DTU Wind Energy Masafumi Azuma

Downloaded from orbit.dtu.
dk on: Nov 11, 2020
Structural control of void formation in dual phase steels
Azuma, Masafumi
Publication date:
2013
Document Version
Publisher's PDF, also known as Version of record
Link back to DTU Orbit
Citation (APA):
Azuma, M. (2013). Structural control of void formation in dual phase steels. DTU Wind Energy. DTU Wind
Energy PhD, No. 0015
General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright
owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.
 Users may download and print one copy of any publication from the public portal for the purpose of private study or research.
 You may not further distribute the material or use it for any profit-making activity or commercial gain
 You may freely distribute the URL identifying the publication in the public portal
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately
and investigate your claim.
PhD Rapport 2013
DTU Vindenergi
Masafumi Azuma
DTU Wind Energy PhD-0015 (EN)
April 2013
Author: Masafumi Azuma DTU Wind Energy PhD-0015 (EN)
Title: Structural control of void formation in dual phase steels April 2013
Division: Section for Materials Science and Advanced Characterization
Department of Wind Energy
This thesis is submitted in partial fulfillment of the requirements for the Ph.D.
Degree at the Technical University of Denmark.
Abstract ISSN: 0106-2840

ISBN: 978-87-92896-40-7
The objective of this study is to explore the void formation mechanisms and
to clarify the influence of the hardness and structural parameters (volume fraction,
size and morphology) of martensite particles on the void formation and
mechanical properties in dual phase steels composed of ferrite and martensite.
Two dual phase steels (Fe-0.099mass%C-1.63mass%Mn and Fe-0.148mass%-
1.60mass%Mn) with martensite particles of different hardness values, volume
fractions, sizes and shapes were produced by hot rolling and annealing.
Mechanical properties were characterised by Vickers hardness and nanohardness
measurements, tensile tests and hole-expansion tests. The initial microstructure
and the deformed microstructure were characterized by means of scanning
electron microscopy (SEM) and transmission electron microscopy (TEM). In situ
tensile tests in a SEM were applied for direct observation of the void formation
behaviour.
In this study it is found that most voids form in martensite particles and few
voids form in the ferrite matrix. The void formation in the martensite is related to
plastic deformation and cracking of martensite particles. The voids in ferrite Sponsorship:
predominantly form near the end of the martensite particles but are not caused by Nippon Steel Corporation
the decohesion of martensite/ferrite interfaces. Three key factors that control the
void formation behaviour in dual phase steels have been established: (i) a critical
strain for void formation in the martensite, (ii) strain partitioning between the
martensite and ferrite and (iii) strain localization. The critical strain for void
formation depends on hardness of the martensite, but is independent of the volume
fraction, shape, size and distribution of the martensite. The strain partitioning
between the martensite and ferrite depends on the volume fraction and hardness of
martensite particles. The strain localization is related to the morphology of the
martensite particles and the strength of the ferrite.
Softening of the martensite significantly retards void formation in the

martensite, but does not change the void formation mechanisms. The increase of
volume fraction of the martensite accelerates the void formation in the martensite
by enlarging the size of voids both in the martensite and ferrite. It is suggested
that controlling the hardness and structural parameters associated with the
martensite particles such as morphology, size and volume fraction are the
essential approach to retard the void formation in the dual phase steels.
Danmarks Tekniske Universitet

DTU Vindenergi
Nils Koppels Allé
Bygning 403
2800 Kgs. Lyngby
www.vindenergi.dtu.dk
2
Masafumi Azuma
Supervised by Xiaoxu Huang and Grethe Winther
Section for Materials Science and Advanced Characterization

Department of Wind Energy
Technical University of Denmark
April 2013
3
4
Contents
Contents....................................................................................................................................................... 5
Preface ......................................................................................................................................................... 8
Publications ................................................................................................................................................. 9
Chapter 1 Introduction ......................................................................................................................... 10
Chapter 2 Background .......................................................................................................................... 11

2.1 Formability for automobiles ............................................................................................................. 11
2.2 Advanced High Strength Steels ........................................................................................................ 11
2.3 Voids in metals ................................................................................................................................. 13
Chapter 3 Materials and techniques .................................................................................................... 16

3.1 Mechanical tests ............................................................................................................................... 16
3.1.1 Hardness test ................................................................................................................................ 16
3.1.2 Tensile test ................................................................................................................................... 17
3.1.3 Hole-expansion test ..................................................................................................................... 18
3.2 Scanning electron microscopy .......................................................................................................... 19
3.3 In-situ mechanical tests in a scanning electron microscope ............................................................. 22
3.4 Transmission electron microscopy ................................................................................................... 23
3.5 Strains in specimens ......................................................................................................................... 24
3.6 Materials ........................................................................................................................................... 26
Chapter 4 Void formation mechanism ................................................................................................ 32

4.1 Introduction ...................................................................................................................................... 32
4.2 Experimental procedures .................................................................................................................. 34
4.3 Results .............................................................................................................................................. 35
4.3.1 Deformed microstructure after a tensile test ................................................................................ 35
4.3.2 Void formation behaviour during in-situ testing ......................................................................... 36
4.4 Discussion ......................................................................................................................................... 43
4.4.1 Plastic deformation of martensite ................................................................................................ 43
4.4.2 Cracks in martensite .................................................................................................................... 46
4.4.3 Voids in martensite ...................................................................................................................... 47
4.4.4 Voids in ferrite ............................................................................................................................. 47
4.4.5 Void formation behaviour at inclusions....................................................................................... 49
5
4.4.6 Void formation behaviour in the bulk ........................................................................................ 50
4.5 Summary......................................................................................................................................... 52
Chapter 5 Effect of martensite hardness ............................................................................................. 53

5.1 Introduction ...................................................................................................................................... 53
5.3 Results .............................................................................................................................................. 54
5.3.1 Influence of tempering treatments on mechanical properties ...................................................... 54
5.3.2 Initial microstructures .................................................................................................................. 57
5.3.3 Deformed microstructure after a tensile test ................................................................................ 60
5.3.4 Void formation behaviour during in-situ testing.......................................................................... 64
5.4 Discussion ......................................................................................................................................... 68
5.4.1 Strain partitioning between martensite and ferrite ....................................................................... 68
5.4.2 Strain localization ........................................................................................................................ 70
5.4.3 Critical strain for void formation in martensite ........................................................................... 78
5.4.4 Void formation behaviour in the bulk .......................................................................................... 79
5.5 Summary........................................................................................................................................... 82
Chapter 6 Effect of volume fraction of martensite ............................................................................. 83

6.1 Introduction ...................................................................................................................................... 83
6.3 Results .............................................................................................................................................. 83
6.3.1 Mechanical properties .................................................................................................................. 83
6.3.2 Void formation behaviour ............................................................................................................ 85
6.4 Discussion ......................................................................................................................................... 91
6.4.1 Critical strain for void formation in martensite ........................................................................... 91
6.4.2 Strain partitioning between martensite and ferrite ....................................................................... 92
6.5 Summary........................................................................................................................................... 95
Chapter 7 Effect of the shape of martensite particles on deformation ............................................. 96

7.1 Introduction ...................................................................................................................................... 96
7.3 Results and Discussion ..................................................................................................................... 96
7.4 Summary........................................................................................................................................... 99
Chapter 8 Conclusions ........................................................................................................................ 100
Acknowledgement ................................................................................................................................... 102
6
Outlook .................................................................................................................................................... 102
References ............................................................................................................................................... 103
7
Preface
This report is submitted in partial fulfilment of the requirements for the PhD degree at the Technical
University of Denmark. The project was started within the Center for Fundamental Research: Metal Structures
in Four Dimensions (M4D), Materials Research Division, Risø DTU, under the supervision of senior scientists,
Drs. Xiaoxu Huang, Niels Hansen and Grethe Winther. The study was conducted during the period from
September 2009 to November 2012. From January 2012, my main supervisor, Dr. Xiaoxu Huang, joint the
Department of Wind Energy, after the reorganization of DTU.
I am grateful to Head of the Materials Research Division Dr. Dorte Juul Jensen and Head of the Center Dr.
Henning Friis Poulsen who gave me the opportunity to work within the M4D Center. I am thankful to my
supervisors Xiaoxu Huang, Niels Hansen and Grethe Winther, who have provided me with constant discussion,
encouragement and much useful support through my PhD study. In particular, Niels Hansen and Xiaoxu Huang
are encouraging me all the time. I am also thankful to Stergios Goutianos, who has helped me to do in-situ
observations and discussed my PhD work with me. I would like to thank Professors Brian Ralph, Johannes
Weertman, Julia Weertman and Nobuhiro Tsuji for their help and discussion. I would like to thank Drs.
Wolfgang Pantleon, Oleg Mishin, Leon Mishnaevsky Jr., Yubin Zhang, Xiaodan Zhang and Chuanshi Hong,
and PhD students Tianbo Yu and Fengxiang Lin and Jacob Kidmose for their help. Special thanks to the very
skilful technicians, Gitte Christiansen, Preben Olesen, Lars Lorentzen, Steen Bang, Frank Adrian, Ove
Rasmussen and Erik Vogeley.
Finally, I would like to thank Nippon Steel Corporation for financial support and for producing the steel
sheets used in this study and to thank Japanese colleagues, Hiroyuki Kawada, Naoki Maruyama and Naoki
Yoshinaga who supported producing the steel sheets and carried out some experiments.
8
Publications
Paper 1 Azuma M., Hansen Goutianos, S., Hansen N., Winther G. and Huang X. In-situ observation of void
formation in dual phase steel. In 30th Risø International Symposium on Material Science (eds N. Hansen et al.),
2010, 243-251. Roskilde, Denmark: Risø National Laboratory.
Paper 2 Azuma M., Hansen Goutianos S., N., Winther G. and Huang X. Effect of hardness of martensite and
ferrite on void formation in dual phase steel. Mater. Sci. Tech., 2012, 28, 1092-1100.
Paper 3 Azuma M., Hansen N., Winther G. and Huang X. In-situ observation of void formation in dual phase
steel. CAMP-ISIJ, 2012, Vol.25, 417.
Paper 4 Azuma M., Hansen N., Winther G. and Huang X. Effect of martensite hardness on void formation
during plastic deformation of a dual phase steel. CAMP-ISIJ, 2012, Vol.25, 1132.
Paper 5 Matsuno T., Azuma M., Maeda D., Uenishi A. and Suehiro M. Mechanism Analysis of Micro Voids
Nucleation in Ferrite-Martensite Dual Phase Steel by FEM. CAMP-ISIJ, 2012, Vol.25, 1065.
Paper 6 Azuma M., Hansen N., Winther G. and Huang X. In-situ observation of void
formation during plastic deformation of dual phase steel (in preparation).
9
Chapter 1
Introduction
Metals have been widely used as structural materials in buildings, ships and automobiles for the last two
hundred years. Modern society requires light strong materials for economical, environmental and safety reasons.
Therefore, a great effort has been put into developing light strong materials. Advanced high strength steels are
one of the answers to the complicated question. As a carbon addition into iron enables to apply phase
transformations, precipitation and recrystallization to control microstructures and textures, a wide variety of
high strength steels have been produced. Whereas the strength is an important issue, the formability is also
required to apply high strength steels for automobiles and electrical appliances. In particular, as components for
automobile and electrical appliances are often deformed by stamping and bending, formability is required.
However, as strengthening generally reduces the formability such as elongation in tension poor formability
should be improved in order to apply for complicated components. Advanced high strength steels such as dual
phase steels consisting of a soft ferrite matrix and hard martensite particles, are one of the answers to the
complicated question as they have a good balance of elongation, in particular uniform elongation, and high
strength. The soft ferrite matrix ensures uniform elongation and hard martensite particles contribute high
strength and large work hardening rate. Whereas the uniform elongation is one of the most important
parameters for the formability, formability for hole-expansion tests and bending tests is also required to apply
steels for complicated automobile components. On the other hand, it is well-known that void formation leads to
ductile fracture of metals. For instance, many dimples related to void formation are observed on fracture
surfaces of metals after tensile tests at ambient temperature. In general, most voids start to nucleate in single-
phase metals after the onset of necking. However, many voids form in metal matrix composites, which consist
of a metal matrix and carbides, nitrides, sulphides and oxides, due to a weak bonded interface and brittle
reinforcements before necking. As strain and stress encourage voids to form in metals and composites, void
formation has a larger effect on post-necking elongation in tension, the formability for hole-expansion tests and
bending tests. As dual phase steels have poor formability for hole-expansion tests and bending tests, the
formability should be improved. However, even though void formation has a large effect on the formability, the
void formation mechanism and effect of metallurgical parameters on void formation in dual phase steels are still
ambiguous.
This project is to study the void formation mechanism and the influence of metallurgical parameters
including hardness, volume fraction and size of martensite particles in dual phase steels. This thesis is structured
as following: Chapter 2 gives a short review of the background; Chapter 3 introduces the materials and
experimental procedures used in this study; Chapter 4 demonstrates void formation behaviour in a dual phase
steel during in-situ loading; Chapter 5 reports the influence of martensite hardness on the void formation
behaviour and main parameters for void formation in a dual phase steel; Chapter 6 suggests the influence of
volume fraction; Chapter 7 shows the influence of the morphology of martensite particles on the deformation
behaviour of martensite particles in the dual phase steel; and Chapter 8 gives conclusions and outlook.
10
Chapter 2
Background
2.1 Formability for automobiles

Most steel sheets are cut mechanically, deformed through stamping and applied as automobile components.
Therefore, the steel sheets need to have a good formability for the stamping. Typically the following formability
for i) deep drawing, ii) stretching, iii) stretch flanging and iv) bending is required due to complicated shapes. In
particular, applications for outer panels such as doors, side panels and roof panels require the formability for
deep drawing and stretching as shown in Fig. 2.1. In order to improve the formability for deep drawing, textures
of steel sheets are controlled by chemical composition, thermomechanical treatments and heat treatments. In
addition, smoothness of the surface is also an important property and mild steels are often used for panels due to
low yield stress and excellent deep drawability. On the other hand, structural components such as cross-
members, wheels and pillars are deformed through bending, hole-expansion and stretch flanging after cutting. It
is suggested that the formability for stretch flanging, hole-expansion and bending is strongly related to void
formation, growth and coalescence in steel sheets as a large number of dimples related to void formation are
observed on the fracture surface of the specimens (e. g. Nishimoto et al. 1981, Hasegawa et al. 2004, Toji et al.
2009, Tasan et al. 2009, Hayashi et al. 2012). In particular, during hole-expansion and bending tests
deformation localizes at the edge of the sheets and strains at the edge are much larger than uniform elongation
in tensile tests as stress gradients retard the onset of necking. Even though uniform elongation is an important
parameter, the formability for stretch flanging and bending is also important. As voids form with strain (e. g.
Maire et al. 2008, Landron et al. 2010, Avramovic et al. 2010) and reduce the area of cross section in specimens,
reduction in area and thickness is also related to void formation. Therefore, a great effort to clarify void
formation mechanisms and control void formation is explored.
2.2 Advanced High Strength Steels

Typically the following high strength steels; i) solid solution hardened steel, ii) precipitation hardened steel,
iii) dual phase steel, iv) TRIP (Transformation Induced Plasticity) steel, v) bainite (martensitic) steel, are often
applied for automobiles components. Solution hardened steels are typically strengthened by a solid solution of
phosphorus (P), manganese (Mn) and silicon (Si). However, the strengthening is limited as P has a negative
effect on weldability (e. g. Saito et al. 1987) and Si leads to defects on the surface (e. g. Takada et al. 2009,
Shimada et al. 2009). The solid solution hardening contributes to 270-540MPa grade steel sheets (Fig. 2.2).
Precipitation hardened steels are strengthened by titanium carbides, niobium carbides and vanadium carbides.
The precipitations increase the yield stress and ultimate tensile strength as precipitations disturb dislocation
motion, counterbalanced by a reduction of elongation. As shown in Fig. 2.2, the precipitation hardened steels
have poor elongation compared with other high strength steels. Dual phase steels consisting of a soft ferrite
phase and hard a martensite phase are more attractive for automobile components due to better combination of
strength and elongation. As the hard martensite contributes to the strength and a soft ferrite matrix ensures
elongation and it follows that the grain size of ferrite, the volume fraction and hardness of martensite are
important parameters. These are controlled through the chemical composition of the steel and the conditions
under which it is heat treated. TRIP steels have more complicated microstructures consisting of ferrite, bainite
and retained austenite. Retained austenite attributes to a large work hardening rate and strength as plastic
deformation assists retained austenite to transform to martensite strengthening the steels. After deformation, the
microstructures consist of the ferrite matrix, bainite and martensite phases. Dual phase steels and TRIP steels
11
have a large elongation, in particular, large uniform elongation (Fig.2.2). On the other hand, bainitic steels and
martensitic steels have homogeneous microstructures composed of a single microstructure. Bainite and
martensite containing a large number of dislocations have a large strength, which is larger than 980MPa (Morito
et al. 2003), but the elongation is the smallest of all (Akisue and Hata 1995, Kishida 2000). While the uniform
elongation is an important parameter, the formability of the material depends on the response to localized
deformation. Parameters to characterize this include the local elongation in the necking zone of a tensile test,
the hole-expansion ratio determined by pushing a cone through a punched hole and the elongation through
bending (e. g. Hasegawa et al. (2004), Yamazaki et al. (1995)). Fig. 2.2 exemplifies the formability for hole-
expansion tests of high strength steels. The hole-expansion tests are used to estimate formability for automobile
components. In general, a strain after hole-expansion tests by a conical punch is much larger than uniform
elongation and the better formability at a large strain is required. It is well-known that bainitic and martensitic
steels have a better formability for hole-expansion tests due to the homogeneous microstructures. On the other
hand, dual phase steels and TRIP steels have a poor formability and the formability should be improved to
apply dual phase steels for more complicated components.
Blank Blank
Stretch Shrink Punch
Die
Flow
Die Flow
Deep Stretching
drawing Stretch
Blank Stretch
Flow flanging Bending
Stretch
Bending
Stretching
Fig. 2.1 Typical formation process for automobile components: 1) deep drawing, 2) stretching, 3) stretch
flanging and 4) bending. The formability for deep drawing and stretching is more important for outer panels
such as doors and front panels. Structural components such as pillars and wheels require more the formability
for stretch flanging and bending.
12
Fig. 2.2 Relationship between tensile ultimate tensile strength and formability of hot rolled high strength steel
sheets. Reproduced from Kishida (2000).
2.3 Voids in metals

It is well-known that void formation in metals and metal matrix composites affects the strength, uniform
elongation and total elongation. Voids are empty space, void formation leads to a reduction of a cross section of
specimens and thereby affects mechanical properties. For instance, Vasudevan et al. (1989) showed that in a
13
metal matrix composite (Al-20vol.%SiC), where voids start to appear before the onset of necking, void
formation decreases the flow stress, ultimate tensile strength, uniform elongation and fracture elongation
through tensile tests under hydrostatic pressure as shown in Fig. 2.3. On the other hand, Fig. 2.4 shows that the
yield stress, ultimate tensile stress and uniform elongation do not depend on hydrostatic pressure in aluminium
as voids form after necking. In addition, González et al. (2002) demonstrated that hydrostatic pressure affects
the work hardening rate due to void formation and thereby retards the onset of necking in a metal matrix
through a finite element modelling (FEM) analysis (Fig. 2.5). It is obvious that void formation accelerates
ductile fracture. For instance, a large number of voids are observed in the necking zone of tensile specimens and
near the fracture surface of the specimens. As mentioned above, in metals without hard particles and inclusions,
void formation mainly affects the post-uniform elongation, which is defined as an elongation after necking, as
voids form in the necking zone of tensile specimens. It is well-known that decohesion at interfaces between
metals and particles and particle cracking leads to void formation in metal matrix composites. For instance,
Babout et al. demonstrated that decohesion at interfaces between Al and ceramics particles, which consisted of a
mixture of zirconia (70% in weight) and silica (30% in weight), and particle cracking of the ceramics dominates
in Al alloy-ceramic composites through X-ray tomography (2001, 2004). Decohesion often occurred at the
interface facing to the tensile axis and cracks in the particles propagated perpendicular to the tensile axis. As
cracks at the interface and in the particles developed into voids and grew in the tensile direction, the
morphology of voids depended on strain applied to specimens. Finally, they coalesced with each other and
macro-crack formation led to fracture of the specimens. It appears that particle cracking and decohesion are
important mechanisms for void formation in composite structures.
Fig. 2.3 Influence of hydrostatic pressure on stress-strain Fig. 2.4 Influence of hydrostatic pressure on the yield stress
curves of reinforced 2014 aluminum alloy. Reproduced and ultimate tensile strength of the unreinforced 2014
from Vasudevan et al. (1989). aluminum alloy. Reproduced from Vasudevan et al. (1989).
14
Fig. 2.5 Influence of hydrostatic pressure on the flow stress () and the work hardening rate (d/d) in a metal
matrix composite. Reproduced from González et al. (2002).
In dual phase steels, void formation has also been an important objective as a large number of voids are
observed near the fracture surface of specimens after tensile tests, bending tests and hole-expansion tests. Many
investigations have shown that voids can form in the martensite, in the ferrite and at inclusions and that void
formation related to the martensite particles dominates. Therefore, void formation behaviour in dual phase steels
is investigated in order to clarify void formation mechanisms and the influence of metallurgical parameters,
such as hardness, volume fraction and shape of martensite particles, on void formation in dual phase steels.
15
Chapter 3
Materials and techniques
3.1 Mechanical tests
3.1.1 Hardness test

A micro-hardness test was carried out to evaluate the hardness of the martensite and ferrite phases in the dual
phase steels. Hardness is one of the important methods to evaluate mechanical properties of materials even
though it is an intrinsic material property. The hardness increases linearly with tensile strength of materials
insofar as the sizes are in a same order. In particular, as dual phase steels are a mixture of a soft ferrite matrix
and hard martensite particles, it is difficult to evaluate the mechanical properties of each phase by a tensile test.
On the other hand, micro-hardness tests enable one to measure the hardness of each grain/particle easily.
Therefore, a micro-hardness test was carried out.
136°
d2
d1
Fig. 3.1 An illustration of Vickers hardness test and an indent by the test.
In this study, the hardness of the martensite and ferrite was characterized by a Vickers hardness test. The
Vickers hardness test consists of indenting the test materials with a diamond indenter, which has a pyramidal
shape and an angle of 136°. After the indentation, the diameters (d1 and d2 in Fig. 3.1) of the indent were
measured and the Vickers hardness (kgf/mm2) is given by the following equation:
136
Hv  2 F sin 2 Eq. [1]
2
 d1  d 2 
 
 2 
The specimen was polished and etched with nital to distinguish the martensite with ferrite. For the micro-
hardness test, a Struers DuraScan 70 Hardness test was used with a load of 10gf applied for 10 seconds. For
each phase 10 particles/grains were measured and the values were averaged. The indent size was about 5-10µm
and were often larger than the ferrite grains and martensite particles as shown in Fig. 3.2. In particular, in
deformed specimens martensite particles were stretched in the tensile direction and were thin (Fig. 3.2c). The
large grains/particles were chosen to evaluate the hardness by the indents.
16
a) b) c)
10µm 10µm 10µm

Fig. 3.2 Indents by Vickers hardness tests: a) in the martensite before a tensile test, b) in the ferrite before a
tensile test and c) in the martensite after a tensile test.
It is well-known that the hardness depends on the indent size (e.g. Stelmashenko et al. 1993, Miyahara et al.
2001). The hardness measured by smaller indents tends to be large due to relatively large elastic deformation.
On the other hand, if the indent size is smaller than the grain size of materials, the hardness tends to be smaller
than the macro-scale hardness as the hardness does not contain the effect of grain boundary hardening. In the
present study, the hardness of the ferrite does not contain effect of grain boundary hardening, but that of the
martensite is affected by grain boundary hardening as the martensite consists of small laths of martensite (the
width are about 200nm).
3.1.2 Tensile test

Flat dog-bone specimens of dimensions 3 mm × 35 mm × 180 mm were deformed in tension. The gauge
length was 50 mm and the gauge width was 25 mm. The strain rate was controlled by the motion of a cross head
and the cross head speed was 3.3×10-3 (s-1). The tensile direction (TD) was parallel to the transverse direction of
the rolled sheet. During tensile testing, after the stress overcame the ultimate tensile strength, the extensometer
was removed. The total elongation was instead estimated based on measurement of gauge length changes after
the tensile test. The following parameters have been determined: yield stress (YS), ultimate tensile strength
(UTS), uniform elongation (U-El.), total elongation (T-El.) and post-uniform elongation (P-El.) in the necking
zone. The uniform and total elongations were estimated based on the nominal strain. Void formation mainly
affects the P-El. which is the elongation of the specimen from the start of necking to fracture.
The work hardening rate of the stress-strain curves in Fig. 3.3 was analysed in more detail in order to clarify
which phenomenon affect the work hardening rate in the dual phase steels. For the analysis, the modified
Crussard-Jaoul analysis (e.g. Lian et al. 1991, Jiang et al. 1995) which is based on the Swift formula, was
carried out. In the formula, a stress-strain relationship is given by the following:
   0  c m Eq. [2]
where 0 is an initial true strain, m is a work-hardening exponent and c is a material constant. An integrated
form of Eq. [2] with respect to , is transformed into a logarithmic form:
 d 
ln   1  m ln   lncm Eq. [3]
 d 
The slope of the ln(d/d) vs ln() curve equals to (1-m) and the intersection with ln(s)=( is given by ln(cm).
In general, as the stress-strain relationship of dual phase steels is affected by the deformation of the ferrite
phase and martensite phase, the work hardening behaviour of dual phase steels is divided into two regions (Fig.
3.3); (i) plastic deformation of the ferrite phase and elastic deformation of martensite phase, and (ii) plastic
deformation of ferrite and martensite phases. In the first region, the work hardening rate is relatively large as a
17
large number of dislocations pile up at the martensite/ferrite interfaces because the martensite can deform
elastically. In the second region, the work hardening rate is relatively small due to strain accommodation by
plastic deformation of the martensite phase. It is also suggested that the slope of Eq. [3] changes suddenly in
dual phase steels when the plastic deformation of the martensite contained in the dual phase steel starts. Based
on the modified Crussard-Jaoul analysis, the start of the martensite phase contained in bulk specimens is
estimated.
martensite
Dual phase steel

