APSC 279

Introduction to Lab 2: Deformation and Recrystallization

Work Hardening

During the plastic elongation of steel or brass in lab 1 we saw that in the plastic deformation
region the materials became stronger as they were deformed. (There was a drop in engineering stress
with strain past the point where a neck formed, but true stress continues to rise up to failure.) This
property can be used to increase the strength of a metal or alloy. This is illustrated in Figure 1. On the left
a sample is plastically deformed to some point then the load is removed. The elastic strain is recovered as
indicated by the downward pointing arrow, but the sample retains its plastic elongation. Now if the
sample is pulled again as shown at right, it follows the same path of elastic extension to a new, higher
yield point (y2) as indicated at right. The material thus has higher yield strength.

y2

y1 y1
Stress

Stress

Strain

Elastic Recovered Strain

Figure 1. Illustration of increase in tensile strength of a ductile metal sample after linear plastic
deformation.

In work hardening (also called cold working and strain hardening) a metal or alloy is mechanically
deformed, for example by passage through heavy duty rolls (rolling) to reduce its cross sectional area.
Other methods include extrusion and drawing, which also reduce cross section, and forging. As a result
the yield strength and ultimate tensile strength increase. (Note that the stresses now are not just uniaxially
tensile. The tensile stress-strain curve illustrates the point that yield strength increases with plastic
deformation. But, in rolling, extrusion, etc. the ultimate tensile strength also is increased.) The reference
to cold work comes from the fact that if a pure metal is at a temperature less than about 40% of its melting
point (on the Kelvin scale) the grains (microstructure) remain intact after deformation. If the temperature
is greater than about 40% of the melting point of the metal, the grains can (in principle) rapidly revert

1
back to their original size and characteristics. Hot working then changes dimensions without changing
grain structure and strength characteristics of the deformed material as compared to the material fed to
the machine. For an alloy the transition from cold to hot working may differ from 40% of the melting
point.

During cold working the grains of the metal or alloy are deformed. It the sample is rolled, the
grains get elongated along the direction of rolling. This is illustrated in Figure 2. The grains retain about
the same volume. This deformation creates many more dislocations within the grains; the crystals have

Figure 2. Illustration of how cold working deforms grains. The main point is that after cold working initially
more or less equiaxed grains (similar dimensions in all directions) become misshapen – elongated.

been massively deformed at the atomic level. This can be observed in micrographs of strained metal
samples. For the deformed material the movement of dislocations under load depends on their
orientation relative to the load direction and the crystallographic planes in which they are located.
Crystallographic planes with a high packing density of atoms move past each other more easily than those
with a low packing density of the atoms. Hence dislocations that don’t move easily acts as barriers to
those that could otherwise move easily. When a grain is plastically deformed the number of dislocations
per unit volume (dislocation density) increases with increasing extent of plastic deformation. (For a highly
deformed metal dislocation densities can increase by factors of up to 106-107 times.) This in turn will create
increasing numbers of dislocations that act as obstacles to dislocation movement under load. There is a
lot more to it than this, and the effects are quantifiable, but this provides a suitable rationale for the
strengthening effect of cold work. Figure 3 schematically illustrates the effects of cold work on yield
strength and ultimate tensile strength.

The extent of deformation is indicated by extent of reduction of cross sectional area:

%CW = 100  (Ao – A)/Ao (1)

where %CW is percent cold work, Ao is the initial cross sectional area and A is the area after deformation.
If the material is rectangular in section and fairly thin then the width does not increase very much upon
reduction of thickness. Then the extent of cold work can be estimated from the change in thickness.

%CW = 100  (to – t)/to (2)

2
where to is the initial thickness and t is the thickness after rolling. If the thickness and width are similar
(e.g. a square rod or a cylindrical wire) then equation 2 does not apply.

Increasing %CW

Stress

Strain

Figure 3. Illustration of the effects of cold work on yield strength and tensile strength.

