Hot Working Guide—A Compendium of Processing Maps, Second Edition Copyright © 2015 ASM International®

Y.V.R.K. Prasad, K.P. Rao, and S. Sasidhara, editors All rights reserved
asminternational.org

1. Introduction
Mechanical processing is an essential step in shaping replaced by modeling techniques, which are
materials into engineering components which require developed on the basis of science-based principles.
not only dimensional accuracy but also specified These techniques address the following design and
microstructures and mechanical properties. The manufacturing issues involved.
techniques of mechanical processing involve bulk
The design requirements are:
metal working using rolling, forging or extrusion
which are generally conducted at elevated  arriving at optimum processing conditions
temperatures in order that large strains may be  controlling the microstructure in the component
imposed in a single step of the operation without the  designing optimum die shapes or preform
onset of fracture. The secondary metal working geometry without resorting to shop floor trials
processes generally use cold working which ensures  obtaining the process limits for the design of
good surface finish, high dimensional tolerance and control systems
better strength. However, these involve smaller
strains and require a large number of steps with The manufacturing issues revolve around:
intermediate annealing to restore the ductility.  the reduction of lead time in manufacturing
Processes like sheet metal working, cold forging,
 increasing the productivity without sacrificing
impact extrusion, coining, wire and tube drawing are
the product quality
some examples of this category. In recent years, with
 reducing the rejects to improve yield
the advent of rapid solidification processing and
atomization techniques for producing powders of  ensuring the repeatability in manufacturing
desired shape and size, powder metallurgy (PM) has The starting condition of the material has an immense
assumed a significant role in shape making. Using effect on its behavior during mechanical processing
this technique it is now possible to make complicated and, in particular, the ingot should be free from
shapes in exotic alloys for many critical applications macro and micro structural defects. The following
like gas turbine components. processes are used to produce input materials for
Among all the mechanical processing methods, the metal working:
bulk metal working stage is considered to be of Metal Casting: Conventionally cast ingots have
primary importance for two reasons: Firstly, in this columnar and dendritic microstructures with heavy
stage, major microstructural changes occur and these segregation of alloying elements. There could be a
have a profound influence on the subsequent non-uniformity of microstructure from center
processing steps. Secondly, in view of the large (equiaxed) to the surface (columnar). Unless careful
tonnage of material being processed by bulk metal control is exercised in casting, macro and micro
working, any improvement in processing techniques defects may occur. The importance of
has a multiplying effect on the overall productivity in homogenization of the cast ingot need not be
manufacturing. Thus, considerable effort has gone overemphasized. In recent years, ferrous materials
into developing techniques for the design and are being produced by continuous casting route and
optimization of bulk metal working processes. The the billets or slabs are continuously rolled in order to
ultimate objective is to manufacture components with increase the productivity and save energy by avoiding
controlled microstructure and properties, without reheating of ingots. In such a case, it is important to
macro or microstructural defects, on a repeatable note that there could be steep temperature gradients
basis in a manufacturing environment. Hitherto, this from surface to center. The surface temperature will
is done using trial and error techniques which are be lower than the interior, since solidification
expensive as well as time consuming and may not proceeds in that direction. This is in contrast to billets
always lead to a successful solution or optimization, which are reheated. Another common casting process
particularly for advanced materials like superalloys, used for non-ferrous alloys is direct chill (DC)
intermetallics and metal matrix composites. In recent casting. When long freezing range alloys are DC cast,
years, however, the trial and error techniques are microporosity may occur unless a hot top is used.

1
Hot Working Guide

Wrought Structures: These are either recovered temperature, strain rate and strain but also by the
(stress relieved) or recrystallized (fully annealed) stress state in the deformation zone. For example, the
microstructures and have better workability than as- tensile elongation of a material may be enhanced (or
cast microstructures. However, improper primary necking may be delayed) by the application of an
metal working may introduce microstructural external hydrostatic compression (classical
damage, flow instabilities or peripheral grain growth Bridgeman experiment) or by slow speed
in the wrought microstructures. Also, materials may deformation of a specimen with a stable fine grained
develop preferred orientation (crystallographic structure at higher temperatures (superplastic
texture) or mechanical fibering which needs to be deformation). It is therefore convenient to consider
controlled for further processing, for example, sheet workability to consist of two independent parts: state-
metal working. of-stress (SOS) workability and intrinsic workability.
Heat Treated Microstructures: In several alloy 1.1.1 State-of-Stress Workability
systems like titanium alloys and zirconium alloys, a
variety of preform microstructures may be produced SOS workability depends upon the geometry of the
ranging from acicular (-quenched) to equiaxed ( deformation zone in which the work-piece is
annealed). The preform microstructure has a subjected to a three-dimensional stress state. This is
profound influence on the workability of the material. represented as a stress tensor with nine components
or six independent components of which three shear
Powder Compacts: With advances in the rapid stress components contribute to the plastic flow of
solidification processing of alloy powders and the material while the hydrostatic components decide
mechanical alloying, it is possible to put in more the workability. For example, if the hydrostatic
alloy content without causing heavy segregation. components are tensile, any weak interface in the
Various compacting routes like cold isostatic material will open up and cause internal fractures.
pressing and sintering, hot isostatic pressing, vacuum For good SOS workability, therefore, the hydrostatic
hot pressing and compaction by blind extrusion have components should be essentially compressive. The
been developed. However, the presence of prior SOS is controlled by the nature of the applied stress
particle boundary (PPB) defects, discrete particle and the geometry of the deformation zone both of
effects (effect of hard and soft particles in a statistical which are different for different metal working
distribution of particle sizes and their individual processes [3]. The SOS workability is thus specific to
microstructures) and the high surface reactivity of the mechanical working process and is independent
powders have made the mechanical processing of of the material behavior. For example, it may be
PM compacts a highly specialized technology. For optimized in rolling by roll pass design, in forging by
advanced materials, processes like gatorizing (Pratt preform (blocking die) design and in extrusion by the
and Whitney Co.) and billet conditioning (Wright- design of the geometry of the die cavity. For a given
Patterson Air Force Base) have been patented for this geometry of the component, the available variations
purpose. for the roll pass design or forging preform design are
restricted. However, in extrusion, for a given
1.1 What is Workability? container geometries and a product geometry there is
The engineering parameter that is of importance in considerable scope for innovation in the die design
mechanical processing is commonly termed as and thus die geometry like shear, conical, parabolic
“workability” which refers to the ease with which a and streamlined dies are developed [4]. A suitable die
material can be shaped by plastic flow without the design may be selected for a controlled SOS in the
onset of fracture. This general term includes all other deformation zone. For example, in case of difficult-
terms like forgeability, rollability, extrudability and to-work materials like metal matrix composites or
formability (sheet metal working). A fundamental PM superalloys, streamlined die design for extrusion
understanding of workability is essential for is recommended since this avoids rigid body rotation
developing science-based techniques in mechanical and ensures hydrostatic compressive state of stress in
processing. Detailed reviews on the influence of the die cavity. One of the accepted methods of die
metallurgical parameters on workability and the design for metal working processes is to use
various standard workability tests are available [1,2]. CAD/CAM techniques [5] incorporating a realistic
It is clear that the workability is influenced not only simulation model so that expensive and time
by the microstructures of the material, applied consuming shop floor trials are minimized.

