University of Pretoria etd – Tang, Z (2007)
Optimising the transformation and yield to ultimate strength ratio of
Nb-Ti micro-alloyed low carbon line pipe steels through alloy and
microstructural control
By Zhenghua Tang
Submitted in partial fulfilment of the requirements for the degree
Philosophiae Doctor (Metallurgical Engineering)
in the
Department of Materials Science and Metallurgical Engineering
Faculty of Engineering, Built Environment and Information Technology
University of Pretoria
Republic South Africa
22 May 2006
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
To my daughter Mingyi
谨此献给我的女儿唐铭艺
-i-
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Acknowledgments
Firstly, I would like to express my sincere gratitude and appreciation to Professor
Waldo Stumpf* for his supervision, invaluable guidance and discussions and professor
Chris Pistorius, Head of Department, and the University of Pretoria for the provision
of facilities and financial support.
*Membership of national and international bodies:
z
Fellow of the South African Academy of Engineering
z
Past Chairman of the UN-International Atomic Energy Agency’s Standing
Advisory Group on Technical Assistance and Co-operation (SAGTAC) in Vienna
for the years 2002 to 2005
z
Past Member of an Expert Group of the UN-IAEA in 2004/2005 on the nonproliferation considerations of Multi-National Arrangements on Uranium
enrichment and spent fuel reprocessing
z
Past Honorary President of the South African Branch of the Institution of Nuclear
Engineers
I would like to specially thank the following kind people:
Dr N.G. van den Bergh for technical support in shadowing replicas.
Professor G.T.van Rooyen for useful discussions and technical support in the tensile
tests.
Mr CF van der Merwe for technical support in much of the work on transmission
electron microscopy.
Mr Charl Smal for technical support in tensile tests and others.
Mr Johann Borman for his assistance before his untimely death.
I also wish to thank
- ii -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Dr Kevin Banks for providing materials for the alloys and making many helpful
suggestions.
Mr Francois Verdoorn for training on the Gleeble and Dilatometer.
Mrs Havenga Sarah for her assistance in administrative matters.
Mr Charles Siyasiya for his assistance in editing my English.
I would like also to thank others who gave me assistance.
The provision of financial support from Mittal Steel SA and the NRF through the
THRIP program is also greatly appreciated.
Finally, I would like to also specially thank my wife and my daughter Mingyi for they
gave me a wonderful and memorable life of four years in far SA and, for their
assistance and encouragement.
致谢
首先,我要衷心的感谢我的导师, 南非工程院院士, 2002-2005 维也纳联合国原子
能机构技术支持与协助顾问团主席, 2004/2005 联合国多国铀废料回收专家组成
员,核工程协会南非分会前名誉主席, 比勒陀利亚大学教授, Stumpf 先生对我的精
心指导! 和比勒陀利亚大学材料科学与冶金工程系主任 Pistorius 教授以及大学所
给予的设备和资金方面的支持!
特别感谢如下的南非友好人士:
N.G. van der Berg 博士在电镜投影复型方面的技术支持;
G.T.van Rooyen 教授的有益讨论和特殊拉伸试验方面的帮助;
- iii -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
CF van der Merwe 先生在透射电镜实验方面提供的方便;
Charl Smal 先生在拉伸试验和其他方面的帮助;
Johann Borman 先生生前非常友好的帮助;
还要感谢
Kevin Banks 博士提供了实验所需的材料及一些有益的建议;
Verdoorn Francois 先生在 Gleeble 和 Dilatometer 的培训;
Havenga Sarah 女士在行政事务方面的帮助;
Mr Charles Siyasiya 在校正英语方面的帮助;
以及其他给予帮助的人士；
还要感谢南非 Mittal 钢公司及 THRIP 项目所给予的资金支持；
此外,我特别要感谢我的妻子和女儿陪伴我在遥远的南非度过了令我终身难忘的
非常快乐的四年海外生活，同时她们也给了我无数的帮助和鼓励。
- iv -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Optimising the transformation and yield to ultimate strength ratio of
Nb-Ti micro-alloyed low carbon line pipe steels through alloy and
microstructural control
by Zhenghua Tang (唐正华)
Supervisor: Professor Waldo Stumpf
Department: Materials Science and Metallurgical Engineering
University: University of Pretoria
Degree: Philosophiae Doctor
Abstract
Thinner walled (about 6 mm thickness) line pipe steels for smaller diameter pipelines
tend to have a relatively high ratio of yield strength to ultimate tensile strength
(YS/UTS) of 0.93 or higher. This study focused on the effect of the microstructures,
prior deformation in the austenite, cooling rate, coiling simulation and the additions of
some micro-alloying elements on the YS/UTS ratio of a currently produced Nb-Ti and
some experimental Nb-Ti-Mo line pipe steels. The experimental research included the
design of the chemical compositions for five experimental alloys, simulation of the
controlled hot rolling process, the determination of the strain-free as well as the strain
affected continuous cooling transformation (CCT) diagrams, phase identification and
quantitative assessment of the microstructures by optical microscopy, scanning
electron microscopy (SEM) and transmission electron microscopy (TEM), the latter
especially on shadowed carbon extraction replicas and, tensile tests etc.
This study indicated that the transformed microstructures of the alloys were a mixture
of acicular ferrite plus polygonal ferrite and the volume fraction of acicular ferrite
varied from 46.3 to 55.4%. Molybdenum additions did not markedly affect the
-v-
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
formation of acicular ferrite after hot rolling and rapid cooling. The microstructural
details of the acicular ferrite were successfully revealed by TEM on shadowed
extraction replicas. This technique was useful to distinguish the acicular ferrite from
the polygonal ferrite through a more smooth surface relief after etching in 2% Nital of
the little etched polygonal ferrite whereas the deeper etched acicular ferrite showed
parallel sets of internal striations. This made it possible to measure the volume
fraction of acicular ferrite in the mixed microstructures of acicular ferrite and
polygonal ferrite.
The continuous cooling transformation behaviors of two alloys with no molybdenum
and with 0.22% Mo were constructed with no prior deformation as well as with prior
deformation of the austenite. Molybdenum additions shifted the strain-free CCT
diagram towards longer times and expanded the region in which acicular ferrite
formed from a cooling rate range of 0.3 to 5 ºCs-1 (Mo-free) to 0.1 to 15 ºCs-1 (with
0.22% Mo). However, its effect was significantly overshadowed by prior deformation
in the austenite. The strain affected CCT diagrams for both alloys appear to be similar.
The prior deformation had a stronger effect on the CCT diagram than molybdenum
additions and promoted acicular ferrite formation, whereas it suppressed the
formation of bainite. The prior deformation had two effects in acicular ferrite
formation: it promoted nucleation and suppressed its growth and, therefore, resulted
in a finer overall grain size.
The effect on the YS/UTS ratio at various cooling rates ranging from 1 to 34, 51, 54
or 60 ºCs-1 was investigated in three cases: (i) without prior deformation and coiling
simulation, (ii) with no prior deformation but with coiling simulation at 575 and 600
ºC and, (iii) with prior deformation of 33% reduction in the austenite below the Tnr
followed by coiling simulation at 575 ºC for 1 hour. It was determined that the
YS/UTS ratio was a function of the microstructure and cooling rate in the case
(treatment (i)) without any coiling simulation and prior deformation. The coarse
bainite or acicular ferrite, which was formed at high cooling rates, raised the YS/UTS
ratio under conditions of no deformation prior to the transformation. The yield
strength and ultimate tensile strength also increased with an increase in cooling rate.
- vi -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
With coiling conditions (treatment (ii)), the ratio was not sensitive to the cooling rate
or the microstructure for the reference Mo-free alloy #6 because the coiling allows
recovery of dislocations, thereby decreasing the difference in dislocation density that
had arisen between a low and a high cooling rate. The YS/UTS ratio ranged from 0.75
to 0.8 after a simulated coiling at 575 ºC and from 0.76 to 0.78 after a coiling
simulation at 600 ºC.
Prior deformation (treatment (iii)) in the austenite raised the ratio from 0.81 to 0.86.
However, the YS/UTS ratio was not sensitive to cooling rate after coiling at 575 ºC
for 1 hour in the cases with and without prior deformation in the austenite.
Deformation with a 33% reduction below the Tnr prior to the transformation increased
the yield strength more than the ultimate tensile strength, leading to a high YS/UTS
ratio that ranged from 0.81 to 0.86. The prior deformation, therefore, had a stronger
effect on the YS/UTS ratio than the microstructure.
Key words:
line pipe steel, acicular ferrite, microstructure, ratio of yield strength to ultimate
tensile strength,, Nb-Ti micro alloyed steel, controlled hot rolling process, CCT
diagram, non-recrystallisation temperature, nucleation of acicular ferrite
- vii -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Table of Contents
Acknowledgments……………………………………………………………………ii
Abstract……………………………………………………………………………….v
CHAPTER 1 INTRODUCTION…………………………………………………….1
1.1 Strengthening mechanisms in line pipe steels……………………………………..1
1.2 Chemical composition of line pipe steels……………………………………….…1
1.3 Steel making of line pipe steel…………………………………………………….2
1.4 The controlled rolling of strip steel………………………………………………..3
1.4.1 Rolling schedule………………………………………………………………3
1.4.1.1 Reheating………………………………………………………………….3
1.4.1.2 Rough rolling……………………………………………………………...3
1.4.1.3 Finish rolling……………………………………………………………...4
1.4.1.4 Heavy reduction…………………………………………………………..4
1.4.2 Cooling rate…………………………………………………………………...5
1.4.3 Coiling temperature…………………………………………………………...5
1.5 Pipe forming and welding process ……………………………………………….5
CHAPTER 2 MICRO-ALLOYING ELEMENTS AND THEIR EFFECT ON
PRECIPITATION……………………………………………………….7
2.1 Vanadium…………………………………………………………………………..7
2.2 Niobium……………………………………………………………………………7
2.3 Titanium……………………………………………………………………….…...9
2.4 Molybdenum………………………………………………………………….…..10
2.5 Carbon………………………………………………………………………….…11
2.6 Manganese………………………………………………………………………..11
2.7 Copper, Chromium and Nickel…………………………………………………...12
CHAPTER 3 THE CONTROLLED ROLLING PROCESS OF LINE PIPE STEELS
…………………………………………………………………….…...13
3.1 Three stages of deformation for controlled rolling process……….……………...13
3.1.1 Deformation in the austenite recrystallisation region…………………….…..14
3.1.2 Deformation in the non-recrystallisation region……………………….……..14
3.1.3 Deformation in the (α+γ) two-phase region………………………………….15
- viii -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
3.2 Reheating temperature and undissolved particles………………………………..16
3.3 Rough rolling…………………………………………………………………….18
3.4 Finish rolling……………………………………………………………………..19
3.5 Heavy reductions…………………………………………………………………21
3.6 Strip thickness……………………………………………………………………22
3.7 Cooling rate after finish rolling…………………………………………………..23
3.8 Finish temperature of accelerated cooling………………………………………..25
3.9 Coiling temperature………………………………………………………………27
3.10 Non-recrystallisation temperature (Tnr)…………………………………………28
3.10.1 Effect of alloying elements…………………………………………………28
3.10.2 Effect of the controlled rolling process…………………………………….29
CHAPTER 4 MICROSTRUCTURE AND MECHANICAL PROPERTIES………..32
4.1 Acicular ferrite……………………………………………………………………32
4.1.1 Nucleation and growth of acicular ferrite……………………………………32
4.1.2 Two types of acicular ferrites: upper and lower acicular ferrite……………..33
4.1.3 Effect of the hot rolling process on acicular ferrite formation………………35
4.2 Acicular ferrite and bainite……………………………………………………….35
4.3 Mechanical properties of line pipe steel………………………………………….36
4.3.1 The ratio of yield strength to ultimate tensile strength (YS/UTS)…………..36
4.3.2 Toughness……………………………………………………………………38
4.3.3 U-O pipe forming and the Bauschinger effect………………………………38
CHAPTER 5 BACKGROUND OF CURRENT SOUTH AFRICA LINE PIPE
PRODUCTION………………………………………………………..42
5.1 Line pipe steel composition of Mittal Steel (South Africa)………………………42
5.2 Parameters of the hot rolling process at Mittal Steel (SA).………………………43
5.3 Typical microstructures and existing developments within Mittal Steel for line
pipe steel………………………………………………………………………….44
5.4 The hypothesis for this study……………………………………………………..45
5.4.1 Design of chemical compositions of the investigated alloys…………….…..45
5.4.2 Design of the controlled hot rolling process……………………….………...46
- ix -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
CHAPTER 6 EXPERIMENTAL PROCEDURES…………………………………...49
6.1 Alloy design………………………………………………………………………49
6.2 The melting of the experimental alloys…………………………………………..50
6.3 The effect of reheating temperature and soaking time on the austenite grain size
……………………………………………………………………………………51
6.4 Measuring the presence and composition of undissolved particles……………...53
6.5 Non-recrystallisation temperature (Tnr)………………………………………….54
6.5.1 Testing schedule for the determination of the Tnr……………………………55
6.5.2 The determination of the non-recrystallisation (Tnr)…………………………57
6.6 CCT diagram…………………………………………………………………….59
6.6.1 The Ac1 and Ac3 test………………………………………………………….59
6.6.2 CCT diagram without prior deformation…………………………………….60
6.6.3 Strain affected CCT diagram…………………………………………………61
6.7 The thermo-mechanical process………………………………………………….63
6.7.1 Cooling unit………………………………………………………………….63
6.7.2 Hot rolling process of the laboratory ingots…………………………………64
6.7.2.1 Reheating before laboratory hot rolling………………………………….64
6.7.2.2 Rough rolling of the laboratory hot rolling………………………………64
6.7.2.3 Finish rolling of the laboratory hot rolling………………………………65
6.7.2.4 Cooling rate after laboratory finish rolling………………………………65
6.7.2.5 Simulation of coiling after laboratory hot rolling………………………..65
6.7.2.6 Hot-rolling process curve…………………………………………….…..66
6.8 The identification of acicular ferrite………………………………………….…..66
6.8.1 Observation with optical microscopy and by SEM…………………………67
6.8.2 Observation of replicas by TEM……………………………………...…….68
6.8.2.1 Preparing replicas without shadowing……………………………………68
6.8.2.2 Preparing shadowed carbon extraction replicas …………………………68
6.8.3 Thin foil TEM samples………………………………………..……………68
6.9 Test of subsize samples on the Gleeble with various cooling rates, coiling
temperatures and prior deformation………………………………………..…….69
6.9.1 Hot rolling plates for Gleeble samples………………………………….…..69
6.9.2 Tests on the Gleeble…………………………………………………….…..70
6.9.3 Tensile tests…………………………………………………………….…...72
6.10 Test of mechanical properties on the as-hot rolled alloys…………………..…..73
-x-
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
CHAPTER 7 RESULTS……………………………………………………………..74
7.1 The effect of the austenitisation temperature and holding time on the presence of
undissolved particles in the V-Nb-Ti-containing alloys………………………....74
7.2 Austenite grain size and reheating temperature…………………………………..85
7.3 The non-recrystallisation temperature (Tnr) and deformation parameters………..88
7.3.1 The Tnr and pass strain……………………………………………………….89
7.3.2 The Tnr and inter-pass time……………………………………………….….93
7.3.3 The Tnr and pass strain rate……………………………………………….….99
7.4 Continuous cooling transformation (CCT diagrams) under strain-free conditions
…………………………………………………………………………………..104
7.4.1 CCT diagram for alloy #6 without molybdenum and without prior deformation
………………………………………………………………………..……..104
7.4.2 CCT diagram for alloy #5 with 0.22% molybdenum and without prior
deformation……………………………………………………………..…..109
7.5 Strain enhanced continuous cooling transformation (CCT diagram) under
deformed conditions………………………………………………………..….113
7.5.1 Strain affected CCT diagram of the Mo-free alloy #6…………………..…..114
7.5.2 Strain affected CCT diagram of alloy #5 (with 0.22% Mo)………….……..118
7.6 The results of the laboratory hot rolling process on the YS/UTS ratio…………124
7.7 Volume fraction of acicular ferrite ……………………………………………..125
7.8 Mechanical properties…………………………………………………………..125
7.8.1 Results of experimental alloys………………………………………………125
7.8.2 Results of mechanical properties for different cooling rates, coiling
temperatures and deformation values……………………………………….126
7.8.2.1 Effect of cooling rate with no coiling and no prior deformation………..126
7.8.2.2 Effect of cooling rate with coiling but without deformation……………128
7.8.2.3 Effect of cooling rate and 575 ºC coiling with a 33% prior reduction below
the Tnr……………………………………………………………………129
7.9 Transformed microstructures of the alloys……………………………………..129
7.9.1 Optical micrographs………………………………………………………..129
7.9.2 Microstructures examined by SEM…………………………………….…..131
7.9.3 TEM studies of acicular ferrite on carbon replicas………………………...135
- xi -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
CHAPTER 8 STUDIES OF ACICULAR FERRITE BY THIN FOIL TEM..….….140
8.1 Acicular ferrite morphology in experimental alloys……………………………140
8.1.1 Acicular ferrite and polygonal ferrite in alloy #6 (Mo-free)………………...141
8.1.2 Acicular ferrite and polygonal ferrite in alloy #1……………………………145
8.1.3 Acicular ferrite in alloys #2 to #5……………………………………………148
8.2 Two types of acicular ferrite…………………………………………………….156
8.2.1 Structure with parallel laths………………………………………………….156
8.2.2 Structure with interwoven laths………………………………………….…..159
8.3 Nucleation of acicular ferrite……………………………………………………162
8.3.1 Nucleation on non-metallic inclusions………………………………………162
8.3.2 Type of non-metallic inclusion as nucleants……………………………...…168
8.3.3 Nucleation mechanisms of acicular ferrite………………………………….168
CHAPTER 9 DISCUSSION……………………………………………………….174
9.1 Effect of molybdenum additions on the continuous cooling transformations.…174
9.1.1 Effect of molybdenum on polygonal ferrite formation……………………...174
9.1.2 Effect of molybdenum on acicular ferrite formation…………………….….175
9.2 Effect of deformation in austenite on acicular ferrite formation………….……176
9.3 Ratio of yield strength to ultimate tensile strength and its effect…………….....179
9.3.1 The effect of cooling rate……………………………………………………179
9.3.2 The effect of coiling temperature……………………………………………183
9.3.3 The effect of prior deformation in the austenite and coiling simulation…….187
9.3.4 The effect of acicular ferrite on the ratio of yield strength to ultimate tensile
strength……………………………………………………………………...190
CHAPTER 10 CONCLUSIONS………………………………………………...…..192
CHAPTER 11 RECOMMENDATIONS FOR FUTURE WORK……………….….195
References………………………………………………………………………...…196
Appendix…………………………………………………………………….………205
- xii -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
List of figures:
Figure 3.1 Schematic illustration of the three stages of the controlled rolling
process[11].
Figure3.2 Metallurgical mechanisms during thermo mechanical hot rolling[8].
Figure 3.3 Correlation between the increase in yield strength ΔYS and the content of
soluble Nb[24] (0.075% C, 043% Si, 1.58% Mn, 0.013% P, 0.005% S, 0.51% Ni,
0.021% Al and 0.106% Nb, and 0.077% C, 0.36% Si, 1.56% Mn, 0.017% P,
0.005%S, 0.47% Ni, 0.020% Al and 0.130% Nb).
Figure 3.4 Comparison between the calculated and measured quantities of microalloying elements as precipitates[37].
Figure 3.5 Change in austenite grain size during reheating process[39].
Figure 3.6 Influence of rolling conditions on the mechanical properties of plate and
strip of different thicknesses[24].
Figure 3.7 Influence of rolling conditions on the average ferrite grain size[24].
Figure 3.8 Improvement in yield strength and Charpy ductile to brittle transition
temperature[10].
Figure 3.9 Plot of the volume fraction of M/A constituents versus the degree of prior
deformation[13].
Figure 3.10 Strip and pipe properties for various X80 pipe sizes)[6]. (Steel G: 0.075%
C, 1.59% Mn, 0.31% Si, 0.057% Nb, 0.22% Mo, 0.013% Ti and 0.006% N; Steel N:
0.070% C,1.53% Mn, 0.19% Si, 0.045 Nb,0.20% Mo, 0.012% Ti and 0.0045% N.
Figure 3.11 Schematic representation of the cooling pattern on the run-out table of a
Hot Strip Mill[2].
Figure 3.12 Influence of reduction and cooling rate on the ferrite grain diameter[10].
Figure 3.13 Effect of finish temperature of accelerated cooling on the average ferrite
grain diameter[39].
Figure 3.14 Comparison of mechanical properties between the observed and the
calculated strengths of micro-alloyed steels[39].
Figure 3.15 Effect of coiling temperature on the YS of hot rolled strip for various
types of alloys[2].
Figure 4.1 Schematic stress-strain curves for the outer (top) and inner (bottom)
material during the U-O pipe forming process, with (left) at 180º and (right) at 30º
from the welding line[93].
- xiii -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Figure 4.2 The change of the Bauschinger effect factor with carbon and manganese
content[94].
Figure 4.3 The Bauschinger effect in micro-alloyed steel. The upper two curves are
for steels with 0.2% C, 0.4% Mn, unalloyed or alloyed respectively with Al, V or
Nb. The lower two curves are for low-pearlite steels with less than 0.1% C, 2% Mn,
and alloyed with Mo, Nb and Ti[94].
Figure 5.1 The optical microstructure of cast #521031, Mittal Steel line pipe
Figure 6.1 Schematic of the modified McQuaid-Ehn carburising process of the
samples directly after reheating.
Figure 6.2 Schematic schedule employed in the multi-pass compression tests for the
Tnr.
Figure 6.3 The curves of flow stress versus strain in a multi-pass compression test on
alloy #6.
Figure 6.4 Determining Tnr from the mean flow stress in MPa versus the inverse pass
temperature in K, during a multi-pass compression test on alloy #6.
Figure 6.5 Schematic dilation as a function of testing temperature[126].
Figure 6.6 Schematic determination of the Ac1 and Ac3 temperatures on the heating
curve[126].
Figure 6.7 Dilatometer chamber
Figure 6.8 Schematic schedule of the test for the CCT diagram on the THETA
Dilatometer.
Figure 6.9 Chamber of the Gleeble 1500D DSI
Figure 6.10 Schematic schedule of the test for the strain affected CCT diagrams on the
Gleeble.
Figure 6.11 Experimental arrangement of the cooling unit for controlled cooling: (a)
overall view, (b) controller for mixing of gas and water, (c) valves for the nozzles
and, (d) cooling spray in the chamber from the spraying jets.
Figure 6.12 Schematic schedule of the hot rolling process on the experimental alloys.
Figure 6.13 Preliminary samples on the Gleeble of (a) type A and (b) type B
Figure 6.14 Graph of the heating and cooling process on the Gleeble for samples
#A124 (the Mo-free alloy #6) and sample #AF3F (the 0.09% Mo alloy #3).
Figure 6.15 Graphs of the heating and cooling cycles in the Gleeble on the Mo-free
alloy #6 for samples (a) #A113 and (b) #B113.
- xiv -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Figure 6.16 Graph of heating, cooling and deformation process in the Gleeble for
sample #TEN06 (the Mo-free alloy #6).
Figure 6.17 Tensile test samples of (a) type A and (b) type B. T is original thickness of
the plates and 6 mm for the as-rolled alloy and the Gleeble samples, respectively.
Figure 7.1 Extraction replicas without shadowing with undissolved particles for alloy
#6 after reheating at 1200 ºC for 15 min. (Most of the darker spots are not particles
but are etching debris on the replicas).
Figure 7.2 Extraction replicas with Au-Pd shadowing for alloy #6 after reheating at
1225 ºC for 120 min.
Figure 7.3 TEM micrograph of particles on the shadowed replicas of alloy #6 reheated
at 1150 ºC for (a) 15 min, (b) 60 min and, (c) 120 min.
Figure 7.4 TEM micrograph of particles on the shadowed replicas of alloy #6 reheated
at 1200 ºC for (a) 15 min, (b) 60 min and, (c) 120 min.
Figure 7.5 TEM micrograph of particles on the shadowed replicas of alloy #6 reheated
at 1225 ºC for (a) 15 min, (b) 60 min and, (c) 120 min.
Figure 7.6 TEM micrograph of particles on the shadowed replicas of alloy #6 reheated
at 1250 ºC for (a) 15 min, (b) 60 min and, (c) 120 min.
Figure 7.7 TEM-EDS results of undissolved particles of alloy #6 after the treatments
of (a) as-hot rolled, (b) as-hot rolled, (c) 1150 ºC for 120 min and, (d) 1200 ºC for
15 min.
Figure 7.8 The effect of reheating temperature and time on the volume fraction of
undissolved particles for alloy #6.
Figure. 7.9 Pro-eutectoid cementite decorates the original austenite grain boundaries
in alloy #6 for soaking times of 60 min at different austenitisation temperatures.
Figure. 7.10 The relationship between the austenitisation temperature and the
austenite grain size for alloy #6. The broken line is from published data[39].
Figure. 7.11 The weak effect of soaking time on the austenite grain size for alloy #6.
Figure 7.12 Stress-strain curves of multi-pass compression tests of alloy #6 at the
same strain rate of 1 s-1, inter-pass time of 8 seconds and different pass strains.
Figure 7.13 Determination of the Tnr on the mean flow stress versus inverse
temperature curves of alloy #6, all deformed at the same strain rate of 1 s-1 and an
inter-pass time of 8 seconds but at different pass strains.
Figure 7.14 The relationship between pass strain (ε) and the non-recrystallisation
temperature for alloy #6. Strain rate ε& =1.0 s-1, inter-pass time tip=8 s.
- xv -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Figure 7.15 Stress-strain curves of multi-pass compression tests on alloy #6 at a c
onstant pass strain of 0.20, a constant strain rate of 1 s-1 and a series of inter-pass
times ranging from 4 to 50 seconds.
Figure 7.16 The mean flow stress versus inverse temperature curve of alloy #6 during
multi-pass compression testing at a constant pass strain of 0.20 and a constant strain
rate of 1 s-1 but with a variation of the inter-pass times between 4 and 50 seconds.
Figure 7.17 The Tnr as a function of inter-pass time (tip) for alloy #6. Strain rate ε& =1.0
s-1, pass strain ε = 0.2.
Figure 7.18 Stress-strain curves of compression deformation tests of alloy #6 at a
constant pass strain of 0.20 and inter-pass time of 8 s but with a series of strain
rates from 0.1 to 2.22 s-1.
Figure 7.19 The mean flow stress versus inverse test temperature of alloy #6 at a
constant pass strain of 0.20 and inter-pass time of 8 s but at a series of strain rates
from 0.1 to 2.22 s-1.
Figure.7.20 Strain rate ( ε& ) versus the non-recrystallisation temperature for alloy #6.
pass strain ε = 0.2, inter-pass time tip=8 s.
Figure 7.21 The optical micrographs (etched in 2% Nital) of the Mo-free alloy #6 and
with no prior deformation after continuous cooling. PF-polygonal ferrite, P-pearlite
and AF-acicular ferrite microstructure.
Figure 7.22 The CCT diagram of the Mo-free alloy #6 and no prior deformation. PFpolygonal ferrite, P-pearlite, AF-acicular ferrite microstriucture and B-bainite.
Figure 7.23 The optical micrographs (etched in 2% Nital) for alloy #5 (with 0.22%
Mo) and with no prior deformation after continuous cooling. PF-polygonal ferrite,
P-pearlite and AF-acicular ferrite microstructure.
Figure 7.24 The CCT diagram of alloy #5 (with 0.22% Mo) and with no prior
deformation. PF-polygonal ferrite, P-pearlite, AF-acicular ferrite microstructure and
B-bainite.
Figure 7.25 The optical micrographs (etched in 2% Nital) of the Mo-free alloy #6
after compression testing with a single pass strain of 0.6 at 860 ºC (which is below
the Tnr), a strain rate of 0.5 s-1 and cooling down to room temperature at different
cooling rates. PF-polygonal ferrite, P-pearlite and AF-acicular ferrite microstructure.
Figure 7.26 The strain affected CCT diagram of the Mo-free alloy #6 after a single
pass compression strain of 0.6 at 860 ºC with a strain rate of 0.5 s-1. PF-polygonal
ferrite, P-pearlite and AF-acicular ferrite microstructure.
- xvi -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Figure 7.27 The optical micrographs (etched in 2% Nital) of alloy #5 (with 0.22% Mo)
after a single pass compression of 0.6 strain at 860 ºC and at a strain rate of 0.5 s-1
and then cooling at different rates.
Figure 7.28 The strain affected CCT diagram of alloy #5 (with 0.22% Mo) after a
single pass compression of 0.6 strain at 860 ºC at a strain rate of 0.5 s–1. PFpolygonal ferrite, P-pearlite and AF-acicular ferrite microstructure.
Figure 7.29 The optical microstructure, etched in a 2% Nital solution for 5 seconds,
after a rapid cooling rate of 47 ºCs-1 for the experimental alloys (a) #1, (b) #2, (c)
#3, (d) #4, (e) #5 and, (f) the reference Mo-free alloy #6 cooled at a rate of 39 ºCs-1.
Figure 7.30 The SEM micrographs after a rapid cooling rate of 47 ºCs-1(etched in 2%
Nital for 5 seconds) for the experimental alloys (a) #1, (c) #2, (c) #3, (d) #4, (e) #5
and, (f) the reference alloy #6 cooled at a rat of 39 ºCs-1.
Figure 7.31 The micrographs in the as-rolled condition by high resolution SEM after a
rapid cooling rate of 47 ºCs-1(etched in 2% Nital for 5 seconds) for the
experimental alloys (a) #1, (c) #2, (c) #3, (d) #4 and, (e) #5.
Figure 7.32 The high resolution SEM micrographs in the as-rolled condition after a
rapid cooling rate of 47 ºCs-1 for the experimental alloy #1 etched in 2% Nital for (a)
10seconds, (c) 15 seconds, (c) 30 seconds, (d) 60 seconds and, (e) 120 seconds.
Figure 7.33 TEM micrographs of carbon extraction replicas without shadowing for
the reference alloy #6 after hot rolling and rapid cooling at a rate of 39 ºCs-1.
Figure 7.34 The TEM micrograph from a shadowed replica of the Mo-free alloy #6
after hot rolling and rapid cooling at a rate of 39 ºCs-1. (AF-acicular ferrite, PFpolygonal ferrite and, GB-grain boundary).
Figure 7.35 TEM micrographs from shadowed replicas of the as-hot rolled and
rapidly cooled (at a rate of 47 ºCs-1) experimental alloys (a) #1, (b) #2, (c) #3, (d)
#4 and, (e) #5 (AF-acicular ferrite, PF- polygonal ferrite and, GB-grain boundary).
Figure 8.1 Thin foil TEM micrographs of alloy #6 (Mo-free) after a rapid cooling rate
of 39 °Cs-1 after hot rolling, (a) polygonal ferrite + laths, (b) polygonal ferrite with
dislocations and, (c),(d) laths with dislocations. PF-polygonal ferrite, AF-acicular
ferrite, A and B-dislocations in polygonal ferrite and an acicular ferrite ,
respectively.
Figure 8.2 Dislocations within the polygonal ferrite in thin foil of the Mo-free alloy
#6. The area M shows a high density of dislocations possibly being emitted from
the moving PF interface while the central regions L of the PF have less dislocations.
- xvii -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Figure 8.3 Polygonal ferrite in a TEM thin foil micrograph from the experimental
alloy #1 after a rapid cooling rate of 47 ºCs-1 after the hot rolling process.
Figure 8.4 TEM thin foil micrograph with laths from alloy #1 after a rapid cooling
rate of 47 ºCs-1 after the hot rolling process.
Figure 8.5 TEM thin foil micrograph of the lath plus PF structure in alloy #1 after a
rapid cooling rate of 47 ºCs-1 after the hot rolling process.
Figure 8.6 Thin foil TEM micrographs of the lath structure in alloy #2 after a rapid
cooling rate of 47 ºCs-1 after the rolling process.
Figure 8.7 Thin foil TEM composite micrographs of parallel laths of an acicular
ferrite in alloy #2 after a rapid cooling rate of 47 ºCs-1 after the hot rolling process.
Figure 8.8 Polygonal ferrite (with a few isolated laths) in alloy #3 with a rapid cooling
rate of 47 ºCs-1 after the hot rolling process.
Figure 8.9 The parallel lath morphology in alloy #3 after a rapid cooling rate of 47
ºCs-1 after the hot rolling process.
Figure 8.10 Thin foil TEM micrographs of a mixture of polygonal ferrite and an
acicular ferrite in alloy #3 after a rapid cooling rate of 47 ºCs-1 after the hot rolling
process.
Figure 8.11 Thin foil TEM composite micrographs of the acicular ferrite in alloy #4
after a rapid cooling rate of 47 ºCs-1 after the hot rolling process.
Figure 8.12 Thin foil TEM micrographs from alloy #5 with 0.22% Mo (a) polygonal
ferrite, (b) and (c) acicular ferrite with interwoven laths.
Figure 8.13 Parallel lath morphology in a colony in alloy #3 after rapid cooling at a
rate of 47 ºCs-1 after the hot rolling process.
Figure 8.14 Interwoven arrangement between lath colonies in alloy #4 after a fast
cooling rate of 47 ºCs-1 after the hot rolling process.
Figure 8.15 Interwoven laths micrographs in alloy #3 after fast cooling of 47 ºCs-1
after hot rolling process.
Figure 8.16 Acicular ferrite morphology in alloy #5 after fast cooling of 47ºCs-1 after
the hot rolling process.
Figure 8.17 (a) TEM image of acicular ferrite and a large non-metallic inclusion in
alloy #5 after a rapid cooling rate of 47 ºCs-1 after the hot rolling, (b) EDS analysis
on the inclusion in this figure (a).
Figure 8.18 Laths nucleated on non-metallic inclusions (a) in alloy #1 after a rapid
cooling rate of 47 ºCs-1 after the hot rolling, (b) in alloy #3 after a rapid cooling rate
- xviii -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
of 40 ºCs-1 from 980 ºC down to room temperature without deformation, (c) EDS
analysis on the inclusion in this figure (a), (d) and (e) EDS analysis on the
inclusions A and B in this figure (b), respectively.
Figure 8.19 (a) Nucleation of interwoven laths of acicular ferrite in sample #AF3F of
alloy #3 after a cooling rate of 20 ºCs-1 from 980 ºC down to room temperature
without deformation, (b) EDS analysis of red peak was from on the inclusion
indicated by an arrow in this figure (a), while blue peak was from the matrix steel.
Figure 8.20 (a) Non-metallic inclusion and acicular ferrite in alloy #3 after a rapid
cooling rate of 47 ºCs-1 after the hot rolling, (b) EDS analysis on the inclusion
indicated by an arrow in this figure (a).
Figure 9.1 The yield strength of the reference Mo-free alloy #6 as a function of the
cooling rate from 980 ºC with no prior deformation before the transformation and
with no coiling simulation.
Figure 9.2 The ultimate tensile strength of the reference alloy #6 as a function of
cooling rate from 980 ºC with no prior deformation before the transformation and
with no coiling simulation.
Figure 9.3 The YS/UTS ratio of the reference alloy #6 as a function of cooling rate
from 980 ºC with no prior deformation before the transformation and with no
coiling simulation. PF-polygonal ferrite, AF-acicular ferrite and P-pearlite.
Figure 9.4 The yield strength of alloy #3 as a function of cooling rate from 980 ºC
under conditions of no prior deformation to the transformation and no coiling
simulation.
Figure 9.5 The ultimate tensile strength of alloy #3 as a function of cooling rate from
980 ºC under conditions of no prior deformation to the transformation and no
coiling simulation.
Figure 9.6 The YS/UTS of alloy #3 as a function of cooling rate from 980 ºC
underconditions of no prior deformation to the transformation and no coiling
simulation.PF-polygonal ferrite, AF-acicular ferrite, B-bainite and P-pearlite
Figure 9.7 The yield strength of alloy #6 as a function of the cooling rate from 980 ºC
to 600 ºC under conditions of no prior deformation to the transformation but with a
coiling simulation at 600 ºC for 1 hour.
Figure 9.8 The ultimate tensile strength of alloy #6 as a function of cooling rate from
980 ºC to 600 ºC under conditions of no prior deformation to the transformation but
with a coiling simulation at 600 ºC for 1 hour.
- xix -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Figure 9.9 The YS/UTS ratio of alloy #6 as a function of the cooling from 980 ºC to
600 ºC under conditions of no prior deformation to the transformation but with a
coiling simulation at 600ºC for 1 hour. PF-polygonal ferrite, AF-acicular ferrite, Bbainite and P-pearlite.
Figure 9.10 The yield strength of alloy #6 as a function of the cooling rate from 980
ºC to 575 ºC under conditions of no prior deformation to the transformation but
with a coiling simulation at 575 ºC for 1 hour.
Figure 9.11 The ultimate tensile strength of alloy #6 as a function of the cooling rate
from 980 ºC to 575 ºC under conditions of no prior deformation to the
transformation but with a coiling simulation at 575 ºC for 1 hour.
Figure 9.12 The YS/UTS ratio alloy #6 as a function of the cooling rate from 980 ºC
to 575 ºC under conditions of no prior deformation to the transformation but with a
coiling simulation at 575 ºC for 1 hour. PF-polygonal ferrite, AF-acicular ferrite, Bbainite and P-pearlite.
Figure 9.13 Effect of the cooling rate on the yield strength of the reference alloy #6
after prior deformation of 33 % reduction in the austenite below the Tnr, cooling to
575 ºC at different cooling rates and simulation of the coiling at 575 ºC for 1 hour.
Figure 9.14 Effect of the cooling rate on the ultimate tensile strength of the reference
alloy #6 after prior deformation of 33 % reduction in the austenite below the Tnr,
cooling to 575 ºC at different cooling rates and simulation the coiling at 575 ºC for
1 hour.
Figure 9.15 Effect of the cooling rate on the YS/UTS ratio of the reference alloy #6
after prior deformation of 33 % reduction in the austenite below the Tnr, cooling to
575 ºC at different cooling rates and simulation the coiling at 575 ºC for 1 hour. PFpolygonal ferrite, AF-acicular ferrite and P-pearlite.
Figure 9.16 Relationship between the YS/UTS ratio (longitudinal specimens) and the
measured volume fraction of acicular ferrite in the experimental alloys #1 to #5
after laboratory hot rolling with an 86% reduction in total and rapid cooling at a
rate of 47 ºCs-1.
- xx -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
List of tables:
Table 3.1 Reduction/pass of hot rolling
Table.4.1 Specification of line pipe steels of API[12]
Table 5.1 Typical chemical composition of the current 11 mm line pipe steel of Mittal
Steel, (wt%)
Table 5.2 The parameters of the hot rolling process at Mittal Steel
Table 5.3 Design of chemical composition ranges of alloys that were investigated (in
wt%)
Table 6.1 Chemical compositions of the experimental alloys, in wt.%
Table 6.2 The composition of the etchant solutions
Table 6.3 Temperatures and soaking time of the treatment for undissolved particles
Table 6.4 Calculation equilibrium Nb carbonitride solution temperature
Table 6.5 Testing parameters for Tnr at strain ranging from 0.15 to 0.32
Table 6.6 Testing parameters for Tnr at inter-pass times ranging from 4 to 50 seconds
Table 6.7 Testing parameters for Tnr at strain rate ranging from 0.1 to 2.22 s-1
Table 7.1 Measured volume fraction of particles on replicas with/without shadowing
in alloy #6
Table 7.2 Undissolved particles: types and sizes after reheating treatments of alloy #6
Table 7.3 Intercept length austenite grain size, in µm, versus reheating temperature
and soaking time of alloy #6
Table 7.4 The non-recrystallisation temperature and pass strains of alloy #6
Table 7.5 The non-recrystallisation temperature of alloy #6 as affected by different
inter-pass times
Table 7.6 The non-recrystallisation temperature and strain rates of alloy #6
Table 7.7 The laboratory hot rolling parameters for alloy #1
Table 7.8 Measured results of volume fraction of acicular ferrite
Table 7.9 Mechanical properties of the experimental alloys
Table 7.10 Mechanical properties of samples #A124 of the Mo-free alloy #6 with no
coiling and no prior deformation
Table 7.11 Mechanical properties of samples #AF3F of alloy #3 (with 0.09% Mo)
and with no coiling and no prior deformation
- xxi -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Table 7.12 Mechanical properties of samples #A113 of the Mo-free alloy #6 with 60
min coiling at 600 ºC without prior deformation
Table 7.13 Mechanical properties of samples #B113 of alloy #6 with 60 min coiling at
575 ºC without prior deformation
Table 7.14 Mechanical properties of samples #TEN06 for the Mo-free alloy #6 with
60 min coiling at 575 ºC and a 33% prior reduction below the Tnr
- xxii -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 1 Introduction
CHAPTER 1 INTRODUCTION
The demand for high-performance line pipe steels has resulted in extensive research
being conducted towards increasing, firstly, performance of high-strength low alloy
steels (HSLA) and secondly, to improve large diameter forming processes to increase
the capacity of pipelines used at higher operating pressures. The trends in the
technology of line pipe are towards the use of thinner wall thicknesses and UOE
pipe-forming, (plate-U-forming-O-bending and shrinking-E-expanding). These
demands include the introduction of higher strength and higher toughness steels with
lower ductile-to-brittle transition temperatures and higher impact energies. The grades
of line pipe steels have, therefore, been developed from X65 grade towards X80 and
X100 to meet these requirements.
1.1 Strengthening mechanisms in line pipe steels
The main strengthening mechanisms used in line pipe steels are strengthening by solid
solution hardening, by dislocation substructure, by phase transformation strengthening,
by precipitation hardening and by grain refinement. The solid solution strengthening
result from elements such as manganese and molybdenum, the phase transformation
from lower transformation temperature phases such as acicular ferrite and bainitic
ferrite or martensite, resulting in finer microstructures with a higher dislocation
density. On account of the micro-alloying element additions, carbonitrides of
vanadium, niobium and titanium contribute to the precipitation strengthening. Besides
dispersion hardening niobium has an added benefit on the refinement of the ferrite
grains. Higher pass strains below the non-recrystallisation temperature during the
controlled rolling process, also contributes to good ferrite grain refinement.
