SECOND EDITION

A
X
B

C
A
Y

Physical
Metallurgy
PRINCIPLES AND PRACTICE

V. RAGHAVAN
PHYSICAL METALLURGY
PHYSICAL
METALLURGY
PRINCIPLES AND PRACTICE
[Second Edition]

V. RAGHAVAN
Formerly Professor
Indian Institute of Technology Delhi

New Delhi-110001
2012
Rs. 195.00

PHYSICAL METALLURGY: Principles and Practice, Second Edition

V. Raghavan

2006 by Prentice-Hall of India Private Limited, New Delhi. All rights reserved.
No part of this book may be reproduced in any form, by mimeograph or any other
means, without permission in writing from the publisher.

ISBN-978-81-203-3012-2

The export rights of this book are vested solely with the publisher.

Twenty-third Printing (Second Edition) June, 2012

Published by Asoke K. Ghosh, Prentice-Hall of India Private Limited, M-97,

Connaught Circus, New Delhi-110001 and Printed by Mohan Makhijani at Rekha
Printers Private Limited, New Delhi-110020.
 CONTENTS

Preface ix
Preface to the First Edition xi
Physical Constants xiii
Conversion Factors xv

1. STRUCTURE OF METALS AND ALLOYS 128

1.1 Nature of Metallic Bonding 1
1.2 Crystal Structures of Metals 2
1.3 Structure of Alloys 10
1.4 Imperfections in Crystals 14
1.5 Experimental Study of Structure 22
References 27
Questions 27

2. PHASE DIAGRAMS 2964

2.1 The Phase Rule 29
2.2 Binary Phase Diagrams 30
2.3 Free Energy Composition Curves for Binary Systems 37
2.4 Microstructural Changes during Cooling 40
2.5 The Iron Carbon Equilibrium Diagram 42
2.6 Effect of Alloying Elements on the Fe-C Diagram 46
2.7 The Copper-Zinc Phase Diagram 51
2.8 Ternary Phase Diagrams 52
2.9 Quaternary Phase Diagrams 57
2.10 Experimental Determination of Phase Diagrams 58
References 63
Questions 64

3. PHASE CHANGES 6599

3.1 Types of Phase Changes 65
3.2 Diffusion in Solids 66
3.3 Nucleation and Growth Kinetics 67
3.4 Solidification 71
3.5 Pearlitic Transformations 79

v
vi Contents

3.6 Bainitic Transformations 83

3.7 Martensitic Transformations 84
3.8 Precipitation and Age Hardening 92
3.9 Recovery, Recrystallization and Grain Growth 96
References 99
Questions 99

4. HEAT TREATMENT OF STEELS 100126

4.1 The T-T-T Diagrams 100
4.2 The C-C-T Diagrams 104
4.3 Function of Alloying Elements 106
4.4 Formation of Austenite 107
4.5 Typical Heat Treatment Processes 109
4.6 Hardenability 112
4.7 Quench Cracks 119
4.8 Case Hardening 122
References 125
Questions 126

5. MECHANICAL PROPERTIES 127160

5.1 The Tensile Test 127
5.2 Plastic Deformation Mechanisms 134
5.3 Methods of Strengthening Metals against Yield 137
5.4 Creep Resistance 143
5.5 Fracture 147
5.6 Fatigue Failures 153
References 159
Questions 159

6. CORROSION AND ITS PREVENTION 161184

6.1 The Galvanic Cell 161
6.2 The Electrode Potentials 164
6.3 Polarization 167
6.4 Passivation 169
6.5 General Methods of Corrosion Prevention 172
6.6 Cathodic Protection 173
6.7 Coatings 175
6.8 Corrosion Prevention by Alloying 177
6.9 Stress Corrosion Cracking 179
References 183
Questions 183
 Contents vii

7. ENGINEERING ALLOYS 185225

7.1 Steel Specifications 185
7.2 Low Carbon Steels 188
7.3 Mild Steels 191
7.4 Medium Carbon Steels 195
7.5 High Strength Structural Steels 197
7.6 Tool Materials 199
7.7 Stainless Steels 203
7.8 High Temperature Alloys 205
7.9 Cast Irons 207
7.10 The Light Alloys 213
7.11 Copper and Its Alloys 218
7.12 Bearing Alloys 223
References 224
Questions 225

INDEX 227232
 PREFACE

In the Second Edition of this book, several new sections such as thermoelastic
martensite, shape memory alloys, rapid solidification processing and quaternary
phase diagrams have been included at appropriate places. The text has been
updated and rewritten for greater clarity. The quality of diagrams has been
improved and their numbers increased. The list of references at the end of each
chapter has been updated.
I wish to thank Ms R. Malar for her assistance in preparing the diagrams.
Thanks are due to the editorial and production team of the Publishers,
Prentice-Hall of India, for their assistance and support.

V. RAGHAVAN

ix
 PREFACE TO THE FIRST EDITION

Physical Metallurgy deals with the relationship between structure and

properties of metals and alloys. It forms a vital link in the processes of making,
shaping and heat treating of metals. It thus interfaces with the other areas of
metallurgy such as process metallurgy, mechanical metallurgy and engineering
metallurgy. Voluminous books are available which deal with the fundamentals
of this subject in a detailed and rigorous manner. In addition, there are books
dealing with specific topics such as the structural aspect at an advanced level.
At the other end, some books present just factual data on the properties and
uses of various grades of steels and nonferrous alloys. There is thus a clear need
for a book that gives a broad introduction to the subject dealing with the
principles at a reasonable depth and at the same time discussing their relationship
to practice. This book aims to do that.
All undergraduate students of engineering and metallurgy study phase
diagrams, heat treatment principles, mechanical behaviour of alloys, corrosion
phenomena, and properties and uses of specific engineering alloys. Undergraduate
students of metallurgy have an introductory course on physical metallurgy in
their early years. Students of engineering, specially mechanical and chemical
engineering, study these topics in one course or the otherwhether it be
metallurgy, engineering materials or materials science and metallurgy. This book
is intended for such students. In addition, the book is designed to be used by
students doing self-study and preparing for the associate membership
examinations of the Indian Institute of Metals (AMIIM), the Institution of
Mechanical Engineers, the Institution of Engineers, etc. The simple style of the
book should make it attractive to them.
The opening chapter covers the structural aspects with a discussion of the
unit cells of metallic crystals, solid solutions, intermediate phases and structural
imperfections. Equilibrium diagrams are introduced in the second chapter, with
emphasis on the binary diagrams. Short sections on ternary systems and on
experimental methods of determining phase diagrams are included. The major
topics under phase changes are solidification, steel transformations and
precipitation processes in nonferrous alloys. A separate chapter is devoted to the
heat treatment of steels, with a detailed discussion of the concept of hardenability.
The chapter on mechanical properties not only covers the basic concepts on
deformation and strengthening mechanisms, but also deals with applications to
creep, fatigue and fracture. Even though corrosion is a peripheral topic in
physical metallurgy, in view of its importance, it is discussed in a separate

xi
xii Preface to the First Edition

chapter. The last chapter is totally devoted to practice, giving information on

structure, treatment, properties and uses of individual alloys with their
specifications, covering most types of steels and common nonferrous alloys.
The author wishes to thank his colleagues in the metallurgy group of the
Department of Applied Mechanics for their assistance. Support received from
IIT Delhi for preparing the manuscript and the diagrams is acknowledged.
Encouragement from Dr. D.P. Antia, Shri C.V. Sundaram and Professor A.K.
Seal of the Indian Institute of Metals is appreciated.

New Delhi V. RAGHAVAN

July 20, 1983
 PHYSICAL CONSTANTS

Avogadros number N = 6.023 1023 mol1

Boltzmanns constant k = 1.380 1023 J K1
= 8.614 105 eV K1
Gas constant R = 8.314 J mol1 K1
Plancks constant h = 6.626 10 34 J s
Electronic charge e = 1.602 1019 C
Electron rest mass m0 = 9.109 1031 kg
Velocity of light c = 2.998 108 m s1
Bohr magneton (magnetic moment) mB = 9.273 1024 A m2
Permittivity of free space e0 = 8.854 1012 F m1
Permeability of free space m0 = 4p 107 H m1
= 1.257 106 H m1
Faradays constant F = 96.49 kC mol1 (of electrons)
Atomic mass unit (amu) 1/(103N) = 1.660 1027 kg
Acceleration due to gravity g = 9.81 m s2

xiii
 CONVERSION FACTORS
1 inch = 25.4 mm
1 nm = 109 m
1 = 1010 m = 0.1 nm
1 = 1/57.3 rad

T C = (T + 273.15) K
T F = 5/9(T + 459.67) K
1 per F = 9/5 K1

1 kgf = 9.81 N
1 lb = 4.45 N
1 dyne = 105 N
1 dyne/cm = 103 N m1
1 atmosphere = 0.101 325 MN m2
1 bar = 101 MPa
1 psi = 6.89 kN m2
1 ksi (103 psi) = 6.89 MN m2
1 ton/sq.in. = 15.46 MN m2
1 kgf/cm2 = 98.1 kN m2
1 kgf/mm2 = 9.81 MN m2
1 dyne/cm2 = 0.1 N m2
1 torr (mm of Hg) = 133.3 N m2
1 kgf/mm3/2 = 0.310 MN m3/2
1 ksi in = 1.10 MN m3/2

1 eV = 1.602 1019 J
1 erg = 107 J
1 calorie = 4.18 J
1 eV/entity = 96.49 kJ mol1
1 erg/cm = 105 J m1
1 erg/cm2 = 103 J m2
1 erg/cm3 = 0.1 J m3
1 lb-in/in2 = 175 J m2

1 lb/cu.in = 27 680 kg m3
1 cm/cm3 = 104 m m3

xv
xvi Conversion Factors

1 mole/cm2/sec = 104 mol m2 s1

1 mole/cm3 = 106 mol m3
1 cm2/sec = 104 m2 s1
1 mA/cm2 = 10 A m2
1 A hr = 3.6 kC
Chapter 1
STRUCTURE OF METALS
AND ALLOYS

Physical metallurgy deals with the relationship between structure and properties
of metals and alloys. The structure is influenced significantly by the process
adopted for metal extraction from the ores and subsequent fabrication. The
properties of interest are mainly those demanded by engineers such as strength,
ductility, toughness and corrosion resistance. Thus physical metallurgy inter-
faces with areas of process or chemical metallurgy, mechanical metallurgy and
engineering metallurgy. It provides a vital link in making, shaping, treating and
using metals and alloys. In this introductory textbook, the principles and the
corresponding practice of physical metallurgy are outlined in simple terms,
without going into a rigorous mathematical treatment of the structure-property
relationships.

1.1 NATURE OF METALLIC BONDING

The atoms in a metal are held together by chemical bonds known as metallic
bonds. The metallic bonding has certain unique characteristics as listed below:
1. A metallic crystal can be visualized as an array of positive ions, with a
common pool of electrons to which all the metal atoms have contributed
their outermost electrons. The electrons in this pool have freedom to move
through the crystal to any part and virtually in any direction. They are
known as free electrons. The attractive force between the positive ions and
the negatively-charged free electrons is what binds the metal atoms together
in the crystal.
1
2 Physical Metallurgy

2. The free electrons carry both electric current and heat through the crystal.
They are responsible for the excellent electrical and thermal conductivity
of metals.
3. The bonds formed by the free electrons are nondirectional. There are no
constraints about bond angles. An atom in a metallic crystal is bonded to
as many neighbours as space permits. This provides considerable flexibility
in the directions along which the metal atoms can be displaced during
plastic deformation. This is mainly responsible for the unique properties
of high ductility and malleability that metals possess. For example, gold
can be drawn into extremely thin wires; aluminium can be rolled into foils
less than 0.1 mm thick.

1.2 CRYSTAL STRUCTURES OF METALS

Approximately three-fourths of the elements in the periodic table are metals.
More than two-thirds of them (more than 50) possess relatively simple crystal
structures. Understanding these structures and their characteristics is a first step
in the study of physical metallurgy.
Body centred cubic (BCC) crystal The reference unit (called the space lattice)
for this crystal is a cube, with lattice points at the eight corners and at the body
centre, as shown in Fig. 1.1a. In the unit cell of the crystal, atoms are positioned
one at each lattice point, i.e. at the eight corners and at the body centre of the
cube, as shown in Fig. 1.1b. The corner atoms are shared by other neighbouring
units. As eight units meet at any corner, the share of the unit under consideration
for a corner atom is one-eighth, see Fig. 1.1c. The atom at the body centre
belongs entirely to the unit.

Figure 1.1 The body centred cubic crystal: (a) the BCC space lattice, (b) the unit
cell of the BCC crystal, and (c) the sharing of corner atoms by the unit cell.

So, the effective number of atoms in the unit cell is

8 (corner atoms) 1/8 + 1 (body centred atom) = 2
The body centred atom is touching all the corner atoms. Each corner atom
in turn is touching eight body centred atoms of the neighbouring cells. The
 Structure of Metals and Alloys 3

number of nearest neighbours to a given atom is called the co-ordination number

Z. Z = 8 for the BCC crystal.
The distance between nearest atoms is used to define the radius r of the
atom. For the BCC crystal, the centre-to-centre distance of nearest neighbours
is clearly half of the body diagonal, a 3 / 2 , where a is the length of the cube
edge known as the lattice parameter. Noting from Fig. 1.1b that the nearest
neighbours touch each other, we have
r = a 3/4
for the BCC crystal.
The packing efficiency (P.E.) of atoms in the unit cell is defined as
volume of atoms in the unit cell
P.E. =
volume of the unit cell
2 (4 / 3)Q (a 3 / 4)3
=
a3
= 0.68
The density r of the crystal can be calculated as follows :
mass of atoms in the unit cell
S=
volume of the unit cell
For example, for BCC iron at room temperature,
lattice parameter a = 2.87 (2.87 1010 m)
mass of the Fe atom = 55.85 atomic mass units (amu)
1 amu = 1.66 1027 kg
2 55.85 1.66 1027
SFe =
(2.87 1010 )3
= 7840 kg m3 = 7.84 g/cm3
Typical examples of metals which form BCC crystals are given below.

Metal Cr Fe K Li Mo Na Nb Ta V W
Lattice 2.88 2.87 5.25 3.51 3.15 4.29 3.30 3.30 3.03 3.16
parameter,

Face centred cubic (FCC) crystal The face centred cubic space lattice is
shown in Fig. 1.2a. In the FCC crystal, there are eight atoms at the eight corners
of the cube, with six more atoms located at the six face centres, Fig. 1.2b. A
face centred atom is shared by two neighbouring unit cells, for which that face
is common, as can be seen from Fig. 1.2c. The effective number of atoms in
the FCC unit cell is
8 (corner atoms) 1/8 + 6 (face centred atoms) 1/2 = 4
4 Physical Metallurgy

In the FCC crystal, each atom is in contact with twelve neighbours. For
example, the face centred atom of the front face in Fig. 1.2b is in contact with
the four corner atoms on that face and the four face-centred atoms just behind
it (on the top, the bottom and the two side faces of the unit cell). In addition,
it is also in contact with four face-centred atoms of the unit cell in front of it
(not seen in Fig. 1.2). So the coordination number of the FCC crystal Z = 12.

Figure 1.2 The face centred cubic crystal: (a) the FCC space lattice, (b) the unit
cell of the FCC crystal, and (c) the sharing of atoms by the unit cell.

The distance between nearest neighbours is half of the face diagonal,

a 2 / 2 . The radius of the atom is r = a 2 / 4 for the FCC crystal.
The packing efficiency can be shown to be 0.74. Examples of metals form-
ing the FCC crystal are listed below.

Metal Ag Al Au Cu Fe(912- Ni Pb Pt
1394C)
Lattice 4.09 4.05 4.08 3.61 3.65 3.52 4.95 3.92
parameter,

Note that iron possesses more than one crystal form. It is BCC at low
temperatures, changes over to the FCC at 912C and back again to BCC at
1394C. This phenomenon is called polymorphism.
Hexagonal close packed (HCP) crystal The hexagonal space lattice is shown
in Fig. 1.3a. The atom arrangements within the HCP unit cell are shown in
Fig. 1.3b. There are six atoms at the six corners of the two hexagonal faces and
one each at their face centres. There are three atoms in the middle layer (where
there are no lattice points). Each corner of the hexagonal prism is shared by six
other prisms and the top and the bottom hexagonal faces are shared by two
cells, Fig. 1.3c. The effective number of atoms in the HCP unit cell is
12 (corner atoms) 1/6 + 2 (face centred atoms) 1/2 +
3 (middle layer atoms within the cell)* = 6.
*Thin slices of the three middle-layer atoms fall outside the unit cell, but these are exactly
compensated by three thin slices of outside atoms falling within the unit cell.
 Structure of Metals and Alloys 5

As in the FCC crystal, each atom here is in contact with twelve nearest
neighbours. For example, referring to Fig. 1.3b, the central atom on the top
hexagonal face touches the six surrounding atoms on the same face, three in the
middle layer and three in the plane above (not shown), yielding Z = 12.

Figure 1.3 The hexagonal close packed crystal: (a) the hexagonal space lattice,
(b) the HCP unit cell, and (c) the sharing of atoms by the unit cell.

The distance between nearest neighbours is the side of the hexagon a. So,
r = a/2 for the HCP crystal.
The height of the prism gives the second lattice parameter c. The ideal
packing efficiency can be shown to be the same as for the FCC crystal (0.74).
Typical examples of HCP crystals and their lattice parameters are listed below.

Metal Be Cd Mg Ti (below 882C) Zn Zr

a, 2.29 2.98 3.21 2.95 2.66 3.23
c, 3.58 5.62 5.21 4.68 4.95 5.15
c/a 1.56 1.89 1.62 1.59 1.86 1.59
6 Physical Metallurgy

Other less-commonly occurring metallic crystal structures include the tet-

ragonal, the orthorhombic and the rhombohedral crystals. In the tetragonal cell,
two sides of the unit cell are equal, but the third side is of a different length.
In the orthorhombic unit cell, all the three sides are of different lengths. The
angles between the axes in these two unit cells are the same as in the cubic
crystal (90). In the rhombohedral cell, the three axes are of the same length and
the angles between axes are equal to one another, but not 90.
Geometry of close packing For a fuller understanding of the FCC and HCP
crystalline arrangements, we need to examine the geometry of close packing
(meaning closest packing). Figure 1.4a shows a close packed plane A of equal-
sized spheres. The spheres cannot be put any closer than they are in Fig. 1.4a.
In building a three-dimensional structure, another close packed plane can be
placed over the first, such that its spheres fit into the triangular cavities between
the spheres of the first layer. There are two sets of such cavities, designated B
and C in Fig. 1.4a. The second plane of atoms, Fig. 1.4b or c, can rest vertically
over either the B positions or the C positions. If we assume that they rest over
B positions, then the third close packed plane can be stacked on the second in
two ways. The atoms of the third plane can be vertically over the C positions
or alternatively over the first-plane atoms, i.e. over A positions. These two
arrangements are designated ABC and ABA stackings respectively.

Figure 1.4 The geometry of close packing.

The FCC crystal is a stacking of close packed planes in the sequence

ABCABC.... In Fig. 1.5, the correspondence between the ABCABC... stacking
and the atoms of the unit cell is shown. This figure should be compared with
Fig. 1.2b, where the atoms are drawn to their correct sizes. The close packed
planes are inclined at an angle to the cube faces and are known as octahedral
planes.
The HCP crystal is a stacking of close packed planes in the sequence
...ABABA. Referring to Fig. 1.3b, if we call the close packed plane
corresponding to the bottom hexagonal face A, it is easily seen that the atoms
of the top hexagonal layer are vertically over those of the bottom layer and are,
 Structure of Metals and Alloys 7

therefore, to be designated also as A. The three atoms in the middle fit into the
triangular cavities between the top and the bottom layers and are B plane atoms.
The C positions are not used in this stacking.

Figure 1.5 The correspondence between ...ABCA stacking and the FCC unit cell.
The body diagonal XY is perpendicular to the close packed planes.

The ideal c/a ratio of HCP stacking is 1.633. The ratios in real crystals,
however, vary significantly from the ideal value, see table on page 5.
Miller indices The method devised by Miller is widely used to denote direc-
tions and planes in crystals. Let the vector r in Fig. 1.6a represent a direction
in a crystal (not cubic in this case). To find the Miller indices of this direction,
the components of the vector along the three axes are determined as multiples
of the unit vector corresponding to each direction.
The vector r has a component of 2 along the a-axis, 1 along b and 1 along
c. Note that the absolute magnitudes of the components along b and c are not
the same, but they are unity, when expressed as multiples of the corresponding
unit vectors. The Miller indices of this direction is given by [211], where the
square brackets stand for a direction. If there is a negative component, it is
denoted by a bar placed on top of the appropriate number.
In cubic crystals, it is convenient to have a notation to specify a family of
directions. The family is denoted by pointed brackets, e.g. <211> stands for
[211], [21 1 ], [121], [11 2 ], [121], etc. giving all combinations of the indices
both positive and negative. The cube edges in the unit cell are represented by
<100>, the face diagonals by <110> and the body diagonals by <111>.
8 Physical Metallurgy

Figure 1.6 The Miller indices of (a) a crystal direction and (b) a crystal plane.

For determining the Miller indices of a plane in a crystal, the following

procedure is used. Referring to Fig. 1.6b, the steps are :
1. Find the intercepts of the plane on the three coordinate axes 2 1
2. Take reciprocals 1 2
3. Convert into smallest integers in the same ratio 1 2 4
4. Enclose in parentheses (124)
A family of planes is denoted by curly brackets { }. For example, the faces
of a cube are denoted by {100}. The octahedral planes in Fig. 1.5 are {111}.
Example
1. Draw a close packed plane within the unit cell of an FCC crystal and give
its Miller indices.
2. Mark the close packed directions that lie on this plane and give their
Miller indices.
3. Draw a (110) plane within the same unit cell and find the line of intersec-
tion of the two planes.
Solution
1. Referring to Fig. 1.5, the close packed planes in the FCC crystal are of
the {111} type. Using D as the origin in Fig. 1.7, the Miller indices of the
plane sketched is (1 1 1) .
2. The atoms are touching one another along the face diagonals of an FCC
crystal. There are three face diagonals that lie on the above plane, corre-
sponding to the three close packed directions. Their Miller indices are :
Taking origin at A, AB = [011]
at B, BC = [1 0 1]
at C, CA = [1 1 0]
 Structure of Metals and Alloys 9

Figure 1.7 Planes and directions within the cubic unit cell.

3. Taking C as the origin, the (1 1 0) plane is drawn. The line of intersection

of the two planes is BE.
Taking origin at B, BE = [1 1 2]
Note: As the choice of origin is arbitrary, it is chosen such that the
directions start from the origin and the planes make intercepts on the three
axes within the unit cell.
Miller-Bravais indices For hexagonal crystals, a four-digit notation hkil, known
as Miller-Bravais indices, is used. The use of such a notation enables
crystallographically equivalent planes or directions in a hexagonal crystal to be
denoted by the same set of indices. Three of the axes a1, a2 and a3 are
coplanar and lie on the basal plane of the hexagonal prism, Fig. 1.3, with a
120 angle between them. The fourth axis is the c axis perpendicular to the
basal plane.
The determination of the Miller-Bravais indices is illustrated in Fig. 1.8 for
the two examples given below. The indices of a prismatic plane (one of the
vertical faces of the prism) is of the type (10 1 0) . It makes an intercept of 1
and 1 along a1 and a3 axes respectively and is parallel to the a2 and c axes.
Note that the use of three coplanar vectors results in the condition that h + k
= i.
10 Physical Metallurgy

Figure 1.8 The Miller-Bravais indices for hexagonal crystals.

The indices of a direction that lie along one of the three a axes (parallel to
a side of the hexagonal face) are of the type [2 1 1 0] . For this example, the
direction is parallel to the a1 axis and is resolved into components along a2 and
a3. Each of these components being 1, the value of the first index corresponding
to the a1 axis can be obtained as (1 1) = 2 from the condition that h + k
= i. This yields [2 1 1 0] .

1.3 STRUCTURE OF ALLOYS

Types of solid solutions When two metals are melted together and crystal-
lized, a single crystal structure may form. In the unit cell of this crystal, both
the metal atoms are present in proportion to their concentration. This structure
is known as a solid solution.
Figure 1.9 illustrates three types of solid solutions. Figure 1.9a is a
random substitutional solid solution. Here the solute atoms (shown shaded)
have substituted randomly for the parent atoms (not shaded) on their sites.
Figure 1.9b shows an ordered (substitutional) solid solution. Here, the solute
and the solvent atoms are arranged in a regular fashion on the atomic sites.
Perfect order is possible, only when the two metals are mixed in some fixed
proportion like 1:1, 3:1, etc.
Figure 1.9c shows an interstitial solid solution. Here, the solute atoms are
much smaller than the parent atoms and have occupied randomly the interstitial
voids between parent atoms. In the FCC unit cell, one set of interstitial voids
is located at the mid-points of the cube edges, (0, 0, ), (0, , 0), (, 0, 0) etc.
and at the body centre (, , ). These voids are called octahedral voids. The
size of the atom that can fit into an octahedral void without distortion is 0.414r,
where r is the radius of the parent atoms that surround the void. Smaller
 Structure of Metals and Alloys 11

tetrahedral voids are located on the body diagonals at (, , ), (, , ) and

other equivalent positions in the FCC unit cell. The size of the atom that can
fit the tetrahedral void is 0.225r. In the HCP crystal, the same types of octahedral
and tetrahedral voids are present. In the BCC crystal, the tetrahedral voids at
(, , 0) positions are larger (0.29r), as compared to the octahedral voids at
(0, 0, ) of size 0.15r. Carbon atoms in BCC iron are located at the smaller
(0, 0, ) voids.

Figure 1.9 (a) Random substitutional solid solution, (b) ordered substitutional
solid solution, and (c) interstitial solid solution.

Hume Rotherys rules Hume Rothery studied a number of alloy systems and
formulated the conditions that favour extensive substitutional solid solubility.
These are known as Hume Rothery rules and are listed below.
1. The size difference between the parent atom and the solute atom must be
less than 15%.
2. The electronegativity difference between the metals must be small.
3. The crystal structure of the metals and the valency of the atoms must be
the same.
The more is the size difference, the more intense is the stress field around
the solute atom. This is valid for both large and small atoms. A large atom
induces a compressive stress field around it, whereas a small atom induces a
tensile stress field. The stress field increases the potential energy of the crystal.
So, the solid solubility becomes more limited, as the size difference increases.
The electronegativity difference between two elements can be computed
from Paulings electronegativity numbers. This difference is usually small between
typical metallic elements.
Complete solid solubility, i.e. the mixing of the two atoms in all proportions
can evidently occur only when the crystal structures are identical. In the unit
cell of one of the metals, the atoms can be continuously substituted by the other,
till the pure crystal of the substituting atom is obtained.
The valency must be the same for extensive solid solubility. When the
valencies are different, it is generally found that an excess of electrons is readily
tolerated rather than a deficiency of bonding electrons. For example, zinc with
12 Physical Metallurgy

two valence electrons dissolves up to 38% in copper, whereas copper with one
valence electron dissolves only about 3% in zinc.
Free energy of solid solutions When two metals A and B are mixed to form
a solid solution, the Gibbs free energy G changes. The change in free energy
(the free energy of the solid solution minus the free energies of the two pure
metals) is called the free energy of mixing DGmix. It can be written in terms of
the enthalpy of mixing DHmix and the entropy of mixing DSmix:
DGmix = DHmix TDSmix (1.1)
1. For an ideal solution, DHmix = 0. Here, the A-A, B-B and A-B bonds are
all equally favoured.
2. If DHmix > 0 (positive deviation from ideality), the formation of like bonds
(A-A and B-B bonds) is preferred. For small positive values of DHmix,
there is a tendency for like atoms to cluster together in the solid solution.
On an average, an A atom has more A neighbours than is the case for a
perfectly random solid solution. At low temperatures or for large positive
values of DHmix, the clustering tendency increases. Eventually, immiscibility
results. The solid solution splits into two separate phases, one rich in A
and the other rich in B.
3. If DHmix < 0 (negative deviation from ideality), the formation of unlike
bonds (A-B bonds) is preferred. Here, there is a tendency for short range
ordering, i.e. A and B atoms show a preference to become nearest
neighbours. At low temperatures, long range order may set in and an
ordered solid solution may form. If DHmix is only slightly negative, the
ordering tendency may be so weak that long range ordering does not
occur at any temperature. If DHmix has a large negative value, the alloy
may remain in the fully ordered state right up to the melting point, behaving
like a chemical compound.
On mixing two metals at a constant temperature, a large number of distin-
guishable configurations w of equal potential energy result. The configurational
entropy Sc is given by
Sc = k ln w (1.2)
where k is Boltzmanns constant. The entropy of mixing XA mole fraction of A with
XB mole fraction of B (XA + XB = 1) in a random solid solution turns out to be
DSmix = R (XA ln XA + XB ln XB) (1.3)
XA and XB being fractions, the entropy of mixing is always positive. It tends to
lower the free energy according to Eq. (1.1).
An ideal solid solution is random. For small deviations from ideality (either
positive or negative), the solid solution can still be taken to be random for the
purpose of computing the entropy of mixing. This is called a regular solution
approximation. The free energy of mixing is then given by
DGmix = DHmix + RT(XA ln XA + XB ln XB) (1.4)
 Structure of Metals and Alloys 13

The variation in the free energy of a solid solution as a function of com-

position is sketched schematically in Fig. 1.10. Figure 1.10a is for an ideal
solution, where DHmix = 0. The free energy decreases with increasing concen-
tration of either element reaching a minimum at the mid-composition, where the
entropy of mixing is a maximum. Figure 1.10b is a case where DHmix < 0. The
free energy curve has the same shape as in Fig. 1.10a, but more U-shaped.
Figure 1.10c is for a solution where DHmix > 0. The free energy curve decreases
initially at the two pure ends due to the dominance of the entropy term. With
increasing concentration, however, the positive DHmix term dominates and the
free energy curve has two minima and a maximum. There is no complete solid
solubility in this case. The maximum solid solubilities are obtained by drawing
the common tangent xy to the free energy curve, Fig. 1.10c. The compositions
corresponding to the points of tangency c1 and c2 give the limit of solubility at
the two ends respectively. In between c1 and c2, the alloy is stable only as a
mixture of two solid solutions of composition c1 and c2. With increasing tem-
perature, however, the entropy term may dominate over the enthalpy term for
all compositions, the free energy curve may become U-shaped and a single solid
solution may become stable over the entire composition range.

Figure 1.10 Free energy of solid solutions: (a) enthalpy of mixing DHmix is zero
(ideal solution), (b) DHmix is negative, and (c) DHmix is positive.

Intermediate phases Intermediate phases refer to alloy crystal structures which

are different from either of the parent crystals. Some intermediate phases have
a fixed composition and are called intermetallic compounds. Quite often, however,
the intermediate phases have a significant composition range.
Only a few intermetallic compounds obey the normal rules of valency.
Many of these are based on the cubic structures of sodium chloride (NaCl) or
calcium fluoride (CaF2). In the NaCl structure, the chlorine ions are at the FCC
positions of the cubic unit cell and the sodium ions are at the octahedral voids.
Examples of such intermediate phases are MgSe, PbSe and CdTe. In the CaF2
structure, the calcium ions are at the FCC positions and the fluorine ions are at
the tetrahedral void positions. Examples are Mg2Si, Mg2Sn, Cu2S and Be2C.
The nickel arsenide (NiAs) structure is often formed by the transition metals
alloyed with the metalloids of the V and VI columns such as As, Bi, Sb, Se and
14 Physical Metallurgy

Te. The structure has the HCP stacking sequence with alternate layers of metal
and metalloid atoms. These phases have a fairly wide range of solubility.
Examples are CrSb, CuSn and MnSe.
Electron phases (also called electron compounds or Hume Rothery com-
pounds) have a wide range of solubility. They occur at certain definite values
of valence electrons to atom ratio in the alloy such as 3:2, 21:13 and 7:4.
Typical examples are CuZn (3:2), Cu5Zn8 (21:13) and CuZn3 (7:4).
Laves phases possess one of the three related structures with the general
formula AB2. Examples of the three prototype structures based on magnesium
are: MgCu2 (cubic), MgZn2 (hexagonal) and MgNi2 (hexagonal).
The sigma phase has a complex crystal structure and is very brittle. The
unit cell is tetragonal with 30 atoms in the cell. This phase is a source of
embrittlement in stainless steels.
Alloy carbides and nitrides form very hard and brittle compounds. They are
often cubic with the formula MX, with the metal atoms (M) in FCC positions
and the carbon or nitrogen (X) in interstitial positions (octahedral or tetrahe-
dral). Examples are TiC, VC, NbC, TiN, VN and NbN. Examples of carbides
with other structures are Cr23C6, Cr7C3, MoC and WC.
Like other solid solutions, intermediate phases may remain disordered at
high temperatures and become ordered at low temperatures. The b (BCC) phase
of the Cu-Zn alloy system is a typical example. The composition of bis around
50Cu50Zn. At temperatures above 465C, the copper and zinc atoms are ran-
domly distributed on the BCC sites. At lower temperatures, an ordered structure
forms with the zinc atoms at the body corners and the copper atoms at the body
centres (or vice versa). When this occurs, extra lines appear in the diffraction
pattern, as the ordered structure corresponds to a simple cubic space lattice
(with lattice points at the eight corners of the cube only) rather than the BCC
space lattice. The ordered compound is known as a superlattice, as it gives rise
to additional diffraction lines.
In general, intermetallic compounds and ordered solid solutions possess
little ductility and are relatively brittle phases. They are useful for increasing the
strength of a ductile matrix, but at the cost of reduced ductility.

1.4 IMPERFECTIONS IN CRYSTALS

Imperfections in crystals can be classified according to their geometry as point,
line and surface imperfections. The characteristics of different imperfections are
briefly considered below.
Point imperfections There are point-like regions of distortion in the crystal.
The common types are :
(a) vacancy
(b) interstitialcy
 Structure of Metals and Alloys 15

(c) interstitial impurity, and

(d) substitutional impurity
A vacancy is an atomic site from which the atom is missing, a in Fig. 1.11.
An interstitialcy refers to an atom displaced from a regular site into an inter-
stitial site, b in Fig. 1.11. An interstitial impurity is a small atom occupying an
interstitial void space between parent atoms of the crystal, c in Fig. 1.11. A
substitutional impurity is an impurity atom that has substituted for one of the
parent atoms on a regular atomic site, d in the figure.

Figure 1.11 Point imperfections in crystals: (a) vacancy, (b) interstitialcy,

(c) interstitial impurity, and (d) susbstitutional impurity.

Point imperfections increase significantly the configurational entropy of the

crystal and thereby lower its free energy. So, a certain number n of point
imperfections is in thermodynamic equilibrium in the crystal at a given tempera-
ture. For example, the fraction of vacancies n/N in a crystal at temperature T is
given by
n % Hf
= exp (1.5)
N kT
where N is the total number of atomic sites and DHf is the enthalpy of formation
of a vacancy and is equal to the reversible work done at constant pressure in
bringing an atom from inside the crystal to the surface.
Dislocations Line imperfections are known as dislocations. The geometry of
dislocations can be understood with reference to the two limiting straight line
types:
1. the edge dislocation, and
2. the screw dislocation
16 Physical Metallurgy

Figure 1.12a shows an edge dislocation extending from the front to the
back face of the crystal. It is the distorted region all along the edge of an
incomplete plane in the crystal. The atoms just above the edge are in compres-
sion, whereas those just below are in tension. A dislocation is characterized by
a vector known as the Burgers vector. Figure 1.12a illustrates the procedure for
determining the Burgers vector b. A Burgers circuit is drawn around the dislo-
cation line. The vector needed to close the circuit, QP, is the Burgers vector of
the dislocation. It is perpendicular to the edge dislocation line. A positive or a
negative edge dislocation is defined, depending on whether the incomplete
plane is from the top or from the bottom of the crystal. They are symbolically
denoted by ^ and respectively.
Figure 1.12b shows a screw dislocation extending from the right to the left
side face of the crystal. It is characterized by a shear distortion. The Burgers
vector is parallel to the screw dislocation line, as illustrated in the figure.
Positive and negative screw dislocations are symbolically denoted by and
.

Figure 1.12 Edge and screw dislocations. The direction of the

dislocation line lies along the direction vector t.

In general, the dislocations in real crystals are some combination of the

edge and the screw types. These are called mixed dislocations. The Burgers
vector of a mixed dislocation remains invariant along the length of the dislo-
cation line. It can be resolved into two components, one parallel to the dislo-
cation line (screw component) and the other perpendicular to the line (edge
component). Two dislocations of opposite sign (either edge or screw) lying on
the same plane attract and can mutually annihilate each other.
The elastic strain energy E per unit length of a dislocation line is
Nb2
E= (1.6)
2
where is the shear modulus of the crystal and b is the magnitude of the
Burgers vector. As the elastic energy is proportional to the square of the Burgers
vector, dislocations tend to have as small a Burgers vector as possible. The
smallest distance between any two atomic sites lies along the close packed
 Structure of Metals and Alloys 17

directions. Accordingly, the Burgers vector of full dislocations lies along close
packed directions in typical crystals as listed below:

Crystal BCC FCC HCP

Burgers vector <111> <110> <11 2 0>

Dislocations are present in crystals due to accidents during the growth of

the crystal. The dislocation density in an annealed crystal ranges from 106 to
1010 m m3. If the growth of the crystal is done very slowly and carefully and
if the volume of the grown crystal is small, many of the accidents can be averted.
In a few cases, perfect crystals that are totally free of dislocations can be grown.
Dislocations play an important role in the plastic deformation of metals and
alloys. Very thin perfect crystals that are totally free of dislocations known as
whiskers possess very high strength. In ordinary crystals, however, the presence
of even a few dislocations is sufficient to make the crystal weak. Dislocations
move at relatively low stresses and cause plastic deformation. The methods of
strengthening crystalline materials against this inherent weakness aim at in-
creasing the resistance to dislocation motion.
The motion of a dislocation in response to an externally applied stress on
a plane containing the Burgers vector and the dislocation line (called the slip
plane) is known as glide motion. The atoms near the dislocation configuration
are conserved during glide motion. Edge dislocations and mixed dislocations
have uniquely defined slip planes. As the Burgers vector and the dislocation line
are parallel in the case of a screw dislocation, there is no uniquely defined
slip plane in this case. A screw dislocation can cross-glide or cross-slip from
one slip plane to another, as illustrated in Fig. 1.13. The screw dislocation first

Figure 1.13 Cross-slip of a screw dislocation.

The hatched areas are the slipped regions.
18 Physical Metallurgy

glides on a (111) plane causing slip over the hatched region. It then cross-slips
on to a nonparallel (1 1 1) plane. In the double cross-slip process illustrated, it
again cross-slips back to a (111) plane, which is parallel to the initial plane.
Edge dislocations cannot cross-slip. When they move on a plane other than
the slip plane, the motion involves addition or subtraction of atoms to the edge.
This motion is nonconservative and is called climb motion. Addition or subtraction
of atoms generally occurs at an atomic step called jog on the edge of the
incomplete plane, Fig. 1.14. Edge dislocations climb up or climb down, depending
on whether the incomplete plane shrinks or increases in extent. Addition of
atoms is required for climbing down, whereas subtraction of atoms (or
equivalently addition of vacancies) is required for climbing up. In Fig. 1.14,
moving of the jog to the left is equivalent to climbing down; movement to the
right is climbing up. Climb can clearly occur only when there is sufficient
thermal energy for the diffusion of atoms to or away from the edge of the
incomplete plane. It is, therefore, an important process in high temperature
deformation known as creep, discussed under Sec. 5.4.

Figure 1.14 Jog is a step on the edge of the incomplete

plane of an edge dislocation.

Partial dislocations and stacking faults Full dislocations described above

have an integral lattice translation as the Burgers vector. For partial (or imper-
fect) dislocations, the Burgers vector is a fraction of an interatomic distance. As
the elastic energy of dislocations is proportional to the square of their Burgers
vector, the splitting of a full dislocation b1 into two partial dislocations b2 and
b3 is favoured energetically, when
 Structure of Metals and Alloys 19

b12 > b22 + b32

A typical example of such a dislocation is the splitting of a full dislocation
in an FCC crystal into two Shockley partials:
[110] 1/ 6[2 1 1] + 1/ 6[121]
The Burgers vectors of the Shockley partials lie on the slip plane (which
is a close-packed plane in FCC crystals). They repel and tend to move away
from each other. In the region of the slip plane between the two partials, the slip
displacement is not a full interatomic distance. The region becomes a planar
surface imperfection known as a stacking fault. The stacking sequence across
the slip plane in the region between the two partials becomes ...ABCACABC...,
i.e. a thin region of HCP stacking CACA occurs in the FCC crystal. Locally the
stacking is improper or faulty, and so this region is called a stacking fault.
Stacking faults possess surface energy. The surface tension due to this tends to
minimize the area of the faulted region. Stacking fault energies for typical
metals and alloys are given below.

Metal A1 Cu Ni Cu25Zn Fe18Cr8Ni

Stacking fault energy, J m2 0.2 0.04 0.03 0.007 0.002

The distance between the two Shockley partials is determined by a balance

between the repulsive force acting on the partials and the attractive force due
to the surface tension tending to shrink the area of the faulted region. The
equilibrium distance of separation between the partials thus depends on the
stacking fault energy. The distance is less when stacking fault energy is high.
This fact has important consequences in the plastic deformation mechanisms. In
order to cross-slip, a dissociated screw dislocation has to recombine over a
critical distance and this is easier if the distance of separation between the
partials is less. So, cross-slip is easier in metals with a high stacking fault
energy.
High angle grain boundaries When the orientation difference between two
neighbouring crystals is more than 1015, the boundary between them is
generally a curved surface known as a high angle grain boundary, Fig. 1.15.
During solidification or recrystallization, crystals nucleate at different points of
the parent phase independently of one another and grow until the neighbouring
crystals impinge. The atoms that are caught in the boundary region take up
uncoordinated positions, belonging to neither of the neighbouring crystals. In
close packed crystals, the average number of nearest neighbours to an atom
in the boundary region is 11, as compared to 12 inside the crystal. This
missing bond accounts for the grain boundary energy, which is in the range of
0.51.0 J m2 for most metals.
20 Physical Metallurgy

Figure 1.15 Grain boundaries in a polycrystalline metal. The atom

arrangements at the boundaries are distorted and unrelated.

The size of the grains in a polycrystalline metal greatly influences a number

of its mechanical and chemical properties. Grain sizes are specified using an
ASTM (American Society for Testing and Materials) designation. If n is the
ASTM grain size number, 2n1 grains are present per square inch in the micro-
structure of the metal viewed under a magnification of 100 X (linear). Noting
that 1 sq.in. = 645 mm2, for an ASTM grain size number of 7, 2n1 = 64. There
are 64 104 grains per sq.in. or ~ 103 grains per mm2. Along a length of one
mm, there are 103 grains. The average grain diameter is thus approximately
1/ 103 = 0.03 mm.
Planar boundaries One type of planar boundary, the stacking fault, has already
been discussed in relation to partial dislocations. The other types are described
below.
When the orientation difference between neighbouring crystals is less than
10, the boundary is described as a low angle boundary. Tilt boundaries are
arrays of parallel edge dislocations of the same sign one below the other as
shown in Fig. 1.16. The tilt angle q is given by
b
tan R = (1.7)
h
where b is the Burgers vector of the edge dislocations and h is the vertical
distance of separation between two neighbouring edge dislocations on the
boundary. Tilt boundaries form during a process called recovery, when excess
dislocations of the same sign arrange themselves one below the other. Twist
boundaries refer to similar low angle boundaries formed by arrays of screw
dislocations.
 Structure of Metals and Alloys 21

Figure 1.16 A tilt boundary consists of an array of edge dislocations

of the same sign vertically one above the other.

When the arrangement of atoms on one side of a boundary is a mirror

image of that on the other side, the boundary is known as a twin boundary,
Fig. 1.17. Twin boundaries occur in pairs, where the orientation difference
introduced by one twin of a pair is nullified by the other. Twin boundaries
form during annealing (called annealing twins) or during deformation
(deformation twins). The energy of twin boundaries is in the range of
0.010.05 J m2.
When the crystal structures of the two metals are similar and the lattice
parameters are nearly equal, the boundary between the two crystals (called an
interface) is coherent, with a one-to-one correspondence of atoms across
the interface. The surface energy of this interface is small, in the range of
0.010.05 J m 2. When there is no such similarity or matching of the two
metals, the interface is incoherent and is similar in structure and energy to large
angle grain boundaries.
Interaction between crystal imperfections Different types of imperfections
in crystals can interact with one another. The interaction is always such that the
total distortional energy of the interacting imperfections is lowered.
22 Physical Metallurgy

Figure 1.17 The atomic arrangement on one side of a twin boundary

is a mirror reflection of that on the other side.

A large substitutional atom can interact with a vacancy in the same crystal
to lower the distortional energy. The two defects stay together in close proxim-
ity. Point imperfections interact with the distortional geometry associated with
dislocations. For example, a large substitutional atom can lower its distortional
energy by staying in the tensile region just below the edge of the incomplete
plane of an edge dislocation. An interstitial atom, that is too large for the void
it occupies, can similarly lower its energy by staying in the core of an edge
dislocation. Such an effect is particularly marked in low carbon and mild steels,
where the oversized (interstitial) carbon atoms tend to segregate in voids around
the tensile region of edge dislocations. Such solute atmospheres around dislo-
cations are known as Cottrell atmospheres. Point imperfections also segregate
at grain boundaries as the structure at the boundary is more open than that in
the interior of the crystal. Such segregation results in several metallurgically
important effects such as the solute drag effect on a migrating grain boundary
and intergranular cracking.
Dislocations lie on slip planes within the crystal. As the crystal orientation
across a grain boundary (or a twin boundary) changes abruptly, a dislocation on
a slip plane in any one grain cannot cross the boundary and move on to the
same slip plane of the next grain. Because of this, moving dislocations get piled
up against grain boundaries. This results in an important influence of grain size
on the yield strength of metals and alloys.