2
 Introduction

Commercial FEM codes are available for this activation energy, deformation mechanisms may be
purpose and these accept experimental constitutive evaluated and microstructural correlations obtained
equations for a given work-piece material. with a temperature compensated strain rate
parameter, Z defined as:
1.1.2 Intrinsic Workability
The intrinsic workability depends upon the initial Z   / A exp( Q / RT)   n (2)
microstructure as decided by the alloy chemistry and
The kinetic rate equation (Eq.1) is valid over a
prior processing history and its response to the
narrow range of temperature and strain rate, and
applied temperature, strain rate and strain in
when considered over a wide range n and Q become
processing. This response is embedded implicitly in
temperature and strain rate dependent. The kinetic
the flow stress variation with temperature, strain rate
analysis is applicable for pure metals and dilute
and strain and is represented mathematically as a
alloys but when extended to commercial alloys with
constitutive equation. However, as a part of the
complex microstructures, the apparent activation
explicit response of the material to the imposed
energy values become too complex to interpret in
process parameters, certain microstructural changes
terms of a single mechanism. Further, the kinetic
(mechanisms) occur within the material and these
model does not specifically lead to optimization of
will have to be characterized. For example, under
intrinsic workability nor can it be applied universally
certain conditions, the response may be in terms of
for microstructural control without knowing the
microstructural damage as the flow may be unstable
specific ranges or domains where it is valid.
or localized. Alternately, the microstructure may
undergo a favorable reconstitution like dynamic
Frost and Ashby [7] were the first to represent the
recrystallization. For obtaining good intrinsic
materials response in the form of a Deformation
workability, it is essential to choose processing
Mechanism Map. These are plots of normalized
conditions that avoid microstructural damage and
stress vs. homologous temperature showing the area
instability during processing.
of dominance of each flow mechanism, calculated
The shapes of stress-strain curves implicitly contain using fundamental parameters. The emphasis in the
information related to the mechanisms of hot Deformation Mechanism Maps has been essentially
deformation. For example, flow softening type of on the creep mechanisms applicable to lower strain
stress-strain behavior with an initial peak stress or rates and the maps are very useful for alloy design.
oscillations suggests dynamic recrystallization However, mechanical processing is done at strain
(DRX). However similar stress-strain behavior can rates orders of magnitude higher than those observed
also be due to flow instability. Likewise, DRX may during creep deformation and therefore involves
occur in cases where the behavior is steady state. It is different microstructural regimes. Considering strain
therefore not advisable to conclude on the rate as one of the direct variables and temperature as
deformation mechanism from the shapes of the the other, Raj [8] extended the concept of Ashby’s
stress-strain curves alone. maps to construct a processing map which is shown
schematically in Fig. 1.1. The Raj map represents the
One of the early attempts [6] to evaluate the limiting conditions for two damage mechanisms: (i)
mechanisms of hot working was to use a kinetic rate cavity formation at hard particles in a soft matrix
equation basically of the type: occurring at lower temperatures and higher strain
rates, and (ii) wedge cracking at grain boundary triple
  A  n exp (  Q RT) (1)
junctions occurring at higher temperatures and lower
where  : strain rate, : flow stress, Q: activation strain rates. At very high strain rates, a regime
energy, R: gas constant, T: temperature, A and n: representing adiabatic heating was identified. In
constants. In order to fit the experimental data more principle, there is always a region which may be
termed “safe” for processing where neither of the two
closely and over a wider range, several functions of 
damage mechanisms nor adiabatic heating occurs.
like the hyperbolic sine function have been used [6]
Using an atomstic approach, processing maps were
but apparently these do not have any other
developed by Raj for pure metals as well as dilute
significance. For use in finite element models, simple
alloys.
polynomial fits are found to be more convenient
mathematically. On the basis of an apparent

3
Hot Working Guide

Fig. 1.1 Raj map for aluminum showing limiting conditions for damage nucleation (schematic) [8].

1.2 What is a Processing Map? pack), a store of power (tools like anvil, ram and die)
and a dissipator of power (the workpiece). Energy is
A processing map is an explicit representation of the generated by the source, transmitted to the tools to
response of a material, in terms of microstructural store the power and transferred to the workpiece
mechanisms, to the imposed process parameters and through an interface (lubricant). The workpiece itself
consists of a superimposition of a power dissipation dissipates the energy while it undergoes plastic flow
and an instability map. These are developed on the to take the shape imposed by the deformation zone.
basis of the Dynamic Materials Model (DMM) [9] The response of each of the above system elements
which is essentially a continuum model using the depends upon their individual constitutive equations
concepts of systems engineering [10], extremum which should be evaluated for modeling their
principles of irreversible thermodynamics with behavior. If the constitutive behavior of the system
application to continuum mechanics of large plastic elements could be modeled accurately, they may be
flow [11] and those describing the stability and self linked together such that suitable process controls
organization of chaotic systems [12]. DMM may be may be designed for optimization. In this system, it is
viewed as a bridge between the continuum mechanics important to note that power or energy per second is
of large plastic deformation and the development of to be considered and not energy per se, since the
dissipative microstructures in the material. The response of the system depends on how fast or slow
different aspects of processing maps and the the energy is input, bringing in “time” as an
principles and applications have been reviewed from independent variable to make the system “dynamic”.
the point of view of microstructural control [13-15]. While the integration of all the system elements has
1.2.1 Processing System not been achieved so far, the characteristics of the
dissipator element (work-piece) are considered to be
The model considers mechanical processing as a most important in designing the control system.
system, an example of which is shown in Fig. 1.2
with reference to the forging process. The system
consists of a source of power (e.g. a hydraulic power

4
 Introduction

Fig. 1.2 Processing system elements with forging as an example.

The constitutive equation of the workpiece describes (iv) Irreversible: Since the workpiece is subjected to
the manner in which energy is converted at any large plastic deformation, the system is irreversible.
instant into two forms - thermal and microstructural, The principles of irreversible thermodynamics as
which are not recoverable by the system. The applied to the continuum mechanics of large plastic
following definitions apply to the dissipator element of flow are relevant in describing the entropy changes
the metal processing system [13,14]: occurring within the material.
(i) Dissipator: During hot deformation the workpiece (v) Away from Equilibrium: Since the strains are not
dissipates all the power that is input to it since the imposed in infinitesimally small increments and the
stress strain curves exhibit either steady-state or flow strain rates normally encountered in metal forming
softening behavior. are large, the system is considered to be away from
equilibrium. It is relevant to mention that the laws of
(ii) Non-linear: The relation between the flow
equilibrium are universal while the behavior of the
variable (strain rate ) and the effort variable (flow
system may be very specific away from equilibrium.
stress ) is non-linear when considered over a wide
range. The workpiece is therefore a non-linear (vi) Sensitivity to initial conditions: Small changes in
dissipator of power. the chemistry, initial microstructure, temperature,
strain rate and strain can cause a large change in the
(iii) Dynamic: Time taken to achieve a particular
response of the system or lead to different
strain depends upon the strain rate and is an
mechanisms of deformation. In view of the non-
independent variable in the constitutive relation
linearity and sensitivity to initial conditions, the
between flow stress and strain rate.
material system exhibits deterministic chaos similar
to that occurring in other dissipative systems.

5
Hot Working Guide

1.2.2 Power Dissipation Map d ( i) S

strain rate,  is the temperature and is the
The non-linear dynamics of a dissipative system are dt
generally analyzed using state space variables [3]. rate of internal entropy production. The total rate of
For a materials system with a given chemistry entropy production consists of two complementary
(composition) and processing “history”, one set of parts [10]. The first part (generally larger) consists of
physical state (or phase) space variables for hot “conduction entropy” which is due to the conduction
working consists of (1) temperature of deformation, of heat from where it is generated (due to plastic
(2) strain rate, (3) strain, and (4) the dissipative state flow) to the colder parts of the body. The second part
of the microstructure. The state-control variables are is due to a microstructural dissipation which lowers
temperature, ram velocity and the extent of the flow stress for plastic flow (dislocation
deformation. While these are related to the first three movement). Ziegler [11] represented these two in
state space variables respectively, there is no simple terms of dissipative functions in the velocity and
or direct method of representing the dissipative state force space and showed that the instantaneously
of microstructure. For this purpose, the extremum
dissipated total power (  . ) is given by:
principles of irreversible thermodynamics of the
 
quasi-static processes of large plastic deformation are
helpful. Ziegler [11] has shown that the behaviour of P   .d    .d  G  J (2)
such a system follows the principle of maximum rate 0 0
of entropy production, which is equivalent to the where  is the effective stress and  is the
principles of least irreversible force or least velocity effective strain rate. In terms of physical systems
corresponding to the velocity and force spaces terminology [11], the first integral is called G content
respectively. At a given temperature in the hot and the second one a J co-content since it is a
working regime, the rate of dissipation work (power) complementary part of G content. The constitutive
is directly proportional to the rate of internal entropy equation decides the relative values of power
production [9] which is always positive since the dissipation through the heat conduction and
process is irreversible. microstructural dissipation since the origin of
viscoplasticity is in the microstructural dissipation.
d (i ) S For plastically deforming materials, the power law:
P        0 (1)
dt   K (T ,  ,  ) m(T , , )

(3)
where  is the effective stress,  is the effective

Fig. 1.3 (a) Schematic representation of the constitutive equation in a non-linear power dissipator (b) Ideal linear
dissipator.