(i) Plastic deformation of ferrite phase
 (ii) + Elastic deformation of martensite
(i) ferrite
(ii) Plastic deformation of ferrite phase
+ Plastic deformation of martensite

Fig. 3.3 Work hardening behaviour of a dual phase steel. The work hardening behaviour depends on the start of
plastic deformation of the martensite phase.
3.1.3 Hole-expansion test

A hole-expansion test was used to estimate the formability for stretch-flanging and hole-expansion. The
hole-expansion tests are engineering tests to estimate the formability for automobile components and two types
of hole-expansion test are suggested. The first one is used for panels of automobiles during deep drawing (Fig.
3.4a) and another is used for structural components and under car body components (Fig. 3.4b). In both tests
steel sheets are punched and a hole is expanded mechanically by a cylindrical punch or a conical punch.
The difference between the two tests is the stress distribution around a hole. In the test by the cylindrical
punch (Fig. 3.4a), stress concentrates at the edge of hole, but it is assumed that the specimen at the edge of the
hole is under a uniaxial tensile stress during loading. Necking occurs at the edge of the specimen and thereby
accelerates fracture of the specimen. Therefore, the formability tested by the cylindrical punch depends on
uniform elongation more even though macro crack formation affects the formability.
On the other hand, in the test by the conical punch (Fig. 3.4b), stress distributes in the radial and thickness
direction. As the large stress gradient generally retards necking, the specimen can be deformed without necking
and the strain at the edge of the specimen is much larger than the uniform elongation. Voids appear at the edge
of the specimen and lead to fracture of the specimen. Therefore, the formability strongly depends on void
formation in a material. As dual phase steels are applied for structural components, the formability is estimated
by the hole-expansion test in Fig. 3.4b.
At first, steel sheets are punched, clamped by a holder and the hole is expanded by the conical punch. In the
present study, the diameter of the punch was 10mm and the diameter of the blank was10.6 mm. During loading
the edge of the hole is observed and the test is stopped when a crack passes through the sheet. The formability is
given as a hole-expansion ratio according to the following equation:
Hole-expansion ratio = (d – d0) / d0 × 100% Eq.[4]
where, d0 is the initial diameter of the hole and d is the diameter of the hole after the hole-expansion test.
In addition to void formation during loading, the formability also depends on damage induced by punching.
As the specimen is deformed heavily during punching, the specimen often contains voids. Voids induced by
punching and plastic deformation by punching lead to early fracture of the specimen. Besides, a rough fracture
surface, which often contains dimples and scratches, after punching accelerates crack formation due to stress
18
concentration. Therefore, machining and cutting by a laser are suggested as a method to make a hole in
components, as they damage less the specimens. The study is a future experiment.
Fig. 3.4 Hole-expansion tests for a) panels (stress distributes in the radical direction.) and b) structural
components (stress distributes in the radical direction and the thickness direction).
3.2 Scanning electron microscopy

Scanning electron microscopy (SEM), which images by scanning with a high energy electron beam, is one of
the most appropriate methods to observe materials, whereas an optical microscope is also a helpful instrument
for structural evolution. The SEM has a much higher spatial resolution than observations by optical microscopy
(OM) due to the smaller wavelength of an electron beam. The spatial resolution of the SEM is typically 1-
100nm and 100 times as much as optical microscopes. In the SEM observations, some signals generated from
specimens when an electron beam interacts with specimens are used to follow structural evolution. There are
two familiar signals; (i) secondary electrons (SEs) and backscattered electrons (BSEs). The contrast of SE
images mainly depends on the specimen topography and it is, therefore, suitable to observed deformation and
fracture behaviour on the surface of specimens as during plastic deformation steps appear on the surface due to
plastic deformation and crack formation. On the other hand, the contrast of BSE images predominantly depends
on the atomic number and crystal orientation of constituents in specimens. The orientation contrast in BSE
images originates in electron channelling contrast (ECC) and is frequently used for lattice defects and boundary
structure.
In this study, a Zeiss EVO60 was used for structural evolution of the initial and deformed microstructures.
The Zeiss EVO60 was operated at an accelerated voltage of 20kV. In order to characterize the initial
19
microstructures, the steel sheets were cut parallel to the rolling direction, polished and etched by nital
(96%C2H5OH + 4%HNO3), the surfaces are observed from the transverse direction. Nital can help to distinguish
the martensite particles from the ferrite matrix. Five areas were observed at a magnification of 2000×. However,
the microstructure at the centre of the specimen was not included to avoid a possible effect of Mn segregation.
The grain size was measured in the transverse direction and in the direction normal to the rolling direction by a
linear intercept method, and the average value was defined as the grain size. In addition, the size distribution
and the aspect ratio of the martensite particles were estimated, because the morphology of the particles was
somewhat irregular and they were elongated in the transverse direction. The aspect ratio was determined as the
length in the transverse direction divided by the length in the normal direction. The volume fraction of each
phase was measured by a point counting method. The deformed microstructures after testing were studied from
the transverse direction to estimate the characteristics of voids including (i) initiation site, (ii) number density
and (iii) size. In order to detect void initiation sites, the microstructure was etched by nital. Comparisons of the
etched and unetched microstructures established that etching does not generate etch pits or affect the shape and
size of the voids. Void initiation sites were observed in areas with a limited elongation because a large
deformation resulted in microstructures which were too complex to characterize.
Electron backscattered diffraction (EBSD) is also frequently used for the structural evolution of crystalline
materials in the last two decades. In an EBSD observation, a Kikuchi diffraction pattern is generated from the
specimen when the electron beam hits on the specimen, is automatically analysed by computers and determined
the crystal orientation of each point as the pattern strongly depends on crystal orientation of constituents in
specimens. Tungsten and lanthanum hexaboride (LaB 6) used to be common as filament materials and field
emission guns (FEGs) are more frequent recently due to the better special resolution. In the EBSD with a FEG,
the angular resolution is about 20nm and 1°. It is suitable to study crystalline materials. In particular, the FEG-
SEM is more useful to estimate deformation behaviour in dual phase steels as a large number of dislocations
induced by martensite transformation and plastic deformation leads to large rotations of crystal lattices.
In the present study, a Zeiss Supra-35 FEG scanning electron microscope, equipped with a HKL Channel 5
EBSD system, was used for the observation of initial microstructures and deformed specimens. The specimens
were observed from the transverse direction. For the observation of deformed microstructure, the specimen was
deformed in tension. The applied strain was 0.05 and smaller than the uniform elongation. The SEM was
operated at an accelerated voltage of 20kV and a step size of 50nm. EBSD data were pre-processed using the
HKL 5 EBSD software to reduce noise and to smooth the orientation. A critical angle for boundary detection
was set to 2° defined as low angle boundaries and grain boundaries, which misorientations over 7°, was defined
as high angle boundaries. Based on the experimental data, the strain distribution is estimated macroscopically
and locally in specimens deformed in tension.
First of all, the macroscopic strain distribution was evaluated based on the deviation of the Euler angle at
each point form the average value of each ferrite grain. For the analysis, a program, which is established by
Andy Godfrey (e. g. Li et al. 2002), was used. Ferrite grains and martensite particles were detected by the HKL
5 EBSD software and then the average Euler angle in each ferrite grain was calculated but the martensite phase
was neglected. For boundary detection, a critical angle for boundary detection was set to 7°. At each point, the
deviation from the average angle was calculated and exemplified in Fig. 3.5. In the map, yellow regions have a
small misorientation of 0°, blue regions have a larger misorientation and green regions have a misorientation of
over 7°. Since the difference of orientation in a grain is related to geometrically necessary dislocations, it is
assumed that the difference means plastic strain in ferrite grains. Based on the contour map, the macroscopic
plastic strain distribution; where strain localized and how frequently strain localized, was estimated.
20
Fig. 3.5 Measurement of strain distribution by a SEM-EBSD method: a) Image Quality map, b) Orientation map
and c) a contour map of strain distribution in the ferrite phase. In the contour map, a yellow region has a small
strain, a blue region has a larger strain and the strain in green regions is the largest of all. The tensile direction is
marked.
In addition, the local strain distribution is evaluated at each martensite/ferrite interface and ferrite/ferrite
grain boundary in more detail. In this study, accumulated misorientation profiles from the martensite/ferrite
interface and the ferrite/ferrite grain boundaries were measured. As strain distributes heterogeneously in the
ferrite phase and martensite phase, typical areas, where strain localizes, were chosen based on the macroscopic
observation of strain localization. The accumulated misorientation profiles were measured in the tensile
direction and perpendicular to the tensile direction. As the strain tends to localize very close to the
martensite/ferrite interface, the accumulated misorientation at 1µm away from the interface is represented as a
local strain. About 20 interfaces/boundaries were measured.
21
3.3 In-situ mechanical tests in a scanning electron microscope
In-situ mechanical tests in a SEM were used to follow void formation behaviour in dual phase steels. For the
tests, a load cell (Fig. 3.6b-d) was installed in a scanning electron microscope (Zeiss EVO60) as shown in Fig.
3.6a. Tensile tests and three point bending tests were carried out in the SEM.
a) b)
c) d)
Fig. 3.6 In-situ mechanical tests: a) a SEM (Zeiss EVO60), b) a load cell and sample holders for c) a tensile test
and d) a bending test.
For in-situ tensile testing, 0.15 mm × 2.8 mm × 30 mm specimens were machined and a hole with a diameter
of 1 mm was made in the centre of the specimen. For bending tests, 2 mm × 3 mm × 26 mm specimens with a
notch in the middle were used. The hole and notch allows one to detect void and crack initiation easily. In
addition, since the strain distribution around the hole and notch is similar to that at an edge of a tensile specimen
after diffusive necking starts and at an edge of a stretch flanging part, the shape of specimens is reasonable to
quantify void formation behaviour in the specimens with a hole or a notch. All specimens were polished and
etched before deformation. Fig. 3.7 illustrates an in-situ loading test. As shown in Fig. 3.7a, the sample was
clamped at both ends and stretched under displacement control in a screw-driven load cell, which was installed
inside the microscope. The specimen was strained gradually and the surface structure was characterized during
loading. Fig. 3.7b illustrates the gradual bending by a pin of a sample without unloading. In the tests, the
deformation concentrates at the edge of the hole and at the bottom of the notch. During the tests the following
procedure has been related to nucleation and growth of voids; (i) nucleation density, (ii) size distribution, (iii)
length and (iv) area fraction of voids in a area of 25000µm 2. Under each condition, in-situ tests were carried out
three times. Void formation behaviour was estimated statistically. The observations have been related to the
strain which is defined in the following section.
22
a) Tensile test
TD
1mm 1mm
Before deformation During deformation
b) Bending test
Pin
1mm 1mm
Fig. 3.7 In-situ mechanical tests: a) tensile test in which deformation concentrates near the hole (the horizontal
arrow marks the tensile direction, TD), b) bending test in which deformation concentrates near the notch.
3.4 Transmission electron microscopy

Transmission electron microscopy (TEM), which images by transmission and diffraction with a high energy
electron beam, is an important method to observe microstructures at a higher resolution, below 0.1 nm for the
best instruments. In a TEM observation, a wide variety of signals is generated from specimens when an electron
beam interacts with specimens and is used for structural evolution. In particular, bright field images which are
produced by a parallel beam, are often used to observe precipitations and dislocations in materials. A diffraction
pattern was used to analyse structures of constituents and crystal orientations. A specimen is so thin that an
electron beam is transmitted through the specimens which were typically from 100 to 500nm for steels.
In this study, a transmission electron microscope (JEM 2000FX) with a LaB 6 filament and a CCD camera
was used. The accelerating voltage was 200kV. The specimens for TEM were prepared by electro chemical
polishing. Bright field images were used to observe initial microstructures and deformed microstructures. Dark
field images and diffraction patterns were used to characterize precipitations in tempered specimens. In addition,
Kikuchi patterns generated convergent electron beam diffraction (CBED) was applied to adjust the tilt angle of
the specimen.
The dislocation density in the ferrite phase before deformation, which contains a number of dislocations due
to the martensite transformation, was measured by TEM. For all specimens, dislocations were observed from
<200> zone axis close to the electron beam direction, making all dislocations visible. In the images many lines
are drawn and dislocation number density was counted. Then the dislocation density is given by the following
equation:
23
2N L
d  Eq. [5]
t
where NL is the number of intersections at dislocations per unit line and t is a thickness of the observed area.
About 10 areas were observed.
3.5 Strains in specimens

Whereas uniform elongation in tensile tests is an important parameter to evaluate the formability, the local
strain at the necking zone of specimens both during tensile testing and during in-situ testing is also a key
parameter as it is related to macro-crack formation and fracture of the specimen in the necking zone. Generally,
since more voids form with strain, void formation has a large effect on the local strain in the necking zone. Post-
uniform elongation (P-El) is used to evaluate the formability from the start of necking to fracture of the
specimen in tensile tests. However, P-El is an engineering strain and depends on the gauge length, the shape and
thickness of specimens as deformation localizes in the narrow necking zone after the start of necking. On the
other hand, a thickness strain, which is estimated based on a thickness reduction of specimens, is another
parameter to evaluate the strain at each point in the necking zone. In this study the thickness strain (t) in the
tensile direction was measured in order to evaluate void formation behaviour in the bulk as shown in Table 3.1a.
An initial thickness of a tensile specimen (t0) is compared with a thickness at each point in the necking zone of
the specimen (t1) after a tensile test. Void formation behaviour in the bulk is evaluated as a function of the
thickness strain (t).
During in-situ testing a local strain at the bottom of a hole (h) is estimated based on measurement of a gauge
length at the bottom of a hole (Table 3.1b). The hole is helpful to detect void initiation due to stress
concentration and retards the onset of local necking due to the stress gradient at the bottom of the hole. The
strain in the tensile direction is defined as a local strain during in-situ testing as deformation localizes at the
edge of the hole. It seems that the local strain is more related to a thickness strain because deformation localizes
at the necking zone in tensile tests. A gauge at the position was defined based on SEM observations. At first, a
100m gauge and the notch at the bottom of the hole was defined by the SEM observation, in which the
distance was estimated based on the microstructure which was visible. The length change of the gauge was
directly followed during in-situ tests. If the gauge length is too small, few grains exist in the gauge and a strain
does not represent an average value of the specimen because each grain has a different strain due to different
crystal orientations and the distribution of martensite particles. On the other hand, since deformation
concentrates at the bottom of the hole or notch, a large gauge length might lead to an underestimation of the
strain.
24
Table 3.1 Strains estimated after tensile tests and during in-situ tensile tests.
a) Thickness strain of the specimen after a tensile test: t = (t 0 – t 1 ) /t 1 Eq.[6]
where, t 0 is an initial thickness of the specimen and t 1 is a thickness at each point after a tensile test.
Before deformation After deformation
t0 t1
b) Local strain near the hole and at a bottom of a notch during an in-situ test:
h = (Lh 1 – Lh 0 ) / Lh 0 Eq.[7]
where, Lh 0 is an initial gauge length (=100µm) and L h 1 is a gauge length during an in-situ test.
Lh 1
Lh 0
c) Average strain in ferrite grains during an in-situ test: F =S{ (LF1 – LF0 ) / LF0 } Eq.[8]
d) Average strain in martensite particles during an in-situ test: M =S{ (LM1 – LM0 ) / LM0 } Eq.[9]
where, LF0 and LM0 are initial length of ferrite grains and martensite particles and LF1 and LM1 are
length of the grains and particles during an in-situ test.
e) Local strain at a necked region of a martensite particle during an in-situ test:

L = (LL1 – LL0 ) / LL0 Eq.[10]
where, LL is an initial length of a necked region of a martensite particle and L L1 is the length of a
0
necked region of a martensite particle during an in-situ test.

LF0 LF1
ferrite
martensite
LL 0 LL1
LM0 LM1
In addition, strains in the martensite and ferrite (M and F) are also important parameters to estimate the
effects of metallurgical parameters, such as hardness, volume fraction and shape of the martensite, on void
formation in the martensite and ferrite as these parameters represent strain partitioning between the martensite
and ferrite (e.g. Shen et al. 1986, Su and Gurland 1987, Tasan et al. 2010). In dual phase steels, as the
martensite is much harder than the ferrite, martensite particles are less deformable than ferrite grains. Hardness,
volume fraction and morphology of the martensite affect strain in the martensite and ferrite, and are, thereby,
related to mechanical properties of dual phase steels. As shown in Tables 3.1c and d and Fig. 3.8, the length
25
change of ferrite grains and martensite particles were followed during in-situ testing. About 200 grains/particles
were measured and the strains were averaged as strains in the martensite and ferrite (M and F).
a) b)
TD
LM0 LM1
LF0 5m LF1 5m

Fig. 3.8 Deformation behaviour in a dual phase steel during an in-situ tensile test: a) before deformation and b)
during deformation.
Moreover, a local strain at the necked region of a martensite particle (L) is another key parameter to
estimate void formation in the martensite. During tensile testing and in-situ testing deformation localized at the
concave part of the martensite particles (Fig. 3.9) and a crack appeared at the necked zone. The local strain at
the necked regions is much larger than the average strain in the martensite due to strain localization. Even
though a martensite particle does not contain a concave part and a concave part often appeared in the particle
during deformation due to necking. Local strain at about 30-180 necked regions was estimated for each
condition.
a) b)
TD
LL1
LL0 5m 5m
Fig. 3.9 Deformation behaviour at a concave part of a martensite during an in-situ tensile test: a) before
deformation and b) during deformation.
3.6 Materials
Four alloys were melted in a vacuum furnace and cast as 20 kg ingots into a crucible to produce the materials
which contain different amount of martensite phase (Table 3.2). The chemical compositions are given in Table
3.2. Carbon is one of the most important alloying elements when producing dual phase steels as it affects not
only the volume fraction of martensite but also the hardness of the martensite. Dual phase steels with 10-50
vol.% of martensite contain about 0.1-0.2mass% C (e.g. Erdogan 2002, Avramovic et al. 2009 and 2010). In the
present steels 0.1-0.2mass% carbon were added to produce a dual phase steel whose matrix is a ferrite phase.
Manganese was added to retard the pearlite transformation during cooling, since most of the austenite phase in
steels without Mn rapidly transforms to pearlite. However, a large amount of Mn generates zones of Mn
segregation in the centre of the ingot, which makes the microstructure heterogeneous and also leads to different
void formation behaviour in this zone [Avramovic et al. 2009]. Therefore, the Mn contents were set to 1.6 and
2.0 mass%. Sulphur and oxygen often exist as sulphides and oxides in steels and they have a negative effect on
the mechanical properties because of void formation around inclusions (e.g. Nutt and Needleman 1987,
Christman et al. 1989, Clyne and Withers 1993, Maire et al. 2008). For this reason, an electrolytic iron with a
sulphur content less than 0.001mass% was used. The ingot was annealed at 1200˚C for 1 hour in an argon
atmosphere, hot-rolled from 45 to 3 mm thickness at 1100-950˚C, cooled slowly from a hot-rolling finishing
temperature of 950˚C to 650˚C to promote ferrite formation, followed by water cooling to ambient temperature
and transformation of austenite into martensite. Some specimens were tempered at 300-600˚C for 100s to
evaluate the effect of a softened martensite phase.
26
Table 3.2 Chemical composition / mass%
C Si Mn P S N O
Steel1 0.099 0.01 1.63 0 0.0007 0.0008 <0.0010
Steel2 0.148 0.01 1.60 0 0.0007 0.0016 <0.0010
Steel3 0.200 0.01 1.59 0.001 0.0006 0.0018 <0.0010
Steel4 0102 0.01 1.99 0.001 0.0008 0.0013 <0.0010
The microstructures after hot rolling are shown in Fig. 3.11. The microstructures were composed of ferrite
grains (F) and martensite particles (M) and neither pearlite nor bainite was observed. The martensite was lath
martensite, which had hierarchic structure as shown in Fig. 3.10. In general, lath martensite is composed of
packets, which consist of parallel blocks. There are three blocks with different orientation in a packet. Each
block consists of laths of martensite with Kurdjumov-Sachs (K-S) relationship to prior austenite. Specially, each
block consists of two specific K-S groups (sub-block), whose misorientation is about 10 degrees. The initial
microstructure was free from voids and cracks. The microstructures strongly depended on the chemical
compositions. As shown in Fig. 3.11 and Table 3.3, a ferrite phase occupied majority of the area in Steel 1, 2
and 4, the matrix was the ferrite phase. On the other hand, Steel 3 contained a large amount of martensite
particles due to retardation of ferrite formation by the 0.2mass% carbon addition, which stabilizes the austenite
phase and retards ferrite formation. In addition, the mechanical properties of Steel 3 were different from others
(Table 3.4), Steel 3 was not used for in-situ mechanical tests. In addition, Steel 4 has a band-like structure
consisting of the martensite phase at the centre of a specimen due to Mn segregation zones (Fig. 3.12).
Avramovic et al. (2009) and Tasan et al. (2010) pointed out that the band-like structure has a negative effect on
the formability and void formation behaviour in the zones is different from other regions. Table 3.4 also shows
that as Steel 1 and 2 had high ultimate tensile strength and a low yield ratio (0.2%YS/UTS) of about 0.5, the
measured tensile behaviour were typical for dual phase steels. Therefore, Steel 1 and 2 were chosen as
specimens to estimate the void formation behaviour in dual phase steels. Whereas Steel 1 and 2 have different
martensite fraction, the size of ferrite grains and aspect ratio of martensite particles are similar. The steel
containing an 0.21 martensite is named SVM (Small Volume fraction of Martensite) and the steel containing an
0.38 martensite is named LVM (Large Volume fraction of Martensite).
Fig. 3.10 Schematic illustration showing lath martensite structure in (a) a low carbon (0-0.4mass%C) and b)
high carbon (>0.6mass%C) alloys. Reproduced from Morito et al. (2003).
The shape of the two phases is also different. Ferrite grains nucleate at triple junctions and grain boundaries
of the austenite during cooling and have a polygonal shape. Martensite particles are, on the other hand,
transformed from the remaining austenite and they are less regular in shape with concave and convex profiles
and some particles are elongated in the rolling direction. As shown in Fig. 3.13, the sizes range from less than 1
to over 20µm and the average aspect ratio of SVM and LVM are 2.1 and 2.2.
27
A microhardness test shows that the hardness of the martensite and ferrite depends on tempering treatments,
but the hardness of the specimens are almost the same at the same tempering temperature, as shown in Fig. 3.14.
In particular, the tempering treatments strongly affected the hardness of the martensite due to a large number of
carbide precipitates. The standard deviations of the hardness of the martensite were relatively large due to the
small size. As these steels (SVM and LVM) have different amounts of martensite, but the hardness of the ferrite
and martensite is similar at each tempering temperature, the steels are suitable to estimate the effect of the
hardness and volume fraction of the martensite on void formation in the dual phase steels.
The stress-strain curves of SVM and LVM also show that these steels were typical dual phase steels. The
stress increases rapidly after yielding and the work hardening rate is very high. 0.2% yield stress (0.2%YS),
which is defined as a yield stress in the dual phase steel because the steel never shown yield point phenomenon
as shown in Fig. 3.15, is relatively small. They had a larger ultimate tensile strength (UTS). The steels had a
low yield ratio, which is given by 0.2%YS/UTS. The steels were typical dual phase steels because it has low
yield ratio and high strength.
a) b) TD
F
M F
20m 20m
c) d)
M F
M
20m 20m
Fig. 3.11 Initial microstructure of dual phase steels: a) Steel 1, b) Steel 2, c) Steel 3 and d) Steel4. Equiaxed
grains are a ferrite phase (F) and martensite particles (M) have concave and convex profiles (marked by arrows).
In Steel 3 cementite precipitated in the martensite particles due to a high Ms temperature. The rolling direction
(RD) is marked.
Table 3.3 Grain size, volume fraction and aspect ratio of martensite particles in the dual phase steels.
Grain size / µm Volume fraction Aspect ratio Microhardness
Ferrite Martensite Ferrite Martensite of martensite Ferrite Martensite
Steel 1 (SVM) 7.3 3.6 0.79 0.21 2.1 162 398
Steel 2 (LVM) 7.1 6.3 0.62 0.38 2.2 163 387
Steel 3 4.8 - 0.34 0.66 - 164 350
Steel 4 6.8 4.3 0.69 0.31 2.4 168 386
SVM: Small Volume fraction of Martensite, LVM (Large Volume fraction of Martensite)
28
Mn segregation zone
20m
Fig. 3.12 The microstructure at the center of Steel 4. A band-like structure consisiting of martensite particles
was observed.
Table 3.4 Mechanical properties of the dual phase steels