1.5 EXPERIMENTAL STUDY OF STRUCTURE

The structure seen with the naked eye or under low magnification by acid
etching of steel bars, billets, castings and forgings is called macrostructure. It
 Structure of Metals and Alloys 23

reveals pipes, segregation, flakes, cracks, inclusions, porosity and decarburized

layers. A solution consisting of one part of HCl and one part of water is the
most widely used etchant. For studying the structure at a finer level, a number
of physical techniques are used.
The optical microscope As metals are opaque, they can be viewed under an
optical microscope only in reflected light. In the optical microscope, Fig. 1.18a,
light from a lamp gets reflected by a semi-silvered glass kept at a 45 tilt, passes
through the objective and then on to the specimen surface. It is reflected back
by the metal surface and partly transmitted through the semi-silvered glass to
the eyepiece. A metallograph is an instrument in which the structure as seen
under the microscope called the microstructure is photographed.
To examine the metal specimen properly, the surface should be polished to
a high degree of reflectivity. This is done by polishing the surface on successively
finer grades of emery papers. Final polishing is done on cloth impregnated with
a very fine powder of alumina or diamond. After polishing, the specimen is
etched with a dilute acid to selectively attack the features of the structure.
Figure 1.18b shows the profile of a polished surface of a polycrystalline sample
after etching. The grain boundaries are selectively attacked by the etchant, as they

Figure 1.18 (a) The optical microscope. (b) Etching produces grooves along grain
boundaries and the light that falls on the grooves is not reflected back into the objective.
(c) The network of dark lines are grain boundaries as seen under the optical microscope.
24 Physical Metallurgy

possess a higher energy than the interior of the crystal. No light is reflected
from the grooves that form along grain boundaries and they appear as a network
of dark lines under the microscope, Fig. 1.18c. Sometimes, the grains are also
attacked at different rates depending on their orientation with respect to the
plane of observation. They may reflect different amounts of light and the grain
structure can be identified.
A number of accessories and attachments are available to the optical
microscope. It continues to be an important tool for the study of the structure,
despite the advances in recent years in high resolution microscopes.
The electron microscope The source of illumination in an optical
microscope is the ordinary visible light with a wavelength of about 5000 .
In an electron microscope, the source of illumination is a beam of electrons
with a wavelength of about 0.05 . As the resolving power of a microscope
is inversely proportional to the wavelength of the illumination used, the
electron microscope has a much better resolving power and can show up
details as small as 10 . In addition to this, the electron microscope
provides facilities for selected area diffraction, i.e. a diffraction pattern
can be obtained from a very small particle in the structure along with its
micrograph.
The illuminating system is the electron gun, Fig. 1.19a, consisting of a
heated filament held at a high accelerating potential of 100 200 kV.
Magnetic lenses are used in the electron microscope. The condenser
lenses (usually two in number) focus the electron beam on to the specimen.
The objective lens produces a first magnified image, which is further
magnified by an intermediate lens. The final magnification is by the
projector lens, which produces the image on a fluorescent screen. For
recording the image, a photographic plate or film is put in place of the
screen.
A dark field image can be formed from one or more of the diffracted beams
(instead of from the main beam), by moving the objective aperture to transmit
the diffracted beam and block the main beam. This enables the diffracting
precipitate particles to be seen more easily than in the normal bright field image.
For metallurgical work, a tilting stage is indispensable. Tilting should be about
two orthogonal axes with at least 25 of tilt about each axis.
Metals may be examined in two ways: by making a replica of a suitably
etched surface or by making a piece of metal thin enough to be transparent to
electrons. In earlier years, the replica was made from a plastic called formwar.
The film about 1000 thick is stripped and examined under the electron
microscope. In carbon extraction replica, a film of about 100200 is evaporated
on the etched surface, stripped off and examined. The resolution of the carbon
replica is much better than the plastic replica. In addition, the particles in the
microstructure that stand out in relief are transferred to the replica and may not
 Structure of Metals and Alloys 25

be dissolved during the subsequent step of dissolving out the metal matrix. By
obtaining a diffraction pattern, the particles can be identified. Extremely fine
particles of NbC (about 20 in size) in microalloyed steels have been identified
this way.

Figure 1.19 Line diagram of (a) the transmission electron microscope, and
(b) the scanning electron microscope. Typical electron micrographs are also shown.

Thin foils of the metal (less than 2500 thick) can be directly examined
in a transmission electron microscope (TEM). The thin foil is obtained by
26 Physical Metallurgy

electropolishing. In the window technique, pointed electrodes are used. The

metal directly in front of the pointed electrode is dissolved more rapidly and
perforates. The metal adjacent to the perforation is thin enough to transmit the
electrons. Diffraction patterns can be obtained over any selected area of the
matrix or of particles embedded in it.
The scanning electron microscope In the scanning electron microscope,
Fig. 1.19b, a beam of electrons is moved over the specimen surface in a sequence
of closely spaced lines. The electrons interact with the surface atoms and the
output depends on the composition and topography. The output signal is picked
up and used to modulate the brightness of a cathode ray beam, which is moving
synchronously with the scanning search beam. The depth of focus of the scanning
electron microscope is so high that a fracture surface can be directly examined
without any polishing. In metallurgical studies, fractographic examination and,
to a lesser extent, examination of polished and etched surfaces are carried out
with the scanning electron microscope.
The electron probe microanalyzer The composition of alloys can be
determined by conventional chemical analysis and other modern instrumental
methods such as spectroscopy or x-ray fluorescence. These give only the average
composition. The electron probe microanalyzer (EPMA) is a tool for determining
the composition on a micron scale relevant to metallurgical problems such as
identification of very small inclusions and precipitates and composition gradients
on a micron scale. In this instrument, a beam of electrons about mm diameter
strikes the specimen surface, interacting with the atoms of the specimen to
produce x-rays. By measuring the wavelength and intensity of the x-rays
produced, it is possible to determine which elements are present and their
respective concentrations.
X-ray diffraction The crystal structure of metals and alloys is determined by
x-ray diffraction. The Bragg law describes the effect produced by a beam of x-
rays of wavelength l incident at an angle of q (called the Bragg angle) on to
a set of parallel atomic planes with an interplanar spacing of d:
nl = 2d sin q (1.8)
where n is an integer.
There are a number of methods for determining the crystal structure. In the
powder method, a monochromatic beam of x-rays is incident on the powder
sample of the unknown metal. The diffracted beam is recorded in a diffracto-
meter as a function of 2q. Figure 1.20 shows a tracing from a diffractometer,
where the intensity peaks correspond to positions, where the Bragg condition
given in Eq. (1.8) is satisfied. Computer software packages are now available
for analyzing the diffraction data and ascertaining the crystal structure.
 Structure of Metals and Alloys 27

Figure 1.20 Tracing from an x-ray diffractometer.

REFERENCES

1. Barrett, C.S. and T.B. Massalski, The Structure of Metals, 3rd revised
edition, Pergamon Press, Oxford (1980).
2. Weinberg, F. (Ed.), Tools and Techniques in Physical Metallurgy, Vols. 1
and 2, Marcel Dekker (1970).
3. Smallman, R.E., Modern Physical Metallurgy, Butterworths, London
(1985).

QUESTIONS

1. The resistivity of copper is 109 ohm m, whereas that of germanium is

101 ohm m and of diamond is 1011 ohm m. What is the reason for such
large variations in resistivity of different materials?
2. The density of BCC chromium is 7190 kg m3. Find its lattice parameter.
Answer: 2.88 .
3. Show that the packing efficiency of an FCC crystal is 0.74.
28 Physical Metallurgy

4. What is the difference between the hexagonal space lattice and the HCP
crystal ? Hint: Compare Fig. 1.3a and 1.3b.
5. Explain why an ... ABBA ... stacking is not stable.
6. Show that the ideal c/a ratio of an HCP crystal is 1.633.
7. Draw a cubic unit cell and sketch the following planes and directions
within the unit cell: (132), (322), [312] and [102].
8. Draw a (1 1 0) plane in the unit cell of a BCC crystal and show the close
packed directions that lie on it.
9. If two metals show only limited solid solubility in each other, is it positive
or negative deviation from ideality ? Explain.
10. Explain why complete solid solubility is not possible in an interstitial solid
solution.
11. Explain the differences between a chemical compound, an ordered solid
solution and a random solid solution.
12. The enthalpy of formation of vacancies in aluminium is 68 kJ mol1.
Calculate the fraction of vacancies in equilibrium at room temperature and
just below the melting point.
Answer: at 300 K, 1.44 1012 and at 930 K, 1.52 104.
13. A cold worked copper sample has a dislocation density of 1015 m m3
Estimate the strain energy stored in the metal.
Answer: 1.48 MJ m3.
14. Calculate the spacing between edge dislocations in a tilt boundary in iron,
when the angle of misorientation between the subgrains is 2.
Answer: 71.
15. Calculate the average grain diameter for ASTM grain size number 14. By
how much would this change, if the ASTM number is 12?
Answer: 2.8 mm; increase by 100%.
Chapter 2
PHASE DIAGRAMS

The relationships between phases in a system as a function of temperature,

pressure and composition are depicted in the form of maps. Generally, these are
equilibrium relationships in that they correspond to the state of the minimum
free energy of the system. Such maps are called equilibrium diagrams (also phase
diagrams or constitutional diagrams). Sometimes, metastable phases may also be
shown on a phase diagram, especially if they are present in alloys of engineering
interest. In such cases, the phase diagram depicts the metastable equilibrium.

2.1 THE PHASE RULE

The Gibbs phase rule can be stated as follows :
F=CP+2 (2.1)
where F is the degrees of freedom,
C is the number of components, and
P is the number of phases in the system.
Phases Different states of aggregation of matter such as solid, liquid and gas
form separate phases. The gaseous phase is always a single phase, as atoms (or
molecules) in the gaseous phase are mixed at the atomic level. A liquid solution
is by the same token a single phase. For example, in salt solution, i.e. salt
dissolved in water, the sodium ions, the chlorine ions and the water molecules
are mixed at the atom level. A liquid mixture of water and oil forms separate
phases. A solid solution of an alloy with the component atoms mixed at the unit
cell level constitutes a single phase.
29
30 Physical Metallurgy

Components Components refer to the independent chemical species that con-

stitute an alloy. For plain carbon steel, iron and carbon are the chemical species;
so, the components are Fe and C (graphite). However, carbon almost always
occurs in the combined form as cementite (Fe3C) in steels, even though cement-
ite is metastable and has a higher free energy as compared to graphite. So, it
is common to show the phase relationships in the Fe-C system on a metastable
phase diagram with Fe and Fe3C as the components. The choice of all the three
(Fe, C and Fe3C) as components does not constitute a set of independent chemi-
cal species.
Degrees of freedom The degrees of freedom refer to the number of indepen-
dent variables associated with the system. The total number of variables is the
external variables (temperature and pressure) plus the composition variables.
The number of composition variables required to express the concentration of
C components in a phase is (C 1)* . For a system with P phases,
total variables = P (C 1) + 2 (2.2)
The degree of freedom F is equal to or less than the total number of
variables :
F = C P + 2 P (C 1) + 2 (2.3)
When there is only one phase in a system, F is equal to the total number
of variables. For P > 1, F is less than the total number. The minimum value for
the degree of freedom is zero and this sets an upper limit on the number of
phases that can coexist in a system under equilibrium.

2.2 BINARY PHASE DIAGRAMS

The most commonly used phase diagrams are those that depict the equilibrium
between two components. These are known as binary phase diagrams. Many
engineering alloys contain more than two components. Nevertheless, the binary
diagram between the two major components is still useful in interpreting the
phase relationships. For example, the Fe-Fe3C phase diagram is used in the
study of carbon steels, which usually contain small quantities of S, P, Si and
Mn, in addition to Fe and C. In other cases, more complex phase relationships
have to be studied. For equilibrium in a system containing three components,
ternary phase diagrams are used. Further, it is possible to depict the more
complicated phase relationships between a number of components on a pseudo-
binary diagram.
For graphical representation of the binary phase relationships, we need a
map with three axes corresponding to temperature, pressure and composition.
As three-dimensional diagrams are difficult to draw on paper, for binary phase
diagrams it is customary to ignore the pressure variable and the vapour phase.

*Note that every phase in equilibrium in a system will have all the components in it, however
small may be the concentration of a component.
 Phase Diagrams 31

As one of the variables has been arbitrarily omitted, the phase rule for con-
densed phases (i.e., liquids and solids) only is written as
F=CP+1 (2.4)
The binary phase diagram is drawn with temperature on the y-axis and
composition on the x-axis. For a binary system of A and B, the x-axis reads %
B from left to right (or % A from right to left). The composition is either weight
% or atom %. It is usually in weight %, unless stated otherwise. The weight %
of A in an alloy of A and B can be converted into atom % A as follows:
wt.% A
Atom% A = atomic wt. of A 100
wt.% A wt.% B
+
atomic wt. of A atomic wt. of B
Isomorphous systems The simplest binary phase diagram is for a system
exhibiting complete liquid solubility as well as solid solubility. This is called an
isomorphous system, to indicate that the crystal structures of the two components
and the solid solution are the same. As an example, the copper-nickel phase
diagram is shown in Fig. 2.1. Pure copper (FCC) melts at 1083C. Pure
nickel (FCC) melts at 1455C. The liquid solution extends over all compositions.

Figure 2.1 The copper-nickel phase diagram, showing complete solubility

both in the liquid and in the solid states.

Similarly, the solid solution (a), which is also FCC, covers the entire composition
range.* The liquid and the solid regions are separated by a two-phase region,
where both the phases coexist. The boundary between the two-phase region and
the liquid is called the liquidus. The boundary between the two-phase region
and the solid is called the solidus.
*The a solid solution splits into two different solid phases at temperatures below 322C.
32 Physical Metallurgy

In certain systems, the liquidus and the solidus may coincide at some com-
position. The alloy of this composition melts and solidifies at a constant tem-
perature like a pure metal without passing through the two-phase region. Such
a composition is said to be congruently melting.
The tie-line rule The compositions of the phases in the two-phase region are
not equal. Nickel (or copper) redistributes or partitions itself in different
concentrations in the two phases. The compositions of the co-existing phases
are determined from the tie-line rule. A horizontal line (called the tie-line) is
drawn at the temperature of interest T, say, 1300C in Fig. 2.1. The intersection
of this line with the liquidus gives the composition of the liquid cl and the
intersection with the solidus gives the composition of the solid cs. Referring to
Fig. 2.1, at 1300C,
composition of the liquid cl = 46% Ni (54% Cu)
composition of the solid cs = 58% Ni (42% Cu)
This is applicable for any overall composition of the alloy that lies along
the tie-line within the two-phase region. In the above example, as the overall
composition of the alloy is varied between 46 and 58% Ni, the liquid and the
solid compositions remain constant. A little reflection shows that this will be
possible only if the relative amounts of the coexisting phases also change, as
the overall composition is varied.
The lever rule The relative amounts of the coexisting phases at a specified
temperature are given by the lever rule. The tie-line is taken to be a lever arm
CA, with the fulcrum at the overall composition co (point B), Fig. 2.1. The solid
and the liquid compositions lie at the two ends C and A of the lever arm. The
fraction of the liquid fl for the overall composition co is given by the ratio of
the lever arm length on the other side of the fulcrum (CB) to the total length
of the lever arm (CA):
CB cs co
fl = =
CA cs cl
Similarly, the fraction of the solid fs is given by
BA co cl
fs = =
CA cs cl
Referring to Fig. 2.1, for the overall composition co = 50% Ni at 1300C,
58 50
fl = = 0.67
58 46
50 46
fs = = 0.33
58 46
It must be clearly understood that the tie-line rule gives the concentration
of the components in each phase, whereas the lever rule gives the fractions of
the phases. Stating the two rules together, for the above example, at 1300C,
 Phase Diagrams 33

fraction of liquid of composition 46% Ni = 0.67

fraction of solid of composition 58% Ni = 0.33
Total = 1.00
0.67 46 + 0.33 58 = 50% Ni
which is the overall composition of the alloy.
Phase rule applied to the isomorphous system In the liquid or in the solid
region of the Cu-Ni phase diagram, only one phase exists. Noting that C = 2
and P = 1 here, from Eq. (2.4), the degrees of freedom F = 2 1 + 1 = 2. Both
the temperature and composition can be varied within limits in the single phase
region. In the two-phase region, we have three variables,
temperature,
composition of the liquid phase, and
composition of the solid phase.
From Eq. (2.4), F = 2 2 + 1 = 1. So, only one of these three variables
can be changed independently. For example, if we choose the temperature ar-
bitrarily and thereby exercise one degree of freedom, the other two variables,
namely, the compositions of the two phases are automatically fixed, as per the
tie-line rule.
The eutectic phase diagram All systems do not exhibit complete solid solu-
bility. The Hume-Rotherys conditions for extensive solid solubility are not
ideally obeyed for many pairs of elements. This usually results in limited solid
solubility with two terminal solid solutions, each based on the crystal structure
of a pure component. Two types of phase diagrams result :
(i) the eutectic diagram, and
(ii) the peritectic diagram
When the melting points of the components are comparable, the eutectic
phase diagram is usually obtained. The Ag-Cu eutectic phase diagram is shown
in Fig. 2.2. A maximum of 8.8% of copper dissolves in silver at 780C, forming
the terminal solid solution a (FCC). A maximum of 7.9% of silver dissolves in
copper at 780C, forming the other terminal solid solution b (also FCC). The
solubility in both cases decreases with decreasing temperature along the solvus
line. The solvus is the solid-state phase boundary between the terminal solid
solution and the two-phase region, see Fig. 2.2.
The alloy with 28.1% Cu has the minimum melting point among all the
alloys of this system. This composition is known as the eutectic composition ce
and the corresponding temperature (780C) is the eutectic temperature Te. The
liquid of eutectic composition solidifies isothermally as a mixture of a and b
crystals on cooling through Te. The characteristics of this eutectic reaction are
summarized as follows:
34 Physical Metallurgy

Figure 2.2 The silver-copper phase diagram is a simple eutectic system.

cool
reaction : L a + b
structure : FCC FCC
composition : ce ca e cb e
wt% Cu : 28.1 8.8 92.1
Plate-like crystals (or rod-like crystals) of a and b arranged in an alternating,
parallel fashion are the characteristic structure of the eutectic mixture as shown
in Fig. 2.3a. Sometimes, an abnormal or divorced eutectic structure is observed.
Here, the eutectic is present in so small a volume fraction that the plate-like
structure does not form. Instead the minor phase of the eutectic forms and
grows as islands that appear unconnected on a polished surface. The other
phase forms the continuous matrix. Abnormal eutectic structures are observed,
when one of the phases forms strong covalent bonds, e.g. the Al-Si and Fe-
graphite systems. Chinese script is a structure between the lamellar (normal)
eutectic and the abnormal eutectic. Here, the lamellar morphology has degenerated
into a pattern that looks like a Chinese script, Fig. 2.3b.
An alloy to the left of the eutectic is known as a hypoeutectic alloy. An
alloy to the right is called a hypereutectic alloy. In both these cases, on cooling
below the liquidus boundary, the proeutectic phase (a or b) separates out first.
On further cooling, the fraction of the proeutectic phase increases, till the eutectic
temperature is reached. Just above the eutectic temperature Te, the tie-line indicates
that the composition of the proeutectic phase is cae (or cbe) and the liquid is of
 Phase Diagrams 35

the eutectic composition ce. On cooling through Te, the eutectic reaction occurs
in that part of the system which is in the liquid form. The proeutectic phase
appears as equiaxed crystals in the eutectic matrix and is clearly distinguishable
from the plate-like crystals of the eutectic mixture. This structure for a
hypereutectic alloy is schematically shown in Fig. 2.2.

Figure 2.3 (a) Microstructure of a normal eutectic. (b) Microstructure of a eutectic

mixture that lies between normal and abnormal, known as the Chinese script.

At the eutectic temperature, all the three phases, a, b and liquid are
present. Here, the total variables are 4 in number: temperature (1) and com-
positions of the three phases (2-4). From Eq. (2.4), F = 2 3 + 1 = 0. With
the zero degree of freedom, none of the four variables can be independently
changed. The eutectic temperature is a fixed value for a given system and
the compositions of the three phases at the eutectic temperature are also
fixed. To denote the zero degree of freedom, the eutectic reaction is called
the invariant reaction and the eutectic temperature is known as the invariant
temperature. Note that, for all compositions which undergo the eutectic
reaction (the eutectic, the hypoeutectic and the hypereutectic alloys), three-
phase equilibrium prevails at the eutectic temperature and the degree of
freedom is zero.
The lever rule cannot be applied at an invariant temperature, where three
phases coexist. However, it can be applied just above or just below the eutectic
temperature to determine the fraction of
(i) a proeutectic phase,
(ii) the eutectic mixture, and
(iii) a phase that forms part of the eutectic mixture.
The peritectic phase diagram When the melting points of two components
showing limited solid solubility are quite different, a peritectic diagram
usually results. Figure 2.4 illustrates the simple peritectic system between
nickel (Ni) and rhenium (Re), the respective melting points being 1455C
and 3186C.
36 Physical Metallurgy

Figure 2.4 The nickel-rhenium phase diagram is a simple peritectic system.

The peritectic reaction occurs at the peritectic temperature Tp (1622C).

cool
reaction : L + b a
structure : HCP FCC
composition : clp cbp cap
wt.% Re : 36 94 41
Here, one liquid and one solid react to form a single solid phase on cooling
through the peritectic temperature. Three phases are in equilibrium at the peritectic
temperature and zero degree of freedom prevails. Accordingly, the peritectic
reaction is invariant. Note that the peritectic temperature Tp lies between the
melting points of the two components, whereas in the eutectic reaction, Te is
lower than either of the melting points.
Other invariant reactions We have considered above the simple phase dia-
grams of the eutectic or the peritectic type. Binary diagrams may have other
invariant reactions, a number of which are described below.
In the eutectoid reaction, a solid phase decomposes isothermally on cooling
through a fixed temperature into two other solid phases:
cool
g a+b
The initial liquid phase of the eutectic reaction is replaced by the solid
phase in this reaction. Corresponding to the peritectic reaction, we have the
peritectoid reaction that occurs entirely in the solid state:
 Phase Diagrams 37

cool
g+a b
In the monotectic reaction, a liquid phase on cooling decomposes to an-
other liquid and a solid phase:
cool
LI LII + a
In the syntectic reaction, two liquids combine to yield a single solid phase
on cooling:
cool
LI + LII a
Three-phase equilibrium and zero degree of freedom prevail at the invariant
temperature of all the above reactions. The following table is a summary of the
invariant reactions.
Name Invariant reaction Phase boundaries near the
direction of cooling invariant composition

L
Eutectic L a + b
a+b
g
Eutectoid g a + b
a+b

Monotectic LI LII + a
L+a
Peritectic L+ a b
b
g+a
Peritectoid g + a b
b

Syntectic LI + LII a

2.3 FREE ENERGY COMPOSITION CURVES FOR

BINARY SYSTEMS
The free energy of each of the phases present on a phase diagram can be
plotted as a function of composition at a series of temperatures. The phases
that coexist in equilibrium are those that correspond to the lowest free energy
of the system.
For a binary system exhibiting complete solid solubility, the free energy-
composition curves for the liquid and the solid phases are shown at three
temperatures in Fig. 2.5. In Fig. 2.5a (the phase diagram), the temperature T1
lies above the liquidus across the entire composition range. Figure 2.5b shows
the free energy curves for this temperature. The liquid phase has a lower free
energy than the solid at all compositions; so, it is the stable phase. The next
38 Physical Metallurgy

lower temperature T2 cuts across the two phase region in Fig. 2.5a. The
corresponding free energy curves intersect each other, Fig. 2.5c. There is a
certain range of composition over which the free energy of a mixture of the
liquid and the solid phases is lower than either of the phases. This composition
range is obtained by drawing a common tangent xy to the two curves, see
Fig. 2.5c. The compositions at points x and y are cl and cs respectively. Over
the range of composition to the left of cl, the free energy of the liquid is lower;
so, it is the stable phase. For compositions that lie between x and y, two-phase
equilibrium prevails. The mixture of the two phases in the proportion given by
the lever rule has the lowest free energy. For compositions to the right of cs, the
solid phase is the stable phase. At temperature T3, Fig. 2.5d, the solid phase has
lower free energy and is stable over the entire composition range.

Figure 2.5 Free energy-composition relationships for an isomorphous

system at three temperatures.

For a eutectic system, three free energy-composition curves are drawn, one
for each of the phases: a, b and liquid. The sketches in Fig. 2.6a-d illustrate the
relationships. At T1, Fig. 2.6a, the liquid phase is stable over the entire
composition range, as it has a lower free energy. At T2, the liquid free energy
curve intersects both the a and b free energy curves, Fig. 2.6b. Two common
 Phase Diagrams 39

tangents are drawn to delineate the two two-phase regions, on either side of the
eutectic composition. At the eutectic temperature Te, all three phases have the
same free energy. A single straight line is tangential to all the three free energy
curves, Fig. 2.6c. At T3, the liquid phase has a higher free energy than either of
the solid phases at all compositions and there is no liquid present at this
temperature. The a and b free energy curves intersect each other, the common
tangent delineating the (a + b) region of the phase diagram, Fig. 2.6d.

Figure 2.6 Free energy versus composition relationships for a

simple eutectic system at different temperatures.

The common tangent principle described above serves to explain why the
solubility is more in a phase which is in equilibrium with a metastable phase
than with a stable phase. In Fig. 2.7, the free energy-composition curves for the
terminal solid solution a, the stable intermediate phase q, and the metastable
intermediate phase q are shown. If the a solid solution is in equilibrium with
q, the maximum solubility is ca as given by the common tangent. If it is in
equilibrium with the metastable q , the maximum solubility is ca. It is easily
seen that ca > ca.
40 Physical Metallurgy

Figure 2.7 The solubility in the matrix (a) is more, when the matrix is in equilibrium
with a metastable precipitate (q ), as compared to a stable precipitate (q).

2.4 MICROSTRUCTURAL CHANGES DURING COOLING

Slow equilibrium cooling It is instructive to consider first the changes that
occur microstructurally during extremely slow cooling, approximating the
equilibrium conditions. Consider the hypereutectic alloy of composition co in
the Ag-Cu system, Fig. 2.2. As this alloy is cooled below the liquidus temperature,
solidification starts. The first solid to crystalize has a composition cb1 as given
by the first tie-line. As the temperature falls further, more and more of the liquid
crystallizes. At the same time, the compositions of the solid and the liquid
phases continuously change with the tie-line, which occupies successive
horizontal positions with decreasing temperature. At T2, the solid composition
is cb2. Just above the eutectic temperature, the composition of the solid is cbe
and that of the liquid is ce. On cooling through Te, the eutectic reaction occurs
in the liquid region, yielding the eutectic mixture of a and b. On further cooling
below Te, the solubilities of copper in silver and silver in copper decrease along
the solvus lines and more a and b form by precipitation from the supersaturated
phases. Figure 2.2 also shows the successive changes in the microstructure
during cooling.
Nonequilibrium cooling In practice, the cooling rates are rarely slow enough
to obtain equilibrium conditions. The continuous adjustments in compositions
of the phases required for equilibrium may not take place.
Consider an alloy of composition co cooled from above the liquidus, Fig. 2.8a.
The first solid to crystallize has a composition ca1. As further solidification
occurs, the composition of the solid phase should shift along the solidus to ca2,
ca3, etc. Diffusion of atoms inside the solid phase is required to bring about
these compositional changes. Diffusion in the solid phase is many orders of
 Phase Diagrams 41

magnitude slower than in the liquid phase. So, only the solid that crystallizes
at any instant is in local equilibrium with the liquid with which it is in contact.
The solid that formed earlier at a higher temperature does not change its
composition according to the new tie-line, as diffusion is slow through the solid.
Thus a gradient of composition develops from ca1 at the centre of the crystal
which solidifies first to ca5 at the periphery of the crystal which solidifies last,
Fig. 2.8b. The average composition of the crystal is shown by the dotted line
in Fig. 2.8a. This lags behind the equilibrium composition, so that some liquid
is left over, even after crossing the equilibrium solidus at Ts. Solidification is
complete at a lower temperature, Ts, where the average composition of the solid
becomes equal to the overall composition of the alloy. This phenomenon is
called coring and is seen under the microscope through an etching effect produced
by the compositional gradient, as schematically shown in Fig. 2.8b. Figure 2.8c
shows the coring effect, when the growing crystals have the shape of a dendrite.
The extent of coring increases, with increasing separation between the liquidus
and the solidus, i.e. in alloys with a wide freezing range.

Figure 2.8 Cooling under nonequilibrium conditions results in coring with

compositional gradients from the centre to the periphery of the growing crystals.

Under nonequilibrium conditions of cooling of an alloy of a eutectic system,

the proeutectic crystals are cored. Due to this, more liquid is left over on
reaching the eutectic temperature. The fraction of the eutectic mixture formed
is correspondingly larger in the final microstructure. The coring effect may also
produce some eutectic mixture in an alloy of overall composition that lies
outside the eutectic horizontal.
42 Physical Metallurgy

Deviation from equilibrium is a serious problem in alloy systems in which

a peritectic reaction occurs. Recall that this reaction is between a liquid and a
solid. The product of the reaction is a solid, which forms at the boundary
between the two parent phases. As soon as some solid product (a) forms, the
parent phases (L and b) get separated and are no longer in contact. This
phenomenon is known as surrounding or enveloping, Fig. 2.9. Further reaction

Figure 2.9 In the peritectic reaction L + b a, surrounding or enveloping occurs,

when the product a phase surrounds the unreacted b phase.

can occur only through solid state diffusion through the intervening product
layer. Solid state diffusion being very slow, peritectic alloys solidify with large
deviations from the equilibrium composition and structure.
A homogenizing treatment is required to remove the compositional variations
that accompany coring or surrounding. The alloy is heated to a high temperature
and held there for a prolonged period, during which diffusion levels out the
compositional gradients and yields a homogeneous solid solution. Extreme care
is required during heating not to cross the solidus. Note that, for a cored structure,
the effective solidus is lower than the equilibrium solidus. If this solidus is
crossed, melting occurs near the grain boundaries. Liquation is often accompanied
by grain boundary oxidation and the phenomenon is called burning. A burnt
alloy is permanently damaged, as the oxidized grain boundaries cannot be restored
to the initial unoxidized condition by means of a heat treatment or by mechanical
working. Overheating refers to heating close to the solidus, but without causing
liquation. Overheating impairs the properties of steels through loss of ductility
and toughness. It is associated with the segregation of impurities and subsequent
precipitation along grain boundaries. The effects of overheating are not permanent
and can be removed by a subsequent heat treatment.

2.5 THE IRON CARBON EQUILIBRIUM DIAGRAM

In Fe-C alloys, carbon may be present either in the form of free carbon (graphite)
or as iron carbide (Fe3C). The most frequently occurring form in steels is Fe3C,
known as cementite. As already pointed out, cementite is metastable with respect
 Phase Diagrams 43

to graphite. The phase diagram in Fig. 2.10 shows Fe and Fe3C as the two
components. In cast irons, carbon is present either as graphite or cementite. To
denote this possibility, Fig. 2.10 has dotted phase boundaries corresponding to
the graphite equilibrium.
Phases in the Fe-C system a-ferrite is an interstitial solid solution of carbon
in BCC iron. It is stable over the temperature range from 273C to 912C.
Carbon is present in a few of the (0, 0, ) and other equivalent interstitial sites
in the BCC unit cell. The size of the largest atom that can fit in this site is 0.19,
which is much smaller than the size of the carbon atom (0.71). So, the solubility
is extremely limited. The maximum solubility is 0.02 wt% (0.10 at %) at 727C.
Austenite (g) is an interstitial solid solution of carbon in FCC iron. It is
stable from 912 to 1394C. Carbon is present in some of the (0, 0, ) and other
equivalent positions which includes the body centre (, , ). The size of the
largest atom that can fit this void is 0.52 . Correspondingly, the solubility is
larger here as compared to ferrite. The maximum solubility is at 1146C and is
2.1 wt% (9.1 at %)
d-ferrite (BCC) is stable from 1394 to 1539C, the melting point of iron.
The maximum solubility of carbon here is 0.09 wt%.

Figure 2.10 The Fe-Fe3C metastable phase diagram. The dotted phase
boundaries correspond to the stable graphite equilibrium.
44 Physical Metallurgy

Cementite is an intermetallic compound with the formula Fe3C. The atomic

percent of carbon in cementite is 25 and the weight percent is 6.67. Cementite
has a complex orthorhombic crystal structure, with twelve iron atoms and four
carbon atoms in the unit cell.
Graphite has a hexagonal crystal structure, in which sheets of primarily-
bonded carbon atoms are held together by secondary bonds.
Invariant reactions Three invariant reactions occur in the Fe-Fe3C system.
A peritectic reaction occurs at 1495C:
reaction : d + L g
structure : BCC FCC
composition, wt.% C : 0.09 0.53 0.17
On cooling through the peritectic temperature, d-ferrite reacts with the liquid to
yield austenite (g).
A eutectic reaction occurs at 1146C:
reaction : L g + Fe3C
austenite cementite
structure : FCC orthorhombic
composition, wt.% C : 4.3 2.1 6.67
On cooling through the eutectic temperature, the lowest melting liquid of the
system decomposes to two solid phases, austenite and cementite. This eutectic
mixture is known as ledeburite.
A eutectoid reaction occurs at 727C:
Reaction : g a + Fe3C
austenite ferrite cementite
structure : FCC BCC orthorhombic
composition, wt.% C : 0.77 0.02 6.67
On cooling through the eutectoid temperature, austenite decomposes into a
lamellar mixture of ferrite and cementite crystals, called pearlite.
Critical temperatures Critical temperatures are those at which phase changes
occur during heating or cooling an alloy. Certain symbols are used to denote the
critical temperatures in steels.
The eutectoid reaction during heating and cooling sets in at the critical
temperatures Ac1 and Ar1 respectively. The letters stand for the French words:
A for arrt (means arrest)
c for chauffage (means heating), and
r for refroidissement (means cooling).
If extremely slow rates of heating or cooling are employed, the corresponding
critical temperatures are nearly equal, i.e. Ac1 = Ar1 = A1 = 727C (eutectoid
temperature). At ordinary rates of heating or cooling, there is a thermal hysteresis,
and Ac1 > Ar1.
 Phase Diagrams 45

The next higher critical temperature A2 is the Curie temperature of iron

(768C), where ferrite undergoes the magnetic transition from the ferromagnetic
to the paramagnetic state on heating. There is no hysteresis in this electronic
transition.
The temperature corresponding to the (g + a)/g phase boundary for
hypoutectoid steels is termed Ac3 or Ar3. Unlike A1, the A3 temperature is a
function of carbon content. It decreases from 912C at 0% C to 727C. In
hypereutectoid steels, the g/(g + Fe3C) phase boundary is called Acm. The Acm
line is considerably steeper than the A3 line.
The Curie temperature of cementite (210C) is called A0.

Microstructure of Slowly Cooled Steels

Eutectoid steel The microstructure of eutectoid steel (0.77% C) is all pearlite.
Pearlite is a lamellar arrangement of alternate, parallel plates of ferrite and
cementite, Fig. 2.11. The relative amounts of ferrite and cementite in pearlite
can be determined from lever rule:
6.67 0.77
fB = = 0.89
6.67 0.02
f Fe3C = 1 0.89 = 0.11
The weight fractions of the phases are in the ratio 8:1. The densities of
ferrite and cementite are comparable (7870 and 7700 kg m3 respectively). The
volume fractions of ferrite and cementite in pearlite are also approximately in
the ratio of 8:1. Correspondingly, the cementite plates are much thinner than the

Figure 2.11 The microstructure of pearlite, showing the lamellar

arrangement of ferrite and cementite plates.

ferrite plates. When nital (a dilute solution of nitric acid in alcohol) is used as
the etchant, the boundaries between ferrite and cementite are selectively attacked.
46 Physical Metallurgy

As the two boundaries of a cementite plate are close together, they may not be
resolved as separate lines. Cementite often appears as a single dark line. At
higher magnifications or in a coarse distribution in pearlite, the cementite
boundaries show up as separate lines.
Hypoeutectoid steels The composition of hypoeutectoid steels is less than
0.77% C. In addition to pearlite, the proeutectoid ferrite (pro a) is present,
Fig. 2.12. The fraction of the pro a can be estimated from lever rule. For
example, in mild steel of 0.2% C,
0.77 0.2
f pro B = = 0.76
0.77 0.02
fpearlite = 1 0.76 = 0.24
The proeutectoid ferrite forms at the austenite grain boundaries between A3
and A1 temperatures.
Hypereutectoid steels Here, the proeutectoid phase is cementite. It separates
between Acm and A1 temperatures as a thin network along the austenite grain
boundaries. In a 1.2% C steel, applying lever rule, we have
1.2 0.77
f pro Fe3C = = 0.07
6.67 0.77
fpearlite = 1 0.07 = 0.93

Figure 2.12 Scanning Electron Micrograph of mild steel,

showing proeutectoid ferrite and pearlite.

2.6 EFFECT OF ALLOYING ELEMENTS ON THE

Fe-C DIAGRAM
The effect of alloying elements on the Fe-C phase relationships is best under-
stood by first considering the two types of phase diagrams obtained, when iron
is alloyed with a substitutional element.
 Phase Diagrams 47

Ferrite stabilizers Some alloying elements tend to stabilize the ferrite phase
in preference to austenite. Many of these elements have the same crystal struc-
ture as ferrite, i.e. BCC. They reduce the extent of the austenitic area on the
equilibrium diagram, by forming a gamma loop, as illustrated for the Fe-Cr
system in Fig. 2.13. Austenite (g) is enclosed within the loop. At chromium
contents greater than 12.7%, the d-ferrite and the a-ferrite regions merge and
become continuous. Such compositions exist only in the BCC form from the
lowest temperature up to the melting point. As austenite cannot be produced in
such compositions, they are not heat treatable. The critical compositions at
which the austenite phase disappears are given below for some typical ferrite-
stabilizing elements.

Element Cr Si Mo W V
Crystal structure BCC DC BCC BCC BCC
Critical composition, wt% 12.7 3 3 6 2
DC = Diamond cubic.

Figure 2.13 The iron-chromium phase diagram illustrates the

formation of a gamma loop.

Austenite stabilizers Manganese and nickel are the common austenite-

stabilizing elements. They enlarge the area of the austenite phase on the phase
diagram, as shown for the Fe-Ni system in Fig. 2.14. When sufficient amount
of nickel (or manganese) is present, austenite may be present at room temperature,
as the low temperature reactions seen in Fig. 2.14 may be too slow to occur.
48 Physical Metallurgy

Figure 2.14 The iron-nickel phase diagram, showing the

stabilization of the austenite phase by nickel.

The interstitial elements carbon and nitrogen are also austenite stabilizers.
Referring to the Fe-C diagram in Fig. 2.10, carbon is seen to extend the tem-
perature range over which austenite (g) is stable, as compared to the range for
pure iron.
Effect on eutectoid temperature and composition Figure 2.15a shows the
effect of different alloying elements on the eutectoid temperature. When the
alloying element is a ferrite stabilizer, the eutectoid temperature is raised above
727C, as the concentration of the elements is increased. Ti and Mo are among
the most effective elements in raising the eutectoid temperature. With an aus-
tenite stabilizer, the eutectoid temperature is depressed below 727C, see the
effect of Ni and Mn in Fig. 2.15a.
Figure 2.15b shows the effect of alloying elements on the eutectoid com-
position. Both ferrite and austenite stabilizers decrease the eutectoid composi-
tion from 0.77% C to lower values.
 Phase Diagrams 49

Figure 2.15 The effect of common alloying elements on

(a) the eutectoid temperature, and (b) the eutectoid composition.

The Fe-C-X phase relationships The effect of alloying elements on the

important phase boundaries of the Fe-C system is illustrated below using
pseudobinary plots. Figure 2.16 shows the A3 and Acm lines for different
manganese contents. As the manganese content increases, the austenite phase
field is enlarged, the eutectoid temperature moves down, and the eutectoid
composition shifts to lower carbon contents. Figure 2.17 shows the effect of
a ferrite stabilizer (chromium) on the Fe-C diagram. As the concentration of
chromium increases, the four phase boundaries enclosing the austenite field
50 Physical Metallurgy

move closer, thus restricting the extent of the austenite phase. To the left of
the field are austenite and ferrite and to the right austenite and carbide. The
eutectoid temperature moves up. The eutectoid composition shifts to lower
carbon contents. With Cr > 19%, the austenite phase field disappears altogether.
The critical composition at which this happens are given below for a few
typical cases.

Ferrite stabilizer Cr Si Mo W V
Critical composition, wt.% 19.5 8.1 6.4 10.8 4

Figure 2.16 The shifting of the A3 and Acm phase boundaries

in the presence of manganese.

Carbon being an austenite stabilizer, a higher concentration of a ferrite

stabilizing element (19.5% Cr for Fe-C-Cr as compared to 12.7% Cr for Fe-Cr)
is required in the presence of carbon to eliminate the austenite field.
For low alloy steels containing C between 0.08 and 1.4 wt%, Ac1 and Ac3
critical temperatures are given by the following empirical equations:

Ac1 (C) = 723 10.7Mn 16.7Ni + 29.1Si

+ 16.9Cr + 290As + 6.38W
Ac3 (C) = 910 203 C 30Mn 20Cu 15.2Ni
11Cr 700P + 44.7Si + 31.5Mo
+ 104V + 460Al + 13.1W + 120As
 Phase Diagrams 51

Figure 2.17 The shrinking of the austenite phase field with

increasing concentration of chromium in steels.

2.7 THE COPPER-ZINC PHASE DIAGRAM

A number of important binary phase diagrams of nonferrous alloys are known.
As an example, the Cu-Zn phase diagram is discussed below. The relevant parts
of many other diagrams such as Al-Zn, Cu-Sn, Cu-Al and Cu-Be are introduced
at appropriate places, when discussing engineering alloys in Chap. 7.
The melting points of copper (1083C) and of zinc (419C) are quite different.
The Cu-Zn phase diagram shown in Fig. 2.18 has a number of peritectic reactions
and intermediate phases. The a terminal solid solution is based on the crystal
structure of copper (FCC), with zinc substitutionally dissolved. It is a commonly
occurring phase in commercially important Cu-Zn alloys. It has excellent
mechanical properties, with both strength and ductility increasing with increasing
zinc content.
The intermediate phase b has the BCC crystal structure. It undergoes an
order-disorder transformation on heating above 465C. In the disordered bphase,
the copper and zinc atoms are randomly distributed on the BCC sites. In the
ordered b phase, copper occupies the body corners and zinc the body centres
of BCC (or vice versa). The b phase has little ductility, but on disordering, the
b phase becomes quite ductile. The g, d and e phases of the Cu-Zn system have
complex crystal structures and are too brittle to be of any commercial use.
52 Physical Metallurgy

Figure 2.18 The copper-zinc phase diagram has a number of

peritectic reactions and intermediate phases.

Within the a and b composition ranges, only one peritectic reaction occurs.
reaction : L + a b
structure : FCC BCC
composition, wt% Zn : 37.5 32.5 36.8
The slopes of the a/(a + b) and (a + b)/b boundaries are unusual. Between
900 and 400C, the solubility of zinc in copper increases with decreasing tem-
perature (as opposed to the normal decrease). A 37.5% Zn alloy is all b at
900C. On slow cooling, the alloy gradually transforms into a mixture of (a + b).
Below 550C, it is fully converted to a. On further cooling below 300C, under
equilibrium conditions, the alloy should become (a + b) mixture, even though
this may not happen at such low temperatures.
Commercial alloys of copper and zinc are discussed in Chap. 7.

2.8 TERNARY PHASE DIAGRAMS

Ternary phase diagrams depict the equilibrium between three components. As
in binary diagrams, it is common to omit the pressure variable. The remaining
 Phase Diagrams 53

variables are two compositions and temperature, requiring three axes for graphical
representation. As perspective three-dimensional diagrams are not easy to draw
or understand on paper, complex ternary relationships are usually presented as
isothermal sections, i.e. each diagram corresponds to the phase equilibrium at
a specified temperature. A number of such isothermal sections describe fully the
ternary system.
The composition variables at a constant temperature are best represented in
the form of an equilateral triangle (called the Gibbs triangle), the corners of
which correspond to the three pure components, Fig. 2.19a. Each side of the
triangle shows the variation in the concentration of the two components at the
two ends of that side, just as the x-axis of a binary phase diagram. All
compositions that lie on a side consist of two components only. Compositions
that lie within the triangle contain all the three components. For reading the
compositions of an alloy at point o within the triangle, lines xy, xy and xy
parallel to the three sides BC, CA and AB of the triangle are drawn through the
point o as in Fig. 2.19a. The length Bx or Cy is equal to the percentage of
component A which is at the apex with respect to the lines BC and xy. Thus,
% A = Bx = Cy
% B = Cx = Ay
% C = Ax = By

Figure 2.19 (a) The reading of composition on a ternary phase diagram. (b) Use of the
tie-line and the tie-triangle to calculate the relative amounts of coexisting phases.

The level rule to determine the relative proportions of two coexisting phases
is applied in the same way as in the binary diagram. For example, if the overall
composition p in Fig. 2.19b exists as a mixture of two phases a and b of
compositions q and r, then the relative amounts of a and b are given by:
54 Physical Metallurgy

fa = pr/qr
fb = qp/qr
If three phases coexist, their relative proportion is determined from a tie-
triangle. For example, if the overall composition o in Fig. 2.19b exists as a
mixture of g, d and e, the respective compositions being s, t and u, the relative
proportions of the three phases are given by (see Fig. 2.19b):
fg = ol/sl
fd = om/tm
fe = on/un
An example of an isothermal section of the Fe-Cr-Ni system at 650C is
shown in Fig. 2.20. Two tie-triangles are seen in this section. Three phases are
present in each of the tie-triangles.

Figure 2.20 Fe-Cr-Ni ternary isothermal section at 650C.

Another common way of presenting the ternary phase relationships is to

draw vertical sections (called isopleths), keeping the concentration of one com-
ponent constant. The x-axis then gives the variation of one of the other two
components. Temperature is plotted on the y-axis. Figure 2.21 shows a vertical
section of the Fe-Cr-Ni system at a constant Fe content of 74 wt%.
 Phase Diagrams 55

Figure 2.21 Fe-Cr-Ni vertical section at a constant iron content of 74 wt.%.

Figure 2.22 The three-dimensional view of the ternary phase diagram

for the case of complete liquid and solid solubility.
56 Physical Metallurgy

Perspective views of ternary phase diagrams can be drawn on a triangular

prism. They are useful in visualizing the phase relationships in simple cases.
The vertical axis of the prism is the temperature axis and any section parallel
to the base is an isothermal section. Figure 2.22 shows a perspective view of
the ternary system, in which all the three pairs of binaries exhibit complete solid
solubility. Within the triangle, the (liquid + solid) two-phase field is a three-
dimensional space enclosed by the liquidus surface at the top and the solidus
surface at the bottom. Above the liquidus surface, a ternary liquid solution
exists. Below the solidus surface, it is a ternary solid solution.
Figure 2.23 shows a ternary phase diagram formed by three components,
which pair up as simple eutectic systems. There are four phases in the system,
corresponding to the liquid and the three terminal solid solutions a, b and g.
Three-phase eutectic reactions occur over a range of temperatures, along liquidus
lines e1e, e2e and e3e. For example, the three-phase eutectic reaction:
L a+b
occurs along line e1e. Note that this eutectic reaction is not invariant. For C = 3
and P = 3, the phase rule Eq. (2.4) gives F = 1.

