered constant. That means D does not change with
concentration or position.
However, this condition meets only in very limited cases.
In most of the practical examples, D is a function of concentration. In this case, solution to
the Fick’s second law is complicated.
So in the beginning, we shall solve the Fick’s second law for constant D.
Solutions are mainly for two different types of conditions, small and large time values.
When diffusion annealing time is small, solution is expressed with integrals or error
functions. For long annealing time, it is expressed in terms of trigonometrical series.
Note that the long or short annealing time is rather relative. By saying long annealing time,
we mean that the complete sample is affected by the diffusion process and may lead to
homogenization.
By saying short annealing time, we mean that experiments are conducted such that whole
material is not affected by the diffusion process.
119
Solution for a thin film source
Let us consider that D is constant. That means D
does not change with the composition. Since at a
particular location C changes continuously with
annealing time, t or a particular C changes its
location continuously. From the assumption, we
can state that D is the same at any location.
The meaning of the above statement will be more
clear, when we shall consider the change of D
with the change in C concentration.
Let us consider the situation, when a very thin
film of element B is sandwiched between material
A and then annealed at elevated temperature.
Note: One might ask, how thin it is?
By saying “thin” we mean that the amount of
material B is very low and even after total mixing
(full homogenization) element B can be
considered as impurities in the material A.
That means, after deposition of B on A, the
chemical potential gradient is negligible.
120
In other sense, we can consider this as an example of diffusion in the absence of any driving
force (although small driving force will be there from entropy of mixing)
Actual meaning will be clear, when we discuss about the atomic mechanism of diffusion!
We can consider that D is constant, since the material in which it diffuses has fixed
composition.
∂C ∂  ∂C  ∂ 2C
For constant D, the Fick’s II law can be written as = D =D 2
∂t ∂x  ∂x  ∂x
As shown in the previous slide, it is seen that the element distribution after the diffusion can
be expressed by exponential relation.
Following the Fick’s second law the distribution of element B in A can be expressed as
C0  x2 
C B ( x ) = 1 2 exp −  where Co is a constant
t  4 D t
B 
This relation is developed based on the fact that composition profile shows exponential
decay from thin film source.
The correctness of the solution can be checked by differentiation of the equation with
respect to t and x and then using them in the Fick’s second law to find the equal values on
both the sides.
121
Further, the boundary condition that CB = 0, at x = ∞ at t = 0 and C B = ∞, at x = 0 at t = 0
also meet.
Now one might get confused, when we say C B = ∞ , since concentration of element B (=
XB / Vm = 1/ Vm) can’t be infinite any time and will have some definite value even at
completely pure state.
Here C B = ∞ is notional and means that the element B is not diluted by any other element,
which further means that it is in pure state and for system it means that it has infinite source
of element B before starting the experiment.
We shall show that the absolute values of CB (x) or C0 are not important but the ratio
CB(x)/C0 is important, where this solution is actually used to calculate the tracer or impurity
diffusion coefficient of the species.
Total material B (mole/m2) that was sandwiched
+∞
before annealing can be found following.
M B = ∫ CB
−∞
+∞
CB  x2 
So we can write M B = ∫ 1 2 exp − dx
−∞
t  4 D t
B 
x
Further, we consider λ = means dx = (2 DB t )dλ
2 DB t
122
Since integration gives √Π, we get
+∞
M B = 2C0 DB ∫ exp(−λ2 ) dλ = 2C0 πDB
−∞
Replacing this for MB we find

MB  x2 
C B ( x) = exp − 
2 πDB t  4 D t
B 
CB vs. x describes the distribution of elements B.

dCB /dx describes (following the Fick’s law) the
change in flux with respect to distance at a
particular annealing time.
d 2C B dC B It explains the rate of

2
= vs.x
dx dt change of element B
The negative values indicate the region, where it

loses the element B and positive value indicates that
the region where it gains the element B. Note that
the region where it is losing or gaining the element
B depends on the time of annealing.
123
The change in distribution of elements B with the increase in time is shown in the figure.
MB  x2 
C B ( x) = exp − 
2 πDB t  4 DB t 
Factor 2 comes from the fact that elements diffuse both the sides from the source
124
If the element B diffuses to one direction then the factor 2 should not be considered.
MB  x2 
C B ( x) = exp − 
πDB t  4 DB t 
MB x2
ln C B ( x) = ln −
πDB t 4 DB t
From the calculated slope, one can

determine the diffusion coefficient from
the known annealing time.
125
Solution in semi infinite diffusion couples (error function analysis)
Let us consider now the semi infinite

diffusion couple of two blocks with
concentration of
C B = 0 and CB = C B+
It means that, in a A-B binary system, it
is bonding between two blocks made of
pure A and an alloy of AB.
Unlike the previous case, here, because
of the difference in the composition,
diffusion will be driven by the chemical
potential gradient.
However, we shall show later that the solution we are going to derive, can be used only in
the case where the concentration and the chemical potential difference of the end members
is not much. That means diffusion coefficients do not vary significantly with the
composition.
Semi infinite means that, we anneal for a certain annealing time such that the end of the
initial materials are not affected by the diffusion of elements.
126
Although the meaning of the semi infinite indicates that a good part of the end of the
diffusion couple should not be affected but actually even with one unaffected atomic layer
in the end is sufficient to consider the system as semi infinite diffusion couple!
This is important since otherwise we can’t apply the relation derived here to determine the
diffusion parameter or to calculate the concentration profile from the known diffusion
parameter.
If this kind of situation shown below, one can calculate no. of B atoms are diffusing in a
metal is following way
CB ( x) 1   x 
= 1 + erf  
CB +
2   2 D t 
 B  127
CB ( x) 1   x  C B ( x) 1   x 
= 1 + erf   = 1 − erf  
+
2    −  
CB  2 DB t  CB 2   2 DB t 
It is apparent that the sign of x will depend on which side of the x = 0, we are interested to
calculate.
Further, if the composition profile is just the opposite compared to the first example, it can
be shown that the relation becomes C ( x) 1   x 
B
= 1 − erf  
−
2   
CB  2 DB t 
Note: We always use “+” at superscript for the concentration to denote right hand side of the
128
couple and “-” for the left hand side of the couple.
Tabulated error function values
2 z
erf ( z ) = ∫ exp(−η 2 )dη
π 0
2  z3 z5 z7 
=  z − + − .....
π  3 ×1! 5 × 2! 7 × 3! 
erf (− z ) = −erf ( z )
Advantage of using this error

function is that one can
determine the values from a
table.
129
In the previous case, we considered the diffusion between the diffusion couple of two
−
different blocks with composition CB = 0, and CB+ or otherwise CB = CB and 0.
That means in both the cases, it was a couple between one pure element and an alloy of AB.
Error function solutions given previously can be rewritten interms of normalized
concentration profiles as
C B ( x) − 0 1   x 
  C B ( x) − 0 1   x 
= 1 + erf  = 1 − erf  
−
CB − 0 2    −
CB − 0  
 2 DB t  2   2 DB t 
In some cases, it is possible that two alloys of AB are couples.
− + + −
Now if we consider the diffusion couple between CB and CB , where C B > C B the relation
can be written as shown in Fig a.
C B ( x) − C B− 1   x  or C B+ + C B− C B+ − C B−  x 
= 1 + erf   C B ( x) = + erf  
+ −  2 D t   2 DB t 
CB − CB 2   B 
2 2 
130
− + − +
Further, the relation in the case of diffusion couple, where CB and CB and C B > C B as sho-
wn in figure b. (in the previous slide), it can be expressed as
C B ( x) − C B+ 1   x  C B− + C B+ C B− − C B+  x 
= 1 − erf   Or C B ( x) = − erf  
C B− − C B+  2 D t  2 2  2 DB t 
2   B  
In the case of carburizing, we can see that the relation will be the same, irrespective of the
side on which carburizing is done, since in the case of Fig. a, x is positive but in Fig. b, x is
negative
C ( x ) − C0  x  C ( x ) − C0  −x   x 
= 1 − erf   = 1 + erf   = 1 − erf  
C S − C0  2 Dt  C S − C0  2 Dt   2 Dt 
 x 
C ( x ) = CS − (CS − C0 )erf  
 2 Dt 
  x 
C ( x ) = CS 1 − erf   for C0 = 0 131
  2 Dt  
Few important notes
So from the previous relations, it is apparent that the sign of x (whether positive or
negative) will depend on in which side of the x = 0, we are interested to calculate.
Note again that D will be more or less the constant at any position it is calculated. It was
also one of the assumptions for this derivation.
From the error function analysis above, we can see that, at x =0
C B ( x) − C B− 1   x 
  it comes
C B ( x = 0) − C B− 1
= ⇒ = =
1 +
+ −
from the relation = 1 + erf + −
C ( x 0 ) (C C )
  −
+ − B B
CB − CB C C 2 2 B
2   2 DB t  B B
 x 
from the relation C ( x) = CS − (CS − C0 )erf   it comes C ( x = 0) = C
 2 Dt  S
It indicates that in a system, where two blocks with different concentrations are coupled, at
x =0, which is basically the initial contact plane, the concentration will be always average of
the concentration of the initial materials.
Note that we need to locate anyway the initial contact plane after the diffusion process,
since x in the equation is actually measured from the location of the initial contact plane.
This also indicates that this equation can only be used when end parts of the couple is not
affected by the diffusion process, since otherwise at x =0, it will have another average value
depending on how much of the end members are affected.
132
Few important notes
So from the previous relations, it is apparent that the sign of x (whether positive or
negative) will depend on in which side of the x = 0, we are interested to calculate.
Note again that D will be more or less the constant at any position it is calculated. It was
also one of the assumptions for this derivation.
From the error function analysis above, we can see that, at x =0
C B ( x) − C B− 1   x 
  it comes
C B ( x = 0) − C B− 1
= ⇒ = =
1 +
+ −
from the relation = 1 + erf + −
C ( x 0 ) (C C )
  −
+ − B B
CB − CB C C 2 2 B
2   2 DB t  B B
 x 
from the relation C ( x) = CS − (CS − C0 )erf   it comes C ( x = 0) = C
 2 Dt  S
It indicates that in a system, where two blocks with different concentrations are coupled, at
x =0, which is basically the initial contact plane, the concentration will be always average of
the concentration of the initial materials.
Note that we need to locate anyway the initial contact plane after the diffusion process,
since x in the equation is actually measured from the location of the initial contact plane.
This also indicates that this equation can only be used when end parts of the couple is not
affected by the diffusion process, since otherwise at x =0, it will have another average value
depending on how much of the end members are affected.
133
Solution for homogenization
Let us consider a system with a relatively long time of annealing such that system may
reach to homogenization.
Time is rather relative. In a small system, relatively small time can be considered as long
time because it will not take much time to homogenize. On the other hand in a large system,
even longer time also may not be sufficient.
The coring effect generally seen as cast alloys during solidification, the first solid that forms
is poor in solute. As the solidification progress, the new solid formed becomes richer &
richer in solute.
This give rise to coring i.e., continuous change of solute concentration from center of a
dendrite towards its surface. Thus, the concentration profile in a cored dendritic structure
can be approximated to a sinusoidal wave form as shown in figure.
134
∂C ∂ 2C
=D 2
∂t ∂x
 Πx 
C( x , 0 ) = C + β 0 sin  
 l 
135
 nΠ x   − n 2Π 2 
C( x ,t ) = C + β 0 sin  . exp 2
Dt  or
 l   l 
 Πx   −t  l2
C( x ,t ) = C + β 0 sin  . exp  where τ = 2
 l  τ  n D
 −t 
β (t ) = β 0 exp 
τ 
1
at t = τ , β (τ ) = × β0 136
e
Summary on Fick’s II Law
Process Solution
 x  Cs = Surface concentration
Carburisation C = CS − (C S − C0 )erf  
 2 Dt  C0 = Bulk concentration
 x 
Decarburisation C = C0 erf   C0 =Initial Bulk concentration
 2 Dt 
 C + C2   C1 − C2   x  C1 = Concentration of steel 1
Diffusion couple C = 1 − erf   C2 = Concentration of steel 2
 2   2   2 Dt 
 πx   t Cmean = Mean concentration

C = Cmean + β 0 sin   exp −  β0 = Initial concentration amplitude
Homogenisation λ  τ λ = half wavelength
t = relaxation time
137
An FCC iron–carbon alloy initially containing 0.55 wt% C is exposed to an
Solved oxygen-rich and virtually carbon-free atmosphere at 1325 K (1052ºC).
Example - 3 Under these circumstances the carbon diffuses from the alloy and reacts at
the surface with the oxygen in the atmosphere; that is, the carbon
concentration at the surface position is maintained essentially at 0 wt% C.
(This process of carbon depletion is termed decarburization.) At what
position will the carbon concentration be 0.25 wt% after a 10-h treatment?
The value of D at 1325 K is 4.3 × 10-11 m2/s.
This problem asks that we determine the position at which the carbon concentration is 0.25 wt%
after a 10-h heat treatment at 1325 K when Co = 0.55 wt% C.
C x − Co 0.25 − 0.55  x 
= = 0.5455 = 1 − erf  
C s − Co 0 − 0.55  2 Dt  Z Erf (Z)
 x  0.40 0.4284
erf   = 0.4545
 2 Dt  Z 0.4545
Using tabulation of error function values and linear interpretation
0.45 0.4755
Z − 0.40 0.4545 − 0.4284
= Z = 0.4277
0.45 − 0.40 0.4755 − 0.4284
 x 
  = 0.4277 x = 2(0.4277) Dt = (0.8554) (3.6 × 10 4 s )(4.3 × 10 −11 m 2 / s )
 2 Dt 
= 1.06 ×10 −3 m = 1.06mm 138
Solved Nitrogen from a gaseous phase is to be diffused into pure iron at 675ºC. If
Example - 4 the surface concentration is maintained at 0.2 wt% N, what will be the
concentration 2 mm from the surface after 25 h? The diffusion coefficient
for nitrogen in iron at 675ºC is 1.9× 10-11 m2/s .
This problem asks us to compute the nitrogen concentration Cx at the 2 mm position after a 25
h diffusion time, when diffusion is non steady-state.
C x − Co C x − 0  x 
= = 1 − erf  
Cs − Co 0.2 − 0  2 Dt 
 2 × 10 −3
m 
= 1 − erf   = 1 − erf (0.765)
 2 (1.9 × 10 −11 m 2 / s )(25h)(3600 s / h) 
 
Using tabulation of error function values and linear interpretation Z Erf (Z)
0.750 0.7112
0.765 − 0.750 y − 0.7112 y = erf (0.765) = 0.7205
=
0.800 − 0.750 0.7421 − 0.7112 0.765 y
Cx − 0 0.800 0.7421
= 1.0 − 0.7205
0.2 − 0
C x = 0.056wt % N
139
Atomic mechanism of diffusion
Till now, we discussed the diffusion process without going to the atomic level. By
measuring concentration profile, one can measure the diffusion coefficients.
However, to get further insights on the diffusion process, we need to understand the atomic
mechanism.
It is almost impossible to track the jump of any particular atom. However, based on the
experimental results, we can use logical arguments to explain the process in the atomic
level.
As we have seen before that all elements will have some impurities. C, O, H etc. are
present in most of the metals in interstitial voids. So, here diffusion occurs by interstitial
diffusion mechanism.
Similarly vacancies are also always present. So substitutional diffusion is possible because
of presence of vacancies.
Let us first consider random jump of atoms that is diffusion without the presence of any
driving force.
140
Activation energy for diffusion
In a crystal, consider two adjacent interatomic planes separated by a distance α and
perpendicular to the diffusion direction x. Let there be n1 moles of the diffusing specie per
unit area in plane 1 and n2 in plane 2, with n1 > n2 .
If ν’ is the frequency with which atoms jump from one plane to a neighboring plane (the
jump may be in either the forward direction or the backward direction)
1
J1→2 = n1ν '
2
1
J 2→1 = n2ν '
2
1
J1→2( net ) = ν ' (n1 − n2 ) Equation → 1
2
n n
C1 = 1 , n1 = C1α and C2 = 2 , n2 = C2α (put in Eq.1)
α α
1
J net = ν ' (C1 − C2 )α Equation → 2
2
dc C2 − C1 dc dc
= ⇒ C2 − C1 = α (or) C1 − C2 = −α (put in Eq. 2)
dx α dx dx
1 dc 1  dc 
J net = − ν 'α 2 ⇒ J net = ν 'α 2  −  → Equation 3
2 dx 2  dx 
141
 ∆Gm   ∆G f 
ν ' = Zν exp −  exp −  → Equation 5
 RT   RT 
1   ∆Gm + ∆G f 
D = α 2 Zν exp −   142
6   RT 
1 2
α Z = a02
6
  ∆Gm + ∆G f 
D = a ν exp − 
2
0  → Equation 6
  RT 
 ∆S m + ∆S f   ∆H m + ∆H f 
D = a ν exp
2
0  exp −   → Equation 7
 R    RT 
 Q 
D = D0 exp −  → Equation 8
 RT 
 ∆S m + ∆S f 
D0 = a 0ν exp
2

 R 
143
The probability that the adjacent site will be vacant is almost unity. Then the following
expression for D as a function of temperature results:
 − ∆Gm 
D = a02ν exp 
 RT 
 − ∆Gm   −Q   Q 
D = νa02 exp  = νa0 exp
2
 = D0 exp− 
 RT   RT   RT 
Now Q = ∆Gm = ∆H m − T∆S m activation energy for migration
 ∆H m  where  ∆S 
. D = D0 exp −  D0 = νa02 exp − m  pre exponential factor
 RT   R 
∆Hm activation enthalpy for migration
Activation energy can be determined from the diffusion

coefficients calculated at different temperatures.
 Q  Q
D = D0 exp −  ⇒ ln D = ln D0 −
 RT  RT
So if we plot lnD vs. 1/T, we can determine the activation energy
for diffusion, Q.
144
145
Factors affecting Diffusion
Ease of a diffusion process is characterized by the parameter D, diffusivity. The value
of diffusivity for a particular system depends on many factors as many mechanisms
could be operative.
Diffusing species
If the diffusing species is able to occupy interstitial sites, then it can easily diffuse
through the parent matrix. On the other hand if the size of substitutional species is
almost equal to that of parent atomic size, substitutional diffusion would be easier.
Thus size of diffusing species will have great influence on diffusivity of the system.
Lattice structure
Diffusion is faster in open lattices or in open directions than in closed directions.
Presence of defects
As mentioned in earlier section, defects like dislocations, grain boundaries act as

short-circuit paths for diffusing species, where the activation energy is diffusion is
less. Thus the presence of defects enhances the diffusivity of diffusing species. 146
Factors affecting Diffusion
Temperature
Temperature has a most profound influence on the diffusivity and diffusion rates.
It is known that there is a barrier to diffusion created by neighboring atoms those
need to move to let the diffusing atom pass. Thus, atomic vibrations created by
temperature assist diffusion.
Empirical analysis of the system resulted in an Arrhenius type of relationship
between diffusivity and temperature.
 Q 
− 
D = D0 e  kT 
Where D0 is a pre-exponential constant, Q is the activation energy for diffusion, R

is gas constant (Boltzmann’s constant) and T is absolute temperature.
From the above equation it can be inferred that large activation energy means
relatively small diffusion coefficient. It can also be observed that there exists a
linear proportional relation between (lnD) and (1/T). Thus by plotting and
considering the intercepts, values of Q and D0 can be found experimentally (see in
next slide for clear understanding). 147
Diffusion paths with lesser resistance
Experimentally determined activation energies for diffusion…!
Qsurface < Qgrain boundary < Qpipe < Qlattice

Lower activation energy automatically implies higher diffusivity
Core of dislocation lines offer paths of lower resistance PIPE DIFFUSION
Diffusivity for a given path along with the available cross-section for the path will
determine the diffusion rate for that path
Comparison of Diffusivity for
Log (D) →
self-diffusion of Ag → Single Polycrystal
crystal vs. polycrystal
Single
Qgrain boundary = 110 kJ /mole crystal
QLattice = 192 kJ /mole 1/T →

← Increasing Temperature
148
Solved Using the following diffusion data, compute the value of D for the
Example - 5 diffusion of magnesium in aluminum at 400ºC.
Do(Mg in Al) =1.2 × 10-4 m2/s Qd=131 KJ/mol
This problem asks us to compute the magnitude of D for the diffusion of Mg in Al at

400ºC (673K).
−4  131,000 J / mol 
D = (1.2 × 10 m / s ) exp −
2

 (8.31J / mol − k )( 673k ) 
= 8.1×10 −15 m 2 / s
149
Solved At what temperature will the difusion coefficient for the diffusion of
Example - 6 zinc in copper have a value of 2.6 × 10-16 m2/s
Do =2.4 × 10-5 m2/s Qd=189 KJ/mol
We are asked to calculate the temperature at which the diffusion coefficient for the
diffusion of Zn in Cu has a value of 2.6 × 10-16 m2/s. Solving for T from below
equation
Qd
T =−
R (ln Do − ln D )
By using the given data we can get
 189,000 J / mol 
T = − −16 −5 
 (8.31J / mol − k )[ln(2.6 × 10 m 2
/ s ) − ln( 2.4 × 10 m 2
/ s ) 
= 901K = 6280 C
150
Solved The diffusion coefficients for nickel in iron are given at two
Example - 7 temperatures:
At 1473K 2.2 × 10-15 m2/s
At 1673K 4.8 × 10-14 m2/s
a) Determine the values of Do and the activation energy Qd.
b) What is the magnitude of D at 1300ºC (1573K)
Qd
T =− From this equation we can compute two simultaneous
R (ln Do − ln D ) equations they are
ln D1 − ln D2
Q 1 Q 1 Qd = − R
ln D1 = ln D0 − d { } ln D2 = ln D0 − d { } 1 1
−
R T1 R T2 T1 T2
[ln(2.2 × 10 −15 ) − ln(4.8 × 10 −14 )

Qd = −(8.314 J / mol − K )
1 1
−
1473 K 1673 K
= 315,700 J / mol
151
 Qd 
Now we can solve Do from this equation  − 
D0 = D1 e  RT1 
−15  315,700 J / mol 

= (2.2 ×10 2
m / s ) exp 
 (8.31J / mol − k )(1473k ) 
= 3.5 ×10 −4 m 2 / s
(b) Using these values of Do and Qd, D at 1573K is just
−4  315,700 J / mol 
D = (3.5 ×10 m / s ) exp −2

 (8.31J / mol − k )(1573k ) 
= 1.1× 10 −14 m 2 / s
152
Concept of random walk
We related the jump frequency without going into details of the possibility of next jump
after making a successful jump.
However, we find a very strange fact when we try to relate diffusion coefficient with the
jump distance.
The diffusion coefficient of carbon in γ- iron with FCC structure at 1100°C is in the order of
10-10 m2/s.
The jump distance can approximately be considered as Δx≈ 10-10 m
From the relation derived between diffusion coefficient and the jump frequency, we can
write Γ≈ 1010 /s
This means, atoms change their position in the order of 1010 times per second!
This number sounds very high, however, if we consider Debye frequency, it sounds
reasonable.
That means atoms make successful jump one out of 103 attempts only, since vibration
frequency or Debye frequency is roughly in the order of 1013/s.
Now suppose for the sake of explanation, we consider that atom can jump on a straight line
and always goes forward.
Then after 1 hour, jump distance will be 10-10 x1010x 60x60 = 3.6 kms! The same after 10
hrs will be 36 kms!
153
However, previously we have seen that the typical diffusion length is in micron.
This indicates that jump of atom is not linear and possibly many times jumps back to
previous position to have resultant jump distance zero.
Jump is random, as we are considering the jump of atom without any driving force, the
jump of atoms will be random.
This must be the reason to find much smaller diffusion length in practice.
This is the reason to say that when there is no driving force atoms go through a random
walk.
So we need to relate this random walk with the actual jump distance and the diffusion
coefficient of atoms.
This sounds very difficult, but can be done following simple approach, without making too
many assumptions.
154
For the sake of explanation, let us consider first atomic arrangement in two dimension.
Host atoms are shown as blue balls and the interstitial atom as small red ball.
As mentioned previously, we are considering very small concentration of interstitial atoms
(one in few thousands voids) and we may assume that no other interstitial atom is present in
the close vicinity.
Let us consider that the interstitial atom reaches to the point Q from P after certain number
of jumps.
Now question is after how many jumps one atom will reach to that point.
One can easily count that there are few paths (shown by red arrows) through which it can
reach to Q after 14 jumps.
However, actually, chance is very less that atoms will follow this route.
It can follow a very torturous long route, as shown by green arrows, to arrive at Q.
For a number of random jumps, n, the mean distance x, after time ‘t’ is given by
x = λ n ⇒ λ νt
Hence the diffusion distance is proportional to √t

Since the jump of atoms is random in nature, there can be huge difference in the number of
jumps that different atoms will make to reach to a particular distance.
155
Now let us go back to our discussion on diffusion of carbon in iron.
We have calculated that the average jump of carbon atoms in g-iron at 1100°C is 1010 /s.
If we consider that each jump distance is about 10-10 m, then each carbon atom travels total
distance of 3.6 kilometers in 1 hr .
However, because of random nature of jump, on average, atoms will travel the effective
distance of
x = 1010 × 3600 × 10 −10 m = 6 × 10 6 × 10 −10 m = 0.6mm = 600 µm
Note that this is the order of diffusion length we actually see in carburized steel.
156
Summary
Diffusion Faster for Diffusion Slower for

Open crystal structures Close packed structures
Lower melting temperature Higher melting temperature
materials. materials
Smaller diffusing atoms Larger diffusing atoms
Cations Anions
Materials with secondary bonding Materials with covalent bonding
Lower density materials Higher density materials
157
Questions?
1. What is the number of atoms in a cubic meter of copper? The gram-atomic weight of copper
is 63.54 gm/mole and the atomic volume of copper is 7.09 cm3 per gram-atom. Next,
compute the number of copper atoms per m3 given that the lattice constant, a, of copper is
0.36153 nm and that there are 4 atoms per unit cell in a face-centered cubic crystal.
2. A diffusion couple, made by welding a thin one centimeter square slab of pure metal A to
similar slab of pure metal B, was given a diffusion anneal at an elevated temperature and
then cooled to room temperature. On chemically analyzing successive layers of the
specimen, cut parallel to the weld interface, it was observed that, at one position, over a
distance of 5000 nm, the atom fraction of metal A, NA, changed from 0.30 to 0.35. Assume
that the number of atoms per m3 of bothe pure metals is 9 × 1028. First determine the
concentration gradient dna/dx. Then if the diffusion coefficient, at the point in question and
annealing temperature, was 2×10-14 m2/s, determine the number of A atoms per second that
would pass through this cross-section at the annealing temperature.
3. On the assumption that the self-diffusion coefficient of a simple cubic metal whose lattice
constant , a, equals 0.300 nm is given by the equation
D = 10 −4 e −200,000 / RT , m 2 / s
Determine the value of the diffusion coefficient at 1200 K and use this to determine the
mean time of stay, τ, of an atom at a lattice site.
158
Questions?
4. Do you expect any difference in room temperature self diffusion coefficients of Al just
quenched from 6000C to room temperature and the one slowly cooled to room temperature?
Explain.
5. The diffusivity of gallium in silicon is 8×10-17 m2/s at 11000C and 1×10-14 m2/s at 13000C.
Determine D0 and Qd for diffusion of gallium in silicon and calculate diffusivity at 12000C.
6. Using the data given below, make a plot of log D versus 1/T, and estimate, by eye, the best
straight line through the points.
Calculate ΔH and D0 for this line.
Calculate ΔH and D0 using a least squares procedure and assuming all error to be in the
values of D. Plot least squares line on the graph.
D (m2/s) 10-12 10-13 10-14 10-15

T(K) 1350 1100 950 800
7. Concentration of copper in an aluminium slab decreases linearly from 0.4 at % Cu at the
surface to 0.2 at% Cu at 1 mm below the surface. Calculate the flux of copper atoms across
a plane 0.5 mm below the surface at 5000C. Lattice parameter of Al is 0.405 nm.
8. Explain why activation energy for the grain boundary diffusion is lower than the activation
energy for the lattice diffusion.
159
Questions?
9. Derive Fick’s second law from the mass conservation with the help of Fick’s first law.
10. Take derivative of the thin film solution and replace in the Fick’s second law to show the
correctness of the relation.
11. A small amount (M) of component i was deposited in the form of an infinitesimally thin
layer onto another element j. The assembly was then annealed at a constant temperature (T)
in vacuum so that i diffused into j. We need to predict the concentration (Ci) as a function of
depth (x) into the material and time (t). Answer the following question for this diffusion
problem.
a) Write the diffusion equation for Ci
b) Describe the initial condition
c) Describe the boundary condition
d) Describe the constraint on the total mass (M) of the component i.
e) Verify that each of the conditions in (a), (b), (c) and (d) satisfy the following equation
for Ci. (D is the constant diffusion coefficient at T).
M  − x2 
Ci ( x, t ) = exp 
Π dt  4 Dt 
For answering part(e), you can use the following information
Z
2
∫ exp(−η 2 )dη = erf ( z ), where erf (z) is ca; ; ed error function z
Π 0
erf(0) = 0, erf(∞) = 1 160
Questions?
12. A binary solid-solid diffusion couple is assembled with two alloys having initial
concentrations of 80 atom% and 20 atom% for component i and annealed at a temperature T
for 1 day. Assuming a constant binary interdiffusion coefficient . of 10-8 cm2/s and constant
molar volume of 8 cm3/mole, calculate the position of the plane having concentration of i to
be 4.8 cm3/mole.
13. A thick steel part with initial carbon content of 0.18 wt% was exposed to a carburizing
atmosphere at 820°C with constant surface concentration of carbon at 0.8 wt%. a) If the
case depth of the carburized steel is taken as the depth at which carbon concentration drops
to 0.4 wt%, determine the time required for achieving a case depth of 0.8mm. Use the
following data for diffusion of carbon in austenite. What will be the case depth if
carburizing time is doubled?
Activation Energy (Q) (KJ/mol) Frequency Factor (D0) m2/s
136 1×10-5
14. Pure aluminium was quenched from 5000C to 250C to retain all the vacancy concentration.
What should be the self diffusion coefficient for Al at 250C immediately after this
quenching? The enthalpy of migration and vacancy formation for Al are respectively 0.68
eV and 0.72eV. Frequency factor for Al self diffusion is 0.047 cm2/s. Assume no divacancy
concentration.
15. Briefly explain the concept of a driving force. And what is the driving force for steady-state
diffusion? 161
Questions?
16. A steel of eutectoid composition was found to have the lamellar spacing of pearlite to be 2l.
The average thickness of each cementite layer is 2h. This steel was heated to 850°C for
austenitization, which proceeds by diffusion of carbon atoms from cementite into austenite.
Assuming that the transformation of ferrite to austenite does not take much time, derive the
equation for concentration profile of carbon developed in a pearlite colony during the
austenitization. You can assume square wave type initial profile in pearlite with zero%
carbon in ferrite before the dissolution starts. Assume no discontinuity of concentration at
the interface.
Hint i) No flow conditions exists i.e. the concentration gradients are zero at the boundaries
and at the middle of the profile i.e. at x = 0, x = l, and x = -l.
Hint ii) If a periodic function f(x) is defined between –l and +l, it can be expresses as a
fourier series as follows:
162
Questions?
17. A thick steel slab with 0.2 wt% carbon concentration is exposed to a carburizing atmosphere
at 900°C. Carbon concentration at the surface of the slab is kept constant at 1.3 wt%. Case
depth is considered as the depth from the surface at which carbon concentration drops to 0.6
wt%. (See the error function tables and the interdiffusion data for C in steel given at the
end).
a) Calculate the time required for achieving a case depth of 0.5mm
b) What should be the temperature of carburizing if the same case depth is required to be
achieved in exactly half the time
c) At 900°C, a particular plane of composition C’ was observed to be at a depth of 0.7mm
at the end of carburizing cycle. What should be the position of the plane C’ after
double the carburizing time at 900°C?
18. A sheet of steel 2.5 mm thick has nitrogen atmospheres on both sides at 9000C and is
permitted to achieve a steady-state diffusion condition. The diffusion coefficient for
nitrogen in steel at this temperature is 1.2×10-10m2/s, and the diffusion flux is found to be
1.0×10-7kg/m2-s. Also, it is known that the concentration of nitrogen in the steel at the high-
pressure surface is 2 kg/m3. How far into the sheet from this highpressure side will the
concentration be 0.5 kg/m3? Assume a linear concentration profile.
19. If iron is kept at 1200˚K in a carburizing atmosphere for 8hrs to obtain a carbon
concentration of 0.75 at a depth of 0.5mm. Find the time it would take to reach same carbon
concentration at depth of 7.5mm at 1250˚K. (Given D0 = 0.2x10-4 m2/s & Q =
163
143kJ/mole/˚K)
Questions?
20. The concentration of carbon on the surface of iron is maintained at 1.00% at 1175˚K for
2hours. Estimate the depth at which % C would be 0.5%. Use the diffusivity values given
D0 = 0.2x10-4 m2/s & Q = 143kJ/mole/˚K. Assume initial carbon content of iron to be
negligible.
164
Avala Lava Kumar*
Introduction
Let us start understanding phase transformations using the example of the solidification of a
pure metal.
Three states of matter are distinguishable: gas, liquid, and solid
In the gaseous state the metal atoms occupy a great deal of space because of their rapid
motion. The atoms move independently and are usually widely separated so that the
attractive forces between atoms are negligible. The arrangement of atoms in a gas is one of
complete disorder.
At some lower temperature, the kinetic energy of the atoms has decreased so that the
attractive forces become large enough to bring most of the atoms together in a liquid. And
there is a continual interchange of atoms between the vapor and liquid across the liquid
surface.
The attractive forces between atoms in a liquid may be demonstrated by the application of
pressure. A gas may be easily compressed into a smaller volume, but it takes a high pressure
to compress a liquid. There is, however, still enough free space in the liquid to allow the
atoms to move about irregularly.
As the temperature is decreased, the motions are less vigorous and the attractive forces pull
the atoms closer together until the liquid solidifies. Most materials contract upon
solidification, indicating a closer packing of atoms in the solid state.
The atoms in the solid are not stationary but are vibrating around fixed points, giving rise to
the orderly arrangement of crystal structures. 166
Mechanism of Crystallization
Crystallization is the transition from the liquid to the solid state and occurs in two stages:
Nuclei formation
Crystal Growth
Although the atoms in the liquid state do not have any definite arrangement, it is possible
that some atoms at any given instant are in positions exactly corresponding to the space
lattice they assume when solidified.
These chance aggregates or groups are not permanent but continually break up and reform
at other points.
The higher the temperature, the greater the kinetic energy of the atoms and the shorter the
life of the group. When the temperature, of the liquid is decreased, the atom movement
decreases, lengthening the life of the group, and more groups will be present at the same
time.
Atoms in a material have both kinetic and potential energy. Kinetic energy is related to the
speed at which the atoms move and is strictly a function of temperature. The higher the
temperature, the more active are the atoms and the greater is their kinetic energy. Potential
energy, on the other hand, is related to the distance between atoms. The greater the
average distance between atoms, the greater is their potential energy.
167
Mechanism of Crystallization
Metals
Thermodynamic ↑ ∆Hfusion High → (10-15) kJ / mole
1
∆G * ∝
∆H 2fusion
Crystallization favoured by
Kinetic ↓ ∆Hd ≈∝ Log [Viscosity (η)] Low → (1-10) Poise
Enthalpy of activation for

diffusion across the interface
Difficult to amorphize metals
Very fast cooling rates ~106 K/s are used for the amorphization of alloys
→ splat cooling, melt-spinning.
168
Solidification (or) Freezing
When metal is poured into the mould,

Pouring Temperature the temperature will be as high as its
inversion temperature.
Temperature →
Superheat
T/t = cooling rate Equilibrium
Freezing
Temperature
Thermal Arrest
Local Solidification time
Total Solidification time
Time →
It gets cooled when poured into the
mould and molten metal in the liquid The solidified metal in the mould (
form will solidify. This time is called called casting) gets cooled in the mould
local solidification time to the temperature of the surroundings
169
In dealing with phase transformations, we are often
concerned with the difference in free energy Solid stable Liquid stable
between two phases at temperatures away from the
equilibrium temperature.
∆Gv
For example if a liquid metal is under cooled by
∆T below Tm before it solidifies, solidification will Solid (GS)
be accompanied by a decrease in free energy ∆G
G →
(J/mol) as shown in figure.
This free energy decreases provides the driving ∆G → −ve
force for solidification. The magnitude of this
change can be obtained as follows. Liquid (GL)
The free energies of the liquid and solid at a ∆T
temperature T are given by Undercooling ∆G → +ve
G = H − TS
L L L
Tm
G S = H S − TS S
Therefore, at a temperature T ∆G = ∆H − T∆S →①
Where ∆H = H L − H S and ∆S = S L − S S 170