A quantitative example of how cold working affects tensile properties is shown in Figure 4. This
figure correlates the yield strength, ultimate tensile strength and ductility (% elongation) as a function of
%CW. Note that initially there is a large difference between yield strength and ultimate tensile strength.
This gap narrows as the extent of cold working increases. For some metals and alloys the yield strength
eventually approaches the ultimate tensile strength at high %CW. Then the material fails shortly after
yielding. It starts to become almost brittle.

Strain hardening is often used in fabrication processes to improve mechanical properties. It is also
a cost-effective means of strengthening a metal or alloy since no costly alloying elements are needed. An
example is cold rolling of steel, which can increase the yield strength by a factor of 10. As can be seen
from Figure 3, there is a price to pay for increased strength; ductility and toughness decrease as %CW
increases. Electrical and thermal conductivities as well as corrosion resistance are also affected by strain
hardening.

Hardness

As the term work hardening implies, cold work also affects the hardness of the metal or alloy. As
%CW increases, so does the hardness of the material. Hardness is the ability of a material to resist localized
plastic deformation, like indentation or a scratch. Obviously this is closely related to yield strength.
Hardness is measured by using an indenter which impinges on a surface under a specified load and at a
specified rate. The depth and size of the indentation is related to the hardness of the material. Obviously

3
the lower the depth of indentation, the higher the hardness. Hardness values for a given technique are
relative measures, not absolute. Different techniques for measuring hardness are not necessarily directly
comparable either. Some methods create an indentation, the size of which is then determined under a
microscope.

Figure 4. Relationship between yield strength, ultimate tensile strength and ductility (% elongation) with
cold work for a low carbon (<0.06%) high nickel-chromium-molybdenum steel. Retrieved from:
http://cartech.ides.com/datasheet.aspx?i=103&e=213&FMT=PRINT Feb. 16, 2016, Technical datasheet,
20Cb-3 Stainless.

Rockwell hardness tests are the most common since the tests are easily performed. The extent of
indentation is determined directly by the instrument without the need for microscopic examination.
Hence this method is quick and well suited to production operations where measurements need to be
quick. The indenters may be hardened steel balls ranging in diameter from 1/16 inch to 1/2 inch. There
are also tungsten carbide indenters and diamond indenters for very hard materials. The Rockwell hardness
test utilizes a small preload followed by a larger major load. The preload establishes a zero point. The
difference between the two indentations represents the hardness. Preload may use loads that correspond
to 3 kg (superficial Rockwell hardness) or 10 kg (regular Rockwell hardness, or just Rockwell hardness)
masses. In this lab we use the latter. Various scales use differing larger final loads. There are several scales
depending on the material to be tested. A few examples are listed in the table below. Note that the masses
used are not actually 60 kg, etc. These are a proxy for the applied forces. The maximum on the various
Rockwell scales that use steel balls is 130; 100 for scales using diamond indenters. Every decrease in depth

4
of indentation by 0.002 mm corresponds to an increase in hardness of 1 unit. The hardness on any
particular scale then is given by:

Rs = M – h/0.002 (3)

where Rs is the Rockwell hardness scale in use, M is the maximum reading (130 for steel ball indenters)
and h is the difference in depths of penetration (mm) for major load and preload. Hence an indentation

Table 1. Some common Rockwell hardness scales that use 10 kg preload.

Rockwell scale Indenter Major load (kg)

h of 0.16 mm corresponds to a hardness number of 50 for a scale using a steel ball. Diamond indenters
have sharp points suitable for harder materials. Higher major loads are also needed for harder materials.
Obviously a very hard material will need a diamond indenter with a large major load to achieve the same
depth of penetration as a soft material using a steel ball and a lighter major load. In this lab we will be
using brass, which is a fairly soft copper-zinc alloy and the Rf scale will be used. Hardness values below
about 20 and above 100 (when M = 130) become increasingly inaccurate. For instance, at low hardness
values the hardness is obtained from a small difference between two similar numbers, each with its own
uncertainty, and measuring small indentations has a large relative error. The resulting uncertainty for the
difference then is relatively large. A common source of error is strain hardening of the material during
indentation, and this may be part of the reason why high hardness values are less accurate. Errors can
also occur if the sample being tested is too thin, if the steel ball becomes non-spherical, if the test is
performed too close to the edge of a sample or if indentations are made too close to each other. Another
common problem can occur if the sample is curved. If the sample is placed so that the indenter impinges
on the concave-down surface then some of the applied load may act to push down on and deform (flatten)
the surface. For any given sample several measurements should be taken and averaged after examining
the data for outliers. Each indentation should be at least three indentation diameters away from others.
If the sample is curved, it should be placed so that the indenter comes down on the surface concave-up,
although the sample should be as flat and smooth as possible. The accuracy of a hardness tester and
indenter can be checked using standard samples of known hardness on the specified scale.