Figure 2.23 The three-dimensional view of the ternary phase diagram between
three components, which pair up as simple binary eutectic systems.
 Phase Diagrams 57

e is an invariant point in the ternary system, where all the four phases
coexist, with zero degree of freedom. When the liquid has this composition, a
four-phase eutectic reaction occurs isothermally:
L a+b+g

2.9 QUATERNARY PHASE DIAGRAMS

With increasing complexity of industrial alloys, multi-component systems, in
which the number of components may be more than three, are gaining impor-
tance. A quaternary system comprises four components A, B, C and D. After
omitting the pressure variable, there are still four variables to be shown on a
graphical representation. If the temperature is kept constant, a three-dimensional
regular tetrahedron (Fig. 2.24) can show the phase relations. The four compo-
nents are located at the four corners A, B, C and D, the six sides of the
tetrahedron represent the six binaries AB, AC, AD, BC, BD and CD, and the
four faces of the tetrahedron are isothermal sections of the four ternaries ABC,
ABD, ACD and BCD. Any composition inside the quaternary space has all the
four components in it. The equilibrium between four phases in the quaternary
space is represented by a tie-tetrahedron (similar to the tie-triangle in ternary
systems). A tie-tetrahedron can be of any irregular shape. The equilibrium be-
tween a, b, t, and s phases is schematically sketched in Fig. 2.24.

Figure 2.24 The composition tetrahedron for the quaternary system A-B-C-D.
A tie-tetrahedron depicting four-phase equilibrium lies within.
58 Physical Metallurgy

For two-dimensional representation, the concentration of one of the

components and temperature is to be kept constant. Figure 2.25 shows an
isothermal section of the Fe-Cr-V-C system at 700C and at a constant C content
of 0.2 wt%. M23C6, M7C3 and MC are alloy carbides, with M standing for
(Fe, Cr, V). ABC is the triangular section of a tie-tetrahedron, showing a four-
phase equilibrium.

Figure 2.25 Fe-Cr-V-C quaternary isothermal section at 700C and

at a constant carbon content of 0.2 wt%.

2.10 EXPERIMENTAL DETERMINATION OF PHASE

DIAGRAMS
A variety of physical techniques are used for the experimental determination of
phase diagrams. Among these, the more important are:
(i) thermal analysis,
(ii) microscopic examination, and
(iii) x-ray diffraction.
Other auxiliary techniques are dilatometry, electrical resistivity and mag-
netic measurements.
 Phase Diagrams 59

Determination of the liquidus and the solidus Thermal analysis is by far

the most important method employed for determining the liquid-solid transfor-
mations. The method is based on the fact that evolution or absorption of heat
occurs during a phase change. When the temperature of an alloy sample is
plotted as a function of time during heating or cooling, abrupt changes in slope
of the plot occur at points corresponding to the start or finish of a phase change.
Figure 2.26 illustrates some typical cases. Figure 2.26a is the cooling curve
for a pure metal. As the liquid is cooled to the freezing point, solidification
starts with the evolution of latent heat. The heat evolved compensates for the
heat lost to the surroundings, so that the temperature of the sample remains
constant. The cooling curve becomes horizontal and remains so, till the solidi-
fication is complete. Thereafter, the temperature drops again with time, with an
abrupt change in slope at the point of completion of the solidification. Often,
a liquid may supercool below the freezing point. But once solidification starts,
heat evolution raises the sample temperature back to the freezing point as
shown in Fig. 2.26b.

Figure 2.26 Thermal analysis. Cooling curves (a) for a pure metal without
supercooling, (b) for a pure metal with supercooling,
(c) for a solid solution alloy, and (d) for a hypoeutectic alloy.
60 Physical Metallurgy

When heating a pure solid metal, similar breaks in the heating curve occur.
In the horizontal part of the curve, the heat input is used up for the latent heat
absorbed during melting. No superheating is known to occur. The temperature
corresponding to the horizontal part of the cooling and heating curves must
coincide. This gives the true freezing/melting point of the metal.
Figure 2.26c is the cooling curve for the solidification of a solid solution.
The first arrest indicating an abrupt decrease in the cooling rate occurs as the
liquidus temperature Tl is crossed. There is no horizontal portion in the cooling
curve here, as solidification occurs over a range of temperature between the
liquidus and the solidus. The second arrest is observed at the solidus tempera-
ture Ts, when the cooling rate starts to increase again.
Figure 2.26d is the cooling curve for a proeutectic binary alloy. The first
arrest occurs at the liquidus temperature and the cooling rate decreases. At the
eutectic temperature Te, the cooling curve becomes horizontal, as the invariant
eutectic reaction occurs isothermally. The cooling rate increases again, when the
eutectic solidification is complete. If the alloy has exactly the eutectic compo-
sition, only the horizontal part in the cooling curve is seen, corresponding to the
eutectic solidification. In principle, a peritectic alloy should show a similar
behaviour. However, the phenomenon of surrounding prevents the peritectic
reaction from going to completion isothermally. The horizontal part in the cool-
ing curve may not be clearly defined.
The following precautions are taken in determining cooling curves:
1. Local fluctuations of temperature in the furnace should be avoided.
2. The rate of cooling must be as slow as feasible for near equilibrium
conditions to be achieved.
3. The molten alloy should be stirred for uniform temperature and to reduce
supercooling.
4. The analysis for composition must be accurate.
5. No contamination must occur from the crucible, thermocouple sheath,
etc.
The solidus is determined only approximately from the break in the cooling
curve, as coring tends to lower the solidus temperature. A more accurate
determination can be carried out by holding the solidified alloy for a long time
for homogenization just below the approximate solidus temperature determined
from the cooling curve. The alloy is heated again to find the break in the heating
curve. This break is at a higher temperature than that determined during cooling,
as coring effects are removed by homogenization. The thermal analysis is
supplemented with microscopic examination. The alloy is heated close to the
solidus and quenched to ascertain microscopically the appearance of the first
chilled liquid.
 Phase Diagrams 61

Using the breaks in the cooling and heating curves for a series of
compositions covering the entire binary range, the liquidus and the solidus
boundaries can be fully determined, as illustrated in Fig. 2.27 for an isomorphous
system.

Figure 2.27 The construction of liquidus and solidus phase

boundaries from cooling curves. Each cooling curve
(shown in perspective view) corresponds to a single composition.

Determination of the solvus The common methods for determining the solvus
are microscopic examination and x-ray diffraction. A series of small ingots of
alloys of different compositions are prepared and homogenized. They are then
annealed at various temperatures for a prolonged time (a few days) and quenched.
In some cases, the high temperature phases may be retained on quenching. In
such cases, subsequent metallographic and x-ray studies at room temperature
disclose the appearance of the second phase at some known composition. In
other cases, the high temperature phase may decompose on quenching. If this
happens, room temperature x-ray studies are not suitable. A high temperature
x-ray camera may be used. Even after the decomposition during quenching, the
metallographic method may be useful, if the transformed phase is in an easily
recognizable form. A phase boundary is first bracketed between two compositions,
as illustrated in Fig. 2.28. The exact location of the boundary is determined by
studying a few more alloys of closely varying composition in the boundary
region.
62 Physical Metallurgy

Figure 2.28 The determination of solvus boundaries by the

metallographic and x-ray methods.

Auxiliary techniques Dilatometry is based on the volume change associated

with most phase changes. It is common to record the length changes of a
specimen as a function of temperature during heating or cooling. Figure 2.29

Figure 2.29 The fractional change in length of a low carbon steel on heating
and cooling. The abrupt changes in slope correspond to phase changes.
 Phase Diagrams 63

shows the length changes in a low carbon steel during heating and cooling. The
interpretation of the curve is similar to that for thermal analysis. Note the
appreciable thermal hysteresis that occurs during this solid state phase change.
The electrical resistivity of a solid solution changes nonlinearly with the
concentration of the solute. However, the resistivity of a mixture of two phases
changes linearly with the volume fraction of the phases. The resistivity of the
alloys of a simple eutectic system is shown in Fig. 2.30a. The abrupt change in
slope of this plot indicates the location of a phase boundary (solvus). The
electrical resistivity of an ordered solid solution is low as compared to that of
the random solid solution. Figure 2.30b shows the electrical resistivity as a
function of composition for the Au-Cu system. The resistivity falls to a minimum
at compositions corresponding to the ordered compounds, AuCu and AuCu3.
An order-disorder transition can thus be identified.

Figure 2.30 Resistivity changes as a function of composition (a) in a eutectic

system, and (b) in a solid solution, where ordering occurs.
The saturation magnetic intensity of a ferromagnetic phase is very large as
compared to a paramagnetic phase. This fact can be used to detect the appearance
or disappearance of a phase, if one of the phases involved is ferromagnetic. This
method is suitable only when it is used in conjunction with one or more of the
other methods described above.

REFERENCES

1. Prince, A., Alloy Phase Equilibria, Elsevier, Amsterdam (1966).

2. Metallography and Microstructures, Metals Handbook, Vol. 9, ASM Inter-
national, Materials Park, Ohio (2000).
64 Physical Metallurgy

QUESTIONS

1. In a three component system, what is the maximum number of phases that

can coexist in equilibrium?
Answer: 5 (taking pressure variable into account).
2. Find the atomic % of carbon in mild steel containing 0.2 wt.% C.
Answer: 0.92
3. A binary eutectic freezes at a fixed temperature, whereas in a ternary
system with three-phase equilibrium, freezing is over a range of tempera-
ture. Explain why this is so.
4. Explain why coring occurs to a greater extent in Cu-Sn alloys than in
Cu-Zn alloys.
5. Explain why a eutectic composition does not show coring, whereas a solid
solution does.
6. Compare and contrast the phenomena of overheating and burning.
7. Give the invariant reactions (and their names) in the Cu-Zn phase
diagram.
8. Explain the relative positions of the Acm line and the corresponding graphite
line on the Fe-C diagram, on the basis of free energy composition
relationships.
9. Calculate the fractions of proeutectoid ferrite, eutectoid ferrite and total
ferrite in a 0.2% C steel.
Answer: 0.76, 0.21, 0.97.
10. What is the fraction of pearlite in a slowly cooled 0.6% C steel? If the
steel contains in addition 1% Mn and 0.5% Si, what is the fraction of
pearlite?
Answer: 0.77, ~ 0.9.
11. Sketch the microstructural changes that occur in a brass of 37.5% Zn,
when cooled slowly from 900C to room temperature.
12. Draw a set of cooling curves obtained from thermal analysis of a simple
binary eutectic system.
13. Draw heating curves for a proeutectic and a eutectic composition.
Chapter 3
PHASE CHANGES

The properties of engineering alloys are optimized by a careful control of the

microstructure. Phase changes (also called phase transformations) that occur in
an alloy determine its microstructural features. For understanding the underlying
principles in the control of microstructure through thermal or thermomechanical
treatments, a knowledge of the mechanism and kinetics of phase changes is
essential.

3.1 TYPES OF PHASE CHANGES

A phase change may involve a change in the state of aggregation (e.g. liquid
solid), in the crystal structure (e.g. BCC FCC) or in composition (e.g.
0.77%C in austenite 0.02%C in ferrite + 6.67%C in cementite). Compositional
changes require long-range diffusion, where the atoms move through several
thousands of interatomic distances to bring about the change. Phase
transformations without a change in composition can occur by short-range
diffusion, where the atoms move only over one or two interatomic distances
from the parent phase to the product phase across an interface. Such
transformations can also occur by a shear process, where a coordinated movement
of atoms (each by a fraction of an interatomic distance only) can change the
parent crystal structure to the product structure.
There are two basic mechanisms of phase transformations. In the nucleation-
and-growth mechanism, a tiny particle identical in structure and composition to
the product phase nucleates and then grows in size. In spinodal decomposition,
compositional fluctuations occur in the parent phase. These fluctuations grow
65
66 Physical Metallurgy

in intensity with time, until two discrete phases corresponding to the equilibrium
compositions form. Many metallurgical transformations occur by nucleation and
growth. We will consider only this mechanism.

3.2 DIFFUSION IN SOLIDS

Diffusion is the mass flow process in which atoms change their positions rela-
tive to their neighbours in a phase under the influence of thermal energy and
a concentration gradient.
Ficks laws Ficks first law describes steady state diffusion and is written as
dc
J =D (3.1)
dx
where J is the diffusional flux in the x-direction (number of atoms transported
per unit area per unit time), dc/dx is the concentration gradient, and D is the
diffusion coefficient.
Under steady-state conditions, the flux remains independent of time t and
position x.
Ficks second law is applicable to nonsteady state diffusion. Here, the flux
is changing as a function of position and time. When the diffusion coefficient
is not a function of concentration, the second law is of the form:
c 2c
=D 2 (3.2)
t x
Solutions to the second law for different initial and boundary conditions are
available, but will not be covered here. A quick but approximate estimate of the
depth of penetration x during nonsteady state diffusion is given by
x = Dt (3.3)
Atomic model of diffusion In interstitial diffusion, Fig. 3.1a, the diffusing
atom jumps from one interstitial site to a neighbouring interstitial site in execut-
ing a unit step. The neighbouring site is usually vacant in dilute solutions. The
activation energy for interstitial diffusion is simply the energy barrier along the
path from one interstitial site to the next. Substitutional diffusion occurs with
the aid of vacancies, Fig. 3.1b. A substitutional atom jumps from a regular site
into a neighbouring vacant site in executing the unit step. Here, in addition to
considering the probability of an atom crossing the activation barrier along the
path, the probability of the neighbouring site being vacant has also to be taken
into account. Because of this, substitutional diffusion is generally 46 orders of
magnitude slower than interstitial diffusion.
The unit step in diffusion is executed with the help of thermal energy.
So, the diffusion coefficient D is strongly dependent on temperature and is of
the form
 Phase Changes 67

Q
D = D0 exp (3.4)
RT
where D0 is the pre-exponential constant and Q is the activation energy for
diffusion.

Figure 3.1 Atomic model of (a) interstitial diffusion, and

(b) substitutional diffusion by the vacancy mechanism.

In several phase changes, diffusion occurs along special paths such as grain
boundaries or across interfaces separating two phases. The activation barrier for
diffusion along such special paths is lower and D is correspondingly larger than
that for lattice diffusion described above. However, the cross-sectional area
across which mass transfer can occur is limited for grain boundaries as compared
to the lattice. Due to this, grain boundary diffusion tends to dominate only at
low temperatures. At higher temperatures, lattice diffusion becomes more
important.

3.3 NUCLEATION AND GROWTH KINETICS

When the free energy of the parent phase becomes higher than that of the
product phase through a change in temperature or pressure, a driving force for
a phase change arises. Figure 3.2 shows the free energy of a pure metal per unit
volume as a function of temperature for the liquid phase gl and for the solid
phase gs. If Tm is the melting point of the metal, the free energy change Dg for
the liquid solid transformation is:
above Tm , gs gl = Dg > 0,
at Tm , gs gl = 0, and
below Tm , gs gl < 0.
Thus, the free energy change becomes negative only below the melting
point. Only then, the liquid can spontaneously transform to the solid.
When Dg < 0, thermal fluctuations result in the formation of very tiny
particles (containing only a few atoms) of the product phase within the parent
volume. Such a tiny particle has an interface that separates it from the parent
matrix. It grows by transfer of atoms across this interface. The transformation
can be divided into two steps that occur sequentially:
68 Physical Metallurgy

Figure 3.2 A driving force for the liquid-to-solid transformation

arises below the melting point Tm.

1. the formation of tiny particles that are stable to further thermal fluctuations
and will not dissolve; and
2. the increase in the size of these particles.
Step 1 is called nucleation and step 2 is growth.
The interface that separates the two phases has a surface energy. This
energy has to be supplied during nucleation. It acts as an inhibiting factor for
the phase change to occur. A very small particle has a large surface area-to-
volume ratio and can, therefore, be unstable.
Homogeneous nucleation In homogeneous nucleation, the probability of nucle-
ation remains constant throughout the volume of the parent phase. If Df is the free
energy change during nucleation of a spherical particle of radius r, we can write
Df = (4/3)pr3 Dg + 4pr2s (3.5)
where s is the specific surface energy of the interface. When a driving force
exists for the phase change, Dg is negative, but s is always positive. When the
particle is initially very small, Df increases with r. As r increases further, Df
reaches a maximum and then decreases, Fig. 3.3. The value of Df corresponding
to the maximum is called the critical nucleation barrier (or the activation free
energy of nucleation) Df* and is found by setting dDf/dr = 0:
16QT 3
'f * = (3.6)
3( 'g ) 2
The corresponding radius of the nucleating particle r* is given by
2T
r* = (3.7)
'g
Dg is zero at the equilibrium temperature; so Df* = . When Dg < 0, the
nucleation barrier becomes finite. With increasing driving force and super cooling,
both the barrier energy and the critical radius decrease, see Fig. 3.3.
 Phase Changes 69

Figure 3.3 The free energy of nucleation Df as a function of radius r

of a spherical nucleating particle.

The barrier to nucleation is overcome with the aid of thermal fluctuations.

At the equilibrium temperature T0, Df* is infinite and the nucleation rate I is zero.
At 0 K, there are no thermal fluctuations and the nucleation rate is again zero. It
passes through a maximum at some intermediate temperature, as shown in Fig. 3.4.

Figure 3.4 The nucleation rate I and the growth rate U as a function
of temperature below T0, which is the equilibrium temperature.
70 Physical Metallurgy

When both the parent and the product phases are solids, volume or shape
changes introduce elastic strains during nucleation. The strain energy is added
as a positive energy term to the nucleation Eq. (3.5). This has an additional
inhibiting effect on the transformation. To minimize the total energy, the particle
usually assumes a nonspherical shape such as a disc or a plate.
The increase in the size of the particle beyond the critical size is called
growth. Like nucleation, growth is dependent on thermal fluctuations. The growth
rate U is zero at the equilibrium temperature T0, as well as at 0 K. It passes
through a maximum at some intermediate temperature, which turns out to be
higher than that for the nucleation maximum, as shown in Fig. 3.4.
Heterogeneous nucleation Here, there are certain preferred sites in the parent
phase, where the probability of nucleation is much greater than that at other
sites. The preferred sites are the walls of the container (in the case of a liquid),
inclusions, grain boundaries, stacking faults and dislocations (in the case of
solids).
In Fig. 3.5, the nucleation of the new phase b on the surface of an inclusion
d embedded in the parent a phase is schematically shown. The b particle is in
the form of a spherical cap. The contact angle q is determined by equilibrium
between surface tension forces (g) acting at the periphery of the b particle. The
energy of the inclusion surface is utilized during nucleation. The barrier for
heterogeneous nucleation Df *het turns out to be
1
Df *het = Df *homo (2 3 cos q + cos3 q) (3.8)
4
The key to the reduction of the nucleation barrier is a small value of q.
When the crystal structures of d and b are similar and the lattice parameters are
nearly equal, a low energy interface forms between them. The contact angle q
is then small and the nucleation barrier is very effectively lowered.

Figure 3.5 The nucleation of a b phase particle on the surface of a nucleating

agent d. The surface tension forces gab , etc. act along the indicated directions.

Overall transformation rate The overall transformation rate dX/dt, where X

is the fraction of the product formed, is a function of both the nucleation rate
and the growth rate. It has the same temperature dependence as nucleation and
 Phase Changes 71

growth, i.e. it is zero at T0, increases with decreasing temperature (or increasing
supercooling), reaches a maximum and then decreases to zero at 0 K. The data
on transformation rate are usually plotted in the form of a T-T-T (temperature-
time-transformation) diagram. The time for a fixed fraction of transformation is
plotted as a function of temperature. This has a C-shape, where the nose of the
C-curve corresponds to the minimum time (or the maximum in the transformation
rate).
The grain size of the product phase depends on the relative rates of nucleation
and growth. Each nucleating particle becomes a grain in the final product. So,
a high nucleation rate means a large number of grains. Also, when this is
combined with a low growth rate, more time is available for further nucleation
to take place in the parent phase that lies between slowly growing particles. So,
the combination of a high nucleation rate and a low growth rate yields a fine
grain size, see Fig. 3.4. On the other hand, a low nucleation rate combined with
a high growth rate yields a coarse grain size.
In Fig. 3.4, the temperature of maximum rate of nucleation is lower than
the temperature of maximum growth rate. An increase in cooling rate lowers the
effective transformation temperature and results in the combination of a high
nucleation rate and a relatively slow growth rate and, therefore, yields a fine
grain size. A progressive increase in cooling rate may show the following
phenomena.

Cooling rate Structure Typical industrial

C/sec process
102 Fine grain size Chill casting
103104 Very fine grain size Rapid solidification
processing
105106 Metastable crystal structures

107 Metallic glass Splat quenching
(if the parent phase is liquid)

3.4 SOLIDIFICATION
Solidification refers to the phase change from the liquid state to the solid
crystalline state. Its study is important in physical metallurgy, as it is directly
related to the industrial production of castings and ingots and the growth of
single crystals from the melt.
Grain structure of an ingot In addition to aspects of nucleation and growth,
the rate of heat flow from the cast liquid to the surroundings through the
solidified layer and the mould wall is equally important in industrial processes.
Figure 3.6 shows the typical grain structure of a cross-section of a large ingot.
The outer zone called the chill zone consists of small, equiaxed crystals that
form under the large supercooling that occurs as soon as the liquid metal comes
72 Physical Metallurgy

into contact with the cold mould walls. Some of the crystals of the chill zone
grow preferentially in a direction perpendicular to the mould wall (opposite to
the direction of heat flow) into long columnar crystals. The liquid at the centre
undergoes solidification under a relatively small undercooling, forming large
equiaxed crystals.

Figure 3.6 The grain structure of a large ingot, showing the chill zone,
the columnar grains and the central equiaxed zone.

Nucleation and grain size The nucleation principles described in the previous
section can be applied to solidification. No strain energy is usually involved in
solidification, as the volume change is accommodated by the flow of the liquid
around. An increase in cooling rate means more supercooling and the
solidification occurs nearer the temperature of the maximum nucleation rate,
resulting in a fine grain size. For example, chill casting (casting in a metal
mould) yields a finer grain size as compared to a sand casting.
In industrial practice, nucleating agents are added to promote
heterogeneous nucleation. Titanium and boron are added to aluminium alloy
melts, ferrosilicon to cast iron (to promote the nucleation of graphite) and
zirconium to magnesium alloys. The lattice parameter of zirconium (HCP:
a = 3.23 ) is close to that of magnesium (HCP: a = 3.21 ), suggesting
a small contact angle during heterogeneous nucleation, when close-packed
planes of the two crystals can have a one-to-one correspondence of atoms
at the interface. This effectively lowers the nucleation barrier, increases the
nucleation rate and produces a fine grain size in the casting. In addition to
 Phase Changes 73

lowering the contact angle by good lattice matching, a successful nucleating

agent should be stable in the molten metal and possess a large surface area
and roughness.
Another way to stimulate nucleation in undercooled melts is to introduce
vibrations in the liquid. Vibrations can be applied by mechanical, sonic or
ultrasonic means. A pulse of sufficient intensity causes cavitations (formation of
voids in the liquid). When the voids collapse, large pressure differences are
created and this promotes nucleation.
In recent years, in addition to nucleation, grain multiplication is also con-
sidered important in controlling the final grain size. In grain multiplication,
some arms of the dendrite crystals break off by melting and move over to other
parts of the melt. If there is undercooling at the new location, these broken arms
grow into new separate crystals. Vibrations can aid in the break away of the
dendrite arms.
The solidified grain structure may be further modified by solid state phase
changes during cooling of the casting or ingot. This has an important grain
refining effect in steels. Solidification starts as d-ferrite. This changes to auste-
nite on further cooling, which in turn transforms into a-ferrite and pearlite.
Where no such solid state changes are available to exploit, grain refinement of
a casting, which cannot be cold worked subsequently, is possible only by the
methods outlined earlier.
Constitutional supercooling Referring to the schematic phase diagram in
Fig. 3.7a, at temperature T, liquid of composition cl is in equilibrium with solid
of composition cs as given by the tie-line. During solidification, if the interface
that separates the solid and the liquid phases is at temperature T, the liquid near
it is of composition cl, which is greater than the average composition c0 of the
melt. The liquid composition decreases with increasing distance from the inter-
face, as shown in Fig. 3.7b. With the aid of the phase diagram, the equilibrium
liquidus temperature can be plotted as a function of distance, Figs. 3.7c and d.
In Fig. 3.7c, the actual temperature gradient in the liquid ahead of the interface
is such that the liquid at every point is at a temperature above the liquidus
temperature for that point. Therefore, there is no supercooling ahead of the
interface, which is stable and flat, Fig. 3.7e. Any protrusions that may form on
the flat interface feels superheated and melts back.
In Fig. 3.7d, the actual temperature profile in the liquid is such that the
temperature at every point between x and y is lower than the equilibrium
liquidus temperature corresponding to that point. Any protrusion formed in this
region feels supercooled and does not melt back. As the supercooling here is
due to a compositional effect, this phenomenon is known as constitutional
supercooling. This effect occurs, when the temperature gradient in the liquid
ahead of the interface is small or when the interface velocity is large. Under
unstable conditions of growth, an interface with protrusions develops, see
Fig. 3.7f.
74 Physical Metallurgy

Figure 3.7 Constitutional supercooling: (a) the phase diagram, (b) the composition
profile in the liquid ahead of the interface, (c) and (d) the temperature profiles
ahead of the interface, and (e) and (f) the profiles of the interface.

Formation of dendrites Under conditions of constitutional supercooling, a

cell structure with protrusions forms. If the velocity of the interface (the growth
rate of the cells) increases further, the cells tend to grow along certain preferred
directions such as <100> for cubic metals. Then a dendrite structure develops,
where the growing crystal is in the form of a Christmas tree, Fig. 3.8. Recall
that the <100> directions are three mutually perpendicular directions. One of
them coincides with the growth direction of the central (primary) arm of the
dendrite. The other two perpendicular <100> directions correspond to the
secondary and tertiary arms of the dendrite. A number of primary dendrites
might have branched off from the same nucleus and thus be part of the same
grain, see Fig. 3.8. The spacing between dendrite arms decreases with increasing
cooling rate.
 Phase Changes 75

Figure 3.8 A dendrite grows in three mutually perpendicular directions,

giving rise to primary, secondary and tertiary arms.

Directional solidification As described above, the dendrites grow along certain

preferred crystallographic directions. When the dendrites of neighbouring grains
grow in the same direction, we obtain the columnar structure. The direction of
columnar growth is always the preferred dendritic direction, i.e. one of the
<100> directions for cubic metals. The grains are randomly oriented in the chill
zone of an ingot. When columnar growth begins, the most favourably oriented
grains crowd out the less favoured ones and grow preferentially. An additional
factor that may favour selective growth is that such grains provide the most
favourable path for heat flow.
The improved properties of unidirectionally solidified alloys have been
exploited in a number of applications. The alignment of the boundaries of the
columnar grains in relation to the direction of stress leads to better creep resistance
in turbine blades. The columnar structure of directionally solidified Alnico
permanent magnets nearly doubles the magnetic energy product. Aligned
composites with finely-distributed phases can be obtained by directional
solidification of eutectics.
Microsegregation During solidification of an alloy, the solute atom partitions
itself in different proportions in the liquid and the solid. Under nonequilibrium
conditions of cooling, coring manifests itself and the solute gets segregated in
the volume of the liquid that solidifies last. During dendritic growth, the liquid
to solidify last is in the spaces between the dendritic arms. This segregation of
the solute in the solid that forms last is known as microsegregation. Compositional
variations due to this segregation are seen in a composition-distance profile
across dendritic arms obtained from an electron probe microanalyzer. The
wavelength of the compositional variation gives the dendritic arm spacing.
76 Physical Metallurgy

Homogenization is the process of heating the casting for a prolonged time

at a high temperature. This allows diffusion to occur in the solid state and tends
to wipe out or reduce microsegregation. The distance over which diffusion is
to occur and the time of annealing during homogenization are determined by the
dendritic arm spacing. Interstitial elements such as carbon in steel become fully
homogeneous, whereas substitutional elements, which diffuse much more slowly,
may be only partly homogenized. Homogenization does not remove
macrosegregation described below, where the diffusion distances are much larger.
Macrosegregation Figure 3.9 illustrates some common types of macro-
segregation in a large ingot. The chill zone which solidifies first is usually purer.
The central part of the ingot has a concentration of solutes higher than the
average. Macroetched sections show channel segregates arranged in a V pattern
along the centre line. Vertical channels of segregates also occur in the upper
part of the ingot. A cone type of segregation is present at the base of the ingot.
In large steel ingots, it is not uncommon to find variations in carbon and other
alloying elements up to 30% of the mean value.

Figure 3.9 The macrostructure of a large ingot shows different types of segregation.
 Phase Changes 77

Macrosegregation is caused by the physical movement of the liquid and the

solid in the semi-solidified mushy region. The liquid flows as a result of
thermal contraction, solidification shrinkage and density differences. In addi-
tion, solid crystals can either settle down or float depending on the density. A
segregation known as kishing of graphite in hypereutectic cast irons is due to
the floating of the lighter graphite flakes. The cone segregation at the bottom
of the ingot in Fig. 3.9 is due to the settling down of the heavier crystallites.
When an ingot of wide freezing range (e.g. Cu-Sn alloys) is poured against
a chill mould, a solute-rich region (instead of the usual pure, solute-poor region)
may be obtained in the vicinity of the chill. This phenomenon is called inverse
segregation. The shrinkage during solidification causes the solute-rich liquid to
flow through the interdendritic channels in a direction opposite to the interface
motion. Inverse segregation does not occur in alloys that expand on solidifica-
tion. Another form of inverse segregation known as exudation refers to the
sucking out of the liquid by a pressure difference created by the separation of
the metal from the mould due to shrinkage or by the pressure build up inside
the metal due to gas evolution. The exuded solute-rich liquid solidifies on the
surface of the metal in the form of small drops known as sweat. Tin sweat occurs
as a coating of tin-rich white layer on the yellow surface of a bronze casting.
Porosity and inclusions As the fraction of the solid in the mushy region
increases, the liquid is not able to flow freely and compensate for shrinkage.
This results in microporosity. The strains generated by shrinkage can fracture
the weak solid. This phenomenon is known as hot tearing.
When a deoxidizer is added to a melt, the deoxidation product is often a
solid. When aluminium or silicon is used to deoxidize molten steel, Al2O3 or
SiO2 particles form in the melt. These are called primary inclusions, as they
form before solidification starts. Secondary inclusions form during or after
solidification, e.g. MnS in steels. Secondary inclusions are usually present in
interdendritic regions. Primary inclusions are present within the dendrites, but
sometimes found in interdendritic regions, if they have been pushed by the
thickening dendrites.
A troublesome class of impurities in cast metals is the dissolved gases. In
aluminium, there is an abrupt decrease in the solubility of hydrogen, when
solidification occurs. The gas precipitates in the form of H2 bubbles. The decrease
in solubility of oxygen in steel results in the reaction between oxygen and carbon
in the steel to produce bubbles of CO. These are examples of gas porosity.
Killed and rimmed steels In a fully-killed steel, the residual oxygen in the
melt is removed by reaction with aluminium. No gas evolution occurs during
solidification. The normal shrinkage gives rise to pipes in the ingot, Fig. 3.10a.
In a semi-killed steel, the oxygen is partly removed by the deoxidants. The
evolution of CO is just enough to compensate for shrinkage and avoid the
formation of pipe, Fig. 3.10b. The CO bubbles are closed during the subsequent
working.
In a rimming ingot, Fig. 3.10c, no deoxidants are added. Gas evolution
occurs with effervescence during solidification. Large bubbles of CO are swept
78 Physical Metallurgy

out of the ingot. The left-over CO is in the form of bubbles or wormholes

between columnar grains. The turbulence during gas evolution in a rimmed
ingot physically transports the metal to different parts, causing macrosegregation
to a greater extent. Capping is a variation of rimming. After the ingot has
rimmed for some time, a cast iron cap is placed on top of the mould to freeze
the top and stop the rimming action.

Figure 3.10 (a) Killed ingot, showing the shrinkage cavity at the top,
(b) semikilled ingot, with CO bubbles near the top half, and
(c) rimmed ingot with CO bubbles throughout the volume.

The upper part containing the exposed pipe in killed steels has to be rejected
and this decreases the yield to about 80%. The yield from a rimmed ingot is higher.
Continuous casting Continuous casting is now an established method of
producing semi-finished steel products like billets, blooms and slabs. In this
method, the liquid steel is first poured from a ladle into a refractory-lined
tundish, which directs the flow into a vertical water-cooled copper mould. The
strand of steel which comes out of the mould is semi-solid, with a liquid core
encased in a thick solidified shell. After solidification is complete, a welding
torch cuts the strand into the desired lengths. The structure of the continuously-
cast steel is similar to that of an ingot: a chill zone on the surface, followed by
columnar grains and a central equiaxed zone.
Continuous casting offers several advantages. It yields a semi-finished shape
(billet, bloom or slab), in contrast to the unfinished shape obtained in ingot
casting. It provides better productivity and yield, reduced energy and man-
power, improved product quality and consistency of quality, as compared to
ingot casting. Only a killed steel can be continuously cast. The yield here is
about 10% more, as there is no rejection in the form of the exposed pipe of an
ingot. A rimmed steel cannot be continuously cast, as the rimming action can
puncture holes in the solidified shell, with liquid pouring out.
Rapid solidification processing The cooling rate in rapid solidification pro-
cessing (RSP) is in excess of 103 C per second. In the melt spinning method,
 Phase Changes 79

a jet of liquid metal is ejected by an applied pressure through an orifice on to

a rotating wheel, Fig. 3.11. Typically, ribbons of 5 mm wide and 20100 mm
thick are produced. The thickness is proportional to the orifice diameter and
inversely proportional to the speed of the rotating wheel. The structural features
of the product include an extremely fine grain size, metastable extension of
solid solubility limits, formation of metastable or glassy products and retention
of disordered high-temperature structures. Microsegregation is reduced and the
manufacture of near-net shape products is made possible. Some examples of RSP
products are soft magnetic ribbons and metallic glasses of Fe-B based alloys.

Figure 3.11 The melt spinning method of rapid solidification processing.

Gas jets can also be directed on to the liquid metal stream to atomize the
liquid into fine droplets, which solidify by impinging on a substrate forming a
dense coating.

3.5 PEARLITIC TRANSFORMATIONS

On cooling a steel below the eutectoid temperature, austenite transforms to
pearlite, consisting of a mixture of alternate parallel plates of ferrite and cementite:
Reaction : g a + Fe3C
austenite ferrite cementite
Structure : FCC BCC orthorhombic
Composition, : 0.77 0.02 6.67
wt%C
In the T-T-T diagram for a eutectoid steel, Fig. 3.12, the upper half of the
C-curve pertains to pearlitic transformation. In alloy steels, the T-T-T diagram
may have two separate C-curves, the upper one for the pearlitic transformation
and the lower one for the bainitic transformation.
80 Physical Metallurgy

Figure 3.12 The T-T-T diagram for a eutectoid steel.

Nucleation and growth of pearlite Pearlite nucleates at the austenite grain

boundaries. The finer is the austenite grain size, the more is the grain boundary
area per unit volume and also more is the number of potential nucleation sites.
The nucleation rate of pearlite increases with decreasing austenitic grain size.
This explains the increase in the transformation rate with decreasing grain size
(see Fig. 3.14). The C-curve shifts to the left, i.e. to shorter times in fine grained
austenite.
Nucleation of pearlite occurs in two stages. First, a cementite plate nucleates
at the grain boundary and grows inwards into one of the austenite grains. As the
cementite plate thickens, the carbon content of the austenite on either side of
the cementite plate decreases. When it falls to a critical value, two ferrite plates,
one on either side of the cementite plate, nucleate. They grow by rejecting
carbon in excess of ~ 0.02% into the adjacent austenite, thereby enriching it. As
the carbon content of the austenite increases to a critical value, two cementite
 Phase Changes 81

plates nucleate on either side and grow. Thus the side-wise nucleation and
growth of pairs of lamellae continues in a pearlitic colony.
The edgewise growth of pearlite is illustrated in Fig. 3.13. In the austenite
ahead of the growing plates, the carbon content opposite to a ferrite plate is
higher than at a point opposite to the next cementite plate. Carbon diffuses
down this concentration gradient, as indicated by arrows in Fig. 3.13.

Figure 3.13 The edgewise growth of a pearlite colony takes place by carbon
diffusion in the austenite ahead, along directions indicated by arrows.

Several colonies form together in an approximately spherical volume called

a pearlitic nodule. The radius of a nodule increases linearly with time. The
volume of the nodule increases initially as the cube of time, until impingement
of neighbouring nodules occurs. The transformation fraction plotted as a func-
tion of time has a sigmoidal shape as shown in Fig. 3.14.

Figure 3.14 The plot of % pearlite against time has a sigmoidal shape.
82 Physical Metallurgy

The interlamellar spacing If the steel is cooled slowly, the effective transfor-
mation temperature is just below the eutectoid temperature. This corresponds to
a combination of a high growth rate and a relatively low nucleation rate, see
Fig. 3.4. The crystals of ferrite and cementite are relatively coarse. If the steel
is cooled more rapidly, the effective transformation temperature is lower and
corresponds to a high nucleation rate and a relatively low growth rate. Note that
carbon diffusion is required for growth and that the diffusion rate of carbon
decreases with decreasing temperature. The crystals formed are, therefore, finer.
The size of the crystals in pearlite is measured by a parameter called the inter-
lamellar spacing.
The interlamellar spacing of a pearlite colony is defined as the distance
from the centre of a ferrite (or cementite) plate to the centre of the next ferrite
(or cementite) plate. The true spacing is seen under the microscope, when
lamellae of a colony are perpendicular to the plane of observation. When they
are inclined at an angle other than 90, the measured spacing is larger than the
true spacing. Experimentally, it is observed that the true interlamellar spacing
is constant for a constant transformation temperature. It decreases with decreas-
ing reaction temperature. In Fig. 3.15, the logarithm of interlamellar spacing in
is plotted against the degree of supercooling DT below the eutectoid tempera-
ture. Transformation very near the eutectoid temperature yields coarse pearlite,
with the interlamellar spacing of about 12 mm. Coarse pearlite is soft, with a
hardness of about 5 Rc. The pearlite formed near the nose of the C-curve
(550C) is very fine, the spacing being less than 0.1 mm. Hardness of very fine
pearlite can be as high as 40 Rc.

Figure 3.15 The variation of the interlamellar spacing with the degree
of supercooling DT below the eutectoid temperature Te.
 Phase Changes 83

Cooling of austenite in a furnace yields coarse pearlite, whereas air cooling

gives fine pearlite of hardness of 2025 Rc. The extremely fine pearlite (unre-
solved under the optical microscope)) formed near the nose of the C-curve is
sometimes called troostite and the air-cooled fine pearlite as sorbite. However,
in modern usage, only the terms coarse, fine and very fine pearlite are used, as
all these are lamellar structures with different interlamellar spacings.
Effect of alloying elements Common alloying elements such as Ni, Cr, Mn,
Si and Mo (with the exception of Co) shift the nose of the C-curve to the right.
They markedly slow down the growth rate of pearlite. For example, the addition
of 0.5% Mo decreases the growth rate of pearlite by a factor of 100.
Austenite stabilizing elements such as Ni and Mn lower the eutectoid tem-
perature. In the presence of these elements, for the same cooling rate, the
effective transformation temperature is lowered and a finer pearlite results.

3.6 BAINITIC TRANSFORMATIONS

When austenite is cooled rapidly past the nose of the C-curve in Fig. 3.12 and
kept isothermally in a constant temperature bath, it transforms to an extremely
fine nonlamellar mixture of ferrite and carbide known as bainite. Several defi-
nitions and descriptions of bainite are in use. The subject is complicated by
experimental difficulties and the lack of general agreement. We will follow the
classical definition, which divides bainite into upper bainite and lower bainite.
Upper bainite forms between about 550 and 350C. Lower bainite forms below
350C. In both upper and lower bainite, the ferrite units grow by rejecting the
excess carbon into the surrounding regions, where the carbides eventually nucle-
ate. In order to proceed, the bainitic transformation requires the diffusion of
carbon.
Upper bainite has the morphology of a feathery-shaped ferrite. The feathery
appearance arises from clusters of fine parallel ferrite laths, Fig. 3.16. Cementite
platelets precipitate at the austenite grain boundaries, at ferrite/austenite interfaces
or between ferrite laths in a direction approximately parallel to the length of the
laths. The cementite plates are resolved only by an electron microscope. The
hardness of upper bainite in a 0.8%C steel is in the range of 4050 Rc.
In lower bainite, the ferrite is a plate-like unit, inside which transition
carbides such as e carbide (Fe2.4C) have precipitated at an angle of 55 to the
axis of the ferrite plate. The carbide units are even finer than those in upper
bainite. The hardness of lower bainite is in the range of 5055 Rc. Generally,
the mechanical properties of lower bainite are better than those of upper bainite.
Sometimes, two temperatures Bs and Bf are used to denote the start and
finish of the bainitic transformation. No bainite forms above Bs. The amount of
bainite formed increases with decreasing temperature below Bs. At Bf, the aus-
tenite transforms to 100% bainite, if held long enough isothermally.
84 Physical Metallurgy

Figure 3.16 The feathery morphology of upper bainite.

Ferrite nucleation is believed to be the first step in bainite formation, in

contrast to cementite nucleation in pearlite. The ferrite plate may form by a
shear mechanism, as in martensitic transformations discussed in the next sec-
tion. On the other hand, like in pearlite reactions, diffusion of carbon is also
essential for bainite to form. As the diffusion rate of carbon is slow, the distance
over which diffusion occurs is very small. This explains the submicroscopic size
of the crystals in bainite.

3.7 MARTENSITIC TRANSFORMATIONS

Below the bainitic transformation range, the diffusion rate of carbon becomes
so negligible that the austenite transforms without a change in composition to
a new product called martensite. For eutectoid steel,
quench
austenite martensite
FCC BCT
0.8%C 0.8%C
The shear mechanism The FCC crystal of austenite is converted to the BCT
structure of martensite in a diffusionless manner by means of a shear. The basic
mechanism is illustrated in Fig. 3.17. A body centred tetragonal (BCT) unit
cell is outlined within two adjacent FCC unit cells. The interstitial carbon
occupies the octahedral voids of FCC, located at the body centre and at the
twelve edge centres of the FCC unit cell. During the shear transformation, the
vertical c axis of the BCT unit cell (which is the same as the cube edge a)
contracts by about 20% and the horizontal, a1 and a2 axes (which are half of
the face diagonal of the cube) expand by about 12% each, so that the c/a ratio
decreases to the value characteristic of the martensitic crystal. The iron atoms
 Phase Changes 85

move in a coordinated fashion, by a fraction of an interatomic distance. As the

carbon atoms are all located at the middle of the c-axis (or equivalent positions),
the shear displacement of the iron atoms is obstructed along the c-axis by the
carbon in between. There are no carbon atoms along the a1 and a2 axes. The
obstruction along the c-axis results in a tetragonal product, with the c/a ratio
slightly greater than unity. The c/a ratio is a function of the carbon content and
increases from ~ 1.00 at 0.2%C to 1.04 at 0.8%C. In pure iron, where there are
no carbon atoms, rapid quenching converts the FCC crystal to BCC by a shear
mechanism. Here, the c-axis contracts by a sufficient amount to make all three
axes exactly equal, as in the BCC unit cell.

Figure 3.17 The correspondence between the BCT unit cell (martensite)
and the FCC unit cell (austenite) in the shear transformation.

Progress of the martensitic reaction Martensite starts to form at a well-

defined temperature (below the bainitic range) called Ms (martensite start).
The transformation progresses essentially by the nucleation of new plates on
cooling. The amount of martensite that forms is a function of the tempera-
ture to which the steel is cooled below Ms and not of the time of holding
at that temperature. The fraction of martensite formed increases slightly at
first just below Ms and then more rapidly with falling temperature. The
transformation tails off again and is virtually complete at Mf (martensite
finish) temperature.
Both Ms and Mf decrease with increasing carbon (Fig. 3.18) and other
alloying elements in the steel (with the exception of cobalt). The following
empirical equation gives Ms as a function of composition (wt%).
Ms (C) = 561 474C 33Mn 17Ni 17Cr 21Mo (3.9)
Mf decreases in a similar manner. With increasing carbon and alloying
elements, M f shifts below room temperature. Such steels have some
untransformed austenite, when quenched to room temperature. This is called
retained austenite. Further cooling to subzero temperatures is necessary to convert
the retained austenite to martensite. As a fully martensitic condition is desirable
after hardening, subzero treatment is a standard practice in the heat treatment
of alloy steels such as tool steels.
86 Physical Metallurgy

Figure 3.18 The effect of carbon on the Ms and Mf temperatures and

on the morphology of martensite in plain carbon steels.

Other characteristics Two morphologies of martensite are commonly

observed:
1. lath martensite, and
2. plate martensite.
At low concentrations of carbon and alloying elements and at high tempera-
tures, lath martensite forms (see Fig. 3.19). The lath unit has the shape of a
strip. A number of laths form in a parallel fashion, more or less completely
filling austenite, Fig. 3.19a. At higher concentrations of carbon and alloying
elements and at lower transformation temperatures, plate martensite forms. The
plates are arranged in a non-parallel fashion, partitioning the austenite grains
into several smaller pockets, Fig. 3.19b.
The hardness of martensite is a function of its carbon content, as shown in
Fig. 3.20. The hardness values at 0.1 wt% C intervals are also listed below.

Carbon, wt.% 0.1 0.2 0.3 0.4 0.5 0.6 0.8

Hardness, Rc 38 44 50 57 60 63 65

To realize the full hardness, the steel should have at least 0.6% C. Alloying
elements have no significant effect on the hardness of martensite. Martensite
with hardness below 55 Rc has some amount of ductility. Those with hardness
more than 60 Rc are generally brittle.
 Phase Changes 87

Figure 3.19 Microstructure of (a) lath martensite and (b) plate martensite.

Figure 3.20 The hardness of martensite is a function

of the carbon content of a steel.

Tempering of martensite Reheating of martensite to higher temperatures is

called tempering. Tempering is done to eliminate quenching stresses and to
restore some ductility in the quenched steel, at the expense of some hardness.
With the increase in temperature, carbon diffusion becomes appreciable and the
metastable martensite decomposes to more stable products. Five stages of tem-
pering are known. In the preliminary stage, recent studies indicate that ageing
effects occur around room temperature and are associated with the clustering of
carbon atoms in martensite prior to precipitation.
Stage I of tempering extends from room temperature to 200C. During this
stage, the martensite decomposes to two phases: a low-carbon martensite with
0.2%C and e (HCP, Fe2.4C), a transition carbide, which is metastable with
respect to cementite. With increase in the carbon content of the steel, more e
carbide forms. It precipitates in a very fine form and is resolved only under the
88 Physical Metallurgy

electron microscope. The hardening effect due to this precipitation is usually

offset by the softening effect associated with the loss of carbon in martensite.
At high carbon levels, e.g. in a 1.4%C steel, the hardness may initially increase
on tempering due to the dominance of the precipitation of e carbide. Other
transition precipitates such as h and c carbides have also been reported.
Stage II tempering occurs in the range of 100300C. The structural change
here is the decomposition of retained austenite to bainite as a function of time
at the tempering temperature. The hardness increase due to this reaction may
not be evident due to the simultaneous occurrence of other softening effects.
Stage III tempering consists of the decomposition of martensite to ferrite
and cementite (Fe3C). This reaction occurs in the range of 250400C. The
cementite particles that form are submicrocopic in size. If the steel had already
been tempered at a lower temperature, the reaction consists of the decomposi-
tion of the low carbon martensite into ferrite and cementite, dissolution of e
carbide and its reprecipitation as cementite.
Stage IV of tempering is the coarsening of cementite particles, which results
in further softening of the steel. The coarsened particles become visible under
the optical microscope, when tempering is done above 500C.
The hardness change as a function of tempering temperature is depicted
in Fig. 3.21. The carbon steels show a continuous decrease in hardness with
increasing temperature. The time at a given tempering temperature is a less
important parameter. When hardness changes are to be specified as a function
of both temperature and time of tempering for carbon steels, the Hollomon-
Jaffee parameter T(c + log t) is used as the x-axis in Fig. 3.21. Here, T is
absolute temperature, t is time in seconds and c is a constant.
Effect of alloying elements on tempering The influence of alloying ele-
ments on the tempering behaviour of steels is of two types. Noncarbide forming
elements may slow down the decrease in hardness during tempering and also
strengthen the ferrite by solid solution. Silicon is particularly effective in this
respect, see Fig. 3.21a. For example, at 550C, the carbon steel has a tempered
hardness of 28 Rc, whereas the silicon steel has a hardness of 45 Rc (both of
the same carbon content).
The carbide forming alloying elements give rise to an important effect
known as secondary hardening (sometimes termed Stage V of tempering). The
effect of molybdenum is shown in Fig. 3.21b. At about 550C, a peak in
hardness appears. Precipitation of very fine alloy carbide particles is responsible
for the peak. The hardness of the molybdenum steel tempered at 550C is about
55 Rc, as compared to 26 Rc of the carbon steel (both of the same carbon
content).
 Phase Changes 89

Figure 3.21 The variation of hardness with tempering temperature for different steels.

