University of Southern California Dissertations and Theses

Evaluation of the superplastic potential in commercial aluminum alloys through equal -channel angular pressing

(USC Thesis Other)

Evaluation of the superplastic potential in commercial aluminum alloys through equal -channel angular pressing

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI fiims
the text directly from the original or copy submitted. Thus, som e thesis and
dissertation copies are in typewriter face, while others may be from any type of
computer printer.
The quality of this reproduction is dependent upon the quality of the
copy submitted. Broken or indistinct print, colored or poor quality illustrations
and photographs, print bleedthrough, substandard margins, and improper
alignment can adversely affect reproduction..
In the unlikely event that the author did not send UMI a complete m anuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by
sectioning the original, beginning at the upper left-hand com er and continuing
from left to right in equal sections with small overlaps.
Photographs included in the original manuscript have been reproduced
xerographically in this copy. Higher quality 6” x 9" black and white
photographic prints are available for any photographs or illustrations appearing
in this copy for an additional charge. Contact UMI directly to order.
ProQ uest Information and Learning
300 North Zeeb Road, Ann Arbor, Ml 48106-1346 USA
800-521-0600
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
EVALUATION OF THE SUPERPLASTIC POTENTIAL IN
COMMERCIAL ALUMINUM ALLOYS THROUGH EQUAL-
CHANNEL ANGULAR PRESSING
© 2000
by
SUNGWONLEE
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(Materials Science)
May 2000
Sungwon Lee
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UM! Number: 3018014
___ ®
UMI
UMI Microform 3018014
Copyright 2001 by Bell & Howell Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
Bell & Howell Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UNIVERSITY OF SOUTHERN CALIFORNIA
THE GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANGELES, CALIFORNIA 90007
This dissertation, written by
Sungwon Lee
under the direction of h . . . .... Dissertation
Committee, and approved by all its members,
has been presented to and accepted by The
Graduate School, in partial fulfillment of re
quirement^ for the degree of
DOCTOR OF PHILOSOPHY
Dean of Graduate Studies
April 7, 2000
DISSERTATION CO]
Chairperson
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ACKNOWLEDGEMENTS
I would like especially to thank my advisor, Professor Terence G. Langdon, for
his support and guidance throughout the program. He has accepted me into his research
group, guided me constantly during my research and helped me to finish my study. His
encouragement and advice made my achievement today possible and will always be
appreciated and remembered.
I would like to thank my thesis committee for taking the time to review and
evaluate my work. The committee members include Professor Edward Goo and Professor
Charles G. Sammis. In addition, I would like to thank Professor Steve R. Nutt for
granting me access to some of his equipment.
I would like to thank Professor Minoru Nemoto and Professor Zenji Horita at the
Department of Materials Science in Kyushu University, Fukuoka, Japan and Professor
Minoru Furukawa at the Department of Technology in Fukuoka University of Education,
Munakawa, Japan, for inviting me as a Visiting Fellow, for providing the equipment and
materials, and for supporting my research with the transmission electron microscope
work. Their guidance was an essential part in this program.
I would like to thank Professor Ruslan Z. Valiev at the Institute of Physics of
Advanced Materials in Ufa State Aviation Technical University, Ufa, Russia for
providing materials and for very helpful discussions throughout the program. I would
also like to thank Professor Terry R. McNelley at the Department o f Mechanical
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Engineering in Naval Postgraduate School, Monterey, California for providing the SEM
equipment.
I would like to thank Kan Lee and John Ushida for their work in the USC
machine shop on the ECA pressing facilities and the specimens.
I would like to thank Dr. Yong Li and Dr. Bing Q. Han for their support during
my research. I would like to thank my seniors, Dr. Patrick B. Berbon and Dr. Min Z.
Berbon for their guidance and devotion. I would also like to acknowledge my fellow
graduate students in our group, Jeff Wang, Siari Sosa, Sangmok Lee, Cheng Xu and
Hyunchul Kim. It was a pleasure to work with these fine young students.
I express my gratefulness to Korean graduate students and researchers of the
Materials Science Department, Chanman Park, Dr. Samsoo Kim, Zonghoon Lee,
Jaeyoung Kim, Daegye Chang, Euitae Kim, Yongwon Song, Byungsang Choi and first
year students.
Finally, I would like to thank my parents in Korea for their constant support and
confidence throughout my study . I wish to thank my wife, Youngmi, and my son, Alex,
for their endless devotion, patience and support. Their love and sacrifice made my
achievement today possible.
This work was supported in part by the National Science Foundation of the
United States under Grants No. DMR-9626969 and EMT-9602919 and in part by the US
Army Research Office under Grant No. DAAH04-96-1-0322.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS ii
LIST OF FIGURES vii
LIST OF TABLES xii
GLOSSARY xiii
ABSTRACT xv
1. INTRODUCTION 1
2. LITERATURE REVIEW 4
2.1 CONVENTIONAL SUPERPLASTICITY 4
2.1.1 Definition 4
2.1.2 High temperature deformation mechanisms 4
2.1.3 The relationship between stress and strain rate 6
2.1.4 Requirements for superplasticity 7
2.1.5 Physical mechanisms for superplastic flow 10
2.2 HIGH STRAIN RATE SUPERPLASTICITY 12
2.2.1 Introduction 12
2.2.2 High strain rate superplasticity in aluminum matrix composites 13
2.2.3 Origin of high strain rate superplasticity in aluminum matrix
composites 15
2.2.4 High strain rate superplasticity in aluminum alloys 20
2.2.5 Origin of high strain rate superplasticity in aluminum alloys 22
2.3 PRINCIPLES OF EQUAL-CHANNEL ANGULAR PRESSING 27
2.3.1 Introduction 27
2.3.2 Fundamental concept of EC A pressing 28
2.3.3 Deformation behavior during EC A pressing 31
2.4 FACTORS ASSOCIATED WITH EQUAL-CHANNEL ANGULAR
PRESSING 32
2.4.1 The angles between the two channels (O and TQ 32
2.4.2 The number o f EC A pressings 34
2.4.3 The pressing routes 35
2.4.4 The pressing speed 37
2.4.5 The pressing temperature 39
3. EXPERIMENTAL PROCEDURES 41
3.1 MATERIALS 41
iv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.2 EXPERIMENTAL PROCEDURES 43
3.3 EQUAL-CHANNEL ANGULAR PRESSING 45
3.3.1 Die and plunger 45
3.3.2 Specimen preparation for EC A pressing 47
3.3.3 Procedure for EC A pressing 49
3.4 MECHANICAL TESTING 51
3.4.1 Tensile testing 51
3.4.2 Microhardness testing 54
3.5 MICROSTRUCTURAL OBSERVATION 55
3.5.1 Optical microscopy 55
3.5.2 Transmission electron microscopy 56
3.5.3 Scanning ion microscopy 58
4. FUNDAMENTAL FACTORS IN ECA PRESSING 59
4.1 STRUCTURE OF ECA PRESSED BILLETS 59
4.1.1 Introduction 59
4.1.2 Test material and experimental procedures 59
4.1.3 Variation of microhardness in the pressed billets 60
4.1.4 Variation of microhardness with the number of pressings 65
4.2 SHEARING CHARACTERISTICS ASSOCIATED WITH ECA
PRESSING 67
4.2.1 Introduction 67
4.2.2 The pressing routes A, Ba, Be and C 68
4.2.3 New pressing routes C-Bc and Bc-C 71
4.2.4 Effects of pressing routes for pure aluminum 76
4.2.5 Effects of pressing routes for Al-2024 alloy 80
5. SUPERPLASTICITY IN ALUMINUM ALLOYS 83
5.1 Al-Sc-Zr AND Al-Mg-Sc-Zr ALLOYS 83
5.1.1 Introduction 83
5.1.2 Materials and experimental procedures 83
5.1.3 Results on Al-0.2% Sc-0.12% Zr alloy 84
5.1.4 Results on Al-3% Mg-0.2% Sc-0.12% Zr alloy 84
5.1.5 Discussion 94
5.1.5.1 The effect of addition of Sc and Zr on pure aluminum 94
5.1.5.2 The effect of addition of Sc and Zr on the Al-Mg binary
alloy 98
5.1.5.3 Deformation mechanism for Al-Sc-Zr alloy 101
5.1.5.4 Deformation mechanism for Al-Mg-Sc-Zr alloy 105
5.2 Al-2024 ALLOY 110
5.2.1 Introduction 110
5.2.2 Materials and experimental procedures 11 1
5.2.3 Results of tensile testing at room temperature 112
5.2.4 Results of tensile testing at high temperatures 116
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5.2.5 Results of annealing at high temperatures 127
5.2.6 Discussion 133
5.2.6.1 Superplasticity in Al-2024 alloy 133
5.2.6.2 The effect of pressing routes on superplasticity 136
5.2.6.3 The effect o f pressing temperatures on superplasticity 138
5.3 SUPRAL 100 ALLOY (Al-2004) 142
5.3.1 Introduction 142
5.3.2 Materials and experimental procedures 143
5.3.3 Microstructures before and after ECA pressing 144
5.3.4 Results of tensile testing at high temperatures 150
5.3.5 Micro structures after tensile testing 156
5.3.6 Discussion 158
5.3.6.1 The effect of ECA pressing on optimum superplastic
conditions 158
5.3.6.2 The effect of ECA pressing on superplasticity of Al-
2004 alloy 160
5.4 Al-1420 ALLOY 162
5.4.1 Introduction 162
5.4.2 Materials and experimental procedures 163
5.4.3 Microstructures after ECA pressing 163
5.4.4 Results of tensile testing at high temperatures 165
5.4.5 Micro structures after tensile testing 172
5.4.6 Discussion 175
5.4.6.1 Superplasticity in Al-1420 alloy 175
5.4.6.2 HSR SP through ECA pressing 183
6. SUMMARY AND CONCLUSIONS 185
REFERENCES 188
vi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
LIST OF FIGURES
Figure 2.1
Figure 2.2
Figure 2.3
Figure 2.4
Figure 3.1
Figure 3.2
Figure 3.3
Figure 3.4
Figure 3.5
Figure 4 .1
Figure 4.2
Figure 4.3
Figure 4.4
Figure 4.5
Figure 4.6
Schematic illustration of elongation to failure (upper) and flow
stress (below) vs. strain rate for typical superplastic metals.
Schematic illustration o f the effect of grain refinement on strain rate
[84].
Schematic illustration of equal-channel angular pressing.
Schematic illustration o f the four pressing routes.
Flow chart o f the experimental procedures.
Dimensions o f the die for ECA pressing at room temperature.
Dimensions o f (a) plunger and (b) billet for ECA pressing at room
temperature.
Schematic illustration o f the high temperature pressing facility.
Dimensions o f tensile test specimen.
(a) Schematic illustration of the three planes designated X, Y and Z
[108].
(b) Schematic illustration of the sectioning and the location of
microhardness indentations.
Variation of microhardness with the number o f plane for pure
aluminum ECA pressed at 298 K via route Be.
Schematic illustration o f shearing up to four pressings through the
die via route C [107,128],
Schematic illustration of shearing up to four pressings through the
die via route Be [107],
Pictorial compilation of the shearing associated with four different
processing routes.
Schematic illustration o f shearing up to four pressings through the
die via route C-Bc.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 4.7
Figure 4.8
Figure 4.9
Figure 5.1
Figure 5.2
Figure 5.3
Figure 5.4
Figure 5.5
Figure 5.6
Figure 5.7
Figure 5.8
Figure 5.9
Figure 5.10
Schematic illustration o f shearing up to eight pressings through the
die via route Bc-C.
Variation of microhardness with the number o f pressings for pure
aluminum ECA pressed at 298 K via six different routes.
Variation of microhardness with the number o f pressings for Al-
2024 alloy ECA pressed at 298 K via three different routes.
The stress vs. strain curves of Al-0.2% Sc-0.12% Zr alloy ECA
pressed at 298 K via route Be up to 8 passes.
Micro structures o f AI-3% Mg-0.2% Sc-0.12% Zr alloy after ECA
pressing at (a) low magnification and (b) high magnification.
The stress vs. strain curve o f Al-3% Mg-0.2% Sc-0.12% Zr alloy
ECA pressed at 298 K via route Be up to 6 passes.
Examples of tensile ductility in samples ECA pressed at 298 K via
route Be up to 6 passes and pulled at 773 K.
Microstructure o f a sample prepared by ECA pressing and then
tested to failure in tension at 773 K with a strain rate o f I O'2 s'1 :
(a) grip region; (b) the fracture tip within the gauge length.
Micro structure of a sample prepared by ECA pressing and then
tested to failure in tension at 773 K with a strain rate of 10-4 s'1 :
(a) grip region; (b) the fracture tip within the gauge length.
Variations of grain size and 0.2% proof stress with annealing
temperature for pure aluminum, Al-0.2% Zr alloy and Al-0.2% Sc
alloy.
Elongation to failure vs. strain rate for pure aluminum, Al-0.2% Sc
alloy and Al-0.2% Sc-0.12% Zr alloy.
Elongation to failure vs. strain rate for a number of Al-Mg-Sc alloys
and Al-3% Mg-0.2% Sc-0.12% Zr alloy.
The logarithm of flow stress as a function of the logarithm o f strain
rate for Al-0.2% Sc-0.12% Zr alloy.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.11
Figure 5.12
Figure 5.13
Figure 5.14
Figure 5.15
Figure 5.16
Figure 5.17
Figure 5.18
Figure 5.19
Figure 5.20
Figure 5.21
Figure 5.22
Figure 5.23
Figure 5.24
The logarithm o f temperature-compensated stress as a function of
the reciprocal o f absolute temperature for Al-0.2% Sc-0.12% Zr
alloy.
The logarithm of flow stress as a function of the logarithm of strain
rate for Al-3% Mg-0.2% Sc-0.12% Zr alloy.
The logarithm o f temperature-compensated stress as a function of
the reciprocal of absolute temperature for Al-3% Mg-0.2% Sc-
0.12% Zr alloy.
The stress vs. strain curves o f Al-2024 alloys tested at 298 K.
The stress vs. strain curves o f the ECA pressed Al-2024 alloys when
testing at 298 K.
The stress vs. strain curves o f Al-2024 alloys tested at 673 K and at
10‘ 3 s'1 .
Examples of tensile ductility in Al-2024 alloys ECA pressed at 298
K via route Be up to 8 passes and pulled at 673 K and 10*3 s'1 .
Examples of tensile ductility in Al-2024 alloys ECA pressed at 373
K via route Be up to 8 passes and pulled at 673 K.
Variations of microhardness with annealing temperature for
unpressed and ECA pressed AJ-2024 alloys.
Microstructures of Al-2024 ECA pressed via route Be up to 8 passes
at (a) 298 K and (b) 373 K.
Elongation to failure vs. strain rate for a number of Al-Cu-Mg
alloys.
Elongation to failure vs. strain rate under optimum conditions for
Al-2024 alloy.
The logarithm of flow stress as a function of the logarithm of strain
rate for Al-2024 alloys ECA pressed at 298 K and 373 K.
Microstructure and associated SAED pattern of as-received Al-2004
alloy
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.25
Figure 5.26
Figure 5.27
Figure 5.28
Figure 5.29
Figure 5.30
Figure 5.31
Figure 5.32
Figure 5.33
Figure 5.34
Figure 5.35
Figure 5.36
Figure 5.37
Microstructures of Al-2004 alloy after pressing (a) at 298 K up to 8
passes, (b) at 373 K up to 8 passes and (c) at 473 K up to 12 passes.
Microstructures of Al-2004 alloy after pressing (a) at 573 K up to 12
passes and (b) at 673 K up to 6 passes.
Variations of grain size and microhardness as a function of ECA
pressing temperature for Al-2004 alloy.
Examples of tensile ductility in samples ECA pressed at 573 K via
route Be up to 12 passes and pulled at 673 K.
Examples o f tensile ductility in samples ECA pressed at 673 K via
route Be up to 6 passes and pulled at 723 K.
The stress vs. strain curves of Al-2004 alloys tested at 623 K and
673 K.
Microstructure of Al-2004 alloy prepared by ECA pressing and then
tested to failure in tension at 673 K with a strain rate of 10'2 s'1 :
(a) grip region; (b) the fracture tip within the gauge length.
Elongation to failure vs. strain rate under optimum conditions for
Al-2004 alloy.
Microstructure and associated SAED pattern of Al-1420 alloy after
ECA pressing to a strain o f— 12.
Micro structure at a high magnification of Al-1420 alloy after ECA
pressing to a strain o f— 12:
(a) bright field image and (b) (100) dark field image.
Examples of tensile ductility in samples ECA pressed via route Be
up to 8 passes at 673 K and an additional 4 passes and pulled at 673
K.
The stress vs. strain curves of Al-1420 alloy tested at 673 K.
Microstructure of Al-1420 alloy prepared by ECA pressing and then
tested to failure in tension at 673 K with a strain rate o f 10'1 s‘l:
(a) grip region; (b) the fracture tip within the gauge length.
146
148
149
153
154
155
157
159
164
166
170
171
173
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.38
Figure 5.39
Figure 5.40
Figure 5.41
Figure 5.42
Figure 5.43
Figure 5.44
Micro structure of Al-1420 alloy prepared by ECA pressing and then
tested to failure in tension at 673 K with a strain rate of 1 s'1 :
(a) grip region; (b) the fracture tip within the gauge length. 174
Elongation to failure vs. strain rate for Al-1420 alloy ECA pressed
via route Be up to 4 passes at 673 K. 176
Elongation to failure vs. strain rate for Al-1420 alloy ECA pressed
via route Be up to 4 passes at 673 K and an additional 2 passes at
473 K. 177
Elongation to failure vs. strain rate for Al-1420 alloy ECA pressed
via route Be up to 4 passes at 673 K and an additional 4 passes at
473 K. 178
Elongation to failure vs. strain rate for Al-1420 alloy ECA pressed
via route Be up to 8 passes at 673 K. 179
Elongation to failure vs. strain rate for Al-1420 alloy ECA pressed
via route Be up to 8 passes at 673 K and an additional 2 passes at
473 K. 180
Elongation to failure vs. strain rate for Al-1420 alloy ECA pressed
via route Be up to 8 passes at 673 K and an additional 4 passes at
473 K. 181
xi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
LIST OF TABLES
Table 2.1
Table 2.2
Table 2.3
Table 4.1
Table 4.2
Table 5.1
Table 5.2
Table 5.3
Table 5.4
Table 5.5
Table 5.6
Table 5.7
Table 5.8
Table 5.9
Table 5.10
High strain rate superplastic behavior in aluminum matrix
composites
Characteristic temperatures o f aluminum alloys and their composites
High strain rate superplastic behavior in aluminum alloys
Microhardness values of pure aluminum ECA pressed at 298 K via
route Be
Hardness test results for pure aluminum ECA pressed using six
different routes
The superplastic properties of Al-Sc-Zr alloy ECA pressed at 298 K
The superplastic properties of Al-Mg-Sc-Zr alloy ECA pressed at
298 K
Tensile test results at room temperature for Al-2024 alloys
Tensile test results at high temperatures for unpressed and ECA
pressed Al-2024 alloys
Tensile test results at high temperatures for T351 heat-treated Al-
2024 alloy
Tensile test results at high temperatures for Al-2024 alloys ECA
pressed at 298 K via route C-Bc
Tensile test results at high temperatures for Al-2024 alloys ECA
pressed at 298 K via route Bc-C
Tensile test results at high temperatures for Al-2024 alloys ECA
pressed at 373 K via route Be
Tensile test results at high temperatures for unpressed and ECA
pressed Al-2004 alloys
Tensile test results at high temperatures for ECA pressed Al-1420
alloys
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
GLOSSARY
Abbreviations
ECA Equal-Channel Angular
EF Elongation to Failure
FEM Finite Element Modeling
FIB Focused Ion Beam
GB Grain Boundary
GBS Grain Boundary Sliding
H-P Hall-Petch Relationship
HCP Hexagonal Closed Packed
HSR SP High Strain Rate Superplasticity
IM Ingot Metallurgy
IS Interfacial Sliding
LEDS Low Energy Dislocation Structure
LTS Low Temperature Superplasticity
MA Mechanically Alloyed
MMC Metal Matrix Composite
ODS Oxide Dispersion Strengthened
PM Powder Metallurgy
PVD Physical Vapor Deposition
RSP Rapid Solidification Processing
SAED Selected Area Electron Diffraction
SIM Scanning Ion Microscopy
SMG Submicrometer-grained
SP Superplasticity
SPD Severe Plastic Deformation
SPF Superplastic Forming
TEM Transmission Electron Microscopy
TMP Thermo-Mechanical Processing
TS Torsion Straining
UFG Ultrafine-grained
UTS Ultimate Tensile Strength
YS Yield Strength
Symbols
A Dimensionless constant or Cross-section of the gauge area
A0 Initial gauge area (unit m )
xiii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
b Burgers vector (unit m)
d Grain size (unit m)
dh Mean value o f the diagonals of the indentation (unit mm)
dp Particulate size (unit m)
D Diffusion coefficient (unit m^s'1 )
D 0
" 7 1
Frequency factor (unit m -s*)
D g b Grain boundary diffusion coefficient (unit m2-s'1 )
A
Interfacial diffusion coefficient (unit m2-s'1 )
D i Lattice diffusion coefficient (unit rn^s'1 )
E Young’s modulus (unit Pa)
F Test load (unit N)
G Shear modulus (unit Pa)
Hv Vickers microhardness
Hv Mean Vickers microhardness
k Boltzmann’s constant (= 1.38X10'2 3 J-K'1 )
k y constant of yielding
K Material constant
L a Initial gauge length (unit m)
L t Total line length over which the measurement is made (unit m)
M Magnification
*fi
Average number of grain boundaries intercepted
m Strain rate sensitivity
n Stress exponent (= m* l)
P
Grain size exponent
P load (unit lb)
P n
Total number of intersections
O Activation energy (unit J-mol'1 )
Qgb
Activation energy for grain boundary diffusion (unit J-mol'1 )
Qi Activation energy for lattice diffusion (unit J-mol'1 )
QP
Activation energy for dislocation pipe diffusion (unit J-mol’1 )
R Gas constant (= 8.31 J-mol^-BC1 )
T Absolute temperature (unit K)
T n , 1 m Melting temperature (unit K)
S Strain
S Strain rate (unit s'1 )
X Subgrain size (unit m)
a Stress (unit MPa)
< y Standard deviation
O b
Friction stress (unit MPa)
O -y Yield stress (unit MPa)
XIV
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ABSTRACT
Since Equal-Channel Angular (ECA) pressing offers significant advantages for
grain refinement and meets the requirements for superplasticity, it has been chosen as a
basic tool of material processing for this study. High temperature ECA pressing facilities
were successfully installed at USC. The effect of ECA pressing on superplasticity of
commercial aluminum alloys was investigated by changing some of the basic parameters
for ECA pressing such as the pressing routes, the number of pressings and the pressing
temperatures.
The ECA pressing was applied to pure aluminum and it is determined that the
variations of microhardness are large throughout the length o f the pressed billet for lower
numbers of passes but a very uniform micro structure is obtained throughout the billet
after higher numbers o f passes.
Two new combination pressing route, C-Bc and Bc-C, were designed, applied to
both pure aluminum and an Al-2024 alloy and compared to route Be. It is concluded that
route Be is the most reliable pressing route because the two combination routes have no
clear advantage in terms of superplasticity over route Be.
The effect of small amounts o f addition of Sc and Zr into pure aluminum and an
Al-Mg binary alloy was investigated. It is determined that the addition of both Sc and Zr
at the same time is more effective in promoting superplasticity than the addition of only
one element.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The ECA pressing was applied to commercial aluminum alloys such as Al-2024,
Al-2004 and Al-1420 alloys. The ECA pressing at high temperatures is necessary to press
strong commercial aluminum alloys to a higher strain successfully. The ECA pressing at
high temperatures always produced superior superplasticity over the ECA pressing at
room temperature. After ECA pressing of the commercial Al-2024 and Al-2004 alloys,
furthermore, the optimum superplastic conditions were shifted to both a lower testing
temperature and a higher imposed strain rate. An exceptional high strain rate
superplasticity (HSR SP) was obtained for every aluminum alloy. For example, an
elongation of -950% was recorded at a temperature of 673 K with a strain rate of 1 s'1
when the Al-1420 alloy was ECA pressed to a strain o f-12.
XVI
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1. INTRODUCTION
Strong, lightweight materials are necessary ingredients for many modem
structural applications such as automobiles, flight, building and construction, and
aluminum alloys have been the predominant choice for those applications for over 50
years. The selection of materials for today’s structural applications depends primarily on
strength-to-weight ratio, resistance to corrosion, damage tolerance and economic
considerations that include producibility, material cost, maintainability and availability.
Historically, aluminum alloys have been leaders in satisfying these requirements and
currently are the dominant materials used for aircrafts and other applications. However,
because of increased performance requirements and the use of complex shaped
components, conventional aluminum forming techniques have been severely restricted in
commercial applications until the late 1960’s. The new forming technique using the
remarkable high plastic ductility characteristic of superplastic metals was developed
commercially over the past thirty years [1,2].
Superplasticity is the capability o f certain polycrystalline materials to undergo
extensive tensile plastic deformation, often without the formation of a neck, prior to
failure. The subject of superplasticity in metal alloys has been reviewed extensively [3-9],
In certain metal alloy systems, tensile elongations of thousands of percent have been
documented. For example, a total elongation of 4850% in a Pb-62wt%Sn alloy [10] was
exceeded in a commercial aluminum bronze (Cu-10wt%Al alloy) by a value reported as
5500% [11], Despite these large elongations, of many thousand percent, most
I
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
superplastic alloys exhibit optimum tensile elongations o f about 300 to 1000%. This
range of values is more than sufficient to make, using superplastic forming technology,
extremely complex shapes, many of which only require 200 to 300% elongation.
However, one o f the major concerns of current superplastic forming technology is
the slow forming rate which is typically — 1 C T 4 to 10~ 3 s’1 . For example, the optimum
superplastic conditions o f the Pb-62wt%Sn alloy are an initial strain rate of 1.33 x I O '4 s'1
and 413 K [10], and the optimum superplastic conditions o f commercial aluminum
bronze are an initial strain rate of 6.3 x 10’3 s'1 and 1073 K [11]. Recent studies have
demonstrated, however, that superplasticity can be found at considerably higher strain
rates than I O ’3 s’1 , i.e. at strain rates of up to 10’2 to 10'1 s*1 . This high strain rate
superplasticity (HSR SP) phenomenon was originally observed in metal matrix
composites [12] and then found in mechanically alloyed materials [13]. The subject of
HSR SP has received much attention in recent years because an increase in the operating
strain rate may lead to a significant reduction in the total time for processing in
superplastic forming and thereby provide the capability for the mass production of a large
number of components.
Recent reports have established that a reduction in the grain size has the potential
of both decreasing the temperature and increasing the strain rate associated with optimum
superplastic flow [14,15]. Among the methods of fabricating ultrafine-grained (UFG)
materials, severe plastic deformation techniques such as torsion straining (TS) and equal-
channel angular (ECA) pressing are probably the most promising because they have
several advantages, including the possibility of economically producing large, fully dense
?
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
samples [16,17]. The ECA pressing was invented by a scientist from the former Soviet
Union, V.M. Segal, and uses repeated shear deformation to refine the grain size in
metallic materials [18]. Furthermore, there is a big possibility that ECA pressing may be
employed for a wide variety o f materials including commercial metals and alloys,
superalloys, intermetallic compounds, ceramics and composites. Considering the
advantages and versatility of the process, therefore, ECA pressing can be considered an
important deforming process of the future.
Since ECA pressing offers significant advantages for grain refinement and meets
the requirements for superplasticity and even for HSR SP, it has been chosen as a basic
tool of material processing for this study. The following materials were used in this work:
pure aluminum, experimental aluminum alloys, commercial cast aluminum alloys and
commercial wrought aluminum alloys. The study on pure aluminum and experimental
aluminum alloys permits a better understanding of the deformation mechanisms involved
during deformation of submicrometer-grained (SMG) materials and permits an evaluation
o f the effect of ECA pressing parameters on the resultant microstructure and properties.
The work on commercial aluminum alloys offers specific data for HSR SP and shows
clear possibilities of using ECA pressing for industrial applications. Therefore, the
ultimate goals of this research are to achieve micro structures suitable for superplastic
deformation conditions in aluminum alloys by adjusting the processing parameters of
ECA pressing and to develop materials showing HSR SP.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2. LITERATURE REVIEW
2.1 CONVENTIONAL SUPERPLASTICITY
2.1.1 Definition
Superplasticity is defined as the ability of a polycrystalline material to exhibit, in
a generally isotropic manner, very high tensile elongations prior to failure [19]. Early
experimental investigations of superplasticity were confined exclusively to metallic
alloys for many years. Since superplasticity basically requires a very small and stable
grain size at a high testing temperature, usually higher than 0.5 Tm , where T ,„ is the
absolute melting point of material, it is necessary to prevent significant grain growth
during the deformation process, and this is achieved either by using metallic alloys
having two-phase eutectic or eutectoid compositions or by introducing a fine dispersion
of a second phase to act as a grain refiner [5].