Hardness is often used as a quality control measurement to ensure a particular processing method
is giving the right results. Hardness may also correlate with other mechanical properties such as ultimate
tensile strength or yield strength and thus be a suitable proxy for them. Hardness values obtained on one
Rockwell scale can be converted to other Rockwell scales. Interconversion between differing scales of
different techniques depends on the material. Textbooks on the subject can be consulted for this.

5
Annealing

Annealing is a word for a heat treatment that restores grains from deformed (by strain hardening)
to equiaxed grains again. The process requires increased temperatures to well above the temperature
where the cold working process was done. This is a complex process, but it can be broken down into three
sequential steps. These are illustrated in Figure 5. The first stage is called recovery. In this stage there is
little visible change in the grains. They are still the same deformed shape, for the most part. In the
deformed grains there is strain energy that was built up by the cold working process. There are a large
number of dislocations and the strain energy resides herein. Many of the atoms are out of their most

(a) (b) (c) (d)

Strength
Tensile strength

Ductility

Ductility

Recovery Recrystallization Grain growth

Figure 5. Illustration of the stages in annealing a cold worked sample and the effects on tensile strength
and ductility. The small black spots in part (b) indicate new crystal nuclei. The figure indicates changes
with temperature for specific times of a sample at temperature.

favourable crystallographic positions and this required energy input to achieve. In the recovery stage due
to the increased temperature there is enhanced atomic diffusion so some of this strain energy is relieved.
However, much of the strain energy is still present. Tensile strength decreases slightly; ductility increases
slightly. One author states that the large grains form numerous smaller sub-grains, while these aggregates
still together retain the shape of the strained grains. In terms of physical properties affected by the
recovery stage, electrical and thermal conductivities return to the pre-strained values.

6
 In the next stage recrystallization begins. Tiny new grains emerge within the large strained grains.
This is called nucleation. These very small grains (nuclei) grow at the expense of the large, deformed
grains. The driving force for this is alleviation of the unstable, high-energy strained states of the deformed
grains. High temperature makes it happen at a reasonable rate. At low temperatures where the material
was cold worked this driving force also exists, but, the activation energy needed to allow the process to
occur at a measurable rate is too high. It is the high temperatures that facilitate this. At first a new set of
fully recrystallized small grains is produced as in part (c) of the figure. During recrystallization ductility
begins to increase and tensile strength starts to decrease, beginning to approach the pre-deformed values
for the material as recrystallization approaches completion. Hardness follows tensile strength, decreasing
along with it. The new set of fully recrystallized grains is smaller than the original pre-deformed grains.
Ductility will still be somewhat less than that of the original unstrained sample; tensile strength will still
be somewhat higher.

The more strained the grains are, the more likely they are to undergo recovery and nucleation to
form the new nuclei that will become the recrystallized grains. In most metals there is a minimum extent
of cold work below which recovery and nucleation will not occur (ranging from 2-20 %CW). Below the
threshold the built-up strain energy is too low to drive recrystallization.

With increasing temperature, or further time at a high enough temperature grain growth occurs.
The small grains continue to grow; larger grains consume smaller grains. This is illustrated in part (d) of
the figure. In this stage there is a moderate increase in ductility and a moderate decrease in tensile
strength and hardness. Note that grain growth does not require prior recovery and recrystallization
stages. Grain growth can occur in any polycrystalline material. Grains have surface energy. This can be
lowered by lowering the surface area per unit volume, that is, by growth of the crystals. (Thermodynamics
favors reduction of surface area.) This necessarily occurs at the expense of the smaller crystals.