Thermoelastic martensite In alloys, where the tempering reaction does not

intervene, the martensite phase may revert back to austenite on heating. This is
illustrated in Fig. 3.22 for Fe-30% Ni, Au-47.5% Cd and Fe3Pt alloys. Corre-
sponding to Ms, there is As (the austenite start temperature) during the reversal.
The temperature interval between the forward and reverse transformations (called
thermal hysteresis) in Fe-Ni is more than 400C, whereas it is only 16C for the
Au-Cd alloy (Fig. 3.22a). In Fe3Pt (Fig. 3.22b), the hysteresis is large when the
transformation occurs in the disordered alloy and is small when the alloy is
ordered.
The alloys exhibiting a small hysteresis show thermoelastic behaviour. Here,
the martensite plates nucleate and grow continuously on cooling. If cooling is
stopped, growth ceases. If cooling is resumed, further growth of the martensitic
plates occur. On heating, the reverse transformation takes place. The plates
shrink and disappear exactly in the reverse order of their appearance. An
90 Physical Metallurgy

externally applied stress above Ms (in place of a temperature change) can cause
the same cyclic phenomenon, i.e. the plates appear and grow on increasing the
applied stress; they shrink on decreasing the stress and disappear when the
stress falls to zero. Some of the factors which favour the thermoelastic behaviour
are a small hysteresis, a small shear strain associated with the martensitic
transformation and accommodation of transformation strains elastically, without
the occurrence of the irreversible plastic deformation.

Figure 3.22 The reverse transformation of martensite to austenite starts at As.

Shape memory alloys The fact that the martensitic transformation in

thermoelastic alloys can be brought about by thermal and/or mechanical means
has led to the development of shape memory alloys. A well-known shape memory
alloy NiTi (called Nitinol) has nearly equiatomic proportions of Ni and Ti. The
stress-strain curve of nitinol at 10C is compared with the normal elastic behaviour
of stainless steel in Fig. 3.23. The stainless steel exhibits a usual linear elastic
region as per Hookes law. The elastic strain is less than half of a percent. The
 Phase Changes 91

deformation beyond the elastic region (not shown in Fig. 3.23) is permanent
plastic deformation in stainless steel. In contrast, nitinol exhibits a nonlinear

Figure 3.23 The stress-strain curve in the elastic region of a shape

memory alloy (Nitinol) is compared with that of stainless steel.

elastic behaviour with elastic strains up to 8%. On removal of the stress, the
strain is more or less completely recovered. The forward martensitic transfor-
mation takes place during loading and the reverse transformation occurs on
unloading, with a hysteresis. A simple example of a combination of thermal and
mechanical application is as follows. A straight rod of nitinol can be bent by
hand into an L-shape just above Ms. On release of the hand load, the rod comes
back to the initial straight form. If, in the bent condition, the rod is cooled below
Ms, the rod remains L-shaped even after the removal of the load. If the bent rod
is now heated above As, it springs back to the straight form remembering its
original shape.
Manipulation of the martensitic transformation within specified limits of
temperature and stress levels has resulted in the development of numerous
devices such as dental arch wires in orthodontics, heat activated fasteners or
couplings (eliminating welding) in inaccessible locations like nuclear equip-
ment, deep sea and space systems, medical instruments such as catheters and
guidewires used in surgery.
Smart materials are materials systems that combine the use of sensors,
signal processing and actuation functions into the material structural system.
The smart materials have the ability to change physically in response to tem-
perature, load, pressure or other environmental factors. In recent years, numer-
ous international symposia have been held on shape memory alloys, with more
than 10,000 patents of shape memory devices on record.
92 Physical Metallurgy

3.8 PRECIPITATION AND AGE HARDENING

In many alloy systems, the solubility in a terminal solid solution decreases with
decreasing temperature. An alloy that is undersaturated or just saturated with
solute at a high temperature is in the form of a single phase. On cooling, the
alloy may become supersaturated with respect to the solute. Then a driving
force arises for rejection of the excess solute in the form of a precipitate.
Several industrially important alloys such as Al-Cu and Cu-Be alloys undergo
precipitation. Precipitation of carbides occurs during tempering of a quenched
steel (as discussed in the last section) or during ageing of rapidly cooled low
carbon steel. When the precipitates are in a very fine form, an increase in
hardness and strength of the alloy usually results. Careful control of the precipi-
tation process is an important step in the heat treatment of such alloys.
Precipitation in Al-4.5% Cu alloy The precipitation process can be described
with reference to the Al-4.5% Cu alloy known as duralumin. Duralumin contains
small quantities of other alloying elements, but, for our present purposes, we
can assume it to be a binary alloy of Al and Cu. The relevant portion of the
Al-Cu phase diagram is shown in Fig. 3.24. The solubility of copper in aluminium
decreases from 4.5% at 500C, to less than 0.2% at room temperature, see the
q solvus in Fig. 3.24.

Figure 3.24 The aluminium-rich end of the Al-Cu phase diagram. The positions of
the solvus corresponding to the metastable transition precipitates are also shown.
 Phase Changes 93

If the Al-4.5%Cu alloy is cooled slowly from 500C, it becomes supersatu-

rated and tends to form particles of q precipitates (CuAl2):
asupersat asat + q
FCC FCC tetragonal
4.5%Cu 0.2%Cu 52%Cu
During slow cooling, the precipitates that form are coarse and do not cause
any improvement in mechanical properties. If the alloy is quenched to room
temperature, no time is available for diffusion to occur to bring about the
compositional changes. If the alloy is left at room temperature for a prolonged
period or reheated to slightly higher temperatures, precipitation occurs. Due to
the limited diffusion rates at these temperatures, the solute atoms move through
only a few tens of interatomic distances, giving rise to extremely fine precipi-
tates. This process is called ageing.
Steps in the ageing process The nucleation barrier for the formation of the
equilibrium precipitate q at room temperature is too high. So, precipitation
occurs in steps involving several transition phases. These are metastable with
respect to q, but still their formation is favoured because of the favourable
nucleation conditions.
First, the Guinier-Preston (GP) zones form. These zones are ordered, copper-
rich clusters of atoms, which are only one or two atoms thick. They are coherent
with the matrix and may cause considerable elastic strains. GP zone formation
occurs by copper diffusion, aided by quenched-in vacancies over relatively
short distances. The GP zone solvus is below and to the right of the q solvus
(see Fig. 3.24). So, GP zones do not form above 200C, as the supersaturation
with respect to the GP zone solvus disappears. If the alloy is heated above
200C after the formation of GP zones, the zones redissolve and the process is
called reversion.
Overlapping with the GP zone formation, the transition precipitates q and
q form. These are much larger than GP zones and are partially coherent with
the matrix. They have a definite composition and crystal structure, which are
slightly different from those of the equilibrium precipitate q. The transition
precipitates may nucleate at GP zones or on dislocations.
Complete loss of coherency at the interface occurs when the final equilib-
rium precipitate q forms. It forms only at higher temperatures in a coarsely
dispersed form.
Age hardening Controlled precipitation hardens the alloy. As hardness in-
creases as a function of time of ageing, this process is known as age hardening.
The heat treatment consists of several steps:
1. Solutionizing is the process of heating above the solvus to produce a
single-phase solid solution. It is essential that the alloy is not heated
94 Physical Metallurgy

above the solidus temperature. If coring persists in the cast alloy, even
heating close to the equilibrium solidus may cause burning, i.e. melting
and oxidation at the solute-rich grain boundaries, with an adverse effect
on ductility. Special problems are encountered in Al-Cu-Mg alloy where
effective solutionizing requires heating to within a few degrees of the
solidus.
2. Quenching or rapid cooling to retain the single phase solid solution at
room temperature. Cold water quenching is very effective in achieving
a maximum supersaturation. It is frequently necessary for thicker cross-
sections to obtain the required cooling rate at the centre of the section.
Slower cooling rates are achieved by quenching in hot or boiling water
or even by aircooling in some cases.
3. Controlled decomposition of the solution by ageing at one or two ageing
temperatures. Natural ageing refers to holding the alloy at room
temperature, whereas artificial ageing is holding at slightly elevated
temperatures (100 200C). Figure 3.25 shows the increase in hardness

Figure 3.25 The variation in hardness as a function of ageing time at 130C for an
A1-4%Cu alloy. The appearance of the various transition precipitates is indicated.

of an Al-4%Cu alloy at 130C. The hardness increases with the formation

of GP zones and the intermediate transition precipitates. Maximum
hardening results in an alloy when there is a critical dispersion of GP
zones or a transition precipitate (q or q ) or both.
 Phase Changes 95

After the peak hardness, further ageing tends to decrease the hardness. This
phenomenon is called overageing. The precipitate particles coarsen during
overageing. During coarsening,
1. the average particle size increases,
2. the number of particles decreases, and
3. the interparticle distance increases.
The dislocations moving in the matrix face less resistance in an over-aged
alloy, which becomes softer.
The service temperature of an age-hardened alloy has to be appreciably less
than the ageing temperature. This will ensure that there is no overageing during
service. For example, some supersonic aircrafts operate in service with the skin
temperature attaining 100110C. The ageing temperature for the aluminium
alloy used for the body of this aircraft is 190C.
Quench ageing The solubility of carbon in ferrite decreases from 0.02% at
727C to less than 0.001% at room temperature. On quenching a low carbon
steel from above 500C, a supersaturated ferrite is obtained. Quench ageing
refers to the precipitation as a function of ageing time of iron carbide (or iron
nitride) from the supersaturated ferrite. The phenomenon is the same as in age
hardenable aluminium alloys.
Between 20200C, the transition phase e carbide precipitates in a submi-
croscopic form. Above 200C, cementite precipitates. Figure 3.26 shows the
increase in hardness as a function of ageing time at different temperatures.
Overageing occurs beyond the hardness peak, as the carbide particles coarsen
and the inter-particle distance increases. If an appreciable amount of nitrogen is
present in the steel, similar precipitation of iron nitrides can cause hardening.

Figure 3.26 Ageing of a quenched low-carbon steel at different temperatures.

Quench ageing has significance in industrial processes such as galvanizing.

When the low carbon steel sheet emerges from the molten zinc bath at 500C,
it is cooled with a blast of cold air. Supersaturation occurs at room temperature
96 Physical Metallurgy

and precipitation ensues. An annealing treatment after galvanizing to coarsen

the precipitate particles may be necessary to have the maximum ductility in the
sheet.

3.9 RECOVERY, RECRYSTALLIZATION AND

GRAIN GROWTH
These processes refer to the microstructural changes that take place on heating
a cold worked metal. In none of these phenomena, a change in composition or
crystal structure is involved, unlike a typical phase change. However, the boundary
migration processes are quite similar to the growth process.
On cold working, the density of point imperfections and dislocations
increases. These structural imperfections have strain energy associated with
them. On annealing, the metal tends to lose the excess energy acquired during
cold working and revert back to the initial state. This loss of the excess strain
energy provides the driving force for these relaxation processes. On heating,
recovery, recrystallization and grain growth occur in that order at successively
higher temperatures.
Recovery Recovery is the annihilation and rearrangement of point imperfec-
tions and dislocations without the migration of high angle grain boundaries. As
the annealing temperature is increased, the excess point imperfections anneal
out in different ways. A vacancy and an interstitialcy pair can mutually annihi-
late each other. The point imperfections also find a sink at high angle grain
boundaries and at edge dislocations. Excess vacancies can disappear at edge dislo-
cations making them to climb up. On further increase of the annealing temperature,
dislocation recovery processes start. A positive and a negative edge (or screw)
dislocation on the same slip plane can mutually annihilate each other. Excess dis-
locations of the same sign left over after this, lower their energy by arranging
themselves in low angle boundaries. This process is known as polygonization.
Recrystallization Recrystallization is the nucleation and growth of new, strain-
free crystals from the cold worked metal. The driving force for recrystallization
is the stored energy in the cold worked metal. Nucleation in the usual sense may
not occur in recrystallization. An existing grain boundary with local differences
in dislocation density on either side may simply migrate into the region of
higher dislocation density. The recrystallization temperature can be defined as
that temperature at which 50% of the metal recrystallizes in 1 hour.
Some well-known laws of recrystallization are the following:
1. The higher is the amount of cold work, the lower is the recrystallization
temperature.
2. The finer is the initial grain size, the lower is the recrystallization
temperature. This is so, as the prior grain boundaries act as sites from
where the recrystallized grains start to grow.
 Phase Changes 97

3. Increasing amount of cold work and decreasing initial grain size produce
finer recrystallized grains as shown in Fig. 3.27.

Figure 3.27 Recrystallized grain size as a function of initial grain size

and amount of prior deformation in a 70/30 brass.

4. The higher is the temperature of cold working, the less is the strain
energy stored in the metal. The recrystallization temperature is
correspondingly higher.
5. The recrystallization rate increases exponentially with temperature.

Solute drag and pinning effects The recrystallization temperature is strongly

dependent on dissolved impurities (solutes) in the metal. Commercial purity
aluminium recrystallizes at 275C, whereas high purity aluminium recrystallizes
at 75C. The impurity atoms in commercial aluminium segregate at the grain
boundaries and retard the migrating boundaries during recrystallization. This is
known as the solute drag effect.
Recrystallization is also slowed down in the presence of fine second phase
particles such as alloy carbides in steels. When the particle lies on the migrating
boundary during recrystallization, the grain boundary area is reduced by an
amount equal to the cross-sectional area of the particle. When the boundary
98 Physical Metallurgy

moves further, it has to pull away from the particle and thereby create new
boundary area equal to the cross-section of the particle. This increases the
energy and manifests itself as a pinning action of the particle on the boundary.
Consequently, the rate of recrystallization decreases.
Grain growth Grain growth is the increase in the average grain size follow-
ing recrystallization. The grain size distribution does not change during normal
grain growth. During abnormal grain growth called secondary recrystallization,
the grain size distribution may radically change, i.e. some very large grains may
be present along with fine grains.
When the average grain size increases, the grain boundary area per unit
volume of the metal decreases, with a corresponding decrease in grain boundary
energy per unit volume. This provides the driving force for grain growth, which
is about an order of magnitude smaller than that for recrystallization.
The solute drag effect and the pinning action of second-phase particles
retard the movement of a migrating boundary during grain growth as well.
Thus, aluminium killed steels tend to remain fine grained during austenitization
in the presence of aluminium oxide and nitride particles.
The effect on mechanical properties of cold work, recovery, recrystalliza-
tion and grain growth processes are summarized in Fig. 3.28.

Figure 3.28 Microstructural and property changes during cold working and
subsequent annealing, when recovery, recrystallization and grain
growth occur at successively higher temperatures.
 Phase Changes 99

REFERENCES

1. Shewmon, P.G., Transformations in Metals, McGraw-Hill Book Co., New

York (1969).
2. Sinha, A.K., Physical Metallurgy Handbook, McGraw-Hill, New York
(2003).

QUESTIONS

1. In order to increase the diffusion penetration distance by a factor of two,

by how much should the diffusion time be changed and why?
2. Why is a nickel steel heated in air prone to decarburization and not to
denickelization?
3. Derive the results given in Eqs. (3.6) and (3.7).
4. Explain how the grain size is determined by the nucleation and growth
rates.
5. List the methods available to you for refining the grain size of a casting.
6. Distinguish between:
(i) microsegregation and macrosegregation, and
(ii) normal segregation and inverse segregation.
7. Compare the characteristics of pearlite, bainite and martensite transforma-
tions in a tabular form.
8. Arrange the following microconstituents in order of decreasing free en-
ergy: pearlite, martensite, tempered martensite, ferrite and graphite,
spheroidized cementite and ferrite.
9. What is the purpose of tempering a quenched steel? How can the tem-
pered properties be optimized? Discuss with examples.
10. Duralumin alloy rivets used in aircraft construction are kept at 40C
after the solution treatment. Whenever required, they are taken out and
riveted on the spot. Explain the reason for keeping them at 40C and
how the riveting operation at room temperature strengthens the rivet.
(Hint: Strain-induced precipitation is faster than normal precipitation).
11. Very fine alloy carbide particles effectively pin down the austenite grain
boundaries during recrystallization. Explain how this results in grain
refinement.
Chapter 4
HEAT TREATMENT OF STEELS

The age hardening treatments based on the precipitation reaction were discussed
in the last chapter. In this chapter, the heat treatments given to steels, which
form an important area of application of physical metallurgy, are discussed.
Some heat treatment processes such as annealing and stress relieving discussed
here are based on the same general principles applicable to both ferrous and
nonferrous alloys. Specific details of heat treatment applicable to individual
ferrous and nonferrous alloys are discussed in Chap. 7.

4.1 THE T-T-T DIAGRAMS

The Fe-C phase diagram (Fig. 2.10) depicts phases that form at extremely slow
rates of cooling. In industrial practice, the cooling rates may not be very slow.
More importantly, a cooling rate may be deliberately chosen to obtain a desired
transformation product in the steel. For understanding the effect of cooling
rates, we need to follow the transformation as a function of both temperature
and time. The T-T-T (temperature-time-transformation) diagrams, also known as
I-T (isothermal-transformation) diagrams, depict the relationships between phases,
temperature and time. The T-T-T diagram for a 0.8%C eutectoid steel was given
in Fig. 3.12. Recall the relevant descriptions of it from Chap. 3.
Determination of the T-T-T diagram The T-T-T diagram is experimentally
determined as follows. A steel specimen is first heated to the austenitic range
and then quenched into a molten salt or metal bath, maintained at a constant
temperature, e.g. 625C in the pearlite region, Fig. 4.1. The specimen is allowed
100
 Heat Treatment of Steels 101

to transform partially to pearlite by holding for a certain length of time in the

bath, after which it is quenched to room temperature. The untransformed aus-
tenite is converted to martensite on quenching. As pearlite and martensite can
be readily distinguished under the microscope, the fraction of pearlite formed
corresponding to that holding time in the bath is estimated. By reacting a series
of about ten specimens in a similar fashion, it is possible to bracket the times
for 1% and 99% pearlite formation. Similar tests are conducted at other tem-
peratures in the pearlitic and bainitic regions to determine the full C-curves for
the start and finish of these two transformations.

Figure 4.1 Experimental determination of the T-T-T diagram. The numbers 1, 2, 3 and
4 indicate increasing times of holding in the pearlitic region. 1', 2' and 3' indicate
quenching to successively lower temperatures within the martensitic region.

Following this, the austenitized specimens are quenched to different

temperatures in the martensitic range. They are then heated to temper the
martensite formed and cooled back to room temperature, see Fig. 4.1. In the last
102 Physical Metallurgy

step, the remaining austenite is converted to martensite. Tempered martensite is

dark and can be distinguished under the microscope from untempered martensite,
which etches lightly. So, the fraction of martensite corresponding to the first
quenching temperature can be determined. In this manner, the amount of
martensite formed at different temperatures as well as the Ms and Mf temperatures
are obtained.
T-T-T diagrams for hypo- and hypereutectoid steels In a hypoeutectoid
steel, proeutectoid ferrite forms between A3 and A1. An additional curve that
starts from A3 depicts the beginning of the proeutectoid ferrite formation on the
T-T-T diagram in Fig. 4.2a, which should be compared with Fig. 4.1. The
equilibrium amount of ferrite forms only at temperatures near A3. At lower
temperatures, the amount of ferrite is less. Near the nose of the C-curve, the
proeutectoid ferrite curve merges with the pearlite curve. Under such condi-
tions, the transformation is mostly to pearlite but the carbon content of pearlite
is less than 0.77%.

Figure 4.2 The T-T-T diagrams for (a) 0.35%C hypoeutectoid steel, and
(b) 1.1%C hypereutectoid steel.

In hypoeutectoid steels, the nose of the C-curve (shown dotted) is much

closer to the temperature axis than in the eutectoid steel. Compare the nose
position in Figs. 4.2a and 4.1. The nose time is 1 sec for the eutectoid steel and
about 0.2 sec for the 0.35%C steel. The hardenability of a hypoeutectoid steel
is correspondingly less and very drastic quenching is required to produce all
 Heat Treatment of Steels 103

martensite. The reason for the shift of the C-curve to the left is that the
proeutectoid ferrite provides sites for pearlite nucleation and enhances its rate
of formation.
A hypereutectoid steel has a similar T-T-T diagram, Fig. 4.2b. The upper
limiting temperature here is the Acm temperature and the proeutectoid phase is
cementite. The shift of the nose to the left occurs here as well.
Effect of alloying elements All common alloying elements (except cobalt)
shift the nose of the C-curve to the right, thus making it easier to quench the
steel past the nose without any transformation on the way. They are said to
increase the hardenability of the steel. Also, in a number of cases, in the
presence of an alloying element, separation of the pearlitic and the bainitic
noses occurs. Figure 4.3 is the T-T-T diagram for a Ni-Cr-Mo low alloy steel
with 0.4%C, 1.8%Ni, 0.8%Cr and 0.3%Mo. The pearlite nose is at 650C and
the nose time is ~100 s. The bainite nose is at 450C and at ~10 s. The shorter
time at the bainitic nose determines the hardenability of the steel. This time is
still an order of magnitude larger than the nose time for a eutectoid steel.

Figure 4.3 The T-T-T diagram for a low-alloy Ni-Cr-Mo steel, depicting two
separate C-curves for the pearlitic and the bainitic transformations.
104 Physical Metallurgy

4.2 THE C-C-T DIAGRAMS

The T-T-T diagrams discussed above depict the isothermal transformation of
austenite. In industrial practice, a steel is rarely quenched to a constant temperature
and isothermally transformed. Rather the steel is continuously cooled from the
austenitizing temperature to room temperature at different cooling rates.
Continuous-cooling-transformation (C-C-T) diagrams depict the transformation,
temperature and time relationships during continuous cooling.
Determination of the C-C-T Diagram Figure 4.4 shows the C-C-T relation-
ships for a eutectoid steel. It is determined as follows. Specimens are cooled
from the austenitic range at a constant cooling rate. The pearlite start (point a)
and the pearlite finish (point b) are determined, by the metallographic method

Figure 4.4 The C-C-T diagram for a eutectoid steel. Different cooling rates
are indicated. The corresponding T-T-T diagram is shown dotted.

outlined for the determination of the T-T-T diagram. Points a and b are at a
lower temperature and a longer time than the corresponding points x and y on
 Heat Treatment of Steels 105

the T-T-T diagram (shown dotted). This is so, because the specimen has spent
much of the cooling time at higher temperatures, where the times for the pearl-
ite start are longer. Experiments with different cooling rates yield the locus of
points a and b. The C-C-T diagram is thus constructed.
The C-C-T diagram for a eutectoid steel A very slow cooling rate of
0.01C s1 crosses both the pearlite start and finish curves of the C-C-T diagram
at fairly high temperatures, Fig. 4.4. The resulting microstructure is coarse
pearlite. A cooling rate of 3C s1 gives fine pearlite. A higher cooling rate
of 35C s1 crosses the pearlite finish near the nose, yielding very fine
pearlite. A cooling rate of 140C s1 just misses the pearlite start curve, does
not enter the bainitic region and transforms to all martensite. This critical
cooling rate must be exceeded to obtain full hardening of the steel. For
cooling rates between 35 and 140C s1, the steel crosses the pearlite start
curve, but not the pearlite finish curve. Such cooling rates convert a part of
the austenite to pearlite. Even though the untransformed austenite passes
through the bainitic range of the T-T-T diagram, it turns out that the bainite
start curve for continuous cooling shifts so far to the right that very little
austenite transforms to bainite. It passes untransformed into the martensitic
range. Thus in a eutectoid steel, continuous cooling does not yield bainite.
The transformation products as a function of cooling rate are summarized
below.

Cooling rate, C s1 Transformation product

< 35 all pearlite

35140 pearlite and martensite
>140 all martensite

The C-C-T diagram for the Ni-Cr-Mo Steel The C-C-T diagram for the
Ni-Cr-Mo steel given in Fig. 4.5 should be compared with the corre-
sponding T-T-T diagram in Fig. 4.3. For cooling rates less than 0.006C
s 1, the structure is proeutectoid ferrite and pearlite. Between 0.006 and
0.02C s 1 , partial transformation to pearlite occurs. The untransformed
austenite transforms to bainite and martensite at lower temperatures.
Between 0.02 and 0.3C s 1 , no pearlite forms. The final structure is
ferrite, bainite and martensite. Between 0.3 and 8C s 1, the products are
bainite and martensite. Note that bainite forms on continuous cooling in
this steel, whereas it does not in the eutectoid steel. The critical cooling
rate to produce all martensite is 8C s 1, which is 1/20th of that for the
eutectoid steel.
106 Physical Metallurgy

Figure 4.5 The C-C-T diagram for a low-alloy Ni-Cr-Mo steel.

Different cooling rates are shown.

4.3 FUNCTION OF ALLOYING ELEMENTS

An alloy steel contains, in addition to Fe and C, one or more substitutional
alloying elements such as Mn, Ni, Cr, Mo, W and V. When the total alloy
content is small (510%), the steel is known as a low alloy steel. When the alloy
content is more than 10%, it is a high alloy steel.
The alloying elements may be present in the steel in the following forms:
1. Substitutional solutes in ferrite or austenite.
2. Nonmetallic inclusions such as sulphides, oxides, and silicates, e.g.
MnS, Al2O3 and MnSiO3.
3. Dissolved in cementite as part of the orthorhombic structure, e.g.
(Fe,Mn)3C.
4. Fine alloy carbides and nitrides, e.g. NbC, VN, WC and Cr23C6.
5. Insoluble metals, e.g. Cu and Pb.
As substitutional solutes in ferrite or austenite, the alloying elements
strengthen the steel. However, they are rarely used for improving the strength
this way, as it is too expensive a method.
 Heat Treatment of Steels 107

Chromium and nickel in solution in ferrite or austenite improve the corrosion

resistance of iron remarkably. Stainless steels, which contain chromium with or
without nickel, are discussed in Chaps. 6 and 7.
An important function of alloying elements in steels is to increase the
hardenability, by shifting the nose of the C-curve to the right on the T-T-T
diagram. To be effective, the alloying element must be in solution in austenite.
Second phase particles have both desirable and undesirable effects in steels.
They are the initiators of voids during deformation, which coalesce ultimately
to cause fracture. Silicates and sulphides get elongated during hot rolling, which
result in poor toughness, when tested transverse to the direction of rolling. MnS
removes hot shortness (brittleness at elevated temperatures). Sulphides, phos-
phides and insoluble metals improve the machinability of steels.
Fine spherical particles of alloy carbides pin down the migrating boundaries
during recrystallization and grain growth. Thus they help in maintaining a fine
grain size. In the presence of alloy carbides, the abrasion and wear resistance
of steels is improved. Some alloy carbides such as MoC and VC enhance the
high temperature (creep) strength, by providing a fine stable dispersion of par-
ticles in the matrix.
The order of increasing stability of carbides in a steel is as follows:
Ni Si Al Fe Mn Cr W Mo V Ti Nb

noncarbide formers increasing carbide stability

When nitrogen is present, many of the carbide formers listed above form
carbonitrides or nitrides, which are also very stable and abrasion resistant.

4.4 FORMATION OF AUSTENITE

In many heat treatment processes, the first step consists of heating the steel into
the austenitic range, the process known as austenitization. Austenite starts to
form when the steel is heated above Ac1. The kinetics of its formation depends
to some extent on the initial microstructure. With tempered martensite as the
initial structure, austenitization occurs more rapidly than with coarse pearlite.
The rate of austenitization can be shown on a T-T-T diagram, Fig. 4.6. The
carbon content is 0.8% and the initial microstructure is fine pearlite. The rate
of austenite formation increases rapidly with temperature. For example, full
austenitization occurs in less than 5 s at 800C. However, some residual car-
bides are left over, as cementite dissolves more slowly than ferrite. The third
curve on the right in Fig. 4.6 indicates the completion of cementite dissolution.
Even after this, the austenite is not homogeneous, as it has less carbon in
regions which were initially ferrite and more carbon in regions of initial cement-
ite. The fourth and the last curve on the right shows the homogenization of
austenite. It is seen that the time for complete homogenization is much longer
than for the formation of 99% of austenite (second curve in Fig. 4.6)
108 Physical Metallurgy

Figure 4.6 The kinetics of austenite formation on heating above A1. Even though
99% austenite forms rapidly, the time for complete homogenization is very long.

In alloy steels, alloy carbides dissolve much more slowly than cementite.
The difference is marked, when the carbide is very stable. For example, vanadium
carbide starts to go into solution at around 1100C, whereas niobium carbide
does so only at around 1150C. These temperatures are much higher than those
required for cementite solution. Even after the alloy carbides have dissolved,
homogenization of austenite is a very slow process, as substitutional diffusion
of alloying elements is involved.
Austenite undergoes grain growth, when held at high temperatures and for
long austenitizing times. A coarse austenite grain size is not desirable, as it
results in poor impact properties of the heat treated steel. Figure 4.7 illustrates
the grain growth characteristics of two low carbon steels. In the aluminium-
killed steel, a fine dispersion of Al2O3 and AlN particles is present. Here, grain
growth is much slower at the normal austenitizing temperatures as compared to
the unkilled steel. This difference arises from the pinning action of Al2O3 and
A1N particles on the migrating boundaries during grain growth of austenite. At
temperatures above 1100C, the pinning action is not effective, as the particles
dissolve. The grain size of the austenite increases abruptly. Note that the differ-
ence between killed and unkilled steels is significant only around 950C, the
temperature at which carburization is carried out. At the hardening temperatures
(800900C) and at the forging temperature (above 1000C), the grain size is
approximately the same in both the steels.
 Heat Treatment of Steels 109

Figure 4.7 The grain growth characteristics of aluminium-killed and unkilled steels.

Effect of undissolved carbides The undissolved alloy carbides produce the

following effects:
1. They help to keep the austenite grain size fine. The fine grain size
results in better mechanical properties in the heat treated steel. With fine
grain size, however, the C-curve shifts to the left, decreasing the ability
of the steel to harden.
2. When undissolved carbides are present, both the alloying element and
carbon are out of solution and this decreases the hardenability of the
austenite.
3. The surface of the carbide particles can provide sites for ferrite or
pearlite nucleation, resulting in a further decrease of hardenability.

4.5 TYPICAL HEAT TREATMENT PROCESSES

Typical heat treatments given to steels are described in this section. Figure 4.8
shows the heating temperatures for various treatments on the Fe-C phase
diagram.
Annealing The general purpose of annealing is to reduce the hardness, facilitate
machining and to relieve internal stresses. Different types of annealing treatments
are described below.
Full annealing consists of heating the steel above the Ac3 temperature in the
case of hypoeutectoid steels and the Ac1 temperature in the case of hypereutectoid
steels, Fig. 4.8. The steel is then cooled very slowly (in the furnace) at the rate
of a few tens of degrees per hour. The austenite transforms to coarse pearlite
110 Physical Metallurgy

within 50C of the eutectoid temperature. As prolonged heat treatment is involved,

full annealing tends to be an expensive process. The time of annealing can be
cut down by transferring the austenitized component to a constant temperature
furnace kept just below the eutectoid temperature. After the transformation is
completed isothermally, the steel is cooled in air. The microstructure is still
coarse pearlite.

Figure 4.8 The heating temperatures for various heat treatment processes.

Spheroidizing anneal is heating to just below Acl or thermal cycling in a

narrow temperature internal around A1. The cementite lamellae in pearlite colo-
nies tend to spheroidize (become spheres) and thereby reduce their surface area.
With prolonged anneal, the spheroidized particles coarsen, becoming fewer in
number and more widely spaced. This corresponds to the softest state of the
steel. For carbon and low-alloy structural steels, the optimum machinability
corresponds to 50% spheroidized50% lamellar carbide in the microstructure.
For high alloy tool steels, the fully spheroidized condition is preferred.
Recrystallization anneal is done at subcritical temperatures in the range of
600 650C, Fig. 4.8. The cold worked ferrite recrystallizes during this anneal.
Stress-relief anneal is a subcritical treatment given to hypoeutectoid steels up to
0.3%C. The temperature used is up to 600C. Recovery processes occur without
recrystallization. The residual stresses due to cold working or machining are
 Heat Treatment of Steels 111

removed by this treatment. Process anneal is a subcritical treatment given to

metals to soften them during mechanical processing. It may or may not involve
full recrystallization of the cold worked metal.
Normalizing Normalizing is done for refining the grain structure and for
improving the mechanical properties. The hardness and strength obtained after
normalizing are higher than those obtained after annealing.
Normalizing consists of heating the steel above the Ac3 temperature for
hypoeutectoid compositions and between Ac1 and Acm temperatures for
hypereutectoid compositions, see Fig. 4.8. This is followed by cooling in still
air. Fine pearlite and proeutectoid ferrite result from this treatment. In
hypereutectoid compositions, heating between Ac1 and Acm helps to break down
any network structure of proeutectoid cementite around the austenite grain
boundaries. The subsequent aircooling is fast enough to prevent the reformation
of the continuous network.
Normalizing is applied to castings and forgings. In castings, normalizing is
essential to eliminate the dendritic structure. The microstructure of hot rolled
products cooled in air approximates to the normalized condition.
Patenting This is a special heat treatment process adopted for wire products
of medium carbon content. The wire from the austenitizing temperature is passed
through a molten lead bath maintained at 500C. Transformation to very fine
pearlite occurs at the nose of the C-curve. During subsequent wire drawing, this
microstructure enables a greater reduction in area to be achieved than the annealed
structure.
Quenching and tempering Quenching of a steel is done from the austenitic
range to produce martensite. The cooling rate at the centre of the cross-section
should exceed the critical cooling rate for the steel to get full hardening.
Hypoeutectoid steels are heated above the Ac3 temperature to ensure full
austenitization and to avoid the soft ferrite in the final microstructure. For
hypereutectoid steels, this is not necessary, as the cementite that remains
undissolved on heating to just above Acl is itself very hard. Moreover, by heating
above Acm, the steel becomes susceptible to grain growth. Recall that the Acm
line is quite steep. In addition, the steel will crack due to the steeper temperature
gradient when cooled from above Acm and also due to the larger carbon content
in solution in austenite.
Tempering is almost always necessary to remove the residual stresses and
to reduce the brittleness of martensite. The actual tempering temperature chosen
depends on the final properties desired. For example, a carbon tool for cutting
is tempered only slightly to relieve the quenching stresses, so that a large part
of the hardness achieved by quenching is retained. A shock-resistant tool should
possess adequate toughness. So, it is tempered at a higher temperature with
some loss of hardness. For structural applications, adequate ductility is required,
necessitating a still higher temperature of tempering.
112 Physical Metallurgy

Temper brittleness During tempering, certain temperature ranges may have

to be avoided. Temper brittleness is an embrittling phenomenon that usually
occurs during tempering in the range of 350550C. Brittleness also occurs during
slow cooling through this critical range, after tempering at a higher temperature.
Temper brittleness is a reversible process. It disappears on heating to higher
temperatures, but returns on holding in the critical range. It does not occur in
plain carbon steels. Also when P, Sn, Sb, As and N are absent, temper brittleness
does not occur. It is believed to be due to the interaction of common alloying
elements such as Cr, Mn and Ni with the above-listed residual elements, many
of which enter the steel with the scrap during steel making. The fracture path
of the embrittled steel is intergranular (along the grain boundaries). The
phenomenon appears to be associated with the segregation of impurities at the
prior austenite grain boundaries without actual precipitation. Molybdenum added
to steel removes temper embrittlement.
Protective atmospheres When heated in an open furnace, steel is prone to
both oxidation and decarburization. Oxidation is caused by oxygen, carbon
dioxide and moisture in the air through the following reactions.
O2 + Fe 2FeO
CO2 + Fe FeO + CO
H2O + Fe FeO + H2
In decarburization, the carbon near the surface layers of the steel combines
with oxygen, carbon dioxide or moisture in the air.
Inert atmospheres prevent oxidation and decarburization. Nitrogen can be
used as an inert atmosphere, provided it is purified to remove residual oxygen
and moisture. Use of gases such as argon or helium or having a vacuum are
expensive methods for protection. They are used for heat treating titanium
alloys and some special stainless steels, where the cost is justified. By burning
producer gas partially, some industrial atmospheres are obtained. With a controlled
composition, such gaseous mixtures can prevent oxidation or decarburization.
Baths used in heat treatment Molten salt or metal baths offer protection
against oxidation and decarburization. In addition, they provide for quick and
uniform heating or cooling during heat treatment. Chloride baths consisting of
mixtures of NaCl and KCl provide a working range of 700 900C and can be
used for austenitizing. Nitrate baths have a working range of 250 600C and
are used for subcritical heat treatments.

4.6 HARDENABILITY
The cooling rates during quenching vary across the cross-section of a steel. The
surface cools most rapidly on quenching and hardens by forming martensite. At
the centre of the cross-section, the cooling rate is slower, which can result in
the formation of pearlite at the centre. If drastic cooling is done, the cooling rate
 Heat Treatment of Steels 113

at the centre will also exceed the critical cooling rate and the steel will have
hardened throughout the cross-section. However, drastic quenching causes other
undesirable effects such as warping and cracking of the steel. The hardenability
of a steel is defined as its ability to harden (by forming martensite) throughout
its cross-section, without having to resort to drastic quenching. Alloying ele-
ments play a crucial role in increasing the hardenability. If this were not so, we
will be building machine components as layer structures, where each layer is
thin enough to harden fully!
Grossmans critical diameter In Grossmans experiments, a number of steel
rods of different diameters, Fig. 4.9, are austenitized and quenched in the same
bath. To avoid end effects, the length to the rod must be at least five times the
diameter. The smallest diameter rod hardens right through. If C 0.6%, a
hardness of ~65 Rc is obtained across the entire cross-section. As the diameter
of the rod increases, the cooling rate at the centre decreases and becomes less than
the critical cooling rate. Then, the hardness falls short of the full value of 65 Rc.

Figure 4.9 Rods of different diameters 1, 2, 3, 4 harden to different depths on

quenching in the same medium. The critical diameter is that of rod 3.

The largest diameter rod 4 in Fig. 4.9 has only fine pearlite at the centre.
The hardness changes most rapidly at a value of 54 Rc, corresponding to a
microstructure of 50% martensite50% pearlite. This inflection point is taken as
the transition zone between the hardened and not-hardened parts of the cross-
section. For a certain diameter (3) less than the largest, the microstructure at the
centre is 50% martensite50% pearlite with a hardness of 54 Rc. This diameter
is called the critical diameter. Diameters more than the critical will not harden
throughout, whereas smaller diameters will. The critical diameter is thus a
measure of the hardenability of a steel.
Severity of quenching media If the critical diameter is to be defined as a
material property that depends only on its composition and the austenitizing
treatment, the effect of the quenching medium used should be eliminated. For
this purpose, Grossman defined an ideal quenching medium and an ideal critical
diameter that corresponds to quenching in that medium. The severity of quench-
ing of a medium is defined by the parameter H.
heat transfer coefficient between steel and medium
H=
thermal conductivity of the steel
114 Physical Metallurgy

The ideal medium is one which brings the surface of the steel instantly to
the bath temperature and maintains it at that temperature. The parameter H is
infinite for such a medium. Note that the ideal medium brings only the surface
of the steel instantly to the bath temperature. The heat flow from the centre to
the surface of the steel still occurs at a finite rate, as determined by the thermal
conductivity of the steel.
In real quenching media, the ideal situation described above does not exist.
Usually a vapour blanket forms between the steel and the liquid. The heat
transfer through this layer is slow. The viscosity of the liquid also influences the
quenching efficiency. For example, oil which is more viscous than water is a
less efficient quenchant. In a brine solution (e.g. 812% NaCl dissolved in
water), the cooling rate is faster than in water. The salt particles cause little
explosions to occur near the vapour blanket, thereby puncturing it. In air, of
course, the cooling rate is slower than in any liquid medium.
Agitation of the quenching bath or shaking of the steel object helps to break
the vapour blanket and increase the quenching efficiency. The H values for
typical media are listed below.

State of the medium Air Oil Water Brine

Still 0.02 0.2 1.0 2.0
Agitated 0.04 0.4 0.8 2.0 4.0 3.0 5.0

Grossman constructed a master graph for correlating the critical diameter

Dc determined in a quenching medium of known H to the ideal critical diameter
DI, Fig. 4.10.

Figure 4.10 The relationship between the ideal critical diameter DI and
the critical diameter Dc for different H values.
 Heat Treatment of Steels 115

For different values of H, we can write from the figure.

H = DI = Dc
H < DI > Dc
H = 0.2 DI ~ 2Dc
(still oil) (at Dc ~ 100 mm)
Jominy end quench test Grossmans method of determining the hardenability
is cumbersome and time-consuming, as a large number of different diameters of
steel bars are to be prepared and tested. Jominy and his coworkers devised a
single-specimen test known as the Jominy end quench test for determining the
hardenability.
In this standardized test, a steel rod, 25.4 mm dia and 102 mm long, is
austenitized and then transferred quickly to a fixture, Fig. 4.11a. A water jet is
opened at the same time and the end of the rod is quenched. The variables
associated with the water jet are standardized. The diameter of the water pipes
is 12.7 mm. The distance of the orifice from the bottom end of the steel rod is
12.7 mm. The free jet height is 64 mm. The temperature of the water should
be between 21 and 27C.
Only the lower end of the sample is quenched directly by the water jet. The
cooling rate decreases progressively, with increasing distance from that end.
Near the top end, the cooling rate is approximately that of air cooling. The
water jet is kept on for about 20 minutes to bring the sample to room temperature.
Then, two flat surfaces are ground opposite to each other along the length of
the rod. The hardness is measured at intervals of 1.6 mm from the quenched
end, Fig. 4.11b. Near the end, the interval may be reduced to 0.8 mm, as the

Figure 4.11 The Jominy end quench test: (a) the experimental set up, and
(b) the hardness variation as a function of distance from the quenched end.
116 Physical Metallurgy

hardness impressions are smaller and the hardness may vary rapidly. The hardness
changes most rapidly, at the location where the microstructure is 50% martensite.
This critical distance measured from the end in units of 1.6 mm (1/16 in the
original specification) is called the Jominy distance. Figure 4.12 is a master plot
relating the distance from the quenched end to the ideal critical diameter. For
example, when the distance is 6.35 mm (4 Jominy units), the ideal critical
diameter is 45 mm.

Figure 4.12 The relationship between the Jominy distance and the
critical diameters for various quenching media and H values.

Computation of the ideal critical diameter The ideal critical diameter DI of

a steel can be computed from its composition. Figure 4.13 gives the base
hardenability in terms of DI for a carbon steel as a function of carbon content
and grain size. The hardenability as measured by DI decreases with decreasing
carbon content and decreasing grain size. Recall that the nose of the C-curve
shifts to the left and the hardenability decreases, as the carbon content decreases
 Heat Treatment of Steels 117

from the eutectoid value or as the grain size becomes finer. Note that the DI
values decrease when the carbon content is increased beyond 0.8%.

Figure 4.13 The ideal critical diameter as a function of carbon content and grain size.

The effect of alloying elements on the hardenability is shown in Fig. 4.14.

Manganese, molybdenum and chromium are among the most effective. The
ideal critical diameter obtained from Fig. 4.13 is multiplied by the factor given
in Fig. 4.14 for each of the alloying elements present. For example, for a steel
of grain size ASTM 7, with 0.4%C, 0.9%Mn and 2%Si,
DI = 15.5 (from Fig. 4.13) 2.05 (Mn factor from Fig. 4.14)
1.9 (Si factor from Fig. 4.14)
= 60 mm
Noting that Dc for still oil is about 0.5DI, 1.2 dia rods of this spring steel
is just about the size that will harden in oil through the entire cross-section. It
must be pointed out here that somewhat conflicting data exist in the literature
on base hardenability and multiplying factors. It should also be noted that the
hardenability is not based solely on composition and grain size, as it also
depends on the austenitizing treatment, which determines the extent of carbide
dissolution and homogeneity of the austenite.
118 Physical Metallurgy

Figure 4.14 The hardenability multiplying factors for different alloying elements.

Figure 4.15 Hardenability multiplying factor for boron.

As phosphorous and sulphur (present in small concentrations in steels) have

opposite effects on hardenability, their combined effect can be ignored. Boron
 Heat Treatment of Steels 119

in extremely small concentrations (0.0010.003%) has a remarkabe effect of

increasing the hardenability of steels. Boron, when in solution in austenite, is
believed to segregate to the grain boundaries and strongly retard the nucleation
of proeutectoid ferrite. The boron multiplying factor at the above concentration
is about 3 for a 0.2%C steel (see Fig. 4.15), but decreases to 1.0 (no effect) at
0.8%C.
As steel compositions are specified over a range, the hardenability can also
vary over a range. The Jominy test results for a steel are usually presented in
the form of hardenability bands, corresponding to the range of composition of
the steel.