2.1.2 High temperature deformation mechanisms
Plastic deformation at high temperatures is associated with three discrete
mechanisms that can occur at the atomic level: slip by dislocation movement, grain
boundary sliding (GBS), and directional diffusional flow. All three mechanisms are
generally believed to occur independently of one another, but these are all thermally
activated and are governed by the diffusion of atoms. Thus, superplastic deformation can
be described by the standard equation for high temperature creep [7]:
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
. ADGb
£ —---------
k T
(2.1)
where D is the diffusion coefficient, G is the shear modulus, b is the Burgers vector, k is
Boltzmann’s constant, T is absolute temperature, d is the grain size, cris the flow stress,
and A, p and n are constants. The stress exponent n is the reciprocal of the strain rate
sensitivity m. The diffusion coefficient D can be rewritten as follows:
where D0 is the frequency factor, O is the activation energy and R is the gas constant. For
any selected material and testing conditions, the critical parameters in equation (2.1) are
the values of n (or m), p and O.
mechanism can be defined uniquely. For example, plastic flow by slip is associated with
a low strain rate sensitivity (m < 0.3) and an activation energy which can be related either
to lattice diffusion, Oi or to dislocation pipe diffusion Qp. Plastic deformation by GBS is
characterized by a high strain rata sensitivity (m ~0.5), and an activation energy which is
equal to Oi or to the activation energy for grain boundary diffusion, Ogb - Plastic
deformation by diffusional flow is characterized by a strain rate sensitivity of unity and
an activation energy equal to Oi or to Ogb [1,20].
The term superplasticity is used to refer to the large tensile elongations, typically
more than 400%, achieved in polycrystalline materials under certain conditions o f strain
rate and temperature. These high elongations are directly related to the high strain rate
(2.2)
Each mechanism of plastic flow has specific values of m and O by which the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
sensitivity which results in the inhibition o f the growth o f necks during tensile
deformation. Typical values of m for superplastic materials are 0.5 or greater so that GBS
or diffiisional flow dominates the deformation processes. Since difliisional creep
generally becomes important at very high temperatures (-0.9 Tm ) and very low strain rates
(< IC T 4 s'1 ), this process is of little importance and GBS is the principal mechanism for
superplasticity [1,7,20],
2.1.3 The relationship between stress and strain rate
Most conventional superplastic metals exhibit large tensile strains when pulled at
high temperatures and at strain rates which are typically in the vicinity of about 10'3 to
IC T 4 s'1 . When the strain rate is either significantly decreased or increased, however, the
maximum elongation to failure decreases continuously. In order to understand the
optimal superplasticity, which occurs within this limited strain rate range, the basic
relationship between stress and strain rate should be considered. Therefore, we can
simplify and rewrite equation (2.1) as follows:
a = K s m (2.3)
where K is a material constant which includes the temperature dependence. The value of
the strain rate sensitivity determines the rate at which necking proceeds after the start of
localized plastic flow. It is known from experiment that, as m increases, the elongation to
failure increases [21], The value of m for ideal superplastic materials, such as hot glass
and tar, is equal to 1 and these materials follow the Newtonian-viscous relationship.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
At a given temperature, the steady state flow stresses, a, are then measured as a
function of the imposed strain rates, s , and the data are plotted logarithmically as cr vs.
e , as shown schematically in Fig. 2.1 [5]. The value of m is about 0.5 in regions II, and
m decreases in regions I and III. The maximum elongations occur in region II and there
are significant decreases in the total elongations in both regions I and HI. This is
consistent with the anticipated relation between m and the total ductility.
In general, the activation energy, O is high in region III and close to the
anticipated value for lattice self diffusion, whereas O tends to be low and probably close
to the value for grain boundary diffusion in region II. There is also evidence that the
value of O increases again in region I [9]. In the superplastic region n, therefore, GBS is
the major deformation mechanism and the process of GBS involves the relative
displacements of individual grains without any change in the grain shape.
2.1.4 Requirements for superplasticity
A high value of m is a necessary condition for superplasticity. Although a material
is then highly sensitive to strain rate (m > 0.4), it is found that a high value of m is not a
sufficient condition because other factors can cause premature failure before reaching a
high elongation to failure. Therefore, it is very important to satisfy the various structural
requirements for structural superplasticity in fine-structure superplastic alloys. These
prerequisites are summarized below [1,5,6,9],
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
c
o
C O
O)
e
LU
C O
Q_
m
b
O)
o
log 8 (s1 )
Figure 2.1 Schematic illustration of elongation to failure (upper) and flow stress (below)
vs. strain rate for typical superplastic metals.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1. Fine grain size
Typically, a grain size of less than 10 pm is required for superplastic flow to
occur. Fine grains increase the fraction of grain boundaries inside the material and this
can reduce dislocation activity inside the grains and promote GBS.
2. High testing temperature
Superplasticity takes place at high temperatures, typically o f the order of 0.5 Tm ,
because GBS and its accommodation process are all thermally activated processes.
3. Presence o f second phase
Superplasticity cannot be observed in very fine-grained but single-phase materials
because of the occurrence of rapid grain growth at high temperatures. For this reason,
many conventional superplastic materials are based on either two-phase alloys (e.g. Al-
Ca, Fe-FesC, and Zn-Al) or alloys containing fine particles.
4. Strength o f second phase
Many fine-grained aluminum and copper based alloys are susceptible to cavitation
during superplastic deformation and this is probably due to the large difference in
strength between the matrix and the second hard phase. Since stress concentrations can
occur easily at matrix/second phase interfaces, therefore the strength of the second phase
should generally be of the same order as that of the matrix.
5. Size and distribution o f second phase
In general, large and hard particles cannot be easily refined by conventional
thermo-mechanical processing (TMP) and cavitation can occur easily during
deformation. When fine but hard particles are distributed uniformly, however, cavitation
9
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
during superplastic flow is inhibited by various recovery mechanisms occurring in the
vicinity o f the particles. If the second phase is considerably harder than the matrix, it
should be distributed uniformly and finely.
6. Nature o f the grain boundaries
The grain boundaries between adjacent matrix grains should be high angle (i.e.
disordered) because low angle boundaries do not slide readily under shearing stresses.
7. Grain boundary m obility
Grain boundaries should be mobile. During GBS, stress concentrations can
develop easily at triple points and near other obstacles. If grain boundaries are mobile,
these stress concentrations can be relieved by grain boundary migration.
8. Shape o f grains
Materials having textured grains show limited ductility because GBS does not
occur readily in the longitudinal direction. Even after hundreds and thousands of percent
strain, the grains remain essentially equiaxed for superplastic materials.
9. Resistance to tensile separation
Grain boundaries in the matrix should not be prone to tensile separation. This is
the reason why many ceramics having fine grains and high m value show poor ductility at
high temperatures.
2.1.5 Physical mechanisms for superpiastic flow
The exact mechanism of superpiastic flow has not been fully understood up to
now. However, there is no doubt that the most commonly considered mechanisms for
10
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
superplasticity all involve GBS. Thus, it is essential to consider an accommodation
process for GBS. The proposed models fall into two distinct types according to the
accommodation process: GBS accommodated by diffusional flow and GBS
accommodated by slip [ 1 ],
1. Grain boundary sliding accommodated by diffusional flo w
The model based on diffusional flow, developed by Ashby and Verrall [22],
clearly shows a topological description of the superpiastic process. The model assumes
that the grains change their shapes and rearrange together by diffusion through the lattice.
This type of grain rearrangement is attractive because it retains the equiaxed grains at
high elongations without any grain growth. However, some topological problems in
developing this model, such as diffusion paths, have been pointed out and the model has
failed to predict quantitatively the creep rates actually observed in fine-grained
superpiastic materials [2 0 ],
2 . Grain boundary sliding accommodated by slip
The model based on dislocation movement was initially proposed by Ball and
Hutchison [23] and then a number of models were presented [24-26], In these models,
groups of grains in reasonable alignment slide as units to create a stress concentration on
an obstacle and this generates dislocations that pile up at the next grain boundary. The
rate controlling process is the rate of removal of dislocations from the head of the pile-up
by climb into and along the grain boundary.
Another interesting model based on GBS accommodated by dislocation slip is the
‘core and mantle’ theory proposed by Gifkins [27]. In tills model, GBS accommodated by
11
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
slip occurs in the mantle region (adjacent to the grain boundaries) and slip occurs within
the core o f each grain. When the former process dominates superplasticity can occur, and
when the latter dominates normal ductility is expected.
Recently a unified model for creep and superplasticity was proposed by Langdon
[28]. He considered that GBS is a characteristics of high temperature deformation in both
creep when the grain size is large (d > A) and superplasticity when the grain size is small
(d < A), where d and A are the grain size and the subgrain size, respectively. The model
leads to n = 3, p = 1 and D = D i for large grain sizes ( d > A) and n = 2, p — 2 and D — Dgb
for small grain sizes {d < A) in equation (2.1).
2.2 HIGH STRAIN RATE SUPERPLASTICITY
2.2.1 Introduction
In 1984, it was demonstrated that commercial SiCw /2124Al composite, following
TMP treatments, can exhibit 300% elongation at relatively high strain rates of 3.3 x I O ' 1
s' 1 [12]. This result is important because one of the major concerns of superpiastic
forming technology is the slow forming rate which is typically I O ' 4 to 10' 3 s'1 . After 1984,
many studies have been focused on high strain rate superplasticity (HSR SP) in a number
of materials including alloys and composites.
It should be noted that HSR SP was not only observed in aluminum matrix
composites but also in many other material systems such as mechanically alloyed (MA)
aluminum alloys, magnesium matrix composites, oxide dispersion strengthened (ODS)
12
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
alloys, and various aluminum alloys. However, this report is concerned only with
aluminum-based composites and alloys.
2.2.2 High strain rate superplasticity in aluminum matrix composites
In producing HSR SP aluminum matrix composites, any of the commonly used
structural aluminum alloys, such as 2xxx, 5xxx, 6 xxx and 7xxx series, can be utilized as
matrices. However it should be noted that it is not always possible to produce HSR SP
aluminum matrix composites by using any combination of matrix and reinforcement.
Therefore, there is a relationship between the chemical composition o f the matrix alloy
and the types o f reinforcement in order to show HSR SP. It is revealed that there are
several important parameters in controlling HSR SP in aluminum matrix composites: the
reinforcement-matrix combination, the processing method, the testing temperature, the
testing strain rate, the strain rate sensitivity and the grain size. Table 2.1 gives some of
the results on HSR SP behavior in aluminum matrix composites.
The HSR SP features in aluminum matrix composites can be briefly summarized
as follows:
1) The aluminum matrix is o f fine grain size, typically 1 to 5 pm.
2) Superplasticity is observed at very high strain rates, typically IC T 1 to 10 s'1 .
3) The largest elongation is obtained when the value of m is more than 0.3.
4) The optimum superpiastic temperature may be above the incipient melting point of
the matrix.
13
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 2.1 High strain rate superpiastic behavior in aluminum matrix composites
Composites T
(K)
e
(s'1 )
EF
(%)
m d
(pirn)
Ref.
SiCw/2124 798 3.3 x 10'1 300 0.33 1 [12]
Si3 N4 w /2124(1 ) 798 1.7 x 10'1 250 0.5
-
[29]
Si3 N4 w /2124(2 ) 818 4 x 10'2 280 0.3 4 [30]
Si3 N4 p /2124(3 ) 788 4 x 10'2 830 0.3 2 [31]
Si3 N4 p /2124(4 ) 773 3 x 10'1 280 0.3 1 [30]
SiCw /2024(5 ) 823 I 150 0.5 2.4 [32]
SiCw /2024(6 ) 783 9 x 10‘2 140 0.38 - [33]
Si3 N4 w /2024 773 1.7 x 10'1 175 0.3 - [33]
Si3 N4 p /5052 818 1 700 0.3 < I [34]
SiCw/6061 823 1.7 x 10'1 300 0.3 - [35]
SiCp/6061(7 ) 853 1.3 x 10'1 350 0.5 - [36]
SiCp/6061(8 ) 853 3 x 10'1 207 0.3 - [37]
Si3 N4 w /6061(9 ) 818 2 x 1 0 '1 260 0.5 3.1 [38]
Si3 N4 w /6061(l0) 818 2 x 10'1 600 0.5
- »
j [39]
Si3N4 w /6061(1 1 ) 818 2 x 10’2 85 0.3 - [40]
Si3 N4 w /6061(l2) 818 2 x 10'2 173 0.3 - [40]
Si3 N4 w /6061(l3) 818 3 x 10'1 150 0.3 3 [38]
Si3 N4 p /6061(l4) 818 1 x 10*1 450 0.3
• * >
j [41,42]
Si3 N4 p /6061(l5) 833 2 500 0.3 1.9 [41,43]
Si3 N4 p /6061(I6) 833 1 620 0.3 1 [41,44]
AlNp/6061(1 7 ) 873 1 250 0.5 - [45]
AlNp/606l(I8) 873 9 x 10'1 509 0.5 - [45]
Si3 N4 v v /7064(l9) 798 8 x 10 '1 160 0.4 3.5 [46]
Si3 N4 w /7064(2 0 ) 818 2 x 10'1 230 0.4 3.8 [46]
Si3 N4 w /7064(2 1 ) 818
5 x 10'1 240 0.5 3.5 [47]
Si3 N4 w /7064(2 2 ) 833 1 x 10'1 380 0.3 <3 [48]
Si3 N4 w /7075 773 1.8 x 10 '1 260 0.3 - [49]
AiNp/TN90 913 2.7 x 10'1 246 0.3
2
[50]
• Extrusion ratio of (1) = 44:1 and (2) = 100:1.
• Diameter of Si3N4 for (3) = I pm and (4) = 0.2 pm.
• (5) was produced by PM and (6 ) was produced by IM.
• (7) was produced by PM and (8 ) was produced by IM.
• Extrusion ratio of (9) & (11) = 44:1 and (10) & (12) = 100:1.
• (9) & (10) were produced by PM and (11) & (12) by IM.
• (13) was produced by using cx-Si3N4 w and extrusion ratio = 44:1.
• Diameter of AIN for (17) = 1.42 pm and (18) - 1.78 pm.
• (19) was produced by using a-Si3N4 w and extrusion ratio of ( 19) & (20) = 44:1 .
• (21) was extruded by 100:1 at 773 K and (22) was extruded by 100:1 at 793 K ..
14
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5) The fact that the fracture surfaces from the superplastically deformed composites
exhibit considerable fiber pullout indicates that extensive interfacial sliding (IS), in
addition to GBS, takes place during superpiastic deformation.
6 ) The chemical composition of the aluminum matrix plays an important role in HSR
SP.
7) The morphology and dimensions of the ceramic reinforcement and the structure and
chemical composition of the reinforcement/matrix interface are also important.
8 ) Greater amounts of prior work (100:1 extrusion compared to 44:1) lead to faster
optimum strain rates, lower stresses and greater elongations.
9) Whiskers having a-Si3 N4 structure are not as beneficial as p-Si3 N4 for HSR SP.
10) The composites made by PM route are superior to those made by IM route.
11) A fine matrix grain size is a necessary but insufficient condition for HSR SP in
aluminum matrix composites.
2.2.3 Origin of high strain rate superplasticity in aluminum matrix composites
At the present time, the exact physical mechanisms associated with HSR SP in
metal matrix composites (MMCs) have not been fully established. From the previously
mentioned features on HSR SP in aluminum matrix composites, however, it is obvious
that the origin of HSR SP in aluminum matrix composites may be very different from
that of superplasticity in conventional fine structure superpiastic metals. The most
interesting fact shown in Table 2.1 is that the optimum superpiastic temperatures for
aluminum matrix composites are quite high when compared to those for conventional
1 5
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
superpiastic metals, and a maximum elongation is often achieved at a temperature very
close to or slightly above the partial melting temperature for the aluminum matrix. The
first reasonable effort to understand these abnormally high optimum superpiastic
temperatures in HSR SP aluminum matrix composites was made by Higashi et al.
[51,52], and they focused on the fact that the presence of a liquid can affect the
accommodation process for GBS during superpiastic deformation o f the materials
exhibiting HSR SP. They further suggested that the presence of a liquid phase at the grain
boundaries and interfaces plays an important role in superpiastic flow at high strain rates,
and not only to enhance the strain rate but also to assist strain accommodation and thus to
delay the fracture process.
Therefore, it is worth comparing the test temperature for the composite with the
incipient melting point o f the matrix. The characteristic temperatures for aluminum
matrix composites and aluminum alloys are shown in Table 2.2 [51,56]. Below the
incipient melting point, a material is entirely in the solid state. Above the incipient
melting point, but at temperatures near the incipient melting point, a small amount of
liquid phase is present. This liquid phase is expected to segregate to either the
reinforcement/matrix interfaces in composites or to the grain boundaries in the alloy. The
maximum elongation to failure appears to occur at a temperature slightly higher than the
incipient melting point. Following this maximum elongation, however, the elongation
decreases sharply with a further increase in the test temperature. It is evident that there is
a critical amount o f liquid phase for the optimization of GBS or IS during superpiastic
deformation [51],
16
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 2.2 Characteristic temperatures of aluminum alloys and their composites
Materials Test temp. Incipient
melting point
Solidus temp. Ref.
(K) (K)
(K)
SiCw/2124 798 783
-
[51]
Si3 N4 w /2124 798 - 775 [56]
Si3 N4 p /2124 788 784 -
[51]
Si3 N4 w /6061 833 843 855 [51,56]
Si3 N4 p /6061 833 830 -
[51]
I T
a.
< N
o'
I I
Si3 N4 p /6061 833 829 -
[51]
{dp = 0.5 pm)
Si3 N4 p /6061 818 822 -
[51]
{dp = I pm)
Si3 N4 w /7064 798 - 798 [56]
SiCp/IN9021 823 751 8 6 6 [51,56]
Al-Ni-Mm 885 873 897
[51]
Al-Ni-Mm-Zr 873 - 898
[51]
PVD Al-Cr-Fe 898 -
896
[51]
IN9021 823 754 852 [51,56]
EN90211 748 - 768 [56]
IN9052 863 837 8 6 6
[51]
EN905XL 848 818 851
[51]
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Even though the test temperature is slightly below the incipient melting point of
the material, a small amount of the liquid phase may be still present. It is well known that
the chemical composition at the SiC/Al interface is substantially different from that in the
bulk matrix. Strangwood et al. [53] studied solute segregation at interfaces in SiC/2xxx
aluminum composite and shown that the segregation of both magnesium and copper is
expected to reduce the local melting point o f the aluminum matrix near the interface
region. Another point is that, due to the extremely fast strain rates, the effective
temperature is expected to be significantly higher than the test temperature as a result of
adiabatic heating [54]. It is also known that the presence of a liquid phase or a low
melting point region at grain boundaries in a crystalline solid can greatly enhance the
deformation rate via GBS [55],
An optimum superpiastic elongation was found at the temperature where local
melting of interfaces was confirmed by in situ transmission electron microscopy (TEM)
observation [57,58], Electron energy loss spectroscopy also indicated the segregation of
solute elements along the grain boundaries and interfaces, which are considered to be a
cause of partial melting [58]. From the above direct evidence, Higashi et al. [52,57,58]
concluded that the presence of a liquid phase relaxes high stress concentrations and
promotes IS without excessive cavity nucleation at the interfaces, so that large
elongations can be obtained at the temperatures close to the melting point. However, a
much larger volume fraction of a liquid phase does not contribute to large elongations.
Although this model can explain the significance of a liquid phase physically,
however, the complete deformation mechanism involving the presence of the liquid phase
1 8
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
on HSR SP should be constructed to give the parameters such as p and O for aluminum
matrix composites. Very recently, Higashi et al. [59] reported that the superpiastic
deformation in Si3N4W /7064 composites could be divided into two regions. Region I, the
low temperature region, shows a low activation energy o f ~142 kJ/mol, which
corresponds to lattice self-diffusion of aluminum, but Region II, the high temperature
region, shows a high activation energy of more than ~800 kJ/mol. The transition point
from region I to II corresponds to the incipient melting points o f the composites. This
high apparent activation energy above the incipient melting point may be attributed to the
fact that the presence of the liquid phase greatly accelerates the strain rate. From the
above, it is very clear that the deformation mechanism in HSR SP should involve both a
solid and a liquid phase.
Alternatively, another analysis on HSR SP was made by Mishra et al. [60] and
they obtained « = 2, (2 = 313 kJ/mol, D = D, (interfacial diffiisivity) and p = 1 . The
activation energy is comparable with the activation energy for grain boundary diffusion
in SiC and it was concluded that the mechanism for HSR SP in composites is probably
interface diffusion controlled GBS. Therefore, it is suggested that a liquid phase is not
necessary to account for HSR SP in MMCs. They also suggested that the activation
energy for HSR SP changes significantly with respect to the different levels of
reinforcement volume fraction, such as from grain boundary diffusion controlled sliding
in composites with less than 1 0 vol% of reinforcement to interfacial diffusion controlled
sliding in composites with more than 1 0 vol% o f reinforcements.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.2.4 High strain rate superplasticity in aluminum alloys
Fine-grained aluminum alloys have attracted considerable interest in the last
decade. This includes modified conventional aluminum alloys, mechanically alloyed
(MA) aluminum alloys, nanocrystalline aluminum alloys and severely deformed
aluminum alloys. The modification of conventional aluminum alloys by Zr, Cr, and Mn
additions leads to a fine grained microstructure since the intermetallic phases of Zr, Cr,
Mn pin the grain boundaries [61]. The mechanically alloying process involves the
formation of both fine 6 precipitate up to 5 vol% and insoluble fine oxide and carbide
dispersions (-30 nm) up to 5 vol% as grain refiners [62]. A new aluminum-based
crystalline alloy, basically an as-extruded Al-Ni-Mm (Mm = Misch metal) alloy, is
fabricated by warm consolidation of its amorphous or nano crystalline powders [63,64],
and both Al3 Mm and AI3 N 1 particulates of 70 nm in size having a high volume fraction of
about 40% are uniformly distributed. Finally, severe plastic deformation (SPD)
techniques, such as torsion straining (TS) and equal-channel angular (ECA) pressing, are
capable of producing materials with grain sizes at the submicrometer level [16,17], Some
of the results on HSR SP in aluminum alloys are shown in Table 2.3.
The HSR SP features in aluminum alloys can be summarized as follows:
1) A very fine grain size is essential, typically 0.1 to 4 pun. The advanced aluminum
alloys consolidated by amorphous or nanocrystalline powders have the smallest grain
size of about 100 nm and MA alloys also have very small grains o f about 500 nm.
2) Superplasticity is observed at very high strain rates, typically 10' 2 to 50 s*1 . The MA
alloys shows the highest optimum strain rates of 1.3 to 50 s'1 .
20
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 2.3 High strain rate superpiastic behavior in aluminum alloys
Alloys T
(K)
e
(s'1 )
EF
(%)
m d
(fim)
Ref.
IN9021 723 1.3 300 0.3 - [13]
IN9021 823 50 1250 0.5 0.51 [65,66]
IN90211 748 2.5 525 0.3 0.5 [54,67]
IN9052 863 1 0 330 0.5 0.5 [6 8 ]
EM905XL 848 2 0 180 0.5 0.4 [69]
IN90 913 1.5 x 10’ 2 230 0.49 2 [50]
Al-Ni-Mm 885 1 650 0.5 0 . 1 [70]
Al-Ni-Mm-Zr(1 ) 873 1 650 0.5 0.05-0.08 [70]
Al-Ni-Mm-Zr(2 ) 873 1 600 0.3 0 . 1 2
[71],
PVD Al-Cr-Fe 898 1 505 0.5 0 . 1 [70]
RSP Al-Li 843 1.4 x 10' 1 250 0.3 < 4 [72]
RSP 7475-Zr 723 8 x 1 0 ‘ 2 390 0.5 1 [61]
IM 2618A 803 2 . 8 x 1 0 ‘ 1 240 0 . 2 <3 [73]
PM Al-Mg-Cr 793 2 x 1 0 ’ 2 1 0 0 0 0.5 3 [74]
PM 2124-Zr 748 3 x 10’ 1 490 0.5 1 [75]
PM 7475-Zr 793 3 x 10’ 1 900 - 1.9 [75]
PM 7475-Zr 788 1 x 1 0 ' 1 1 0 0 0 0.3 2 [76]
PM Al-Ti-Fe 908 8 x 1 0 ' 1 500 0.3 0 .8 - 1 . 2 [77]
IM AI-Mg-Sc(3 ) 672 1 x 1 0 ' 2 1 0 2 0 0.4 - [78]
IM Al-Mg-Sc(4 ) 748 1.4 x 10' 2 1130 0.58 0 .2 - 1 [79]
Al-Cu-Zr(5 ) 573 1 x 1 0 ' 2 970
-
0.5 [15]
Al-Mg-Li-Zr(6 ) 623 1 x 1 0 ‘ 2 > 1180 - 1 . 2 [15,80]
Al-Mg-Li-Zr(7 ) 673 1 x 1 0 ' 1 1 2 1 0 - 0 . 8 [81]
Al-Mg-Sc(8 ) 673 3.3 x 10' 2 1030 - 0 . 2 [82,83]
Al-Mg-Sc(9 ) 673 3.3 x 10‘ 2 1560 - 0 . 2 [83]
• Al- 14Ni-7Mm-lZr for (1) and Al-l4.8Ni-6.6Mm-2.3Zr for (2).
• Al-4Mg-0.5Sc for (3) and Al-6Mg-0.3Sc for (4).
• Ail alloys from (5) to (9) were ECA pressed.
• Route A was applied for (6) and route Bc was applied for (7).
• (8) was ECA pressed up to 8 passes and (9) was ECA pressed up to 12 passes.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3) The largest elongation is obtained when the value of m is more than 0.3.
4) The optimum superpiastic temperature may be above the incipient melting point for
some alloys. Severely deformed alloys produced by ECA pressing exhibit the lowest
optimum superpiastic temperatures.
5) Both PM and IM can produce aluminum alloys which exhibit HSR SP.
2.2.5 Origin of high strain rate superplasticity in aluminum alloys
The optimum superpiastic strain rate range in HSR SP aluminum alloys is about
two to four orders of magnitude higher than for conventional aluminum alloys. There is
no doubt that a fine grain size is essential for HSR SP because a decrease in grain size d
leads directly to an increase in the maximum strain rate for superpiastic flow as seen in
equation (2.1), typically aocd~2 or a oc d ~ l . In Fig. 2.2, the logarithmic strain rate is
plotted as a function of logarithmic flow stress and the effect of grain refinement on the
optimum strain rate range is illustrated schematically [84].
Two independent processes, namely GBS and slip, are represented as straight
lines, and the point of intersection, a ( 1 ), represents the maximum strain rate for a given
micro structural condition. For the case of independent mechanisms, such as GBS and
dislocation creep, the faster process will control the overall deformation. Grain
refinement can push the transition from GBS to slip mechanisms into high strain rates,
and the dashed lines in Fig. 2.2 show the significant increase in the maximum strain rate,
a (2 ), for superpiastic flow.
2 2
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
8 dV\^
Slip
Grain
refinement
Therefore, it seems that HSR SP in aluminum alloys is a direct result of grain
refinement. The average grain size of modified conventional aluminum alloys showing
HSR SP, such as 2124-0.6wt% Zr alloy [75] and 7475-0.45wt% Zr alloy [61], is about 1
pirn. Furthermore, the average grain size o f MA aluminum alloys is about 0.5 pm and that
o f nanocrystalline aluminum alloys, such as Al-Ni-Mm alloy, is less than 0.1 pm. On- the
other hand, the conventional superpiastic aluminum alloys showing low strain rate
superplasticity have an average grain size o f 5 to 10 pm.
However, it appears that a fine grain size is a necessary but insufficient condition
to produce HSR SP. For example, both PM64 and 2024 Al alloy produced by TMP
method have a very fine grain size of 6 p.m and 2.5 pm but they exhibit relatively poor
ductility over the strain rate range from 10" 2 to 1 s’ 1 [85,86], Another example is the MA
aluminum alloy, IN905XL alloy, which has a very fine grain size o f 0.4 pm and shows a
relatively small maximum elongation of 180% at a high strain rate of 10 s' 1 at 863 K.
This illustrates the critical effects of micro structure and chemical composition of the
alloy on the development o f HSR SP.
When the maximum elongation data of nanocrystalline aluminum alloys (i.e. Al-
Ni-Mm or Al-Ni-Mm-Zr alloys) are compared with those of MA aluminum alloys (i.e.
IN9021 alloy), as shown in Table 2.3, the former has smaller initial grain size but they
are inferior in their maximum elongations. In spite of the nano-scale grained structures,
less than 100 nm in size for as-extruded Al-Ni-Mm alloys, these fine extruded structures
are unstable at high temperatures. The value of the grain size in Al-Ni-Mm alloys
annealed at high temperatures is relatively larger (~1 pm) than that in MA aluminum
24
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
alloys (~0.5 pm) at the same condition. Therefore, the optimum strain rate range of
nanocrystalline aluminum alloys (10" 1 - I s'1 ) is slower than that of MA aluminum alloys
(1-50 s'1 ) and a stable grain size at high temperatures is needed to obtain HSR SP.
Until now, two different explanations have been put forward to explain the HSR
SP mechanism in MA aluminum alloys. The first one is that HSR SP is an extension of
conventional superplasticity and it is a direct result of the fine and stable microstructure
o f MA alloys [60,62,67,87], Bieler and Mukheijee [60,87] have described the rate-
limiting step of superpiastic deformation of IN90211 in the following terms. The
operative mechanism for superpiastic behavior in MA aluminum alloys is grain boundary
diffusion controlled GBS. Dislocations within the grain interior are glide controlled due
to solute atom drag effects. They arrive at the boundaries and dissociate into extrinsic
grain boundary dislocations that determine the rate of GBS. This model predicts « = 2, p
= 2 and grain boundary diffusion as rate controlling.