At the equilibrium melting temperature Tm the free energies of solid and liquid are equal,
i.e., ∆G = 0. Consequently
∆G = ∆H − Tm ∆S = 0
∆H L
And therefore at Tm ∆S = = →②
Tm Tm
This is known as the entropy of fusion. It is observed experimentally that the entropy of
fusion is a constant ≈ R (8.3 J/mol . K) for most metals (Richard’s rule).
L
For small undercoolings (∆T) the difference in the specific heats of the liquid and solid ( C p
- C pS ) can be ignored.
L
Combining equations 1 and 2 thus gives ∆G ≅ L − T
Tm
i.e., for small ∆T L∆T
∆G ≅
Tm This is called Turnbull’s approximation
171
Solidification of pure metal : Supercooling
In a pure metal at its freezing point where both the liquid and solid states are at the same
temperature. The kinetic energy of the atoms in the liquid and the solid must be the same,
but there is a significant difference in potential energy.
The atoms in the solid are much closer together, so that solidification occurs with a release
of energy. This difference in potential energy between the liquid and solid states is known
as the latent heat of fusion.
However, energy is required to establish a surface Cooling Curve for a pure metal
between the liquid and solid. In pure materials at the
freezing point insufficient energy is released by the
Temperature →
heat of fusion to create a stable boundary, and some
under cooling is always necessary to form stable
nuclei.
Subsequent release of the heat of fusion will raise
the temperature to the freezing point. The amount of
undercooling required may be reduced by the Under cooling
presence of solid impurities which reduce the

amount of surface energy required.
Time → 172
Solidification of pure metal
When the temperature of the liquid metal has dropped sufficiently below its freezing point,
stable aggregates or nuclei appear spontaneously at various points in the liquid. These
nuclei, which have now solidified, act as centers for further crystallization.
As cooling continues, more atoms tend to freeze, and they may attach themselves to
already existing nuclei or form new nuclei of their own. Each nucleus grows by the
attraction of atoms from the liquid into its space lattice.
Crystal growth continues in three dimensions, the atoms attaching themselves in certain
preferred directions, usually along the axes of the crystal this gives rise to a characteristic
treelike structure which is called a dendrite.
Since each nucleus is formed by chance, the crystal axes are pointed at random and the
dendrites growing from them will grow in different directions in each crystal. Finally, as
the amount of liquid decreases, the gaps between the arms of the dendrite will be filled and
the growth of the dendrite will be mutually obstructed by that of its neighbors. This leads
to a very irregular external shape.
The crystals found in all commercial metals are commonly called grains because of this
variation in external shape. The area along which crystals meet, known as the grain
boundary, is a region of mismatch.
173
1 2
Liquid to Solid Phase Transformation : Solidification
Growth
of crystal
Solid
3 ↑t
Two crystal going to join
to from grain boundary
Grain boundary
5 6
Growth of nucleated crystal
Solidification Complete 174

“For sufficient Undercooling”
Solidification of pure metal
This mismatch leads to a noncrystalline (amorphous) structure at the grain boundary with
the atoms irregularly spaced.
Since the last liquid to solidify is generally along the grain boundaries, there tends to be
higher concentration of impurity atoms in that area. Figure (previous page) shows
schematically the process of crystallization from nuclei to the final grains.
Due to chilling action of mold wall, a thin skin of solid metal is formed at the wall surface
immediately after pouring.
Grain structure in a casting of a pure metal, showing randomly oriented grains of small size
near the mold wall, and large columnar grains oriented toward the center of the casting.
175
Solidification of Alloys
Most alloys freeze over a temperature range

Phase diagram for a Cu-Ni alloy system and
cooling curve for different alloy systems.
Characteristic grain structure in an alloy
casting, showing segregation of alloying
components in center of casting.
1455 ºC Liquid Solution Pouring Temperature

Liquidus Liquid cooling
Freezing begins
L+S
Temperature
Freezing completed
Solidus
1083 ºC Solid cooling
Total Solidification
Solid solution time
Ni % Cu → Cu Time →
176
Cast Structure
Chill Zone: A region of small, randomly

oriented grains that forms at the surface
of a casting as a result of heterogeneous
nucleation.
Columnar Zone: A region of elongated
grains having a preferred orientation
that forms as a result of competitive
growth during the solidification of a
casting. Nucleation begins Chill zone forms
Equiaxed Zone: A region of randomly
oriented grains in the center of a casting
produced as a result of widespread
nucleation.
Figure: Development of the ingot structure

of a casting during solidification: preferred growth produces additional nucleation creates
the columnar zone 177
the equiaxed zone
Remember….!
We have seen phase transformations using the example of the solidification of a pure metal.
There is no change in composition involved as we are considering a pure metal. If we solidify an alloy
this will involve long range diffusion.
When a volume of material (V) transforms three energies have to be considered :
reduction in G (assume we are working at constant T & P),
increase in γ (interface free-energy),
increase in strain energy.
In a liquid to solid phase transformation the strain energy term can be neglected as the liquid melt can
flow to accommodate the volume change (assume we are working at constant T & P).
The process can start only below the melting point of the liquid (as only below the melting point the
GLiquid < GSolid). I.e. we need to Undercool the system. As we shall note, under suitable conditions (e.g.
container-less solidification in zero gravity conditions), melts can be undercooled to a large extent
without solidification taking place.
New interface created Bulk Gibbs free energy ↓
Energies involved Interfacial energy ↑
Volume of transforming material Strain energy ↑ Solid-solid

178
transformation
Nucleation
Solidification = Nucleation + Growth

Nucleation: localized formation of a distinct thermodynamic phase.
Nucleation an occur in a gas, liquid or solid phase. Some examples of phases that may
form via nucleation include:
In gas-creation of liquid droplets in saturated vapor
In liquid-formation of gaseous bubbles crystals (e.g., ice formation from water) or
glassy regions.
In solid-Nucleation of crystalline, amorphous and even vacancy clusters in solid
materials. Such solid state nucleation is important, for example, to the semi
conductor industry.
Most nucleation processes are physical, rather than chemical.
There are two types of nucleation : homogeneous and heterogeneous. This distinction
between them made according to the site at which nucleating events occur. For the
homogeneous type, nuclei of the new phase form uniformly throughout the parent phase,
where as for the heterogeneous type, nuclei form preferentially at structural
inhomogeneities such as container surfaces insoluble impurities grain boundaries,
dislocations and so on 179
Nucleation
It occurs spontaneously and

Homogenous randomly, but it requires
superheating or supercooling of
the medium.
Nucleation
Liquid → solid
Heterogenous walls of container, inclusions
Solid → solid
inclusions, grain boundaries,
dislocations, stacking faults
The probability of nucleation occurring at point in the parent phase is same

throughout the parent phase
In heterogeneous nucleation there are some preferred sites in the parent phase
where nucleation can occur
180
Homogeneous Nucleation
Let us consider L→S transformation taking place by homogenous nucleation. Let the
system be undercooled to a fixed temperature ∆T. Let us consider the formation of a
spherical crystal of radius ‘r’ from the melt. We can neglect the strain energy contribution.
Let the change in ‘G’ during the process be ∆G. This is equal to the decrease in bulk free
energy + the increase in surface free energy. This can be computed for a spherical nucleus
as below.
Neglected in L → S
Free energy change on nucleation = transformations
Reduction in bulk free energy + increase in surface energy + increase in strain energy
ΔG = (Volume).(∆G ) + (Surface).(γ )
∆Gv = f (∆T ) 4 3
ΔG =  πr .(∆Gv ) + 4πr 2 .(γ ) ( )
f(r) →
r3 3 
r2
Note that below a value of ‘1’ the lower power of ‘r’
dominates; while above ‘1’ the higher power of ‘r’ dominates.
In the above equation these powers are weighed with other
‘factors/parameters’, but the essential logic remains.
r → 1 181
4 
( )
ΔG =  πr 3 .(∆Gv ) + 4πr 2 .(γ )
3 
In the above equation, the r3 term is +ve and the r2 term is -ve. Such kinds of equations are often
encountered in materials science, where one term is opposing the process and the other is
supporting it. Example of such processes are crack growth (where surface energy opposes the
process and the strain energy stored in the material supports crack growth).
In the current case it is the higher power is supporting the phase transformation. Since the higher
power dominates above ‘1’, the function will go through a maximum as in fig. below. This
implies the ∆G function will go through a maximum. I.e. if the process just even starts it will lead
to an increase in ∆G!
On the other hand the function with -ve contribution from the lower power (to ∆G) will go
through a minimum (fig. below) and such a process will take place down-hill in G and stop.
182
As we have noted previously ∆G vs r plot will go through a maximum (implying that as a small
crystal forms ‘G’ will increase and hence it will tend to dissolve). The maximum of ∆G vs r plot is
obtained by, by setting d∆G/dr = 0. The maximum value of ∆G corresponds to a value of ‘r’ called
the critical radius (denoted by superscript *).
If by some ‘accident’ (technically a ‘statistical random fluctuation’) a crystal (of ‘preferred’ crystal
structure) size > r* (called supercritical nuclei) forms then it can grow down-hill in ‘G’. Crystals
smaller than r* (called embryos) will tend to shrink to reduce ‘G’. The critical value of ∆G at r* is
called ∆G*. Reduction in G (below the liquid state) is obtained only after r0 is obtained (which can
be obtained by setting ∆G = 0).
Trivial
r1* = 0 d∆G
=0
d∆G dr
=0 2γ
dr r2* = − As ∆Gv is −ve, r*is +ve
∆Gv
∆G = 0
2γ 16 γ 3
r =−
*
∆G = π*
∆Gv 3 ∆Gv2
∆G →
r* r0
3γ Embryos Supercritical nuclei
∆G = 0 r0 = −
∆Gv 183
r →
What is the effect of undercooling (∆T) on r* and ∆G*?
We have noted that ∆GV is a function of undercooling (∆T). At larger undercooling ∆GV
increases and hence r* and ∆G* decrease. This is evident from the equations for r* and ∆G*
as below
At Tm , ∆GV is zero and r* is infinity! → That the melting point is not the same as the
freezing point!! This energy (G) barrier to nucleation is called the ‘nucleation barrier’.
∆Gv = f (∆T )
G*
Tm
Decreasing ∆G
The bulk free energy reduction is a function of undercooling
2γ 16 γ 3
r =−
*
∆G = π
*
∆Gv 3 ∆Gv2
r →
Using the Turnbull approximation
(linearizing the G-T curve close to Tm),
Decreasing r* we can get the value of ∆G interms of the
∆G →
enthalpy of solidification.
16 3 Tm2
∆G = πγ
∗
3 ∆T 2 ∆H 2 184
How are atoms assembled to form a nucleus of r* →“Statistical Random Fluctuation”
To cause nucleation (or even to form an embryo) atoms of the liquid (which are randomly moving
about) have to come together in a order, which resembles the crystalline order, at a given instant of
time. Typically, this crystalline order is very different from the order (local order), which exists in the
liquid.
This ‘coming together’ is a random process, which is statistical in nature →i.e. the liquid is exploring
‘locally’ many different possible configurations and randomly (by chance), in some location in the
liquid, this order may resemble the preferred crystalline order.
Since this process is random (& statistical) in nature, the probability that a larger sized crystalline order
is assembled is lower than that to assemble a smaller sized ‘crystal’.
Hence, at smaller undercoolings (where the value of r* is large) the chance of the formation of a
supercritical nucleus is smaller and so is the probability of solidification (as at least one nucleus is
needed → which can grow to cause solidification). At larger undercoolings, where r* value is relatively
smaller, the chance of solidification is higher.
185
What is meant by the ‘Nucleation Barrier’ →an alternate perspective
Here we try to understand: “What exactly is meant by the nucleation barrier?”.

It is sometime difficult to fathom out as to the surface energy can make freezing of a small
‘embryo’ energetically ‘infeasible’ (as we have already noted that unless the crystallite size
is > r0 the energy of the system is higher). Agreed that for the surface the energy lowering is
not as much as that for the bulk*, but even the surface (with some ‘unsaturated bonds’) is
expected to have a lower energy than the liquid state (where the crystal is energetically
favored). I.e. the specific concern being: “can state-1 in figure below be above the zero
level (now considered for the liquid state)?” →“Is the surface so bad that it even negates the
effect of the bulk lowering?”
We will approach this mystery from a different angle - by first asking the question: “what is
meant by melting point?” & “what is meant by undercooling?”.
186
Melting point, undercooling, freezing point
The plot below shows melting point of Au nanoparticles, plotted as a function of the particle
radius. It is to be noted that the melting point of nanoparticles decreases below the ‘bulk melting
point’ (a 5nm particle melts more than 100 0C below TmBulk). This is due to surface effects
(surface is expected to have a lower melting point than bulk!?*) →actually, the current
understanding is that the whole nanoparticle melts simultaneously (not surface layer by layer).
Let us continue to use the example of Au.

Suppose we are below TmBulk (1337K=1064 0C,
i.e. system is undercooled w.r.t the bulk melting
point) at T1 (=1300K →∆T = 37K) and suppose
a small crystal of r2 = 5nm forms in the liquid.
Now the melting point of this crystal is ~1200K
→this crystal will ‘melt-away’. Now we have to
assemble a crystal of size of about 15nm (= r1)
for it ‘not to melt’. This needless to say is much
less probable (and it is better to undercool even
further so that the value of r* decreases). Thus
the mystery of ‘nucleation barrier’ vanishes and
we can ‘think of’ melting point ≡ freezing point
(for a given size of particle)! Other materials like Pb, Cu, Bi, Si show similar trend lines
Tm is in heating for the bulk material and in cooling if we take into account the size dependence
187
of melting point everything ‘sort-of’ falls into place .
Homogenous Nucleation Rate
The process of nucleation (of a crystal from a liquid melt, below TmBulk) we have described
so far is a dynamic one. Various atomic configurations are being explored in the liquid
state - some of which resemble the stable crystalline order. Some of these ‘crystallites’ are
of a critical size r*∆T for a given undercooling (∆T). These crystallites can grow to
transform the melt to a solid→by becoming supercritical. Crystallites smaller than r*
(embryos) tend to ‘dissolve’.
As the whole process is dynamic, we need to describe the process in terms of ‘rate’ →the
nucleation rate [dN/dt ≡ number of nucleation events/time].
Also, true nucleation is the rate at which crystallites become supercritical. To find the
nucleation rate we have to find the number of critical sized crystallites (N*) and multiply it
by the frequency/rate at which they become supercritical.
If the total number of particles (which can act like potential nucleation sites - in
homogenous nucleation for now) is Nt , then the number of critical sized particles given by
an Arrhenius type function with a activation barrier of ∆G*.
 ∆G * 
− 
 kT 
N = Nt e
*  
188
The number of potential atoms, which can jump to make the critical nucleus supercritical
are the atoms which are ‘adjacent’ to the liquid→let this number be s*.
If the lattice vibration frequency is ν and the activation barrier for an atom facing the
nucleus (i.e. atom belonging to s*) to jump into the nucleus (to make in supercritical) is
∆Hd, the frequency with which nuclei become supercritical due atomic jumps into the
nucleus is given by: '  ∆H 
−d

ν = S νe
*  kT 
No. of critical sized x Frequency with which they

Rate of nucleation = particles become supercritical
 ∆G *   ∆H d 
dN −
 kT 
 − 
I= N * = Nt e   ν ' = s *ν e  kT 
dt
No. of particles/volume in L ν → lattice vibration frequency (~1013 /s)
s* atoms of the liquid facing the nucleus
Critical sized nucleus
Jump taking particle to supercriticality

→ nucleated (enthalpy of activation = ∆Hd)
189
Critical sized nucleus
Homogenous Nucleation Rate
The nucleation rate (I = dN/dt) can be written as a product of the two terms as in the equation below.
How does the plot of this function look with temperature?
At Tm , ∆G* is ∞→I = 0 (as expected if there is no undercooling there is no nucleation).
At T = 0K again I = 0
This implies that the function should reach a maximum between T = Tm and T = 0.
A schematic plot of I(T) (or I(∆T)) is given in the figure below. An important point to note is that the
nucleation rate is not a monotonic function of undercooling.
Tm T = Tm → ∆G* = ∞ → I = 0
Increasing ∆T
T
∆G* ↑ ⇒ I ↓
 ∆G* +∆Hd 
−  Note: ∆G* is a function of T

I = Nt s* ν e  kT 
T↑ ⇒I ↑
T (K) →
0 T=0→I=0 190
I →
Heterogenous Nucleation
Heterogeneous nucleation can be considered as a surface catalyzed or assisted nucleation
process. The extent of how a surface can catalyze or facilitate the nucleation depends on the
contact angle of the nucleus with respect to the substrate. The smaller the angle (or the
stronger the wetting of the surface), the lower the free energy change, and the lower the
nucleation barrier will be.
Critical radius of the nucleus (r*) for a heterogeneous nucleation is the same as that for a
homogeneous nucleation, whereas the critical volume of the nucleus (like the droplet for
liquid nucleated from gas/vapor phase) is usually smaller for heterogeneous nucleation than
for homogeneous nucleation, due to the surface wetting (spreading).
Heterogeneous nucleation occurs much more often than homogeneous nucleation.
Heterogeneous nucleation applies to the phase transformation between any two phases of
gas, liquid, or solid, typically for example, condensation of gas/vapor, solidification from
liquid, bubble formation from liquid, etc.
In the solidification of a liquid this could be the mold walls. For solid state transformation
suitable nucleation sites are: non-equilibrium defects such as excess vacancies, dislocations,
grain boundaries, stacking faults, inclusions and surfaces.
One way to visualize the ease of heterogeneous nucleation → heterogeneous nucleation at a
defect will lead to destruction/modification of the defect (make it less “‘defective’”). This
will lead to some free energy ∆Gd being released →thus reducing the activation barrier.
191
Heterogenous Nucleation : Derivation
There are three different surface energies of interest:
γ αβ = Surface of gas/liquid
γ βδ = Surface of liquid/solid
h
γ αδ = Surface of gas/solid
∆G = V. ∆G V + A.γ x r
Surface energy = Aαβ γ αβ + Aβδ γ βδ + Aαδ γ αδ θ
Aαβ = 2πrh ⇒ 2πr (r − r cos θ ) ⇒ 2πr 2 (1 − cos θ )
r −h
Cosθ =
r
h = r − r cos θ
Aβα = πr 2 ⇒ π (r sin θ ) ⇒ πr 2 sin 2 θ

2
Aαδ = Total area - A βα ⇒ Ao − πr 2 sin 2 θ (∴ Ao = Total area)

Surface energy = γ αβ 2πr 2 (1 − cos θ ) + γ βδ πr 2 sin 2 θ + γ αδ Ao − πr 2 sin 2 θ
Neglect the γ αδ Ao term, then above equation becomes
Surface energy = γ αβ 2πr 2 (1 − cos θ ) + γ βδ πr 2 sin 2 θ − πr 2 sin 2 θ ⇒ Eq.1
192
θ
Volume of spherical cap = Area × Height = πr ε ) sin ε dε
∫ −
3 2
(1 cos
θ 0
Volume = ∫ π (rsinε ) × dH
2
θ
0 = πr 3
∫ (sin ε − sin ε cos 2
ε )dε
we know H = r − r cos ε ⇒ dH = r sin ε dε 0
θ θ
θ
= ∫ πr 2sin 2ε × r sin ε dε = πr 3 ∫ sin ε − πr 3 ∫ sin ε cos 2 εdε
0 0
0
θ dt − dt
let us take t = cosε ⇒ = − sin ε ⇒ dε =
= ∫ πr sin ε × sin ε dε
3 2
dε sin ε
0 θ
θ dt
= πr (− cos ε ) 0 − πr ∫ sin ε .t 2 . −
3 3
0
sin ε
θ
θ
= πr (− cos ε ) 0 + πr
3 3
∫ dt
t 2
0
θ
3 t 
3
= πr (− cos θ + 1) + πr  
3
 3 0
193
πr 3
= πr (1 − cos θ ) +
3
(cos3 θ − cos 3 0) α γαβ
3
πr 3 β
= πr (1 − cos θ ) +
3
(cos3 θ − 1) θ
3
3  3 − 3 cos θ + cos θ − 1 
3
γαδ γβδ
= πr  
 3 
πr 3 δ
= [2 − 3 cos θ + cos θ ] 3
γ αδ − γ βδ
3 Cos θ =
γ αβ
Surface tension force balance
Cosθγ αβ = γ αδ − γ βδ
γ αδ − γ βδ
Cosθ = ⇒ Eq.2
γ αβ
πr 3
∆G = [2 − 3 cos θ + cos3 θ ]∆GV + γ αβ 2πr 2 (1 − cos θ ) + πr 2 sin 2 θ (γ βδ − γ αδ )
3
put γ βδ − γ αδ = −γ αβ cos θ
194
πr 3
= [ 2 − 3 cos θ + cos 3 θ ]∆GV + γ αβ 2πr 2 (1 − cos θ ) − γ αβ πr 2 sin 2 θ cos θ
3
4πr 3 [ 2 − 3 cos θ + cos 3 θ ]
= ∆GV + πr 2γ αβ ( 2 − 2 cos θ − sin 2 cos θ )
3 4
4πr 3 [ 2 − 3 cos θ + cos 3 θ ]
= ∆GV + πr 2γ αβ ( 2 − 2 cos θ − (1 − cos 2 θ ) cos θ )
3 4
4πr 3 [ 2 − 3 cos θ + cos 3 θ ] [ 2 − 3 cos θ + cos 3 θ ]
= ∆GV + 4πr γ αβ
2
3 4 4
[ 2 − 3 cos θ + cos 3 θ ] 4πr 3
= [ ∆GV + 4πr 2γ αβ ]
4 3
4πr 3 [ 2 − 3 cos θ + cos 3 θ ]
=[ ∆GV + 4πr γ αβ ] f (θ ) → where f (θ ) =
2
3 4
∆G Hetero = ∆G Homo . f (θ )
∆G *Hetero = ∆G Hetero ( r*) = ∆G Homo ( r*). f (θ ) = ∆G *Homo f (θ )
Nucleation barrier can be significantly lower for heterogeneous nucleation due to wetting
angle affecting the shape of the nucles
195
Heterogenous Nucleation
Using the procedure as before (for the case of the homogenous nucleation) we can find r*
for heterogeneous nucleation. Using the surface tension balance equation we can write the
formulae for r*and ∆G* using a single interfacial energy γαβ (and contact angle θ).
Further we can write down ∆G*hetero in terms of ∆G*homo and contact angle θ.
2γ αβ 4 γ αβ
( )
3
d∆ G =−
=0 ∆G = π − θ + θ
* * 3
r 2 3Cos Cos
∆GV 3 ∆GV2
hetero hetero
dr
∆Ghetero
( )
*
∆G*
hetero
1 *
= Ghom
4
o (
2 − 3Cos θ + Cos 3
θ ) *
Ghom o
=
1
4
2 − 3Cosθ + Cos 3θ Just a function of θ →
The contact angle
θ = 00 → f(θ) = 0 Complete wetting

∆Ghetero
( )
*
1
*
= 2 − 3Cosθ + Cos 3θ = f (θ ) θ = 900 → f(θ) = 1/2 Partialwetting
Ghom o 4
θ = 1800 → f(θ) = 1
196
The plot of ∆G*hetero / ∆G*homo is shown in the next page No wetting
Plot of ∆G*hetero /∆G*homo is shown below. This brings out the benefit of heterogeneous
nucleation vs homogenous nucleation
If the β phase nucleus (lens shaped) completely wets the substrate/inclusion (δ-phase) (i.e. θ
= 0º) →then ∆G*hetero = 0 →there is no barrier to nucleation.
On the other extreme if β-phase does not we the substrate (i.e. θ = 180º) → ∆G*hetero =
∆G*homo → there is no benefit of the substrate. In reality the wetting angle θ is somewhere
between 0º-180º.
Hence, we have to chose a heterogeneous nucleating agent with a minimum ‘θ’ value
1
∆G*hetero (180o) = ∆G*homo
no benefit
0.75
∆G*hetero (0o) = 0
∆G*hetero / ∆G*homo →
no barrier to nucleation ∆G*hetero (90o) = ∆G*homo/2
0.5
γ αδ − γ βδ
Cosθ =
γ αβ
0.25
Complete wetting Partial wetting No wetting

0
0 30 60 90 120 150 180
θ (degrees) → 197
Choice of heterogeneous nucleating agent
α γαβ
Heterogeneous nucleation has many practical applications.
During the solidification of a melt if only a few nuclei β
θ
form and these nuclei grow, we will have a coarse grained
γαδ γβδ
material (which will have a lower strength as compared to
a fine grained material- due to Hall-Petch effect).
Hence, nucleating agents are added to the melt (e.g. Ti for δ γ αδ − γ βδ
Cosθ =
Al alloys, Zr for Mg alloys) for grain refinement. γ αβ
How to get a small value of θ? (so that ‘easy’ heterogeneous nucleation).

Choosing a nucleating agent with a low value of γβδ (low energy βδ interface)
(Actually the value of (γαδ − γβδ) will determine the effectiveness of the heterogeneous
nucleating agent → high γαδ or low γβδ)
How to get a low value of γβδ
We can get a low value of γβδ
o Crystal structure of β and δ are similar
o lattice parameters are as close as possible
Examples of such choices are
Seeding rain-bearing clouds → AgI or NaCl → nucleation of ice crystals
Ni (FCC, a = 3.52 Å) is used a heterogeneous nucleating agent in the production of artificial
diamonds (FCC, a = 3.57 Å) from graphite 198
Heterogeneous nucleation rate
The rate of heterogeneous nucleation can be expressed in a form similar to that of
Homogeneous nucleation rate. In addition to the difference in the ΔG* term and the pre-
exponential term will include only the number of preferred nucleation sites, which is many
orders of magnitude smaller than the number of atoms per unit voulme used in the
homogeneous case
In order of magnitude , the number of nucleation sites for various situations is typically as
folllows:
If Ni is the number of sites per unit volume having a nucleation barrier ΔGi *, then we can
write expression for heterogeneous nucleation rate will be
Nhomogeneous ~1029 m-3
 ∆G* +∆Gd  Ngrain boundary ~1023 m-3
− 

Ihet = Ni si ν e ~1017 m-3
*  RT  Ngrain edge
Ngrain corner ~1011 m-3
Ndislocation ~1013 m-3
Usually, I het >I homo, because ΔG* het < ΔG* homo dominates
the result through the pre-exponential factor may be much smaller for heterogeneous
nucleation as compared to the homogeneous case.
199
Why does heterogeneous nucleation rate dominate?
To understand the above questions, let us write the nucleation rate for both cases as a pre-
exponential term and an exponential term. The pre-exponential term is a function of the
number of nucleation sites.
However, the term that dominates is the exponential term and due to a lower ∆G* the
heterogeneous nucleation rate is typically higher.
 ∆Ghomo
*   ∆Ghetero
* 
− 
 − 

I homo ∝ I homo
0
e  kT 
I hetero ∝ I 0
hetero e  kT 
= f(number of nucleation sites) = f(number of nucleation sites)

~ 1042 ~ 1026
BUT
the exponential term dominates
Ihetero > Ihomo 200

Growth of a Pure solid
At transformation temperature the probability of jump of atom from α → β (across the
interface) is same as the reverse jump
Growth proceeds below the transformation temperature, wherein the activation barrier for
the reverse jump is higher.
There are basically two different types of solid/liquid interface: an atomically rough or
diffuse interface associated with metallic systems, and an atomically flat or sharply defined
interface often associated with nonmetals.
Because of the differences in atomic
structure these two types of interface
migrate in quite different ways . Rough
interface migrate by a continuous growth ∆Hd
processes while flat interface migrate by ∆Hd – vatom ∆Gv
lateral growth (Spiral growth and
surface nucleation) process involving α phase
ledges.
Trasformation Nucleation Growth β phase

= of till
+
α→β β phase α is
exhausted 201
Summary
Compared to the heterogeneous nucleation (which starts at nucleation sites on surfaces)
homogeneous nucleation occurs with much more difficulty in the interior of a uniform
substance. The creation of a nucleus implies the formation of an interface at the boundaries
of a new phase.
Liquids cooled below the maximum heterogeneous nucleation temperature (melting
temperature) but which are above the homogeneous nucleation temperature. (pure substance
freezing temperature) are cooled super cooled.
An example of supercooling: pure water freezes at -42ºC rather than at its freezing
temperature 0ºC.
Nucleation - The physical process by which a new phase is produced in a material.

Critical radius (r*) - The minimum size that must be formed by atoms clustering together in the
liquid before the solid particle is stable and begins to grow.
Undercooling - The temperature to which the liquid metal must cool below the equilibrium
freezing temperature before nucleation occurs.
Homogeneous nucleation - Formation of a critically sized solid from the liquid by the clustering
together of a large number of atoms at a high undercooling (without an external interface).
Heterogeneous nucleation - Formation of a critically sized solid from the liquid on an impurity
surface.
202
Questions?
1. (a). Suppose that 0.100 kg of liquid copper is supercooled 250 K, where it is allowed to
nucleate and solidify adiabaticall (no heat is lost to the surroundings). Calculate how much
copper will solidify until the temperature recalesces to its melting point of 1356 K. The
specific heats of solid and liquid copper in the temperature range of interest are,
respectively:
C P ( s ) = 22.64 + 5.86 ×10 −3 T , J / mol.K
C P (l ) = 31.4 J / mol.K
Where T is the temperature in degrees Kelvin. The heat of fusion of copper is 13.20 kJ/mol
and its molar weight is 0.0635 kg/mol.
(b). How much supercooling would be necessary in order to solidify the entire sample
adiabatically?
2. Derive the expression for Homogeneous nucleation (ΔG* and r*) and it’s rate, and
heterogeneous nucleation (ΔG* and r*) and its rate (neglect the strain energy).
3. Briefly explain Heterogeneous nucleation on grain boundaries, Grain edges grain corners,
inclusions and dislocations.
4. Derive the expressions for Homogeneous and heterogenerous nucleation by considering the
strain energy term.
5. Derive an expression for the critical size and activation energy for a cubic nucleus. Under
what conditions is such a nucleus most likely to occur?
203
Questions?
6. Assume for the solidification of nickel that nucleation is homogeneous, and the number of
stable nuclei is 106 nuclei per cubic meter. Calculate the critical radius and the number of
stable nuclei that exist at the following degrees of supercooling: 200 K and 300 K. and
What is significant about the magnitudes of these critical radii and the numbers of stable
nuclei? [rNi – 0.255 J/m2, ΔHf = -2.53×109 J/m3, Super cooling value for Ni = 319oC]
7. Calculate the homogeneous nucleation rate in liquid copper at under coolings of 180, 200,
and 220K, using the given data: L =1.88×109 J m-3, Tm = 1356K, γsL = 0.177 J m-2, f0 = 1011
s-1, C0 = 6×1028 atoms m-3, k = 1.38×1
8. Explain the concept constitutional Supercooling. When does this take place? And explain
dendritic formation
9. The surface energy of pure metal liquid , γ, is 600 dynes/ cm; The volume of an atom of this
metal in the liquid is 2.7×10-25cm3; and the free-energy difference between an atom in the
vapor and liquid , ΔGvl, is -2.37J. Under these conditions, what would be the critical radius
of a droplet, r0 in nm and the free energy of the droplet, ΔGr0, in J?
10. Write a short note on ‘Grwoth of a pure solid’ ?
11. Differentiate between Homogeneous and Heterogeneous nucleation? In which case
nucleation rate will be high? Why.
12. Explain the effect of undercooling on Nucleation rate, growth rate, r* and G*? And what is
the glass transition temperature. And how it is related to undercooling.
204
Questions?
13. The following data concern liquid magnesium at a temperature close to its boiling point of
1380 K: the surface energy of the liquid-vapor interface is 0.440 J/m2; the density of the
liquid phase is 1.50 ×103 kg/m3; and the atomic weight is 0.02432 kg/mol.
a) First determine vl, the volume of the liquid per atom. To do this, use the density and the
atomic weight.
b) Now compute the number of embryos in the vapor at a temperature just above the
boiling point, that contain 10 magnesium atoms.
14. The accompanying diagram is for a hypothetical embryo of silver growing against an
arbitrary mold wall. With the aid of this diagram,
a) Compute the angle of contact,θ of the embryo with the mold wall.
b) Determine the magnitude of the factor that may be used to convert the homogeneous
free energy needed to obtain a nucleus into that of the corresponding heterogeneous
free energy.
15. An alloy having solute concentration C is held at a temperature T within solid – liquid range
and the liquid is filtered out. If this process is repeated by heating & filtering out the liquid
will it ultimately give pure A? Is this a practical method of purification? 205
Questions?
16. Estimate entropy change during solidification of the following elements and comment on the
nature of the interface between solidifying crystal and liquid. The latent heat and melting point
are given with brackets. (a) Al [10.67kJ/mole, 660C] (b) Si [46.44kJ/mole, 1414⁰C]
17. What is partition coefficient? Derive Scheil equation for solidification of binary alloys. State the
assumptions made during its derivation. What is the composition of the last solid that forms
during solidification of a terminal solid solution of a binary eutectic system?
18. Estimate the temperature gradient that is to be maintained within solid aluminum so that the
planar solidification front moves into liquid aluminum maintained at its melting point at a
velocity of 0.001m/s. Given thermal conductivity of aluminum = 225 W/mK, latent heat of fusion
= 398 KJ/kg and density = 2700kg/m3 .
19. In cobalt, a coherent interface forms during HCP to FCC transformation. The lattice parameter of
the FCC phase is 3.56 Ao. The distance between nearest neighbours along a close packed
direction in the basal plane of the HCP phase is 2.507Ao. show that the misfit to be
accommodated by coherency strains is small.
20. Suppose that an iron specimen containing 0.09 atomic percent carbon is equilibriated at 720°C
(993 K) and then rapidly quenched to 300°C (573 K). Determine the length of the time needed for
one side of a plate shaped carbide precipitate to grow out by 103 nm.
p) How wide a layer of the matrix next to a plate would have its carbon concentration lowered
from 0.09 percent carbon to that corresponding to nαe to form a layer of cementite 103 nm
thick?
q) How long would it take to increase one side of the plate by 10 nm? 206
Avala Lava Kumar*
207
History
Figure 1
Iron – carbon (Fe-C) diagram and
first phase diagram of any alloy to be
established, done by Roberts –Austen
in 1898 after whom austenite came to Figure 2
be named. (Figure 1)
The Iron Carbon diagram, adjusted by Bakhuis-Roozeboom in 1900. Notice line A-a and
the temperature range between 1000 and 1100ºC, where carbide formation was suppose to
take place as the result of a chemical reaction between graphite and austenite (at that time
called martensite! (Figure 2) 208
Allotropic Transformation in Iron
Iron is an allotropic metal, which means that it can exist in more than one type of lattice
structure depending upon temperature. A cooling curve for pure iron is shown below:
Liquid
1539ºC
δ (delta) Fe - B.C.C
1400ºC
Temperature ºC →
γ (gamma) Fe - F.C.C
Non Magnetic
910ºC
Non Magnetic α (alpha) Fe - B.C.C
768ºC
α (alpha) Fe - B.C.C
Magnetic
Time →
209
Can be other allotropic structures are possible….?
Effect of pressure on allotropy of Iron
This line slopes upward as at constant
T if we increase the P the gas will
liquefy as liquid has lower volume
(similarly the reader should draw
horizontal lines to understand the effect
of pressure on the stability of various
phases- and rationalize the same).
Phase fields of non-close packed
structures shrink under higher pressure
Phase fields of close packed structures
expand under higher pressure
These lines slope downward as: Under

higher pressure the phase with higher
packing fraction (lower volume) is
preferred
The face centered tetragonal (FCT) iron is coherently deposited iron grown as thin film on a
{100} plane of copper substrate. Growing Trigonal iron on mis-fiting {111} surface of a face
210
centered cubic copper substrate.
Iron - Cementite phase diagram
The Fe-C (or more precisely the Fe-Fe3C) diagram is an important one. Cementite is a
metastable phase and ‘strictly speaking’ should not be included in a phase diagram. But
the decomposition rate of cementite is small and hence can be thought of as ‘stable
enough’ to be included in a phase diagram. Hence, we typically consider the Fe-Fe3C part
of the Fe-C phase diagram.
In the phase diagram, temperature is plotted against composition. Any point on the
diagram therefore represents a definite composition and temperature. The phase diagram
indicates the phases present and the phase changes that occur during heating and cooling.
The relative amounts of the phases that exist at any temperature can also be estimated with
the help of lever rule.
A portion of the Fe-C diagram – the part from pure Fe to 6.67 wt.% carbon (corresponding
to cementite, Fe3C) – is technologically very relevant.
Cementite is not a equilibrium phase and would tend to decompose into Fe and graphite.
This reaction is sluggish and for practical purpose (at the microstructural level) cementite
can be considered to be part of the phase diagram. Cementite forms as it nucleates readily
as compared to graphite.
Compositions upto 2.1%C are called steels and beyond 2.1% are called cast irons. In
reality the calssification should be based on ‘castability’ and not just on carbon content.
Heat treatments can be done to alter the properties of the steel by modifying the
microstructure → we will learn about this in coming chapters. 211
Fe-Fe3C metastable phase diagram
0.1 0.16 0.51
Peritectic Liquid (L)
Eutectic
L+δ→γ 1493ºC
L → γ + Fe3C
δ (Ferrite)
L+ γ L + Fe3C
Temperature →
δ+ γ γ (Austenite) 1147ºC
Hypo Eutectic Hyper Eutectic
910ºC 2.11
[γγ(Austenite) + Fe3C(Cementite)] = Ledeburite
723ºC
α (Ferrite) Eutectoid 4.3
Eutectoid
Hyper
Hypo
0.025 %C
Eutectoid α (Ferrite) + Fe3C (Cementite)] = Pearlite