6
 Introduction

is widely used to define a constitutive relation. In value of m increases, the microstructural dissipation
Eq.3, T is the temperature and  is the effective will increase.
strain. This is represented schematically in Fig. The value of m less than one (commonly known as
1.3(a). The instantaneous total power dissipated will negative strain rate sensitivity) is observed for the
be given by a rectangle of area (  .  ). For fixed process of dynamic strain aging [13] where moving
dislocations are repeatedly locked and unlocked by
values of T and  we are interested in studying
faster moving solute atoms. Deforming solids in
(  J/  G) and (  J/  P) for small variations of  , which m is more than one are described as “locking
to obtain an indicator of changes in the dissipative solids” in continuum mechanics [16]. Both the above
mechanism [14]. Assuming that dependence of K, m extremes represent flow instabilities. Since the
on  is weak (not necessarily zero) over a narrow power dissipation and the rate of entropy production
are directly related (Eq.1), the efficiency parameter
range, the instantaneous values of  J,  G and  P
(Eq.9) describes the relative rate of entropy
are given by the following integrals:
  
production occurring during hot deformation due to a
microstructural change in the system. It can thus be
J   d (4) used as a state space parameter representing the
 microstructural evolution in the system. It may be
 0   0 noted that a higher efficiency of power dissipation
G   d
 0
(5) corresponds to a lower dissipative energy state of the
material.
 P  K (    )m+1 (6)
The efficiency parameter (η) may be plotted as a
and function of temperature and strain rate to obtain the
 J/  G = m (7) power dissipation map. The efficiency represents the
relative rate of internal entropy production during hot
 J/  P = m/(m+1) (8)
deformation and characterizes the dissipative
microstructure under different temperature and strain
We further define the efficiency of power dissipation
rate conditions. An example of a three dimensional
() with respect to a linear dissipator (m=1) [Fig.
(3D) map of the efficiency of power dissipation with
1.3(b)] to be
temperature and strain rate is shown in Fig. 1.4(a). In
view of the non-linear variation of the flow stress
J / P m /( m  1) 2m with strain rate and temperature, the map will have
= =   (9)
(J / P)linear 1/ 2 m 1 hills and valleys. An easier method of representation
is in the form of an isoefficiency contour map,
obtained by sectioning the 3D map at constant
Maximum possible dissipation occurs when m=1 for efficiency levels, which is shown in Fig. 1.4(b). It
which the J co-content is Jmax = (  .  )/2. From this may be noted that in dissipative systems, the strain
analysis, we see that m is the key parameter defining rate and the initial conditions determine the motion of
the relative (not absolute) partitioning of power the microsystem and the entropy production within
between heat generation and microstructural change, the microsystem is a function of strain rate alone.
i.e., constant values of m imply a fixed dissipation Maps at different strains do not show many
mechanism, while changing m values imply changing differences and the strain components only define the
mechanisms. The value of m for stable flow in a frame of the microsystem [11].
viscoplastic solid is between 0 and 1 [12] and as the

7
Hot Working Guide

1.2.3 Instability Map kink bands, mechanical twinning and flow rotations.
The presence of these instabilities in the
The stability condition described in the Dynamic microstructure of the component will have to be
Materials Model is that considered by Ziegler [11]. avoided by keeping away from the processing
Stable flow will occur if the differential quotient conditions of the unstable regimes.
satisfies the inequality :
1.2.4 Processing Map
dD D
 (10) The instability map (Fig. 1.5) may be superimposed
dR R
on the power dissipation map [Fig. 1.4(b)] to obtain a
where R =  .  and D is the dissipative function processing map in which the limits for the unstable
flow regime are clearly marked. An example of such
which represents the constitutive behavior of the
a processing map for Nimonic AP-1 alloy is shown in
material. Since J determines the dissipation through
Fig. 1.6. The map exhibits several efficiency “hills”
metallurgical processes, the dissipation function
which appear as “domains” with successively
related to metallurgical stability is given by J and by
increasing isoefficiency contours in the contour map.
putting D = J in Eqn. (10) one gets the condition for
The domains are separated by “valleys” which are
microstructural stability at a constant temperature in
referred to as “bifurcations”. This behavior of the
terms of a dimensionless parameter. (  ) , given by dissipative system is akin to the self organization of
chaotic systems suggested by Prigogine [12] and the
 ln( m / m  1) efficiency of power dissipation represents the relative
(  )  m  0 (11)
 ln  rate of entropy production, the highest being that for
a linear dissipator. Each of the domains in the
The (  ) parameter may be evaluated as a function of processing map represents a microstructural
temperature and strain rate to obtain an instability mechanism that contributes to the power dissipation
map, where metallurgical instability during plastic and is deterministic in the sense that kinetic laws are
flow occurs in regimes where (  ) is negative. A obeyed. Processing under conditions in a
typical instability map for Nimonic AP-1 alloy is deterministic domain will ensure microstructural
shown in Fig. 1.5. The well known manifestations of control and reproducibility.
flow instabilities are adiabatic shear bands, flow
localization, dynamic strain aging (Lüder’s bands),

8
 Introduction

Fig. 1.5 Instability map for Nimonic

AP-1 superalloy showing variation of
(  ) parameter with temperature
and strain rate. Flow instability occurs
when (  ) is negative (within
contour H).

Fig. 1.6 Processing map for Nimonic

AP-1 superalloy obtained by
superimposition of power dissipation
map and instabiliy maps. Contour
numbers represent percent efficiency
of power dissipation. Shaded region
corresponds to flow instability.

9
Hot Working Guide

On the other hand, processing in the neighborhood of the process is called superplastic deformation [22].
bifurcation points will result in a combination of The basic mechanisms involved are the grain
microstructures that characterize the neighboring boundary sliding and diffusion accommodated flow
domains. The occurrence of bifurcations is “history” at grain boundary triple junctions mitigating the
dependent and is not entirely predictable. The formation of wedge cracks. Superplasticity is also
microstructural combinations resulting at bifurcations observed in coarse grained materials in some alloy
will probably be related and may not be exactly systems like -brass and Mg-Li alloys [23].
reproducible in the component. However, for large strain deformation, superplasticity
may cause micro porosity at grain boundaries,
The instability criterion used in Eqn. (11) for
because of which the process is safe at smaller
delineating regimes of flow instability may be
strains. In comparison with the DRX process, the
physically interpreted in simple terms as follows: If
following differences exist from the view point of
the material system does not produce entropy
bulk metal working:
constitutively at a rate that at least matches the rate of
entropy input through imposed process parameters, (a) Superplastic deformation occurs at strain rates
the flow becomes localized and causes a flow which are usually several orders of magnitude lower
instability. than those for the DRX process and hence is a very
slow process for manufacturing.
1.3 Hot Deformation Mechanisms (b) In view of the slow speeds involved, superplastic
deformation requires isothermal conditions where the
The following microstructural processes occur, in die has to be heated to the same temperature as that
general, in metals and alloys during hot deformation. of the workpiece.
Dynamic Recrystallization (DRX): This process (c) There is no large scale reconstitution of microstructure
refers to the occurrence of simultaneous during superplasticity unlike in the DRX process and for
recrystallization during deformation by nucleation preform microstructure like -transformed Ti alloys,
and growth processes [17-21]. DRX is different from superplasticity is restricted since the interfaces cannot easily
static recrystallization in the sense that DRX slide due to their crystallographic relationship with the
characteristics are decided by the rate of nucleation matrix.
versus rate of growth under given imposed
conditions of temperature and strain rate. In static (d) In PM compacts, superplastic deformation may
recrystallization on the other hand, a fixed amount of cause interface cracking at hard particles since the
stored energy (dependent on cold work) is released interfaces slide and do not transfer mechanical energy
by thermally activated dislocation recovery and grain across the interface to refine the hard particles. Also
boundary migration and may be termed a kinetic the PPB defects are not redistributed since there is no
process. DRX is a beneficial process in hot reconstitution of microstructure.
deformation since it not only gives stable flow and
good workability to the material by simultaneously (e) Since superplastic forming is a slow high
softening it but also reconstitutes the microstructure. temperature process, it may be combined with
For example, DRX breaks down the as-cast processes like diffusion bonding and thus very useful
microstructure to produce wrought microstructure, or in the manufacture of sheet metal components used
globularizes the acicular preform microstructure as in for advanced applications like aerospace structures.
titanium and zirconium materials, or redistributes the Dynamic Recovery: Thermal recovery of dislocations
prior particle boundary defects in PM compacts to due to their climb causes dynamic recovery which
facilitate further processing, or eliminates discrete occurs generally in the homologous temperature scale
particle effects by transferring mechanical energy of 0.4-0.6 and is therefore relevant to warm working
across the hard particle interfaces to refine them. of materials. The process that occurs during dynamic
Thus DRX is a chosen domain for optimizing hot recovery is the diffusion of the rate controlling
workability and controlling the microstructure and is atomic species. The dynamically recovered
a “safe” domain for bulk metal working. microstructure has well defined subgrains with
Superplastic Deformation: Materials with stable fine relatively dislocation free interiors. Dynamic
grained structure when deformed at slow speeds and recovery causes work hardening of the material, the
high temperatures exhibit abnormal elongations and