The stages are not perfectly distinct; there is some overlap. The phenomena are kinetic in nature,
governed by rate equations that depend on extent of cold working, temperature and time. Some degree
of recovery continues into the recrystallization stage; recrystallization may be occurring to some small
degree during recovery; grain growth will begin to occur during recrystallization. The recrystallization
temperature is conventionally defined as the temperature at which recrystallization is just complete after
one hour. As the preceding discussion suggests this temperature will also depend on the extent of prior
cold work. Recrystallization is more rapid in pure metals than in alloys; the recrystallization temperature
is higher in an alloy than in the pure metal constituents, or its main constituent. For pure metals the
recrystallization temperature is about 0.3Tm, where Tm is the melting point on the absolute temperature
scale (Kelvin). For alloys the recrystallization temperature can be up to 0.7Tm.

The formation of new crystals involves two distinct processes, as indicated above. The first is
nucleation and the second is crystal growth from these nuclei. The nuclei form, then they grow into the
new set of crystals by consuming the material from the strained grains. (This growth process is distinct
from the grain growth process indicated in Figure 4d, which is growth of bigger crystals from the smaller,
already recrystallized ones.) Nucleation is the spontaneous formation of a new microscopic crystallite.
Nucleation and crystal growth both have their own separate rates. If the nucleation rate is large relative
to the crystal growth rate then there will be many nuclei available and the final result will be a large
number of small crystals or grains. This appears to be what is happening with the brass samples used in
Lab 2. And perhaps is often the case with recrystallization of cold worked metals and alloys. (It need not

7
be the case with phase transformations from one solid phase into another.) If this pertains, then we can
say that the recrystallization rate of the strained grains is governed by the crystal growth stage; it will be
slowest and therefore the rate-determining step. (We probably can’t completely separate the two
processes, but to a first approximation this will be true.) Crystal growth rate is indicated by the expression:
dFx Q⁄
Rx = dt
= Ae− RT (4)

where Rx is the recrystallization rate indicated by dFx/dt, i.e. the rate of change of fraction recrystallized
with time, Fx is the mass fraction of the material that has been recrystallized, A is a constant (In chemical
kinetics it is called the frequency factor or pre-exponential factor. It can be thought of as related to the
frequency of collisions or interactions between reacting chemical species.), Q is the activation energy in
J/mol (Q is often given in more convenient units of kJ/mol and must be converted to J/mol.), R is the ideal
gas constant (8.314 J/mol K) and T is the absolute temperature (K). A in this case has units of time -1. In
chemical kinetics equation 4 is essentially the Arrhenius equation, except that it is the rate constant that
is equal to Ae-Ea/RT. But, in chemical kinetics the rate refers to changes in concentrations of chemical
species with time. Fraction recrystallized makes more sense in this context. We can write an equation for
the fraction recrystallized with time:
F t −Q⁄ −Q⁄
∫0 x dFx = Fx = ∫0 Ae RT dt = A (e RT ) t (5)

−Q⁄
Fx = A (e RT ) t (6)

since A, Q, R and T are either constants or independent of time. Again, this equation applies so long as the
nucleation rate is rapid compared with crystal growth rate. Then this equation can be used to estimate
times for different extents of recrystallization at specified temperatures, so long as A and Q are known.
These parameters can be found from kinetics studies of recrystallization rates.

Hot Working

When deformation occurs at temperatures above that at which recrystallization occurs the
process is called hot working. During hot working extensively strain-deformed grains rapidly recrystallize
so there is not net strain deformation. The grain structure after deformation is the same as that before.
Hot working requires much less mechanical energy, since the material is much more easily deformed than
in cold working. In hot working large deformations are readily achieved. Generally metals and alloys have
to be heated to quite high temperatures in order to be hot worked. This often results in significant surface
oxidation. This may result in material losses and a poor quality surface finish. Cold working allows for
better control of final dimensions, surface finish and mechanical properties. An example where hot
working is practiced is in steel mills for forming large coils of steel sheet. A hot steel billet is rapidly rolled
to a thin and very long sheet that is then rapidly coiled. The coiled steel then is shipped to customers.