The predictions of the Bieler and Mukheijee’s model [87] show a good agreement
with their experimental data but experimental data of Higashi el al. [65,68,69] lie about
an order of magnitude faster than their model. They attribute this disagreement to the
difference of grain size because a small grain size can account for this faster rate due to
the grain size dependence on strain rate [60], However, the IN9021 alloy used by Higashi
et al. [65] has a similar grain size of about 0.5 pm to that used by Bieler and Mukherjee
[60], and therefore it seems that their model fails to predict the exact mechanism for the
MA aluminum alloys.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Alternatively, the second model proposed by Nieh and Wadsworth [56] arose
from the study of aluminum matrix composites. They focused mainly on the deformation
temperature and suggested that the faster deformation mechanism leading to HSR SP is a
unique one that depends on the presence of a liquid phase at grain boundaries or
interfaces. Very recently, another but similar model was proposed by Mabuchi and
Higashi [52,57] that superpiastic flow is controlled by a GBS mechanism accommodated
with a relaxation of the stress concentration by an isolated liquid phase at interfaces
between the matrix and reinforcements of the composites or MA alloys exhibiting HSR
SP.
The above models based on liquid phases at grain boundaries or
matrix/reinforcement interfaces cannot only be applied to account for HSR SP in
aluminum matrix composites but also they apply to advanced aluminum alloys, such as
nanocrystalline aluminum alloys and MA aluminum alloys because the optimum
superpiastic temperatures for these alloys are very close to or above the incipient melting
point of the matrix, as shown in Table 2.2. Higashi el al. [8 8 ] reported that superpiastic
flow in the Al-Ni-Mm alloy is governed by GBS accommodated by dislocation climb
below the incipient melting point but by GBS accommodated by the liquid phase at grain
boundaries or interfaces above the incipient melting point. This result is quite similar to
the case o f aluminum matrix composites [59],
The PM aluminum alloys showing HSR SP generally contain a significant volume
fraction o f second phase particles. In addition to the intentionally added second phase
constituents (Zr) or the precipitates that form because of the alloy chemistry, AI2O 3
26
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
particles are invariably present. It was suggested that the mechanism o f HSR SP in the
modified conventional aluminum alloys produced by PM method, such as Zr-modified
aluminum alloys, is quite similar to the mechanism of superplasticity in many
conventional systems. The m value of 0.5, an inverse grain size dependence of 2 and an
activation energy close to that for grain boundary diffusion have been predicted [75,76],
The HSR SP behavior in heavily deformed aluminum alloys is quite unique and
very attractive in terms o f superpiastic forming (SF) technology [15,80-83], Since the
optimum superpiastic temperatures for severely deformed aluminum alloys are
significantly lower than those for other aluminum alloys showing HSR SP, a liquid phase
cannot play any role in their superpiastic flow. In addition, although the grain size of the
severely deformed alloys is not as fine as that o f MA alloys or nanocrystalline alloys,
excellent elongations at high strain rate range can be obtained in many systems.
However, it remains in debate whether HSR SP in severely deformed aluminum alloys is
an extension of conventional fine structure superplasticity.
2.3 PRINCIPLES OF EQUAL-CHANNEL ANGULAR PRESSING
2.3.1 Introduction
Among the methods of fabricating ultrafine-grained (UFG) materials, severe
plastic deformation (SPD) techniques, such as torsion straining (TS) and equal-channel
angular (ECA) pressing, are probably the most promising [16,17], The ECA pressing
technique was invented in 1972 by a scientist from the former Soviet Union, V.M. Segal
27
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
[18], and it uses the principle o f repeated shear deformation to refine the grain size in
metallic materials.
The ECA pressing has a number of advantages [89]: First, very large deformation
strain can be obtained after repeated passes without changing the shape of billet. Second,
very uniform and homogeneous deformation can be applicable throughout the cross
section of the billet. Third, no residual porosity is found in the deformed billet. Fourth,
since the size of the billet is only limited by the size of the die and the pressing facility, it
is possible to produce massive samples. Fifth, the areas exposed to tensile stresses are
limited during deformation.
2.3.2 Fundamental concept of ECA pressing
In ECA pressing, as shown in Fig. 2.3, a sample is pressed through a die in which
two channels of equal cross-section intersect at an angle of < f > and an additional angle of
'F defines the arc of curvature at the outer point of intersection of the two channels [90],
In the case of no friction between the die walls and the billet, the strain associated with a
total o f N pressings through the die having a round-cornered channel, e , v , is given by the
following relationship [90]:
2 c o t 1
U 2 )
+ 1 ? cos ec (2.4)
Therefore, it is possible to press the same billet through the die a number of times in
order to achieve a high total strain because the cross-section of the pressed billet is not
changed after pressing.
2 8
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
plunger
test material
Figure 2.3 Schematic illustration of equal-channel angular pressing.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
When no frictional effect is also considered, on the other hand, Segal [91] showed
that the total strain after N pressings through a die having a sharp-cornered channel is
given by the following equation:
£l V = — j= ^ cot
-s/3
(2.5)
v 2 y
It is evident from the above equations that the total strain accumulated in the pressed
billet is only a function o f the number o f passes and angles of a die. These theoretical
relationships were verified experimentally by using either billets made from plasticine
layers o f two different colors [92] or divided billets where grids were drawn on one
surface o f one half-billet [93]. It was concluded that the above relationships can be
directly applied to the center of the billets but they may not be valid at areas away from
the center o f the billets due to friction.
As another approach to understanding the deformation process during ECA
pressing, results of finite element modeling (FEM) simulations were reported [94,95],
The FEM analysis is very useful because the effect of friction and the thermal gradient
during pressing can be considered. The important results are as follows: When friction is
taken into account, the above relationships are valid in the middle o f the billet but the
portion and the intensity of the inhomogeneous deformation which occurs on the bottom
surface o f the pressed billet increases with die angle [94]. The thermal gradients induced
by die chill and deformation heating during pressing play an additional role in producing
flow localization; thus, larger billets are beneficial to produce uniform structure because
the chilled zone can be minimized [95].
30
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.3.3 Deformation behavior during ECA pressing
The exact deformation mechanism during ECA pressing has not been determined
but one reasonable explanation was made by Berbon [89]. He considered the low energy
dislocation structure (LEDS) theory as a basic deformation model during ECA pressing.
The LEDS theory can be summarized briefly as follows [96,97]: During plastic
deformation, the LEDS readily occurs as a form of cells with a low internal dislocation
density and cell boundaries composed of dislocation tangles inside grains. The driving
force for the formation o f the LEDS is the strain energy decrease associated with the
formation of tangles because of the mutual screening o f the stress fields of the
dislocations. At the initial stage of ECA pressing, a very large number of lattice
dislocations is created within the large grains but these dislocations cannot rearrange into
perfect LEDS because a very high strain of ~ l is introduced in a very short time. These
locally rearranged dislocations possess high stresses and these high stresses can stabilize
the subgrain boundaries. All actions involving cell boundary formation and
rearrangement as subgrains occur within the first pass. Further pressing leads to an
increase in misorientation angles of subgrain boundaries by absorption of newly created
lattice dislocations. Therefore, it can be assumed that the grain size is saturated after the
initial pressing.
Although the above model can explain the microstructural evolution during ECA
pressing very simply, a number of factors should be considered. For example, when the
pressed billet is rotated, different slip systems within grains become operative. This can
form very complex dislocation boundaries and thus further grain refinement may be
31
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
possible after the initial pressing due to the interaction of dislocation cell boundaries.
Other factors, such as pressing temperature and precipitates in alloys, may also affect the
deformation mechanism during ECA pressing and these other factors make it complex to
establish the deformation mechanism during pressing.
2.4 FACTORS ASSOCIATED WITH EQUAL-CHANNEL ANGULAR PRESSING
2.4.1 The angles between the two channels (O and *P)
The amount of strain introduced in the material through ECA pressing is a direct
function of two angles, < t> and ¥ , as shown in equation (2.4). In the case of ECA pressing
through a die having a sharp-cornered channel, = 0 °, and the strain accumulated
through a die is given by equation (2.5). In fact, the effect of angle, *?, on the strain for a
round-cornered die is quite negligible because equation (2.4) is not a strong function of ¥
but a strong function of O. Only a small deviation in the strain accumulated in a single
passage can be obtained when the values of'F are changed at fixed values of < t > [98],
It was pointed out that a higher pressure is required to press the sample through a
die having a sharp-cornered channel when compared to a die having a round-cornered
channel [98] and it was also demonstrated that there is some space between the billet and
the wall at the comer where the two channels intersect in the case of ¥ = 0° [92].
Furthermore, it was shown by using billets made from plasticine layers of two different
colors that the sharp-cornered channels result in larger and more non-uniform
deformation than the round-comered ones [92], Therefore it may be beneficial to use a
32
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
die with a finite value for 'F so that the resultant strain in a single passage will be
approximately the same as that for a die with sharp-cornered channel.
The effect of the angle, < t> , on the strain is clearly seen in equations (2.4) and (2.5)
and, typically, a single passage through a die with < t > = 90° will always give a strain close
to — 1.0 regardless of the value o f v j/. This was investigated experimentally by using four
different dies with angles from 90 to 157.5° and pure aluminum billets were rotated by
90° between each consecutive pressing up to a strain of 5 [98]. The grain or subgrain
sizes after ECA pressing were a minimum when using O = 90° and a maximum when
using O = 157.5°. An ultrafine microstructure, with equiaxed grains separated by high
angle grain boundaries, was developed most readily by ECA pressing through the die
with d> = 90°. Alternative procedures for introducing the same cumulative strain through
smaller strain increment, using dies with higher values of O, cannot produce the same
micro structure which is achieved when using < f > = 90°.
Another interesting approach was performed using the FEM analysis [94], Two
dies of < 3 > = 90 and 100° with sharp-cornered channels were modeled with and without
friction. It was revealed that the friction between the sliding surfaces was more severely
developed for a die of = 90° and more uniform deformation o f the billets was obtained
for the case of < t > = 100°. However it should be noted that this simulation was performed
only for an initial pressing so that no micro structural development was discussed.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.4.2 The number of ECA pressings
This factor is also directly related to the amount of strain introduced in the billet
as shown in equations (2.4) and (2.5), and this is one of the practical parameters which
can be changed during ECA pressing without any difficulties while the angles o f the die
cannot be modified during the operation.
It was revealed that a single pressing leads to a substantial reduction in the grain
size and the microstructure consists of parallel bands of subgrains further divided by
many boundaries having very low angles of misorientation. These bands were always
formed essentially parallel to the side faces of the pressed sample and lying at 45° to the
top and bottom faces. The same observations were made for both pure aluminum
[99,100] and Al-Mg binary alloys [101], It was also shown that the additional pressings
produced microstructural evolution into an array of equiaxed grains having high angle
grain boundaries [99-101], However, further detailed microstructural development cannot
be explained only by the number o f pressings and other variable, such as the pressing
routes, should be combined with the number of pressings to explain the results
completely.
Another interesting point was that the grain sizes were often saturated after a
certain number o f pressings and this number of pressings to establish a homogeneous
micro structure was dependent on the pressed material [101]. For example, for pure
aluminum, the number of pressing is four and the stable grain size is ~1.3 pm while the
numbers of pressing is six and eight, and the stable grain sizes are -0.45 and -0.27 pm,
for AI-1% Mg and Al-3% Mg alloys, respectively.
34
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.4.3 The pressing routes
The nature of the microstructural evolution depends strongly upon whether the
samples are pressed with or without rotation between each pressing because of the
modification of the shear planes and shear directions in the samples during ECA pressing
[91,99], Two pressing routes were initially introduced by Segal [91]. In route A, the billet
is removed from the die and the pressing is repeated without any rotation of the billet and
in route C the billet is rotated through 180° between each pressing. Another route, route
B, denotes the situation where the sample is rotated by 90° between each pass, and this
can be divided into two different routes. The two routes of BA and Be denote a rotation
by 90° in alternate directions between each pass and by 90° in the same direction between
each pass, respectively [100,102,103]. The schematic illustration for the four routes is
shown in Fig. 2.4.
The first detailed research on the effect of the pressing routes was conducted by
Ferrasse et al. [104] and it was reported that route C produced a more stable, equiaxed
and uniform submicrostructure than the other two routes, routes A and BA , when
aluminum alloy 3003 was subjected to ECA pressing at room temperature. After multiple
pressings with route A, by using pure aluminum, it was also reported that the evolution of
the elongated subgrains into an array o f high angle grain boundaries was slowest, the
evolution was less rapid after multiple pressings with routes BA and C, and the
development of a homogeneous microstructure of grains separated by high angle grain
boundaries occurred most rapidly when the sample was pressed via route Be [ 1 0 0 , 1 0 2 ].
35
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Route A
Route B a
i » N
ISSi
IfillllS iK
Route B q
H i t i
tt8t§
i s i i
Route C
« « i i l ^
MMxAs! I l l l t l l t i t i S
Figure 2.4 Schematic illustration of the four pressing routes.
36
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
294898
457283
Another interesting pressing route was introduced by Huang et al. [105] and they
designated this new route as route C2 . This route is also named as a modified route C and
the route C was modified in a way that before each rotation the billet was repeatedly
pressed twice through the channel without any direction change. Compared with route C,
route C2 resulted in more equiaxed grains and they concluded that the amount of strain
before each rotation in route C has a significant effect on the microstructure.
A FEM analysis on the effect o f the pressing routes has not been conducted yet
because it may be too difficult to demonstrate the shearing in two-dimensional figures
when the rotations of samples are involved. Alternatively, a simple mechanical analysis
was developed to understand the shearing characteristics for different pressing routes
within the sample during ECA pressing [106]. It was also shown that the shearing models
developed for the various pressing routes were in good agreement with the corresponding
microstructures investigated by optical microscope and TEM [107], This will be
discussed fully in chapter 4.
2.4.4 The pressing speed
Since the pressing speed directly governs the deformation rate of the material
subjected to ECA pressing, it may play an important role in developing microstructure,
especially accommodation mechanisms after severe plastic deformation, within the
pressed sample. In terms of grain size, however, it was revealed that the microstructure
after ECA pressing has little or no significant dependence on the pressing speed, at least
in the testing range from ~10‘ 2 to ~10 mm-s' 1 [108]. Further detailed inspection at higher
37
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
magnification TEM revealed that recovery occurred more easily when pressing at the
slower speed. Therefore, it was concluded that the additional time available at the slower
pressing speed permitted the development of a more equilibrated micro structure [108].
The above experiment used very ductile materials of pure A 1 and an Al-1% Mg
binary alloy. For soft materials, the pressing speed may not be important because the
numerous dislocations introduced within the grains can be accommodated very easily.
For hard materials such as commercial aluminum alloys, however, the pressing speed is
one of the most important parameters controlling the soundness o f the pressed billets in
ECA pressing because some accommodation mechanisms should be followed by pressing
in order to prevent the premature breaking of the billets in earlier pressings. For example,
when pure titanium was subjected to ECA pressing at 598 K by one pass, uniform billets
were obtained at a pressing speed between 0.25 mm-s' 1 and 2.5 mm-s'1 , but a segmented
billet was obtained at the high pressing speed of 25 mm-s'1 . The same result was obtained
for 4340 steel after ECA pressing at 598 K by one pass [109],
Another interesting effect of the pressing speed is the adiabatic heating developed
during ECA pressing. It was reported that there was no significant adiabatic heating when
using the slower pressing speed of 0.18 mm-s’ 1 whereas there was an abrupt increase in
the temperature on passing through the shearing plane when using the faster pressing
speed of 18 mm-s' 1 [110]. In order to minimize the effect o f adiabatic heating on the
microstructure, therefore, it was suggested that ECA pressing should be conducted at
reasonably slow pressing speeds.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.4.5 The pressing temperature
There has been no systematic investigation of the effect of the pressing
temperature. However, this is a very important parameter because it not only determines
the soundness of the pressed billet after multiple pressings macroscopically but it also
affects the characteristics of dynamic recovery of dislocation networks within the billet
during ECA pressing microscopically.
The maximum allowable number o f pressings is dependent on the pressing
temperature. When pure aluminum was pressed at room temperature and a low
temperature of 123 K, the ECA pressing at room temperature was able to continue up to
s H— 20.9 without breaking whereas the sample was pressed at 123 K only up to 15
passes (sN= 13.6) and then fractured at pass number 16 [111]. This is also consistent for
a hard material. When pure titanium was subjected to ECA pressing at 598 K by one pass
a uniform billet was obtained, but a segmented billet was obtained when it was pressed at
room temperature [109], Therefore, in order to obtain sound billets after multiple
pressings, the ECA pressing should be conducted at relatively high temperatures
especially for hard and brittle materials. For example, commercial aluminum alloys, Al-
1420 and Al-6061, were pressed at 673 K + 473 K and at > 383 K, respectively [81,112].
A very strong material, 0.15% C steel, was pressed at 473 K [113] and brittle hexagonal
close packed (HCP) materials, such as Mg-Al-Zn (AZ91) alloy and Ti-Al-V alloy, were
pressed at 448 K and at > 1118 K, respectively [114,115]. For HCP based materials,
higher pressing temperatures should be desirable in order to achieve a large strain
39
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
produced via ECA pressing because more slip systems can be operated at high
temperatures.
The pressing temperature may influence the microstructure during ECA pressing.
In terms of grain structure, it was reported that the 5056 aluminum alloy subjected to
ECA pressing at 373 K consisted of very fine grains o f ~0.5 jam and a dislocation.cell
structure, but for the alloy processed at 523 K most grains were recrystallized and coarser
(~2 pm) [116]. Therefore, it is obvious that the ECA pressing at elevated temperatures
can assist dynamic recovery or dynamic recrystallization during operation.
From the above mentioned five parameters influencing ECA pressing, it should be
noted that any one factor cannot fully control the deformation behavior during ECA
pressing. All parameters are correlated with each other so that the actual deformation
mechanism involved in ECA pressing is rather complex.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3. EXPERIMENTAL PROCEDURES
3.1 MATERIALS
Several materials were chosen for use in this program:
1. Pure A l
Pure Al (99.99%) was selected as a model material because all microstructural
complexity can be completely removed. It was processed to investigate the soundness of
the pressed billet and the effect of the pressing routes on the resultant micro structure and
on the mechanical properties.
2. Al-Sc-Zr and Al-Mg-Sc-Zr alloys
The two alloys, Al-0.2% Sc-0.12% Z r and Al-3% Mg-0.2% Sc-0.12% Zr,
containing Sc and Zr as grain refining elements were cast in Japan and both ingots were
homogenized at 753 K for 24 hours to produce an uniform microstructure through the
entire ingots. Finally the former alloy was annealed at 903 K for 1 hour and the latter
alloy was annealed at 883 K for 1 hour, and then both alloys were iced water quenched
immediately.
Research on these alloys is an extension of previous results [89] and two alloys
were used to investigate the effect of the combination of Sc and Zr additions on grain
refinement, thermal stability at elevated temperatures and superplasticity. The tests were
also intended to obtain very high elongations to failure and determine the possibility of
achieving HSR SP.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3. Al-2024 alloy
This alloy was provided by Alcan Aluminum Corporation, Cleveland, Ohio and
its chemical composition is Al-4.4% Cu-1.5% Mg-0.6% Mn (in percent by weight) [117].
This alloy is a commercial aluminum wrought alloy widely used for aircraft structures
and has high strength, very good machinability, fair corrosion resistance but limited
formability. It was received in the form o f an extruded bar and was initially fully
annealed (“O” - by the Aluminum Association’s Temper Designation System). Since the
as-annealed alloy is not used in practical applications, 2024-T3 51 heat-treated alloy was
also selected for comparison. The T351 heat treatment denotes that an alloy is solution
heat-treated, strain hardened and then naturally aged [117,118].
The Al-2024 alloy was originally chosen to improve the limited ductility in
commercial aluminum alloys by changing many variables of ECA pressing, such as the
number of passes, the pressing routes and the pressing temperatures. Furthermore, after
selecting the optimum condition for ECA pressing, it was tested to evaluate the
possibility of HSR SP in commercial aluminum alloys.
4. SUPRAL 100 alloy (Al-2004)
A SUPRAL 100 alloy, also known as Al-2004, with a chemical composition of
Al-6 % Cu-0.4% Zr was received from Superform USA, Riverside, California. The alloy
was cast, homogenized at 648 K for 5 hours and then hot rolled. This alloy is known as a
typical superplastic aluminum alloy and has been used extensively in the superplastic
forming industry.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Since the optimum superplasticity for this alloy occurred at relatively high
temperatures and at low strain rates, this alloy was tested to investigate the effect of ECA
pressing on the optimum superplastic conditions. It was also considered that there is a
possibility of achieving exceptional superplasticity at very high strain rate range. Finally,
this alloy was chosen to verify the effect of pressing temperature on the microstructure
and the resultant superplastic behavior.
5. Al-1420 alloy
This commercial hot rolled cast alloy was fabricated in Russia as a lightweight
and high strength alloy and its chemical composition is Al-5.5% Mg-2.2% Li-0.12% Zr.
This alloy has been studied extensively by Valiev et al. [15] and Langdon et al. [80] and
it was reported that Al-1420 alloy showed exceptional HSR SP. Therefore, this alloy was
chosen due to the great potential for HSR SP by changing some of the variables for ECA
pressing, such as the pressing routes, the number of pressings and the pressing
temperatures.
3.2 EXPERIMENTAL PROCEDURES
The brief overview of the experimental procedures is shown schematically in Fig.
3.1. Before ECA pressing, if necessary, the initial material was heat-treated and then cut
into small pieces to measure the grain size by optical microscopy. Also tensile tests and
microhardness tests were preformed in order to compare with data obtained after ECA
pressing.
43
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Initial material
Heat treatment
Pressed billet
Tensile test Microhardness test
Optical microscopy TEM/SIM
ECA pressing
at high temperatures
(373, 473, 573 and 673 K)
ECA pressing
at 298 K
Figure 3.1 Flow chart of the experimental procedures
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The ECA pressing was conducted at room temperature or at elevated
temperatures. The specific pressing variables, such as the pressing routes and the number
o f passes, were selected for each material. Before high temperature deformation, some of
the pressed materials were annealed to investigate the thermal stability o f the ECA
pressed microstructures. After ECA pressing, a number o f tensile tests and microhardness
tests were conducted to reveal the effect o f ECA pressing on the mechanical properties of
materials and to find the optimum superplastic conditions of the tested materials. The
microstructure was observed by transmission electron microscopy (TEM) and scanning
ion microscopy (SIM) before and after tensile testing to measure the grain size and to
investigate the effect o f grain growth during deformation.
3.3 EQUAL-CHANNEL ANGULAR PRESSING
3.3.1 Die and plunger
In principle, two types o f dies were used for ECA pressing. The main operating
mechanisms for the two dies are identical: The two channels were drilled in a bulk piece
of steel, with two angles 0 = 90° and 'F = 45°. The second channel was separated into
two parts, the first half part having a diameter slightly smaller than the first channel to
suppress the elastic expansion during deformation, and the second part having a diameter
slightly larger than the first channel to minimize the friction between die and billet [89].
A very small hole was also drilled in the opposite side of the second channel to put a
thermocouple to check the temperature o f the die. This consideration is shown in Fig. 3.2.
45
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Scale = 1:2
(unit = mm)
101.6
< j > 9.5
50.8
a '
'T
co
41.3
23.0
C M
L O
4 ) 9.7
41.3
CO
82.6
Figure 3.2 Dimensions o f the die for ECA pressing at room tempeature.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The main difference between the two dies was the heat treatment condition of the
materials used for dies. The first die, as shown in Fig. 3.2, was made of hardened steel in
the machine shop in USC and it was heat-treated for service at room temperature. The
second die was made o f SKD1 1 tool steel in Japan and it was heat-treated for service at
elevated temperatures. Therefore, the first die was used for ECA pressing at room
temperature and the second die at high temperatures. Another difference is the diameter
of the channels. The size o f the channel of the first die (9.5 mm) is slightly smaller than
that of the second die (10 mm). Therefore all dimensions of the first die in Fig. 3.2 are
slightly modified for the second die.
The plunger was also made o f the same material for each die. As shown in Fig.
3.3 (a), the diameter o f the plunger is slightly smaller than the first channel o f the die and
the length of the plunger was designed not to touch the wall of curvature. The threaded
end of the head of the plunger was designed to fit to the plunger protection cap for easy
removal from the die.
3.3.2 Specimen preparation for ECA pressing
Billets were cut from the initial materials parallel to the extrusion direction or the
rolling direction. The size and shape of the billet for ECA pressing is illustrated in Fig.
3.3 (b). The diameter o f the billet is slightly smaller than the first channel and exactly the
same as the plunger. The length of the billet is not important but the length should be
-55-75 mm. These billets were polished on the longitudinal faces prior to pressing by
using SiC abrasive papers from 180 to 600 mesh size in order to give a smooth surface.
47
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Scale =1:2
(unit = mm)
T hreaded e n d
4)25.4
R 8
120.7 25.4
146.1
(a)
Scale = 1:1
(unit = mm)
4 > 9.3
(b)
Figure 3.3 Dimensions of (a) plunger and (b) billet for ECA pressing at room
temperature.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.3.3 Procedure for ECA pressing
The ECA pressing facility was set up using a Dake hydraulic press having a .150
tons capacity as shown in Fig. 3.4. The press mainly consists of two plates: the upper
plate is fixed and the lower plate is moving. The maximum ram speed of the press is
about 19 inch/min (or 8 mm/sec) and the minimum ram speed is about 6 inch/min (or 2.5
mm/sec). The material can be pressed constantly at the maximum speed by setting the
pressing lever to the uppermost position, however, it is difficult to set the intermediate or
minimum speed constantly because the pressing lever cannot be fixed constantly during
operation.
As an initial step, the die was fixed firmly on the bottom plate by using a pair of
die supporters. Then the die and the plunger were aligned perfectly to avoid bending of
the plunger during the pressing. A specially designed plunger protection cap was used to
work as a removing mechanism, the extract of the plunger from the die. The next step
was to increase the pressing temperature if the ECA pressing was conducted at a high
temperature. The press has built-in heating elements, the upper and lower plates. Since
the thermal expansion o f the plunger should be minimized during heating, however, the
lower plate was chosen as a heating source. The temperature was controlled by the built-
in thermometer and it was always checked by a thermocouple put directly into the die in
order to read the temperature inside the die. It was possible to reach 523 K after 4 hours
of heating. For ECA pressing at room temperature, the same procedure as described
above was used without heating of the plate.
49
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Heat source
Fixed upper plate
Plunger
protection cap
Temperature
controller
Plunger
i
Die
j i
Moving lower plate
Heat source
Control lever
Multimeter
Figure 3.4 Schematic illustration of the high temperature pressing facility.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In order to reduce the friction between the die and the billet, two types of
lubricant were applied. For ECA pressing at room temperature, Sumico #150 containing
5 vol% of M0 S2 was used. For ECA pressing at high temperatures, spray-type Crown
#9105 containing anti-seize compound was applied. Before every pressing, the lubricant
was applied to both billet and plunger at the same time. After deforming a billet with one
pass, the billet was cleaned, lubricated and inserted again for the next pass. In some
cases, when a very strong material was pressed, the pressed billet was punched to fit into
the die for the next pressing. For ECA pressing at high temperatures, the pressed billet
was put on the lower plate in order to maintain the service temperature but the holding
time was minimized because the other billet is inside the die and a long holding time may
cause a change in the microstructure.
3.4 MECHANICAL TESTING
3.4.1 Tensile testing
Tensile specimens were machined from both the unpressed samples and ECA
pressed billets. These specimens were always cut along the rolling or extrusion direction
for unpressed samples and parallel to the pressing direction for the ECA pressed billets.
The size and shape of the tensile specimens is shown in Fig. 3.5. The gauge length of the
specimen is 4 mm and the width and thickness of the gauge area are 3 mm and 2 mm,
respectively. In some cases, a different gauge length of 5 mm was used for comparison
with other data.
51
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Scale = 4:1
(unit = mm)
Thickness = 2 mm
4.25
L O
CO
R 1.5
8.5 8.5
24
Figure 3.5 Dimensions of tensile test specimen.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The machine used for tensile tests was an Instron model 1125 having a 1000 lb
capacity load cell. By using this load cell, up to 85% of the load cell capacity was used
for the aluminum samples tested at room temperature while less than 30% of the load cell
capacity was used for samples tested at elevated temperatures. The constant cross head
speeds were set to give initial strain rates within the range of 10- 4 to 1 s'1 . Tensile tests at
elevated temperatures varying from 523 K to 823 K were performed using a single zone
resistance furnace in air and the temperature was carefully controlled to a precision of ± 2
K.
Initially, the specimens were tightened by specially designed grips and positioned
at the center o f the furnace. After setting the specific temperature and strain rate, the
output of the test was recorded in the form of load versus displacement of the cross head
on a chart recorder or into a computer. Finally, the relationship of stress versus strain was
calculated directly from the relationship between load and displacement. The engineering
strain, which can be transformed easily as the elongation to failure in percent, and the
true stress were usually adopted as the main components in the stress versus strain curves
in all cases. The elongation to failure (EF) was initially calculated from the chart, and
then always compared with the deformed tensile specimen and corrected if necessary.