10
 Introduction

rate of which is lower than that obtained in cold not conducted away to the cooler regions of the body
working. since the time available is too short. The flow stress
in the deformation band will get lowered and further
Wedge Cracking: Under conditions of low strain rate
plastic flow will get localized. The band gets
and higher temperature where grain boundary sliding
intensified and nearly satisfies adiabatic conditions.
occurs under shear stress, wedge cracks are produced
Such bands are called adiabatic shear bands (ASB)
at triple junctions to relieve the stress concentration.
which may exhibit cracking, recrystallization or
If the stress concentration is relieved by diffusion-
phase transformation along macroscopic shear planes
aided flow, superplasticity occurs and so
[24] and hence have state-of-stress dependent
superplasticity is a subset of wedge cracking. Wedge
manifestation. Their intensity depends on physical
cracking is reduced by increasing the strain rate or
properties of the material like specific heat and
decreasing the temperature [8] or grain growth.
conductivity in addition to the deformation
Void Formation: If hard particles are present in a characteristics.
soft matrix, deformation causes the interface to crack
The second common flow instability manifestation in
and debond since the matrix undergoes plastic flow
the microstructure is the flow localization [25]. This
while the particles do not deform. When the
is less intense than adiabatic shear band formation
accumulated stresses become large, the interface may
and gives microstructural inhomogeneity where
separate or the particle itself may crack, which may
localized shear bands may be curved or wavy. Flow
lead to the creation of microstructural damage due to
localization also occurs at high strain rates.
cavity formation, ultimately contributing to ductile
fracture. This process dominates at lower There are other flow instability manifestations in
temperatures and higher strain rates [8]. materials like Lüder’s bands when dynamic strain
aging occurs (lower strain rates) or kink bands in
Intercrtystalline Cracking: The propagation of
textured materials or mechanical twinning at high
cracks along the grain boundaries during deformation
strain rates.
may occur when low melting compounds are formed
due to segregation of certain alloying additions. At Out of all the above mechanisms, DRX and
very high temperatures (closer to melting point) superplastic deformation are “safe” mechanisms for
intercrystalline cracking can occur at high strain rates hot working while dynamic recovery is preferred for
due to incipient melting. At intermediate warm working. All other mechanisms either cause
temperatures generally encountered in creep (slow microstructural damage or inhomogeneities of
strain rate deformation), intercrystalline failure is a varying intensities and hence are to be avoided in the
common mode. microstructure of the component. The processing
map will be a useful guide in designing the hot
Prior Particle Boundary (PPB) Cracking: The
working processes such that the process parameters
surfaces of the powder particles would have picked
are controlled to be within the “safe” processing
up impurities or formed oxides, hydroxides nitrides
domains or at least away from the undesirable
and carbides in view of their high reactivity and
regions. The DRX domain is a chosen domain for
would appear as clearly visible PPB defects in the
bulk metal working in view of its advantages over the
powder compact. Being weaker links, these provide
superplastic deformation, as discussed earlier.
ready paths for crack propagation and reduce the
workability of the material. Deformation of PM 1.4 How to Generate a Processing Map
compacts at lower temperatures and strain rates
reveal cracking along the PPB defects. While hot The input to generate a processing map is the
isostatic pressing does not eliminate the PPB defects, experimental data of flow stress as a function of
high temperature deformation involving dynamic temperature strain rate and strain. As the map
recrystallization is required to redistribute the PPB generated will be only as good as the input data, it is
defects and reduce their effect on workability. important to use an accurate, reliable and yet simple
experimental technique for generating them. While
Flow Instability Processes: The microstructural hot tensile, hot torsion or hot compression testing
manifestations of flow instabilities are many but the techniques may be used for this purpose, a hot
most common process is the occurrence of adiabatic compression test has decisive advantages over others.
shear bands. At high strain rates, heat generated due First of all, in a compression test on a cylindrical
to the local temperature rise by plastic deformation is specimen, it is easy to obtain a constant true strain

11
Hot Working Guide

rate using an experimental decay of the actuator A matrix of 6 temperatures and 6 strain rates is
speed. It is convenient to measure the adiabatic generally selected for testing. Test temperatures are
temperature rise directly on the specimen and chosen in the homologous temperature range 0.6-0.8
conduct the tests under isothermal conditions. In the with 50oC gap between tests and the true strain rate
following section the experimental technique of hot range is 0.001-100 s-1 with each test done at strain
compression testing used for generating most of the data rate values one order of magnitude apart. At each test
given in the subsequent chapters is described. The maps temperature, a 15 minutes soaking time is allowed for
generated by other techniques are not very different as equilibration of temperature before compressing the
far as the workability domains are concerned but the specimen to about half the height. From the data
cracking regimes are wider for maps obtained in tension acquisition system, the load-stroke data are obtained
and the flow instability regimes extend wider in maps and the adiabatic temperature rise is recorded on the
obtained in compression [26]. transient recorder.
1.4.1 Hot Compression Testing 1.4.2 Data Analysis
Cylindrical specimens with a diamter of 10 mm and a The load-stroke curves were converted using a
height of 15 mm are generally used for hot computer program into the true stress-true plastic
compression testing. The specimen height to diameter strain curves, first by subtracting the elastic portion
ratio (aspect ratio) is 1.5 and has chamfered edges to of both the material and the machine from the stroke
avoid fold-over in the initial stages of compression. values at each of the loads, and then by using the
The top and bottom faces of the specimen are standard equations:
machined parallel and have concentric grooves of 0.5
mm depth to retain lubricant. The specimen has a 0.8  = P/Ao (1 - e) (13)
mm diameter hole reaching to the center of the  = - ln (ho/h) (14)
specimen for inserting a fine thermocouple to
measure the actual specimen temperature as well as where  = true stress (or flow stress), P = load, Ao =
the adiabatic temperature rise. original area of cross section of the specimen, e =
engineering strain [(ho - h)/ho],  = true plastic strain,
Testing is done on a computer controlled ho = original height of the specimen, h =
servohydraulic testing machine of 100 kN capacity instantaneous height.
and with a capability to give strain rates in the range
0.0003 to 100 s-1. The machine is equipped with a 1.4.3 Computational Procedure
control system that can give constant true strain rates
by an experimental decay of the actuator speed with Flow stress data obtained (at constant temperature,
time as per the equation: strain rate and strain) using Eqn. (13) is corrected for
adiabatic temperature rise by applying linear
X(t) = -ho [1-exp(  t)] (12) interpolation between log() and (1/T) values. Cubic
spline interpolation is carried out to compute the flow
where X(t) = actuator position at any time ‘t’, ho =
stress values at finer temperature and strain rate
initial sample height and  = strain rate. intervals, using the experimental data points as knots.
The control system makes the actuator speed The value of strain rate sensitivity (m) in each of
correction 150 times for the true strain rates in the these subintervals is calculated using Eqn. (6).
range 0.0003-1 s-1, 138 times for the strain rate of 10 Efficiency power dissipation () and instability
s-1 and 13 times for the strain rate of 100 s-1. parameter () values are computed using Eqns. (9)
and (11) and are plotted in the temperature-strain rate
The machine is equipped with two interchangeable split
plane to obtain power dissipation and instability maps
furnaces, one for testing at temperature up to 1000oC
respectively.
and the other up to 1250oC. The temperature control is
within 2oC and the adiabatic temperature rise is
measured using a transient recorder. The loading system
has superalloy push rods and replaceable platens.
Lubrication is done using molybdenum disulfide with
graphite at temperatures up to 500oC and molten glass at
temperatures higher than 900oC.