4.7 QUENCH CRACKS

When thick cross-sections are quenched drastically, cracks form on the surface
of the steel. Distortions of shapes of thin sections due to plastic deformation
known as warping may also occur. Warping and cracking are certainly not
desirable. The presence of surface cracks reduces drastically the fatigue life of
machine components.
Origin of quench cracks A steep temperature gradient is produced from the
surface to the centre of the steel on quenching. Superposed on this are two
effects that occur on cooling.
1. thermal contraction of the steel, and
2. volume expansion due to the transformation.
The coefficient of linear thermal expansion (or contraction) for steels is
about 11 106 K1. For a change in temperature from 820 to 25C, the total
contraction is about 0.9%. The % volume expansion during the austenite-to-
martensite transformation is + 4.64 0.53 (%C). The linear expansion is about
1/3 of the volume expansion. For a 0.6%C steel, the linear expansion due to
transformation is about + 1.4%.
When the steel is quenched from the austenitizing temperature, the surface
of the steel comes more or less immediately to the bath temperature, whereas
the centre is still near the austenitizing temperature. The transformation to
martensite occurs at the surface layers and is accompanied by a volume in-
crease. This produces stresses at the centre. The yield strength of austenite is
low at high temperature. So, the stresses are accommodated by plastic flow of
the austenite. Eventually, the centre also cools down to room temperature,
undergoing the martensitic transformation. The centre now expands and pro-
duces tensile stresses at the surface. These stresses cannot be accommodated by
plastic flow of the surface layers, which already contain martensite. Martensite
is too brittle to flow plastically. The residual stress pattern obtained is shown
in Fig. 4.16a. The outer layers are in tension and the centre is in compression.
Figure 4.16b shows that the residual stress distribution is opposite, when only
the surface layers are hardened.
120 Physical Metallurgy

The stress distribution due to thermal contraction is opposite in sense to that

generated by the volume expansion due to the transformation. In most cases, the
transformation stresses are dominant and the net residual stresses at the surface
are tensile in nature. Brittle martensite, which cannot support tensile stresses,
cracks. This is the origin of surface cracks in drastically quenched steels.
It should be noted that the residual stresses arise from the steep temperature
gradient and the lack of simultaneous transformation throughout the cross-
section. If the cooling rates are slower, the temperature gradient is less and the
transformation occurs more or less simultaneously, with little residual stresses.
Thus, increasing the hardenability of the steel by adding alloying elements,
which permit a slower cooling rate, becomes important in reducing the stresses
during quenching.

a b
Figure 4.16 Residual stress pattern on quenching: (a) after through hardening,
and (b) after shallow hardening or surface hardening.

To remove the residual stresses, a stress relieving operation is usually done

by reheating the steel to a relatively low temperature. About 85% of the stresses
are estimated to be relieved on heating to 250C. Tempering the steel at higher
temperatures, of course, relieves all the stresses.
Martempering and austempering These are special heat treatment processes
that are adopted for reducing residual stresses and minimizing the tendency for
cracking. Figure 4.17 compares these processes with the conventional quenching
and tempering process. In the conventional process, Fig. 4.17a, the surface and
the centre cool at different rates and transform to martensite at different times.
The residual stresses are set up due to this as discussed above. Figure 4.17b
depicts martempering (also called marquenching). Here, the steel is quenched
into a bath kept just above Ms. After allowing sufficient time for the temperature
to become uniform throughout the cross-section, it is aircooled through the
martensitic range. The transformation to martensite occurs more or less
 Heat Treatment of Steels 121

simultaneously across the section. Residual stresses induced are a minimum.

The steel is tempered in the usual way to the desired hardness. As the cooling
rate through the martensitic range is less here, more austenite tends to be
retained due to thermal stabilization.
Figure 4.17c illustrates the austempering process. Here, the steel is quenched
into the bainitic bay above Ms and kept isothermally till all the austenite is
converted to lower bainite.
The pearlitic nose must be further to the right for austempering and
martempering than for conventional quenching, to prevent any transforma-
tion near the nose. Also, for martempering, the time at the bainitic bay
should be long enough for temperature equalization before any transforma-
tion occurs.

Figure 4.17 Heat treatment processes to reduce quenching stresses: (a) conventional
quenching and tempering, (b) martempering, and (c) austempering.

The impact energy and elongation of a 0.95%C steel heat treated to the
same final hardness by these three methods are compared as follows:
122 Physical Metallurgy

Treatment Hardness, Izod impact, Elongation,

Rc J %
Water quenched and tempered 53 16 0
Martempered 53 35 0
Austempered 52 56 9

Both martempered and austempered steels have better toughness. The

austempered steel has the best ductility. However, austempering tends to be
slow and expensive, as prolonged holding above Ms is required.

4.8 CASE HARDENING

In case hardening, the surface of the steel is made hard and wear resistant, but
the core remains soft and tough. Such a combination of properties is desired in
applications such as gears. A hard and wear resistant surface can be produced
in two ways. In steels with more than 0.35%C, the surface can be preferentially
hardened by a heat treatment. If the carbon content is lower (0.15 0.2%), the
chemistry of the surface needs to be changed by adding more carbon or nitrogen
to give a hard martensite after the heat treatment. Recall that the hardness of
martensite is a function of its carbon content. In the first category of treatments,
we have induction hardening, flame hardening and laser hardening. In the sec-
ond category, we have carburizing, cyaniding, carbonitriding and nitriding.
Induction hardening Here, an alternating current of high frequency passes
through an induction coil enclosing the steel part to be heat treated. The induced
emf heats the steel. The depth up to which the heat penetrates and raises the
temperature above Ac3 is inversely proportional to the square root of the ac
frequency. Correspondingly, the hardened depth decreases with increasing
frequency. In induction hardening, the heating time is usually a few seconds.
Immediately after heating, water jets are activated to quench the surface.
Martensite is produced at the surface, making it hard and wear resistant. The
microstructure of the core remains unaltered. Induction hardening is suitable for
mass production of articles of uniform cross-section.
Flame hardening For large work pieces, such as mill rolls, large gears and
complicated cross-sections, induction heating is not easy to apply. In such cases,
flame hardening is done by means of an oxyacetylene torch. Heating should be
done rapidly by the torch and the surface quenched, before appreciable heat
transfer to the core occurs.
Laser hardening A laser beam can be used for surface hardening. As laser
beams are of high intensity, a lens is used to reduce the intensity by producing
a defocused spot of size ranging from 0.5 to 25 mm. Proper control of energy
input is necessary to avoid melting. Laser hardening has the advantage of precise
control over the area to be hardened, an ability to harden reentrant surfaces,
 Heat Treatment of Steels 123

very high speed of hardening and no separate quenching step (the quench is
effected by the mass of the unheated material). The disadvantage is that the
hardening is shallower than in induction and flame hardening.
Carburizing Carburizing is the most widely used method of surface hardening.
Here, the surface layers of a low carbon steel (C 0.25%) is enriched with
carbon up to 0.81.0%. The source of carbon may be a solid medium, a liquid
or a gas. In all cases, the carbon enters the steel at the surface and diffuses into
the steel as a function of time at an elevated temperature. Carburizing is done
at 920 950C. This fully austenitic state is essential. If carburizing is done in
the ferritic region, the carbon, with very limited solubility in ferrite, tends to
form massive cementite particles near the surface, making the subsequent heat
treatment difficult. For this reason, carburizing is always done in the austenitic
state, even though longer times are required due to the diffusion rate of carbon
in austenite being less than that in ferrite at such temperatures.
In pack carburizing, the articles to be carburized are packed in a box,
embedding them in a powdery mixture of 85% charcoal and 15% of energizers
such as BaCO3. The box is sealed with fireclay and loaded into the furnace kept
at 930C. The residual air in the box combines with carbon to produce CO. The
energizer decomposes as below:
BaCO3 BaO + CO2
CO2 + C 2CO
The carbon enters the steel through the following reaction.
2CO + Fe Fe (C) + CO2
The depth of penetration increases as the square root of time, recall the
diffusion equation (3.3). Typical carburizing times are 6 8 hours. The case
depth obtained is 12 mm.
If selective carburization is to be done, copper is electroplated to a thickness
of ~ 0.05 mm in regions where carburization is not desired. Alternatively, a
refractory paste of fireclay mixed with asbestos can be applied. Control of
temperature and penetration depth is less in pack carburizing as compared to
liquid and gas carburizing. Also, direct quench from the carburizing temperature
to harden the surface is not possible.
In gas carburizing, a mixture consisting of 515% methane (or propane) in
a neutral carrier gas is used. The methane decomposes according to the follow-
ing reaction:
CH4 + Fe 2H2 + Fe (C)
The carbon potential of the gas mixture increases with increasing
concentration of methane. Too large a concentration or too high a gas velocity
releases carbon faster than it can be absorbed and may result in soot formation
on the surface. Closer control of temperature and case depth is possible in gas
carburizing, as compared to pack carburizing. Also, post quenching can be done
directly.
124 Physical Metallurgy

Liquid carburizing is an outgrowth of the cyaniding process discussed below.

A typical bath composition is: 8% NaCN, 82% BaCl2 and 10% NaCl. The
following reactions take place:
BaCl2 + 2NaCN Ba(CN)2 + 2NaCl
Ba(CN)2 + Fe Fe(C) + BaCN2
Rapid heat transfer and accurate temperature control are the advantages of
the liquid bath.
Post-carburizing heat treatments After carburizing at 930C, the grain
structure of the steel is usually coarse throughout the cross-section. Two separate
heat treatments may be necessary to obtain the optimum properties of the core
and the case. The steel is first heated above Ac3 and cooled in air to refine the
grain size in the core. The purpose of the second heat treatment is to refine and
harden the case. Here, the heating is done above the Ac1 temperature. Fine
grained austenite forms and any proeutectoid cementite in the case remains
undissolved. On quenching, the austenite transforms to martensite at the case.
The core, rendered fine-grained by the first heat treatment, has a low hardenability
and does not harden during the second heat treatment. A direct quenching from
the carburizing temperature is adopted only when the steel is inherently fine-
grained (e.g. aluminium-killed). A coarse grained steel, if quenched directly
from the carburizing temperature, tends to harden in the core as well. These
heat treatments are schematically shown in Fig. 4.18.

Figure 4.18 Post-carburizing heat treatments: (1) direct quenching,

(2) core refining, (3) case refining, and (4) core and case refining.

The volume expansion during the martensitic transformation on the surface

tends to induce compressive stresses, which are beneficial for improving the
fatigue strength.
Cyaniding Cyaniding is done in a liquid bath of NaCN, with the concentra-
tion varying between 30 and 97%. Both carbon and nitrogen enter the steel via
the following reactions:
2NaCN + O2 2NaCNO
3NaCNO NaCN + Na2CO3 + C + 2N
 Heat Treatment of Steels 125

The temperature used for cyaniding is lower than that for carburizing and
is in the range of 800 870C. The time of cyaniding is 0.53 hour to produce
a case depth of 0.25 mm or less.
Carbonitriding This process is also known as dry cyaniding or gas cyaniding.
It is the gas carburizing process modified by the addition of anhydrous ammo-
nia. The decomposition of ammonia provides the nitrogen, which enters the
steel along with carbon. A typical gas mixture consists of 15% NH3, 5% CH4
and 80% of neutral carrier gas. The temperature used is 750 900C. With
increasing temperature, a greater proportion of carbon enters the steel.
Nitriding In contrast to the processes described before, nitriding is carried
out in the ferrite region. Consequently, no phase change occurs after nitriding.
The part to be nitrided should possess the required core properties prior to
nitriding. If necessary, a prior heat treatment may be given to develop these.
Pure ammonia decomposes to yield nitrogen which enters the steel:
2NH3 2N + 3H2
The solubility of nitrogen in ferrite is small. Most of the nitrogen that enters
the steel forms hard nitrides (e.g. Fe3N). A typical nitriding steel contains alloying
elements of 1%Al, 1.5%Cr and 0.2%Mo. Aluminium, chromium and
molybdenum form very hard and wear resistant nitrides.
The temperature of nitriding is 500590C. The time for a case depth of
0.02 mm is about 2 hour. An extremely hard case is produced with hardness in
the range of 1000 2000 VPN. Compare this with the hardness of martensite
which is 830 VPN (= 65 Rc). In addition to providing outstanding wear resis-
tance, the nitride layer increases the resistance of a carbon steel to corrosion in
moist atmospheres.
A common problem encountered in nitriding is the formation of g nitride
(Fe4N) on the outer layers of the case, known as the white layer, as it looks
white under the microscope. This layer is very brittle and tends to crack. It must
be removed by a final grinding operation. Its formation can be minimized by
maintaining the correct ratio of NH3/H2 in the gas mixture during the heat
treatment.

REFERENCES

1. Totten, G.E. and M.A.H. Howes, Eds., Steel Heat Treatment Handbook,
Marcel Dekker Inc., New York (1997).
2. Davis, J.R., Ed., Surface Hardening of Steels, ASM International, Materials
Park, Ohio (2002).
3. Rajan, T.V., C.P. Sharma, and A. Sharma, Heat Treatment: Principles and
Techniques, Prentice-Hall of India, New Delhi (1988).
126 Physical Metallurgy

QUESTIONS

1. An austenitized steel, after holding for some time at a constant temperature

below the eutectoid, is quenched. The microstructure was found to be
ferrite and martensite. What is the possible composition of the steel and
from which phase field of the T-T-T diagram was it quenched? Explain
with a sketch.
2. A continuously-cooled carbon steel shows pearlite and martensite in its
microstructure. Indicate the composition of the steel and the possible cooling
route on a C-C-T diagram.
3. List the three most important functions of alloying elements in steel,
explaining each one briefly.
4. A high austenitizing temperature increases hardenability. Can this method
be used to improve the hardenability in place of alloying elements? Explain.
5. Explain why grain refinement of a hypoeutectoid steel requires heating
above Ac3.
6. Explain the difference between ideal critical diameter and critical diameter.
7. Calculate the ideal critical diameter of the following steel: Grain size
ASTM No. 8, 0.4%C, 0.85%Mn, 0.25%Si, 0.5%Ni, 0.5%Cr, 0.2%Mo,
0.04%P (max), 0.04%S (max).
Answer: 71 mm
8. Distinguish between austempering and martempering.
9. Explain why a 0.2% carbon steel carburizes much more easily at 875C
than armco iron.
10. In what way is the nitriding process different from the other case harden-
ing processes? Explain.
11. Explain how quenching a carburized steel introduces residual compressive
stresses at the surface.
12. Are post-carburizing heat treatments necessary? Explain.
Chapter 5
MECHANICAL PROPERTIES

Typical mechanical properties of metals and alloys are strength, ductility, creep
strength, fatigue and fracture resistance. In this chapter, the details of a tensile
test are first described. The role of dislocations in plastic deformation and the
methods of strengthening metals against plastic yield are discussed. The mecha-
nisms of creep and the design of creep-resistant materials are then described. At
the end, the phenomena of fracture and fatigue are discussed from a fracture-
toughness viewpoint.

5.1 THE TENSILE TEST

The engineering stress-strain curve Specimens used in a tensile test are
prepared according to standard specifications. The test pieces can be cylindrical
or flat. Figure 5.1a shows the standard dimension of a typical cylindrical
specimen. It is gripped at the two ends and pulled apart in a machine by the
application of a load. The stress-strain curve obtained from the tensile test of
a typical ductile metal is shown in Fig. 5.1b. On the y-axis, the engineering
stress, s, defined as the load P divided by the original cross-sectional area Ao
of the test piece, is plotted. The engineering strain e, defined as the change in
length DL divided by the initial gauge length Lo, is plotted on the x-axis. The
% elongation is obtained by multiplying the engineering strain by 100.
The stress-strain curve starts with elastic deformation. The stress is
proportional to strain in this region, as given by Hookes law. At the end of the
elastic region, plastic deformation starts. The engineering stress corresponding
to this transition is known as the yield strength (YS), an important design
127
128 Physical Metallurgy

parameter. Many metals exhibit a continuous transition from the elastic region
to the plastic region. In such cases, the precise determination of the yield
strength is difficult. A parameter called proof strength (or offset yield strength)
that corresponds to a specified permanent set is used. After loading up to the
proof stress level and unloading, the specimen shows a permanent elongation
of 0.1 or 0.2%.
The stress-strain curve has a positive slope in the plastic region, indicating
that the stress required to cause further deformation increases with increasing
strain, a phenomenon known as work hardening or strain hardening. If the load
is removed when the specimen is in the plastic region, it retraces a straight line
path parallel to the initial line and reaches zero stress at a finite value of
permanent elongation, see Fig. 5.1b. Thus, the elastic part of the deformation
is recovered. On reloading, plastic deformation starts only on reaching the stress
level prior to unloading.

Figure 5.1 The tensile test: (a) a standard cylindrical test specimen
(dimensions in mm), and (b) the engineering stress-strain curve.
 Mechanical Properties 129

The engineering stress reaches a maximum and then decreases. The maximum
value is known as the ultimate tensile strength (UTS) or simply the tensile
strength. Up to the UTS, the strain is uniformly distributed along the gauge
length. Beyond UTS, somewhere near the middle of the specimen, a localized
decrease in cross-section known as necking develops. Once the neck forms,
further deformation is concentrated in the neck. The strain is no longer uniform
along the gauge length. The cross-sectional area of the neck continuously
decreases, as the % elongation increases. Voids nucleate in the necked region
at the interface of hard second-phase particles in the material. These voids
grow and coalesce, as the strain increases. The true cross-section bearing the
load becomes very small, as compared to the apparent cross-section, due to the
growth of these internal voids. At this stage, the specimen may fracture.
Figure 5.2 shows that ductility measured in terms of the true strain at fracture
(see below for definition of true strain) decreases with increasing concentration

Figure 5.2 Effect of second-phase particles on the true strain at fracture.

Elongated and plate-like particles reduce ductility more than spherical particles.

of second-phase particles, which are responsible for the nucleation of voids in

the neck. It is seen that spherical-shaped second phase particles are less harmful
than plate-like carbides or elongated sulphides. If there are no second phase
particles to nucleate the voids, the specimen may neck down to a point before
fracture occurs. The % elongation corresponding to the fracture point is called
the total elongation. The % elongation corresponding to the maximum load is
called the uniform elongation, as it corresponds to the point up to which the
cross-sectional area decreases uniformly along the gauge length. The % reduction
130 Physical Metallurgy

in area (% RA) is defined as the decrease in cross-sectional area of the neck

at fracture, divided by the initial area Ao (expressed as a percentage). Both %
elongation and % RA are used as a measure of ductility.
The maximum point on the engineering stress-strain curve is called the
point of plastic instability. At this point, the fractional increase in stress ds/s
due to work hardening is exactly balanced by the fractional decrease in cross-
section dA/A due to elongation of the specimen:
dT dA
= (5.1)
T A
The true stress-true strain curve The fall in the engineering stress beyond
UTS is not due to a softening of the material. If we define the true stress s as
the load P divided by the instantaneous (minimum) cross-sectional area A and
the true strain et as the integral of the incremental change in length dl divided
by the current value of the gauge length l, we can replot the tensile test data as
a stet curve. As the true stress is always larger than the engineering stress and
the true strain is smaller than the engineering strain, the stet curve lies above
and to the left of the engineering stress-strain curve. Figure 5.3 shows the true
stress-true strain curves for a number of metals and alloys. The curves have a
positive slope right up to the point of fracture, showing that work hardening
occurs up to fracture, even though at a decreasing rate. The decreasing rate of
work hardening is called dynamic recovery and is attributed to the cross-slip of
screw dislocations, which initially get piled up against obstacles.

Figure 5.3 The true stress-true strain curves for a number of metals and alloys.
 Mechanical Properties 131

The rate of work hardening The true stress-true strain curve in the plastic
region is often described by the Hollomon equation:
st = Ketn (5.2)
where K is the strength coefficient and n is the work hardening exponent. For
some metals, n is constant throughout the plastic region, whereas in others n
changes (generally decreases) with increasing strain. Typically, n is in the range
of 0.10.3. A large value of n means a high work hardening rate and a high
tensile strength relative to the yield strength. Combining Eqs. (5.1) and (5.2),
it can be shown that the uniform true strain etu corresponding to the point of
plastic instability is equal to n. So, as n increases, the uniform strain increases,
Fig. 5.4, and the material exhibits more ductility.

Figure 5.4 The true uniform strain is equal to the work-hardening exponent n.

In austenitic stainless steels, the work hardening rate is high. This is attrib-
uted to the low stacking fault energy (~ 0.005 J m2) and the corresponding
difficulty in the cross-slip of screw dislocations. This delays the onset of dy-
namic recovery and increases the extent of uniform strain along the gauge
length. The tensile strength of 18/8 stainless steel is more than twice its yield
strength. The % elongation is also high (~ 60%). Aluminium has a high stacking
fault energy (0.2 J m2), as compared to copper (0.04 J m2) and hence work
hardens less, as seen in Fig. 5.3. Zinc, when added to copper, decreases the
stacking fault energy. Correspondingly, alpha brass is stronger and more ductile
than copper.
132 Physical Metallurgy

The flow stress Flow stress is a term used to define the stress level at which
plastic deformation starts in a specimen that might have undergone a specified
amount of prior plastic strain. It is easily seen that the flow stress is equal to
the yield stress, if the prior plastic strain is zero; otherwise, it is greater than the
yield stress, due to prior work hardening. The flow stress decreases with
increasing temperature T, but increases with increasing strain rate F. The
combined effect of temperature and strain rate on flow stress is denoted by the
Zener-Hollomon parameter Z, which is defined as
Z = F exp ( DH/RT) (5.3)
where DH is the activation energy for plastic deformation. It is clear that, for
the same Z, more than one combination of temperature and strain rate is pos-
sible. For example, in copper the flow stress at e = 0.3 is the same (about 70
MPa) when
T = 600C and F = 2 103
T = 800C and F = 2 101
The stress-strain curve for mild steel Figure 5.5a shows the tensile stress-
strain curve for an annealed 0.2%C steel. The end of the elastic region is
marked by the upper yield point (UYP). Plastic deformation begins at this point.

Figure 5.5 The engineering stress-strain curve for mild steel. The inset
shows the passage of a Luders band during yield point elongation.
 Mechanical Properties 133

Immediately thereafter, the stress falls to a lower value known as the lower yield
point. A horizontal part of the stress-strain curve called the yield point elongation
ensues, where deformation bands known as Luders bands pass from one end to
the other end of the test specimen, as illustrated in Fig. 5.5b. At the end of the
horizontal region, work hardening begins.
Interstitial elements, carbon and nitrogen, are too large for the voids they
occupy in ferrite. So, they tend to segregate to dislocations and lower their
distortional energy in the process. They are said to form Cottrell atmospheres
around the dislocations. Most of the dislocations are locked up on account of
this and cannot move when a stress is applied. Very few dislocations are free
to move at the onset of plastic deformation near the upper yield point. As soon
as they start to move, they rapidly multiply and lower the stress required for
continued plastic deformation to the level of the lower yield point.
Strain ageing If the deformation of the steel is interrupted after the work
hardening region is entered and the specimen is allowed to rest at room
temperature or for a shorter time at 100 150C, on reloading a new yield point
is observed, see Fig. 5.5a. This return of the yield point is referred to as strain
ageing. Interstitial atoms diffuse to form atmospheres and lock up the newly
generated dislocations during ageing. This behaviour should be contrasted with
quench ageing, discussed in Sec. 3.8, where actual precipitation (not just impurity
atmosphere formation) takes place. Another point of distinction is that plastic
deformation is not required to produce quench ageing, whereas it is necessary
for strain ageing.
Strain ageing can cause embrittlement. For example, cold sheared mild steel
plates undergo strain ageing near the sheared edge during the subsequent gal-
vanizing treatment and may become brittle.
On increasing the temperature of test on mild steel, the single yield point
is replaced by fine serrations all along the stress-strain curve, Fig. 5.6. These
arise, because the temperature is high enough for the interstitial atoms to diffuse
during deformation and form atmospheres around the newly generated disloca-
tions. This phenomenon is called dynamic strain ageing, as it occurs during the
test itself, as opposed to static strain ageing, which occurs on holding after
deformation.
Note from Fig. 5.6 that the elongations have decreased at the higher
temperatures. During dynamic strain ageing, locking of newly generated
dislocations occurs more readily with increasing temperature of deformation.
So, for the same deformation, the dislocation density at 200C is found to be
an order of magnitude more than that at 20C. This high density of locked
dislocations causes the decrease in ductility. This phenomenon is often referred
to as blue brittleness, blue being the interference colour of the steel, when it
gets oxidized in the temperature range of 200 300C.
134 Physical Metallurgy

Figure 5.6 The stress-strain curves for mild steel at different temperatures,
showing serrations due to dynamic strain ageing.

5.2 PLASTIC DEFORMATION MECHANISMS

Slip and twinning Plastic deformation permanently changes the shape and
dimensions of a metal. X-ray analysis of a deformed metal shows that, during
plastic deformation, there is no change in the crystal structure or the lattice
parameters. Only the number of imperfections increases. The macroscopic change
of shape and dimensions can be brought about without any change in the unit
cell dimensions by two processes:
(i) slip, and
(ii) twinning.
Figure 5.7 illustrates the processes of slip and twinning. In slip, atoms move
over a number of interatomic distances relative to their initial positions. The
magnitude of each step in the displacement is one full interatomic distance. The
orientation of the displaced region is the same as the original undeformed
region.
 Mechanical Properties 135

Figure 5.7 The processes of slip and twinning: (a) before deformation,
(b) after slip, and (c) after twinning.

In twinning, each atom moves by only a fraction of an interatomic distance

relative to its neighbours. The orientation of the twinned region is different from
the untwinned region. Generally, slip is the operative mechanism of plastic
deformation at higher temperatures and lower strain rates, whereas twinning is
operative at lower temperatures and higher strain rates.
It is experimentally observed that slip and twinning occur on well-defined
crystallographic planes along specific crystal directions. Typical slip planes, slip
directions, twin planes and twin directions for common metallic crystals are
listed below.

Crystal Slip Slip Twin Twin Examples

plane direction plane direction

HCP (0001) [1120] (10 12) [10 12] Cd, Mg, Zn

FCC (111) [110] (111) [112] Ag, Cu, Au
BCC (110) [111] (112) [111] Fe, Mo, Ta

Slip produces steps on a prepolished surface which appear as parallel

lines under the microscope. Figure 5.8 shows the slip lines in copper de-
formed at room temperature. In contrast to the straight slip lines seen in
Fig. 5.8, BCC iron typically produces wavy slip lines, as slip occurs on (112)
and (123) planes in addition to (110) planes, with a common slip direction of
[111].
136 Physical Metallurgy

Figure 5.8 Slip lines on a prepolished surface of copper after plastic deformation.

Role of dislocations in slip The stress required to cause slip in a perfect,

ideal crystal can be estimated. It is in the range of /6 to /30, where is
the shear modulus of the crystal. The experimentally determined shear stress
at which plastic deformation starts in a real (imperfect) crystal is much
lower, in the range of /5000 to /100, 000. This large difference in the
observed and calculated yield stresses arises from the presence of dislocations
in real crystals.
Slip does not take place in the way visualized in Fig. 5.7. The dislocations
already present in the crystal move in response to the externally applied
stress. The dislocation configuration is moved through the crystal in a
sequential fashion by small adjustments in the bond length, as illustrated in
Fig. 5.9. The figure also shows the analogy of the motion of a caterpillar.
The stress required to move a dislocation is much smaller than to shear a
perfect crystal. Whiskers are very thin hair-like crystals, which are free of
dislocations. These possess very high strengths, approaching the calculated
value for perfect crystals.
Slip usually starts from sources within the crystal. A typical source is
the Frank-Read source, which under favourable conditions of operation can
produce a very large number of dislocation loops. Many of these loops are
left back in the metal after deformation, as they often get piled up against
obstacles in the crystal such as sessile (immobile) dislocations and grain
boundaries.
 Mechanical Properties 137

Figure 5.9 Successive positions of an edge dislocation under a shear stress t.

The analogy of a caterpillar movement is also shown.

5.3 METHODS OF STRENGTHENING METALS

AGAINST YIELD
As noted above, metals are inherently weak because of the presence of
dislocations. Metals of high strength can be obtained by eliminating dislocations,
i.e. by making whiskers. However, this approach cannot be adopted in practice,
as metal whiskers are too thin to serve as structural and machine components.
Moreover, if a dislocation is introduced accidentally due to some stress
concentration, the strength of the whisker falls to a much lower value. So, we
have to live with dislocations and to adopt methods by which the stress
required to move a dislocation is increased. Based on this approach, there are
four main strengthening methods, which are described below.
(i) Strain hardening: When metals are cold worked, their strength and
hardness increase. Annealed crystals have a typical dislocation density of
138 Physical Metallurgy

108 m2. This can be increased to 10101012 m2 by moderate cold working of

the annealed metal and to 10141016 m2 by heavy cold working. At first sight,
it appears that such striking increases in dislocation density will make the
crystal more prone to slip, as many more dislocations are available, but this is
not so. As the dislocation density increases, the movement of any one disloca-
tion becomes more difficult due to the interfering effect of the stress fields of
the surrounding dislocations. This phenomenon is the basis of work hardening.
It can be described in terms of the following equation:
U =Uo + A S (5.4)
where t is the stress to move a dislocation in a matrix of dislocation density r,
to is the stress to move the dislocation in the same matrix with zero dislocation
density and A is a constant. A ~ b/2, where is the shear modulus of the
crystal and b is the Burgers vector of the moving dislocation.
Torsteel used as reinforcing bars for concrete is cold twisted to increase the
yield strength. Under suitable conditions, very heavy cold working can increase
the strength of iron to nearly the strength of a perfect crystal. The main disad-
vantage of cold working is the accompanying decrease in ductility, as illustrated
in Fig. 5.10. Cold working is also detrimental, as it raises the ductile-brittle
transition temperature of steels.

Figure 5.10 The tensile stress-strain curve as a function of prior cold work.

(ii) Refinement of the grain size: As the orientation changes at a grain

boundary, the slip plane in a crystal does not continue in the same direction
beyond the boundary. So, dislocations gliding on a slip plane are unable to cross
the boundary but get piled up against it. The number of dislocations in a pile
up is more for a larger grain, as the average distance from a dislocation source
to the boundary is larger. The magnitude of the stress concentration produced
at the leading edge of the pile up (i.e. next to the boundary) is proportional to
the square root of the number of dislocations in the pile up. Thus, a greater
 Mechanical Properties 139

stress concentration is produced at the boundary of a larger grain. This is able

to activate more readily dislocation sources in the neighbouring grain. The
propagation of the deformation from grain to grain (in the form of Luders bands
in steels) occurs at a lower stress, when the grain size is larger.
This effect increases the macroscopically observed yield strength of fine
grained metals and is described by the Hall-Petch equation:
sy = si + kd1/2 (5.5)
where sy is the yield strength of the polycrystalline metal,
si is the yield strength at an infinite grain size,
k is the Hall-Petch constant, and
d is the mean grain diameter.
When d is very large, sy ~ si. The additional contribution to the yield
strength from the kd term increases, as the grain size is reduced, Fig. 5.11.
It is seen from the figure that si term (intercept on the y-axis) is temperature
dependent, whereas k (the slope of the lines) is independent of temperature.

Figure 5.11 The Hall-Petch plot for mild steel at three different temperatures.
The yield strength increases with decreasing grain size.
140 Physical Metallurgy

The strengthening effect for a given grain size depends on the magnitude
of the constant k. In general, k tends to be larger for BCC metals, as compared
to FCC or HCP metals. For example, k = 0.71 MN m3/2 for BCC iron, whereas
it is 0.11 and 0.07 MN m3/2 for copper and aluminium (both FCC). A similar
strengthening effect arises from the subgrain boundaries in deformed metals.
Dislocation motion can be impeded by subgrain boundaries as well.
In contrast to the cold working method, grain refinement does not decrease
the ductility. Further, the ductile-brittle transition temperature is lowered in
steels by grain refinement, making the steel tough over a wider range of
temperatures. The fatigue resistance also improves by grain refinement. Thus,
grain refinement is perhaps the most desirable of the four methods of
strengthening metallic materials.
The full potential of this method is being realized only in recent years. The
advent of microalloyed steels, which contain small quantities of strong carbide
forming elements such as Nb, V or Ti, is an example. These elements go into
solution in austenite during reheating of the billet in the soaking pit. When the
steel is hot rolled, solubility of the microalloying elements decreases with de-
creasing temperature. Strain-induced precipitation of alloy carbides takes place
on a very fine scale. These precipitates effectively pin down the migrating grain
boundaries during repeated recrystallization of the deformed austenite between
passes at the successive stands of the rolling mill. The decreased growth rate
due to pinning yields a fine grained austenite. A fine grained austenite provides
more potential sites at the grain boundaries for the nucleation of ferrite, the end
result being a very fine-grained ferrite in the mild steel. Thereby the yield
strength increases considerably. The finest ferrite grain size is obtained by
controlled rolling, wherein the nucleation of ferrite takes place from a heavily-
deformed, unrecrystallized austenite. This process is sometimes called the high
temperature thermomechanical treatment (HTMT). The excellent weldability of
mild steel remains unaffected in microalloyed steels. Note that no heat treatment
step (that adds to the cost) is involved in increasing the strength.
The other methods of grain refinement are the increase in the cooling rate
and the use of nucleating agents during solidification. The cooling rates in rapid
solidification processing are much higher than those in conventional methods,
the result being a substantial grain refinement.
(iii) Solid solution strengthening: Solute atoms in general are either smaller
or larger in size than the solvent (parent) atoms of a solid solution. Smaller
atoms produce a local tensile stress field in the crystal, whereas larger atoms
produce a local compressive stress field. In both cases, the stress field of a
moving dislocation interacts with the stress field of the solute atom. This increases
the stress required to move the dislocation. In general, therefore, solid solutions
are stronger than pure metals. The yield strength of single crystals of Au-Ag
alloys (showing complete solid solubility) is shown in Fig. 5.12a. The yield
strength increases with increasing solute at the two pure ends, reaching a
maximum at the mid-composition.
 Mechanical Properties 141

Figure 5.12 (a) The shear yield stress of single crystals of Au-Ag solid solutions.
(b) The effect of different solutes on the yield stress of iron.

The strengthening effect due to a solute atom depends on several factors:

(a) A large concentration of the solute means more frequent obstacles to
dislocations moving in the matrix. The strength increases in proportion
of c where c is the concentration. For dilute solutions, a linear
dependence on concentration is equally valid.
(b) As the size difference between the solute and the solvent increases, the
intensity of the stress field around solute atoms and the resistance to
dislocation motion increase.
142 Physical Metallurgy

(c) The nature of the distortion produced by the solute atoms is also
important. Spherical distortion produced by substitutional solutes is much
less effective than the non-spherical distortion produced by interstitial
atoms. Figure 5.12b shows the linear increase in yield strength of ferrite
with increasing concentration of different solutes. Interstitial elements
(C and N) have the maximum effect per unit concentration, due to the
tetragonal distortion they produce. Substitutional solutes have a much
smaller effect.
In solid solutions with long-range order, the dislocations corresponding to
an integral lattice translation of the superlattice possess large Burgers vectors.
As the energy of a dislocation is proportional to the square of its Burgers vector,
such superdislocations tend to split into two ordinary dislocations separated by
an anti-phase domain boundary between them. This boundary is a surface that
separates two domains of the ordered structure that do not match at the boundary.
The situation is very similar to the formation of a stacking fault by the splitting
of an ordinary dislocation into two Shockley partials, recall Sec. 1.4. In order
not to increase the anti-phase domain boundary area, dislocations will have to
move in pairs in ordered structures. This increases the strength of ordered alloys
as compared to disordered alloys. In some cases, the ordered structure may even
be brittle.
The strengthening effect due to solutes can be increased by forcing in-
creased solubility by nonequilibrium methods. A supersaturated solution can be
produced by quenching from a high temperature, where the solubility is more.
Austenite holds 0.8% carbon in solution just above the eutectoid temperature.
On quenching, it produces a supersaturated martensitic structure. The intense
stress field around carbon atoms in martensite effectively hinders dislocation
motion, so much so that martensite is brittle and requires tempering to restore
some ductility.
(iv) Precipitation Hardening: Closely spaced fine precipitates obstruct the
motion of dislocations moving in the matrix. A moving dislocation has two
options: it can
(a) cut through the precipitates, or
(b) bend around and bypass the precipitates.
Cutting through occurs, when the precipitate particles are coherent with the
matrix and less than about 50 in size. The increased resistance to dislocation
motion in cutting through the particles is due to the creation of a step at the
interface, Fig. 5.13a, and other surface imperfections like stacking faults within
the particles.
When the particles are larger in size (100 500 ), the dislocations bend
around and bypass them, Fig. 5.13b. The stress to do so varies inversely as the
interparticle spacing. Since finer particles are more closely spaced, a fine
 Mechanical Properties 143

precipitate distribution effectively increases the stress to bend and bypass. This
microstructural state corresponds approximately to the optimum aged condition
of a precipitation hardening alloy. Particles of larger sizes (>1000 ) are bypassed
with ease and little hardening occurs. This state corresponds to the overaged or
overtempered condition.

Figure 5.13 Interaction of a moving dislocation with precipitate particles:

(a) cutting-through, and (b) bending-and-bypassing mechanisms.

Large nondeforming particles are present in many two-phase alloys, e.g.

pearlite in steels. As the dislocations moving in the matrix easily bypass them,
the yield strength tends to be low in such two-phase materials. However, the
work hardening rate can be high, as each dislocation that bypasses leaves behind
a small loop around the particles, see Fig. 5.13b. The stresses from the dislocation
loops tend to elastically strain the non-deforming particles to conform to the
plastically deforming matrix. This results in a high tensile strength as compared
to the yield strength.

5.4 CREEP RESISTANCE

At temperature above 0.4Tm (where Tm is the melting point of the metal in
kelvin), plastic deformation of the metal occurs as a function of time at constant
load or stress. This phenomenon is known as creep. Lead creeps at ambient
temperature, but iron and copper do not. The fact that creep is a function of
temperature and time indicates that it is a thermally activated process.
The creep curves Figure 5.14 shows typical creep curves, where the creep
strain is plotted as a function of time at constant load and constant temperature.
The curve shows three stages. In stage I, known as primary creep, the creep
rate decreases with time. The work hardening due to plastic deformation is more
144 Physical Metallurgy

than the recovery (softening effect) during creep. In stage II (secondary creep),
the rate of work hardening and creep softening exactly balance each other. The
slope of strain-time curve is constant and a minimum in stage II. Stage III
(tertiary creep) is characterized by an accelerating rate of creep, where softening
effects dominate. In this stage, necking occurs and voids form at grain junctions.
These voids coalesce to cause intergranular fracture at the end of this stage.
Microstructural softening effects such as recrystallization or coarsening of
precipitate particles also contribute to tertiary creep. The three stages shift to
shorter times, if either the stress or temperature is increased, see Fig. 5.14. At
low temperatures, the creep may be transient without the three stages, the creep
strain following a logarithmic dependence on time.

Figure 5.14 Creep curves for a low alloy steel at different

temperatures at a constant tensile stress of 70 MPa.

Creep data are also presented in the form of stress-rupture curves for different
temperatures of test, Fig. 5.15. Here, the time for creep rupture is plotted against
the applied stress at various temperatures. The data can also be presented in the
form of creep rate (% strain per 1000 hr) as a function of stress. Such data are
useful in the selection of materials for high temperature use. In designs, where
creep deformation can be tolerated but rupture must be prevented, the stress-
rupture data are directly useful. The creep rate-stress data are useful, when the
structure cannot undergo more than a specified amount of creep during the
entire service life, e.g. jet engine components and steam turbine blades.
 Mechanical Properties 145

Figure 5.15 Stress-rupture and stress-creep rate curves for different

temperatures for an austenitic stainless steel.

Mechanisms of creep Dislocations piled up against obstacles can bypass

them during creep, with the aid of thermal energy and continue to glide and
cause further deformation. Screw dislocations overcome obstacles by cross-slip,
using thermal energy. As the screw dislocation is frequently dissociated into
partial dislocations, recombination over a certain length is required before cross-
slip can start and this requires thermal energy. Recombination is easier, if the
distance of separation between the partials is less.
An edge dislocation climbs out of the slip plane in which it lies against the
obstacle and continues gliding on another parallel slip plane. Climb requires the
diffusion of atoms to or away from the dislocation. The diffusion rate increases
exponentially with temperature. So, the higher is the temperature, the faster will
be the diffusion and the climb. The activation energy for creep by the climb
mechanism is the same as the activation energy for self-diffusion.
A jog on a screw dislocation has edge character. Non-conservative climb
motion of the jog may be required during the glide of such a screw dislocation.
If this climb motion is rate controlling during the cross-slip of the screw
dislocation, the activation energy for creep will again be the same as the activation
energy for self-diffusion.
146 Physical Metallurgy

Figure 5.16 compares these two activation energies for a number of metals
and shows them to be equal.

Figure 5.16 For metals, the activation energy (Q) for creep is equal to the
activation energy for self-diffusion, both of these increasing with
increasing melting point of the metal.

Diffusional creep occurs, when a metal crystal under a tensile stress elon-
gates by the stress-assisted motion of vacancies. Vacancies move from faces
perpendicular to the tensile axis to faces parallel to them. The flow of atoms is
in the opposite direction. The activation energy for diffusional creep is the same
as the activation energy for self-diffusion. In a polycrystalline metal, the va-
cancy flow occurs in each grain. Creep that occurs by diffusion of vacancies
through the grain is called Nabarro-Herring creep. When the diffusion is along
grain boundaries, the process is called Coble creep.
At a temperature called the equicohesive temperature (~ 0.50.6Tm), the
grain boundaries reverse their role from being a source of strengthening to that
of weakening the metal. Below this temperature, they act as effective obstacles
to dislocation motion. Above this temperature, the boundaries undergo sliding,
i.e. they behave as if they were a very viscous liquid. Creep deformation is
caused by this grain boundary sliding.
Creep-resistant materials Creep-resistant materials are required for structural
and machine components used at elevated temperatures. They should be capable
of withstanding these temperatures, without undergoing creep, which may cause
dimensional changes not permissible in the design.
 Mechanical Properties 147

The first requirement for a material to be creep-resistant is that it should

have a high melting point. Recall that creep becomes significant only at
temperatures above 0.4Tm. Consistent with the factors such as cost, availability
and fabricability, metals such as iron, nickel and cobalt with moderately high
melting points (around 1500C) have been used. These metals can be made
more creep-resistant by various methods. We can reexamine the strengthening
methods against plastic yield described in Sec. 5.3 to determine their applicability
to creep. Strain hardening cannot generally be used to improve creep resistance,
as recrystallization and the associated softening can occur above 0.4Tm. Solute
strengthening is effective here in the same manner as in low temperature
strengthening.
Grain refinement is not applicable for creep. In fact, the opposite is required
to minimize grain boundary sliding. A fine grain size means more grain bound-
ary area per unit volume. Coarse grained alloys, with the crystals oriented in
such a way that voids are not generated at grain junctions, are adopted. In the
limit, turbine blades made out of single crystals, where there are no grain
boundaries, have been used.
Precipitation hardening is applicable, only when there is no risk of particle
coarsening and consequent softening during service. The driving force for coars-
ening is the decrease in the interfacial energy per unit volume. If the interfacial
energy is low, the driving force is less and the tendency for coarsening is
minimized. In nickel-base and iron-nickel-base superalloys, precipitates of
Ni3(Al,Ti) form a coherent interface with the matrix, with an interfacial energy
as low as ~ 0.005 J m2. The lattice parameter of Ni3Ti(g precipitate) decreases
from 3.59 to 3.56 , as titanium is completely replaced by aluminium. The
optimum ratio of Ti/A1 in the alloy is chosen to match the lattice parameter of
the matrix with that of the precipitate for coherency.
In dispersion hardening, insoluble particles are embedded in a metal matrix.
The components are manufactured by powder metallurgy, by mixing the metal
powder and the insoluble particles (usually an oxide), compacting and then
sintering it to make the metal matrix continuous. During coarsening, small
particles dissolve, the solute atoms diffuse through the matrix and reprecipitate
on larger particles. As oxides do not dissolve in metals, no coarsening occurs
in dispersion hardening and the metal-oxide dispersion remains stable at elevated
temperatures. TD nickel (thoria dispersed nickel) retains an appreciable fraction
of its strength even at temperatures as high as 0.9 Tm.

5.5 FRACTURE
Fracture refers to the breaking of a component into two or more pieces either
during service or during fabrication. Ductile fracture occurs after a considerable
amount of plastic deformation. The nucleation, growth and coalescence of voids
during necking, as discussed in Sec. 5.1, are responsible for ductile fracture.
148 Physical Metallurgy

Brittle fracture occurs without any apparent sign of plastic deformation. However,
plastic deformation on a microscopic level at the tip of a crack frequently occurs
during brittle fracture.
The Griffith criterion Estimates show that the tensile stress required to break
all the atomic bonds simultaneously across the perfect cross-section of a solid
is of the order of Y/6, where Y is the Youngs modulus of the solid. But real
brittle materials fracture at a much lower stress of the order of Y/500 to Y/1000.
This discrepancy is explained on the basis of the presence of cracks in real
materials, which propagate at relatively low stress levels and cause fracture.
Griffith proposed an energy criterion for the spontaneous propagation of a
preexisting crack in a brittle solid. The critical tensile sense sf, at which a flat
elliptical crack propagates and causes fracture, corresponds to the condition that
the elastic energy released per unit area of the crack face is equal to the surface
energy consumed in creating the unit area of the crack face. The Griffith con-
dition is
2H Y
Tf = (5.6)
Qc
where g is the surface energy of the crack faces per unit area, and
c is the half-length of the crack.
Note that the fracture strength sf is inversely related to the square root of the
crack length c.
In order for a crack to propagate after the Griffith condition is satisfied,
there must exist at the crack tip sufficient stress concentration. In brittle mate-
rials such as a silicate glass, the cracks are sharp and a high stress concentration
exists at the tip. Therefore, the cracks propagate, as soon as the Griffith con-
dition is satisfied.
In less brittle (crystalline) materials, the stress concentration at the tip is
relieved by plastic deformation and the crack becomes blunt. In such cases, the
crack propagation is more difficult. Also, the Griffith energy balance should
include the plastic work. Often, the plastic work per unit area of crack face
created turns out to be much larger than the surface energy g. The term gp, used
in place of g in the modified Griffith equation, includes the plastic work.
Limitations of the classical approach
1. Most structural materials are crystalline in nature and invariably undergo
plastic deformation at the tip. The value of the modified surface energy
gp at the critical condition is not easily determined.
2. The fracture strength sf of a material as defined by the Griffith equation
is not material property like yield strength or tensile strength, as it
depends on the crack length in a given component and at a given stage
of the service life.
3. Structures and machine components invariably contain cracks that are
introduced during solidification, fabrication or heat treatment. A material
 Mechanical Properties 149

parameter is required that will allow the design stress to be computed,

if the size of the most damaging defect is known. Alternatively, the
material parameter enables the computation of the maximum size of the
crack that can be tolerated for a given design stress. Periodic
nondestructive inspection can then be done to ensure that the most
damaging defect is smaller than the tolerable size.
4. According to the Griffith criterion, a crack propagates only when the
critical condition is satisfied. But in practice fracture may occur after the
certain lapse of time or after a certain number of stress reversals, even
though the original design stress is less than the critical stress. This is
so, because slow subcritical crack growth occurs during the time interval
before fracture. The crack eventually attains the critical size and fast
fracture ensues. Clearly, the rate of subcritical crack growth should be
known, if the life of the structural part is to be predicted. Such
measurements of crack growth rates have been made both for statically
loaded structures (where crack growth occurs due to the combined action
of the stress and the environment) and for dynamically loaded components
(where stress reversals cause slow crack growth).
Stresses at the crack tip Figure 5.17 illustrates the three modes of loading.
In the opening mode, known as mode I, the displacement is perpendicular to the
crack faces, Fig. 5.17(I). In the sliding mode (mode II), the displacement is
parallel to the crack faces and perpendicular to the leading edge of the crack,
Fig. 5.17(II). In the tearing mode (mode III), Fig. 5.17(III), the displacement is
parallel to the crack faces and the leading edge. In practical situations, the
opening mode is the most important. The stress distribution for the opening
mode at the tip of an elastic crack is considered below.