0.2%YS UTS El. U-El. P-El.
Steel 1 361 683 19.1 10.4 8.7
Steel 2 456 868 13.7 8.4 5.3
Steel 3 603 1077 9.4 6.5 2.9
Steel 4 513 825 20.2 11.0 9.2
29
a)
1
0.8 SVM
LVM
Frequency
0.6
0.4
0.2
0
<2 2-4 4-6 6-8 8-10 10-12 12-14 14-16 16-18 18-20
Size / m
b)
1
SVM
0.8
LVM
Frequency
0.6
0.4
0.2
0
<2 2-4 4-6 6-8 8-10 10-12 12-14 14-16 16-18 18-20
Aspect ration
Fig. 3.13 Size distribution (a) and aspect ratio (b) of the martensite particles.
30
a) b)
500 500
Ferrite Ferrite
Microhardness, Hv 10gf
Microhardness, hv 10gf
400 400 Martensite
Martensite
300 300
200 200
100 100
0 0
0 200 400 600 0 200 400 600
Tempering temperature / °C Tempering tempeature / °C
Fig. 3.14 Tempering temperature dependence of microhardness of the martensite and ferrite in a) SVM and b)
LVM. The hardness of the martensite mainly depends on the tempering temperature.
1000
Nominal stress / MPa
800
600
400
SVM
200 LVM
0
0 0.05 0.1 0.15 0.2
Nominal strain
Fig. 3.15 Nominal stress-nominal strain curves of SVM and LVM.
31
Chapter 4
Void formation mechanism
4.1 Introduction
Void formation in metals affects the strength and the formability for tensile tests, hole-expansion tests and
bending tests. Dual phase steels are applied for automobile components as dual phase steels have a high strength
and large uniform elongation. The uniform elongation and strength are important parameters but critical to
formability is also the response to the crack formation and growth. Parameters to characterize this response
include the elongation in the necking zone of a tensile test, the hole-expansion ratio determined by pushing a
cone through a punched hole and the elongation during a bending test (e.g. Yamazaki 1995 et al. Hasegawa et al.
2004). To clarify in detail the material behaviours, the focus in the present work is a characterization of the
local deformation behaviour during loading the nucleation and growth of voids as studied by in-situ scanning
electron microscopy.
It has been suggested that void formation, growth and coalescence lead to fracture of specimens in dual
phase steels. Nucleation of voids has been an important objective in many investigations which have shown that
voids can nucleate in the martensite, in the ferrite and at inclusions and that void nucleation related to the
martensite particles is dominating as shown in Table 4.1. In the previous studies in Table 4.1, voids are
classified based on observation of deformed microstructures. Voids elongating perpendicular to a tensile
direction tend to be classified as voids in the martensite phase, which form due to cracking of martensite
particles. On the other hand, equiaxed voids and voids elongating in a tensile direction are often classified as
voids in the ferrite, which form at the martensite/ferrite interface due to decohesion. As a ferrite/ferrite grain
boundary often exists between two martensite particles and martensite particles are connected by a void after
deformation, decohesion at a ferrite/ferrite grain boundary is suggested as another void formation mechanism.
Such observations have been followed by suggestions and discussion of void nucleation mechanism where four
are illustrated in Fig. 4.1. The experimental basis for these mechanisms is the observation that most voids in
broken specimens are formed between two martensite particles in the necking zone which is illustrated in the
last column of Fig. 4.1. Four mechanisms which may lead to this end-scenario are also illustrated in Fig. 4.1a,
where the first row shows that voids have their origin in a crack in the martensite particle which breaks the
particle into fragments which separate and form the void. Other mechanisms are illustrated in Fig. 4.1b and 1c
where the first columns show decohesion at the martensite/ferrite interface leading to void formation. The bases
for this suggestion are observations showing that voids nucleation by cracking of the martensite particles are
limited whereas large voids are observed between martensite particles near the fracture (e.g. Avramovic et al.
2010). Fig. 4.1d exemplifies that decohesion at the ferrite/ferrite grain boundary leads to void formation
between two martensite particles. However, these mechanisms as well as other mechanisms are still ambiguous
and mechanisms are still explored as an improvement of the dual phase steels is an important industrial issue.
The present study thereby has an objective to study mechanisms behind void nucleation and growth and in-situ
testing in a scanning electron microscopy (SEM) has been chosen as an appropriate technique as it enables one
to follow the void formation behaviour in a specimen directly as a function of the applied strain and stress. Such
observations together with an analysis of cracked specimens have led to classification of cracks and voids
allowing quantification of two defects and an analysis of their importance. The study focuses on two types of
dual phase steels for which strength of the martensite particles, which is an important parameter, has been
varied by tempering treatments at 300-600°C.
32
Table 4.1 Observation of voids in dual phase steels.
authors date Void formation site
Martensite* Ferrite* inclusion Ferrite/ferrite
grain boundary
Koo and Thomas 1977 O
Tomota et al. 1977 ● ●
Rashid 1978 ▲ ▲ - -
Gerbase et al. 1979 O
Speich and Miller 1979 O
Korzekwa et al. 1980 O
Balliger 1982 O ▲ - -
Szewczyk and Gurland 1982 - O
He et al. 1984 O
Ray 1984 ▲ ▲ ▲
Shen et al. 1986 O - - -
Steinbrunner et al. 1988 ●(crack and ●
separation)
Han and Margolin 1989 O ▲
Sarwar et al. 1996 ●(crack and ●
separation)
Gladman 1997 O
Kim et al. 2000 ▲ O
Sun and Pugh 2002 ● ●
Erdogan et al. 2002 ● ● ▲
Hasegawa et al. 2003 ▲ O
Han et al. 2004 O
Poruks et al. 2006 - O
Mazinani and Poole 2007 O ▲
Kang et al. 2007 O - - -
Uthaisangsuk et al. 2008, 2009 O - - -
Maire et al. 2008 ▲ O
Tasan et al. 2009 O -
Kelestemur et al. 2009 O - ▲
Ahmed et al. 2009 ● ●
Avramovic et al. 2009 ▲ O ▲
Avramovic et al. 2009 ▲(crack and O ▲
separation)
Toji et al. 2009 ▲ O
Landron et al. 2010 O
Ghadbeigi et al. 2010 O -
Calcagnotto et al. 2011 ● ●
Kadkhodapour et al. 2011 ▲ O
O: major void formation site, ▲: minor void formation site, ●: observed void site
33
a) Void formation by cracking of martensite
ferrite
martensite crack void

Crack initiation Martensite separation
b) Void formation by decohesion between martensite and ferrite (1)
ferrite
martensite void void

Void initiation Void growth and coalescence
c) Void formation by decohesion between martensite and ferrite (2)
ferrite

Void initiation Void growth
d) Void formation by decohesion between two ferrite grains
ferrite
martensite grain boundary void void

Void initiation Void growth and coalescence
Fig. 4.1 Illustration of four void formation mechanisms between martensite particles in dual phase steels: a)
void initiated by cracking of a martensite particle , b) void initiation by decohesion at the interface between
ferrite and martensite and coalescence of two voids formed at the end of martensite particles, c) void initiation
by decohesion at the interface between ferrite and martensite and growth of a void formed at the end of a
martensite particle and d) void formation by decohesion at the ferrite/ferrite grain boundary.
4.2 Experimental procedures

A dual phase steel, SVM, without tempering treatments (see Section 3.6) is used to follow the void
formation, growth and coalescence. At first, the deformed microstructure after testing was studied from the
transverse direction to estimate the characteristics of voids including (i) initiation site, (ii) number density and
(iii) size. In order to detect void initiation sites, the microstructure was etched by nital. Comparisons of the
etched and unetched microstructures established that etching does not generate etch pits or affect the shape and
size of the voids. Void initiation sites were observed in areas with a limited elongation because a large
deformation resulted in microstructures which were too complex to characterize.
In addition, tensile tests and three point bending tests were carried out in a scanning electron microscope (see
Section 3.3). All specimens were polished and etched before deformation. During the tests the following
procedure has been related to nucleation and growth of voids; (i) nucleation density, (ii) size distribution, (iii)
length and (iv) area fraction. The observations have been related to the strain which is defined in Section 3.5.
34
Voids in the bulk after a tensile test were also characterized based on a classification by mechanisms suggested
through in-situ mechanical tests.
4.3 Results
4.3.1 Deformed microstructure after a tensile test

Fig. 4.2 shows typical microstructures after tensile testing at room temperature. In the figure, the horizontal
axis is the tensile direction. Most voids, which appear black in the figure, are located between martensite
particles (light grey) near the fracture surface or in the bulk. Since the initial microstructure contained no cracks
or voids, these voids had formed during the test. Figs. 4.2a and b demonstrate that, near the fracture surface, the
number of voids is large and they show a variation in size and shape. In particular, voids elongated in the tensile
direction are observed. Figs. 4.2c and d show that in the bulk away from the fracture surface the size and
number of voids are relatively small when necking started. Figs. 4.2a and b also show some voids which have
coalesced near the fracture surface. It is apparent that void formation and coalescence led to the rupture of the
specimen. Since the majority of the voids formed in the necking zone, void formation mainly affects the post-
uniform elongation (P-El).
a) b) TD
50m void 10m
c) d) TD
M M M
void void
50m 10m
Fig. 4.2 Typical microstructures in the dual phase steel (SVM) after a tensile test: a) and b) near the fracture
surface (t =0.429), c) and d) in the bulk (t =0.250). The tensile direction (TD) is marked.
Areas outside the necking zone have also been examined and typical microstructures are shown in Fig. 4.3,
illustrating three important crack and void formation sites. Here a distinction is made between a defect which
can propagate perpendicular to the maximum tensile direction which is called a “crack”, and a defect which can
grow in the tensile direction and the shear direction which is called a “void”. The three sites are the following:
(i) At an interface between ferrite and martensite but located in the martensite phase
35
(ii) At an interface between ferrite and martensite but located in the ferrite phase
(iii) At an inclusion
Examples of all three sites are given in Fig. 4.3:
(i) : A crack in a martensite particle is illustrated in Fig. 4.3a. The crack tends to form at the concave side of the
particle and propagates perpendicular to the tensile direction. Cracks that are not connected with the
martensite/ferrite interface are not observed. This type of crack is called “a crack in martensite”. The crack
often developed as a void due to crack propagation in a martensite particle and separation of the particle
fragments. The void is called “a void in martensite”.
(ii) : A martensite/ferrite interface void in ferrite is shown in Fig. 4.3b. This type of void initiates at the interface
between ferrite and martensite near the end of the martensite particle and grows in the tensile direction without
propagating along the interface or into the martensite. This type of void is called “a void in ferrite”.
(iii) : A void at an inclusion which may be initiated by decohesion between the ferrite matrix and the inclusion
is shown in Fig. 4.3c. In the present samples such voids are few as the sulphur and oxygen content was low.
In general, the three types of voids nucleate and propagate due to the stress and strain which build up at the
martensite particles and inclusions. In this respect the behaviour of the dual phase steel will be discussed in
Chapter 5.
a) b) c) TD
M
Inclusion
M
M M
M
void void
crack
M
5m 5m 5m
Fig. 4.3 Typical voids in the dual phase steel (SVM) deformed in tension to a thickness strain of  t = 0.111: a)
void in martensite, b) void in ferrite, c) void at an inclusion. The strain ( t = 0.111) is larger than the uniform
elongation (U-El = 0.104).
4.3.2 Void formation behaviour during in-situ testing

A typical observation during an in-situ tensile test is shown in Fig. 4.4. The microstructures shown in this
figure has been taken from an area near the hole in the specimen. Fig. 4.4a shows neither voids nor cracks in or
near the martensite particles before deformation. However, after deformation to a strain of h=0.159, a small
crack (crack1) was initiated in a martensite particle (Fig. 4.4b). At a larger strain (h=0.227), an additional crack
(crack 2) has divided the martensite particle in Fig. 4.4c into two, and crack1 has propagated in the martensite
particle. At a strain of h=0.623, crack1 has propagated perpendicular to the tensile direction and has passed
through the martensite particle as shown in Fig. 4.4d. However, once the crack has passed through the particle,
it stops and does not propagate into the ferrite even when the strain is raised further.
After propagation, the morphology of the martensite separation (crack1) in Fig. 4.4d is similar to that shown
in Fig. 4.4c (crack2). It is therefore possible that crack 2 may also have its origin in a crack in the martensite
particle. Fig. 4.4d also shows that the martensite separation around crack2 can lead to void formation at the
martensite. After fracture of the particles the two parts of the martensite particle have moved apart in the tensile
direction and a void has formed. The width of the void (defined as the dimension perpendicular to the tensile
axis) is almost equal to the thickness of the martensite particle after fracture and the width does not depend on
strain as the void growth is predominantly in the tensile direction. Most of the voids in the bulk of the tensile
sample illustrated in Fig. 4.2 have a similar morphology, suggesting that they have formed by the same
36
mechanism, involving cracking and separation of martensite particles and in-situ observations also appear to be
representative of the bulk behaviour.
a) b) TD
necking crack 1
5m 5m
c)
e= d)
0.
22
7
crack 2 crack 1
void crack 1
5m 5m
Fig. 4.4 In-situ observation of void formation in the dual phase steel (SVM) during a tensile test: a) initial
microstructure without cracks or voids, b) crack initiation in martensite (h= 0.159), c) crack propagation in
martensite (h= 0.227) and d) void growth and crack propagation ( h = 0.623). A definition of a local strain (h)
is given in Table 3.1.
Another example of voids formed during in-situ tensile testing is shown in Fig. 4.5. As for Fig. 4.5 the right
column of images are enlarged sections of the left column. Necking of the martensite and multiplicative void
formation in the martensite can be observed. The initial microstructure consists of elongated martensite particles
which are aligned along the tensile direction (Figs. 4.5a and b). The numbering of the ferrite grains in Fig. 4.5a
refers to section 4.4.2 where the strain of each ferrite grain is estimated for further analysis and discussion.
After deformation to a strain of h=0.067, a crack has propagated in a martensite particle between the ferrite
grains labelled 5 and 7 as shown in Figs. 4.5d and e. At the larger deformation in Figs. 4.5g-h, multiple voids
have formed in parallel with necking of the martensite. The voids at the martensite tend to form at the concave
parts or the necked regions and multiple void formation in the martensite is often observed. Local strains
measured in the necked regions, L1 and L2, are 0.23 to 0.28, i.e. larger than the average strain (h) because the
deformation concentrates at the concave parts of the martensite particles.
37
a) b) c)
2 3
1 4 6
5
8
7
10
9 11
10m 5m necking 1
d) e) f)
L1
void
10m 5m
g) h) i)
void necking 2
void L2
10m 5m
j) k) l)
void
void
10m 5m
Fig. 4.5 In-situ observation of void formation in the dual phase steel (SVM) during a tensile test at low (left
column) and high (right column) magnifications: a) and b) initial microstructure without cracks or voids, d) and
e) necking of a martensite particle and void formation in a martensite particle (h= 0.067), g) and h) necking,
formation and growth of voids (h= 0.200), j) and k) formation and growth of voids in martensite particles ( h =
0.422). L1 and L2 show local strain in the necking zone, c), f), i) and l) sketches from b), e), h) and k) of the
martensite particles, necked regions and voids. A definition of a local strain at the necked region of a martensite
particle (L) is given in Table 3.1.
38
a) b) c)
2m 2m 2m

d) e) f)
crack
block
void
Fig. 4.6 Crack propagation behaviour in the dual phase steel (SVM): a) before deformation (h=0), b) after
deformation to h =0.376 and c) h =0.471. d), e) and f) sketches from a), b) and c) of the martensite particle. A
crack passed through a block illustrated in d)-f) during the in-situ test.
During crack propagation, cracks often cross block boundaries consisting of a lath of martensite as shown in
Fig. 4.6. The specimen was tempered at 300°C for 100s to observe block boundaries. Before deformation, a
block was located at a concave part of a martensite particle. Figs. 4.6d-f highlight the deformation and crack
propagation behaviour in the block. At a strain of h =0.376, a crack initiated at the martensite/ferrite interface
on the martensite side after plastic deformation of the concave part and crossed the block with strain. After
deformation to a strain of h =0.471, the crack divided the martensite particle into two but the crack stopped
propagation into ferrite grains. As lathes of martensite generally align along the long axis of a block, it appears
that the crack passed through lathes of martensite. Steinbrunner et al. and Sarwar et al. suggested that separation
of prior austenite grain boundaries leads to crack formation in martensite particles due to segregation of
phosphorus on the prior austenite grain boundaries. However, in-situ observations have revealed that cracks in
the martensite are not related to prior austenite grain boundaries, but the shape of the martensite particles. In
addition, prior austenite grain boundaries are not located in martensite particles in dual phase steels produced
through hot-rolling processes as ferrite grains nucleate at prior grain boundaries.
A large deformation leads to void coalescence as shown in Fig. 4.7 observed in a bending test similar to the
one shown in Section 3.4. Figs. 4.7a and b show martensite particles without cracks and voids in the
undeformed structure. Deformation to a strain of 0.366 led to fracture and separation of the martensite particles
at the concave part as shown in Fig. 4.7e. After deformation to a strain of 0.545 in Figs. 4.7g and h, the voids
have grown in the shear direction (SD) and coalesced to form an extended void with a length of about 14m.
39
a) b) c)
10m 5m
d) e) SD f)
void
10m 5m void

g) h) SD i)
10m 5m coalescence

Fig. 4.7 In-situ observation of void formation in the dual phase steel (SVM) during a bending test (shear
directions (SD) are marked by arrows): a) and b) initial microstructure without cracks or voids, d) and e) void
formation in martensite particles (h= 0.366), g) and h) void coalescence (h= 0.545, c), f) and i) sketches from
b), e) and h) of the martensite particles and voids.
Although void formation in the ferrite matrix was rarely observed, Fig. 4.8 shows an example. Before
deformation, ferrite grain boundaries are seen but the ferrite matrix does not contain voids (Fig. 4.8a).
Deformation to a strain of 0.045 did not produce voids in the ferrite matrix but created slip bands in a ferrite
grain at a tip of a martensite particle (Fig. 4.8b). As the strain was increased in Fig. 4.8c, a void has appeared
near the end of a martensite particle on the ferrite side. Even though a further strain was applied, the void grew
in the tensile direction without propagating along the martensite/ferrite interface and it did not reach another
martensite particle on the left side. In the ferrite voids often formed at a triple junction consisting of two ferrite
grains and a martensite particle as a ferrite/ferrite grain boundary was frequently located near the tip of the
particles. However, decohesion between the ferrite/ferrite grain boundary, which was suggested as a main void
formation mechanism in a dual phase steel by Kadkhodapour et al. (2011), was not observed.
a) b) c)
TD
5µm 5µm 5µm

Fig. 4.8 In-situ observation of void formation in the ferrite during a tensile test: a) initial microstructure without
cracks or voids, b) slip band formation in the ferrite (h= 0.045) and c) void formation in the ferrite (h= 0.262).
The void (marked by an arrow) initiated at the martensite particle on the right side did not reach another
martensite particle on the left side. 
40
3000
Number density / 1/mm2

in martensite
in ferrite
2000
at inclusion
1000
0
0 0.2 0.4 0.6 0.8
Local strain, h
Fig. 4.9 Void number density in the martensite and the ferrite as a function of the local strain (see Table 3.1) in
an in-situ test.
To evaluate the effect of void formation during deformation, the number density, size distribution, average
length, maximum length and area fraction of voids have been quantified. As an example, Fig. 4.9 shows the
void number density in the martensite and the ferrite during an in-situ tensile test. Voids form predominantly in
the martensite and already at a low strain whereas few voids form in the ferrite and at a larger strain. This agrees
with previous observations showing that cracks can form in martensite before necking of the specimen while
voids appear in ferrite after necking (e.g. Erdogan 2002, Avramovic et al. 2009 and 2010, Calcagnotto et al.
2011). In addition, voids started to appear at inclusions at a very low strain and the number density saturated at
a strain of h=0.067 due to weak bonded interfaces and a low density.
41
a)
1500
εh=0.159
1000
εh=0.227
εh=0.357
500
εh=0.623
0
<1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 9-10
Length range of void / m
b)
1500
εh=0.159
1000
εh=0.227
εh=0.357
500
εh=0.623
0
<1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 9-10
Length range of void / m
Fig. 4.10 Size distribution of voids formed in the dual phase steel (SVM) during an in-situ tensile test: a) in the
martensite, b) in the ferrite.
The size distribution of voids is shown in Fig. 4.10. For all strains, in both the martensite and the ferrite, the
number of small voids dominates. This indicates that voids nucleate when the local strain overcomes the critical
strain for void formation. The voids grow steadily with strain until the specimen breaks. A continuous void
formation was observed in previous works through an observation of a deformed specimen and X-ray
tomography (e.g. Maire et al. 2008, Landron et al. 2010, Avramovic et al. 2010). Figs. 4.11a and b show the
influence of local strain on void length in the tensile direction. Both the average void length and the maximum
void length in the martensite are larger than in the ferrite. However, the difference in average length between
the martensite and ferrite is relatively small because many small voids are generated in the martensite during
deformation. Also, the strain dependence of the void length in the martensite is larger than in the ferrite. The
area fraction of voids is estimated as a function of strain in Fig. 4.11c. The area fraction in both the martensite
42
and the ferrite increases with strain because of void nucleation and growth. Obviously, voids in the martensite
also occupy a relatively large area at a large local strain leading to fracture of the specimen.
a) b)
12 12
Maximum length / mm
Average length / mm
10 in martensite 10 in martensite
8 in ferrite 8 in ferrite
6 6
4 4
2 2
0 0
0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8
Local strain, h Local strain, h
c)
1.2
1 in martensite
Area fraction / %
0.8 in ferrite
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8
Local strain, h
Fig. 4.11 Influence of local strain on void formation behaviour in the dual phase steel during an in-situ tensile
test: a) average length, b) maximum length and c) area fraction.
4.4 Discussion
The structure of the tensile deformed specimens and the structural evolution during in-situ testing
show that after an initial small strain cracks in the martensite particles start to appear and develop into
voids during further straining. At small strains voids also form in the ferrite, however, with a much
smaller frequency (see Fig. 4.12). Voids are observed at inclusions but their frequency is low due to
the purity of the steel. Based on these microstructural observations, cracks and voids related to
martensite particles will be discussed in more detail also considering the hardening behaviour of
martensite and ferrite and how this behaviour can be affected through annealing of the dual phase
steel. In a final section void formation in ferrite is discussed. The structural evolution has also been
studied in bulk specimens tested in tension and the observations indicate that the results obtained by
in-situ testing are representative for the bulk behaviour of the dual phase steel.
4.4.1 Plastic deformation of martensite

In an analysis of the deformation behaviour of the dual phase steel an important issue is the individual
behaviour of the martensite phase and the ferrite matrix and how interaction takes place in the composite
structure. In particular, deformation of the martensite is related to void formation in the martensite. For
43
example, plastic deformation of martensite particles starts at small strains (h≈0.05) during in-situ loading and
takes place preferentially at the concave part of the particles before cracks are initiated (Fig. 4.5). A hardness
test also demonstrates that martensite and ferrite can deform plastically due to the hardening behaviour (Fig.
4.12). However, the hardening rate of the martensite is smaller than the ferrite due to a small work hardening
behaviour of the martensite phase and strain partitioning between the martensite and ferrite (e.g. Shen et al.
1986, Su and Gurland 1987, Tasan et al. 2010, Ghadbeigi et al. 2010). Similar findings are reported after a
bending test of a dual phase steel (e.g. Hayashi et al. 2012). It is also found that in the bulk the martensite
started to deform plastically at a small strain through an analysis of the work hardening behaviour of the bulk
specimen during tensile testing (e.g. Lian, Jiang et al. 1991) and in-situ observation by neutron diffraction (e. g.
Morooka et al. 2012) and X-ray diffraction (e.g. Jia et al. 2009). It is obvious that deformation of the martensite
is related to void formation in the martensite. In this study the onset of plastic deformation of the martensite in
the bulk specimen was estimated based on the stress-strain curve of the dual phase steel (Fig. 3.15) and the
modified Crussard-Jaoul analysis (e.g. Lian et al. 1991, Jiang et al. 1995). Fig. 4.13 shows that the work
hardening behaviour rate is divided into three regions; (i) the work hardening rate is the highest of all, (ii) the
work hardening rate decreases gradually with deformation and (iii) the work hardening rate drops rapidly. The
large work hardening rate is related to a large number of dislocations piled up at the martensite/ferrite interfaces
because the martensite deforms only elastically. On the other hand, the change of the slope shows that plastic
deformation starts in the martensite as pile-up dislocations lead to yielding of the martensite and plastic
deformation of the martensite accommodates the work hardening in the ferrite. A transition strain (trans) from
elastic to plastic deformation of the martensite is 0.03. In previous works, it is shown that the martensite phase
starts to deform plastically less than a strain of 0.06 even though a martensite phase is very hard. Before the
start of necking, the work hardening rate has already decreased rapidly. The drop of the work hardening rate is
related to a reduction of cross section by void formation. It is obvious that void formation before the onset of
the necking reduces the work hardening rate and reduces uniform elongation. In parallel with the plastic
deformation, deformation of the ferrite grains takes place as illustrated by the observation of coarse slip bands
(Figs. 4.4b, 4.5c and h). The observations also show that a fairly large strain is required before the first void
appears in the ferrite (Fig. 4.8).
It is also pointed out that plastic deformation of martensite particles is proceeded to separation of the
deformed particles (e.g. Steinbrunner et al. 1988). Whereas plastic deformation of the martensite phase has an
important role on void formation in dual phase steels, brittle fracture of martensite particles is sometimes
assumed to occur in dual phase steels and only plastic deformation of a ferrite matrix is taken into account in
order to simulate void formation in dual phase steels (e.g. Kadkhodapour et al. 2011, Vajragupta et al. 2012).
The present study has demonstrated that plastic deformation of the martensite is a key factor to improve the
formability of dual phase steels. In addition, as shown in Fig. 4.5, the critical strains at the necked region of the
martensite particles for void formation in the martensite were about 0.23-28 and the local stain in the martensite
for crack formation was much larger than the local strain of the specimen (h), even though the critical local
strain for void formation is about 0.05, which is also reported by Steinbrunner et al. (1988). The critical local
strain for void formation in the martensite did not depend on volume fraction and morphology of the martensite,
but the hardness of the martensite (e.g. Azuma et al. 2012). This result indicates that plastic deformation of the
martensite is a key factor to control the formability of dual phase steels. Strain partitioning between soft ferrite
and hard martensite depends on the volume fraction of martensite phase and hardness of martensite phase (e.g.
Kadkhodapour et al. 2011, Lian et al. 1991, Jiang et al. 1995 and Azuma et al. 2012).
44
500
Microhardness, Hv 10gf
400
300
200
martensite
100
ferrite
0
0 0.1 0.2 0.3 0.4 0.5
Strain, t
Fig. 4.12 Hardness of martensite and ferrite in the dual phase steel (SVM) after tensile testing.
11
Region Ⅰ Region Ⅱ Region Ⅲ
10
Transition strain (trans )
9
ln(d/d)
6
6 6.2 6.4 6.6 6.8
ln()
Fig. 4.13 An analysis of the work hardening behaviour in the dual phase steel (SVM) based on the modified
Crussard-Jaoul method. In region Ⅰ, the ferrite phase deformed plastically but the martensite deformed
elastically. Region Ⅱ means that plastic deformation occurred in the ferrite and martensite. Region Ⅲ is related
to a reduction in the area due to void formation.4.4.2 Plastic deformation of ferrite.
In an analysis of the deformation behaviour of the dual phase steel an important issue is the individual
behaviour of the martensite particles and the ferrite matrix and how interaction takes place in the composite
structure.
In parallel with the plastic deformation of the martensite particles, plastic deformation of the ferrite grains
takes place as illustrated by the observation of coarse slip bands (Figs. 4.5b, 6c and h). The void formation is
proceeded by localized glide representing slip band formation around the concave part and the tip of the
martensite particles. Similar findings have been reported that slip bands formed in the ferrite composing of dual
phase steels before cracking of martensite particles. It is apparent that the deformation of soft ferrite grains is
related to deformation of hard martensite particles if the grains are constrained by the particles. The application
of the in-situ technique has also allowed a monitoring of the strain in the ferrite grains in an area where voids
45
have been observed. This has been done by following the length change of the ferrite grains (Fig. 4.5). The
strains are given in Table 3.1, which reveals that the strains in individual grains increase with the strain of the
specimen (h). This shows that the effects of the grain size and the crystallographic orientation of the ferrite
grains are of minor importance. However, Table 4.2 also shows that a number of grains have deformed to
relatively large strains and that these grains are located next to the broken martensite particles, although some
grains, which are not constrained by martensite particle, were heavily deformed. Once a large deformation was
induced in the ferrite grains due to crack formation in the particles, other voids formed in other martensite
particles locating to opposite side of the broken martensite. It appears that larger strain in the ferrite grains next
to broken martensite leads to larger plastic deformation of other martensite particles and multiple void
formation. The observation illustrates the complex deformation pattern of the dual phase steel and that the
deformation behaviour of both phases as well as the interaction between martensite and ferrite is important
since the multiple void formation leads to fracture of specimens.
Table 4.2 Strains of ferrite grains in an area with broken martensite particles as shown in Fig. 4.5.
Grain h=0.067  h =0.200  h =0.422
1 0.050 0.125 0.150
2 0.125 0.188 0.625
3 0.130 0.261* 0.522*
4 0.067 0.200 0.467
5 0.104* 0.313* 0.750*
6 0.053 0.263 0.947
7 0.111* 0.286* 1.029*
8 0.024 0.122 0.308
9 0.063 0.191* 0.468*
10 0.056 0.167 0.556*
11 0.080 0.240 0.760
* Formation of voids in neighbouring martensite particles
In addition, it is apparent that the deformation of soft ferrite grains is related to deformation of hard
martensite particles if the grains are constrained by the particles. The application of the in-situ technique has
also allowed a monitoring of the strain in the ferrite grains in an area where voids have been observed. This has
been done by following length change of the ferrite grains (Fig. 4.5). The strains are given in Table 4.2, which
reveals that the strains in individual grains increase in relatively good accordance with the strain of the
specimen (h). This shows that the effects of the grain size and the crystallographic orientation of the ferrite
grains are of minor importance. However, Table 4.2 also shows that a number of grains have deformed to
relatively large strains and that these grains are localized next to the broken martensite particles, although some
are also found in other locations. A more detailed analysis requires a larger experimental basis but the
observation illustrates the complex deformation pattern of the dual phase steel and that the deformation
behaviour of both phases as well as the interaction between martensite and ferrite is important. Landron et al.
pointed out that reduction of a cross section by void formation accelerates void formation based on an
engineering model and an observation by X-ray tomography (2010). It supplements that the multiple void
formation is an important process in dual phase steels.
4.4.2 Cracks in martensite