1.2 Chemical composition of line pipe steels
A special alloying design philosophy is used for line pipe steels. Micro-alloying
element additions are necessary to obtain the high strengths, and raise the temperature
-1-
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 1 Introduction
above which complete recrystallisation during hot rolling occurs and the
recrystallisation of austenite is, therefore, retarded below this temperature. Generally
speaking, the relatively low levels of the main chemical elements contribute to a low
carbon equivalent (CE) for good weldability, and approximately an 0.15% total
micro-alloying element content of vanadium + niobium + titanium is usual[1]. The
micro-alloying elements may be only one of these three or a combination of any two
or three of the micro-additions (V, Nb and Ti). Molybdenum and niobium are
beneficial to the formation of acicular ferrite and they suppress the formation of
pearlite[ 2, 3, 4 ]. Boron additions are used to form bainitic ferrite and also to suppress
the formation of pearlite. Lower carbon contents (below 0.05% or less) are considered
to improve weldability, increase the toughness, form less pearlite and increase the
strength due to more effective dissolution of niobium.
1.3 Steel making of line pipe steel
The main steel-making procedure for line pipe steels is a combination of converter
blowing and ladle treatment[5] or the Electric Arc Furnace (EAF) or the Basic Oxygen
Furnace (BOF) route. The cleanliness of the steel and the optimum chemical
composition are important for obtaining the specified properties of the line pipe. The
main considerations concerning the cleanliness of the steel are the requirements for a
low sulphur content for high fracture toughness of the pipe body and the avoidance of
clustered alumina inclusions in the vicinity of the weld line. CaSi ladle injection
treatment (~0.8 kg/t
[2]
is often added for this purpose) is vital to reduce the sulphur
content and to change the type and morphology of sulphide inclusions. The preferred
inclusion composition and morphology is generated when the Ca:Al ratio in the steel
is about 0.20[2]. The inclusion shape is also controlled by Ti and rare earth element
additions. Vacuum degassing is another method to reduce the impurity content of
these steels.
-2-
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 1 Introduction
1.4 The controlled rolling of strip steel
The continuously-cast slab is processed after melting by Electric Arc Furnace (EAF)
or the Basic Oxygen Furnace (BOF). The major production route of strip for line pipe
steels is that of controlled rolling and is most effective with line pipe steels containing
micro-alloying elements that raise Tnr. Briefly, the entire process involves slab
reheating, rough rolling, finish rolling, cooling on the run-out table and coiling.
1.4.1 Rolling schedule
1.4.1.1 Reheating
Of importance here, are the reheating temperature and the presence of undissolved
carbide particles: The objective of reheating is to achieve uniformly heated slabs with
the complete dissolution of micro-alloyed carbonitrides (Nb,V)(C,N). Slabs are heated
uniformly within either walking beam furnaces or pusher-type furnaces. The reheating
temperature depends on which micro-alloying elements are present in the steel. It
ranges generally from 1150 to 1250 ºC for Nb-containing steels. TiN particles will
remain undissolved at these temperatures and serve to inhibit the growth of γ-grains
during reheating.
1.4.1.2 Rough rolling
The objective of rough rolling is to achieve the finest possible recrystallised austenite
grain size before the Tnr is reached[6]. In this stage both recrystallisation and strain
hardening will take place[7,8]. The rough-rolling phase is completed above the
non-recrystallisation temperature (Tnr), typically above about 1030 ºC for X60 to X80
ERW line pipe steels [2, 9]. Specially developed reduction schedules together with TiN
particles that inhibit growth of recrystallised austenite grains, are employed for this
purpose[6]. For most line pipe steels the temperature of rough rolling is from 1200
down to about 950 ºC[7, 9, 10, 11]. A minimum pass strain of 0.2% is needed for the
required austenite grain refinement[12].
-3-
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 1 Introduction
1.4.1.3 Finish rolling
The objective of this phase is to accumulate rolling strain within the
non-recrystallisation region of the austenite grains so that on subsequent ferrite
transformation, ferrite nucleation sites are greatly multiplied in number and a very
fine ferrite grain size can be achieved after rolling and during controlled cooling[6].
Finish rolling is, therefore, undertaken in the non-recrystallisation austenite region in
the lower temperature range above the Ar3 temperature (i.e. Tnr>T>Ac3)[9]. NbN also
precipitates in this stage due to decreasing solubility of Nb in the austenite. The
austenite grains elongate and become “pan-caked” in shape. The starting temperature
for controlled rolling ranges from 1050 to 950 ºC and finishes around 730 to 1000 ºC
[1]
depending on alloying and processing conditions. The total finishing reduction
below the Tnr should be about 64 to 80% when the initial and final thicknesses of the
strip are 110 and 18mm respectively [1,6].
1.4.1.4 Heavy reduction
The heavy reduction (60~85%) during controlled rolling increases the yield strength
without an adverse effect on the Charpy ductile-brittle transition temperature (DBTT).
Moreover, the heavy reduction followed by rapid cooling on the run-out table
remarkably improves both the yield strength and the DBTT[10]. On their own, the
heavy reduction results in grain refinement and the rapid cooling increases the
fraction of bainitic structure. However, when the heavy reduction is followed by rapid
cooling, the microstructure with the increased fraction of bainite becomes
significantly grain-refined[10]. These results also showed that the volume fraction of
martensite/austenite (M/A) constituents decreased on increasing the total deformation
for a steel containing 0.03%C, 0.05%Nb and 0.024%Ti[13].
-4-
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 1 Introduction
1.4.2 Cooling rate
The cooling rate (Vc) and the finishing temperature (Tc) of accelerated cooling after
finish rolling are important parameters of thermo-mechanically controlled processing
for line pipe steels to achieve optimum strength through controlling the influence of
transformation microstructure, ferrite grain refinement and precipitation hardening.
The effect of Tc and Vc may be attributed to the control of the precipitation of
carbonitrides during accelerated cooling[9]. The availability of high cooling rates
immediately after finish rolling provides greater microstructural adjustment
possibilities, especially for the case of Mo-Nb steel types. The possibilities for
adjustment of the relative volume fractions of polygonal and acicular ferrite (and
grain size) through the setting of the cooling rate and coiling temperature, are evident
in principle[2]. Rapid cooling increases the volume fraction of bainite or M/A islands
or acicular ferrite.
1.4.3 Coiling temperature
There is no phase transformation during coiling, but it is very important for
precipitation strengthening of these steels, so it requires careful control of the coiling
temperature if the maximum degree of precipitation strengthening is to be attained.
TiC[5], NbC[12] and V(C,N) precipitate in the ferrite during coiling. The temperature
should be sufficiently low (i.e. a high degree of under-cooling) for fine precipitation
of these carbides and carbonitrides in the ferrite but not so low that the volume
fraction of TiC is reduced through too slow diffusion[5]. When the coiling temperature
is too low, precipitation is suppressed, resulting in low strength.
1.5 Pipe forming and welding process
The process of pipe making is through cold forming and cold expansion. There are
two possible seam lines for the manufactured pipes, one is a spiral seam and the other
longitudinal. Most of the pipelines manufactured globally belong to the former, but, as
the quality of the plate has increased, pipelines with longitudinal seams have
-5-
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 1 Introduction
increased lately in many manufacturing processes.
The last operation of pipe making is to either weld the seam by submerged arc
welding, or by electrical resistance welding as has been preferred in the recent
pipeline manufacturing process [6]. Low carbon equivalence improves the weldability.
-6-
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 2 Micro-alloying elements and their effects on precipitation
CHAPTER 2 MICRO-ALLOYING ELEMENTS AND THEIR EFFECTS ON
PRECIPITATION
2.1 Vanadium
Vanadium is used in line pipe steels to form vanadium carbonitrides. Its content
ranges from 0.03 to 0.05%V[14]. The main action of vanadium in these steels is
dispersion hardening, refinement of ferrite grains and inhibiting grain growth of
austenite up to 1050 ºC, but it is detrimental to the formation of acicular ferrite[4]. As
the solubility of vanadium in austenite is high there is a weak influence of vanadium
on the non-recrystallisation temperature (Tnr ) of austenite. The solubility of vanadium
in ferrite, however, is low, and its addition is extremely effective in the precipitation
strengthening of ferrite.
The precipitation starting temperature of VN particles is 1005 ºC through the
solubility equation of the form log(X%)(Y%)=(-A/T)+B, where X and Y in the
equation are mass percentages of the dissolved micro-alloys, A and B are constants,
(A and B are 8330 and 3.46 for VN, respectively), T is the absolute temperature[15].
The solubility limit of vanadium in typical line pipe steels is about 0.2% at 1150
ºC[14].
Vanadium carbonitrides will mostly dissolve in austenite at about 900 ºC[9] .
Vanadium also inhibits grain growth of austenite up to 1050 ºC. It passes completely
into solid solution above 1050 ºC for typical line pipe compositions. Vanadium has no
effect on the recrystallisation of hot-deformed austenite above 900 ºC[16].
2.2 Niobium
Niobium carbide is fully dissolved between 1225 and 1250 ºC. It has a greater
refinement effect on ferrite grain size, dispersion hardening and reduction of pearlite
content. Niobium has a stronger effect on recrystallisation of hot-deformed
-7-
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 2 Micro-alloying elements and their effects on precipitation
austenite[17-19]. Niobium in solution is also considered to suppress the γ/α
transformation due to pining of the γ/α interfaces by NbC[20]. The solute niobium
strongly segregates to γ/α boundaries and reduces ferrite growth kinetics because of a
solute drag effect. Accordingly, niobium additions suppress or delay the formation of
polygonal ferrite and pearlite, and promote the formation of acicular ferrite[4] or
bainite/martensite-austenite constituents and increases the bainitic hardenability[21].
The solubility temperature of niobium increases from about 1170 ºC in a typical
niobium steel to about 1250 ºC, in a Nb-Ti steel. In relation to titanium, the presence
of niobium leads to a higher amount of titanium precipitating in the low temperature
range, up to the dissolution of complex Ti-Nb carbonitrides. Consequently, it can be
concluded that high titanium concentrations present in Nb-Ti steels, exert a significant
influence on the grain coarsening behaviour[22]. Some strengthening will result from
NbN or Nb(C,N) precipitation in the austenite. However, the most effective
precipitation strengthening comes from NbC precipitation in ferrite [12].
The precipitation start temperature of Nb(C,N)0.87 is 1181 ºC. The solubility limit is
about 0.05% at 1150 ºC[14]. Higher niobium contents enable finish rolling to be carried
out at higher temperatures because undissolved NbC provides a finer austenite grain[23]
on soaking, which reduces the critical deformation for austenite to dynamically
recrystallise. The most rapid precipitation of niobium occurs between 800 and 950 ºC
in austenite. The intermediate finish rolling temperature (FRTs ) of 800~850 ºC leads
to lower soluble niobium levels than do lower or higher FRTs[24].
The content of niobium in line pipe steels is usually less than 0.05%. The main effects
of niobium in these steels are as follows:
(i) Much greater refinement[25,26] of ferrite grains due to increasing the inhibiting
effect of the growth of austenite grains with Nb%;
(ii) Dispersion hardening;
(iii) Reduction of pearlite content due to the binding of carbon in Nb(C,N); and
-8-
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 2 Micro-alloying elements and their effects on precipitation
(iv) Stronger inhibiting effect on recrystallisation[27] of hot-deformed austenite below
950 ºC [16], thereby raising the non-recrystallisation temperature (Tnr)[12] due to the
precipitation of Nb(C,N) during hot rolling[28].
2.3 Titanium
Titanium can refine ferrite grains, suppress recrystallisation of hot-deformed austenite
and reduce pearlite formation. Titanium will bind with free N during solidification of
the steel. TiN is very stable at high reheating temperatures (such as 1250 ºC), thus
retarding grain growth during the reheating process. TiN will also retard grain growth
between rolling passes. If it binds all of the free N, this will result in a significant
increase in niobium available in the ferrite to precipitate as NbC. The stoichiometric
ratio of Ti to N in TiN, is 3.4/1. Titanium can be used to control the sulphide inclusion
shape, thus preventing MnS stringer inclusions. The stoichiometric value for Ti/S in
TiS ratio is 1.5/1. Therefore, the titanium addition is calculated from[12]
Ti=3.4N+1.5S
(2.1)
The precipitation start temperature for TiN is 1633 ºC[15] while that of TiC is less than
that of TiN’ (<1000 ºC[5] for TiC). Therefore, the alloy needs sufficient titanium
content for TiC to form (Ti>0.4N)[14]. The stoichiometric ratio Ti/N is 3.4/1 but it may
change depending on several factors, such as the reheating temperature before hot
rolling and the content of other micro-alloying elements[29].
The main effects of titanium in high-strength low alloy steels are as follows:
(i) Refining of ferrite grains[30,31]: The optimum Ti/N ratio is close to 2 for Nb-free
and V-free steels[29]. Titanium at levels of 0.02 to 0.03% has a much greater
refinement effect on ferrite grain size. It inhibits the growth of austenite grains by the
difficult-to-dissolve TiN and Ti(C,N) during heating[14]. It inhibits grain growth most
effectively at 1150 ºC and an addition of 0.015% Ti has a much greater refinement of
austenite grains than the addition of the same amount of Nb[16]. Titanium in excess
after binding with free nitrogen will be available to bind with sulphur.
-9-
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 2 Micro-alloying elements and their effects on precipitation
(ii) Suppression of recrystallisation of hot-deformed austenite: Titanium has a
stronger inhibiting effect on recrystallisation at 1050 ºC than niobium. There is a
contradictory result, however, that 0.03% Ti does not have a substantial effect on the
recrystallisation of the austenite of low alloy steels due to differences in the
composition of these steels[16].
(iii) Dispersion hardening: Because TiC forms at lower temperatures, the titanium
content must exceed the bound content from TiN (Ti>0.4N).
(iv) Reduction of pearlite formation
2.4 Molybdenum
The addition of molybdenum in Nb-containing steels can improve transformation
hardening (increased volume fracture of acicular ferrite and M/A islands), grain
refinement and precipitation hardening. It is mainly the metastable coherent carbide
Mo2C that provides hardening in Mo-containing steels, although the formation of this
carbide requires relatively higher levels of addition. Molybdenum also greatly
suppresses or delays the formation of polygonal ferrite and pearlite[4] and promotes
the formation of acicular ferrite[32]. When the carbon content of the steel is sufficiently
low, martensite formation is avoided and fine structures of bainite and acicular ferrite
are formed during air-cooling after hot rolling. This will result in very good ductility.
Line pipe steels with a discontinuous stress-strain curve lose some yield strength as a
result of U-O pipe forming (U-bending and O-bending) and do not completely recover
this loss after expansion of the pipe. For steel containing-molybdenum, the
stress-strain curve of the as-rolled plate is continuous without an upper yield point.
This results in complete control of the detrimental Bauschinger effect during pipe
forming and contributes to an increase in the yield strength from plate to pipe. The
impact-transition temperature of traditional grades is raised by 10 to 20 ˚C after pipe
forming, but does not change for steel with a molybdenum content [33].
The addition of molybdenum in Nb-containing steels has the following purpose in line
- 10 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 2 Micro-alloying elements and their effects on precipitation
pipe steels:
(i)Transformation
hardening
due
to
increased
acicular
ferrite
(AF)
and
martensite/austenite (MA) formation;
(ii) Grain refinement; and
(iii) Precipitation hardening[2].
The addition of molybdenum provides a considerable strengthening advantage over
Nb-V steels[2].
2.5 Carbon
Carbon level in these steels is maintained below 0.06% for the following purposes:
(i) Lowering the carbon equivalent (CE) for improved weldability：The maximum CE
specification is 0.45;
(ii) High toughness;
(iii) Less micro- and macro-segregation;
(iv) More effective dissolution of niobium: increasing the strength of steels; and
(v) Less pearlite: improving toughness, formability, SSCC (sulphide stress corrosion
cracking).
2.6 Manganese
An increase in the manganese content prolongs the incubation time for polygonal
ferrite formation. The manganese addition extends the polygonal ferrite curve of
continuous cooling transformation (CCT) diagrams. Accordingly, acicular ferrite
structures can be obtained at a slower cooling rate with an increase in the manganese
content. The addition of manganese delays the precipitation of titanium and niobium,
and increases the solubility of NbC by decreasing the diffusivity of niobium in
austenite[4]. Manganese also increases the solubility for nitrogen in austenite
significantly.
- 11 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 2 Micro-alloying elements and their effects on precipitation
2.7 Copper, Chromium and Nickel
These elements can improve resistance to HIC (hydrogen-induced cracking) and
SSCC (sulphide stress corrosion cracking). Chromium and copper have little effect on
the mechanical properties of as-rolled plate. Chromium and copper promote the
formation of very fine precipitates when used in combination with niobium or
vanadium[33]. Steels with a small amount of copper tend to have a somewhat higher
corrosion resistance than alloys without copper. As little as 0.05 wt% Cu has been
shown to have a significant effect and, usually, the addition of 0.2% Cu can provide
increased resistance to atmospheric corrosion. Copper can increase the strength of
ferrite through solid solution strengthening and, with 0.5% or more, is precipitated as
elemental Cu-particles when the steel is aged within the temperature range of about
425 to 650 °C, thus providing a degree of precipitation hardening by virtue of ferrite
strengthening[34].
- 12 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 3 The controlled rolling process of line pipe steels
CHAPTER 3 THE CONTROLLED ROLLING PROCESS OF LINE PIPE
STEELS
Line pipe steels with high strength and good toughness properties are produced by
using a controlled rolling process. Controlled rolling appears to be vital to produce
these steels with optimum mechanical properties. The controlled rolling technique is
most effective with steels containing micro-alloying elements that provide a region
below which non-recrystallisation will take place during the finishing stages of hot
rolling. Micro-alloying additions increase this temperature above where complete
recrystallisation occurs and the recrystallisation of austenite below this temperature is
retarded during deformation.
3.1 Three stages of deformation for controlled rolling process
Controlled rolling may be divided in principle into three stages[11]:
(1) Deformation in the austenite recrystallisation temperature region;
(2) Deformation in the austenite non-recrystallisation region; and
(3) Deformation in the two-phase austenite-ferrite region.
This is illustrated in figure 3.1[11] along with the structural changes, which occur
during the controlled rolling process. Optimum mechanical properties of HSLA steels
can be obtained only by the careful control of microstructural changes in each stage of
controlled rolling. The principal variables are deformation temperature, amount of
deformation and strain rate.
- 13 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 3 The controlled rolling process of line pipe steels
Figure 3.1 Schematic illustration of the three stages of the controlled rolling
process[11].
3.1.1 Deformation in the austenite recrystallisation region
The reheating temperature and the solubility of micro-alloying precipitates have a
strong influence on the grain size of the initial austenite after reheating and the grain
size of the recrystallised austenite before the Tnr is reached. The effects of
micro-alloying elements depend on the reheating temperature[11]. The recrystallised
grain size decreases rapidly as a function of both an increased reduction per pass and
a decreasing deformation temperature. The smallest recrystallised austenite grain size
can be attained by using deformations above the critical amount required for the
initiation of dynamic recrystallisation during hot rolling.
3.1.2 Deformation in the non-recrystallisation region
It is commonly believed that micro-alloying elements retard the recrystallisation of
austenite either by solid-solution effects through solute drag or by the pinning effect
of strain-induced precipitation. Below the non-recrystallisation temperature,
recrystallisation of austenite grains is sufficiently suppressed so that rolling produces
deformed elongated austenite grains (also called ‘pancake’ grains) and deformation
- 14 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 3 The controlled rolling process of line pipe steels
bands within the austenite. Deformation bands also serve as nucleation sites for
ferrite[11]. With an increasing amount of deformation, the austenite grains become
more elongated, thereby increasing their length-to-thickness ratio, and the number of
deformation bands increases and their distribution becomes uniform, giving rise to a
fine and uniform final austenite structure and also to the final ferrite structure after
transformation on the run-out table.
3.1.3 Deformation in the (α+γ) two-phase region
Deformation in this two-phase region is generally difficult to control because there is
a differing deformation resistance in the two-phase region and this type of hot rolling
is generally not considered for line pipe purpose. A time-temperature diagram
indicating the microstructural processes during hot rolling is shown in figure 3.2[8].
Numerous metallurgical processes take place during hot rolling and subsequent
cooling, which determine the final grain size and thus the mechanical properties of the
strip.
Figure3.2 Metallurgical mechanisms during thermo mechanical hot rolling[8].
- 15 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 3 The controlled rolling process of line pipe steels
3.2 Reheating temperature and undissolved particles
The objective of reheating is to achieve uniformly heated slabs with full dissolution of
the micro-alloyed carbonitrides (Nb,V)(C,N). Continuously cast slabs are heated
within walking beam furnaces or pusher-type furnaces. The reheating temperature
depends on which micro alloying elements are present in the steel. It is generally from
1150 to 1250 ºC for Nb-containing steels, so that most of the alloy carbides and
nitrides (except for TiN) are dissolved in the steels in order to obtain maximum
dispersion hardening later. The soaking time is usually from 30 to 60 minutes for a
slab thickness of 26 to 110 mm[1, 10, 14, 35, 36].
At higher reheating temperatures, most micro-alloying elements will be in solution in
the austenite and will precipitate at lower temperatures both in the austenite and in the
ferrite as carbides, nitrides, or carbonitrides, thereby increasing the strength of these
phases (figure 3.3[24]). However, high reheating temperatures will allow considerable
grain growth to occur. At lower reheating temperatures, a larger proportion of
micro-alloying precipitates will remain undissolved in the austenite. These
precipitates will act as barriers to grain movement. Precipitates formed during
deformation will inhibit grain growth of the recrystallised austenite in the same
manner as undissolved precipitates[11].
Figure 3.3 Correlation between the increase in yield strength ΔYS and the content
of soluble Nb[24] (0.075% C, 043% Si, 1.58% Mn, 0.013% P, 0.005% S, 0.51% Ni,
0.021% Al and 0.106% Nb, and 0.077% C, 0.36% Si, 1.56% Mn, 0.017% P,
0.005%S, 0.47% Ni, 0.020% Al and 0.130% Nb).
- 16 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 3 The controlled rolling process of line pipe steels
Figure 3.4 Comparison between the calculated and measured quantities of
micro-alloying elements as precipitates[37].
Vanadium carbonitride passes completely into solid solution if the reheating
temperature is above 1050 ºC[16]. Niobium carbonitride is not fully dissolved in
austenite at 1250 ºC and it has a fine spherical shape[38]. Figure 3.4[37] shows that the
amount of Nb and (Nb+Ti) present within precipitates in three Nb-Ti bearing steels,
decreases with increasing heating temperature[37]. It also shows that there are some
undissolved Nb carbonitrides in the matrix even at ~1250 ºC. The measurements also
show that only 4 to 27% of Nb was dissolved in the austenite at 1100 ºC for the Nb-Ti
steel. This suggests that the mixing of niobium and titanium increases the stability of
the precipitates, so that the solubility of precipitates decreases in Nb-Ti bearing steels
through the formation of complex precipitates[35,37]. The complex (Nb+Ti) precipitates
are more stable than binary Nb carbonitrides so that the coarse precipitates are not
dissolved at 1250 ºC[35].
The austenite grain size increases and the quantity of undissolved precipitates
decreases with increasing reheating temperature. Accordingly, the grain size after
reheating is directly linked to the presence of micro-alloying elements, the dissolution
- 17 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 3 The controlled rolling process of line pipe steels
behaviour of carbonitrides and the heating time and temperature (figure3.5)[39]. The
higher the quantity of undissolved precipitates, the more stable the carbonitrides are.
So, the austenite grain size becomes finer due to the retarding effect of undissolved
particles on the recrystallisation of austenite. The complex micro-alloying precipitates
are more stable than binary precipitates. The reheating temperature may, therefore,
have to be higher for those steels with complex precipitates.
Figure 3.5 Change in austenite grain size during reheating process[39].
3.3 Rough rolling
The objective of rough rolling is to achieve the finest possible recrystallised austenite
grain size at the point where the Tnr is reached[6]. The rough-rolling stage is completed
at a temperature above about 1030 ºC[2], i.e. in the recrystallised austenite region for
the X60 to X80 ERW line pipe steels[9] where recovery and recrystallisation of
austenite grains take place[7]. Specially developed reduction schedules together with
TiN particle control to inhibit growth of recrystallised grains, are employed for this
purpose[6]. The transfer bar thickness is typically about 26 to 30 mm for line pipe coils
in the final strip thickness range of 5 to 10 mm and effectively sets the
non-recrystallisation finish rolling reduction level at about 70 to 80%[2, 6]. However,
for most line pipe steels the temperature of rough rolling varies from 1200 to 950
ºC[7,9-11]. Strain hardening also takes place besides recrystallisation in this stage[8].
- 18 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 3 The controlled rolling process of line pipe steels
3.4 Finish rolling
The objective of this stage is to accumulate rolling strain within the austenite grains so
that on subsequent ferrite transformation, ferrite nucleation sites are greatly multiplied
in number and a very fine ferrite grain size can be generated during controlled
cooling[6].
The stage is confined to the non-recrystallised austenite region in the lower
temperature range above Ar3[9]. The starting temperature for this stage is in the region
of about 1050 to 950 ºC and the finish rolling temperature is in the region of 850 to
950 ºC[6, 9 ,10]. The total finishing reduction is usually about 64 to 80% when the initial
slab thickness and finishing thickness are 110 and 18 mm, respectively[1]. While
further strengthening of the hot strip can be achieved when finish rolling is continued
into the intercritical region of austenite and ferrite, there is little resultant benefit to
the final pipe strength since the Bauschinger effect is increased in magnitude.
Lowering the finish rolling temperature below Ar3 was effective in one instance, in
significantly increasing the strength of the hot strip (figure 3.6[22]). However, almost
no further improvement in the corresponding pipe strength was recorded, suggesting
that two-phase rolling exacerbates the Bauschinger drop[2]. This “diminishing return”
effect led to the consideration of molybdenum additions as a means in which to
comfortably reach higher strength levels in X70 and X80 line pipe steels whilst
avoiding the use of an excessively high carbon equivalent in the interests of
preheat-free welding[2].
- 19 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 3 The controlled rolling process of line pipe steels
Figure 3.6 Influence of rolling conditions
Figure 3.7 Influence of rolling
on the mechanical properties of plate
conditions on the average ferrite
and strip of different thicknesses[24].
grain size[24].
The role of the finish rolling temperature on mechanical properties of the steel may be
attributed to the acceleration of transformation to ferrite and the refinement of grains
by the decrease in the finishing temperature[9]. Therefore, the decrease in finishing
temperature will be beneficial to both strength and ductility as shown in the following
regression equations determined for a steel of composition: 0.07%C, 0.25%Si,
0.9%Mn, 0.046%Nb, 0.04%V, 0.015%Ti and 40ppm N[9]:
YS=0.508Ts-0.231Tf-0.334Tc+1.905Vc+323.6
R2=0.94
(3.1)
EL=-0.002Ts-0.064Tf-0.086Tc+0.325Vc+121.8
R2=0.98
(3.2)
where EL is the elongation in %, Ts is the start rolling temperature (ºC), Tf is the
finish rolling temperature (ºC), Tc is the finish cooling temperature (ºC), Vc the
cooling rate (ºCs-1) and R2 is the regression coefficient. In such an approach, however,
one needs to strictly control the lowering of the finish rolling temperature as higher
mill loads are encountered. It has been reported[9] that when Tf is above a certain
lower limit temperature, both the yield strength and the toughness of the steel increase
with a decrease of Tf. If Tf is below this critical temperature, a banded microstructure
appears which greatly reduces toughness.
- 20 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 3 The controlled rolling process of line pipe steels
Figure 3.7
[24]
shows the effect of finish rolling temperature on the final ferrite grain
size, showing that the ferrite grain size can be reduced at lower finish rolling
temperatures[24, 36]. Some phenomena that appear are strain hardening, precipitation of
micro alloying carbonitrides and/or γ-α transformation[8].
3.5 Heavy reduction
A heavy reduction of 60 to 85% in total, increases the yield strength without adversely
affecting the Charpy brittle to ductile transition temperature. Moreover, a heavy
reduction followed by rapid cooling remarkably increases the yield strength and
decreases the Charpy brittle to ductile transition temperature (figure 3.8
[10]
). The
heavy reduction on its own results in a slight grain refinement and the rapid cooling
again increases the fraction of the bainitic microstructure. However the two together,
when the heavy reduction is followed by rapid cooling, the microstructure with the
increased fraction of bainite becomes grain-refined[10]. The result also showed that the
volume fraction of M/A constituents decreased on increasing the deformation for a
steel containing 0.03%C, 0.05%Nb and 0.024%Ti (figure 3.9)[13].
Figure 3.8 Improvement in yield strength
Figure 3.9 Plot of the volume fraction of
and Charpy ductile to brittle transition
M/A constituents versus the degree of
temperature[10].
prior deformation[13].
- 21 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 3 The controlled rolling process of line pipe steels
The total or cumulative reduction of HSLA steels in the hot rolling process is
generally between 67 to 84%[1,6,9, 10,14,35,38], while the maximum cumulative reduction
is up to 95%[40]. The number of passes of thermo-mechanically controlled processing
are generally between 4 to 6 with a maximum of 7 to 9 for an ~0.06%C steel[41]. Table
3.1 shows the typical reduction and passes of hot rolling for line pipe steels.
Table 3.1 Reduction/pass of hot rolling
Reduction, %
Steel
No
Pass 1
Pass 2
Pass 3
Pass 4
Pass 5
1
16
23
29
27
2
--
--
--
--
--
--
3
15
25
25
25
25
25
4
21
21
31
30
30
5
13
30
30
30
6
33
40
40
7
16
23
8
25
33.3
--
Pass 6
Rough
Finishing
rolling
rolling
--
Ref.
Total
67
38,42
64
84
1
--
--
80
10
25
--
--
84
10
30
25
--
--
84
10
35
--
--
--
--
84
10
29
27
--
--
--
--
67
35
40
33.3
--
--
--
--
73
9
51
3.6 Strip thickness
The thickness of the final strip has a significant effect on the strength of these steels,
especially on the ratio of yield strength to ultimate tensile strength or YS/UTS.
Experimental results on X80 line pipe steels showed that the yield strength and tensile
strength increased and the YS/UTS ratio also increased from 0.86 to 0.95 with
decreasing strip thickness from 9 to 3 mm for an 0.07% C, 0.19% Si, 1.53% Mn,
0.045% Nb, 0.012% Ti and 0.20% Mo steel (figure.10)[6]. A similar relationship
between the thickness and the YS/UTS ratio that dropped from 0.89 to 0.76 as the
thickness was increased from 13.5 to 50 mm, was found in a X80 steel of composition:
0.075% C, 0.43% Si, 1.58% Mn and 0.106% Nb)[24]. The ratio also dropped from 0.80
- 22 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 3 The controlled rolling process of line pipe steels
to 0.73 as the thickness increased from 10.3 to 32mm for a steel with composition of
0.09% C, ~0.20% Si, ~1.20% Mn, ~0.043% V, ~0.04% Nb and 0.007% Ti[38].
Figure 3.10 Strip and pipe properties for various X80 pipe sizes)[6]. (Steel G: 0.075%
C, 1.59% Mn, 0.31% Si, 0.057% Nb, 0.22% Mo, 0.013% Ti and 0.006% N; Steel N:
0.070% C,1.53% Mn, 0.19% Si, 0.045 Nb,0.20% Mo, 0.012% Ti and 0.0045% N.
3.7 Cooling rate after finish rolling
The cooling rate (Vc) and the finishing temperature (Tc) of the accelerated cooling
process after finish rolling are important parameters to achieve optimum strength of
thermo-mechanically controlled processing for line pipe steels. Controlling the Tc and
Vc may lead to the control of precipitation of carbonitrides during accelerated
cooling[9]. The availability of high cooling rates immediately after finish rolling
provides greater microstructural adjustment possibilities, especially for the case of
Mo-Nb steel types. The possibilities for adjustment of the relative volume fraction of
polygonal and acicular ferrite (and grain size) through the setting of the cooling rate
and coiling temperature are schematically evident in figure 3.11[2]. The rapid cooling
increases the fraction of bainitic structure or M/A islands or acicular ferrite. This is
because polygonal and pearlite transformations are governed by slow diffusion and
have some difficulty in fully transforming during accelerated cooling. The austenite
may remain as retained austenite or transform to an acicular ferrite structure because
- 23 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 3 The controlled rolling process of line pipe steels
the transformation to acicular ferrite is a mixed mode of shear and diffusion[4,9,11,43].
The M/A islands are carbon-rich due to the replacement of pearlite by these small
islands. The M/A islands also refine the ferrite grain size structure. The ferrite grain
size becomes finer when a heavy reduction is followed by rapid cooling (figure
3.12[10]) because the nucleation rate increases and there is not enough time for growth
of nuclei of ferrite grains due to increasing under-cooling at a rapid cooling rate. The
heavy reduction during finish rolling results in more crystal defects and these provide
the nucleation sites for the γ→α transformation.
Figure 3.11 Schematic representation
Figure 3.12 Influence of reduction and
of the cooling pattern on the run-out
cooling rate on the ferrite grain
table of a Hot Strip Mill[2].
diameter[10].
The cooling rate after controlled rolling also has a significant effect on the degree of
precipitation[11]. With slower cooling rates the coarsening of precipitates makes them
less effective for strengthening while with faster cooling rates, no precipitation occurs
and the micro-alloying elements remain in solution. Correctly designed accelerated
cooling leads to fine precipitation of Nb(C,N) and also provides ferrite grain
refinement[4]. Roberts[5] indicated that a higher yield strength can be obtained at a low
carbon equivalent by employing accelerated cooling. In general, an increasing degree
of precipitation hardening as the cooling rate is raised, has no detrimental effect on
fracture toughness because of the simultaneous refinement of the microstructure and
- 24 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 3 The controlled rolling process of line pipe steels
the replacement of pearlite by small islands of bainite.
An increase in cooling rate from 15 to 30 ºCs-1 in a Nb-bearing steel (0.1% C, 0.20%
Si, 0.09% Mn, 0.046% Nb and 0.0018% N) enhanced the yield strength by 20 to 30
MPa[10]. This leads to a guideline equation (3.1) above. The yield strength of a line
pipe steel can increase by an average of about 2 MPa for every increase of 1 ºCs-1 in
the cooling rate[9].
The cooling rate on run-out tables for line pipe strip rolling in the laboratory is
generally between 10 to 60 ºC s-1[9,10,44].
3.8 Finish temperature of accelerated cooling
The finish temperature of accelerated cooling (Tc) is important in controlling the
precipitation of carbonitrides during this cooling. At a temperature just above 500 ºC,
V4C3 which is the main vanadium carbide in these steels, has the highest nucleation
rate and its contribution to precipitation strengthening is most prominent[45]. Therefore,
within a certain temperature range just above 500 ºC, a lower Tc is beneficial to both
strength and ductility as predicted by formulae (3.1) and (3.2). However, Tc cannot be
decreased too low, because it will be detrimental to the fracture toughness of the
steel[9].
The pearlite transformation will also be retarded as the finish temperature of
accelerated cooling is decreased. Optical microscopy showed that some pearlite is
formed when Tc is at 781 ºC, only a little pearlite is formed at 660 ºC and no pearlite
is formed below 580 ºC in an 0.08% C, 0.007% Nb, 0.009% Ti steel[39]. When
lowering the Tc , ferrite grains become finer and the volume fraction of acicular ferrite
increases. Below 580 ºC, however, the ferrite grain size remains largely unaffected
because the ferrite transformation is already complete (figure 3.13)[39].
- 25 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 3 The controlled rolling process of line pipe steels
Figure 3.13 Effect of finish temperature
Figure 3.14 Comparison of mechanical
of accelerated cooling on the average
properties between the observed and the
ferrite grain diameter[39].
calculated strengths of micro-alloyed
steels[39].
The relationship between Tc and the tensile strength is shown in figure 3.14[39]. It
indicates that the yield strength and ultimate tensile strength increase with decreasing
Tc until about 580 ºC. But almost no further change in strength is found at Tc < 580
ºC.
The finishing temperature of accelerated cooling (Tc) is, therefore, maintained
generally between 550 and 670 ºC[7,8,10,44] although sometimes it may even be as high
as 800 ºC (~0.10% C)[36] or even below 500 ºC [9,10].
From the above, it appears that the optimum mechanical properties for a line pipe
steel can be obtained when the reheating temperature Ts is approximately 1100 ºC, the
finishing temperature Tf is about 890 ºC, the finishing cooling temperature Tc is about
520 ºC and the cooling rate Vc from the Tf to Tc is about 30ºCs-1[9].
- 26 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 3 The controlled rolling process of line pipe steels
3.9 Coiling temperature
The coiling temperature will influence the effectiveness of Nb-precipitation but
especially that of V(C,N) precipitation in the ferrite, thus controlling the precipitation
strengthening of the steel. The coiling temperature requires careful control if the
maximum degree of precipitation strengthening is to be attained. High coiling
temperatures result in few and coarse precipitates, which will add very little to the
strength, while lower temperatures should be sufficiently low for a fine precipitation,
adding to the strength depending on the levels of micro-alloying additions, especially
the vanadium and free N in solution. Low coiling temperatures will result in finer
ferrite grains because of less ferrite grain growth taking place after transformation,
adding to the strength and fracture toughness. Lower finish cooling and coiling
temperatures, will suppress pearlite formation and microstructural banding[46]. The
temperature, however, should not be too low so that the volume fraction of TiC is
reduced by too slow diffusion[5]. Too low a coiling temperature results in low strength
because precipitation may be retarded. The optimum coiling temperature for a given
composition should preferably be determined experimentally.
Production experience with Nb-Ti-, Nb-V-Ti- and Mo-Nb-Ti-type line pipe steels
shows somewhat different sensitivities of coiling temperature with strip strength
properties as indicated in figure 3.15[2]. There was little effect in the relationship
between coiling temperature and yield strength for these micro-alloying steels. The
same tendency is that the yield strength increases with decreasing coiling temperature.
Decreasing the coiling temperature has no effect on the yield strength below about
620 ºC for a Ti-Nb-V steel with composition 0.085% C, 0.050% V, 0.045% Nb and
0.013% Ti[2].
- 27 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 3 The controlled rolling process of line pipe steels
Figure 3.15 Effect of coiling temperature on the YS of hot rolled strip for various
types of alloys[2].
3.10 Non-recrystallisation temperature (Tnr)
The non-recrystallisation temperature (Tnr) is also an important parameter for the
controlled rolling process of line pipe steels. The rough-rolling process is generally
carried out above the Tnr temperature, which is associated with the finishing
temperature of the rough-rolling stage. During finishing below the Tnr, a total of at
least 0.8 strain should be applied for effective grain refinement (effective controlled
rolling)[46].
A coarse, polygonal ferrite plus pearlite microstructure can be ascribed to insufficient
strain below Tnr, coupled with higher coiling temperatures, while a coarse acicular
ferrite can be ascribed to insufficient strain below Tnr coupled with lower coiling
temperatures[46].
3.10.1 Effect of alloying elements
Micro-alloying elements have some effect on the Tnr. Niobium binds with free
nitrogen and precipitates as niobium nitrides at higher temperatures, e.g. during rough
rolling. Thus, most of the niobium then becomes ineffective in retarding austenite
recrystallisation during finish rolling at lower temperatures. Titanium is, therefore,
often used in combination with niobium to tie the free nitrogen in Nb-Ti
- 28 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 3 The controlled rolling process of line pipe steels
micro-alloyed steels and this generally forms (Ti,Nb)N precipitates. Thus, titanium
will increase the effect of the niobium on the Tnr. Thus titanium and niobium
micro-alloying elements will raise the Tnr (inhibiting recrystallisation)[47], and cause
more effective grain refinement. Titanium appears to accelerate the precipitation of
Nb(C,N), and therefore, contributes to the retardation of recrystallisation. Titanium is
only 0.31 times as effective as niobium in retarding recrystallisation and aluminium
0.15 times[48]. The Tnr temperatures at a strain rate of 3.63 s-1 were found to be 918
and 937 ºC for a V-micro-alloyed (0.075% V) and a V-Ti (0.055% V+0.024% Ti)
micro-alloyed steel, respectively, while the Tnr was about 1042 ºC for a Nb (0.026%)
steel deformed at the same strain rate[47]. The Tnr decreases with decreasing
super-saturation ratio or the [Nb][C] solubility product[48,49]. The addition of Mo, Nb,
and (Mo+Nb) increases the Tnr in this order[48]. The Tnr is increased by boron
additions because it accelerates the precipitation kinetics of Nb(C,N), but a maximum
in the Tnr at a boron concentration of 48 ppm occurs. The results showed that
increasing the boron content beyond 48 ppm leads to a decrease in the Tnr[48]. This
decrease in the Tnr was possibly due to an acceleration in the rate of coarsening of the
Nb(C,N), as precipitate coarsening of Nb(C,N) is accelerated by the presence of
boron[48] and such coarsened particles lose their effectiveness with regard to retarding
the recrystallisation. When less than 48 ppm boron was added, the equilibrium
segregation of boron to austenite boundaries and the acceleration of Nb(C,N)
precipitation can both contribute to increasing the Tnr.
3.10.2 Effect of the controlled rolling process
At the same solution reheating temperature, the non-recrystallisation temperature was
found to decrease with an increase in the strain rate from 1.09 to 3.63 s-1. This is
because a higher strain rate and strain will cause a consequent increase in the density
of dislocations, which are beneficial to the onset of recrystallisation through
increasing its driving force. Increasing the pass strain and strain rate was found to lead
to
a
decrease
in
the
Tnr
for
- 29 -
Nb-containing,
(Nb+Ti)-containing,
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 3 The controlled rolling process of line pipe steels
(Nb+Ti+Mo)-containing and B-containing steels. The strain rate had a smaller effect
on the Tnr than the pass strain. These relationships were found to be as follows[48]:
Tnr ∝ ε −0.12
(3.3)
Tnr ∝ ε& -0.01
(3.4)
The effect of inter-pass time on the Tnr is complicated by the transition from a solute
drag effect to retardation by precipitation. In a short inter-pass time range (generally
less than 12.5 seconds), the Tnr decreases with increasing inter-pass time because
niobium
solute
drag
appears
to
be
responsible
for
the
retardation
of
recrystallisation[48]. The full Tnr relationship to process conditions can be expressed
as:
Tnr =A ε −0.12 t ip
−0.1
ε& -0.01
(tip<12.5 s)
Tnr =(Alog[Nb]eq+B) ε −0.12 t ip
−0.1
or
(3.5)
ε& -0.01
(3.6)
where [Nb]eq is the equivalent niobium content given by [Nb]eq=[Nb+0.31Ti+0.15Al],
A=88.1 ºC per wt% and B=1156 ºC.