[α
γ → α + Fe3C Steels Cast Irons
0.008 0.8 6.7
Fe % Carbon → Fe3C
212
Solubility of carbon in Fe = f (structure, temperature)
Where is carbon located in iron lattice ?
Octahedral
sites
Octahedral
sites
6 faces sharing with two sides (6/2)=3 One interstitial site in center plus
12 edges sharing with four sides (12/4)=3 12 edges sharing with four sides (12/4)=3
Total sites is (3+3), 6 per unit cell Total sites is (1+3), 4 per unit cell
Every one Fe atom we have 3 interstitial sites Every one Fe atom we have 1 interstitial site
213
Why concentration of carbon in α-Fe with BCC structure
is less than γ-Fe with FCC structure?
FIRST LET US CONSIDER FCC STRUCTURE (γ-Fe)
Packing factor of FCC lattice 0.74
This indicates the presence of voids. Let us discuss it more elaborately.
In a FCC crystal there are mainly two types of voids:
Tetrahedral: coordination number (CN) is 4. That means the void is surrounded by 4 atoms.
Octahedral: CN is 6. That means the void is surrounded by 6 atoms.
There are 8 tetrahedral voids in a unit cell. That means 2 voids per atom.
There are 4 octahedral voids in a unit cell. That means 1 void per atom.
However, impurities prefer to occupy octahedral voids.
Because the ratio of the radius of the tetrahedral void to atom is 0.225 and the same for the
octahedral void is 0.414.
The ratio of the radius of the carbon atom (size is 77 pm) to Fe (when it has FCC crystal) is
0.596.
So when a carbon atom occupies any void, lattice will be distorted to increase the enthalpy.
Distortion will be less if it occupies the octahedral voids.
Although it increases the enthalpy, carbon atoms will be present up to a certain extent
because of the gain in entropy, as explained previously, which decreases the free energy.
214
FCC Void (Oct) Void (Tet)
o
Size of Fe atom
CCP crystal r Fe
FCC = 1.292 A FeFCC C
o
Size of the OV
Fe
xFCC (oct ) = 0.534 A Relative sizes of voids w.r.t atoms
o
Size of Carbon atom
r = 0.77 A
C
Note the difference in size of the atoms
BCC
o
Size of Fe atom
BCC crystal r Fe
BCC = 1.258 A Void (Oct) Void (Tet)
o
(d .tet ) = 0.364 A
Fe FeBCC C
Size of the TV
x BCC
o Relative sizes of voids w.r.t atoms
Size of the OV x Fe
BCC (d .oct ) = 0.195 A
Remember
FCC Size of the largest atom which can fit into the tetrahedral is 0.225 and octahedral void is 0.414
BCC Size of the largest atom which can fit into the d.tetrahedral is 0.29 and d.octahedral void is 0.154
215
NOW LET US CONSIDER THE BCC UNIT CELL (α - Fe)
Packing factor of BCC lattice 0.68. So total void in a BCC unit cell is higher than FCC cell.
However, there are 12 (6 per atom) tetrahedral and 6 (3 per atom) octahedral voids present.
This number is higher than the number of voids present in a FCC unit cell.
Unlike voids in FCC lattice, in a BCC lattice the voids are distorted. That means if an atom sits
in a void, it will not touch all the host atoms.
The ratio of the radius of tetrahedral void to atom is 0.29 and the radius of octahedral void to
atom is 0.155.
The ratio of the radius of the C atom (size is 77 pm) to Fe (when it has BCC crystal) is 0.612.
So it is expected that in a BCC unit cell, impurities should prefer tetrahedral voids.
However, although the octahedral void size is small, planar radius which has 4 atoms on the
same plane is 79.6 pm, which is larger that the C atom size. That means it needs to distort only
other two atoms.
On the other hand if C sits in the tetrahedral void it has to distort all four atoms. So in α – Fe
with BCC unit cell C occupies the octahedral voids
Now the octahedral void size in g-Fe (FCC) is higher than a-Fe (BCC). So naturally the
distortion in a BCC cell will be higher and the activation energy for impurities to occupy a
void in a BCC cell also will be higher.
216
This is the reason that we find much lower solubility limit of C in a-Fe.
Why carbon preferentially sits in the apparently smaller octahedral void in BCC ?
Ignoring the atom sitting at B and assuming the interstitial atom touches the atom at A
o
2a
OA = r + x A = r Fe
BCC = 1.258 A
2
2 6r BCC : 3a = 4r
r + xA =
3
xA  2 6 
=  − 1  = 0.6329
r  3 
o o o
OX = x A = 0.796 A OY = xB = 0.195 A x Fe
BCC ( d .tet ) = 0.364 A 217
Characteristics of phases appeared in
Fe-Fe3C phase diagram
218
Ferrite (α)
It is an interstitial solid solution of a small amount of carbon dissolved in α iron. The
maximum solubility is 0.025%C at 723ºC and it dissolves only 0.008%C at room
temperature. It is the softest structure that appears on the diagram.
Ferrite is ferromagnetic at low temperatures but loses its magnetic properties with the rise
of temperatures with major loss at curies temperatures, 768ºC and above this temperature, it
becomes non magnetic (paramagnetic).
The crystal structure of ferrite (α) is B.C.C
Tensile strength – 245 Mpa, Yield strength 118 Mpa
Elongation – 40-50% in 2 in.
Hardness - 95 VPN
α(Ferrite) contains
B.C.C structure α (Ferrite) – 90X 219
Cementite (Fe3C)
Cementite or iron carbide, chemical formula Fe3C, contains 6.67%C by weight and it is a
metastable phase.
It is typically hard and brittle interstitial compound of low tensile strength (35 Mpa) but
high compressive strength and high hardness ~800VPN.
It is the hardest structure that appears on the diagram.
It has a complex orthorhombic crystal structure with 12 iron atoms and 4 carbon atoms per
unitcell.
It is slightly ferromagnetic up to 210ºC and paramagnetic above it. Melting point around
1227ºC.
Cementite network
Carbon
(4 atoms)
Iron
(12 atoms)
a≠b≠c
220
Orthorhombic α = β = γ = 90º
Pearlite (α+Fe3C)
Pearlite is the eutectoid mixture containing 0.80 %C
and is formed at 723ºC on very slow cooling.
It is very fine platelike or lamellar mixture of ferrite
and cementite. The fine fingerprint mixture called
pearlite is shown in below figure.
Eutectoid
The weight % of these two phases are thus in ratio
8:1 Hypo Hyper
Tensile strength – 120,000 psi or 825 Mpa Eutectoid Eutectoid
Elongation – 20 percent in 2 in.
Hardness – HRC 20, HRB 95-100, or BHN 250-300 α + Fe3C
Fine Pearlite Coarse

Remember…!
Pearlite is a not a
phase but combination
of two phases (ferrite
+ cementite)
221
Austenite (γ)
It is an interstitial solid solution of a small amount of carbon dissolved in γ iron and has
FCC crystal structure. The maximum solubility is 2.1%C at 1147ºC.
Austenite is soft, ductile tough and malleable (FCC structure) and non magnetic
(paramagnetic).
Steels are commonly rolled and forged above about 1100ºC when they are in austenite state
due to its high ductility and malleability, which is also due to its FCC structure.
Tensile strength – 150,000 psi or 1035 Mpa
Elongation – 10% in 2 in.
Hardness - 395 VPN and Toughness is high.
Fe
γ(Austenite) – 325X
222
Ledeburite (γ+Fe3C)
Ledeburite is the eutectic mixture of austenite and

cementite. It contains 4.3%C and is formed at 1147ºC
Structure of ledeburite contains small islands of
austenite are dispersed in the carbide phase.
Not stable at room temperature
6.7
The pearlite is resolved in some regions

where the sectioning plane makes a
glancing angle to the lamellae. The
ledeburite eutectic is highlighted by the
arrows. At high temperatures this is a
mixture of austenite and cementite formed
from liquid. The austenite subsequently
decomposes to pearlite.
223
Ferrite (δ)
Interstitial solid solution of carbon in iron of body centered
cubic crystal structure. (δ iron ) of higher lattice parameter
(2.89Å) having solubility limit of 0.09 wt% at 1495°C with
respect to austenite. The stability of the phase ranges between
1394-1539°C.
It is a high temperature phase and is a high temperature
manifestation of α ferrite.
This is not stable at room temperature in plain carbon steel. However it can be present at
room temperature in alloy steel specially duplex stainless steel.
δ-ferrite in dendrite form

in as-cast Fe-0.4C-2Mn-
0.5Si-2 Al0.5Cu
224
Invariant Reactions in Fe-Fe3C
Phase Diagram
225
Peritectic Reaction
The invariant peritectic reaction in Fe-Fe3C diagram is given by
L δ-ferrite Cool
Austenite (γ)
+
0.51%C 0.1 %C 1495ºC
0.17%C
0.16 − 0.1
Thus Liquid, wt% is L= × 100 = 14.63%
0.51 − 0.1
0.51 − 0.16
Thus δ ferrite, wt % is δ ( ferrite) = × 100 = 85.37%
0.51 − 0.1
Fe-0.16%C steel is a peritectic steel because only this steel undergoes above reaction
completely.
Peritectic reaction is of some importance during freezing of steels (carbon from 0.1 to
0.51% particularly under fast cooling conditions, when micro segregation may result,
otherwise no commercial heat treatment is done in this region.
Unfortunately these temperatures are attained during heating of steels for forging or rolling
etc., then severe overheating and burning results in steels turning them to scrap form.
226
Eutectic Reaction
The invariant Eutectic reaction in Fe-Fe3C diagram is given by
Liquid (L) Cool Austenite (γ) Cementite
+
0.17%C 1147ºC 2.11%C 6.67 %C
6.67 − 4.3
Thus Austenite, wt% is γ = ×100 = 51.97%
6.67 − 2.11
4.3 − 2.11
Thus cementite, wt % is Fe3C = × 100 = 48.03%
6.67 − 2.11
Fe-4.3%C alloy is called eutectic cast iron as it is the lowest melting point alloy, which is
single phase liquid (100%) of 4.3% carbon at the eutectic temperature, 1147°C just attained
and undergoes eutectic reaction completely at this constant eutectic temperature to give a
mixture of two different solids, namely austenite and cementite, solidifying simultaneously.
The eutectic mixture called Ledeburite.
As Fe-C alloys having more than 2.11% carbon are classed as cast irons, the Fe-C alloys
having carbon between 2.11 and 4.3% are called hypo eutectic cast irons, where as those
having carbon between 4.3% and 6.67% are called hypereutectic cast irons. Alloys of Fe
with 4.3% carbon is called eutectic cast iron. 227
Eutectoid Reaction
The invariant Eutectoid reaction in Fe-Fe3C diagram is given by
Austenite (γ) Cool Ferrite (α) Cementite
+
0.8%C 727ºC 0.02%C 6.67 %C
6.67 − 0.8
Thus Ferrite, wt% is α= × 100 = 88%
6.67 − 0.02
0.8 − 0.02
Thus cementite, wt % is Fe3C = × 100 = 12%
6.67 − 0.02
During cooling austenite of 0.8% at constant eutectoid temperature, 727°C undergoes
eutectoid transformation to form a mixture of ferrite (C%=0.02%) and cementite i.e., there
are alternate lamellae of ferrite and cementite .
This eutectoid mixture of ferrite and cementite is called PEARLITE, because of its pearly
appearance under optical microscope.
The weight % of these phases are thus 8:1. The densities are (α-7.87 gm/cm3) and (Fe3C- -
7.70 gm/cm3) are quite comparable. Thus the Volume % also approx 8:1. Thus ferrite
lamilla is 8 times thicker than cementite lamilla. as the two boundaries of cementite plate
are close together, they may not resolved separately under the microscope, instead of two
lines, it appears a single dark line. 228
Eutectoid Reaction
Phase changes that occur upon passing from the γ
region into the α+ Fe3C phase field.
Consider, for example, an alloy of eutectoid
composition (0.8%C) as it is cooled from a temperature
within the γ phase region, say 800ºC – that is,
beginning at point ‘a’ in figure and moving down
vertical xx’. Initially the alloy is composed entirely of
the austenite phase having composition 0.8 wt.% C and
then transformed to α+ Fe3C [pearlite]
The microstructure for this eutectoid steel that is slowly
cooled through eutectoid temperature consists of
alternating layers or lamellae of the two phases α and
Fe3C
The pearlite exists as grains, often termed “colonies”;
within each colony the layers are oriented in essentially
the same direction, which varies from one colony to
other.
The thick light layers are the ferrite phase, and the
cementite phase appears as thin lamellae most of which
appear dark.
229
Eutectoid Structure
Pearlite in SEM
Schematic representation of the formation of pearlite from

austenite; direction of carbon diffusion indicated by arrows
Pearlite in AFM
230
Pearlite in optical microscope 500X
Hypo Eutectoid Region
Hypo eutectoid region – 0.008 to 0.8 %C
Consider vertical line yy’ in figure, at
about 875ºC, point c, the microstructure
will consist entirely of grains of the γ
phase.
In cooling to point d, about 775ºC, which
is within the α+γ phase region, both these
phases will coexist as in the schematic
microstructure Most of the small α
particles will form along the original γ
grain boundaries.
Cooling from point d to e, just above the
eutectoid but still in the α+γ region, will
produce an increased fraction of the α
phase and a microstructure similar to that
also shown: the α particles will have
grown larger.
231
Just below the eutectoid temperature, at point f, all the γ phase that was present at
temperature e will transform pearlite. Virtually there is no change in α phase that existed at
point e in crossing the eutectoid temperature – it will normally be present as a continuous
matrix phase surrounding the isolated pearlite colonies.
Thus the ferrite phase will be present both in the pearlite and also as the phase that formed
while cooling through the α+γ phase region. The ferrite that is present in the pearlite is
called eutectoid ferrite, whereas the other, is termed proeutectoid (meaning pre- or before
eutectoid) ferrite.
Pearlite
Pro-eutectoid ferrite
232
Hyper Eutectoid Region
Hyper eutectoid region – 0.8 to 2.1 %C
Consider an alloy of composition C1 in figure
that, upon cooling, moves down the line zz’. At
point g only the γ phase will be present and the
microstructure having only gamma grains.
Upon cooling into the γ+ Fe3C phase field –
say to point h – the cementite phase will began
to form along the initial γ grain boundaries,
similar to the α phase in point d. this cementite
is called proeutectoid cementite that which
forms before the eutectoid reaction.
As the temperature is lowered through the
eutectoid to point I, all remaining austenite of
eutectoid composition is converted into
pearlite; thus the resulting microstructure
consists of pearlite and proeutectoid cementite
as microconstituents.
233
Pearlite
Cementite network 234

Concept of Primary Ferrite & Primary Cementite
When the carbon content of steels is much away from the eutectoid carbon, then distinction
can easily be made between hypoeutectoid steels and the hypereutectoid steel. Nital is the
common etching agent.
Shape of the phases
Proeutectoid ferrite appears as grains which are quite wide, polyhedral and the grain
boundaries in between neighboring ferrite grains can be seen.
The films of pro eutectoid cementite generally are much thinner, have irregular outlines and
bounded by sharp lines. These are present as network of needles or platelets. Cementite
looks much brighter and sharp because of its hardness and etching characteristics.
Major difference is cementite is present as network at the grain boundaries of pearlite (at
RT), whereas ferrite is present as grains (equiaxed polyhedral grains) with grain boundaries
in between ferrite grains (if etched properly).
Microstructu
res of (a) b
Hypereutecto
id steel (b)
Hypoeutectoi
d steel
235
a
Concept of Primary Ferrite & Primary Cementite
Relative Hardness
Cementite is very hard (~800 VPN) and ferrite is (~95 VPN). Micro hardness testing can be
done to distinguish between ferrite and cementite.
A simple scratch test can be done. Make a scratch on the polished and etched surface of the
steel and then, examine the point of the scratch where it enters the white proeutectoid phase
from the pearlite.
If the scratch widens on entering, it is the soft phase ferrite, and if it thins in white phase,
then the white phase is much harder than pearlite, i.e, it is cementite.
a b
Relative hardness of ferrite
and cementite distinguishes
them (a) Hypereutectoid
steel (b) Hypoeutectoid steel
Special Etchants
Nital etching causes cementite as well as ferrite to look white under microscope
A sodium picrate solution used either boiling or electrolytically, darkens Fe3C but not ‘α’
Another etchant based on sodium thiosulphate and ammonium nitrate gives colors ferrite
but not cementite. 236
Fe-Fe3C phase diagram (microstructural aspects)
L=liquid, Cm=cementite, LB=ledeburite, δ=delta ferrite, α=
1539˚C δ+L L alpha ferrite, α’= alpha ferrite(0.00005 wt%C) γ=austenite,
δ A5=1495˚C P=pearlite, eu=eutectic, ed=eutectoid, I=primary,
0.09 0.53 II=secondary, III=tertiary
δ+γ 0.17 1227˚C
1394˚C
L+γI 4.30 L+CmI
γ 2.11 γI+LB LB+CmI A4=1147˚C
Temperature, °C
LB’ (γeu(γII
γI’(γII+CmII)+LB’ (γ’eu(γII
+CmII)+Cmeu)
+CmII)+Cmeu)
+CmI
A B C
Ledeburite=LB(γγeu+Ceu)
910˚C D E F
A2 A3
=668/ Cm
0.77
770˚C αI+γ γII+CmII
0.0218 A1=727˚C
α αI+ (P(αed P(αed+Cmed) (P(αed+Cmed)+CmII)+ LB’ (P(αed+Cmed) LB’ (P(αed+Cmed)
Pearlite
+CmII +CmII+Cmeu) +CmII)+Cmeu)+CmI

+Cmed)
αI(α’+CmIII)+ P(α (α’ (P(αed(α’ed+CmIII)+Cmed) +CmII)+ LB’

ed ed
(P(αed(α’ed+C +Cm )+Cmed) LB’ ((P(αed(α’ed+CmIII)+Cmed) ((P(αed(α’ed+CmIII)+Cmed)
mIII)+Cmed) +CmIII +CmII)+Cmeu) +CmII)+ Cmeu)+CmI
II
Ao=210˚C
α’+CmIII
6.67
0.00005
Weight percent carbon
237
Evolution of Microstructures on equilibrium cooling
Sequence of evolution of microstructure can be described by the projected cooling on
compositions A, B, C, D, E, F.
At composition A
L → δ + L → δ → δ + γ → γ → γ + α I → α → α ' + CmIII
At composition B
L → δ + L → L + γ I → γ → α I + γ → α I + ( P(α ed + Cmed ) → α I (α ' + CmIII ) + P(α ed (α ed' + CmIII ) + Cmed )
At composition C
L → L + γ I → γ → γ II + CmII → P (α ed + Cmed ) + CmII → P (α ed (α ed' + CmIII ) + Cmed ) + CmII
At composition D
L → L + γ I → γ I + LB → γ I' (γ II + CmII ) + LB ' (γ eu' (γ II + CmII ) + Cmeu )
→ ( P(α ed + Cmed ) + CmII ) + LB ' ( P(α ed + Cmed ) + CmII + Cmeu

→ ( P(α ed (α ed' + CmIII ) + Cmed ) + CmII ) + LB ' (( P(α ed (α ed' + CmIII ) + Cmed ) + CmII ) + Cmeu )
238
Evolution of Microstructures on equilibrium cooling
Sequence of evolution of microstructure can be described by the projected cooling on
compositions A, B, C, D, E, F.
At composition E
L → L + CmI → LB (γ eu + Cmeu + CmI ) → LB ' (γ eu (γ II + CmII ) + Cmeu ) + CmI
→ LB ' ( P(α ed + Cmed ) + CmII ) + Cmeu ) + CmI
→ LB ' (( P (α ed (α ed' + CmIII ) + Cmed ) + CmII ) + Cmeu ) + CmI
At composition F
L → Fe3C
239
Application of Lever rule in
240
Solved For a 99.6 wt% Fe-0.40 wt% C at a temperature just below the
Example eutectoid, determine the following:
a) The amount of Fe3C, ferrite (α) and pearlite
b) The amount of pearlite and proeutectoid ferrite (α)
0.025 0.4 6.67

a) The amount of Fe3C and ferrite (α)
Ferrite Cementite
Percentage of Fe3C =
Percentage of Fe3C in 0.4 %C steel is 5.64 %
Percentage of Ferrite (α) in 0.4 %C steel = (100- 5.64)%
Percentage of ferrite in 0.4 %C steel = 94.36%

or
Percentage of ferrite= = 94.36%
241
b) Phase fraction of pearlite and proeutectoid ferrite (α)
0.025 0.4 0.8

Percentage of pearlite = 48 %
Ferrite Pearlite
Percentage of proeutectoid ferrite (α) in 0.4 %C steel = (100- 48)%
Percentage of proeutectoid ferrite (α) = 52 % or
242
Critical Temperature lines
A5
A4
A3
A2 Acm A0
767ºC
A1,3
210ºC
A1
In general, A0 – Subcritical temperature, A1 - lower critical temperature, A3 - upper critical

temperature, A4 – Eutectic temperature, A5 – Peritectic temperature and Acm - γ/γ+cementite
phase field boundary.
While heating we denoted as Ac1, Ac2, Ac3 etc., ‘c’ stands for chauffage (French word),
which means heating and while cooling we denoted as Ar1, Ar2, Ar3 etc., ‘r’ stands for
refroidissement, (French word) which means cooling. 243
Critical Temperature lines
The upper – and lower critical temperature lines are shown as single lines under
equilibrium conditions and are sometimes indicated as Ae3, Ae1 etc. When the critical lines
are actually determined, it is found that they do not occur at the same temperature.
The critical line on heating is always higher than the critical line on cooling. Therefore, the
upper critical line of a hypo eutectoid steel on heating would be labeled AC3 and the same
line on cooling Ar3. The rate of heating and cooling has a definite effect on the temperature
gap between these lines.
The results of thermal analysis

of a series of carbon steels with
an average heating and cooling
rate of 11ºF/min are shown in
figure.
Final word…! with infinitely slow

heating and cooling they would
probably occur at exactly the same
temperature.
440
Effect of alloying elements on Fe-Fe3C phase diagram
Based on stabilizing Austenite
Mn, Ni, Co, Cu, Zn increase the range in which γ-phase, or austenite is stable [by raising A4
and lowering A3 temperature and also tend to retard the separation of carbides.
These elements have γ-phase FCC crystal structure (or similar structure) in which these
elements are more soluble than ferrite, and that is why, in the (α+γ) two phase equilibrium,
these segregate in austenite in preference to ferrite.
Elements like carbon and nitrogen (interstitial solid solution forming elements) are also
austenite stabilizers.
Based on stabilizing Ferrite
Cr, W, Mo, V, Si, Al, Be, Nb, P, Sn, Ti, Zr increase the range of α-phase (by lowering A4 and
raising A3 temperatures).
These elements have α phase BCC crystal structure (or similar structure) and thus in (α+γ)
two phase equilibrium , these elements segregate in ferrite in preference to austenite. These
elements decrease the amount of carbon soluble in austenite, and thus tend to increase the
volume of the free carbide in the steel for a given carbide content.
Chromium is a special case of these elements as at low concentrations , chromium lowers
A3 temperature and raises A4 , but at high concentrations raises A3 temperature. Overall, the
stability of austenite is continuously decreased. 245
Effect of alloying elements on Austenite phase region Mn, Cr
Temperature →
0.35% Mn
Mn is Austenite stabilizer
Expansion of γ phase field with ↑ Mn
Outline of the γ phase field

6.5% Mn
0 0.4 0.8 1.2 1.6
C (%) →
Temperature →
Outline of the γ phase field

15% Cr 12% Cr
5% Cr 0% Cr
Cr is Ferrite stabilizer
Shrinking γ phase field with ↑ Cr
0 0.4 0.8 1.2 1.6 246

C (%) →
Effect of alloying elements on Fe-Fe3C phase diagram
Carbide forming elements
Important elements, in this class, are arranged in order of increasing affinity for carbon, and
thus the carbide forming potential of the element :
Fe → Mn → Cr → W → Mo → V → Ti → Nb → Ta → Zr
For example, vanadium is added in steel having chromium and molybdenum with
insufficient carbon, then vanadium first removes carbon from chromium carbide, the
remaining vanadium then removes carbon from molybdenum carbide and forms its own
carbide. The released Cr and Mo dissolve to form solid solution in austenite.
Graphitising elements
Si, Ni, Cu, Al are common graphitizes. Small amount of these elements in steel can
graphitise it and thus, impair the properties of steel unless elements of austenite stabilizers
are present to counteract the effect.
Neutral element
Co is the only element which neither forms carbide, nor causes graphitisation.
247
Effect of alloying elements on Fe-
Effect on Eutectoid composition
All the elements lower the eutectoid carbon content. Titanium and molybdenum are the
most effective in lowering it.
For example, a steel with 5% Cr has its eutectoid point at 0.5%C as compared to 0.8% in
carbon steels. High speed steel has eutectoid point at 0.25% carbon.
Effect on Eutectoid temperature
Elements like Ni, Mn i.e., the austenite stabilizers lower the eutectoid temperature (727°C).
Ferrite stabilizers like Cr, V, W etc. raise the eutectoid temperature.
248
Effect of alloying element on eutectoid temperature Effect of alloying element on effective carbon content
Limitations on Fe-Fe3C phase diagram
Fe-Fe3C diagram represents behavior of steels under equilibrium conditions, whereas the
actual heat treatments of steels are normally under non-equilibrium conditions.
The diagram does not indicate the character of transformation of austenite such as to
bainite, or martensite.
The diagram does not indicate the presence of metastable phases like martensite, or bainite.
It does not indicate the temperature of start of martensite Ms or bainite Bs .
It does not indicate the kinetics of the transformation of austenite to martensite, bainite or
even pearlite.
It does not indicate the possibilities of suppressing the pearlitic or bainitic transformations.
249
Classification of steels
Compositions, such as carbon (or non alloy ), low -alloy, and alloy steels
Manufacturing methods , such as converter, electric furnace, or electroslag
remelting methods
Application or main characteristic, such as structural , tool, stainless steel, or
heatresistant steels.
Finishing methods, such as hot rolling, cold rolling, casting, or controlled rolling
and controlled cooling.
Product shape, such as bar, plate, strip, tubing, or structural shape
Oxidation practice employed, such as rimmed, killed, semikilled, and capped
steels.
Microstructure, such as ferritic, pearlitic, martensitic, and austenitic
Required strength level, as specified in the American Society for Testing and
Materials (ASTM) standards.
Heat treatment, such as annealing, quenching and tempering, air cooling
(normalization), and thermo-mechanical processing.
Quality descriptors and classifications, such as forging quality and commercial
quality.
250
Classification of steels
Steels
Low Alloy High Alloy
low carbon Med carbon high carbon
<0.25wt% C 0.25-0.6wt% C 0.6-1.4wt% C
heat austenitic
Name plain HSLA plain plain tool
treatable stainless
Cr,V Cr, Ni Cr, V,
Additions none none none Cr, Ni, Mo
Ni, Mo Mo Mo, W
Example 1010 4310 1040 4340 1095 4190 304
Hardenability 0 + + ++ ++ +++ 0
TS - 0 + ++ + ++ 0
EL + + 0 - - -- ++
Uses auto bridges crank pistons wear drills high T
struc. towers shafts gears applic. saws applic.
sheet press. bolts wear dies turbines
vessels hammers applic. furnaces
blades V. corros.
resistant
increasing strength, cost, decreasing ductility 251
Avala Lava Kumar*
Time-Temperature-Transformation
(TTT) Diagrams
The temperature of transformation controls the nature of decomposed product (of austenite)
which in turn decides the resultant properties of steel.
The kinetics of austenitic transformation can be studied best at a constant temperature rather
than by continuous cooling.
The constant temperature transformation is also referred to as isothermal transformation
which is studied by the following experiment.
Davenport and Bain were the first to develop the TTT diagram of eutectoid steel. They
determined pearlite and bainite portions whereas Cohen later modified and included MS
and MF temperatures for martensite.
There are number of methods used to determine TTT diagrams. The most popular method
is salt bath techniques combined with metallography and hardness measurement with
addition of this we have other techniques like dilatometry, electrical resistivity method,
magnetic permeability, in situ diffraction techniques (X-ray, neutron), acoustic emission,
thermal measurement techniques, density measurement techniques and thermodynamic
predictions.
TTT diagrams, also called as Isothermal (temperature constant) Transformation diagrams.
For every composition of steel we should draw a different TTT diagram.
253
Determination of TTT diagram for eutectoid steel
For the determination of isothermal transformation (or) TTT diagrams, we consider molten
salt bath technique combined with metallography and hardness measurements.
In molten salt bath technique two salt baths and one water bath are used.
Salt bath I is maintained at austenising temperature (780˚C for eutectoid steel).
Salt bath II is maintained at specified temperature at which transformation is to be
determined (below Ae1), typically 700-250°C for eutectoid steel.
Bath III which is a cold water bath is maintained at room temperature.
In bath I number of samples are austenite at A1+20-40°C for eutectoid, A3+20-40°C for
hypo-eutectoid steel and ACm+20-40°C for hyper-eutectoid steels for about an hour.
Then samples are removed from bath I and put in bath II and each one is kept for different
specified period of time say t1, t2, t3, t4,…..........,tn etc.
After specified times, the samples are removed and quenched in cold water.
The microstructure of each sample is studied using metallographic techniques. The type, as
well as quantity of phases, is determined on each sample.
Transformation of austenite to ferrite-cementite mixtures occurs after a definite time (say t1)
This time during which transformation does not proceed is known as incubation period.
The magnitude of incubation period provides a qualitative idea about the relative stability
of supercooled austenite. Smaller incubation period corresponds to lesser stability of
austenite. 254
Equipments for determination of TTT Diagrams
Salt bath I Salt bath II

Austenitisation heat treatment Low-temperature for isothermal treatment
Sample and fixtures

for dilatometric measurements Dilatometer equipment 255
Simple experimental arrangement for determining

The kinetics of isothermal austenitic transformation Eutectoid section of the iron - carbon diagram
The progress of austenite transformation

A-austenite, P-pearlite, M-martensite 256
TTT curve of austenite to pearlite for 1080 steel
% of pearlite 100
T2 At T1, incubation
50% T1
period for pearlite=t2,
Pearlite finish time =t4
0
Ae1 Minimum incubation

period t0 at the nose of
T2
the TTT diagram,
T1
t1 t2 t3 t4 t5 Important points to be
t0 Nose noted:
α + Fe3C The x-axis is log scale.
‘Nose’ of the ‘C’ curve is
Temperature
in ~sec and just below

TE transformation times
may be ~day.
The starting phase

(left of the C curve) has
MS, Martensite start temperature to γ.
M50,50% Martensite
To the right of finish C
Metastable austenite +martensite curve is (γ+ Fe3C) phase
MF, Martensite finish temperature
field. This phase field
Martensite has more labels included.
257
Log time
Important points to be
800
noted: Eutectoid temperature
The x-axis is log 723
scale. ‘Nose’ of the Austenite
‘C’ curve is in ~sec Pearlite
and just below TE 600
transformation times
may be ~day. α + Fe3C
Temperature →
500 Pearlite + Bainite
The starting phase
(left of the C curve)
has to γ.
400 Bainite
To the right of
finish C curve is (γ+
300
Fe3C) phase field. Ms
This phase field has 200
more labels included.
Mf
100
Martensite
0.1 1 10 102 103 104 10 5
t (s) → 258
Possible phases in TTT diagram for eutectoid steel
Ae1
Coarse Pearlite R/C 15

Medium Pearlite R/C 30
Nose Fine Pearlite (F.P) R/C 40

50% F.P + 50% U.B
Temperature
Hardness
Upper bainite (U.B) R/C40
50% Transformation
Lower bainite R/C 60
MS, Martensite start temperature

M50,50% Martensite
Metastable austenite +martensite
MF, Martensite finish temperature
Martensite R/C 64
Log time
259
Isothermal Transformation diagram for eutectoid steel
As pointed out before one of the important utilities of the TTT diagrams comes from the
overlay of micro-constituents (microstructures) on the diagram.
Depending on the T, the (γ+ Fe3C) phase field is labeled with micro-constituents like
Pearlite, Bainite.
The time taken to 1% transformation to, say pearlite or bainite is considered as
transformation start time and for 99% transformation represents transformation finish.
We had seen that TTT diagrams are drawn by instantaneous quench to a temperature
followed by isothermal hold.
Suppose we quench below (~225°C, below the temperature marked Ms), then Austenite
transforms via a diffusionless transformation (involving shear) to a (hard) phase known as
Martensite. Below a temperature marked Mf this transformation to Martensite is complete.
Once γ is exhausted it cannot transform to (γ + Fe3C).
Hence, we have a new phase field for Martensite. The fraction of Martensite formed is not a
function of the time of hold, but the temperature to which we quench (between Ms and Mf).
Strictly speaking cooling curves (including finite quenching rates) should not be overlaid on
TTT diagrams (remember that TTT diagrams are drawn for isothermal holds!).
260
Isothermal hold at: (i) 800
T1 gives us Pearlite, (ii) Eutectoid temperature Eutectoid steel (0.8%C)
T2 gives Pearlite+Bainite, 723
(iii) T3 gives Bainite. Note
that Pearlite and Bainite
Austenite Coarse
are both α+Fe3C (but their T1 Pearlite
morphologies are 600 Fine
different).
Temperature →
500 T2 Pearlite + Bainite
To produce Martensite
we should quench at a rate
such as to avoid the nose
of the start ‘C’ curve.
400 T3 Bainite
Called the critical cooling
rate.
300
If we quench between Austenite
Ms and Mf we will get a Ms
mixture of Martensite and 200
γ (called retained
Austenite). Mf
100
Martensite
Not an isothermal 1 102 103 104
0.1 10 105
transformation t (s) → 261
In principle two curves exist for Pearlitic and Bainitic transformations

→ they are usually not resolved in plain C steel
(In alloy steels they can be distinct)
262
TTT diagram for Hypo-eutectoid steel
In hypo- (and hyper-) eutectoid steels (say composition C1) there is one more branch to the
‘C’ curve-NP (next slide: marked in red).
The part of the curve lying between T1 and TE (marked in fig : next slide) is clear, because
in this range of temperatures we expect only pro-eutectoid α to form and the final
microstructure will consist of α and γ. (E.g. if we cool to Tx and hold).
The part of the curve below TE is a bit of a