Most voids are formed in the martensite particles following cracking of the particles. In the tensile deformed
microstructure, cracks initiated at the martensite/ferrite interface on the martensite side after plastic deformation
of martensite particles as illustrated in Figs. 4.4-6. In all cases, decohesion and void formation at the
martensite/ferrite interface was not observed prior to crack initiation in the martensite. It is apparent that the
dual phase steel before deformation is free of cracks and voids but already at a small strain cracks in martensite
start to appear following necking of the particles (Fig. 4.4). Many of these cracks form in martensite particles
elongated in the tensile direction and they are typically observed where the surface of the particles is concave
pointing to an increased stress and strain in such regions. In addition, crack propagation across a block
boundary suggests that the shape of the martensite particles has a larger effect on crack formation in the
martensite whereas the block boundaries and packet boundaries contribute to the strength and the ductility of
46
the martensite. In parallel with the initiation of small cracks in the martensite, deformation localizes more in the
ferrite grains next to cracks as illustrated by the observation of coarse slip bands (Figs. 4.4c, d and 4.5h). These
observations point to a combined effect of strain, particle size and morphology on the crack initiation in the
martensite. The findings supplement previous observations showing that cracks form in martensite particles at a
smaller strain in a dual phase steel containing large martensite particles, which have elongated shapes and
concave profiles, than in a steel with small martensite particles (e.g. He et al. 1984, Erdogan 2002). Similarly
cracks in a spheroidized steel, consisting of a ferrite matrix and cementite particles, form in large cementite
particles with concave parts and propagate in the particles perpendicularly to the tensile stress direction during
tensile tests, compression tests and torsion tests (e.g. Gurland 1972), which has led to the suggestion that large
particles are preferred crack formation sites. However, in the present study, cracks did not form in large
martensite particles but in elongated particles or at the concave parts of the particles even though the
microstructure is composed of both small and large martensite particles. It appears that it is related to
deformation of the martensite as the large martensite particles were deformed less than the thinner particles.
Tasan et al. (2010) also pointed out that thinner martensite bands are deformed more than thicker martensite
bands in a dual phase steel during an in-situ tensile test.
4.4.3 Voids in martensite

The present observations show in agreement with previous studies that voids form in the martensite particles
as shown in Table 4.1 (e.g. Koo et al. 1977, Balliger 1982, He et al. 1982, Shen et al. 1986 and Gladman 1997).
It is also found that both in the tensile test of the bulk specimen and in-situ testing voids start to form in the
martensite at relatively small strains about t , h = 0.05 (Fig. 4.9) which agrees with the findings for a dual
phase steel deformed in tension that voids form at a strain of 0.034 (e.g. Calcagnotto et al. 2011). However,
application of the in-situ techniques has illustrated void nucleation related directly to plastic deformation of the
martensite particle and crack formation in the particles. On the other hand, Avramovic et al. (2011) have
suggested that elongated voids between two martensite particles originate in void formation in ferrite phase as
voids at the martensite/ferrite interface should grow more than in the martensite. In addition, they have
suggested that as fracture occurs in the specimen after void formation in the ferrite, voids in the ferrite have a
more negative effect on the formability in the dual phase steel. However, the in-situ test has demonstrated that
cracks in the martensite particles develop into voids and voids in the martensite grow in the tensile direction
more than in the ferrite. Moreover, the observation has clarified that influence of voids in the ferrite is minor
due to the small area fraction even though some voids form in the ferrite in advance of the fracture.
Based on the observations it is suggested that cracks and voids in the martensite particles evolve in four
steps: (i) plastic deformation of the martensite, (ii) crack initiation at the martensite/ferrite interface, (iii) crack
propagation leading to fracture of the martensite particles and (iv) void formation by separation of the particle
fragments. Besides, the in-situ observations have also allowed quantification of the size and number of voids
and their area fraction as a function of a given strain. Fig. 4.10 shows that both the fraction of large voids and
the area fraction of voids increase rapidly for strains at 0.36 and above. This illustrates the importance of voids
in the martensite and also that routes to reduce such voids must be explored in order to increase the formability
of a dual phase steel. One such route is discussed in the next section.
4.4.4 Voids in ferrite

The voids in ferrite are typically observed near the end of the martensite particles (Figs. 4.3b and 4.8). It is
also found that voids never propagate along the martensite/ferrite interface and the ferrite/ferrite grain boundary
but grow in the tensile direction. The absence of decohesion points to a high strength of the martensite/ferrite
interface and the ferrite/ferrite interface as decohesion is an indication of a weak interface, for example,
between inclusions and ferrite or martensite in steel (e.g. Roberts et al. 1976, Qiu et al .1999, Sabirov and
Kolendnik 2005, León-Garcia et al. 2010, Hosseini et al. 2007) and between reinforcement and matrix in
metal-matrix composites (e.g. Nutt et al. 1987, Christman et al. 1989, Clyne and Withers 1993, Whitehouse and
Clyne 1993). In the present specimens the voids in ferrite may have their cause in a high stress concentration
near the end of the elongated hard martensite particle as the interface is faced perpendicular to the tensile
direction and deformation concentrates in ferrite grains between martensite particles. The concentrated stress
may be partly relaxed by plastic deformation of the ferrite, creating the dislocations and vacancies required for
void growth composites (e.g. Lubarda et al. 2004, Traiviratanav et al. 2008, Meyers et al. 2009).
47
Finally, the void growth has been quantified in more detail based on a structural evolution. Fig. 4.14 shows
that the void initiated at the interface, which was shown in Fig. 4.8, grew in the tensile direction, the ferrite
grain between two martensite particles was elongated until fracture of the specimen and the void did not reach
another martensite particle. Voids at a martensite particle did not reach another particle even though the length
between two martensite particles was about 1 µm (Fig. 4.15) and voids formed at a triple junction of two ferrite
grains and a martensite particle (Fig. 4.16). In addition, the ferrite grain remained between two martensite
particles during the tensile test and decohesion at ferrite/ferrite grain boundaries did not occur in the dual phase
steels as shown in Fig. 4.16. The characteristic void formation behaviour in the ferrite also supplements that
void formation in the martensite is dominant in the dual phase steels. Moreover, the complicated void formation
behaviour in dual phase steels; voids formed in the ferrite are still isolated at the martensite/ferrite interface but
cracks initiated in the martensite exist as a void between two martensite particles, makes deformed
microstructures confusing. The in-situ scanning electron microscopy is a powerful technique to reveal fracture
behaviour in dual phase steels.
25
20
Length / µm
15 M-M
ferrite
10 void
0
0 0.2 0.4 0.6 0.8
Strain, h
Fig. 4.14 Length change between two martensite particles, of a ferrite grain between two the particles and a
void in the ferrite in Fig. 4.8.
a) b) c)
TD
5m 5m 5m


Fig. 4.15 In-situ observation of void formation at a ferrite/martensite interface in SVM tempered at 500°C
during a tensile test: a) initial microstructure without cracks or voids, b) slip band formation in the ferrite (h=
0.481) and c) void formation in the ferrite (h= 0.880). The void (marked by an arrow) initiated at the martensite
particle on the right side did not reach another martensite particle on the left side.
48
a) b) c)
TD
2m 2m 2m

Fig. 4.16 In-situ observation of void formation at a triple junction of two ferrite grains and a martensite particle
in LVM during a tensile test: a) initial microstructure without cracks or voids, b) slip band formation in the
ferrite (h= 0.118) and c) void formation in the ferrite (h= 0.310). The void (marked by an arrow) initiated at
the martensite particle on the right side did not reach the other martensite particle on the left side.
4.4.5 Void formation behaviour at inclusions

Few voids formed at inclusions at smaller strains even though the inclusion density is low. Fig. 4.17
demonstrates typical void formation by decohesion at the martensite/inclusion interface. At a strain of h= 0.046,
cracks had already propagated along the interface and voids surround the inclusions. Some voids existed at
inclusions before deformation. A crack tended to initiate at the interface facing perpendicular to the tensile
direction, propagated along the interface and then plastic deformation of the surrounding ferrite grains allow the
crack to develop as a void. It appears that a large stress concentration is at the interface perpendicular to the
tensile direction, weak bonded interfaces and stress concentration at a tip of the cracks contributes void
formation at inclusions. Kadkhodapour et al. (2011) suggested that stress is large at a metal/inclusion interface
facing perpendicular to a tensile axis, once a crack forms at the metal/inclusion, stress concentrates at the tip of
the crack and the crack propagates along the interface very rapidly through a FEM analysis. On the other hand,
voids at the martensite/ferrite interface did not propagate along the interface, but grew in the tensile direction.
The different void growth behaviour indicates that the interfacial strength between the martensite and ferrite is
relatively large. In addition, Table 4.3 demonstrates interfacial strength, which is estimated by comparison
experimental results with a physical model, between ferrite matrixes and inclusions, ferrite grains and
martensite particles. Interfacial strength between martensite and bainite (lathes of ferrite) is about 2-2.4GPa and
interfacial strength between ferrite and inclusions is about 1-2GPa. It is obvious that the ferrite/inclusion
interfaces, such as carbide, nitride, sulphide and oxide, are weak, but the martensite/ferrite interface is strong.
Therefore, the large interfacial strength between the martensite and bainite contributes to the absence of
decohesion between the martensite and ferrite. Plastic deformation of the surrounding ferrite grains allow a
crack to develop as a void. It, therefore, appears that the voids in Fig. 4.16 and Fig. 4.3c surrounded inclusions.
The rapid void formation at inclusions leads to saturation of the number density of voids at inclusions at a small
strain (h<0.05) as shown in Fig. 4.9.
49
Void
Inclusion
martensite
1m
Fig. 4.17 Void formation at inclusions (h= 0.046) in the dual phase (SVM). The voids, inclusions and
martensite are marked by arrows. The inclusions are surrounded by voids as the voids have already formed at a
small strain, propagated along the interface between the inclusions and martensite.
Table 4.3 Interfacial strength between ferrite and martensite, inclusions.

Particle Volume fraction nuc Interfacial strength / GPa Reference
martensite 0.10 0.9 2.4-2.5 Poruks P. Et al.
Y 2O 3 0.02-0.10 0.02-0.13 1-1.6 Kosco J. B. et al.
MnS 0.0023 0.36 1.1-1.4 Qiu H. Et al.
Fe3C 0.02-0.14 0.45-0.62 1.2-2.0 LeRoy G. Et al., Kwon D.
4.4.6 Void formation behaviour in the bulk

Void formation behaviour in the bulk after the tensile test was also estimated. Each void in the bulk was
characterized based on the classification by the in-situ observation; voids induced by a crack in the martensite
exist between two broken martensite particles, and voids formed in the ferrite touch only a martensite particle. It
is assumed that voids touches only one martensite particle are characterized as a void in the ferrite because
simultaneous void formation was not observed at the tip of two martensite particles, which faced each other,
due to heterogeneous morphology of the particles and stress relaxation by single void formation. Fig. 4.18
demonstrates that, in the bulk specimen, voids in the martensite are more frequent and larger than in the ferrite,
and occupy the majority of the area. It is apparent that the voids in the martensite mainly contribute to the
formability in dual phase steels. Hosseini et al. (2007) and Roberts et al. (1976) pointed out that in steels with
MnS same mechanisms such as crack formation in MnS or decohesion between steels and MnS dominate void
formation on the surface of specimens and in the bulk. However, crack formation and decohesion occurs more
frequently on the surface due to out of plane deformation. On the other hand, X-ray observation has exemplified
that voids form more frequently in the centre of specimens than near the surface after the start of necking (e. g.
Maire et al. 2008). It appears that necking affects the stress distribution in the bulk specimen and more voids
form at the centre of the specimen. Specially, a stress distribution, such as stress triaxiality and hydrostatic
stress, is one of the most important parameters to affect void formation in materials.
50
a) b)
5000 2
Number d ensity / 1/mm

in marten site
2 4000 in martensite
Area f raction / %
in f errite 1.5
in f errite
3000
1
2000
1000 0.5
0
0
0 0.2 0.4 0.6 0 0.2 0.4 0.6
Strain, t
Strain Strain, t
Strain
c) d)
5 40
Maximum length / mm
Average lengt h / mm
in martensite in martensite
4
30 in f errite
in f errite
3
20
2
1 10
0 0
0 0.2 0.4 0.6 0 0.2 0.4 0.6
Strain,
Straint Straint
Strain,
Fig. 4.18 Void formation behaviour in the bulk of the dual phase steel: a) number density, b) maximum length,
c) average length and d) area fraction of voids. Each void is characterized by the following classification. Voids
induced by cracking of martensite particles exist between two broken martensite particles and voids formed in
the ferrite touch at only a martensite particle.
51
4.5 Summary
Crack and void formation during deformation of a dual phase steel (SVM) has been followed by tensile
testing of bulk specimens and by in-situ loading in a scanning electron microscope. The following conclusions
are reached.
1. Cracks and voids have been observed at different sites; (i) in the martensite phase, (ii) in the ferrite
phase and (iii) at inclusions. At a given strain both the frequency and the area fraction of voids in
martensite dominate.
2. The cracks and voids in the martensite particles evolve in four steps; (i) plastic deformation of the
martensite, (ii) crack initiation at the martensite/ferrite interface, (iii) crack propagation leading to
fracture of martensite particles and (iv) void formation by separation of particle fragments. The in-situ
observations have allowed void nucleation to be related directly to plastic deformation and cracking
of martensite particles.
3. The in-situ observations have revealed the absence of decohesion at the martensite/ferrite interface
and ferrite/ferrite grain boundary. The voids on a martensite particle do not reach another martensite
particle even though voids grow in the tensile direction.
4. The voids in ferrite are predominantly formed at the tip of elongated martensite particle interfaces
perpendicular to the tensile axis and grow in the tensile direction.
5. Characteristic void morphology enables one to distinguish voids in the ferrite from voids in the
martensite in bulk specimens. Martensite cracks exist as elongated voids between martensite
fragments and voids in the ferrite are still isolated.
52
Chapter 5
Effect of martensite hardness
5.1 Introduction
As mentioned in the last chapter, a majority of voids formed in the martensite, but voids rarely formed in the
ferrite by the tensile tests of bulk specimens and in-situ tensile tests in a SEM. In particular, as more voids form
after the onset of necking, void formation strongly affects post-uniform elongation. In composite structures, as a
matrix interacts with reinforcements, a number of parameters related to martensite have been suggested which
may affect the crack and void formation, especially the hardness of the martensite and the grain size of the
ferrite based on a concept that void formation has an origin in the difference of mechanical properties of the
martensite and ferrite.
Mazinani and Poole (2007) suggested that softening of the martensite through intercritical annealing at high
temperature, where the carbon content in the austenite is low and thereby makes the martensite soft, improves
the true fracture strain. However, as intercritical annealing affects the volume fraction, shape and size of the
martensite, which might affect void formation, the effect of the hardness of the martensite is still ambiguous.
Tempering treatments are also suggested as a method to improve the formability, counterbalanced by a
reduction of the ultimate tensile strength. Fang et al. (2003), Samuel (1985), Joarder et al. (1990) and Hasegawa
et al. (2004) reported that softening of the martensite through tempering treatments increases the elongation and
hole-expansion ratio. In particular, post-uniform elongation is significantly improved by the treatments.
However, they did not observe the void formation behaviour in the dual phase steels in detail even though void
formation leads to fracture of the specimen. Jardim et al. (1984) and Tomota et al. (1977) suggested that
softening of the martensite assists plastic deformation of the martensite and retards void formation in the dual
phase steels. However, criteria for void formation and how the hardness of the martensite affects void formation
are still ambiguous.
On the other hand, Calcagnotto et al. (2011) focused on the hardness of the ferrite phase to improve the
formability in dual phase steels. They suggested that hardening of the ferrite through grain refinement made the
martensite deform more, further deformation of the martensite reduced deformation concentration at the
martensite/ferrite interface and the grain refinement contributed to retardation of voids in the dual phase steels
based on a SEM observation. The ferrite grain size was controlled by hot-rolling and annealing and the heat
treatment also strongly affected the shape of martensite particles. However, they did not show unambiguous
evidence. Specifically, in the coarse grained dual phase steel martensite particles surrounding ferrite grains were
deformed as much as the ferrite grains, whereas in the fine grained dual phase steels equiaxed martensite
particles were stretched in the tensile direction after a tensile test. The effect of martensite shape was not taken
into account, even though the shape of martensite particles is one of the most important parameters to affect the
interaction between the matrix and reinforcements in composite structures.
The present study therefore aims at clarifying the effect of the martensite hardness on void formation in a
dual phase steel. The void formation behaviour was followed by in-situ tensile tests in a SEM where the
structural evolution can be studied as a function of the given strain. These observations have revealed criteria
for void formation and effect of the hardness on strain distribution in the dual phase steel.

A dual phase steel (SVM) was applied to estimate the influence of tempering treatments on void formation in
the dual phase steel. In order to control the hardness of the martensite, tempering treatments at 300 to 600°C
were carried out because tempering does not affect the volume fraction, size and shape of ferrite and martensite.
In addition, since decohesion at the interface between the ferrite and large carbides and cracking of large
carbides lead to void formation, a short holding time was chosen in order to prevent carbides from coarsening.
Then some steel sheets were tempered at 300-600°C for 100s.
53
After tempering, the hardness of the martensite and ferrite was characterized through a Vickers hardness test
and the initial microstructure was characterized by SEM and TEM observations. Void formation behaviour in
the dual phase steel was followed by observation of the bulk specimens after tensile tests and in-situ tensile tests
in a SEM. In addition, strain distribution in the dual phase steel was measured by TEM and EBSD method after
tensile deformation to a strain of t=0.05.
5.3 Results
5.3.1 Influence of tempering treatments on mechanical properties

Tempering treatments strongly affect the mechanical properties of the dual phase steel (SVM). Figs. 5.1 and
5.2 show the nominal stress-nominal strain curves and the mechanical property of the dual phase steel. Fig. 5.1
does not show the stress-strain curve after necking because the extensometer was removed when the nominal
strain became larger than the uniform elongation. The total elongation is estimated based on measurement of the
gauge length after the tensile test. Before tempering, the stress increases rapidly after yielding and the work
hardening rate is high at a low strain. The 0.2% offset stress is defined as a yield stress (YS) as the steel does
not show a yield point phenomenon before tempering. On the other hand, as tempering treatments lead to a
yield point phenomenon, an upper yield stress is defined as a yield stress. To estimate the post-uniform
elongation (P-El) from the onset of necking to fracture, the post-uniform elongation was calculated as the
difference between the total elongation (T-El) and the uniform elongation; P-El = T-El – U-El. The ultimate
tensile strength was about 680MPa before tempering treatments. Fig. 5.2 shows that tempering treatments
increased all elongations, a reduction in thickness and hole-expansion ratio (), counterbalanced to reduction of
the UTS. A tempering treatment at 600°C led to a 30% reduction of the ultimate tensile strength in SVM.
However, the yield stress does not depend on tempering treatments due to a yield point phenomenon. Figs. 5.2c
and d also show that tempering treatments improve the P-El more than the U-El. It appears that tempering
treatments have a large effect on void formation in the dual phase steel as more voids are generally observed in
a necking zone of deformed specimens. Besides, the effect of the treatments on reduction in thickness and hole-
expansion ratio is also large as strains at the fracture of the specimens deformed in tension and at the edge of the
hole of the hole-expanded specimens are often larger than the U-El. Whereas the effect of the treatments on P-
El was relatively large, tempering treatments also affected the U-El and work hardening behaviour of the dual
phase steel. Fig. 5.3 shows true stress-true strain curves and the work hardening rate (d t /d t) of the dual phase
steel. Here the specimens without tempering, tempered at 300 and 500°C, were picked up as characteristic work
hardening behaviours appeared in the specimens. It general, a criterion for the onset of necking is given by the
following equation.
dt/dt <     Eq.[11]
where  t and  t are true stress and true strain. d t /d t is a work hardening rate at each strain. In general, the
work hardening rate is large after yielding, gradually decreases with strain until necking and decreases rapidly
after the start of necking due to a reduction of cross sections. However, the work hardening rate decreased
rapidly in the specimen without tempering before the onset of necking. The rapid reduction of the work
hardening rate might be related to reduction of the cross section due to void formation before the onset of the
necking. Therefore, the effect is investigated in Section 5.4.4 in more detail.
54
without tempering
1000 tempered at 300°C
temperedat 400°C

tempering at 600°C
600
400
200
0
0 0.05 0.1 0.15 0.2 0.25
Nominal strain
Fig. 5.1 Stress-strain curves of the dual phase steel before and after tempering treatments. Tempering treatments
led to a yield-point phenomena in the tempered specimens.
55
a) b)
600 700
Ultimate tensile stress / MPa

Yield stress / MPa
500 600
400 500
300 400
0 200 400 600 0 200 400 600
Tempering temperature / °C Tempering temperature / °C
c) d)
25 25
Post-uniform elongation / %
Uniform elongation / %
20 20
15 15
10 10
5 5
0 200 400 600 0 200 400 600
e) f)
0.8 80
60
Thickness reduction
0.6
/%
40
0.4 20
0
0 200 400 600
0.2
0 200 400 600
Tempering temperature / °C
Tempering temperature / °C
Fig. 5.2 Influence of tempering treatments on tensile properties in the dual phase steel (SVM) containing: a)
yield stress, b) ultimate tensile strength, c) uniform elongation, d) local elongation, d) reduction in thickness and
f ) hole-expansion ratio (). 0.2% flow stress was defined as tempering treatments lead to yield point
phenomenon.
56
1500
True stress, dt/dt / MPa
1000 without tempering
tempered at 300℃
500 tempered at 500℃
0
0 0.05 0.1 0.15
True strain, t
Fig. 5.3 Influence of tempering treatments on the work hardening rate (dt/d) of the dual phase steel (SVM).
In the specimen without tempering, rapid reduction of the work hardening rate is observed before onset of
necking (marked by arrow). In the other specimens the work hardening rate decreases gradually with strain.
5.3.2 Initial microstructures

Tempering treatments significantly improved the formability, counterbalanced by a reduction of ultimate
tensile strength as shown in Fig. 5.2. To clarify the effect of the tempering treatments on the tensile properties,
initial microstructures and deformed microstructures were studied in more detail. The initial microstructures
before deformation are shown in Fig. 5.4. In the figure, the rolling direction is a lateral direction. All
microstructures consisted of a ferrite matrix (F) and martensite particles (M). After tempering the martensite
phase looks brighten due to carbide precipitation. There was neither pearlite nor bainite in all specimens.
Tempering treatments did not affect the volume fraction, the grain size and the shape of the ferrite and
martensite. The volume fraction of martensite was 21% as reported in Table 5.1. The initial microstructures
were free from voids and cracks. The shape of the two phases was also different. Ferrite grains nucleated at
triple junctions and grain boundaries of austenite during cooling and have the shape of a polygon.
a) b) c) RD
F F
M M M
F
F
20m 20m 20m