Creep

Alloys may be put in service at high temperatures while being subject to load. A good example is
turbine rotors in jet engines and power generators where the vanes are subject to high temperatures and

8
centrifugal forces. The constant high load at high temperature results in gradual deformation called creep.
This may limit the service life of a part in such applications. Creep may occur in almost all materials. For
metals it becomes significant when the temperature exceeds 0.3 - 0.6 of Tm. An intriguing example of
creep is a tungsten filament in an incandescent light bulb. The filament may be as hot as 3300 K. Under
its own weight the filament can sag and the wire coils can come into contact, causing a short circuit and
failure. For metals a creep test can be performed by hanging a mass on a fixed tensile sample and
measuring the strain (called creep strain) with time. A plot for a typical test is shown in Figure 6.

Upon application of the load there is an immediate deformation which is mainly elastic in nature.
Then three discrete stages follow. Note that the strain reported here is engineering strain. The sample
narrows in cross sectional area with time, and in some cases a neck may form. In the first stage strain
decreases with time. This is attributed to strain hardening. At first glance this seems contradictory to the

Rupture

Primary
Creep strain e

Secondary
Tertiary
Instantaneous
deformation

Time tr

Figure 6. Illustration of strain versus time for a creep test. The three stages are indicated. Time to failure
(rupture) is indicated as tr.

fact that above about 0.4Tm deformation by hot working would be expected. However, recall that
recrystallization requires a significant amount of strain to be introduced in order for there to be sufficient
driving force for the process to begin. Evidently the amount of strain energy introduced by creep in this
stage in metals is not enough for them to undergo recrystallization, despite the high temperature.

The next stage is secondary creep or steady-state creep. The deformation with time is often linear.
This is also often the longest stage, depending in part on temperature. (The creep rate accelerates with
temperature.) The linear behavior is attributed to a balance between strain hardening and recovery (the
same recovery discussed above). Recall that recovery acts to relieve some of the strain energy. Again, the
strain energy that occurs with creep is probably small enough to prevent recrystallization, but, apparently
a suitable measure of recovery does occur.

In the third stage the creep rate increases with time. There may be several reasons for this. Grain
boundaries may separate and cracks or voids may form in the metal. If a neck forms then the cross
sectional area decreases significantly and the true stress then is greater than the engineering stress.

9
Indeed, the formation of voids and cracks also represents a decrease in cross sectional area. For some
metals/alloys when true stress is plotted against time there is no tertiary rise prior to rupture. Then the
true stress is close to constant from the end of the primary stage to rupture.

When temperature or load are increased the initial instantaneous deformation is greater, the
slope of the linear secondary stage is greater, the secondary stage is shorter and rupture occurs sooner.
The steady-state creep rate (secondary stage) may often be described by an empirical relationship:

es = K1n (7)

where es is the steady-state creep strain (engineering strain), K1 is a constant that depends on the material
and temperature,  is the engineering stress, and n is likewise a material constant. If a plot of loges vs.
log is linear then the slope is the value of n. Thus,

lnes = lnK1 + n ln (8)

and a plot of lnes versus ln is linear with a slop of n and an intercept of lnK1.

References

The following books provide different emphases on the topics presented in this document. The
approaches are helpfully complementary.

1. W.D. Callister Jr., Fundamentals of Materials Science and Engineering, 2nd ed., Wiley, 2005, Chpt. 8 and
9, or see the same topics in later editions of the same book, W.D. Callister and D. G. Rethwisch, Materials
Science and Engineering, an Introduction. Figures 1, 3 and 5 adapted from this reference.

2. J. Newel, Essentials of Modern Materials Science and Engineering, Wiley, 2009, Chpt. 4.

3. C. M. Gilmore, Materials Science and Engineering Properties, 2015, Cengage Learning (through Nelson
in Canada), Chpt. 7 and 9.

4. N. E. Dowling, Mechanical Behavior of Materials, Prentice Hall, 1993, pp. 4-5, 44-45, 117-118 and 680-
687. Figures 2 and 6 adapted from this reference.

10