12
 Introduction

1.4.4 Microstructural Examination

Specimens after testing are water quenched or air structure (Fig.1.8), the extent of which is dependent
cooled as the case may be and carefully retained after on strain, and (iii) PPB structure is eliminated in PM
proper identification of the deformation conditions. compacts.
Selected specimens, for example at the peaks of the
(b) The grain boundaries in the DRX microstructure
domains or in the instability regime, are sectioned at
are wavy in nature (Fig.1.7(b)).
mid plane parallel to the compression axis and the cut
sections are prepared for metallographic examination. (c) The average grain diameter increases sigmoidally
For examining the DRX microstructures, the central with temperature at DRX strain rate and a 50%
region of the section may be selected while for change in grain size and efficiency occur at the DRX
observing cracking features, the bulge portion may be temperature. The grain size decreases with increase
viewed. For identifying adiabatic shear bands, macro in strain rate at DRX temperature.
examination of the specimen may be appropriate.
(d) The tensile or torsional ductility variation at the
1.5 How to Interpret and Validate DRX strain rate is similar to that of the efficiency
variation and reaches a peak value at DRX
Once the map is obtained, the microstructural temperature.
mechanisms occurring in each of the domains and the
manifestations of flow instabilities will have to be
carefully identified and fully validated. The following
general guidelines will help in interpreting the
B. Superplasticity/Wedge Cracking Domain
processing map. However detailed microstructural The superplasticity/wedge cracking domain occurs at
examination is needed for confirmation. temperatures of 0.7-0.8 Tm and strain rates lower than
0.01 s-1. Both processes are characterized by a high
1.5.1. Microstructural Examination of efficiency of power dissipation (> 60%) and a steep
Deformed Specimens rise in efficiency with decrease in strain rate
(contours in the map occur closer). The grain
A. DRX Domain structure has a diamond configuration with respect to
(a) DRX domain generally occurs in the homologous the compression axis, with many of the grain
temperature range 0.7-0.8 and at intermediate strain boundaries oriented at 45° angle with respect to the
rates (0.1-1 s-1) in the maps for low stacking fault compression axis. An example is shown in Fig. 1.9. If
energy materials. The strain rate range is lower for the domain represents superplasticity, the tensile
high stacking fault energy materials (0.01-0.001 s-1). ductility will be abnormally high (several hundreds of
percent) at the temperature and strain rate
(b) The maximum efficiency of power dissipation in corresponding to the peak efficiency, but will be poor
the DRX domain is about 30-40% for low stacking if wedge cracking occurs. Wedge cracking may be
fault energy materials (Ni, Cu, Zn, Ti, Zr-base alloys) identified at grain boundary triple junctions of the
and is about 50-55% in high stacking fault energy bulge region of the compression specimen
materials (Al and Cd). (Fig.1.10(a)). Wedge cracks may get repaired in
(c) The contours in the DRX domain are widely compressive state-of-stress and it may not be possible
spaced representing a less steep hill and present a to obtain microstructural evidences for these cracks.
fairly wide window in most materials. If a tension test is carried out under the peak
conditions of this domain, fracture by intercrystalline
DRX domain may be validated with the help of the mode is a clear indication of wedge cracking.
following general microstructural features in the
deformed samples.
(a) There is a reconstitution of the microstructure in
the DRX domain in three possible ways: (i) as-cast
structure changes to wrought microstructure
(Fig.1.7), (ii) acicular preform microstructure (-
transformed Ti or Zr alloys) becomes spheroidized

13
Hot Working Guide

(a) (b)

Fig.1.7. (a) As-cast microstructure of AZ31 showing dendrites, and (b) DRX microstructure of AZ31 specimen
deformed at 550 oC/0.001 s-1.

(a) (b)
Fig.1.8. (a) β-treated or acicular microstructure of Ti-6Al-4V alloy, (b) Spheroidized microstructure produced by
hot forging of β-treated Ti-6Al-4V at 900 oC at 0.001 s-1 .

70

60

50
Grain boundaries

40

30

20

10

0
0-10 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90
o
(a) Angle with respect to compressioin axis,
(b)
Fig 1.9. (a) Microstructure of AZ31-1Ca (DMD) deformed at 500 C/0.0003 s showing "diamond" configuration
o -1

of grain boundaries with respect to the compression axis (vertical). (b) Statistical distribution of the grain
boundaries with respect to the angle with respect to the compression axis.

14
 Introduction

(a) (a)

(b) (b)
Fig.1.10. (a) Wedge cracking in IN625 deformed Fig.1.11. (a) Adiabatic shear band in ABaX422 at
at 1200 oC/0.001 s-1. (b) Void formation hard 260 oC/1s-1. (b) Flow localization in AZ31-2Ca
particles observed in Tx32-Al-Si Mg alloy. alloy deformed at 400 oC and 10 s-1

C. Ductile Fracture Domain

The void formation or ductile fracture domain occurs axis (Fig.1.11(a)).
at lower temperatures and higher strain rates and is (b) Within the band, cracking or recrystallization or
characterized by a steep increase in the efficiency phase transformation may occur. It is also possible
with increasing strain rate and decreasing that several parallel bands may occur intersected by
temperature. This domain is relevant to metal matrix other similar bands. These bands can be confirmed
composites deformed at higher strain rates. The with the help of macro or micro structures.
microstructures exhibit interface debonding and/or (c) Flow localization is less intense than adiabatic shear
particle fracture (Fig. 1.10(b)). bands and appears as inhomogenous deformation with
wavy localized bands (Fig.1.11(b)). Under intense flow
D. Flow Instability Regions localization conditions, the bands may have a 35°
The manifestation of flow instabilities were given orientation with respect to the compression axis.
earlier and the following microstructural features are
(c) The kink bands have their axis along the applied
helpful in identifying them.
compressive stress while the tensile fractures of the
(a) The adiabatic shear bands occur as intense flow specimens deformed in the dynamic strain aging
localization at 45° with respect to the compression region occur at 45° with respect to the tensile axis.

15
Hot Working Guide

1.5.2. Physical Validation Experiments been used to confirm the conclusions from processing
maps of electrolytic copper [27], AZ31 [28] and Mg-
Forging, rolling or extrusion processes may be used 4Al-2Ba-2Ca [29] alloys.
for validating processing maps. The processes may
be conducted at temperature and strain rate 1.6 How to Apply to Industrial Processes
conditions corresponding to the peak efficiency
The information from processing maps may be
conditions in the domain identified for optimum
applied to industrial metal working processes in two
workability (e.g. DRX) and also in the instability
ways: Firstly, for newer materials the process may be
regimes. The adiabatic temperature rise and the local
designed to suit the constitutive requirements of the
strain rate distributions in the deformation zone may
material from view point of optimum workability and
be estimated using finite element simulation software
microstructural control. Secondly, the existing
[3] applicable to large plastic flow problems. The
processes may be optimized and proper process
scheme of physical validation methodology is
controls may be established such that the product
depicted in Fig. 1.12. The simulation methods not
quality and yield improves. The application of
only give estimates of load-stroke curves but also
processing maps to design and optimize bulk metal
predict the state-of-stress existing in the deformation
working processes like rolling, forging and extrusion
zone if the relevant constitutive equations are
is described below along with few examples.
introduced in the finite element formulation.
Laboratory scale isothermal forging validation
experiments producing cup-shaped component have

Forging of Rolled AZ31 in RD, TD & ND

Constitutive Eqns FEM Simulation Die Geometry Forging Experiments

(300-500
(300-550C/0.01
oC/0.01 –
– 10
10
σ= f (T,  , ε) Model
mm/s)

Materials Models Local Values of Load-Stroke Load-Stroke

Processing Maps σ, ,ε in the Curves Curves
Kinetic Analysis component
Validation

Microstructural Prediction of Grain Microstructural

Development as a Size in the Observations & Grain
f(T,  ) component shape
Validation Size Measurements

Fig. 1.12. Physical modeling methodology for validation of simulation models and processing maps.