‘mystery’ (since we are instantaneously undercooling wrt A cm w.r.t
cooling to below TE, we should get a mix (formation of pro-eutectoid α)
of α+ Fe3C ⇒ what is the meaning of a
‘pro’-eutectoid phase in a TTT diagram?
(remember ‘pro-’ implies ‘pre-’)
Suppose we quench instantaneously an
hypo-eutectoid composition C1 to Tx we
should expect the formation of α+Fe3C
undercooling wrt A 1 line
(and not pro-eutectoid α first). (formation of α+ Fe3C)
The reason we see the formation of pro-
eutectoid α first is that the undercooling
C1
w.r.t to Acm is more than the undercooling
w.r.t to A1. Hence, there is a higher
propensity for the formation of pro-
eutectoid α. 263
TTT diagram for Hypo-eutectoid steel
T1
P
TE
264
TTT diagram for Hyper-eutectoid steel
Similar to the hypo-eutectoid case, hyper-eutectoid compositions C2 have a γ+Fe3C branch.
For a temperature between T2 and TE (say Tm (not melting point- just a label)) we land up
with γ+Fe3C.
For a temperature below TE (but above the nose of the ‘C’ curve) (say Tn), first we have the
formation of pro-eutectoid Fe3C followed by the formation of eutectoid γ+Fe3C.
T2
TE
C2
265
Transformation Mechanism
TTT diagrams involve nucleation and growth. Nucleation involves the formation of a
different phase from a parent phase (e.g. crystal from melt). Growth involves attachment of
atoms belonging to the matrix to the new phase (e.g. atoms ‘belonging’ to the liquid phase
attach to the crystal phase).
Nucleation we have noted is ‘uphill’ in ‘G’ process, while growth is ‘downhill’ in G.
Growth can proceed till all the ‘prescribed’ product phase forms (by consuming the parent
phase).
Transformation rate
As expected transformation rate (Tr) is a function of nucleation rate (I) and growth rate (U).
In a α→β transformation, if Xβ is the fraction of β-phase formed, then dXβ/dt is the
transformation rate.
The derivation of Tr as a function of I & U is carried using some assumptions (e.g. Johnson-
Mehl and Avarami models).
We have already seen the curve for the nucleation rate (I) as a function of the undercooling.
The growth rate (U) curve as a function of undercooling looks similar. The key difference
being that the maximum of U-∆T curve is typically above the I-∆T curve.
This fact that T(Umax) > T(Imax) give us an important ‘handle’ on the scale of the
transformed phases forming. 266
Transformation Rate
dX β
Transformation rate = f(Nucleation rate, Growth rate) T= = f (I , U )
dt
Tm Maximum of growth rate usually

U at higher temperature than
maximum of nucleation rate
∆T
Increasing ∆
I
T (K) →
0
I, U, Tr → [rate ⇒ Sec-1 ] 267
Transformation Rate
Fraction of the product (β) phase forming with time →the sigmoidal growth curve
Many processes in nature (etc.), e.g. growth of bacteria in a culture (number of bacteria with
time), marks obtained versus study time(!), etc. tend to follow a universal curve the
sigmoidal growth curve.
In the context of phase transformation, the fraction of the product phase (Xβ) forming with
time follows a sigmoidal curve (function and curve as below).
1.0 Saturation phase → decreasing growth

 π I U3 t4 
−  
 rate with time
Xβ = 1 − e  3 
0.5 Linear growth regime → constant high

Xβ →
growth rate
Incubation period → Slow growth (but

with increasing growth rate with time)
0
t → 268
From ‘Rate’ to ‘time’ : The Origin of TTT diagrams
The transformation rate curve (Tr -T plot) has hidden in it the I-T and U-T curves.
An alternate way of plotting the Transformation rate (Tr ) curve is to plot Transformation
time (Tt) [i.e. go from frequency domain to time domain]. Such a plot is called the Time-
Temperature-Transformation diagram (TTT diagram).
High rates correspond to short times and vice-versa. Zero rate implies ∞ time (no
transformation).
This Tt -T plot looks like the ‘C’ alphabet and is often called the ‘C-curve. The minimum
time part is called the nose of the curve.
Tm
Rate = f (T , t ) Tm
Small driving
force for nucleation
T (K) →
T (K) →
Tr Time for transformation
Replot
Tt
Sluggish
growth
0 Nose of the ‘C-curve’ 0

Tr (rate ⇒ sec−1) → Tt (rate ⇒ sec−1) →
269
Understanding the TTT diagram
Though we are labeling the transformation temperature Tm , it represents other
transformations, in addition to melting.
Clearly the Tt function is not monotonic in undercooling. At Tm it takes infinite time for
transformation.
Till T3 the time for transformation decreases (with undercooling) [i.e. T3 < T2 < T1 ]
→due to small driving force for nucleation.
After T3 (the minimum) the time for transformation increases [i.e. T3 < T4 < T5]→due
to sluggish growth.
The diagram is called the TTT diagram because it

plots the time required for transformation if we hold
the sample at fixed temperature (say T1) or fixed
undercooling (∆T1). The time taken at T1 is t1.
To plot these diagrams we have to isothermally hold
at various undercoolings and note the transformation
time. I.e. instantaneous quench followed by
isothermal hold.
Hence, these diagrams are also called Isothermal
Transformation Diagrams. Similar curves can be
drawn for α→β (solid state) transformation. 270
Transformation Rate
This is a phase diagram where the blue region is the Liquid (parent) phase field and purplish region is
the transformed product (crystalline solid).
Clearly the picture of TTT diagram presented before is incomplete → transformations may start at a
particular time, but will take time to be completed (i.e. between the Liquid phase field and solid phase
field there must be a two phase region L+S).
This implies that we need two ‘C’ curves → one for start of transformation and one for completion. A
practical problem in this regard is related to the issue of how to define start and finish (is start the first
nucleus which forms? Does finish correspond to 100%?) . Since practically it is difficult to find ‘β%’
and ‘100%’, we use practical measures of start and finish, which can be measured experimentally.
Typically this is done using optical metallography and a reliable ‘resolution of the technique is about
1% for start and 99% for finish.
Another obvious point: as x-axis is time any

Increasing % transformation
‘transformation paths’ have to be drawn such that
it is from left to right (i.e. in increasing time).
β
How do we define the fractions transformed?
T (K) →
99% = finish α '→α +β

α Fraction volume fraction of β at t
f →
transformed final volume of β
1% = start α →β
271
t (sec) → TTT diagram α → β phase transformation f → volume fraction of β
Overall Transformation Kinetics
The ‘C’ curve depends on various factors as listed in diagram below.
Nucleation rate
Growth rate
f(t,T) determined by Density and distribution of nucleation sites
Overlap of diffusion fields from adjacent transformed volumes
Impingement of transformed volumes
Some common assumptions used in the derivation are

constant number of nuclei
constant nucleation rate
constant growth rate.
272
Constant number of nuclei
Generally constant number of nuclei will form at the beginning of the transformation.
One assumption to simplify the derivation is to assume that the number of nucleation sites
remain constant and these form at the beginning of the transformation.
This situation may be approximately valid for example if a nucleating agent (inoculant) is
added to a melt (the number of inoculant particles remain constant).
In this case the transformation rate is a function of the number of nucleation sites (fixed)
and the growth rate (U). Growth rate is expected to decrease with time.
In Avrami model the growth rate is assumed to be constant (till impingement).
f = F (number of nucleation sites, growth rate) growth rate ↓ with

273 time
Constant growth rate
Generally in cellular transformations constant growth rate is observed.

Cellular Transformations → Constant growth rate
All of the parent Pearlitic transformation

phase is consumed by Cellular Precipitation
the product phase
Massive Transformation
α → β +γ α →β Recrystallization
Termination of transformation does not occur by a gradual reduction in the growth rate but
by the impingement of the adjacent cells growing with a constant velocity.
274
Constant nucleation rate
Another common assumption is that the nucleation rate (I) is constant.

In this case the transformation rate is a function of both the nucleation rate (fixed) and the
growth rate (U).
Growth rate decreases with time.
If we further assume that the growth rate is constant (till impingement), then we get the
Johnson-Mehl model.
f = F (number of nucleation sites, growth rate) growth rate ↓ with time

275
Derivation of f(T,t) : Avrami Model
Parent phase has a fixed number of nucleation sites Nn per unit volume (and these sites are
exhausted in a very short period of time
Growth rate (U = dr/dt) constant and isotropic (as spherical particles) till particles impinge
on one another.
At time t the particle that nucleated at t = 0 will have a radius r = Ut
Between time t = t and t = t + dt the radius increases by dr = Udt
The corresponding volume increase dV = 4πr2 dr
Without impingement, the transformed volume fraction (f) (the extended transformed
volume fraction) of particles that nucleated between t = t and t = t + dt is:
f = N n 4π r 2 ( dr ) = N n 4π [Ut ] (Udt ) = N n 4π U 3t 2 dt
2
This fraction (f) has to be corrected for impingement. The corrected transformed volume
fraction (X) is lower than f by a factor (1−X) as contribution to transformed volume fraction
comes from untransformed regions only:
dX X t =t  4π N n U3 t 3 
f = dX dX − 
⇒ = N n 4π U 3t 2 dt ⇒ ∫0 1 − X t =∫0
= N 4π U 3 2
t dt
Xβ = 1 − e
3
1− X 1− X
n  
Based on the assumptions note that the growth rate is not part of the equation →it is only
the number of nuclei.
276
Derivation of f(T,t): Johnson-Mehl Model
Parent phase completely transforms to product phase (α → β)
Homogenous Nucleation rate of β in untransformed volume is constant (I)
Growth rate (U = dr/dt) constant and isotropic (as spherical particles) till particles impinge
on one another
At time t the particle that nucleated at t = 0 will have a radius r = Ut

The particle which nucleated at t = τ will have a radius r = U(t − τ)
Number of nuclei formed between time t = τ and t = τ + dτ → Idτ
Without impingement, the transformed volume fraction (f) (called the extended transformed
volume fraction) of particles that nucleated between t = τ and t = τ + dτ is:
4 4
f = π r 3 ( Idτ ) = π [U (t − τ ) ] ( Idτ )
3
3 3
This fraction (f) has to be corrected for impingement. The corrected transformed volume
fraction (X) is lower than f by a factor (1−X) as contribution to transformed volume fraction
comes from untransformed regions only:
dX 4 3 4
= π r ( Idτ ) = π [U (t − τ )] ( Idτ )
dX 3
f = ⇒
1− X 1− X 3 3
277
Derivation of f(T,t): Johnson-Mehl Model
X τ =t For a isothermal transformation
dX 4
∫0 1 − X τ ∫=0 3 [
= π − τ ] ( Idτ )
3
U (t ) 1.0
 π I U3 t 4 
− 
 0.5
Xβ = 1 − e  3 
Xβ →
Note that Xβ is both a function of I and
U. I & U are assumed constant
0
t →
 π I U3 
  → is a constant during isothermal transformation
 3 
278
Effect of alloying elements on TTT diagram
Almost all alloying elements (except, Al, Co, Si) increases the stability of supercooled
austenite and retard both proeutectoid and the pearlitic reaction and then shift TTT curves
of start to finish to right or higher timing. This is due to
low rate of diffusion of alloying elements in austenite as they are substitutional
elements,
reduced rate of diffusion of carbon as carbide forming elements strongly hold them.
Alloyed solute reduce the rate of allotropic change, i.e. γ→α, by solute drag effect on
γ→α interface boundary.
Additionally those elements (Ni, Mn, Ru, Rh, Pd, Os, Ir, Pt, Cu, Zn, Au) that expand or
stabilise austenite, depress the position of TTT curves to lower temperature. In contrast
elements (Be, P, Ti, V, Mo, Cr, B, Ta, Nb, Zr) that favour the ferrite phase can raise the
eutectoid temperature and TTT curves move upward to higher temperature.
However Al, Co, and Si increase rate of nucleation and growth of both ferrite or pearlite and
therefore shift TTT diagram to left. In addition under the complex diffusional effect of
various alloying element the simple C shape behaviour of TTT diagram get modified and
various regions of transformation get clearly separated. There are separate pearlitic C
curves, ferritic and bainitic C curves and shape of each of them are distinct and different.
279
Effect of alloying elements on TTT diagram
The effect of alloying elements is less pronounced in bainitic region as the diffusion of only
carbon takes place (either to neighboring austenite or within ferrite) in a very short time
(within a few second) after supersaturated ferrite formation by shear during bainitic
transformation and there is no need for redistribution of mostly substitutional alloying
elements.
Therefore bainitic region moves less to higher timing in comparison to
proeutectoid/pearlitic region. Addition of alloying elements lead to a greater separation of
the reactions and result separate C-curves for pearlitic and bainitic region. Mo encourage
bainitic reaction but addition of boron retard the ferrite reaction. By addition of B in low
carbon Mo steel the bainitic region (almost unaffected by addition of B) can be separated
from the ferritic region.
However bainitic reaction is suppressed by the addition of some alloying elements. BS
temperature (empirical) has been given by Steven & Haynes
Bs ( 0C ) = 830 − 270(%C ) − 90(% Mn) − 37(% Ni ) − 70(%Cr ) − 83(% Mo)

Most alloying elements which are soluble in austenite lower MS, MF temperature except Al,
Co. Andrews gave best fit equation for MS:
M s ( 0C ) = 539 − 423(%C ) − 30.4 Mn − 17.7 Ni − 12.1Cr − 7.5Mo + 10Co − 7.5Si

Effect of alloying elements on MF is similar to that of MS. Therefore, subzero treatment is
must for highly alloyed steels to transform retained austenite to martensite. 280
Various types of TTT diagrams for alloy steels
(a) (b) (c) (d)
With resembalance of plain With both pearlitic and With out bainitic bay With out pearlitic bay
carbon steel bainitic bays
The (a) type of TTT The (b) type of TTT The (c) type of TTT The (d) type of TTT
diagram is similar to that diagram consists of two diagram is peculiar in the diagram does not exhibit
of carbon steel. In the minima with respect to sense that bainitic region pearlitic bay. Under
presence of carbide the stability of austenite. is not present. Such a normal conditions either
forming elements super The upper bay (γ → TTT diagram is obtained bainite or martensite can
cooled austenite pearlite), lower bay (γ→ , in general, for high be formed.
decomposes to a mixture bainite). Very few steels alloy steels and Ms
of ferrite and carbides exhibit such a TTT region shifted to sub zero
rather than to aggregate diagram. region
of ferrite and cementite.
281
Factors affecting on TTT diagram
Composition of steel
Carbon and alloying elements affect the transformation of austenite in many ways. As the
amount of carbon and most alloying elements increase in steel, the lower part of the curves
is progressively lowered because except for aluminium, all the elements lower the Ms
temperature.
The austenite stabilizers lower the Ac3 as well as Ac1 temperatures, i.e., these elements (Ni,
Mn, C etc) lower the upper part of the TTT curve. This probably is the reason of having
overlapping ‘C’ curves for pearlitic and bainitic transformations in plain carbon steel as
well as in steels having Ni, etc.
Ferrite stabilizers raise Ac3 as well as Ac1 temperatures. As the TTT curve for such steels
get raised upward as well as gets lowered down ward, there are invariably two ‘C’ curves-
one for pearlitic and the other for bainitic transformations. For example, presence of 0.8%
Cr and 0.33% Mo in steel yield two ‘C’ curves.
Andrew suggests the effect of the elements on Ac3 and Ac1 by equations :
Ac3 = 910 − 203 %C − 15.2(% Ni ) + 44.7(%Si ) + 104(%V ) + 31.5(%Mo) + 13.1(%W )
Ac1 = 727 − 10.7(% Mn) − 16.9(% Ni ) + 29.1(% Si ) + 290(% As) + 16.9(%Cr ) + 6.38(%W )
282
Effect of grain size Ae3 For finer austenite
Ae1
γ=austenite
α=ferrite α+CP
CP=coarse pearlite α+P
P=pearlite FP
Hardness
Temperature, T
FP=fine pearlite
UB=upper Bainite
LB=lower Bainite 50% FP + 50% UB
M=martensite UB
MS=Martensite start
temperature
M50=temperature at Metastable γ
which 50% martensite is
obtained
MF= martensite finish MS LB
temperature
M50 Metastable γ + M
MF
M
Log(time, t) 283
Effect of grain size

All the composition products of austenite nucleate heterogeneously preferentially at the
grain boundaries.
A fine grained steel has larger grain boundary area than a coarse grained steel, and
consequently favors nucleation of pearlite, bainite, ferrite, cementite and thus, reduces the
incubation period, that means the TTT curve of the fine grained steel is more towards left,
significantly in the pearlitic range, than a coarse grained steel of same composition.
Effect of Heterogeneity of Austenite
Heterogeneous austenite increases transformation time range, start to finish of ferritic,

pearlitic and bainitic range as well as increases the transformation temperature range in case
of Martensitic transformation and bainitic transformation. Undissolved cementite, carbides
act as powerful inoculants for pearlite transformation. Therefore heterogeneity in
austenite increases the transformation time range in Diffusional transformation and
temperature range of shear transformation products in TTT diagram.
This heterogeneity of austenite, when the steel is heated to temperatures not higher than
50°C above A1 is put to advantage to obtain spheroidized pearlite by letting it decompose
within 50°C below A1 to improve the ductility and the machinability (high carbon steels).
284
Applications of TTT diagrams
MARTEMPERING
This heat treatment is given to oil hardenable and air hardenable steels and thin section of
water hardenable steel sample to produce martensite with minimal differential thermal and
transformation stress to avoid distortion and cracking.
The steel should have reasonable incubation period at the nose of its TTT diagram and
long bainitic bay.
The sample is quenched above MS temperature in a salt bath to reduce thermal stress
(instead of cooling below MF directly) Surface cooling rate is greater than at the centre.
The cooling schedule is such that the cooling curves pass behind without touching the nose
of the TTT diagram. The sample is isothermally hold at bainitic bay such that differential
cooling rate at centre and surface become equalize after some time.
The sample is allowed to cool by air through MS-MF such that martensite forms both at
the surface and centre at the same time due to not much temperature difference and thereby
avoid transformation stress because of volume expansion. The sample is given tempering
treatment at suitable temperature.
The rate of cooling from austenising temperature should avoid formation of pearlitic and
bainite. Thus the success of Martempering depends on the incubation period at the pearlitic
nose, and also at the bainitic bay, and thus the steel suitable for Martempering should have
sufficient hardenability. Plain carbon steels with dia. > 10mm are difficult to mar-temper
and thus, this process is restricted to alloy steels. 285
Martempering
Ae3
Ae1 γ=austenite
α=ferrite
α+CP CP=coarse pearlite
α+P P=pearlite
t0 FP FP=fine pearlite
t0=minimum incubation
50% FP + 50% UB period
Temperature
UB=upper bainite
UB LB=lower bainite
M=martensite
MS=Martensite start
Tempering temperature
M50=temperature at which
MS LB 50% martensite is
obtained
M50 MF= martensite finish
Metastable γ + martensite temperature
MF
Martensite Tempered martensite
286
Log time
AUSTEMPERING
Austempering heat treatment is given to steel to produce lower bainite in high carbon steel
without any distortion or cracking to the sample.
The heat treatment is cooling of austenite rapidly in a bath maintained at lower bainitic
temperature (above Ms) temperature (avoiding the nose of the TTT diagram) and holding it
here to equalize surface and centre temperature and till bainitic finish time.
At the end of bainitic reaction sample is air cooled. The microstructure contains fully lower
bainite. This heat treatment is given to 0.5-1.2 wt%C steel and low alloy steel.
The product hardness and strength are comparable to hardened and tempered martensite
with improved ductility and toughness and uniform mechanical properties. Products do not
required to be tempered.
TTT diagram is great help in scheduling Austempering as it fixes
Temperature of holding of the bath
Duration of holding time
Section which can be quenched to avoid pearlite formation
Whether Austempering is worth while , or not, as the time at the bay may be too long.
287
Austempering
Ae3
Ae1
γ=austenite
α+CP α=ferrite
α+P CP=coarse pearlite
t0 FP P=pearlite
FP=fine pearlite
t0=minimum incubation
Temperature
50% FP + 50% UB
period
UB UB=upper bainite
LB=lower bainite
Metastable γ M=martensite
MS=Martensite start
temperature
MS LB M50=temperature at which
50% martensite is
M50 obtained
Metastable γ + martensite MF= martensite finish
MF temperature
Martensite
Lower bainite
288
Log time
ISOTHERMAL ANNEALING
Isothermal annealing consists of four steps.

The first step is heating the steel components similar as in the case of full annealing.
The second step is slightly fast cooling from the usual austenitizing temperature to a
constant temperature just below A1.
The third step is to hold at this reduced temperature for sufficient soaking period for
the completion of transformation and the final step involves cooling the steel
component to room temperature in air.
Figure (nextslide) depicts the heat treatment cycles of full annealing and isothermal
annealing. The terms α, γ, P, Ps and PF refer to ferrite, austenite, pearlite, pearlite starting
and pearlite finish, respectively.
Isothermal annealing has distinct advantages over full annealing which are given below.
Reduced annealing time, especially for alloy steels which need very slow cooling to
obtain the required reduction in hardness by the full annealing.
More homogeneity in structure is obtained as the transformation occurs at the same
time throughout the cross section.
Improved machinability and surface finish is obtained after machining as compared to
that of the full annealed components.
289
Isothermal Annealing
Isothermal annealing is primarily used for medium carbon, high carbon and some of the alloy
steels to improve their machinability. 290
PATENTING
Patenting is an isothermal heat treatment processes used for producing high strength ropes,
springs, and piano wires of normally 0.45% to 1.0% carbon steel. In fact, the strongest
material in commercial quantities is the patented and cold drawn wire made from 0.80% to
1.0% carbon steel, containing no martensite. Wires having high ultimate tensile strength of
4830MN/m2 with 20% elongation have been reported.
Patenting consists of austenitising steel in continuous furnace to temperature 150-200°C
above Ac3 , then cooling rapidly in and holding in a lead, or salt bath maintained at a
temperature of 450-550°C (near the pearlitic nose of its TTT curve, i.e., in the lower
temperature limit of pearlitic transformation) for sufficient time for austenite to completely
transform to finest pearlite (some upper bainite may also form), and then wound on to
driven drum
The eutectoid steel after patenting may have interlamellar spacing as small as 40nm with
strength as 1240-1450 Mpa. The process of patenting helps in two ways to obtain high
strengths. Patented wires can be cold drawn by large extent (80-90%) without fracture as
the soft , weak primary ferrite or brittle cementite is absent and the interlamellar spacing is
same every where.
TTT diagram help to fix the time and temperature of holding. The speed of motion of the
wire through the bath should be such that it remains in bath for a time slightly greater than
the time of completion of pearlitic transformation to avoid any chance of untransformed
austenite to bainite or martensite later. 291
Patenting
292
Figure : Patenting (a) Schematic process of patenting, (b) TTT diagram of 0.6% carbon steel
Continuous Cooling Transformation (CCT) diagrams
The TTT diagrams are also called Isothermal Transformation Diagrams, because the
transformation times are representative of isothermal hold treatment (following a
instantaneous quench).
In practical situations we follow heat treatments (T-t procedures/cycles) in which (typically)
there are steps involving cooling of the sample. The cooling rate may or may not be
constant. The rate of cooling may be slow (as in a furnace which has been switch off) or
rapid (like quenching in water).
Hence, in terms of practical utility TTT curves have a limitation and we need to draw
separate diagrams called Continuous Cooling Transformation diagrams (CCT), wherein
transformation times (also: products & microstructure) are noted using constant rate cooling
treatments.
A diagram drawn for a given cooling rate (dT/dt) is typically used for a range of cooling
rates (thus avoiding the need for a separate diagram for every cooling rate).
However, often TTT diagrams are also used for constant cooling rate experiments- keeping
in view the assumptions & approximations involved.
Important difference between the CCT & TTT transformations is that in the CCT case
Bainite cannot form.
The CCT diagram for eutectoid steel is considered next.
293
Determination of CCT diagram for eutectoid steel
CCT diagrams are determined by measuring some physical properties during continuous
cooling. Normally these are specific volume and magnetic permeability. However, the
majority of the work has been done through specific volume change by dilatometric
method. This method is supplemented by metallography and hardness measurement.
In dilatometry the test sample is austenitised in a specially designed furnace and then
controlled cooled. Sample dilation is measured by dial gauge/sensor. Slowest cooling is
controlled by furnace cooling but higher cooling rate can be controlled by gas quenching
Sample and fixtures

for dilatometric measurements Dilatometer equipment
294
Cooling data are plotted as temperature versus time (Fig. a). Dilation is recorded against
temperature (Fig. b). Any slope change indicates phase transformation. Fraction of
transformation roughly can be calculated based on the dilation data as explained below.
a
X
c
Temperature
Dilation
b
V
d Z
IV
I II III
TF TS
Time Temperature
Fig. b : Dilation-temperature plot

Fig. a : Schematic cooling curves
for a cooling curve
295
In Fig. a curves I to V indicate cooling curves at higher cooling rate to lower cooling rate
respectively.
Fig. b gives the dilation at different temperatures for a given cooling rate/schedule.
In general slope of dilation curve remains unchanged while amount of phase or the relative
amount of phases in a phase mixture does not change during cooling (or heating) however
sample shrink or expand i.e. dilation takes place purely due to thermal specific volume
change because of change in temperature.
Therefore in Fig. b dilation from a to b is due to specific volume change of high
temperature phase austenite. But at TS slope of the curve changes.
Therefore transformation starts at TS. Again slope of the curve from c to d is constant but is
different from the slope of the curve from a to b.
This indicates there is no phase transformation between the temperature from c to d but the
phase/phase mixture is different from the phase at a to b.
Slope of the dilation curve from b to c is variable with temperature. This indicates the
change in relative amount of phase due to cooling.
The expansion is due to the formation of low density phase(s). Some part of dilation is
compensated by purely thermal change due to cooling.
Therefore dilation curve takes complex shape. i.e first slope reduces and reaches to a
minimum value and then increases to the characteristic value of the phase mixture at c.
296
Therefore phase transformation start at b i.e. at temperature TS and transformation ends or

finishes at c or temperature TF. The nature of transformation has to be determined by
metallography.
When austenite fully transforms to a single product then amount of transformation is
directly proportional to the relative change in length. For a mixture of products the
percentage of austenite transformed may not be strictly proportional to change in length,
however, it is reasonable and generally is being used.
Cumulative percentage of transformation at in between temperature T is equal to
YZ/XZ*100 where X, Y and Z are intersection point of temperature T line to extended
constant slope curve of austenite (ba), transformation curve (bc) and extended constant
slope curve of low temperature phase (cd) respectively.
So at each cooling rate transformation start and finish temperature and transformation
temperature for specific amount (10 %, 20%, 30% etc.) can also be determined.
For every type of transformation, locus of start points, is percentage points and finish points
give the transformation start line, is percentage lines and finish line respectively and that
result CCT diagram.
Normally at the end of each cooling curve hardness value of resultant product at room
temperature and type of phases obtained are shown.
297
CCT diagram for eutectoid steel
800
Eutectoid temperature
723
600 Pearlite
Temperature →
500
Original TTT lines
400
Austenite
300
Ms
200
Mf
100
Martensite
0.1 1 10 102 103 104 105
t (s) → 298
Different cooling rates for eutectoid steel
800
723
Process Cooling rate (K/s)
Annealing 10−5 – 10−3 600

Air Cooling 1 – 10
Oil Quenching ~100 500

Temperature →
Water Quenching ~500
Splat Quenching 105

400
Melt-Spinning 106 – 108
Evaporation,
109 (expected)
300
sputtering
200
Coarse P
100
M + +P
M Fine P
0.1 1 10 102 103 104 105

M = Martensite P = Pearlite t (s) → 299
Questions?
1. The Johnson-Mehl equation, in most cases adequately describes the kinetics of the austenite
to pearlite transformations. This can be demonstrated by taking several sets of N and G
values, which contains data from a 0.8% C plain carbon steel. This figure indicates that at
7000C, N = 6.31×10-4 nuclei per mm3 per sec and G = 3.16 × 10-4 mm per sec. Substitute
these values of N and G into J-M equation, and obtain the pearlite reaction curves for this
steel at 700 and 5000C. Plot the fraction transformed , f(t), versus log(t) as in fig b.
(a) (b)
300
Questions?
2. Determine, using the curves obtained in problem 1, the times

required to obtain 1% and 99% pearlite at 700 and 550oC. (c)
Compare these times with the experimental data for an
eutectoid plain carbon steel shown in the T-T-T in fig c. is the
agreement reasonable?
3. In a slowly cooled hypereutectoid iron-carbon steel, the
pearlite colonies are normally separated from each other by a
more or less continuous boundary layer of cementite. Explain
how this microstructure develops. Use simple sketches to
illustrate your answer.
4. In slowly cooled hypoeutectoid steels one normally finds a microstructure in which the
pearlite colonies are surrounded by ferrite grains. Explain.
5. After being slowly cooled from the austenite region, a simple iron-carbon steel exhibits a
microstructure consisting of 40 percent pearlite and 60 percent ferrite.
a) Estimate the carbon concentration of the steel.
b) Describe the equilibrium microstructure that would be obtained if the steel were heated
to 730°C and held there for a long period of time.
c) What would be the equilibrium structure of this steel if it were heated to 850°C?
d) Make sketches of all of these microstructures.
301
Questions?
6. Consider an iron-carbon alloy containing 1.0 percent carbon.

a) If this composition is slowly cooled from the austenite region, what would be the
respective percentages of the constituents and the phases in its microstructure?
b) Now assume that it has been cooled at a rate rapid enough to yield a proeutectoid
constituent of only 1.2 percent. What would be the percentages of the constituents and
the phases in this case?
7. Explain in detail how bainite differs from martensite and from pearlite. How do upper and
lower bainite differ?
8. Describe the meaning of the term partitioning coefficient.
9. Explain why C occupies an octahedral void not the tetrahedral void in α−Fe.
10. In the Fe-C system Fe3C is only a metastable phase, while graphite is the most stable carbon
rich phase. By drawing schematic free energy-composition diagrams show how the Fe-
graphite phase diagram compares to the Fe-Fe3C phase diagram from 0 to 2 wt% Fe.
11. An iron-carbon steel containing 0.5 percent of carbon has a microstructure consisting of 85
percent pearlite and 15 percent ferrite.
a) Are these the amounts of these constituents that one would expect to find in the steel if
it had been slowly cooled from the austenite region?
b) In a slowly cooled microstructure, pearlite normally has a 7 to 1 ratio of the widths of
the ferrite and cementite lamellae. What would this ratio be in the present case?
302
Questions?
12. Answer the following with regard to the T-T-T diagram of a steel of eutectoid composition,
and assume that the specimens involved in the various cooling paths were cut from a thin
sheet and austenitized at 750°C before cooling. Describe the microstructure resulting from
being:
a) Cooled to room temperature in less that 1 s.
b) Cooled to 160°C in less than 1 s and then maintained at this temperature for several
years.
c) Quenched to 650°C and held at this temperature for 1 day, then quenched to room
temperature.
d) Quenched to 550°C and held at this temperature for 1 day, then quenched to room
temperature.
13. A high-carbon steel containing 1.13 percent carbon, with a (d)

microstructure similar to that in Fig. d, is heated to 730°C and
allowed to come to equilibrium. It is then quenched to room
temperature. Make a sketch of the resulting microstructure,
identifying the constituents.
Fig. d: Hypereutectoid-steel microstructure. Notice the band of cementite plates outlining the pearlite colony in the center of
the photograph. 1000X 303
Questions?
14. Assume that the 0.52 percent carbon steel of Fig. (e) is (d)
slowly cooled to the temperature of point c and then
quenched to 450°C and held at this latter temperature for a
period of a day. Describe the microstructure that one might
find. Determine the percentages of the constituents.
15. Transpose paths 1, 2, and 3 of Fig. g to Fig. f and determine
the microstructures corresponding to these paths in the
hypereutectoid steel.
(g)
(f)
Fig. f: Isothermal transformation diagram for a

hypereutectoid steel: 1.13 percent carbon, 0.30 percent Fig. g: Arbitrary time-temperature paths304on the
manganese. isothermal diagram of a hypoeutectoid steel
Questions?
16. With regard to the data for eutectoid steels in (h)

Figure h.
a) What is the interlamellar spacing in the pure
Fe-C alloy at 660°C?
b) How much smaller is it at this temperature
when the steel contains 1.8 percent
chromium?
c) How much larger is it in the steel with 1.8 %
manganese?
Fig. h: Minimum interlamellar spacing data for

eutectoid steel containing additions of manganese
and chromium
17. Estimate %Cm in Ledeburite just below eutectic and just above eutectoid temeratures. What
is its structure at room temperature?
18. Use the concept of Bain distortion to estimate maximum displacement experienced by iron
atom during martensitic transformation. Lattice parameters of austenite and ferrite are
0.356nm & 0.286 respectively. Assume c/a ratio of martensite to be 1.15.
19. Carbon atoms occupy octahedral interstitial sites in austenite and ferrite. Estimate fraction
of these sites that are occupied in these if carbon contents are 0.1 and 0.01wt% respectively.
305
Questions?
20. Show that the inter-lamellar spacing of pearlite is inversely proportional to the degree of
under cooling.
21. Which is the more stable, the pearlitic or the spheroiditic microstructure? Why?
22. How are the critical temperatures in steel designated? What is the critical range?
23. Draw the Fe-Fe3C phase diagram and label the phase fields. Discuss in brief the different
reactions that takes place in this system.
24. Compute the following
a) % pearlite and cementite in asteel containing 1.2% C.
b) % austenite and ledeburite in cast iron containing 3.5% C.
c) % ferrite and cementite in a steel containing 0.8% carbon.
25. Discuss in brief the effect of alloying elements in Fe-Fe3C phase diagram.
26. Discuss in brief the effect of along elements in TTT & CCT diagrams
27. Draw the TTT diagram for Hypo eutectoid, Eutectiod and Hyper eutectoid steels.
28. Draw the CCT diagram for Hypo eutectoid, Eutectiod and Hyper eutectoid steels.
29. CCT curve of eutectoid steel is displaced towards right and lowered as compared to its TTT
curve. Why?
30. What are the limitations of Fe-Fe3C diagram, TTT diagram and CCT diagram?
306
Avala Lava Kumar*
Introduction
The majority of phase transformations that occur in the solid state takes place by thermally
activated atomic movements.
The transformations that will be dealt with in this chapter are those that are induced by a
change of temperature of an alloy that has a fixed composition.
1-phase
Many-phases
C0
Usually we will be concerned with the transformations caused by a temperature change
from a single phase region of a (binary) phase diagram to a region where one or moreother
phases are stable.
The different types of phase transformations that are possible can be roughly divided in to
several groups (see in next slide).
308
Classification of Diffusional transformations
Precipitation α '→α + β
α ' → metastablesupersaturated solid solution
Eutectoid γ →α + β
Diffusional
Polymorphic β crystal −1 → α crystal − 2
Transformations
Ordering α (disordered ) → α '(ordered )
Original phase decomposes to one or more

new phases which have the same composition
Massive β →α as the parent phase but different crystal
structures 309
Precipitation
α ' →α + β
Eutectic System Eutectoid System Peritectic System
Polymorphic
Eutectoid
γ → α +β
β crystal −1 → α crystal − 2 310
Ordering
α ( disordered ) → α '(ordered )
Massive
β →α
311
Precipitation Hardening
The most important heat treating process for non ferrous alloys is age hardening, or
precipitation hardening.
In order to apply this heat treatment, the equilibrium diagram must show partial solid
solubility, and the slope of the solvus line must be such that there is greater solubility at a
higher temperature than at a lower temperature.
The purpose of precipitation hardening treatment is to improve strength of the materials. It
can explained by with respect to dislocations.
The presence of dislocation weakens the crystal → easy plastic deformation
Putting hindrance to dislocation motion increases the strength of the crystal
Fine precipitates dispersed in the matrix provide such an impediment.
For example: Strength of Al → 100 MPa
Strength of Duralumin (Al + 4% Cu + other alloying elements) → 500 Mpa
Two stages are generally required in heat treatment to produce age hardening:
Solution treatment
Aging
312
L
600
α
T ((ºC) →
400
θ
Sloping Solvus line
⇒ high T → high solubility
200
low T → low solubility
of Cu in Al
Al 15 30 45 60
% Cu →
Al rich end of the Al-Cu phase diagram

313
On Equilibrium cooling
α→α+θ
Slow equilibrium cooling gives rise to
α coarse θ precipitates which is not good
in impeding dislocation motion.*
α
α+θ
θ
4 % Cu
α ( FCC )   α ( FCC )  θ CuAl2 (Tetragonal ) 

  slow cool    
 4 % Cu    → 0.5 % Cu  +  52 % Cu 
 550o C   RT   RT 
      314
To obtain a fine distribution of precipitates the
cycle A → B → C is used α
B
Solution treating at high temperature, then
quenching (second phase is in solid solution)
→Cycle A and B
A C
α+θ
A Heat (to 550oC) → solid solution α
4 % Cu
supersaturated solution
B Quench (to RT) →
Increased vacancy concentration
Ageing at room temperature or slightly higher temperature (Precipitation of second phase,
giving strengthening effect) →Cycle C
C Age (reheat to 200oC) → fine precipitates 315