Figure 5.17 The three modes of loading: mode I (opening mode),

mode II (sliding mode) and mode III (tearing mode).

The normal stress sn on an element located at the radial coordinates (r, q)

(taking the crack tip as the origin) in the direction of the applied stress s is
given by
150 Physical Metallurgy

KI
Tn = f (R ) (5.7)
2Q r

where KI = T Q c and f(q) is some function of q. KI is called the stress

intensity factor. I stands for mode I (opening mode). KI describes the stress field
at the tip of the crack. When the crack tip stress field reaches a critical condi-
tion, fast fracture occurs. The corresponding KI is called the critical stress
intensity factor KIc.
Plastic zone at the crack tip Taking f(q) = 1 (which actually corresponds to
q = 0), sn is plotted against r in Fig. 5.18. As r 0, sn . This is obviously
not possible. When sn exceeds the yield strength sys, plastic deformation occurs
at the tip. The stress distribution at the tip is modified as shown in Fig. 5.18,
the maximum value being sys. The size of the plastic zone rp* can be deter-
mined by substituting sn = sys:
K I2 QT 2 c
rp * = = (5.8)
2T ys
2
2T ys
2

The size of the plastic zone increases with increasing crack size c and
increasing applied stress s. But it decreases, as the yield strength of the material
increases. For cases of limited plastic deformation at the crack tip, KI can be
taken to describe all stresses and strains (both elastic and plastic) around the tip.

Figure 5.18 The stress distribution at the tip of a crack. When this stress
exceeds the yield stress, plastic deformation occurs near the tip.
 Mechanical Properties 151

Experimental determination of KIc Test specifications have been developed

for determining KIc. The test specimen contains a notch of known geometry. It
is fatigue cracked, i.e. it is subjected to a limited number of stress reversals,
which introduces a crack at the root of the notch. Then it is loaded statically and
the load is increased until fracture occurs. The size of the initial crack can be
estimated from the differences in appearance of the regions of the fracture
surface. The region of slow crack growth during fatigue loading is smooth and
polished due to the constant rubbing action of the two crack faces against each
other during stress reversals. The fast crack growth region corresponding to
static loading is rough and granular. The Kc values determined from the crack
size and the stress at fracture are plotted as a function of specimen thickness in
Fig. 5.19. The value of Kc reaches a limiting value at large thicknesses, corres-
ponding to the plane strain condition. This limiting value is termed KIc, so that
it becomes a reproducible material property.

Figure 5.19 The critical stress intensity factor Kc as a function of

plate thickness. The thickness must be sufficiently large for
achieving the plane strain condition.

KIc values for three typical alloys are listed below:

Property Ni-Cr-Mo Steel Maraging Steel Al-alloy

KIc, MN m3/2 46 90 32
TS, MN m2 1820 1850 560

The maraging steel, which has the largest KIc, will tolerate the longest crack
for a given applied stress or with an existing crack of known length, will bear
the highest stress.
Strain energy release rate G Another fracture toughness parameter is the
strain energy released per unit extension of the crack known as G (named after
Griffith). When the crack extends over unit area, two crack faces are created.
152 Physical Metallurgy

At the critical condition for an elastic crack, it can be shown from the Griffith
equation (5.6) that
Gc = 2g (5.9)
When a limited amount of plastic flow occurs at the crack tip, g can be replaced
by gp. Then, at the critical condition,
Gc = 2gp (5.10)
Both KIc and Gc are material parameters valid for high strength alloys, where
plastic deformation at the crack tip is limited.
Crack opening displacement (COD) Frequently, structural components are
built out of low strength and medium strength materials. For steels, the low
strength range is 250 450 MPa and the medium range is 500 800 MPa. The
yield strengths being low for such steels, extensive plastic deformation occurs
at the crack tip. Recall that the size of the plastic zone is inversely proportional
to the square of the yield strength. For such cases, KIc and Gc are not applicable
and different fracture toughness parameters called COD and Jc are used. COD
stands for crack opening displacement. Jc is an energy integral fracture tough-
ness parameter. We will discuss COD only.
Figure 5.20a compares a fully elastic crack and a crack with plastic flow
at the tip. The ideally elastic crack is atomically sharp. The crack with plastic
flow is blunted. Here, at the tip of the crack, the two crack faces are separated
by a distance d, which is defined as the crack opening displacement (COD).
The more extensive is the plastic flow, the more will be d. The criterion used
for the fracture event is that d attains a critical value dc at the time of fast
fracture.

Figure 5.20 (a) A sharp crack and a blunt crack, where plastic deformation has
occurred near the tip. (b) The experimental set up for determining COD.
 Mechanical Properties 153

COD can be measured experimentally, by fixing a clip gauge at the mouth

of the crack as illustrated in Fig. 5.20b. In a three-point bending type of loading,
as the load is increased, the displacement D measured by the clip gauge increases.
By knowing the depth of the crack c and D, d or dc can be calculated.

5.6 FATIGUE FAILURES

About 90% of engineering failures are attributed to fatigue of the material under
cyclic loading. Machine components undergo reversals of stresses millions of
time during their lifetime. Connecting rods are subjected to pull (tension) and
push (compression) type of loading during each cycle. Leaf springs bend back
and forth. Aircraft wings flutter in turbulent air. In a single loading, a metal
withstands a stress up to its tensile strength. But under repeated loading, it may
fail at a stress level much lower than the tensile strength.
Fatigue tests Cyclic loading can be of three types:
1. tension-tension,
2. tension-zero, and
3. tension-compression.
As illustrated in Fig. 5.21, in the tension-tension type of loading, both the
maximum and the minimum stresses are tensile in nature. In the tension-zero
case, the minimum stress is zero. In the tension-compression loading (Fig. 5.21c
and d), the minimum stress is compressive. The stress amplitude S is defined
as half of the difference between the maximum and the minimum stress levels.
For example, in the tension-zero cases, S = smax.

Figure 5.21 Various types of fatigue loading.

+ stands for tension and for compression.
154 Physical Metallurgy

Figure 5.22 shows a typical experimental set up for a fatigue test. The
specimen is in the form of a cantilever loaded at one end. It is rotated at the
same time, by means of a high-speed motor to which it is connected. At any
instant, the upper surface of the loaded specimen is in tension and the lower
surface in compression, with the neutral axis at the centre. During each revo-
lution, the surface layers pass through a full cycle of tension and compression.

Figure 5.22 The experimental set up for a cantilever fatigue test.

The maximum stress amplitude S that the material withstands decreases

with increasing frequency of stress reversals. The results of a fatigue test are
reported in the form of SN curves, where N is the number of stress cycles.
Figure 5.23 shows the SN curves for a 0.45%C steel and for an aluminium
alloy. The steel exhibits a fatigue limit, i.e. if the stress amplitude is below a
certain value, failure does not occur, however large is N. In the case of the
aluminium alloy, there is no fatigue limit. Here, the fatigue strength can be
specified only by giving the corresponding value of N.

Figure 5.23 S-N curves for a carbon steel and an aluminium alloy.
 Mechanical Properties 155

The fatigue strength of a material is lower than its tensile strength and
sometimes, even lower than its yield strength. Even when the cyclic stress is
lower than the yield strength, microscopic plastic deformation occurs on a
localized scale during stress reversals. Slip at the microscopic level may be
evident, at a very early stage of the fatigue life, at about 1% of the total life.
The cyclic nature of the stress causes the slip to appear as extrusions and
intrusions on the surface, as illustrated in Fig. 5.24. During the tensile cycle,
slip occurs on a plane with the maximum shear stress on it. During the com-
pressive part of the cycle, slip may occur on a nearby parallel slip plane, with
the slip displacement in the opposite direction. This results in extrusions and
intrusions at the surface, Fig. 5.24. These act as sites for fatigue crack nucle-
ation. Once nucleated, the crack grows inwards by a small increment during
each cycle. As the critical condition is attained, the crack propagates rapidly
causing fracture.

Figure 5.24 (a) Slip steps appear during forward loading. During reverse
loading, they become either (b) intrusions or (c) extrusions.

A typical fatigue-fracture surface looks like the one shown in Fig. 5.25. The
fatigue crack nucleated at the stress concentration in the keyhole of the shaft.

Figure 5.25 The fatigue fracture surface showing various regions.

The crack initiated at the root of the keyhole, a site of stress concentration.

The rough granular part corresponds to the stage of fast growth, after the
critical condition is attained. The smooth and polished surface corresponds to
the slow growth of crack, when the crack faces smoothen out by constant
156 Physical Metallurgy

rubbing against each other. Concentric lines known as ripples or striations are
often seen on this part of the fracture surface. The relatively widely spaced
ripples are caused by variations in the stress amplitude during the life of the
component. On a much finer level, a large number of ripples may sometimes
be seen. The width of each ripple here is equal to the distance by which the
crack grows during one cycle.
Low cycle fatigue The fatigue failures referred to before occur when the
number of stress cycles is in excess of 105. In low cycle fatigue, failure occurs
after a relatively small number of cycles. The stresses in low cycle fatigue are
usually thermal in origin, the expansion and contraction on heating and cool-
ing of structures such as pressure vessels and turbines. For a given range of
temperature variation, the strains induced by thermal expansion or contraction
are constant. Low cycle fatigue data are presented in terms of Dep (the plastic
strain range) as a function of the number of cycles. The number of cycles to
failure decreases with increasing Dep, i.e. with larger thermal variations,
Fig. 5.26.

Figure 5.26 In low cycle fatigue, the number of cycles to failure

decreases with increasing plastic strain range.

Slow crack growth rate The kinetics of crack growth in the subcritical stage
can be experimentally determined. One of the commonly used empirical equa-
tions correlates the crack growth rate per cycle, dc/dN, to the range of the stress
intensity factor, 'K (= 'T Q c) :
dc
= C ( ' K )n (5.11)
dN
 Mechanical Properties 157

Such a relationship is usually applicable when smin ~ 0. Figure 5.27 shows

a plot of log (dc/dN) against log DK. The constants C and n in Eq. (5.11) can be

Figure 5.27 The crack growth rate during fatigue, as a function

of the stress intensity range, shows a threshold stress DKi,
a region of subcritical growth and a region of fast fracture.

determined from the slope and intercept. It is seen that a threshold intensity
range (DK)i exists below which the crack does not grow. At very large values
of DK, the crack growth rate becomes larger than predicted by Eq. (5.11).
Typical experimental values are:
Martensitic steels : dc/dN = 0.66 108 (DK)2.25 (5.12)
Ferrite-pearlite steels: dc/dN = 0.36 109 (DK)3 (5.13)
For tension-tension type of loading, where smin > 0, the range of K as well
as Kmax influences the growth rate. For tension-compression type of loading,
only Kmax is relevant, as no crack growth occurs in compression.
158 Physical Metallurgy

Factors affecting fatigue life

1. The frequency of cyclic loading has only a small effect on the fatigue
strength.
2. The form of the stress cycle such as square, triangular or sinusoidal
wave has no effect on the fatigue life.
3. The environment in which the component undergoes stress reversals has
a marked effect on fatigue life. The fatigue life in vacuum is some 10
times more than that in moist air. Clearly, the environment has a corrosive
effect. Fatigue occurring under specific corrosive environments is called
corrosion fatigue.
4. The thickness of the test specimen has an effect on the fatigue properties.
Thinner samples show a decrease in the crack growth rate.
5. An increase in temperature around room temperature causes an increase
in the crack growth rate.
Protection methods The methods of increasing fatigue life are primarily aimed
at preventing or delaying the initiation of the crack at the surface.
1. Shot peening introduces compressive residual stresses near the surface
layers. Here, the surface is plastically deformed by repeated impingement
of hard steel balls.
2. Carburizing and nitriding introduce compressive residual stresses at the
surface, as well as increase the hardness and strength. Figure 5.28 shows

Figure 5.28 Nitriding and shot peening improve fatigue life.

 Mechanical Properties 159

the improvement in fatigue life achieved by nitriding and shot peening.

Decarburization lowers the fatigue strength.
3. A fine grain size improves fatigue life.
4. The crack initiation invariably occurs from a site of stress concentration.
In design, fillets of adequate radius of curvature should be provided at
places where a sudden change of cross-section exists. Shafts are usually
highly polished to reduce the chances of surface irregularities providing
sites of stress concentration. Sometimes, stress concentration is introduced
accidentally. A fatigue crack in a plane propeller was traced to the stress
concentration at the root of the stamp mark, bearing the name and
address of the manufacturer!

REFERENCES

1. Dieter, G.E., Mechanical Metallurgy, 3rd Edition, McGraw-Hill, New York

(1988).
2. Honeycombe, R.W.K., The Plastic Deformation of Metals, Edwin Arnold,
London (1968).

QUESTIONS

1. Using Eqs. (5.1) and (5.2), prove that n = etu.

Hint: Assume the volume to be constant during plastic deformation.
2. Explain the role of stacking fault energy on the deformation characteris-
tics of FCC metals and alloys.
3. Distinguish between strain ageing and quench ageing.
4. Discuss the relative advantages and disadvantages of the four main meth-
ods of strengthening metals and alloys against plastic yield.
5. Ordered alloys are generally stronger than disordered alloys. Explain why
this is so.
6. If the grain size of mild steel is changed from ASTM No. 8 to 13 by
microalloying additions, estimate the increase in yield strength.
Answer: 207 MPa
7. The presence of pearlite in the microstructure of carbon steels has little
effect on the yield strength, but tends to increase the tensile strength.
Explain why this is so.
160 Physical Metallurgy

8. Describe the mechanisms of creep that become important, as the tempera-

ture of deformation is progressively increased.
9. From the data of KIc values, make a plot of sc vs. c for the three alloys:
Ni-Cr-Mo steel, maraging steel and aluminium alloy.
10. A connecting rod is made of a martensitic steel with KIc = 46 MN m3/2.
It is subjected to an alternating stress of amplitude 200 MPa and has an
initial crack of length 0.2 mm. Calculate the number of cycles at which
the rod will fail. Assume that Eq. (5.12) is applicable.
Answer: 1.0 105 cycles.
Chapter 6
CORROSION AND ITS
PREVENTION

Metals and alloys corrode in environments encountered during service. Rusting

of iron is a common example of atmospheric corrosion. One estimate puts the
loss due to corrosion at about 3% of the GDP (gross domestic product) of
countries. In addition to direct losses such as replacement cost, many indirect
losses can also arise from corrosion. Shut down of machinery for replacement
results in loss of productivity. Leakages developed in pipelines lead to loss of
the liquid product and can have a hazardous effect on the surrounding environ-
ment and populace. Deposition of the corrosion products inside a tube reduces
the efficiency of heat transfer, resulting in overheating and rupture of the tube.
Contamination of foodstuffs by toxic (poisonous) products can occur.
In this chapter, the electrochemical principles governing corrosion are first
described. After introducing the equilibrium potentials that describe the ten-
dency of various metals to corrode, the phenomena associated with the corro-
sion kinetics such as polarization and passivation are described. In the second
half of the chapter, the methods of corrosion prevention are discussed. These
include cathodic and anodic protection, coatings and alloying. In the end, stress
corrosion cracking is briefly discussed.

6.1 THE GALVANIC CELL

The familiar galvanic cell used in calculators and transistor radios comprises
two electrodes in a solution called the electrolyte, Fig. 6.1. Chemical reactions
occur at the electrodes and the energy liberated due to these reactions is con-
verted into electrical energy.
161
162 Physical Metallurgy

Anode is that electrode at which the conventional current (passage of imagi-

nary or real positive charge carriers) flows from the electrode into the electro-
lyte. Cathode is that electrode at which the conventional current enters from the
electrolyte to the electrode. When the cell is short circuited, electrons (negative
charge carriers) flow in the external metallic circuit in a direction opposite to
that of the conventional current, i.e. from the anode to the cathode. In the
electrolyte, the current is carried by cations and anions flowing in opposite
directions. At any instant, the total current flowing through the cell is the same
in every part of the circuit.

Figure 6.1 The galvanic cell comprises an anode, a cathode and an electrolyte.

The anode is defined as the electrode at which an oxidation reaction occurs.

Removal of electrons (e) from neutral atoms to create positive ions is an
oxidation process. Examples of anodic processes are:
Zn 2e Zn++
Al 3e Al3+
In the above reactions, the metal dissolves as ions into the electrolyte or in
other words, corrodes. Oxidation reactions can also occur in the following
ways.
Fe++ 1e Fe3+
2OH 2e O2 + H2O
Thus oxygen evolution is an anodic reaction.
The cathode is the electrode at which a reduction reaction occurs, by the
addition of electrons to the reactant. Examples are:
 Corrosion and its Prevention 163

Cu++ + 2e Cu
H+ + 1e H2
In the first example, copper is plated out on the cathode. In the second,
hydrogen evolution occurs.
Types of corrosion cells Two dissimilar metals in contact with each other
form a potential galvanic cell. The atmosphere containing oxygen and moisture
can act as the electrolyte. Thus a steel shaft in a bronze bearing, a bronze
propeller in contact with the ship hull, or a copper tube connected to a steel pipe
are combinations of dissimilar metals in contact. They form galvanic cells, in
which the anode may corrode. Two phases in an alloy can form galvanic cells
at the microscopic level. Cementite is cathodic to ferrite in steel, so that ferrite
corrodes, Fig. 6.2a.
If there are variations in the state of stress within the same metal, stress
cells may be set up. The atoms in the grain boundary region are in a distorted
(stressed) state, as compared to the atoms inside the grain. Thus, the grain
boundaries behave anodic to the grains. Similarly, a cold worked region is
anodic to an annealed region of a metal. A bent nail may corrode at the bend,
which is cold worked and anodic to the adjacent undeformed region, Fig. 6.2b.

Figure 6.2 Types of corrosion: (a) microgalvanic cells between different phases,
(b) a stress cell, (c) a concentration cell, and (d) a differential aeration cell.
Arrows in (c) and (d) indicate the direction of the conventional current.
164 Physical Metallurgy

When the concentration of ions in the electrolyte varies from region to

region, concentration cells may be set up. Thus, in an electrolyte containing
cupric ions with two identical copper electrodes immersed in it, the region with
a small concentration of cupric ions is anodic to the region that is rich in cupric
ions, Fig. 6.2c.
When two electrodes of the same metal are in contact with the same elec-
trolyte, if oxygen is bubbled through the electrolyte near one electrode and
nitrogen near the other electrode, corrosion takes place at the electrode deficient
in oxygen supply. This is known as a differential aeration cell, Fig. 6.2d.
Oxygen and water take part in the cathodic reaction as follows:
O2 + H2O + 2e 2OH
(Note that this reaction written in the reverse direction was introduced as
an anodic reaction earlier.) Thus, oxygen deficient regions such as threaded
connections, interfaces between pipes, the metal covered by rust and the part of
the water tank just under the water line act as anode and corrode.

6.2 THE ELECTRODE POTENTIALS

Free energy change and emf of a cell The tendency of a metal to combine
with oxygen and water is measured by DG, the Gibbs free energy change under
standard conditions. For the following reaction, we have
Mg + O2 + H2O Mg(OH)2 DG = 600 kJ mol1
This large negative free energy change indicates that magnesium has a
strong and spontaneous tendency to react in this manner. The above reaction
can be split into two electrode reactions:
Anode reaction: Mg 2e Mg++
Cathode reaction: O2 + H2O + 2e 2(OH)
The emf of a corrosion cell in which the above reactions take place is
related to DG as follows:
DG = nFe (6.1)
where n is the number of electrons taking part in the reaction (it is 2 in the
above case), F is Faradays constant (= 96,490 coulomb per mole of electrons)
and e is the emf measured under standard conditions.
Standard electrode potentials As it is useful to know the tendency of a metal
to go into solution by the anode reaction, irrespective of what the cathode
reaction is, we can split the emf into two half-cell potentials. Since the potential
of a half-cell cannot be determined independently, it is measured with respect
to a reference electrode known as the hydrogen electrode, the potential of which
is taken to be zero. For convenient laboratory work, the secondary reference
electrode used is the calomel electrode, consisting of mercurous chloride mixed
in liquid Hg and in contact with a saturated KCl solution. The calomel electrode
has a standard potential of + 0.241 V.
 Corrosion and its Prevention 165

In the above reaction of magnesium oxidation, DG is negative and so,

from Eq. (6.1), the emf e is positive. However, the standard potentials have
been defined as reduction potentials by international agreement. So, the standard
potential E = e. The table of standard potentials is known as the emf series
and is shown in Table 6.1. The potential of the Li/Li+ electrode at the bottom
of the series is 3.05 V (lithium is the most active metal) and that of the Au/
Au3+ electrode at the top is + 1.50 V (gold is the most noble metal). If, for
example, zinc and copper form a galvanic cell, zinc ( 0.76 V) will be anodic
to copper (+ 0.34 V). The emf of this cell under standard conditions will be
+ 0.34 ( 0.76) = 1.10 V.
Table 6.1 Standard Electrode Potentials of Metals

Electrode Standard Electrode Standard

system Potential E system Potential E
at 25C, volt at 25C, volt
Noble end
Au/Au3+ + 1.50 Zn/Zn2+ 0.76
Ag/Ag+ + 0.80 Mn/Mn2+ 1.18
Cu/Cu2+ + 0.34 Zr/Zr4+ 1.53
H2/H+ 0.0 Ti/Ti2+ 1.63
Pb/Pb2+ 0.13 Al/Al3+ 1.66
Sn/Sn2+ 0.14 U/U3+ 1.80
Mo/Mo3+ 0.20 Be/Be2+ 1.85
Ni/Ni2+ 0.25 Mg/Mg2+ 2.37
Co/Co2+ 0.28 Na/Na+ 2.71
Cd/Cd2+ 0.40 Ca/Ca2+ 2.87
Fe/Fe2+ 0.44 K/K+ 2.93
Cr/Cr3+ 0.74 Li/Li+ 3.05
Active end

The Nernst equation The Nernst equation gives the potential E under non-
standard conditions, when the activity (or concentration) of the metal in the
electrode or the metal ions in the electrolyte is not unity. For the zinc electrode,
RT [Zn 2+ ]
EZn = EZn
+ log (6.2)
2.303 nF [Zn]
At room temperature, RT/2.303 nF = 0.059 V. When the zinc ion concen-
tration in the electrolyte is less than unity, EZn < EZn, i.e. the zinc electrode
becomes more active with a greater tendency to dissolve.
In the reference hydrogen electrode, with hydrogen gas bubbling at one
atmospheric pressure, the potential as a function of hydrogen ion concentration
in the electrolyte is given by
EH 2 = EH 2 + 0.059 log [H + ] (6.3)

By definition, log [H+] = pH. Noting that EH 2 = 0, we have

166 Physical Metallurgy

EH 2 = 0.059 pH (6.4)
In acidic solutions (pH < 7), the electrode is more noble as compared to
alkaline solutions (pH > 7).
The Pourbaix diagrams The emf series gives the electrode potentials under
standard conditions, where the metal ion concentration in the electrolyte is
unity. In aqueous solutions (most common in corrosion problems), the metal ion
concentration is usually much less than unity, due to the limited solubilities of
metal oxides and hydroxides. It is useful to know the equilibrium potentials
between a metal, its ions and compounds as a function of pH in an aqueous
solution. Pourbaix diagrams show such relationships. The diagram for iron and
its hydroxides is given in Fig. 6.3. Line a, which is independent of pH, gives
the equilibrium between iron and ferrous ions:
Fe2 + + 2e Fe

Figure 6.3 The Pourbaix diagram for iron showing the three
main regions: immunity, corrosion and passivity.

This line at a potential of E = 0.617 V is drawn at [Fe2 +] concentration

of 106. Lower concentrations than this can be ignored, as they correspond to
negligible corrosion rates. So, below the line, we have immunity against corrosion.
Above this line, corrosion can occur. Lines b, c, d and e represent the equilibrium
between species on either side of a line as shown in the figure. Except for line
e, the other lines involve either H+ ions or OH ions in the equilibrium reaction
and so are functions of pH. When Fe(OH)3 is stable, protection by the passive
film formation occurs. The dotted lines f and g correspond to hydrogen and
oxygen evolution respectively. Below line f, hydrogen is evolved. Above line g,
oxygen evolution occurs. Between f and g, water is stable.
Other such equilibrium relationships in aqueous systems are listed in the
atlas of Pourbaix diagrams.
 Corrosion and its Prevention 167

Galvanic series in sea water In view of the complex nature of engineering

alloys and the difficulties in quantifying a particular environment, tables of
galvanic series have been developed for commercial alloys in a given
environment. These give the qualitative positions of different alloys in that
environment. The galvanic series in sea water is shown in Table 6.2. Magnesium
at the bottom of the list is the most active and 18/8 stainless steel (passive
condition) at the top of the list is the least active. It may be noted that the order
in this table is not the same as in the emf series (Table 6.1). For example,
titanium and nickel which are more active in the emf series are near the noble
end in the galvanic series, by virtue of their ability to form a passive film.
Table 6.2 Galvanic Series in Sea Water

Noble end 188 stainless steel (type 316) (passive)

188 stainless steel (type 304) (passive)
Titanium
Nickel (passive)
Copper
Brass
Tin
Lead
188 stainless steel (type 316) (active)
188 stainless steel (type 304) (active)
PbSn solder
Cast iron
Mild steel
Alclad
Aluminium
Zinc
Active end Magnesium

6.3 POLARIZATION
The potentials given in the above section are measured under equilibrium con-
ditions, when no current is flowing through the cell. When a net current flows,
there is a shift in the potential of the electrodes; the phenomenon is known as
polarization. The potential shift is always in a direction so as to oppose the flow
of the corrosion current. The polarization of a zinc-copper galvanic cell is
shown in Fig. 6.4. When no current flows, the open-circuit potential of the cell
is ECu EZn = 1.10 V. With increasing flow of current, the potential of the
copper electrode shifts downwards to more active values, whereas the zinc
potential shifts upwards to more noble values. Thus, the potential difference
decreases, as the current increases. If R is the resistance of the circuit, a steady
current I will flow, such that IR is the potential drop. When the cell is short
circuited with negligible resistance in the circuit, the corrosion current will
correspond to the point of intersection of the two polarization curves in Fig. 6.4.
The corresponding potential is called Ecorr.
168 Physical Metallurgy

Causes of polarization
1. Concentration polarization arises from the changes in the concentration
of the reacting ions near an electrode, as soon as a current starts to flow.
At the cathode, if the reaction involves the plating out of the metal ions,
the metal ion concentration decreases. It is easy to see from the Nernst
equation (6.2) that this decrease shifts the potential of the cathode to
more active values. The metal ion concentration never drops to zero, as
the ions from the surrounding electrolyte continue to replenish those
lost by deposition. When the potential shifts sufficiently, some other
electrode reactions such as hydrogen evolution replace the initial reaction.
Similarly, with metal dissolution at the anode, there is an increase in
metal ion concentration there. This build-up near the anode shifts its
potential towards more noble values. If the electrolyte near the anode
gets saturated with metal ions, the excess ions may form a precipitate.

Figure 6.4 The polarization of a zinc-copper galvanic cell.

2. Activation polarization is due to the slowness of an electrode reaction.

Metal deposition or dissolution is usually fast, but gas evolution reactions
are slow. Hydrogen evolution at the cathode occurs in two steps:
H + + e Hads
Hads + Hads H2
The second step, where the two adsorbed atoms combine to form a
molecule, is slow and requires an activation potential to keep it going at the
rate required by the magnitude of the current flowing through the cell. The
extra potential h known as the activation polarization (called the hydrogen
overvoltage in this case) aids to reduce the activation barrier for the reaction,
as schematically illustrated in Fig. 6.5. Metal surfaces which require a large
 Corrosion and its Prevention 169

hydrogen overvoltage retard the combination of the adsorbed hydrogen atoms

into molecules. The atomic hydrogen may then diffuse into the metal and can
cause hydrogen cracking.
The activation polarization follows the Tafel relationship over a wide range
of current density i:
h = a + b log i (6.5)
where a and b are constants. Note that, for a given corrosion current, a small
area of the electrode means a high current density and a greater extent of
polarization.

Figure 6.5 The overpotential h lowers the activation barrier DG*

for an electrode reaction.

6.4 PASSIVATION
When an iron rod is immersed in dilute nitric acid, it reacts vigorously. On the
other hand, if it is immersed in concentrated nitric acid, it stops reacting soon
after immersion. This difference in behaviour is explained on the basis of the
oxidizing capacity of the two acids. The strong acid effectively oxidizes the iron
to form a thin protective film. Once it forms, the film prevents further reaction
of the iron with the acid. The iron is said to become passive. The dilute acid
is not strong enough to oxidize and produce the passive film, so the metal
continues to dissolve.
A passive metal is defined as one that is active in the emf series, but
nevertheless corrodes at a very low rate, behaving like a noble metal. Nickel,
titanium and stainless steels are passive in the atmosphere and therefore do not
rust like carbon steels. Some metals are passive only in specific electrolytes, e.g.
iron is passive in conc. HNO3 and in chromate solutions.
The potential E versus log I for iron in 1N H2SO4 is shown in Fig. 6.6. EA
is the open circuit potential of the metal (acting as the anode). As the current
density increases, the anode becomes polarized along the line AB. At a critical
current density, icrit, the passive film forms and further dissolution of the metal
is drastically reduced. So, the current density abruptly falls to a much lower
170 Physical Metallurgy

value, ipassive. For iron, ipassive is smaller than icrit by a factor of 30,000. The
corrosion rate at this current density is negligible. With further increase in the
potential, the current density remains constant in the passive region of Fig. 6.6.
Eventually, the transpassive region is reached, when the current density starts
to increase again with the evolution of oxygen.
Clearly, a metal that can be passivated and retained in the passive state
during service can reduce the corrosion rate to negligible values. The formation
and retention of a stable passive film are promoted by the following factors:
1. A low critical current density.
2. A more active potential at which passivation occurs.
3. A high potential at which the change to the transpassive region occurs.

Figure 6.6 The passivation of iron in 1N H2SO4.

There are three main methods that use passivity as a means of corrosion
protection:
1. In anodic protection, an external current, icrit, is initially impressed on
the metal so as to passivate it. Then the current density is reduced to
ipassive and maintained at that value to ensure that the passive film does
not dissolve. The necessary electrical equipment is complex and expensive
to install and maintain. If control is lost temporarily and the potential
strays into the active region, the corrosion can be disastrously high.
Because of this, anodic protection is not as widely used as the other
methods.
 Corrosion and its Prevention 171

2. Passivators are inorganic oxidizing anions such as chromates (CrO4)4

and nitrites (NO2)2. Solutions containing these anions promote the
formation of the passive film.
3. Alloying is done to reduce the critical current density for passivation
and also to shift the passivating potential to more active values. 12%
chromium added to iron reduces icrit from 200 mA/cm2 to 2 mA/cm2 and
the passivating potential from 0.6 V to 0.2 V.
Pitting The passive film may be destroyed in the presence of certain anions
in the electrolyte. The presence of Cl ions, e.g. in marine atmospheres, destroys
the passive film on iron, chromium, nickel, aluminium and stainless steels.
Titanium is the only metal that is immune to the Cl ions. The Cl ions
act mainly by lowering the potential at which the transpassive region sets in.
Figure 6.7 shows the decrease in this potential for stainless steels, with
increasing concentration of NaCl in a solution of Na2SO4. The passive film is
usually destroyed in a localized region. This region becomes anodic with respect
to the surrounding passive region and corrodes at a fast rate due to the high
current density over the small area. Thus a pit forms and the continued presence
of the Cl ions in the pit does not allow repassivation. The passive-active cell
continues to corrode and produces a deep pit. The potential at which the current
density starts to increase on entering the transpassive region is called the critical
pitting potential.

Figure 6.7 The presence of chlorine ions reduces the extent of the passive region.
172 Physical Metallurgy

6.5 GENERAL METHODS OF CORROSION PREVENTION

Some methods of corrosion prevention are understood readily. From Table 6.1,
one can deduce that use of noble metals will prevent corrosion. Although this
is true, the choice of a metal for an engineering application is dependent on
many other factors such as cost, availability, fabricability, strength, and so on.
Avoidance of contact between dissimilar metals will prevent corrosion. Yet,
we frequently come across designs such as a steel screw in brass marine equip-
ment, a Pb-Sn solder on copper wire, a steel shaft in a bronze bearing and
aluminium fins attached to copper tubes. When contact between dissimilar metals
is unavoidable, the design should be such that the metal that acts as the anode
does not form a small surface area as compared to the cathode. A small anodic
area means a high current density and consequently a high rate of corrosion. A
copper nut or bolt on a large steel plate is acceptable, but not a steel bolt on
a copper article of large area.
In two-phase alloys, the two phases may form a galvanic couple at the
microstructural level and electrical contact between them is inherent in the
structure. Untempered martensite is a single phase and corrodes at a slower rate
as compared to tempered martensite, where fine carbide particles and ferrite
form innumerable galvanic cells. If the martensite is tempered at a relatively
high temperature (e.g. 550C), coarsening of cementite particles occurs and the
number of potential galvanic cells per unit area of the exposed surface de-
creases. Correspondingly, the corrosion rate decreases after reaching a peak as
a function of the tempering temperature, as shown in Fig. 6.8.

Figure 6.8 The corrosion rate as a function of tempering

temperature of a quenched steel.

Unfortunately, the optimum mechanical properties of many alloys

correspond to a sub-microscopic distribution of two phases. So, such alloys
 Corrosion and its Prevention 173

have a lower corrosion resistance than pure metals. For example, aluminium
has a much better corrosion resistance than duralumin, in which very fine
precipitates are present. In the aircraft industry, alclad is used as a composite
sheet of duralumin covered on both sides with thin sheets of pure aluminium
for corrosion protection.
Removing a cathodic reactant such as oxygen dissolved in water can
reduce the corrosion rate. The feed water for boilers and cooling systems is
treated with sodium sulphite to remove the dissolved oxygen, according to the
reaction:
Na2SO3 + O2 Na2SO4
The dissolved oxygen can also be removed by deaeration, where the water
is just heated to remove oxygen, by using the fact that the solubility of oxygen
in water decreases with increasing temperature.
Inhibitors are additives that reduce the rate of corrosion, by slowing down
or preventing an electrode reaction. Anodic inhibitors include the passivators.
The oxidizing additives such as chromates and nitrites promote the formation
of the passive film on the metal surface, by themselves readily undergoing
reduction. The nonoxidizing anions such as silicates, phosphates and carbonates
do not supply their own oxygen but catalyze the oxidation of the metal in the
presence of oxygen. In the case of iron, the anodic inhibitors are effective only
in neutral or alkaline solutions (recall the pH range of the passive region on the
Pourbaix diagram in Fig. 6.3). Cathodic inhibitors slow down the cathodic
reaction, e.g. hydrazine (N2H4) scavenges the residual oxygen in water by the
following reaction:
N2H4 + O2 N2 + 2H2O
This action is similar to that of sodium sulphite discussed before.

6.6 CATHODIC PROTECTION

Principle of cathodic protection The principle of cathodic protection can be
understood from the polarization diagram in Fig. 6.9. In cathodic protection, the
polarization of the cathode is done to an extent that its potential becomes equal
to the open circuit potential EA of the anode. This can be achieved either by
means of an externally impressed current or by employing a sacrificial anode.
An electrolyte is needed to ensure the passage of current through the part to be
protected. So, cathodic protection is effective only in soils or aqueous media,
wherein the part to be protected is immersed. It is not effective in the
atmosphere.
When the potential of the two electrodes is equal to EA, no corrosion
current passes through the cell. All the current bc is Iimpressed, Fig. 6.9. When
no current is externally supplied, the potential is Ecorr and all current ab is Icorr.
For all potentials between Ecorr and EA, the current passing through the cathode
has two components: ab is the corrosion current and bc is the impressed current.
174 Physical Metallurgy

Under such conditions, the metal is partially protected. When the potential is
EA, the metal is fully protected. When the potential is lower than EA, the metal
is overprotected. Overprotection is usually not desirable, as it means a waste of
electrical energy from the external source as well as the danger of hydrogen
cracking due to excessive hydrogen liberation on the metal surface.

Figure 6.9 The principle of cathodic protection.

Practice of cathodic protection Cathodic protection is one of the most widely

used methods of corrosion protection. It extends the service life of buried steel
pipelines, offshore oil-drilling structures, ship hulls and water tanks.
In the impressed current cathodic protection (ICCP), the externally im-
pressed current comes from a DC source, the positive terminal of which is
connected to an auxiliary anode and the negative terminal to the part to be
protected. Iron scrap can be used as the anode in some cases and it needs to be
replaced periodically. The auxiliary anode in sea water for the protection of ship
hulls is usually platinized Pb-2%Ag alloy. The current density required for full
protection in moving sea water is about 0.01 mA/cm2.
In the sacrificial anode cathodic protection (SACP), a sacrificial anode that
has a much more active potential than the metal to be protected is used. For the
protection of a steel hull in sea water, the potential difference for various
sacrificial anodes is as follows.
Anode Mg Zn Al0.1%Sn
Potential difference 1.0 0.5 1.2

The aluminium alloy does not become passive in sea water. Aluminium can
be used in other applications, such as protection of underground pipes, only if
it is kept active by surrounding it with a solution rich in C1 ions.
Both ICCP and SACP are used frequently in conjunction with a coating
(paint) on the structure to be protected. The coating need not be pore free. The
 Corrosion and its Prevention 175

surface exposed through the pores is cathodically protected by the external

current. The current requirement in the presence of a coating is much less than
that for an uncoated structure. For example, an uncoated oil pipeline buried
underground requires an anode every 30 m of length, whereas a coated pipeline
can be protected over a distance of 30 km with one anode in SACP or 80 km
in ICCP!

6.7 COATINGS
Metallic coatings The principal methods of applying metal coatings for cor-
rosion protection are listed below:
1. Hot dipping: dipping the part to be protected in a molten bath of the
coat metal.
2. Electroplating: the part to be coated is made the cathode in a plating
bath with an electrolyte containing the ions of the coat metal.
3. Spraying: the coat is melted into small drops of liquid and propelled by
a blast of air on to the metal surface.
4. Diffusion coating: the coat metal is diffused into the part through the
surface at an elevated temperature.
According to their electrochemical behaviour, metal coatings are classified
as:
1. noble coatings, and
2. sacrificial coatings.
In noble coatings, the coat is noble with respect to the underlying metal. If
a pore is present in the noble coating, the underlying base metal is exposed and
is attacked continuously, as illustrated in Fig. 6.10a. For effective protection,
such pores are to be sealed by an organic coating on top of the noble metal
coating.

Figure 6.10 (a) A pore in a noble coating causes the underlying metal to corrode,
and (b) a base coating protects by sacrificial action even in the presence of a pore.

In sacrificial coatings, the underlying metal is protected by the sacrificial

action of the coating, which becomes the anode, Fig. 6.10b. Protection contin-
ues as long as the resistance of the electrolyte is small enough for sufficient
176 Physical Metallurgy

corrosion current to flow. With a zinc coating on iron immersed in soft water,
a 3 mm exposed area begins to show rust at the centre, whereas this does not
happen even on a 30 mm exposed area, when the electrolyte is sea water of
higher conductivity.
Nickel coatings are usually applied by electroplating. Nickel is sensitive to
attack by SO2 found in industrial atmospheres and loses lustre by the formation
of NiSO4. A chromium coating on nickel is given to improve the appearance in
applications such as car bumpers.
Zinc coatings on iron are sacrificial. The most common method of appli-
cation is by hot dipping, the process known as galvanizing. Low carbon steel
sheets coated with zinc are called galvanized iron (GI). In aqueous environ-
ments with pH in the range of 712, Zn(OH)2 forms as a protective layer on
top of the zinc. Thus, galvanized iron is especially suitable for use in aqueous
environments.
Tin coatings are applied by electroplating, which produces a uniform layer.
To save tin, which is expensive, coatings may be as thin as 0.002 mm. Tin with
standard potential of 0.136 V is cathodic to iron (E0 = 0.440 V). This
situation prevails on the outer surface of a tin can. On the inside, the stannous
ions form complexes with the juices in foodstuffs and their activity is drastically
lowered, so much so that tin becomes anodic to iron! In the absence of a readily
occurring cathodic reaction, the corrosion rate is negligible. The stannous ions
are nontoxic.
Aluminium coatings are obtained by hot dipping or spraying. Passive alu-
minium is cathodic to iron and a coating needs sealing of pores with an organic
lacquer. Active aluminium (e.g. in sea water) protects by sacrificial action.
Nonmetallic coatings Enamel and porcelain are inorganic nonmetallic coatings.
They act by simply excluding contact of the metal with water and air and also
by providing a high resistance path. They are essentially alkali borosilicate
(Na2O.B2O3.SiO2) glasses. The composition is adjusted to give a matching
thermal expansion coefficient with the metal. Brittleness and poor thermal
shock resistance are the main disadvantages of such coatings. Portland cement
coatings used on underground water pipes have similar characteristics and
they are cheaper.
Organic coatings refer mainly to paints applied to metals. Here also,
protection arises from the exclusion of air and water and from the high electrical
resistance of the paint layer. Paints are a mixture of particles of a pigment in
a continuous organic medium. The common pigments are red lead (Pb3O4) and
zinc chromate (ZnCrO4). The PbO44 and CrO44 ions are passivators and promote
the formation of a passive film on iron. The medium is either linseed oil or other
drying oils, which polymerize by oxidation with air and solidify. Paints are
applied in several coats, as this helps in sealing up pores and reduces the
tendency to crack. The number of coats given increases with the severity of the
exposure.
 Corrosion and its Prevention 177

A good surface preparation is a must for the application of paint. Chemical

solvents can be used to clean out dirt, oil and grease from the surface. Scales
(oxide layers) are removed by dissolving them with dilute sulphuric acid or
hydrochloric acid. A final dip in phosphoric acid is desirable to remove ions like
C1 which prevent passivation. Oxide layers can also be removed by blasting
the surface with sand, steel balls or silicon carbide particles.

6.8 CORROSION PREVENTION BY ALLOYING

Stainless steels Alloying is an important method of corrosion prevention. As
a typical example, stainless steels are discussed here. As already pointed out
under passivation, chromium, when alloyed with iron, changes the passivating
potential of iron to more active values and drastically lowers the critical current
density required for passivation. At very small icrit values, self passivation occurs,
i.e. the corrosion current during atmospheric exposure is sufficient for the
formation of the passive film. Steels with more than 10 12%Cr are self-
passivating and are known as stainless steels.
The passive film on stainless steels is attacked under reducing conditions.
In addition, C1 ions are destructive to the passive film. Nickel and molybdenum
improve the corrosion resistance imparted by chromium. The 304 type austenitic
stainless steel with 18%Cr, 8%Ni and 0.08%C is resistant to neutral chloride
solutions. Molybdenum improves resistance to pitting as well as to sulphuric
acid. Type 316 stainless steel, which contains in addition 2%Mo, is widely used
in the chemical industry.
Intergranular corrosion Stainless steels are prone to intergranular corrosion
(corrosion along grain boundaries) under certain conditions. The pseudobinary
phase diagram for 18/8 stainless steel as a function of carbon content is shown
in Fig. 6.11. With about 0.08%C, the steel is in the fully austenitic state only
when quenched from above 800C. If cooled slowly from this temperature,
chromium carbide particles precipitate preferentially along the austenitic grain
boundaries. With a thin carbide film at the grain boundaries, the steel is said to
be in the sensitized condition.
The composition profile near a grain boundary in the sensitized state is
schematically shown in Fig. 6.12. At the boundary, where the chromium carbide
is present, the chromium content may exceed 70%. In regions immediately
surrounding the carbide on either side of the boundary, a thin region is depleted
of chromium, as it has diffused to the boundary to form the carbide. If the
chromium content in this thin region falls below 12%, they lose their self-
passivating property and become anodic to the grain interiors, which contain
18%Cr and are passive. The galvanic cell thus set up attacks the thin region
selectively, resulting in intergranular corrosion. Under extreme conditions, the
steel crumbles into a fine powder, with each particle in the powder correspond-
ing to one or more grains of the steel before the attack!
178 Physical Metallurgy

Figure 6.11 The pseudo-binary phase diagram for an 18/8 stainless steel as a function
of carbon content. The precipitation of the a phase below the dotted line rarely occurs.

Figure 6.12 Chromium carbide precipitation at a grain boundary and the

composition profile near the boundary.

A sensitized steel can be restored to its original state of excellent corrosion

resistance in the following ways.
 Corrosion and its Prevention 179

1. The steel is reheated into the single austenitic region, Fig. 6.11, and then
quenched. The carbides go into solution during austenitizing and have
no time to reprecipitate during quenching. This procedure cannot be
adopted, if the steel is to be used in service at temperatures between 500
and 800C, as grain boundary precipitation will again occur during
service.
2. A low carbon grade of stainless steel such as 304L or 316L can be used.
In these grades, the carbon content is limited to a maximum of 0.03%.
So, there is no precipitation of carbides, note the solubility of carbon in
austenite in the phase diagram in Fig. 6.11.
3. The stabilized grades 321 and 347 can be used. In type 321, titanium
greater than 5 times the wt%C is added. In type 347, niobium greater
than 10 times the wt%C is present. Both titanium and niobium form
more stable carbides than chromium, recall the carbide stability scale
given on p. 107. So, carbon combines preferentially with titanium or
niobium. Chromium remains in solution and sensitization does not occur.
The above description is valid for austenitic stainless steels. Because of the
low solubility of carbon in ferrite, ferritic stainless steels sensitize much more
rapidly and at lower temperatures. Here, the intergranular corrosion cannot be
prevented, by a solutionizing treatment and water quenching. They must be
annealed at about 800C to replenish the Cr-depleted regions near the grain
boundaries by diffusion of Cr from the interior of the grains. Reducing C to
0.03% is also not applicable here. Duplex stainless steels, which contain meta-
stable ferrite within the austenite grains, are known to be remarkably resistant
to intergranular corrosion.

6.9 STRESS CORROSION CRACKING

Stress corrosion cracking (SCC) refers to the cracking of a material under static
load by the combined action of a stress and a corrosive environment. The
related phenomenon of hydrogen cracking or hydrogen embrittlement is due to
the absorption of atomic hydrogen into the metal.
General characteristics The experimentally known characteristics of stress
corrosion cracking are summarized below.
1. Stress corrosion cracking is found in alloys and not in pure metals.
However, there is some uncertainty as to what level of purity can be
classified as pure!
2. SCC occurs only in a specific environment for a given alloy. Some
examples of environments and the corresponding alloys are given
below.
180 Physical Metallurgy

Alloy Environment
Mild steel Conc. NO3, OH
High strength steel (T.S. > 1200 MPa) H2O
Stainless steel C1
a-brass Traces of NH3 or NH4+
Aluminium alloys H2O, NaC1 solutions

3. The presence of a tensile component of stress is necessary. The stress

may be externally applied. Or it can be residual stresses in the alloy.
Early examples of SCC were due to residual stresses: the season cracking
of brass and the caustic embrittlement of boilers. When cold drawn
cartridge cases of a-brass are exposed to traces of ammonia in the
atmosphere, they crack spontaneously. When the cold worked brass is
given a stress-relieving anneal at 250300C, the susceptibility to SCC
disappears. Boiler water is treated with NaOH to control pH. In the
crevices between bolts and the boiler plates, evaporation results in a
concentration of OH ions, leading to SCC of the bolts having residual
stresses induced by plastic deformation during tightening. Welded boilers
are also prone to SCC, in the presence of residual welding stresses.
4. The cracks in SCC move on a very narrow front. In hydrogen cracking,
the cracks are known as hairline cracks often difficult to identify. If
cracking is due to anodic dissolution at the crack tip, extremely localized
corrosion must be taking place.
5. The crack path is often intergranular in SCC, Fig. 6.13. Transgranular
crack propagation, i.e. propagation across the grains, has also been
found in some cases.