At long inter-pass times, however, the Tnr increases with increasing inter-pass time.
This is because the volume fraction of precipitates that are formed is time dependent,
so that the retardation effect increases with the holding time before precipitate
coarsening sets in. The effect of the deformation parameters on the Tnr is then as
follows:
Tnr =A′ ε −0.12 t ip
0.04
ε& -0.01
(12.5 s< tip<30 s)
Tnr =(A′log([Nb][C])+B′) ε −0.12 t ip
where A′=63.5 ºC per wt%,
0.04
ε& -0.01
or
(3.7)
(3.8)
B′=885 ºC
Strain has the following effect on the Tnr[47]:
Tnr=Ts−Pεa ε& bDc
(3.9)
where Ts is the solution temperature, ε is the strain, ε& is the strain rate and D is the
austenite grain size. From the equation above, it appears that a smaller grain size
- 30 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 3 The controlled rolling process of line pipe steels
contributes to raising the Tnr.
The effect of micro alloying elements on the Tnr is described by the following solute
retardation parameter (SRP):
⎛ t
SRP = log⎜ x
⎜t
⎝ ref
⎞⎛ 0.1 ⎞
⎟⎜
100
⎟⎝ X ⎟⎠
⎠
(3.10)
where tx is the time required for the start of static recrystallisation in the steel
containing the element x, tref is the equivalent time for a reference plain carbon steel
and X is the alloy content of x in at%. The SRP of V, Mo and Nb is 10, 37 and 409,
respectively for a static condition or 8.4, 25 and 135, respectively for a dynamic
condition with ε& =2 s-1 at 1000 ºC[50 ].
- 31 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 4 Microstructure and mechanical properties
CHAPTER 4 MICROSTRUCTURE AND MECHANICAL PROPERTIES
4.1 Acicular ferrite
4.1.1 Nucleation and growth of acicular ferrite
Microstructures with a significant proportion of acicular ferrite present an optimised
combination of mechanical properties if compared with mainly bainitic structures. It
is well documented that acicular ferrite formation is enhanced by the presence of
non-metallic inclusions in studies in weld pools[51-64], low carbon steels[65] and
medium carbon forging steels[66-69], and is characterized by elongated grains that are
“chaotically” arranged[4]. These second-phase particles act as point sites on which the
intragranular nucleation[70,71] of ferrite units develops. There may also be some M/A
islands present with a high dislocation density[11]. Acicular ferrite is a non-equiaxed
structure phase with an interior that contains a dense substructure of dislocations[4,72].
The carbon content in the M/A islands is higher than that in the surrounding matrix.
Accordingly, the M/A islands are carbon-enriched, whose formation may be attributed
to the partitioning of carbon during the transformation to acicular ferrite and the
post-transformation of carbon-enriched austenite. When the specimen is deformed in
the non-recrystallisation austenite region, high densities of substructure and
dislocations will be formed in the austenite, which increase the nucleation rate of
acicular ferrite, impedes the growth of the coherent and/or semi-coherent γ/α
interfaces and accelerates diffusion of carbon to these γ/α interfaces, which leads to
carbon-enriched austenite. During the accelerated cooling after finish rolling and
followed by the coiling process, part of the carbon-enriched austenite transforms to
martensite and the retained austenite coexists with the martensite[73]. The
transformation model is, therefore, a mix of diffusion and shear transformation[4,9,11,43].
The start temperature of the transformation to acicular ferrite is slightly higher than
that of an upper bainite[4,11].
Acicular ferrite always has an orientation relationship with the austenite grain, such
- 32 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 4 Microstructure and mechanical properties
that one of its close packed {110}AF planes is nearly parallel to the close-packed
{111}γ plane of the parent austenite. Within these parallel planes, a close-packed
−
−
<11 1 >AF direction of the acicular ferrite is found to be near to a close-packed < 101>γ
direction of the austenite[74]. This demonstrates that the growth of acicular ferrite
occurs by a displacive transformation and its growth, therefore, takes place without
carbon diffusion. The excess carbon in ferrite is probably rejected into the austenite
soon after nucleation. Acicular ferrite plates are never found to grow across austenite
grain boundaries and this is also consistent with the displacive transformation
mechanism, since the necessary co-ordinated movements cannot be sustained across
austenite grain boundaries. TEM work has revealed that the ferrite units belonging to
the same sheaf have the same crystallographic orientation in most cases[75].
As regards the carbon concentration of acicular ferrite structures during
transformation, experiments and thermodynamic theory have demonstrated that the
growth of acicular ferrite is diffusionless with the ferrite inheriting the chemical
composition of the parent austenite. The excess carbon in the acicular ferrite is
rejected into the retained austenite after transformation and can apparently occur
within a few seconds[74].
4.1.2 Two types of acicular ferrites: Upper and lower acicular ferrite
There are two different microstructural morphologies of acicular ferrite (AF) in
medium carbon micro-alloyed steels, depending on the isothermal treatment
temperature[75]. One is upper acicular ferrite, the typical acicular ferrite with an
interlocked microstructure (plate morphology) that is formed at high isothermal
treatment temperatures of typically 450 ºC. The secondary, is of acicular plates of
ferrite that had nucleated at the interface between the primary ones and the austenite
and are inclined at a high angle with respect to the substrate unit.
Another sheaf morphology of lower acicular ferrite, is composed of packets of plates
- 33 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 4 Microstructure and mechanical properties
following the same growth direction at lower isothermal treatment temperatures of
typically 400 ºC. A significant change in the morphology of the acicular ferrite is,
therefore, clearly apparent with a lowering of the formation temperature. It is
observed that the nucleation of the primary plates takes place intragranularly on the
same second-phase particles. These significant differences between the morphologies
of the two types of acicular ferrite can be distinguished in the early stages of the
transformation. At a high nucleation temperature of 450 ºC, single plates form at
second phase particles while at temperatures lower than 400 ºC, individual parallel
platelets are formed with residual phases in between them.
There may be two reasons for the formation of parallel AF units at low temperatures.
Firstly, the lower stability of the austenite close to the tip of the ferrite plate and
secondly, the strain field produced by the invariant plane-strain shape transformation,
both favour the formation of the same variant as that of the primary plate at these sites.
Further growth of these subunits seems to be possible parallel to the primary plate,
leaving a thin layer of carbon-enriched retained austenite between the different
subunits. Afterwards, these regions of austenite lead to the precipitation of interlath
cementite between the ferrite plates[75].
The autocatalytic formation of new plates of acicular ferrite is expected to depend
strongly on the carbon concentration profile of the parent austenite ahead of the
interface with the primary AF plates. This concentration profile will become more
pronounced as the transformation temperature decreases and the diffusion in the
austenite of the carbon rejected from the ferrite, becomes slower. Close to the acicular
ferrite tips, the carbon enrichment of the austenite could be low enough to allow the
transformation to proceed, leading to the formation of elongated sheaves. At high
transformation temperatures, the diffusion of carbon is more rapid and plate formation
on interfaces is more likely[75].
- 34 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 4 Microstructure and mechanical properties
4.1.3 Effect of the hot rolling process on acicular ferrite formation
Hot deformation will promote the subsequent formation of acicular ferrite. This is
because high densities of substructure and dislocations are formed in the austenite
during deformation of austenite in the non-recrystallisation region, which increases
the nucleation sites for acicular ferrite and promotes the acicular ferrite
transformation[4]. The temperature range for the nucleation of acicular ferrite moves
slightly towards higher temperatures and shorter times with hot deformation of
austenite if compared to an equivalent austenite without hot deformation. The growth
of acicular ferrite, however, is retarded in plastically deformed austenite[76]. With an
increase in the cooling rate after hot rolling, the fraction of polygonal ferrite decreases
and the fraction of acicular ferrite increases in volume and the grain size of the ferrite
becomes smaller[73].
4.2 Acicular ferrite and bainite
Bainite forms typically at temperatures between pearlite and martensite and the
transformation model is also a displacive mechanism. There are three kinds of
microstructure: upper bainite, lower bainite and granular bainite. Carbides precipitate
in-between the laths within upper bainite (which often results in a lower toughness),
while carbides are finely distributed within the bainite sheaths at a fixed orientation
within lower bainite together with some minor interlath formation of carbides also
here. Both forms of carbides have a specific orientation relationship between the
bainite lath and the carbides in lower bainite[11].
Bainitic ferrite is very different from acicular ferrite in its shape. It possesses largely
parallel sheaves of ferrite with a lath-like structure with some granule-like or rod-like
cementite particles alongside or within the laths and along the prior austenite grain
boundary network that can be seen clearly. The ferrite in bainite nucleates at the
austenite grain boundaries[66,77-80], forming sheaves of parallel plates with the same
- 35 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 4 Microstructure and mechanical properties
crystallographic orientation with respect to the parent austenite. Interlath carbide
particles in upper bainite precipitate from the carbon-enriched retained austenite
trapped in-between the platelets, or in the lower bainite, from within the
supersaturated ferrite within the bainite lath[74].
Acicular ferrite micrographs, on the other hand, have been studied[81] in which the
acicular ferrite transformation starts through the nucleation of the primary plates at
non-metallic inclusion particles[33,52,54,56,82,84] and progresses by the formation of a
new generation of secondary plates of ferrite nucleated at the interfaces of the
austenite/AF primary plates[78]. Therefore, inclusions play an important role in the
formation of acicular ferrite in low alloy welded metals[85,86] because they provide
preferential sites for the nucleation[87-89] of the AF. The acicular ferrite matrix is
characterized by a fine non-equiaxed ferrite or interwoven nonparallel ferrite
laths[66,73,78,90], which have various sizes distributed in a random manner, very often
described as a “chaotic arrangement” of plates showing fine-grained interlocking
morphologies[78]. The prior austenite grain boundary network is completely eliminated
and some fine M/A island constituents are scattered throughout the matrix[4]. This is
due to the partitioning of carbon near the austenite/ferrite interface during the growth
of acicular ferrite. The carbon content in the austenite will be increased and
accordingly, the austenite’s stability will be increased. As a result, the partial austenite
that is carbon-enriched, remains and transforms to martensite during the subsequent
cooling process, resulting in the M/A island constituent.
4.3 Mechanical properties of line pipe steel
4.3.1 The ratio of yield strength to ultimate tensile strength (YS/UTS)
A low YS/UTS ratio a very important parameter in the API specifications for line pipe
steels as a high work hardening rate is required for this application. The lack of strain
hardening in high YS/UTS steels means that there is a reduced potential for strain
redistribution in thinned areas (thinned by corrosion or weld dressing) during service.
- 36 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 4 Microstructure and mechanical properties
A high strength is, of course, required for line pipe to transport oil or natural gas at
higher pressures. The American Petroleum Institute (API) specifies a YS/UTS ratio
not greater than 0.93 for an application involving pipelines. The 11 mm line pipe strip
steel currently produced by MITTAL Steel (South Africa) tends to have a slightly high
YS/UTS ratio of typically 0.93. This ratio is affected by the microstructure of the steel
and an optimised microstructure (such as acicular ferrite[32]) is, therefore, beneficial in
achieving a lower YS/UTS ratio by carefully controlling the hot rolling, cooling and
coiling schedules. This ratio is also an important issue in the development of higher
grades of line pipe steels. The specification of line pipe steels of the America
Petroleum Institute (API) is shown in table 4.1.
Table.4.1 Specification of line pipe steels of API[12]
Grades of
Minimum yield
Minimum tensile
Minimum
Maximum
steels
strength (MPa)
strength (MPa)
Elongation
YS/UTS
(%)
X65
448
530
20.5
0.93
X70
482
565
19
0.93
X80
551
620
17.5
0.93
X90
601
650
17.5
0.93
Some results showed that there is a slightly higher volume fraction of about 7% of
martensite/austenite (M/A) constituent with a higher finish rolling temperature and,
therefore, a more rounded stress-strain curve and a higher strain-hardening rate on this
curve[1]. It is beneficial to lower the YS/UTS ratio and that author found that when the
finish rolling temperature was 720 ºC the ratio was 0.69 and at 780 ºC, the YS/UTS
was 0.65 and the volume fractions of M/A constituents were 4.6 and 7.0%
respectively for a steel with composition 0.057% C, 0.27% Si, 2.04% Mn, 0.040% Nb,
0.112% Ti and 0.001% B[1].
- 37 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 4 Microstructure and mechanical properties
4.3.2 Toughness
The Charpy toughness specification for the control of fracture initiation normally does
not prove too onerous a requirement for higher grade line pipe. Toughness is also
important for line pipe steels. Higher toughness can be obtained by lowering the
carbon content and refining the ferrite grain size[38]. Niobium can improved Charpy
toughness[25,26] at lower finishing temperature below 980 ºC[91]. A lower vanadium
content is also useful to increase the toughness[38]. Acicular ferrite microstructures
resulted in an improvement of the Charpy toughness with no deterioration of the
strength[65,92] whereas bainite resulted poor toughness[66].
4.3.3 U-O pipe forming and Bauschinger effect
During U-O pipe forming, the plate materials are subjected to different cyclic strains,
depending on the location along the circumference and in the wall. Typical examples
of cyclic strain at locations 180˚ and 30˚ from the longitudinal seam are illustrated in
figure 4.1[93].
At the 180˚ location where mill tensile test specimens are usually taken, the outer
layer receives a tensile force during U-bending, a compressive force during
O-bending, a compressive force during shrinking, and a tensile force during
expansion. At the same time, the inner layer is subjected to a cyclic strain of
compression, tension, compression and tension.
The total process from U-bending to expansion is not a simple work-hardening
process but is actually very complex. On the other hand, during the flattening of a
curved pipe section for tensile specimens, the Bauschinger effect and work hardening
occur in the outer and inner layer respectively.
- 38 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 4 Microstructure and mechanical properties
Figure 4.1 Schematic stress-strain curves for the outer (top) and inner (bottom)
material during the U-O pipe forming process, with (left) at 180º and (right) at 30º
from the welding line[93].
The Bauschinger effect is a characteristic material behavior that is highly dependent
on testing conditions[94]. It results in the lowering of a proof-stress value after a
- 39 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 4 Microstructure and mechanical properties
previous single uniaxial initial loading in the opposite direction. When hot rolled strip
is converted into ERW line pipe, the pipe forming and sizing strains can significantly
modify the pipe yield strength by virtue of the Bauschinger phenomenon. Various
steels have different responses to the Bauschinger effect due to different stress-strain
curves. Steels with a yield plateau have a Bauschinger strain equal to that
corresponding to the strain equal to the yield elongation[95].
The Bauschinger effect is largely controlled by the carbon content and, to a
considerably smaller degree, by the manganese content (figure 4.2). Grain size
appears to have a minor influence, while the influence of residual-stress conditions is
strong (figure 4.3)[94].
Figure 4.2 The change of the Bauschinger Figure 4.3 The Bauschinger effect in
effect factor with carbon and manganese
micro-alloyed steel. The upper two curves
content[94].
are for steels with 0.2% C, 0.4% Mn,
unalloyed or alloyed respectively with Al,
V or Nb. The lower two curves are for
low-pearlite steels with less than 0.1% C,
2% Mn, and alloyed with Mo, Nb and
Ti[94].
- 40 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 4 Microstructure and mechanical properties
Work-hardening and the Bauschinger effect occur during the pipe-forming process
and during the flattening of the tensile test pieces before the tensile test. In the
pipe-forming process, work hardening by pipe expansion is more important than the
Bauschinger effect, while the reverse is true during sample flattening. The
ring-expansion test is used in measuring work hardening. The pipe after forming, has
a considerably higher yield strength than the plate, which indicates that work
hardening has taken place during pipe forming[96].
- 41 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 5 Background of current South Africa line pipe production
CHAPTER 5 BACKGROUND OF CURRENT SOUTH AFRICA LINE PIPE
PRODUCTION
5.1 Line pipe steel composition of Mittal Steel (South Africa)
The chemical composition of the current Mittal Steel (SA) line pipe steel is provided
in table 5.1.
Table 5.1 Typical chemical composition of the current 11 mm line pipe steel of Mittal
Steel, (wt%)
C
Si
Mn
P
S
Cr
Ni
Mo
Cu
Al
0.066
0.258
1.583
0.011
0.004
0.021
0.007
0.001
0.007
0.037
V
Nb
Ti
Sn
B
0.062
0.037
0.017
0.001
0.0002
- 42 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 5 Background of current South Africa line pipe production
5.2 Parameters of the hot rolling process at Mittal Steel (SA)
The parameters of hot rolling of line pipe steels at Mittal Steel are shown in table 5.2.
Reheating
tempe`rature, ºC)
Table 5.2 The parameters of the hot rolling process at Mittal Steel
Pass No
R1
R2
R3
R4
R5
R6
F4
F5
F6
1552
1150
1076
1053
1042
1008
938
915
896
1050
1008
969
915
896
879
Force (tons)
in
Temperature
(ºC)
1200
1150
out
Gauges
(mm)
F1
F2
1860
F3
in
240
195
160
120
85
60
40
28.87
21.02
15.76
13.51
out
195
160
120
85
60
40
28.87
21.02
15.76
13.51
11.70
0.21
0.20
0.29
0.34
0.35
0.40
0.32
0.32
0.29
0.15
0.14
16.1
22.6
20.9
26.2
1.8
2.7
3.3
3.9
Strain/pass, ε
Total strain
1.79
1.22
Strain rate
9.4
(s-1)
Roll speed,
~1.5
V (ms-1)
Inter-pass time, t, (s)
Inter-pass cooling
rate (ºCs-1)
~1.5
~1.5
~1.5
~1.5
10
10
10
10
10
3
3
3
3
3
2.7
1.8
1
1.5
40 ºCs-1 —for 6mm of the final thickness of strip
Cooling
finishing
rate
(ºCs-1)
after
20 ºCs-1 —for 11.5mm
50 ºCs-1 —for 5mm
~2.4%
t/D (Thickness/Diameter)
NB: F3-dummy for rolling
- 43 -
1.3
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 5 Background of current South Africa line pipe production
5.3 Typical microstructures and existing developments within Mittal Steel for line
pipe steel
The optical microstructure of the current 11 mm wall thickness line pipe steel for
Mittal Steel is shown in figure 5.1. This is the alloy that was used for a major part of
the pipe-line for large-scale gas transportation from Mozambique to Secunda in a 2.5
meters diameter pipe line. The microstructure is a mixture of polygonal ferrite,
acicular ferrite and pearlite.
Figure 5.1 The optical microstructure of cast #521031, Mittal Steel line pipe
Smaller thin walled (about 6 mm thickness) pipelines may be used in future within
South Africa for smaller scale gas distribution and reticulation to consumers. The
current steel produced by Mittal Steel tends to have a slightly high YS/UTS ratio of
0.93, which is on the maximum limit of the specification of the American Petroleum
Institute (API) of 0.93 for an application involving pipelines. The current line pipe
steel consists typically of about 0.06% carbon with micro-alloying elements of
titanium, niobium and vanadium (generally less than 0.05% for each) and is produced
either via the Electric Arc Furnace (EAF) or Basic Oxygen Furnace (BOF) route.
- 44 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 5 Background of current South Africa line pipe production
5.4 The hypothesis for this study
The objective of this study is to establish the relationship between micro-alloying
elements, the microstructure and deformation in austenite on the ratio of YS/UTS and
other mechanical properties. Therefore, a study is to be undertaken on the austenite to
acicular ferrite transformation with particular emphasis on the kinetics of the acicular
ferrite formation and as affected by the above process variables. The experimental
research would include a redesign of the chemical composition, dilatometer analyses,
simulation of the controlled rolling process on a Gleeble 1500 (initially the work was
started on a Gleeble 1500TM model but halfway through the study, the Gleeble was
upgraded to a Gleeble 1500D DSI), microstructural observations, tensile tests, SEM
and TEM investigations, etc.
5.4.1 Design of the chemical compositions of the investigated alloys
The V-content in the steel was decreased in this study because it only contributes to
dispersion hardening as V(C,N) in ferrite. Its dispersion hardening is less than that of
Nb in steels.
Niobium is a very important micro-alloying element in line pipe steels and formed
one of the main-alloying elements considered in this study. It contributes to dispersion
hardening (NbC in ferrite), promotes acicular ferrite formation (reduction of pearlite)
and raises the Tnr (“pancake” of austenite). The niobium concentration was increased
to about 0.045%wt (which is more than the 0.037% used currently by Mittal Steel).
Titanium is another micro-alloying element that retards austenite grain growth during
reheating. The reheating temperature may be as high as 1225 ºC due to undissolved
TiN. If titanium binds the free nitrogen in the steel, more niobium will be available in
the ferrite to precipitate as NbC and will increase the precipitation hardening.
Titanium can also control the shape of sulphide inclusions (TiS). Accordingly, the
titanium addition was kept at levels of 0.017 to 0.022% in this study.
- 45 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 5 Background of current South Africa line pipe production
Molybdenum contributes to phase transformation hardening and can be used instead
of V in the hardening of steels. It might benefit the increase in volume fraction of
acicular ferrite and M/A islands that are useful to lower the ratio of YS to UTS.
Molybdenum can diminish the Bauschinger effect during pipe forming. Thus,
molybdenum was added to the experimental alloys in this study. The molybdenum
additions considered here were 0.10%, 0.15% and 0.25%, respectively.
Carbon was slightly decreased to 0.05% C in this study for the purpose of improving
the weldability of steel.
Summarising the target analysis above, the chemical compositions of the experimental
alloys were designed as follows (Table 5.3):
Table 5.3 Design of chemical composition ranges of alloys that were investigated (in
wt%)
C
Si
Mn
S
P
V
Nb
Ti
0.05
0.2-0.25
1.0-1.2
<0.005
<0.01
0
0.045
0.022
N
Cu
Ni
Al
Cr
Ca
Mo
0.006
0.007
0.007
0.03
0.02
0.002
0.1-0.25
5.4.2 Design of the controlled hot rolling process
The austenite grain size of the current Mittal Steel alloy, reference alloy #6 (Mo-free)
was found to be 57 and 63 µm at 1225 ºC for 60 and 120 minutes respectively.
Therefore, the reheating temperature of 1225 ºC was chosen for this study to provide
almost complete dissolution of the niobium carbonitrides to achieve maximum
precipitation hardening, but little austenite grain size coarsening.
- 46 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 5 Background of current South Africa line pipe production
In this study, the finish rolling temperature was maintained at about 870 ºC, i.e. above
the Ar3 and, therefore, with no deformation in the (α+γ) two-phase region.
The temperature range for rough rolling was chosen to be from about 1190 to 1000 ºC
in this study, which is above the non-recrystallisation temperature Tnr. The total or
cumulative true strain in the rough rolling stage was chosen to be about 1.4 with
individual pass strains of more than 0.2.
The final temperature for the finish rolling stage was between 840 to 870 ºC. The
finish rolling temperature in this stage was maintained above the Ar3 temperature. The
deformation for this stage was, therefore, carried out in the austenite
non-recrystallisation region. The total strain in the finish rolling stage was about 0.54
in order to accumulate enough rolling strain within the austenite grains for the
subsequent ferrite transformation, leading to a pass strain of more than 0.2 as well.
The initial and final thicknesses of the ingot and plate for laboratory hot rolling were
planned to be 43 and 6 mm respectively, with a total heavy reduction of 86%.
The cooling rate after finish rolling has a significant effect on the subsequent
microstructure of the line pipe steel. Rapid cooling is useful to increase the volume
fraction of acicular ferrite and this contributes to good mechanical properties. It
results in a continuous stress-strain curve, decreases the Bauschinger effect during
pipe-forming and leads to a low ratio of YS/UTS. In this study, various cooling rates
and with/without prior deformation in the austenite on the Gleeble were used to
establish the effect of these parameters on the ratio of YS/UTS. For this purpose, the
experimental challenge was how to measure the volume fraction of acicular ferrite?
The coiling temperature is also very important to the degree of precipitation hardening
of line pipe steels. If the coiling temperature is too high, the precipitates become too
- 47 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 5 Background of current South Africa line pipe production
coarse. At low temperatures acicular ferrite may form, but the temperature cannot be
too low because the precipitation requires diffusion, and in this study two coiling
temperatures of 600 and 575 ºC were chosen. These are also the previous and the
current coiling temperatures respectively used by Mittal Steel in their 11 mm strip
steel for line pipe.
In this study a hypothesis that the acicular ferrite or an optimised mixture of acicular
ferrite and polygonal ferrite, is the most suitable microstructure for decreasing the
ratio of YS/UTS of steels, was, therefore, tested.
A series of tests on the Gleeble were also carried out to study the effect of cooling rate,
coiling temperatures and deformation prior to transformation on the ratio of yield
strength to ultimate tensile strength.
- 48 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
CHAPTER 6 EXPERIMENTAL PROCEDURES
This chapter describes the experimental procedures for the investigation, including the
alloy design, the melting and casting of the ingots, the hot-rolling process, the testing
of the austenite grain size and the presence and identification of undissolved particles,
the determination of the non-recrystallisation temperature, the determination of the
strain-free and the strain affected CCT diagrams and the determination of the
mechanical properties, etc. Distinguishing between the acicular ferrite and polygonal
ferrite in the microstructures of the samples was done by TEM through shadowed
carbon extraction replicas and thin foils. The chemical compositions of the
experimental alloys were typical of commercial line pipe steels with Nb-V-Ti
micro-alloying elements with the current line pipe steel from Mittal Steel as the
reference steel.
6.1 Alloy design
As indicated earlier and briefly summarised here again, the design of the experimental
alloys was based on the considerations set out below.
1. The chemical compositions of HSLA line pipe steels are normally low in carbon
and contain some micro-alloying elements that may be only one, or a combination
of any two or three of the micro-additions (vanadium, niobium and titanium). A
low carbon level was selected for improving the weldability and toughness, to
provide less pearlite and more effective dissolution of niobium during reheating
that will increase the precipitation hardening of these steels.
2. Niobium has strong dispersion hardening characteristics due to the formation of
NbC in ferrite. It promotes the transformation to acicular ferrite that can be
beneficial to lower the ratio of yield strength to ultimate tensile strength of these
steels. Niobium also causes refinement of the austenite grains during the rolling
process by raising the non-recrystallisation temperature (Tnr). Accordingly, a little
more niobium was considered than in the current Mittal Steel alloy while
- 49 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
vanadium was reduced or left out, which decreases the non-recrystallisation
temperature.
3. Titanium can retard the austenite grain growth during reheating of slabs due to the
presence of TiN particles. As described above, considering the stoichiometric ratio
of Ti to N (3.4 /1) and preventing MnS stringer inclusions, (which requires Ti/S in
TiS of 1.5/1), the ideal titanium addition was calculated from equation (2.1).
4. The addition of molybdenum in Nb-containing steels can improve transformation
hardening (increased volume fraction of acicular ferrite and M/A islands), can
provide grain refinement and precipitation hardening. It also greatly suppresses or
delays the formation of polygonal ferrite and pearlite[4]. Additions of molybdenum
were, therefore, considered instead of the usual vanadium for enhancing the
strength of the steels. In steels with molybdenum, the stress-strain curve of the
as-rolled plate is usually continuous, without an upper yield points[33]. This may
provide control of the Bauschinger effect and contribute to an increase in yield
strength from plate to pipe.
Thus, the newly designed alloys that were made up, all had the same low carbon,
niobium and titanium levels but were made with and without vanadium, and also had
varying amounts of molybdenum.
6.2 The melting of the experimental alloys
Sections from the currently produced line pipe steel from Mittal Steel were used as
feed stock material for the melting of the new alloys in a 50 kg vacuum induction
melting furnace at Mintek. The liquid steel was cast into ingots of 43 × 66 × 235 mm.
The chemical compositions of the five new alloys are listed in table 6.1. The Mittal
Steel line pipe steel is included in the table as a reference.
- 50 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
Table 6.1 Chemical compositions of the experimental alloys, in wt.%
Alloy #
1
2
3
4
5
6(Mittal Steel)
Cu
0.02
0.03
0.03
0.02
0.02
0.007
Al
0.004
0.055
0.054
0.055
0.065
0.037
C
0.05
0.06
0.05
0.05
0.05
0.066
Co
0.006
0.009
0.008
0.009
0.01
--
Si
0.29
0.25
0.23
0.24
0.25
0.26
Mn
1.21
1.29
1.05
1.31
1.14
1.583
V
0.06
<0.005
<0.005
<0.005
<0.005
0.062
P
0.014
0.018
0.019
0.019
0.019
0.011
Nb
0.055
0.05
0.051
0.052
0.055
0.037
S
0.011
0.01
0.011
0.011
0.011
0.004
Ti
0.017
0.019
0.019
0.019
0.021
0.017
Cr
0.07
0.05
0.04
0.05
0.05
0.021
Sn
-0.003
0.003
0.003
0.003
0.001
Ni
0.07
0.04
0.04
0.04
0.04
0.007
Mo
0.01
0.09
0.09
0.12
0.22
0.001
B
Free-N
0.0001 0.0068
0.0003 0.0035
0.0002 0.0032
0.0003 0.0032
0.0003 0.0027
0.0002
--
6.3 The effect of reheating temperature and soaking time on the austenite grain
sizes
The samples from the cast ingots were machined into cubes of about 10 × 10 × 10 mm.
Two methods were used to process these samples in order to measure the austenite
grain size. The first was to reheat the samples at temperatures of 1150, 1200, 1225 and
1250 °C, respectively, and then quench them into water. The samples were etched
using many different etchants (see table 6.2), but the results revealed that these
etchants were not suitable to reveal the prior austenite grain boundaries for the alloys
studied. The carbon content of the alloys was probably too low for this. The more
successful technique was a modified McQuaid-Ehn method by carburising the
samples after reheating in argon, in-situ within the austenite region in a dry carbon
monoxide gas atmosphere at 927 ºC for up to 5 hours directly after reheating at the
above four austenitisation temperatures, i.e. without going through the ferrite
transformation. Pro-eutectoid cementite formed on the prior austenite grain
boundaries during very slow cooling from the carburisation temperature at 927 ºC
down to 690 ºC at which temperature the sample was removed from the furnace. Thus,
- 51 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
the cementite layers indicated the sites of the original austenite grain boundaries,
which then became easy to measure. The carburising process is illustrated in figure
6.1. After carburisation, the samples were polished and etched in a 2% Nital solution
and then the original austenite grain sizes were measured by the mean linear intercept
method[97].
Table 6.2 The composition of the etchant solutions
Solution Number
#1
#2
#3
#4
#5
#6
Chemical
Quantity
Picric acid
1g
Hydrochloric acid
5 ml
Ethanol alcohol
95%
FeCl3
1g
H2O
100 ml
FeCl3
1g
Hydrochloric acid
5 drops
H2O
100 ml
Sodium bisulphite
34 g
H2O
100 ml
Hydrochloric acid
40 ml
Sulphuric acid
10 ml
H2O
50 ml
10% aq. Oxalic acid
28 ml
H2O2(30%)
4 ml
H2O
80 ml
- 52 -
University of Pretoria etd – Tang, Z (2007)
Temperature, °C
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
Figure 6.1 Schematic of the modified McQuaid-Ehn carburising process of the
samples directly after reheating.
6.4 Measuring the presence and composition of undissolved particles
During reheating, any undissolved small particles will retard the austenite grain
growth. Before entering the first rough rolling pass, a fine austenite grain size is
beneficial to the later strength and toughness of the steels. The quantity and the size of
the undissolved particles are related to the reheating temperature. Austenitisation
temperatures ranging from 1150 to 1250 °C were employed, together with soaking
times of 15 to 120 minutes for this part of the investigation. The samples were
quenched into water or were fast cooled in helium gas. The details of these treatments
are listed in table 6.3 below. The samples were mechanically ground and polished
after the treatment, then were lightly etched in a 2% Nital solution without an
apparent visible optical microstructure of the matrix, and vacuum coated with carbon.
The shadowed carbon extraction replicas were similarly made, also after light etching
in 2% Nital, then the vacuum application of Au-Pd shadowing at an angle between 20
to 40º before vertical coating of the carbon. The size of the particles was measured on
the micrographs obtained from the transmission electron microscope (TEM) and their
volume fraction calculated using the following equation[98]:
f =
π
6
N s ( x + σ Al )
2
2
Al
- 53 -
(6.1)
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
σ
where
x
2
2
Al
=
n ∑ x Al − (∑ x Al ) 2
(6.2)
n(n − 1)
is the planar arithmetic mean of the particle diameter;
Al
f is the volume fraction of particles;
Ns is the total number of particles intersecting a unit area;
σ
x
2
Al
Al
is the standard deviation of the particle size distribution;
is the diameter of a particle;
n is the total number of particles measured.
Table 6.3 Temperatures and soaking time of the treatment
for undissolved particles
Temperature (ºC)
Time (min)
1150
15
60
120
1200
15
60
120
1225
15
60
120
1250
15
60
120
6.5 Non-recrystallisation temperature (Tnr)
The finishing temperature of rough rolling is associated with the non-recrystallisation
temperature (Tnr) and heavy reductions must take place within this recrystallisation
region at temperatures higher than the Tnr in order to obtain a fine recrystallised
austenite grain size at the start of finish rolling below the Tnr. This is beneficial to high
strength and good toughness of the steels. Accordingly, the non-recrystallisation
temperature is an important parameter that should be considered when the hot rolling
process is designed. Many researchers have studied the recrystallisation of
austenite[99-107]
and
some
mechanisms
- 54 -
have
been
proposed[108-110].
The
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
non-recrystallisation temperature may be a function of the parameters[99,111-113] of the
rolling process, such as pass strain[114,115], strain rate and inter-pass time[48], and may
also depend on the content of micro-alloying elements[116], etc. The hot torsion test
has been widely used to simulate industrial hot rolling processes[117-121]. In this study,
the recrystallisation behaviour of the steel was investigated during multi-pass
compression deformation on a Gleeble hot working facility on cylindrical samples of
8 mm in diameter and 15 mm in length that were machined from the as-cast ingots.
6.5.1 Testing schedule for the determination of the Tnr
Reheating temperatures should be high enough to dissolve all the precipitates (mainly
the Nb precipitates, but except the TiN) because the micro-alloying elements affect
the non-recrystallisation temperature. The reheating temperature for Nb precipitates
can be determined from the following equation[122,123]:
12 ⎞
6770
⎛
log[ Nb]⎜ C + N ⎟ = 2.26 −
14 ⎠
T
⎝
(6.3)
The calculated solution temperatures for Nb(C,N) in the experimental alloys are listed
in table 6.4.
Table 6.4 Calculation equilibrium Nb carbonitride solution temperature
Alloy number
Solution temperature of Nb(C,N) (ºC)
1
1145
2
1149
3
1179
4
1130
5
1137
- 55 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
Laasraoui[124] reported that the niobium carbonitrides remained undissolved at
reheating temperature of 1100 ºC for an 0.04% Nb steel. Thus, a maximum reheating
temperature of 1225 ºC was selected here.
Samples of alloy #6 (the current Mo-free Mittal Steel alloy) were heated to 1225 °C at
a heating rate of 100 °Cmin-1 and held at this temperature for 15 minutes. The
multi-pass compression tests were carried out using the test parameters shown in
tables 6.5 and 6.6. Pass strains, ranging from 0.15 to 0.32, and inter-pass times
ranging from 4 to 50 seconds were employed, at a constant strain rate of 1 s-1. In two
particular tests, the inter-pass time and pass strain were held constant. Another test of
strain rate ranging from 0.1 to 2.22 s-1, was also employed at a constant pass strain of
0.2 and constant inter-pass time of 8 seconds (see table 6.7). Such a multi-pass
compression testing schedule is illustrated schematically in figure 6.2 below.
Table 6.5 Testing parameters for Tnr at strain ranging from 0.15 to 0.32
Inter-pass time (s)
8
8
8
8
8
8
Strain rate (s-1)
1
1
1
1
1
1
0.15
0.2
0.24
0.28
0.30
0.32
Strain/pass
Table 6.6 Testing parameters for Tnr at inter-pass times ranging from 4 to 50 seconds
Inter-pass time (s)
4
6
8
15
20
30
35
40
50
Strain rate (s-1)
1
1
1
1
1
1
1
1
1
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
Strain/pass
Table 6.7 Testing parameters for Tnr at strain rate ranging from 0.1 to 2.22 s-1
8
8
8
8
8
8
8
8
8
Pass strain, ε
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
Strain rate (s-1)
0.1
0.47
0.9
1.22
1.38
1.67
1.80
2.0
2.22
Inter-pass time (s)
- 56 -
University of Pretoria etd – Tang, Z (2007)
Temperature, °C
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
100 ºCmin-1
Time, min
Figure 6.2 Schematic schedule employed in the multi-pass compression tests for the
Tnr.
6.5.2 The determination of the non-recrystallisation (Tnr)
A typical set of curves of the true flow stress versus true strain from a multi-pass
compression test is illustrated in figure 6.3.
160
866℃
interpass time:4s pass strain:0.20
strain rate:1/s
140
895℃
s tre s s ,M P a
120
928℃
100
80
1108℃
1064℃
1018℃
1153℃
60
40
20
0
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
strain
Figure 6.3 The curves of flow stress versus strain in a multi-pass compression test on
alloy #6.
The mean flow stress of each pass was calculated from the following equation and the
flow stress-strain curve (see figure 6.3):
- 57 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
−
σ=
1 ε
∫ε σdε
ε b −ε a
b
(6.4)
a
The non-recrystallisation temperature was determined from the relationship between
the mean flow stress (MFS in MPa) of each pass and the inverse temperature (1/T in
K-1) of the compression deformation, as illustrated in figure 6.4. This typical curve is
divided into two stages by a slope change in the two straight line sections. In the
lower slope stage (which corresponds to a high temperature deformation), full
recrystallisation takes place during the pass because there is no strain accumulation
and the increase in the mean flow stress is solely due to the decrease in temperature of
the inherent strength of a well annealed microstructure. However, in the higher slope
stage (deformation below the Tnr), there is only partial dynamic recrystallisation, or no
recrystallisation at all, indicating that the strain is accumulated from pass to pass, and
the mean flow stress increases more rapidly with decreasing temperature[125]. The
intersection of two straight lines provides the Tnr.
140
M e an flo w stre s s, M P a
120
i n t e r p a s s t i m e: 4 s p a s s s t r a i n : 0.2 0
s t r a i n r a t e : 1/s
100
80
60
40
20
0
0.65
0.7
0.75
0.8
1000/T
0.85
0.9
(K-1)
Figure 6.4 Determining Tnr from the mean flow stress in MPa versus the inverse pass
temperature in K, during a multi-pass compression test on alloy #6.
- 58 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
6.6 CCT-diagram
The alloys used for the CCT diagrams, were the as-cast alloy #5 (with 0.22% Mo) and
alloy #6 (Mo-free) (Mitall Steel reference alloy). The chemical compositions of alloys
#5 and #6 were listed in table 6.1. Samples of with a size of a 15 to 16 mm long
cylinder with a diameter of 7 mm were used for the strain affected CCT tests, and a 8
to 9 mm long cylinder with a diameter of 7 mm for the strain-free CCT test were used.
The temperatures of phase transformations were measured by the C-strain facility on
the Gleeble which measures the change in diameter of the sample during cooling for
the strain affected CCT tests.
6.6.1 The Ac1 and Ac3 test
The Ac1 and Ac3 are the important critical equilibrium temperatures for starting and
completion of the austenitisation transformation during phase transformation of low
carbon hypoeutectoid steels (less than 0.77% C). The determination of the Ac1 and Ac3
temperatures were made on a single-LVDT THETA dilatometer facility. Phase
transformations are normally associated with a non linear volume change in the
temperature range of the transformation. The linear thermal expansion or contraction
of the samples, therefore, takes place in a manner that allows the subtraction of the
linear relationship between dilatometry and temperature from the non-linear
transformation portions. A schematic sketch of the dilation with temperature is
represented in figure 6.5[126]. Figure 6.6 shows the schematic determination of the Ac1
and Ac3 temperatures of steels on the heating curve of dilation versus the testing
temperature. In order to completely dissolve all the Nb-alloyed precipitates in this
study, e.g. the Nb(C,N), the reheating temperature was chosen as 1225 ºC for 15
minutes with a heating or cooling rate to and from this temperature of 3 ºCmin-1 for
equilibrium conditions. Samples for the THETA dilatometer were cylinders with a 7
mm diameter and a 10 mm length. The chamber containing the samples was kept at a
vacuum of 10-4 torr, to prevent any significant oxidation of the samples.
- 59 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
Figure 6.5[126] Schematic dilation as a
Figure 6.6[126] Schematic determination
function of testing temperature.
of the Ac1 and Ac3 temperatures on the
heating curve.
6.6.2 CCT diagram without prior deformation
The continuous cooling transformation (CCT diagram) without prior deformation was
also made on the THETA Dilatometer. The reheating rate was taken as 100 ºCmin-1 up
to 1225 ºC and held for 15 minutes at this temperature for the purpose of complete
dissolution of the Nb-precipitates. The samples were subsequently cooled down to
980 ºC at a rate of 5 ºCs-1 and held for 5 minutes at this temperature before finally,
cooling down to 25 ºC at the various almost linear cooling rates of 0.1, 0.2, 0.5, 1, 2, 5,
8, 10 and 20 ºCs-1, respectively. The sample chamber was evacuated before cooling
after 980 ºC and the cooling media was either flowing argon or helium gas, depending
on the required cooling rate. The chamber of the THETA dilatometer and the
schematic test schedule are illustrated in figures 6.7 and 6.8, respectively. After
cooling, samples were polished and etched in 2% Nital and the microstructures
examined with an Olympus PMG3 optical microscope. A combination of the optical
micrographs and the cooling curves of dilation with the test temperature, were used to
determine the phase transformation temperatures. The CCT diagram was then
constructed from the various curves of temperature on a linear scale with the test time
on a log scale and the phase transformation temperatures indicated on these.
- 60 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
Thermocouple
Sensor
sample
Figure 6.7 Dilatometer chamber
Figure 6.8 Schematic schedule of the test for the CCT diagram on the THETA
Dilatometer.
6.6.3 Strain affected CCT diagram
The stain affected CCT diagram with prior deformation, was carried out on Gleeble
1500D DSI hot simulator. The reheating rate was also 100 ºCmin-1 up to 1225 ºC and
held for 15 minutes at this temperature for the complete dissolution of precipitates in
an argon atmosphere. The samples were then cooled down to 860 ºC at 10 ºCs-1, and
held for 5 minutes at this temperature. The samples were then compression deformed
with a strain of 0.6 (45% reduction below the Tnr) at 860 ºC at a strain rate of 0.5 s-1.