16
 Introduction

Fig. 1.13 Processing map for

microalloyed steel. Contour
numbers represent percent
efficiency of power dissipation.
Shaded region represents flow
instability.

1.6.1 Information from Processing Maps and decreases logarithmically with increase in strain
rate at the DRX temperature. Thus the lower
(a) To begin with, the domain of DRX may be temperature, higher strain rate combination may be
identified in the processing map and validated used for finishing operations since it gives fine
microstructurally. The temperature and strain rate grained structures.
corresponding to the peak efficiency in this domain
are chosen as the optimum parameters for hot (c) The limiting temperature and strain rate
working the material. For example, for the processing conditions for avoiding flow instabilities and
map of a microalloyed steel shown in Fig. 1.13, the cracking processes may be obtained to avoid these
optimum temperature is 1090°C and the strain rate is regimes in the process design. For example, in the
about 0.3 s-1, since DRX domain occurs with a peak map cited above (Fig. 1.13), flow instabilities occur
efficiency of 33% under these conditions. at strain rates higher than about 10 s-1 on the higher
temperature side of the domain and about 1s-1 on the
(b) The temperature and strain rate ranges for the lower temperature side. These high strain rate limits
DRX domain may be identified. For the above map are important for design since manufacturing is done
(Fig. 1.13) these are 1000-1150°C and 0.01-5 s-1 at the maximum allowable speeds for enhanced
respectively. The trajectory for the control system productivity. Likewise, the wedge cracking domain at
should be within this temperature and strain rate temperatures higher than 1150°C and strain rates
range for obtaining good microstructure control. It lower than 0.01 s-1 and intercrystalline cracking
may be noted that the grain size increases above 1150°C and 10 s-1 are undesirable.
sigmoidally with temperature at the DRX strain rate

17
Hot Working Guide

Fig. 1.14 Strain rate ranges for bulk metal working processes (schematic).

Fig. 1.14 Strain rate ranges for bulk metal working processes (schematic).

1.6.2 Information on Machines (a) On the basis of the strain rate range chosen for
processing, the machinery may be selected with the
Bulk metal working processes like rolling, forging help of Fig. 1.14 in the first instance and the strain
and extrusion are performed on a variety of machines rate controls may be designed from a more accurate
[30]. The strain rates achievable depend on the speed estimates of strain rates from process simulation
of the machine, the geometry of the deformation zone models. Furthermore, the machine capacity may be
and the geometry of the workpiece. Also the local estimated using simple techniques like slab method
values of strain rates in the deformation zone may for first approximation or in more detail using FEM
vary and may be estimated by simulation using FEM simulation models, which also give load-stroke
techniques as described earlier [3]. However, using curves for the process.
simpler analytical methods, the average strain rate
values may be estimated for the process and a range (b) One-step processes like extrusion and forging
assigned to each type of machine. These estimates may be conducted at the optimum temperature for
will be approximate and nonetheless help in choosing DRX. However for multi-step processes like rolling,
the type of machine required for a particular strain the process may be started at the higher temperature
rate range. These approximate ranges for the bulk and lower strain rate side of the domain and finished
metal working processes are given in Fig. 1.14. at the lower temperature and higher strain rate end.
1.6.3 Process Design The end product will have fine grained structure
The information from processing maps and which will enhance workability for further
machinery may be integrated to design and optimize processing. It may be noted that the actual
the bulk metal working processes. The following temperature during processing may be slightly
aspects may be considered for this purpose: different, since heat is generated during plastic

18
 Introduction

deformation and is also simultaneously lost by rolling mill, the average strain rate (  ) is calculated
conduction to tools and by radiation. For a very using the equation [31]:
accurate microstructural control, these aspects will
also have to be taken into account in the process 2N R 1 t  1 
design.   ln   (15)
60 t1 t 1 t 
(c) In designing the process, care should be taken
such that the temperature-strain rate trajectory where N = speed of the rolls (16 rpm), R =
adopted for the control does not enter either the undeformed roll radius (210 mm), t 1=thickness of
instability regime or cracking domains. plate after rolling (15 mm) t = thickness strain (to-
t1)/to, to= initial thickness (25 mm).
(d) When the temperature range for DRX domain is
very narrow or the strain rate range is slower than The calculated average strain rate is about 2 s-1. As
0.01s-1, it is necessary to design an isothermal process marked on the processing map (Fig. 1.13) the rolling
where the die or tools are heated to the same may be started at 1150°C and finished at 1000°C . At
temperature as that of the workpiece. Alternately, it is lower temperatures flow instability starts when the
helpful sometimes to enclose the workpiece in a can strain rate is 2 s-1. However, the hot shearing process
made out of a suitable material, for example, pure may be done in the instability regime since it forms
aluminum for aluminum alloys or stainless steel for adiabatic shear bands which facilitate crack
superalloys. propagation and fracture, to complete the shearing
process.
(e) For materials which have limited intrinsic
workability (narrow workability domain) like metal- (B) Continuous Rolling of Aluminum
matrix composites (MMCs), intermetallics or PM
Electrical conductor wire rods are generally produced
compacts, it may be necessary to optimize the state-
by the continuous rolling process and the wire rods
of-stress workability with the help of proper die or
are expected to meet certain ductility levels
preform designs. For example, streamlined dies are
consistently. This property requirement may be
useful for extruding MMCs.
achieved by the application of processing maps for
(f) Taking advantage of the effect of prior processing the process design.
history on the processing maps, connected processes
The processing map for commercial aluminum is
may be designed. Each step of the process may refine
shown in Fig. 1.15. The optimum temperature and
the grain structure and expand the workability
strain rate parameters are 600°C and 0.001 s-1. This
domain for the next step of processing while moving
strain rate is too slow for a manufacturing process
it to higher strain rates without the onset of
and hence the rolling has to be done away from the
instabilities to facilitate higher productivity.
optimum values. Continuous rolling using a 8-stand
The following typical examples illustrate the above mill has the characteristics shown in Fig. 1.16, where
aspects: in the first stand the material is rolled at 510°C and
1.5 s-1 and in the last stand the parameters are 200°C
(A) Hot Rolling of HSLA Steel and 25 s-1. The wire rod is then coiled at temperatures
Microalloyed steel (High Strength Low Alloy steel) lower than 200°C. With these process parameters, the
is hot rolled into plates which require close control of ductility of the material was lower and inconsistent
microstructure and properties. Often the yield in which may be attributed to the instabilities occurring
rolling is reduced due to rolling defects. The at temperatures lower than 250oC and strain rates
processing map may be used to optimize the hot higher than about 1 s-1. To improve the process, it is
rolling process. necessary to conserve heat between stands and raise
the temperature in the last stand to 250oC so that
The processing map for the microalloyed steel (Fig. instability regime is avoided. The suggested new
1.13) indicates that the optimum temperature is schedule is indicated in Fig. 1.15 and processing as
1090°C and the workability range is 1000-1150°C. per this new design higher and consistent ductility is
The optimum strain rate is 0.3 s-1. For the particular achieved.

19
Hot Working Guide

Fig. 1.15 Processing map for commercial

aluminum (99.5% pure). Contour numbers
represent percent efficiency of power
dissipation. Shaded regions correspond to
flow instability. Dashed line corresponds to
rolling path followed in plant and solid line is
the suggested optimum route.

Fig. 1.16 Continuous rolling mill characteristics.