Figure 6.13 Intergranular fracture in stress corrosion cracking.

 Corrosion and its Prevention 181

6. Many of the alloys which are susceptible to SCC form passive films.
Examples are aluminium alloys and stainless steels. However, this is not
always the case.
7. Cracking consists of two stages: (i) initiation and (ii) propagation.
Titanium alloys are immune to crack initiation in chloride solutions. But
when the alloys are precracked, crack propagation is found to occur.
Mechanisms of crack growth In the electrochemical mechanism, corrosion
by anodic dissolution is believed to occur at the crack tip. In alloys that form
a passive film, the crack tip is not covered by the film, because the plastic
deformation at the tip exposes fresh metal or the stress concentration at the tip
is sufficient to break the brittle oxide film. An active-passive cell is set up
between the tip and the crack faces that are covered with the film. Only a very
small area of the tip is exposed, resulting in a high current density at the tip.
Intergranular corrosion is due to the galvanic cells set up at the grain
boundaries by segregation or precipitation. In this mechanism, cathodic protection
should be effective in preventing SCC, as is found in some cases.
In hydrogen cracking, the adsorption of atomic hydrogen at the crack tip
faces can lower the surface energy g in the Griffith equation (5.6). This
explanation is probably valid for the time-dependent cracking observed in non-
metallic materials, where an electrochemical mechanism is not possible. In
ductile alloys, gp is much larger than g and an adsorption mechanism is unlikely.
Hydrogen atoms presumably diffuse into the metal near the crack tip and lock
up the mobile dislocations thus preventing plastic flow. The yield strength
increases locally and the lack of plastic flow at the tip causes embrittlement.
Alternatively, supersaturated hydrogen atoms can precipitate as H2 molecules
at a crack nucleus or interface. The stresses from the build-up of hydrogen
pressure can induce the nucleation and/or growth of cracks. At room temperature,
hydrogen being a small interstitial atom can diffuse over a number of interatomic
distances in a matter of seconds. A high hydrogen overvoltage slows down the
combination of hydrogen atoms into molecular hydrogen, thus facilitating their
inward diffusion. Sulphur containing oil increases the hydrogen overvoltage and
thereby causes embrittlement of oil pipelines.
The stress intensity factor KIscc The time dependent fracture under a corrosive
environment has been studied experimentally by measuring the stress intensity
factor KI . After precracking, a cantilever specimen that is surrounded by the
corrosive environment is subjected to a constant load. The measured time for
fracture is plotted in Fig. 6.14 as a function of KI. If the applied stress corresponds
to KIc, fast fracture occurs instantly. When KI < KIc, fracture occurs after a delay
period, during which the crack grows slowly. When the stress intensity factor
is below a critical value known as KIscc, no fracture occurs even after prolonged
holding, see Fig. 6.14.
182 Physical Metallurgy

Figure 6.14 The critical stress intensity factor decreases with increasing
time in a corrosive environment, reaching a constant value of KIscc,
below which no cracking occurs.

Figure 6.15 shows the measured rates of crack growth as a function of KI.
Below KIscc, the crack growth rate is zero. Above KIscc, in region II, the crack
grows at a constant rate independent of KI. The growth rate is probably con-
trolled by an electrochemical process in this region. Region III corresponds to
fast growth that culminates in fracture.
The energy balance for crack growth in the Griffith equation (5.6) needs to
be modified to take into account the electrochemical energy released at the
crack tip due to anodic dissolution. For unit extension of the crack, we can write
the condition for spontaneous growth as:
surface energy + plastic work = elastic energy + electrochemical
of the crack at the tip released energy released

2gp = ps2c/Y + (nFrd/M)e

where s is the tensile stress,
c is half-length of the crack,
Y is Youngs modulus of the alloy,
n is the number of electrons taking part in the anodic reaction,
F is Faradays constant,
r is the density of the alloy,
d is the opening at the crack tip,
M is the atomic weight, and
e is the oxidation potential (= E).
 Corrosion and its Prevention 183

Figure 6.15 The crack growth rate in a Ni-Cr-Mo steel in a NaCl environment
as a function of the stress intensity factor, depicting the three stages.

Rewriting the critical condition in terms of KIscc, we have

KIscc = [2gpY (nFrd/M)e]1/2 (6.6)
Equation (6.6) shows that KIscc decreases with an increase in yield strength
(which decreases gp) and with an increase in the corrosion tendency as given
by e.

REFERENCES

1. Uhlig, H.H., Corrosion and Corrosion Control, John Wiley, New York (1971).
2. Jones, D.A., Principles and Prevention of Corrosion, Prentice Hall, New
Jersey (1996).

QUESTIONS

1. Calculate the amount of zinc (anode) that corrodes, when a current of

100 mA passes for 5 hr through a galvanic cell.
Answer: 0.61 g.
184 Physical Metallurgy

2. Calculate the corrosion rate in mm per year, when iron dissolves as ferrous
ions at a current density of 0.15 mA/cm2.
Answer: 1.75 mm.
3. Calculate the stannous ion concentration required to shift the standard
potential of tin to that of iron inside a tin can.
Answer: 5 1011.
4. Ordinary brown rust is ferric hydroxide, Fe(OH)3. Write down the chemi-
cal reaction which converts iron to Fe(OH)3, splitting it into an anode and
a cathode reaction.
5. Explain the similarities and differences between the table of standard
electrode potentials and the table of galvanic series.
6. Is a steep polarization curve an advantage or a disadvantage in corrosion
control? Explain.
7. Explain the significance of the terms: icrit, ipassive, Epassive, and Etranspassive.
8. A pit in a 18/8 stainless steel exposed to sea water grows to a depth of
50 mm in a year. To what average current density does this correspond?
Answer: 4.3 mA/cm2.
9. Explain why aluminium used as a sacrificial anode needs to be kept in a
solution of Cl ions.
10. Explain the differences in the electrochemical mechanism of protection
offered by (i) zinc, and (ii) tin, applied as a coating.
11. A high strength bolt of a low alloy steel is heat treated to a hardness of
42 Rc, instead of the specified hardness of 35 Rc. The bolt fails after some
time in a moist atmosphere. Explain the reason for this behaviour.
Chapter 7
ENGINEERING ALLOYS

In this concluding chapter, the structure, treatment, properties and uses of indi-
vidual engineering alloys are considered. The first part deals with tonnage steels
such as low carbon steels and mild steels and their more recent modifications.
This is followed by a brief discussion of high strength steels and spring steels.
Tool steels are then described in some detail. Different types of cast irons are
considered. Among the light alloys, emphasis is on aluminium. We close the
chapter with a discussion of copper and its alloys and bearing metals.

7.1 STEEL SPECIFICATIONS

To select the right steel for a particular application, some familiarity with stan-
dard specifications and the associated notation is necessary. Specifications vary
from country to country and can be based on differing criteria. The majority of
specifications are in terms of the chemical composition of the steel. For an
experienced engineer, the composition indicates the probable treatment given to
the steel and the corresponding mechanical properties obtainable. Alternative
criteria based on mechanical properties or hardenability are also used to specify
steels.
Indian standard specifications The Indian standard specifications for steels
are based mainly on chemical composition. For example, IS: C07 is a carbon
steel with 0.07%C. Similarly, C15 stands for 0.15%C. The concentrations of the
other alloying elements may be expressly stated. For example, 55Mn1stands for
0.55%C, 1%Mn. 15Mn75 refers to 0.15%C, 0.75%Mn. The underlining of 75
185
186 Physical Metallurgy

indicates that the manganese content is 0.75% and not 75%. Even if the
underlining is missing, it is to be understood this way, as there is no steel with
75% Mn! For low alloy steels, the same notation is applicable. The student
should write down the compositions of the following steels: 25Mn1S14,
55Si2Mn90, 50Cr1V23, 40Ni2Cr1Mo25.
For tool steels, the letter T is used at the start of the notation. For example,
T118 stands for 1.18%C tool steel. Examples of alloy tool steels are
T105Cr1Mn60 and T75W18Cr4V1, the latter being the 18-4-1 high speed tool
steel.
AISI-SAE classification As an engineer may come across imported steels or
may have to order a steel from abroad, some familiarity with foreign specifica-
tions is useful.
The US designations for steels are those of the American Iron and Steel
Institute (AISI) and the Society of Automotive Engineers (SAE). In both of
them, the composition is denoted by a four digit number. The last two digits
stand for the carbon content. For example, xx45 stands for 0.45%C. The first
two digits denote the other alloying elements. The table below gives the mean-
ing of these digits for some typical cases. For example, 1040 stands for a 0.4%C
steel. AISI 4340 means a Ni-Cr-Mo steel, with 1.8%Ni, 0.8%Cr, 0.25%Mo, and
0.4%C. 9260 is a spring steel with 2%Si, 0.9%Mn, and 0.6%C. 52100 is a ball
bearing steel with 1.45%Cr and 1.0%C.
AISI-SAE Type of steel
Classification
10xx Plain carbon steels
11xx Resulphurized grades
13xx Manganese steels (1.6 1.9%Mn)
40xx Mo steels
41xx Cr-Mo steels
43xx Ni-Cr-Mo steels
51xx 0.8%Cr steels
52xx 1.45%Cr steels
61xx Cr-V steels
92xx Si-Mn steels
50B60 boron treated steel (0.5%Cr, 0.6%C)

British specifications The older British specifications for steels have the let-
ters En followed by a number. For example, En 2 stands for a carbon steel with
0.15%C. En 8 is a 0.4%C steel. En 8A is the resulphrized grade of En 8. En
24 is a Ni-Cr-Mo low alloy steel. In the later British specification BS970, the
first three digits stand for the alloy content. The middle letter A, M or H stands
for analysis (A), mechanical properties (M) or hardenability (H) based specifi-
cations. The last two digits give the carbon content. The following table gives
some examples of nearest equivalents of different specifications.
 Engineering Alloys 187

Indian British (old) British (new) AISI-SAE

55Cr 70 En 11 526 M 60 5155
40Ni2Cr1Mo25 En 24 817 M 40 4340
C100 En 44 060 A 96 1095
55Si2Mn90 En 45 250 A 53 9260
50Cr1V23 En 47 735 A 50 6150

The German specification starts with the letters DIN, the Japanese with JIS
and the Russian with GOST. Details of these can be found in reference books.
Classifications of stainless steels and tool steels are described under those sections
in this chapter.
Unified numbering system (UNS) The Unified Numbering System for metals
and alloys provides a means of correlating many internationally used numbering
systems currently administered by societies and trade associations. A UNS
designation is not, in itself, a specification, because it establishes no requirements
for form, condition, property or quality. It is a unified identifier of a metal or
an alloy for which controlling limits have been established in specifications
published elsewhere. The UNS designation consists of a single letter followed
by 5 digits. The letter is suggestive of the family of metals: G for carbon and
low alloy steels, S for stainless steels, T for tool steels, A for aluminium alloys
and P for precious metals. For example, UNS G10200 = AISI-SAE 1020 and
T11301 = AISI T1.
Common residual elements in steels The four common impurities or re-
sidual elements in a steel are the following: sulphur, phosphorous, manganese
and silicon. In addition, the steels may contain residual aluminium or nitrogen.
Normally, sulphur should not exceed 0.04 0.05%. However, much lower
sulphur contents are obtained nowadays by special processing techniques, as
removal of residual sulphur appreciably enhances the properties. If no manganese
is present, the sulphur forms FeS, which is present in the form of a thin film
along the grain boundaries. This compound has a low melting point and may
melt, when the steel is heated to high temperatures, resulting in hot shortness,
i.e. brittleness at high temperatures. In the presence of manganese, sulphur
combines preferentially with manganese to form globules of MnS. The steel is
free of hot shortness in this form. It is necessary to have manganese at least five
times the weight of sulphur to ensure globule formation.
Normally, the phosphorus content of a steel does not exceed 0.04%. Even
though phosphorus is an effective strengthener of ferrite (see Fig. 5.12b), it
drastically lowers the ductility. Phosphorus is said to induce cold shortness, in
contrast to the hot shortness effect of sulphur.
The beneficial action of manganese on sulphur has already been referred
to. In addition, the solute strengthening effect of manganese is appreciable. It
also lowers the critical temperatures and promotes the formation of finer pearl-
ite. The ductile-brittle transition temperature of a steel decreases in the presence
188 Physical Metallurgy

of manganese. As manganese increases the hardenability markedly (see Fig. 4.14),

its presence in excess of 2% may result in the formation of martensite, when
it is not desired.
Silicon is a good strengthener of ferrite. It retards softening during temper-
ing (see Fig. 3.21) and thus aids in the removal of quenching stresses without
appreciable decrease in hardness.
Figure 7.1 shows the general variation in mechanical properties of ferrite-
pearlite steels as a function of carbon content. The detailed description of steels
of specific carbon ranges is given in several sections below.

Figure 7.1 The variation in typical mechanical properties as a function

of carbon content in ferrite-pearlite steels.

7.2 LOW CARBON STEELS

Conventional low carbon steels Low carbon steels contain about 0.1% car-
bon. They are not heat treated, as the hardenability is too low to produce
martensite. They are used in the form of cold rolled sheets. The microstructure
consists essentially of ferrite, with a small amount of pearlite. Typical mechani-
cal properties are:

Y.S., MPa T.S., MPa % Elongation

200 300 300370 2840
 Engineering Alloys 189

The excellent formability of these steels makes them suitable for cold-
formed applications such as stampings of automobile and refrigerator bodies,
tin cans and corrugated sheets.
In the tensile test, low carbon steels show a pronounced yield point, followed
by plastic deformation at constant load (yield point elongation), recall Fig. 5.5.
This plastic deformation is inhomogeneous and occurs by the passage of Luders
bands. This produces a rough surface, called the orange peel effect. Surface
roughness mars the appearance of the final product, and hence is not desirable.
The effect can be minimized by having a fine ferrite grain size, in the range of
ASTM 79. Also, temper rolling eliminates surface roughness. Here, the sheets
are given a skin pass, i.e. a small reduction of about 1% by passing through
rolls. The steel is thereby deformed beyond the yield point, prior to the stamping
operation.
Low carbon steels are susceptible to strain ageing. Even though strain
ageing can serve as a method of strengthening, it can lead to a dramatic increase
in the impact transition temperature and consequent embrittlement at ambient
temperatures.
Cold forming characteristics Low carbon steels are ideally suited for cold
forming operations such as deep drawing and stretch forming. The deep draw-
ing quality of a steel is primarily controlled by the plastic anisotropy parameter
R which is defined as
Fw
R= (7.1)
Ft
where ew is the true strain in the width direction and et is the true strain in the
thickness direction of a strip tensile specimen. A large value of R is indicative
of good resistance to thinning of the walls of a sheet during deep drawing. A
proper recrystallized texture must be developed for increasing R. Figure 7.2
shows that R increases with increasing grain size and is more for A1-killed and
Ti-added steels than for rimmed steels.
Stretch forming, where the ends of a sheet are firmly gripped during stretch-
ing, is primarily influenced by the work hardening exponent n. The larger is n,
the more is the uniform elongation and this helps in delaying localized necking
during forming.
Dual phase steels In recent years, a good deal of effort has been put in the
development of higher strengths in low carbon steels, without an appreciable
loss of formability. Dual phase steels are characterized by:
1. The absence of a discontinuous yield point,
2. A low ratio of yield strength to tensile strength,
190 Physical Metallurgy

3. A high work hardening rate,

4. A high tensile strength, and
5. A large uniform elongation.

Figure 7.2 The effect of grain size and Al/Ti additions on the
plastic anisotropy parameter R.

Low carbon sheets are annealed in the inter critical (a + g) region of the
Fe-C phase diagram and then cooled rapidly enough to convert the austenite
part to lath martensite. The final microstructure consists of 1520% of martensite
islands in a fine-grained ferrite matrix, Fig. 7.3. Typical mechanical properties
attained are:
Y.S., MPa T.S., MPa % Elongation
350 650 30

The formability of dual phase steels is only slightly inferior to the conven-
tional low carbon steels, but they have the advantage that no Luders bands form
during deformation. Due to the high work hardening rate, the strength increases
to 550 MPa after only 3% of plastic strain.
The dual phase steels are used in various stamped automobile components
such as bumpers, wheel discs and door panels. Weight saving by as much as
10% is possible. A 45 kg reduction in weight saves about 2% of fuel in city
driving. Besides automobiles, the energy savings can be significant in other
modes of ground transportation.
 Engineering Alloys 191

Figure 7.3 The microstructure of a dual phase steel consists of

1520% of lath martensite dispersed in a matrix of ferrite.

7.3 MILD STEELS

Conventional mild steels Mild steel used in large tonnages for structural
purposes contains 0.150.25% carbon. It is not heat treated, but used in the as-
rolled and air-cooled condition. The microstructure is about 25% of fine pearlite
interspersed in a ferrite matrix. Typical mechanical properties are:

Y.S., MPa T.S., MPa % Elongation

300 350 400 450 26 30

Due to higher carbon contents, the strengths are higher here, as compared to the
low carbon steels, but the ductility is lower.
Both mild steels and low carbon steels possess excellent weldability. The
heat affected zone (HAZ) near the weld attains a temperature above A3 and
becomes austenite. When the welding is complete, this region cools more rapidly
than in air cooling, due to the quenching effect of the surrounding cold metal.
If the carbon content does not exceed 0.25%, the hardenability (taking into
account the residual manganese) is sufficiently low for non-martensitic products
to form in the HAZ. Even if martensite is to form, its hardness is about 45 Rc
and so, it is not brittle. Structural components (that are often welded in situ)
such as plates for ship hulls, boilers, oil pipelines, I beams, H beams, angles,
channels and girders are some of the numerous applications of mild steel.
192 Physical Metallurgy

Banding, frequently encountered in mild steel, refers to the microstructural

condition that consists of alternate layers of ferrite and pearlite, Fig. 7.4. It is
caused by the interdendritic segregation of manganese during solidification.
Manganese reduces the activity of carbon in austenite and so, carbon also
segregates with manganese. At such concentrated regions, pearlite forms during
cooling after rolling. For many applications, banding is not harmful. Since it is
easier for a crack to propagate parallel to a band than normal to it, banded steels
are to be avoided in applications where lamellar tearing is possible. Figure 7.5
illustrates lamellar tearing (crack propagation parallel to the band) in members
constrained due to heavy welding, as in very large ships and off-shore plat-
forms.

Figure 7.4 Banding in mild steel.

The anisotropic properties of hot-rolled steels are not so much due to banding
as due to nonmetallic inclusions. These inclusions get elongated during rolling
and give rise to anisotropy. The ductility and toughness are lower in directions
normal to the rolling direction. To obtain uniform mechanical properties in all
directions, the sulphur and oxygen content of the steel should be reduced as
much as possible to prevent the formation of sulphides and oxides. Any inclusion
present must be small and equiaxed or globular. To obtain this condition, it is
necessary to precipitate equiaxed particles that resist deformation during
subsequent rolling. This is done by adding elements such as Ca, Zr or Ce (or
mixed rare earths). Calcium additions globurize the alumina-type oxides by
forming calcium aluminates. Zirconium dissolves in manganese sulphide and
decreases its plasticity. Rare earths form a non-deformable sulphide or
oxysulphide. The major effect of these additions is to increase the transverse
ductility and toughness to levels comparable to those obtained in the longitudinal
direction.
 Engineering Alloys 193

Figure 7.5 Lamellar tearing (parallel to banding) occurs in

members constrained by welding.

Microalloyed steels Attempts have been made to improve the strength of

mild steel, without impairing its excellent weldability. This has led to a new
class of steels known as microalloyed steels. To the basic composition of mild
steel, small amounts of strong carbide forming elements such as niobium,
vanadium and/or titanium are added. The total concentration of the microalloying
elements is less than 0.2%.
As discussed under Sec. 5.3, the microalloyed steels have a very fine ferrite
grain size, in the range of ASTM 1214. The yield strength increases with
decreasing grain size, recall the Hall-Petch equation (5.5). In addition, the fine
dispersion of alloy carbides results in precipitation hardening. Typical contribu-
tions from various strengthening mechanisms to the total yield strength are as
follows:
Strengthening mechanism Contribution, MPa
Iron lattice 35
Solid solution 135
Precipitation 45
Dislocations 50
Grain size 260
Total 525

It is evident that the major contribution comes from grain refinement. It has
the additional advantage in that the impact transition temperature (ITT) decreases
194 Physical Metallurgy

with decreasing grain size, whereas it increases with other strengthening methods.
For example, the increase in the transition temperature per unit increase in yield
strength is compared below for precipitation hardening and grain refinement.
The factor is negative for grain refinement.

Strengthening mechanism DITT/Ds, C/MPa

Precipitation hardening + 0.3
Grain refinement 0.8

Typical mechanical properties of microalloyed steels are:

MPa Y.S. MPa T.S. % Elongation

400 500 600 650 20 22

It may be noted that the increase in the strength of microalloyed steels

results in lower elongation values. Unlike dual phase steels, which have good
formability in spite of an increase in the tensile strength, microalloyed steels
have poorer formability than the conventional low carbon steel. The stress-
strain curves of a low carbon steel, a dual phase steel and a microalloyed steel
are compared in Fig. 7.6.

Figure 7.6 Comparison of the engineering stress-strain curves of three steels.

 Engineering Alloys 195

Weathering steels If copper and phosphorus are added to mild steel, the
resistance to atmospheric corrosion is vastly improved. These are known as
weathering steels. Cu2+ and PO43 ions promote the formation of a cohesive
crack-free layer of FeOOH (Fe2O3.H2O), thereby limiting the access of oxygen
and water to the metal surface. Copper in conjunction with chromium is also
used for improving the corrosion resistance of mild steels. The life of many
paint systems is doubled on weathering steel surfaces.
Free cutting steels When machinability is the main criterion for selection,
free cutting grades are specified. These steels generally contain intentionally
added sulphur and are known as resulphurized grades. Sufficient amount of
manganese (58 times the weight of sulphur) is also added to form globular
MnS particles in the ferritic matrix.
A very hard material such as high carbon martensite is difficult to machine.
At the other extreme, a very soft steel is also difficult to machine, because
during turning on the lathe a soft phase like ferrite forms continuous curls,
which produce a rough surface and also heats the tool. If the curls were to break
into chips, the tearing and heating can be avoided. In the resulphurized grades,
the presence of an adequate number of globular MnS particles breaks the
continuity of the matrix and promotes the tendency for chip formation. The
machinability index of resulphurized grades is about twice that of a medium
carbon steel.
Machinability can also be improved by adding insoluble metals like lead
and copper to the ferrite matrix. Addition of phosphorus up to 0.12% results in
the formation of Fe3P particles, which serve a function similar to that of MnS
particles.

7.4 MEDIUM CARBON STEELS

Forged components such as agricultural implements have carbon contents in
the range of 0.30.5%. After hot forging, the components are air cooled. The
microstructure is fine ferrite-pearlite mixture.
Rail structurals Structural parts used by the railways such as rails, wheels,
wheel tyres and axles are either forged or hot rolled and have a carbon content
of 0.50.65%. The higher level of carbon combined with about 1% Mn shifts
the eutectoid composition sufficiently to yield a mostly-pearlitic structure. The
lowering of the transformation temperature by Mn results in a finer pearlite.
Figure 7.7 shows the weight loss due to wear as a function of hardness and
interlamellar spacing in pearlite.
For service conditions with exceptionally high rates of wear in rails, an
alloy steel (sometimes called Hadfield steel) with 0.75 0.9%C and 1214%Mn
is used. The steel is austenitic in structure and has a high rate of work hardening
associated with the formation of stacking faults in the austenite.
196 Physical Metallurgy

Figure 7.7 The weight loss due to wear of rail steels decreases with increasing
hardness of the steel and decreasing interlamellar spacing of pearlite.

Spring steels Spring steels have carbon contents in the range of 0.50.65%.
They are quenched and tempered to a high level of yield strength. A spring must
possess a high resilience. It must be capable of storing a large amount of elastic
energy, so that it can spring back without permanently deforming. The quantity
sy /Y should be maximized for this purpose. The Youngs modulus Y being a
structure-insensitive property, a large increase in the yield strength is the key
to achieve a high resilience. Tempering is done to give a yield strength of about
1500 MPa.
The alloying elements in a spring steel have several functions. They increase
the hardenability so that martensite is obtained without resorting to drastic
quenching. The presence of silicon in the 55Si2Mn90 spring steel serves the
additional purpose of retarding softening during tempering, so that residual
stresses are relieved without much loss in hardness and strength. Vanadium in
the 50Cr1V23 steel prevents grain coarsening during austenitizing and improves
the toughness of the steel. A fine grain size and prevention of decarburization
during heat treatment ensure a good fatigue strength.
 Engineering Alloys 197

7.5 HIGH STRENGTH STRUCTURAL STEELS

High strength steels have often tensile strengths in excess of 600 MPa. Those
with strengths up to 1000 MPa are called high strength steels. Those with
strengths greater than 1000 MPa are called ultra high strength steels.
High strength low alloy (HSLA) steels These steels contain carbon in the
same range as mild steels (0.15 0.25%), but in addition have small quantities
of alloying elements. The microalloyed steels discussed earlier are sometimes
called HSLA steels. In many HSLA steels, the alloying elements serve the
purpose of increasing the hardenability to some extent. The microstructure may
consist of ferrite and bainite or ferrite and tempered martensite. Typical
mechanical properties are:

Y.S., MPa T.S., MPa % Elongation

400 700 500 800 18 25

These steels have satisfactory weldability, as the carbon content of marten-

site (if present) is low.
Ni-Cr-Mo low alloy steels Among the common alloying elements, nickel is
the only one which increases the toughness of ferrite. Chromium, in addition
to increasing the hardenability, increases the strength and wear resistance, but
at the expense of toughness. For structural alloys, the Ni/Cr ratio should be
about 2.5. To counter the temper brittleness inherent in Ni-Cr steels (recall Sec.
4.5), about 0.25%Mo is added. One of the well-known low alloy structural
steels is the Ni-Cr-Mo steel corresponding to AISI 4340.
A 0.4%C steel, when quenched and tempered at 450C, has the following
properties:

Y.S., MPa T.S., MPa % Elongation Izod impact, J

650 800 15 30

A 4340 steel, when quenched and tempered for the same elongation and
impact value, has the following strength properties:

Y.S., MPa T.S., MPa % Elongation Izod impact, J

1200 1350 15 32

It is clear that, for the same ductility and toughness, the low alloy steel
possesses superior strength. Conversely, for the same strength, the low alloy
steel would have larger ductility and toughness. This superiority arises from the
presence of nickel among the alloying elements, as also from the increased
hardenability, which enables slower quenching rates with less residual stresses
in the steel. The ideal critical diameter for the 4340 steel is 65 mm, as compared
to 15 mm for a 1040 steel. The high hardenability, however, makes welding
difficult in the case of 4340 steel.
198 Physical Metallurgy

Ausformed steels Thermomechanical treatments (TMT) refer to a combina-

tion of thermal treatment and mechanical deformation of steels. Controlled
rolling of microalloyed steels discussed under Sec. 5.3 is TMT carried out on
stable austenite. TMT of metastable austenite is called ausforming, as discussed
below.
In ausforming, the steel after austenitizing is cooled to a temperature that
lies in the bainitic bay (the temperature interval between the bainitic nose and
the Ms temperature) and plastically deformed. It is then quickly oil quenched to
room temperature before any non-martensitic products form. The martensitic
units that form in the deformed austenite are smaller and finely dispersed. The
dislocation substructure of the deformed austenite is inherited by martensite.
The properties after tempering at 100C are:

Y.S., MPa T.S., MPa % Elongation

2300 2700 8

This is a remarkable combination of strength and ductility. A plain carbon

steel tempered to the same hardness level will be brittle with poor tensile
strength.
Maraging steels Another group of ultra high strength steels are the maraging
(martensite plus ageing) steels. A typical composition is 0.03%C, 18%Ni, 38%
each of Co and Mo, a fraction of a percent each of Ti and A1. The steel
transforms on cooling to martensite, which is soft due to the low carbon content.
The steel is then cold worked as desired and aged around 500C. During
ageing, strain induced precipitation of intermetallic compounds such as
Ni3(Ti,A1) and Ni3Mo form, resulting in precipitation hardening. A very high
yield strength of ~1800 MPa is combined with an excellent fracture toughness
KIc = 120 MN m3/2. Full weldability is another advantage, due to the softness
of martensite. By virtue of their high cost, maraging steels are used mainly in
special applications such as rocket casing and other aerospace applications.
Trip steels TRIP stands for Transformation Induced Plasticity. A typical
composition of trip steels is 0.25%C, 2%Mn, 2%Si, 8%Ni, 9%Cr, and 4%Mo.
The composition is so adjusted that the Ms is below room temperature and Md
is above room temperature (Md is the highest temperature up to which deformation
of austenite can induce martensite). The steel is first heavily deformed above
Md, where no transformation occurs. The deformation produces the right degree
of metastability so that a small plastic strain at the tip of the crack is sufficient
to induce the austenite to transform to martensite. The plastic zone size at the
crack tip is thus enlarged, so that more work is done during crack growth. The
result is a high yield strength 1400 MN m2, combined with a high fracture
toughness KIc = 100 MN m3/2. Like maraging steels, TRIP steels are expensive
and used for specialized applications.
 Engineering Alloys 199

7.6 TOOL MATERIALS

Classification and property requirements The production of almost every
object in industry requires tools. In addition to the familiar hand tools, there is
a wide variety of machine tools used by industry. Tools play a vital role in
industrial advancement. Most tools have been developed on the basis of shop
floor experience. Nevertheless, the structure-property relationships in tool steels
are better understood today, due to increasing availability of new data on phase
relationships and T-T-T diagrams.
Tools can be classified into the following types:
1. Cutting tools such as those used on a lathe,
2. Shearing tools: shears and blanking dies,
3. Battering tools: chisels, rail track tools and punches, and
4. Forming tools: forging dies, piercing tools, extrusion dies, die-casting
dies and cold forming dies.
The general properties required of a tool are:
1. Wear and abrasion resistance, a property that is met by the presence of
hard alloy carbides,
2. Toughness that is obtained with lower carbon contents,
3. Red hardness is the ability to retain hardness at elevated
temperatures, a property that is imparted by secondary hardening
during tempering,
4. Nondeforming property is the resistance to distortion or dimensional
changes during heat treatment, a property that is achieved by a high
hardenability that permits air cooling during hardening, and
5. Minimum cost, the steel should be no more expensive than is needed to
meet the particular requirements.
Heat treatment of high speed steel The general principles involved in the
heat treatment of a tool steel can be illustrated with the example of a tungsten-
based high speed steel: 0.7%C, 18%W, 4%Cr, and 1%V. A pseudo-binary phase
diagram for this steel is shown in Fig. 7.8. The left end corresponds to Fe-18W-
4Cr-lV and percent carbon is plotted along the x-axis. The eutectoid temperature
is 840C. The maximum solubility of carbon in austenite is only 0.7% and this
occurs at the eutectic temperature of 1330C. In the early stages of development
of high speed steel, these phase relationships were not well understood. An
austenitizing treatment of 850C followed by quenching produced a hardness of
45 Rc, that is characteristic of martensite with 0.22%C. Clearly, higher
austenitizing temperatures are required to put more carbon in solution in austenite.
Chromium carbide (Cr23C6 or Cr7C3) begins to dissolve above 900C and is
fully in solution at 1100C. The M6C-type carbides (Fe3(Mo,W)3C) begin
dissolving above 1100C. Vanadium carbide (VC) is the last to dissolve. To put
200 Physical Metallurgy

most of the carbon in solution, heating close to the eutectic temperature is

required. This introduces additional problems in heat treatment such as large
temperature gradients, oxidation, decarburization and austenitic grain growth.
To minimize these, the heating is done in two stages: first to about 850C,
followed by rapid heating for only a few minutes at 1250 1300C before
quenching. The short time at this high temperature minimizes oxidation and
decarburization. Much of vanadium carbide is still undissolved after a few
minutes of soaking and aids in restraining the austenitic grain growth.

Figure 7.8 The pseudo-binary phase diagram for Fe18W4Cr1V tool

steel as a function of the carbon content.

The T-T-T diagram for the same composition and austenitizing treatment is
shown in Fig. 7.9. The steel is quenched in two stages to reduce thermal
stresses: first to 500C, where the time for transformation is very large. After
equalization of temperature across the cross-section, it is quenched to room
temperature. As Ms is below room temperature, the steel is further cooled to
80C (subzero treatment) to convert the retained austenite to martensite. It is
then tempered at 550C, corresponding to the peak of secondary hardening.
Very fine precipitates of tungsten and chromium carbides form. Retained austenite
that is still left over does not transform, but gets conditioned by probably losing
carbon to the surroundings and transforms to martensite on cooling. A second
tempering treatment is necessary to temper the newly formed martensite.
Sometimes, as many as four tempering treatments are given.
 Engineering Alloys 201

Figure 7.9 T-T-T diagram for a 18W4Cr1V0.75C tool steel.

Medium duty tools The cheapest tool steel is the water hardening type denoted
by W in the AISI classification. It is a plain carbon tool steel, with carbon in
the range of 0.6 1.4%. A small quantity of vanadium is present in some cases
to keep the austenite grain size small. The austenitizing temperature is above
Ac3 for hypoeutectoid steels and above Ac1 for hypereutectoid compositions.
Tempering is done in the range of 150 300C and the tempered hardness
ranges from 40 55 Rc. These steels are used for general purpose medium
duty applications such as woodworking tools, files, taps, reamers and most
drill bits.
The shock resisting tool steels denoted by S contain lower carbon, ~0.5%.
Small quantities of chromium, tungsten and molybdenum are present to increase
hardenability. These are oil quenched to reduce quenching stresses and to impart
shock resistance. Typical uses include chisels, hammers and punches.
Tool for cold and hot forming Cold work tools fall under three categories,
depending on the severity of duty. The oil hardening type denoted by O has
typically 0.9%C, with small quantities of alloying elements. They are tempered
202 Physical Metallurgy

at a relatively low temperature to obtain a high hardness, that will withstand

without deforming the higher loads involved in cold forming operations. The O
type is used for light duty applications such as moulding plastics. Medium duty
applications are met by the air hardening type denoted by A. The concentration
of alloying elements is higher to ensure adequate hardenability for air harden-
ing. For heavy duty applications, such as coining dies and cold forging dies, the
high-chromium, high-carbon type denoted by D is used. It contains 1.52.25%C
and 12%Cr. To put sufficient quantity of chromium in solution, a higher
austenitizing temperature is used. Still an appreciable quantity of undissolved
chromium carbide persists, which adds to the final hardness.
Hot work tools used in hot forming and hot extrusion dies are of two types.
The chromium-based ones denoted by the numbers H1 to H19 contain 57%Cr
and 0.350.55%C. A high austenitizing temperature is used to put almost all
carbon in solution during austenitizing. The tempering temperatures are also
high, in the range of 500650C. Hot forming service temperature can be up to
400C. Note that the service temperature is at least 100C lower than the
tempering temperature, for the properties of the steel to remain stable during
service. The higher tempering temperature means a lower hardness. This is
usually adequate, as loads in hot forming are less as compared to cold forming.
Hot shears, aluminium and magnesium extrusion dies and die-casting dies are
some of the applications.
The tungsten-based hot forming steels are denoted by H20 to H39 and
contain 918%W and 0.350.5%C. The austenitizing and tempering tempera-
tures are higher than those for chromium-based hot work tools. Hot extrusion
dies for brass, nickel or steel are made out of these steels.
Tools for high speed cutting High speed steels have the characteristic prop-
erty of red hardness. They cut at twenty times the speed of a plain carbon tool
without losing their sharpness even when they are hot enough to glow with a
dull red colour in a dark room! The tungsten-based ones denoted by T contain
typically 0.75%C, 18%W, 4%Cr, and 1%V (this composition is designated T1).
Tungsten provides for toughness, wear resistance and cutting ability, chromium
serves to increase the hardenability and vanadium is for grain refinement. The
heat treatment steps for this steel have already been discussed.
In the molybdenum-based high speed steel denoted by M, part or all of the
tungsten is replaced by molybdenum. To provide the same atom ratio for carbide
formation, the wt% of molybdenum required is less, as the atomic weight of
molybdenum is only half of that of tungsten. M2 with 0.85%C, 6%W, 5%Mo,
4%Cr and 2%V is very common high speed steel, ranking higher than T1 in
popularity.
 Engineering Alloys 203

In sintered carbide tools, tungsten carbide particles are embedded in a

matrix of 520% cobalt. The carbide and metal powders are mixed, compacted
and sintered close to the melting point of cobalt. Cobalt undergoes incipient
fusion during sintering and forms a continuous matrix with interspersed WC
particles. The hardness of this tool material is extremely high, about 1500 on
the Vickers scale (note that 65 Rc = 830 VPN). This tool is used for very heavy
duty applications such as cutting cast iron and drilling rocks. It can cut at five
times the speed of high-speed steel.
Other tool materials (which are not steels) are stellites, cermets and ceramic
tools. Stellites are Co-Cr-W-C alloys with excellent red hardness. Cermets are
generally titanium and chromium carbide particles embedded in a nickel-base
matrix. Ceramic tools are made primarily from Al2O3.

7.7 STAINLESS STEELS

Stainless steels are generally chosen for service conditions where corrosion and
oxidation resistance is important. In addition, they often have good creep strength,
as cross-slip of screw dislocations is difficult due to the low stacking fault
energy. They contain not less than 10 12%Cr and often Ni in addition. Chromium
forms a protective passive film on the steel surface. Referring to the Fe-Cr
phase diagram, Fig. 2.13, we note that, for Cr > 12.7%, the ferrite phase
becomes stable over the entire temperature range up to the melting point. In the
presence of carbon, to produce an all-ferrite microstructure, we need (Cr 17
%C) > 12.7. If this is less than 12.7%, we have the martensitic stainless steel,
as martensite forms on cooling the high temperature austenite to room
temperature. In austenitic stainless steels, the gamma stabilizers such as Ni
and Mn are present in sufficient amounts to make austenite stable at room
temperature.
The effect of other elements present in Cr-Ni steels can be expressed as
nickel equivalent if they stabilize the austenite and as chromium equivalent if
they stabilize ferrite:
Ni equivalent = Ni + Co + 0.5Mn + 0.3Cu + 30C + 25N
Cr equivalent = Cr + 2Si + 1.5Mo + 5V + 5.5Al + 1.75Nb + 1.5Ti + 0.75W
The Schaeffler diagram (Fig. 7.10) depicts the phases present in an alloy
as a function of nickel and chromium equivalents. More recent thermodynamic
calculations by this author indicate that the Ni and Cr equivalents are also a
function of the annealing temperature.
In discussing the various stainless steels, we can follow the AISI classifi-
cation, as it is widely understood. Three digits are used in this classification.
The AISI 500 series steels contain 46%Cr and so are not true stainless steels.
However, they are 510 times more resistant to sulphide corrosion than a car-
bon steel and 3 times more oxidation resistant. Typical uses are in the petroleum
refining industry for valve bodies and cracking still tubes.
204 Physical Metallurgy

Figure 7.10 The Schaeffler diagram depicting the phases present

as a function of the nickel and chromium equivalents.

Ferritic stainless steels When (%Cr 17 %C) > 12.7, the stainless steel
is in the ferrite state and cannot be heat treated, as austenite does not form at
any temperature. The strength can be increased only by cold working. Grain
refinement is done by recrystallization following cold work. AISI 430 with
16%Cr and 0.12%C max. and AISI 446 with 25%Cr and 0.20%C max. are
typical examples. By virtue of the high chromium content, these steels possess
excellent corrosion and oxidation resistance. Typical mechanical properties of
430 in the annealed condition are:
Y.S., MPa T.S., MPa % Elongation Izod impact, J
350 550 30 75

The formability of ferritic stainless steels is poorer than that of the auste-
nitic steels. In recent years, the formability has been improved considerably, by
lowering the level of the interstitial carbon and nitrogen to 100 ppm (0.01%)
by special processing techniques. Addition of 2%Mo to the basic compositions
of 18%Cr and 25%Cr improves the resistance to pitting corrosion. By virtue of
their lower cost (expensive nickel is not present!), the ferritic stainless steels
with improved properties are competing with austenitic stainless steels in house-
hold and transportation uses.
Martensitic stainless steels AISI 410 and 416 are typical of the low carbon
martensitic varieties, with 1214%Cr and 0.15%C max. Here, (%Cr 17
 Engineering Alloys 205

%C) < 12.7. They respond to heat treatment, but due to the low carbon, the
martensite is not brittle. Typical as-quenched properties are:

Hardness, Rc Y.S., MPa T.S., MPa % Elongation

42 1200 1300 5

The cutlery grade AISI 440B is typical of high carbon martensitic stainless
steels, with 17%Cr and 0.75%C. The as-quenched hardness is 60 Rc. This grade
is used for cutlery items, surgical instruments and high quality ball bearings.
Austenitic stainless steels Austenitic stainless steels of the 300 series contain
nickel as the austenite stabilizing element. When nickel is partly or fully replaced
by the cheaper manganese, we have the 200 series.
In the 300 series, the most commonly used type is 304, with 18%Cr, 8%Ni,
and 0.08%C max. The significance of the carbon content in relation to inter-
granular corrosion was discussed in Sec. 6.8. The low carbon variety with
0.03%C max. is much less susceptible to intergranular corrosion. Addition
of 2%Mo gives the 316 variety, which has better resistance to pitting and
sulphuric acid.
The formability of the 18/8 austenitic stainless steels is excellent. The work
hardening rate is high and is attributed to the very low stacking fault energy of
0.002 J m2. The uniform elongation is also large. Typical tensile properties of
304 stainless steel are:
Y.S., MPa T.S., MPa % Elongation
240 600 65

By moderate working, the yield strength can be increased up to 1000 MPa.

It is used in a very wide variety of applications in the chemical industry and for
household and sanitary fittings.
Precipitation-hardened or controlled-transformation stainless steels contain,
in addition to the basic composition, Mo, Nb, Ti or Al. Precipitates such as
Ni3Ti form on ageing and increase the strength. 17-7PH grade, for example,
contains 1.15%Al in addition to 17%Cr and 7%Ni.

7.8 HIGH TEMPERATURE ALLOYS

The properties required in a high temperature alloy are:
1. Good oxidation resistance.
2. Adequate creep strength.
3. Microstructural stability, i.e. no deterioration of microstructure and
properties with time at the service temperature.
4. Depending on the application, other properties such as corrosion
resistance, fatigue strength and impact toughness may be required. For
206 Physical Metallurgy

example, turbine blades undergo stress reversals and must have sufficient
fatigue strength at the operating temperature.
5. The alloy must be capable of easy fabrication and should not undergo
large property changes on heating and cooling. (Jet engine parts
repeatedly cool down and heat up, as the engine is shut off and restarted.)
A number of alloys which can serve at high temperatures have already been
introduced in earlier sections and chapters. Here, only a short summary (includ-
ing such alloys as are not discussed up to now) is given.
In low alloy steel, molybdenum and vanadium are the two alloying ele-
ments which significantly improve the creep resistance. 0.5%Mo steel is used
for pressure vessels and superheater tubes for use up to 450C. The service
temperature can be increased by about 100C by the addition of 12%Cr, which
improves the resistance to graphitization.
Stainless steels possess many of the properties required of high temperature
alloys. Martensitic stainless steels have adequate creep strength and good
corrosion and oxidation resistance up to 650C. For higher temperatures, the
excellent oxidation resistance of ferritic stainless steels such as 430 and 446 can
be used. The creep strength and corrosion resistance of austenitic stainless steels
are the best. They are used up to 900C in applications such as furnace linings
and exhaust systems.
Superalloys For superior creep strength in sophisticated aircraft applications,
superalloys are used. There are three main types; iron-nickel-base, nickel-base
and cobalt-base superalloys. In the first two, the major creep strengthening
constituent is the intermetallic compound Ni3(Ti,Al), recall Sec. 5.4. An additional
strengthening effect arises from the precipitation of carbides, e.g. Cr23C6. In the
iron-nickel superalloys, some typical names are discaloy and incoloy, the
compositions being 15%Cr, 20 40%Ni, and smaller concentrations of
molybdenum, titanium and aluminium and balance iron. These are the least
expensive among superalloys. Nickel-base superalloys have the best creep strength
and some of the names are nimonics, inconel, and waspalloy. A typical
composition of a nimonic alloy is 0.2%C, 10%Cr, 20%Co, 5%Mo, 5%Al,
1.3%Ti and balance nickel. The creep resistance of cobalt-base superalloys lies
between the above two types. A typical composition is 25%Cr, 10%Ni, 8%W,
2%Fe, 0.5%C and balance cobalt. Here, the major strengthening comes from
tungsten and chromium carbide precipitates. Vitallium is a precision cast cobalt-
base superalloy. Figure 7.11 compares the creep strength of a number of high
temperature alloys discussed above.
Superalloys rapidly lose their strength above 1000C. Special alloys for use
at higher temperatures based on the refractory metals of molybdenum, niobium,
tantalum and tungsten are being developed. Ceramic materials such as silicon
nitride have a very high melting point and are being explored for high temperature
use. They have no problem of oxidation but their brittleness is the major obstacle
in development. Cermets are refractory ceramics such as Al2O3 and TiC embedded
 Engineering Alloys 207

in a Ni-Cr metal matrix and have good high temperature properties. Recall that
cermets are also used as cutting tools as discussed in Sec. 7.6.

Figure 7.11 The creep strength as a function of temperature for 0.01%

creep strain per 1000 hr for a number of alloys.

7.9 CAST IRONS

White versus gray cast irons Alloys of iron and carbon with more than
2.11%C are called cast irons. Referring to the Fe-C phase diagram in Fig. 2.10,
we note that iron-carbon alloys with more than 2.11%C undergo the eutectic
reaction at 1146C, whereas compositions with less carbon (steels) do not. The
eutectic mixture is known as ledeburite. Compositions to the left of the eutectic
composition (4.3%C) contain proeutectic austenite in addition to ledeburite just
below the eutectic temperature. The solubility of carbon in austenite decreases
from 2.1% at 1146C to 0.77% at 727C. The excess carbon precipitates as
cementite both from the proeutectic and the eutectic austenite. On cooling
through the eutectoid temperature (727C), the austenite decomposes to pearlite.
The microstructure of a hypoeutectic cast iron shown in Fig. 7.12 consists of
cementite (white regions) and pearlite (dark regions). In the above description,
the carbon is present in the combined form and the alloy is called white cast
iron, as the fracture surface of such irons looks white. White cast iron has
excellent wear resistance, but it is very brittle. It finds uses as balls for grinding
mills, liners for cement mixers and rolls for paper manufacture.
In the Fe-C phase diagram (Fig. 2.10), in addition to the cementite
equilibrium, some important graphite lines are also shown as dotted lines. The
graphite eutectic line is at 1152C, six degrees above the cementite line. When
carbon is present in the free form as graphite, the fracture looks gray (also spelt
as grey) and the alloy is called gray cast iron. Graphite has a lower free
energy than cementite. Yet, cementite often forms, because only 6.67%C has
to segregate to form cementite, whereas 100% segregation is required to
nucleate graphite.
208 Physical Metallurgy

Figure 7.12 The microstructure of white cast iron.