Samples were finally cooled down to room temperature after deformation, at the
various linear cooling rates of 0.1, 0.2, 0.5, 1, 2, 5, 8, 10, 20 and 40 ºCs-1, respectively.
The cooling media was also argon (for a slow cooling rate) or helium gas (for a rapid
- 61 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
cooling rate). The steps for preparing the micrographs were the same as for the
strain-free CCT diagram above. The chamber of the Gleeble hot simulator and the
schematic test schedule are illustrated in figures 6.9 and 6.10, respectively. The
C-gauge facility of the Gleeble was used to measure the dilation of the sample during
cooling after compression deformation.
Thermocouple
C-gauge
Piston
sample
Figure 6.9 Chamber of the Gleeble 1500D DSI
Figure 6.10 Schematic schedule of the test for the strain affected CCT diagrams on the
Gleeble.
- 62 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
6.7 The thermo-mechanical process
6.7.1 Cooling unit
A specially constructed cooling unit was used to cool the laboratory melted ingots
after hot rolling in a controlled manner. The cooling unit consists of nozzles, a water
pump, control valves, and a water box etc. The experimental arrangement is shown in
figure 6.11. The coolant is fed with compressed air and sprayed onto the cooling
samples through nozzles. The linear distance between the nozzles was about 70 mm,
so that uniform cooling of samples of about 100 × 300 mm could be attained. The
cooling rate, for instance, was 21 °Cs-1 using fresh water at 24 °C as the coolant and
with compressed air spraying of the water with an air pressure of 580 MPa. A higher
figure of 47 °Cs-1 was achieved with a 10% NaCl aqueous solution instead of water.
Cooling Box
Air gauge
Air switch and
Water Switch
Water gauge
Water out
Water container
Pump Switch
(a)
Water pump
Water in
(b)
Water valve
Water in
Water in
(c)
(d)
Figure 6.11 Experimental arrangement of the cooling unit for controlled cooling: (a)
overall view, (b) controller for mixing of gas and water, (c) valves for the nozzles and,
(d) cooling spray in the chamber from the spraying jets.
- 63 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
6.7.2 Hot rolling process of the laboratory ingots
The schedule for the laboratory hot rolling of the ingots, i.e. the pass strain, the total
reduction and inter-pass times etc., were controlled as far as possible to optimise the
process for austenite grain refinement[114,115]. The experimental casts were hot rolled
on a two-stand laboratory hot rolling mill equipped with 10 inch rolls. The ingots of a
machined 43 × 90 × 66 mm size for the alloys #1 to #6 were reheated and hot rolled to
6 × 100 × 300 mm plates with the final thickness of 6 mm chosen to simulate a
possible future reduced gauge for Mittal Steel from their current 11 mm strip.
6.7.2.1 Reheating before laboratory hot rolling
From the results of the study of the effects of temperature and time on the austenite
grain size, a reheating temperature of 1225 °C and a time of 60 minutes were taken.
This is quite safe as it has been reported that austenite grains will not coarsen unduly
before an austenitisation temperature of about 1250 °C is reached in V-Nb-Ti micro
alloyed steels[35].
6.7.2.2 Rough rolling of the laboratory hot rolling
As indicated before, the metallurgical function of roughing is to refine the coarse
austenite grains after the reheating and soaking to achieve the finest possible
dynamically recrystallised grains before entering the finish rolling below the Tnr. Pass
strains should, therefore, be at least 0.2 or higher to promote Dynamic
Recrystallisation (DRX) during rough rolling to produce finer recrystallised grains.
The starting temperature for rough rolling of the laboratory cast ingots was between
1148 and 1190 °C and five passes in this roughing stage were undertaken, with a total
roughing strain of 1.43.
- 64 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
6.7.2.3 Finish rolling of the laboratory hot rolling
As before, the objective of this stage is to accumulate sufficient rolling strain within
the austenite grains to promote a finer ferrite transformation after rolling. Ferrite
nucleation sites are, therefore, greatly multiplied in number and a very fine ferrite
grain size can be generated during the subsequent controlled cooling[6]. In this study
three passes were used with a total strain of 0.54 and the finish rolling temperature
was maintained between 857 and 865 °C.
6.7.2.4 Cooling rate after laboratory finish rolling
The cooling rate (Vc) and the finishing temperature (Tc) of the accelerated cooling
after finish rolling, are important parameters of the thermo-mechanically controlled
processing for the experimental alloys to achieve their optimum strength and
controlling the Tc and Vc may lead to the control of the precipitation of carbonitrides
during the accelerated cooling[9]. A rapid cooling rate helps to promote finer
precipitation of Nb(C,N), ferrite grain refinement and acicular ferrite formation. The
latter is preferred for a low ratio of yield strength to ultimate tensile strength[4], and it
also avoids the development of any pearlite in the microstructure. A rapid cooling rate
of 47 °Cs-1could be achieved in the cooling box by using an aqueous solution of 10%
NaCl for alloys #1 to #5 while a rate of 39 °Cs-1could be obtained for the Mo-free
alloy #6.
6.7.2.5 Simulation of coiling after laboratory hot rolling
The coiling temperature will influence the effectiveness of Nb-carbonitride but
especially V(C,N) precipitation in the ferrite, thus controlling precipitation
strengthening of these steels. If the coiling temperature is high, the precipitates will
become coarser during coiling. A coiling temperature of 600 °C was selected for the
study of these alloys as this was also the temperature used in the past by Mittal Steel
before they recently lowered it to 575 ºC. After reaching 600 °C in the cooling box,
the small plates were placed in a furnace at 600 °C until the temperature became
- 65 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
uniform throughout the plate. Thereafter they were well insulated for 24 hours using a
generous covering of vermiculite for a slow cool simulating the actual coiling process
in the plant.
6.7.2.6 Hot-rolling process curve
The hot-rolling process schedule is illustrated schematically in figure 6.12. The
Temperature, °C
symbols R and F in the figure refer to the rough and finishing passes, respectively,
Figure 6.12 Schematic schedule of the hot rolling process on the experimental alloys.
6.8 The identification of acicular ferrite
The optimum microstructure of line pipe steels appears to be one that contains
acicular ferrite and some polygonal ferrite. Acicular ferrite is very different from
polygonal ferrite and bainitic ferrite. Polygonal ferrite, as the name implies, has
polygonal boundaries with a carbon concentration in the ferrite that is almost uniform
together with a lower dislocation density. Acicular ferrite, on the other hand, is
characterised by fine non-equiaxed or interwoven nonparallel ferrite laths, which have
various grain sizes and are arranged in a random manner[78]. Some M/A
(martensite/austenite) islands and a high density of dislocations in-between the AF are
common[11]. The carbon content in the M/A islands is higher than that in the
surrounding
matrix
and
accordingly,
- 66 -
these
islands
are
carbon-enriched
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
martensite/austenite islands whose formation may be attributed to the partitioning of
carbon during the transformation to acicular ferrite and the post-transformation of
carbon-enriched austenite. During accelerated cooling followed by the coiling process,
part of the carbon-enriched austenite transforms to martensite and the remaining
austenite will coexist with the martensite[73]. The accepted transformation model of
acicular ferrite is a mix of diffusion and shear transformation[ 4,9,11,43].
Acicular ferrite contributes to a low yield strength and a higher tensile strength due to
the lower carbon in the matrix and the M/A islands having a high density of
dislocations and this, therefore, leads to a lower YS/UTS ratio. It was found initially
that it was easy to confuse acicular ferrite with polygonal ferrite under an optical
microscope when samples were etched in 2% Nital because the grain boundaries
between them do not become clearly etched. Therefore, one of the key aspects of the
experimental techniques in this study was how to distinguish between acicular ferrite
and polygonal ferrite in low carbon, Nb-Ti micro-alloyed steels. Various techniques
were initially attempted to identify the acicular ferrite, including the use of optical
microscopy,
Scanning
Electron
Microscopy
(SEM),
Transmission
Electron
Microscopy (TEM) using carbon extraction replicas with and without shadowing and
finally, thin foil TEM samples.
6.8.1 Observation with optical microscopy and by SEM
Samples taken after hot rolling that were etched in a 2% Nital solution, were
examined with an Olympus PMG3 optical microscope, both JEOL 5800LV and 6000F
high resolution SEM, respectively. Acicular ferrite could not be distinguished from
polygonal ferrite on the micrographs after etching with a light (5 seconds) or a deep
etch (up to 120 seconds) under the optical microscope and SEM.
- 67 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
6.8.2 Observation of replicas by TEM
6.8.2.1 Preparing replicas without shadowing
Firstly, the polished samples were deeply etched from 30 to 60 seconds in a 2% Nital
solution and thoroughly washed to remove all loose etching debris from the surface.
Carbon coating was done under vacuum on the etched surface of the samples and the
coating separated from the sample’s surface in a mixture of 7 ml nitric acid and 75 ml
ethanol before floating-off in distilled water.
6.8.2.2 Preparing shadowed carbon extraction replicas
The etched surface of the samples was first shadowed through vacuum evaporation
with a gold-palladium alloy before the vertical deposition of carbon. Deep etching
from 30 to 60 seconds was employed before the shadowing and the shadowing angle
with respect to the surface was varied from 20° to 40°. The carbon was then coated
vertically onto the shadowed layer. The technique used for separating the carbon film
from the sample was the same as described above for the replicas without shadowing.
6.8.3 Thin foil TEM samples
Thin foil samples were used to further distinguish between acicular ferrite and
polygonal ferrite and to also validate the results of the shadowed replicas in the TEM.
Thin slices of material were cut by electro-discharge in a wire-cutting machine in
order to reduce any adverse effect of deformation on the dislocation density in the
samples that could have formed from mechanical cutting. The disc size was 3 mm in
diameter and 0.6 mm in thickness and the procedure for preparing the thin foil
samples was as follows:
•
The original cylinder of material with dimensions of 3 mm diameter and 15
mm length was machined in a lathe;
•
Five discs of 3 mm diameter and 0.6 mm thickness were cut from the cylinder
with the electro-discharge wire-machine;
•
The disc samples were carefully and lightly polished to between 50 to 80 μm
- 68 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
in thickness on fine grinding paper;
•
The final thinning was done by electro-polishing with a solution of 1.25 l
acetic acid, 0.08 l perchloric acid and 0.7 g chromium oxide at 25 °C.
All samples for replicas and thin foils were examined in a Philips CM200 TEM,
operated at 60 or 200 KV, respectively.
6.9 Test of subsize samples on the Gleeble with various cooling rates, coiling
temperatures and prior deformation
In order to study how some parameters of the controlled hot rolling process influence
the YS/UTS ratio of steels, a series of experiments were planned. These parameters
include cooling rates, coiling temperatures and deformation in the austenite.
The cooling rate after hot rolling has a strong effect on the fraction of acicular ferrite
in Nb-Ti micro-alloyed steels and particularly, accelerated cooling after finish rolling
helps to increase the volume fraction of acicular ferrite which may, in turn, influence
the YS/UTS ratio. The coiling process controls the precipitation of particles in the
matrix of these steels which affects dispersion hardening. The reduction during hot
rolling also has an effect on the CCT diagram of these steels as well. A series of tests
were designed for this purpose.
6.9.1 Hot rolling plates for Gleeble samples
Alloys #3 and #6 were selected for these tests. Casts of these alloys, firstly, were hot
rolled to 6 mm thickness (the parameters of the hot rolling for plates are shown in
Appendix H). Two types of preliminary samples were machined: the first type A to a
rectangular size of 6 ×10 × 100 mm was used to study the effect of cooling rates and
coiling temperature without deformation (figure 6.13-(a)), while the other type B
shown schematically in figure 6.13-(b), was used to study the effect of prior
compression in austenite, from there the shorter gauge length. At this stage, the gauge
- 69 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
lengths had not been machined into the rectangular samples, hence their being called
“preliminary”.
Figure 6.13 Preliminary samples on the Gleeble of (a) type A and (b) type B
6.9.2 Tests on the Gleeble
The first two groups of type A samples were tested at different cooling rates after
austenitisation in which the cooling rates ranged from 1 to 51 ºCs-1 for the Mo-free
alloy #6 (sample #A124) and 1 to 54 ºCs-1 for the 0.09% Mo alloy #3 (sample #AF3F).
The process graph is shown schematically in figure 6.14.
- 70 -
University of Pretoria etd – Tang, Z (2007)
Temperature, ºC
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
Time
Figure 6.14 Graph of the heating and cooling process on the Gleeble for samples
#A124 (the Mo-free alloy #6) and sample #AF3F (the 0.09% Mo alloy #3).
The second two groups of type A samples were subjected to a 60 minutes coiling
simulation at 575 (sample #B113) and 600 ºC (sample #A113) for the Mo-free alloy
#6, respectively, after cooling in Gleeble. The process graphs are illustrated
Temperature, ºC
Temperature, ºC
schematically in figures 6.15.
(b)
(a)
Time
Time
Figure 6.15 Graphs of the heating and cooling cycles in the Gleeble on the Mo-free
alloy #6 for samples (a) #A113 and (b) #B113.
The last group, which consisted of type B samples (sample #TEN06 from the Mo-free
alloy #6) was tested on the Gleeble with a prior 45% reduction in the austenite (only
33% reduction below the Tnr) before cooling and coiling simulation at 575 ºC for 60
minutes. Figure 6.16 shows the test process.
- 71 -
University of Pretoria etd – Tang, Z (2007)
Temperature, ºC
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
Time
Figure 6.16 Graph of heating, cooling and deformation process in the Gleeble for
sample #TEN06 (the Mo-free alloy #6).
6.9.3 Tensile tests
All of these rectangular and preliminary samples that were subjected to the above tests
on the Gleeble, were thereafter machined with their gauge lengths to subsize tensile
samples of type A (see figure 6.17-(a)) for the tensile test which was done on an
INSTRON-8500 Digital Control tensile testing machine for the first four groups of
samples and the type B (see figure 6.17-(b)) on a smaller instrumented Hounsfield
tensile testing machine.
- 72 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 6 Experimental procedures
(a)
(b)
Figure 6.17 Tensile test samples of (a) type A and (b) type B. T is original thickness of
the plates and 6 mm for the as-rolled alloy and the Gleeble samples, respectively.
6.10 Test of mechanical properties on the as-hot rolled alloys
The specimens for the tensile tests from the hot rolled plates were cut from the middle
of the rolled plates in the longitudinal and transverse directions and were machined to
the subsize tensile of type A (see figure 6.17-(a)). The tensile tests were carried out at
room temperature on an INSTRON-8500 Digital Control tensile testing machine with
an initial cross-head speed of 0.25 mm min-1 until the elongation of 0.5 mm was
reached and then a second cross-head speed of 2 mm min-1 thereafter.
- 73 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
CHAPTER 7 RESULTS
7.1 The effect of the austenitisation temperature and holding time on the
presence of undissolved particles in the V-Nb-Ti-containing alloys
Stable undissolved particles are important in the thermo-mechanically controlled
process (TMCP) of line pipe steels. The finer austenite grain size after reheating and
before entering the rough rolling stage, result in a good balance of strength and
toughness of steels due to the existence of fine undissolved particles. These
undissolved particles coarsen with increasing austenitisation temperature, which
results in a coarse austenite grain size which is harmful to the final strength and
toughness of the steel. The higher austenitisation temperature, however, is beneficial
to the thermo-mechanical controlled process as it leads to a low deformation force in
the rolling mill. Micro-alloying element additions to high-strength low carbon low
alloy line pipe steels are necessary to, firstly, retard the growth of austenite grains at
higher austenitisation temperatures and secondly, to retard deformation-induced
recrystallisation in the final passes of hot rolling. In order to analyse these particles
after reheating, two types of replicas were made in this work: carbon extraction
replicas with and without Au-Pd shadowing.
The micrographs of the particles without shadowing are shown in figure 7.1 in which
few and relatively small particles can be seen. This suggests that there is a low
contrast between the small particles and the carbon film on the extraction replicas
made without shadowing. In other words, this technique is not sensitive to small
particles of less than 100 nm in diameter. For example, the small dark dot (shown by
an arrow in figure 7.1-(b)) can not simply be identified as a particle. Some of the
small particles, therefore, could not be included in the count for the calculation of the
volume fraction of particles on normal unshadowed carbon replicas, which may have
resulted in an error for the measured volume fraction of particles. In order to decrease
this error, the technique of shadowed carbon extraction replicas was used to more
clearly reveal the smaller particles. The carbon extraction replica with Au-Pd
shadowing of alloy #6 after reheating at 1225 ºC for 120 min is illustrated in figure
7.2.
- 74 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
(b)
(a)
Figure 7.1 Extraction replicas without shadowing with undissolved particles for alloy
#6 after reheating at 1200 ºC for 15 min. (Most of the darker spots are not particles
but are etching debris on the replicas).
2µm
Figure 7.2 Extraction replicas with Au-Pd shadowing for alloy #6 after reheating at
1225 ºC for 120 min.
Comparing figures 7.1 (without shadowing) and 7.2 (with shadowing), the very small
particles can be seen more clearly with shadowing, as shown by the small particle of
about 52 nm in diameter (shown by an arrow in figure 7.2). It was easy to identify it
as an undissolved particle and not etching debris as it contains a shadow beside it,
- 75 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
which indicates that the small particle was protruding from the surface during
shadowing and before carbon evaporation. It was, therefore, concluded that the
shadowed carbon extraction replica is superior in revealing undissolved particles than
carbon extraction replicas without shadowing and this reduced errors in the measuring
volume fraction of particles. Comparing the results of the volume fraction
measurement of particles with and without shadowing, the measured volume fractions
on shadowed replicas were higher than those without shadowing (see table 7.1 below).
The volume fraction of particles was calculated in equations (6.1) and (6.2).
Table 7.1 Measured volume fraction of particles on
replicas with/without shadowing in alloy #6
Volume fraction (%)
Treatment
No shadowing
With shadowing
As hot rolled
0.31
0.34
1200 ºC 15 min
0.26
0.27
1200 ºC 120 min
0.21
0.22
As may be seen in table 7.1, some smaller particles may not have been counted on
unshadowed carbon replicas due to lack of adequate contrast with the carbon film.
Therefore, the measured results on the shadowed replicas have a superior precision
although both suffer from the effects of a lack of a planar surface after etching[127],
which is a pre-requisite in the equation used to calculate the volume fractions.
Some TEM micrographs of particles on the shadowed replicas of alloy #6 after light
etching in 2% Nital, are shown in figures 7.3, 7.4, 7.5 and 7.6 for the reheating
treatments of 1150, 1200, 1225 and 1250 ºC for different times, respectively. Most
undissolved particles are intragranular (marked with “A” in figure 7.6-(a)) and a few
particles are on grain boundaries (marked with “B” in figure 7.6-(a)).
- 76 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
2µm
(a)
0.9µm
(b)
1.5µm
(c)
Figure 7.3 TEM micrograph of particles on the shadowed replicas of alloy #6 reheated
at 1150 ºC for (a) 15 min, (b) 60 min and, (c) 120 min.
- 77 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
1.2µm
(a)
4.2µm
(b)
5.3µm
(c)
Figure 7.4 TEM micrograph of particles on the shadowed replicas of alloy #6 reheated
at 1200 ºC for (a) 15 min, (b) 60 min and, (c) 120 min.
- 78 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
1.2µm
(a)
0.9µm
(b)
0.5µm
(c)
Figure 7.5 TEM micrograph of particles on the shadowed replicas of alloy #6 reheated
at 1225 ºC for (a) 15 min, (b) 60 min and, (c) 120 min.
- 79 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
A
B
B
A
2.7µm
(a)
1.2µm
(b)
3.2µm
(c)
Figure 7.6 TEM micrograph of particles on the shadowed replicas of alloy #6 reheated
at 1250 ºC for (a) 15 min, (b) 60 min and, (c) 120 min.
Table 7.2 is a summary of the measured results of undissolved particles observed on
the shadowed replicas in the V-Nb-Ti micro-alloyed line pipe steel of Mittal Steel
(alloy #6) whose chemical composition was given in table 6.1 in section 6.1.
- 80 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
Table 7.2 Undissolved particles: types and sizes after reheating treatments of alloy #6
Treatment
As-hot rolled
15 min
1150 ºC
60 min
120 min
15 min
1200 ºC
Type
Size in
diameter
(nm)
Shape
Nb/Ti
Ratio of peak
of EDS
(Ti,Nb)(C,N)
22~132
Square, Ellipsoid
0.31~0.39
(Ti,Nb)C
33~313
Square, Ellipsoid
0.22~0.8
(Nb,Ti)C
--
Square, Ellipsoid
4.11
VN
--
Square, Rectangle
--
(Ti,Nb)N
132
Ellipsoid
0.13
(Ti,Nb)(C,N)
24~94
Square, Rectangle
0.21~0.28
(Ti,Nb)(C,N)
25~132
Square
0.23~0.26
(Ti,Nb)C
56
Square, Rectangle
0.31
(Ti,Nb)(C,N)
56~185
Square, Round
0.21~0.43
(Ti,Nb)N
73~400
Round, Square
0.21~0.27
(Nb,Ti)C
26
Round, Rectangle
10.7
Ti(C,N)
46~132
Square
---
(Ti,Nb)N
182
Square, Rectangle
0.18
(Ti,Nb)(C,N)
27~313
Square
0.15~0.29
29~330
Square, Rectangle
(Ti,Nb)(C,N)
22~182
Square, Rectangle
0.12~0.16
(Ti,Nb)N
94 ~382
Square, Rectangle
0.11~0.21
27~322
Square, Rectangle
--
0.22
25~396
Square, Rectangle
--
0.21
30~375
Square, Rectangle
--
0.19
37~428
Square, Rectangle
--
0.21
34~403
Square, Rectangle
--
0.20
46~439
Square, Rectangle
--
0.19
(Ti,Nb)(C,N)
60 min
120 min
(Ti,Nb)N
(Ti,Nb)(C,N)
15 min
1225 ºC
Ti(C,N)
(Ti,Nb)(C,N)
60 min
120 min
Ti(C,N)
(Ti,Nb)(C,N)
(Ti,Nb)(C,N)
15 min
Ti(C,N)
Volume
fraction
(fv,%)
0.34
0.32
0.28
0.26
0.27
0.24
0.22
1250 ºC
60 min
(Ti,Nb)(C,N)
(Ti,Nb)(C,N)
120 min
Ti(C,N)
As can be seen from table 7.2, the types of undissolved particles are (Ti,Nb)(C,N),
(Ti,Nb)C, (Ti,Nb)N, Ti(C,N) and (Nb,Ti)C. Their respective shapes are cubic or
squared, rectangular, rounded and ellipsoidal. There were some VN particles in the ashot rolled condition but no VN particles were observed after reheating treatments
- 81 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
above 1150 ºC in alloy #6. VN particles were completely dissolved in the austenite
above 1150 ºC in this steel and, therefore, can not inhibit the grain growth of the
austenite above this temperature.
(Ti,Nb)C and (Nb,Ti)C particles of ellipsoidal and rounded shapes appeared to
dissolve at 1200 ºC after 120 min. (Ti,Nb)(C,N) and (Ti,Nb)N remained undissolved
at this temperature as indicated by the fact that only square and rectangular particles
remained in the steel. It was observed that these squared and rectangular particles
were more stable than the ellipsoidal and rounded ones with increasing solution
temperature. Some EDS spectra from the transmission electron microscope analyses
are presented in figure 7.7.
(a) (Nb,Ti)C
(b) (Ti,Nb) (C,N)
(c) (Nb,Ti) C
(d) (Ti,Nb)N
Figure 7.7 TEM-EDS results of undissolved particles of alloy #6 after the treatments
of (a) as-hot rolled, (b) as-hot rolled, (c) 1150 ºC for 120 min and, (d) 1200 ºC for 15
min.
- 82 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
Some copper peaks in figure 7.7 were probably from the copper grid that supports the
carbon coating while a large part of the carbon from the carbon peak (marked with
blue) was from the background of the carbon replicas films.
As may be expected in figure 7.8, the volume fractions (fv) of undissolved particles
decreases with an increase in the temperature of solution treatment for all the
investigated soaking times, i.e. 15, 60 and 120 min. The curve fitted equations are as
follows for the times of 15, 60 and 120 min, respectively.
Volume fraction of particles,fv,%
0.35
0.3
0.25
0.2
0.15
0.1
15min
60min
0.05
0
1140
120min
1160
1180
1200
1220
Temperature,
1240
1260
ºC
Figure 7.8 The effect of reheating temperature and time on the volume fraction of
undissolved particles for alloy #6.
fv=10-6T2–0.0038T+3.2258
R2=0.97
for 15min soaking time
(7.1)
fv=10-6T2–0.0043T+3.3251
R2=0.99
for 60min soaking time
(7.2)
fv=4×10-6T2–0.0103T+6.862
R2=0.97
for 120min soaking time
(7.3)
where fv is volume fraction of undissolved particles after the soaking time in %.
T is temperatures of solution or reheating treatment in ºC.
R2 is the regression fitting coefficient.
As can be seen from equations (7.1) to (7.3), there is a similar trend between the
volume fraction of undissolved particles and the solution or reheating temperature at
different soaking times. The undissolved particles become less stable at higher
- 83 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
solution temperatures and, therefore, reduce their ability to inhibit any growth of the
austenite grains. It can, therefore, be concluded that at high solution temperatures,
small particles dissolve and large ones coarsen and the grain boundary pinning effect
is reduced or eliminated, leading to austenite grain growth.
There are two effects in the present work: one is coarsening at constant temperature
by small particles dissolving and large particles growing by the Lifshitz-SlyosovWagner process, such as with the more stable (Ti,Nb)(C,N) and Ti(C,N) that will
coarsen above 1250 ºC for 120 min (see table 7.2). The second effect arises from a
larger solubility at higher temperatures in which all sizes of particles dissolve, i.e.
both small and large ones dissolve, such as with the less stable VN at 1150 ºC and
(Ti,Nb)C, (Nb,TI)C at 1200 ºC. This latter type of particle is less stable than the
former ones, so that they lose any effect to inhibit the growth of austenite grains.
Table 7.2 shows that the size of the smallest undissolved particles does not
significantly change below 1225 ºC. The smallest sizes for the as-received as-hot
rolled steel of alloy #6 was 22 nm and after soaking at 1225 ºC for 120 min, this size
increased to about 30 nm. The particles started coarsening at a soaking temperature of
1250 ºC for 120 min when a mean particle diameter of 46 nm was observed. It can be
concluded that the undissolved particles in alloy #6 containing Nb-Ti-V microalloying elements, still had a strong inhibition on the growth of the austenite grains
and, therefore, fine austenite grains can still be obtained at reheating temperatures up
to about 1225 ºC. The particles of Ti- or (Ti+Nb)-carbonitrides, however, were still
undissolved at temperatures above 1200 ºC. All of the V- and Nb-carbides and Vnitrides were dissolved completely at temperatures above 1200 ºC, meaning that only
Ti- and (Ti+Nb)-carbonitrides or nitrides contribute to a grain boundary pinning effect
to limit austenite grain growth above the temperature of 1200 ºC. In order to get much
greater dispersion hardening (Nb- and V-carbides in ferrite) and stronger retarding of
austenite grain growth (undissolved particles of Ti- or (Ti+Nb)-carbonitrides in
austenite), enough titanium in the steel is needed to bind all free nitrogen, leading to a
significant increase in niobium available in the ferrite. In conclusion: if the reheating
temperature in the hot rolling process is not more than 1225 ºC, a combination of
- 84 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
greater dispersion hardening and smaller austenite grains after reheating may still be
achieved in Nb-Ti micro-alloyed steels.
7.2 Austenite grain size and reheating temperature
The reheating temperature should be high enough to dissolve most of the microalloying elements without significant growth of the austenite grains, but also allowing
dispersion hardening after the TMCP. The in-situ carburising technique by carbon
monoxide gas was used to reveal austenite grain boundaries. The austenite grain size
for alloy #6 is shown in figures 7.9 (a)-(d) for reheating conditions at 1150 to 1250 ºC,
respectively.
1150 ºC
(a)
1225 ºC
(c)
1200 ºC
(b)
1250 ºC
(d)
Figure. 7.9 Pro-eutectoid cementite decorates the original austenite grain boundaries
in alloy #6 for soaking times of 60 min at different austenitisation temperatures.
- 85 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
The measured intercept length of the austenite grain size for each treatment is shown
in table 7.3 below as well as in figures 7.10 and 7.11 while some data from the
literature[39] for a steel with composition of 0.1% C, 0.2% Si, 1.4% Mn, 0.005%N,
0.03% Al and 0.01% Ti, have been included in figure 7.10 as a reference.
Table 7.3 Intercept length austenite grain size, in µm, versus
Austenite grain size, µm
reheating temperature and soaking time of alloy #6
Soaking time
1150 ºC
1200 ºC
1225 ºC
1250 ºC
15 min
32.3
41.6
50.4
65.5
30 min
33.9
49.6
52.9
75.0
60 min
48.9
53.0
57.1
76.5
120 min
54.0
58.7
62.9
82.7
90
80
70
60
50
15 min
40
30 min
30
120min
20
1140
60 min
literature
1160
1180
1200
1220
1240
1260
Temperature, ºC
Figure. 7.10 The relationship between the austenitisation temperature and the
austenite grain size for alloy #6. The broken line is from published data[39].
- 86 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
Austenite grain size, µm
90
80
70
60
50
40
1150 Deg C
30
1200 Deg C
1225 Deg C
20
1250 Deg C
10
0
0
20
40
60
80
100
120
140
Soaking time, min
Figure. 7.11 The weak effect of soaking time on the austenite grain size for alloy #6.
As may be seen from figure 7.10, the austenite grain size for alloy #6 increased with
increasing austenitisation temperature from 1150 to 1250 ºC, at a constant soaking
time. The austenite grains grow slowly, however, below 1225 ºC due to the pinning of
the grain boundaries by the stable micro-particles, i.e. (Ti,Nb)(C,N) or (Ti,Nb)N as
described in section 7.1, as there is less coarsening of these carbonitrides in the alloy
below 1225 ºC. Significant coarsening begins, however, when the carbonitrides
dissolve at 1250 ºC (see figure 7.10). As a result, the austenite grains grow sharply
with an increase in temperature above 1225 ºC. The growth in austenite grain size is
given by an increase of ∆D=19.8 µm from 1225 to 1250 ºC for a soaking time of 120
minutes, compared to a growth by an increase of only ∆D=4.2 µm from 1200 to 1225
ºC. Accordingly, the reheating temperature of the alloys in the present study should
not be higher than 1225 ºC for a relatively fine austenite grain size.
Considering much greater dispersion hardening in the ferrite from Nb-bearing
precipitates and raising the non-recrystallisation temperature, the niobium should
preferably be dissolved completely into the matrix of the steel. The reheating
temperature before the hot rolling process should, therefore, not be too low also, but
so high that all of the niobium could dissolve into the matrix of the steel. The most
effective precipitation strengthening, therefore, can be obtained from NbC
precipitation in ferrite[12]. A higher non-recrystallisation temperature (Tnr) can also be
obtained[12] due to the precipitation of Nb(C,N) during hot rolling[28]. This will, on the
- 87 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
other hand, result in “pancake” austenite grains during hot rolling, providing more
nucleation sites as well as sufficient strain accumulation for the following ferrite
formation, which results in finer ferrite. Cuddy[114] reported that the broad range of
initial austenite grain sizes (53 to 325 µm) converged to a narrow range of grain sizes
of 43 to 53 µm after multiple recrystallisation that was produced by a five passes
reduction schedule. It was also found that the final austenite grain size before reaching
the Tnr was dependent mainly on the deformation parameters and was largely
independent of the steel’s composition, i.e. similar for C-Mn, Nb-bearing and Vbearing steels. Considering a combination of finer austenite grains and greater
dispersion hardening, the limit of 1225 ºC was taken as the most appropriate reheating
temperature in this study. On the other hand, niobium also raises the Tnr because of
solute drag by niobium on austenite boundaries if niobium is completely dissolved
into the steels. A higher Nb addition, therefore, results in more deformation below the
Tnr during the hot rolling (if compared to that of lower Nb additions) which induces a
finer ferrite grain size.
The soaking time at the austenitisation temperature influences the austenite grain size
as well although to a lesser extent than the temperature. The longer the soaking time,
the larger the austenite grain size (see figure 7.11). The soaking time is, therefore,
largely determined by the slab’s section size to achieve uniform temperatures and also
to a lesser extent, by the need to dissolve the niobium micro-alloying elements.
7.3 The non-recrystallisation temperature (Tnr) and deformation parameters
The non-recrystallisation temperature (Tnr) is one of the most important factors in the
design of the thermo-mechanically controlled process (TMCP) schedule for line pipe
steels. The Tnr is the critical temperature below which no dynamic or static
recrystallisation will take place during or immediately after hot rolling and this leads
to the formation of so-called “pancake” grains. The austenite grains become flattened
with consecutive pass reductions as the strain accumulates, up to the Finishing Mill
Head where after transformation to ferrite takes place on the run-out table. The
finishing rolling is normally carried out below the Tnr, which is a function of the alloy
content (particularly the Nb alloying addition) as well as parameters of the TMCP, i.e.
pass strain (ε), strain rate( ε& ) and inter-pass time (tip).
- 88 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
7.3.1 The Tnr and pass strain
Figure 7.12 illustrates for alloy #6 the compression deformation cycles used at various
pass strains ranging from 0.15 to 0.32 at the same strain rate of 1 s-1 and inter-pass
times of 8 seconds.
160
140
869℃
p a s s s t r a i n : 0 .1 5 s t r a i n r a t e : 1/s
interpass time:8s
120
1045℃
1087℃
s tre s s ,M p a
80
120
1006℃
1130℃
60
904℃
100
40
993℃
1034℃
80
1130℃
60
40
20
(b)
20
(a)
0
0
0
0.2
0.4
0.6
0.8
1
1.2
0
0.2
0.4
0.6
strainn
0.8
1.2
1.4
160
849℃
140
p a s s s t r a i n : 0. 2 4 s t r a i n r a t e :1/s
interpass time:8s
120
877℃
100
993℃
1038℃
80
passs t r ai n: 0. 28 st r ai n r at e:1/s
i nt er pass t ime: 8s
120
908℃
1 1 3 7℃
60
898℃
976℃
100
40
837℃
865℃
140
stress,Mpa
s tre s s ,M p a
1
strainn
160
1042℃
80
1128℃
60
40
(c)
20
(d)
20
0
0
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
0
0.2
0.4
0.6
0.8
strainn
1
1.2
1.4
1.6
1.8
st r ai nn
120
140
pass st r ai n: 0. 30 st r ai n r at e: 1/s
i nt er pass t i me: 8s
120
877℃
p a s s s t r a i n : 0. 3 2 s t r a i n r a t e : 1/s
interpass time:8s
849℃
100
908℃
80
1137℃
1038℃
844℃
858℃
s tre s s ,M p a
100
stress,Mpa
852℃
878℃
interpass time:8s
916℃
100
s tre s s , M p a
140 p a s s s t r a i n : 0. 2 0 s t r a I n r a t e : 1/s
887℃
993℃
60
80
1068℃
60
976℃
880℃
1143℃
40
40
20
20
(e)
0
(f)
0
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
st r ai nn
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
strainn
Figure 7.12 Stress-strain curves of multi-pass compression tests of alloy #6 at the
same strain rate of 1 s-1, inter-pass time of 8 seconds and different pass strains.
- 89 -
2.2
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
The number of passes in these compression tests was seven at 0.15 strain per pass and
6 passes at more than 0.2 strain per pass. The stress-strain curves for each pass are
presented in figure 7.12 with the temperatures of each pass marked on its stress-strain
curve. The mean flow stress of each pass was then calculated according to equation
(6.4) and the curves of the mean flow stress with the inverse pass deformation
temperature plotted as shown in the following figure 7.13. Two straight lines can be
obtained here: one has a lower slope at the higher temperature range and the other has
a higher slope at the lower temperature range. When the deformation takes place
below the Tnr, the slope in figure 7.13 increases because of the retained strain
hardening within the non-recrystallisation region. As may be seen in figure 7.13, a
point of intersection is obtained, which is the non-recrystallisation temperature
marked by an arrow (Tnr).
140
120
100
M e a n flo w stre ss, M P a
M e a n flo w s tre s s , M P a
120
140
p a s s s t r a i n : 0. 15 s t r a i n r a t e : 1/s
interpass time:8s
80
60
40
100
(a)
20
0
0.65
0.7
0.75
0.8
0.85
p a s s s t r a i n : 0. 2 0 s t r a i n r a t e : 1/s
interpass time:8s
80
60
40
(b)
20
0
0.65
0.9
0.7
0.75
-1
1 0 0 0 / T (K )
M e a n flo w stre ss, M P a
M ean flow stress, M P a
140
p a s s s t r a i n : 0. 24 s t r a i n r a t e : 1/s
interpass time:8s
100
80
60
40
(c)
20
0
0.65
0.85
0.9
0.95
160
140
120
0.8
-1
1 0 0 0 / T (K )
0.7
0.75
0.8
1000/T
0.85
0.9
p a s s s t r a i n : 0. 2 8 s t r a i n r a t e : 1/s
interpass time:8s
120
100
80
60
40
(d)
20
0.95
0
0.65
0.7
0.75
0.8
0.85
1 0 0 0 / T (K-1)
(K-1)
- 90 -
0.9
0.95
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
120
120
p a s s s t r a i n : 0. 3 0 s t r a i n r a t e : 1/s
interpass time:8s
100
M e a n flo w s tre s s , M P a
M e a n flo w s tre s s , M P a
100
80
60
40
p a s s s t r a i n : 0. 3 2 s t r a i n r a t e : 1/s
interpass time:8s
80
60
40
20
20
(f)
(e)
0
0.65
0.7
0.75
0.8
0.85
0.9
0
0.65
0.95
0.7
0.75
0.8
0.85
0.9
0.95
-1
1 0 0 0 / T (K )
-1
1 0 0 0 / T (K )
Figure 7.13 Determination of the Tnr on the mean flow stress versus inverse
temperature curves of alloy #6, all deformed at the same strain rate of 1 s-1 and an
inter-pass time of 8 seconds but at different pass strains.
As may be seen in figure 7.13, the slope of the straight line below the Tnr is lower at a
low pass strain of 0.15 than at a strain of 0.32 per pass. The slope of the straight line
below the Tnr, therefore, appears to increase with an increase in pass strain. This is
because the accumulation of retained strain from pass to pass during nonrecrystallisation, is higher at higher pass strains. The higher pass strain induces a
higher density of dislocations in the flattened austenite grains, which will remain
below the Tnr, until transformation to ferrite occurs on the run-out table. The results of
the Tnr at various pass strains are listed in table 7.4 for alloy #6 and are also shown in
figure 7.14.
Table 7.4 The non-recrystallisation temperature and pass strains of alloy #6
Sample number
A15
A20
A24
A28
A30
A32
Inter-pass time, tip (s)
8
8
8
8
8
8
Strain rate (s-1)
1
1
1
1
1
1
Pass strain ε
0.15
0.2
0.24
0.28
0.30
0.32
Tnr (ºC)
937
931
922
916
918
893
- 91 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
950
Tnr, ºC
940
930
920
910
900
890
0.15
0.2
0.25
0.3
0.35
pass strain
Figure 7.14 The relationship between pass strain (ε) and the non-recrystallisation
temperature for alloy #6. Strain rate ε& =1.0 s-1, inter-pass time tip=8 s.
The strain rate and inter-pass time were held constant at ε& =1.0 s-1 and tip=8 s in this
series of tests. The Tnr in the Mo-free reference alloy #6 decreases with increasing
pass strain in an approximately linear relationship. The tendency of the Tnr versus pass
strain is consistent with the results of Cuddy et al[128] and Bai et al[48,125], i.e., the Tnr
decreases with an increase in pass strain. This quantitative relationship in alloy #6 can
be described by the following equation:
R2=0.79
Tnr= –210 ε+972
(7.4)
where Tnr is in ºC and the strain ε expressed as the true strain.
Above the Tnr grain refinement significantly increases the dislocation density during
hot working and this encourages the coarsening of precipitates[111]. As the pass strain
increases, austenite grains become finer. The higher dislocation density is associated
with a higher stored energy in the austenite, which constitutes the driving force for
recrystallisation.
Therefore,
increasing
the
pass
strain
promotes
austenite
recrystallisation and lowers the Tnr. Furthermore, the coarser particles have little effect
on retarding the austenite recrystallisation[125]. Weiss and Jonas[112], Bai et al[125] and
Speer and Hansen[113] reported that precipitate coarsening during hot deformation
takes place as the pass strain increases and these larger particles lose their ability to
retard recrystallisation.
- 92 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
7.3.2 The Tnr and inter-pass time
The inter-pass time also affects the non-recrystallisation temperature during the hot
rolling process. The hot compression results at various inter-pass times ranging from 4
to 50 seconds, with the same pass strain of 0.2 and strain rate of 1 s-1 for alloy #6, are
shown in figure 7.15.
160
180
866℃
interpass time:4s pass strain:0.20
strain rate:1/s
140
895℃
140
120
903℃
928℃
s tre s s ,M P a
120
80
1018℃
1064℃
1108℃
s tre s s ,M P a
100
1037℃
1081℃
988℃
100
1128℃
80
1153℃
60
60
40
40
20
(a)
(b)
20
0
0
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
0
0.2
0.4
0.6
strain
0.8
1
1.2
1.4
1.6
strain
160
160
interpass time:8s pass strain:0.20
strain rate:1/s
140
852℃
902℃
s tre s s ,M P a
100
993℃
1034℃
80
846℃
858℃
120
904℃
100
interpass time:15s pass strain:0.20
strain rate:1/s
140
878℃
120
80
1130℃
60
1087℃
1140℃
1037℃
1007℃
60
40
40
20
20
(c)
(d)
0
0
0
0.2
0.4
0.6
0.8
1
1.2
0
1.4
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
strain
strain n
120
160
interpass time:30s pass strain:0.20
strain rate:1/s
841℃
interpass time:20s pass strain:0.20
strain rate:1/s
140
863℃
100
896℃
120
1029℃
1078℃
80
s tre s s ,M P a
100
837℃
857℃
874℃
898℃
80
s tre s s ,M P a
s tre s s ,M p a
843℃
868℃
interpass time:6s pass strain:0.20
strain rate:1/s
160
983℃
1074℃
1034℃
987℃
1124℃
60
1130℃
60
40
40
20
(f)
(e)
20
0
0
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
0
0.2
0.4
0.6
0.8
1
strain
strain
- 93 -
1.2
1.4
1.6
1.8
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
120
120
interpass time:35s pass strain:0.20
strain rate:1/s
100
840℃
863℃
885℃
835℃
interpass time:40s pass strain:0.20
strain rate:1/s
100
899℃
857℃
875℃
898℃
80
1085℃
1035℃
s tre s s , M P a
s tre s s ,M P a
80
996℃
1138℃
60
1073℃
40
1034℃
992℃
1123℃
60
40
20
20
(g)
(h)
0
0
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
0
0.2
0.4
0.6
0.8
strain
1
1.2
1.4
1.6
1.8
strain
140
interpass time:50s pass strain:0.20
strain rate:1/s
120
837℃
859℃
880℃
s tre s s , M P a
100
901℃
80
60
1085℃
1039℃
1001℃
1159℃
40
20
(i)
0
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
strain
Figure 7.15 Stress-strain curves of multi-pass compression tests on alloy #6 at a
constant pass strain of 0.20, a constant strain rate of 1 s-1 and a series of inter-pass
times ranging from 4 to 50 seconds.