The EF and the true stress were obtained by the following equations:
EF(%) = — - 1 0 0 = £ ^ • 1 0 0 (3.1)
L . j _ . r l+ ™ ) (3.2)
A(in2) 145 A0(Jn2) 145 ^ 100 J
53
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
where L0 is the initial gauge length, L is the instantaneous gauge length after
deformation, P is the instantaneous applied load and A0 is the initial cross section o f the
gauge area. The yield stress (YS) and the ultimate tensile stress (UTS) were taken as the
0 .2 % offset stress which produce a small amount of plastic deformation and the
maximum stress in the stress versus strain curve, respectively.
3.4.2 Microhardness testing
The specimens for microhardness testing were mounted on a low curing
temperature epoxy resin in order to prevent any microstructural changes from the latent
heat during curing. They were polished using SiC abrasive papers through 180, 240, 400,
600, 2400 and 4000 mesh size and the final polishing was conducted by using 1 p.m MgO
powder on a polishing cloth in order to get a mirror-like surface.
Microhardness measurements were conducted using a digital microhardness tester
FM -le from Future-Tech Corporation. The value of Hv was calculated according to the
following equation from the test load at the time when the test surface was indented and
the surface area was obtained from the lengths o f diagonals of the indentation by using a
diamond indenter in the form of a right pyramid with a square base having an angle
between the opposite faces of 136°:
Hv = 0.1891— (3.3)
d h~
where Hv is Vickers hardness with no units, F is the test load (N) and d is the mean value
of the diagonals of the indentation in mm. Various kinds of load, such as 50, 100, 200
54
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
and 500 g, were selected separately for each material system, but a fixed dwelling time of
15 s was applied for every measurement.
At least nine separate measurements were taken for each specimen, the maximum
and the minimum values of the Vickers microhardness were removed and then the mean
value and the standard deviation were obtained from seven measurements for each
specimen. The mean value and the standard deviation were calculated as follows:
where N is the number of the hardness measurements for each specimen.
3.5 MICROSTRUCTURAL OBSERVATION
3.5.1 Optical microscopy
An Olympus AH-3 research photomicrographic microscope was used to observe
the microstructure in order to measure the initial grain size before ECA pressing. As in
the microhardness tests, the specimens were mounted on a low curing temperature epoxy
resin in order to prevent any microstructural changes from the latent heat during curing.
They were polished using SiC abrasive papers through 180, 240, 400, 600, 2400 and
4000 mesh size and the final polishing was conducted by using I pirn MgO powder on a
polishing cloth to a mirror-like surface.
(3-4)
(3-5)
i = i
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
A . number of etchants for microexamination were tried. For Al-2024, AJ-Mg-Sc-
Zr and SupraI-100 alloys, only two etchants were successful [119]. The Keller’s reagent
having a composition of 2 ml HF (48%), 3 ml HC1 (cone), 5 ml HNO3 (cone) and 190 ml
H 2 O was applied on the surface of samples for 8 to 15 s at room temperature and then
removed in a stream of warm water. The Graff/Sargent reagent having a composition of
15.5 ml HNO3 (cone), 0.5 ml HF (48%), 3.0 g CrC>3 and 84.0 ml H 2 O was used for 20 to
60 s. Consecutive etching by using the above two etchants was often conducted in order
to further develop the grain boundary lines.
In order to measure the grain size o f a material, the linear intercept method was
used [119]: The mean grain intercept, J , was obtained from the average distance between
grain boundaries along a line laid down on a photomicrograph. These measurements were
repeated several times and the straight lines were placed on the photograph in a random
fashion. The mean intercept length is defined as:
] = - _L = _ ^ Z _ (3.6)
N t P„ -M
where A /- , is the average number of grain boundaries intercepted, L t is the total line
length over which the measurement was made, Pm is the total number of intersections and
M is the magnification in the photograph.
3.5.2 Transmission electron microscopy
Samples for transmission electron microscopy (TEM) were prepared by the
following procedures:
56
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1) Thin slices of the bulk specimen with a thickness of 2 mm were cut by using a micro
cutting machine Buehler ISOMET 1000.
2) Both sides of the thin slices were polished using SiC abrasive papers from 180 to 600
mesh size and the final thickness was in the range of 1 0 0 to 2 0 0 |j.m.
3) Several disks with a diameter of 3 mm were trimmed.
4) A double-jet electropolishing facility by Struers TENUPOL was finally used for
thinning the samples. The samples were located exactly at the center of the polishing
cell and thinned on both sides simultaneously by pumping the polishing solutions to
the center of the samples. Polishing stopped automatically after a very small hole was
detected in the center o f the samples by the attached optical sensor. The conditions of
electropolishing were as under:
For pure A 1 and dilute aluminum alloys, the solution consisted o f 10% HCIO4 , 20%
C3 H8 O3 and 70% C2 H5 OH. This solution was kept either in the tank surrounded by
iced water or at room temperature. Initially, a voltage of 15 V was applied for about 1
minute to remove the surface oxide and later this was increased up to 5 to 6 V to
polish the surface of the samples. It should be noted that the higher the
electropolishing temperature, the faster the polishing rate.
For Al-2024 and Supral-100 alloys, a solution with 20% HNO3 and 80% CH3OH was
kept at 253 K. This was done by pouring liquid nitrogen in the outer container and the
applied voltage and duration was then almost the same as for pure aluminum.
For Al-1420 alloy, a solution with 10% HclC> 4 , 20% C3H8O3 and 70% C2H5OH was
used at 278 K and a voltage of 8 V.
57
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5) Finally, the TEM samples were washed using pure methanol and stored safely.
The TEM observations were conducted using two microscopes: a Philips EM 420
operating at an accelerating voltage of 120 KV and an Hitachi H-8100 TEM operating at
200 KV. The structural parameters, such as grain size and subgrain size, were estimated
using the linear intercept technique. And also selected area electron diffraction (SAED)
patterns were used in order to obtain information on grain boundary orientations.
3.5.3 Scanning ion microscopy
After the tensile testing, microstructures near the fracture tips of selected samples
were examined by scanning ion microscopy (SIM). The samples for SIM observation
were prepared by using a focused ion beam (FIB) facility (Hitachi FB-2000) in which a
very small area of the specimen (typically — 30 x 40 pm2 ) is thinned with a Ga ion beam
oriented parallel to the specimen surface. The microstructure within this region was then
examined using an SIM attachment within the FIB apparatus.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4. FUNDAMENTAL FACTORS IN ECA PRESSING
4.1 STRUCTURE OF ECA PRESSED BILLETS
4.1.1 Introduction
It has been generally accepted that ECA pressing can develop uniform structure
and properties throughout the deformed materials [91]. In actual pressing, however, this
may not be true because the ideal simple shear cannot be performed completely during
deformation. It has not been investigated whether the microstructure of materials after
ECA pressing is homogeneous over the whole area of the pressed samples. This is very
important because often only a very small portion of pressed billets can be used for
investigations, such as optical microscope observation, TEM, hardness and tensile tests.
If an ECA pressing does not give the uniform microstructure throughout the samples, a
very small section of samples cannot be representative of the whole area of samples and
any experimental data may not be reliable. In order to investigate this problem, a pure
aluminum was selected as a model material and microhardness tests were conducted.
4.1.2 Test material and experimental procedures
Microhardness measurements can be affected by a number of microstructural
variations, such as impurities, second phases, precipitates, grain size and grain boundary
characteristics. For alloys, especially aluminum alloys, all o f the above variables may be
involved in the resultant hardness. For pure metals, however, microhardness only
depends on the grain size and grain boundary characteristics because there are no second
59
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
phases and precipitates. In order to remove any microstructural complexity, therefore,
high purity aluminum (99.99%) was chosen.
Before ECA pressing, samples were cut along the rolling direction and annealed
at 773 K for I hour in order to give a uniform microstructure over the whole area and to
remove the effect of machining. These samples were then pressed at room temperature
using route Bc and the selected numbers of passes were 1 , 2 , 4, 8 and 1 2 .
It is very convenient to investigate the pressed billets by designating three
orthogonal planes as shown in Fig. 4.1 (a) [100,108]; The X plane denotes the plane
perpendicular to the longitudinal axis of the pressed billets, the Y and Z planes are the
planes parallel to the side faces or to the top face at the point of exit from the die,
respectively. In this investigation, after ECA pressing, the X plane was chosen for all
pressed samples and the detailed sectioning is described in Fig. 4.1 (b). In Fig. 4.1 (b), the
pressed billet is schematically depicted and the billet was cut at uniform distances. The
plane number © represents the head of the pressed billet and the plane number ©
represents the tail of the pressed billet. After polishing, the Vickers microhardness, Hv,
was measured under a load of 1 0 0 g applied for 15 seconds for all samples.
4.1.3 Distribution of microhardness along the pressed billets
All hardness measurements were conducted 9 times at regularly positioned points
for every plane, as shown in Fig.4.1 (b), and the results are summarized in Table 4.1 and
are also represented graphically in Fig. 4.2.
60
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
♦
plunger
sample
die
(a)
pressed sample
Pressing
© © © @ ® © ©
(b)
Figure 4.1 (a) Schematic illustration o f the three planes designated X, Y and Z [108].
(b) Schematic illustration o f the sectioning and the location of
microhardness indentations.
61
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 4.1 Microhardness values o f pure aluminum ECA pressed at 298 K via route Bc
1
Number
o f plane
Microhardness (Hv)
"1" pass "2" passes "4" passes "8" passes "12" passes
42.8 ±3.2 46.5 ± 3.2 45.4 ±2.2 41.5 ± 1.7 38.5 ± 1.3
41.9 ± 4.4 47.4 ± 2.7 44.7 ±2.9 41.6 ± 1.3 37.8 ±0.9
■
41.0 ±3.2 47.1 ±2.1 43.7 ±2.2 41.2 ± 1.8 38.0 ±0.9
41.2 ± 2.2 44.2 ± 1.8 44.0 ± 1.5 40.3 ± 1.3 38.2 ±1.3
®
40.3 ±3.8 44.4 ±3.3 45.2 ± 1.9 40.2 ± 1.6 38.5 ±1.1
i
©
39.4 ±3.6 46.4 ± 1.8 43.3 ± 1.7 41.4 ± 1.0
1
38.8 ±1.3
© 39.2 ±3.0 44.4 ± 2.6 44.7 ±3.6 40.4 ± 1.3 38.6 ± 2.1
j
Average 1 40.8 ±3.3
1
45.8 ±2.5
1
44.4 ±2.3 | 40.9 ± 1.4 38.3 ±1.3 I
... i
62
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Microhardness (Hv)
48
46
44
42
40
38
36
1 pass
2 passes
4 passes
8 passes
12 passes
34
High purity Al
X p lan e
ECA p re s s e d at 2 98 K via route B(
32
30
2 1 3 4 5 7 6
Number of plane
H ead of billet Tail of billet
Figure 4.2 Variation o f microhardness with the number of plane for pure aluminum ECA
pressed at 298 K via route Be-
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In Fig. 4.2, all error-bars are removed in order to give a clear view because some
o f the error-bars at lower passes are very large. For lower passes, such as 1 and 2 passes,
the variations of microhardness are large throughout the length of the pressed billets,
therefore it is revealed that the pressed billets at lower number of passes do not possess a
uniform microstructure over the whoie area. However, as the number of pressings is
increased, the variations in the hardness get progressively smaller. Very uniform
micro structure is obtained throughout the length of the pressed billet at 12 passes. This
can be also confirmed by the fact that the magnitude o f the error-bar decreases with
increasing in the number of pressings in Table 4.1.
The hardness for the head of the billet is often higher than that for the tail of the
billet and this may be related to the cutting position of the samples. The billet receives a
large compressive force at both ends in addition to the shearing deformation during
pressing and this force is exerted by the pre-existed another billet inside the die and a
plunger. After 1 pressing, as shown in Fig 4.1 (b), the cutting position o f the head of the
billet is close to the front end of the sample but the cutting position of the tail of the billet
is comparatively far to the rear end of the sample. If the shearing deformation is applied
uniformly throughout the billet during pressing, the difference of cutting position can
give the variation of the amount of applied compressive force between two positions.
It was reported that the distribution of Hv values along the longitudinal direction
of a PM 2024 Al-3% Fe- 5% Ni specimen after ECA pressing at 573 K up to 3 pressings
was uniform [120]. In order achieve a uniform microstructure, therefore, higher pressing
64
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
temperature is also desirable due to a diminishing in the level of the applied compressive
force.
4.1.4 Variation of microhardness with the number of pressings
As shown in Table 4.1, the average hardness on all planes is increased by almost
60% after only 1 pressing (before pressing, Hv = 25.4 ± 0.4). The hardness reaches a
maximum value between 2 and 4 passes, and then, the hardness decreases continuously
up to 12 passes. The same result was reported for pure aluminum ECA pressed via routes
A and C [99],
The Hall-Petch relationship is a very useful tool to relate the grain size, d, to the
yield stress of a polycrystalline material, ay, and can be given by [121,122]:
where cr0 is the friction stress and ky is a constant of yielding. This relationship is well
established for grain sizes larger than 1 pm and it demonstrates that the yield stress
increases as the grain size is decreased. In the absence o f appreciable work hardening, the
Vickers hardness, Hv, is proportional to the yield stress through the expression Hv = 3 < j y
[123], and the above equation may be rewritten as [124-126]:
where H 0 and kn are appropriate constants associated with the hardness measurements.
In this investigation, this H-P relationship can be directly applicable because grain
boundaries are the only obstacle to dislocation movement for pure aluminum and the
(4.1)
Hv = H a + kHd~w2 (4.2)
65
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
grain size after ECA pressing was — 1.0-1.3 (im [99,100], Even for severely deformed
aluminum alloys, it was proven that the H-P relationship was valid within a limited range
o f grain sizes [124-127]. When pure aluminum was subjected to ECA pressing via routes
A and C, it was reported that the grain size or the subgrain size decreased up to 3 passes,
and then, it saturated or slightly increased up to 10 passes [99]. The same trend was also
found for route Be [100],
As shown in Table 4.1, therefore, the increase of hardness up to 3 passes is the
direct result of the H-P relationship. After 3 passes, if the grain size is slightly increased
as reported earlier [99,100], the H-P relationship still can be applicable. However, it still
remains in question whether further severe plastic deformation can make the grain size
larger or not. The deformation mechanism of ECA pressing proposed earlier does not
involve any grain growth mechanism during repetitive pressings [89]. On the other hand,
it may be reasonable to conclude that the grain refinement has almost stopped at 3 passes
and the microstructural evolution into high angle grain boundaries may play an important
role in the decrease of hardness if the grain size is saturated. Since the hardness is a
measure o f the resistance to penetration under load by the indenter, it may be assumed
that a material having high angle grain boundaries shows lower hardness than the
counterpart due to the easy movement of dislocations when chemical composition and
grain size are the same for the two materials. This can be confirmed by the fact that
repetitive pressings produce a microstructural evolution from low angle grain boundaries
into high angle grain boundaries [99,100].
6 6
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The magnitude of the error-bars on the average values of microhardness in Table
4.1 should be noted. At lower passes, such as 1, 2 and 4 passes, the range o f error-bar is
comparatively large and this becomes smaller with an increase in the number of passes.
This illustrates that the microstructure at lower numbers of pressings is heterogeneous.
This is consistent with TEM investigations showing a mixed microstructure of equiaxed
grains having high angle grain boundaries and subgrains having low angle boundaries for
Al-3% Mg alloy [125] and Al-1420 alloy [127] ECA pressed via route Be up to 4 passes
at 298 K and 673 K, respectively. At higher passes, however, the microstructure is
homogeneous and this indicates that most subgrains evolve into equiaxed grains having
high angle grain boundaries. Another evidence is the shape of the indentation. At lower
passes, some of the observed indentations showed wavy and twisted shapes. At higher
passes, such as 8 and 12 passes, however, most indentations were very close to the right
pyramidal shape and the wavy and twisted indentations almost disappeared.
4.2 SHEARING CHARACTERISTICS ASSOCIATED WITH ECA PRESSING
4.2.1 Introduction
The variety of different pressing routes permits an opportunity to change the
shearing characteristics, such as the shearing planes and shear directions, within the
material during ECA pressing. By choosing the proper shearing characteristics, it is
possible to control microstructure during ECA pressing. In order to understand the nature
67
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
o f shearing, the approach by Furukawa et al. [106] has been adopted as a basic tool and
will be used for further discussion in the following sections.
4.2.2 The pressing routes A, BA , Be and C
The assumptions for Furukawa’s analysis [106] were made to simplify the ECA
pressing conditions, such as a die having O = 90° and = 0°, an ideal frictionless
pressing condition and shearing of a cubic element within the test sample. A single
pressing through the die shears the cubic element into a rhombohedral shape
[106,107,128], In subsequent pressings, therefore, the shearing can be modified by
rotation o f the sample.
In route A, the multiple pressings continuously increase the distortion of the
rhombohedron and the original cubic element is not recovered [106,107,128], This is also
the same in the case of route BA - The difference between these two routes is that there is
no deformation in the Z plane when using route A. However, it is evident that the grains
become more elongated when using these two routes. In terms of superplasticity,
processing by using these two routes may not be desirable because the grains should be
equiaxed and the formation of texture should be suppressed.
The shearings up to four pressings with rotation of 180° (route C) and with
rotation of 90° in the same direction between each pressing (route Be) are illustrated in
Fig. 4.3 and Fig. 4.4, respectively. In both figures, the shaded plane at the intersection of
two channels represents the theoretical shear plane and each shear plane after pressings is
also depicted inside the elements [106,107,128],
68
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ute q
m
k m
r° u te ra riciIh
c ^ % r ° ° o fs/lean.
8 U P to
four
PressL
ln§s thr.
OUsh^ < i h
W a
ofthe
°°Pyri(
m
°Wner
Furthf
erre^ c , ion
69
Pr°hibitecj
m 0ut
Perrbission
1 / f e a
l§Ure 4 4
r ^ c ? ; 0 ^ o n
' § U P to
four
Press,
lnS$ thn
° uSh the
die
via
ofthe
c°Pyright
°Wner
Furtherra
reProdl
70
uotion
Pr°hib,
iteci
W th,
out
Perm is*L
ISsion,
In route C, as shown in Fig. 4.3, the cubic element is restored after In pressings,
where n is an integer, and each grain contains a single shear plane but shearing directions
are reversed for each pressing. In route Be, as shown in Fig. 4.4, the cubic element is
restored after 4n pressings, two shearing planes are involved up to four pressings and
shearing directions are reversed after 2n pressings.
A . pictorial form is very useful because the shearing associated with ECA pressing
can be simply shown by changing the shape of the initial cubic element [106], In Fig. 4.5,
the initial square sections are shown on the left and subsequent distortions are illustrated
for each plane of sectioning from I to 8 pressings for the pressing routes C and Be. It is
evident that there is a restoration of the initial cubic element for routes C and Be and no
deformation in the Z plane for route C.
4.2.3 New pressing routes C-Bc and Bc-C
Both routes Be and C can make materials superplastic because route Be is the
optimum processing route for pure A 1 in order to most rapidly achieve a homogeneous
microstructure of equiaxed grains separated by boundaries having high angles of
misorientation [100,102] whereas route C restores a cubic element every second pass.
Regardless o f the fast recovery of the initial cubic element in route C, however, there is
no deformation in the Z plane within the deformed sample as shown in Fig. 4.5.
Therefore, two new combination routes, C-Bc and Bc-C, were designed and their
schematic illustrations of shearing are shown in Fig. 4.6 and Fig. 4.7, respectively. Their
pictorial forms are also added in Fig. 4.5.
71
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Route Plane
Number of pressings
0 1 2 3 4 5 6 7 8
Bc
X
□
□
0
□ □
n n
0 0 □
Y
□
□ □ □ □
Z
□
□ □ □ □
C
&
C-Bc
X
□
CD
□
□
□
i= i
□
□
□
Y
□ □ □ □ □
Z
□ □ □ □ □ □ □ □ □
B c C
X
□
c n
0 □ =
nn
□ □
Y
□ □
—
□ □
Z
□ □ □ □ □
Figure 4.5 Pictorial compilation of the shearing associated with four different processing
routes.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
O e o ° e S S / n 9 s
m m m
( i 8 o % S S 'n 9 s
on)
,0" 0f ^ear,-
four
Pres,
SinSsthn
' O u sh o ,e d . e
Wa
°^/je
C O p yright
°wner
Fm ,
73
er
reProat
uction
Pr°hib,
ited W h
Without
Pern'*Si0„ ,
ofthe
U P to
ei8ht
Pressj,
lnSs thri
°Ush0>edie
b r ig h t
°Wner
Fu rth ,
er
74
reProat
ucti0r)
Prohibite d l ■
d w'thout
Perm ,
,ission.
The pressing sequence of new combination route C-Bc is shown in Fig. 4.6, and
the initial cubic element is restored after 2n pressings and this is the same as route C. This
is further confirmed in Fig. 4.5 that the shearing characteristics for route C-Bc are
identical as those for route C and there is also no shearing deformation in the Z plane.
However, the initial Y plane at 1 and 2 pressings is changed to Z plane at 3 and 4
pressings and the initial Z plane at 1 and 2 pressings is also changed to Y plane at 3 and 4
pressings due to the 90° rotation o f the sample after 2 pressings. Therefore, it can be
assumed that there is a shearing deformation for every plane. The mixed rotation further
develops additional shearing planes. When Fig. 4.3 and Fig. 4.6 are compared, it is clear
that two additional shearing planes are working at 3 & 4 pressings in the case of route C-
Bc whereas there is only one shearing plane in route C. Therefore, when considering
shearing planes, the initial cubic element is fully recovered after 4n pressings. After 4
pressings, the same shearing procedures are repeated for the next passage.
Another new combination route Bc-C develops totally different shearing
characteristics compared with route C-Bc and its pressing sequence is illustrated in Fig.
4.7. The initial cubic element is restored completely after a total of 8n pressings and this
is the same as the previously reported route Bc-A [106]. Inspection of Fig. 4.7 shows that
there are six different shearing planes up to 8 pressings and these shearing planes have
their own shearing directions. Shearing planes at 1 and 2 pressings are the same as those
at 7 and 8 pressings. Furthermore, considerable shear deformation occurs in every plane
as shown in Fig. 4.5. Some interesting facts are that shearing characteristics of 7 and 8
pressings for route Bc-C are the same as those of 3 and 4 pressings for route Be and some
75
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
symmetries in every plane are found for route Bc-C in Fig. 4.6 whereas route Bc-A
shows limited symmetry in the shearing diagram.
4.2.4 Effects of pressing routes for pure aluminum
Samples of pure A 1 were ECA pressed at 298 K using six different routes, namely
routes A, BA , Be, C, C-Bc and Bc-C up to 12 passes (samples were pressed up to 20
passes for routes A and Be). All pressed samples were oriented in mutually perpendicular
planes, such as the X, Y, and Z planes, and they were cut using a precision saw machine.
Microhardness measurements were conducted for all samples and the same indentation
condition was used as shown in section 4.1.2. All data are summarized in Table 4.2.
The hardness is increased greatly by almost 60% only after the initial pressing for
all pressing routes, and the hardness for most pressing routes reaches a maximum value at
2 to 4 pressings. After that, the values o f the hardness decrease continuously up to 12
pressings. Furthermore, a saturation in hardness occurs after 20 pressings for route A but
the hardness decreases slightly up to 20 passes for route Be- These trends are clearly
shown in Fig. 4.8 where the X plane is selected.
An inspection of Fig. 4.8 shows that the values of microhardness for routes A C
and C-Bc are lower than those for routes BA , Be and Bc-C after 2 pressings. For large
numbers o f passes, as with 12 pressings, the values of hardness for routes Be and C are
higher than those for routes A and BA , and the two combination routes show intermediate
values of hardness. However, when samples were pressed up to 20 passes by using routes
A and Be, the difference in hardness between the two routes almost disappears.
76
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 4.2 Hardness test results for pure aluminum ECA pressed using six different routes.
Pressing
routes
Number of
pressings
Microhardness (Hv)
X plane Y plane Z plane
Unpressed - 25.4 ± 0 .4 25.6 ±0.5 25.6 ±0.5
Route A
1 39.7 ± 1.8 41.4 ±4.1 40.0 ± 1.8
2 43.1 ± 1.3 43.4 ±2.0 42.6 ±1.3
4 41.9 ± 0.7 41.0 ±0.8 38.6 ± l.l
8 37.5 ± 1.3 36.6 ±0.9 39.6 ± 1.7
12 36.6 ± 1.3 38.1 ± 1.8 37.7 ± 1.9
20 36.6 ± 0.6 36.3 ±0.4 36.8 ±0.7
Route Ba
1 39.7 ± 1.8 41.4 ±4.1 40.0 ± 1.8
2 46.3 ± 1.3 47.0 ±0.8 45.3 ±2.0
4 43.1 ± 3 .4 44.5 ± 4.7 41.9 ± 1.4
8 38.6 ± 1.4 39.4 ±0.5 38.2 ±0.3
12
36.9± 1.1 37.0 ±0.5 37.8 ±0.6
Route Be
1 39.7 ± 1.8 41.4 ± 4.1 40.0 ± 1.8
2 46.3 ± 1.3 47.0 ±0.8 45.3 ±2.0
4 43.0 ± 1.4 43.1 ± 1.4 43.4 ± 1.2
8 41.1 ±0.9 40.5 ±0.6 40.0 ± 1.0
12 39.9 ± 1.2 38.6 ±0.7 39.0 ±0.7
20 36.7 ± 0.7 36.9 ±0.4 36.9 ±0.3
Route C
1 39.7 ± 1.8 41.4 ±4.1 40.0 ± 1.8
2 43.7 ± 1.5 45.4 ± 1.6 43.8 ± l.l
4 42.5 ± 1.4 43.7 ±0.7 43.5 ± 1.7
8 41.3 ± 1.7 41.6 ± 1.1 40.3 ± 1.3
12 39.7 ± 1.5 38.6 ±0.7 39.0 ±0.5
Route C-Bc
1 39.7 ± 1.8 41.4 ±4.1 40.0 ± 1.8
2 43.7 ± 1.5 45.4 ± 1.6 43.8 ± 1.1
4 45.1 ± 1.2 44.9 ± 1.4 43.8 ±0.5
8 39.8 ± 1.1 40.3 ±2.2 40.9 ± 1.5
12 38.6 ±0.9 38.3 ±0.7 38.7 ± l.l
Route Bc-C
1 39.7 ± 1.8 41.4 ±4.1 40.0 ± 1.8
2 46.3 ± 1.3 47.0 ± 0.8 45.3 ±2.0
4 43.4 ± 1.3 44.4 ± 1.1 44.2 ± 1.6
8 40.8 ± 1.3 40.2 ± 1.4 39.2 ±1.0
12 38.3 ± 0.4 38.9 ±0.7 38.9 ±0.9
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Microhardness (Hv)
50
High purity Al
X plane
45
40
Pressing routes
30
20
20 5 15 0 10
Number of pressings
Figure 4.8 Variation o f micro hardness with the number of pressings for pure aluminum
ECA pressed at 298 K via six different routes.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The present results are very consistent with the data reported by Iwahashi el al.
[99] where the values of the hardness are slightly decreased as the number of pressings is
increased and the hardness for route C is slightly higher than that for route A for most
numbers of pressings. It was also reported that the grain size or the subgrain size
decreased up to 3 passes, and then it saturated and increased slightly up to 10 passes for
route C and route A, respectively [99]. From the above report, it appears that the decrease
in hardness is a direct result of the H-P relationship for every route if the grain size is
increased as the number o f pressings is increased after 3 pressings. However, the above
explanation seems unlikely because it is not reasonable to increase the grain size with the
numbers of pressings.
Another alternative explanation may be as follows. After the initial pressing, a
large number of lattice dislocations created inside the grains cannot rearrange into perfect
LEDS i.e. dislocation cell boundaries. Up to 2 to 3 pressings, depending on the pressing
routes, these incomplete boundaries evolve into perfect LEDS i.e. subgrain boundaries.
Further pressing leads only to an increase in misorientation angles of the subgrain
boundaries. Therefore, the grain size is saturated after the initial pass, the hardness is
increased as the number o f pressings is increased up to 2 to 3 pressings, and the hardness
is decreased as the number of pressings is increased after 3 pressings due to the easy
movement of dislocation along high angle grain boundaries.
It is evident that the difference in microhardness for the different pressing routes
is caused by the different shearing characteristics and the resultant microstructural
development during ECA pressing. It was reported that the predictions of shearing
79
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
patterns for routes A, Be and C are very consistent with the microstructural observations
for pure Al pressed after 2 pressings [107]. However, the relationship between the
shearing characteristics and the resultant microstructure may not agree well at high
numbers of pressings. For example, it is believed that route A produces more elongated
grains or texture as the number of pressings is increased, as predicted in Fig. 4.5, but it
was reported that a reasonably homogeneous microstructure with equiaxed grains was
obtained after 10 pressings via route A [99]. Another point is that the values of hardness
are almost the same for every plane after multiple pressings via route C in Table 4.2 but
the shearing pattern predicts no deformation in the Z plane in Fig. 4.5.
4.2.5 Effects of pressing routes for Al-2024 alloy
Samples of fully annealed Al-2024 alloy were ECA pressed at 298 K using three
different routes, namely route Be, C-Bc and Bc-C up to 8 passes (samples were pressed
up to 10 passes for route Be)- Further pressings cannot be conducted at room temperature
because of the occurrence o f an unsound surface and propagation of a crack. The Vickers
microhardness, Hv, was measured under a load of 500 g applied for 15 seconds. Other
experimental procedures are the same as before.