Fig. 5.4 Initial microstructure of the dual phase steel (SVM) showing equiaxed ferrite grains (F) and martensite
particles (M) with concave and convex profiles (marked by arrows). The rolling direction (RD) is marked: a)
without tempering, b) tempered at 300°C and c) tempered at 500°C.
Table 5.1 Volume fraction and grain size of the dual phase steel (SVM).
Volume fraction Grain size (m) Aspect ratio of martensite
Ferrite Martensite Ferrite Martensite particles
0.79 0.21 7.3 3.5 2.1
57
On the other hand, tempering treatments affected the hardness of the martensite and ferrite as shown in Fig.
5.5. The data at 25°C are in the specimen before tempering treatments. A microhardness test shows that
tempering mainly softened the martensite, but the martensite was still harder than the ferrite (Fig. 5.5.a). Fig.
5.5b shows that the difference of the hardness between the martensite and ferrite depended on the tempering
temperature. It is obvious that softening of the martensite mainly leads to a reduction of ultimate tensile strength
of the specimen. According to Pickering (1978), tempering treatments decrease the solution hardening by
excess carbon and dislocation hardening induced by martensite transformation due to carbide precipitation and
recovery of dislocations. In the present study, cementite precipitation in the ferrite and martensite was observed.
Fig. 5.6 shows TEM micrographs of the microstructures of SLM before and after tempering treatments at 300
and 500°C. As shown in Figs. 5.6a-c, the martensite consisted of lathes of martensite and contained huge
amounts of dislocations. The ferrite also contains a number of dislocations induced by the martensite
transformation. The dislocations in the ferrite contribute to a low yield stress before tempering treatments. After
tempering treatments, the martensite and ferrite contain a number of precipitates as shown in Figs. 5.6e and f.
TEM observations have revealed that the precipitates in the martensite and ferrite were cementite. Fig. 5.7
exemplifies that cementite particles precipitated in the martensite after a tempering treatment at 300°C.
Cementite particles were located within lathes of martensite and at the lath boundaries. Cementite particles in
the martensite were smaller than in the ferrite as shown in Figs. 5.6e and f due to a large driving force for
cementite precipitation in the martensite, which is related to the carbon content in the martensite. As the
martensite phase had more excess carbon, the martensite was softened by tempering more than the ferrite.
a) b)
5 2.5
bulk
ΔHv (=HvM - HvF) / GPa
4 Ferrite
Microhardness / GPa
Martensite 2
3
1.5
2
1
1
0 0.5
0 200 400 600 0 200 400 600
Fig. 5.5 Effect of tempering treatments on microhardness of martensite and ferrite in the dual phase steel
(SVM): a) microhardness by the Vickers hardness tests and b) difference of the hardness between the martensite
phase and ferrite phase. The load was 10gf to measure the hardness of the martensite and ferrite phase.
In the ferrite grains more dislocations tended to be located near the martensite/ferrite interface compared
with the interior of the ferrite grains. Fig. 5.8 shows dislocations in the interior of a ferrite were less and more
dislocation were located near the interface as the dislocations were induced by the martensitic transformation.
Before tempering, the dislocation density near the martensite/ferrite interface (<500nm) was about
2.4×1014(1/m2) and that in the interior of the ferrite was about 1.3×1014(1/m2). A tempering treatment at 500°C
slightly reduced the dislocation density in the ferrite, which near the interface was about 9.3×1013(1/m2) and in
the interior of the ferrite was about 6.2×1013(1/m2). Calcagnotto et al. (2011) also suggested that more
dislocations exist near the interface through an EBSD observation. On the other hand, the tempering treatments
did not lead to recrystallization of the martensite due to the short holding time and the low temperature. The
tempering treatments mainly affected the carbon content and dislocation density in the ferrite and martensite.
Similar findings are reported by Morito et al. (1999) and Natori et al. (2005). They suggested that dislocations
induced by the martensite transformation are less effective on recrystallization compared with dislocations
induced by cold rolling. A cold-rolled low carbon steel is completely recrystallized, but lathes of martensite still
exist in a low carbon martensite after a tempering treatment at 1023K for 0.9ks even though the average
dislocation density induced by cold rolling is as many as that by the phase transformation.
58
As tempering treatments do not affect the volume fraction, size and shape of the martensite and ferrite, but
can control the hardness of the martensite, and thereby improve the formability, the steels are suitable for a
study of effect of the martensite hardness on the void formation behaviour in dual phase steels.
a) M c) e)
M
F
F M
b) d) f)
M
Fig. 5.6 TEM micrographs of SVM showing ferrite (F) and martensite (M) which contain huge amounts of
dislocations: a) without tempering, b) tempered at 500°C, c)-f) tempered at 300°C. d), e) and f) are
enlargements of c). e) and f) show that the martensite and ferrite contain carbides.
a) b)
c) d)
+
+
+ +
+ +
+ + 121 +
+ + +
+ + 110 +123 +
+ +
+ + 000 + +
+ + 002+
+ + + 301
+ +
+ + 211
+
+
hkl : martensite
+ hkl : cementite
Fig. 5.7 Cementite particles in the martensite tempered at 300°C: a) bright field image of tempered martensite,
b) dark field image of tempered martensite, c) diffraction pattern from lath martensite and d) diffraction patterns
 
form lath martensite and cementite ( 113 M //1.8330  0.9999  0.0768C ).
59
a) b)
c) d)
Fig. 5.8 Dislocations in a ferrite grain in SLM: a) dislocations in the interior of the ferrite grain and b)
dislocations near the martensite/ferrite interface before tempering treatments, and c) dislocations in the interior
of the ferrite grain and d) dislocations near the martensite/ferrite interface after tempering at 500°C.
5.3.3 Deformed microstructure after a tensile test

Figs. 5.9-11 demonstrate deformed specimens and void distribution in the specimens. Voids appear as black
spots, light grey particles are the martensite phase and dark grey grains are the ferrite phase. Fig. 5.9 shows that
the specimen without tempering treatments was broken after slight necking and the specimen contained a large
number of voids. Fig. 5.9d shows that the specimen contained some voids at a strain of t=0.111 and the voids
were short. As shown in Fig. 5.9b, voids were more frequent and some voids were elongated in the tensile
direction near the fracture (t=0.429). It is obvious that voids formed and grew with strain. In addition, void
coalescence was observed (Fig. 5.9b). Tempering treatments made the specimen deform more before fracture.
Fig. 5.10a demonstrates that the specimen tempered at 300°C was broken after necking. Voids were also more
frequent near the fracture (Fig. 5.10b) than at a strain of t=0.429 (Fig. 5.10c). At a strain of t=0.124, voids at
inclusions were observed. After tempering at 500°C, the specimen was necked heavily and broken due to void
formation (Fig. 5.11a). Fig. 5.11b shows that many voids existed near the fracture, but the size was smaller than
in the specimen without tempering and tempered at 300°C. At a smaller strain (t=0.153), voids existed at
inclusions (Fig. 5.11d). At a same strain (t=0.429), the number and size of voids depended on the tempering
temperature. In addition, the elongated martensite particles near the fracture show that the martensite particles
were deformed plastically.
60
a)
b) c) d)
20m 20m 20m

Fig. 5.9 Typical void distribution in the specimen without tempering treatments after a tensile test. a) a
deformed specimen, b) t=0.429, c) t=0.250 and d) t=0.111. The strain was calculated based of the thickness
reduction of the specimen.
a)
b) c) d)
20m 20m 20m

Fig. 5.10 Typical void distribution in the specimen tempered at 300°C after a tensile test. a) a deformed
specimen, b) t=1.000, c) t=0.429 and d) t=0.124. The strain was calculated based of the thickness reduction
of the specimen.
61
a)
b) c) d)
20m 20m 20m

Fig. 5.11 Typical void distribution in the specimen tempered at 500°C after a tensile test. a) a deformed
specimen, b) t=2.500, c) t=0.429and d) t=0.153. The strain was calculated based of the thickness reduction of
the specimen.
Void formation sites were investigated in all specimens. Fig. 5.12 demonstrates that regardless of tempering
treatments, voids formed at the three different sites; (i) in the martensite, (ii) in the ferrite and (iii) at inclusions.
However, further straining led to void formation in the martensite and ferrite contained in the tempered
specimens. For instance, at a strain of  t = 0.250 a void formed at the concave part of the martensite in the
specimen tempered at 300°C (Fig. 5.12d). In the specimen tempered at 500°C, voids formed in the heavily
deformed martensite particle at  t = 0.429. On the other hand, voids at inclusions were observed frequently at a
small strain of  t = 0.05. In particular, voids started to appear at inclusions in the ferrite earlier than in the
martensite.
62
Fig. 5.12 Typical voids in the dual phase steel (SVM) deformed in tension: voids a) in the martensite, b) in
ferrite and c) at an inclusion in the specimen without tempering ( t = 0.111), voids d) in the martensite ( t =
0.250), e) in ferrite ( t = 0.250), and f) at an inclusion ( t = 0.124) in the specimen tempered at 300°C
martensite ( t = 0.429), voids g) in the martensite ( t = 0.429), h) in ferrite ( t = 0.429) and i) at an inclusion( t
= 0.143) in the specimen tempered at 500°C.
Fracture surfaces also depended on the tempering temperature. Fig. 5.13 shows the fracture surfaces of the
deformed specimens without tempering, tempered at 300 and 500°C. The fracture surfaces consisted of two
types of dimples. The first one is an elongated dimple, which appears black due to depth. The second one is a
shallow dimple. It is found that the elongated dimples were related to voids in the martensite. Fig. 5.14
exemplifies that deep dimples are related to voids in the martensite and at the ferrite/martensite interface
because these voids grow with strain in the tensile direction as shown in Fig. 4.4 and 4.5. On the other hand, the
shallow dimples were related to secondary void formation and internal necking. The fracture surface in the
tempered specimen contained more elongated dimples and shallow dimples as the fracture strain of the
tempered specimens is much larger than the specimen without tempering. Plastic deformation made the
martensite particles and ferrite grains thin and voids formed in the heavily deformed particles/grains tiny.
63
a) c) e)
10µm 10µm 10µm
b) d) f)
10µm 10µm 10µm
Fig. 5.13 Fracture surface of SVM after tensile tests: a) and b) without tempering, c) and d) tempered at 300°C,
e) and f) tempered at 500°C. Lower row is enlargement of upper row.
10µm
Fig. 5.14 Deformed microstructure of SVM without tempering after the tensile test. Deep dimples in Fig. 5.13
are related to voids in the martensite and at the ferrite/martensite interfaces.
5.3.4 Void formation behaviour during in-situ testing

Tempering treatments significantly affected void formation in the dual phase steel. Fig. 5.15 demonstrates
the void formation behaviour induced by cracks in the martensite in the specimen tempered at 500°C.
Deformation to a strain of 0.428 and 0.844 does not generate cracks and voids in the martensite and ferrite (Figs.
5.15d-i), but the martensite particle was stretched heavily. After deformation to a strain of 1.104, voids formed
in the martensite particle due to cracking of the particle. It is obvious that voids formed in the martensite after
the martensite particle was deformed heavily compared with voids in the specimen without tempering (Figs. 4.4
and 5). This means that the tempering treatment increases the critical strain for crack formation in the martensite
and thereby retards void formation in the tempered dual phase steel. This is discussed in Section 5.4.3 in more
detail.
To evaluate the effect of the tempering treatments on void formation during in-situ tensile tests, (i) number
density, (ii) maximum length and (iii) area fraction of voids were measured. As an example, Fig. 5.16 shows
64
that tempering treatments affect void formation behaviour in SVM during in-situ observation. Here, the data in
the specimen without tempering, tempered at 300 and 500°C were picked up as the data clearly showed the
effect of tempering treatments. Here voids at inclusions are neglected as the void formation behaviour did not
depend on tempering treatments. In the specimen without tempering, voids form predominantly in the
martensite and already at a strain of t=0.05 whereas few voids form in the ferrite at a strain of 0.1. This agrees
with previous observations (e.g. Avramovic et al. 2009; Avramovic et al. 2010, Calcagnotto et al. 2011)
showing that cracks can form in martensite before necking of the specimen while voids appear in ferrite after
necking. The void number densities increase with strain before and after tempering treatments. Smaller voids
were more frequent at all strains and deformation increases the number of larger voids as shown in Fig. 5.17.
This indicates that voids form when local strain and stress overcome criteria for void formation and grow with
strain. This continuous void formation behaviour was observed in previous works. Landron et al. (2010)
observed void formation behaviour in a commercial dual phase steel by X-ray tomography. Avramovic et al.
(2010) estimated the number density and area fraction of voids as a function of a strain after tensile tests. The
criteria for crack formation in the martensite particles are discussed in Section 5.5.3 in more detail. Once voids
formed in the ferrite and martensite, the voids grew steadily with strain until the specimen was broken. Figs.
5.16c and d show the influence of tempering treatments on void length in the tensile direction. Voids in both the
martensite and ferrite grew with strain, but the maximum size of voids in the martensite was larger than in the
ferrite and voids in the martensite grew more rapidly than in the ferrite. Figs. 5.16e and f also show that
tempering treatments decrease the area fraction of voids due to the number density and size of voids and thereby
improve the formability. Obviously, voids in the martensite dominate the number density and area fraction.
65
a) b) c)
10m 5m
d) e) f)
10m 5m
g) h) i)
10m 5m
j) k) l)
void
void
10m 5m
Fig. 5.15 In-situ observations of void formation in SVM tempered at 500°C during a tensile test: a) and b) initial
microstructure without cracks or voids, d) and e) crack initiation in martensite (h= 0.428) g) and h) crack
propagation in martensite (h= 0.844), j) and k) void growth and crack propagation ( h = 1.104), c), f), i) and l)
sketches from b), e), h) and k) of the martensite particles and voids.
66
a) b)
5000 5000

4000 4000
3000 Without tempering Without tempering

3000
Tempered at 300°C Tempered at 300°C
2000 2000 at 500°C
Tempered Tempered at 500°C
1000 1000
0 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
Strain, h Strain, h
c) d)
25 25
Maximum length m
Maximum length / m
20 20
/
15 Without
15 tempering Without tempering
Tempered at 300°C Tempered at 300
°C
10 10 Tempered at 500
°C
Tempered at 500°C
5 5
0 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
Strain, h Strain,h
e) f)
2 2
1.5
Area fraction / %
1.5
Area fraction / %
Without tempering Without tempering

1 Tempered
1 at 300°C Tempered at 300°C
Tempered at 500°C Tempered at 500°C
0.5 0.5
0 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
Strain, h Strain, h
Fig. 5.16 Influence of tempering treatments on void formation behaviour in the dual phase steel during in-situ
tensile tests: number density of voids a) in the martensite and b) in the ferrite, maximum length of voids c) in
the martensite and d) in the ferrite, area fraction of voids e) in the martensite and f) in the ferrite.
67
a) b)
1500 1500

2
εh=0.159
1000 εh=0.159 1000
εh=0.227
εh=0.227
εh=0.357 εh=0.357
500 500
εh=0.623 εh=0.623
0 0
<1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 9-10 10-11 <1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 9-10 10-11
Length range of void / m void length/m
c) d)
1500 1500

εh=0.194 εh=0.194
1000 εh=0.340 1000 εh=0.340
εh=0.622 εh=0.622
500 εh=1.111 500 εh=1.111
0 0
<1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 9-10 10-11 <1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 9-10 10-11
Length range of void / m Length range of void / m
e) f)
1500 1500
εh=0.223 εh=0.223
1000 εh=0.392 1000 εh=0.392
εh=0.887 εh=0.887
500
500 εh=1.435 εh=1.435
0
0
<1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 9-10
<1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 9-10
Length range of void / m Length range of void / m
Fig. 5.17 Size distribution of voids formed in the dual phase steel during an in-situ tensile tests: a) in the
martensite without tempering, b) in the ferrite without tempering, c) in the martensite after tempering at 300°C,
d) in the ferrite after tempering at 300°C, e) in the martensite after tempering at 500°C, f) in the ferrite after
tempering at 500°C.
5.4 Discussion
In-situ tensile tests have revealed that, regardless of tempering treatments, most voids formed in the
martensite due to i) plastic deformation of martensite particles, ii) crack formation in the particles, iii) crack
propagation in the particles and iv) separation of the particle fragments. Voids formed at the ferrite/martensite
interface are minor. Regardless of tempering treatments, same mechanisms operate void formation in the dual
phase steel. The tests have also shown that softening of the martensite through tempering treatments retard void
(crack) formation in the dual phase steel and thereby improve the formability. As plastic deformation of the
martensite particles proceeded void formation in the martensite, plastic deformation of the martensite has an
important role on void formation in the martensite. In this section effect of softening of the martensite on
deformation behaviour of the martensite and ferrite, and then a criteria for void formation in the martensite are
discussed.
5.4.1 Strain partitioning between martensite and ferrite

Interaction between a matrix and reinforcements in composite structures affects deformation behaviour of
the matrix and reinforcements. In the present study deformation behaviour of each phase and the effect of
tempering treatments were investigated during in-situ tensile tests. Specifically, strains in the martensite phase
and ferrite phase were estimated based on the length change of each particle/grain in and the influence of
tempering treatments on the strains were discussed. Fig. 5.18 shows a distribution of average strain of
martensite particles and ferrite grains in the specimen without tempering at a strain of h=0.422. The strain in
the martensite phase is much smaller than the ferrite, but martensite particles were deformed plastically.
68
Martensite particles and ferrite grains have different strains even though the strain given is the same. It appears
that the wide distribution is related to the shape and distribution of the martensite particles as the ferrite matrix
was constrained by martensite particles in the dual phase steels. For example, strains in the ferrite grains
surrounded by the particles were relatively small and isolated ferrite grains were deformed more. The equiaxed
martensite particles were deformed less and the particles elongated in the tensile direction were deformed more.
The effect of martensite shape on strain in the martensite is discussed in Section 5.4.4. The strains were
averaged at each strain.
0.4
martensite
0.3 ferrite
Frequency
0.2
0.1
0
0-0.05
0.1-0.15
0.2-0.25
0.3-0.35
0.4-0.45
0.5-0.55
0.6-0.65
0.7-0.75
Strain of the martensite particles and ferrite grains
Fig. 5.18 Strain distribution in martensite particles and ferrite grains in the specimen without tempering at a
strain of h=0.422. Strains in the martensite particles and ferrite grains were measured as shown in Table 3.1c
and d and Fig. 3.8.
Fig. 5.19 shows that strains in each phase increases with strain and tempering treatments affected strains in
the martensite and ferrite. Softening of the martensite through tempering treatments assisted the martensite to
deform more and decreased the strain in the ferrite due to strain compatibility. It appears that softening of the
martensite led to early yielding and further straining of the martensite. Strain partitioning between the
martensite and ferrite was normalized as a function of the hardness difference between the martensite and ferrite.
For calculation, strains in the martensite and ferrite at each applied strain were divided by the applied strain and
normalized. Fig. 5.20 shows that the strain partitioning between the martensite and ferrite depends on the
hardness difference between the martensite and ferrite. In previous works the strain partitioning between the
martensite and ferrite in dual phase steels is related to the hardness of the martensite (e.g. Shen et al. 1986; Su
and Gurland 1987). This might affect void formation and the formability. However, an increase of the strain in
the martensite by softening of the martensite overcomes criteria at a smaller strain and thereby accelerates void
formation in the martensite. It contradicts to the result that tempering treatments retard void formation in the
martensite. Then, local strains at the concave part of the martensite, where cracks/voids formed, were
investigated in the following section.
69
2
Strain, M and F 1.5 martensite (WT)

ferrite (WT)
martensite (300°C)
1
ferrite (300°C)
martensite (500°C)
0.5 ferrite (500°C)
0
0 0.5 1 1.5 2
Local strain, h
Fig. 5.19 Strain partitioning between the martensite and ferrite during in-situ tensile tests. Softening of the
martensite through tempering treatments increases strain in the martensite, but deceases strain in the ferrite due
to strain compatibility.
1.5
1
M/h , F/h
martensite
ferrite
0.5
0
50 100 150 200 250
ΔHv(=HvM-HvF),10gf
Fig. 5.20 Influence of hardness difference between the martensite and ferrite on strain partitioning between the
martensite and ferrite. HvM: microhardness of the martensite, HvF: microhardness of the ferrite.
5.4.2 Strain localization

Plastic deformation of the martensite has an important role in void formation in the martensite as plastic
deformation occurs at the concave part of the martensite in advance of crack formation. Strain distribution
around martensite particles was estimated by EBSD measurements after tensile tests. To estimate the strain
distribution before void formation, the specimen was deformed to a strain of 0.05 in tension. Fig. 5.21 shows
the strain distribution in the dual phase steels before and after tensile tests. The contour maps show deviation of
Euler angles at each point from the average value in each ferrite grain. The martensite was black and neglected.
70
The yellow region shows that the deviation is small, the deviation in the blue region is larger and the deviation
in the green region is over 7°. As accumulated dislocations lead to a rotation of the crystal lattice, the large
deviation means that a large number of dislocations have accumulated. Figs. 5.21a-c show strain distributions in
the SVM before deformation Figs. 5.21d-f show strain distributions in the deformed specimens. Even though
dislocations induced by the martensitic transformation existed in the ferrite near the martensite/ferrite interface,
strains were distributed in the undeformed specimens more homogeneously than in the deformed specimens. To
estimate dislocation distribution around martensite particles, local misorientation changes at the
martensite/ferrite interfaces were measured in more detail. Figs. 5.22 and 23 show local strain distributions at
the concave part and the tip of a martensite particle in the specimen without tempering before deformation. Figs.
5.22a and 23a show orientation maps around a martensite particle and Figs. 5.22b and 23b show band contrast
maps (image quality maps) in the same area. In the band contrast map martensite particles appear black due to a
large dislocation density. Figs. 5.22c and 23c show accumulated misorientation profiles at the concave interface
(C-M/F) and the flat interface (F-M/F) between the martensite and ferrite, at the tip of the martensite (T-M/F)
and the ferrite/ferrite grain boundary (F/F), which were measured along the lines in Figs. 21a and 22a. Figs.
5.22c and 23c show that the misorientation changed more at all martensite/ferrite interfaces than at the
ferrite/ferrite grain boundaries. The misorientation changes were about 1-2.2° at 1µm away from the interface.
On the other hand, the changes were less than 0.5° at 1µm away from the boundary. This is because dislocations
in the ferrite were induced by the martensitic transformation during heat treatments. Tempering treatments
tended to reduce the misorientation change at the martensite/ferrite interface as well as TEM observations (Fig.
5.8). After tensile deformation to a strain of 0.05, strain localized at the concave part and near the tip of the
martensite particles as shown in Figs. 5.21d-f. It is obvious that the strain localization is related to void
formation as voids also form at the concave part and the tip of the martensite particles. This also shows that
tempering treatments retarded strain localization at the concave part and near the tip as the area fraction of the
green region was reduced in the tempered specimens. In addition, green regions were often aligned across some
ferrite grains and martensite particles in the 30-60° direction from the tensile direction. The strain distribution
was observed in more detail. Figs. 5.24 and 25 exemplify typical misorientation change at the interface between
the martensite and ferrite in the marked area in Fig. 21d. In the deformed specimen the misorientation increased
at the martensite/ferrite interface. In particular, the change at the concave interface (8°/1µm) was larger than the
flat interface of the martensite particle (2.7°/1µm). On the other hand, the misorientation changed less at the
ferrite/ferrite grain boundary (<1°/1µm). As dislocations lead to crystal orientation changes, it is obvious that a
large number of dislocations accumulated at the concave interface and the built-up stress might lead to local
plastic deformation at the concave part of the martensite. As shown in Figs. 4.4-8 and 5.14, slip band formation
around the concave part of the martensite particles and necking of the martensite particles also mean that
deformation localized at the concave part of the martensite. It is apparent that the strain localization at the
concave part of the martensite accelerated void formation in the martensite. This implies that the shape of the
martensite particles is one of the most important factors to retard void formation in dual phase steels. In addition,
the distribution of the martensite particles is also important because the heavily deformed areas (green regions)
aligned across some ferrite grains and martensite particles.
Strain also localized near the tip of the particles as shown Fig. 5.25. Here accumulated strain was measured
in the tensile direction. Fig. 5.25c shows that strain localized at the tip of the martensite particles elongated in
the tensile direction. The strain localized strongly at the interface (T-M/F1) and the accumulated misorientation
was 8°/1µm. On the other hand, the misorientation at another side (T-M/F2) was relatively small (2.3°/1µm) as
the distribution of the particles is also important for strain localization. Specifically, deformation localized areas
were often connected a tip of a martensite particle to another tip of a martensite or a concave part of a
martensite particle. Therefore, whereas the shape of the martensite particle is an important parameter to control
void formation in dual phase steels, the distribution of martensite particles is also important. .
A FEM analysis has also shown that deformation localizes at the martensite/ferrite interface, which faces
perpendicular to the tensile direction, in a dual phase steel as hard martensite particles constraint other ferrite
grains (e.g. Kadkhodapour et al. 2011).
71
a) b) c)
10µm 10µm 10µm