The number of passes in these tests is 6, 7 or 8 (some results were obtained on the
Gleeble before its upgrading, others after the upgrading. From there a difference in
number of passes arose). The pass temperature is also marked on the respective
figures. The mean flow stress values calculated from figure 7.15 is plotted against the
inverse pass temperature (in K) in figure 7.16.
- 94 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
140
160
i n t e r p a s s t i m e: 4 s p a s s s t r a i n : 0.2 0
s t r a i n r a t e : 1/s
i n t e r p a s s t i m e: 6 s p a s s s t r a i n : 0.2 0
s t r a i n r a t e : 1/s
140
M e a n flo w s tre s s , M P a
M e an flo w stre s s, M P a
120
180
100
80
60
40
120
100
80
60
40
(a)
20
0
0.65
0.7
0.75
0.8
0.85
(b)
20
0
0.65
0.9
0.7
0.75
1 0 0 0 / T (K-1)
140
80
60
40
80
60
40
(c)
(d)
0.7
0.75
0.8
0.85
0.9
0
0.65
0.95
0.7
0.75
160
0.85
0.9
0.95
140
i n t e r p a s s t i m e: 2 0 s p a s s s t r a i n : 0.2 0
s t r a i n r a t e : 1/s
120
M e a n flo w s tre s s , M P a
120
M e a n flo w s tre s s , M P a
0.8
-1
1 0 0 0 / T (K )
1 0 0 0 / T (K-1)
100
80
60
40
20
0
0.65
0.95
100
20
20
140
0.9
i n t e r p a s s t i m e: 15 s p a s s s t r a i n : 0.2 0
s t r a i n r a t e : 1/s
120
100
0
0.65
0.85
140
i n t e r p a s s t i m e: 8 s p a s s s t r a i n : 0.2 0
s t r a i n r a t e : 1/s
M e a n flo w s tre s s ,M P a
M e a n flo w s tre s s , M P a
120
0.8
-1
1 0 0 0 / T (K )
(e)
0.7
0.75
0.8
0.85
0.9
0.95
-1
1 0 0 0 / T (K )
i n t e r p a s s t i m e: 3 0 s p a s s s t r a i n : 0.2 0
s t r a i n r a t e : 1/s
100
80
60
40
(f)
20
0
0.65
0.7
0.75
0.8
1000/T
- 95 -
0.85
(K-1)
0.9
0.95
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
120
140
i n t e r p a s s t i m e: 3 5 s p a s s s t r a i n : 0.2 0
s t r a i n r a t e : 1/s
100
M e a n flo w s tre s s , M P a
M ea n flo w stre ss, M P a
120
100
80
60
40
i n t e r p a s s t i m e: 4 0 s p a s s s t r a i n : 0.2 0
s t r a i n r a t e : 1/s
80
60
40
20
20
(h)
(g)
0
0.65
0
0.65
0.7
0.75
0.8
0.85
0.9
0.95
0.7
0.75
0.8
0.85
1000/T
-1
1 0 0 0 / T (K )
0.9
0.95
(K-1)
120
M e a n flo w s tre s s , M P a
100
i n t e r p a s s t i m e: 5 0 s p a s s s t r a i n : 0.2 0
s t r a i n r a t e : 1/s
80
60
40
20
(i)
0
0.65
0.7
0.75
0.8
0.85
0.9
0.95
-1
1 0 0 0 / T (K )
Figure 7.16 The mean flow stress versus inverse temperature curve of alloy #6 during
multi-pass compression testing at a constant pass strain of 0.20 and a constant strain
rate of 1 s-1 but with a variation of the inter-pass times between 4 and 50 seconds.
The values for Tnr obtained from figure 7.16 at various inter-pass times are
summarised in table 7.5 and figure 7.17.
Table 7.5 The non-recrystallisation temperature of alloy #6 as affected by different
inter-pass times
Sample
number
D4a/4b
D6
D8
D15a/15b
D20
D30
D35
D40 D50
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
Strain rate (s )
1
1
1
1
1
1
1
1
1
Inter-pass
time, tip (s)
4
6
8
15
20
30
35
40
50
938
931
930
914
916
921
913
Pass strain, ε
-1
Tnr (ºC)
934/
956
941/
921
- 96 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
960
955
Tnr, ºC
950
945
940
935
930
925
920
915
910
0
10
20
30
40
50
60
tip, s
Figure 7.17 The Tnr as a function of inter-pass time (tip) for alloy #6. Strain rate ε& =1.0
s-1, pass strain ε = 0.2.
It can be seen that the Tnr decreases with increasing inter-pass time from 4 to 50
seconds in alloy #6. Although a large degree of scatter was present, this dependence
can be described approximately by the following equation:
Tnr=961tip-0.0128
R2=0.68
(7.5)
The result is different from that of Bai[48,125]. Bai reported that the dependence of the
Tnr can be divided into three distinct regions: short inter-pass times (less than 12.5
seconds), medium inter-pass times (from 12.5 to 80 seconds) and long inter-pass
times (more than 80 seconds). Within the short inter-pass times, the Tnr decreases with
an increase in inter-pass time up to 12.5 seconds because solute drag by Nb on
austenite grain boundaries retards recrystallisation prior to precipitation of NbC.
Within the medium inter-pass times, precipitation of NbC occurs and retards
recrystallisation and the Tnr then increases with an increase in inter-pass time. With a
further increase in inter-pass time, the precipitates become coarse and become less
effective in retarding the recrystallisation, so that the Tnr decreases again with interpass time. Some researchers[17,27,129] supported this finding of Bai while another[106]
supported the key mechanism of a process of strain induced precipitation.
In this study, it was found that the Tnr decreases with increasing inter-pass time within
the range of 4 to 50 seconds, probably with both solute drag and precipitation that
- 97 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
affect the recrystallisation. The dependence of the Tnr on inter-pass time can be
divided into two regions: short inter-pass times (tip<20 s) and long inter-pass time
(20<tip<50 s). In the short inter-pass times, deformation induces a high density of
dislocations and point defects[130] that are beneficial to the nucleation of austenite
recrystallisation. Solute drag by Nb on austenite grain boundaries prior to
precipitation plays only a retarding role in recrystallisation, which probably is weaker
in this range, while dislocations and point defects provide enough driving force for
recrystallisation. This means that more recrystallisation will takes place with
increasing inter-pass time, leading to a sharp decrease in the Tnr with an increase in
inter-pass time. With a further increase in inter-pass time, i.e. more than 20 seconds,
the effect of inter-pass times on the Tnr becomes less. Precipitation takes place and the
precipitates are still fine (coarsening is only significant after 80 seconds[125]) in this
stage as there is enough time for Nb(C,N) or NbN nucleation on the dislocations
introduced from the deformation[130]. These finer precipitates pin the grain boundaries,
so that the retarding effectiveness of Nb(C,N) or NbN particles on the recrystallisation
is significant. This means that the effect of inter-pass time could become weaker at
longer times. Figure 7.17 showed that the effect of the precipitation on the
recrystallisation is, therefore, probably stronger than that of solute drag. However, the
effect of the precipitates of Nb(C,N) or NbN appears to be not greater than that found
by Bai[125].
- 98 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
7.3.3 The Tnr and pass strain rate
The Tnr can also be affected, besides by the pass strain and inter-pass time, also by the
pass strain rate during compression deformation. A group of tests at a pass strain rate
ranging from 0.1 to 2.22 s-1 was carried out at a constant pass strain of 0.20 and a
constant inter-pass time of 8 seconds for alloy #6 and the results are shown in figure
7.18.
180
160
842℃
140
strain rate:0.1/s
interpass time:8s pass strain:0.20
strain rate:0.47/s
interpass time:8s pass strain:0.20
160
862℃
899℃
120
s tre s s , M P a
s tre s s ,M P a
892℃
100
100
977℃
1076℃
1131℃
846℃
140
120
80
866℃
1073℃
80
1030℃
60
1033℃
984℃
1136℃
60
40
40
20
20
(a)
0
0.2
0.4
0.6
0.8
1
1.2
1.4
(b)
0
0
0
1.6
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
strain
strain
160
160
843℃
strain rate:0.9/s
interpass time:8s pass strain:0.20
140
120
850℃
877℃
120
901℃
s tre s s , M P a
905℃
s tre s s , M P a
strain rate:1.22/s
interpass time:8s pass strain:0.20
140
872℃
100
100
80
991℃
1042℃
1086℃
1143℃
60
990℃
80
1039℃
1083℃
1123℃
60
40
40
20
20
(c)
0
(d)
0
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
0
0.2
0.4
0.6
0.8
strain
1
1.2
1.4
1.6
strain
160
180
160
843℃
strain rate:1.38/s
interpass time:8s pass strain:0.20
140
852℃
strain rate:1.67/s
interpass time:8s pass strain:0.20
140
872℃
873℃
120
901℃
905℃
s tre s s , M P a
s tre s s , M P a
120
100
100
1143℃
80
1086℃
991℃
80
1042℃
1090℃
1129℃
1038℃
988℃
60
60
40
40
20
20
(e)
0
0.2
0.4
0.6
0.8
1
1.2
1.4
(f)
0
0
1.6
0
0.2
0.4
0.6
0.8
strain
strain
- 99 -
1
1.2
1.4
1.6
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
160
160
842℃
strain rate:1.8/s
interpass time:8s pass strain:0.20
140
120
1032℃
1127℃
80
903℃
100
983℃
1072℃
846℃
862℃
120
s tre s s , M P a
s tre s s , M P a
898℃
100
strain rate:2.0/s
interpass time:8s pass strain:0.20
140
857℃
80
60
1036℃
1078℃
984℃
1127℃
60
40
40
20
20
(g)
0
(h)
0
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
0
0.2
0.4
0.6
strain
0.8
1
1.2
1.4
1.6
strain
140
strain rate:2.22/s
interpass time:8s pass strain:0.20
120
853℃
874℃
100
s tre s s , M P a
901℃
989℃
80
1036℃
1082℃
1126℃
60
40
20
(i)
0
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
strain
Figure 7.18 Stress-strain curves of compression deformation tests of alloy #6 at a
constant pass strain of 0.20 and inter-pass time of 8 s but with a series of strain rates
from 0.1 to 2.22 s-1.
- 100 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
The number of passes in this series of tests was seven. The pass deformation
temperatures are also marked on figure 7.18. The mean flow stresses calculated from
figure 7.18 are plotted against the inverse of pass temperature (in K) in figure 7.19.
140
180
s t r a i n r a t e : 0 .1/s
i n t e r p a s s t i m e : 8 s p a s s s t r a i n : 0.2 0
160
s t r a i n r a t e : 0 .4 7 / s
i n t e r p a s s t i m e : 8 s p a s s s t r a i n : 0.2 0
140
M e a n flo w s tre s s , M P a
M e a n flo w s tre s s , M P a
120
100
80
60
40
120
100
80
60
40
20
0
0.65
0.7
0.75
0.8
0.85
0.9
(b)
20
(a)
0
0.65
0.95
0.7
0.75
0.8
strain rate:1.22/s
i n t e r p a s s t i m e : 8 s p a s s s t r a i n : 0.2 0
120
M e a n flo w s tre s s, M P a
M e a n flo w s tre ss , M P a
s t r a i n r a t e : 0 .9 / s
i n t e r p a s s t i m e : 8 s p a s s s t r a i n : 0.2 0
100
80
60
40
0.7
0.75
0.8
0.85
0.9
80
60
40
20
(c)
20
(d)
0
0.65
0.95
0.7
0.75
160
120
M e a n flo w s tre s s , M P a
M e a n flo w s tre s s , M P a
0.85
0.9
0.95
140
strain rate:1.38/s
i n t e r p a s s t i m e : 8 s p a s s s t r a i n : 0.2 0
120
100
80
60
40
strain rate: 1.67/s
i n t e r p a s s t i m e : 8 s p a s s s t r a i n : 0.2 0
100
80
60
40
20
20
0
0.65
0.8
1 0 0 0 / T (K-1)
-1
1 0 0 0 / T (K )
140
0.95
140
100
0
0.65
0.9
1 0 0 0 / T (K )
140
120
0.85
-1
1 0 0 0 / T (K-1)
(e)
0.7
0.75
0.8
0.85
0.9
0.95
0
0.65
(f)
0.7
0.75
0.8
0.85
-1
1 0 0 0 / T (K )
1 0 0 0 / T (K-1)
- 101 -
0.9
0.95
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
140
160
strain rate:1.8/s
i n t e r p a s s t i m e : 8 s p a s s s t r a i n : 0.2 0
120
M e a n flo w s tre s s , M P a
120
100
80
60
40
20
0
0.65
0.7
0.75
0.8
0.85
0.9
strain rate:2.0/s
i n t e r p a s s t i m e : 8 s p a s s s t r a i n : 0.2 0
100
80
60
40
20
(g)
(h)
0
0.65
0.95
0.7
0.75
1 0 0 0 / T (K-1)
0.8
0.85
0.9
0.95
-1
1 0 0 0 / T (K )
120
strain rate:2.22/s
i n t e r p a s s t i m e : 8 s p a s s s t r a i n : 0.2 0
100
M e a n flo w s tre s s , M P a
M e a n flo w s tre s s , M P a
140
80
60
40
20
(i)
0
0.65
0.7
0.75
0.8
1000/T
0.85
0.9
0.95
(K-1)
Figure 7.19 The mean flow stress versus inverse test temperature of alloy #6 at a
constant pass strain of 0.20 and inter-pass time of 8 s but at a series of strain rates
from 0.1 to 2.22 s-1.
Table 7.6 The non-recrystallisation temperature and strain rates of alloy #6
Sample number
C01
C05
C1
C15
C2
C3
C4
C6
C8
Inter-pass time,
tip (s)
8
8
8
8
8
8
8
8
8
Pass strain, ε
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
Strain rate (s-1)
0.1
0.47
0.9
1.22
1.38
1.67
1.80
2.0
2.22
Tnr (ºC)
914
915
921
929
941
918
920
920
923
The results of the Tnr at different stain rates (with ε and tip constant) are listed in table
7.6 and are plotted in figure 7.20. The austenite should strain harden more at higher
- 102 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
strain rates during a compression test and this should result, theoretically at least, in a
larger driving force for recrystallisation and, therefore, should decrease the Tnr.
However, the opposite tendency was found in alloy #6 (see figure 7.20) and is also
different from the results of Bai et al.[125] and Laasraoui[124]. They reported that the Tnr
decreases with an increase in strain rate when the pass strains are 0.2, 0.3, 0.4 and 0.5.
This inconsistency between these results and those obtained by others, may have to be
examined further by torsion tests which allow higher strain rates than the highest
value of 2.22 s-1 used here.
990
970
950
930
Tnr, ºC
910
890
870
850
0
0.5
1
1.5
strain rate, s
2
2.5
-1
Fig.7.20 Strain rate ( ε& ) versus the non-recrystallisation temperature for alloy #6.
Pass strain ε = 0.2, inter-pass time tip=8 s.
- 103 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
7.4. Continuous cooling transformation (CCT diagrams) under strain-free
conditions
The aim in manufacturing technology of line pipe steel is to achieve a balance of high
strength and toughness. Many researchers[32,65,66,92] have found that the optimal
microstructure consists mainly of an acicular ferrite microstructure for optimised
mechanical properties in micro-alloyed HSLA steels. Nb and Mo as micro-alloying
elements[3,4,32] and thermo-mechanical processes[4,131] can affect the transformation to
an acicular ferrite microstructure in line pipe steels. It is important to understand the
continuous cooling transformation behaviour of Nb-Mo-Ti micro-alloyed steels, in
particular whether an optimal fraction of an acicular ferrite microstructure can be
obtained through micro-alloying additions and the hot rolling process. The following
sections discuss the CCT diagrams for the experimental alloys #5 (with 0.22% Mo)
and #6 (Mo-free) under conditions with and without prior deformation.
7.4.1 CCT diagram for alloy #6 without molybdenum and without prior
deformation
The transformed microstructures of alloy #6 without molybdenum and prior
deformation are given in figure 7.21.
Cooling rate= 0.1 ºCs-1
Cooling rate= 0.1 ºCs-1
PF
PF
P
P
(a) Polygonal ferrite +pearlite
(b) Polygonal ferrite +pearlite
- 104 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
Cooling rate= 0.2 ºCs-1
Cooling rate= 0.2 ºCs-1
PF
P
PF
P
(c) Polygonal ferrite +pearlite
(d) Polygonal ferrite +pearlite
Cooling rate= 0.5 ºCs-1
Cooling rate= 0.5 ºCs-1
P
PF
PF
P
AF
AF
(e) Polygonal ferrite +pearlite + acicular (f) Polygonal ferrite +pearlite + acicular
ferrite
ferrite
Cooling rate= 1 ºCs-1
Cooling rate= 1 ºCs-1
PF
PF
P
AF
AF
(g) Polygonal ferrite + pearlite +
acicular ferrite
P
(h) Polygonal ferrite +pearlite + acicular
ferrite
- 105 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
Cooling rate= 2 ºCs-1
Cooling rate= 2 ºCs-1
AF
PF
PF
B
(i) Polygonal ferrite +bainite
Cooling rate= 5 ºCs-1
(j) Polygonal ferrite +acicular ferrite
Cooling rate= 5 ºCs-1
B
B
(k) Bainite
(l) Bainite
Cooling rate= 8 ºCs-1
Cooling rate= 8 ºCs-1
B
B
(m) Bainite
(n) Bainite
- 106 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
Cooling rate= 10 ºCs-1
Cooling rate= 10 ºCs-1
B
B
(o) Bainite
Cooling rate= 20 ºCs-1
(p) Bainite
Cooling rate= 20 ºCs-1
B
B
(q) Bainite
(r) Bainite
Figure 7.21 The optical micrographs (etched in 2% Nital) of the Mo-free alloy #6 and
with no prior deformation after continuous cooling. PF-polygonal ferrite, P-pearlite
and AF-acicular ferrite microstructure.
The transformed microstructures are affected by the different cooling rates and
contain polygonal ferrite, pearlite, an acicular ferrite microstructure or bainitic ferrite.
The microstructures contained polygonal and pearlite at low cooling rates, while an
acicular ferrite microstructure was obtained at medium cooling rates ranging from
about 0.5 to 5 ºCs-1. The bainite phase occurred at a cooling rate above about 2 ºCs-1
while bainite was practically the only phase above a rate of 5 ºCs-1.
The CCT diagrams (with no prior deformation) were constructed through dilatometry
and the optical micrographs (figure7.21) after cooling. The CCT diagram of the Mofree alloy #6 is given in figure 7.22.
- 107 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
The areas of polygonal ferrite, pearlite, an acicular ferrite microstructure and bainite
are identified as PF, P, AF and B in the CCT diagram (figure 7.22), respectively. As
may be seen, a 100% acicular ferrite microstructure could not be obtained for the
reference alloy #6 without prior deformation. A mixture of polygonal ferrite, pearlite
and an acicular ferrite microstructure was found within the range of cooling rates
between 0.3 and 2 ºCs-1, and polygonal ferrite, an acicular ferrite microstructure plus
bainite in the range 1.5 to 5 ºCs-1. The acicular ferrite microstructure constituent was
formed within the cooling rate range of 0.3 to 5 ºCs-1 for this Nb-bearing alloy #6
while a 100% bainite transformation occurred above 5 ºCs-1.
1000
Temperature,
ºC
800
Ac3=871ºC
PF
Ac1=762ºC
P
AF
600
B
400
200
Cooling rate: (ºCs-1)
20
10 8
5
2
1
0.5
0.2 0.1
0
100
101
102
103
104
Time, s
Figure 7.22 The CCT diagram of the Mo-free alloy #6 and no prior deformation. PFpolygonal ferrite, P-pearlite, AF-acicular ferrite microstriucture and B-bainite.
- 108 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
7.4.2 CCT diagram for alloy #5 with 0.22% molybdenum and without prior
deformation
The microstructures for alloy #5 with 0.22% Mo and without deformation and
transformed at different cooling rates are given in figure 7.23.
Cooling rate=0.1 ºCs-1
Cooling rate=0.1 ºCs-1
PF
AF
P
PF
(a) polygonal ferrite + pearlite
(b) polygonal ferrite + acicular ferrite
Cooling rate=0.2 ºCs-1
Cooling rate=0.2 ºCs-1
PF
PF
AF
P
(c) polygonal ferrite + pearlite
Cooling rate=0.5 ºCs-1
(d) polygonal ferrite + acicular ferrite
Cooling rate=0.5 ºCs-1
B
PF
AF
(e) polygonal ferrite + acicular ferrite
(f) bainite + polygonal ferrite
- 109 -
PF
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
Cooling rate=1 ºCs-1
Cooling rate=1 ºCs-1
PF
AF
B
(g) polygonal ferrite + bainite
(h) acicular ferrite
Cooling rate=2 ºCs-1
Cooling rate=2 ºCs-1
PF
(i) polygonal ferrite + bainite
Cooling rate=5 ºCs-1
PF+AF
B
(j) polygonal ferrite + acicular ferrite
Cooling rate=5 ºCs-1
PF+AF
B
(k) bainite
(l) polygonal ferrite + acicular ferrite
- 110 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
Cooling rate=8 ºCs-1
Cooling rate=8 ºCs-1
B
(m) bainite
PF+AF
(n) polygonal ferrite + acicular ferrite
Cooling rate=10 ºCs-1
Cooling rate=10 ºCs-1
B
AF
(o) acicular ferrite
(p) bainite
Cooling rate=20 ºCs-1
Cooling rate=20 ºCs-1
B
B
(q) bainite
(r) bainite
Figure 7.23 The optical micrographs (etched in 2% Nital) for alloy #5 (with 0.22%
Mo) and with no prior deformation after continuous cooling. PF-polygonal ferrite, Ppearlite and AF-acicular ferrite microstructure.
- 111 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
1000
800
Ac1=782ºC
ºC
Temperature,
Ac3=894ºC
PF
P
AF
600
B
400
200
Cooling rate: (ºCs-1)
20
10 8
5
2
1
0.5
0.2 0.1
0
100
101
102
103
104
Time,s
Figure 7.24 The CCT diagram of alloy #5 (with 0.22% Mo) and with no prior
deformation. PF-polygonal ferrite, P-pearlite, AF-acicular ferrite microstructure and
B-bainite.
The CCT diagram for alloy #5 with 0.22%Mo and without deformation is shown in
figure 7.24. By comparing figure 7.22 without molybdenum and figure 7.24 with
0.22% Mo and with no applied deformation in both cases, the two CCT diagrams
appear very similar. In both cases, a combination of polygonal ferrite and pearlite was
obtained at lower cooling rates. An acicular ferrite microstructure could be obtained in
the microstructures with increasing cooling rates. Bainite was the predominant
constituent at higher cooling rates, replacing the acicular ferrite microstructure. But
closer inspection of the two diagrams reveals some differences between them.
z
An acicular ferrite microstructure could be obtained at a cooling rate above
approximately 0.3 ºCs-1 for the Mo-free alloy #6 while it was obtained already
above about 0.1 ºCs-1 for alloy #5 (0.22% Mo). The presence of molybdenum,
therefore, appears to promote the formation of an acicular ferrite microstructure
to some degree, even without prior deformation.
z
A bainite phase transformation occurred at lower cooling rates of 0.3 ºCs-1 in
alloy #5 with molybdenum than in alloy #6, in which a cooling rate of about 1.5
ºCs-1 was required. This indicates that molybdenum micro-alloying in the alloy
- 112 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
delays the bainite formation to lower cooling rates. Molybdenum, therefore,
shifted the CCT diagram to longer times in the Nb-bearing low carbon alloys
without prior deformation.
z
An cicular ferrite microstructure formed within a wider cooling rate range from
0.1 to 10 ºCs-1 in alloy #5 while in the Mo-free alloy #6, this range was narrower
from 0.3 to 5 ºCs-1. This means that the cooling rate range for an acicular ferrite
microstructure formation for the experimental alloy #5 with 0.22% Mo, is wider
than for the Mo-free alloy #6. Molybdenum can, therefore, expand the region of
an acicular ferrite microstructure formation to some degree. The molybdenum
addition to Nb-containing alloys, therefore, promotes to some degree the
formation of an acicular ferrite microstructure in the microstructures without
prior deformation. An acicular ferrite microstructure containing microstructure is
an optimum microstructure in low carbon low alloyed steels due to its excellent
balance of strength and toughness[92].
z
Therefore, molybdenum additions to steels can result in superior mechanical
properties under conditions of no prior deformation.
z
Molybdenum is an alloying element that also promotes the formation of
polygonal ferrite. The polygonal ferrite region in the CCT diagram of alloy #5
(figure 7.24) is larger than that of the Mo-free alloy #6 (figure 7.22).
z
In contrast, however, the pearlite region in figure 7.24 for alloy #5 is smaller than
that for the Mo-free alloy #6 in figure 7.22 and molybdenum additions, therefore,
appear to act against the formation of pearlite in these steels without any prior
deformation.
7.5 Strain enhanced continuous cooling transformation (CCT diagram) under
deformed conditions
Besides chemical composition, deformation in the austenite phase prior to its
transformation to ferrite, changes the appearance of the continuous cooling
transformation diagram of these alloys. The final phases may vary with deformation
parameters, i.e. temperature, strain or reduction, and cooling rate. It is important,
therefore, to determine the strain affected CCT diagram after prior deformation.
- 113 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
7.5.1 Strain affected CCT diagram of the Mo-free alloy #6
The optical microstructures for alloy #6 (without molybdenum) are shown in figure
7.25 after a single pass strain of 0.6 at 860 ºC with strain rate of 0.5 s-1 and cooling
down to room temperature after prior deformation.
Cooling rate=0.1 ºCs-1
P
Cooling rate=0.1 ºCs-1
PF
P
PF
Deformation direction
(a) polygonal ferrite + pearlite
(b) polygonal ferrite + pearlite
Cooling rate=0.2 ºCs-1
Cooling rate=0.2 ºCs-1
PF
P
PF
P
(c) polygonal ferrite + pearlite
Cooling rate=0.5 ºCs-1
(d) polygonal ferrite + pearlite
Cooling rate=0.5 ºCs-1
P
PF
PF
P
(e) polygonal ferrite + pearlite
(f) polygonal ferrite + pearlite
- 114 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
Cooling rate=1 ºCs-1
Cooling rate=1 ºCs-1
P
AF
AF
PF
PF
(g) polygonal ferrite + acicular ferrite +
(h) polygonal ferrite + acicular ferrite
pearlite
Cooling rate=2 ºCs-1
Cooling rate=2 ºCs-1
AF
PF
P
PF
P
AF
(i) polygonal ferrite + acicular ferrite +
(j) polygonal ferrite + acicular ferrite +
pearlite
pearlite
Cooling rate=5 ºCs-1
Cooling rate=5 ºCs-1
AF
AF
PF
PF
(k) polygonal ferrite + acicular ferrite
(l) polygonal ferrite + acicular ferrite
- 115 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
Cooling rate=8 ºCs-1
Cooling rate=8 ºCs-1
AF
AF
PF
PF
(m) polygonal ferrite + acicular ferrite
(n) polygonal ferrite + acicular ferrite
Cooling rate=10 ºCs-1
Cooling rate=10 ºCs-1
AF
AF
PF
(o) polygonal ferrite + acicular ferrite
PF
(p) polygonal ferrite + acicular ferrite
Cooling rate=20 ºCs-1
Cooling rate=20 ºCs-1
AF
PF
PF
AF
(q) polygonal ferrite + acicular ferrite
(r) polygonal ferrite + acicular ferrite
- 116 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
Cooling rate=40 ºCs-1
Cooling rate=40 ºCs-1
PF+AF
PF+AF
(s) polygonal ferrite + acicular ferrite
(t) polygonal ferrite + acicular ferrite
Figure 7.25 The optical micrographs (etched in 2% Nital) of the Mo-free alloy #6
after compression testing with a single pass strain of 0.6 at 860 ºC (which is below the
Tnr), a strain rate of 0.5 s-1 and cooling down to room temperature at different cooling
rates. PF-polygonal ferrite, P-pearlite and AF-acicular ferrite microstructure.
1000
Temperature, ºC
800
PF
P
600
AF
400
200
Cooling rate: (ºCs-1)
40
20
10 8
5
2
0.5
1
0.2
0.1
0
100
101
102
103
104
Time, s
Figure 7.26 The strain affected CCT diagram of the Mo-free alloy #6 after a single
pass compression strain of 0.6 at 860 ºC with a strain rate of 0.5 s-1. PF-polygonal
ferrite, P-pearlite and AF-acicular ferrite microstructure.
- 117 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
The strain affected CCT diagram of alloy #6 after a single pass strain deformation of
0.6 (equivalent to a 45% reduction) at 860 ºC with strain rate of 0.5 s-1 is shown in
figure 7.26.
The micrographs of the Mo-free alloy #6 after compression to a strain of 0.6 show a
combination of two or three of the phases of polygonal ferrite, pearlite and an acicular
ferrite microstructure. This is different from the case of the same alloy but without
prior deformation (figure 7.22) where bainite could be obtained at high cooling rates
without deformation in this alloy. With prior deformation, however, no bainite is
found at all cooling rates ranging from 0.1 to 40 ºCs-1 (see figures 7.25 and 7.26). This
observation is consistent with the results of others[4,131] and it indicates that bainite
transformation is restrained from forming in deformed austenite[132,133]. When a
sample is deformed in the austenite region, more dislocations are formed in the
austenite and these may act as obstacles in retarding the growth of a bainite packet or
sheath in the deformed austenite. In undeformed austenite, on the other hand, there are
less of these obstacles during bainite phase formation and as a result, bainite can be
found at high enough cooling rates (figure 7.22).
7.5.2 Strain affected CCT diagram of alloy #5 (with 0.22% Mo)
Figure 7.27 shows the results for alloy #5 (with 0.22% Mo) after a prior deformation
with a single pass strain of 0.6 at 860 ºC and at a strain rate of 0.5 s-1.
Cooling rate= 0.1 ºCs-1
Cooling rate= 0.1 ºCs-1
PF
P
P
PF
Deformation direction
(a) polygonal ferrite + pearlite
(b) polygonal ferrite + pearlite
- 118 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
Cooling rate= 0.2 ºCs-1
Cooling rate= 0.2 ºCs-1
PF
PF
P
P
(c) polygonal ferrite + pearlite
(d) polygonal ferrite + pearlite
Cooling rate= 0.5 ºCs-1
Cooling rate= 0.5 ºCs-1
PF
PF
P
AF
(e) polygonal ferrite + pearlite
P
(f) polygonal ferrite + acicular ferrite +
pearlite
Cooling rate= 1 ºCs-1
Cooling rate= 1 ºCs-1
AF
P
AF
P
PF
(g) polygonal ferrite + pearlite +
PF
(h) polygonal ferrite + pearlite +
acicular ferrite
acicular ferrite
- 119 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
Cooling rate= 2 ºCs-1
Cooling rate= 2 ºCs-1
AF
PF
P
AF
PF
(i) polygonal ferrite + pearlite +
(j) acicular ferrite + polygonal ferrite
acicular ferrite
Cooling rate= 5 ºCs-1
Cooling rate= 5 ºCs-1
AF
PF
PF
AF
(k) acicular ferrite + polygonal ferrite
(l) acicular ferrite + polygonal ferrite
Cooling rate= 8 ºCs-1
Cooling rate= 8 ºCs-1
AF
AF
PF
PF
(m) acicular ferrite + polygonal ferrite
(n) acicular ferrite + polygonal ferrite
- 120 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
Cooling rate= 10 ºCs-1
Cooling rate= 10 ºCs-1
AF
PF
AF
PF
(o) acicular ferrite + polygonal ferrite
Cooling rate= 20 ºCs-1
(p) acicular ferrite + polygonal ferrite
Cooling rate= 20 ºCs-1
PF
PF
AF
AF
(q) acicular ferrite + polygonal ferrite
Cooling rate= 40 ºCs-1
(r) acicular ferrite + polygonal ferrite
Cooling rate= 40 ºCs-1
PF
PF
AF
AF
(s) acicular ferrite + polygonal ferrite
(t) acicular ferrite + polygonal ferrite
Figure 7.27 The optical micrographs (etched in 2% Nital) of alloy #5 (with 0.22% Mo)
after a single pass compression of 0.6 strain at 860 ºC and at a strain rate of 0.5 s-1 and
then cooling at different rates.
- 121 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
1000
Temperature, ºC
Deg C
800
PF
P
600
AF
400
200
Cooling rate, ºCs-1
40
20
10 8 5
2
1
0.5
0.2
0.1
0
100
101
102
103
104
Time,s
Figure 7.28 The strain affected CCT diagram of alloy #5 (with 0.22% Mo) after a
single pass compression of 0.6 strain at 860 ºC at a strain rate of 0.5 s–1. PF-polygonal
ferrite, P-pearlite and AF-acicular ferrite microstructure.
Figure 7.28 shows that the continuous cooling transformation for alloy #5 (with
0.22% Mo), is similar to alloy #6, the Mo-free alloy. Comparing figures 7.26 and 7.28,
an acicular ferrite microstructure is formed at high cooling rates after prior
deformation instead of bainite if compared to the case of no prior deformation. It
appears that deformation in the austenite region is beneficial to the formation of an
acicular ferrite microstructure for Mo-bearing alloys as it impedes the transformation
to bainite. It seems, furthermore, that molybdenum additions did not markedly affect
the formation of an acicular ferrite microstructure while the prior deformation in
austenite had a dominant contribution. This may be from more dislocations resulting
from the prior deformation that offered more intragranular nucleation sites for an
acicular ferrite microstructure.
- 122 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
Therefore, the following conclusions can be drawn:
•
Deformation in austenite promotes an acicular ferrite microstructure
formation and hinders bainite formation.
•
The effect of molybdenum is less than the effect of prior deformation on
promoting an acicular ferrite microstructure formation.
•
Bainite could be observed at high cooling rates in steels without prior
deformation.
- 123 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
7.6 The results of the laboratory hot rolling process on the YS/UTS ratio
The five alloys #1 to #5 were hot rolled with the same pass numbers, pass strains, pass
strain rates and the same total strain in rough and finish rolling on a laboratory mill
while the same reheating conditions of 1225 ºC for 60 min were also used. The first
rough rolling temperature and the finish rolling temperature, respectively, were very
similar for all five experimental alloys. The accelerated cooling rate after finish
rolling was also the same at about 47 ºCs-1, for all five experimental alloys, and was
39 ºCs-1 for the Mittal steel alloy #6. The simulated coiling after cooling was at a
temperature of 600 ºC for 24 hrs. The detailed variables in the hot rolling process for
alloys #1 are listed in tables 7.7.
Table 7.7 The laboratory hot rolling parameters for alloy #1
Pass No.
R1
R2 reheating R3
R4
R5
F1 reheating F2 reheating F3
Tempera- in 1190 1060 1225 1140 1070 1020 910
5 min
ture (ºC)
out
tip (s)
Gauge
(mm)
Pass ε
23
--
16
14
43
37
28
out 37
28
in
0.15 0.28
16
930
5 min
--
890
930
5 min
890
860
--
--
20 13.6 10.3
8.3
6.9
20 13.6 10.3 8.3
6.9
6
0.34 0.38 0.28 0.22
0.18
0.14
Total ε
1.43
0.54
Reduction(%)
76
42
-1
ε& (s )
1.67 2.43
3.15 3.92 4.00 4.07
4.00
3.89
N.B.: The R and F in the table stand for roughing and finishing passes, respectively.
The precise details for the other five alloys #2 to #6, which were very similar to those
of alloy #1, are shown in Appendix A. Sample #M1-11 of alloy #6 was taken as a
reference to study the microstructures and carry out TEM work.
As can be seen in the table above, the pass strain in the rough rolling stage is more
than 0.2 (except for the first pass) in order to obtain a fine recrystallised austenite
grain size. The heavy total reduction for alloys #1 to #5, i.e. 76% in rough rolling,
42% in finish rolling and 86% in total, contributes to an optimum combination of
strength and toughness through a fine size. Rapid cooling of 47 ºCs-1 is useful for the
- 124 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
formation of an acicular ferrite microstructure. The alloy #6 was also used to study
the effect of cooling rate, coiling temperature and deformation vis-à-vis the ratio of
yield strength to ultimate tensile strength (see Appendix H).
7.7 Volume fraction of acicular ferrite
The volume fraction of acicular ferrite was measured by image-analysis software on
shadowed carbon extraction replicas from TEM micrographs taken at 980 times
magnification on the experimental alloys #1 to #5. The results are shown in table 7.8
below. Alloy #1 is similar to the Mo-free reference alloy #6 in chemical composition,
whereas alloys #2 to #5 contained varying amounts of molybdenum.
Table 7.8 Measured results of volume fraction of acicular ferrite
Alloy #
#1
#2
#3
#4
#5
Molybdenum (%wt)
0.01
0.09
0.09
0.12
0.22
Volume fraction of AF (%)
55.4
46.3
49.4
52.0
46.8
Volume fraction of PF (%)
44.6
53.7
50.6
48.0
53.2
NB. AF-acicular ferrite and PF-polygonal ferrite
Although there appears to be a small effect of molybdenum in affecting the volume
fractions of AF and PF, the effect is relatively small and probably needs further study
to confirm this.
7.8 Mechanical properties
7.8.1 Results of experimental alloys
The mechanical properties of the five experimental alloys that were hot rolled in the
laboratory mill (see section 7.6 above) are given in table 7.9 and the individual curves
of strength versus elongation are shown in Appendix G. The longitudinal and
transverse (to the rolling direction) values in the table were the average of five
samples per direction per alloy except for only four samples for alloy #5 (both
directions) and the transverse direction only for alloy #1. The yield strength was taken
as the proof strength at 0.5% permanent strain for continuous yielding.
- 125 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
Table 7.9 Mechanical properties of the experimental alloys
Sample
YS,
UTS,
direction
MPa
MPa
Longitudinal
463
Transverse
Yield
YS/UTS
Elongation
(%)
555
0.83
33
D
454
547
0.83
31
D
Longitudinal
504
589
0.86
33
D
Transverse
478
591
0.81
27
D
Longitudinal
467
558
0.84
36
D
Transverse
470
558
0.84
30
D
Longitudinal
472
559
0.84
33
D
Transverse
494
581
0.85
30
D
Longitudinal
492
573
0.86
33
D
Transverse
500
592
0.84
27
C
API
X65
≥448
≥530
≤0.93
≥20.5
specification[12]
X70
≥482
≥565
≤0.93
≥19
X80
≥551
≥620
≤0.93
≥17.5
Alloy #
1
2
3
4
5
type
NB. YS-yield strength, UTS-ultimate tensile strength, D-discontinuous, C-continuous
7.8.2 Results of mechanical properties for different cooling rates, coiling
temperatures and deformation values
7.8.2.1 Effect of cooling rate with no coiling and no prior deformation
The mechanical properties are shown in tables 7.10 and 7.11 for samples subjected to
a series of cooling rates after austenitisation and with no prior deformation and no
simulated coiling process. Controlled cooling was done by means of helium gas
cooling on the Gleeble testing machine at zero load and with conditions of no
deformation. The curves of strength versus elongation are shown in Appendix B and
C for alloys #6 and #3, respectively.
- 126 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
Table 7.10 Mechanical properties of samples #A124 of the Mo-free alloy #6 with no
coiling and no prior deformation
Cooling rate
YS,
UTS,
(ºCs-1)
MPa
MPa
1
439
5
YS/UTS
Yield type
573
0.77
C
461
587
0.78
C
10
491
613
0.80
C
21
551
663
0.83
C
40
603
732
0.82
C
51
648
763
0.85
C
Table 7.11 Mechanical properties of samples #AF3F of alloy #3 (with 0.09% Mo)
and with no coiling and no prior deformation
Cooling rate
YS,
UTS,
(ºCs-1)
MPa
MPa
1
475
5
YS/UTS
Yield type
593
0.80
C
483
579
0.83
C
10
508
606
0.84
C
20
534
625
0.85
C
40
542
640
0.85
C
54
547
648
0.84
C
- 127 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
7.8.2.2 Effect of cooling rate with coiling but without deformation
Results are given in tables 7.12 and 7.13 for the Mo-free alloy #6 treated similarly as
above after austenitisation and without prior deformation but with various cooling
rates and with subsequent coiling simulations at 600 (sample #A113) and 575 ºC
(sample #B113). The individual curves of strength versus elongation are shown in
Appendix D and E for 600 and 575 ºC coiling, respectively.
Table 7.12 Mechanical properties of samples #A113 of the Mo-free alloy #6 with 60
min coiling at 600 ºC without prior deformation
Cooling rate
YS,
UTS,
(ºCs-1)
MPa
MPa
1
494
5
YS/UTS
Yield type
635
0.78
C
478
619
0.77
C
10
492
650
0.76
C
20
502
650
0.77
C
40
500
651
0.77
C
60
500
646
0.77
C
Table 7.13 Mechanical properties of samples #B113 of alloy #6 with 60 min coiling at
575 ºC without prior deformation
Cooling rate
YS,
UTS,
(ºCs-1)
MPa
MPa
1
435
5
YS/UTS
Yield type
566
0.77
C
518
647
0.80
C
10
502
643
0.78
C
20
477
638
0.75
C
40
503
651
0.77
C
60
484
638
0.76
C
- 128 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
7.8.2.3 Effect of cooling rate with 575 ºC coiling and a 33% prior reduction below
the Tnr
Table 7.14 shows the mechanical properties of the Mo-free alloy #6 with a 575 ºC
coiling simulation and a 45% prior reduction in total but 33% reduction below the Tnr
while the curves of strength versus elongation are shown in Appendix F.