Representative results are shown in Fig. 4.9 where the X plane was measured and
similar results were obtained from the Y and Z planes. For every route, the hardness is
increased greatly by almost 30% after the initial pressing. After that, the degree of the
increase becomes smaller as the number of pressings is increased but the hardness
increases continuously up to 8 or 10 pressings without any saturation.
80
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Microhardness (Hv)
130
Ai-2024
120
X p lan e
1 1 0
100
90
80
Pressing routes
70
60
0 2 6 4 8 10
Number of pressings
Figure 4.9 Variation o f microhardness with the number of pressings for Al-2024 alloy
ECA pressed at 298 K via three different routes.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In Fig. 4.9, the hardness for route Be is always higher than the two combination
routes, and the hardness for route Bc-C is higher than that for route C-Bc. This is similar
to the previous results for pure aluminum.
The microstructure of the aluminum alloy is fundamentally different from that for
pure aluminum because second phases (i.e. solid solutions) or precipitates are always
involved in aluminum alloys, so it is evident that the source o f grain refinement is much
more than that for the pure material. For Al-2024 alloy, precipitates such as CuAk and
CuMgAk exist inside the matrix or along the grain boundaries [117,119], Up to a high
number of pressings, therefore, grain refinement is very active and the evolution into high
angle grain boundaries is in part suppressed by second phases or precipitates. This is
very consistent with the fact that the grain sizes of pure A 1 [100], Al-3% Mg alloy [125]
and Al-Mg-Li-Zr alloy [127] after EC A pressing by the same amount of strain are 1.3,
0.3 and 1.2 pm, respectively, although the initial grain sizes are different (the last alloy
was pressed at 673 K). This may be the reason for the increase in hardness up to 8 or 10
pressings for the Al-2024 alloy.
The effect of pressing routes on the microstructure and the superplastic properties
of Al-2024 alloy will be presented and discussed in Chapter 5.
82
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5. SUPERPLASTICITY IN ALUMINUM ALLOYS
5.1 Al-Sc-Zr AND AI-Mg-Sc-Zr ALLOYS
5.1.1 Introduction
Major requirements for superplasticity are that the grain size should be small and
the small grains should be stable at high temperatures. It is known that submicrometer-
grains can be produced using the technique of ECA pressing and the addition of Zr or Sc
can stabilize the small grains up to high temperatures [89], In fact, this investigation is
the latest development o f earlier research on the effect of the addition of scandium
[82,83,89,129,130].
Therefore, the present investigation was conducted to evaluate the potential for
using ECA pressing to achieve superplastic ductilities in pure aluminum and in an Al-3%
Mg alloy through the introduction of small additions o f scandium and zirconium at the
same time.
5.1.2 Materials and experimental procedures
The Al-0.2% Sc-0.12% Zr and the Al-3% Mg-0.2% Sc-0.12% Zr alloy were both
pressed at 298 K via route Be up to 8 passes and 6 passes, respectively. Since it was
found that the second alloy tended to break after more than ~6 pressings, the 6 pressings
were chosen. After pressing, samples were cut into tensile specimens having a gauge
length of 5 mm (normally the gauge length is 4 mm) in order to compare with the
previous data.
83
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5.1.3 Results on Al-0.2% Sc-0.12% Zr alloy
All tensile testing results are summarized in Table 5.1. In addition, typical stress-
strain curves for samples tested at 773 K are shown in Fig. 5.1. A maximum elongation to
failure of — 590% was found when testing at 773 K using a strain rate of 10'3 s‘l after ECA
pressing at 298 K via route Be up to 8 pressings. When samples were tested at 773 K,
furthermore, elongations of more than 300 % were obtained over a wide range o f strain
rates from 3.3 x 10'2 to 10-4 s'1 . This result demonstrates that a small amount of Sc and Zr
can be very beneficial to superplastic properties of pure aluminum. After a short strain
hardening, as shown in Fig. 5.1, extensive strain softening up to failure occurs for every
sample, even when testing at different temperatures and strain rates.
5.1.4 Results on Al-3% Mg-0.2% Sc-0.12% Zr alloy
Fig. 5.2 shows microstructures at low and high magnifications for a sample ECA
pressed for 6 passes at 298 K. Inspection of Fig. 5.2 shows that the grains are reasonably
equiaxed and the average grain size is -0.3 pm. It was also found that many of the grain
boundaries are poorly delineated and the presence o f irregular extinction contours along
the boundaries or inside the grains suggests evidence for significant strain within the
sample. The selected area electron diffraction (SAED) pattern with an aperture size of
12.3 p.m consists o f rings of diffraction spots showing that the grain boundaries have high
angles of misorientation.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 5.1 The superplastic properties of Al-Sc-Zr alloy ECA pressed at 298 K
Temp. Strain rate YS UTS E F
(K ) (sec'1 ) (MPa) (MPa)
(% )
623 1 62 9 4 76
623 3 .3 x 1 0 '1 66 90 91
623 1 0 '1 59 73 109
623 3.3 x 10'2 53 68 144
623 10'2 39 50 235
623 3.3 x 10'3 23 36 240
623 I O'3 21 28 174
623
3.3 x IO"4
14 17 258
623 IO"4 11 14 133
673 1 43 74 91
673 3.3 x 1 0 '1 51 68 111
673 1 0 '1 34 51 228
673
3.3 x 10'2
30 42 232
673 10'2 17 28 223
673 3.3 x 10'3 13 20 399
673 10‘3 8 12 259
673 3.3 x lCT4 7 11 220
673 IO -4 6 8 206
698 10*4 3 5 228
723 1 36 55 105
723 3.3 x 10'1 33 49 163
723 icrl 29 39 148
723 3.3 x 10'2 20 30 204
723 10'2 11 18 500
723 3.3 x 10'3 8 13 344
723 10'3 5 8 352
723 3.3 x IO -4 4 7 230
723 lO"4 3 5 355
748 i o -4 2 4 281
773 1 37 46 84
773
3.3 x IO '1 25 39 127
773 IO '1 19 29 279
773 3.3 x 10'2 12 21 349
773 10'2 8 13 351
773 3.3 x 10'3 5 9 495
773 10'3 4 7 585
773 3.3 x IO-4 3 5 378
773 io-4 2 3 306
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
a (MPa)
l —i 1 —i —j —i
AI-0-2% Sc-0.12% Zr
T = 773 K
3.3 x 10
10'1
3.3 x 10
3.3 x 10
3.3 x 10
200 400 500 600 700
S (%)
Figure 5.1 The stress vs. strain curves of Al-0.2% Sc-0.12% Zr alloy ECA pressed
at 298 K via route Be up to 8 passes.
8 6
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.2 Microstructures o f Al-3% Mg-0.2% Sc-0.12% Zr alloy after ECA pressing at
(a) low magnification and (b) high magnification.
87
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Ail tensile testing results are summarized in Table 5.2 and the typical stress-strain
curves for samples tested at 773 K are also shown in Fig. 5.3. The maximum elongation
to failure of — 1680% was recorded when testing at 773 K using a strain rate of 10' s'
after ECA pressing at 298 K via route Be up to 6 pressings. Elongations of more than
1000 % were obtained over a wide range of temperatures from 673 to 798 K and the
largest ductilities were often found near 10'2 s'1 . It is clearly shown in Fig. 5.3 that the
overall stress vs. strain curves are very similar to those in Fig. 5.1 but well developed
steady state regions exist, where the strain hardening and the strain softening are
balanced, when the elongations to failure exceed -1000%.
If the optimum superplastic condition is defined as the combination of testing
temperature and initial strain rate where the maximum elongation of failure is obtained,
the optimum superplastic condition for Al-3% Mg- 0.2% Sc-0.12% Zr alloy ECA pressed
at 298 K via route Be up to 6 passes can be determined as a testing temperature of 773 K
and an initial strain rate of 10'2 s'1 . In addition, Fig. 5.4 shows the appearance of the
tensile specimens tested to failure at 773 K and at various strain rates. It is apparent that
there is uniform deformation within the gauge length without any localized necking
especially for highly elongated specimens.
Fig. 5.5 and Fig. 5.6 show the microstructures of samples, after tensile testing to
failure where, for each condition, (a) is a TEM image taken from within the undeformed
grip region of the sample and (b) is an SIM image taken from the fracture tip. For each of
the SIM images, the surface of the sample lies close to the right hand edge o f the
photomicrograph and the direction of the tensile axis is indicated.
88
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 5.2 The superplastic properties of Al-Mg-Sc-Zr alloy ECA pressed at 298 K
Temp.
(K)
Strain rate
(sec'1 )
YS
(MPa)
UTS
(MPa)
EF
(%)
573 IO'1 104 156 219
573 3.3 x 10'2
58 110 494
573 IO'2 40 79 561
573 3.3 x IO *3 25 49 841
573 10'3 16 30 757
573 io -4 1 1 15 449
623 IO'1 60 98 312
623 3.3 x IO’2 33 62 472
623 10'2 24 38 578
623 3.3 x 10'3 16 29 666
623 10’ 3 12 20 574
623 io - 4 8 9 287
673 10'* 24 54 616
673 3.3 x 10'2 15 33 1235
673 IO’2 10 20 1397
673
3.3 x 10'3
7 13 975
673 10'3 5 10 489
673 io -4 3 6 351
723 10'1 23 44 612
723
3.3 x 10'2 12 28 918
723 IO * 2 7 13 1419
723
3.3 x 10'3
5 10 957
723 10'3 3 6 621
723 io - 4
2
3 219
748 I O '2 7 14 1418
773 3.3 x 10'1 25 45 358
773 IO'1 14 31 728
773 3.3 x 10'2 8 19 1228
773 10'2 5 11 1682
773
6.6 x 10'3
5 9 1444
773
3.3 x 10'3
3 7 1159
773 IO'3 2 4 774
773
3.3 x IO * 4 1 3 675
773 io - 4 1 2 427
798 10'2 5 10 1435
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
a (MPa)
At-3% Mq-0.2% Sc-0.12% Zr
200 400 600 800 1000 1200 1400 1600 1800
Figure 5.3 The stress vs. strain curves of Al-3% Mg-0.2% Sc-0.12% Zr alloy ECA
pressed at 298 K via route Be up to 6 passes.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Al-3%Mg-0.2%Sc-0.12%Zr
Initial sample
Figure 5.4 Examples o f tensile ductility in samples ECA pressed at 298 K via route Be up
to 6 passes and pulled at 773 K.
91
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.5 Microstructure of a sample prepared by ECA pressing and then tested to failure in tension at 773 K
with a strain rate of 10'2 s'1 : (a) grip region; (b) the fracture tip within the gauge length.
V O
w
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
edge surface
I n i
Figure 5.6 Microstructure o f a sample prepared by ECA pressing and then tested to failure in tension at 773 K
with a strain rate o f IO'4 s'1 : (a) grip region; (b) the fracture tip within the gauge length.
V O
u >
Fig. 5.5 was taken on the sample pressed at 298 K up to 6 passes and tested at 773
K at a strain rate o f 1 O'2 s'1 to an elongation o f -1680%. It is evident from Fig. 5.5 (a) that
the grain size has increased to — 2.6 pm in the grip region although the grains remain
essentially equiaxed. Near the fracture tip, shown in Fig. 5.5 (b), the grains are also
equiaxed with an average size o f -6.6 pm. Thus, as in normal superplasticity, grain
growth is enhanced during superplastic deformation within the gauge length. Since some
evidence of intergranular cracking is found within the sample, as shown in Fig. 5.5 (b), it
is believed that the fracture mechanism of Al-3% Mg-0.2% Sc-0.12% Zr alloy ECA
pressed at 298 K via route Be to a strain o f -6 is the nucleation, growth and coalescence
of internal cavities along the grain boundaries.
Fig. 5.6 was taken on the sample pressed at 298 K up to 6 passes and tested at 773
K at a strain rate of IO "4 s'1 to an elongation o f -430%. The grains are equiaxed both in
the grip region and in the gauge length with average sizes of -3.8 and — 18.5 pm,
respectively. There is extensive intergranular cracking near the left-hand edge o f Fig. 5.6
(b). In terms of grain size in the grip region, although the tensile test was conducted at a
very high temperature of 773 K for a long time of about 43000 sec, the grain size only
increased from -0.3 to -3.8 pm.
5.1.5 Discussion
5.1.5.1 The effect of addition of Sc and Zr on pure aluminum
It is known that an alloying addition by scandium or zirconium to aluminum
matrix produces a significant strength increment [61,78], This is mainly achieved by the
94
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
precipitation hardening mechanism where scandium or zirconium combines with A 1 to
form a stable coherent L I 2 phase, such as AI3 SC or Al3 Zr precipitates. These precipitates
are also very effective in stabilizing the microstructure mostly by pinning the grain
boundaries or subgrain boundaries at elevated temperatures. Of these two precipitates,
however, it has not been reported which one is more beneficial to strengthen the A 1
matrix and to stabilize the microstructure at high temperatures.
Fig. 5.7 shows the variations of the grain size and 0.2% proof stress with
annealing temperature for pure Al, Al-0.2% Zr alloy and Al-0.2% Sc alloy. The data of
grain size for pure Al and Al-0.2% Sc alloy were obtained from earlier reports [83,130],
the data of grain size for Al-0.2% Zr alloy from unpublished results [131] and the data of
0.2% proof stress from Berbon [89]. After annealing at selected temperatures for 1 hour,
it is found that there is a significant grain growth in pure Al whereas there is only very
limited grain growth in the Al-0.2% Sc alloy. Since the same amount of both Sc and Zr is
added to the aluminum matrix, it can be concluded that Sc is more effective in stabilizing
gains at high temperatures than Zr. Furthermore, the corresponding strength of Al-0.2%
Sc alloy is maintained after annealing at high temperatures up to 673 K but the strength
of Al-0.2% Zr alloy drops abruptly after annealing at 623 K for 1 hour.
Fig 5.8 shows the elongation to failure against the imposed strain rate for pure
aluminum, Al-0.2% Sc alloy and Al-0.2% Sc-0.12% Z r alloy. The first and second
materials are the same as the materials used in Fig. 5.7 [83,130] and were ECA pressed at
298 K via route Be up to 4 passes and 8 passes, respectively. The data for the Al-0.2%
Sc-0.12% Zr alloy were obtained from Table 5.1.
95
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Number of passes
Pure Al
Al-0.2% Zr alloy
Al-0.2% Sc alloy
102
E
ECA pressing at 298 K via route B ,
C D
180
160
C O
Q _
140
1 2 0
C O
C O
2
to
( 4 —
O
2
100
80
Q - 60
C M 40
O
20
200 300 400 500 600 700 800
Annealing temperature (K )
Figure 5.7 Variations of grain size and 0.2% proof stress with annealing temperature for
pure aluminum, Al-0.2% Zr alloy and Al-0.2% Sc alloy.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Elongation t o failure (%)
700
l iK l
673
673
alloy 673
alloy 773
• — Pure Al
v — Al-0.2% S c alloy
« — Al-0.2% Sc-0.12%
O - Al-0.2% Sc-0.12%
600
500
400
300
200
100
10-4 10'3 IO2 10‘1 10°
Strain rate (s'1 )
Figure 5.8 Elongation to failure vs. strain rate for pure aluminum, Al-0.2% Sc alloy and
Al-0.2% Sc-0.12% Zr alloy.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
If the elongations between pure Al and Al-Sc alloy tested at 673 K are compared,
as shown in Fig. 5.8, it is evident that the addition o f a small amount of Sc can give
improved superplasticity over all o f the strain rate range. The values of elongation for the
Al-Sc-Zr alloy tested at 673 K are superior to those for the Al-Sc binary alloy especially
at high strain rates and the maximum elongation to failure o f -590% was recorded when
testing at 773 K using a strain rate of 10'3 s"L after ECA pressing at 298 K via route Be up
to 8 pressings. Therefore, it is concluded that the addition of both elements at the same
time is more effective on ductility than the addition o f only one element.
S. 1.5.2 The effect of addition of Sc and Zr on the AI-Mg binary alloy
It is well known that the primary reason for Mg addition to aluminum is solid
solution strengthening by reducing the dislocation mobility. Solid solution strengthening
may be an effect in simple alloy systems but it was also reported that Mg additions can
change the subsequent microstructural evolution for severely deformed aluminum alloys
[101]. First, the number of pressings required to establish a homogeneous micro structure
increases with an increasing Mg content in aluminum matrix. Second, the ultimate
equiaxed equilibrium grain size attained by ECA pressing decreases with increasing Mg
content. For example, for pure Al, Al-1% Mg alloy and Al-3% Mg alloy, the numbers of
pressings to obtain ultimate stable micro structure were 4, 6 and 8, and the grain sizes
were — 1.3, — 0.45 and -0.27 pm, respectively [101].
In this investigation, the maximum tolerable number o f pressings at 298 K for Al-
3% Mg-0.2% Sc-0.12% Zr alloy was 6 and the resultant grain size was -0.3 pm. When
98
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
these conditions are compared with those for Al-3% Mg binary alloy mentioned above,
the following conclusion can be made. The resultant grain size of ~0.3 (j.m is not the
ultimate equiaxed equilibrium grain size and a homogeneous microstructure has not been
completely obtained after ECA pressing up to 6 pressings for the Al-3% Mg-0.2% Sc-
0.12% Zr alloy. In order to achieve better conditions for superplasticity, higher
temperature pressing may be the only solution because of the limited number of pressings
at room temperature.
In spite of the low number of pressings, an exceptional elongation o f -1680% was
recorded when testing at 773 K using a strain rate of 10*2 s'1 . Fig. 5.9 shows the
elongation to failure against the strain rate in order to compare the present results with
reported data for Al-Mg-Sc alloys. The elongations of the Al-Mg-Sc-Zr alloy are much
higher than any other alloys in this figure and exceptional elongations can be also
obtained over a wide range of temperatures, 673 to 798 K, as shown in Table 5.2. For Al-
3% Mg-0.2% Sc alloys [83,129], the initial grain size was reported as -0.2 pirn after both
8 and 12 passes and there was a significant increase in ductility, from -1030% to 1560%
at 673 K with a strain rate of 3.3 x I O'2 s'1 , when the number of pressings was increased
from 8 to 12. Although the present Al-Mg-Sc-Zr alloy possesses slightly larger grains of
-0.3 pun and is pressed only up to 6 passes, the higher ductility of this alloy over Al-Mg-
Sc alloy indicates the greater effectiveness in superplasticity when there are both Sc and
Zr additions at the same time. The fact that the maximum possible numbers of pressings
at 298 K was 8 for Al-Mg-Sc alloy and intermediate annealings after every 2 passes were
applied to press up to 12 passes is also consistent with the present investigation [83].
99
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
E longation t o fa ilu r e (% )
2800
2600
2400
2200
2000
1800
1600
1400
1200
1000
800
600
400
200
0
~ i - - - - - - - - - - - !- - - - - - - i— i— i— i— n - j - - - - - - - - - - - - - - - - - - - - i- - - - - - - - - - - r- - - - - - - i — i— i— i— n - j - - - - - - - - - - - - - - - - - - - i- - - - - - - - - I i i i l l
T (K)
■ i i i i i i i .
Treatment
Al-3Mg-0.2Sc (Berbon ef a/,1999) 573 ECAP 8 passes
AI-3Mg-0.2Sc (Berbon et a/,1999) 673 ECAP 8 passes
AI-3Mg-0.2Sc (Berbon et a/,1999) 673 ECAP 12 passes
AI-4Mg-0.5Sc (Sawtell & Jensen,1990) 672 TMP
AI-4Mg-0.5Sc (Sawtell & Jensen,1990) 811 TMP
AI-6Mg-0.3Sc (Nieh etal.1998) 748 TMP
AI-4Mg-1Mn-0.5Sc (Vetrano etal.1998) 823 TMP
AI-3Mg-0.2Sc-0.12Zr 773 ECAP 6 passes
1 0 - 4 10-3 10-2 IQ - 100
Strain rate (s'1 )
Figure 5.9 Elongation to failure vs. strain rate for a number of Al-Mg-Sc alloys and
Al-3% Mg-0.2% Sc-0.12% Zr alloy.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Other alloys in Fig. 5.9 shows much less elongations than the two alloys which
have been subjected ECA pressing. These alloys were all thermomechanically heat-
treated and these TMP techniques usually involve solution treatment, warm or cold
rolling and aging [78,79,132]. Therefore it is proven that the simple ECA pressing at
room temperature can produce very high tensile ductilities in these alloys without any
complex TMP technique.
5.1.5.3 Deformation mechanism for Al-Sc-Zr alloy
In order to understand the deformation behavior at high temperatures, at first, it is
essential to establish the relationship between stress and strain rate. The logarithm of flow
stress is plotted as a function o f the logarithm of strain rate for Al-0.2% Sc-Al-0.12% Zr
alloy in Fig. 5.10 where UTS is chosen as flow stress. The strain rate sensitivity value, m
(or ri), is obtained from equation (2.3). As shown in Fig. 5.10, there is a clear transition
near the strain rate range of 3.3 x 10'2 to 10'1 s"1 and each region appears to correspond to
regions II and HI in conventional superplasticity theory [5]. The measured average values
of m are 0.30 and 0.16 in regions II and HI, respectively. It is confirmed that high m
values result in high tensile elongations because high elongations, more than 300%, were
often achieved in region II, especially at 10'2 to 10'3 s'1 , at all testing temperatures listed
in Table 5.1. It is known that high temperature deformation is controlled by the climb of
dislocations over obstacles in pure metals and the anticipated m value is 0.2 to 0.25 [20],
In this investigation, a very dilute aluminum alloy was used and the m value was ~0.3 in
region II which is higher than the reported value o f-0.23 for pure aluminum [133].
101
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
a (MPa)
100
: Al-0.2% Sc-0.12% Zr
10
0.30
m
623
673
723
773
1
10° 10'1
8 (S 1)
Figure 5.10 The logarithm o f flow stress as a function of the logarithm of strain rate
Al-0.2% Sc-0.12% Zr alloy.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Ductilities higher than 300% have not been reported in pure metals and this high
value of m o f-0.30 leads to an exceptional elongation o f -590% for an ECA pressed Al-
activation energy is given by the following relationship at constant strain rate [134]:
where the data o f G are given by the following equation and the true activation energy
was obtained from a slope of QI2.3R in Fig. 5.11.
In the climb-controlled creep mechanism for pure metals, in class M, the value of
O corresponds to the activation energy for lattice self-diffusion in pure metals [20], For
pure aluminum, the values of O for lattice self-diffusion and for grain boundary diffusion
are — 143.4 kJ/mol and — 86 kJ/mol, respectively [135]. The value of Q obtained in this
investigation, 116 ± 5 kJ/mol, does not correspond to either of these two activation
energies and it may be assumed that the deformation behavior of the ECA pressed Al-
0.3% Sc-0.12% Zr alloy is not conventional class M type behavior.
It was reported that O for grain growth in an ECA pressed Al-Mg binary alloy
was very low and this was due to the existence of non-equilibrium grain boundaries
[136]. Therefore, the following explanation can be put forward to explain the deformation
behavior of Al-Sc-Zr alloy. The fact the maximum elongation of - 590% was obtained at
a very high temperature of 773 K, which corresponds to > 0.8 Tm , indicates lattice
diffusion in Al is more dominant than grain boundary diffusion at this high temperature.
Sc-Zr alloy without any solute elements when tested at 773 K and 10'3 s'1 . The true
(5.1)
G = 3.022 x 10"4 - 16T(MPa) (5.2)
103
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
70-s
1 Q - 7
^ 70-s
^ 70-s
C O
^rj 10M
b 70-Jj/
o-
70-t 2 /
70-73/
1-2S
'tgure 5
7-3o
7-35
“ Th
rec'P r o c ^ ^ ° f tempa
'/5 ct
5 fc//j
niol
-X 7
1-40
1-4S
1 0 ° o r r (^ } S ° »■« ) eo
7 -© S
*uperat ure- compens at ed stress as a f uncti - o f absol ute t emperat ure for Al -0.2 % ~
S c - ° ^ Z z * '
c°P I,rm B U ,n. 104
Furthe
°W ner.
1er^ o au„,„
However, the activation energy for superplastic deformation can be lower than Qi
because the high atomic mobility of non-equilibrium boundaries introduced by ECA
pressing may be another important diffusional source along with lattice diffusion.
Although there is no solute element in this alloy, the fine grains after ECA pressing can
be retained effectively at high temperatures by the addition of Sc and Zr. From the high m
value, therefore, it may be concluded that GBS is a major deformation mechanism for
ECA pressed Al-0.2% Sc-0.12% Zr alloy and this can be confirmed by the fact that the
sample showing the highest elongation was deformed uniformly up to failure without the
formation of local necking.
5.1.5.4 Deformation mechanism for Al-Mg-Sc-Zr alloy
The logarithm o f flow stress as a function of the logarithm o f strain rate for Al-
3% Mg-0.2% Sc-Al-0.12% Zr alloy is plotted in Fig. 5.12 where UTS is also chosen as
flow stress. The measured average m value is — 0.36 and m is increased with increasing
the testing temperature except for the case o f 673 K. For example, m is increased from
0.35 to 0.40 as the temperature is increased from 573 K to 773 K. This is very consistent
with the fact that the maximum elongation o f 1680% was found at 773 K where rn is the
highest in the tested temperature range. It is also very consistent with the fact that the
largest elongations for HSR SP aluminum matrix composites and alloys are obtained
when the value of m is more than 0.3, as shown in Table 2.1 and 2.3. The present Al-Mg-
Sc-Zr alloy clearly shows HSR SP over the whole testing temperatures from 573 to 773
K and its elongations are quite exceptional at very high strain rates.
105
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
M issio n
Fig u
re 5.12 T he lo g arith m o f flow stre ss a s a function o f th e logarithm o f s f c
A i -3% M g-0. 2% S c-0. 12% Z r alloy.
ra/n fGr
of the con, ■ 106
^Pyright owner
Fmherreprm
l0nPm^ awith0
1 M ission.
The present result of m -0.36 is consistent with the reported results for other Al-
Mg-Sc systems. For example, the m value o f an Al-4% Mg-0.5% Sc alloy tested at 672 K
by Sawtell and Jensen [78] was -0.4 when the maximum elongation of >1020% was
obtained at 10"2 s'1 , that of an Al-6% Mg-0.3% Sc alloy tested at 748 K by Nieh et al.
[79] was -0.6 when the maximum elongation of 1130% was obtained at 1.38 x 10'2 s'1 ,
and those of AI-4% Mg-0.5% Sc alloys ECA pressed at 473 K and 673 K by Mukai et al.
[137] were -0.5 and -0.33, respectively.
As shown in Fig. 5.5, it is found that there is grain growth within the deformed
region by dynamic recrystallization. It is believed that grain growth occurs in the early
stage of deformation and may affect the measured m value because the UTS, which also
occurs in the early stage of deformation as shown in Fig. 5.3, is chosen as the value o f the
flow stress in calculating the m value. In order to verify this, the values of stress in the
steady state region in Fig. 5.3 were selected to calculate the m value when tested at 773
K. The measured m value at 773 K is -0.4, which is the same as the previous value using
UTS at 773 K. Therefore, it is reasonable to conclude that m = 0.36 is obtained for
superplastic deformation in the Al-3% Mg-0.2% Sc-0.12% Zr alloy.
For Al-Mg substitutional binary solid solution alloys, the principle deformation
mechanism at high temperatures is dislocation glide-controlled creep (solute-drag
mechanism) in which the strain rate sensitivity is 0.3 [135], The solute drag mechanism is
controlled by the glide of dislocations with solute atom atmospheres and can produce
relatively large elongations up to -400% with the occurrence of necking [20], In the
present investigation, however, the Al-3% Mg-0.2% Sc-0.12% Zr alloy does not exhibit
107
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
class A behavior because the value o f m is higher than 0.3 and the tested tensile samples
show no visual necking within the gauge length as shown in Fig. 5.4. Therefore, it is
evident that the dominant deformation mechanism is GBS.
Furthermore, the activation energy for the solute drag mechanism corresponds to
the activation energy for the diffusion o f the solute atom in the alloy [20] and the
activation energy for the impurity diffusion o f Mg in Al is -130.5 kJ/mol [138]. The true
activation energy is extrapolated in Fig. 5.13 and the measured average activation energy
is 100 ± 11 kJ/mol. This is similar to the case of the Al-Sc-Zr alloy where the measured
O is lower than the anticipated O for lattice self-diffusion of aluminum. This value is
significantly higher than the reported value o f 35 ± 3 kJ/mol for the Al-4% Mg-0.5% Sc
alloy [78] and lower than the value of — 190 kJ/mol for the Al-6% Mg- 0.3% Sc alloy
[79].
It was proposed that the mechanism for superplasticity in Al-Mg-Sc alloy is
continuous recrystallization in the early stage of deformation and subsequent GBS
accommodated by dislocation glide across the grains with a uniform dispersion of AljSc
precipitates [79], For the present Al-Mg-Sc-Zr alloy, however, the continuous
recrystallization which involves the conversion of low angled subgrain boundaries to
high angle grain boundaries in the beginning of the tensile deformation may not occur
because the initial grain boundaries after ECA pressing up to 6 pressings at 298 K have
already high angles of misorientation as shown in Fig. 5.2 (a). Furthermore, the model
predicts an m value o f 0.5 but a value of — 0.36 is obtained in the present Al-Mg-Sc-Zr
alloy.