d) e) f)
10µm 10µm 10µm

Fig. 5.21 Strain distribution, which is represented by the deviation of Euler angle at each point from the average
value, in the ferrite at a strain of t=0.05: initial microstructures a) without tempering, b) tempered at 300 and c)
500°C, and deformed microstructure d) without tempering treatments, e) tempered at 300 and 500°C. In the
contour maps black regions are martensite and yellow, blue and green regions are ferrite. In the yellow regions
the deviation was 0°, in the blue region the deviation is larger and in the green regions the deviation is over 7°.
a) b) c)
F/F 5
Accumulated misorientation / °
C-M/F
4 F-M/F
F/F
3
F-M/F C-M/F 2
0
0 1 2 3 4 5
Length from interface / µm
Fig. 5.22 EBSD maps of the SVM without tempering treatments before deformation: a) orientation map and b)
band contrast map of a martensite particle containing concave parts and ferrite grains, and c) accumulated
misorientation profiles from the concave interface between the ferrite and martensite (C-M/F), the flat interface
between the ferrite and martensite (F- M/F) and the ferrite/ferrite grain boundary (F/F) to another interface. In
the orientation, the colours represent the crystallographic direction. In the band contrast map, dark contrast is
related to the martensite due to the large dislocation density. The tensile direction is horizontal. The step size is
50nm.
72
a)
F2
c)
5
F1
T-M/F1
4
T-M/F2
3
2
b)
1
0
0 1 2 3 4 5
Length from the interface / µm
Fig. 5.23 EBSD maps of the SVM without tempering treatments before deformation: a) orientation map and b)
band contrast map near the tip of a martensite particle, and c) accumulated misorientation profiles in a ferrite
grain 1(F1) and 2 (F2) from the left side to the right side. The tensile direction is horizontal. The step size is
50nm.
a) b) c) C-M/F
15
F/F F-M/F
F/F
10
C-M/F F-M/F
5
0
0 1 2 3 4 5 6
Fig. 5.24 EBSD maps of the SVM without tempering treatments after deformation to a strain of 0.05: a)
orientation map and b) band contrast map of a martensite particle containing concave parts and ferrite grains,
and c) accumulated misorientation profiles from the concave interface between the ferrite and martensite (C-
M/F), the flat interface between the ferrite and martensite (F- M/F) and the ferrite/ferrite grain boundary (F/F) to
another interface. In the orientation, the colours represent the crystallographic direction. In the band contrast
map, dark contrast is related to the martensite due to the large dislocation density. The tensile direction is
horizontal. The step size is 50nm.
73
a)
c)
T-M/F1
25
20
T-M/F2
15
10
b)
T-M/F1
5
T-M/F2
0
0 1 2 3 4 5 6
Length from the interface / µm
Fig. 5.25 EBSD maps of the SVM without tempering treatments after deformation to a strain of 0.05: a)
orientation map and b) band contrast map near the tip of a martensite particle, and c) accumulated
misorientation profiles in a ferrite grain 1(F1) and 2 (F2) from the left side to the right side. The tensile
direction is horizontal. The step size is 50nm.
Tempering treatments decreased the accumulated misorientation related to strain localization at the interface.
Fig. 5.26 and 27 show accumulated misorientation in the SVM tempered at 500°C after tensile testing. In the
tempered specimen strain localized at the martensite/ferrite interface, but the maximum misorientation was
relatively small. For example, the accumulated misorientation at the concave part of the martensite was
2.7°/1µm in Fig. 5.25c and the accumulated misorientation at the tip of the martensite was 3.5°/1µm in Fig.
5.27c. The effect of tempering treatments on the accumulated misorientation is shown in Fig. 5.28. Fig. 5.28a
shows that misorientation at the martensite/ferrite interface before deformation depends on the tempering
temperature, but the effect is very small. On the other hand, the misorientation after tensile tests strongly
depended on the temperature. In the specimen without tempering the misorientation at the concave interface
between the martensite and ferrite was about 8.5°/1µm and in the specimen tempered at 500°C the
misorientation was about 2°/1µm. As the accumulated misorientation is related to dislocations accumulated at
the interface, a reduction of the misorientation through tempering treatments retards void formation at the
concave part and the tip of the martensite. It appears that strain partitioning between the martensite and ferrite
attributes to strain localization. When the martensite is hard, strain in the martensite becomes small but strain in
the ferrite becomes large. The large strain in the ferrite contributes to a large strain in ferrite grains between two
martensite particles. The hard martensite allows locally the stress to overcome the yield stress of the martensite
due to stress concentration at the concave part of the martensite.
74
a) b) c)
15
C-M/F
F-M/F
C-M/F F-M/F
10 F/F
F/F 0
0 2 4 6 8 10
Fig. 5.26 EBSD maps of the dual phase steel temped at 500°C after deformation to a strain of 0.05: a)
orientation map and b) band contrast map of a martensite containing concave parts and ferrite grains, and c)
accumulated misorientation profiles from the concave interface between the ferrite and martensite (C-F/M), the
flat interface between the ferrite and martensite (F-F/M) and the ferrite/ferrite grain boundary (F/F) to another
interface. In the band contrast map, dark contrast is related to the martensite due to the large dislocation density.
The tensile direction is horizontal. The step size is 50nm.
a)
F2
c)
F1 15
T-M/F1
10 T-M/F2
b) 5
0
0 1 2 3 4
Fig. 5.27 EBSD maps of the dual phase steel temped at 500°C after deformation to a strain of 0.05: a)
orientation map and b) band contrast map near the tip of a martensite particle, and c) accumulated
misorientation profiles in ferrite grains 1(F1) and 2 (F2) from the left side to the right side. The tensile direction
is horizontal. The step size is 50nm.
75
a) b)
Accumulated misorientation / °/µm
15 15
Accumulated misorientation / °/µm

M/F C-M/F
F/F F-M/F
10 10 T-M/F
F/F
5 5
0 0
0 200 400 600 0 200 400 600
Fig. 5.28 Influence of tempering treatments on accumulated misorientation at the martensite/ferrite interfaces
and ferrite/ferrite grain boundaries: a) before tensile tests and b) after tensile tests. After tensile tests, strain
often localized at the concave interface between the martensite and ferrite and at the tip of the martensite
particle. As strain localized heterogeneously due to the distribution of martensite particles, martensite/ferrite
interfaces where strain localized were chosen. The data before tempering is shown at 25°C.
TEM observations have also revealed that deformation is localized at the concave part and the tip of the
martensite particle. Fig. 5.29 exemplifies that dislocation walls representing pile-up dislocations are observed at
the concave and the tip of the martensite particles in the SVM without tempering. Fig. 5.29a shows that
dislocation walls were observed in the ferrite near the concave part of the martensite particle and small cells
were located at the concave part of the martensite particle. The cell size became small near the concave part and
the smallest size was about 0.3 µm. Dislocation walls and cell structures were observed at the tip of a martensite
particle. In particular, a number of dislocation walls often crossed at the tip of martensite particles when the
distance between two martensite particles was small. It appears that pile-up dislocations might lead to plastic
deformation at the concave part of the martensite particles. Moreover, the minimum size of dislocation cells was
about 0.9µm after tempering at 500°C and relatively large. Fig. 5.30 exemplifies that cell-block structures were
observed in the SVM tempered at 500°C, but the size was relatively large. These TEM observations supplement
the EBSD observation showing that strain localizes at the concave part and tip of the martensite particles,
softening of the martensite through tempering treatments reduced localized strain and retarded void formation in
the martensite and ferrite.
a) b)
TD M TD
Fig. 5.29 TEM micrographs of dislocations in the SLM without tempering after deformation to a strain of 0.05:
a) at the concave part of a martensite and b) the tip of a martensite particle. Dislocation walls representing the
misorientation change and small cells are located at the concave part and the tip of martensite particles. The cell
size often became smaller at the martensite/ferrite interface.
76
Vickers hardness tests also reinforced that a small built-up stress retarded void formation in the tempered
dual phase steel (Fig. 5.31). The mechanical properties of martensite and ferrite have also been measured
directly by Vickers hardness tests. In these tests only the hardness of large ferrite grains and martensite particles
were measured because the indent size is larger than the small particles as demonstrated in Fig. 3.2. One source
of error may therefore be an effect of size on the hardness suggesting that a next step could be the introduction
of nano-indentation tests (e.g. Delincé et al. 2006). The hardness data for martensite and ferrite before and after
tempering are illustrated in Fig. 5.31 which shows (i) that the strength of the martensite decreases significantly,
thereby reducing the difference between the strength of the martensite and ferrite and (ii) that this difference is
maintained at all strains. This change in mechanical properties by tempering may significantly reduce the stress
which builds up at the martensite particles during straining thereby reducing the tendency for crack and void
formation in the martensite.
TD
Fig. 5.30 A TEM micrograph of dislocations in the SLM tempered at 500°C after deformation to a strain of 0.05.
Cell block structures were observed at the martensite/ferrite interface. The cell size in the specimen tempered at
500°C was much larger than in the specimen without tempering.
5
Microhardness / GPa
4
martensite (without tempering)
ferrite (without tempering)
3
martensite (tempered at 300°C)
ferrite (tempered at 300°C)
2
martensite (tempered at 500°C)
ferrite (tempered at 500°C)
1
0
0 0.5 1 1.5
Thickness strain, t
Fig. 5.31 Hardness of the martensite and ferrite in the dual phase steel before and after tempering treatments as
a function of the thickness strain.
77
5.4.3 Critical strain for void formation in martensite
Tempering treatments affected not only strain partitioning, but also a criterion for void formation
in the martensite as further straining of the martensite led to void formation in the tempered
martensite. Gladman et al. (1971) and Melander and Steninger (1982) have suggested that a critical
strain or stress is required for void formation. A strain of specimens when voids formed is often used
as a critical strain for void formation. However, since a matrix and reinforcements interact with each
other in composite structures, the hardness of the martensite affects strain partitioning between the
martensite and ferrite. On the other hand, a critical strain for void formation in the martensite might
depend on a property of the martensite. As the evaluation of the strain of the specimens contains two
effects, it is difficult to estimate individual effects of the metallurgical parameters on plastic
deformation and crack formation in the martensite. Therefore, in the present work, the local strain at
the necked region of the martensite was defined as a critical strain for void formation in the
martensite when a crack or a void formed in the martensite. The critical strains in about 20 martensite
particles were directly measured during in-situ tensile tests. Fig. 5.32 shows that the critical strain for
void formation in the martensite does not depend on the volume fraction of the martensite, but
depends on the hardness of the martensite. In addition, it is also found that the critical strain did not
depend on the shape, size and distribution of the martensite particles. It, therefore, appears that the
critical strain is used as an intrinsic parameter to describe void formation in dual phase steels.
In the specimen without tempering treatments, the critical strain was about 0.27 and the smaller criterion led
to early void formation in the martensite. On the other hand, a tempering treatment at 300°C improved it to 0.84
and the large critical strain contributed to the retardation of void formation and thereby led to the improvement
of the formability. The small criterion for void formation in the martensite leads to early void formation in the
martensite contained in the dual phase steel. Hadianfard (2009) has reported that cracks initiate in martensite
particles contained in a dual phase steel in low cycle fatigue tests, in which the applied stress is so large that the
martensite particles are deformed plastically. This result reinforces the idea that a martensite phase has a smaller
criterion for void formation than in a ferrite phase.
4
Critical strain for void formation
0
2 2.5 3 3.5 4 4.5
Martensite hardness / GPa
Fig. 5.32 Influence of the martensite hardness on the critical strain for crack formation in the martensite. The
local strain before cracking is observed is defined as a critical strain for crack formation in the martensite.
78
5.4.4 Void formation behaviour in the bulk
The influence of softening of the martensite on plastic deformation of the martensite and void formation in
the bulks was investigated. To estimate the effect of softening of the martensite on work hardening behaviour in
the dual phase steels, the Crussard-Jaoul analysis was carried out (e. g. Lian et al. 1991, Jiang et al. 1995). Fig.
5.33 shows that a logarithmic work hardening rate (ln/ln) depends on a logarithmic stress (ln) and the work
hardening behaviours are divided into two or three regions. The first region means that the martensite is
deformed elastically and the ferrite is deformed plastically. And slope changes of the work hardening rate are
related to the start of plastic deformation of the martensite. However, the transition from elastic to plastic
deformation of the martensite was not observed in the specimen tempered at 500°C due to a Luders deformation.
In addition, as shown in Section 4.4.1, a drop of the work hardening rate related to void formation was observed
in the SVM without tempering. However, the drop was not observed in the specimens tempered at 300 and
500°C due to late void formation. This means that softening of the martensite through tempering treatments
improves the uniform elongation as the softening keeps the work hardening rate high close to uniform
elongation. Softening of the martensite also affected the onset of plastic deformation of the martensite as shown
in Fig. 5.34. It is obvious that softening of the martensite leads to early plastic deformation of the martensite.
The early plastic deformation agrees with an increase of strain in the martensite (Fig. 5.19).
11 without tempering
tempered at 500°C
9
ln(d/d)
Transition strain (trans )

8
6
6 6.2 6.4 6.6 6.8
ln()
Fig. 5.33 An analysis of the work hardening behaviour in the SVM.
Finally, the void formation behaviour in bulk specimens after tensile tests was estimated based on the
classification in Section 4. Fig. 5.3.3 shows that the majority of voids formed in the martensite, voids in the
martensite were large and thereby the area fraction of voids in the martensite dominated. In the bulk specimens,
tempering treatments also retarded void formation in the bulk specimens, reduced void growth rate and thereby
reduced the area fraction of voids in the dual phase steel. The retardation of void formation by softening of the
martensite contributes to improvement of the formability in the dual phase steel.
79
0.04
strain(trans)
0.03
Transition strain
0.02
0.01
0.5 1 1.5 2 2.5
ΔHv(=HvM-HvF) / GPa
Fig. 5.34 Influence of the hardness of the martensite on the transition strain (trans). Softening of the martensite
leads to early plastic deformation of the martensite.
80
a) b)
3000 3000
without tempering without tempering
Number density / 1/mm 2
Number density / 1/mm 2

tempered at 300°C tempered at 300°C
2000 tempered at 500°C 2000
tempered at 500°C
1000 1000
0 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
Thickness strain, t Thickness strain, t
c) d)
40 40
without tempering
Maximu length / µm
Maximu length / µm
30 30 tempered at 300°C
tempered at 500°C
20 20
without tempering
10 tempered at 300°C 10
tempered at 500°C
0 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
e) f)
2 2
without tempering without tempering
1.5 tempered at 300°C 1.5 tempered at 300°C
Area fraction / %
Area fraction / %
tempered at 500°C tempered at 500°C

1 1
0.5 0.5
0 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
Fig. 5.35 Influence of tempering treatments on void formation in bulk specimens: number density of voids a) in
the martensite and b) in the ferrite, maximum length of voids c) in the martensite and d) in the ferrite, area
fraction of voids e) in the martensite and f) in the ferrite. In the bulk specimens tempering treatments retarded
void formation, reduced void growth rate and thereby reduced the area fraction of voids.
81
5.5 Summary
In a dual phase steel, the influence of softening of the martensite through tempering treatments at 300-
600°C on void formation in the martensite and ferrite has been investigated by tensile testing of bulk specimens
and by in-situ tensile tests in a scanning electron microscope. Based on the microstructural observations,
deformation and void formation behaviour have been related to the properties of the martensite and ferrite.
1. The tempering treatments significantly increase all elongations, counterbalanced by a reduction in strength.
In particular, the post-uniform elongation is significantly improved as more voids form in the necking zone. On
the other hand, the yield stress does not depend on the treatments due to a yield point effect.
2. The tempering treatments at from 300 to 600°C control the hardness of the martensite from 4 to 2.2GPa
without any effect on the grain size, shape and volume fraction of the martensite due to cementite precipitation
and recovery of dislocations in the martensite. On the other hand, the tempering treatments affect less the
hardness of the ferrite due to the small excess carbon.
3. Cracks and voids have been observed at different sites (i) in the martensite phase, (ii) in the ferrite phase and
(iii) at inclusions before and after tempering. At a given strain the majority of voids formed in the martensite,
voids in the martensite were larger than in the ferrite and the area fraction of voids in the martensite thereby
dominated.
4. Three factors to describe the void formation behaviour in dual phase steels have been suggested based on in-
situ tensile tests and EBSD measurements; (i) strain partitioning between the martensite and ferrite, (ii) strain
localization and (iii) a critical strain for void formation in the martensite.
5. In-situ tensile tests have revealed that in the dual phase steel, strain partitioned between the martensite and
ferrite. Softening of the martensite allows the martensite to deform more, but makes the ferrite deform less due
to strain compatibility. In addition, strain in the martensite significantly depends on the shape of the martensite
particles. A decrease of the strain in the ferrite contributes to reduce the void size in the martensite and ferrite as
plastic deformation of the ferrite leads to void growth in dual phase steels.
6. EBSD measurements have revealed that strain localizes at the concave part and tip of the martensite particles
in dual phase steels and strain localization leads to void formation in the martensite and ferrite in the dual phase
steel. Softening of the martensite retarded strain localization at the concave part and tip of the martensite
particles and thereby improves the formability.
7. Local strain at the necked region of the martensite particles when a void form is defined as a critical strain for
void formation in the martensite. The critical strain depends on the hardness of the martensite, but is
independent of the shape and size of the martensite particles. Softening of the martensite retards void formation
in the martensite due to improvement of the critical strain.
82
Chapter 6
Effect of volume fraction of martensite
6.1 Introduction
In this chapter, it is suggested that softening of the martensite through tempering treatments retards void
formation in dual phase steels and thereby improves the formability. However, as softening of the martensite
also leads to a reduction of the ultimate strength, a solution ensuring the strength should be suggested. In this
chapter, an increase of the volume fraction of the martensite is focused as a method to ensure the strength.
The volume fraction of the martensite is one of the important parameters to control the strength of dual phase
steels. Mazinani and Poole (2007) showed that an increase of the volume fraction ensures the tensile strength,
counterbalanced to a reduction of all elongations. As the ferrite contributes to uniform elongation, the reduction
of the volume fraction of the ferrite decreases the uniform elongation in dual phase steels. Whereas uniform
elongation is an important parameter, post-uniform elongation is also important to ensure the formability for
bending and hole-expansion. However, the effect of the martensite fraction on the void formation contributing
to the formability is not shown.
On the other hand, as the hardness of the martensite also affects the strength and formability of dual phase
steels. In particular, it is well-known that the post-uniform elongation, which is the elongation of the specimen
from the start of necking to fracture, relates strongly to void formation and depends on the hardness of the
martensite. It has been suggested that the hardness of the martensite affects void formation behaviour in dual
phase steels (e.g. Mazinani and Poole 2007). However, as the hardness of the martensite is controlled through
intercritical annealing, the volume fraction and the shape of the martensite are also affected by the intercritical
annealing. The volume fraction of the martensite might affect void formation behaviour in dual phase steels.
Therefore, the present study focuses on the effect of the volume fraction of the martensite in dual phase
steels. In the present work, the effect of the volume fraction of the martensite on void formation is investigated
by using two different dual phase steels. The volume fraction of the martensite is controlled by the chemical
composition and the hardness of the martensite is controlled through a hot-rolling condition and tempering
treatments at 300-600°C. The void formation behaviour and deformation behaviour of the martensite and ferrite
are directly followed during in-situ tests.

Two dual phase steels (SVM and LVM), which contain different volume fraction of martensite as listed in
Table 3.3, were used to estimate the influence of the tempering treatments on void formation in dual phase
steels. In order to control the hardness of the martensite, tempering treatments from 300 to 600°C were carried
out because tempering does not affect the volume fraction and shape of the ferrite and martensite. In addition,
since decohesion at the interface between the ferrite and large carbides and cracking of large carbides leads to
void formation, a short holding time was chosen in order to prevent carbides from coarsening. Then some steel
sheets were tempered at 300- 600°C for 100s.
6.3 Results
6.3.1 Mechanical properties

The volume fraction of the martensite affects mechanical properties of dual phase steels. Fig. 6.1 shows
tensile properties of the dual phase steels and hole-expansion ratio. In is obvious that an increase of the volume
fraction of the martensite increases yield stress and ultimate tensile strength, but reduced all elongations, a
reduction in thickness and hole-expansion ratio. It is well-known that uniform elongation depends on the
volume fraction of the ferrite as the ferrite ensures uniform elongation in dual phase steels.
83
a) 600 b) 900
SVM SVM
Ultimate tensile stress / MPa

800 LVM
LVM
Yield stress / MPa
500
700
600
400
500
300 400
0 200 400 600 0 200 400 600
c) d)
25 25
Post-uniform elongation / %
SVM SVM
Uniform elongation / %
20 LVM 20 LVM
15 15
10 10
5 5
0 200 400 600 0 200 400 600
e) f)
0.8 80
SVM SVM
Thickness reduction
0.6 LVM 60 LVM

/ %
40
0.4
20
0.2
0
0 0 200 400 600
0 200 400 600
Tempering temperature °C
/
Tempering temperature °C
/
Fig. 6.1 Influence of tempering treatments on tensile properties in the dual phase steels: a) yield stress, b)
ultimate tensile strength, c) uniform elongation, d) local elongation, d) reduction in thickness and f) hole-
expansion ratio (). 0.2% flow stress was defined as tempering treatments lead to yield point phenomenon.
On the other hand, it appears that void formation is related to the volume fraction of the martensite as post-
uniform elongation was affected by the volume fraction and more voids form in the necking zone. Whereas
softening of the martensite affects the post-uniform elongation more than the uniform elongation, increase of
the volume fraction of the martensite reduced the post-uniform elongation as much as uniform elongation. It
appears that the different behaviour of post-uniform elongation is related to a different effect on void formation
in the dual phase steels. Fig. 6.2 shows the nominal stress-nominal strain curves and the work hardening rate of
LVM. After tempering, Lüders deformation was observed. In the LVM, a drop of the work hardening rate was
84
observed. The influence of the volume fraction of the martensite on void formation and plastic deformation of
the martensite is discussed in Section 6.3.2 and 6.4.1 in more detail.
a) b)
1000
2000
without tempering
800
True stress, dt/dt / MPa


tempered at 500°C
600
1000
without tempering
400
tempered at 300°C
tempered at 400°C 500
tempered at 600°C
0
0
0 0.05 0.1 0.15
0 0.05 0.1 0.15 0.2 0.25
True strain, t
Nominal strain
Fig. 6.2 Influence of tempering treatments on a) nominal stress-nominal strain curves and b) the work hardening
rate (dt/d t) in LVM. In the specimen without tempering, a drop of the work hardening rate is observed
(marked by arrows).
6.3.2 Void formation behaviour

In LVM as well as SVM, voids formed (i) in the martensite, (ii) in the ferrite and (iii) at inclusions, as shown
in Fig. 6.3. Cracks often formed at the concave part of martensite particles and developed into voids in the dual
phase steel (Fig. 6.3a)). On the other hand, voids rarely formed near the tip of martensite particles (Fig. 6.3b). In
addition, few voids formed at inclusions due to decohesion and voids were often covered with inclusions (Fig.
6.3c).
a) b) c)
TD
M
void
M
M M M
void Inclusion
crack
5m 5m M 5m
Fig. 6.3 Typical voids in LVM deformed in tension: voids a) in the martensite, b) in ferrite and c) at inclusion in
the specimen without tempering ( t = 0.085).
85
a) b) TD
necking
5m 5m
c) d)
void
crack
crack
crack
5m 5m
Fig. 6.4 In-situ observation of void formation in LVM without tempering during a tensile test: a) initial
microstructure without cracks or voids, b) necking of a martensite particle and void formation in a martensite
particle (h= 0.112), c) necking, formation and growth of voids (h= 0.210), d) formation and growth of voids in
martensite particles ( h = 0.302).
Fig. 6.4 demonstrates the void formation behaviour in LVM without tempering treatments. In LVM, plastic
deformation of the martensite (Fig. 6.4c) and crack formation in the martensite (Fig. 6.4c) also led to void
formation in the martensite (Fig. 6.4d). Further straining made the voids grow in the tensile direction. Plastic
deformation often occurred at the concave part of the martensite particles. It is obvious that the same void
formation mechanism led to void formation in the dual phase steel with the large amount of the martensite.
However, more voids tended to form in the LVM at a smaller strain. In addition, multiple void formation; void
formation assists plastic deformation of the ferrite grains next to the broken martensite and further deformation
in the ferrite grains induces crack/void formation in other martensite particles, was also observed as well as in
Fig. 4.5.
After tempering treatments, plastic deformation and crack formation in the martensite led to void formation
in the martensite in LVM. Fig. 6.5 demonstrates the typical void formation behaviour in LVM tempered at
500°C. Here the martensite appears bright due to cementite precipitation. The initial microstructure is free from
cracks and voids in Fig. 6.5a. Fig. 6.5b shows that at a deformation to a strain of h =0.364 martensite particles
and ferrite grains were elongated in the tensile direction and thinned. In particular, the concave part of the
martensite was heavily deformed due to deformation localization. In addition, slip bands representing strain
localization were also observed at the concave part. After deformation to a strain of h=0.471, a crack initiated
at the concave part of the martensite particle and propagated in the martensite (Fig. 6.5c). It is apparent that
plastic deformation of the martensite had a key role for crack formation in the martensite. Fig. 6.5d shows that
further straining (h= 0.994) made the crack develop into a void and the void grew in the tensile direction. A
number of slip bands were located at the void and the outline of the void was zigzag. The zigzag profile is
evidence that dislocations contribute to void growth in dual phase steels at ambient temperature.
86
a) b) TD
2m 2m
c) d)
2m 2m
Fig. 6.5 Void formation behaviour in the LVM tempered at 500°C during in-situ testing: a) initial
microstructure without cracks or voids, b) plastic deformation of the martensite (h= 0.364), c) crack formation
and growth in the martensite (h= 0.471) and d) void growth in the tensile direction (h= 0.994). The martensite
appears bright due to cementite precipitation.
a) b) TD
2m 2m
c) d)
2m 2m
Fig. 6.6 In-situ observation of void formation in LVM without tempering during a tensile test: a) initial
microstructure without cracks or voids, b) void formation and slip-band formation near the end of a martensite
particles (h= 0.118), c) void growth in the ferrite (h= 0.210) and d) void growth in the ferrite (h= 0.302).
87
a) b) TD
5m 5m
c) d)
5m 5m
Fig. 6.7 In-situ observation of void formation in LVM tempered at 300°C during a tensile test: a) initial
microstructure without cracks or voids, b) void formation and slip-band formation near the end of a martensite
particles (h= 0.481), c) void growth in the ferrite (h= 0.642) and d) void growth in the ferrite (h= 0.880).
In addition, void formation in the ferrite was rarely observed in LVM. Figs. 6.6 and 7 demonstrate void
formation behaviour at a martensite/ferrite interface on the ferrite side in LVM without tempering and tempered
at 300°C. During in-situ tensile tests, voids initiated at the martensite/ferrite interface and grew in the tensile
direction with strain. However, voids did not reach other martensite particles as well as in SVM. In addition,
voids formed at the triple junction of a martensite and two ferrite grains (Fig. 6.6) and the martensite/ferrite
interface, where a ferrite/ferrite grain boundary did not touch. Kadkhodapour et al. (2011) suggested that
decohesion between the ferrite/ferrite grain boundaries leads to void formation in a dual phase steel based on a
structural evolution and a FEM analysis. However, the in-situ tests have revealed that decohesion at the
ferrite/ferrite grain boundary was not dominant in the dual phase steels. Fig. 6.8 shows quantitative analyses of
void growth in the ferrite. Here length change between two martensite particles, of the ferrite between the
martensite particles and voids were followed. As mentioned in Section 4.4.5, void formation and growth did not
shorten the ferrite between two martensite particles and the void, thereby, did not lead to void formation
between two martensite particles. These observations also supplement the idea that most voids form in the
martensite in dual phase steels.
88
a) b)
4 12
M-M M-M
10
3 ferrite ferrite
8
Length / mm
Length / m
2 void length 6
void length
4
1
2
0 0
0 0.1 0.2 0.3 0.4 0 0.2 0.4 0.6 0.8 1
Local strain, h Local strain, h
Fig. 6.8 Void growth behaviour in the ferrite: a) in LVM without tempering treatments (in Fig. 6.6) and b) in
LVM tempered at 500°C (in Fig. 6.7). Voids formed at the martensite/ferrite interface did not reach another
martensite particle.
In LVM, tempering treatments also affected the void formation and growth. Fig. 6.9 exemplifies the effect of
tempering treatments on void formation in LVM. As shown in Figs. 6.9a and b, tempering treatments retarded
void formation in the martensite and ferrite and most voids formed in the martensite, regardless of tempering
treatments. Figs. 6.9c and d show that tempering treatments also decreased void growth rates in the martensite
and ferrite. The length in the martensite was much larger than in the ferrite. Therefore, tempering treatments
reduced the area fraction of voids in the martensite and ferrite (Figs. 6.9e and f). As shown in Fig. 5.16,
tempering treatments also had a similar effect on void formation in SVM. It is obvious that the volume fraction
and hardness of the martensite affects void formation and growth in the dual phase steels.
To clarify the effect of the martensite volume fraction on the void formation behaviour in the martensite and
ferrite, (i) number density, (ii) maximum length, (iii) average length and (iv) area fraction were directly
measured during in-situ tensile tests. Fig. 6.10 shows typical void formation behaviour in LVM without
tempering. Voids frequently formed in the martensite and an increase of the volume fraction of the martensite
accelerated void formation in the martensite (Fig. 6.10a), made voids in the martensite and ferrite larger (Figs.
6.10c and d) and area fraction of voids in the martensite thereby dominates (Fig. 6.10c). It appears that the
maximum length of voids is related to the void growth rate in the martensite and ferrite. After tempering
treatments, an increase of the volume fraction of the martensite made more voids form in the martensite and the
length of voids larger.
The large increase of void number density in the martensite implies that not only void initiation site density,
but also other parameters affected void formation in the martensite. For instance, the volume fraction in LVM
(=0.38) is about twice as large as in SVM (=0.21) but the number density of voids in the martensite in LVM is
about four times as large as in SVM. In addition, it is obvious that the void growth rate represented to the slope
of the maximum length change is affected by the volume fraction of the martensite. These points are discussed
in more detail in the following section.
89
a) Without tempering b)
5000 5000
Tempered at 300
°C