20
 Introduction

Fig. 1.17 Processing map for

IN600 alloy. Contour numbers
represent percent efficiency of
power dissipation. Shaded region
corresponds to flow instability.

C) Forging of IN600 Superalloy Fig. 1.17 which matches well with the indications
from the map. It may be noted that the processing
In this example, the commercial practice of forging window is away from flow instabilities.
IN600 (Ni-Fe-Cr alloy) is compared with the results
of the processing map for this alloy.
(D) Extrusion of 304L Stainless Steel
The processing map for IN600 is shown in Fig. 1.17.
The optimum processing parameters are 1200 oC and Hot extrusion of 304L stainless steel is used to
0.3 s-1. As per the map, the ingots may be cogged in a manufacture tubes for critical structural applications.
forging press at a starting temperature of 1232 oC and The productivity is often low because of surface
during the last step, the temperature is about 1100 oC defects produced in the extrusion. Optimization of
which produces a fine grained structure. The the process using processing map is described in this
finishing temperature is 950oC which will avoid example.
nonuniform grain structure from surface to center due The processing map for 304L stainless steel is shown
to temperature gradients and very fine grained in Fig. 1.18 according to which the optimum hot
structures are produced due to the presence of working temperature is 1100oC and the strain rate is
carbides in the matrix. The average strain rate for the 0.1 s-1. Actual extrusions were performed using an
cogging operation may be in the range 0.1-1 s-1. The extrusion ratio of about 15:1 using different
forging schedule described by Johnston [32] is combinations of processing parameters given in
marked on the processing map for IN600 given in Table I.

21
Hot Working Guide

The mean strain rates (  ) are calculated using the Extrusion trials were also performed on Zircaloy-2
equation [33] using processing maps and the results similar to the
above were documented [34].
6Vo D 2o tan 
  ln( ER ) (17)
D 3o  D 3f Table I Extrusion data for 304L stainless steel

where Vo = ram speed, Do = initial diameter, Df =

Sample Temperature, Ram Mean
final diameter, ER = extrusion ratio and  = die
No. o
C speed, strain
semiangle.
mm/sec rate, s-1
From the processing map (Fig. 1.18) it may be noted
that the sample "a" is extruded under conditions of flow a 1080 300 5.6
instability while the others are in the workability domain b 1150 200 2.8
though not at their optimum. Extrusion "a" exhibited
surface bands and cracks (Fig. 1.19) while extrusion "b", c 1120 180 2.4
"c" and "d" gave superior mechanical properties,
d 1090 160 2.1
although "d" is the best of the three. It may be noted that
the extrusion temperature is close to the optimum value
for "d".

Fig. 1.18 Processing map for

304L stainless steel. Contour
numbers represent percent
efficiency of power dissipation.
Shaded region corresponds to
flow instability.

22
 Introduction

Fig. 1.19 A view of the cut bits of

extruded tubes of 304L stainless
steel. The processing conditions of
the tubes (a)-(d) are given in Table I.

Fig. 1.20 Processing map for hot rolled

Zn-Pb alloy. Contour numbers represent
efficiency of power dissipation. Shaded
region corresponds to flow instability.

23
Hot Working Guide

(E) Impact Extrusion of Zn-Pb Alloy

Cans for dry cell application are manufactured by temperature of about 175oC at 100 s-1 and is
impact extrusion of Zn-Pb alloy at ambient responsible for thickness variations in the extruded
temperatures. The cans are often observed to have can. It is therefore desirable to conduct the impact
non-uniform thickness which may cause leakage of extrusion at temperatures higher than 175oC . Plant
batteries. The punch in impact extrusion is of primary trials have shown that cans with uniform thickness
importance and its enhanced life is beneficial in and bright surface finish may be manufactured by
manufacturing. The average strain rate for this this process with an additional advantage of
process is estimated to be about 100 s-1. In this improved punch life since the loads were lower.
example, the application of a processing map to
optimize this process is discussed. The processing
map for Zn-Pb alloy is given in Fig. 1.20 which
shows the workability window at 250-300oC and 100
s-1. However, flow instability occurs up to a

Fig. 1.21 Processing map for 2124-

20v/o SiCw composite. Contour
numbers represent percent efficiency
of power dissipation. Shaded portion
corresponds to flow instability.

24
 Introduction

(F) Extrusion of MMC Structurals

Metal matrix composites with 2124-20 v/o SiCw are 400-550oC and in the strain rate range 0.0001-0.001
used for aerospace structural components for greater s-1. The strain rate range is too slow for
specific stiffness and are manufactured by extrusion. manufacturing and hence higher strain rates will have
In this example, the design of the extrusion process to to be used but not too high to enter the instability
achieve the required properties in the structural regime. The optimum average strain rate of 0.01 s-1
shapes is described. For obtaining the stiffness was selected for extrusion at 500oC using
properties, it is essential that the aspect ratio of the streamlined dies to produce structural sections
whiskers is maintained as high as possible by including T, Z and wingspar caps. A comparison of
preventing their breakage during processing. For this the extrusion produced by shear, conical and
purpose, the matrix should be kept in a soft condition streamlined dies is shown in Fig. 1.22. Successful
and the rotational flow in the extrusion die is extrusion using optimum design of streamlined die is
avoided. For this purpose, the intrinsic workability as shown in Fig. 1.22(a), which clearly demonstrates the
well as the state of stress workability will have to be application of processing maps in process design for
optimized, the former by using a processing map to this composite material.
select a T-  window for processing and the latter by
using a streamlined extrusion die design [4]. The
processing map for this MMC is shown in Fig. 1.21
which shows a DRX domain in the temperature range

Fig. 1.22 T section extrusions

of 2124-20v/o SiCw composite
using (a) streamlined die
designed using FEM
simulation, (b) streamlined die
of non-optimum length, (c)
conventional die design used
in industry (shear die) [4].

25
Hot Working Guide

(G) Strain-Inducted Porosity in ELI in hydraulic press forging, the hot deformation
grade Ti-6Al-4V mechanism is large-grained superplasticity (LGSP),
which nucleates pores at grain boundary triple
Cogging of ELI grade Ti-6Al-4V with starting β- junctions which lower the tensile ductility
transformed microstructure is conducted to produce (Fig.1.24(a)). The pores then grow during
the blanks for engine bulk-heads in aircraft. When intermediate annealing steps in order to relieve the
press forging is done at a temperature of 962 oC, residual tensile stress component generated at the
porosity is generated at mid-plane of the slab and that mid-plane during forging. Typical pore morphology
is considered to be unacceptable. This problem has is shown in Fig. 1.24(b). A large population of such
been analysed [35,36] using a processing map pores are generated during forging and annealing
developed for ELI Ti-6Al-4V for which the (α+β) cycles causing unacceptable porosity in the plate.
→β transus temperature is in the range 960-985 oC. The solution to this problem is to reduce the forging
The map is shown in Fig. 1.23 (a) and its temperature to 925 oC so that DRX (globularization)
microstructural interpretation is shown in Fig. 1.23 is the hot working mechanism.
(b). When forged at 962 oC at slower strain rates as

ASB
Cracking Flow
Localization

Lamellae
Kinking
Void
Nucleation

Large-Grained
Shear Superplasticity
Globularization
Cracking

(a) (b)
Fig.1.23. (a) Processing map for ELI Ti-6Al-4V β-transformed preform microstructure. (b) Metallurgical
interpretation of the processing map.

26
 Introduction

(a) (b)
Fig.1.24 (a) Variation of tensile ductility with temperature in ELI Ti-6Al-4V β-transformed microstructure. (b)
Typical pore morphology at triple junction.