108
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1 0 - 3
1 0 - 4
1 0 - 5
1 0 - 6
C O
Q _
S 1 0 - 7
S l « — * *
h—
GO
h - 1 0 - 8
O
GO
1 0 - 9
CM
1 0 " 1 0
1 0 - ”
1 0 - 1 2
I I I I I I I
_ 1 ---------,---------,---------,---------,---------j---------r -
Al-3% Mg-0.2% Sc-0.12% Zr
Q^= 100 ±11 kJ/mol
8 (S -1)
•
1.0 x 10-4
o
1.0 x 10'3
T 3.3 x 10'3
V 1.0 x 10‘2
■
3.3 x 10’2
□
1.0 x 10*1
1.2 1.3
1.4 1.5 1.6 1.7
1000/T (K'1 )
Figure 5.13 The logarithm of temperature-compensated stress as a function of the
reciprocal of absolute temperature for Al-3% Mg-0.2% Sc-0.12% Zr
alloy.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Therefore, it is concluded that the deformation mechanism for ECA pressed Al-
3% Mg-0.2% Sc-0.12% Zr alloy is GBS accommodated by dislocation glide and this is in
good agreement with the previous models [79,83]. Since very small grains o f — 0.3 pm
and non-equilibrium boundaries result from ECA pressing, however, the measured m
value of -0.36 is higher than 0.33 for the solute drag mechanism and the activation
energy is slightly smaller than the anticipated value for the impurity diffusion of Mg in
Al.
5.2 Al-2024 ALLOY
5.2.1 Introduction
A 2024 aluminum alloy, one of the Al-Cu-Mg systems (2xxx series), is a heat-
treatable wrought aluminum alloy widely used for aircraft parts [117], This commercial
aluminum alloy is not a standard superplastic alloy and one of the main limitations is the
limited formability. This limited formability is caused primarily by the large grain size of
the industrial products. Therefore, much effort has been dedicated to developing very fine
grain structures. One method to refine the grain size is by the TMP techniques, which
often involve solution treatment, over-aging, deformation (rolling or extrusion) and
recrystallization [86,139], and another method is using a PM technique [75]. However,
both methods often involve very complicated procedures. In order to improve the
superplastic properties of 2024 Al alloy, therefore, a very simple ECA pressing technique
was applied to the 2024 Al alloy under various pressing conditions.
110
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5.2.2 Materials and experimental procedures
The AI-2024 alloy is a heat-treatable aluminum alloy and the basic requirement of
age hardening is a decrease in solid solubility of one or more o f the alloying elements
with decreasing temperature. This heat treatment normally involves the following stages
[117,118], Solution treatment at a relatively high temperature within the single phase
region to dissolve the alloying elements, quenching to obtain a supersaturated solid
solution of these elements in aluminum, and then controlled decomposition of the
supersaturated solid solution to form finely dispersed precipitates. Typically, in Al-Cu-
Mg systems, two types of precipitates are mainly formed, namely CuA12 (9 phase) and
CuMgAl2 or Cu2 Mg2Al5 (S phase). Therefore, the strengthening mechanism of Al-2024
alloy is mainly precipitation hardening in addition to the solid solution strengthening.
In the present investigation, two types of heat-treatment were applied to the Al-
2024 alloy. First, 2024-0 indicates an annealed Al-2024 alloy and this alloy has the
lowest strength and highest ductility at room temperature so that it can be successfully
ECA pressed at room temperature. Since as-annealed aluminum alloy is seldom used in
practical application, however, 2024-T351 heat-treated alloy was chosen as a reference
for the ECA pressed material.
Initially, billets of Al-2024 were pressed at 298 K using route Be up to 8
pressings. The selected numbers o f pressings were 1, 2, 4, 6 and 8. In order to investigate
the effect of the pressing routes on superplasticity, billets were also pressed at 298 K
using Bc-C and C-Bc combination routes up to 8 pressings. Furthermore, in order to
investigate the effect of the pressing temperatures, billets were pressed at 373 K using
111
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
route Be up to 12 pressings. It should be noted that the maximum applicable number of
pressings at 298 K was 8 and it was possible to press up to 12 passes at 373 K. For 2024-
T351 alloy, it was determined that ECA pressing at 298 K led this strong heat-treated
alloy to breakage at lower numbers of pressings.
After pressing, all pressed samples were tested in tension up to failure at room
temperature or at elevated temperatures. For comparison, unpressed Al-2024 and 2024-
T351 alloys were also tested in tension at the same testing conditions. In order to
investigate the thermal stability of the submicrometer-grained structure after ECA
pressing at 298 K via route Be up to 8 passes, samples were annealed for 1 hour over a
wide range of temperatures and then the Vickers microhardness was measured by using
the same testing conditions mentioned in section 4.2.5.
5.2.3 Results of tensile testing at room temperature
All tensile test results o f unpressed, T-351 heat-treated and ECA pressed Al-2024
alloys when testing at room temperature are summarized in Table 5.3. It is found that the
elongations to failure of unpressed materials are much larger than those o f T-351 heat-
treated and ECA pressed materials, however, the YS and UTS o f unpressed materials are
much smaller than those of T-351 heat-treated and ECA pressed materials. This suggests
that the additional heat-treatment to form finely distributed precipitates or ECA pressing
to refine grains make the Al-2024 alloy brittle in terms of room temperature deformation
behavior.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 5.3 Tensile test results at room temperature for Al-2024 alloys
Treatment
Strain rate
(sec'1 )
YS
(MPa)
UTS
(MPa)
EF
(%)
10'1 130 295 45
annealed
10'2
10'3
119
122
297
289
48
45
1 0 - 4 111 309 50
T-351 heat-
treated
10'2
10'3
10"4
556
575
578
783
812
791
26
26
25
ECA pressed up
to 1 pass
10'2
10'3
1 0 - 4
241
245
204
406
397
366
19
17
25
ECA pressed up
to 2 passes
10'2
10'3
10-4
274
222
211
419
410
417
17
20
26
ECA pressed up
to 4 passes
10'2
10'3
xo-4
259
259
246
458
479
422
16
19
27
ECA pressed up
to 8 passes
10'2
10'3
io- 4
373
357
286
510
492
461
10
13
19
113
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The strengthening effect of heat-treatment is superior to that of ECA pressing and
this is due to the fine precipitates uniformly distributed in the aluminum matrix of 2024-
T351 alloy and also the reduced content of solute elements in the aluminum matrix in
addition to the coarse precipitates for the ECA pressed Al-2024 alloy regardless of the
much smaller grain size. This deformation behavior is clearly seen in Fig. 5.14. After
considering the effect of grip slippage and Young’s modulus o f about 70 GPa, the elastic
part in stress vs. strain curves in Fig. 5.14 were removed to get reliable data when testing
at room temperature.
In Fig. 5.14, the stress-strain curves for the annealed and T351 heat-treated alloys
show an extensive strain hardening stage up to failure and this is typical for alloys when
testing at low temperature. However, the Al-2024 alloy, when subjected to ECA pressing
up to 8 passes via route Be, not only exhibits an extensive strain hardening stage but also
shows a short straining softening region, especially when testing at low strain rates. The
same trends were also obtained for other pressing conditions, such as for 1 , 2 and 4
passes. Another important fact is that the stress vs. strain behavior of unpressed and T351
heat-treated alloys is not sensitive to the strain rate used and all curves can match each
other as indicated in Fig. 5.14. However, the stress vs. strain behavior of pressed Al-2024
alloy is very sensitive to strain rate and larger elongations and lower strength are often
obtained when testing at lower strain rates, as clearly shown in Table 5.3 and Fig. 5.14. It
is believed that the microstructure after ECA pressing consists of very small dislocation
tangled subgrains and subgrain boundaries whereas the microstructure of unpressed
alloys mainly is fully developed large grains and texture.
114
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
a (MPa)
1000
. Al-2024 Treatment
annealed 10
T = 2 9 8 K
800
10
T351 heat-treated
10
10
600
ECA pressed up to 8 passes
10
400
200
10 40 60 0 20 30 50
8 (%)
Figure 5.14 The stress vs. strain curves of Al-2024 alloys tested at 298 K.
115
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
These results are different from those o f Valiev et al. [140]. They used torsion
straining (TS) to achieve submicrometer-grained structure for Al-Cu-Zr and Al-1420
alloys. After TS, the YS, UTS and EF were all increased and the strain hardening was
reduced significantly. However, the same results with the present results were obtained
by Xu [141] using ECA pressing for an Al-3% Mg alloy. After pressing, the YS and UTS
were increased and EF was decreased significantly.
Deformation behavior at room temperature and at the fixed strain rate of I O '4 s'1
for the ECA pressed Al-2024 alloy is depicted in Fig. 5.15. The strength increases as the
number of pressings is increased. This is consistent with the previous results of
microhardness measurements.
5.2.4 Results of tensile testing at high temperatures
ECA pressing at 298 K was conducted for annealed Al-2024 alloy and route Be
was used for all pressed samples. The tensile testing results at high temperatures for both
the unpressed and ECA pressed Al-2024 alloy are summarized in Table 5.4. For the
unpressed Al-2024 alloy, the maximum elongation to failure of — 280% was recorded
when testing at 723 K using a strain rate of 10"3 s'1 . Therefore, the unpressed material
already shows superplastic-like properties without ECA pressing, but the EF values for 1 ,
2 and 4 pressed materials were less than — 200%. Above an applied strain of -6, some of
the EF values are larger than 300% at specific temperatures and strain rates. The largest
elongation to failure of -460% was obtained when testing at 673 K using a strain rate of
10'3 s'1 after ECA pressing at 298 K via route Be up to 8 pressings.
116
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
a (MPa)
500
450
400
350
300
250
200
Number of passes
150
Al-2024
100
50
T = 298 K
20 25 30 0 5 10 15
8 (%)
Figure 5.15 The stress vs. strain curves of the ECA pressed Al-2024 alloys when testing
at 298 K
1
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 5.4 Tensile test results at high temperatures for unpressed and ECA pressed
Al-2024 alloys
Treatment
Temp. Strain rate YS UTS EF
(K) (sec'1 ) (MPa) (MPa)
(%)
573 10'3 43 74 91
623 10'3 35 55 115
673 1 75 116 129
673 10'1 49 82 135
673 10'2 36 63 158
Unpressed
673
673
10'3
io- 4
30
14
43
27
189
150
723 10'2 31 44 221
723 10'3 15 25 280
723 IQ4 10 13 267
773 10'2 16 25 175
773 10'3 8 12 189
573 10'2 103 120 72
573 10'3 44 79 120
623 10'2 46 74 126
ECA pressed 623 10'3 23 53 138
up to 1 pass 673 10*2 38 54 120
673 1 0*3 32 41 135
723 I O '3 22 32 96
773 10'3 17 19 62
573 10'2 65 112 112
573 10'3 39 76 152
623 10'2 39 73 122
623 10'3 32 49 130
ECA pressed
up to 2 passes
673
673
673
10'1
10'2
10'3
51
30
25
73
49
37
133
127
115
673 10-4 22 28 92
723 10'2 35 41 141
723 10'3 17 26 113
773 10'3 12 16 68
(continued)
118
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 5.4 Tensile test results at high temperatures for unpressed and ECA pressed
Al-2024 alloys
Treatment
Temp.
(K)
Strain rate
(sec'1 )
YS
(MPa)
UTS
(MPa)
EF
(%)
523 10'3 69 121 135
573 10'2 55 105 118
573 10'3 53 74 169
623 10'2 39 79 149
623 10'3 27 50 159
673 1 60 100 135
ECA pressed 673 10'1 50 78 148
up to 4 passes 673 10'2 28 46 199
673 10'3 19 33 176
673 10*4 18 23 140
723 10'2 28 35 154
723 10'3 14 22 137
773 icr2 21 25 76
773 10'3 5 7 103
ECA pressed 623 10'3 15 38 312
up to 6 passes 673 10'3 9 23 279
573 10'3 28 74 211
623 10'2 31 73 254
623 10'3 18 40 344
673 1 72 104 165
ECA pressed
673 10'1 37 77 185
673 10'2 17 52 304
up to 8 passes
673 10'3 8 23 456
673 3.3 x 10'3 15 36 318
673 io- 4 7 13 277
723 10'3 4 15 284
773 10'3 5 7 156
119
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The tensile testing results at high temperatures for T351 heat-treated Al-2024
alloy are summarized in Table 5.5. The maximum elongation o f -150% was obtained
when testing at 723 K using a strain rate of 10'1 s'1 . Although this EF value is not high as
that found in the ECA pressed Al-2024 alloy, it should be noted that the maximum
elongation was obtained at a very high strain rate.
In general, elongations to failure for all materials were not good at very high
temperatures (— 773 K) probably due to grain growth and they were also bad at very low
temperatures (— 573 K) due to insufficient thermal energy. This is the same result as in
conventional superplastic data in which there are specific optimum ranges of temperature
and strain rate. Inspection of all stress vs. strain curves for annealed, T351 heat-treated
and ECA pressed Al-2024 alloys shows that the shape of all stress vs. strain curves are
almost identical regardless of the processing conditions and all deformation behavior is
sensitive to the strain rate.
The high temperature mechanical behavior can be compared to each other by
using Fig. 5.16 for annealed, T351 heat-treated and ECA pressed Al-2024 alloys. The
strain rate of 10'3 s'1 and the temperature of 673 K are chosen because this test condition
gives the best results for the ECA pressed materials. When the flow stress (YS or UTS) is
considered, stresses for T3 51 heat-treated material are the highest, and then stresses for
annealed materials are higher than those for ECA pressed materials. For the ECA pressed
materials, the values of stress decrease as the number of pressings increases. When
comparing with deformation behavior at room temperature, this trend is reversed except
for the case of the T351 heat-treated material, as shown in Fig. 5.14 and Fig. 5.15.
120
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 5.5 Tensile test results at high temperatures for T351 heat-treated Al-2024 alloy
Temp. Strain rate YS UTS EF
(K) (sec'1 ) (MPa) (MPa) (%)
623 10'1 205 239 48
623 10'2 134 153 58
623 10*3 103 128 58
623 1 0 - 4 84 93 49
673 10'1 96 120 87
673 10'2 79 90 88
673 10'3 63 70 72
673 1 0 - 4 44 49 65
723 1 71 119 131
723 10'1 66 83 146
723 10'2 46 57 130
723 10'3 33 41 105
723 1 0 - 4 23 28 75
773 10'1 43 56 90
773 10'2 27 36 118
773 10'3 18 23 116
773 io- 4 10 13 82
1 2 1
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
a (MPa)
80
I l l “ T I---------- 1 I---------- (-----------1-----------1 ----------- 1 ----------- 1 ----------- 1 ----------- 1 ----------- 1 ----------
Al-2024
T = 673 K & s = 10*3 s '1
treatm ent____________
annealed
T351 heat-treated
ECA pressed up to 1 pass
ECA pressed up to 2 passes
ECA pressed up to 4 passes
ECA pressed up to 6 passes
ECA pressed up to 8 passes
500
s (% ]
Figure 5.16 The stress vs. strain curves o f Al-2024 alloys tested at 673 K and at 10‘3
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
As already stated, ECA pressing produces non-equilibrium sub-grain or grain
boundaries. As the number of pressings increases, the portion o f these boundaries is
increased and then saturated. Since these boundaries possess higher atomic mobility (i.e.
diftusivity) than ordinary grain boundaries, the applied stress can be greatly relaxed
during superplastic deformation. As shown in Fig. 5.16, therefore, the values of the stress
decrease as the number of pressings is increased but they become saturated after 6
pressings. This also indicates that the optimum superplastic condition for GBS can only
be satisfied by a sufficient amount of strain induced by ECA pressing because
elongations of — 300% are achieved only after 6 pressings.
Fig. 5.17 shows the appearance of the tensile specimens tested to failure at 673 K
and at various strain rates after ECA pressing at 298 K via route Be up to 8 passes. It is
apparent that there is uniform deformation within the gauge length without any localized
necking especially for highly elongated specimens.
In order to investigate the effect of pressing routes on superplasticity, again ECA
pressing at 298 K was conducted for annealed Al-2024 alloy and two combination routes,
C-Bc and Bc-C, were used for all pressed samples. The tensile testing results at high
temperatures for ECA pressed Al-2024 alloy by routes C-Bc and Bc-C are summarized in
Table 5.6 and Table 5.7, respectively. The largest elongations to failure of -330% and
-350% were obtained when testing at 623 K using a strain rate of 10'3 s'1 after ECA
pressing at 298 K up to 8 pressings via routes C-Bc and Bc-C, respectively. From the
results for three different routes, it is clear that the higher number of pressings produces
the larger elongations to failure.
123
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.17 Examples o f tensile ductility in Al-2024 alloys ECA pressed at 298 K via
route Be up to 8 passes and pulled at 673 K and 1 O'3 s'1 .
124
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 5.6 Tensile test results at high temperatures for Al-2024 alloys ECA pressed
at 298 K via route C-Bc
Number of Temp. Strain rate YS UTS EF
pressings (K) (sec*1 ) (MPa) (MPa)
(%)
573 10'2 72 108 96
573 I O ’ 3 46 71 128
623 10’ 2 53 74 162
2 passes
623 10‘3 38 48 119
673 10'2 45 51 130
673 I0*3 35 37 113
723 10'2 34 39 141
723 10'3 25 29 98
573 I O '2 70 106 105
573 10‘ 3 41 72 192
623 10‘2 37 77 221
4 passes
623 10'3 26 51 266
673 I O '2 41 48 135
673 10'3 30 37 121
723 10'2 33 38 139
723 10’3 23 25 126
573 10'2 59 106 144
573 10'3 35 68 208
623 10-2 37 67 223
6 passes 623 10° 17 42 321
673 10‘2 38 51 139
673 10’ 3 27 35 124
723 10'3 22 23 109
573 10'2 64 97 140
573 10'3 39 70 222
623 10‘1 73 98 139
623 10’2 37 66 233
8 passes 623 10'3 18 42 332
623 10-4 11 23 278
673 I O '2 41 50 143
673 10‘3 25 36 131
723 10*3 20 23 110
125
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 5.7 Tensile test results at high temperatures for Al-2024 alloys ECA pressed
at 298 K via route Bc-C
Number of Temp. Strain rate YS UTS EF
pressings (K) (sec'1 ) (MPa) (MPa)
(%)
573 10'2 65 112 112
573 10'3 39 76 152
623 10'2 39 73 122
623 10'3 32 49 130
673 10'1 51 73 133
2 passes 673 10'2 30 49 127
673 10'3 25 37 115
673 10*4 22 28 92
723 10'2 35 41 141
723 10'3 17 26 113
773 10'3 12 16 68
573 10'2 76 107 111
573 10'3 41 76 171
4 passes 623 10'2 42 74 203
623 10'3 23 48 264
673 10'3 28 37 132
573 10'2 67 106 127
573 10'3 39 73 200
6 passes 623 10'2 38 72 227
623 10'3 19 47 300
673 10'3 21 34 171
573 10'2 72 108 120
573 10'3 37 72 229
8 passes
623 10'2 40 73 225
623 10'3 17 44 347
673 10'2 36 52 172
673 10'3 16 30 223
126
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The deformation behavior (i.e. stress vs. strain curves) for two combination routes
is the same as those for unpressed and ECA pressed using route Be Al-2024 alloy. Since
the maximum applicable number of pressings for these two combination routes was also
8 which is the same as that for route Be, ECA pressing at high temperature (i.e. 373 K)
was conducted for the annealed Al-2024 alloy. The selected number o f pressings was
either 8 or 12 and route Be was used. The tensile test results at high temperatures are
summarized in Table 5.8.
For the Al-2024 alloy ECA pressed at 373K up to 8 passes, a maximum
elongation of — 500% was obtained at a temperature of 673 K and a strain rate of 10'2 s'1 .
And for the Al-2024 alloy ECA pressed at 373K up to 12 passes, a maximum elongation
of -500% was obtained at a temperature of 673 K and a strain rate of 10'3 s'1 . Fig. 5.18
shows the appearance of the tensile specimens tested to failure at 673 K and at various
strain rates after ECA pressing at 373 K via route Be up to 8 passes. It is apparent that
there is uniform deformation within the gauge length without any localized necking
especially for highly elongated specimens.
5.2.5 Results of annealing at high temperatures
In order to investigate the stability of the micro structure after ECA pressing, all
pressed samples including unpressed Al-2024 alloy were annealed for 1 hour over a wide
range of temperatures (from 373 to 773 K by a 50 K increment) where they were held
constant to within ± 2 K, and then air-cooled, finally, the Vickers microhardness on the X
plane was measured for each specimen. The results are represented in Fig. 5.19.
127
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 5.8 Tensile test results at high temperatures for Al-2024 alloys ECA pressed
at 373 K via route Be
Number of Temp. Strain rate YS UTS EF
pressings (K) (sec'1 ) (MPa) (MPa)
(%)
623 10'j 9 23 380
8 passes
673 10'1 37 58 288
673 10'2 15 29 497
673 10'3 7 15 409
623 10'2 34 69 302
623 10'3 16 43 412
12 passes
673 10'1 42 68 254
673 10'2 19 46 383
673 10'3 8 24 501
723 10'3 5 13 299
128
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
A l2024
Al-4.4%Cu-1.5%Mg
I n itia l s a m p le
1 0 '1 s '1 290%
500 %
410%
10 s '
10’3 s’1
Figure 5.18 Examples o f tensile ductility in Al-2024 alloys ECA pressed at 373 K via
route Be up to 8 passes and pulled at 673 K.
129
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Microhardness (Hv)
160
Al-2024 U npressed
ECA pressed up to 1 pass
ECA pressed up to 2 passes
ECA pressed up to 4 passes
ECA pressed up to 8 passes
140
120
100
80
60
40
600 300 400 500 800 700
Annealing temperature (K )
Figure 5.19 Variations of microhardness with annealing temperature for unpressed and
ECA pressed Al-2024 alloys.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The hardness values are very consistent up to 623 K for the unpressed material, it
is believed that no abrupt microstructural change has occurred because the initial
unpressed Al-2024 alloy possesses large grains. It was determined that the as-annealed
Al-2024 alloy has plate-like grains with average dimensions o f -500 x 300 x 10 pm3 ,
whereas the grains of the T351 heat-treated Al-2024 alloy are also highly elongated and
the average longitudinal grain sizes are -300-400 pm and the average transverse grain
sizes are -40-150 pm.
For pressed specimens, however, the hardness values are initially consistent up to
423 K, and they decrease in the temperature region from 423 K to 623 K, especially for
the 8 pressed sample. This can be attributed to grain growth in the pressed specimens in
this temperature range. After 673 K, the hardness of all tested materials increases very
rapidly. Fig. 5.20 shows microstructures of the Al-2024 alloys EC A pressed via route Be
up to 8 pressings at 298 K and 373 K. The measured grain sizes of the sample pressed at
298 K is -0.2-0.3 pm and that of the sample pressed at 373 K is -0.4-0.6 pm. This is
consistent with the fact that the higher pressing temperature produces the larger initial
grain sizes.
The increase in hardness after high temperature annealing may be due to natural
aging o f Al-2024 alloy regardless of any grain growth. The solid solution retained at the
instant of quenching or air-cooling is unstable at room temperature, and the hardening
phases or precipitates begin to form immediately. It is known that Al-2024 alloy age-
hardens quite rapidly at room temperature and hardening is almost completed after about
four days [117,118],
131
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.20 Microstructures of Al-2024 ECA pressed via route Be up to 8 passes
at (a) 298 K and (b) 373 K.
132
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In industry, the usual practice for full annealing is to heat for 2-3 hours at 686 K,
and then to cool slowly in the furnace at a controlled rate to 533 K in order to form coarse
equilibrium precipitates. Thereafter, the cooling rate is unimportant. In the case of partial
annealing below 616 K, it is also known that the cooling rate is not important [117,118],
Therefore, it is concluded that the increase in hardness after annealing at 623 K is a direct
result o f the formation of precipitates by natural aging of the Al-2024 alloy.
5.2.6 Discussion
5.2.6.1 Superplasticity in Al-2024 alloy
The commercial Al-2024 alloy, Al-4.4% Cu-1.5% Mg alloy, is not often regarded
as a superplastic aluminum alloy and its poor formability prevents its use in the
superplastic forming industry. However, heat-treated Al-2024 alloys are now widely used
for aircraft construction and a development of forming technique by using superplastic
properties of this material is essential. Therefore, many attempts have been devoted to
improve the superplastic properties of this material.
One of the methods is using TMP techniques. A number o f TMP techniques are
now available for materials processed by IM or PM, and the specific TMP route is only
effective on selected alloy systems. For example, an Al-2014 alloy after subjecting TMP
(the cross upsetting-drawing process) showed a maximum elongation o f — 380% at 693 K
with the strain rate of 4.17 x 10'3 s'1 [139] and an Al-2024 alloy after subjecting TMP
(hot stamping and extrusion) showed a maximum elongation of -1830% at 773 K with
the strain rate of 1.3 x 10'3 s'1 [86], Regardless of the superior elongation, however, the
i o n
1 33
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
chemical composition of the latter alloy was modified by adding a small amount of Ti
and the values o f elongation were dropped abruptly with increase in the strain rates (i.e.
EF o f— 350% was obtained at 5 x 10'3 s'1 ).
Another approach to improve superplasticity of Al-Cu-Mg alloys is the addition
of alloying elements, especially zirconium. For example, an Al-2124-0.6% Zr alloy
showed a maximum elongation to failure o f -500% at 748 K with a strain rate of 3.3 x
10'1 s'1 [75] and an Al-2024-3% Fe-5% Ni alloy showed a maximum elongation to failure
of -400% at 773 K with a strain rate o f 3 s'1 [142], It should be noted that both alloys
were processed by PM and also treated by their own TMP technique before testing.
In this investigation, the maximum elongation to failure of the Al-2024 alloy,
when subjected to EC A pressing at 373 K up to 8 passes via route Be, was -500% when
testing at 673 K with a strain rate of 10'2 s'1 . In order to compare the present results with
those previously mentioned for other AI-Cu-Mg alloys, Fig. 5.21 shows the elongations
to failure against the strain rate.
The two alloys without the additional elements show good ductility in the low
strain rate region but the values of elongation to failure decreased abruptly at high strain
rates of > 10'2 s'1 . This may be due to the relatively larger grain size than in other listed
alloys. For the alloys processed by PM and TMP, exceptional HSR SP was obtained for
both alloys with Zr and a combination of Fe and Ni. However, their ductilities were
significantly inferior at low strain rates of < 10'2 s'1 . Since HSR SP forming in industry is
often conducted at — 10'2 s'1 , the ECA pressed Al-2024 alloy, which shows moderate
elongations of more than -300% at 10'1 to 10'2 s'1 , is more than enough to be utilized.
134
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
E longation t o fa ilu r e (% )
i i i l i r ~
1000
T(K ) Treatment
— # — AI-2014 (Li e t a/,1991) 693 TMP
Al-2024 (Wei & Zhang ,1994) 773 TMP
AI-2124-0.6Zr (Nieh & Wadsworth,1993) 748 PM + TMP
—O— AI-2024-3Fe-5Ni (Matsuki efa/,1999) 773 PM + TMP
— A —
Al-2024 673 ECAP 8 passes
100
1 0 - 4 1 0 - 3 10-2 10-1
Strain rate (s'1 )
10° 101
Figure 5.21 Elongation to failure vs. strain rate for a number o f Al-Cu-Mg alloys.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Furthermore, the testing temperature o f the ECA pressed Al-2024 is the lowest
among the alloys listed in Fig. 5.21 and this is another advantage over the TMP
techniques. Thus, it can be concluded that ECA pressing is a very effective procedure for
achieving very fine submirometer-grains in alloys without complex TMP techniques.
Since the heat-treated Al-2024 alloy possesses much higher strength at both room
temperature and high temperatures, for the practical application of ECA pressing
technique on commercial aluminum alloys, the following simple procedures may be
applicable in the forming industry. First, the softest condition (i.e. as-annealed condition)
should be chosen in order to apply high strains by ECA pressing. Second, the ECA
pressed billets can be rolled to form a plate for the next forming operation. Third,
superplastic forming can be conducted at low temperatures and high strain rates to save
cost and operation time. And last, the final formed parts can be heat-treated to give the
desired mechanical and physical properties.
5.2.6.2 The effect of pressing routes on superplasticity
If the optimum superplastic conditions for three different pressing routes, such as
Be, C-Bc and Bc-C, are compared when ECA pressing was conducted at 298 K, the test
condition at a temperature of 623 K and at a strain rate of 10'3 s'1 usually produced the
best superplastic results for the two combination routes, but the test condition of a higher
temperature o f 673 K and the same strain rate of 10'3 s'1 gave the best superplastic results
for route Be, as shown in Tables 5.4, 5.6 and 5.7. These results were also consistent for
every number of pressings, such as 2, 4, 6 and 8 passes. When the values of elongation
136
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
are compared with respect to the number of pressings, for 2 passes there is no variation in
elongation o f -150-160% for all routes. For 4 passes, however, the values of elongation
for routes C-Bc and Bc-C are superior to those for route Be. For 8 passes, the maximum
elongation (-460%) for route Be is much higher than those (-330% and -350%) for
routes C-Bc and Bc-C.