2
Tempered at 500
°C Without tempering
2
4000 4000
Tempered at 300
°C
°C
2000 2000
1000 1000
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
Local strain,h Local strain,h
c) d)
Without tempering
25 25
Tempered at 300
°C
Maximum length / mm
Maximum length / mm
20 Tempered at 500
°C 20 Without tempering
Tempered at 300
°C
°C
10 10
5 5
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
e) f)
Without tempering
2 2
Tempered at 300
°C
Tempered at 500
°C Without tempering
1.5 1.5
Area fraction / %
Area fraction / %
Tempered at 300
°C
Tempered at 500
°C
1 1
0.5 0.5
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
Fig. 6.9 Effect of tempering temperature on void formation in LVM: void number density a) in the martensite
and b) in the ferrite, maximum length of voids c) in the martensite and d) in the ferrite, area fraction of voids e)
in the martensite and f) in the ferrite.
90
a) b)
4000 1.5
Number density / mm
Area fraction / %
2
3000
1 (HM)
in m a rtensite in martensite (S
in ferrite (HM) in ferrite (SVM
2000
in m a rtensite (Ba se) in martensite (H
0.5
in ferrite (Ba se)
1000 in ferrite (HM)
0 0
0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8
Local strain,h
c) Local strain,h
20
Maximum length /m
15
in martensite (LVM)
in ferrite (LVM)
10
in martensite (SVM)
5 in ferrite (SVM)
0
0 0.2 0.4 0.6 0.8
Local strain,h
Fig. 6.10 Effect of volume fraction of the martensite on void formation behaviour in the dual phase steels
without tempering treatments: a) number density of voids in the martensite and ferrite, b) area fraction of voids
in the martensite and ferrite, c) maximum length and d) average length of voids in the martensite and ferrite.
6.4 Discussion
6.4.1 Critical strain for void formation in martensite

Gladman et al. (1971) and Melander et al. (1982) suggested that a critical strain or stress is required for void
formation. In this study, the local strain at the necked region of the martensite was defined as a critical strain for
void formation in the martensite when a crack or a void formed in the martensite. The critical strains in about 20
martensite particles were directly measured during in-situ tensile tests. Fig. 6.11 shows that the critical strain for
void formation in the martensite does not depend on the volume fraction of the martensite, but depends on the
properties of the martensite. In addition, it is also found that the critical strain did not depend on the shape, size
and distribution of the martensite particles. It, therefore, appears that the critical strain is used as an intrinsic
parameter to describe void formation in dual phase steels.
91
4
Critical strain for crack formation

SVM
3
LVM
0
200 250 300 350 400 450
Martensite hardnes, Hv 10gf
Fig. 6.11 Critical strain for void formation in the martensite. The critical strain does not depend on the volume
fraction of the martensite, but depends on the hardness of the martensite.
6.4.2 Strain partitioning between martensite and ferrite

Plastic deformation of the martensite and ferrite depended on the volume fraction of the martensite. Fig. 6.12
shows the work hardening behaviour of LVM and influence of the volume fraction on the transition strain
related to the onset of plastic deformation of the martensite. Fig. 6.12a demonstrates the work hardening
behaviour of LVM is also divided into two or three regions based on the modified Crussard-Jaoul analysis (e.g.
Lian et al. 1991, Jiang et al. 1995). In all specimens the martensite was deformed elastically and a small
deformation makes the martensite deform plastically. Fig. 6.12b demonstrates that the volume fraction of the
martensite affected the onset of plastic deformation of the martensite. In the dual phase steel containing a large
amount of the martensite plastic deformation started at less than trans=0.02, which is a strain of tensile
specimens. In addition, the specimens without tempering and tempered at 300°C show a rapid drop of the work
hardening rate due to void formation. It appears that void formation affected the work hardening rate at a large
strain more due to large area faction of voids.
a) b)
11 0.04
LVM-WT VM=0.21
(trans)
10 VM=0.38
LVM-300°C
strain
0.03
Transitionstrain
9 LVM-500°C
ln(d/d)
Transition
8
0.02
7
6 0.01
6.2 6.4 6.6 6.8 7 0.5 1 1.5 2 2.5
ln() ΔHv(=HvM-HvF) / GPa
Fig. 6.12 Influence of volume fraction of the martensite on plastic deformation of the martensite; a) Work
hardening behaviour of the LVM a) and b) influence of volume fraction of the martensite on the transition strain
(trans) related to the onset of plastic deformation of the martensite in the dual phase steels.
Strains in the martensite and ferrite were measured at each strain based on the length changes of about 200
grains/particles and the averaged values were normalized by applied local strains (h) of the specimens. Fig.
92
6.13 shows that the strain in the martensite was relatively small and an increase of the volume fraction of the
martensite and ferrite deformed more. It appears that an increase of the hard martensite led to further
deformation of the ferrite due to strain compatibility, further deformed ferrite contributed to straining of the
martensite due to the large work hardening in the ferrite. Therefore, the large volume fraction increases the
strain in the martensite, allows the strain to overcome the critical strain shown in Fig. 6.11 at smaller applied
strains of the specimens and thereby leads to early void formation. The previous works supplement that strain
partitioning between the martensite and ferrite depends on the volume fraction of the martensite. Tasan et al.
(2010) showed that during an in-situ tensile test the strain in the martensite is much smaller than in the ferrite
contained in a dual phase steel, where the martensite fraction is less than the SVM. Ghadbeigi et al. (2010)
pointed out that the martensite is deformed as much as the ferrite in a dual phase steel containing 50 vol.%
martensite during an in-situ tensile test. The tendencies in their works are the same as in the present study even
though the hardness of the martensite and ferrite is not shown in their papers. Besides, it is also found that in
bulk specimens, the volume fraction of the martensite affected strain partitioning between the martensite and
ferrite, which was estimated based on the thickness reduction of the martensite particles and ferrite grains.
In addition, strain partitioning between the martensite and ferrite contributes to void growth in the martensite
and ferrite. As shown in Fig. 6.13, the strain in the ferrite was also affected by the volume fraction of the
martensite. As plastic deformation of the ferrite matrix leads to separation of the broken martensite particles and
void growth in the ferrite, the increase of the strain in the ferrite assists void growth in the martensite and ferrite.
Moreover, deformation localization at voids accelerates void growth in dual phase steels as void formation leads
to local softening. Therefore, the void growth rate in the LVM is larger than in the SVM.
Finally, the deformation behaviour of each grain/particle also depended on the shape and distribution of the
martensite particles. Specifically, ferrite grains surrounded by martensite particles were deformed less, whereas
isolated ferrite grains, which did not touch martensite particles, were deformed more. Strains in equiaxed
martensite particles were smaller than in elongated particles and particles with concave parts. It, therefore,
appear that the shape of the martensite particles is one of the most important parameters to affect the strain
partitioning between the martensite and ferrite.
This result also implies that the effect of the martensite hardness, which is controlled by intercritical
annealing, on void formation is underestimated. The increase of volume fraction of the martensite makes the
martensite soft due to a reduction of the carbon content in the martensite, and thereby increases the critical
strain for void formation in the martensite. On the other hand, the increase contributes to further straining of the
martensite leading to early void formation. Therefore, the volume fraction of the martensite should be constant
in order to estimate accurately the effect of the martensite hardness on void formation in dual phase steels. A
tempering treatment is a better method to estimate the effect of the martensite hardness on void formation in
dual phase steels.
The difference between the tempering treatments, which affect the critical strain and strain partitioning, and
volume fraction of the martensite, which affects only strain partitioning, leads to different effects on void
formation in dual phase steels and strength of specimens. The combination of softening of the martensite and
increasing of the martensite fraction enables one to develop advanced dual phase steels with high strength and
better formability.
93
1.5
M(without tempering)
M(tempered at 300°C)
1 M(tempered at 500°C)
M/h , F/h
M(Tasan et al.)
M(Ghadbeigi et al.)
0.5 F(without tempering)
F(tempered at 300°C)
F(tempered at 500°C)
0 F(Tasan et al.)
0 0.2 0.4 0.6 0.8 F(Ghadbeigi et al.)
Volume fraction of Martensite
Fig. 6.13 Influence of the volume fraction of the martensite on strain in the martensite and ferrite. The strain
partitioning between the martensite and ferrite depends on the volume fraction of the martensite. The hardness
of the martensite also affects the strain partitioning. Strain partitioning behaviour in other dual steels, which
were carried out by Tasan et al. (2010) and Ghadbeigi et al. (2010), is also compared.
94
6.5 Summary
The influence of the volume fraction of the martensite on void formation has been investigated by tensile
testing of bulk specimens and by in-situ tensile tests in a scanning electron microscope. The hardness of the
martensite was controlled through tempering treatments at 300-600°C. Based on the microstructural
observations, influence of the volume fraction and hardness on void formation and growth has been discussed.
The following conclusions are reached.
1. Increasing of the martensite fraction from 21 to 38 vol.% increases the yield stress and ultimate
tensile strength, and is counterbalanced by a reduction of all elongations. However, the tempering
treatments did not affect the yield stress due to a yield point phenomenon.
2. Regardless of volume fraction of the martensite, a majority of voids formed in the martensite and
the area fraction of voids in the martensite dominated in the dual phase steels. Voids rarely formed at
the martensite/ferrite interface, but did not lead to void formation between two martensite particles.
3. The increase of the volume fraction of the martensite significantly accelerated void formation in
the martensite, enlarged voids in the martensite and ferrite in dual phase steels and thereby increased
the area fraction of voids in the martensite.
4. A critical strain for void formation in the martensite depends on the properties of the martensite,
but is independent of the volume fraction, shape, size and distribution of the martensite. The strain
partitioning between the martensite and ferrite depends on the volume fraction, hardness and shape of
the martensite.
5. The increase of the volume fraction of the martensite makes the martensite and ferrite deform more and
further straining of the martensite contributes to early void formation in dual phase steels. Further straining of
the ferrite increases the void growth rate in the ferrite and martensite. Therefore, an increase of the volume
fraction of the martensite has a negative effect on the formability.
95
Chapter 7
Effect of the shape of martensite particles on deformation
7.1 Introduction
In the last chapter, it is suggested that the morphology of the hard particle is a key parameter for void
formation in dual phase steels and equiaxed martensite particles are preferable to retard void formation and
thereby, improve formability for tensile tests, hole-expansion tests and bending tests. The morphology of the
martensite particles also affects the void formation behaviour in dual phase steels. As mentioned above, voids
predominantly formed at the concave part of the martensite particles and in the elongated particles after plastic
deformation of the particles. This obviously shows that martensite shape is one of the most important
parameters for void formation in dual phase steels. Then, the shape of the martensite particles was characterised
and deformation behaviour of each particle was followed during in-situ loading.

SVM was used to estimate the influence of particle morphology on the deformation behaviour of martensite
particles in the dual phase steel. In order to estimate the straining behaviour of equiaxed particles and elongated
particles in the dual phase steel, length change of the particles were directly followed during in-situ loading and
the strain in each particle was estimated as a function of applied strain on specimens. As mentioned the chapter
3, the morphology of martensite particles in the dual phase steel was complicated. Fig. 7.1 exemplifies typical
morphologies of martensite particles in the dual phase steel and are categorized as; a) equiaxed particles, b)
elongated particles in the tensile direction, c) elongated particles perpendicular to the tensile direction and d)
less regular particles. The martensite particles whose aspect ratio is from 0.5 to 2.0 were defined as equiaxed
particles (Fig. 7.1a). If the aspect ratio is over 2.0 and aligned in the tensile direction, the particles are defined as
elongated particles in the tensile direction (Fig. 7.1b). The population was dominant in the dual phase steel
because the sheet was deformed in the rolling direction during the hot-rolling process. When the aspect ratio of
the particles is less than 0.5 in the tensile direction, the particles are called elongated particles perpendicular to
the tensile direction (Fig. 7.1c). On the other hand, some particles have branches and large concave parts. They
are defined as a less regular shape (Fig. 7.1d). To estimate the deformation behaviour of martensite particles
with different shapes, SVM tempered at 500°C was chosen to follow the strain at larger given strains as the
critical strain for void formation in the martensite is large and it is easy to follow the length change. The
straining behaviour of each particle was estimated as a function of given strains during in-situ loading.
7.3 Results and Discussion

The typical deformation behaviour of 4 types of martensite particles is shown in Figs. 7.2 -5. Deformation
behaviour of each martensite particle strongly depends on the shape. Fig. 7.2 shows the deformation behaviour
of an equiaxed particle during the in-situ tensile test. After deformation to strains of h=0.33 and 0.887, the
shape was almost identical to the initial shape and the equiaxed particle was deformed less. On the other hand,
the shape of elongated particles depended on the applied strain. Fig. 7.3 shows that an elongated particle was
stretched in the tensile direction and became thinner with strain. However, when the particle was aligned
perpendicular to the tensile direction, the particle was not deformed much (Fig. 7.4). Some elongated particles
were bent and rotated by plastic deformation of the surrounding ferrite grains. A less regular shape of the
martensite led to more complicated deformation as shown in Fig. 7.5. The thinner part aligned in the tensile
direction was deformed more than the other parts perpendicular to the tensile direction. Even though not only
the shape but also the distribution affects the strain distribution in dual phase steels as shown in Table 4.1, strain
of the elongated and less regular shaped martensite particles tend to be larger than of the equiaxed particles (Fig.
96
7.6). Therefore, the strain of the particles elongated in the tensile direction overcomes a critical strain for void
formation in the martensite at a smaller strain and voids form in the elongated particles at a smaller strain than
in the equiaxed particles even though a dual phase steel contains equiaxed particles and elongated particles. The
results also imply that spheroidizing of the particles retards void formation in the martensite and thereby
improves the formability in dual phase steels as it appears that the critical strain for void formation in the
martensite does not depend on the martensite shape. As shown in Figs. 4.4 and 4.5, rapid void formation in
elongated particles shows that martensite morphology is a key parameter for void formation in dual phase steels.
It seems that martensite morphology is not important only for the formability, but also for strength of dual phase
steels due to strain/stress partitioning between hard particles and soft matrix related work hardening behaviour
of composite materials. He et al. (1984) and Sun et al. (2002) pointed out that flow stress in a dual phase steel
containing elongated martensite particles is larger than in a dual phase steel with fine spherical particles, but
voids form at a smaller strain in the steel with elongated particles and the elongation is small. In metal matrix
composites, mechanical properties and void formation behaviour depends on shape of reinforcements. Clyne
and Withers (1993) pointed out that elliptical particles often contain cracks and decohesion tends to occur
between metals and spherical particles. Mishnaevsky Jr (2007) suggested that flow stress in the composites with
elliptical particles is larger than in the composite containing spherical particles and the elliptical particles are
broken at a smaller strain through a FEM analysis.
a) equiaxed b) Elongated along the tensile direction
-Aspect ration is larger than 2.

-Aspect ration is from 0.5 to 2. -The particles are elongated in the tensile direction.
-Most elongated particles have concave parts.
c) Elongated perpendicular to d) Less regular

the tensile direction
-Aspect ration is smaller than 0.5. -The particles branch.

-The particles are elongated perpendicular to -Most particles have concave parts.
the tensile direction.
Fig. 7.1 Morphologies of martensite particles in the SVM: a) equiaxed particles, b) elongated particles in the
tensile direction, c) elongated particles perpendicular to the tensile direction and d) less regular particles.
97
a) b) c)
10µm 10µm 10µm

h =0 h =0.330 h =0.887
Fig. 7.2 Deformation behaviour of equiaxed martensite particles: a) before deformation, b) h=0.330, c)
h=0.887 and d) schematic illustration of martensite shape at each strain. The tensile direction is horizontal.
a) b) c)
5µm 5µm 5µm

h=0 h=0.330 h=0.887
Fig. 7.3 Deformation behaviour of elongated particles in the tensile direction: a) before deformation, b)
eh=0.330, c) eh=0.887 and d) schematic illustration of martensite shape at each strain. The tensile direction is
horizontal.
a) b) c)
5µm 5µm 5µm

h =0 h =0.330 h =0.887
Fig. 7.4 Deformation behaviour of elongated particles perpendicular to the tensile direction: a) before
deformation, b) eh=0.330, c) eh=0.887 and d) schematic illustration of martensite shape at each strain. The
tensile direction is horizontal.
a) b) c)
10µm 10µm 10µm

h =0 h =0.330 h =0.887
Fig. 7.5 Deformation behaviour of elongated particles perpendicular to the tensile direction: a) before
deformation, b) eh=0.428, c) eh=0.844 and d) schematic illustration of martensite shape at each strain. The
tensile direction is horizontal.
98
1.2
1
equiaxed
0.8
Strain, M
elongated in the tensile direction

0.6
elongated perpendicular to the tensile direction
0.4
less regular
0.2
0
0 0.2 0.4 0.6 0.8 1
Strain, h
Fig. 7.6 Deformation behaviour of martensite particles with different shapes.
In addition, the complicated 3D-structure in a dual phase steel makes morphological control more important.
Landron et al. (2010) and Adachi et al. (2012) showed that martensite particles in dual phase steels were
elongated in the rolling direction, often connected to other particles and had many concave parts through X-ray
holotomography and serial sectioning even though martensite particles were equiaxed in 2D-images. In
particular, band-like structures are often observed in higher strength dual phase steels due to Mn segregation
zones (Avramovic-Cingara et al. 2010). However, since 3D-structures of dual phase steels are still ambiguous,
estimation of 3D-structures in commercial dual phase steels, non-destructive observation of void formation and
deformation behaviour in bulk samples and analysis of the deformation behaviour of the constituents in dual
phase steels are future works.
7.4 Summary
The influence of the morphology of the martensite particles on deformation has been investigated by in-
situ tensile tests in a scanning electron microscope. The morphology of martensite particles was categorized as
a) equiaxed particles, b) elongated particles in the tensile direction, c) elongated particles perpendicular to the
tensile direction and d) less regular particles. The straining behaviour of each particle was estimated as a
function of given strains. Based on the microstructural observation, the influence of the martensite morphology
on void formation has been discussed. The following conclusions are reached.
1. The strain partitioning between the martensite and ferrite strongly depends on the shape of the martensite
particles. Martensite particles elongated in the tensile direction are deformed more than the equiaxed particles
and particles aligned perpendicular to the tensile direction. Equiaxed particles are less deformable and voids
rarely form in these particles. On the other hand, the critical strain for void formation in martensite depends on
the hardness.
2. The strain partitioning behaviour in each particle and the intrinsic critical strain for void formation lead to
rare void formation in equiaxed martensite particles even though martensite particles have various kinds of
morphology.
99
Chapter 8
Conclusions
The purpose of the present work is to classify void formation mechanisms, characterise which metallurgical
factors are important to describe void formation and reveal the effect of metallurgical parameters on void
formation in dual phase steels. The deformation behaviour of the constituents and void formation behaviour in
the matrix and reinforcements were followed by tensile testing of bulk specimens and by in-situ tensile tests in a
scanning electron microscope.
At first, the structural evolution has revealed that there are three typical void formation sites; (a) in the
martensite phase, (b) in the ferrite phase and (c) at inclusions, as shown in Fig. 8.1. At a given strain both the
frequency and the area fraction of voids in martensite dominate. The void formation in the martensite evolves in
four steps; (i) plastic deformation of martensite, (ii) crack initiation at the martensite/ferrite interface, (iii) crack
propagation leading to fracture of martensite particles and (iv) void formation by separation of particle
fragments (Fig. 8.1a). The in-situ observations have allowed void nucleation to be related directly to plastic
deformation and cracking of martensite particles. The voids in ferrite are predominantly formed near the end of
the martensite particles and were proceeded by voids in the martensite (Fig. 8.1b). Void formation in the ferrite
does not lead to void formation between two martensite particles. Decohesion of the martensite/ferrite interfaces
and ferrite/ferrite grain boundaries is not observed. Void formation at inclusions is minor due to the high purity,
but voids formed at the smallest strain at the metal/inclusion interface due to a weakly bonded interface (Fig.
8.1c). This characteristic void formation has revealed that void formation in the martensite dominates in bulk
samples. Whereas void formation affects post-uniform elongation more as voids form with strain, early void
formation decreases the work hardening rate of the dual phase steels and uniform elongation.
a) void in martensite
ferrite
martensite necking crack void

Plastic deformation Crack initiation & propagation Martensite separation
b) void in ferrite
ferrite

Void initiation Void growth
c) void at an inclusion
void void void void

Decohesion Crack propagation Void growth
Fig. 8.1 Void formation behaviour in dual phase steels; a) in the martensite, b) in the ferrite and c) at an
inclusion.
Second, three factors related to void formation are suggested based on structural evolutions by in-situ testing
and EBSD observations; (i) strain partitioning between the martensite and ferrite, (ii) strain localization and (iii)
a critical strain for void formation in the martensite. During deformation, the strain often localizes at the
concave part and tip of the martensite particles and voids form in the martensite when a localized strain
100
overcomes a critical value for void formation in the martensite. The critical strain for void formation in the
martensite is an intrinsic parameter to describe void formation in the martensite and depends on the properties
of the martensite, but is independent of the volume fraction, morphology and size of the martensite particles. As
plastic deformation of the martensite leads to crack formation in the martensite, strain in the martensite is
important. In particular, strain partitions between the martensite and ferrite due to the different mechanical
properties of the martensite and ferrite. The strain partitioning between the martensite and ferrite was estimated
as a function of the hardness and volume fraction of the martensite. Whereas strain partitioning is an important
parameter, strain localization related to the shape and distribution of the martensite particles is another
important parameter. Strain localizes at the concave part of the martensite and necking of the martensite
particles leads to void formation in the martensite during tensile tests. Localized strain at the tip of a martensite
particle also contributes to void at the martensite/ferrite interface in dual phase steels due to stress concentration
and stress relaxation by void formation. These parameters are bases for both analyzed and numerical modelling
in order to optimize forming performance and related to metallurgical parameters such as the hardness, volume
fraction and shape of the martensite particles.
Thirdly, softening of the martensite is suggested as a method to improve the formability of dual phase steels.
This is exemplified that softening of the martensite through tempering treatments at 300-600°C significantly
improves the formability, counterbalanced by a reduction of ultimate tensile strength in the dual phase steels. In
particular, the improvement of post-uniform elongation is remarkable. The structural evolution has revealed a
relationship between the hardness of the martensite and the three factors related to void formation. Softening of
the martensite increases the strain in the martensite and decreases the strain in the ferrite due to strain
compatibility. Reduction of strain in the ferrite contributes to slow void growth in the martensite and ferrite.
The more mild strain partitioning by softening of the martensite reduces the localized strain at the concave part
and tip of the martensite and thereby retards void formation in the martensite and ferrite. In addition,
improvement of the property of the martensite representing to the hardness significantly retards void formation
in the martensite. Reduction of localized strain and the large critical strain contributes to retardation of void
formation in the martensite and ferrite and thereby improves the formability in dual phase steels.
Moreover, it is shown that an increase of volume fraction of the martensite (0.21->0.38) mainly affects strain
partitioning between the martensite and ferrite, and thereby affects void formation in dual phase steels and
mechanical properties of dual phase steels. In particular, a larger strain in the martensite phase accelerates the
void formation in the martensite in dual phase steels containing a large amount of martensite particles. On the
other hand, the critical strain for void formation in the martensite does not depend on the martensite fraction,
but martensite hardness. A combination of softening of the martensite and an increase of the martensite fraction
enables one to develop advanced dual phase steels with high strength and better formability due to the different
effects on the three factors due to a different effect of martensite hardness and martensite fraction on the critical
strain and the strain partitioning in dual phase steels. For instance, a combination of tempering treatment at
400°C and volume fraction change (0.21->0.38) increases the post-uniform elongation from 8.7 to 14.8% and
hole-expansion ratio from 20 to 40% without changing the ultimate tensile strength.
Finally, control of martensite morphology is suggested as a method to retard void formation in dual phase
steels. In particular, martensite particles elongated in the tensile direction deform more and can lead to void
formation at a smaller strain in dual phase steels.
101
Acknowledgement
The authors thank Nippon Steel Corporation for financial support and for producing the steel sheets used in
this study, as well as acknowledging support from the Danish National Research Foundation and the National
Natural Science Foundation of China (Grant No. 50911130230) for the Danish-Chinese Centre for Nanometals,
within which part of this work was performed.
Outlook
In this section, the hope of non-destructive observation of the deformation and fracture behaviour in
advanced materials is discussed. As shown in the present study, to clarify void and crack formation mechanisms
is significantly important to improve the formability. In addition, to reveal strain partitioning between a matrix
and hard particles in multi-phase materials is of great importance to understand the stress-strain relationship in
advanced materials. As Landron et al. (2010) suggest that stress-triaxiality affects macroscopic void formation
behaviour, void formation behaviour such as formation rate and growth rate of voids in bulk samples might be
different from on surfaces. Therefore, non-destructive observations of deformation and fracture behaviour in
advanced materials is of great importance. However, to reveal the mechanisms is limited by lack of non-
destructive observation of deformation and fracture behaviour in bulk samples. This causes the deformed
microstructures of dual phase steels are very complicated and thereby there is no suitable observation technique.
Figs. 5.20-26 show that deformation localizes at the tip of and at the concave part of martensite particles and the
change of the accumulated misorientation is about 3-14° per 1m. On the other hand, as spatial resolution of X-
ray tomography and 3D-XRD (X-Ray Diffraction), which are well known as typical non-destructive
observations, is about 1m×1m×1m, it might be difficult to detect each deformed ferrite grain and martensite
particle, and strain localization related to martensite particles during deformation. In addition, advanced high
strength steels consist of ultra-fine grains and particles. As reported by Calcagnotto et al. (2011), the grain size
is often less than 1m. Void formation and deformation behaviour in bulk samples are still puzzling; where
voids form and interact with the constituents, even though Maire et al. (2008) estimate the size, number density
and volume fraction of voids in a dual phase steel by X-ray tomography. Moreover, computational analysis such
as finite element modelling is also helpful to optimize the formability of advanced materials based on 3D-
structure measured by X-ray tomography, 3D-XRD and serial sectioning. Therefore, it is hoped that non-
destructive observation such as X-ray tomography and 3D-XRD have been developed more and applied more
frequently for the characterization for deformation and fracture mechanism in advanced metals.
Table 3D/4D observation techniques