Table 7.14 Mechanical properties of samples #TEN06 for the Mo-free alloy #6 with
60 min coiling at 575 ºC and a 33% prior reduction below the Tnr
YS,
UTS,
MPa
MPa
1
434
5
Cooling rate (ºCs-1)
YS/UTS
Yield type
536
0.81
D
528
626
0.84
D
10
514
605
0.85
D
19
568
664
0.86
D
34
528
639
0.83
D
7.9 Transformed microstructures of the alloys
7.9.1 Optical micrographs
The optical microstructures of the five experimental alloys after a rapid cooling rate
of 47 ºCs-1 and the reference alloy #6, are shown in figures 7.29.
(a)
(b)
- 129 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
(c)
(d)
(e)
(f)
Figure 7.29 The optical microstructure, etched in a 2% Nital solution for 5 seconds,
after a rapid cooling rate of 47 ºCs-1 for the experimental alloys (a) #1, (b) #2, (c) #3,
(d) #4, (e) #5 and, (f) the reference Mo-free alloy #6 cooled at a rate of 39 ºCs-1.
As may be seen in figure 7.29, the optical microstructures for all five experimental
alloys #1 to #5 in the as-rolled condition, are similar. They appear to be typical of
acicular ferrite, consisting of a “white” phase and an irregular “grey” phase with no
clearly etched grain boundaries in the microstructure. In a low carbon low alloyed
steel when etched with 2% Nital, polygonal ferrite usually appears as a white coloured
phase. From figure 7.29, it is uncertain, however, to establish without any doubt the
presence of polygonal ferrite or whether the microstructure shows a 100% acicular
ferrite or a mixture of both acicular ferrite and polygonal ferrite. If the polygonal
ferrite is mixed within the acicular ferrite structure, it can not be recognised with any
confidence. The microstructure of reference alloy #6 consists of a similar
microstructure as the five experimental alloys. Accordingly, it was necessary to
confirm by other means whether any polygonal ferrite exists in these microstructures.
- 130 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
7.9.2 Microstructures examined by SEM
The Scanning Electron Microscopy (SEM) (Model JEOL JSM-5800LV) was used
first in an attempt to identify the presence of polygonal ferrite in the microstructure.
The SEM microstructures for the experimental alloys etched in a 2% Nital, are
illustrated in figure 7.30 below.
It seemed that there were two different types of microstructures or matrix phases, one
is raised above the mean level of the etched plane (marked in “A”), while the other is
sunken below this mean plane (marked in “B”). No fine details within each matrix
phase can be seen, however, and the SEM was, therefore, not the final answer as to
identify the phases which were present.
Next, high resolution SEM (Model JEOL JSM-6000F) micrographs were taken after
varying the etching time from 10 to 120 seconds in an attempt to identify these two
phases. Figure 7.31 represents the structures of the as-rolled experimental alloys
under high resolution SEM after only 5 seconds etching. The microstructures appear
relatively similar to those in figure 7.30 although the differences in etching depth
between the “raised” and the “sunken” matrix were accentuated. The SEM
micrographs after various etching times ranging from 10 to 120 seconds, are shown in
figure 7.32. Although the differences in etching depth between the two types of matrix
areas are once more, evident, the lack of any finer details of the internal structure
within each of the two types of matrix areas did not assist in identifying them
conclusively (although one may be tempted to conclude that the deeper etched matrix
areas could possibly be acicular ferrite due to its higher dislocation content while the
lesser etched matrix areas could possibly be polygonal ferrite with a lower dislocation
content). Accordingly, high resolution SEM analysis was also found not to be a totally
satisfactory technique to fully identify the matrix microstructures in the present alloys.
- 131 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
(a)
(b)
A
B
A
B
(d)
(c)
A
A
B
(e)
(f)
Figure 7.30 The SEM micrographs after a rapid cooling rate of 47 ºCs-1(etched in 2%
Nital for 5 seconds) for the experimental alloys (a) #1, (c) #2, (c) #3, (d) #4, (e) #5
and, (f) the reference alloy #6 cooled at a rat of 39 ºCs-1.
- 132 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
(a)
(b)
(c)
(d)
(e)
Figure 7.31 The micrographs in the as-rolled condition by high resolution SEM after a
rapid cooling rate of 47 ºCs-1(etched in 2% Nital for 5 seconds) for the experimental
alloys (a) #1, (c) #2, (c) #3, (d) #4 and, (e) #5.
- 133 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
(a)
(b)
(c)
(d)
(e)
Figure 7.32 The high resolution SEM micrographs in the as-rolled condition after a
rapid cooling rate of 47 ºCs-1 for the experimental alloy #1 etched in 2% Nital for (a)
10seconds, (c) 15 seconds, (c) 30 seconds, (d) 60 seconds and, (e) 120 seconds.
- 134 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
7.9.3 TEM studies of acicular ferrite on carbon replicas
Carbon extraction replicas with or without Au-Pd shadowing and thin foils were
finally used to identify any differences between the matrix phases of acicular ferrite
and polygonal ferrite. An etching time from 30 to 60 seconds in a 2% Nital solution
was used to obtain a relief on the etched surface on the specimens. A deeply etched
relief surface before they are carbon coated, enhances the shadowing of samples.
The TEM micrograph of an unshadowed carbon extraction replica for alloy #6
(reference alloy and Mo-free) is shown in figure 7.33. It appears once more that there
are two different matrix phases in the as-rolled reference alloy #6. The quality of the
image is a little better than that of the SEM micrographs in figures 7.30 and 7.32.
Hence it seemed that carbon extraction replicas may be an improved technique to
reveal the finer details of these matrix phases in the alloys, particularly if shadowing
by Au-Pd is used before applying the carbon coating to accentuate the relief on the
etched surface. Figure 7.34 shows such a shadowed microstructures of alloy #6.
Figure 7.34 shows the marked improvement introduced by shadowing if compared to
the unshadowed case in figure 7.33 on the same alloy. The “raised” matrix area (note
the direction of the shadow at its edge which is the same as that of protruding carbide
particles) has a relatively smooth etched surface revealing some individual carbide
particles after being etched in 2% Nital, while the “sunken” matrix areas in the same
figure have a very “rough” etched surface which consists of a fine internal and
parallel striated structure that has clearly etched very much differently from the
“raised” smooth matrix areas.
It may, therefore, be concluded that the raised matrix areas appear to be one phase
whereas the sunken matrix areas are clearly another phase or even a multiphase
(except for the carbide particles which are present in both types of areas).
Consequently, there are apparently two types of matrix phases in alloy #6. Taking all
of the above evidence of optical, SEM and TEM observations together, it was
concluded that the smooth and raised matrix phase in figure 7.34 is most likely
polygonal ferrite (marked with “PF”), whereas the striated and deeper etched one is
likely to be acicular ferrite (marked with “AF”). This result was further validated with
- 135 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
thin foil samples in sections 8.3 and 8.4 in the next chapter. As can be seen in figure
7.34, the shadowing is also effective to reveal the smaller precipitates and grain
boundaries (marked with “GB”) by enhancing their presence.
2.4µm
(a)
1.5µm
(b)
13 µm
(c)
Figure 7.33 TEM micrographs of carbon extraction replicas without shadowing for
the reference alloy #6 after hot rolling and rapid cooling at a rate of 39 ºCs-1.
Summarising these results, it was concluded that the shadowing of extraction carbon
replicas by applying Au-Pd shadowing at an angle to the etched surface, is an
- 136 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
improved technique to identify the matrix structures in these alloys as both acicular
ferrite and polygonal ferrite can be clearly distinguished from each other. The
microstructure of reference alloy #6 after hot rolling and rapid cooling at a rate of 39
ºCs-1 is, therefore, one of polygonal ferrite plus acicular ferrite with some individual
smaller carbides in both these matrix phases.
By this technique, it was also confirmed that the microstructures of the experimental
alloys #1 to #5 equally consisted of both polygonal ferrite and acicular ferrite (see
figure 7.35) while no structure in the entire study was found with 100% acicular
ferrite. The mixed microstructures of the five experimental alloys were, therefore,
very similar to that found in the reference Mo-free alloy.
AF
GB
GB
“raised”
“sunken”
PF
0.6 µm
Figure 7.34 The TEM micrograph from a shadowed replica of the Mo-free alloy #6
after hot rolling and rapid cooling at a rate of 39 ºCs-1. (AF-acicular ferrite, PFpolygonal ferrite and, GB-grain boundary).
- 137 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
GB
AF
AF
PF
PF
1.4µm
(a)
2 µm
(b)
PF
GB
PF
AF
AF
(c)
3.7µm
(d)
- 138 -
2µm
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 7 Results
PF
GB
AF
3.8µm
(e)
Figure 7.35 TEM micrographs from shadowed replicas of the as-hot rolled and
rapidly cooled (at a rate of 47 ºCs-1) experimental alloys (a) #1, (b) #2, (c) #3, (d) #4
and, (e) #5 (AF-acicular ferrite, PF- polygonal ferrite and, GB-grain boundary).
- 139 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
CHAPTER 8 STUDIES OF ACICULAR FERRITE BY THIN FOIL TEM
8.1 Acicular ferrite morphology in experimental alloys
The optical micrographs in figure 7.29 for the alloys after rapid cooling at a rate of 47
ºCs-1, show a typical acicular ferrite microstructure, also according to the
literature[4,65,131,134]. No researcher in any of these references reported the presence of
any polygonal ferrite in an acicular ferrite microstructure, however, even though their
optical microstructures appear very similar to those in this study (figure 7.29).
Polygonal ferrite is of a “polygonal” shape with a lightly etched colour of “white” in
its microstructures under an optical microscope after being etched in a 2% Nital
solution. But acicular ferrite also appears as a white plus grey or striated phase in
microstructures under some etching conditions. Previous researchers[4,65,131,134] have
accepted that optical micrographs like figure 7.29-(b) to (f), consist of only acicular
ferrite. As mentioned in section 7.9.3 above, however, the TEM observations on
carbon extraction replicas in this study with Au-Pd shadowing, as represented by the
microstructures of alloys #1 to #5, are obviously a mixture of polygonal ferrite plus
acicular ferrite. A final conclusion based only on the somewhat ill-defined optical
microscopy but more clearly confirmed by the observations on shadowed extraction
replicas is, of course, not satisfactory in itself and TEM observations on thin foils
were done to confirm more details of this apparent anomaly with conclusions reached
by other researchers in these microstructures.
Furthermore, it should be recognised that the latest understanding on the nature and
origin of acicular ferrite in wrought line pipe steels, is that it may not necessarily be a
“unique singular phase” (for instance, the “chaotic morphology” found in welds) but
that acicular ferrite should rather be understood as a collective term describing a
microstructure containing possibly more than one type of ferrite phase, all of which
have nucleated intragranularly. It will be seen that this broader concept of what
acicular ferrite in line pipe steels really is, is supported by the results of this work here
below where two different morphologies of intragranularly nucleated ferrite laths
were found, i.e. a “chaotic” arrangement (as in the “classical” acicular ferrite in welds)
as well as an arrangement of “parallel laths” somewhat similar to classical bainite
except for its nucleation intragranularly.
- 140 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
8.1.1 Acicular ferrite and polygonal ferrite in alloy #6 (Mo-free)
(i) Acicular ferrite and polygonal ferrite
Figure 8.1 below represents the thin foil TEM micrographs of the Mo-free alloy #6
after a rapid cooling rate of 39 °Cs-1 after the hot rolling process. A “polygonal” shaped
phase could be observed in figure 8.1-(a) (marked with “PF”). According to its shape
it appears to be polygonal ferrite with a size of approximately 5 µm in diameter. Many
laths in the structure (marked “lath” in the figure) were found beside the polygonal
ferrite. The shape of the “PF” is quite different from that of the laths in this figure
with the latter more elongated in their shape. This is consistent with the observation
from figure 7.34 and discussed in the introduction in section 8.1 above.
- 141 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
B
PF
PF
Lath
A
A
2 µm
(a)
B
(b)
2 µm
Lath
B
Lath
B
2 µm
(d)
(c)
Figure 8.1 Thin foil TEM micrographs
2 µm
of alloy #6 (Mo-free) after a rapid cooling
-1
rate of 39 °Cs after hot rolling, (a) polygonal ferrite + laths, (b) polygonal ferrite
with dislocations and, (c),(d) laths with dislocations. PF-polygonal ferrite, AF-acicular
ferrite, A and B-dislocations in polygonal ferrite and an acicular ferrite , respectively.
Another polygonal ferrite grain with size of about 5.8 µm in figure 8.1-(b) is also
evident. No cementite was found within the laths or on inter-lath boundaries. This
means that the laths are in all likelihood part of acicular ferrite instead of bainite
where cementite would be present within the laths in lower bainite and on inter-lath
- 142 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
position in upper bainite[97]. The carbon rejected from the acicular ferrite laths may
possibly be in martensite/austenite islands enriched in carbon as no cementite was
found in the thin foils. A further study of the martensite/austenite islands in these
steels probably needs further thin foil TEM work to confirm it.
According to the CCT diagrams of alloy #6 (see figure 7.22), the polygonal ferrite
probably forms during cooling first at higher temperature and the acicular ferrite later.
The acicular ferrite is, therefore, mixed in between the polygonal ferrite. Polygonal
ferrite can not be fully distinguished from an acicular ferrite on optical micrographs
alone such as in figure 7.29-(f), but combining optical and TEM observations of alloy
#6 in figure 7.29-(f), it appears that the laths (see figure 8.1) are most likely from an
acicular ferrite and not bainite because of the less visible grain boundaries[131,135] and
no cementite within or between the laths was observed after deep etching. A further
study on this lath structure is made below. It is, therefore, concluded that the
microstructure of alloy #6 is most probably a mixture of acicular ferrite and polygonal
ferrite.
(ii) Dislocations within the polygonal ferrite
Dislocations were also observed within the polygonal ferrite (see figures 8.1 and 8.2)
in these line pipe steels. This is somewhat different, for instance, from polygonal
ferrite found in plain low carbon steels where generally dislocation-free polygonal
ferrite is usually found. This has led to some researchers even referring to
“quasi-polygonal ferrite” in wrought low alloy steels. The reason for this difference
caused by alloying elements, requires further investigation but may be associated with
solute drag of dislocations by alloying elements, thereby hindering their movement
and recovery whereas this is largely absent in plain low carbon steels.
Figure 8.2 showed that there were dislocations apparently being emitted (marked with
“M”) from the area close to the moving interface of the PF. There were also more
dislocations near this moving interface than within the central regions of the PF
(marked with “L”). These dislocations were possibly generated during the formation
of polygonal ferrite within the austenite. Firstly, according to the CCT diagram of
alloy #6 (see figure 7.26), the polygonal ferrite was formed in a higher temperature
- 143 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
range than that of acicular ferrite. Secondly, there was a volume expansion of about
3.6% (calculated for a 0.06% C steel) during polygonal ferrite formation from
austenite as the lattice changed from face-centered cubic (austenite) to body-centered
cubic (polygonal ferrite). This will lead to the transformed polygonal ferrite that was
surrounded by the parent austenite, to undergo a compression strain of about 1.2%
from the austenite. Because of the relatively high temperatures at this stage and hence
a low flow stress in the ferrite, this linear strain will probably be a plastic one, thereby
creating interface dislocations in the ferrite near the interface as it advances into the
austenite.
In addition, however, a second volume expansion again of about 3.6% occurs at a
slightly lower temperature as the surrounding austenite now transforms to the bcc
acicular ferrite, with the less dense acicular ferrite now creating a further strain
field[74,76] on the already formed polygonal ferrite. Such a strain field from acicular
ferrite formation leads to a high dislocation content in the acicular ferrite from the
displacive transformation but also from some plastic strain within the adjacent
austenite[75] or the already existing polygonal ferrite due to the lattice volume change.
Both of these volume expansion processes will result in strain or deformation to be
concentrated within the outer edges of the polygonal ferrite rather than in the centre.
Therefore, there were more dislocations found in the edge region (marked by “M”)
than in the centre region of the polygonal ferrite (marked by “L”). Figures 8.1-(a) and
8.2 show the acicular ferrite that was situated around the polygonal ferrite.
- 144 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
M
L
200 nm
Figure 8.2 Dislocations within the polygonal ferrite in thin foil of the Mo-free alloy
#6. The area M shows a high density of dislocations possibly being emitted from the
moving PF interface while the central regions L of the PF have less dislocations.
8.1.2 Acicular ferrite and polygonal ferrite in alloy #1
Alloy #6 was the current Mittal Steel’s line pipe steel, i.e. V-Nb-Ti micro-alloyed. The
main difference between the experimental alloy #1 and the reference alloy #6 was in
the niobium and carbon contents. The niobium content in alloy #1 had been increased
to 0.055%wt Nb and the carbon content lowered to 0.05%wt C in order to obtain a
higher degree of dispersion hardening in the ferrite. TEM thin foil micrographs of
polygonal ferrite in this study are shown in figure 8.3, in which a high density of
dislocations was also observed with a high magnification.
Figure 8.4 represents an interwoven lath structure from alloy #1 and a high density of
dislocations was found within laths B and C. Lath A is not parallel to laths B and C.
This is a typical interwoven structure of acicular ferrite, also reported in the
literature[4,65,131,134]. Another interwoven lath structure with PF present in this alloy #1,
is also shown in figure 8.5. Consequently, it is confirmed that the microstructures of
alloy #1 with rapid cooling after hot rolling, also consists of a mixture of an acicular
ferrite plus polygonal ferrite.
- 145 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
4 µm
(a)
1.5 µm
(b)
Figure 8.3 Polygonal ferrite in a TEM thin foil micrograph from the experimental
alloy #1 after a rapid cooling rate of 47 ºCs-1 after the hot rolling process.
Lath B
Lath A
Lath C
2 µm
Figure 8.4 TEM thin foil micrograph with laths from alloy #1 after a rapid cooling
rate of 47 ºCs-1 after the hot rolling process.
- 146 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
1 µm
Figure 8.5 TEM thin foil micrograph of the lath plus PF structure in alloy #1 after a
rapid cooling rate of 47 ºCs-1 after the hot rolling process.
- 147 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
8.1.3 Acicular ferrite in alloys #2 to #5
The lath structures of alloy #2 with an 0.09% Mo + 0.05% Nb addition, are shown in
figures 8.6 and 8.7 from TEM thin foils after a rapid cooling rate of 47 ºCs-1 after the
hot rolling. Parallel laths were found in figure 8.7.
Lath C
Lath A
Lath B
(a)
1.5µm
(b)
2.0 µm
Figure 8.6 Thin foil TEM micrographs of the lath structure in alloy #2 after a rapid
cooling rate of 47 ºCs-1 after the rolling process.
- 148 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
2µm
Figure 8.7 Thin foil TEM composite micrographs of parallel laths of an acicular
ferrite in alloy #2 after a rapid cooling rate of 47 ºCs-1 after the hot rolling process.
Figure 8.8 shows the polygonal ferrite (plus a few isolated laths) found in alloy #3
after a rapid cooling rate of 47 ºCs-1 after the hot rolling process. Some dislocations
can be found within the polygonal ferrite as well although of a lesser density than in
the AF laths. The micrographs with a parallel lath structure of alloy #3 are shown in
figure 8.9.
- 149 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
Lath
2.2µm
(a)
(b)
3.9µm
Lath
Lath
Lath
(c)
(d)
3.9µm
2µm
Figure 8.8 Polygonal ferrite (with a few isolated laths) in alloy #3 with a rapid cooling
rate of 47 ºCs-1 after the hot rolling process.
- 150 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
(a)
1µm
(b)
3.2µm
Figure 8.9 The parallel lath morphology in alloy #3 after a rapid cooling rate of 47
ºCs-1 after the hot rolling process.
Figure 8.10 shows a composite micrograph of a mixed polygonal ferrite and an
acicular ferrite microstructure in alloy #3. Polygonal ferrite is marked with PF in the
figure. The interwoven laths are clearly seen, where lath A is crossed by laths B and
C.
Figure 8.11 represents the lath structure in alloy #4. Polygonal ferrite with
dislocations, can also be found in alloy #5 (with 0.22% Mo) after a rapid cooling rate
of 47 ºCs-1 (figure 8.12-(a)). An interwoven lath structure of an acicular ferrite in alloy
#5 can also be observed in figures 8.12-(b) and (c). It is, therefore, confirmed that the
microstructure of alloy #5 (with 0.22%Mo) is also one of polygonal ferrite plus
acicular ferrite.
- 151 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
Lath A
Lath C
Lath B
Lath A
PF
PF
1µm
Figure 8.10 Thin foil TEM micrographs of a mixture of polygonal ferrite and an
acicular ferrite in alloy #3 after a rapid cooling rate of 47 ºCs-1 after the hot rolling
process.
- 152 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
2µm
Figure 8.11 Thin foil TEM composite micrographs of the acicular ferrite in alloy #4
after a rapid cooling rate of 47 ºCs-1 after the hot rolling process.
- 153 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
Lath B
Lath A
Lath C
(a)
1µm
(c)
(b)
2µm
2µm
Figure 8.12 Thin foil TEM micrographs from alloy #5 with 0.22% Mo (a) polygonal
ferrite, (b) and (c) acicular ferrite with interwoven laths.
- 154 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
Conclusions
Summarising these thin foils results shown above, the following conclusions can be
drawn:
•
All experimental alloys cooled at a rapid cooling rate of 47 ºCs-1 after hot
rolling, had a mixed microstructure of polygonal ferrite and an acicular ferrite,
as confirmed by thin foil TEM micrographs.
•
This confirmed the tentative observation made from the optical micrographs in
figure 7.28 that the structure did not appear to be a 100% acicular ferrite but
that it rather consisted of a mixture of polygonal ferrite and an acicular ferrite.
•
There were no visible etched boundaries between polygonal ferrite and the
acicular ferrite microstructure on the optical micrographs. Accordingly, optical
micrographs such as in figure 7.29, can not decisively prove by themselves
whether there is any polygonal ferrite in a largely acicular ferrite
microstructure, as is apparently sometimes inadvertently done in the literature.
It can only be confirmed by TEM work on thin foils and shadowed carbon
replicas.
•
Acicular ferrite has an interwoven lath structure with a high density of
dislocations but contained no cementite, neither in the interlath positions or
within the laths.
•
Polygonal ferrite has a significantly lower dislocation content than the acicular
ferrite but some apparent interface emission of dislocations in regions near to
the interface into the polygonal ferrite has been observed.
- 155 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
8.2 Two types of acicular ferrite
8.2.1 Structure with parallel laths
There appeared to be two types of acicular ferrite laths that were observed in those
alloys cooled with a rapid cooling rate of 47 ºCs-1 after the hot rolling. Diaz-Fuentes[78]
has also found similarly two types of acicular ferrite by isothermal treatment in a
medium carbon steel. Figure 8.13 shows some parallel laths in a lath colony in which
some parallel laths interweave with one another. This can be clearly seen in figures
8.13 and 8.14. The apparent preferred habit plane of the laths in colony 1 appears to
be different from that of the nearby colony 2 in figure 8.13. Combining this
conclusion with the optical micrographs of this alloy (in figure 7.29), these parallel
laths are possibly not from a bainite structure. This conclusion is based on no grain
boundaries that can be observed in optical metallography (figure 7.29), and no
cementite within the laths or on interlath positions by TEM. Acicular ferrite nucleates
intragranularly on inclusions[71,74,136,137] in weld pools and not intergranularly, while
its nucleation may possibly also be enhanced by dislocations resulting from prior
deformation, as was described in section 7.5. The different parallel laths are possibly
primary laths which have the same growth direction[75] and with a secondary lath that
may nucleate at the tip of the primary one but with both having the same habit plane
and orientation, resulting in a parallel sheaf morphology.
Colony 1
Colony of laths
Colony 2
(a)
3µm
(b)
- 156 -
1µm
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
Figure 8.13 Parallel lath morphology in a colony in alloy #3 after rapid cooling at a
rate of 47 ºCs-1 after the hot rolling process.
Madariaga and Bhadeshia[75] reported a similar parallel lath morphology of acicular
ferrite in a medium carbon micro-alloyed steel that was isothermally treated at a
relatively low temperature of 400 ºC. The two reasons given by the authors for the
formation of a parallel lath microstructure are:
(1) The lower stability of the austenite with low carbon enrichment close to the ferrite
tip rather than on the face of the lath, leading to the secondary plate of ferrite
nucleating at the tip of previous one.
(2) Madariaga[75] reported that some retained austenite and cementite was found
between the ferrite platelets. In that case, there was enough time for diffusion of
carbon after the acicular ferrite formation because of the isothermal treatment. The
excess carbon in acicular ferrite may be rejected into the austenite after the
acicular ferrite formation and the cementite can be formed from the carbonenriched austenite during this continued isothermal treatment. The austenite close
to the face of the platelet has more carbon than the tip and as a result, cementite is
distributed differently in the different sections of the boundary between adjacent
platelets.
On the contrary, however, the parallel laths found in this study appear to be different
from those in Madariaga’s study[75]. No cementite was observed here on boundaries
between laths or within the laths in any of the alloys in this study. Furthermore, the
phase transformation in this study must have taken place in a few seconds due to the
rapid cooling rate of 47 ºCs-1, far less than in any isothermal treatment as was used by
Madariaga. Consequently, the nucleation and growth of laths in the alloys studied here
are probably different from those in Madariaga’s study. The nucleation of a secondary
lath is dependent not only on the carbon depletion into the parent austenite in front of
the interface with the primary lath (the excess carbon from the primary lath will be
rejected into the adjacent austenite[75,78], which then is not suited for the formation of
a secondary lath), but also on the defects in the austenite adjacent to the primary lath,
which resulted from the hot rolling process below the Tnr. These defects, therefore,
- 157 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
assist in the nucleation of the ferrite nuclei[134]. On the other hand, the secondary lath
was possibly formed heterogeneously along the primary one in order to decrease the
activation energy of transformation. A set of parallel laths that composes a colony, is
typical of acicular ferrite found in this study. The boundary between the colonies
could not be observed under the optical microscope (figure 7.29) because the size of
the colonies was too small and the orientation differences between colonies were
chaotic. The morphology at the end of the laths in figure 8.14 may not support the
assumption that they had nucleated on the boundary of the austenite.
Colony 1
Colony 2
2µm
Figure 8.14 Interwoven arrangement between lath colonies in alloy #4 after a fast
cooling rate of 47 ºCs-1 after the hot rolling process.
- 158 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
8.2.2 Structure with interwoven laths
Another type of lath structure may be seen in figures 8.15 and 8.16. Laths are
chaotically arranged and interwoven in between each other (marked with A, B, C, D,
E and F). Each lath had its own apparent growth direction. This lath morphology
appears to be the same[69,90,134] as that which other researchers have observed in welds
in which many non-metallic inclusions in the weld pool act as high density nucleation
sites for acicular ferrite[57,81,138]. Such a nucleation process will result in various
orientations, interwoven in the nature of acicular ferrite formation.
Although there were relatively few non-metallic inclusions in the alloys in the present
study, some interwoven laths were still observed even though this was not typical
acicular ferrite morphology in these alloys. Where this did occur, the primary acicular
ferrite (AF) lath may have nucleated on an inclusion with the next lath then nucleated
on the face of the previous lath, i.e. on the interface of the primary AF lath and the
untransformed austenite. The direction of growth will favour the direction in which
the retarding forces of the transformation are the lowest. Figure 8.15 shows that laths
B, C and F apparently nucleated on the face of lath A, while lath G probably nucleated
on lath F.
It, therefore, may be concluded that there are two types of laths of acicular ferrite
present in the alloys studied in this work, namely, parallel and interwoven laths with
the former being the dominant type.
- 159 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
D
B
A
C
G
F
(c)
1µm
E
Figure 8.15 Interwoven laths micrographs in alloy #3 after fast cooling of 47 ºCs-1
after hot rolling process.
- 160 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
2µm
Figure 8.16 Acicular ferrite morphology in alloy #5 after fast cooling of 47ºCs-1 after
the hot rolling process.
- 161 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
8.3 Nucleation of acicular ferrite
8.3.1 Nucleation on non-metallic inclusions
Both acicular ferrite and bainite have a lath-like structure when observed in thin foil
samples by Transmission Electron Microscopy, i.e. both belong to a group of
transformations that proceed basically by a displacive mechanism[139]. The main
differences between the nucleation and growth of these two phases are the nucleation
sites and the presence of cementite inside the laths or at inter-lath positions. Bainite
nucleates on an austenite grain boundary[65,66,78], forming sheaves of parallel plates or
laths with all essentially of the same crystallographic orientation. The boundaries and
the general orientation of the plates are generally visible by optical microscopy (see
figure 7.21-(k) in section 7.4 above).
The nucleation of an acicular ferrite structure, on the other hand, generally occurs
intragranularly and often on non-metallic inclusions[70,136] as found in weld pools. As
described in section 8.4, the morphologies found in this study were either parallel
laths (figure 8.13) or interwoven laths (figure 8.15). Such intricate structural features,
however, could not be identified by optical microscopy (see figure 7.29 in section 7.9)
because of a lack of resolution of visible boundaries[131,135,140]. No precipitation of
cementite was also found by thin foil TEM between or within the laths of the acicular
ferrite. Accordingly, the lath morphologies in the present work differ significantly
from that of bainite. These differences stem basically from differences in their
nucleation mechanisms. Such differences in nucleation sites between bainite and
acicular ferrite formation were also found by other authors [65,66,78,134]. A reduction of
the austenite grain boundary surface area per unit volume, favours the formation of
acicular ferrite and is detrimental to the formation of bainite due to the decrease in the
number of potential bainite nucleation sites[65] on the austenite grain boundaries. A
similar result in enhancing acicular ferrite formation was achieved by increasing the
quantity of inclusions in the steel[141]. It has, therefore, been generally accepted that
inclusions are the favoured sites for the nucleation of acicular ferrite[53,54,84,141,142], at
least in the case of weld pools.
As described in section 8.3 above, the microstructures of the alloys studied here
consisted in general of a mixture of acicular ferrite plus polygonal ferrite. An attempt
- 162 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
was made to find any non-metallic inclusions that existed within the laths. Figures
8.17 to 8.20 show acicular ferrite laths around such non-metallic inclusions. The
analysis by Energy Dispersive Spectroscopy (EDS) on the TEM of the inclusion in
figure 8.17-(a) showed that it consisted of manganese and iron oxides. The iron peak
in figure 8.17-(b) was not thought to arise entirely from the steel matrix as the size of
the inclusion of about 1.1 µm was large enough for the electron beam to strike
primarily on the inclusion during the EDS analysis, although the beam had to
penetrate through some matrix material to arrive at the inclusion. No diffraction
pattern could be obtained on the TEM as the inclusion was too thick. The inclusion
was centered inside the 50 to 80 µm thickness of the thin foil which made it even
more difficult for the electron beam to transmit through the inclusion. Consequently,
any information on the structure or molecular formula of the inclusion could not be
obtained from a diffraction pattern.
(a)
1 µm
(b)
Figure 8.17 (a) TEM image of acicular ferrite and a large non-metallic inclusion in
alloy #5 after a rapid cooling rate of 47 ºCs-1 after the hot rolling, (b) EDS analysis on
the inclusion in this figure (a).
- 163 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
A
B
(a)
1 µm
2 µm
(b)
(d)
(c)
(e)
Figure 8.18 Laths nucleated on non-metallic inclusions (a) in alloy #1 after a rapid
cooling rate of 47 ºCs-1 after the hot rolling, (b) in alloy #3 after a rapid cooling rate of
40 ºCs-1 from 980 ºC down to room temperature without deformation, (c) EDS
analysis on the inclusion in this figure (a), (d) and (e) EDS analysis on the inclusions
A and B in this figure (b), respectively.
- 164 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
(a)
2µm
(b)
Figure 8.19 (a) Nucleation of interwoven laths of acicular ferrite in sample #AF3F of
alloy #3 after a cooling rate of 20 ºCs-1 from 980 ºC down to room temperature
without deformation, (b) EDS analysis of red peak was from on the inclusion
indicated by an arrow in this figure (a), while blue peak was from the matrix steel.
- 165 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
(a)
1µm
(b)
Figure 8.20 (a) Non-metallic inclusion and acicular ferrite in alloy #3 after a rapid
cooling rate of 47 ºCs-1 after the hot rolling, (b) EDS analysis on the inclusion
indicated by an arrow in this figure (a).
- 166 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
Similar nucleation of laths of acicular ferrite on inclusions was also found in alloy #1
(figure 8.18-(a) above). The inclusions in this case were generally round in shape,
with a size of about 1.2 µm. The EDS image in figure 8.18-(c) reveals that the
composition of these inclusions was apparently a complex oxide containing silicon,
aluminium and iron.
Further nucleation of acicular ferrite on non-metallic inclusions is shown in figure
8.18-(b) for sample #AF3F of alloy #3 (0.09% Mo) that was cooled at 40 ºCs-1 from
980ºC down to room temperature (figure 6.14 in section 6.9). The acicular ferrite had
once more, a typical lath structure nucleated around non-metallic inclusions with
rounded shapes. The diameter of the inclusions ranged from 1.9 µm to 2.2 µm (see
figure 8.18-(b)). The EDS image in figure 8.18-(d) and (e) revealed that they were
complex oxides containing silicon, aluminium and iron. From the figure it appears
that the two primary parallel laths of acicular ferrite nucleated on the non-metallic
inclusion. The secondary lath, however, nucleated at the tip of the primary one,
possibly owing to a lower carbon content in this area and then grew in the same
direction as the primary one.
The nucleation of interlocked laths of acicular ferrite is shown in figure 8.19 for
another sample in group #AF3F of alloy #3 (0.09% Mo) (figure 6.14 in section 6.9).
The cooling rate was 20 ºCs-1 from 980 ºC down to room temperature. The inclusions
were about 0.5 µm in diameter and, therefore, quite small to be analysed with
confidence by EDS in the thin foil. The blue peak for iron in figure 8.19-(b) probably
arose from the matrix of the steel surrounding the inclusion. There is a difference in
the iron peak between the red and blue lines in the EDS image. Therefore, it appears
that the non-metallic inclusion shown by an arrow in the figure was a mixture of
manganese, iron and copper oxide.
Figure 8.20 shows the nucleation of an acicular ferrite colony on an inclusion of about
0.35 µm in diameter. For such an inclusion, the iron peak in figure 8.20-(b) is
probably from the steel matrix and not from the inclusion. Thus, the acicular ferrite
nucleated on an inclusion consisting probably of manganese and copper sulphide.
- 167 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
Summarising the results above, a number of inclusions were found that were
associated with acicular ferrite laths. Previous researchers have also found that
acicular ferrite nucleates on non-metallic inclusions[75,77,141] and their nucleation
frequency increases with an increasing quantity of inclusions in a weld pool[141,143]. In
this work the nucleation sites for acicular ferrite were generally oxide and sulphide
inclusions with generally a rounded shape whereas the size of these inclusions ranged
from 0.35 to 2.2 µm. The chemical compositions of these non-metallic inclusions
were complex containing silicon, aluminium, iron, manganese and copper.
8.3.2 Type of non-metallic inclusion as nucleants
Other researchers have reported that acicular ferrite nucleates on inclusions or
particles, such as TiO[141,144,145], BN plus rare earth metal oxysulphides[148],
aluminium-rich inclusions[147,148], and TiN[149]. Most of the work, however, focused on
weld pools where a larger quantity of non-metallic inclusions is readily introduced.
Bhatti[147] has reported that inclusions rich in manganese and inclusions covered by a
skin of sulphide are ineffective nucleants for acicular ferrite. Inclusions which are
covered by or are rich in copper-sulphur or silicon and which were effective as a
nucleant[65,66,70] for acicular ferrite, however, were found by Zhang and Farrar[141],
Dowling et al[144], Court[150], Harbottl[151], Madariaga[69] and Kayali et al[152]. Similar
oxide inclusions of manganese plus iron and containing copper and also manganese
sulphides containing copper have been found in this work (see figures 8.19 and 8.20
respectively). It may, therefore, be concluded that the type of inclusions that had
formed nucleation sites for acicular ferrite in these steels studied here, were most
likely complex oxides or sulphides.
8.3.3 Nucleation mechanisms of acicular ferrite
Three nucleation mechanisms of acicular ferrite on inclusions have been proposed by
previous researchers: (1) the existence of local variations in the chemistry of the
matrix[90]; (2) the generation of high strain fields around the inclusion due to the
different thermal expansion coefficients between austenite and the inclusion[153,154] or
deformation in the austenite region; and (3) the creation of a low energy interface
between acicular ferrite and the inclusion owing to a low lattice mismatch between
them[69, 155-158].
- 168 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
When acicular ferrite is formed, the total system’s Gibbs free energy must be lowered
for the nucleus to become thermodynamically stable. This total energy includes the
decrease in the chemical free energy, the disappearance of an interface between the
inclusion and the austenite, and two new interfaces created of austenite-acicular ferrite
and inclusion-acicular ferrite. This change in total energy can be described by the
following equation:
ΔGtotal= –ΔGv VAF+σAF/γAAF/γ+
σAF/IAAF/I +ΔGεVAF– σI/γAI/γ – ΔGdefectVAF
(8.1)
where ΔGv is the decrease in Gibbs free energy per unit volume owing to the
transformation from austenite to acicular ferrite, also often termed the “chemical free
energy”,
VAF is the volume of transformed acicular ferrite,
σAF/γ, σAF/I, σI/γ
are interface energies per unit area between acicular ferrite and
austenite, acicular ferrite and an inclusion and an inclusion and austenite, respectively;
AAF/γ, AAF/I, AI/γ are the interface areas between acicular ferrite and austenite, acicular
ferrite and an inclusion and, an inclusion and austenite, respectively,
ΔGε is the strain energy per unit volume around the new AF nucleus due to the
volume expansion during the transformation from a face-centered cubic lattice (fcc)
to a body-centered cubic lattice (bcc) and
ΔGdefect is the stored defect elastic energy per unit volume in austenite around the
inclusion due to deformation in the austenite, such as dislocation or point defects.
–ΔGv VAF is the chemical Gibbs driving force for the transformation that is dependent
on the degree of under-cooling of the austenite as well as the chemical composition of
the austenite and is generally independent of the inclusions present in the austenite.
σAF/IAAF/I and σI/γAI/γ are only affected by inclusions where acicular ferrite nucleates
and is primarily determined by the mismatch of the interface between acicular ferrite
and an inclusion. The transformation will proceed more readily through an increasing
absolute value of –ΔGv VAF and become stable when the embryo size of acicular
ferrite (r) reaches the critical size (r*) (r* is determined according to
dΔG total
=0). This
dr
driving force is increased by increasing the under-cooling ΔT which again is increased
by a higher cooling rate on the rising part of the “nose” on a CCT diagram. That this
- 169 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
is so, is confirmed by the results shown on the CCT diagrams in figures 7.26 and 7.28
in section 7.5 where it was found that a faster cooling rate after deformation of the
austenite is more favourable for acicular ferrite formation because the under-cooling
increases with an increase in cooling rate.
σAF/γAAF/γ, σAF/IAAF/I and ΔGεVAF are the barriers to the nucleation of AF. It means
that low values of the surface energies σAF/γ and
of acicular ferrite on inclusions.
σAF/γ
σAF/I are beneficial to the nucleation
is dependent on the crystal lattice mismatch
between acicular ferrite and austenite and will, therefore, determine the
crystallographic orientation of the laths. For instance, a low
σAF/γ will arise from a
low lattice mismatch between them.
In the present case where a lack of sufficient inclusions in the steels may occur, the
typical acicular ferrite will be the parallel lath structure due to a low mismatch
between austenite and acicular ferrite (also leading to a low
which is parallel to the primary one.
σAF/I,
σAF/γ) in this direction
however, is dependent on the lattice
structure of the inclusion. An inclusion that has a lower mismatch with an acicular
ferrite
nucleus,
is
favoured
more
to
nucleate
acicular
ferrite.
Some
researchers[141,144,159] have observed that an inclusion with a manganese sulphide core
covered by a skin of copper sulphide acts as a nucleant for acicular ferrite, instead of
manganese sulphide only.
Considering the surface energy
σI/γ between an inclusion and austenite only (which
effectively becomes an additional driving force as it is removed from the system upon
forming an acicular ferrite nucleus), on the other hand, this energy increases with an
increase in their melting temperature[141]. The melting temperature is 1620 ºC for
manganese sulphide, while it is 1125 ºC for copper sulphide[141], which means that
σMnS/γ is higher than σCuS/γ, and this will result in manganese sulphide being a more
effective nucleant than copper sulphide as it provides a higher additional driving force.
However, Evans[160] has reported that a high lattice mismatch exists between
- 170 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
manganese sulphide and ferrite, leading to a high retarding force
σMnS/AF. Therefore,
copper sulphide may be more favourable to nucleate acicular ferrite than manganese
sulphide owing to the lower mismatch between it and the ferrite[65]. A similar
inclusion has also been found in the present study in alloy #3 (see figure 8.20). It is
believed that the inclusion may possibly also have been manganese sulphide in the
core and copper sulphide on the outer skin although the difference between core and
skin were not resolved by EDS owing to the small size of the inclusion. Accordingly,
it can only be speculated that the surface layer of this inclusion may have been
favourable for the nucleation of acicular ferrite as found by others for such types of
inclusions.
As indicated above, –σI/γAI/γ is another driving force for the transformation because
of its negative sign. A high energy interface would preferably be replaced by a low
energy one during the phase transformation, thus lowering the effective energy barrier
for nucleation according to classical nucleation theory. Using the general assumption
that the interface energy of phases or particles increase with their melting temperature,
Dowling[144] has proposed that Al2O3 and SiO2 would be expected to have high
interface energies. These inclusions would, therefore, be efficient nucleants for
acicular ferrite. In this work some aluminium, silicon and manganese oxides were
found within the acicular ferrite laths (see figures 8.17 to 8.19) which were effective
nucleants because these inclusions are of higher melting temperatures, 2015 ºC for
Al2O3, 1713 ºC for SiO2 and 1650 ºC for MnO[65].
The additional driving force –ΔGdefect arises from any defects in the austenite around
inclusions from the deformation in the parent austenite during hot rolling below the
Tnr. Most of these defects are likely to be dislocations as point defects will probably
already anneal out during the hot rolling, even below the Tnr. The total strain in the
finish rolling process below the Tnr of 0.54 that was applied here, would have resulted
in a relatively high dislocation density within the austenite at the point of deformation.