It is evident that the difference in values o f elongation to failure for the different
pressing routes is caused by the different shearing characteristics and the resultant
micro structural development during ECA pressing. It was reported that the predictions of
shearing patterns for routes A, Be and C are very consistent with the microstructural
observations for pure AI pressed after 2 pressings [107] and this microstructural
difference can produce a different response in high temperature deformation. However,
the relationship between the shearing characteristics and the resultant micro structure may
not agree well at high numbers of pressings. For example, it is believed that route A
produces more elongated grains or texture as the number of pressings is increased, but it
was reported that a reasonably homogeneous micro structure with equiaxed grains was
obtained after 10 pressings via route A [99]. For the three routes in this investigation,
they all include the shearing characteristics o f route Be and it is expected that the
microstructural variations between these routes will be minimized as the number of
pressings is increased.
The HSR SP behavior was not found in any routes and the maximum elongations
to failure were always obtained at 10'3 s'1 . The optimum superplastic temperature for the
two combination routes is 623 K, which is lower than that (673 K) for route Be-
137
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
However, the largest elongation of -460% was recorded when the Al-2024 was pressed
via route Be- From the above results, it still remains in question whether the route Be is
superior to routes C-Bc and Bc-C in terms of superplasticity. However, we cannot say
that route Be is superior to other routes, but we can consider that route Be is the most
reliable pressing route because other routes have no clear advantage in terms of
superplasticity over route Be-
5.2.6.3 The effect of pressing temperatures on superplasticity
The optimum superplastic condition may be defined as the combination of testing
temperature and imposed strain rate which gives the largest elongation to failure. Thus,
inspection of Tables 5.4 and 5.8 shows that the optimum condition in the unpressed Al-
2024 alloy is at a temperature of 723 K with strain rates in the vicinity o f— 10'4 to 10'3 s'1 .
After ECA pressing, however, it is evident that the optimum superplastic conditions tend
to change to lower temperatures and possibly also to faster strain rates. For example, after
ECA pressing through 8 passes at 298 K using route Be, the optimum strain rate is again
~10'3 s'1 but the optimum temperature is reduced to 673 K. Furthermore, the elongations
are higher after ECA pressing with a maximum recorded elongation of close to 500%
after pressing for 8 passes at 373 K using route Be.
Fig. 5.22 shows the optimum superplastic conditions for the Al-2024 alloy in the
unpressed condition and after ECA pressing using route Be at either 298 K or 373 K. It is
apparent that this alloy is essentially on the threshold of exhibiting HSR SP where high
ductilities are achieved at strain rates above 10'2 s'1 .
138
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Elongation t o failure (%)
700
tiki
723
673
673
Al-2024
Unpressed
Route Bc : 8p (298K)
Route Bc : 8p (373K)
600
500
400
300
200
100
10-4 10'3 10-2 10-1 10 °
Strain rate (s'1 )
Figure 5.22 Elongation to failure vs. strain rate under optimum conditions for Al-2024
alloy.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In Fig 5.22, it is clear that ECA pressing at high temperature can give higher
elongations than ECA pressing at room temperature. When a comparison is made
between the values of YS or UTS for Al-2024 alloys ECA pressed up to 8 passes but at
different pressing temperatures, as shown in Tables 5.4 and 5.8, the overall values of
stresses for samples pressed at 298 K is always higher than those for samples pressed at
373 K. This can be attributed to the difference in grain size. The measured initial grain
size o f the sample after pressing at 298 K is -0.2-0.3 pm and that of the sample after
pressing at 373 K is -0.4-0.6 pm.
It seems that there is no significant advantages for higher numbers of passes over
8 in Al-2024 alloy because the optimum strain rate decreases from 10'2 to 10'3 s'1 as the
number of pressings increases from 8 to 12 but similar elongations of — 500% were
obtained for both cases. The maximum elongation to failure o f — 290% was recorded at
673 K with the strain rate of 10'1 s'1 for Al-2024 alloy ECA pressed at 373 K up to 8
passes via route Be. This result is very attractive because an elongation of — 300% is more
than enough for a superplastic forming process in industry.
Fig. 5.23 shows m values for Al-2024 alloys ECA pressed at 298 and 373 K, and
the optimum superplastic temperature o f 673 K is chosen for both materials. As the strain
rate increases, there is a transition in the flow stresses only for alloys ECA pressed at 298
K. In the high strain regime (> 10'1 s'1 ), the m value decrease from -0.3 to -0.13. For the
Al-2024 alloy ECA pressed at 373 K, however, there is no transition and the m value is
-0.3. The reason for this may be related to the extension of the superplastic region II to
higher strain rates by ECA pressing at relatively high temperatures.
140
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
a (MPa)
1 0 0
Al-2024
T = 673 K
0.13
0.3
• Route Bc : 8p (298K )
▼ Route Bc : 8p (373K)
10
10°
6 (S 1)
Figure 5.23 The logarithm of flow stress as a function of the logarithm of strain rate for
Al-2024 alloys ECA pressed at 298 K and 373 K.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The measured m value o f -0.3 in the optimum superplastic regime II is consistent
with other reported data, such as m -0.39 for TMP treated Al-2024 alloy [86], m -0.43
for TMP treated A1-20I4 alloy [139], m — 0.5 for PM Al-2124-0.6Zr alloy [75] and m
-0.33 for PM Al-2024-3Fe-5Ni alloy [142]. For the MA 2024 alloys and Al-2024 based
MMC, where HSR SP occurs, the m values are close to or slightly higher than 0.3
[32,33,143,144], Therefore, the ECA pressed Al-2024 alloy can be regarded as a Class I
solid solution alloy [20]. Meanwhile, the measured m value for the annealed Al-2024 and
T351 heat-treated Al-2024 alloy are -0.26 and -0.16, respectively. This is also consistent
with the fact that the values of elongation of ECA pressed materials are superior to those
of the annealed and heat-treated Al-2024 alloys.
5.3 SUPRAL 100 ALLOY (Al-2004)
5.3.1 Introduction
Much attention has focused recently on procedures which may be used to increase
the strain rate and decrease the temperature for superplasticity in commercial aluminum-
based alloys. This interest arises because of the relatively slow forming rates and high
forming temperatures which are currently in use in conventional superplastic forming
processes. Since superplasticity is dependent upon the grain size of the material, it may
be possible to achieve these objectives by making a substantial reduction in the grain
size. At the present time, grain refinement is generally achieved through TMP but this has
142
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the limitation that different processing procedures must be developed separately for each
alloy and, in addition, it is usually not possible to reduce the grain size below -2-5 pm.
A British Al-2004 alloy, which known as SUPRAL 100, is a standard superplastic
material [1,4,145,146] widely used for commercial forming operations [2], When this
alloy is thermo-mechanically processed, the maximum elongation to failure has been
reported as — 1000% at 753 K with a strain rate of 3 x 10'3 s'1 [4].
The present investigation was therefore initiated in order to verify the effect of
ECA pressing on the subsequent superplastic properties o f the SUPRAL 100 alloy
without TMP and it was also intended to investigate the effect of the pressing
temperatures on micro structure and superplasticity of this commercial aluminum alloy.
5.3.2 Materials and experimental procedures
The chemical composition o f this alloy is Al-6% Cu-0.4% Zr and this alloy
contains extremely fine Al3 Zr precipitates and relatively coarse ALCu precipitates in A 1
matrix. Most of 9 phase (AbCu) dissolved into the matrix at temperatures of 673 to 773
K but the fine Al3 Zr particles are stable. In order to achieve adequate supersaturation of
zirconium, higher casting temperatures (— 1073 K) and faster solidification speed are
often used. After casting, the material is often aged to precipitate Al3 Zr, and then heavily
cold-worked to achieve a microstructure that is very sensitive to recrystallization [4,118].
The alloy used in this investigation was designed to make it suitable for ECA
pressing and it was cast, homogenized at 648 K for 5 hours and then hot rolled. Initially,
billets o f Al-2004 were pressed at 298 K using route Be up to 8 pressings. It was found
143
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
that this alloy tended to break after more than ~8 pressings and this is the same as the Al-
2024 alloy. Furthermore, in order to investigate the effect o f the pressing temperatures,
billets were pressed using route Be at 373 K up to 4, 8 and 12 pressings, at 473 K and 573
K up to 12 pressings, and finally at 673 K up to 6 pressings. After pressing, tensile tests
at high temperatures were conducted for the unpressed and pressed Al-2004 alloys.
5.3.3 Microstructures before and after ECA pressing
The micro structure of Al-2004 alloy before ECA pressing is shown in Fig. 5.24.
The grains are elongated and the average grain size is — 1.4 pm (long length is — 1.8 pm
and short length is — 1 pm). However these grains are in subgrain structure containing
A12C u and Al3 Z r precipitates because the SAED pattern with aperture size o f 12.3 pm is
nearly net and random. The initial average grain size measured by the optical microscope
before ECA pressing was -100 pm.
Fig. 5.25 shows microstructures of samples ECA pressed via route Be (a) at 298
K up to 8 passes, (b) at 373 K up to 8 passes and (c) at 473 K up to 12 passes. The grains
are reasonably uniform and equiaxed for every pressing condition and the average grain
sizes of samples ECA pressed at 298 K up to 8 passes, at 373 K up to 8 passes and at 473
K up to 12 passes are -0.2-0.3 pm, -0.4-0.5 pm and — 0.6-0.7 pm, respectively. The
specific features o f non-equilibrium grain boundaries [147], such as, extinction contour
and diffusive diffraction contrast, are clearly shown in Fig. 5.25 (a) and those features are
rarely seen in Fig. 5.25 (c) and reasonably well defined high angle grain boundaries are
also visible in Fig 5.25 (c).
144
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.24 Microstructure and associated SAED pattern of as-received Al-2004 alloy.
145
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.25 Microstructures of Al-2004 alloy after pressing (a) at 298 K up to 8 passes,
(b) at 373 K up to 8 passes and (c) at 473 K up to 12 passes.
146
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. 5.26 shows the microstructures o f samples ECA pressed via route Be (a) at
573 K up to 12 passes and (b) at 673 K up to 6 passes. The grains are reasonably uniform
and equiaxed for both pressing conditions and the average grain sizes of samples ECA
pressed at 573 K up to 12 passes and at 673 K up to 6 passes are ~0.9 pm and — 2 pm,
respectively. Although the ECA pressings were conducted at relatively high
temperatures, there remains a considerable density of dislocations both within the grains
and along the grain boundaries, as shown in Fig. 5.26 (a) and (b). This may be due to the
strong pinning effect of stable and fine Al3 Zr precipitates for the moving dislocation
during high temperature deformation. The SAED pattern with an aperture size o f 12.3 pm
in Fig. 5.26 (a) consists of rings o f diffraction spots showing that most grain boundaries
have high angles o f misorientation.
The Vickers microhardness, Hv, was measured under a load of 500 g applied for
15 seconds and the X plane was chosen for every samples. A hardness of 77.7 ± 2.4 was
obtained for unpressed Al-2024 alloy and the hardness increases as the number of
pressings increases, such as 116.0 ±1.5 for the sample ECA pressed at 298 K up to 4
passes to 117.4 ±1.9 for the sample ECA pressed at 298 K up to 8 passes. However, a
saturation in hardness occurs after 8 pressings, such as 111.7 ± 1.0 and 111.6 ± 0.9 for the
samples ECA pressed at 373 K up to 8 passes and up to 12 passes, respectively. The
variation of microhardness as a function of pressing temperature is shown in Fig. 5.27
where the measured grain sizes are also included. It is clear that the ECA pressing at high
temperatures can produce a relatively large grain size and a corresponding decrease in
strength by the Hall-Petch relationship.
147
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.26 Microstructures o f Al-2004 alloy after pressing (a) at 573 K up to 12 passes
and (b) at 673 K up to 6 passes.
148
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
< D
N
't o
c
to
C D
3 130
Al-2004
— Grain size
■ o — Microhardness
120
2
100
1
0 L
200 400 500 700 300 600
ECA pressing temperature (K )
Figure 5.27 Variations of grain size and microhardness as a function of ECA pressing
temperature for Al-2004 alloy.
149
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
M icrohardness (H v)
5.3.4 Results of tensile testing at high temperatures
All testing results for unpressed and ECA pressed Al-2004 alloys are summarized
in Table 5.9. For the unpressed material, a maximum elongation o f -450% was obtained
at a temperature of 773 K and a strain rate o f 3.3 x 10-4 s'1 . Therefore, it can be proven
that the as-received Al-2004 alloy in this investigation is initially superplastic without
ECA pressing but at very high temperatures and low strain rates.
After ECA pressing, the elongations to failure in the Al-2004 alloy were
significantly higher than in the unpressed condition, and furthermore, HSR SP occurs for
every ECA pressed sample. Exceptional elongations of -860% and -1070% were
attained at a strain rate o f 10'2 s'1 with a testing temperature of 673 K after ECA pressing
for 12 passes using route Be at 473 K and 573 K, respectively. The largest elongation of
— 1100% was recorded when testing at a relatively high temperature o f 723 K using a
strain rate of 10'2 s'1 after ECA pressing at 673 K via route Be up to 6 pressings.
In addition, Fig. 5.28 and Fig. 5.29 show the appearance o f the tensile specimens
which were ECA pressed at 573 K up to 12 passes and at 673 K up to 6 presses,
respectively. It is apparent that there is uniform deformation within the gauge length
without any localized necking especially for the highly elongated specimens.
Fig. 5.30 shows typical stress vs. strain curves for the Al-2004 alloy ECA pressed
at 573 K via route Be up to 12 passes. After a short strain hardening, extensive strain
softening up to failure occurs for every sample, even when testing at different
temperatures and strain rates. The deformation behavior is almost identical for other Al-
2004 alloys ECA pressed at the different pressing temperature and numbers of pressings.
L50
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 5.9 Tensile test results at high temperatures for unpressed and ECA pressed.
Al-2004 alloys
Treatment
Temp.
(K)
Strain rate
(sec'1 )
YS
(MPa)
UTS
(MPa)
EF
(%)
623 IO * 3 36 47 126
673 10'2 30 47 171
673 10'3 19 31 187
723 10'2 19 38 279
723 10'3 11 21 244
753 3.3 x 10'3 11 22 338
Unpressed
773 10'1 24 43 204
773 10'2 13 25 261
773 3.3 x 10'3 8 19 380
773 10'3 7 12 432
773 3.3 x 10-4 5 8 448
773 io- 4 4 5 317
573 10'2 30 75 254
623 10'1 45 76 222
ECA pressed
673 IO'1 29 57 364
at 298 K up to
673 10'2 13 30 567
8 passes
693 10'1 31 49 321
723 IO'2 11 24 524
623 10'3 11 27 402
673 10'2 15 32 478
ECA pressed
673 10'3 8 17 442
at 373 K up to
723 10'2 14 25 377
4 passes
723 10'3 5 9 719
723 10'3 6 10 237
623 10'2 28 74 287
623 10'3 11 26 380
673 10'1 25 65 419
673 3.3 x 10'2 17 44 627
ECA pressed 673 10'2 13 30 544
at 373 K up to 673 3.3 x 10'3 8 19 774
8 passes 673 10'3 8 22 486
673 IO * 4 4 9 311
723 10'2 17 35 287
723 10'3 7 13 340
773 10'3 4 7 220
(continued)
L 51
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 5.9 Tensile test results at high temperatures for unpressed and ECA pressed
Al-2004 alloys
Temp. Strain rate YS UTS EF
l reaiment
(K) (sec'1 ) (MPa) (MPa)
(%)
623 10'2 18 52 344
ECA pressed 623 IO'3 10 26 438
at 373 K up to 673 10'2 16 37 437
12 passes 673 10'3 8 17 427
723 10'3 8 15 301
623 IO'2 18 50 516
ECA pressed 623 10'3 11 24 453
at 473 K up to 673 IO *1 27 63 434
12 passes 673 I O '2 12 26 858
673 10'3 6 14 665
623 10'2 23 48 418
ECA pressed 673 IO'1 28 63 331
at 573 K up to 673 3.3 x 10'2 19 42 527
12 passes 673 10'2 12 25 1070
673 10'3 6 13 731
623 IO'2 29 54 245
673 10'2 16 31 582
ECA pressed 698 IO'2 14 27 649
at 673 K up to 723 10'1 24 47 376
6 passes 723 3.3 x 10‘2 15 31 553
723 10'2 10 22 1095
723 10'3 5 9 523
152
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Al-6%Cu-0.4%Zr
Initial sample
Figure 5.28 Examples of tensile ductility in samples ECA pressed at 573 K via route Be
up to 12 passes and pulled at 673 K.
153
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Al-6%Cu-0.4%Zr
Initial sample
Figure 5.29 Examples of tensile ductility in samples ECA pressed at 673 K via route Bc
up to 6 passes and pulled at 723 K.
154
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
a (MPa)
1 Al-2004 ECA pressed at 573 K via route Bc up to 12 passes J
70
T(K) s (s'1 )
623 1 .0 x 1 0
673 1 .0 x 1 0
673 3 .3 x 1 0
673
673
60
50
40
1.0 x 10
30
20
10
0
0 200 400 800 1200 1000 600
S (%)
Figure 5.30 The stress vs. strain curves of Al-2004 alloys tested at 623 K and 673 K
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5.3.5 Microstructures after tensile testing
Fig. 5.31 shows the microstructures of samples after tensile testing to failure. In
Fig. 5.31, (a) is a TEM image taken from within the undeformed grip region of the
sample and (b) is an SIM image taken from the fracture tip. For each o f the SIM images,
the surface of the sample lies close to the right hand edge of the photomicrograph and the
direction of the tensile axis is indicated.
Fig. 5.31 was taken on the sample pressed at 573 K up to 12 passes and tested at
673 K at a strain rate o f 1 C T 2 s'1 to an elongation o f— 1070%. It is evident from Fig. 5.31
(a) that the grain size has increased from -0.9 to -2.6 pm in the grip region during tensile
test although the grains remain essentially equiaxed. Near the fracture tip, shown in Fig.
5.31 (b), however, the grains are slightly elongated with an average size o f-4 pm (long
length is -4.6 pm and short length is — 3.5 pm). Thus, it is clear that stress-enhanced grain
growth, namely dynamic recrystallization, occurred within the gauge length but static
recrystallization occurred within the grip region during tensile deformation.
Since some evidence of intergranular cracking is found within the sample, as
shown in Fig. 5.31 (b), it is believed that the fracture mechanism o f Al-2004 alloy ECA
pressed at 573 K via route Be to a strain of -12 is the nucleation, growth and coalescence
o f internal cavities along the grain boundaries. This is also confirmed by the fact that the
value of the stress at fracture is relatively high as shown in Fig. 5.30 and the fracture tip
after testing in tension to failure is not in the form o f a pinpoint as shown in Fig. 5.28.
The same results were obtained for the sample pressed at 673 K up to 6 passes and tested
at 723 K at a strain rate of 10'2 s'1 to an elongation o f-1100%.
156
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission.
edge surface—-
Figure 5.31 Microstructure o f Al-2004 alloy prepared by ECA pressing and then tested to failure in tension at 673 K
with a strain rate of 10'2 s'1 : (a) grip region; (b) the fracture tip within the gauge length.
U l
-sJ
5.3.6 Discussion
5.3.6.1 The effect of ECA pressing on optimum superplastic conditions
In this investigation, the as-received Al-2004 alloy showed the maximum
elongation to failure of -450% when testing at 773 K with the strain rate of 3.3 x I O '4 s'1 .
When this alloy was thermo-mechanically processed, the maximum elongation to failure
was reported as -1000% at 753 K with a strain rate o f 3 x 10'3 s'1 [4]. There is no doubt
that the discrepancy in values of elongation is caused by the different processing
procedures and the resultant different microstructures, especially the grain size. The grain
size of the unpressed Al-2004 alloy is much larger than that o f the reported alloy,
typically — 4-5 pm. However, it is evident that optimum superplasticity was obtained at
very high temperatures and low strain rates for both cases.
Earlier reports demonstrated the potential for using severe plastic deformation to
reduce the temperature for superplasticity [147] and also to achieve superplasticity at
very rapid strain rates [15,81,148], Close inspection of Table 5.9 reveals that the
optimum superplastic conditions after ECA pressing are consistently achieved at a
temperature of 673 K which is 100° lower than in the unpressed conditions. Furthermore,
all samples after ECA pressing exhibit HSR SP with good tensile ductilities at a strain
rate of 10'2 s'1 . This is clearly seen in Fig. 5.32. The elongations are also high at 673 K
with a strain rate of 10'1 s'1 , ranging from — 360% after pressing at room temperature via
route Be to -430% after pressing using the same route at 473 K. Although the strain
applied to the Al-2004 alloy during ECA pressing at 673 K was not enough, it was
remarkable to achieve a maximum elongation o f -1100% at 723 K and 10'2 s'1 .
158
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
E longation t o fa ilu r e (% )
1600
: Al-2004
TflQ
773
673
1400
Unpressed
Route Bc : 8p (373K)
Route Bc : 12p (473K) 673
Route Bc : 12p (573K) 673
1200
1000
800
600
400
-O 200
10-4 10-3 - i 0-2 10-'
Strain rate (s'1 )
Figure 5.32 Elongation to failure vs. strain rate under optimum conditions for Al-2004
alloy.
.159
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
S.3.6.2 The effect of ECA pressing on superplasticity of Al-2004 alloy
Close inspection o f Table 5.9 reveals the following facts. When an Al-2004 alloy
was ECA pressed at 373 K using route Be, the values o f elongation to failure are
increased as the number o f pressings increases up to 8, but the elongation decreased after
ECA pressing for 12 passes at the same testing conditions. The reason for this may be
related to the pressing temperature. The pressing temperature can control the number of
pressings. For example, for ECA pressing at room temperature, the maximum applicable
number of passes was 8 for Al-2004 alloy and further pressings cannot be conducted due
to the occurrence of an unsound surface and propagation of a crack. For ECA pressing at
373 K, the maximum applicable number of pressings may be 12 and it may not be
possible to press up to higher number o f pressings than 12. Therefore, at this stage, it is
possible to introduce cavities or micro-cracks inside the pressed billet, a homogeneous
microstructure was partially destroyed because of the slow recovery rate at the low
pressing temperature, and the material tends to break before reaching a modest ductility.
The development of extensive cavitation during superplastic deformation at high
temperatures is well documented for the SUPRAL 100 alloy. The above consideration is
also consistent with the previous results of Al-2024 alloy in section 5.2. When ECA
pressing was conducted at 373 K, further pressing up to 12 passes showed no significant
advantage over 8 passes.
Another interesting fact found in Table 5.9 is that the largest elongation to failure
of -1100% was obtained when testing at 723 K using a strain rate o f 10'2 s'1 after ECA
pressing at 673 K via route Be up to 6 pressings. The number of pressings, namely the
160
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
strain introduced in the material, may play an important role in determining the optimum
superplastic conditions. Close inspection of the optimum superplastic temperatures in
Table 5.9 shows that 673 K is the optimum temperature if the material is pressed up to 8
or 12 passes but 723 K is the optimum temperature if the material is pressed up to 4 or 6
passes. In order to achieve superplasticity at low temperatures, therefore, the higher strain
should be introduced in the material. For the Al-2004 alloy and many alloys containing
very fine particles, it is known that dynamic recrystallization plays an important role in
achieving superplasticity [150,151]. As the number of pressings increases, dynamic
recrystallization may also occur during ECA pressing or a microstructure, which is very
susceptible to the following dynamic recrystallization in tension, can be achieved during
ECA pressing. During superplastic deformation in tension, therefore, dynamic
recrystallization and subsequent GBS can occur easily at relatively low temperatures if
the number o f pressings is high enough to produce the optimum micro structure for
superplasticity. Thus, for the Al-2004 alloy ECA pressed at 673 K up to 6 pressings, it is
anticipated that the optimum superplastic temperature of the Al-2004 alloy will be shifted
to lower temperatures after ECA pressing up to 8 or 12 passes.
When the Al-2004 alloy was pressed via route Be to a total strain of — 12 with 8
passes at 673 K and 4 additional passes at 473 K, HSR SP was obtained at very low
temperatures o f 573 and 623 K [15], namely a maximum elongation of -970% was
recorded when testing at 573 K with a strain rate of 10'2 s’1 and that of -740% was
obtained when testing at 623 K with a strain rate of 10'1 s'1 . When a direct comparison is
made between the above reported results and the results listed in Table 5.9 (especially,
161
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ECA pressing at 673 K up to 6 passes), HSR SP favorably occurred for the former alloy
regardless o f the higher elongation o f the latter alloy. There is no doubt that the former
alloy has a finer grain size (-0.5 pirn) than the latter alloy (-2 pm) and this difference in
grain sizes can produce the different optimum superplastic conditions for each alloy.
5.4 Al-1420 ALLOY
5.4.1 Introduction
In order to achieve superplasticity at high strain rates and at low temperatures,
earlier reports demonstrated the potential for using ECA pressing to reduce the grain size
of materials [81,128,147], It was also understood that the fine grains achieved by ECA
pressing should be stable at elevated temperatures otherwise a premature fracture occurs
due to extensive grain growth. An Al-1420 alloy, one o f the candidate materials for use in
superplastic forming operations, showed very stable grains up to 700 K after ECA
pressing [80,126,127,152,153] with the highest elongation o f — 1180% without failure at a
testing temperature of 623 K and an initial strain rate o f 10'2 s'1 after ECA pressing via
route A [80,81,128,148], Since detailed investigations have revealed that route Be leads
most rapidly to a uniform micro structure of equiaxed grains separated by high angle grain
boundaries [100,102], the route Be was chosen in this investigation in order to understand
the effect of pressing routes and to evaluate HSR SP and LT SP (Low temperature
superplasticity) potential of this commercial Al-Mg-Li-Zr alloy.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5.4.2 Materials and experimental procedures
Al-1420 alloy, Al-5.5% Mg-2.2% Li-0.12% Zr alloy, is a commercial cast
aluminum alloy and contains fine precipitates of S'-A^Li and fT-A^Zr. The alloy was
received in a non-superplastic condition after hot rolling and its initial grain size was
-400 pm.
Samples were subjected to ECA pressing using route Be under six different
conditions: (1) 4 passes at 673 B C , (2) 4 passes at 673 K and an additional 2 passes at 473
B C , (3) 4 passes at 673 K and an additional 4 passes at 473 B C , (4) 8 passes at 673 B C , (5) 8
passes at 673 K and an additional 2 passes at 473 K and (6) 8 passes at 673 K and an
additional 4 passes at 473 K. Following the ECA pressing, samples were examined by
TEM and SIM and tested in tension at high temperatures.
5.4.3 Microstructures after ECA pressing
Fig. 5.33 shows microstructure o f the Al-1420 alloy after pressing via route Be
for 8 passes at 673 K and a further 4 passes at 473 K. This sample is selected because an
exceptional HSR SP was found in this sample and this will be discussed at the next
section. Inspection o f Fig. 5.33 shows that the grains are reasonably equiaxed but there
was a high density o f dislocations both within the grains and along the grain boundaries.
The measured average grain size was -0.8 pm. The SAED pattern with an aperture size
of 12.3 pm consists of rings of diffraction spots showing that the grain boundaries have
high angles of misorientation.
163
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.33 Microstructure and associated SAED pattern o f Al-1420 alloy after ECA
pressing to a strain o f ~12.
164
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. 5.34 shows the microstructure at a high magnification for the same sample
taken in Fig. 5.33 and the aperture size of the SAED pattern is 1.25 pm. The same region
is shown (a) with a bright field image and (b) with a (100) dark field image. It is obvious
from Fig. 5.34 (a) that there are small P'-AhZr precipitates within the grains and very
fine S'-A^Li precipitates are uniformly distributed within the grains and their measured
average size is ~3 nm. It is believed that 8 '-Al3 Li particles are precipitated during the air
cooling following ECA pressing.
5.4.4 Results of tensile testing at high temperatures
All testing results for ECA pressed Al-1420 alloys are summarized in Table 5.10.
Additional tensile tests at elevated temperatures were conducted for the unpressed Al-
1420 alloy. The test conditions are the same as the tests in Table 5.10, such as
temperatures of 623, 673 and 723 K and strain rates o f 10~2 and 10' 3 s'1 . Most tensile test
results showed that the maximum elongations were less than 2 0 0 %, and the best
elongation was — 225% when testing at 723 K and 10' 3 s'1 . This result is consistent with
the fact that a higher temperature and a slower strain rate are favorable for conventional
superplasticity. Therefore, it is revealed that the unpressed Al-1420 alloy is not
superplastic because its grain size is -400 pm.
In table 5.10, the largest elongations to failure for most materials are often
obtained when testing at 673 K but the optimum superplastic temperature is 723 K for the
Al-1420 alloy ECA pressed at 673 K up to 4 passes. The strain rate of 10' 2 s' 1 can be
regarded as the optimum strain rate for most Al-1420 alloys.