Factors that promote the formation of graphite are:

1. Slow cooling: thick cross-sections or castings in sand moulds tend to
have graphite, as the cooling rate is slow. Chill castings (in metal moulds)
and thin cross-sections tend to have cementite. This effect can also be
seen in the fracture appearance across the cross-section varying from
white at the surface to gray inside. The transition region has the mottled
appearance.
2. Among the alloying elements which influence graphitization most is
silicon. Silicon shifts the graphite eutectic line upwards, so that the
temperature interval between the graphite line and the cementite line
increases from 6C at 0%Si to 35C at 2%Si. So, during cooling from
the liquid state, a larger degree of undercooling is possible with a greater
chance to form graphite, before cementite formation becomes feasible.
3. Increasing carbon in the melt tends to promote graphite formation.
4. The effect of alloying elements other than silicon on graphitization is
described in terms of the silicon equivalent:
Si equivalent = % Si + 3(%C) + %P + 0.3(%Ni) + 0.3(%Cu)
+ 0.5(%Al) 0.25(%Mn) 0.35(% Mo) 1.2(%Cr)
Noncarbide forming elements Cu, Ni and Al have positive factors, i.e. they
promote the formation of graphite. Carbide forming elements Mn, Cr and Mo
have negative factors.
Gray cast ironmicrostructure If all the carbon is in the form of graphite,
we have ferritic gray iron, where the graphite flakes are embedded in a matrix
of ferrite, Fig. 7.13. Flakes are in fact curved plates often interconnected in
three dimensions. If only a part of the carbon is in the form of graphite, we have
the pearlitic gray iron, with pearlite as the matrix. Graphite flakes are sharp at
their tips and act like internal cracks or stress raisers. For this reason, gray iron
is brittle and shows only about 0.5% elongation in tension. As the cracks do not
propagate under compressive loads, the strength of gray iron is more in com-
pression than in tension.
 Engineering Alloys 209

Figure 7.13 The microstructure of gray cast iron consists of graphite

flakes in a matrix of ferrite (or pearlite).

The standards set up by AFA-ASTM (American Foundrymens Association

and the American Society for Testing and Materials) describe different mor-
phologies of graphite. Type A refers to the random orientation of uniformly
sized flakes. Type B is rosette groupings (flakes arranged radially like petals of
a rose). Type C is flakes of mixed sizes with a random orientation. D is
interdendritic flakes randomly oriented and E is the same preferentially ori-
ented. Figure 7.14 shows schematically the graphite flake sizes from 1 to 6.

Figure 7.14 The graphite flake-size chart.

210 Physical Metallurgy

The properties of gray iron depend on the morphology and size of the
graphite flakes and the matrix. The most desirable microstructure is fine and
randomly oriented flakes. Large flakes act as long internal cracks and hence
lower the strength. A pearlitic matrix has a greater strength than a ferritic
matrix. According to strength, class of gray iron is specified as 20, 30, 40, 60,
etc. For example, class 40 stands for a tensile strength of 40,000 psi (280 MPa).
Figure 7.15 shows the stress-strain curves for class 20 and class 40 gray irons
in tension and in compression.

Figure 7.15 The stress-strain curves in tension and compression

for Class 20 an Class 40 cast irons.

Gray cast ironproperties and uses Gray cast iron has certain unique
properties:
1. Cast iron is cheaper than steel, as the temperature to be attained for
making it is several hundred degrees lower than that for casting steel.
Also, control of impurities is not critical here, as in steel making.
 Engineering Alloys 211

2. It has excellent fluidity and takes mould impressions quite well.

3. The wear resistance of gray iron is very good, as graphite flakes act as
lubricant. (White cast iron has also good wear resistance but for an
entirely different reason that cementite is very hard.)
4. Gray iron is easy to machine, as chip formation is promoted by the
graphite flakes. In addition, the flakes serve as a lubricant for the cutting
tool. (White cast iron cannot be machined.)
5. The damping capacity (ability to damp vibrations) of gray iron is superior
to that of steel.
Gray iron has manifold uses. The good damping capacity and the high
compressive strength make it suitable as a base for erection of machinery. Ease
of machining, good wear resistance and damping capacity are utilized in
applications such as locomotive and internal combustion engine cylinder blocks
and heads. Ease of casting and low cost make it suitable for flywheels and
counterweights for lifts.
Alloying elements added to gray iron improve the properties, but the cost
tends to increase. Nihard iron with 4%Ni and 1.5%Cr has excellent wear resis-
tance due to the presence of martensite and alloy carbide. Niresist, with 20%Ni
and 2%Cr has graphite and alloy carbide particles embedded in an austenitic
matrix. It has excellent corrosion resistance and heat resisting properties and is
used for handling alkalis at high temperatures.
The poor tensile ductility of gray iron results from the flaky form of graph-
ite. If the graphite particles are more spherical in shape, the stress concentration
is reduced and the tensile ductility improves. Two such modifications of gray
iron with spheroidal shaped graphite particles are described below.
Malleable cast iron White cast iron, where the carbon is in the form of
cementite, is heat treated to produce malleable cast iron. A typical composition
of the white iron used is 2.5%C and 1%Si. The silicon equivalent is low enough
to prevent the formation of graphite flakes during casting, but adequate to keep
the time of subsequent annealing for graphite precipitation within reasonable
limits. The castings are packed in boxes with sand used as a filler to exclude
air. The sealed boxes are given a prolonged annealing for several days at 900
950C and then cooled very slowly, the cooling time down to 600C also being
a few days. Cementite decomposes during this heat treatment to the more stable
form (graphite). The free carbon precipitates in the form of spheroidal particles
called temper carbon, Fig. 7.16. Some carbon is lost by combination with the
residual oxygen in the box. In white heart malleable iron, most of the carbon
is lost by oxidation, leaving behind an essentially ferrite matrix. In black heart
malleable, temper carbon is present.
212 Physical Metallurgy

Figure 7.16 Temper carbon is present in the form of approximately

spherical aggregates in a malleable cast iron.

Malleable cast iron has tensile strengths up to 700 MPa, with an elongation
of 10 15%. They are more expensive than gray irons because of the heat
treatment involved. They are used in applications such as automobile crank-
shafts, chain links and brackets.
Spheroidal graphite (SG) iron Spheroidal graphite iron also known as nodular
iron or ductile iron is produced by adding small quantities of a modifier such
as magnesium or cerium to the melt. Good control of the melt charge is essential
to keep residual and impurity elements within specified limits. The basic
composition of the melt is 3 4%C and 2.5%Si. This fairly-high silicon equivalent
promotes graphitization during solidification. The modifier has the effect of
making the growth rate of graphite to be approximately the same in all directions,
so that a spherical shape results, Fig. 7.17. Note that the graphite nodules here
are truly spherical, in contrast to the irregular aggregate of temper carbon in
Fig. 7.16. Magnesium is almost universally used as the modifier. Due to its
lightness, it is added in the form of Ni-Mg alloy (note that nickel is a graphitizer).
In the pressure ladle method, the alloy is added in a capsule and the ladle is
sealed with a cover to prevent the magnesium vapour from escaping. The
residual magnesium in the casting is about 0.05%.

Figure 7.17 Spherical nodules of graphite in SG iron.

 Engineering Alloys 213

Nodular iron is a major engineering material, as it combines the engineering

advantages of steel with the processing economics of cast iron. The tensile
strength of nodular iron ranges from 400 to 700 MPa, with elongation between
1018%. Agricultural implements, industrial fan hubs, coke oven doors,
crankshafts and gears are some of the applications of nodular iron.
Cast iron exposed to elevated temperatures may deteriorate through a process
called growth. Oxidation of iron occurs through gases penetrating into the
graphite cavities. Any cementite present decomposes into graphite. Volume
expansion due to these produces fine surface cracks called crazes on the cast
iron, which eventually fails. Among alloying elements, chromium is particularly
effective in reducing growth.

7.10 THE LIGHT ALLOYS

Introduction The light metals are Be, Mg, Al and Ti. Their specific gravities
and Youngs moduli Y are compared with those of other common metals below:

Metal Be Mg Al Ti Fe Cu Pb W
Specific gravity 1.86 1.74 2.71 4.5 7.86 8.9 11.3 19.3
Y, GN m2 289 44 71 106 210 124 15.7 396

In the properly treated condition, the light alloys possess favourable strength-
to-weight and stiffness-to-weight ratios as shown below:

Alloy T.S./specific gravity, Y/specific gravity,

MN m2 GN m2
High strength steel 170 26.5
Duralumin 200 25.9
Mg alloy 190 24.9
Ti alloy 280 26.5
Beryllium 170 102.0

In the following table, the mass of metal required for the same stiffness of
beams of equal length is compared using steel as reference:

Metal Steel Ti Al Mg Be
Comparative mass 1 0.81 0.59 0.48 0.20

In addition to weight saving seen in the above table, with a smaller mass,
the inertial forces are less in reciprocating parts such as connecting rods and
pistons. Also, the elastic buckling of slender columns or wrinkling of thin sheets
bearing loads is less of a problem with light metals and their alloys, as they
have a larger volume (and thickness) for the same stiffness.
214 Physical Metallurgy

The other desirable properties of the light metals are listed below:
High corrosion resistance

Aluminium High electrical conductivity
High thermal conductivity
Magnesium Outstanding machinability
Titanium Outstanding corrosion resistance
The major disadvantage of the light metals is the large energy requirement
for production. For aluminium, the energy required is 75,000 kWh per tonne
from ore to primary metal, which is five times that for steel. The energy require-
ments for magnesium and titanium are even larger. However, aluminium scrap
requires only 5% of the energy required for the production of the primary metal
from the ore. So, recycling is an important process.
Classification of aluminium alloys The AAA (Aluminium Association of
America) classification for wrought aluminium alloys has been adopted by the
International Alloy Development System (IADS) and is now accepted by most
countries. The classification is based on a four-digit system.
1xxx Commercial purity and high purity aluminium
2xxx Al-Cu alloys
3xxx Al-Mn alloys
4xxx Al-Si alloys
5xxx Al-Mg alloys
6xxx Al-Mg-Si alloys
7xxx Al-Zn alloys
In addition to the four digits, specific letters are added to denote the thermal/
mechanical treatment given:
O annealed
H work hardened only
H12 quarter-hard; H14 half-hard; H18 full-hard; H19 > 80%
reduction
H2 and H3 work hardened and recovery annealed
T age hardened
T1 naturally aged after hot working
T4 solution treated, quenched and naturally aged
T6 solution treated, quenched and artificially aged
T8 same as T6, except cold worked before ageing
Non heat-treatable aluminium alloys Starting with aluminium metal, we
note that pure aluminium (1350H16 in Table 7.1) is used in the electrical
industry. Even though the electrical conductivity of aluminium is only 60% of
that of copper, for equal weight, the conductivity is 200% of copper, as aluminium
 Engineering Alloys 215

of the same weight has more than three times the volume of copper. After the
electrical industry, the largest use for aluminium is in the packaging industry.
Aluminium is attractive for this purpose, due to its good corrosion resistance,
high thermal conductivity, and impenetrability to oxygen, moisture and micro-
organisms. In addition, unalloyed aluminium of the 1xxx series is used in chemical
process equipment and for architectural purposes.
In the 3xxx series, Al-Mn and Al-Mn-Mg alloys have moderate strength
and are used for beverage cans, cooking utensils, buildings and roofing sheets.
The composition and properties of 3003H18 are listed in Table 7.1.
4xxx Al-Si alloys are used only as welding and brazing electrodes.
The 5xxx series Al-Mg alloys are widely used in welded applications as
transportation structural plate for truck bodies and large tanks carrying petrol,
milk or grain, and for pressure vessels at cryogenic temperatures. Their high
corrosion resistance makes them suitable for boat hulls and superstructures of
ocean-going vessels. As an example, the properties of 5052H38 are listed in
Table 7.1.
Heat-treatable aluminium alloys The following aluminium alloys can be
given an age hardening treatment.
In the 2xxx series of Al-Cu alloys, CuAl2 and the associated transition
phases form the precipitates, recall Sec. 3.8. Referring to Table 7.1, 2219T62
(Al-6.3Cu) is produced as sheets, plates and extrusions for structural purposes.
2024T6 is used as sheets, plates and forgings for aircraft structures. To improve
the corrosion resistance, Al-Cu alloys (and also Al-Zn-Mg alloys) are sand-
wiched between two pure aluminium sheets and rolled to produce the composite
Alclad.
Nickel improves the creep resistance needed for supersonic aircraft. In
2618T61 alloy (Rolls Royce designation RR58), copper and magnesium con-
tribute to strengthening by age hardening, whereas nickel and iron form the
intermetallic compound FeNiAl9, which causes dispersion hardening.
6xxx series Al-Mg-Si alloys are used as medium strength structural alloys
having the advantage of good weldability, corrosion resistance and immunity to
stress corrosion cracking. 6063T6 in Table 7.1 is a commonly used alloy of this
series. It responds well to colour anodizing and is used in decorative and archi-
tectural applications.
7xxx series offers the greatest potential for age hardening. Note the rapid
decrease in the solubility of zinc with decreasing temperature in the Al-Zn
equilibrium diagram, Fig. 7.18. In addition to the base composition of Al-Zn-
Mg, many of the alloys contain copper, which reduces the susceptibility to SCC.
The best known is 7075T76 (see Table 7.1 for composition and properties),
which is a major aircraft structural material, the British equivalent being
DTD 683.
 Table 7.1 Aluminium Alloys

Alloy Composition 0.2%

Designation Si Fe Cu Mn Mg Zn Cr Ti Others Temper P.S. T.S. %E Applications
MPa MPa

I. Non heat-treatable Alloys

1100 1.00 Si+Fe 0.1 0.05 0.1 H18 152 165 5 Sheets, plates,
216 Physical Metallurgy

tubes
1199 0.006 0.006 0.006 0.002 0.006 0.006 0.002 O 10 45 50 Capacitor foils
1350 0.1 0.4 0.05 0.01 0.05 0.01 H16 110 124 4 Conductors, ACSR
cables, bus bars
3003 0.6 0.7 0.15 1.2 0.1 H18 185 200 7 Rigid containers,
cooking utensils
5052 0.25 0.4 0.1 0.1 2.5 0.1 0.25 H38 255 290 7 Marine fittings, air-
craft fuel tanks
II. Age Hardenable Alloys
2024 0.05 0.5 4.4 0.6 1.5 0.25 0.1 0.15 T6 395 475 10 Aircraft structures,
truck wheels
2219 0.2 0.3 6.3 0.3 0.02 0.1 0.05 0.1V T62 290 415 10 Welded aircraft
parts
2618 0.15 1.1 2.4 1.5 0.1 0.07 1.1Ni T61 310 400 4 Forgings, aircraft
pistons
6063 0.4 0.35 0.1 0.1 0.7 0.1 0.1 0.1 T6 215 240 12 Furniture, pipes,
extrusions
7075 0.4 0.5 1.6 0.3 2.5 5.6 0.23 0.2 T76 470 540 12 High strength
aircraft structures
 Engineering Alloys 217

Figure 7.18 The Al-Zn phase diagram shows the rapid decrease of the
solubility of zinc in aluminium with decreasing temperature.

Cast aluminium alloys Aluminium is a common foundry metal. Apart from

light weight, the low melting temperature is an advantage. Al-12%Si alloy
(AAA 413) is among the most important cast alloys. The advantage is the high
fluidity imparted by silicon. The Al-Si phase diagram is a simple eutectic sys-
tem, with the eutectic temperature at 577C and the eutectic composition at
12.6%Si. Other advantages of Al-Si castings are good weldability, corrosion
resistance and wear resistance. The lower thermal expansion coefficient of Al-
Si alloys makes them more suitable than Al for use as pistons and cylinder
heads in cast iron engine cylinders.
Rapid cooling of the Al-Si alloy refines the eutectic structure, probably by
a transition from the plate-like to the rod-like form of silicon in the eutectic
mixture. Rod-like units are smaller in size compared to plates. The refinement
can also be achieved by the process known as modification, in which small
quantities of metallic sodium or sodium salts are added to the melt before
casting. Sodium either depresses the eutectic temperature or slows down the
growth rate of silicon by poisoning, causing the transition in morphology and
the refinement.
Excessive sodium is known to cause the condition known as
overmodification, where the coarser eutectic structure tends to reappear. More
recently, strontium has been used in place of sodium, as it does not vaporize or
oxidize as sodium does. Overmodification is less of a problem with strontium.
Magnesium and its alloys The production of magnesium is an energy inten-
sive process. Electrolytic reduction of MgCl2 is the standard method. More than
50% of magnesium produced is used to alloy with aluminium and as an additive
for making SG iron. Oxidation and fire hazard are problems with magnesium
218 Physical Metallurgy

melts. The metal melts at 650C. The rate of oxidation increases rapidly with
temperature and, at 850C, the melt spontaneously bursts into flames.
A well-known example of a magnesium alloy is AZ81, with 8%Al and
1%Zn. This alloy gives tough, leak-tight castings and is used for light-weight
crank cases in some cars. With a small quantity of beryllium added, the alloy
is used for pressure die castings.
Titanium and its alloys The density of titanium is midway between that of
aluminium and steel. The critical need for new materials with high strength-to-
weight ratio in aerospace industries provided the stimulus for the development
of titanium alloys. Titanium with a melting point of 1678C has adequate creep
strength to resist aerodynamic heating. Over 80% of titanium produced is used
in the aerospace industry.
Titanium is a highly reactive metal, but due to the formation of a stable,
self-healing passive film, it exhibits outstanding corrosion resistance in a wide
variety of environments. The film resists attack by chloride ions. Its resistance
to sea water is the best among structural alloys. Uncracked titanium is immune
to SCC. The outstanding corrosion resistance of titanium is increasingly utilized
in the chemical industries and in the medical profession (as body implants).
At 882C, titanium changes its crystal structure from HCP (a) to BCC (b)
on heating. Alloying elements with electron/atom ratio of less than 4 stabilize
the a phase, whereas elements with ratio more than 4 stabilize the b phase.
Zirconium, tin and silicon with the ratio of 4 are neutral. Most two-phase (a + b)
alloys contain alloying elements to stabilize and strengthen the a phase, together
with 4 to 6% of b stabilizing elements, which allow substantial amounts of b
phase to be retained on quenching. Strength is enhanced by ageing the quenched
alloy. Tensile strengths in the range of 1400 MPa can be realized. Ti-6%
Al-4%V is the most important commercial alloy. Aluminium, with three valence
eletrons per atom, stabilizes the a phase and vanadium with 5 electrons stabilizes
b. The principal use of this alloy is as forged fan blades in aircraft jet engines.
It is also used for airframe structural components in supersonic aircraft requiring
strength up to 300C.

7.11 COPPER AND ITS ALLOYS

Among the common metals, copper is the best conductor of heat and electricity,
next only to the noble metals silver and gold.
The common impurity in copper, oxygen, is virtually insoluble and forms
a eutectic mixture of Cu-Cu2O. It helps in removing the dissolved impurities
from copper and in indirectly increasing the conductivity. The conductivity of
tough pitch copper with 0.020.05% oxygen is 101% of IACS (International
Annealed Copper Standard). In contrast, OFHC (oxygen free high conductivity)
copper has a conductivity of 95% of IACS. Phosphorus is often used to deoxidize
copper, which leaves some residual phosphorus in solution. As little as 0.04%P
 Engineering Alloys 219

lowers the electrical conductivity of copper by 15%. Among the solutes, cadmium
is the least harmful and can be used for solid solution strengthening, without
impairing the conductivity much.
The high thermal conductivity of copper makes it useful in radiators and for
circulating the coolant fluid in refrigerators and airconditioners. The good cor-
rosion resistance of copper and the nontoxic nature of copper ions make it
suitable for pipes to carry foodstuffs and beverages.
The brasses Brasses are alloys of copper and zinc. The Cu-Zn phase diagram
is discussed under Sec. 2.7 and should be recalled. The copper rich end of the
diagram is reproduced in Fig. 7.19. The Cu-Zn alloys are classified as a-brasses
and (a + b) brasses. a-brasses containing up to 20%Zn are reddish in colour.

Figure 7.19 The strength and ductility of brasses as a function of the zinc content.
220 Physical Metallurgy

Above 20%Zn, the brass is yellow in colour. Yellow a-brasses have zinc content
between 20 and 36%. (a + b) brasses, also yellow in colour, have compositions
between 36 and 46%Zn.
The improvement in mechanical properties of copper with the addition of
zinc is illustrated in Fig. 7.19, both the tensile strength and the elongation
increase with increasing zinc content almost up to the limit of a stability. This
is attributed to a lowering of the stacking fault energy of copper by zinc and
the consequent increase in work hardening rate and uniform elongation. Beyond
36%Zn, in the (a + b) region, the tensile strength continues to increase, but the
ductility decreases. As the b phase content exceeds 50%, the strength also
decreases.
a-brass has excellent cold formability. The 70/30 brass containing 70%Cu
is a single a solid solution, known as cartridge brass. It has excellent deep
drawing property and is used for making cartridge cases, radiator fins and lamp
fixtures. Cold worked a-brass is susceptible to season cracking as discussed
under Sec. 6.9. This susceptibility is removed by a stress-relief anneal at about
250C.
Admirality brass with 29%Zn and 1%Sn has good corrosion resistance and
is used for condenser tubes and heat exchangers in steam power plants. How-
ever, it is susceptible to dezincification in marine environments, where the zinc
corrodes preferentially leaving behind a porous mass of copper. Aluminium
brass with 22%Zn and 2%Al and a trace of arsenic (0.04%) is not susceptible
to dezincification and has largely replaced admirality brass for marine uses.
Red brasses are not susceptible to season cracking or dezincification. Gilding
metal with 5%Zn is used in making coins, medals and emblems. Commercial
bronze with 10%Zn is used for costume jewellery. Leaded brasses contain 0.5
3% lead and have better machinability.
The 60/40 (a + b) brass known as muntz metal is cheaper and stronger than
a-brass, but with limited ductility. In contrast to the excellent cold formability
of a-brass, (a + b) brasses can only be hot worked. The ordered b phase present
at room temperature is brittle. On heating, however, the order-disorder reaction
occurs above 465C, see Fig. 7.19, the b phase becomes ductile and hot working
can be done readily. (a + b) alloys become a single phase b solid solution at
elevated temperatures.
With the addition of 1%Sn to (a + b) brass, the corrosion resistance to
marine environments is vastly improved. Brass with 39%Zn and 1%Sn, called
naval brass, is used in marine hardware such as properller shafts.
The bronzes The name bronze was originally used to denote copper-tin alloys.
When copper is alloyed with other metals such as Al, Si and Be, the term
bronze is still used. For example, Cu-Al alloys are called aluminium bronzes
and Cu-Si alloys are known as silicon bronzes. The Cu-Sn bronzes are
commercially known as phosphor bronzes, as some residual phosphorus used in
deoxidation is always present. As in brasses, both the tensile strength and
elongation of the solid solutions formed by different solutes increase with the
 Engineering Alloys 221

increasing concentration of the solute in the single phase region. In the annealed
condition, the tensile strength is about 400 MPa and the elongation is over 60%
at 10%Sn, or 3%Si or 5%Al as the solute in copper.
The copper rich end of the Cu-Sn equilibrium diagram is shown in Fig. 7.20.
At 798C, a peritectic reaction occurs and the b phase forms. At 586C, the
b phase undergoes a eutectoid reaction to form a mixture (a + g). At a still

Figure 7.20 The copper-rich end of the Cu-Sn phase diagram.

lower temperature of 520C, the g phase undergoes another eutectoid reaction

to form (a + d) phase. The d phase decomposes to (a + e) phases by a third
eutectoid reaction at 350C. However, this reaction is so sluggish that the e
phase is never seen in the microstructure. The solubility of tin in copper is about
17% at 520C and decreases considerably on cooling to room temperature. Here
again, at normal cooling rates, the precipitation of d is not complete, so that
phosphor bronzes with less than 10%Sn are usually single phase solid solutions
at room temperature.
In the Cu-Si system, the solubility of silicon in copper is 5.3% at 852C and
decreases with decreasing temperature. Here also, the precipitation of the excess
silicon does not occur at normal cooling rates so that the single a phase is found
in compositions up to 5%Si.
222 Physical Metallurgy

Figure 7.21 The copper-rich end of the Cu-Al phase diagram.

The copper rich end of the Cu-Al equilibrium diagram is shown in Fig. 7.21.
At 500C, copper dissolves up to 9.5%Al. The Cu-11.5%Al alloy undergoes an
eutectoid reaction at 566C:
Reaction : b a + g2
Structure : BCC FCC complex cubic
Al% : 11.8 9.4 15.6
On furnace cooling this composition, a lamellar structure resembling
pearlite forms. On quenching, the b phase transforms martensitically to b
without a change in composition. The phase can be tempered in the same
manner as in a steel. In this way, tensile strengths up to 700 MPa can be
obtained. Heat treated aluminium bronzes are used as gears, dies, bearings
and pump parts.
 Engineering Alloys 223

In the Cu-Be system, the solubility of beryllium in copper is 2.3% at

800C and decreases to less than 0.2% at room temperature, Fig. 7.22. This
makes the Cu-Be alloys age hardenable. The precipitation process involves
several transition phases shown in Fig. 7.22 and is quite similar to that in an
Al-Cu alloy discussed in Sec. 3.8. A high yield strength of 1300 MPa can be
obtained by optimum ageing. This high strength combined with the lower
stiffness of copper as compared to steel results in a large sy/Y ratio and a high
resilience. Cu-Be age hardened alloys are used for nonmagnetic springs and
nonsparking tools.

Figure 7.22 The copper-rich end of the Cu-Be phase diagram.

7.12 BEARING ALLOYS

Despite the development of ball and roller bearings, the conventional stationary
bearing is still in use. The properties that are required of such a bearing are:
1. The friction between the bearing and the rotating part should be as
small as possible.
2. The wear of the contacting metals should be as low as possible.
3. The bearing must be capable of withstanding pressures that arise from
small misalignments.
4. It must contain a relatively low melting component, which will soften
or melt and prevent seizures.
Usually a thin film of lubricating oil separates the bearing and the rotating
part. This ensures that requirements 1 and 2 above are satisfied. The pressures
224 Physical Metallurgy

that a bearing can withstand depend on the strength and the melting point of the
bearing alloy. Soft and low-melting alloys can withstand comparatively low
pressures only. Bimetal bearings use a strong bearing back (e.g. steel) to which
a softer, relatively thin layer of a bearing alloy is metallurgically bonded. By
this combination, the excellent surface and corrosion-resistant properties of a
weak bearing alloy are combined with the high load-bearing capacity of the
back-up material.
The microstructure of a bearing alloy usually consists of a soft matrix in
which hard particles are embedded. After some use, the soft matrix wears off
more rapidly and the hard particles stand out in relief. The lubricating oil is
retained in the channels between the hard particles. Soft pockets embedded in
a strong matrix are obtained in some bearing alloys like leaded bronze.
The alloys known as babbits or white metals are either lead-based or tin-
based. A typical lead-based white metal consists of 80%Pb, 10%Sb and 10%Sn.
The microstructure contains hard cube-shaped particles (cuboids) of SbSn in a
soft lead-rich matrix. The pressure and speed that the bearing will withstand
increases with the addition of a few percent of arsenic. A typical composition
of a tin-base babbit is 90%Sn, 5%Sb and 5%Cu. In addition of SbSn, needle-
shaped CuSn (or Cu6Sn5) particles are also present in the microstructure. Tin-
base babbits are more expensive but have better corrosion and wear resistance.
For high bearing loads, bronze bearings are specified. In the widely used
10%Sn phosphor bronze, the hard (a + d) eutectoid mixture is embedded in the
softer a matrix. Leaded bronzes contain in addition 510% lead for better
lubricating action. The low-melting lead particles also prevent seizure. Aluminium
and silver based bearings are also known.

REFERENCES

1. Properties and Selection: Irons, Steels and High Performance Alloys,

Metals Handbook, 10th Edition, Vol. 1 (1990); Properties and Selection:
Nonferrous Alloys and Special Purpose Materials, Vol 2 (1990).
2. Metals and Alloys in the Unified Numbering System, 8th Edition, ASTM-
SAE Joint Publication, Warrendale, PA (1999).
3. Worldwide guide to Equivalent Irons and Steels, American Society for
Metals, Metals Park, Ohio (1979).
4. Lllewellyn D.T., and R.C. Hudd, Steels: Metallurgy and Applications,
Butterworth-Heinemann, Oxford (1998)
5. Hoyle, G., High Speed Steels, Butterworths, London (1988).
 Engineering Alloys 225

QUESTIONS

1. Write down the compositions corresponding to the following specifica-

tions of steels: T105, AISI1118, SAE8640, En23, En19, AISI 321.
2. What are the advantages of the residual manganese in a carbon steel?
3. Compare in tabular form the properties and uses of conventional low
carbon steel, mild steel, dual phase steel and microalloyed steel.
4. Distinguish between ausforming and austempering in terms of processes,
microstructures and properties of the product.
5. What are the special problems encountered in heat treating high speed
steel and how are these solved?
6. Discuss the significance of the compositions of the melt used for making
malleable cast iron and nodular cast iron.
7. What is the difference in the heat treatment schedule for production of
pearlitic malleable and ferritic malleable cast iron?
8. For the same weight, calculate the relative stiffness of steel, titanium,
aluminium and magnesium, taking the stiffness of steel to be unity.
Answer: 1:1.53:2.85:4.29.
9. Comment on the relative weldability of age hardened aluminium alloys
and aluminium alloys that are not heat treated.
10. Describe the composition and heat treatment of (a + b) titanium alloys for
optimum mechanical properties.
11. Distinguish between season cracking and dezincification.
12. Describe the role of hard particles, soft matrix and insoluble metals in the
functioning of a bearing alloy.
 INDEX

Abnormal eutectic, 34 British specifications, 186-187

Abrasion resistance, 107 Bronzes, 220-223
Activation polarization, 168 Burgers vector, 16-17
Admirality brass, 220 Burning, 42
Ag-Cu phase diagram, 33-34
Agehardening, 92-95
Ageing, 94 Calomel electrode, 164
Aircraft structurals, 215-216 Capping, 78
AISI specifications, 186-187 Carbide stability, 107
Alclad, 173, 215 Carbonitriding, 125
Al-Cu phase diagram, 92 Carburizing, 123-124, 158
Alloy carbides, 88, 199 Cartridge brass, 220
Alloying elements in steel, 46-51, 186 Case hardening, 122-125
Alpha brass, 220 Cast irons, 207-213
Aluminium, 214-215 Cathode, 162-163
Aluminium alloys, 214-217 Cathodic protection, 173-175
Aluminium brass, 220 Caustic embrittlement, 180
Aluminium bronzes, 222 Continuous cooling transformation (CCT)
Al-Zn phase diagram, 215, 217 diagrams, 104-106
Annealing, 109-111 C-curve, 79, 103
Anode, 162 Cell structure, 74
Anodic protection, 170 Cementite, 42-44
Antiphase domain boundary, 142 Cermets, 203, 206-207
Arrt, 44 Chauffage, 44
Artificial ageing, 94 Chemical metallurgy, 1
ASTM grain size number, 20 Chill casting, 71
Ausformed steels, 198 Chill zone, 71-72
Austempering, 120-122 Chinese script, 35
Austenite, 43 Chip formation, 195
Austenite stabilizers, 47-48, 50 Chromium, 47, 107
Austenitic stainless steels, 205 Climb of dislocations, 18
Austenitization, 107-109 Close packing geometry, 6-7
Clustering, 87, 93
Coarsening, 88, 147
Babbits, 224 Coatings, 175-177
Bainitic transformations, 83-84 Coble creep, 146
Banding, 192 Coherent interface, 21, 147
Bearing alloys, 223-224 Cold shortness, 187
Beryllium bronzes, 223 Cold working, 98, 137-138
Blue brittleness, 133 Cold work tools, 201-202
Body centred cubic (BCC) crystal, 2-3 Columnar growth, 72, 75
Boron in steel, 118-119 Common tangent rule, 38
Brasses, 219-220 Components, 30
Brine solution, 114 Composition tetrahedron, 57
227
228 Index

Concentration cells, 164 Edge dislocation, 16

Concentration polarization, 168 Electrical resistivity, 63
Configurational entropy, 12, 15 Electrode potentials, 164-166
Congruent melting, 32 Electron gun, 24-25
Constitutional supercooling, 73-74 Electron microscope, 24-26
Continuous casting, 78 Electron phases, 14
Coordination number, 3 Electron probe microanalyzer, 26
Copper, 218-219 Electroplating, 176
Copper alloys, 219-223 Elongation, 127, 129
Coring, 41 Enamel, 176
Corrosion fatigue, 158 Energizers, 123
Cottrell atmospheres, 22 Energy integral, 152
Crack opening displacement (COD), 152-153 Engineering metallurgy, 1
Creep, 143-146 Engineering strain, 127-128
Critical diameter, 113 Engineering stress, 127-128
Critical pitting potential, 171 Enveloping, 42
Critical sress intensity factor, 151 Equicohesive temperature, 146
Critical temperatures, 44-45 Eutectic composition, 33
Cross-slip of dislocations, 17-18, 130, 145 Eutectic phase diagram, 33-35
Cu-Al phase diagram, 222 Eutectic temperature, 33
Cu-Be phase diagram, 223 Eutectoid reaction, 36
Cu-Ni phase diagram, 31 Eutectoid steel, 45
Curie temperature, 45 Extrusions, 155
Cu-Sn phase diagram, 221 Exudation, 77
Cu-Zn phase diagram, 51-52, 219
Cyaniding, 124-125
Cyclic loading, 153 Face centred cubic (FCC) crystal, 3-4
Fatigue, 153-159
Fatigue fracture, 155-156
Damping capacity, 211 Fatigue limit, 154
Deaeration, 173 Fatigue strength, 154
Degrees of freedom, 30 Fe-C phase diagram, 42-43
Dendrites, 74-75 Ferrite, 43
Density, 3 Ferrite stabilizers, 47, 50-51
Dezincification, 220 Ferritic stainless steels, 204
Differential aeration cell, 163-164 Ficks first law, 66
Diffusion, 66-67 Ficks second law, 66
Diffusional creep, 146 Flame hardening, 122
Diffusionless transformation, 84-85 Flow stress, 132
Dilatometry, 62-63 Fracture, 147-148
Directional solidification, 75 FrankRead source, 136
Discaloy, 206 Free cutting steels, 195
Dislocations, 15-19 Free electrons, 1
Dispersion hardening, 147, 215 Free energy composition curves, 37-40
Divorced eutectic, 34 Full annealing, 109-110
Dual phase steels, 189-190
Ductile fracture, 147
Duplex stainless steels, 179 Galvanic cell, 161-163
Duralumin, 92, 173, 216 Galvanic series, 167
Dynamic recovery, 130 Galvanized iron, 176
Dynamic strain ageing, 133-134 Galvanizing, 176
 Index 229

Gamma loop, 47 Incoloy, 206

Gas carburizing, 123 Inconel, 206
Gilding metal, 220 Induction hardening, 122
Glide of dislocations, 17 Ingot structure, 71-72
G-P zones, 93 Inhibitors, 173
Grain boundary, 19-20 Intergranular corrosion, 177-179
Grain boundary diffusion, 67 Intergranular fracture, 180-181
Grain growth, 98 Interlamellar spacing, 82-83
Grain multiplication, 73 Intermediate phases, 13-14
Grain size, 20, 138-140 Intermetallic compounds, 13
Graphite, 43-44 Interstitialcy, 14-15
Graphite flakes, 208 Interstitial diffusion, 66-67
Gray cast iron, 208-211 Interstitial impurity, 15
Griffith criterion, 148-149 Interstitial solid solution, 10-11
Grossmans critical diameter, 113 Intrusions, 155
Guinier-Preston zones, 93 Invariant reaction, 35
Invariant temperature, 35
Inverse segregation, 77
Hair line cracks, 180 Isomorphous system, 31
Hall-Petch equation, 139 Isopleths, 54
Hardenability, 112-119 Isothermal section, 54, 58
Hardenability bands, 119 IS specifications, 185-186
Hardness of martensite, 86-87
HAZ, 191
Heterogeneous nucleation, 70 Jog, 18, 145
Hexagonal close packed (HCP) crystal, 4-5 Jominy distance, 116
High speed steel, 199-201 Jominy end quench test, 115-116
High strength low alloy steels, 197
High temperature alloys, 205-207
Hollomon equation, 131 Killed steels, 77-78
HollomonJaffee parameter, 88 KIc, 150-151
Homogeneous nucleation, 68-69
KIscc, 181-183
Homogenization, 76
Hookes law, 90
Hot dipping, 175, 176
Lamellar arrangement, 45
Hot shortness, 107, 187
Lamellar tearing, 192-193
Hot tearing, 77
Laser hardening, 122-123
Hot work tools, 202
Lath martensite, 86-87
HSLA steels, 197
Lattice parameter, 3
Hume Rotherys rules, 11
Laves phases, 14
Hypereutectic alloy, 34
Ledeburite, 44
Hypereutectoid steel, 46
Lever rule, 32
Hypoeutectic alloy, 34
Linseed oil, 176
Hypoeutectoid steel, 46
Liquid carburizing, 124
Hydrogen electrode, 164
Liquidus, 31
Hydrogen overvoltage, 168-169
Long range order, 12
Low alloy steels, 106, 197
Ideal critical diameter, 114-117 Low carbon steels, 188-191
Ideal solid solution, 12 Low cycle fatigue, 156
Immiscibility, 12 Lower bainite, 83
Impact transition temperature, 193-194 Lower yield point, 133
Inclusions, 70, 77, 192 Luders bands, 133
230 Index

Machinability, 195 Off-set yield strength, 128

Macrosegregation, 76-77 Optical microscope, 23
Macrostructure, 22-23 Orange peel effect, 189
Magnesium alloys, 217-218 Ordered solid solutions, 10
Magnetic lenses, 24 Orthorhombic crystal, 6
Malleable cast iron, 211-212 Overageing, 95
Manganese in steel, 187-188 Overheating, 42
Maraging steels, 198 Oxidation reaction, 162
Marquenching, 120
Martempering, 120-122
Martensitic stainless steels, 204-205 Pack carburizing, 123
Martensitic transformations, 84-86 Packing efficiency, 3-5
Mechanical metallurgy, 1 Paints, 176-177
Medium carbon steels, 195-196 Partial dislocations, 18-19
Melt spinning, 78-79 Passivation, 169-171
Metallic bonding, 1 Passivators, 171
Metallograph, 23 Passive-active cell, 171
Microalloyed steels, 140, 193-194 Patenting, 111
Microgalvanic cells, 163 Pearlite, 44
Microporosity, 77 Pearlitic transformations, 79-83
Microsegregation, 75 Peritectic phase diagram, 35-36
Microstructure, 23 Peritectoid reaction, 36-37
Mild steels, 191-193 pH, 165-166
Miller-Bravais indices, 9-10 Phase rule, 29-30
Miller indices, 7-9 Phases, 29
Monotectic reaction, 37
Phosphor bronzes, 220
Muntz metal, 220
Phosphorus in steel, 187, 195
Physical metallurgy, 1
Nabarro-Herring creep, 146 Pigments, 176
Natural ageing, 94 Pinning effect, 97-98
Naval brass, 220 Pitting, 171
Necking, 129, 144 Plane strain condition, 151
Nernst equation, 165-166 Plastic anisotropy parameter, 189-190
Nickel arsenide structure, 13-14 Plastic instability, 130
Nihard iron, 211 Plastic zone at crack tip, 150
Nimonics, 206 Plate martensite, 87
Ni-Re phase diagram, 36 Point imperfections, 14-15
Niresist iron, 211 Polarization, 167-169
Nital, 45 Polygonization, 96
Nitinol, 90-91 Polymorphism, 4
Nitriding, 125, 158-159 Porcelain, 176
Noble coating, 175 Portland cement, 176
Nodular iron, 212-213 Pourbaix diagrams, 166
Nonmetallic inclusions, 192 Powder method, 26
Normalizing, 111 Precipitation hardening, 142-143
Nucleating agent, 72
Primary creep, 143
Nucleation and growth, 67-68
Process anneal, 111
Nucleation barrier, 68, 70
Proeutectic phase, 34
Proof strength, 128
Octahedral planes, 6 Protective atmospheres, 112
Octahedral voids, 10 Pseudobinary, 49
 Index 231

Quaternary phase diagrams, 57-58 Single crystals, 147

Quench ageing, 95-96 Sintered carbide tools, 203
Quench cracks, 119-120 Slip, 17, 134-136
Quenching, 111 Smart materials, 91
S-N curves, 154
Solidification, 71-79
Rail structurals, 195-196 Solid solutions, 10-13, 140-142
Rapid solidification processing, 71, 78-79 Solidus, 31
Recovery, 96 Solute drag effect, 97
Recrystallization, 96-98 Solutionizing, 93-94
Recrystallization anneal, 110 Solvus, 33
Red brasses, 220 Solvus determination, 61-62
Red hardness, 199 Sorbite, 83
Red lead, 176 Spheroidal graphite iron, 212-213
Reduction in area, 130 Spheroidizing anneal, 110
Reduction reaction, 162-163 Spinodal decomposition, 65-66
Refroidissement, 44 Splat quenching, 71
Regular solution, 12 Spring steels, 196
Replicas, 24 Stabilized stainless steel, 179
Residual elements in steel, 187-188 Stacking fault energies, 19
Residual stresses, 119-120 Stacking faults, 19
Resilience, 196, 223 Stages of tempering, 87-88
Retained austenite, 85, 200 Stainless steels, 177-179, 203-205
Rhombohedral crystal, 6 Standard electrode potentials, 164-165
Rimmed steel, 77-78 Steel specifications, 185-187
Stellites, 203
Stiffness-to-weight ratio, 213
Sacrificial coating, 175 Strain ageing, 133-134
SAE specifications, 186-187 Strain hardening, 137-138
Salt baths, 112 Strain rate, 132
Sand casting, 72, 208 Strength coefficient, 131
Scanning electron microscope, 26 Strengthening methods, 137-143
Schaeffler diagram, 203-204 Strength-to-weight ratio, 213
Screw dislocation, 16 Stress amplitude, 153
Season cracking, 180 Stress cells, 163
Secondary creep, 144 Stress corrosion cracking, 179-183
Secondary hardening, 88-89 Stress intensity factor, 150-151
Secondary recrystallization, 98 Stress-relief anneal, 110
Segregation, 75-77 Stress-rupture curves, 145
Selected area diffraction, 24 Stretch forming, 189
Self passivation, 177 Striations, 156
Sensitized stainless steel, 177 Subcritical anneal, 111
Sessile dislocations, 136 Subcritical crack growth, 149
Severity of quenching, 113-114 Subgrain boundaries, 140
SG iron, 212-213 Substitutional diffusion, 66-67
Shape memory alloys, 90-91 Substitutional impurity, 15
Shear mechanism, 84-85 Substitutional solid solution, 10
Shockley partials, 19 Subzero treatment, 200
Short range ordering, 12 Sulphur in steel, 187
Shot peening, 158 Superalloys, 206-207
Sigma phase, 14 Superlattice, 14
Sigmoidal curve, 81 Surrounding, 42
Silicon in steel, 188 Syntectic reaction, 37
232 Index

Tafel equation, 169 Twinning, 134-135

TD nickel, 147 Twist boundary, 20
Temperature-time-transformation (T-T-T)
diagram, 100-103
Temper brittleness, 112 Ultimate tensile strength, 129
Unified numbering system, 187
Tempering of martensite, 87-88
Temper rolling, 189 Uniform elongation, 129
Upper bainite, 83-84
Tensile strength, 129
Tensile test, 127-133 Upper yield point, 132
Ternary phase diagrams, 52-57
Tertiary creep, 144 Vacancy, 15
Tetragonal crystal, 6 Vertical section, 54-55
Tetragonal distortion, 142 Vibrations in nucleation, 73
Tetragonality of martensite, 85 Vitallium, 206
Tetrahedral voids, 11 Voids in ductile fracture, 129, 147
Thermal analysis, 59-61
Thermoelastic martensite, 89-90
Thermomechanical treatments, 140, 198 Warping, 119
Thin foils, 25 Waspalloy, 206
Tie-line rule, 32 Wavy slip, 135
Tie-tetrahedron, 57 Wear resistance, 199, 211
Tie-triangle, 54 Weathering steels, 195
Tilt boundary, 20-21 Weldability, 191, 193
Tin sweat, 77 Whiskers, 136
Titanium alloys, 218 White cast iron, 207
Tool steels, 199-202 White layer, 125
Torsteel, 138 White metals, 224
Total elongation, 129 Work hardening, 128
Transmission electron microscope, 25-26 Work hardening exponent, 131
Transition carbides, 87
Trip steels, 198 X-ray diffraction, 26-27
Troostite, 83
True strain, 130 Yield strength, 127
True stress, 130
Turbine blades, 147
Twin boundary, 21-22 Zener-Hollomon parameter, 132
 SECOND EDITION
Physical Metallurgy
PRINCIPLES AND PRACTICE
V. RAGHAVAN

This well-established book, now in its Second Edition, presents the principles and
applications of engineering metals and alloys in a highly readable form. This new
edition retains all the basic topics such as phase diagrams, phase transformations,
heat treatment of steels and nonferrous alloys, solidification, fatigue, fracture and
corrosion covered in the First Edition. The text has been updated and rewritten
for greater clarity. Also, more diagrams have been added to illustrate the concepts
discussed.

This Edition gives New Sections on

Thermoelastic martensite
Shape memory alloys
Rapid solidification processing
Quaternary phase diagrams
Intended as a text for undergraduate courses
in Metallurgy/Metallurgical and Materials Engineering, this book is also suitable for
students preparing for associate membership examination of Indian Institute of
Metals (AMIIM), as well as other professional examinations like AMIE.

The Author
V. RAGHAVAN, Ph.D. (Sheffield), formerly Professor at the Indian Institute of Technology
Delhi, taught Metallurgy and Materials Science courses at the Massachusetts Institute of
Technology (MIT), Cambridge, Mass., and at IIT Kanpur and IIT Delhi for many years. A
Fellow of the American Society for Metals and an Honorary Member of the Indian Institute
of Metals, Professor Raghavan is the recipient of the Brunton Medal of Sheffield University,
UK (1964), the G.D. Birla Gold Medal of the Indian Institute of Metals (1985), and the
Hume-Rothery Award of the Institute of Materials, UK (2001).

ISBN:978-81-203-3012-2

` 195.00

www.phindia.com 9 788120 330122