Technique Type Parameters issues
In-situ SEM non-destructive Deformation on surfaces Not observing deformation and fracture
Void formation on surfaces behaviour in bulks
Serial destructive 3D microstructure Not following deformation and fracture
sectioning behaviour
X-ray non-destructive Void formation in bulks Not distinguishing the ferrite with the
tomography martensite in dual phase steels
3D-XRD non-destructive Strain distribution Spatial resolution
Too small grains, too many grains
102
References
Adachi1 Y., Sato N., Mayumi Ojima M., Nakayama M. and Wang Y. T. Development automated serial-
sectioning 3D microscope and topological approach to pearlite and dual-phase microstructure in steels.
In proceedings of international conference of 3D materials Science. 2012. 37-42.
Akisue O. and Hada T. Past Development and Future Outlooks of Automotive Steel Sheets. Nippon Steel
Technical Report. 1995, 64, 1-6.
Akbarpour M. R. and Ekrami A. Effect of ferrite volume fraction on work hardening behaviour of high
bainite dual phase (DP) steels. Mat. Sci. Eng. A, 2008, 477A, 206-10.
Akisue O. and Usuda M. New types of steel sheets for automobile weight reduction. Nippon Steel Technical
Report, 1993, 57, 11-15.
Ahmad E., Manzoor T. and Hussain N. Thermomechanical processing in the intercritical region and tensile
properties of dual-phase steel. Mater. Sci. Eng. A, 2009, A508, 259-65.
Avramovic-Cingara G., Saleh C. H. A. R., Jain M. K. and Wilkinson DS. Void nucleation and growth in
dual-phase steel 600 during uniaxial tensile testing. Metal Mater. Trans. A, 2010, 40A, 3117- 27.
Avramovic-Cingara G., Ososkov Y., Jain M. K. and Wilkinson D. S. Effect of martensite distribution on
damage behaviour in DP600 dual phase steels. Mater. Sci. Eng. A, 2009, A516, 7-16.
Babout L., Maire E., Buffiere J. Y. and Fougeres R. Characterization by X-ray computed tomography of
decohesion, porosity growth and coalescence in model metal matrix composites. Acta Matell., 2001, 49,
2055-63.
Babout L., Maire E., Buffiere J. Y. and Fougeres R. Damage initiation in model metallic materials: X-ray
tomography and modelling. Acta Mater., 2004, 52, 2475-87.
Calcagnotto M., Adachi Y., Ponge D. and Raabe D. Deformation and fracture mechanisms in fine- and
ultrafine-grained ferrite/martensite dual-phase steels and the effect of aging . Acta Metall., 2011, 59, 658-
70.
Christman T., Needleman A., Nutt S. and Suresh S. On microstructural evolution and micromechanical
modelling of deformation of a whisker-reinforced metal-matrix composite. Mater. Sci. Eng. A, 1989,
A107, 49-61.
Clyne T. W. and Withers P. J. In An introduction to metal matrix composites. Cambridge University Press;
1993, 231-34.
Davies R.G. In Formable HSLA and Dual Phase Steels. TMS-AIME, (ed. Davenport A. T.)Warrendale,
PA; 1977, 25-39.
Davies R. G. The deformation behaviour of a Vanadium-strengthened dual phase steel. Metall. Trans. A,
1978, 9A, 41-52.
Delincé M., Jacques P. J. and Pardoen T. Separation of size-dependent strengthening contributions in fine-
grained dual phase steels by nanoindentation. Acta Metall., 2006, 54, 3395-04.
Delincé M., Bréchet Y., Embury J. D., Geers M. G. D., Jacques P. J. and Pardoen T. Structure-property
optimization of ultrafine-grained dual-phase steels using a microstructure-based strain hardening model.
Acta. Metall., 2007, 55, 2337-50.
Erdogan E. The effect of new ferrite content on the tensile fracture behaviour of dual phase steel. J. Mater.
Sci.,2002, 37, 3623-30.
103
Fang X., Fan Z., Ralph B., Evans P. and Underhill R. Effects of tempering temperature on tensile and hole
expansion properties of a C-Mn steel. Mater. Proc. Tech., 2003, 132, 215-18.
Gerbase J., Embury J. D. and Hobbs R. M. In Structure and Properties of Dual-Phase Steels, (eds. Kot R. A.
and Morris J.W.) TMS-AIME, New York, 1979, 118–144.
Ghadbeigi H., Pinna C., Celotto S. and Yates J. R. Local plastic strain evolution in a high strength dual-
phase steel. Mater. Sci. Eng., A 2010, 527A, 5026-32.
Gladman T., Holmes B. and Mclvor I. D. In Effect of Second-Phase particles on Mechanical Properties of
Steel.ISI Publication, Londaon, 1971, 68.
González C. and Llorca J. An analysis of the effect of hydrostatic pressure on the tensile deformation of
aluminum-matrix composites. Mater. Sci. Eng., 2002, A341, 256-263.
Gurland J. Observations on the fracture of cementite particles in a spheroidized 1.05%C steel deformed at
room temperature. Acta Metall., 1972, 20, 735-41.
Jardim O. R., Longo W. P. and Chawla K. K. Fracture behaviour of a Tempering dual phase steel.
Metallography, 1984, 17, 123-130.
Jena A. K. and Charturvedi M. C. On the effect of the volume fraction of the martensite on the tensile
strength of dual-phase steel. Mat. Sci. Eng., 1988, 100, 1-6.
Jia N., Cong Z. H., Sun X., Cheng S., Nie Z. H., Ren, Liaw R. K. and Wang Y. D. An in-situ high-energy
X-ray diffraction study of micromechanical behaviour of multiple phases in advanced high-strength
steels. Acta Mater., 2009, 57, 3965-3977.
Jiang Z., Guan Z. and Lian J. Effects of microstructural variables on the deformation behaviour of dual-
phase steel . Mater. Sci. Eng., A 1995, A190, 55-64.
Joarder A., JHA J. N., Ojha S. N. and Sarma D. S. The tempering behaviour of a plain carbon Dual-Phase
steel. Mater. Chara., 1990, 25, 199-209.
Kadkhodapour J., Butz A., Ziaei-Rad S. and Schmauder S. A micro mechanical study on failure initiation of
dual phase steels under tension using single crystal plasticity model. Int. J. Plast., 2011, 59, 2575-2588.
Kim S. and Lee S. Effects of Martensite Morphology and Volume Fraction on Quasi-Static and Dynamic
Deformation Behaviour of Dual-Phase Steels. Metall. Mater. Trans. A, 2000, 31A, 1753-1760.
Kishida K. High strength steel sheets for light weight vehicle. Nippon Steel Technical Report, 2000, 81, 12-
16.
Hadianfard M. J. Low cycle fatigue behaviour and failure mechanism of a dual-phase steel. Mater. Sci.Eng.
A, 2009, 499A, 493-499.
Han S. K. and Margolin H. Void formation, void growth and tensile fracture of plain carbon steel and a
dual-phase steel. Mater. Sci. Eng. A, 1989, A112, 133-41.
Han S., Byoungchul H., Sunghak L., Chang G. L. and Kim S. J. Effect of martensite morphology and
tempering on dynamic deformation behaviour of dual-Phase steels. Metall. Trans. A, 2004, vol.35A,
2371-2382.
Hasegawa K., Kawamura K., Urabe T. and Hosoya Y. Effects of microstructure on stretch-flange-
formability of 980MPa grade cold-rolled ultra high strength steel sheets. ISIJ Int., 2004, 44, 603-09.
Hayashi K., Miyata K. and Katsuki F. Deformation behaviour in high-strength dual-phase steel sheets
during bending test. Tetsu-to-Hagane, 2012, vol.98, No. 6, pp. 82-88.
104
He X. J., Terao N. and Berghezan A. Influence of martensite morphology and its distribution on mechanical
properties and fracture mechanisms of Fe-Mn-C dual phase steels. Mater. Sci., 1984, 18, 367-73.
Hosseini S. B., Temmel C., Karlsson B. and Ingesten N. G. An in-situ scanning electron microscopy study
of the bonding between MnS inclusions and the matrix during tensile deformation of hot-rolled steels.
Metal. Mater. Trans. A, 2007, 34A, 982-89.
Jiang Z., Guan Z. and Lian J. Effects of microstructural variables on the deformation behaviour of dual-
phase steel. Mater. Sci. Eng. A, 1995, 190, 55-64.
Kadkhodapour J., Butz A., Rad S. Z. and Schmauder S. Experimental and numerical study on geometrically
necessary dislocations and non-homogeneous mechanical properties of the ferrite phase in dual phase
steels. Int. J. Plast., 2011, 59, 2575-88.
Kang J., Ososkov Y., Embury J. D. and Wilkinson D. S. Digital image correlation studies for microscopic
strain distribution and damage in dual phase steels. Scripta Mater., 2007, 56, 999-1002.
Kelestemur O., Kelestemur M. H. and Yildiz S. Improvement of Mechanical Properties of Reinforcing

Steel Used in the Reinforced Concrete Structures. J. Iron Steel Res. Inter., 2009, 16, 55-63.
Kim S. and Lee. S. Effects of martensite morphology and volume fraction on quasi-static and dynamic
deformation behaviour of dual phase steels. Metal. Mater. Trans. A, 2000, 31A, 1753-1760.
Koo J. Y. and Thomas G. In Formable HSLA and Dual Phase Steels. (ed. Davenport A. T.) TMS-AIME,
Warrendale, PA, 1977, 25-39.
Korzekwa D. A., Lawson R. D., Matlock D. K. and Krauss G. A consideration of models describing the
strength and ductility of dual-phase steels. Scripta Metal. 1980, 14, 1023-1028.
Kosco J. B. and Koss D. A. Ductile fracture of mechanically alloyed iron-yttria alloys. Metall. Mater., 1993,
A 24, 681-87.
Landron C., Bouaziz O., Maire E. and Adrien J. Characterization and modeling of void nucleation by
interface decohesion in dual phase steels. Scripta Metall.,2010, 63, 973-76.
León-Garcia O., Petrov R. and Kestens L. A. I. Void initiation at TiN precipitates in IF steels during tensile
deformation. Mater .Sci. Eng. A 2010, A527, 4202-09.
Li B. L., Godfrey A. and Liu Q. Investigation of Macroscopic Grain Sub-Division of an IF-Steel during
Cold-Rolling. Mater. Sci. Forum, 2002, 408-412, 1185-1190.
Mishnaevsky Jr L. In Computational Mesomechanics of Composites. John Wiley & Sons. Ltd. 2007, 161-
165.
LeRoy G., Embury J. D., Edward G. and Ashby M. F. A model of ductile fracture based on the nucleation
and growth of voids. Acta Metal., 1981, 29, 1509-22.
Lian J., Jiang Z. and Liu J. Theoretical model for the tensile work hardening behaviour of dual-phase steel.
Mater. Sci. Eng. A, 1991, A147, 55-65.
Lubarda V. A., Schneider M. S., Kalantar D. H., Remington B. A. and Meyers M. A. Void growth by
dislocation emission. Acta Metall., 2004, 52, 1397-08.
Maire E., Bouaziz O., Michiel M. D. and Verdu C. Initiation and growth of damage in a dual-phase steel
observed by X-ray microtomography. Acta Metall., 2008, 56, 4954-64.
Marder A. R. and Bramfitt B. L. In Structure and Properties of Dual-Phase Steels. (eds. Kot R. A., Morris J.
W.) TMS-AIME, Warrendale, PA; 1979, 145-82.
105
Matsuoka S., Hasegawa K. and Tanaka K. Newly-developed ultra-high tensile strength steels with excellent
formability and weldability. JFE Technical Report, 2007, 10, 13-18.
Mazinani M. and Poole W. J. Effect of martensite plasticity on the deformation behaviour of a low-carbon
dual phase steel. Metall. Mater. Trans. A, 2007, 38A, 328-39.
Melander A. and Steninger J. The roles of surfides, oxides and pearlite in the ductile fracture of a niobium
microalloyed steel. Mater. Sci. Eng. 1982, 52, 239-248.
Meyers M. A., Traiviratana S., Lubarda V. A., Benson D. J. and Bringa E. M. The role of dislocations in the
growth of nanosized voids in ductile failure of metals . JOM, 2009, 61, 35-41.
Miyahara K., Matsuoka S. and Hayashi T. Nanoindentation as a Strength Probe-a study on the hardness
dependence of indent size for fine-grained and coarse-grained ferritic steel. Metal. Mater. Trans., 2001,
32A, 761-768.
Morito S., Tanaka H., Furuhara T. and Maki T. In Proc. of Int. Sympo. on in Honer Professor Krauss G.,
ASM, Cincinnati, 1999, 631.
Morito S., Nishikawa J. and Maki T. Dislocation Density within Lath Martensite in Fe-C and Fe-Ni Alloys.
ISIJ-Int., 2003, 43, 1475-.1477
Morito S., Tanaka H., Konishi R.. Furuhara T. and Maki T. The morphology and crystallography of lath
martensite Fe-C alloys. Acta Mater., 2003, 51, 1789-1799.
Morooka, S., Umezawa, O., Harjo, S., Hasegawa, K. and Toji , Y. Quantitative analysis of tensile
deformation behaviour by in-situ neutron diffraction for ferrite-martensite type dual-phase steel. Tetsu-to-
Hagane, Vol. 2012, vol. 98, pp.311-319.
Movahed P., Kolahgar S., Marashi S. P. H., Pouranavari M. and Parvin N. The effect of intercritical heat
treatment temperature on the tensile properties and work hardening behaviour of ferrite-martensite dual
phase steels. Mat. Sci. Eng. A, 2009, A518, 1-6.
Nakagawa A. H. and Thomas G. Microstructure-mechanical property relationships of dual-phase steel wire.

Metall. Trans. A, 1985, 16A, 831-40.
Natori M., Futamura Y., Tsuchiyama T. and Takaki S. Difference in recrystallization behaviour between
lath martensite and deformed ferrite in ultralow carbon steel. Scripta Mater., 2005, 53, 603-606.
Kishida K. High Strength Steel Sheets for Light Weight Vehicle. Nippon Steel Technical Report. 2000, 81,
12-16.
Nishimoto A. Hosoya Y. And Nakaoka K. In Fundamentals of Dual-Phase Steels, (eds. Kot R.A., Bramfitt
B.L.) TMS-AIME, Warrendale, PA, 1981, 447-455.
Nutt Sr. and Needleman A. Void nucleation at fibre ends in Al-SiC composites. Scripta Metall., 1987, 21,
705-10.
Poruks P., Yakubtsov I. and Boyd J. D. Martensite-ferrite interface strength in a low-carbon bainitic steel.
Scripta Mater., 2006, 54, 41-45.
Osokov Y., Wilkinson D. S., Jain M. and Simpson T. In-situ measurement of local strain partitioning in a
commercial dual-phase steel. Int. J. Mater. Res., 2007, 98, 664-673.
Qiu H., Mori H., Enoki M. and Kishi T. Evolution of ductile fracture of structural steels by microvoid
model. ISIJ Int., 1999, 39, 358-64.
Pickering F. B. In Hardenability Concepts with Application to Steel. (ed. Daone D. V. and Kirkaldy J. S.)
AIME, 1978, 179.
106
Rashid M. S. In Formable HSLA and Dual-Phase Steels, (ed. Davenport A. T.) TMS-AIME, New York,
1979, 1-24.
Ray R. K. Tensile fracture of a dual-phase steel. Scripta Metall., 1984, 18, 1205-09.
Roberts W. Lehtinen B. and Easterling K. E. An in-situ SEM study of void development around inclusions
in steel during plastic deformation. Acta Metall., 1976, 24, 745-58.
Sabirov I. and Kolendnik O. The effect of inclusion size on the local condition for void nucleation near a
crack tip in a mild steel. Scripta Metall., 2005, 53, 1373-78.
Saito T., Takahashi M., Miyazaki Y., Mizuhashi N. and Sakiyama T. Welding Technology for Uncoated
and Coated Steel Sheets for Automobiles. Seitetsu Kenkyu, 1987, 34, 327-32.
Samuel F. H. Effect of Dual-Phase treatment and tempering on the microstructure and mechanical
properties of a high strength, low alloy steel. Metall. Trans., 1985, 75, 51-66.
Sarwar M. and Priestner R. Influence of ferrite-martensite microstructural morphology on Tensile

properties of dual-phase steel. J. Mater. Sci., 1996, 31, 2091-95.
Sarwar M., Manzoor T., Ahamd E. and Hussain N. The role of connectivity of martensite on the tensile
properties of a low alloy steel. Material and design, 2007, 28, 1928-33.
Shen H. P., Lei T. C. and Liu J. Z. Microscopic deformation behaviour of martensitic-ferritic dual-phase.
Mater. Sci. Tech., 1986, 2, 28-33.
Shimada S., Takada Y., Lee J. and Tanaka T. Trial of Evaluate Wettability of Liquid Zn with Steel Sheets
Containing Si and Mn. ISIJ-Int., 2008, 48, 1246-1250.
Speich G. R. and Miller R. L. In Structure and Properties of Dual-phase Steels, (eds. Kot R.A., Morris
J.W.) TMS-AIME, New York, NY, 1979, 145-182.
Steinbergunner D. L., Matlock D. K. and Krauss G. Void formation during tensile testing of dual phase
steels. Metal. Trans. A, 1988;19A:579-89.
Stelmashenko N. A., Walls M. G., Brown L. M. and Milman Y. U. V. Microindentations on W and Mo

oriented single crystals: AN STM study. Acta Metall. Mater., 1993, 141, 2855-2865.
Su Y. L. and Gurland J. Strain partition, uniform elongation and fracture strain in dual-phase steels. Mater.
Sci.Eng.,1987, 95, 151-65.
Sun S. and Pugh M . Properties of thermomechanically processed dual-phase steels containing fibrous
martensite. Mater. Sci. Eng. A, A335 (2002) 298-308.
Sugimoto K., Sakaki T., Fukusato T. and Miyagawa O. Influence of Martensite Morphology on Initial
Yielding and Strain Hardening in a 0.11C-1.36Mn Dual phase Steel. Tetsu-to-hagane, 1985, 71, 994-
1001.
Szewczyk A. F. and Gurland J. A study of the deformation and fracture of a dual-phase steel. Metall. Trans.
A,1982, 13A, 1821-26.
Takada Y., Shimada S., Lee J. Kurosaki M. and Tanaka T. The Effect of Si and Mn Content on Dynamic
Wetting of Steel with Liquid Zn. ISIJ-Int., 2009, 49, 100-104.
Tasan C. C., Hoefnagels J. P. M. and Geers M. G. D. Microstructural banding effects clarified through
microscopic digital image correlation. Scripta Metall., 2010, 62, 835-38.
Tasan C. C., Hoefnagels J. P. M. and Geers M. G. D. Experimental analysis of strain path development
ductile damage mechanics and forming limit. Mechanics and Mater., 2009, 41, 1264-1276.
107
Toji Y., Takagi S., Yoshino K., Hasegawa K. and Tanaka Y. Evaluation of hydrogen embrittlement for high
strength steels. Tetsu-to-Hagane, 2009, 95, 81.
Thomas G. and Koo J. Y. In Structure and Properties of Dual phase Steels. (eds. Kot R. A and Morris J.
W.), TMS-AIME, Warrendale, PA, 1979, 183-201.
Tomota Y., Yoshino H. and Kuroki K. Effect of ductility of the second phase on ductile fracture. Scripta
Metall.,1977,11, 853-56.
Traiviratanav S., Bringa E. M., Benson D. J. and Meyers M. A. Void growth in metals: Atomistic
calculations original. Acta Metall., 2008, 56, 3874-86.
Uthaisangsuk V., Prahl U. and Bleck W. Micromechanical modelling of damage behaviour of multiphase
steel. Comp. Mater. Sci., 2008, 43, 27-37.
Yamazaki K., Mizuyama Y., Oka M., Tsuchiya H. and Yasuda H. Recent advance in ultrahigh-strength
sheet steels for automotives. Nippon Steel Technical Report, 1995, 64, 37-44.
Vajragupta N., Uthaisangsuk V., Schmaling B., Munstermann S., Hartmaier A. and Bleck W. A
Micromechanical damage simulation of dual phase steels using XFEM. Comput. Mater. Sci., 2012, 54,
271-279.
Vasudevan A. K. and Richmond O. The influence of hydrostatic pressure on the ductility of Al-SiC
composite. Mater. Sci. Eng., 1989, A107, 63-69.
Whitehouse A. F. and Clyne T. W. Cavity formation during testing starting of particulate and short fiber
metal matrix composites. Acta Metall., 1993, 41, 1701-11.
108
Danmarks Tekniske Universitet
DTU Vindenergi
Nils Koppels Allé
Bygning 403
2800 Kgs. Lyngby
Telefon 45 25 25 25
info@vindenergi.dtu.dk
www.vindenergi.dtu.dk

DTU Wind Energy Masafumi Azuma

Uploaded by

Copyright:

Available Formats

DTU Wind Energy Masafumi Azuma

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

DTU Wind Energy Masafumi Azuma

Uploaded by

Copyright:

Available Formats

Downloaded from orbit.dtu.

dk on: Nov 11, 2020

Structural control of void formation in dual phase steels

Link back to DTU Orbit

Abstract ISSN: 0106-2840

Softening of the martensite significantly retards void formation in the

Danmarks Tekniske Universitet

Supervised by Xiaoxu Huang and Grethe Winther

Section for Materials Science and Advanced Characterization

Chapter 1 Introduction ......................................................................................................................... 10

Chapter 2 Background .......................................................................................................................... 11

Chapter 3 Materials and techniques .................................................................................................... 16

Chapter 4 Void formation mechanism ................................................................................................ 32

Chapter 5 Effect of martensite hardness ............................................................................................. 53

Chapter 6 Effect of volume fraction of martensite ............................................................................. 83

Chapter 7 Effect of the shape of martensite particles on deformation ............................................. 96

Chapter 8 Conclusions ........................................................................................................................ 100

Acknowledgement ................................................................................................................................... 102

References ............................................................................................................................................... 103

2.1 Formability for automobiles

2.2 Advanced High Strength Steels

2.3 Voids in metals

Materials and techniques

3.1 Mechanical tests

3.1.1 Hardness test

10µm 10µm 10µm

3.1.2 Tensile test

Dual phase steel

3.1.3 Hole-expansion test

Hole-expansion ratio = (d – d0) / d0 × 100% Eq.[4]

3.2 Scanning electron microscopy

3.4 Transmission electron microscopy

3.5 Strains in specimens

Before deformation During deformation

e) Local strain at a necked region of a martensite particle during an in-situ test:

necked region of a martensite particle during an in-situ test.

LF0 5m LF1 5m

Table 3.4 Mechanical properties of the dual phase steels

Fig. 3.15 Nominal stress-nominal strain curves of SVM and LVM.

Void formation mechanism

martensite crack void

martensite void void

martensite void void

martensite grain boundary void void

4.2 Experimental procedures

4.3.1 Deformed microstructure after a tensile test

50m void 10m

Examples of all three sites are given in Fig. 4.3:

4.3.2 Void formation behaviour during in-situ testing

2m 2m 2m

10m 5m void

10m 5m coalescence

5µm 5µm 5µm

Number density / 1/mm2

Length range of void / m

Length range of void / m

4.4.1 Plastic deformation of martensite

4.4.2 Cracks in martensite

4.4.3 Voids in martensite