(H) Microstructural Control at [39], AZ31 [40,41], and AZ31-Nano Alumina

Bifurcations composite [42]. The example of rolled AZ31
magnesium alloy compressed parallel to Transverse
The processing map in general consists of Direction (TD) is presented here to illustrate the
deterministic domains where the kinetic rate equation effect of proccessing in the bifurcation region. The
is obeyed and regimes where flow instability occurs. processing map for AZ31 TD is shown in Fig. 1.25
The domains are not separated sharply but over a (a). The map exhibits two domains - one in the lower
region which is referred to as a "bifurcation" similar strain rate range and the other at higher strain rate
to that observed in other dyanmical systems [12,37]. range, and these two are separated by a bifurcation
The situation at the bifurcation may be explained in occurring at strain rates around 0.1 s-1. The two
simple terms where the dissipative energy is highest domains represent dynamic recrystallization which is
in the bifurcation region and lowest in the domains controlled by lattice self diffusion at lower strain
on either side and therefore the bifurcation has a rates and grain boundary self diffusion at higher
saddle-point configuration. Any fluctuations in the strain rates. The grain size variation with strain rate
applied parameters in the bifurcation region can at a temperature of 450 oC is shown in Fig.1.25 (b)
cause abnormal changes in the dissipative energy along with that of efficiency of power dissipation. At
state, which in the present case is represented by the the strain rate of 0.1 s-1, the efficiency of power
microstructure. Microstructural control is difficult if dissipation reaches its lowest point (or highest
the material is hot worked in these regions where dissipative energy) indicating a change-over or
probability plays an important role [12,37]. For bifurcation point. At this point, the grain size reaches
example, the grain size may record abnormal grain an abnormally large value. If processed around this
growth or the microstructure may consist of a point, any small change in temperature or strain rate
mixture of large and small grains which may not be will result in unpredictable distribution of grain sizes
reproducible or predictable. This aspect is observed leading to a problem in microstructural control in the
in a number of alloys like Ni3Al alloy [38], FeAl product.

27
Hot Working Guide

(a) (b)
Fig.1.25.(a) Processing map for rolled AZ31 compressed parallel to Transverse Direction. (b) Variation of grain
size and efficiency of power dissipation with strain rate.

1.7 Caution! transformation temperature, and other metallurgical

effects since these will be reflected in the map in
A Processing map is a powerful tool for the design terms of inflexions in the contours, and ending or
and optimization of metal working processes and is expanding of domains and their movement with
very effective in achieving microstructure and respect to the strain rate axis. Prior knowledge of
property control. However, it is very important to physical metallurgy of the system offers a good way
observe great caution in applying the map to the to cross check with the features of the map and often
industrial processes. The following precautions may some new effects caused by the dynamics of
be strictly observed in developing and applying the deformation are revealed. Furthermore, the
results of processing maps. dissipative microstructures respond to the post
Since the processing maps are only as good as the deformation heat treatments differently and may
data input, considerable care has to be exercised in result in a new set of innovative microstructures. The
generating the experimental data. For example, maps of materials that exhibit very low intrinsic
testing has to be done under constant true strain rate workability (extensive cracking regimes) are less
conditions, and the correction for the adiabatic accurate since input flow stress data itself will not be
temperature rise has to be incorporated while arriving accurate, and so extra care must be exercised during
at the flow stress values. Another important aspect is their interpretation. Finally, it must be emphasized
the characterization of the processing history, prior that the domains and instability regimes exhibited in
heat treatment, and initial microstructure. In the case a map should be validated by detailed microstructural
of as-cast materials, homogenization treatment has a examination before the results are implemented.
significant impact on the processing map and so do
the differences in the initial microstructure like cast (a) As the map is generated on the basis of
(dendritic or cellular) versus wrought, powder experimental data, it will be only as good as the input
metallurgy compacts, and lamellar versus equiaxed data. Hence, great care should be exercised while
microstructures in the case of some two-phase conducting experiments to obtain stress data. A wide
materials. It is also useful to have knowledge of range of matrix of temperature and strain rate may be
solutionizing temperature, particle precipitation, chosen and a minimum of 5x5 matrix is preferable.

28
 Introduction

(b) The map is sensitive to the chemistry, prior strain becomes negligible. It is generally observed
mechanical and processing history and heat treatment that the effect of strain in hot working may be safely
given to the material, and it is dangerous to extend neglected in most of the materials.
the information to materials of unknown ‘history’
(e) Industrial processes are too complex to model
although they may belong to the same class of alloys.
accurately. Hence, it is always advisable to use a
(c) Although the workability domain (DRX) is not laboratory scale validation of the design before
greatly affected by the state-of-stress in the extending it to the actual manufacturing process. The
processing zone, the cracking and instability regimes standard physical modeling techniques may be
are sensitive. For example, the cracking regime will adopted for this purpose [43].
be wider in the regions of high hydrostatic tensile
(f) While attempting to solve an industrial problem or
components while shear stress may enhance the flow
optimize an existing process, it is useful to develop
localization. Hence, SOS modeling which is specific
the material model first before optimization of the die
to process will be useful and can be accurately done
design or geometry aspect is attempted. This
using FEM simulation techniques.
sequence has an added advantage that the actual
(d) In materials whose microstructure drastically constitution equations required for the realistic
changes with strain, the maps respond sensitively to process simulation are available from the material
the strain value in the initial stages of deformation. model.
However, once a steady-state is reached, the effect of

References: [13] Y.V.R.K.Prasad and T.Seshacharyulu, Inter.

Mater. Rev. 44(1998)243.
[1] G.E. Dieter, Metals Handbook, Vol.14, ASM., [14] Y.V.R.K.Prasad, J. Mater. Eng. and
Metals Park, OH(1987)363. Perf.12(2003)638.
[2] H.A. Kuhn, Metals Handbook, Vol.14, ASM, [15] K.P. Rao and Y.V.R.K. Prasad, Handbook on
Metals Park, OH(1987)388. Comprehensive Materials Processing Technology,
[3] S. Kobayashi, S.I. Oh and T. Altan, Metal Forming Editor-in-Chief: S. Hashmi, Vol.3, Advanced
and The Finite Element Method, Oxford University Forming Technologies, Ed. S.T. Button, p.397-
Press, New York, NY (1989). 426, June 2014, Elsevier.
[4] H.L. Gegel, J.C. Malas and S.M. Doraivelu, [16] L.E. Malvern, Introduction to the Mechanics of
Innovations in Materials Processing, Eds. G. Continuous Medium, Prentice-Hall, Eaglewood Cliffs,
Bruggerman and V. Weiss, Plenum Press, New York NJ (1969)197.
(1985)137. [17] W. Roberts, Deformation Processing and
[5] H.L. Gegel, J.C. Malas, S.M. Doraivelu , .and V.A. Structure, Ed.G. Krauss, ASM, Metals Park, OH
Shende, Metals Handbook, Vol.14, ASM, Metals Park, (1984)109.
OH (1987)417. [18] T. Sakai and J.J. Jonas, Acta Met. 32(1984)189.
[6] J.J. Jonas, C.M. Sellars and W.J.McG. Tegart, [19] H.J. McQueen, E. Evangelista, and N.D. Ryan,
Met.Rev. 14(1969)1. Proc. Recrystallization, Ed. T. Chandra, TMS-AIME,
[7] H.J. Frost and M.F. Ashby, Deformation Warrendale, Pa. (1990)89.
Mechanism Maps, Pergamon Press, London (1982). [20] Y.V.R.K. Prasad and N.Ravichandran,
[8] Rishi Raj, Met.Trans. 12A(1989)1089. Bull.Mat.Sci. 14(1991)1241.
[9] Y.V.R.K. Prasad, H.L. Gegel, S.M. Doraivelu, J.C. [21] J.J. Jonas, X. Quelennec, L. Jiang, E. Martin, Acta
Malas, J.T. Morgan, L.A. Lark and D.R. Barker, Mater. 57(2009)2748.
Met.Trans. 15A(1984)1883. [22] A.K. Mukherjee, Plastic Deformation and
[10] P.E. Wellstead, Introduction to Physical Systems Fracture of Materials, Ed.H. Mughrabi, Materials
Modeling, Academic Press, London (1979). Science and Technology, Vol.6, VCH
[11] H. Ziegler, Progress in Solid Mechanics, Vol.4, Verlagsgesellschaft mbH, West Germany (1993).
Eds. I.N. Sneddon and R. Hill, John Wiley and Sons, [23] P. Griffiths and C. Hammond, Acta Met.
New York (1963)93. 20(1972)935.
[12] I. Prigogine, Science 201(1978)777.

29