Furthermore, the density of dislocations immediately around inclusions will be higher
than further away because of a concentration in deformation behaviour in the
austenite in those areas. That this is likely to occur, has been shown by a number of
authors in the theory of Particle Stimulated Nucleation (PSN) around large
- 171 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
precipitates after cold working in aluminium alloys[97]. In the PSN theories[161,162], it
has been shown that the deformation zones around large second phase particles
contain a larger number of “geometrically necessary” dislocations than further away
and this leads to a higher driving force for nucleation during recrystallisation at the
surfaces of these particles.
Similar to the above case of PSN, a higher stored defect energy, therefore, possibly
exists in the area of austenite around an inclusion than further away or in undeformed
austenite. This stored energy provides an additional driving force for acicular ferrite
nucleation. It also means that the nucleation barrier for acicular ferrite may be reduced
in the strain and dislocation field around the inclusion, leading to the observation that
deformation in the austenite accelerates the nucleation of acicular ferrite or generates
suitable nucleation sites[134,141]. This point has been demonstrated clearly by the
results of the strain affected CCT diagrams in this study, in which it was demonstrated
in section 7.5 that deformation in the austenite is beneficial to acicular ferrite
formation instead of bainite.
ΔGεVAF is another retarding force of strain energy due to the volume expansion
during the transformation because there is a lattice change from fcc (austenite) to bcc
(acicular ferrite). The amount of this strain energy is dependent on the temperature of
the transformation and lower transformation temperatures will result in a higher strain
energy because the untransformed austenite around acicular ferrite will strain harden
more than at higher temperatures.
Acicular ferrite, however, is a displacive transformation and it is also possible that
defects in the austenite prior to the transformation that may promote the nucleation,
can also retard the growth[76]. The nucleated primary laths formed around inclusions
may, therefore, grow to a smaller size in deformed austenite (see figure 7.27 of alloy
#5 in section 7.5) than in a well annealed austenite. This will result in an overall finer
ferrite grain size after transformation. A coarse acicular ferrite microstructure was
only observed in alloy #5 without prior deformation (figure 7.23 in section 7.4).
- 172 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 8 Studies of acicular ferrite by thin foil TEM
In summary, therefore, deformation in the austenite prior to the transformation
appears to accelerate the nucleation of acicular ferrite and thereafter possibly limits its
growth, so that the size of the formed acicular ferrite packets or grains becomes finer.
- 173 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 9 Discussion
CHAPTER 9 DISCUSSION
9.1 Effect of molybdenum additions on the continuous cooling transformations
9.1.1 Effect of molybdenum on polygonal ferrite formation
(i) With no prior deformation
From figures 7.22 and 7.24 (in section 7.4 above), it is evident that the polygonal
ferrite region in the CCT diagram is more extended at the expense of the pearlite
region in alloy #5 (with 0.22% Mo) than in the Mo-free alloy #6 for a condition of no
prior deformation. From these figures it may be concluded that the temperature range
for polygonal ferrite formation is enlarged in the case of alloy #5 with an 0.22% Mo
addition, although the cooling rate range for polygonal ferrite formation is very
similar (ranging from 0.1 to 4 ºCs-1 for alloy #6 and from 0.1 to 5 ºCs-1 for alloy #5).
Formation of pearlite takes place within the cooling rate range of 0.1 to 1.5 ºCs-1 for
the Mo-free alloy #6. Pearlite, however, can be observed between the cooling rates of
0.1 and 0.4 ºCs-1 with an 0.22% Mo addition in alloy #5. This is clearly due to the Mo
addition in the alloy. Molybdenum is a ferrite forming element and diminishes the
austenite phase field[97]. It can, therefore, promote polygonal ferrite and hinder
pearlite formation under no prior deformation conditions in austenite.
(ii) With prior 45% reduction deformation at 860 ºC below the Tnr
A similar result was obtained with a prior 45% reduction deformation in alloys #5 and
#6 (see figures 7.28 and 7.26 in section 7.5). Polygonal ferrite was observed after all
cooling rates, but formed within a wider temperature range in alloy #5 (with 0.22%
Mo) than in the Mo-free alloy #6 during continuous cooling. The transformation
temperature range for pearlite formation, however, was suppressed in alloy #5 (with
0.22% Mo) in the case of a prior deformation (see figure 7.28). This was an indication
that the pearlite formation region was reduced by the molybdenum addition. It can,
therefore, be concluded that molybdenum addition to Nb-containing low carbon line
pipe alloys with prior deformation, promotes the formation of polygonal ferrite and
inhibits pearlite formation.
- 174 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 9 Discussion
9.1.2 Effect of molybdenum on acicular ferrite formation
(i) With a prior deformation
Figure 7.26 shows the strain affected CCT diagram for the Mo-free alloy #6 after
solution treatment at 1225 ºC to completely dissolve the Nb(C,N) prior to a
deformation of 45% reduction at 860 ºC. The non-recrystallisation temperature Tnr of
alloy #6 ranged from 900 to 930 ºC and was dependent on the deformation parameters
of pass strain, strain rate and inter pass time etc. The deformation prior to cooling for
the CCT determination was, therefore, applied in the non-recrystallisation region.
Figure 7.26 (in section 7.5) showed that the lowest cooling rate limit for acicular
ferrite formation for alloy #6 (Mo-free) was approximately 0.7 ºCs-1 while almost the
same limit for alloy #5 (0.22% Mo) (figure 7.28 in section 7.5), with a prior 45%
reduction, was about 0.5 ºCs-1. These two limits for the lowest cooling rates for
acicular ferrite formation after prior deformation are, therefore, very close to each
other. Furthermore, there was also no measurable difference in cooling rate range
(almost 1 to 40 ºCs-1) for acicular ferrite formation during cooling between the alloys
#5 and #6 (see figures 7.28 and 7.26). It appears, therefore, that a molybdenum
addition of about 0.22% Mo to Nb-bearing line pipe alloys does neither promote or
retard acicular ferrite formation after a 45% prior reduction at 860 ºC below the nonrecrystallisation temperature.
(ii) With no prior deformation
In contrast with the above effects of a prior deformation, the formation of acicular
ferrite is affected by molybdenum additions with no prior deformation. The difference
can be recognised from figures 7.22 and 7.24 (see section 7.4). Acicular ferrite was
found in a wider cooling rate range from 0.1 to 15 ºCs-1 for alloy #5 (with 0.22% Mo)
if compared to the Mo-free alloy #6 in which the region for acicular ferrite formation
was smaller and was observed only within the narrower cooling rate range from 0.3 to
5 ºCs-1.
(iii) Conclusions
On one hand there is, therefore, clear evidence that molybdenum additions to Nbbearing line pipe steels promote acicular ferrite formation under no prior deformation
conditions, as has been observed by a number of other researchers[78, 131,163-165]. On the
- 175 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 9 Discussion
other hand, the contribution of molybdenum additions appears to be over-shadowed
by the effect of deformation in the 0.22% Mo alloy #5 since there was no difference in
the AF-forming regions in the CCT diagrams from that of the Mo-free reference alloy
#6 (figures 7.26 and 7.28) after a prior deformation below the Tnr. It is evident,
therefore, that both molybdenum additions and retained strain in the austenite can
independently promote acicular ferrite formation. A high density of dislocations
induced by the deformation in the austenite prior to transformation is, therefore,
beneficial to the nucleation of acicular ferrite. This effect of retained strain on
promoting acicular ferrite appears to be stronger than the effect of molybdenum
additions.
9.2 Effect of deformation in austenite on acicular ferrite formation
Comparing figures 7.22 (section 7.4) and 7.26 (see section 7.5), the cooling rate range
in which acicular ferrite formed, shifted from 0.3 to 5 ºCs-1 with no prior deformation,
to 0.7 to 40 ºCs-1 with prior deformation, for the Mo-free alloy #6. It is, therefore,
clear that the acicular ferrite region is expanded due to retained strain in the austenite
prior to the phase transformation. It seems that the high density of dislocations in the
deformed austenite is beneficial to the formation of acicular ferrite instead of
bainite[75]. Bainite formation is completely suppressed in deformed austenite (figures
7.25 and 7.26 in section 7.5). Both bainite and acicular ferrite belong to the group that
transform by a displacive transformation[74,76] that may be suppressed by predeformation in the austenite[166-168]. From these observations, however, it appears that
the bainite displacive transformation must have been suppressed in the deformed
austenite whereas the acicular ferrite formation was enhanced.
The main differences between bainite and acicular ferrite formation are related to their
nucleation sites and growth directions. Bainite nucleates on austenite grain boundaries
and grows as a sheaf of parallel plates with the same growth direction within the
austenite grains (figure 7.21-(q) in section 7.4). Acicular ferrite generally nucleates
intragranularly on non-metallic inclusions[70, 144,159, 169,170] and grows as primary plates
in the same[75,78] or many different orientations[75,76] from these inclusions. Then a new
generation of secondary plates nucleates at the austenite/primary acicular plate’s
interface[75,77]. Consequently, acicular ferrite is characterised by a chaotic arrangement
- 176 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 9 Discussion
of plates showing fine grained interwoven morphologies (figure 7.25-(k) to (t) in
section 7.5). The grain structures of acicular ferrite laths are also smaller than those of
bainite. This may be related to the effect of dislocations adjacent to inclusions/primary
plates in deformed austenite that can act as the nucleation sites for secondary plates of
acicular ferrite. The retained strain energy[74] in the areas with a high density of
dislocations can contribute to the mechanical driving force[51] for acicular ferrite
nucleus formation[134]. Dislocations, however, can also hinder the growth of a nucleus
of acicular ferrite[76] owing to the displacive nature of the transformation. Thus both
of the effects of dislocations of, firstly, promoting nucleation of an acicular ferrite lath
and, secondly, suppressing its growth, will lead to an eventual finer acicular ferrite
lath structures. Shipway and Bhadeshia[171] have found that a large enough dislocation
density in austenite introduced by prior deformation, appears to act as a further
nucleation source for Widmanstätten ferrite which forms also by a displacive
transformation mechanism, leading to a refinement of microstructure. Sugden[74] has
also found that acicular ferrite never grows across austenite grain boundaries.
Figure 7.22 in section 7.4 shows that the cooling rate for the formation of bainite is
higher than that required for acicular ferrite. It means that bainite formation requires a
higher degree of under-cooling for nucleation, but once it has nucleated, it grows
relatively quickly. On the contrary, the undercooling required for the formation of
acicular ferrite is lower than that for bainite and its formation might be rate controlled
by both nucleation and growth for the primary plates or secondary nucleation on the
interface between the plate and austenite. Its transformation model is a mixture of
diffusion and shear transformation[4,9,11,43]. Figure 7.22 also shows that the acicular
ferrite formation temperature is higher than that of bainite, which is in agreement with
the work of Zhao[4] and Kim[11]. This higher transformation temperature is helpful for
the carbon diffusion in its rejection from the acicular ferrite primary laths with its
lower solubility for carbon than the untransformed austenite.
The same result was obtained in alloy #5 (with 0.22% Mo) (see figures 7.24 in section
7.4 and 7.28 in section 7.5). The bainite transformation was completely suppressed in
deformed austenite, while the acicular ferrite formation was promoted, even more
with prior deformation than in alloy #6 (Mo-free).
- 177 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 9 Discussion
In summary, retained strain in the austenite promotes the nucleation of acicular ferrite,
hinders its growth, and at the same time, suppresses the formation of bainite. This
results in a shorter lath size of acicular ferrite, which may be beneficial to the
toughness of the steel. An addition of 0.22% Mo to Nb-bearing low carbon line pipe
steels, therefore, does not markedly affect the formation of acicular ferrite to the same
extent as prior deformation of the austenite does.
- 178 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 9 Discussion
9.3 Ratio of yield strength to ultimate tensile strength and its effect
The ratio of yield strength to ultimate tensile strength (YS/UTS) is one of the
properties sometimes specified for low carbon structural steel. In particular, a low
YS/UTS ratio is a very important parameter in the API specifications for line pipe
steels as a high work hardening rate is required in this application. In the present work
a number of differently alloyed steels treated at various cooling rates, two coiling
temperatures and with and without prior deformation in the austenite, were tensile
tested in order to establish the relationships between their YS/UTS ratios and these
materials and process parameters.
9.3.1 The effect of cooling rate
The cooling rate after austenitisation affects the transformed microstructures that
result in different mechanical properties in line pipe steels, such as yield strength (YS),
ultimate tensile strength (UTS), elongation and impact toughness. The two alloys #3
(0.09% Mo) and the reference Mo-free alloy #6 were selected for studying the effect
of cooling rate on the YS/UTS ratio. The temperature range of varying cooling rates
was from 980 ºC down to room temperature.
The effects of cooling rate on the yield strength and ultimate tensile strength are
shown in figures 9.1 and 9.2, respectively. Both the yield strength and ultimate tensile
strength increase with increasing cooling rate after soaking at 980 ºC, as predicted by
the two fitted equations (9.1) and (9.2). The microstructures also changed with
cooling rate. Polygonal ferrite dominated the microstructure at lower cooling rates,
whereas acicular ferrite plus bainite dominated at high cooling rates. The amount of
bainite and acicular ferrite constituents increased with an increase in cooling rate (see
figure 7.22 in section 7.4.1) and this increased the strength of the alloy.
(in MPa)
R2 = 0.98
(9.1)
UTS = 3.9CR+ 572 (in MPa)
R2 = 0.99
(9.2)
YS = 4.1CR + 445
where CR is cooling rate in ºCs
-1
R2 is the Regression Coefficient
- 179 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 9 Discussion
700
650
648
600
603
550
YS, MPa
491
500
551
461
450
439
400
350
300
250
200
0
10
20
30
40
50
60
Cooling rate, ºCs-1
Figure 9.1 The yield strength of the reference Mo-free alloy #6 as a function of the
cooling rate from 980 ºC with no prior deformation before the transformation and
with no coiling simulation.
850
800
750
763
UTS,MPa
700
732
650
613
663
600
587
550
573
500
450
400
0
10
20
30
40
50
60
cooling rate, ºCs-1
Figure 9.2 The ultimate tensile strength of the reference alloy #6 as a function of
cooling rate from 980 ºC with no prior deformation before the transformation and
with no coiling simulation.
- 180 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 9 Discussion
Figure 9.3 The YS/UTS ratio of the reference alloy #6 as a function of cooling rate
from 980 ºC with no prior deformation before the transformation and with no coiling
simulation. PF-polygonal ferrite, AF-acicular ferrite and P-pearlite.
As may be seen in figure 9.3, it appears that the ratio of YS/UTS increases slowly
with increasing cooling rate under conditions of no coiling simulation and no
deformation prior to the transformation. Equations (9.1) and (9.2) show that the slope
of the YS versus CR line is higher than that of the UTS versus CR, i.e. the cooling
rate may affect the YS more than the UTS. This suggests that the dominant structure
of acicular ferrite plus bainite could improve the yield strength more than the ultimate
tensile strength of these steels. Furthermore, the grain size of the steel will also be
finer with increasing cooling rate because of the higher under-cooling at a faster
cooling rate. A high density of dislocations and other crystal defects could also have
been introduced by an acicular ferrite microstructure or bainite formation due to their
displacive formation mechanisms. Such a high density of dislocations has been found
in sample #AF3F from alloy #3 with a cooling rate of 40 ºCs-1 (seen figure 8.18-(b)).
These “prior” defects may interact with other dislocations during early plastic
straining, leading to a high yield strength. Therefore, the finer grain size and a high
density of dislocations after a rapid cooling rate are useful to raise the YS, thereby
increasing the YS/UTS ratio with increasing cooling rate.
- 181 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 9 Discussion
A similar result (see figures 9.4 to 9.6) between YS, UTS, YS/UTS and the cooling
rate was obtained for alloy #3 containing 0.09% Mo after the same treatment as above
for alloy #6. The YS/UTS ratio in alloy #3, however, was not affected to the same
degree by the cooling rate as was the case in alloy #6, compare figures 9.6 with 9.3.
600
550
547
YS, MPa
534
500
542
508
483
475
450
400
350
0
10
20
30
40
50
60
Cooling rate, ºCs-1
Figure 9.4 The yield strength of alloy #3 as a function of cooling rate from 980 ºC
under conditions of no prior deformation to the transformation and no coiling
simulation.
700
650
648
UTS,MPa
640
593
600
625
606
579
550
500
450
0
10
20
30
cooling rate, ºCs
40
50
60
-1
Figure 9.5 The ultimate tensile strength of alloy #3 as a function of cooling rate from
980 ºC under conditions of no prior deformation to the transformation and no coiling
simulation.
- 182 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 9 Discussion
Figure 9.6 The YS/UTS of alloy #3 as a function of cooling rate from 980 ºC under
conditions of no prior deformation to the transformation and no coiling simulation.
PF-polygonal ferrite, AF-acicular ferrite, B-bainite and P-pearlite
9.3.2 The effect of coiling temperature
Two sets of specimens were treated for coiling simulations at 575 and 600 ºC,
respectively, without deformation prior to the transformation. The latter coiling
temperature of 600 ºC has recently been lowered to 575 ºC by Mittal Steel SA in their
line pipe process. The results of alloy #6 are given in figures 9.7 to 9.9 for the coiling
temperature of 600 ºC. It seems that the cooling rate does not markedly affect the
yield strength and ultimate tensile strength of this alloy after coiling at 600 ºC. It was
observed that the ratio of YS/UTS was also not a function of the cooling rate, with the
ratio almost constant, ranging from 0.76 to 0.78. This could be due to the annealing
out of the dislocations and other crystal defects introduced by the displacive
transformation in acicular ferrite or bainite during the coiling simulation process.
- 183 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 9 Discussion
600
YS, MPa
550
494
500
492
502
500
500
478
450
400
350
0
10
20
30
40
Cooling rate,ºCs
50
60
70
-1
Figure 9.7 The yield strength of alloy #6 as a function of the cooling rate from 980 ºC
to 600 ºC under conditions of no prior deformation to the transformation but with a
coiling simulation at 600 ºC for 1 hour.
690
U TS, M Pa
670
651
650
650
650
635
646
630
619
610
590
570
0
10
20
30
40
50
60
70
Cooling rate, ºCs-1
Figure 9.8 The ultimate tensile strength of alloy #6 as a function of cooling rate from
980 ºC to 600 ºC under conditions of no prior deformation to the transformation but
with a coiling simulation at 600 ºC for 1 hour.
- 184 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 9 Discussion
Figure 9.9 The YS/UTS ratio of alloy #6 as a function of the cooling from 980 ºC to
600 ºC under conditions of no prior deformation to the transformation but with a
coiling simulation at 600ºC for 1 hour. PF-polygonal ferrite, AF-acicular ferrite, Bbainite and P-pearlite.
Figures 9.10 to 9.12 show the results of the tensile tests of alloy #6 for a coiling
temperature of 575 ºC. Similar results were obtained for the coiling simulation at 600
ºC with the ratio of YS/UTS ranging from 0.75 to 0.80. It was concluded that varying
the coiling temperature within the range of 575 to 600 ºC did not substantially
influence the ratio of YS/UTS for the entire cooling range of 1 to 60 ºCs-1.
- 185 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 9 Discussion
650
600
550
518
503
502
YS, MPa
500
484
477
450
435
400
350
300
250
200
0
10
20
30
40
50
Cooling rate, ºCs
60
70
-1
Figure 9.10 The yield strength of alloy #6 as a function of the cooling rate from 980
ºC to 575 ºC under conditions of no prior deformation to the transformation but with a
coiling simulation at 575 ºC for 1 hour.
800
750
700
647
UTS, MPa
650
643
651
638
600
638
566
550
500
450
400
350
300
0
10
20
30
40
50
60
70
Cooling rate, ºCs-1
Figure 9.11 The ultimate tensile strength of alloy #6 as a function of the cooling rate
from 980 ºC to 575 ºC under conditions of no prior deformation to the transformation
but with a coiling simulation at 575 ºC for 1 hour.
- 186 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 9 Discussion
Figure 9.12 The YS/UTS ratio alloy #6 as a function of the cooling rate from 980 ºC
to 575 ºC under conditions of no prior deformation to the transformation but with a
coiling simulation at 575 ºC for 1 hour. PF-polygonal ferrite, AF-acicular ferrite, Bbainite and P-pearlite.
9.3.3 The effect of prior deformation in the austenite and coiling simulation
This group of specimens was subjected to a deformation in the austenite region with
33% reduction below the Tnr, before cooling through the transformation to ferrite
from 860 to 575 ºC and then applying a coiling simulation at 575 ºC for 1 hour. The
results from these tests are given figures 9.13 to 9.15. As may be seen, the yield
strength and ultimate tensile strength were affected by cooling rate as well.
- 187 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 9 Discussion
700
650
600
568
528
YS, MPa
550
528
500
514
450
434
400
350
300
250
200
0
5
10
15
20
Cooling rate,
25
30
35
40
-1
ºCs
Figure 9.13 Effect of the cooling rate on the yield strength of the reference alloy #6
after prior deformation of 33 % reduction in the austenite below the Tnr, cooling to
575 ºC at different cooling rates and simulation of the coiling at 575 ºC for 1 hour.
800
750
700
664
650
639
UTS, MPa
626
605
600
550
536
500
450
400
350
300
0
5
10
15
20
25
30
35
40
-1
Cooling rate, ºCs
Figure 9.14 Effect of the cooling rate on the ultimate tensile strength of the reference
alloy #6 after prior deformation of 33 % reduction in the austenite below the Tnr,
cooling to 575 ºC at different cooling rates and simulation the coiling at 575 ºC for 1
hour.
- 188 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 9 Discussion
Figure 9.15 Effect of the cooling rate on the YS/UTS ratio of the reference alloy #6
after prior deformation of 33 % reduction in the austenite below the Tnr, cooling to
575 ºC at different cooling rates and simulation the coiling at 575 ºC for 1 hour. PFpolygonal ferrite, AF-acicular ferrite and P-pearlite.
Comparing the two cases, in the first case without prior deformation (figure 9.12), a
high cooling rate introduces more dislocations into the microstructures because of the
acicular ferrite or bainite microstructures which undergo displacive transformation
from the austenite. The effect of any difference in dislocation density between a low
and a high cooling rate, however, decreased thereafter due to the 60 minutes coiling
simulation at 575 ºC, in spite of the fact that a higher dislocation density had initially
been introduced by the high cooling rate. In other words, the coiling process decreases
the effects of any initial difference in dislocation density in the microstructures caused
by low and high cooling rates.
In the case of the prior deformed alloy (figure 9.15), the dislocations are not only
from the high cooling rate, but also from the prior deformation. However, the
dislocations introduced from the deformation in the austenite do not lead directly to a
higher dislocation content in the acicular ferrite, but it does induce a finer (one pass
deformation at 1050 ºC above the Tnr) and flattened (two passes deformation at 900
and 860 ºC, respectively, below the Tnr) austenite grain size that will provide a high
nucleation rate of the acicular ferrite during the transformation, leading to a finer
- 189 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 9 Discussion
ferrite grain size. The yield strength is sensitive to ferrite grain size[28,172] through the
Hall-Petch relationship[174], whereas the ultimate tensile strength is not as sensitive to
the grain size[28]. All of these can increase the obstacles for the commencement of
plastic deformation and will result in a high yield strength of the steel. Therefore, the
YS/UTS ratio could be raised by deformation as it is sensitive to deformation.
Thus, it can be concluded that:
•
The YS/UTS ratio was found to be only sensitive to cooling rate or the
transformed microstructures in the case without any prior deformation and
simulated coiling at 575 and 600 ºC;
•
Varying the coiling temperature at 575 and 600 ºC did not affect the YS/UTS
ratio but coiling in general diminishes the effect of cooling rate on this ratio;
•
Coiling decreases the YS/UTS ratio in the case without prior deformation; and
•
A prior deformation of 33% below the Tnr in the austenite strongly increases
the YS/UTS ratio at all cooling rates from 1 to 34 ºCs-1 and overshadows the
effect of microstructure or cooling rate.
9.3.4 The effect of acicular ferrite on the ratio of yield strength to ultimate tensile
strength
The relationship between the volume fraction of acicular ferrite and the YS/UTS
(longitudinal specimens) is shown in figure 9.16. The volume fraction of acicular
ferrite was measured on the experimental alloys #1 to #5 after laboratory hot rolling
with an 86% reduction and cooling at a rate of 47 ºCs-1. These microstructures
revealed mainly acicular ferrite mixed with some polygonal ferrite. It appears that the
YS/UTS ratio was not markedly influenced by the quantity of acicular ferrite,
although the range over which the AF content could be varied, was very limited. This
observation is not in agreement with results of other researchers[173] that reported that
there is a high density of mobile dislocations in acicular ferrite which result in a high
work hardening rate and this lowers the number of obstacles for the commencement
of plastic deformation. For instance, Kim has found[172] that increasing the volume
fraction of acicular ferrite or bainite is useful to lower the YS/UTS ratio. Tither[173]
has also reported that dominating acicular ferrite in a microstructure of line pipe steel
could result in a large work hardening rate.
- 190 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 9 Discussion
Summarising the above results, it is concluded that varying the acicular ferrite content
from both undeformed and deformed austenite in these alloys was not measurably
beneficial to lower the YS/UTS ratio, at least not in the quantities used here. An
unresolved question remains, however, whether the effect of the quantity of acicular
ferrite amount is weaker than that of a prior deformation in austenite? This needs
further study with larger variations in the quantity of acicular ferrite than was found
possible here.
0.95
0.9
0.86
0.86
0.84
0.85
0.83
YS/UTS
0.84
0.8
0.75
0.7
0.65
0.6
44.00
46.00
48.00
50.00
52.00
54.00
56.00
58.00
Volume fraction of AF,%
Figure 9.16 Relationship between the YS/UTS ratio (longitudinal specimens) and the
measured volume fraction of acicular ferrite in the experimental alloys #1 to #5 after
laboratory hot rolling with an 86% reduction in total and rapid cooling at a rate of 47
ºCs-1.
- 191 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 10 Conclusions
CHAPTER 10 CONCLUSIONS
The objective of this study was to establish the relationship between micro-alloying
elements, the microstructure and process variables such as cooling rate, prior
deformation in the austenite and coiling conditions, on the ratio of YS/UTS and other
mechanical properties. Therefore, a study was undertaken on the austenite to acicular
ferrite transformation with particular emphasis on the kinetics of the acicular ferrite
formation and as affected by the above process variables. The following conclusions
could be drawn from the present study:
9.1: The type of undissolved particles are mainly (Ti,Nb)(C,N) and Ti(C,N) after
reheating at 1225 ºC for 120 minutes and the smallest size of 46 nm was found in the
Nb-Ti-containing reference alloy #6. The volume fraction of undissolved particles
decreased with increasing reheating temperatures as follows:
fv =10-6 T2 – 0.0038T + 3.2258
R2=0.97
for 15 minutes soaking time
fv =10-6 T2 – 0.0043T + 3.3251
R2=0.99
for 60 minutes soaking time
-6
2
fv =4×10 T – 0.0103T + 6.862
2
R =0.97
for 120 minutes soaking time
9.2: The austenite grain size increased with increasing austenitisation temperature and
soaking time. For alloy #6, however, the effect of the temperature is larger than the
effect of soaking time. When the temperature reaches above 1225 ºC, coarsening of
the particles sets in.
9.3: The non-recrystallisation temperature (Tnr) for the Mo-free reference alloy #6 was
affected by the pass strain and the inter-pass time. These relationships between the Tnr
and the pass strain ε and the inter-pass time tip can be described by the following
respective equations:
Tnr = –210 ε + 972
Tnr = 961tip- 0.0128
9.4: The strain-free and strain affected CCT diagrams for alloy #5 (with 0.22% Mo)
and alloy #6 (Mo-free), have shown that molybdenum additions can shift the acicular
ferrite region on the strain-free CCT diagram to longer times and expand the region
- 192 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 10 Conclusions
for the formation of acicular ferrite. The dominant microstructures with no
deformation in the austenite, were found to have changed from polygonal to an
acicular ferrite microstructure and bainite with increasing cooling rate. A 45%
deformation in the austenite below the Tnr is beneficial to acicular ferrite formation as
the deformation promoted the nucleation of an acicular ferrite microstructure and
hindered the growth of its constituents at the same time, and substantially suppressed
the transformation to bainite. The effect of molybdenum additions on acicular ferrite
transformation in these steels is overshadowed by the stronger effect of the prior
deformation in alloys #5 (with 0.22% Mo) and the Mo-free reference alloy #6.
9.5: The transformed microstructures of all alloys were found to be a mixture of
polygonal ferrite plus an acicular ferrite microstructure. The technique of TEM
examination of shadowed carbon extraction replicas was found to be a superior
method to identify these microstructures, rather than optical microscopy and SEM.
9.6: The acicular ferrite microstructure in the present study was found to have a lath
morphology with two types: parallel laths and interwoven laths but with the more
typical one of parallel laths. A high density of dislocations was found inside the laths.
However, no cementite was observed between laths and within the laths. The
nucleation sites for acicular ferrite were often oxide and sulphide inclusions with
mostly round shape and the observed sizes of these inclusions ranged from 0.35 to 2.2
µm.
9.7: The YS/UTS ratio ranged from 0.83 to 0.86 for the alloys with microstructures of
polygonal ferrite plus acicular ferrite after hot rolling and rapid cooling. However, an
acicular ferrite microstructure or bainite was found to be not beneficial in lowering
the YS/UTS ratio. This ratio was only sensitive to the microstructure or the cooling
rate in the case with no prior deformation and without any simulated coiling process.
The YS/UTS, yield strength and ultimate tensile strength by themselves, however,
were not sensitive to the cooling rate after a simulated coiling process.
9.8: Varying the temperature of the coiling process between 575 and 600 ºC did not
affect the YS/UTS ratio and the simulated coiling process itself diminished the effect
- 193 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 10 Conclusions
of cooling rate and decreased the ratio in the case with no prior deformation.
9.9: Prior deformation with a 33% reduction below the Tnr in the austenite strongly
increased the YS/UTS ratio at all cooling rates from 1 to 34 ºCs-1 and overshadowed
the effect of microstructures or cooling rate on this ratio.
9.10: Molybdenum additions to Nb-Ti micro alloyed steels did not markedly affect the
YS/UTS ratio after a simulated coiling process.
- 194 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Chapter 11 Recommendations for future work
CHAPTER 11 RECOMMENDATIONS FOR FUTURE WORK
1. The influence of the strain rate on the non-recrystallisation temperature in torsion
tests at higher strain rates of more than 2.5 s-1, to simulate strip rolling mill strain rates
more closely.
2. The use of selected area diffraction pattern analysis and misorientation
measurements on acicular ferrite lath structures in thin foil TEM work to determine
whether the parallel laths have the same crystal orientation and or habit planes.
3. More thin foil TEM work to confirm the role played by non-metallic inclusions as
the preferred nucleation sites for acicular ferrite.
4. Larger variations in the quantity of acicular ferrite in these steels to study the effect
of this on the YS/UTS ratio by changing the cooling rate after hot rolling or changing
the parameters of the hot rolling schedule.
- 195 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
References
References
1. M. Pontremoli et al: Met. Tech., Nov.,11(1984), pp504-514
2. J.G. Williams et al: Micro alloying ’95 Conference Proceeding, pp117-139
3. Jon P. Orton: Micro-alloying 75 Proceedings, Publ. Union Carbide Corp.,
Washington, D.C. USA, 1977, pp334-347
4. Ming-Chun Zhao, Ke Yang, Fu-Ren Xiao and Yi-Yin Shan: Mater. Sci. Eng., A355
(2003), pp126-136
5.William Roberts: Conference Proceedings of International Conference on
Technology and Applications of HSLA Steels, American Society for Metals, Metals
Park, Ohio 44073, 3 - 6 Oct., 1983, pp33-65
6. J.G. Williams et al: Materials Forum, 20(1996), pp13-28
7. R. Mendoza and M. Alanis et al: Mater. Sci. Eng., A337(2002), pp115-120
8. Andreas Kern, Joachim Degenkolbe, Bruno Musgen and Udo Schriever: ISIJ Int.,
32(3)(1992), pp387-394
9. Min-Chun Zhao, Ke Yang and Yi-Yin Shan: Mater. Sci. Eng., A335(2002), pp14-20
10. Akihiko Takahashi and Makio Iino: ISIJ Int., 36(2)(1996), pp235-240
11. N.J. Kim: Journal of Metals, April (1983), pp21-27
12. J.J.Withfield: Options for manufacturing X70-X80 line pipe, Internal report
ISCOR (new Mittal Steel (SA)), South Africa, (1998)
13. C.S. Chiou et al: Materials Chemistry and Physics, 69(2001), pp113-124
14. Yu Matrosov: Metal Science and Heat Treatment of Metals, Issue 28, 3-4(1986),
pp173-180
15. C.O.I.Emenike and J.C.Billington: Mater. Sci. and Tech., May, 5(1989),
pp450-456
16. Matrosov, Yu: Metal Science and Heat Treatment (English) Translation of
Metallovedenie i Termicheskaya Obrabotka, Issue 11-12, Nov-Dec., 26(1984),
pp798-807
17. W.P.Sun, M.Militzer, D.Q.Bai and J.J.Jonas: Acta Metall. Mater., 41(1993),
- 196 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
References
pp3595-3604
18. S.F.Medina: Mater. Sci. Tech., 14(1998), pp217-221
19. S.F.Medina and J.E.Mancilla: Acta Metall. Mater., 42(1994), p3945-3951
20. M.G.Mecozzi, J.Sietsma and S.van der Zwaag: Acta Materialia, 54(2006),
pp1431-1440
21. P.Cizek, B.P.Wynne, C.H.J.Dvies, B.C.Muddle, P.D.Hodgson: Metall. Mater.
Trans., A, 33A(2002), pp1331-1349
22. A.I. Fernandez et al: Mater. Sci. Eng., A361(2003), pp367-376
23. C.I.Smith, R.R.Preston and N.L.Richards: Met. Tech., 5(1978), pp341-350
24. F. Heisterkamp AND k.hULKA: Metals Tech., Dec.,11(1984), pp535-544
25.B.de.Meester: ISIJ Int. 37(6) (1997), pp537-551
26.D.Dormagen: Steel Res., 59(8) (1988), pp368-374
27. J.J.Jonas and I.Weiss: Met. Sci., 13(1979), pp238-245
28. S.N.Prasad, D.S.Sarma: Mater. Sci. Eng., A408(2005), pp53-63
29. S.F.Medina, M.Chapa, P.Valles, A.Quispe and M.I.Vega: ISIJ Int., 39(9)(1999),
pp930-936
30. T.J.George and N.F.Kennon: Aust. Inst. Met., 17(1972), pp73-80
31. T.George and J.J.Irani: J. Aust. Inst. Met., 13(1968), pp94-106
32. Kong Junhua, Zhen Lin, Guo Bin, Li Pinghe, wang Aihua and Xie Changsheng:
Materials and design, 25(2004), pp723-728
33. Michel L. Lafrance, Francis A Caron and Guy R. Lamant “use of Microalloyed
Steels in the Manufacture of Controlled Rolled Plates for Pipe” Micro-alloying 75
Proceedings, Publ. Union Carbide Corp., Washington, D.C. USA, 1977,
pp367-374
34. He Kejian and T.N. Baker: Mater. Sci. Tech., Dec.,8(1992), pp1082-1089
35. C.O.I.EMENIKE: J. Mater. Sci. Letters, 9(1990), pp406-409
36. P.D. Hodgson et al: ISIJ Int., 32(12)(1992), pp1329-1338
37. Shuji OKAGUCHI and Tamotsu HASHIMOTO: ISIJ Int., 32(3)(1992),
pp283-290
- 197 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
References
38. C.O.I. Emenike: Pipes & Pipelines International, Nov. - Dec., (1988), pp15-22
39. Atsuhiko YOSHIE, Masaaki Fujioka, Yoshiyuki Watanabe, Kiyoshi and Hirofumi
Morikawa: ISIJ Inter., 32(3)(1992), pp395-404
40. Ming-Chun Zhao, Ke Yang and Yi-Yin Shan: Materials Letters, 57(2003),
pp1496-1500
41. A.M. Sage : Metals Technology , March (1981), pp94-102
42. C.O.I. Emenike et al: Pipes & Pipelines International Jan. - Feb. (1990), pp24-27
43. Sunghak Lee, Dongil Kwon, Young Kook Lee and Ohjoon Kwon: Metall. Mater.
Trans. A, May, 26A(1995), pp1093-1100
44. M. Charleux et al: Metall. Mater. Trans. A, July, 32A(2001), pp1635-1647
45. A.Yoshie and M.Fujioka: ISIJ Int. 32(1992), pp395-404
46. J.J.Withfield: Thermo-mechanical Processing, Report number: TMP98/040,
Internal report, ISCOR (new Mittal Steel (SA)), South Africa, (1998)
47. S.F.Medina and A.Quispe: Mater. Sci. Tech., June, 16(2000), pp635-642
48. D.Q. Bai et al: ISIJ Int., 36(8) (1996), pp1084-1093
49. C. Ouchi et al: ISIJ. Int., 20 (1980), pp833-842
50. H.L. Andrade, M.G. Akben, and J.J. Jonas: Metall. Trans. A, Oct, 14A(1983),
pp1967-1977
51. S.S.Babu and H.K.D.H. Bhadeshia: Mater. Sci. Eng., A156(1992), pp1-9
52. F.J.Barbaro, P.Kraulis, K.E.Easterling: Mat. Sci. Tech., 5(1989), pp1057-1068
53. A.O.Kluken, O.Grong and G.Rorvik: Metall. Trans. A, 21A(1990), pp2047-2058
54. J.M.Gregg and H.K.D.H.Bhadeshia: Acta Mater., 45(1997), pp739-748
55. F.Ishikawa and T.Takahashi: ISIJ Int., 35(1995), pp1128-1133
56. F.Ishikawa, T.Takahashi and T.Ochi: Metall. Mater. Trans., A, 25A(1994),
pp929-936
57. S.Ohkita, H.Homma, S.Tsushima and N.Mori: Aus. Welding J., (1984), pp29-36
58. S.Liu and D.L.Olson: Weld. J., 65(1986), pp139s-149s
59. D.W.Oh, D.L.Olson and R.H.Frost: Weld. J., 69(1990), pp151s-158s
60. R.B.Oldland: Aust. Weld. Res., 14(1985), pp44-56
- 198 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
References
61. G.M.evans: Weld. J., 70(1991), pp32s-39s
62. H.Terashima and P.H.M.Hart: Weld. J., 63(1984), pp173s-182s
63. S.Ohkita, H.Homma, S.Tsushima and N.Mori: Aust. Weld. J., 29(1984), pp29-36
64. A.G.Glover, J.T.McGRATH, M.J.Tinkler and G.C.Weatherly: Weld. J., 56(1977),
pp267s-273s
65. M.Diaz-Fuentes, I.Madariaga and I.Gutierrez: Materials science Forum,
284-286, (1998), pp245-252
66. I.Madariaga and I.Gutierrez: Materials science Forum, 284-286(1998),
pp419-426
67. M.A.Linaza, J.L.Romero. J.M.Rodriguez-Ibabe and J.J.Urcola: Scripta Metall.
Mater., 29(1993), pp1217-1222
68. M.A.Linaza, J.L.Romero. J.M.Rodriguez-Ibabe and J.J.Urcola: Scripta Metall.
Mater., 32(1995), pp395-400
69. I.Madariaga and I.Gutierrez: Scripta Mater., 37(8), (1997), pp1185-1192
70. I.Madariaga, J.L.Romero and I.Gutierrez: Metall. Trans., 29A(1998),
pp1003-1015
71. S.Hoekstra, M.A.Munnigschmidt-van der Berg and G.Denouden: Met. Contr.,
18(12) (1986), pp771-775
72. S.W.Thompson, D.J.Colvin and G.Krauss: Metall. Tran. A, 21A(1990),
pp1493-1507
73. Ming-chun Zhao, Ke Yang and Yi-Yin Shan : Mater. Sci. Eng., A335 (2002), pp
14-20
74. A.A.B. Sugden and H.K.D.H Bahadeshia: Metall. Trans. A, Sept, 20A(1989),
pp1811-1818
75. Madariaga, I. Gutierrez and H.K.D.H. Bhadesia: Metall. Mater. Trans. A, Sept,
32A(2001), pp2187-2197
76. C.H. Lee and H.K.D.H. Bhadeshia: Mater. Sci. Eng., A360(2003), pp249-257
77. J.R.Yang and H.K.D.H.Bhadeshia: Mater. Sci. Tech., 5(1989), pp93-97
78. M. Diaz-Fuentes et al: Mater. Sci. Eng., A363(2003), pp316-324
- 199 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
References
79. J.M.Gregg and H.K.D.H.Bhadeshia: Acta Mater., 45(2)(1997), pp739-748
80. A.A.B.Sugden and H.K.D.H.Bhadeshia: Metall. Trans.A, 20A(1989),
pp1811-1818
81. J.G.Garland and P.R.Kirkwood: Met. Constr., 7(5), (1975), pp275-283
82. R.A.Farrar and P.L.Harrison: J. Mater. Sci., 22(1987), pp3812-3820
83. T. Mohandas and GM.Reddy: J. Mater Process Techol., 69(1-3)(1998),
pp222-226
84. K.Yamamoto, T.Hasegawa and J.Takamura: ISIJ Int., 36(1996), pp80-86
85. M.Ferrante and R.A.Farrar: J. Mater. Sci., 17(1982), pp3293-3298
86. D.J.Abson and R.E.Dolby: Weld. Inst.Res. Bull., 19(1978), pp202-207
87. R.A.Ricks, P.R.Howell and G.S.Barritte: J. Mater. Sci., 17(1982), pp2218-2226
88. S.A.Court and G.Pollard: Metallography, 22(1989), pp219-243
89. R.C.Cochrane: Weld. World, 21(1983), pp16-24
90. R.A.Ricks, P.R.Howell and G.S.Barritte: J. Mater. Sci., 17 (1982), pp732-740
91. J.M.Chilton and M.J.Roberts: Metall. Trans. A, 11A(1980), pp1711-1721
92. Simith Ye, Coldren AP and Cryderman RL: Met. Sci. Heat Treat., 18(1-2)(1976),
pp59-65
93. Tomo Tanaka: “Bauschinger Effect during Pipe-forming Operations”,
Micro-alloying 75 Proceedings, Publ. Union Carbide Corp., Washington, D.C.