Note: Treatments A, B, C are for the same composition
180oC Peak-aged
100oC
Hardness →
Dispersion of Coarsening
Hardness →
fine precipitates of precipitates

(closely spaced) with increased
180oC interparticle spacing
20oC Underaged Overaged
Log(t) →
Log(t) → Region of solid solution

Region of precipitation
hardening (but little/some
strengthening solid solution
(no precipitation hardening) strengthening)
Higher temperature ⇒ less time of aging to obtain peak hardness

Lower temperature ⇒ increased peak hardness
⇒ optimization between time and hardness required
316
180oC Peak-aged
Hardness →
Section of GP zone parallel to (200) plane
Log(t) →
CRSS Increase →
Particle Particle r = f (t )
shearing By-pass
1 1
∝r 2 ∝
r
317
Particle radius (r) →
180oC Peak-aged
A complex set of events
are happening
Hardness →
parallely/sequentially
during the aging process
→ These are shown
schematically in the
figure ⇒
Log(t) →
Increasing size of precipitates with increasing interparticle (inter-precipitate) spacing
Interface goes from coherent to semi-coherent to incoherent
Precipitate goes from GP zone → θ’’ → θ’ → θ

318
GP Zones
Cu rich zones fully coherent with the matrix → low interfacial energy
(Equilibrium θ phase has a complex tetragonal crystal structure which has incoherent
interfaces)
Zones minimize their strain energy by choosing disc-shape ⊥ to the elastically soft <100>
directions in the FCC matrix
The driving force (∆Gv − ∆Gs) is less but the barrier to nucleation is much less (∆G*)
2 atomic layers thick, 10nm in diameter with a spacing of ~10nm
The zones seem to be homogenously nucleated (excess vacancies seem to play an
important role in their nucleation)
Due to large surface to volume ratio the fine precipitates have a tendency to coarsen →
small precipitates dissolve and large precipitates grow
Coarsening
⇒ ↓ in number of precipitate
⇒ ↑ in interparticle (inter-precipitate) spacing
⇒ reduced hindrance to dislocation motion (τmax = Gb/L)
319
10 nm thick ,100 nm diameter (001)θ '' (001)α
Distorted FCC
Unitcell composition θ '' [100]θ '' [100]α
Al6Cu2 = Al3Cu
Become incoherent as ppt. grows

(001)θ ' (001)α
θ'
Tetragonal
[100]θ ' [100]α
Unitcell composition
Al4Cu2 = Al2Cu
α
θ
Body Centered
Tetragonal Unitcell composition
Al8Cu4 = Al2Cu
320
Schematic diagram showing the lowering of the Gibbs free energy of the system on sequential transformation:
GP zones → θ’’ → θ’ → θ 321
The activation barrier for
θ)
precipitation of equilibrium (θ
phase is large
But, the free energy benefit in each step is small compared to the
overall single step process
Single step
(‘equilibrium’) process
Schematic plot 322

In this diagram additionally information has
been superposed onto the phase diagram (which
strictly do not belong there- hence this diagram
should be interpreted with care)
The diagram shows that on aging at various
temperatures in the α + θ region of the phase
diagram various precipitates are obtained first
At higher temperatures the stable θ phase is

produced directly
At slightly lower temperatures θ’ is produced
first
At even lower temperatures θ’’ is produced

first
The normal artificial aging is usually done in

this temperature range to give rise to GP
zones first
323
Phase and TTT diagrams
324
Phase and TTT diagrams
In previous slide, we show the Al-rich portion of the Al-Cu phase diagram (schematically).
From the phase diagram, it is clear that the Al with a few percent copper is cooled from high
temperature leads to the formation of a microstructure in which the θ phase precipitates out
of the supersaturated α matrix.
However, if an alloy of composition Al - 4 wt. %Cu is solutionised at say, 540°C, and the
resultant phase is rapidly quenched to room temperature, the solid solution is largely
retained; if this alloy is kept at room temperature (or at any temperature below 180°C), a
metastable phase known as Guinier-Preston zones (GP zones) is formed.
Similarly, the aging treatment at other temperatures can produce other precipitates such as
θ’’ and θ’.
The solvus for these metastable phases in shown in Fig. Further, we also show the
corresponding time-temperature-transformation curves for these metastable phase as well as
the stable phase.
we already discussed the crystallography and interface structure of all these phases as well
as the reasons for their formation.
325
Misfit strain effects : Coherent precipitate
The equilibrium shape of a coherent precipitate or zone is, when the misfit between the
precipitate and matrix is small.
When misfit is present is present the formation of coherent interfaces raises the free energy
of the system on accounted of the elastic strain fields that arise.
If the elastic strain energy is denoted by ∆Gs the condition for equilibrium becomes
∑ Aγ i i + ∆Gs = min imum

If the lattice parameters of the unstrained precipitate and matrix are ap and am , respectively,
the unconstrained misfit (δ) is defined by
a p −a m
δ=
am
Schematic showing a
coherent precipitate and
the origin of coherency
strains
326
When the precipitate is incoherent with the matrix, there is an attempt at matching the two
lattices and lattice sites are not conserved across the interface.
Under these circumstances there are no coherency strains. Misfit strains can, however, still
arise if the precipitate is the wrong size for the hole it is located.
In this case the lattice misfit δ has no significance and it is better to consider the volume
misfit ∆ as defined by V −V
∆=
p m
Vm
327
The origin of misfit strain for an incoherent precipitate (no lattice matching
Choose particle shape → minimize the free energy of system
The variation of misfit strain energy with aspect ratio of ellipsoid shape 328
Precipitation Sequence in some precipitation hardening systems
Base Metal Alloy Precipitation Sequence

Aluminium Al-Ag GPZ (spheres) → γ’ (plates) → γ (Ag2 Al)
Al-Cu GPZ (discs) → θ’ ‘ (discs) → θ’ (plates) → θ (CuAl2 )
Al-Cu-Mg GPZ (rods) → S ‘ (laths) → S (CuMgAl2 ) (laths)
Al-Zn-Mg GPZ (spheres) → η ‘ (plates) → η (MgZn2 ) (plates or rods)
Al-Mg-Si GPZ (rods) → β‘ (rods) → β (Mg2 Si) (plates)
Copper Cu-Be GPZ (discs) → γ‘ → γ (CuBe)
Cu-Co GPZ (spheres) → β (Co) (plates)
Iron Fe-C ε-carbide (discs) → Fe3C
Fe-N α‘ ‘ (discs) → Fe4 N
Nickel Ni-Cr-Ti-Al γ‘ (cubes or spheres)
329
Details in ‘practical’ aging curves
In low T aging (130oC)→The aging curves have more detail than the single peak as
discussed schematically before.
In low T aging (130oC)→ the full sequence of precipitation is observed (GPZ →θ''→ θ').
At high T aging (190oC)→ θ'' directly forms (i.e. the full precipitation sequence is not
observed).
Peak hardness increases with increasing Cu%.
For the same Cu%, the peak hardness is lower for the 190oC aging treatment as compared to
the 130oC aging treatment.
Peak hardness is achieved when the microstructure consists of a θ' or combination of (θ' +
θ''). 330
Particle/precipitate Coarsening
There will be a range of particle sizes due to time of nucleation and rate of growth
As the curvature increases the solute concentration (XB) in the matrix adjacent to
the particle increases
Concentration gradients are setup in the matrix → solute diffuses from near the
small particles towards the large particles
⇒ small particles shrink and large particles grow
⇒ with increasing time * Total number of particles decrease
* Mean radius (ravg) increases with time
331
r0 → ravg at t = 0
3
ravg − r03 = kt
D → Diffusivity
ravg Xe → XB (r = ∞)
k ∝ Dγ Xe
r0 D is a exponential function of temperature
⇒ coarsening increases rapidly with T
t 332
dravg k ⇒ small ppts coarsen more
Volume diffusion rate ∝ 2 rapidly
dt ravg
Rate controlling
3
factor Linear ravg versus t relation may break down due to diffusion short-circuits
or if the process is interface controlled
Interface diffusion rate
Diffusion coefficient (D)
Rate of coarsening Interfacial energy
Equilibrium solubility of very large particles (Xe)
Precipitation hardening systems employed for high-temperature applications must avoid

coarsening by having low: γ, Xe or D
333
Rate of coarsening
Low γ
In superalloys, Strength obtained by fine dispersion of γ’ [ordered FCC Ni3(TiAl)]

precipitate in FCC Ni rich matrix
Matrix (Ni SS)/ γ’ matrix is fully coherent [low interfacial energy γ = 30 mJ/m2]
Misfit = f(composition) → varies between 0% and 0.2%
Creep rupture life increases when the misfit is 0% rather than 0.2%
Low Xe
ThO2 dispersion in W (or Ni) (Fine oxide dispersion in a metal matrix)

Oxides are insoluble in metals
Stability of these microstructures at high temperatures due to low value of Xe
The term DγXe has a low value
Low D
ThO2 dispersion in W (or Ni) (Fine oxide dispersion in a metal matrix)

Cementite dispersions in tempered steel coarsen due to high D of interstitial C
If a substitutional alloying element is added which segregates to the carbide → rate of coarsening
↓ due to low D for the substitutional element 334
Spinodal decomposition/Spinodal clustering
Phase diagrams showing miscibility gap correspond to solid solutions which exhibit
clustering tendency.
Within the miscibility gap the decomposition can take place by either
Nucleation and Growth or by
Spinodal Mechanism
If the second phase is not coherent with the parent then the region of the spinodal is called
the chemical spinodal
If the second phase is coherent with the parent phase then the spinodal mechanism is
operative only inside the coherent spinodal domain
As coherent second phases cost additional strain energy to produce (as compared to a
incoherent second phase – only interfacial energy involved) → this requires additional
undercooling for it to occur
Spinodal decomposition is not limited to systems containing a miscibility gap
Other examples are in binary solid solutions and glasses
All systems in which GP zones form (e.g.) contain a metastable coherent miscibility gap →
THE GP ZONE SOLVUS
Thus at high supersaturations it is GP zones can form by spinodal mechanism.
335
Spinodal mechanism
In phase separating systems, at low temperatures, the Gibbs free energy consists of regions
of concave curvature (as shown Fig. 1) and in such regions the binary alloy separated into a
mechanical mixture of A and B-rich regions instead of remaining a solid solution. This is
because such a phase separation into mechanical mixture reduces the free energy of the
system.
Further, in such systems with concavity of free energy, for certain compositions, the
mechanism of phase separation changes from the classical nucleation and growth to
spinodal decomposition. The change over in the mechanism is related to the curvature of the
free energy curve as shown in Fig. 2.
Fig. 1 Fig. 2
The free energy of a system with concave curvature. In Positive curvature (nucleation) and negative curvature
the concave curvature region, the system becomes a (spinodal) regions of the free energy versus composition
mechanical mixture of A rich and B rich phases with the diagram; phase separation mechanism changes 336
from
given compositions marked in this figure. nucleation to spinodal at the point of zero curvature.
d 2G
Spinodal mechanism ⇒ 2 <0
dx
When/How it is possible?
We always know that, towards the two pure metal ends, the slope of free energy verses
composition is minus infinity.
We will understand the Spinodal decomposition better, if we try

to see, what is the first and second derivatives of this free energy. (a)
∂G ∂x is nothing but slope of the curve (a) and it will be
given in Fig. (b).
2
∂G 2 ∂x is slope of the curve (b) and it will be given in Fig.
(c).
(b)
(c)
337
Composition fluctuations
Spinodal mechanism
Consider a system with an interdiffusion coefficient D and undergoing spinodal
decomposition. Within the spinodal region, the composition fluctuations grow as shown in
Fig. (previous slide); the fluctuations grow with a characteristic time constant
τ = −λ2 / 4π 2 D where λ is the wavelength of the composition modulation

(assuming one-dimensional modulations).
Thus, for smaller λ, the rate of transformation becomes high; however, there is a minimum λ
below which spinodal decomposition cannot occur; this is because, during spinodal
decomposition, as A- and B-rich regions are forming, there are also interfaces between these
regions where AB bonds are formed which are energetically costlier; these regions give rise
to an increase in free energy; the `incipient' interfacial energy associated with the formation
of these regions with large AB bonds are the ones which set the lower wavelength limit. The
lower limit on the wavelength λ can be obtained using the following argument.
Consider a homogeneous alloy of composition xB0 decomposing into two parts: one with
composition xB + ∆x and another with composition . It can be shown that the total free
0
energy change associated with this decomposition is
1 d 2G
∆Gchem = ( ∆x ) 2
2 dx 2B
338
Spinodal mechanism
As noted earlier, the AB bonds in the incipient interface regions also contribute to the free
energy; this free energy contribution, thus, is associated with the gradients in composition.
Consider a sinusoidal composition modulation of wavelength λ and amplitude ∆x ; the
maximum composition gradient is thus ∆x/λ and the gradient energy contribution is
 ∆x 
2 where ∆Ggrad is a proportionality constant which is
∆Ggrad = κ  dependent on the difference between AB and AA and BB
 λ  bond energies.
The total change in free energy associated with a composition fluctuation of wavelength λ is
thus given by the addition of the chemical and gradient terms
 d 2G 2κ  (∆x) 2
∆G = ∆Gchem + ∆Ggrad = 2 + 2 
 dx B λ  2
 
From the above expression, it is clear that for spinodal decomposition
d 2G 2κ
− 2 > 2
dx B λ
Or
2κ
λ2 > − 2
d G
dxB2
339
Spinodal mechanism
In this figure we show the phase
separation region along with the
points at every temperature at which
the curvature of the free energy versus
composition plot changes its sign; the
locus of these points is as shown and
is known as chemical spinodal.
In the case of spinodal phase
separation, any small composition
fluctuation grows leading to A-rich
regions becoming richer in A and B-
rich regions becoming richer in B.
This is because such a process leads to
a decrease in free energy.
Thus, the process of spinodal
decomposition is in contrast to the
classical diffusion equation scenario
which predicts that regions with
positive curvature for composition
profile grow in time while those with
negative curvature decay leading to
homogenisation. 340
Spinodal Ordering
Ordering leads to the formation of a superlattice

Ordering can take place in Second Order or First Order (in continuous mode
below Ti) modes
Any change in the lattice dimensions due to ordering introduces a third order
term in the Landau equation
∆G = Aη 2 + Bη 3 + C η 4 ...
Not zero
Continuous ordering as a first order transformation requires a finite supercooling
below the Coherent Phase Boundary to the Coherent Instability (Ti) boundary
These (continuous ordering) 1st order transitions are possible in cases where the
symmetry elements of the ordered structure form a subset of the parent
disordered structure
341
Spinodal Ordering
L+α L
Ordered solid
α’
A B
α + α’
342
Summary
Nucleation & Growth Spinodal

The composition of the second phase A continuous change of composition occurs
remains unaltered with time until the equilibrium values are attained
The interfaces between the nucleating The interface is initially very diffuse but
phase and the matrix is sharp eventually sharpens
There is a marked tendency for random
A regularity- though not simple- exists both
distribution of both sizes and positions
in sizes and distribution of phases
of the equilibrium phases
The separated phases are generally non-
Particles of separated phases tend to be
spherical and posses a high degree of
spherical with low connectivity
connectivity
343
Eutectoid Transformation in Fe-Fe3C phase diagram
0.1 0.16 0.51
Peritectic Liquid (L)
Eutectic
L+δ→γ 1493ºC
L → γ + Fe3C
δ (Ferrite)
L+ γ L + Fe3C
Temperature →
δ+ γ γ (Austenite) 1147ºC
Hypo Eutectic Hyper Eutectic
910ºC 2.11
[γγ(Austenite) + Fe3C(Cementite)] = Ledeburite
723ºC
α (Ferrite)
Eutectoid 4.3
Eutectoid
Hyper
Hypo
0.025 %C
Eutectoid α (Ferrite) + Fe3C (Cementite)] = Pearlite

[α
γ → α + Fe3C Steels Cast Irons
0.008 0.8 % Carbon → 6.7
Fe Fe
344 3C
Iron – Cementite Phase Diagram
Eutectoid Transformation : Nucleation of Pearlite
Pearlite typically forms on the grain boundaries of the austenite phase by the nucleation of
either the ferrite or cementite; the exact phase that nucleates is decided both by the
composition and the structure of the grain boundary.
The nucleation is such that the barrier for nucleation is the lowest; in other words, the nuclei
has a orientation relationship with one of the grains such that the interfacial free energy is
minimized.
Suppose the first phase to nucleate is cementite; this leads to a depletion of the carbon
surrounding this region leading to ferrite nucleation. Ferrite nucleation is also such that the
interfacial free energy is minimized; thus, it also has an orientation relationship with the
cementite. This process is repeated with the entire grain boundary coated with alternating
cementite and ferrite nuclei.
Branching mechanism
345
Nucleation of the eutectoid phases in a system with eutectoid composition
If the composition of the steel is not the eutectoid composition, then, it is possible that
proeutectoid ferrite or cementite is nucleated at the grain boundary. The other phase, be it
cementite or ferrite, then forms on the incoherent boundary of this proeutectoid phase. This
process is shown schematically in Fig.
Orientation Relation:
Kurdyumov-Sachs
−
(100) c (111)γ
Proeutectoid cementite
(010) c (110)γ
−
(001) c (112)γ
Nucleation of the eutectoid phases in a system with pro-eutectoid composition
The development of Pearlitic microstructure requires cooperative growth of Ferrite and

Cementite
If this cooperation is not established then the resulting microstructure is non-lamellar →
Degenerate Pearlite 346
Eutectoid Transformations
Nucleation at GB Pearlite in Fe-C alloys γ → α + Fe3C
Low undercooling (below A1)

Small number of nuclei form which grow without interference
Large undercooling (below A1)

High nucleation rate → Site saturation along GB
347
Decomposition of Austenite
The transformation of austenite on cooling is a complex processes in the sense that
compositional as well as configurational changes are involved. Carbon present in the
austenite adjusts itself in such a way that at one end it leaves behind an almost carbon free
phase (ferrite) and on the other it combines with iron to form cementite.
Similarly large scale redistribution of alloying elements takes place when the solubility of
the element is restricted in alpha iron, and/or there is a strong tendency for the element to
form carbides. Also, the crystal structure changes from γ-iron to α-iron.
The process of decomposition of austenite to ferrite-cementite aggregate is essentially a
diffusion controlled processes and proceeds by nucleation and growth mechanism.
Therefore, any factor, which is able to vary either the rate of nucleation/ the rate of growth
or both is also able to change the nature of ferrite cementite mixture.
Some of These factors are: nature of austenite under consideration, the temperature of
transformation and the presence of alloying elements.
The more the homogeneity of austenite, the better is the probability of getting a lamellar
structure. A heterogeneous austenite may result in the transformation of austenite to
spheroidized structure. With lowering of the transformation temperature, a harder and finer
aggregate of ferrite and cementite results.
The microstructures evolving during austenite decomposition are Ferrite with different
morphologies (grain boundary allotriomorphic ferrite, idiomorphic ferrite, Widmanstatten
ferrite, and intragranular ferrite) , cementite, Pearlite, Bainite and Martensite.
348
Precipitation of Ferrite from Austenite
In this topic we will be concerned with phase transformations in which the first phase to
appear is that given by the equilibrium phase diagram.
Here we will consider to the diffusional transformation of Fe-C austenite into ferrite.
However, many of the principles are quite general and have analogues in other systems
where the equilibrium phases are not preceded by the precipitation of transition phases.
Under these conditions the most important nucleation sites are grain boundaries and the
surface of inclusions.
It can be seen in figure (next slide) that ferrite can also precipitate within the austenite
grains (intragranular ferrite). Suitable heterogeneous nucleation sites are thought to be
inclusions and dislocations. These precipitates are generally equiaxed at low undercoolings
and more platelike at higher undercoolings.
In general the nucleation rate within grains will be less than on grain boundaries. Therefore,
whether or not intergranular precipitates are observed depends on the grain size of the
specimen.
In fine grained austenite for example, the ferrite that forms on grain boundaries will rapidly
raise the carbon concentration with in the middle of the grains, thereby reducing the
undercooling and making nucleation even more difficult.
In a large grained specimen, however, it takes a longer time for the carbon rejected from the
ferrite to reach the centers of the grains and meanwhile there will be time for nucleation to
occur on the less favorable intragranular sites. 349
Ferrite
40s
910oC A3
Acm
a 800oC
A3
b 750oC
723oC 150s
A1
9s
c 650oC 2s
T
d 550oC
0.15 % 0.8 %
Wt% C →
350
800oC
Ferrite nucleates on Austenite grain boundaries and
grows with a blocky appearance
→ Grain Boundary Allotriomorphs Small ∆T
Curved interfaces → incoherent

Faceted interfaces → semicoherent
150s
750oC 40s
Larger undercooling – Ferrite grows as plates from GB
→ Widmannstätten side-plates
Large ∆T
Ferrite can also nucleate within the γ grains heterogeneously at

inclusions and dislocations → equiaxed at low undercoolings
and plate-like at higher undercoolings.
351
650oC 9s
Widmannstätten side-plates → become finer with larger
undercooling
550oC 2s
Very Large undercooling – Ferrite grows as needles GB
→ Widmannstätten needles
352
Eutectoid Transformation : Bainite transformation
At relatively larger supersaturations (austenite cooled below the nose of the pearlite
transformaton), there is another eutectoid product that develops known as bainite; bainite is
also a mixture of ferrite and cementite; however, it is microstructurally quite distinct. In the
next two subsections, we discuss these microstructural features.
Upper bainite
At the higher end of the temperatures (350-550˚C), the
microstructure consists of needles of laths of ferrite
nucleated at the grain boundary and grown into one of the
grains with cementite precipitates between the ferrites; see
in Fig.
The ferrite formed is Widmanstatten; it has a Kurdjumov-
Sachs orientation relationship with the austenite grain into
which it is growing; it is in this respect, namely the
orientation relationship between the ferrite/cementite and
the austenite grain in which they grow, that the bainite
differs from pearlite.
353
Eutectoid Transformation : Bainite transformation
Lower bainite
At low enough temperatures, the bainitic microstructure
changes to that of plates of ferrite and very finely dispersed
carbides; since the diffusion of carbon is very low at these
temperatures (especially in the austenite phase as compared
to ferrite), the carbides precipitate in ferrite (and, with an
orientation relationship). These carbides that precipitate
could be the equilibrium cementite or metastable carbides
(such as ε carbide for example). A schematic of lower
bainite plate that is formed is shown in Fig.
γ → α + Fe3C**
Pearlite is nucleated by a carbide crystal, bainite is nucleated by a ferrite crystal, and
this results in a different growth pattern.
Acicular, accompanied by surface distortions
** Lower temperature → carbide could be ε carbide (hexagonal structure, 8.4% C)
Bainite plates have irrational habit planes
Ferrite in Bainite plates possess different orientation relationship relative to the parent
Austenite than does the Ferrite in Pearlite 354
Formation of Austenite
Formation of austenite is a preliminary step for any
heat treatment process.
Formation of austenite in eutectoid steel differs from
that of hypoeutectoid and hypereutectoid steels in the
sense that in the former case it occurs at a particular
temperature (AC1) where as for the latter it takes place
a over range of temperature.
At 1147°C, the maximum solubility of carbon in FCC
iron is about 2 percent. Therefore, at this temperature,
regions around the cementite layer will be enriched
with carbon because of diffusion.
355
Formation of Austenite
The maximum diffusion of carbon atoms will take place from the cementite at the ferrite-
cementite interface. As sufficient number of interfaces are available, austenite nuclei will be
formed at the interface.
By gradual dissolution of carbon of cementite into gamma iron, primary austenite/grains are
formed. These primary austenitic grains dissolve the surrounding ferrite and austenitic
grains grow at the expense of ferrite.
The growth rate of austenite is higher than the rate of dissolution of the cementite into
austenite because austenitic grain growth takes place because of the transformation of alpha
iron to gamma iron and diffusion of carbon atoms from austenite to ferrite. This explains the
experimentally observed fact that dissolving of ferrite is completed before that of cementite.
The austenite thus formed at eutectoid temperature
is not homogeneous. The carbon concentration is
higher in these regions which are adjacent to the
original cementite lamellae than those which are
adjacent to the ferrite mass.
Chemically homogeneous austenitic grains are
obtained by holding (holding time should be
sufficient to diffuse properly) steel above the
eutectoid temperature
Effect of temperature and time on the austenite
formed from pearlite in 0.3% carbon356steel
Kinetics formation of Austenite
The formation of austenite on heating always occurs at a temperature higher than that
predicted by the Fe-Cementite phase diagram.
To study the kinetics of austenite formation is to heat a number of steel samples to different
temperatures above the eutectoid temperature.
Heating is done by immersing samples in constant temperature baths. A number of samples
are immersed in a constant temperature bath and are taken out one by one after a definite
interval of time followed by immediate quenching which will result in the formation of
austenite from transformed austenite.
The amount of martensite formed will depend on the amount of transformed austenite
which in turn will depend on the temperature at which the steel sample has been heated and
the holding time at that temperature.
From the figure, we can concluded
that the lower transformation
temperature, the more is the time
required to complete the
transformation
Any property which changes with the
formation of austenite, and/or with
martensite formed by quenching of
transformed austenite, can be utilized Transformation of pearlite to austenite as a function of time
357
this purpose
The relationship between transformation temperature and transformation time has been
derived by taking into consideration the effect of superheating. An analysis of figure leads
to the following conclusions:
Transformation is completed in a short
period at high transformation temperature.
For Higher heating rates, transformation
will start at higher temperature, whereas for
slower heating rate, transformation will
start at lower temperature.
For any given practical rate of heating , the
formation of austenite will occur over a
Effect of temperature on the time required for start and
range of temperature, and not at constant completion of transformation of pearlite to austenite
temperature.
For the formation of austenite at a constant temperature, the heating rate should be
extremely slow, and the two curved lines will converge to a single point located at eutectoid
temperature line.
The end of transformation curve does not reveal any information about the nature
(homogeneity) of austenite. The curved line ensures that all the pearlite has been
transformed into austenite. In order to attain homogeneous austenite, the steel has to be
heated to still higher temperatures. 358
The processes of austenite formation on heating proceeds by nucleation and growth
reaction. Therefore, the factors which can vary either the rate of nucleation or the rate of
growth or both will change the kinetics of austenite formation. Two such parameters are
transformation temperature and holding time at transformation temperature.
The kinetics of austenite transformation is governed to great extent, by the nature of
pearlite. The number of possible austenite nuclei will increase with the increase in
interfacial area. The interfacial area can be increased in two ways: by increasing the
cementite contents, and by decreasing the interlamellar spacing. This is the reason why high
carbon steels austenitize more rapidly than low carbon steels.
The closer the ferrite-cementite lamellae, the higher will be the rate of nucleation. Also, the
carbon atoms have to diffuse for smaller distances in order to enrich low carbon regions.
Therefore, the rate of growth of primary austenitic grains will also be higher in this case .
This explains why pearlitic structure with less interlamellar spacing is transformed faster to
austenite.
The kinetics of austenite transformation for coarse pearlitic structure is slow for the reason
given above. The kinetics of austenitic transformation from granular pearlite is slower than
that of lamellar pearlite for same reasons.
The kinetics of transformation will further decrease with increase in the size of globular
cementite particles. Quenched structure will also transform to austenite more rapidly than
the granular pearlitic structure.
359
Austenitic Grain Size
The size of austenitic grains is the most important structural characteristic of heated steel. The
austenitic grain size strongly effects its own transformation behavior and the mechanical
properties of the microstructures formed from austenite.
Austenitic grain boundaries are preferred sites for the nucleation of pro-eutectoid phases and
pearlite, which are diffusion controlled transformation products.
Coarse austenitic grains having less grain boundary area, have fewer nucleating sites, thus
diffusion-controlled transformation of austenite is retarded paving way for easy transformation to
martensite (Diffusionless transformation product).
The impact toughness of steel is most sensitive to the size of the austenite grains in the hardened
and tempered state. The charpy impact value improves with decreasing austenitic grain size to the
extent that its value for a fine-grained steel can exceed several times that of a coarse grained steel
of the same grade.
The reason, partly is due to segregation of impurity atoms to the austenitic grain boundaries
during Austenitisation (more segregation takes place if grain boundary area is less as is the case
in coarse grained steels), and thus the fracture frequently takes place along prior austenitic grain
boundaries.
The co-segregation of impurities like Sb, P, Sn, As along the large angle grain boundaries of
austenitic grains weakens the adhesion at these boundaries to cause fracture along them. Such
intergranular fracture is quite brittle.
In alloys of Fe-Ni, and Fe-Ni-C, the Ms temperature is lowered significantly by decreasing the
austenitic grain size, probably due to higher strength of fine grained austenite , which in turn
increases the shear resistance of austenite to transform martensite. 360
At temperatures just above the upper
critical temperature, when the structure is
fully austenitic, that is just after the
transformation is complete, the initial grain
size of austenite is fine (though more
disperse is the initial structure, the finer is
the austenite grain formed).
As the temperature is raised further, or
holding time at a given temperature is
increased, then grain coarsening occurs.
This behavior is schematically illustrated.
Austenitic grain growth is a natural
spontaneous process and is caused by the
tendency to reduce the surface energy by
reducing the total surface area of the grain
boundaries. A high temperature accelerates
the rate of this processes.
Driving force is the surface energy stored
as grain boundary energy. Certain grains
grow at the expense of smaller grains,
which due to their less stable. 361
Schematic illustration of changes in microstructure of steel on heating
Depending on the tendency of steel to grain
growth, steels can be classified into two broad
groups:
Inherently fine grained steels
Inherently coarse grained steels
Inherently fine grained steel resists the growth
of austenitic grains with increasing
temperature. The kinetics of austenite grain
growth is very slow and the steel remains fine Comparison of austenitic grain growth characteristics
grained even at temperatures as high as of inherently fine grained and course grained eutectoid
steel
1000°C or 1050°C.
On the other hand, grains of inherently coarse grained steel grow abruptly with increasing
temperature, and so a steel with coarse grains is obtained.
In the presence of ultramicroscopic particles of oxides, carbides and nitrides which prevent
grain growth. These particles are refractory in nature, i.e., they can withstand high
temperatures and are generally present at grain boundaries of austenite. Therefore, they act
as barriers to the growth of austenitic grains.
Intensive grain coarsening of inherently fine grained steels above a particular temperature is
due to the dissolution of ultrafine particles of carbides, oxides and nitrides in the matrix
(austenite). After dissolution, no particles are left to offer resistance to grain growth. 362
Cellular Precipitation (Discontinuous Precipitation)
Grain boundary precipitation may also occur by cellular precipitation (apart from
Allotriomorphs, Widmannstätten side plates & needles)
Main Feature → boundary moves with the tip of the precipitates
The mechanism by which GB ppt. develops into cellular precipitation is system specific
called discontinuous precipitation as the composition of the matrix changes discontinuously
as the cell front passes
(Continuous precipitation- not cellular- occurs throughout the matrix and the matrix
composition changes continuously with time
α ' (higher solute concentration) → α (lower solute concentration) + β
363
In the previous section, we discussed the effect of grain boundaries on the nucleation of
precipitates with specific reference to ferrite (α) on austenite (γ). We saw that there exist
two morphologies, namely, grain boundary allotriomorphs and Widmanstatten side plates.
In some cases, the precipitates that are nucleated at the grain boundary, when they grow,
also carry the grain boundary along with them. This process is shown schematically in Fig.
Such a process leads to a cellular microstructure as shown. This kind of process is known as
cellular precipitation.
Partitioning of the solute ahead of

the moving front by GB diffusion
α ' (higher solute concentration) → α (lower solute concentration) + β 364

Cellular precipitation is also known discontinuous precipitation because of the sudden

change in composition of matrix across the moving grain boundary as shown below; on the
other hand, noncellular precipitation is known as continuous because the composition of the
matrix phase decreases continuously at any point. In general, continuous precipitation leads
to much better mechanical properties (due to the more uniform distribution of precipitates,
which nucleate throughout the matrix (on dislocations for example) and hence, much
smaller size distribution) as opposed to discontinuous precipitation.
Cellular precipitation leads to microstructures that are very similar to eutectoid
transformations. In fact, the only difference between cellular precipitation and eutectoid
transformation is that while in eutectoid transformation both the phases that form are
different from the original phase, in cellular precipitation one of the phases remains the
same as the original phase (albeit with a different composition):
α' →α + β
α’ is the supersaturated matrix, and is the matrix phase with a composition closer to
equilibrium as compared to α ’ and β is the precipitate phase.
365
Original position of GB
New position of GB
Grain-II
Grain-I
The cellular precipitation that takes place The composition profiles across the
during the aging treatment leads to boundary as well as parallel to the
microstructures boundary
366
Consider the phase diagram of the Al-Mg system, The β phase is the Mg17 Al12 phase and
the phase is the solid solution of Mg and Al. Consider an alloy of 9 at % Al, which is
solution treated (at, say 410oC) and aged for an hour at 220oC.
The cellular precipitation that takes place during the aging treatment leads to
microstructures as shown schematically in Fig. (next slide).
The composition profiles across the boundary as well as parallel to the boundary are as
shown in Fig. (next slide).
The discontinuous nature of the composition in the phase across the moving boundary
indicates that the mechanism of diffusion which leads to the formation of cellular structures
is the diffusion of solutes through the moving grain boundary (since, if the diffusion took
place through the matrix on either side of the boundary, it would have led to composition
gradients).
The composition of the α in the profile parallel to the moving boundary indicates that
composition is still not equilibrated after the precipitation of β.
When the mechanism of phase transformation changes from nucleation and growth to
spinodal decomposition, it is possible for discontinuous spinodal to take place; in such
cases, phase separation starts near the grain boundary and the boundary moves with the
phase separation (reaction) front.
367
Ordering Transformations
In the thermodynamics & kinetics topic, we considered a system with Ω > 0. In such
systems, at lower temperatures, the free energy develops a concave curvature leading to a
phase separation into mechanical mixture.
Now, consider a system with , that is, the AB bonds are preferred over AA/BB bonds. What
happens to such systems at low temperatures?
In systems with Ω < 0, that is, systems in which AB bonds are preferred over AA/BB bonds
are preferred, at lower temperatures, the system becomes ordered. This ordering is over and
above the crystallographic ordering. In the crystalline lattice, specific lattice positions are
occupied by specific atoms, thus leading to more of the preferred unlike bonds.
Consider for example a bcc lattice occupied by A and B atoms. If it is disordered, then, the
probability of the cube corners and cube centers are occupied by the A or B atoms is 50%
(that is, the same as the alloy composition in at%). However, when this system orders, the
cube corners preferentially occupy one of the positions, say, cube corners while the other
preferentially occupies the cube centers.
That is, the bcc lattice now can be considered to be consisting of two interpenetrating cubic
lattices. Such a structure is known as B2. Notice that in the (ideal) B2 structure, there are
only AB bonds and no AA/BB bonds. NiAl is a system in which, for example, such B2
ordered structure is known. There are also fcc based ordered structures such as L12
(example: Ni3Al) and L10 (example: CuAu).
368
Ordering Transformations
A−B bonds stronger A−A and B−B bonds stronger
than A−A and B−B bonds than A−B bonds
Solid stabilized → Ordered solid Liquid stabilized → Phase
formation separation in the solid state
L+ α L L
L+ α
E.g. Au-Ni
α
α’ α1 α2
A B
α + α’ A α1 + α2 B
Ordered solid α1 & α2 are different only in lattice parameter
369
Variations to the isomorphous phase diagram
Ordered Phases
Orderd Random Short Range Orderd
Part of the Cu-Au phase diagram showing the regions where

370
the Cu3Au and CuAu superlatices are stable.
Ordered Phases
371
Long range and short range order parameters
There are two different order parameters that one can define. The first is the short range
order parameter (S) defined as :
PAB − P AB (random)
S=
PAB (max) − PAB (random)
Here, PAB(max)and PAB(random) refer to the maximum number of bonds that can exist and
the number of bonds in a random solution. As one can see, the short range order parameter
is closely related to the, regular solution parameter; this type of ordering is possible in
alloys of all compositions.
In case the alloy composition is in simple ratios of the constituent atoms, then, one can
define the so called long range order parameter L in terms of the alloy composition as x
given as:
p−x
L=
1− x
Where p is the probability of occupancy of the given site by the right kind of atom.
At absolute zero, the system will choose a state with L = 1; however, as temperature
increases, the effects of configurational entropy come into play; so the value of L decreases
from unity and eventually reaches zero. The temperature at which this change of Long
Range Order (LRO) from unity to zero takes place is known as the critical temperature (Tc )
for the order-disorder transformation. 372
Long range and short range order parameters
In Fig. we show the variation of L and S with temperature in two systems, namely, one that
undergoes an order-disorder transformation from B2 to disordered bcc and another that
undergoes an order-disorder transformation from L12 to disordered fcc.
It is clear from the figures, the changes are of two different types; in the equiatomic case of
B2 to bcc (CuZn type), the variation is continuous; however, in the case of L12 to fcc
(Cu3Au type), the variation is abrupt. These differences in the behavior is a consequence of
the differences in atomic configurations in the two ordered lattices.
Order parameter variation with temperature: continuous (B2 type) and abrupt (L1
3732).
Ordering Transformations : Microstructural features
The order-disorder transformation can take place both through the nucleation and growth
mechanism and spinodal mechanisms. In spinodal mechanism there is continuous increase
in, homogeneously all through the crystal leading to the transformation. In the nucleation
and growth mechanism, small regions form overcoming an energy barrier and these regions
grow.
In ordered alloys, the two phases have near-identical lattice parameters and the interfacial
energies between the ordered and disordered phases is very low. Hence, the barrier for
nucleation is very small. Hence, order-disorder transformation, when takes place through
the nucleation and growth mechanism, takes place through homogeneous nucleation.
Occurring homogenously throughout the crystal

Continuous increase in SRO
Mechanism
Nucleation and Growth
Due to an energy barrier to the formation of ordered domains
374
Antiphase boundary
In ordered alloys, there exist surface defects which are known as antiphase boundaries. This
is because the two different ordered domains formed at different regions, when impinge, it
can so happen that there is no matching across the interface; see the schematic in Figure.
The different ordered domains are possible because in B2 for example, the two sites are
equivalent and hence what is cube center in one domain could be the cube corner in
another.
In the schematic shown in the Fig., for example, in the left upper portion, the cube centers
are occupied by Ni while in the lower right portion cube centers are occupied by Al. So,
when these two domains meet an antiphase boundary is formed. Since the antiphase
boundary consists of NiNi and AlAl bonds, they increase the energy of the system. Hence,
given a chance the system would like to get rid of them.
375
The schematic of the formation of antiphase boundaries in an ordered alloy due to the accidents of growth.
The type of transformation that occurs in cooling (e.g. Cu-38 at% Zn alloy) depends on
the cooling rate
Slow cooling → transformation at small undercoolings ⇒ equiaxed α
Faster cooling → transformation at lower T ⇒ Widmannstätten α needles
Even faster cooling (brine quench) → no time for diffusion ⇒ Transformation of β
→α without change in composition (MASSIVE TRANSFORMATION)
α richer in Cu ⇒ growth of α requires long-range diffusion of Zn away from advancing
α/β interface
MASSIVE Tr.
Nucleation at GB and rapid growth into β
Irregular appearance
No change in composition ⇒ Only short range jumps (thermally activated) across the
interface (α/β) (no long range diffusion required)→ fast growth rate
(thermally activated migration of incoherent interfaces → diffusionless civilian)
Separate transformation curve (in TTT & CCT diagrams)
~ to GB migration during recrystallization → but driving force v.high
376
An example of massive transformation
is shown in Fig.: in Cu-Zn system, Cu-Zn System
when an alloy of 38 at.% Zn is cooled
from 850oC to say 400oC at fast β
enough cooling rates, the structure
changes from that of β to α albeit with
the same composition. Massive
Since the composition is the same,
there is no need for long range
diffusion and hence the transformation
is very fast.
α
Such changes in structure without
changes in composition can be
achieved in two ways: massive which
is through thermally activated jumps
of atoms from regions of one phase to
another and martensitic which is
through diffusionless (military)
377
Massive
378
Typically, systems which undergo massive transformation also undergo martensitic
transformation at higher cooling rates; this is indicated in the schematic CCT diagram in
Fig. The mechanism of formation of massive transformation can be understood with
reference to this CCT diagram.
At slow cooling rates and at
smaller undercoolings,
precipitation and growth of
α leads to equiaxed α. At
higher cooling rates and
larger undercoolings
Widmanstatten α formation
takes place.
In both these cases, since
the growth of α requires
long range diffusion (see
Cu-Zn phase diagram for
example, where it is the
long range diffusion of Zn
that is needed), it requires
long time to form. 379
If Ms below RT → metastable β
However, relatively faster cooling rates would nucleate the α phase at the grain boundaries;
since the growth of this α only requires that atoms jump across the α -β interface, and since
the driving forces for the formation of α are very large (see in Fig.)
380
Free energy composition curves (schematic) for α and β in the Cu-Zn system
V.High cooling rate will

produce Martensite
Massive Tr. can also occur in Fe-C system
381
Questions?
1. What is the condition for the development of a miscibility gap in an isomorphous phase
diagram?
2. Where do the eutectic and peritectic reactions get the driving force for the progress of the
reaction?
3. Show that coupled growth (growth of α and β together as is observed in case of eutectic
solidification) is not possible in peritectic solidification.
4. What is the need for the formation of coherent precipitates at high undercooling?
5. What would be the positions of the GP zones, θ”, θ’ solvus lines in relation to that of θ in
Al-Cu phase diagram? Explain based on thermodynamic principles.
6. How do you know the driving force for the nucleation during precipitation of a second
phase from a supersaturated solid solution using the G-X diagram?
7. Identify the differences between the composition outside and inside the spinodal in terms of
(a) driving force for the transformation and (b) the progress of transformation.
8. Usually massive transformations are observed in single phase field instead of a two phase
field. Why?
9. Why do massive transformations generally occur at lower temperatures but higher rates than
precipitation transformations.
10. How does pressure influence the austenite-martensite transformation in steels?
382
Questions?
11. Define interlamellar spacing. Is it characteristic of a temperature of transformation? How is it
that it appears different in different colonies in a microstructure?
a) Derive an expression to prove that growth rate of pearlite is proportional to square of the
amount of under cooling. Why does the growth rate decrease after becoming a maximum
with the fall of temperature
b) What is the kinetics of pearlitic transformation? How is the volume rate of pearlite
formation related to growth rate and grain size of austenite?
12. Why does the flow stress increase as the carbon content of annealed steels increases?
13. Distinguish between two types of bainites.
a) Discuss at least four characteristics of bainitic transformation
b) Compare the characteristics of bainitic transformation with pearlitic and with martensitic
reaction
c) Discuss the mechanism of formation of bainites.
d) What makes lower bainite to have more applications than upper bainite?
14. Explain the term active nucleus. Name the active nucleus for (i) Pearlitic transformation and
(ii) bainitic transformation. Give reasons in support of your answer.
15. What is cellular precipitation? What is the mechanism involved in it. How it is different from
Precipitation hardening.
16. What is meant by ordering parameter? How it influences the material property? Define SRO
383
and LRO.
Questions?
17. What is super lattice? Give some of examples. And what are the criteria to form a super
lattice. What are the techniques available to find out a super lattice?
18. Write short note on Spinodal decomposition. Compare with Nucleation and Growth.
19. Explain the thermodynamics and kinetics of solid state phase transformation with reference
to the following factors;
Volume free energy, surface energy and strain energy components.
Why does solid state phase transformation generally require under cooling to start
nucleation?
20. Explain the Hull-Mehl mechanism of pearlitic transformation by citing Hultgren’s
extrapolation technique.
What is Johnson-Mehl equation of kinetics of pearlitic transformation
Explain the parameters affecting inter-lamellar spacing of pearlitic structure and also states
its roles on mechanical properties.
384
Avala Lava Kumar*
Martensitic transformations - Introduction
The name martensite is after the German scientist Martens. It was used originally to
describe the hard microconstituent found in quenched steels. Martensite remains of the
greatest technological importance in steels where it can confer an outstanding combination
of strength (> 3500 MPa) and toughness (> 200 Mpa/m2 ).
Martensitic transformation is the name for any transformation that takes place in a
diffusionless and military manner - that is, these transformations take place through atomic
movements which are less than one atomic spacing; and in these transformations atoms
change their positions in a coordinated manner (unlike thermally activated diffusional, or,
so-called, civilian processes).
In shape memory alloys such as Ni-Ti (nitinol), it is the martensitic transformation that is
responsible for the shape memory effect.
In this topic, we describe some characteristic features of the martensitic transformations
(with specific reference to steels in which, this transformation is repsonsible for hardening
by quenching).
Since martensitic transformations are diffusionless, necessarily, the composition does not
change during the transformation. It is only the crystal structure that changes. For example,
in Fe-C alloys, the austenite (fcc) transforms into martensite (bct);
in Ni-Ti, an ordered bcc (called austenite) transforms to another ordered CsCl type structure
(called martensite). Note that since martensitic transformation is diffusionless, if the
austenitic phase is ordered, the martensitic phase is also ordered. 386
Characteristics of martensitic transformation
Occur at high super saturations, without change in composition
No long range diffusion → Movement of atoms a fraction of the inter-atomic distance
Nearest neighbour configuration remains unchanged
Cooperative movement of a large group of atoms → transformation proceeds at the speed of
sound in the material
Thermal activation does not play a role in the growth of the martensitic phase
Thermal activation may play a role in the nucleation of martensite
Usually martensite crystals nucleate and grow across the whole grain (in some cases this
growth is a function of temperature)
In some cases the transformation occurs over a period of some time (due to thermally
assisted nucleation)
Martensitic crystals have a specific orientation relationship (OR) with the parent phase
Planes of the parent lattice on which Martensitic crystals from → Habit planes (irrational
Indices)
Interface between Martensite and parent phase → Coherent or Semi-coherent
At a given temperature (between Ms and Mf) the fraction transformed ↑ with plastic
deformation (in some cases elastic stress also has a similar effect)
With prior plastic deformation the transformation temperature can be increased to Md
387
Shape of the Martensite formed → Lathe and Plate morphologies

Associated with shape change (shear)
But: Invariant plane strain (observed experimentally) → Interface plane between Martensite and
Parent remains undistorted and unrotated
This condition requires
1) Bain distortion → Expansion or contraction of the lattice along certain
crystallographic directions leading to homogenous pure dilation
2) Secondary Shear Distortion → Slip or twinning
3) Rigid Body rotation
388
Martensitic transformation can be understood by first considering an alternate unit cell for
the Austenite phase as shown in the figure below.
If there is no carbon in the Austenite (as in the schematic below), then the Martensitic
transformation can be understood as a ~20% contraction along the c-axis and a ~12%
expansion of the a-axis → accompanied by no volume change and the resultant structure
has a BCC lattice (the usual BCC-Fe) → c/a ratio of 1.0.
γ ( FCC ) α ' ( BCT )
Quench
→
0.8 %C 0.8 %C
In Pure Fe after
FCC Austenite alternate choice of Cell the Martensitic transformation
c=a
FCC → BCC
~20% contraction of c-axis

488 ~12% expansion of a-axis
But shear will distort the lattice!
Slip Twinning
Average shape
remains undistorted
390
How does the motion of dislocations lead to a macroscopic shape change?

(From microscopic slip to macroscopic deformation a first feel!)
Step formed
Dislocation
when dislocation
formed by
leaves the crystal
pushing in
a plane r
b
391
392
Net shape change
393
Properties of 0.8% C steel

Constituent Hardness (Rc) Tensile strength (MN / m2)
Coarse pearlite 16 710
Fine pearlite 30 990
Bainite 45 1470
Martensite 65 -
Martensite tempered at 250 oC 55 1990 394
0.95 1.90 2.85 3.80 4.75 5.70 6.65
3.70
3.66
1000
3.62 a – parameter of fcc austenite
Temperature oF →
Lattice parameter, Ao
800
3.58
3.54 600
Ms
3.50
Mf
400
3.02 c – parameter of bct martensite 200
2.98
0.4 0.8 0.12 0.16
2.94 % Carbon →
2.90 Influence of carbon on the
a – parameter of bct martensite
2.86
martensite range
2.82
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
% Carbon →
Effect carbon on lattice parameters of
austenite and martensite 395
Driving force for martensitic transformation
Diffusionless transformations require larger driving forces than for diffusional

transformations
Why? In order for a transformation to occur without long range diffusion, it must take
place without a change in composition
This leads to the so-called T0 concept, which is the temperature at which the new phase can
appear with a net decrease in free energy at the same composition as the parent (matrix)
phase.
As the following diagram demonstrates, the temperature, T0, at which segregation-less
transformation becomes possible (i.e. a decrease in free energy would occur), is always less
than the solvus (liquidus) temperature.
The driving force for a martensitic transformation can be estimated in exactly the same way
as for other transformations such as solidification.
Provided that an enthalpy (latent heat of transformation) is known for the transformation,
the driving force can be estimated as proportional to the latent heat and the undercooling
below T0.
γ →α ' γ →α ' T0 − M S
∆G = ∆H
. T0
396
397
Driving force for martensitic transformation
α product
∆Gg→a
T1
γ parent
G Common tangent
∆Gg→a
T1>T2
T2
Diffusionless transformation impossible at T1,

Diffusionless transformation possible at T2;
“T0” is defined by no difference in free X

energy between the phases, ∆G=0.
Various ways of showing Martensitic Transformations
(b)
(a)
(c)
equilibrium diffusionless
(d)
(a) Free energy – temperature diagram for Austenite and martensite of

fixed carbon concentration
(b) Free energy – composition diagram for the austenite and martensite
phases at the Ms temperature.
(c) Iron-carbon phase diagram with T0 as defined in (a), Ms and Mf
superimposed.
(d) Ms and Mf in relation to the TTT diagram for alloy C0 in C
398
Why tetragonal Fe-C martensite?
At this point, it is worth stopping to ask why a tetragonal martensite forms in iron. The
answer has to do with the preferred site for carbon as an interstitial impurity in bcc Fe.
Remember: Fe-C martensites are unusual for being so strong (& brittle). Most martensites
are not significantly stronger than their parent phases.
Interstitial sites:
FCC: octahedral sites radius= 0.052 nm
tetrahedral sites radius= 0.028 nm
Carbon atom radius = 0.08 nm.
BCC: octahedral sites radius= 0.019 nm
tetrahedral sites radius= 0.036 nm
Surprisingly, it occupies the octahedral site in the bcc Fe structure, despite the smaller size
of this site (compared to the tetrahedral sites) presumably because of the low modulus in the
<100> directions.
One consequence of the occupation of the octahedral site in ferrite is that the carbon atom
has only two nearest neighbors. Each carbon atom therefore distorts the iron lattice in its
vicinity. The distortion is a tetragonal distortion.
If all the carbon atoms occupy the same type of site then the entire lattice becomes
tetragonal, as in the martensitic structure.
Switching of the carbon atom between adjacent sites leads to strong internal friction peaks
at characteristic temperatures and frequencies.
399
Massive vs. Martensitic Transformations
There are two basic types of diffusionless transformations.
One is the massive transformation. In this type, a diffusionless transformation takes place
without a definite orientation relationship. The interphase boundary (between parent and
product phases) migrates so as to allow the new phase to grow. It is, however, a civilian
transformation because the atoms move individually.
The other is the martensitic transformation. In this type, the change in phase involves a
definite orientation relationship because the atoms have to move in a coordinated manner.
There is always a change in shape which means that there is a strain associated with the
transformation. The strain is a general one, meaning that all six (independent) coefficients
can be different.
CIVILIAN MILITARY
Diffusion Required Precipitation, Spinodal
-
Decomposition
Diffusionless Massive Transformations Martensitic Transformations
400
Morphology of Martensite
Formation of (a)
Lath martensite (b),
(c) plate martensite.
(a) Schematic
illustration of mode
of nucleation and
growth of lathe
martensite. (b) Lens
Shaped martensite
(c) Successive
partioning of
austenite grain with
the formation of
plates of martensite
In steels, two distinct morphologies of martensite are seen Lathe Martensite and Plate
(Lenticular) Martensite.
Low carbon martensite – Lathe martensite
Medium carbon martensite – Mixed martensite –lathe as well as plate martensite
High carbon martensite – Lenticular plates
401
Morphology of Martensite
(a) (b) (c)
(d) (e)
Low C (lathe)
(a, b) Growth of martensite with

increasing cooling below Ms
(c-e) Different martensite
morphologies in iron alloys
Medium C (plate) Fe-Ni: Plate 402

Kinetics of Martensitic transformations
Athermal
Fraction transformed = f(T) only (between Ms and Mf)
At a given T a certain number of nuclei form which grow at
the speed of sound in the material across the whole grain
The number of nuclei forming increases with decreasing T
⇒ more transformation
Ms and Mf temperatures increase with increasing grain size
Burst
At some T (Mb) → sudden transformation of a large fraction

of parent to Martensite (often with a acoustic effect-click)
% transformed during the “burst” → few-over 50%
The ‘burst’ can be thought of as an extreme form of
autocatalytic effect
Remaining transformation is completed at → lower T
longer t
403
Kinetics of Martensitic transformations
Isothermal
Fraction transformed at a given T (between Ms and Mf) = f(t)
→ observed in some alloys at subzero T
At the low T thermal activation of growth does not occur (as usual fast)
“t effect” is due to small thermal activation required for nucleation of Martensitic plates and
autocatalytic effect
Rate of transformation at a given T ↑ with ↑ grain size (fewer nucleation events give more
transformed fraction)
Isothermal Transformation diagram for

Martensitic Transformation in Fe-Ni-Mn alloy 404
Questions?
1. Differentiate between the following transformations
Pearlite, (b) Bainitic (c) Martensitic, based on Diffusion, Temperature range, Nuclei,
Kinetics of reaction, Crystallography, Density of dislocation, Surface relief, and Growth rate.
2. Where the carbon atoms sitting in BCT martensite
a) Why does martensite has BCT structure?
b) Why martensitic transformation is not reversible in steels?
c) Slip in the cause of lattice invariant deformation in low carbon martensite, while
twinning is the cause in high carbon martensite. What is the relationship as the
temperature of transformation drops
3. Give an exact definition of the habit plane of martensite. Describe how this habit plane might
be measured experimentally. Give possible reasons why there is so much scatter of habit
plane measurements in given sample.
4. Draw a diagram to illustrate Bain’s homogeneous deformation model for the fcc → tbcc
diffusionless transformation. Assuming αγ = 3.56 Ao and αα = 2.86 Ao, and that c/a for
martensite is 1.15 calculate the maximum movement experienced by atoms during the
transformation. Assume that c/a = 1.1.
5. What is the role of austenitic grain size in martensitic transformations? Is austenitic grain
size is important to the strength of martensite? What other factors are important to the
strength and toughness in technological hardened steels?
405
Questions?
6. Calculate the volume change associated with austenite to martensite transformation in 1 %

carbon steel
7. Explain why martensite is hard? What is the crystal structure of martensite? Show the
positions of carbon atoms in unit cell of martensite.
8. What is the impact of tetragonal expansion (due to carbon) on the locking of dislocations.
9. Explain the role of applied stress on martensitic transformation and also enumerate a
thermal and isothermal kinetics of martensite formation.
10. Classify martensite based on the morphology and what is the effect of carbon on it?
406
Avala Lava Kumar
407
Classification
HEAT TREATMENT
BULK SURFACE
Annealing Tempering
Normalizing Hardening Thermal Thermo-Chemical
Full Annealing Flame Carburizing
Recrystallization Annealing Nitriding

Induction
Stress Relief Annealing Cyaniding
Laser
Spheroidization Annealing Boronizing
Electron Beam
Diffusion Annealing Chromizing
Partial Annealing
Process Annealing 408

Annealing
Full Annealing
The steel is heated above A3 (for hypo-eutectoid steels) | A1 (for hyper-eutectoid steels)
→(hold) →then the steel is furnace cooled to obtain Coarse Pearlite
Coarse Pearlite has ↓ Hardness, ↑ Ductility
Not above Acm →to avoid a continuous network of proeutectoid cementite along grain
boundaries (→path for crack propagation)
No
910oC Ful
rma
l i za ation
lA tion a liz Acm
nne r m
A3 alin
g No
Full Annealing
723oC
Spheroidization A1
Recrystallization Annealing
Stress Relief Annealing
T
0.8 %
Wt% C 409
Annealing
The Heat below A1 → Sufficient time → Recrystallization
Cold worked grains → New stress free grains

Used in between processing steps (e.g. Sheet Rolling)
No
910oC Ful
rma
liza ation
lA tion iz
al Acm
nne r m
A3 alin
g No
Full Annealing
723oC
Spheroidization A1
T
0.8 %
Wt% C 410
Annealing
Residual stresses → Heat below A1 → Recovery
→ Differential cooling Annihilation of dislocations,

polygonization
→ Martensite formation
No n
rma tio
→ Machining and cold working 910oC Ful
lA
nne
liza
tion
mali
za
Acm
r
A3 aling No
Full Annealing
→ Welding 723oC
Spheroidization A1
T
0.8 %
Wt% C
411
Annealing
Spheroidization Annealing
Heat below/above A1 (Prolonged holding*)

Cementite plates → Cementite spheroids → ↑ Ductility
Used in high carbon steel requiring extensive machining prior to final hardening and
tempering
Driving force is the reduction in interfacial energy
The spheroidized structure is desirable when minimum hardness, maximum ductility, or(in
high-carbon steels) maximum machinability is important.
Low-carbon steels are seldom spheroidized for machining, because in the spheroidized
condition they are excessively soft and “gummy”.
Medium-carbon steels are sometimes spheroidization annealed to obtain maximum
ductility.
*If the steel is kept too long at the spheroidized-annealing temperature, the cementite
particles will coalesce and become elongated thus reducing machinability 412
Annealing
Diffusion Annealing
This process also known as homogenizing annealing, is employed to remove any structural
non-uniformity.
Dendrites, columnar grains and chemical inhomogeneities are generally observed in the
case of ingots, heavy plain carbon steel casting, and high alloy steel castings. These defects
promote brittleness and reduce ductility and toughness of steel.
In diffusion annealing treatment, steel is heated sufficiently above the upper critical
temperature (say, 1000-1200°C), and is held at this temperature for prolonged periods,
usually 10-20 hours, followed by slow cooling.
Segregated zones are eliminated and a chemically homogeneous coarse grain steel is
obtained by this treatment as a result of diffusion.
The coarse grained structure can be refined either by plastic working for ingots or by
employing a second heat treatment for castings.
Hypoeutectoid and eutectoid steel castings are given full annealing treatment, whereas
hypereutectoid steel castings are either normalized or partially annealed for this purpose.
413
Annealing
Partial Annealing
Partial annealing is also referred to as intercritical annealing or incomplete annealing. In this
process, steel is heated between the A1 and the A3 or Acm. It is followed by slow cooling.
Generally, hypereutectoid steels are subjected to this treatment. Resultant microstructure
consists of fine pearlite and cementite. The reason for this is that grain refinement takes
place at a temperature of about 10 to 30°C above Ac1 for hypereutectoid steels.
As low temperature are involved in this process, so it is cost effective than full annealing
Process Annealing
In this treatment steel is heated to a temperature below the lower critical temperature, and is
held at this temperature for sufficient time and then cooled. Since it is a subcritical
annealing, cooling rate is of little importance
The purpose of this treatment is to reduce hardness and to increase ductility of cold-worked
steel so that further working may be carried out easily. It is an Intermediate operation and is
sometimes referred to as in process annealing. The process is less expensive than
Recrystallization annealing.
414
Annealing
Diffusion
annealing
Partial
annealing
Process
annealing
415
Temperature ranges for various types of annealing processes
Normalizing
Heat above A3 | Acm → Austenization → Air cooling → Fine Pearlite ((Higher
Higher hardness)
No
rma on
910oC Ful liza izati
Refine grain structure prior to hardening lA
nne tion al Acm
a o rm
A3 ling N
Full Annealing
723oC
A1
Purposes To harden the steel slightly Spheroidization
To reduce segregation in casting or forgings T
0.8 %
Wt% C
In hypo-eutectoid steels normalizing is done 50oC above the annealing temperature
In hyper-eutectoid steels normalizing done above Acm → due to faster cooling
cementite does not form a continuous film along GB
Annealed Vs Normalized
Annealed Normalized
Less hardness, tensile strength and toughness Slightly more hardness, tensile strength and toughness
Pearlite is coarse and usually gets resolved by Pearlite is fine and usually appears unresolved with
the optical microscope optical microscope
Grain size distribution is more uniform Grain size distribution is slightly less uniform
Internal stresses are least Internal stresses are slightly more 416
Hardening
Heat above A3 | Acm → Austenization → Quench (higher than critical cooling rate)
Certain applications demand high tensile strength and hardness values so that the
components may be successfully used for heavy duty purposes. High tensile strength and
hardness values can be obtained by a processes known as Hardening.
hardening process consists of four steps. The first step involves heating the steel to above
A3 temperature for hypoeutectoid steels and above A1 temperature for hypereutectoid steels
by 500C.
The second step involves holding the steel components for sufficient soaking time for
homogeneous austenization.
The third step involves cooling of hot steel components at a rate just exceeding the critical
cooling rate of the steel to room temperature or below room temperature.
The final step involves the tempering of the martensite to achieve the desired hardness.
Detailed explanation about tempering is given in the subsequent sections. In this hardening
process, the austenite transforms to martensite. This martensite structure improves the
hardness.
In the hardening process, which involves quenching and tempering. During quenching outer
surface is cooled quicker than the center. In other words the transformation of the austenite
is proceeding at different rates. Hence there is a limit to the overall size of the part in this
hardening process. 417
Hardening
A few salient features in hardening of steel
Proper quenching medium should be used such that the component gets cooled at a rate just
exceeding the critical cooling rate of that steel.
Alloy steels have less critical cooling rate and hence some of the alloy steels can be
hardened by simple air cooling.
High carbon steels have slightly more critical cooling rate and has to be hardened by oil
quenching.
Medium carbon steels have still higher critical cooling rates and hence water or brine
quenching is necessary.
418
Hardening
Factors affecting Hardening Processes
Chemical composition of steel
Size and shape of the steel part
Hardening cycle (heating/cooling rate, temp, soak time
Homogeneity and grain size of austenite
Quenching media
Surface condition of steel part
Hardening Methods
Conventional or direct quenching
Quenching in stages in sequence in different media
Spray Quenching
Quenching with self tempering
Austempering or Isothermal Quenching
Martempering
419
Retained Austenite
Austenite that is present in in the ferrous alloys even after the completion of the heat
treatment process is referred to as retained austenite. In simple words, retained austenite is
the untransformed austenite.
Austenite transforms to martensite between Ms and Mf temperatures as it is essentially an
athermal transformation. However, this transformation never goes to completion, i.e., 100%
martensite (Mf temperature line is illustrated as dotted line in TTT diagrams).
This is because at Mf , a small amount of (~1%) of austenite is present in highly stressed
state along with ~99% martensite, and can not transform to martensite because unfavourable
stress conditions.
Both Ms and Mf temperatures decrease with increase in carbon content. Therefore amount of
retained austenite in hardened steels increase with increase in carbon content.
All alloying elements, except Al and Co, lower the Ms temperature and hence enhance the
amount of retained austenite. Therefore, both high carbon steels and high alloy steels are
more prone to the presence of retained austenite.
The substructure of retained austenite differs from that of the original austenite as it has as a
higher density of imperfections like dislocations, stacking faults, etc. which are created by
local plastic deformation of the austenite by martensite crystals.
Tool steels may have retained- austenite in the range of 5-35%. At the surface of a quenched
steel, that restrains are minimum. R.A is less at surface than center of the part. 420
Retained Austenite
Advantages
Ductility of austenite can help to relieve some internal stresses developed due to hardening,
to reduce danger of distortion and cracks. 10% retained austenite along with martensite is
desirable.
The presence of 30-40% retained austenite makes straightening operation of the
components possible after hardening. Straightening increases the hardness slightly.
Non-distorting steels owe their existence to retained austenite. Here enough austenite is
retained to balance the transformational contracting during heating, on the formation of
austenite from ferrite carbide aggregate on the one hand, and the expansion corresponding
to the formation of martensite during cooling, on the other, Here, the basis of dimensional
stability of non-distorting steels is the presence of retained austenite.
disadvantages
The soft austenite if present, in large amounts, decreases the hardness of hardened steels.
As retained austenite may transform to lower bainite, or to martensite, there takes place
increase in dimensions of the part. Not only it creates problems in precision gauges, or dies,
the neighboring parts may be put under stress by it. In the component itself, stresses may be
created to cause distortion or cracking.
Retained austenite decreases the magnetic properties of the steel. 421
Sub-Zero treatment
The retained austenite is generally undesirable, sub-zero treatment is one of the method to
eliminate retained austenite.
As the room temperature lies between Ms and Mf temperatures of steel, quenching to room
temperature results in retained austenite.
Subzero treatment consists in cooling the hardened steel to a temperature below 0°C. The
temperature of the sub zero treatment depends on the position of Mf temperature of the
steel.
A steel can be cooled much below the Mf temperature, but it, evidently achieves nothing,
because it cannot bring about any additional increase of hardness, or any additional increase
of martensite, because the Martensitic transformation ends at Mf temperature.
Sub-zero treatment is more effective, if it is carried out immediately after quenching
operation. Any lapse of time between hardening and the cold treatment causes the
stabilization of austenite, makes the retained austenite resistant to further transformation.
Most steels can be cooled by subzero treatment in a low cooling unit with one of the
mediums as given in table (next page) .
The low-cooling unit consists of two vessels, the interior one of copper, where the parts or
tools to be deep frozen, are placed and the exterior one of steel provided with a good heat
insulation.
422
Sub-Zero treatment
Table : Subzero Coolants with
Temperature of Application
Minimum
Coolant
temperature °C
Dry ice (solid Co2)
-78
+ Acetone
Ice + Salt (NaCl) -23
Ice + Salt (CaCl2) -55
Liquid air -183
Liquid Nitrogen -196
Liquid Pentane -129
Freon -111
Figure : Liquid cooled (liquid N2 ) system.

Components are immersed in a bath of
alcohol, or trichloro ethylene, which is cooled
by a submerged liquid nitrogen spray (-150°C
), cooling rates can be controlled.
423
Sub-Zero treatment
The space in between the vessels is filled with one of the chosen medium, or a system
(figure in previous page) which is inexpensive and can be used.
Usually the temperature range used is in range of -30°C to -150°C, and total time of cooling
and holding at that temperature (Mf) varies from ½ - 1 hour. The hardness increased by 2-4
HRc.
As the amount of martensite increases by sub-zero treatment, it increases hardness, abrasion
resistance, fatigue resistance and eliminates the danger of developing grinding cracks.
As the newly formed martensite may add further to unfavorable stresses to cause distortion
and cracks, the complicated, or intricate shaped components may be first tempered at 150-
160°C immediately after first quenching and then given the sub-zero treatment.
Sub-zero treatment has been most extensively used for…!

Alloyed tool steels – like high speed steel, which now shall need only single stage
tempering.
Tools and components which need exact dimensions – gauges
Carburized steels, especially alloy steels (having elements like Ni in it) to increase their
hardness and wear resistance
Steels having 0.8 to 1.1%C as hardness increases by 1-3 HRc
424
Tempering
Objective
Relieve Internal Restore ductility To improve To improve magnetic

stresses and toughness dimensional stability properties
Structure in as Quenched state
Highly supersaturated martensite
Retained austenite
Undissolved carbides
Rods, or plates of carbide particles produced during ‘auto-tempering’
Segregation of carbon
425
Tempering
The hardened steel is not readily suitable for engineering applications. It possesses
following three drawbacks.
Martensite obtained after hardening is extremely brittle and will result in failure of
engineering components by cracking.
Formation of martensite from austenite by quenching produces high internal stresses in
the hardened steel.
Structures obtained after hardening

consists of martensite and retained
austenite. Both these phases are
metastable and will change to stable
phases with time which subsequently
results in change in dimensions and
properties of the steel in service.
Tempering helps in reduce these problems.
Tempering is the process of heating the
hardened steel to a temperature maximum
up to lower critical temperature (A1),
soaking at this temperature, and then
cooling, normally very slowly. 426
Variation in properties with tempering temperature
Tempering of plain carbon steels
First stage of Tempering
First Stage of tempering temperature extends from room temperature to 200°C. The
tempering reactions in steels, containing carbon less than 0.2%, differ somewhat from the
steels containing more than 0.2% carbon.
In the former, if carbon atoms have not yet segregated (during quenching) to dislocations,
these diffuse and segregate around the dislocations and lath boundaries in the first stage of
tempering. No ε-carbide forms as all the carbon gets locked up to the dislocations (defects).
Martensite in steels with more than 0.2% carbon is highly unstable because of super
saturation, and interstitial diffusion of carbon in BCT martensite can occur. Thus in the first
stage of tempering, the decomposition of martensite into low-tetragonality martensite
(containing ~0.2%C, c/a ~ 1.014) and ε-carbide, Fe2.4C occurs. (There are reports of
precipitation of eta-carbide, Fe2C and Haggs carbide, Fe2.2C.
ε-carbide is a separate phase and is not a preliminary step in the formation of cementite, but
it nucleates and grows more rapidly than cementite. It has HCP structure with c = 4.33A°, a
= 2.73A°, c/a = 1.58A° and forms as small (0.015-0.02 μm) platelets, or needles observed
under electron microscope.
The structure at this stage referred to as tempered martensite, which is double phase mixture
of low tetragonal martensite and ε-carbide.
In this stage volume ↓ because specific volume of martensite ↓ due to rejecting of C atoms.
427
Second stage of Tempering
Second Stage of tempering temperature lies between 200-300°C. The amount of retained
austenite in the as-quenched steel depends mainly on the composition of the steel, and the
temperature to which steel is quenched.
In the second stage of tempering retained austenite transforms to lower bainite (the carbide
in bainite is ε-carbide). The matrix in lower bainite is cubic ferrite (c/a = 1), where as in
tempered martensite, the low tetragonal martensite has c/a ~ 1.014
When retained austenite changes to lower bainite, their takes place increase in volume.
Third stage of Tempering
Third Stage of tempering temperature lies between 200-350°C. In this stage of tempering,
ε-carbide dissolves in matrix, and low tetragonal martensite losses its completely its carbon
and thus, the tetragonality to become ferrite .
Cementite forms as rods at interfaces of ε-carbide and matrix, twin boundaries, interlath
boundaries, or original austenite grain boundaries.
During this stage, volume decreases just as in stage one, due to complete loss of
tetragonality. In a 1% carbon steel , the total decrease in length in the first and third stages
in around 0.25%
428
Fourth stage of Tempering
Fourth Stage of tempering temperature lies between 350-700°C.