USA, 1977, pp350-352
94 Franz M. Oberhauser: “Bauschinger effect under Multi-axial Loading Conditions”,
Micro-alloying 75 Proceedings, Publ. Union Carbide Corp., Washington, D.C. USA,
1977, pp349-349
95. M. Grumach: “Influence of work hardening and Bauschinger effect on
plate-to-pipe yield-strength differences”, Micro-alloying 75 Proceedings, Publ.
Union Carbide Corp., Washington, D.C. USA, 1977, pp348-348
96. Tetsuo Yamaguchi: “Relative Importance of Work Hardening and Bauschinger
Effect on Strength during Pipe Forming” Micro-alloying 75 Proceedings, Publ.
Union Carbide Corp., Washington, D.C. USA, 1977, pp352-353
- 200 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
References
97. Waldo Stumpf: Class notes on Phase Transformations in Metals and Their Alloys,
University of Pretoria, RSA, (2003), p1.9-12, p5.1-1, p5.15-2, p8.1-10, p11.7-3
98. M.F.Ashby and R.Ebeling: Transactions of the Metallurgical Society of AIME,
Oct, 236(1966), pp1396-1404
99. D.R.Barraclough and C.M.Sellar: Met. Sci., 13(1979), pp257-267
100. M.J.Luton, R.A.Petkovie and J.J.Jonas: Acta Metall., 28(1979), pp729-743
101. G.G.Glover and C.M.Sellars: Metall. Trans., 4(1973), pp765-775
102. R.A.P.Djaic and J.J.Jonas: Metall. Trans., 4(1973), pp621-625
103. Y.Iwahashi, Z.Horita, M.Nemoto and T.G.Langdon: Acta Mater., 45(1997),
pp4733-4741
104. S.F.Medina A.Quispe, P.Valles and J.L.Banos: ISIJ Int., 39(1999), pp913-922
105. A.Quispe, S.F.Medina and P.Valles: ISIJ Int., 37(1997), pp783-788
106. S.S.Hansen, J.B.Vandersande and M.Cohen: Metall. Trans. A, 11A(1980),
pp387-402
107.A.LeBon, J.Rofes-Vernis and C.Rossard: Met. Sci., 9(1975), pp36-40
108. T.Sakai and J.J.Jonas: Acta Metall., 32(1984), pp189-209
109. G.Gottstein and U.F.Kocks: Acta Metall., 31(1983), pp175-188
110. A.Belyakov, R.Kaibyshev and T.Sakai: Metall. Trans. A, 29A(1998), pp161-167
111. W.P.Sun, M.Militzer and J.J.Jonas: Metall. Trans. A, 23A(1992), pp3013-3023
112. I.Weiss and J.J.Jonas: Metall. Trans. A, 11A(1980), pp403-410
113. J.G.Speer and S.S.Hansen: Metall. Trans. A, 20A(1989), pp25-38
114. L.J. Cuddy: Metall. Trans., A, 12A(1981), pp1313-1320
115. C.M.Sellars and J.A.Whiteman: Metal Sci., 13(1979), pp187-194
116. L.J.Cuddy: Metall. Trans. A, 15A(1984), pp87-98
117. L.N.Pussegoda, S.yue, and J.J.Jonas: Metall. Trans. A, 21A(1990), pp153-164
118. L.N.Pussegoda and J.J.Jonas: ISIJ. Int., 31(3)(1991), pp278-288
119. F.H.Samuel, S.Yue, J.J.Jonas, and B.A.Zbinden: ISIJ. Int., 29(10)(1989),
pp 878-886
120. F.H.Samuel, S.Yue, J.J.Jonas, and K.R.Barnes: ISIJ. Int., 30(3)(1990),
- 201 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
References
pp 216-225
121. S.Yue and J.J.Jonas: Mater. Forum, 14(1990), pp245-252
122. K.J. Irvine, F.B.Pickering, and T.Gladman: J. Iron Steel Inst., 205(1967),
pp161-182
123. B.Dutta and C.M.Sellars: Mater. Sci. Tech., 3(1987), pp197-206
124. A.Laasraoui and J.J.Jonas: Metall. Trans. A, 22A(1991), pp151-160
125. D.Q.Bai, S.Yue, W.P.Sun and J.J.Jonas: Metall. Trans. A, Oct, 24A(1993),
pp2151-2159
126. G.K.Prior: Materials Forum, 18(1994), pp265-276
127. T.Mukherjee, W.E.Stumpf and C.M.Sellars: J. of Mater. Sci. 3(1968), pp127-135
128. L.J.Cuddy, J.J.Bauwin and J.C.Raley: Metall. Trans. A, 11A(1980),
pp381-386
129. M.G.Akben, I.Weiss, and J.J.Jonas: Acta Met., 29(1981), pp111-121
130. A.Le. Bon,J.Rofes-Vernis and C.Rossard: Metal Sci., 9(1975), pp36-40
131. Furen Xiao, Bo Liao Deliang Ren, Yiyin Shan and Ke yang: Materials
Characterization, 54 (2005), pp205-314
132. P.H.Shipway and H.K.D.H.Bhadeshia: Mater. Sci. Tech., 11(1995), pp1116-1128
133. S.B.Singh and H.D.K.H. Bhadeshia: Mater. Sci. Tech., 12 (1996), pp610-612
134. G.I.Rees and H.K.D.H.Bhadeshia: Mater. Sci. Tech., May, 10(1994),
pp353-358
135. V.Biss and R.L.Cryderman: Metall. Trans., 2(1971), pp2267-2276
136. S.S.Babu and H.K.D.H.Bhadeshia: Mater. Sci. Tech., 6(1990), pp1005-1020
137. H.K.D.H.Bhadeshia, L.E.Svensson and B.Gretoft: Acta Metall., 33(1985),
pp1271-1283
138. J.G.Garland and P.R.Kirkwood: Met. Constr., 7(6)(1975), pp320-330
139. H.K.D.H.Bhadeshia: Scripta Metall., 21(1987), pp1017-1022
140. P.L.Mangonon: Metall. Trans. A, 7A (1976), pp1389-1400
141. Z.Zhang and R.A.Farrar: Mater. Sci. Tech., 12(1996), pp237-260
142. A.R.Mills, G.Thewlis and J.A.Whiteman: Mat. Sci. Tech., 3(1987),
- 202 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
References
pp1051-1061
143. P.L.Harrison and R.A.Farrar: J. Mater. Sci., 16(1981), pp2218-2226
144. J.M.Dowling J.M.Corbett and H.W.Kerr: Mettal. Trans. A, 17A(1986),
pp1611-1623
145. G.M.Evans: Weld. J., 71(1992), pp447s-454s
146. T.Funakoshi, T.Tanaka, S.Ueda, M.Ishikawa, N.Koshizuka and K.Kobayashi:
Transactions ISIJ., 17(1977), pp419-427
147. A.R.Bhatti, M.E.Saggese, D.N.Hawkins, J.A.Whiteman and M.S.Golding:
Welding Journal, 63(1984), pp224s-230s
148. G.M.Evans: Weld. J., 72(1993), pp123s-133s
149. T.N.North, H.B.Bell, A.Koukabi and I.Craig: Welding Journal, 58(1979),
pp343s-354s
150. S.G.Court and G.Pollard: J. Mater. Sci. Letters, 4(1985), pp427-430
151. J.E.Harbottl and S.B.fisher: Nature, 299(1982), pp139-140
152. E.S.Kayali, I.M.Corbett and H.W.Kerr: J. Mater. Sci. Letters, 2(1983),
pp123-128
153. G.S.Barritte and D.V.Edmonds: “Advances in the Physical Metallurgy and
Applications of Steel”, 1982, London, The Metals Society, pp126-135
154. B.Ralph: Mater. Sci. Tech., 6(1990), pp1139-1144
155. D.J.Abson: weld World, 27(1989), pp76-100
156. G.Thewlis: Join. Mater., 2(1989), pp25-32
157. G.Thewlis: Join. Mater., 2(1989), pp125-129
158. A.R.Bhatti, M.E.Saggese, D.N.Hawkins, J.A.Whiteman and M.S.Golding: Weld.
J., 63(1984), pp224s-230s
159. I.Madariaga and Gutierrez: Acta Mater., 47(3), (1999), pp951-960
160. G.M.Evans: Met. Constr., 18(1986), pp631R-636R
161. R.D.Doherty: Mater. Sci. Eng., A238(1997), pp219-274
162. J.D.Embury: Met. Trans., 16A (1985), pp2191-2200
163. C.Garcĺa de Andres, C.Capdevila, I.Madariaga and I.Gutierrez: Scripta
- 203 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
References
Materialia, 45(2001), pp709-716
164. Pavel Morcinek: Micro-alloying 75 Proceedings, Publ. Union Carbide Corp.,
Washington, D.C. USA, 1977, pp272-278
165. E.C.Hamre and A.M.Gilroy-Scott: Micro-alloying 75 Proceedings, Publ. Union
Carbide Corp., Washington, D.C. USA, pp375-381
166. C.Y.huang, J.R.Yang and S.C.Wang: Mater. Trans., JIM, 34(1993), pp658-668
167. R.H.Edwards and N.F.Kennon: Metall. Trans. A, 9A(1978), pp1801-1809
168. R.Freiwillig, J.kudraman and P.chraska: Metall. Trans. A, 7A(1976),
pp1091-1097
169. O.Grong and D.K.Matlock: Int. Met.Rev., 31(1)(1986), pp27-44
170. E.S.Kayali, J.M.Corbett and H.W.Kerr: J. Mater. Sci. Lett., 2(1983), pp123-128
171. P.H.Shipway and H.K.D.H.Bhadeshia: Mater. Sci. Eng., A223(1997), pp179-185
172. Y.M.Kim, S.K.Kim, Y.J.Lim and N.J.Kim: ISIJ Int., 42(12) (2002), pp1571-1577
173. G.Tither and M.Lavite: J. of Met. (JOM), (1975), p15-23
174. Waldo Stumpf: Class notes on Mechanical Metallurgy, University of Pretoria,
South Africa, (2002), p3.3-11
- 204 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Appendix
Appendix A
Parameters of the laboratory hot rolling process
Table A-1 The hot rolling parameters for alloy #2
Pass No
R1
R2 reheating R3
Tempera- in 1165 1050
ture (ºC)
out
--
20
10
in
43
37
28
20
out
37
28
20
0.15 0.28
Pass ε
Total ε
Reduction
(%)
ε& (s-1)
R5
F1 reheating
1080 1030 990 910
23
tip (s)
Gauge
(mm)
1225
5 min
R4
13
930
5 min
--
F2
880
930
5 min
890
865
--
--
13.6 10.3
8.3
6.9
13.6 10.3 8.3
6.9
6
0.34 0.38 0.28 0.22
0.18
0.14
1.43
0.54
76
42
1.67 2.43
reheating F3
3.15 3.92 4.00 4.07
4.00
3.89
Table A-2 The hot rolling parameters for alloy #3
Pass No
R1
R2 reheating R3
Tempera- in 1162 1043
ture (ºC) out
Pass ε
--
18
12
in
43
37
28
20
out
37
28
20
0.15 0.28
Total ε
Reduction
(%)
ε& (s-1)
1.67 2.43
R5
F1 reheating
1045 1032 980 910
24
tip (s)
Gauge
(mm)
1225
5 min
R4
18
--
930
5 min
F2
885
reheating F3
930
5 min
890
861
--
--
13.6 10.3
8.3
6.9
13.6 10.3 8.3
6.9
6
0.34 0.38 0.28 0.22
0.18
0.14
1.43
0.54
76
42
3.15 3.92 4.00 4.07
- 205 -
4.00
3.89
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Appendix
Table A-3 The hot rolling parameters for alloy #4
R1
Pass No
R2 reheating R3
Tempera- in 1165 1065
Ture (ºC) out
--
19
14
in
43
37
28
20
out
37
28
20
0.15 0.28
Pass ε
Total ε
Reduction
(%)
ε& (s-1)
R5
F1 reheating
1082 1050 1000 910
22
tip (s)
Gauge
(mm)
1225
5min
R4
17
930
5min
--
F2
890
930
5min
889
860
--
--
13.6 10.3
8.3
6.9
13.6 10.3 8.3
6.9
6
0.34 0.38 0.28 0.22
0.18
0.14
1.43
0.54
76
42
1.67 2.43
reheating F3
3.15 3.92 4.00 4.07
4.00
3.89
Table A-4 The hot rolling parameters for alloy #5
Pass No
R1
R2 reheating R3
Tempera- in 1148 1058
ture (ºC)
out
Pass ε
--
17
13
in
43
37
28
20
out
37
28
20
0.15 0.28
Total ε
Reduction
(%)
ε& (s-1)
1.67 2.43
R5
F1 reheating
1081 1020 1010 910
22
tip (s)
Gauge
(mm)
1225
5 min
R4
23
--
930
5 min
F2
870
reheating F3
930
5 min
885
857
--
--
13.6 10.3
8.3
6.9
13.6 10.3 8.3
6.9
6
0.34 0.38 0.28 0.22
0.18
0.14
1.43
0.54
76
42
3.15 3.92 4.00 4.07
- 206 -
4.00
3.89
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Appendix
Table A-5 The hot rolling parameters for sample #M1-11 of alloy #6
Pass No
Temperature (ºC)
R1
Pass ε
R3
t-ing
R4
R5
R6
940
901
Rehea
t-ing
1200
F1
F2
925
--
30 s
--
18
22
--
20
--
--
--
--
in
45
37
30
25
20
15
10
8.3
out
37
30
25
20
15
10
8.3
6.9
0.196 0.21 0.182
Total ε
Reduction
(%)
ε& (s-1)
Rehea
in 1140 1108 1008 1200 1085
5 min
out
tip (s)
Gauge
(mm)
R2
1.82
2.08
2.09
0.223 0.288 0.405
0.186 0.18
1.50
0.51
78
40
2.56 3.31
- 207 -
4.66
3.05 4.00
Rehea
t-ing
1200
60 s
F3
920
870
6.9
6
0.14
3.89
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Appendix
Appendix B
Curves of tensile tests
Conditions on the Gleeble:
Reheating temperature: 1225 ºC
No prior deformation and coiling simulation
Alloy #6
700
600
600
500
Alloy #6
Stress,M Pa
Stress,M Pa
500
Sample group: A124
Sample
samplegroup:A124
#AF-124F01A
400
AlloyCooling
#6 rate: 1 篊/s (980 to 25篊)
300
Cooling rate: 1 ºCs-1
200
Sample group:A124
400
Sample group: A124
sample #AF-124F01C
Alloy
Alloy #6 #6
300
Cooling rate: 1 篊/s (980 to 25篊)
Cooling rate: 1 ºCs-1
200
(980 to 25 ºC)
100
(980 to 25 ºC)
100
0
0
0
5
10
15
20
25
30
35
40
0
45
5
10
15
20
700
700
600
600
30
35
40
45
35
40
45
500
Sample group: A124
Sample group:group:A124
A124
Sample
400
Stress ,M Pa
S tre s s,M P a
500
sample #AF-124F05B
Alloy #6
Alloy
#65 篊/s (980 to 25篊)
Cooling rate:
300
Cooling rate: 5 ºCs-1
200
400
sample #AF-124F05D
Sample
group:A124
300
Alloy #6
200
Cooling rate: 5 ºCs-1
Alloy #6
Cooling rate: 5 篊/s (980 to 25篊)
(980 to 25 ºC)
100
(980 to 25 ºC)
100
0
0
5
10
15
20
25
30
35
40
0
45
0
5
10
15
elong,%
20
25
30
elong,%
700
700
600
600
500
500
400
Sample
group:A124
Sample group: A124
300
sample #AF-124F10A
Alloy #6 #6
Alloy
Cooling rate: 10 篊/s (980 to 25篊)
Sample group: A124
Stress, M Pa
Stress, M Pa
25
elong,%
elong,%
Cooling rate: 10ºCs-1
200
(980 to 25 ºC)
100
400
sample #AF-124F10C
Sample
group:A124
300
Alloy #6
200
Cooling rate: 10 ºCs-1
Alloy #6
Cooling rate: 10 篊/s (980 to 25篊)
(980 to 25 ºC)
100
0
0
0
5
10
15
20
25
30
35
40
45
0
elong,%
5
10
15
20
25
elong,%
- 208 -
30
35
40
45
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
700
700
600
600
500
Sample
group:A124
Sample group:
A124
400
sample #AF-124F20A
Alloy #6 #6
Alloy
Cooling rate: 20 篊/s (980 to 25篊)
300
Cooling rate: 21ºCs-1
(980 to 25 ºC)
200
Sample group: A124
500
Stress, M Pa
Stress, M Pa
Appendix
100
sample #AF-124F20D
Sample
group:A124
Alloy #6
Cooling rate: 20 篊/s (980 to 25篊)
400
Alloy #6
300
Cooling rate: 21ºCs-1
200
(980 to 25 ºC)
100
0
0
0
5
10
15
20
25
30
35
40
0
5
10
15
elong,%
30
35
40
700
700
600
500
Sample
group:A124
Sample group:
A124
400
sample #AF-124F40A
Alloy #6
Alloy
#6
Cooling rate: 40 篊/s (980 to 25篊)
300
Stress, M Pa
600
Stress, MPa
25
800
800
Cooling rate: 40ºCs-1
Sample group: A124
500
Sample
group:A124
sample #AF-124F40B
Alloy #6
400
Cooling rate: 40 篊/s (980 to 25篊)
Alloy
#6
300
Cooling rate: 40ºCs-1
200
200
(980 to 25 ºC)
(980 to 25 ºC)
100
100
0
0
0
5
10
15
20
25
30
35
0
40
900
800
700
600
Sample group:A124
Sample group: A124
sample #AF-124F60A
500
400
Alloy
#6
Alloy #6
300
Cooling rate: 51ºCs-1
200
(980 to 25 ºC)
Cooling rate: 60 篊/s (980 to 25篊)
100
0
0
5
10
15
20
25
30
35
5
10
15
20
elong,%
elong,%
Stress, M Pa
20
elong,%
40
elong,%
- 209 -
25
30
35
40
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Appendix
Appendix C
Curves of tensile tests
Conditions on the Gleeble:
Reheating temperature: 1225 ºC
No prior deformation and coiling simulation
Alloy #3
700
700
600
600
500
Sample group:AF3F
400
400
Sample
Alloy
#3group: AF3F
sample #AF-3F01A
Alloy #3
-1
Cooling
rate:
Cooling rate:
1 篊1/s ºCs
(980 to 25篊)
300
Sample group:AF3F
S tre ss , M P a
Stress, M Pa
500
Sample group: AF3F
sample#3
#AF-3F01B
Alloy
300
200
Alloy #3
Cooling rate: 1 篊/s (980 to 25篊)
Cooling rate: 1 ºCs-1
200
(980 to 25 ºC)
(980 to 25 ºC)
100
100
0
0
5
10
15
20
25
30
35
40
45
0
50
0
5
10
15
elongation ,%
30
35
40
45
600
600
500
500
Sample group:AF3F
400
S tre s s, M P a
S tre s s, M P a
25
700
700
Sample group: AF3F
Alloy
#3#AF-3F05A
sample
Alloy #3
Cooling rate:
5 篊/s5 (980
Cooling
rate:
ºCsto -125篊)
300
200
(980 to 25 ºC)
100
400
SampleSample
group:AF3F
group: AF3F
300
Alloy #3
Alloy #3
200
Cooling rate: 5 ºCs-1
sample #AF-3F05B
Cooling rate: 5 篊/s (980 to 25篊)
(980 to 25 ºC)
100
0
0
0
5
10
15
20
25
30
35
40
0
45
5
10
15
20
25
30
35
40
45
30
35
40
45
elongation, %
elongation, %
700
700
600
600
500
500
Sample
group:AF3F
Sample group: AF3F
400
S tre s s, M P a
Stress, M Pa
20
elongation, %
sample #AF-3F10A
AlloyAlloy
#3#3
Cooling rate: 10 篊/s (980 to
25篊)
-1
300
Cooling rate: 10 ºCs
200
Sample group:AF3F
Sample group: AF3F
sample #AF-3F10C
400
AlloyAlloy
#3#3
300
Cooling rate: 10 篊/s (980 to
25篊) rate: 10 ºCs-1
Cooling
200
(980 to 25 ºC)
(980 to 25 ºC)
100
100
0
0
10
20
30
40
50
60
0
0
elongation, %
5
10
15
20
25
elongation, %
- 210 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
700
700
600
600
500
500
Sample
group:AF3F
Sample group: AF3F
400
S tress, M P a
S tre s s, M P a
Appendix
sample #AF-3F20A
Alloy #3 #3
Alloy
Cooling rate: 20 篊/s (980 to 25篊)
300
Cooling rate: 20 ºCs-1
200
Sample group: AF3F
sample#3
#AF-3F20D
Alloy
Alloy #3
Cooling rate: 20 篊/s (980 to 25篊)
300
Cooling rate: 20 ºCs-1
200
(980 to 25 ºC)
100
Sample group:AF3F
400
(980 to 25 ºC)
100
0
0
0
5
10
15
20
25
30
35
40
45
0
5
10
15
elongation, %
700
700
600
600
500
Stress, M Pa
Sample
group:AF3F
Sample group:
AF3F
sample #AF-3F40B
Alloy #3#3
Alloy
Cooling rate: 40 篊/s (980 to
25篊)
300
Cooling rate: 40 ºCs-1
200
25
30
35
40
Sample group: AF3F
Sample
group:AF3F
sample #AF-3F40C
Alloy #3
400
Cooling
Alloy
#3rate: 40 篊/s (980 to
25篊)
300
Cooling rate: 40 ºCs-1
200
(980 to 25 ºC)
(980 to 25 ºC)
100
100
0
0
0
5
10
15
20
25
30
35
40
0
45
5
10
15
20
25
30
35
40
elongation, %
elongation, %
700
700
600
600
500
500
Sample group:AF3F
Stress, M Pa
Stress, M Pa
Stress, M Pa
500
400
20
elongation, %
Sample group: AF3F
sample
Alloy
#3#AF-3F60A
Alloy #3
Cooling rate: 54 篊/s (980 to 25篊)
400
300
Cooling rate: 54 ºCs-1
200
Sample group:
AF3F
Sample
group:AF3F
sample #AF-3F60C
Alloy #3
400
Alloy
#3rate: 54 篊/s (980 to 25篊)
Cooling
300
Cooling rate: 54 ºCs-1
200
(980 to 25 ºC)
(980 to 25 ºC)
100
100
0
0
0
5
10
15
20
25
30
35
40
45
elongation, %
0
5
10
15
20
25
elongation, %
- 211 -
30
35
40
45
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Appendix
Appendix D
Curves of tensile tests
Conditions on the Gleeble:
Reheating temperature: 1225 ºC
Coiling at 600 º for 1 h
No prior deformation
Alloy #6
700
700
600
600
500
500
S tre s s, M P a
S tre s s, M P a
Sample group:A113
Sample group: A113
sample
Alloy
#6#A113F-1A
Alloy #6
Cooling rate: 1 篊/s (980 to-1
Cooling
rate: 1 ºCs
600篊)
Coiling temperature: 600 篊 for 1 h
400
300
200
Sample group: A113
sample#6
#A113F-5A
Alloy
Alloy #6
Cooling rate: 5 篊/s (980 to 600篊)
300
-1
Cooling
rate: 5600ºCs
Coiling temperature:
篊 for 1 h
200
( 980 to 600 ºC)
100
Sample group:A113
400
(980 to 600 ºC)
100
Coiling at 600 ºC for 1 h
Coiling at 600 ºC for 1 h
0
0
0
5
10
15
20
25
30
35
0
40
5
10
15
700
25
30
35
700
600
600
Sample group:A113
500
S tre s s, M P a
Alloy #6Sample group: A113
S trss, M Pa
sample #A113F-10A
Alloy #6
400
-1
CoolingCooling
rate:rate:
1010ºCs
篊/s (980 to
300
(
600篊)
Coiling
600 篊 for 1 h
980
totemperature:
600 ºC)
500
Sample group:A113
400
Alloy Sample
#6 group: A113
300
-1
Cooling
rate:
20篊/sºCs
Cooling
rate: 20
(980 to 600篊)
sample #A113F-20A
Alloy #6
Coiling temperature: 600 篊 for 1 h
( 980 to 600 ºC)
200
200
Coiling at 600 ºC for 1 h
Coiling at 600 ºC for 1 h
100
100
0
0
0
5
10
15
20
25
30
0
35
5
10
15
20
700
700
600
600
500
Sample group:A113
400
Alloy #6
Sample group: A113
300
sample #A113F-40A
#6 40 ºCs-1
CoolingAlloy
rate:
Cooling rate: 40 篊/s (980 to 600篊)
Coiling temperature: 600 篊 for 1 h
25
30
35
40
30
35
45
elongation, %
elongation, %
Sample group:A113
500
Stress, M Pa
Stress, M Pa
20
elongation, %
elongation, %
400
Sample group: A113
Alloy #6
sample #A113F-60A
300
-1
Cooling
rate:
ºCs(980
Cooling
rate:60
60 篊/s
to 600篊)
( 980 to 600 ºC)
Alloy #6
Coiling temperature: 600 篊 for 1 h
200
( 980 to 600 ºC)
200
100
Coiling at 600 ºC for 1 h
100
0
Coiling at 600 ºC for 1 h
0
0
5
10
15
20
25
30
35
40
45
elongation, %
- 212 -
0
5
10
15
20
elongation, %
25
40
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Appendix
Appendix E
Curves of tensile tests
Conditions on the Gleeble:
Reheating temperature: 1225 ºC
Coiling at 575 º for 1 h
No prior deformation
Alloy #6
700
600
600
500
400
AlloySample
#6 group: B113
sample #B113F-1D
Alloy #6
-1
Cooling
1 ºCs
Coolingrate:
rate: 1 篊/s
(980 to 575篊)
Coiling temperature: 575 篊 for 1 h
300
200
Sample group:B113
S tre s s , M P a
S tre s s, M P a
500
Sample group:B113
400
Alloy
Sample#6
group: B113
sample #B113F-5D
300
Alloy #6 rate: 5 ºCs-1
Cooling
Cooling rate: 5 篊/s (980 to 575篊)
Coiling temperature: 575 篊 for 1 h
200
( 980 to 575 ºC)
100
( 980 to 575 ºC)
100
Coiling at 575 ºC for 1 h
Coiling at 575 ºC for 1 h
0
0
0
5
10
15
20
25
30
35
40
0
45
5
10
15
700
7
600
6
Sample #6
group: B113
Alloy
sample #B113F-10D
Alloy #6
-1
Cooling
Cooling rate:rate:
10 篊/s 10
(980 ºCs
to 575篊)
Coiling temperature: 575 篊 for 1 h
300
200
4
Alloysample
#6#B113F-20D
3
Coolingrate:
rate: 20 篊/s
Cooling
20 (980
ºCsto -1575篊)
35
Alloy #6
Coiling temperature: 575 篊 for 1 h
( 980 to 575 ºC)
2
( 980 to 575 ºC)
100
30
Sample group: B113
S tre s s , M P a
400
25
Sample group:B113
5
Sample group:B113
Coiling at 575 ºC for 1 h
1
Coiling at 575 ºC for 1 h
0
0
0
5
10
15
20
25
30
35
40
0
5
10
15
elongation, %
700
700
600
600
500
Sample group:B113
400
Alloy #6
200
Sample group: B113
sample #B113F-40D
-1
Cooling
Alloy #6 rate: 40 ºCs
Cooling rate: 40 篊/s (980 to 575篊)
Coiling(temperature:
for 1 hºC)
980 to575 篊575
100
Coiling at 575 ºC for 1 h
300
20
25
30
35
40
elongation, %
S tre s s, M P a
S tre s s, M P a
S tre s s, M P a
500
20
elongation, %
elongation, %
500
Sample group:B113
400
Alloy
#6group: B113
Sample
300
Alloy #6 rate: 60 ºCs-1
Cooling
sample #B113F-60D
Cooling rate: 60 篊/s (980 to 575篊)
Coiling temperature: 575 篊 for 1 h
( 980 to 575 ºC)
200
Coiling at 575 ºC for 1 h
100
0
0
0
5
10
15
20
25
30
35
40
elongation, %
0
5
10
15
20
elongation, %
- 213 -
25
30
35
40
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Appendix
Appendix F
Curves of tensile tests (Instrumented Hounsfield)
Conditions on the Gleeble:
Reheating temperature: 1225 ºC
45% and 33% prior deformation in total and below the Tnr, respectively
cooling
from 860 down to 575 ºC
Coiling at 575 º for 1 h
Alloy #6
cooling rate : 1 ºCs-1
cooling rate : 1 ºCs-1
cooling rate : 5 ºCs-1
cooling rate : 5 ºCs-1
- 214 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Appendix
cooling rate : 10 ºCs-1
cooling rate : 10 ºCs-1
cooling rate : 19 ºCs-1
cooling rate : 19 ºCs-1
- 215 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Appendix
cooling rate : 34 ºCs-1
cooling rate : 34 ºCs-1
- 216 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Appendix
Appendix G
Curves of tensile tests
Conditions : all experimental alloys #1 to #5
1. Alloy #1:
1.1 Longitudinal specimen:
600
700
600
500
500
S tre ss, M P a
S tre ss , M P a
400
Sample:A1 (Longitudinal)
Alloy: #1
300
400
Sample:A2 (Longitudinal)
Alloy: #1
300
200
200
100
100
0
0
0
5
10
15
20
25
30
35
40
45
0
50
5
10
15
20
600
600
500
500
S tre ss , M P a
S tre ss , M P a
30
35
40
45
400
400
Sample:A3 (Longitudinal)
Alloy: #1
300
Sample:A4 (Longitudinal)
Alloy: #1
300
200
200
100
100
0
0
0
5
10
15
20
25
30
35
40
45
600
500
400
Sample:A5 (Longitudinal)
Alloy: #1
300
200
100
0
0
5
10
15
20
25
0
5
10
15
20
elongation, %
elongation, %
S tre s s , M P a
25
elongation, %
elongation, %
30
35
40
45
elongation, %
- 217 -
25
30
35
40
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Appendix
1.2 Transverse specimen:
500
500
400
400
S tr e s s , M P a
600
S tre s s , M P a
600
Sample:B1 (Transverse)
Alloy: #1
300
Sample:B2 (Transverse)
Alloy: #1
300
200
200
100
100
0
0
0
5
10
15
20
25
30
35
40
45
0
5
10
15
20
elongation, %
600
600
500
500
30
35
40
25
30
35
40
400
S tre s s , M P a
400
Stress, MPa
25
elongation, %
Sample:B3 (Transverse)
Alloy: #1
Sample:B4 (Transverse)
Alloy: #1
300
300
200
200
100
100
0
0
0
5
10
15
20
25
30
35
40
45
elongation, %
0
5
10
15
20
elongation, %
- 218 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Appendix
2. Alloy #2:
2.1 Longitudinal specimen
600
600
500
500
S tre ss , M P a
700
S tre ss , M P a
700
400
Sample:E1 (Longitudinal)
Alloy: #2
300
Sample:E2 (Longitudinal)
Alloy: #2
400
300
200
200
100
100
0
0
0
5
10
15
20
25
30
35
0
40
5
10
15
600
600
500
500
S tr e s s , M P a
700
S tre s s , M P a
700
400
25
30
35
40
25
30
35
40
400
Sample:E3 (Longitudinal)
Alloy: #2
300
300
200
200
100
100
Sample:E4 (Longitudinal)
Alloy: #2
0
0
0
5
10
15
20
25
30
35
40
45
elongation, %
600
500
400
S tre ss , M P a
20
elongation, %
elongation, %
300
Sample:E5 (Longitudinal)
Alloy: #2
200
100
0
0
5
10
15
20
25
30
35
40
elongation, %
- 219 -
0
5
10
15
20
elongation, %
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Appendix
700
700
600
600
500
500
400
S tre ss, M P a
S tre ss , M P a
2.2 Transverse specimen:
Sample:F1 (Transverse)
Alloy: #2
300
400
Sample:F2 (Transverse)
Alloy: #2
300
200
200
100
100
0
0
0
5
10
15
20
25
30
35
0
5
10
15
elongation, %
600
600
500
500
S tre s s , M P a
700
S tre s s , M P a
700
400
25
30
35
40
25
30
35
40
400
Sample:F3 (Transverse)
Alloy: #2
Sample:F4 (Transverse)
Alloy: #2
300
300
200
200
100
100
0
0
0
5
10
15
20
25
30
35
40
elongation, %
600
500
400
Sample:F5 (Transverse)
Alloy: #2
300
200
100
0
0
5
10
15
20
0
5
10
15
20
elongation, %
700
S tre ss , M P a
20
elongation, %
25
30
35
40
elongation, %
- 220 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Appendix
3. Alloy #3:
3.1 Longitudinal specimen:
600
600
500
500
400
S tre ss , M P a
S tre s s , M P a
400
Sample:J1 (Longitudinal)
Alloy: #3
300
Sample:J2 (Longitudinal)
Alloy: #3
300
200
200
100
100
0
0
0
5
10
15
20
25
30
35
40
0
45
5
10
15
20
25
30
35
40
45
30
35
40
45
elongation, %
elongation, %
600
700
600
500
500
S tre ss , M P a
S tre s s , M P a
400
Sample:J3 (Longitudinal)
Alloy: #3
300
400
Sample:J4 (Longitudinal)
Alloy: #3
300
200
200
100
100
0
0
0
5
10
15
20
25
30
35
40
45
700
600
S tre ss, M P a
500
400
Sample:J5 (Longitudinal)
Alloy: #3
300
200
100
0
0
5
10
15
20
25
0
5
10
15
20
25
elongation, %
elongation, %
30
35
40
45
elongation, %
- 221 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Appendix
3.2 Transverse specimen:
600
700
500
600
500
S tr e s s , M P a
S tre ss , M P a
400
Sample:K1 (Transverse)
Alloy: #3
400
300
Sample:K2 (Transverse)
Alloy: #3
300
200
200
100
100
0
0
5
10
15
20
25
30
35
0
40
0
elongation, %
600
600
500
500
15
20
25
30
35
40
25
30
35
40
400
S tre ss, M P a
S tre ss, M P a
10
elongation, %
400
Sample:K3 (Transverse)
Alloy: #3
300
Sample:K4 (Transverse)
Alloy: #3
300
200
200
100
100
0
0
0
5
10
15
20
25
30
35
elongation, %
500
400
Sample:K5 (Transverse)
Alloy: #3
300
200
100
0
0
5
10
15
20
0
5
10
15
20
elongation, %
600
S tre s s, M P a
5
25
30
35
elongation, %
- 222 -
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Appendix
4. Alloy #4
4.1 Longitudinal specimen:
600
600
500
500
400
400
Sample:P2 (Longitudinal)
Alloy: #4
S tre s s , M P a
S tre s s , M P a
Sample:P1 (Longitudinal)
Alloy: #4
300
300
200
200
100
100
0
0
0
5
10
15
20
25
30
35
40
0
45
5
10
15
20
25
30
35
40
30
35
40
elongation, %
elongation, %
600
700
600
500
500
Sample:P4 (Longitudinal)
Alloy: #4
S tre s s , M P a
S tre s s , M P a
400
Sample:P3 (Longitudinal)
Alloy: #4
300
400
300
200
200
100
100
0
0
0
5
10
15
20
25
30
35
40
elongation, %
600
S tress, M P a
500
Sample:P5 (Longitudinal)
Alloy: #4
300
200
100
0
0
5
10
15
20
5
10
15
20
elongation,%
700
400
0
25
30
35
40
elongation, %
- 223 -
25
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Appendix
4.2 Transverse specimen:
700
600
600
500
500
400
S tre s s , M P a
S tr e s s , M P a
Sample:R1 (Transverse)
Alloy: #4
Sample:R2 (Transverse)
Alloy: #4
400
300
300
200
200
100
100
0
0
0
5
10
15
20
25
30
0
35
5
10
15
700
700
600
600
500
25
30
35
40
500
S tr e s s , M P a
S tre s s , M P a
Sample:R3 (Transverse)
Alloy: #4
400
Sample:R4 (Transverse)
Alloy: #4
400
300
300
200
200
100
100
0
0
0
5
10
15
20
25
30
35
700
600
500
400
Sample:R5 (Transverse)
Alloy: #4
300
200
100
0
0
5
10
15
20
0
5
10
15
20
elongation, %
elongation, %
S tre s s , M P a
20
elongation, %
elongation, %
25
30
35
40
elongation, %
- 224 -
25
30
35
40
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Appendix
5. Alloy #5
5.1 Longitudinal specimen:
600
500
500
400
400
Sample:U1 (Longitudinal)
Alloy: #5
S tre s s , M P a
Sample:U2 (Longitudinal)
Alloy: #5
S tre s s , M P a
600
300
300
200
200
100
100
0
0
0
5
10
15
20
25
30
35
0
40
5
10
15
20
25
30
35
40
45
35
40
elongation, %
elongation, %
700
700
600
600
500
500
400
Sample:U4 (Longitudinal)
Alloy: #5
S tre s s , M P a
S tre s s , M P a
Sample:U3 (Longitudinal)
Alloy: #5
400
300
300
200
200
100
100
0
0
0
5
10
15
20
25
30
35
40
0
5
10
15
20
elongation, %
elongation, %
- 225 -
25
30
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Appendix
700
700
600
600
500
500
S tre ss , M P a
S tre ss , M P a
5.2 Transverse specimen:
Sample:V1 (Transverse)
Alloy: #5
400
300
Sample:V2 (Transverse)
Alloy: #5
400
300
200
200
100
100
0
0
0
5
10
15
20
25
0
30
5
10
15
700
700
600
600
25
30
35
500
S tre ss , M P a
500
S tre ss , M P a
20
elongation, %
elongation, %
Sample:V3 (Transverse)
Alloy: #5
400
300
Sample:V4 (Transverse)
Alloy: #5
400
300
200
200
100
100
0
0
0
5
10
15
20
25
30
35
0
5
10
15
elongation, %
elongation, %
- 226 -
20
25
30
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Appendix
Appendix H
Parameters of the laboratory hot rolling process for tests samples on Gleeble:
Tables H-1 to H-4 show the hot rolling details for the test samples of type A and B for
the YS/UTS ratio versus cooling rates, coiling temperatures and deformation values
after compression deformation. These coiling temperatures were inadvertently too
high. These particular plates were, therefore, not used for the YS/UTS tests directly,
but were later used for studying the effect of cooling rate, coiling temperature and
prior deformation on the YS/UTS ratio on the Gleeble (see section 6.9.2).
Table H-1 The laboratory hot rolling parameters for sample #A124 of the Mo-free
reference alloy #6
Pass No
Temperature (ºC)
R1
in 1132 1075
out
Pass ε
1225
5 min
R4
--
18
32
in
45
37
28
20
out
37
28
20
0.196 0.28
R5
F1 reheating
1157 1112 1058 910
26
tip (s)
Gauge
(mm)
R2 reheating R3
58
--
930
5 min
F2
870
880
860
--
13.6 10.3
8.3
6.9
13.6 10.3 8.3
6.9
6
0.34 0.38 0.28 0.22
0.18
0.14
1.48
0.54
Reduction(%)
77
42
1.82 2.43
930
5 min
--
Total ε
ε& (s-1)
reheating F3
3.15 3.92 4.00 4.07
Reheating: 1220 ºC for 60 min
Fast cooling: 35 ºCs-1 (860 down to 650 ºC)
Coiling: 650 ºC for 24 hrs
- 227 -
4.00
3.89
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Appendix
Table H-2 The laboratory hot rolling parameters for sample #AF3F of alloy #3 (with
0.09% Mo)
Pass No
Temperature (ºC)
R1
R2 reheating R3
in 1170 1060
out
R4
R5
F1 reheating
1120 1070 1000 910
1225
5 min
930
5 min
F2
reheating F3
890
880
930
5 min
--
tip (s)
Gauge
(mm)
in
43
37
28
out
37
28
20
0.15 0.28
Pass ε
20
13.6 10.3
8.3
6.9
13.6 10.3 8.3
6.9
6
0.34 0.38 0.28 0.22
0.18
0.14
Total ε
1.43
0.54
Reduction(%)
76
42
ε& (s-1)
1.67 2.43
3.15 3.92 4.00 4.07
4.00
3.89
Reheating: 1225 ºC for 60 min
water quench down to room temperature
Table H-3 The laboratory hot rolling parameters for sample #A113 and B113 of the
Mo-free alloy #6
Pass No
Temperature (ºC)
R1
Pass ε
R3
Reheat
-ing
in 1141 1096 1013
1200
out
5 min
R4
R5
R6
1119 1049 1002
Reheat
-ing
925
5 min
F2
F2
20
--
15
15
--
10
--
in
45
37
30
25
20
15
10
8.3
out
37
30
25
20
15
10
8.3
6.9
0.196 0.21 0.182
0.223 0.288 0.405
0.186 0.18
Total ε
1.50
0.51
Reduction(%)
78
40
ε& (s-1)
1.82
2.08
2.09
2.56 3.31
- 228 -
4.66
Rehea
t-ing F3
880 840 925 870
5 min 840
10
tip (s)
Gauge
(mm)
R2
3.05 4.00
-6.9
6
0.14
3.89
University of Pretoria etd – Tang, Z (2007)
University of Pretoria – Z Tang (2006)
Appendix
Reheating: 1178 ºC for 60 min
Fast cooling: 29 ºCs-1 (840 down to 640 ºC)
Coiling: 640 ºC for 24 hrs
Table H-4 The laboratory hot rolling parameters for sample #TEN06 of the Mo-free
alloy #6
Pass No
R1
R2
Tempera- in 1135 1026
ture (ºC) out
Pass ε
968
-ing
1200
R4
R5
R6
1040
980
940
5 min
Reheat
-ing
1200
30 s
F2
F2
--
15
13
--
14
--
in
45
37
30
25
20
15
10
8.3
out
37
30
25
20
15
10
8.3
6.9
0.196 0.21 0.182
Reduction(%)
1.82
2.08
0.223 0.288 0.405
0.186 0.18
1.50
0.51
78
40
2.09
2.56 3.31
Reheating: 1178 ºC for 60 min
Fast cooling: 33 ºCs-1 (820 down to 620 ºC)
Coiling: 620 ºC for 24 hrs
- 229 -
4.66
Rehea
t-ing F3
870 820 1200 840
60 s 820
18
Total ε
ε& (s-1)
Reheat
15
tip (s)
Gauge
(mm)
R3
3.05 4.00
-6.9
6
0.14
3.89
* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project
advertisement
© 2021