Growth and spheroidisation of cementite, as well as recovery and Recrystallization of ferrite
occur. Though the growth of cementite starts above 300°C, its spheroidisation starts above
400°C to 700°C.
Spheroidisation takes place due to reduction in interfacial energy of ferrite-cementite
interfaces. As quenched martensite has high concentration of lattice defects. Though their
annealing out starts in the third stage of tempering, but the cementite precipitates retard the
recovery processes.
Substantial recovery processes starts occurring only above 400°C. original lathe boundaries
are stable up to 600°C, but above this, these are replaced by equiaxed-ferrite grain
boundaries – the process, which is best described as ‘Recrystallization’.
In the end, the optical microstructure consists of equiaxed ferrite grains with coarse
Spheroidal particles of cementite, and then the structure is called globular pearlite, or
spheroidized cementite.
The structure perhaps is the most stable of all ferrite- cementite aggregates, and is the
softest with highest ductility with best machinability.
429
Effect of carbon on Tempering
Effect of tempering temperature( 1 hour at each temperature) on hardness and reactions

430
Tempering of alloy steels
The presence of alloying elements, steels can change their nature and properties effectively.
Most common elements (except cobalt) shift the CCT curve to longer times, which
essentially result in the increase of hardenability of the steels, so that pearlitic
transformation can be avoided easily to obtain Martensitic structure, even at a slower trate
of cooling and in thicker parts.
Alloying elements also lower Ms and Mf temperatures, increasing further the amount of
retained austenite. The decomposition of retained austenite on tempering, plays quite a
significant role on the properties of tempered steels, specially having high carbon and high
alloying elements.
Some elements, that are not found in carbides, but are present as solid solution in ferrite, are
Al, Cu, Si, P, Ni, and Zr. Some elements arranged in order of increasing tendency to form
carbides are Mn, Cr, W, Mo, V and Ti. These carbide forming elements retard most
effectively the rate of softening during tempering of the steel.
The first stage of tempering does not appear to be effected by the presence of the alloying
elements. However, most of the alloying elements in steels tends to increase the hardness
after tempering than a plain carbon steel with the same carbon content.
At smaller concentration, they merely retard the tempering processes hence the softening,
particularly at higher temperature (> 500°C), where these elements have good diffusivity to
take part in tempering reactions. 431
When alloy carbides are formed, the drop in hardness during tempering is not only retarded
but is significantly increased. The steel is then said to secondary hardening.
Thus, 0.5% chromium , or less than 0.5% Mo resists softening but secondary hardening is
produced by either 12% chromium, or 2 % Mo.. Stronger the carbide, the more potent is the
secondary hardening.
Effect of increasing Mo on tempering of as

432
Effect of increasing chromium in 0.35%C steel on tempering
quenched 0.1%C steel
Element, such as, silicon dissolves in ε-carbide to stabilize it. Steels with 1-2% silicon have
ε-carbide present even after tempering at 400°C, which means that the formation of
cementite is delayed considerably, and thus, resisting the softening. Otherwise, the effect of
silicon is essentially due to solid solution strengthening.
Nickel has a small, but constant effect on tempered hardness at all temperatures due to solid
solution strengthening as it is not a carbide former.
Manganese has little effect on softening at low tempering temperatures, but at high
temperatures, has a strong effect on due to its faster diffusion then, and thus, it resists
cementite coarsening as it is present in cementite as (Fe, Mn)3C.
Martensite in plain carbon steels losses its tetragonality by 300°C, but the tetragonality is
seen at 450°C, or even at 500°C if the steels have elements like Cr, W, Mo, Ti, V and Si.
The basic cause of steep softening in carbon steels on tempering above 400°C, is the
coagulation of the cementite particles. Alloying elements notably Si, Cr, Mo, V, when
present in steels, retard the coalescence and the coarsening of cementite particles., resulting
in enhanced hardening over and above the solid solution hardening effect.
Elements like Cr, Si, Mo, or W delay coarsening to temperature range of 500-550°C. up to
tempering temperature 500°C, the carbides formed are of iron with proportional alloying
elements in it, but above 500°C, alloying elements can form their own carbides and thus,
coarse cementite particles are replaced by fine dispersion of more stable alloy carbides.
433
An Fe-10 Ni alloy shows constant
hardness on tempering up to 450°C and
then, there takes place some decrease in
strength (curve I).
Addition of 0.12% carbon increases the
as quenched strength to almost double,
and slow decrease of hardness occurs on
tempering to fall to 0.7 GPa at 500°C.
A 8% cobalt addition, which doesn’t
enter the carbide, delays the softening to
have strength of 0.8 GPa at 500°C .
Addition of 2% Cr almost continuously
but slowly ↑ hardness to start falling at
above ~450°C to become 1.1 GPa at
500°C by fine dispersion of Cr carbide.
Addition of Mo causes secondary
hardening, as it is very strong carbide
forming element, to attain a hardness of Effect of C, Co, Cr, and Mo on tempering of
1.3 Gpa at 500°C Fe-10Ni steels
434
Tempering of alloy steels : Secondary Hardening
In alloy steels, having larger amounts of strong carbide forming elements like Mo, Ti, V,
Nb, W, Cr etc., and carbon , a peculiar phenomena occurs, the hardness of the as-quenched
martensite (called primary hardness) on tempering, decreases initially, as the tempering
temperatures is raised, but starts increasing again to often become higher than the as
quenched hardness, with much improved toughness, when tempered in the range of 500 to
600°C. This increase in hardness is called secondary hardness (also called red hardness).
This is great importance in high speed steels, as these are able to continue machining, at
high speeds (as these are able to resist fall in hardness and thus, the cutting property ) even
when they become red hot.
Secondary hardening is a process, similar to age hardening, in which coarse cementite
particles are replace by a new and much finer alloy carbide dispersion of V4C3, Mo2C, W2C
(which normally form on dislocations). As in aging a critical dispersion causes a peak in the
hardness and strength of the alloy, and as over aging takes place, i.e., carbide dispersion
slowly coarsens, the hardness decreases.
Secondary hardening is best shown in steels containing Mo, V, W, Ti and also in Cr steels at
high chromium concentrations.
The amount of secondary hardening in an alloy steel is directly proportional to the volume
fraction of the alloy carbides, and thus is directly proportional to the concentration of strong
carbide forming elements present in steels. The alloy carbides must precipitate as fine
dispersion in ferrite matrix rather than massive carbide particles. 435
Time and Temperature relationship in Tempering
For a given steel, a heat treater might like to choose some convenient tempering time, say
over night, otherwise different than 1 hour, and thus, wants to calculate the exact
temperature required to achieve the constant hardness.
Hollomon and Jaffe’s “tempering parameter” may be used for this purpose as it relates the
hardness, tempering temperature and tempering time. For a thermally activated process, the
usual rate equation is : 1 −Q / RT
Rate = = Ae
t
Where, t is the time of tempering to develop a given hardness, and Q is the ‘empirical
activation energy’ . ‘Q’ is not constant in the complex tempering processes but varies with
hardness. Thus, hardness was assumed to be a function of time and temperature:
H = f [te −Q / RT ]
−Q / RT
Interestingly, [te ] is a constant, and let it be t0. Equating activation energies of eq (1)
and (2) gives, Q = T [ln t − ln t0 ] = f ( H )
As t0 constant then
H = f [T (C + ln t )]
Where, C is a constant, whose value depends on the composition of austenite. The single
parameter which expresses two variables time and the temperature i.e., T (C + ln t) is called
the Hollomon and Jaffe tempering parameter. (hardness in vickers is preferable) 436
Temperature and colours for Heating and Tempering of Steel
Colours of Hot Solid metal °C Process of Heat treatment

White 1500
Yellow white 1300 High speed steel hardening (1230-1300°C)
HEAT COLOURS
Yellow 1100
Orange Red 1000 Alloy steel hardening (800-1100°C)
Light-Cherry-Red 900
Cherry-red 800 Carbon steel hardening
Dark-red 700
Vary dark-red 500 High speed steel tempering (500-600°C)
Black red in dull light, or darkness 400
Steel gray 300 Carbon steel tempering (150-575°C)
Colour of Oxide film °C Parts Heat treated
Steel Gray 327 Cannot be used for cutting tools
TEMPER COLOURS
Pale-light blue 310 For springs

Purple 282 Spring and screw drivers
Brown 270 Axes, wood cutting tools
Gold 258 Shear blades, hammer faces, cold chisels
Dark-straw-light-brown 240 Punches and Dies
Light-Straw-Yellow 220 Steel cutting tools, files, paper cutters
437
Embrittlement during Tempering
Normally, as the tempering progresses, the hardness and the strength decreases, but the
ductility and impact strength increase. But in certain steels, there is an unexpected decrease
of the impact strength in certain ranges of temperatures. This indicates that there are two
main types of embrittlement during tempering.
Tempered Martensite Embrittlement (TME)
Temper Embrittlement
Both these embrittlement raise the impact transition temperature (ITT) to higher
temperature. Figure (below) indicates the increase in impact transition temperature, Δ(ITT)
due to TE in SAE 3140 steel.
unembrittled
Temper-embrittlement in SAE 3140
embrittled
steel shifts the impact transition
temperature to higher temperature .
Δ(TTT) indicates the amount of
increase.
438
Embrittlement during Tempering
Normally, the degree of both type of embrittlement is expressed in terms of relative
displacements of ITT i.e., by Δ (ITT) illustrated in figure (in previous slide).
In both the transgranular fracture mode is replaced by an intergranular (IG) mode below the
transition temperature, i.e., these show bright inter crystalline fracture along original
austenite grain boundaries.
Inspite of these similar effects, the two types of embrittlement are two separate phenomena,
which occur in two different temperature ranges. Moreover, TME is a much faster process
usually occurring in one hour, where as TE takes many hours.
Temper embrittlement is of mush greater concern from practical point of view, as the rotors
and shafts of power generating equipment even after tempering above 600°C, develop it
when thick section cool very slowly through the range (450-600°C)
Schematic illustration of effect of

tempering temperature on impact
strength in steels prone to
embrittlement during tempering.
439
Tempered Martensite embrittlement
As TME develops after tempering in range 260°C to 370°C, it is called as “350°C”
embrittlement, or 500°F embrittlement. It is called ‘one-step embrittlement’ as during
heating only in this range, TME develops. It is also called ‘irreversible embrittlement
because a steel embrittled by tempering in this range , if heated further to above 400°C
(above the critical range), becomes tougher, and the tempered martensite embrittlement
does not occur again if cooled down to or tempered in the range of 260°C to 370°C again.
All steels, including the plain carbon steels are prone to irreversible embrittlement to some
extent, and that is why tempering range of 260°C- 370°C is avoided in all steels, though it
is a malady of low alloy steels.
The embrittlement is associated with the change in the structure of carbide from epsilon (ε)
to cementite in the form of a film at the grain boundaries.
On tempering at higher temperatures, this film disappears and can not be restored, on
repeated heating in 260°C-370°C temperature range.
Although, tempered martensite embrittlement is concurrent with the precipitation of
cementite, but such precipitation is not in itself the cause of loss of impact toughness, as the
embrittlement is does not occur if P, Sb, Sn, As, or N are not present in steel.
Addition of sufficient silicon to the steel inhibits the formation of cementite in the critical
range, as silicon dissolved in epsilon carbide, increases its stability, and thus embrittlement
does not occur 440
Temper embrittlement
The sickness of alloy steels occurs when they are tempered in the range 450°C to 600°C. it
is also called reversible embrittlement (as well as two step embrittlement), because it
occurs, when steels are tempered in this range, but gets removed, when heated to high
temperatures, but occurs again on slow continuous cooling through this range from that high
temperature (>600°C). The degree of embrittlement depends on the rate of cooling in the
range 600-450°C.
The phenomena of temper embrittlement results in loss of toughness as measured by
notched impact test (without affecting very much the hardness, Y.S, UTS, elongation and
fatigue properties), and a rise in ductile to brittle transition temperature occurs, with an
intergranular (IG) fracture, below the transition temperature along the original austenitic
grain boundaries.
Carbon steels in general, but with less than 0.5% Mn, do not show temper embrittlement.
Alloy steels of high purity do not show it. It is caused primarily by Sb and P and secondarily
by Sn or As (even in amounts as low as 0.01%) in presence of elements like Ni, Mn, Cr, Si
in steels. The highest effect is in Ni-Cr and Mn-Cr steels. Presence of elements like Mo, Ti,
Zr delay, or remove embrittlement.
The characteristic features of temper embrittlement are best explained by the concept of co-
segregation. The impurity solutes are the surface active elements in iron, i.e., these reduce
the grain boundary energy, and thus reduce the cohesion. Elements like Sb, P, As, Sn,
interact with certain elements like Ni and Mn in steels. 441
Temper embrittlement
These interactions leads to co-segregation of alloying elements and the impurity elements
such as between Ni-Sb, Ni-P-Ni-Sn and Mn-Sb. The reason of co-segregation is the stronger
interaction between them than, between either of these and iron.
If the interaction is very strong then, co-segregation does not occur, but a scavenging effect
is got, as happens between Mo-P, Ti-P, which is the cause of elimination of embrittlement
by 0.5% Mo in such steels. If larger amount of Mo, Ti, Zr are present , then these elements
slowly react with carbon to form stable carbides releasing the impurity atoms to segregate to
the boundaries.
Additional segregation may take place, when two alloying elements are present
simultaneously, such as Ni and Cr. At high temperatures (>600°C), thermal vibrations make
the equilibrium segregation low enough not to cause embrittlement., and at lower
temperature (<450°C), the diffusion of the elements is too low to cause enough co-
segregation with in the normal tempering time.
The following methods are normally recommended to minimize the effect of Temper
embrittlement
Keep the impurities such as Sb, P, Sn, As as low as possible
Alloy the steel with Mo (0.5-0.75%)
Quench from tempering at higher temperatures
442
Heat Treatment Defects
Heat treatment of steels or aluminum can lead to several defects. The principal types of
defects found in quenching of steels are internal and external cracks in the work, distortion
and warping.
CRACK : When the internal tensile stresses exceed the resistance of the steel to separation
the crack occurs. The insertion of the tools in the furnace without preheating for tempering
is one of the main causes of crack propagation. The crack formation is reduced by
preheating the tool between 2000Cto 3000C.
DISTORTION : Distortion occurs due to uneven heating, too fast cooling, part incorrectly
supported in furnace, incorrect dipping in quenching and stresses present before preheating.
Distortion can be prevented by preheating the tool or check furnace capacity, reduce the
hardening temperature, and by reviewing the method of dipping.
WARPING : Asymmetrical distortion of the work is often called warping in heat-treating
practice. Warping is usually observed upon non-uniform heating or over heating for
hardening. It also occurs when the work is quenched in the wrong position and when the
cooling rate is too high in the temperature range of the martensite transformation. An
elimination of these causes should subsequently reduce warping.
The properties required in the heat treated part are obtained without the parts being distorted
beyond the acceptable limits.
The ideal design for a heat treatable part is the shape that when heated (or cooled) would
have the same temperature at every point with in the part. 443
Avala Lava Kumar*
Introduction
Plastic deformation in the temperature range (0.3 – 0.5) Tm → COLD WORK
Point defects and dislocations have strain energy associated with them
(1 -10) % of the energy expended in plastic deformation is stored in the form of
strain energy (in these defects) → The material becomes battery of energy..!
The cold worked material is in a micro structurally metastable state.
Depending on the severity of the cold work the dislocation density can increase 4-6
orders of magnitude more. The material becomes stronger, but less ductile.
The cold worked material is stronger (harder), but is brittle.
Heating the material (typically below 0.5Tm) is and holding for sufficient time is a
heat treatment process called annealing.
Depending on the temperature of annealing processes like Recovery at lower
temperatures) or Recrystallization (at higher temperatures) may take place. During
these processes the material tends to go from a micro structurally metastable state
to a lower energy state (towards a stable state).
Further ‘annealing’ of the recrystallized material can lead to grain growth.
Annealed material Stronger material
 →
Cold work
445
ρ dislocation ~ (10 6 − 109 ) ρ dislocation ~ (1012 − 1014 )
Introduction
↑ point defect density

Material tends to lose
Cold work Anneal
the stored strain energy
↑ dislocation density
Increase in strength Softening of the material

of the material
During cold work the point defect density (vacancies, self interstitials…) and dislocation
density increase. Typical cold working techniques are rolling, forging, extrusion etc.
Cold working is typically done on ductile metals (e.g. Al, Cu, Ni)
Low temperature
Recovery
Cold work Anneal
Recrystallization
High temperature 446
Overview of processes taking place during annealing
Driving force is free energy stored in point defects

Cold work Anneal Recovery
and dislocations
Driving force is free energy stored in dislocations Recrystallization
Driving force is free energy stored in grain boundaries Grain growth
Cold work
↑ Strength ↑ Hardness ↑ Electrical resistance ↓ Ductility
Changes occur to almost all physical and mechanical properties 447

Recovery
Recovery takes place at low temperatures of annealing
“Apparently no change in microstructure”
Excess point defects created during Cold work are absorbed:
► at surface or grain boundaries RECOVERY
► by dislocation climb
Random dislocations of opposite sign come together and annihilate each other
Dislocations of same sign arrange into low energy configurations:
► Edge → Tilt boundaries
► Screw → Twist boundaries
POLYGONIZATION
Overall reduction in dislocation density is small
At the early stage of annealing of cold formed metals, external thermal energy permits the
dislocations to move and form the boundaries of a polygonized subgrain structure while the
dislocation density stays almost unchanged. This process also removes the residual stresses
formed due to cold working significant. The recovering of physical and mechanical
properties varies with the temperature and time.
448
POLYGONIZATION
Bent crystal
Low angle grain boundaries
POLYGONIZATION
449
Recrystallization
Trecrystallization ∈ (0.3 – 0.5) Tm

“Nucleation” and growth of new, strain free crystals
Nucleation of new grains in the usual sense may not be present and grain boundary
migrates into a region of higher dislocation density
∆G (recrystallization) = G (deformed material) – G (undeformed material)
TRecrystallization is the temperature at which 50 % of the material recrystallizes in 1
hour
Region of lower
Region of higher dislocation density
dislocation density
Direction of grain
boundary migration
450
Recrystallization
Deformation ↑ ⇒ recrystallization temperature (Trecrystallization) ↓

Initial grain size ↓ ⇒ recrystallization temperature ↓
High cold work + low initial grain size ⇒ finer recrystallized grains
↑ cold work temperature ⇒ lower strain energy stored
⇒ ↑ recrystallization temperature
Rate of recrystallization = exponential function of temperature
Trecrystallization = strong function of the purity of the material
Trecrystallization (very pure materials) ~ 0.3 Tm
Trecrystallization (impure) ~ (0.5 – 0.6) Tm
► Trecrystallization (99.999% pure Al) ~ 75oC

Trecrystallization (commercial purity) ~ 275oC
The impurity atoms segregate to the grain boundary and retard their motion →
Solute drag (can be used to retain strength of materials at high temperatures)
Second phase particles also pin down the grain boundary during its migration
451
Hot Work and Cold Work
Often the range is further subdivided into Hot, Cold and Warm working as in the figure
Hot Work Plastic deformation above TRecrystallization

Cold Work Plastic deformation below TRecrystallization
Hot Work
− 0.9 Tm
When a metal is hot worked. The conditions of deformation are
− 0.8 Tm such that the sample is soft and ductile. The effects of strain
− 0.7 Tm hardening are negated by dynamic and static processes (which keep
the sample ductile)
− 0.6 Tm
The lower limit of temperature for hot working is taken as 0.6 Tm
Warm − 0.5 Tm
Working − 0.4 Tm Recrystallization temperature (~ 0.4 Tm)
− 0.3 Tm
Cold Work
− 0.2 Tm The effects of strain hardening is not negated. Recovery

mechanisms involve mainly motion of point defects.
− 0.1 Tm
Upper limit ►0.3 Tm
452
Grain growth
The interfaces in a material cost the system energy: the excess free energy associated with
the interfaces is the interfacial free energy.
Grain boundaries are the interfaces between two crystallites which are differently oriented in
space; the excess free energy associated with the grain boundary is the grain boundary
energy.
The grain boundary energy acts as the driving force for the movement of grain boundaries.
Hence, if a recrystallised material is further annealed, then grains growth takes place; bigger
grains grow at the cost of smaller ones. This process is known as grain growth.
Since the driving force for grain growth is the interfacial energy, and since the excess energy
associated with a system due to interfaces is related to the curvature of the interface, the grain
growth is curvature driven. In Figure we shown the direction of movement of grain
boundaries and their relationship to curvature (in 2D systems).
453
Curvature driven growth of grains (in 2D)
Grain growth
It is the velocity of the interface, then ν ∝ κ where κ is the curvature  1  . More specifically
 
2γ r
ν =M
r
where M is the mobility, and ϒ is the grain boundary energy. In terms of the diameter (D) of
the grains, dD 4 Mγ
=
dt D
The solution to this equation is given by
D 2 = D02 + 4 Mγt
where D0 is the mean size of the grain at time t = 0 and D is the mean diameter of the grains
in the system.
Experimentally, it is found that the grain size as a function of time does follow an expression
of the type D = Kt n ,where K is a temperature dependent proportionality constant and n is
a number much less than 0.5. The deviation of the exponent from 0.5 is not yet clearly
understood.
454
Grain growth
Due to the differences in the driving forces for
recrystallisation and grain growth, as shown in
Figure the movement of the interface in these two
cases are different in character; recrystallisation,
as long as the growing grain is free of strain and
eats into the strained grain, will proceed
irrespective of the curvature of the grain;
however, during grain growth, the movement of
the interface is dependent on the curvature of the
grain.
During grain growth, the direction of movement
of the grain boundary is completely decided by
curvature. On the other hand, during
recrystallisation it is decided by the strains in the
grains; the strained one is eaten up by the strain
free crystal; hence, the growth can sometimes be
such that the boundary moves away from its
centre of curvature.
455
Grain growth
Large grains have lower free energy than small grains. This is associated with the
reduction of the amount of grain boundary.
Therefore, under ideal conditions, the lower energy state for a metal would be as a single
crystal. This is driving force for grain growth.
Opposing this force is the rigidity of the lattice. As the temperature increases, the
rigidity of the lattice decreases and the rate of grain growth is more rapid.
At any given temperature there is a maximum grain size at which these two effects are in
equilibrium
↓ Heating Time↓
Grain size →
↓ Degree of prior deformation ↑
Grain growth ↓ Time at temperature↑
↓ Annealing temperature↓
Temperature →
↓ Insoluble impurities↑ Effect of temperature on recrystallized 456
grain size
Recovery Recrystallization Graingrowth
Electical conductivity
Internal stress
Ductility
Tensile strength
Cold work Recovery Recrystallization Grain growth
Change of properties with increased annealing temperature of a cold

worked sample. Note that there are changes in properties beyond
recrystallisation temperatures too due to grain growth.
457
Questions?
1. What is annealing? What are its aims? Discuss the different types of annealing processes
giving the temperature ranges and the aims of each type.
2. Differentiate between:
a) Process annealing and recrystallization annealing
b) Diffusion annealing and spheroidising
c) Normalising and annealing
d) Stress-relieving and tempering
3. With the help of a suitable diagram, show normalizing and hardening temperature ranges for
plain carbon hypereutectoid steels. What is the criteria for the selection of these temperature
ranges?
4. Discuss the merits and demerits of the following hardening processes:
a) Direct quenching
b) Auto-tempering
c) Interrupted quenching
5. Explain why hardening by quenching is followed by tempering treatment? How do
mechanical properties vary with tempering temperature?
458
Questions?
6. Describe the structural changes that take place during tempering. Is there any fourth stage of
tempering? Which type of steels exhibit hardening and why?
7. What is the meant by temper brittleness? Why does it occur? Are there any remedies for it?
Discuss.
8. With the help of suitable diagram, explain the process of martempering. How does it differ
from austempering? What do the microstructures of martempered and austempered steels
consists of? What are limitations.
9. Distinguish between tempered martensite embrittlement (TME) and temper embrittlement
(TE) based on (i) rate of embrittlement (ii) reversibility (iii) no. of steps (iv) types of steels
(v) mechanism (vi) range of temperature.
10. Consider cold worked aluminium in which the dislocation density is increased from 1010 to
1014 per m2. Calcualte the driving force for recrystallisation
11. Most of the materials that we encounter in our daily lives are polycrystalline. If the grain
boundary energy does increase the free energy of the system, why do they not disappear
leaving behind a single crystal?
12. Describe the following processes
a) Sub zero treatment
b) Patenting
459
Questions?
13. What is the role of cold working in the recovery and recrystallization processes? Is it
important in the grain growth process?
14. What is the role of diffusion in the heat treatments applied to the alloy studied in this
experiment?
15. Why do grains grow? What is the upper limit for grain size? Lower limit?
460
Avala Lava Kumar
461
Introduction
Hardenability is one of the most important properties of a steel because it describes the ease
with which a given steel can be quenched to form martensite or the depth to which
martensite is formed on a given quench.
It is an important property for welding, since it is inversely proportional to weldability, that
is, the ease of welding a material.
The ability of steel to form martensite on quenching is referred to as the hardenability.
Hardenability is a measure of the capacity of a steel to be hardened in depth when
quenched from its austenitizing temperature.
Steels with high hardenability form martensite even on slow cooling.
High hardenability in a steel means that the steel forms martensite not only at surface but to
a large degree throughout the interior.
For the optimum development of strength, steel must be fully converted to martensite.
To achieve this, the steel must be quenched at a rate sufficiently rapid to avoid the
decomposition of austenite during cooling to such products as ferrite, pearlite and bainite.
Hardenability of a steel should not be confused with the hardness of a steel.
Hardness ≠ Hardenability
The Hardness of a steel is a measure of a Hardenability refers to its ability to be
sample's resistance to indentation or hardened to a particular depth under a
462
scratching, particular set of conditions.
Hardness profile in a cylinder from case to core
Schematic showing variation in cooling

Typical hardness test survey made along rate from surface to interior leading to
a diameter of a quenched cylinder different microstructures
463
Hardenability
It is a qualitative measure of the rate at which hardness drops off with distance into
the interior of a specimen as a result of diminished martensite content.
Hardenability is more related to depth of hardening of a steel upon heat treat.
The depth of hardening in a plain carbon steel may be 2-3 mm Vs 50 mm in an alloy steel.
A large diameter rod quenched in a particular medium will obviously cool more slowly than
a small diameter rod given a similar treatment. Therefore, the small rod is more likely to
become fully martensitic.
The hardenability of a steel is the
maximum diameter of the rod which will
have 50% martensite even in the core
when quenched in an ideal quenchants.
This diameter is known as Di or ideal
diameter.
Relation between cooling curves for

the surface and core of an oil-
quenched 95 mm diameter bar
464
Determination of Hardenability
Hardenability of steel is determined by the following methods
Grossman’s critical diameter method
Jominy end quench test
Estimation of hardenability from chemical composition
Fracture test
465
In Grossman’s method, we use round bars of different diameters.
These bars are quenched in a suitable quenchants.
Further, we determine the critical diameter (Dc) which is the maximum diameter of the rod
which produced 50% martensite on quenching.
The ideal diameter (Di) is then determined from the curve.
This type of experiment requires multiple Austenitisation and quenching treatments on
specimens of varying diameter just to quantify the hardenability of a single material.
Determination of the critical diameter Dcrit according to Grossmann 466

Severity of Quenching media
Process Variable H Value If the increase in rate of heat
Air No agitation 0.02 conduction is greater than the
Oil quench No agitation 0.2 decrease due to persistence of the
" Slight agitation 0.35 vapor film, the net result will be an
increase in the actual cooling rate.
" Good agitation 0.5
However if the reverse is true, then the
" Vigorous agitation 0.7 result will be decrease in cooling rate.
Water quench No agitation 1.0
" Vigorous agitation 1.5 Severity of Quench as indicated by the heat
transfer equivalent H
Brine quench
No agitation 2.0 f
(saturated Salt water)
H= [m −1 ]
" Vigorous agitation 5.0 K
f → heat transfer factor
Ideal quench ∞
K → Thermal conductivity
Note that apart from the nature of the quenching medium, the vigorousness of the shake determines the
severity of the quench. When a hot solid is put into a liquid medium, gas bubbles form on the surface of
the solid (interface with medium). As gas has a poor conductivity the quenching rate is reduced.
Providing agitation (shaking the solid in the liquid) helps in bringing the liquid medium in direct
contact with the solid; thus improving the heat transfer (and the cooling rate). The H value/index
compares the relative ability of various media (gases and liquids) to cool a hot solid. Ideal quench is a
conceptual idea with a heat transfer factor of ∞ (⇒ H = ∞) 467
The relation between ideal critical diameter DI and critical diameter DC and severity of
quench (H) can be determined from thermodynamic considerations.
These relations are shown in Grossman’s master graph. In this figure the DI is plotted as the
abscissa, and the DC is plotted as ordinate.
A number curves are plotted in this graph and each belongs to different rates of cooling. In
every case, the rate of cooling is measured by the H-value or the severity of quench.
From this graph, by knowing the value of DC, the corresponding value for DI can be found
out. For example, assuming DC = 1inch and H = 5 (for agitated brine quench), the critical
diameter or hardenability works out to 1.2inch with help of figure.
From figure it can be observed that

DI = DC for H = ∞
DI > DC for H < ∞
468
The relation between ideal critical diameter DI and critical diameter DC that can be fully hardened by using a quenching
medium with a given cooling power H.
Jominy End Quench method
The most commonly used method for determining hardenability is the end quench test
developed by Jominy and Boegehold
Grossmans method requires multiple Austenitisation and quenching treatments on
specimens of varying diameter just to quantify the hardenability of a single material.
An alternative approach is to develop a more convenient standard test method that can be
used for relative comparison of hardenability. The Jominy end-quench test is one such
approach.
The Jominy end-quench test is specified in ASTM standard A255 and is a widely used
method for quantifying hardenability. Its wide use adds to its value, since the utility of
empirical relations and data comparison becomes more reliable as more data are
accumulated.
Moreover, Jominy data have been collected on a large enough scale to offer a high degree of
statistical certainty for many steels.
These data have been correlated with measurements and/or calculations of DC.
By using these correlations, a single Jominy test can be used to estimate DC and DI for a
given steel (and austenite grain size).
Information gained from this test is necessary in selecting the proper combination of alloy
steel and heat treatment to minimize thermal stresses and distortion when manufacturing
components of various sizes.
469
Jominy End Quench method - Principle
The hardenability of a steel is measured by a Jominy test with following procedure.
In conducting this test, a 1 inch round specimen 4 inch long is heated uniformly to the
proper austenising temperature. It is then removed from the furnace and placed on a fixture
where a jet of water impinges on the bottom face of the sample.
After 10 min. on the fixture, the specimen is removed, and to cut along the longitudinal
direction.
And then Rockwell C scale hardness readings are taken at 1/16 inch, intervals from the
quenched end. The results are expressed as a curve of hardness values Vs. distance from the
quenched end. A typical hardenability curve shown below for eutectoid steel.
Jominy Sample
1”
flat ground
specimen
(heated to γ 4”
phase field)
24°C water
470
Jominy End Quench method - Principle
A number of Jominy end quench samples are first end- quenched for a series of different
times and then each of them (whole sample) is quenched by complete immersion in water
to freeze the already transformed structures.
Cooling curves are generated putting thermocouple at different locations and recording
temperature against cooling time during end quenching.
Microstructures at the point where cooling curves are known, are subsequently examined
and measured by quantitative metallography. Hardness measurement is done at each
investigated point.
Based on metallographic information on investigated point the transformation start and
finish temperature and time are determined. The transformation temperature and time are
also determined for specific amount of transformation.
These are located on cooling curves plotted in a temperature versus time diagram. The
locus of transformation start, finish or specific percentage of transformation generate CCT
diagram (see next slide).
A, B, C, D, E, F are six different locations on the Jominy sample shown in Figure (before
slide) that gives six different cooling rates. The cooling rates A, B, C, D, E, F are in
increasing order.
The corresponding cooling curves are shown on the temperature log time plot. At the end of
the cooling curve phases are shown at room temperature. Variation in hardness with
distance from Jominy end is also shown in the diagram. 471
Hardness, HRC
Jominy sample
FE D C B A
Distance from quench end
TE
to=Minimum incubation
A period at the nose of the
TTT diagram,
t0 tꞌo=minimum incubation
tꞌ0 period at the nose of the
B CCT diagram
Temperature
F C
E D M – Martensite
P – Pearlite
F.P – Fine pearlite
C.P – Coarse pearlite
MS MS – Martensite start
MF – Martensite finish
M50 TE – Eutectoid temp.
MF
M M P+M F.P P C.P
472
Log time
Determination of Hardenability from Jominy curve
After plotting the Jominy distance Vs Hardness curve, the Jominy distance having hardness
equal to 50 % martensite is determined.
Then the diameter of a rod having cooling rate similar to the cooling rate at the Jominy
distance having 50 % martensite is determined from the graph correlating the Jominy
distance with the diameter of the rod having similar cooling rate for water quenching.
This diameter gives the hardenability of the steel in water quenching (having H value equal
to 1).
Hardenability in any other quenchants can be determined from the same graph.
DI (hardenability in ideal quenching medium) can also be determined in a similar manner.
We can determine hardenability for any other amount of martensite in the core in any
quenchants in a similar way.
Variation of hardness along
a Jominy bar
Grossman chart
used to determine
the hardenability
of a steel bar
473
Estimation of Hardenability from chemical composition
Although the ideal critical diameter is generally determined experimentally, it is also

possible to estimate it from chemical composition and the grain size of steel
This method is based on the fact that the hardening of steel is controlled basically by the
carbon content.
Every steel has a base hardenability which depends only on carbon content and grain size.
Alloying additions change the rate of reaction.
Furthermore, the effect of each alloying element

is independent of other alloying elements. The (a)
effect is also independent of carbon content and
grain size.
Figure gives base hardenability in terms of ideal
critical diameter. It is clear from the figure that,
as the carbon content and grain size decreases,
the base diameter value also declines.
The effect of alloying elements in the
hardenability is shown in Figure (b) (next page).
Mo, Mn and Cr are seen to very effective. The
base diameter obtained from Figure (a) is
multiplied by the multiplying factors F for each
of the alloying elements given in figure (b) 474
Base diameter as function of carbon content and grain size
Estimation of Hardenability from chemical composition
DI = DC (basediameter ) × FMn × FCr × FNi
For example, for a steel of grain size ASTM 8, with 0.5% carbon, 0.6% Mn, 1% Cr and
2%Ni, DI = 0.22 (From figure a) × 3.00 (Mn factor from figure b) × 3.17 (Cr factor from
figure b) × 1.77 (Ni factor from figure b) = 3.70 inches
Sulphur and phosphorous are present in low concentrations in steel as impurity. Their
combined effect can be ignored in most of the cases. Multiplying factors for sulphur and
phosphorus are unity
(b)
475
Multiplying factor F as a function of weight % for a set of alloying elements
Significance of Hardenability Multiplying factor
The Hardenability Multiplying Factor shows the rate at which the hardening depth is
increased with the percentage of the alloying element
The ideal diameter (DI ) is calculated from: DI = DCritical × FMn × FCr × FNi × FSi × FMo
Where DIC is the basic DI factor for carbon and ƒx is the multiplying factor for the alloying
element x.
Carbon grain size Alloying factor, fX
% No. 6 No. 7 No.8 Mn Si Ni Cr Mo
0.05 0.0814 0.0750 0.0697 1.167 1.035 1.018 1.1080 1.15
0.10 0.1153 0.1065 0.0995 1.333 1.070 1.036 1.2160 1.30
0.15 0.1413 0.1315 0.1212 1.500 1.105 1.055 1.3240 1.45
0.20 0.1623 0.1509 0.1400 1.667 1.140 1.073 1.4320 1.60
0.25 0.1820 0.1678 0.1560 1.833 1.175 1.091 1.5400 1.75
0.30 0.1991 0.1849 0.1700 2.000 1.210 1.109 1.6480 1.90
0.35 0.2154 0.2000 0.1842 2.167 1.245 1.128 1.7560 2.05
0.40 0.2300 0.2130 0.1976 2.333 1.280 1.246 1.8640 2.20
0.45 0.2440 0.2259 0.2090 2.500 1.315 1.164 1.9720 2.35
0.50 0.2580 0.2380 0.2200 2.667 1.350 1.182 2.0800 2.50
0.55 0.2730 0.2510 0.2310 2.833 1.385 1.201 2.1880 2.65
0.60 0.284 0.262 0.2410 3.000 1.420 1.219 2.2960 2.80
0.65 0.295 0.273 0.2551 3.167 1.455 1.237 2.4040 2.95
0.70 0.306 0.283 0.260 3.333 1.490 1.255 2.5120 3.10
0.75 0.316 0.293 0.270 3.500 1.525 1.273 2.6200 3.25
0.80 0.326 0.303 0.278 3.667 1.560 1.291 2.7280 3.40
0.85 0.336 0.312 0.287 3.833 1.595 1.309 2.8360 3.55
0.90 0.346 0.321 0.296 4.000 1.630 1.321 2.9440 3.70
0.95 - - - 4.167 1.665 1.345 3.0520 -
1.00 - - - 4.333 1.700 1.364 3.1600 -
476
Fracture Test
There is a contrast in the nature of fracture undergone by martensitic and pearlitic regions.
Whereas martensite formed on the case exhibits brittle fracture, the pearlite formed in the
core undergoes ductile fracture.
Where there is changeover from martensitic to pearlitic structure, there is corresponding
sharp change from brittle to ductile fracture.
It is similar to a sudden change in hardness or microstructure as one passes from martensitic
to pearlite region.
This region of sudden change is the one that contains 50% pearlite and 50% martensite.
The method based on the nature of fractured surface is successful when the transformation
processes is quick and a sharp boundary is formed.
When the transformation is sluggish, the method cannot be applied since the gradient in
hardness is gradual, and it is not possible to get a clear demarcating boundary.
477
Hardenability Band
The industrial products of steels may

change composition and average grain size
from batch to batch, therefore, the
measured hardenability of a given type of
steel should be presented as a band rather
than a single line, as demonstrated by the
Figure at right
Hardenability data now exists for a wide
range of steels in the form of maximum
and minimum end-quench hardenability
curves, usually referred to as hardenability
bands.
This data is, available for very many of
the steels listed in specifications such as
those of the American Society of
Automotive Engineers (SAE), the
American Iron and Steel Institute (AISI)
and the British Standards.
478
Factors effecting Hardenability
Slowing the phase transformation of austenite to ferrite and pearlite increases the
hardenability of steels.
The most important variables which influence hardenability are
Quenching conditions
Austenitic grain size
Carbon content
Alloying elements
Quenching Conditions
The fluid used for quenching the heated alloy effects the hardenability.
Each fluid has its own thermal properties like..
Thermal conductivity
Specific heat
Heat of vaporization
These cause rate of cooling differences
Ideal quenchant : It is one which brings down the surface temperature to room temperature
instantaneously and keeps it at that temperature thereafter.
479
Austenitic grain size
The hardenability increases with increasing

austenite grain size, because the grain
boundary area which act as nucleating site is
decreasing.
This means that the sites for the nucleation of
ferrite and pearlite are being reduced in
number, with the result that these
transformations are slowed down, and the
hardenability is therefore increased.
The more γ-grain boundary surface the easier
it is for pearlite to form rather than
martensite
Smaller γ-grain size → lower hardenability
Larger γ-grain size → higher hardenability
480
Percentage of carbon
Carbon is primarily a hardening agent in steel.
It also increases hardenability by slowing the formation of pearlite and ferrite.
But its use at higher levels is limited, because of the lack of toughness which results in
greater difficulties in fabrication and, most important, increased probability of distortion
and cracking during heat treatment and welding.
481
Effect of alloying elements
Most metallic alloying elements slow down the ferrite and pearlite reactions, and so also
increase hardenability. However, quantitative assessment of these effects is needed.
Chromium, Molybdenum, Manganese, Silicon, Nickel and Vanadium all effect the
hardenability of steels in this manner. Chromium, Molybdenum and Manganese being used
most often.
Boron can be an effective alloy for improving hardenability at levels as low as .0005%.
Boron has a particularly large effect when it’s added to fully deoxidized low carbon steel,
even in concentrations of the order of 0.001%, and would be more widely used if its
distribution in steel could be more easily controlled.
The most economical way of increasing the hardenability of plain carbon steel is to increase
the manganese content, from 0.60 wt% to 1.40 wt%, giving a substantial improvement in
hardenability.
Chromium and molybdenum are also very effective, and amongst the cheaper alloying
additions per unit of increased hardenability.
Hardenability of a steel increases with addition of alloying elements such as Cr, V, Mo, Ni,
W → TTT diagram moves to the right.
482
Cr, Mo, W, Ni
temperature
time
Exceptions
S - reduces hardenability because of formation of MnSand takes Mn out of solution as MnS

Ti - reduces hardenability because it reacts with C to form TiC and takes C out of solution;
TiC is very stable and does not easily dissolve
Co - reduces hardenability because it increases the rate of nucleation and growth of pearlite
483
All steels have 0.4wt% C, but with

different alloying elements. (in the
figure shown)
At the quenched end all alloys have
the same hardness, which is a
function of carbon content only.
The hardenability of the 1040 is low
because the hardness of the alloy
drops rapidly with Jominy distance.
The drop of hardness with Jominy
distance for the other alloys is more
gradual.
The alloying elements delay the
austenite-pearlite and/or bainite
reactions, which permits more
martensite to form for a particular
cooling rate, yielding a greater
Hardness at center of a 3 inch bar is hardness.
different for different steels indicating
different amounts of martensite at the
484
center
Effect of boron on Hardenability
Boron is added to steel for only one reason-to increase hardenability. Boron-treated steels
have certain peculiar characteristics
Boron increases hardenability in hypoeutectoid steel, has no effect on eutectoid steel, and
decreases hardenability in hypereutectoid stee.
As the austenite grain site becomes finer, the hardenability effect of B increases.
Austenitizing at high temperature reduces the hardneability effect of B.
For the maximum B effect, the concentration should be in the range 0.0005 to 0.003 wt%.
Increasing hardenability by adding B does not decrease the Ms temperature.
These arise because of the manner in which B produces its effect and the interactions of B
with other elements in steel.
To be effective, B must be in solid solution in austenite. The solubility of B in austenite is
very low, for example, the solubility at 912°C is about 0.001 wt% , increasing with
temperature to a maximum value of about 0.005 wt% at the eutectic.
In α-iron, the solubility is essentially zero, but is influenced by the impurities present.
Because of its low solubility in austenite, B can be highly concentrated in grain boundaries.
When a boron steel is cooled from the hardening temperature the solubility of boron is
reduced, which results in a still greater concentration of B at the grain boundaries.
Minute grains of boron carbide Fe23(BC)6 are formed there and to some extent they assume
an orientation coherent with one of the two austenite grains between them which separate
out. 485
Effect of boron on Hardenability
Atomic contact is thereby established between Fe23(BC)6 and austenite, resulting in a
reduction in the surface tension and grain-boundary energy.
The presence of boron in solid solution and coherent boron carbide in the grain boundaries
delays the formation of ferrite and pearlite and also to some extent, bainite; hence
increasing the hardenability of the steel
The effect of B may be expressed quantitatively as the boron factor, which is the ratio of the
ideal diameters, Di, (according to Grossmann) for the steel with and without boron
Di with' B'
BF =
Di without ' B'
Di (with boron) is derived from the Jominy end-

quench hardenability curve; Di (without boron) is
calculated from the chemical composition of the
steel.
The optimum B content, near 0.002 wt%, for
increasing hardenability in 0.2% C, 0.65% Mn,
0.55% Mo (wt%) steel is shown in Fig. 15. Boron
contents above about 0.003 wt% lead to a loss in
hardenability, and B in excess of about 0.004 wt%
causes a loss in toughness through precipitation of
Fe2B in austenite grain boundaries. 486
Questions?
1. Define hardenability? What is the common criterion of hardenability of steels, and why?
Enumerate five factors effecting the hardenability of the steel.
2. What is the effect of the presence of a carbide forming elements (on the hardenability of
steel) that is not dissolved in austenite prior to hardening processes? And how do the
nonmetallic inclusions effects the hardenability of steel?
3. Calculate the hardenability (DI) of steel composition:
C = 0.4%, Mn = 0.7, P = 0.04, S = 0.04, Si = 0.3, Ni = 1.8, Cr = 0.8, Mo = 0.25, and ASTM
grain size = 8, what could be critical diameter (DC) in water and oil. What severity of
quench would be required to fully harden the bar of 3” in diameter?
4. Differentiate between hardness and hardenability
5. Define and explain the term severity of quench. Explain the its impact on hardenability
6. Discuss how hardenability is affected by: (i) austenitic grain size (ii) carbon content (iii)
presence of alloying elements
7. Distinguish between
a) Pearlitic hardenability and bainitic hardenability
b) Core and case hardenability
c) Shallow hardening and deep hadening steels
487
Questions?
8. Discuss the effect of boron on hardenability

a) “Boron is added in steels to increase the hardenability” comment
b) What is the effect austenitising temperature, increased carbon, on boron hardenability.
9. Which elements are commonly used to increase hardenability of the steels? And what are
the disadvantages of steels of high hardenability?
10. What are the various methods of determining hardenability?
488
Avala Lava Kumar
E-mail : lavakumar.vssut@gmail.com
489
Iron – Graphite Phase Diagram
The true equilibrium diagram for iron and carbon is generally considered as iron-graphite
phase diagram (next slide).
Earlier we learn about iron-iron carbide phase diagram and it is not a true equilibrium
diagram, generally it is called metastable iron - iron carbide phase diagram.
Cementite (Fe3C) is a metastable compound, and under some circumstances it can be made
to dissociate or decompose to form ferrite and graphite, according to the reaction
Fe3C → 3Fe + C
For explain of cast irons we will refers to both Iron-Iron carbide phase diagram and Iron-
Graphite phase diagram.
Earlier we are studied steel microstructures from Iron-Cementite phase diagram. Same as
earlier now we will learn the characteristic features of CAST IRONS with the help of Iron-
Graphite phase diagram.
Cast irons are a class of ferrous alloys with carbon contents above 2.14 wt%; in practice,
however, most cast irons contain between 3.0 and 4.5 wt% C and, in addition, other alloying
elements .
The ductility of cast iron is very low and brittle, it cannot be rolled, drawn, or worked at
room temperature. However they melt readily and can be cast into complicated shapes
which are usually machined to final dimensions. Since casting is the only suitable process
applied to these alloys, they are known as cast irons. 490
Iron – Graphite Phase Diagram
δ 0.51 Stable
1539 L
1493
1395 L + Graphite
L+γ
1153
γ
Temperature ºC →
0.09 %C 1.98 4.2

Metastable
912 γ + Graphite
740
0.69
α
0.021 α + Graphite (C)
Fe C
0.16 6.7 100
% Carbon →
491
Classification of cast irons
White CI
Grey CI Malleabilize
Ductile/Nodular CI Stress concentration

at flake tips avoided
CAST IRONS
Malleable CI
Compacted Graphite CI
Alloy CI
Good castability ⇒ C > 2.4%

492
Cast Irons Alloying System
Most common alloying element of Cast Irons is Silicon for various reasons, which include
the manipulation of temperatures required to achieve desired microstructures.
For example Increases the stability of solidification of Graphite phases. Decreases the
stability of the solidification of Fe3C. Eutectic and eutectoid temperatures change from
single values to temperature ranges. Eutectic and eutectoid compositions are affected.
Carbon Equivalent (CE): a measure of
the equivalency of Carbon coupled
with other alloying elements to that of
just Carbon. An easier basis for
classifying the properties of a multi
alloy material.
CE = %C + 1/3%Si = 4.3
(hypoeutectic) < CE = 4.3 <
(hypereutectic), with the addition
phosphorus CE = %C + 1/3(%P +
%Si)
493
Solidification of phases in cast irons
Graphite structure is a factored crystal bounded by low index planes.
 − 
Growth occurs along 10 1 0  & (0001) planes (direction A & C). Unstable growth occurs
 
along direction A, giving the Graphite microstructure rough, poorly defined, edges in
certain areas.
When grown from the solidification of liquid Iron Carbon alloys, Graphite takes on a layer
structure. Among each layer covalent chemical bonds with strengths between (4.19 × 105 to
5 × 105 J/mol.) exist. Between layers weaker bonds exist on the order of (4.19 × 103 – 8.37
× 103 )J/mol.
The structure of the Graphite depends on the chemical composition, the ratio of temperature
gradient to growth rate, and the cooling rate. Such structures are:
Crystalline structure of graphite, A and C possible growth directions 494

Nucleation of Flake and Spheroidal Graphite
Such structures are (a) Flake or Plate Graphite, (b) Compacted vermicular graphite, (c)
Coral Graphite, (d) Spheroidal Graphite.
A wide variety of compounds and certain metals have been claimed to serve as either
inoculants or nuclei for Flake Graphite growth. (Silicon dioxide, silicates, sulfides, boron
nitride, carbides, sodium, potassium, calcium, etc..!)
Two methods of growth is
possible. The nucleation of Flake
Graphite Iron occurs mainly on
silicon dioxide particles. Another
one, Salt like carbides containing
the ion Carbon are used as
inoculants. Such carbides include
NaHC2 & KHC2 from Group I,
CaC2, SrC2, BaC2 from group II,
and YC2 & LaC2 from group III. 495
Epitaxial growth of graphite on CaC2 crystal
Cooling Curve Analysis
The solidification of a sample is
represented mathematically by the
relationship dQ
= VρC P dT dt
dT
Where V is the volume of the sample, ρ is
the density, CP is the heat capacity, and
dT/dt is the cooling rate of the liquid.
dT/dt is the slope of the cooling curve
before solidification begins.
When the liquid cools below the liquidus
temperature, crystals nucleate and begin to
grow. The rate is thus re-expressed as
dQ
= (VρC P + ρ∆H df dt ) dT dt
dT
Cooling curve for an Iron-carbon alloy with
Where ΔH is the heat of solidification and 3.2%C. The solidification starts with a primary
df/dt is the volume fraction of solid formed precipitation of austenite, followed by a
at a changing temperature eutectic temperature
496
White Cast Iron
In which all the C is in the combined form as Fe3C (Cementite)
The typical microstructure of white cast iron (see next slide), consisting of dendrites of
transformed austenite (pearlite) in a white interdendritic network of cementite.
Microstructure → Pearlite + Ledeburite + Cementite
White cast iron contains a relatively large amount of cementite as a continuous
interdendritic network, it makes the cast iron hard and wear-resistant but extremely brittle
and difficult to machine.
‘completely white’ cast irons are limited in engineering applications because of this
brittleness and lack of machinability. mainly used in liners for cement mixers, ball mills and
extrusion nozzles.
A large tonnage of white cast iron is used as a starting material for the manufacturer of
malleable cast iron.
Hardness : 375 to 600 BHN, Tensile strength : 135-480 Mpa, Compressive strength: 1380 -
1725 Mpa
White cast irons fall into three major groups: Nickel Chromium White Irons: containing
3-5%Ni, 1-4%Cr. Identified by the name Ni-Hard iron. The chromium-molybdenum
irons (high chromium irons): 11-23%Cr, 3%Mo, and sometimes additionally alloyed with
Ni or Cu. 25-28%Cr White Irons: contain other alloying additions of Molybdenum and/or
Nickel up to 1.5% 497
White Cast Iron
Typical white cast iron contains 2.5-3.5% C, 0.4-1.5% Si, 0.4-0.6% Mn, 0.1-0.4% P, 0.15%
S, and balance Fe.
During solidification, high internal stresses may be developed due to varying cooling rates
across the cross section . These stresses can be relieved by heating white cast iron to about
500-550°C
Pearlite
400 X
498
White cementite network
Grey Cast Iron
Gray cast iron is obtained by cooling the molten metal slowly during solidification. A
typical gray cast iron contains 2.5-3.5% C, 1.4-2.8% Si, 0.5-0.8% Mn, 0.1-0.9% P, and
0.06-0.12% S.
Fractured surface of gray cast iron appears grey because of the presence of graphite. Hence
the alloy is termed gray cast iron.
In the manufacture of gray cast iron the tendency of cementite to separate into graphite and
austenite or ferrite is favored by controlling alloy additions and cooling rates.
These alloys solidify by first forming primary austenite. The initial appearance of combined
carbon is in the cementite resulting from the eutectic reaction. With proper control of carbon
content, temperature and the proper amount of graphitizing elements notably silicon,
magnesium cerium and, alloy will follow the stable iron – graphite equilibrium diagram.
For most of these cast irons, the graphite exists in the form of flakes (similar to corn flakes),
which are normally surrounded by an α-ferrite or pearlite matrix.
Mechanically, gray iron is comparatively weak and brittle in tension as a consequence of its
microstructure; the tips of graphite flakes are sharp and pointed, and may serve as points of
stress concentration when an external tensile stress is applied. Strength and ductility are
much higher under compressive loads.
Gray irons are very effective in damping vibrational energy. Base structures for machines
and heavy equipment that are exposed to vibrations are frequently constructed of this
material. In addition, gray irons exhibit a high resistance to wear. 499
Grey Cast Iron
Furthermore, in the molten state they have a high fluidity at casting temperature, which
permits casting pieces having intricate shapes; also, casting shrinkage is low.
Finally, and perhaps most important, gray cast irons are among the least expensive of all
metallic materials.
Gray irons having different type of microstructures, it may be generated by adjustment of
composition and/or by using an appropriate heat treatment.
For example, lowering the silicon content or increasing the cooling rate may prevent the
complete dissociation of cementite to form graphite. Under these circumstances the
microstructure consists of graphite flakes embedded in a pearlite matrix.
Tensile strength of gray cast iron varies from 100 Mpa to 340 Mpa
Comparison of the relative

vibrational damping capacities
of (a) steel and (b) gray cast
iron.
500
Grey Cast Iron
∈ [2.4% (for good castability), 3.8 (for OK mechanical propeties)]
< 1.25% → Inhibits graphitization
< 0.1% → retards graphitization; ↑ size of Graphite flakes
Fe-C-Si + (Mn, P, S)
→ Invariant lines become invariant regions in phase diagram
Si ∈ (1.2, 3.5) → C as Graphite flakes in microstructure (Ferrite matrix)
↑ volume during solidification ⇒ better castability
Most of the ‘P’ combines with the iron to form iron phosphide (Fe3P).This iron
phosphide forms a ternary eutectic known as steadite, contains cementite and
austenite (at room temperature pearlite).
L → γ + ( Fe3C ) → α + Fe3C + ( Fe3C )

14243 1424 3 suuuuuuuu
Ledeburite Pearlite Si ↑ ⇒ Ceutectoid
501
Grey Cast Iron
Graphite flakes
500 X
Ferrite matrix Pearlite matrix
502
Grey Cast Iron : Heat treatment
The heat treatment given to grey cast iron may be classified is
Stress relieving
Annealing
Ferritizing annealing
Full annealing
Graphitizing annealing
Normalizing
Hardening and Tempering
Stress Relieving
The purpose of stress relieving in grey cast iron is to relieve residual stresses introduced
during solidification, due to different cooling rates prevalent at various section of castings.
Residual stresses adversely affect strength and cause distrotion and may even result in
cracking in some cases.
The temperature of stress relieving is kept much below Ac1 temperature. For maximum
stress relief without changing the microstructure, a temperature range 538-565°C is
recommended.
In this temperature range, about 80% of the residual stresses are removed with holding time
of about one hour. When held at 590°C, more than 85% of the stresses can be removed
503
Annealing
The purpose of annealing is to soften the grey cast iron and to improve its machinability by
minimizing or eliminating massive eutectic carbides.
Three types of annealing treatment are given to grey cast iron : ferritizing annealing, full
annealing, and graphitizing annealing.
FERRITIZING ANNEALING : To improve machinability in unalloyed or low alloy grey
cast iron of normal composition, ferritizing annealing treatment is carried out.
In this treatment, pearlitic carbide transforms to ferritic matrix and graphite. Ferritic matrix
is soft. Above 590°C, the rate of decomposition of iron carbide to ferrite and graphite
increases significantly, and at 760°C the rate of decomposition is maximum.
Therefore for most gray cast irons, the ferritizing annealing temperature is kept between
700°C and 760°C, and the recommended holding time is 1 hour per 25 mm of section.
Normally, After ferritizing annealing, the annealed specimens are cooled at a rate varying
from 100°C/hour to 300°C/hour.
FULL ANNEALING : If alloy content is high in grey cast iron, then it is difficult to
decompose iron carbide into ferrite and graphite at 760°C.
Therefore, under such conditions, full annealing treatment is adopted. For this treatment,
casting is heated to a suitable temperature between 790°C and 900°C and held about 1 hour.
Then it is cooled slowly between 790°C and 680°C.
504
Annealing
GRAPHITIZING ANNEALING : The purpose of graphitizing annealing is to convert
massive iron carbide into pearlite and graphite. The treatment is carried out at 900-955°C.
At this temperature, massive carbide dissolves in austenite which later decomposes to
pearlite and graphite on cooling.
Above 925°C, Fe3P may also melt. Holding time varies from 15 minutes to several hours.
Normalizing
The purpose of normalizing treatment for grey cast is to improve mechanical properties
such as hardness and tensile strength.
This treatment also helps restore as cast properties which have been modified by other
heating processes such as graphitizing or pre-heat and post-heat treatment of welded joints.
Normalizing temperature is kept above transformation range, i.e., 885-925°C. Holding time
recommended for gray cast iron at normalizing temperature is about 1 hour per 25 mm of
maximum thickness.
Heating temperature significantly affects the mechanical properties, e.g., hardness and
tensile strength as also the microstructure.
Alloying elements such as Cr, Mo, and Ni enhance the strengthening due to normalizing.
505
Hardening and Tempering
The purpose of hardening and tempering in grey iron is to improve its strength and wear
resistance. After this treatment, the wear resistance of pearlitic grey cast iron increases four
to five times.
Ordinarily, grey iron is furnace or salt bath hardened from a temperature of about 860-
870°C. The transformation range can extend by more than 55°C above the Ac1 temperature.
Approximately Ac1 temperature of unalloyed grey iron is related to silicon and Manganese
content by the relation
Ac1 (°C) = 730 + 28.0 (%Si) – 25.0 (%Mn)
Tempering treatment increases the toughness and relieves internal stresses which may be
developed during quenching. Due to tempering hardness decreases.
To achieve maximum toughness in grey iron, tempering temperature of about 370°C is
recommended. After tempering at this temperature, the matrix retains a hardness level of
472 BHN.
Increase in wear resistance in grey cast iron is achieved by producing a structure consisting
of graphite embedded in a martensitic matrix through heat treatment.
Flame or induction hardening of grey iron is not so common as furnace hardening because,
for adopting the first two methods, a relatively large content of combined carbon is required
since very little time is available for carbon to dissolve in austenite. Water is often used as a
quenchant with flame or induction hardening where only the outer case is hardened. 506
Ductile Cast Iron : Spheroidal Graphite (SG) iron
Ductile cast iron also called as Nodular cast iron, and Spherulitic cast iron.
Graphite nodules instead of flakes (in 2D section)
Mg (0.03-0.06)%, Ce, Ca (or other spheroidizing) elements are added
The elements added to promote spheroidization react with the solute in the liquid to form
heterogeneous nucleation sites
The alloying elements are injected into mould before pouring
It is thought that by the modification of the interfacial energy the ‘c’ and ‘a’ growth
direction are made comparable leading to spheroidal graphite morphology
The graphite phase usually nucleates in the liquid pocket created by the pro-eutectic γ
As compared to flaky graphite in grey cast iron, spheroidal graphite does not weaken the
matrix considerably. For this reason the mechanical properties of SG iron are superior to
gray iron.
SG iron has tensile strength of 400-700 Mpa, Yield strength 270-390 Mpa, and percentage
of elongation 10-20.
Approx. chemical composition of SG iron is 3.0-3.6%C, 2.0-2.5%Si, 0.6%Mn, 0.04 max P,
0.04 max S, and balance Fe.
507
Ductile Cast Iron : Spheroidal Graphite (SG) iron
Ferrite Graphite nodules
10 µm
With Ferritic Matrix With (Ferrite + Pearlite) Matrix
With Pearlitic matrix

508
Spheroidal Graphite (SG) iron : Heat treatment
The normalizing, hardening, and austempering heat treatment, which involve
Austenitisation, followed by controlled cooling or isothermal reaction, or a combination of
the two, can produce a variety of microstructures and greatly extend the limits on the
mechanical properties of ductile cast iron.
These microstructures can be separated into two broad classes
Those in which the major iron-bearing matrix phase is the thermodynamically stable
body-centered cubic (ferrite) structure.
Those with a matrix phase that is a meta-stable face-centered cubic (austenite)
structure. The former are usually generated by the annealing, normalizing, normalizing
and tempering, or quenching and tempering processes.
The latter are generated by austempering, an isothermal reaction process resulting in a
product called austempered ductile iron (ADI).
Other heat treatments in common industrial use include stress-relief annealing and selective
surface heat treatment. Stress-relief annealing does not involve major micro-structural
transformations, whereas selective surface treatment (such as flame and induction surface
hardening) does involve microstructural transformations, but only in selectively controlled
parts of the casting
509
Stress Relieving: 540 - 595˚C reduces
warping and distortion during subsequent
machining.
Annealing: Full feritizing Annealing used to
remove carbides and stabilized Pearlite, heat
to 900˚C, holding long enough to dissolve
carbides, then cooling at a rate of 85˚C/h to
705˚C, and still air cooling to room
temperature. Also improves low temperature
fracture resistance but reduces fatigue
strength.
Normalizing, Quenching, And Tempering:
heated to 900˚C, held for 3 hours (allowing
1h/25mm or 1h/in of cross section to reach
that temperature). The air blasted or oil
quenched. Followed by tempering between
540 - 675˚C. Used to achieve grade 100-70-
03. Hardness’s as high as 255HB, increase Relationship between austempering
short time fatigue strength, decreases fatigue temperature and the strength and ductility
life. Increases tensile and yield strengths at of a 1.5Ni-03.Mo alloyed ductile iron.
the expense of ductility. Austenitizing temperature was 900˚C.
510
AUSTEMPERING: Used to achieve Austempered Ductile Iron, requires two stages. Stage
1: heating to and holding at 900˚C. Stage 2: quenching and isothermally holding at the
required austempering temperature, usually in a salt bath. austempering temperatures
shown in figure below.
SURFACE HARDENING : SG iron are also flame or induction hardened. Pearlite type of
SG irons are preferred for flame or induction hardening as time required for austenitizing is
comparatively small. in the case pearlitic SG iron, tempering is carried out at 595-650°C for
1 hour plus 1 hour per 25 mm of section thickness to remove virtually all internal stresses
before going in for flame or induction hardening.
The possible applications of S.G Iron are very wide. The use of S.G Iron is suggested where
improved properties are dictate a replacement of other material or where the use of S.G Iron
will permit an improvement in the design.
Some popular uses of S.G Iron for various engineering application are for –Support bracket
for agricultural tractor, Tractor life arm, Check beam for lifting track, Mine cage guide
brackets, Gear wheel and pinion blanks and brake drum, Machines worm steel, Flywheel,
Thrust bearing, Frame for high speed diesel engine, Four throw crankshaft, etc.,
511
Spheroidal Graphite (SG) iron : Nodular cast iron
Ferrite (White)
Graphite (black)
Bull’s Eye
Ferrite
5 µm
Pearlite (grey)
512
Malleable Cast Iron
White Cast Iron 
Malleabilize
To Increase Ductility
→ Malleable Cast Iron
As we discussed earlier that cementite is actually a metastable phase. There is a tendency
for cementite to decompose into iron and carbon. This tendency to form free carbon is the
basis for the manufacture of malleable cast iron.
The reaction of Fe3C 3Fe + C is favored by elevated temperatures, the existence of solid
nonmetallic impurities, higher carbon contents, and the presence of elements that aid the
decomposition of Fe3C.
Categorized into 3 categories: Ferritic, Pearlitic, and Martensitic Malleable Cast Iron.
ferritic malleable Irons require a two stage annealing cycle. first: converts primary
carbides to temper Carbon.
Second: converts Carbon dissolved in Austenite at the first-stage annealing temperature to
temper Carbon and Ferrite. Consists of temper Carbon in a matrix of Ferrite. Contain a
slight amount of controlled Iron.
Pearlitic :1st stage identical to that of Ferrite. Casting is slowly cooled to approx 870˚C.
When the combined Carbon content of the Austenite is reduced to about .75% the castings
are air cooled. Usually air blasted to avoid the formation of ferrite around the temper
Carbon particles. Then, the castings are tempered to specified time.
513
Malleable Cast Iron
• (940-960)°C (Above eutectoid temperature)
Stage I • Competed when all Cementite → Graphite
A: Low T structure (Ferrite + Pearlite + Martensite) → (γ + Cementite)
B: Graphite nucleation at γ/Cementite interface

(rate of nucleation increased by C, Si)
(Si ↓ solubility of C in γ ⇒ ↑ driving force
for growth of Graphite)
C: Cementite dissolves → C joining growing Graphite plates
Spacing between Cementite and Graphite →

↓ spacing ⇒ ↓ time (obtained by faster cooling of liquid)
Time for Addition of Alloying elements

Graphitization → which increase the nucleation rate of Graphite temper nodules
in Stage I
Si ↑ ⇒ t ↓
514
Malleable Cast Iron
• (720-730)°C (Below eutectoid temperature)
Stage II • After complete graphitization in Stage I → Further Graphitization
Slow cool to the lower temperature such that γ does not form Cementite
C diffuses through γ to Graphite temper nodules
(called Ferritizing Anneal)
Full Anneal in Ferrite + Graphite two phase region
Partial Anneal (Insufficient time in Stage II Graphitization)
γ → Ferrite is partial and the remaining γ transforms to Pearlite
⇒ γ → Pearlite + Ferrite + Graphite
If quench after Stage I ⇒ γ → Martensite (+ Retained Austenite(RA))
(Graphite temper nodules are present in a matrix of Martensite and RA)
> 48 hrs
Fe3C (WCI)  
2 stage heat treatment
→ Graphite Temper Nodules (Malleable Iron)
515
Malleable Cast Iron
Pearlitic Matrix Ferrite (White)
Graphite (black)
Pearlite (grey)
Ferritic Matrix
Partially Malleabilized Iron
→ Incomplete Ferritizing Anneal
Ferrite (White)
Graphite (black)
10 µm
Fully Malleabilized Iron
→Complete Ferritizing Anneal
516
Compacted graphite Cast Iron
A relatively recent addition to the family of cast irons is compacted graphite (CGI).
Micro structurally, the graphite in CGI alloys has a worm-like (or vermicular) shape; a
typical CGI microstructure. In a sense, this microstructure is intermediate between that of
gray iron and ductile iron and, in fact, some of the graphite (less than 20%) may be as
nodules.
The chemistries of CGIs are more complex
than for the other cast iron types; compositions
of Mg, Ce and other additives must be
controlled so as to produce a microstructure Vermicular shape
that consists of the worm-like graphite Ferritic Matrix
particles, while at the same time limiting the
degree of graphite nodularity, and preventing
the formation of graphite flakes. Furthermore,
depending on heat treatment, the matrix phase Graphite Nodules
will be pearlite and/or ferrite.
CGI are now being used in a number of
important applications – these include: diesel
engine blocks, exhaust manifolds, gearbox
housings, brake discs for high speed trains and
flywheels. 517
Alloy Cast Iron
Cr, Mn, Si, Ni, Al
↑ the range of microstructures
Beneficial effect on many properties
↑ high temperature oxidation resistance
↑ corrosion resistance in acidic environments
↑ wear/abaration resistance
Graphite free
Alloy Cast Irons
Graphite bearing
518
Chromium addition (12-35 wt %)
Excellent resistance to oxidation at high temperatures
High Cr Cast Irons are of 3 types:
12-28 % Cr matrix of Martensite + dispersed carbide
29-34 % Cr matrix of Ferrite + dispersion of alloy carbides
[(Cr,Fe)23C6, (Cr,Fe)7C3]
15-30 % Cr + 10-15 % Ni stable γ + carbides [(Cr,Fe)23C6, (Cr,Fe)7C3]
Ni stabilizes Austenite structure
29.3% Cr, 2.95% C 519

Nickel Hard Cast Iron
Stabilizes Austenitic structure
↑ Graphitization (suppresses the formation of carbides)
(Cr counteracts this tendency of Ni for graphitization)
↓ Carbon content in Eutectic
Moves nose of TTT diagram to higher times ⇒ easy formation of Martensite
Carbide formation in presence of Cr increases the hardness of the eutectic
structure → Ni Hard Cast Irons (4%Ni, 2-8% Cr, 2.8% C)
Good abrasion resistance
Needles of Martensite
Transformation sequence
Crystallization of primary γ
Eutectic liquid → γ + alloy carbide
4%Ni, 2-8% Cr, 2.8% C γ → Martensite
520
Nickel Resist Cast Iron
Ni Resist Iron: 15-30% Ni + small amount of Cr:

Austenitic Dendrites + Graphite plates/flakes + interdendritic carbides
due to presence of Cr
Resistant to oxidation (used in chemical processing plants, sea water, oil
handling operations…)
Graphite plates
Dendrites of γ
521
Silicon (Silal) Cast Iron
Silal Iron (trade name): Alloy CI with 5% Si
Si allows solidification to occur over larger temperature range →
promotes graphitization
Forms surface film of iron silicate → resistant to acid corrosion
522
Some other types of Cast Irons
Chilled Cast Iron
Chilled –iron castings are made by casting the molten metal against a metal
chiller, resulting in a surface of white cast iron. This hard, abrasion-resistant
white iron surface or case is backed up by a softer gray iron core.
This case-core structure is obtained by careful control of the overall alloy
composition and adjustment of the cooling rate.
Mottled Cast Iron
Solidifying at a rate with extremes between those for chilled and gray irons, thus
exhibiting micro structural and metallurgical characteristics of both
High-Alloy Graphitic Irons
Produced with microstructures consisting of both flake and nodule structures.

Mainly utilized for applications requiring a combination of high strength and
corrosion resistance.
523
Graphite flake - Gf
Graphite nodules - Gn
Graphite rosettes - Gr
524
Questions?
1. Compare the processes of ferritizing annealing and graphitizing annealing
2. How hardening and tempering is carried out for gray cast irons?
3. What heat treatment would you recommend for malleable cast iron and why?
4. What do you undesrstand by malleabilizing of cast irons?
5. Discuss the heat treatment procedures adopted for S.G. iron castings.
6. How austempering is achieved in S.G. iron castings? What are the possible applications of
austempered S.G. iron?
7. Assume that a C clamp is to made of cast iron. Select a suitable a type of cast iron and
explain the reasons for the selection.
8. On the basis of microstructure, briefly explain why gray iron is brittle and weak in tension.
9. Compare gray, malleable, nodular and white cast irons with respect to (a) composition and
heat treatment, (b) microstructure, and (c) mechanical characteristics.
10. Discuss the influence of the following elements on the structure and properties of cast iron.
(i) Si (ii) Mn (iii) S and (iv) P
525
THANK YOU...!
526

Lecture1460085501 2

Uploaded by

Lecture1460085501 2

Uploaded by

LECTURE NOTES ON

PHASE TRANSFORMATIONS & HEAT TREATMENT

6th Semester, MME

For Batches 2013-

Avala Lava Kumar

Chapter Chapter Name Page No

AMORPHOUS QUASICRYSTALS RATIONAL CRYSTALS

Segmented Structures: are Hybrids are

Atom Structure Microstructure Component

Dislocation Stress fields

*Simple Unit Cells 10

Phases Defects Residual stress

Phase Defect structure Stress-state

Phases Transformations Microstructural Transformations

Thermodynamics Order of a phase transformation

Classification of Based on Reconstructive Civilian

Non-quenchable Athermal → Rapid

Heterogeneous Transformations Homogeneous Transformations

GROWTH CONTROLLED GROWTH CONTROLLED BY GROWTH CONTROLLED BY A THERMAL

SHORT RANGE TRANSPORT LONG RANGE TRANSPORT

No heat flows No heat flows No heat flows

First law of Thermodynamics

Entropy S is defined by the equation

If we divide by dT on both sides we get dH δQ dP

HT , H0 , and ∆H298 are the enthalpies at temperature, T, 0, and 298K respectively

Phase fields of non-close packed

Phase fields of close packed structures

These lines slope downward as: Under

From the definition, we know that the slope of

dG = dE + PdV + Vdp − TdS − SdT (dE = δQ − PdV )

If we want to know whether a solid or liquid phase will be stable at a particular

For that we need to know the variation of specific

Therefore, at a temperature T ∆G = ∆H − T∆S →①

Unlike single component system, where we

Total free energy before mixing G0 = X AGA + X BGB

After mixing there will be change in free energy

We take the following assumptions:

The change in internal energy

Situation 1: Enthalpy of mixing is zero

Situation 2: Enthalpy of mixing is less than zero

∆Hmix = ΩXAXB < 0

Situation 3: Enthalpy of mixing is greater than zero

∆Hmix = ΩXAXB > 0

Because of transformation internal energy will

Since we are considering transformation at a particular temperature, the change in entropy

S = l ln ω k is the Boltzmann constant

If we consider the random solid solution, then

So, ∆Smix can be written as

∆Smix = k lnω = k{[(nA + nB ) ln(nA + nB ) − (nA + nB )] −[nA lnnA − nA ] −[nB lnnB − nB ]}

Mixed states with ‘various’ degrees of mixing

Note: the profoundness of the concept of entropy comes

* We assume that all states have equal probability of

= X AGA + X BGB + RT[ X A ln X A + X B ln X B ]

With the increase in temperature, -T∆Smix

The values of GA and GB also will decrease.

Following the slope Go might change since

G= XAGA + XBGB +ΩXAXB + RT[XA lnXA + XB lnXB]

Here both ∆Hmix and -T∆Smix are negative

With increasing temperature G will become even

Note that here also G0 may change the slope

G0 Positive ∆Hmix -T∆Smix

This kind of phase diagram is found

In this system, there are two solid

Further, we can write X A X B = X A2 X B + X A X B2

Further, comparing the expressions for free energy, we can write

So the relations between the chemical potential and activity are

Activities and chemical potentials are

Further, we can write

γi are the activity coefficient of element i.

∆HV is the increase in enthalpy because of one mole of vacancies

Total entropy of mixing ∆S mix = ∆SV X V − R[ X V ln X V + (1 − X V ) ln(1 − X V )]

Note here that G of element A when vacancies are

The relation for equilibrium concentration of vacancies can be written as

Although there is increase in configurational entropy because of the presence of vacancies,

From the equation in previous slide, we can also write

The concentration of vacancies in mole