165
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
b
• 200
100
• 000
0.2 [im
11 1
t n t t W P I R M a S R
Figure 5.34 Microstructure at a high magnification o f Al-1420 alloy after ECA pressing
to a strain o f— 12: (a) bright field image and (b) (100) dark field image.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 5.10 Tensile test results at high temperatures for ECA pressed Al-1420 alloys
Treatment
Temp. Strain rate YS UTS EF
(K) (sec'1 ) (MPa) (MPa)
(%)
573 10'1 108 209 154
573 10'2 69 125 243
573 10'3 36 69 321
623 10'1 76 106 291
ECA pressed
up to 4 passes
at 673 K
623
623
673
10'2
10'3
10'1
47
27
30
73
47
69
366
578
425
673 10‘ 2 16 36 955
673 10'3 6 12 761
723 10'1 18 46 571
723 10'2 6 19 1166
723 10'3 2 7 1165
573 10'L 99 162 187
573 10'2 44 65 489
573 10‘3 32 50 405
ECA pressed
up to 4 passes
at 673 K
+
up to 2 passes
at 473 K
623
623
623
10'1
10'2
10'3
42
22
8
74
39
15
501
646
513
673 10'1 32 64 453
673
673
O O
12
5
25
10
714
701
673 io- 4 3 5 318
723 10'1 36 56 259
723 10'2 19 28 250
723 10'3 11 15 275
573 10'1 99 170 150
573 10'2 31 75 458
573 10'3 22 45 330
ECA pressed 623 10'1 51 73 357
up to 4 passes 623 I O '2 26 42 404
at 673 K 623 I O '3 9 18 448
+ 673 10'1 39 65 296
up to 4 passes 673 10'2 12 24 659
at 473 K 673 10'3 5 11 420
723 10'1 30 40 126
723 IO * 2 16 24 228
723 10'3 9 13 185
(continued)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 5.10 Tensile test results at high temperatures for ECA pressed Al-1420 alloys
Treatment
Temp. Strain rate YS UTS EF
(K) (sec'1 ) (MPa) (MPa) (%)
573 10'1 106 194 207
573 IO'2 39 101 377
573 10'3 26 42 535
623 IO'1 59 105 372
ECA pressed
up to 8 passes
at 673 K
623
623
673
673
10'2
10'3
10'1
I O '2
15
7
23
8
46
18
60
21
852
889
754
1249
673 10'3 4 7 1139
723 10'1 31 47 350
723 IO'2 6 17 1104
723 IO'3 10 15 296
573 IO'1 93 159 211
573 I O '2 37 68 436
573 I O '3 16 32 545
ECA pressed 623 I O'1 69 128 276
up to 8 passes 623 10'2 27 70 417
at 673 K 623 IO'3 13 27 570
+
673 10'1 24 64 458
up to 2 passes 673 10'2 10 24 476
at 473 K 673 1 0*3 6 11 572
723 10'1 20 50 377
723 I O '2 6 25 452
723 10'3 4 12 371
573 10'1 89 127 348
573 I O '2 31 50 756
573 10'3 22 30 407
623 1 59 109 546
ECA pressed
up to 8 passes
at 673 K
+
up to 4 passes
at 473 K
623
623
623
10'1
10'2
10'3
26
11
5
44
18
10
997
936
507
673 1 31 70 945
673
673
I O'1
10‘2
12
9
28
16
1214
837
673 10'3 4 8 680
723 1 51 70 346
723 IO'1 13 30 1088
723 10'2 7 15 917
723 IO'3 3 7 578
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The largest elongation to failure of -1250% was obtained for alloys ECA pressed
up to 8 passes at 673 K when testing at 673 K and 10'2 s'1 . An exceptional HSR SP was
obtained for Al-1420 alloy ECA pressed for 8 passes at 673 K and an additional 4 passes
at 473 K. For example, elongations o f — 1210% and -950% were recorded at a
temperature of 673 K when testing at 10'1 s'1 and at 1 s'1 , respectively. The tested samples
are shown in Fig. 5.35 and it is evident that there is uniform deformation within the gauge
length without the formation o f necking.
Fig. 5.36 shows typical stress vs. strain curves for the Al-1420 alloy ECA pressed
via route Be up to 8 passes at 673 K and an additional 4 passes at 473 K when testing at
673 K with various strain rates. After a short strain hardening, extensive strain softening
or steady state up to failure occurs for every sample, even when testing at different
temperatures and strain rates. The deformation behavior in this investigation is not
consistent with the reported results [154,155] where stress-strain curves for Al-1420 alloy
produced by severe plastic deformation at low temperatures when testing at elevated
temperatures (523, 573 and 623 K) and at a strain rate of 10'1 s'1 showed significant work
hardening and they also demonstrated that the flow stresses were relatively higher than
those of conventional superplastic materials. However, as clearly shown in Fig. 5.36, the
true flow stress-elongation curves are almost identical to those o f other A 1 alloys showing
an exceptional elongation of more than 1000% such as Al-Mg-Sc-Zr and Al-2004 alloys.
Furthermore, the flow stresses at 673 K are not higher than those o f other alloys (i.e. Al-
Mg-Sc-Zr and Al-2004 alloys). Even when the testing condition is the same (i.e. at 623 K
and 10'1 s'1 ), an extensive strain hardening is never found in the present materials.
169
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.35 Examples o f tensile ductility in samples ECA pressed via route Be up to 8
passes at 673 K and an additional 4 passes and pulled at 673 K.
170
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
c r (MPa)
AI-5.5Ma-2.2Li-0.12Zr
Route Bc ; 8p(673K ) + 4p(473K)
T = 673 K
60
x 10
x 10
x 10
x 10'
40
20
800 1200 200 400 600 1000 0 1400
S (%)
Figure 5.36 The stress vs. strain curves o f Al-1420 alloy tested at 673 K.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5.4.5 Microstructures after tensile testing
Fig. 5.37 and Fig. 5.38 illustrate the microstructures o f samples after tensile
testing. In these figures, (a) is a TEM image taken from the undeformed grip area and (b)
is an SIM image taken from the deformed fracture tip. For each o f the SIM images, the
surface of the sample lies close to the right hand edge o f the photomicrograph and the
direction of the tensile axis is indicated.
Fig. 5.37 was taken on the sample pressed up to 12 passes at 673 K and an
additional 4 passes, and tested at 673 K at a strain rate o f IO'1 s'1 to an elongation of
— 1210%. It is obvious from Fig. 5.37 (a) that the grain size has increased from -0.8 to
-1.4 pm in the grip region during tensile test although the grains remain essentially
equiaxed. Near the fracture tip, shown in Fig. 5.37 (b), the grains are also equiaxed with
an average size of — 2.5 pm. Thus, it is apparent that stress-enhanced grain growth
occurred within the gauge length. Furthermore, although very rapid strain rate has been
applied, extensive intergranular cracks are developed within the gauge length.
Fig. 5.38 was taken on the same sample used in Fig. 5.37 but tested at 673 K at a
strain rate of 1 s'1 to an elongation o f 950%. The grains are equiaxed and the measured
grain size has increased slightly from — 0.8 to — 1.2 pm in the grip region during a very
short tensile test as shown in Fig. 5.38 (a). The grains within the gauge length are also
equiaxed with an average size of 1.8 pm. Any evidence of intergranular cracking near the
fracture tip has not been found in Fig. 5.39 (b).
172
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
inside surface
Figure 5.37 Microstructure of Al-1420 alloy prepared by ECA pressing and then tested to failure in tension at 673 K
with a strain rate o f IO"1 s'1 : (a) grip region; (b) the fracture tip within the gauge length.
h - *
U >
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
surface
inside
Figure 5.38 Microstructure of Al-1420 alloy prepared by ECA pressing and then tested to failure in tension at 673 K
with a strain rate of 1 s'1 : (a) grip region; (b) the fracture tip within the gauge length.
5.4.6 Discussion
5.4.6.1 Superplasticity in Al-1420 alloy
In order to examine the superplastic properties o f the ECA pressed Al-1420
alloys, the plots o f elongation to failure vs. strain rate for the Al-2024 alloy pressed for 4
passes at 673 K, pressed for 4 passes at 673 K plus 2 passes at 473 K, pressed for 4
passes at 673 K plus 4 passes at 473 K, pressed for 8 passes at 673 K, pressed for 8
passes at 673 K plus 2 passes at 473 K and pressed for 8 passes at 673 K plus 4 passes at
473 K are shown in Fig. 5.39, Fig. 5.40, Fig. 5.41, Fig. 5.42, Fig. 5.43 and Fig. 5.44,
respectively.
When ECA pressing was conducted up to 4 passes at 673 K, as shown in Fig.
5.39, maximum elongations of > 1100% were obtained at the highest test temperature of
723 K and at strain rates of 10'2 and 10° s'1 . This is a surprising result that even this small
number of passes can make a material highly superplastic. When an additional 2 or 4
passes at 473 K was applied, however, the total elongations dropped by almost 50%, as
shown in Fig. 5.40 and Fig. 5.41. One interesting fact is that the maximum elongation
was obtained at the highest testing temperature of 723 K for the Al-1420 alloy ECA
pressed for only 4 passes at 673 K but the elongations were very poor at this testing
temperature when an additional 2 or 4 passes were conducted at 473 K. Furthermore,
when comparing all data of elongation to failure in Table 5.10 between additional 2
passes at 473 K and 4 passes at 473 K, the Al-1420 alloy ECA pressed for 4 passes at 673
K plus an additional 2 passes at 473 K always exhibited higher elongations than Al-1420
alloy ECA pressed for 4 passes at 673 K plus an additional 4 passes at 473 K.
175
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1600
: AI-5.5Ma-2.2Li-0.12Zr
; Route Bc ; 4p(673K )
1400
T IK I
573
623
673
723
1200
= 3 1000
800
c
600
cn
400
200
1 0 -5 1 0 -4 1 0 -3 i 0 -2 10-1 1Q0 1Q1
Strain rate (s'1 )
Figure 5.39 Elongation to failure vs. strain rate for Al-1420 alloy ECA pressed via route
Be up to 4 passes at 673 K.
176
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1600
AI-5.5Ma-2.2Li-0.12Zr
; Route Bc ; 4p(673K ) + 2p(473K)
1400
i m
573
623
673
723
1200
3 1000
800
c
o
600
200
10°
Strain rate (s'1 )
Figure 5.40 Elongation to failure vs. strain rate for Al-1420 alloy ECA pressed via route
Be up to 4 passes at 673 K and an additional 2 passes at 473 K.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1600
: AI-5.5Mo-2.2Li-0.12Zr
: Route Bc ; 4p(673K ) + 4p(473K )
1400
i m
573
623
673
723
1200
3 1000
800
600
O)
400
200
10°
Strain rate (s'1 )
Figure 5.41 Elongation to failure vs. strain rate for Al-1420 alloy ECA pressed via route
Be up to 4 passes at 673 K and an additional 4 passes at 473 K.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1600
: AI-5.5Ma-2.2Li-0.12Zr
; Route Bc ; 8p(673K )
1400
TIKI
573
623
673
723
1200
o
a >
B
800
C
0
“ c O 600
CD
1 4 0 0
200
10'1 10°
Strain rate (s'1 )
Figure 5.42 Elongation to failure vs. strain rate for Al-1420 alloy ECA pressed via route
Be up to 8 passes at 673 K.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1600
: AI-5.5Ma-2.2Li-0.12Zr
: Route Bc ; 8p(673K) + 2p(473K )
1400
TOO
573
623
673
723
1200
a>
k_
3
1000
o
800
c:
o
600
a s
c n
c
o
L U
400
200
10°
Strain rate (s'1 )
Figure 5.43 Elongation to failure vs. strain rate for Al-1420 alloy ECA pressed via route
Be up to 8 passes at 673 K and an additional 2 passes at 473 K.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1600
: AI-5.5Ma-2.2Li-0.12Zr
: Route Bc ; 8p(673K) + 4p(473K )
TJKl
■ 573
623
673
723
1400
1200
< D
=3 1000
o
800
C
o
(0
O)
600
c:
o
LU
400
200
10°
Strain rate (s'1 )
Figure 5.44 Elongation to failure vs. strain rate for Al-1420 alloy ECA pressed via route
Be up to 8 passes at 673 K and an additional 4 passes at 473 K.
1 8 1
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
For an Al-1420 alloy ECA pressed up to 8 passes at 673 K, as shown in Fig. 5.42,
the largest elongation to failure was obtained at 673 K and 10~2 s'1 . This optimum
temperature is lower than that for the Al-1420 alloy ECA pressed up to 4 passes at 673 K
so that it is revealed that the optimum superplastic temperature is decreased as the
number of passes is increased. This trend is the same as the previous result for Al-2004
alloy. Furthermore, the values of elongation to failure for 8 passes at 673 K are always
larger than those for 4 passes at 673 K.
For an Al-1420 alloy ECA pressed for 8 passes at 673 K and an additional 2
passes at 473 K, as shown in Fig. 5.43, it is surprising that all values of elongation to
failure are very poor. The reason for this may be that the smooth structure of the grain
boundaries formed by 8 pressing at 673 K can be destroyed after a small number of
passes at 473 K.
For an Al-1420 alloy ECA pressed for 8 passes at 673 K and an additional 4
passes at 473 K, however, the values o f elongation to failure were fully improved to the
best results as shown in Fig. 5.44. Especially for three testing temperatures of 623, 673
and 723 K, the maximum elongations to failure of more than 1000% were recorded at a
high strain rate o f 10'1 s'1 , and a maximum elongation of -950% was recorded when
testing at 673 K and at a very rapid strain rate of 1 s'1 . Meanwhile, this result is slightly
different from that for the same material but using a different pressing route [156]. When
ECA pressing was conducted using route A, an elongation o f more than 1180% was
obtained at a temperature of 623 K and at a strain rate o f 10'2 s'1 , and an elongation of
-910% was obtained at the same testing temperature but at a faster strain rate of 10'1 s'1 .
182
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Therefore, the small grain size of — 0.8 pm after pressing by route Be is relatively
favorable to show HSR SP than the slightly larger grain size o f -1.2 pm after pressing by
route A.
S.4.6.2 HSR SP through ECA pressing
An elongation o f -950% was recorded in this investigation for a sample tested at
673 K with a strain rate o f 1 s'1 after ECA pressing via route Be up to 8 passes at 673 K
and an additional 4 passes at 473 K. This result appears to exceed any reported elongation
at this strain rate for the standard MMCs and PM aluminum based alloys showing HSR
SP.
In order to achieve HSR SP, a fine-grained microstructure is necessaiy.
Introduction of an extremely fine grain size through ECA pressing is not a sufficient
condition to accomplish HSR SP because it is also essential that the grains remain
reasonably stable at the high temperatures required for diffusion-controlled superplastic
flow. Since grain growth is restricted in the commercial Al-1420 alloy through the
existence of a fine dispersion of p'-A^Zr precipitates, therefore, this alloy is capable of
exhibiting HSR SP.
It should be noted that TMP has two distinct disadvantages by comparison with
the utilization of an ECA pressing technique. First, the recrystallized grain sizes are
generally considerably higher after TMP by comparison with the ultrafine grain sizes
which are achieved in ECA pressing and this means that the superplastic range of strain
rates will tend to be lower than those required for HSR SP in thermomechanically
183
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
processed materials. Second, the development of an appropriate TMP for any selected
alloy is time consuming since it depends critically upon the precise composition and the
nature of the micro structure in the unprocessed material. On the other hand, ECA
pressing is not only a relatively simple technique which can be applied to many different
materials without any significant changes but it is also a very powerful technique which
can control the microstructure of materials in a favorable manner by adjusting some of
the variables during the pressing operations.
184
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6. SUMMARY AND CONCLUSIONS
1. Since ECA pressing offers significant advantages for grain refinement and meets the
requirements for superplasticity, it has been chosen as a basic tool of material
processing for this study. High temperature ECA pressing facilities were successfully
installed at USC and the upper limit of pressing temperature is about 523 K.
2 . Pure aluminum (99.99%) was chosen as a model material and pressed to investigate
the soundness o f the pressed billets with the number of pressings by measuring
microhardness. The variations o f microhardness are large throughout the length of the
pressed billet for lower numbers of passes but a very uniform microstructure is
obtained throughout the billet after higher numbers of passes.
3. Two new combination pressing routes, C-Bc and Bc-C, were designed, applied to
both pure aluminum and AI-2024 commercial alloy and compared to route Be- It is
concluded that route Be is the most reliable pressing route because the two
combination routes have no clear advantage in terms of superplasticity over route Be-
4. The ECA pressing at high temperatures is necessary to press strong commercial
aluminum alloys to a higher strain and there is also a limitation in the possible
number of pressings when aluminum alloys were pressed at room temperature. The
ECA pressing at high temperatures always produced superior superplasticity by
comparison with the ECA pressing at room temperature.
5. In the experimental Al-0.2% Sc-0.12% Zr and Al-3% Mg-0.2% Sc-0.12% Zr alloys,
the role of Sc is to form a fine dispersion of AI3 SC particles to inhibit significant grain
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
growth at high temperatures where superplasticity is expected. The addition of both
Sc and Z r at the same time is more effective for superplasticity than the addition of
only one element.
6. The deformation mechanism for the ECA pressed Al-Mg-Sc-Zr alloy is grain
boundary sliding accommodated by dislocation glide. The measured strain rate
sensitivity, m, o f — 0.36 is higher than 0.33 for the solute drag mechanism and the
activation energy is slightly smaller than the anticipated value for the impurity
diffusion o f Mg in Al.
7. After ECA pressing of the Al-2024 and the AI-2004 alloys, the optimum superplastic
conditions are shifted to both a lower testing temperature and a higher imposed strain
rate.
8. An exceptional HSR SP was obtained for every aluminum alloy.
• For the Al-Mg-Sc-Zr alloy ECA pressed via route Be up to 6 passes at 298 K,
an elongation to failure of -1680% was recorded at a temperature of 773 K
with a strain rate of 10'2 s'1 .
• For the Al-2024 alloy ECA pressed at via route Be up to 8 passes at 373 K, an
elongation to failure of -500% was recorded at a temperature o f 673 K with a
strain rate of 10'2 s'1 .
• For the Al-2004 alloy ECA pressed via route Be up to 12 passes at 573 K, an
elongation to failure o f — 1070% was recorded at a temperature of 673 K with
a strain rate of 10'2 s'1 .
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
• For the Al-1420 alloy ECA pressed via route Be up to 8 passes at 673 K and
an additional 4 passes at 473 K, elongations to failure o f — 1210% and -950%
were recorded at a temperature o f 673 K with strain rates o f 10'1 s’1 and 1 s'1 ,
respectively.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
REFERENCES
1. O.D. Sherby and J. Wadsworth, Mater. Sci. Tech. 1, 925 (1985).
2. A.J. Barnes, Mater. Sci. Forum, 170-172, 701 (1994).
3. J.W. Edington, Met. Tech. 138 (1976).
4. R. Grimes, M.J. Stowell and B.M. Watts, Met. Tech. 154 (1976).
5. T.G. Langdon, Metall. Trans. 13A, 689 (1982).
6. O.D. Sherby and J. Wadsworth, Prog. Mater. Sci. 33, 169 (1989).
7. T.G. Langdon, Mater. Sci. Eng. A137, 1 (1991).
8. T.G. Langdon, T. Watanabe, J. Wadsworth, M.J. Mayo, S.R. Nutt and M.E.
Kassner, Mater. Sci. Eng. A166, 237 (1993).
9. A.H. Chokshi, A.K. Mukheijee and T.G. Langdon, Mater. Sci. Eng. RIO, 237
(1993).
10. M.M.I. Ahmed and T.G. Langdon, Metall. Trans. 8A, 1832 (1977).
11. K. Higashi, T. Ohnishi and Y. Nakatani, Scripta Metall. 19, 821 (1985).
12. T.G. Nieh, C.A. Henshall and J. Wadsworth, Scripta Metall. 18, 1405 (1984).
13. T.G. Nieh, P.S. Gilman and J. Wadsworth, Scripta Metall. 19, 1375 (1985).
14. R.Z. Valiev, Mater. Sci. Forum, 243-245, 207 (1997).
15. R.Z. Valiev, D.A. Salimonenko, N.K. Tsenev, P.B. Berbon and T.G. Langdon,
Scripta Mater. 37, 1945 (1997).
16. R.Z Abdulov, R.Z. Valiev and N.A. Krasilnikov, J. Mater. Sci. Lett. 9, 1445 (1990).
17. J. Wang, Z. Horita, M. Furukawa, M. Nemoto, N.K. Tsenev, R.Z. Valiev, Y. Ma
and T.G. Langdon, J. Mater. Res. 8, 2810 (1993).
18. V.M. Segal, V.I. Reznikov, A.E. Drobyshevskiy and V.I. Kopylov, Russian Metall.
1,99(1981).
188
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
19. T.G. Langdon and J. Wadsworth, in Superplasticity in Advanced M aterials, Edited
by S. Hori, M. Tokizane and N. Furushiro, JSRS, Osaka, Japan, 847 (1991).
20. T.G. Nieh, J. Wadsworth and O.D. Sherby, in Superplasticity in metals and
ceramics, Cambridge University Press, Cambridge, UK, 1-273 (1997).
21. T.G. Langdon, Scripta Metall. 11, 997 (1977).
22. M.F. Ashby and R.A. Verrall, Acta Metall. 21, 149 (1973).
23. A. Ball and M.M. Hutchison, Met. Sci. J. 3, 1 (1969).
24. T.G. Langdon, Phil. Mag. 22A, 689 (1970).
25. A.K. Mukheijee, Mater. Sci. Eng. 8, 83 (1971).
26. T.G. Langdon, Mater. Sci. Eng. A166, 67 (1993).
27. R.C. Gifkins, Metall. Trans. 7A, 1225 (1976).
28. T.G. Langdon, Acta Metall. Mater. 42, 2437 (1994).
29. T. Imai, M. Mabuchi, Y. Tozawa and M. Yamada, J. Mater. Sci. Lett. 9, 255
(1990).
30. M. Mabuchi, K. Higashi and T.G. Langdon, Acta Metall. Mater. 42, 1739 (1994).
31. M. Mabuchi, K. Higashi, S. Wada and S. Tanimura, Scripta Metall. Mater. 26, 1269
(1992).
32. H.Y. Kim and S.H. Hong, Scripta Metall. Mater. 30, 297 (1994).
33. S.W. Lim and Y. Nishida, Scripta Metall. Mater. 32, 1911 (1995).
34. M. Mabuchi, K. Higashi, K. Inoue and S. Tanimura, Scripta Metall. Mater. 26, 1839
(1992).
35. H. Xiaoxu, L. Qing, C.K. Yao and Y. Mei, J. Mater. Sci. Lett. 10, 964 (1991).
36. T.G. Nieh, T. Imai, J. Wadsworth and S. Kojima, Scripta Metall. Mater. 31, 1685
(1994).
189
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
37. T. Hikosaka, T. Imai, T.G. Nieh and J. Wadsworth, Scripta Metall. Mater. 31, 1181
(1994).
38. M. Mabuchi, T. Imai, K. Kubo, K. Higashi and S. Tanimura, Mater. Lett. 12, 276
(1991).
39. M. Mabuchi, T. Imai, K. Kubo, K. Higashi, Y. Okada and S. Tanimura, Mater. Lett.
11,339(1991).
40. I. Tochigi, T. Imai and K. Ai, Scripta Metall. Mater. 32, 1801 (1995).
41. M. Mabuchi and K. Higashi, Mater. Trans. JIM, 35, 399 (1994).
42. M. Mabuchi, K. Higashi, Y. Okada, S. Tanimura, T. Imai and K. Kubo, Scripta
Metall. Mater. 25, 2003 (1991).
43. H. Iwasaki, M. Takeuchi, T. Mori, M. Mabuchi and K. Higashi, Scripta Metall.
Mater. 31, 255 (1994).
44. M. Mabuchi, K. Higashi, Y. Okada, S. Tanimura, T. Imai and K. Kubo, Scripta
Metall. Mater. 25, 2517 (1991).
45. T. Imai, G. L'Esperance and B.D. Hong, Scripta Metall. Mater. 31, 321 (1994).
46. M. Mabuchi, K. Higashi, S. Tanimura, T. Imai and K. Kubo, Scripta Metall. Mater.
25, 1675 (1991).
47. M. Mabuchi, T. Imai, K. Kubo, K. Higashi and S. Tanimura, Mater. Lett. 12, 330
(1991).
48. M. Mabuchi, K. Higashi, K. Inoue, S. Tanimura, T. Imai and K. Kubo, Mater. Sci.
Eng. A156, L9 (1992).
49. S.W. Lim and Y. Nishida, Scripta Metall. Mater. 32, 1821 (1995).
50. T. Imai, G. L'Esperance, B.D. Hong and S. Kojima, Scripta Metall. Mater. 33, 1333
(1995).
51. K. Higashi, T.G. Nieh, M. Mabuchi and J. Wadsworth, Scripta Metall. Mater. 32,
1079 (1995).
52. M. Mabuchi and K. Higashi, Acta Mater. 47, 1915 (1999).
190
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
53. M Strangwood, C.A. Hippsley and J.J. Lewandowski, Scripta Metall. Mater. 24,
1483 (1990).
54. T.R. Bieler and A.K. Mukheijee, Scripta Metall. Mater. 24, 1003 (1990).
55. B. Baudelet, M.C. Dang and F. Bordeaux, Scripta Metall. Mater. 26, 573 (1992).
56. T.G. Nieh, J. Wadsworth and T. Imai, Scripta Metall. Mater. 26, 703 (1992).
57. M. Mabuchi and K. Higashi, Phil. Mag. Lett. 70, 1 (1994).
58. J. Koike, M. Mabuchi and K. Higashi, Acta Metall. Mater. 43, 199 (1995).
59. M. Mabuchi and K. Higashi, Mater. Trans. JIM, 36, 420 (1995).
60. R.S. Mishra, T.R. Bieler and A.K. Mukheijee, Acta Metall. Mater. 43, 877 (1995).
61. G.H. Strijbos & W.H. Kool, Mater. Sci. Eng. AI94, 129 (1995).
62. T.R. Bieler and A.K. Mukheijee, Mater. Sci. Eng. A128, 171 (1990).
63. K. Higashi, T. Mukai, S. Tanimura, A. Inoue, T. Masumoto, K. Kita, K. Ohtera and
J. Nagahora, Scripta Metall. Mater. 26, 191 (1992).
64. K. EBgashi, T. Mukai, S. Tanimura, A. Inoue, T. Masumoto and K. Ohtera, J.
Alloys Comp. 193, 29 (1993).
65. K. Higashi, T. Okada, T. Mukai and S. Tanimura, Scripta Metall. Mater. 25, 2053
(1991).
66. K. Higashi, T.G. Nieh and J. Wadsworth, Mater. Sci. Eng. A188, 167 (1994).
67. T.R. Bieler, T.G. Nieh, J. Wadsworth and A.K. Mukeijee, Scripta Metall. 22, 81
(1988).
68. K. Higashi, T. Okada, T. Mukai and S. Tanimura, Mater. Sci. Eng. A159, LI
(1992).
69. K. Higashi, T. Okada, T. Mukai and S. Tanimura, Scripta Metall. Mater. 26, 761
(1992).
70. K. Higashi, Mater. Sci. Eng. A166, 109 (1993).
191
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
71. K. Higashi, A. Uoya, T. Mukai, S. Tanimura, A. Inoue, T. Masumoto and K.
Ohtera, Mater. Sci. Eng. A181/A182, 1068 (1994).
72. Z. Cui, W. Zhong and Q. Wei, Scripta Metail. Mater. 30, 123 (1994).
73. Z. Cui, W. Zhong, J. Bao, P. Zhan and L. Yong, Scripta Metall. Mater. 31, 1311
(1994).
74. O.D. Sherby and J. Wadsworth, Mater. Sci. Forum, 233-234, 125 (1997).
75. T.G. Nieh and J. Wadsworth, Scripta Metall. Mater. 28, 1119 (1993).
76. W.J. Kim, K. Higashi and J.K. Kim, Mater. Sci. Eng. A260, 170 (1999).
77. D.W. Kum, W.J. Kim and D. Frommeyer, Scripta Mater. 40, 223 (1999).
78. R.R. Sawtell and C.L. Jensen, Metall. Trans. 21A, 421 (1990).
79. T.G. Nieh, L.M. Hsiung, J. Wadsworth and R. Kaibyshev, Acta Mater. 46, 2789
(1998).
80. T.G. Langdon, M. Furukawa, Z. Horita and M. Nemoto, JOM, 50, June, 41 (1998).
81. S. Lee, P.B. Berbon, M. Furukawa, Z. Horita, M. Nemoto, N.K. Tsenev, R.Z.
Valiev and T.G. Langdon, Mater. Sci. Eng. A272, 63 (1999).
82. S. Komura, P.B. Berbon, M. Furukawa, Z. Horita, M. Nemoto and T.G. Langdon,
Scripta Mater. 38, 1851 (1998).
83. P.B. Berbon, S. Komura, A. Utsunomiya, Z. Horita, M. Furukawa, M. Nemoto and
T.G. Langdon, Mater. Trans. JIM, 40, 772 (1999).
84. O.D. Sherby and J. Wadsworth, in Deformation, Processing and Structure, edited
by G. Krauss, American Society for Metals, Metals Park, OH, p.355 (1984).
85. M.W. Mahoney and A.K. Ghosh, Metall. Trans. 18A, 653 (1987).
86. Z. Wei and B. Zhang, J. Mater. Sci. Lett. 13, 1806 (1994).
87. R.S. Mishra, T.R. Bieler and A.K. Mukheijee, Scripta Metall. Mater. 26, 1605
(1992).
88. K. Higashi, T. Mukai, A. Uoya, A. Inoue and T. Masumoto, Mater. Trans. JIM, 36,
1467 (1995).
192
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
89. P.B. Berbon, PhD thesis, Department o f Materials Science and Engineering,
University of Southern California, Los Angeles, CA, (1998).
90. Y. Iwahashi, J. Wang, Z. Horita, M. Nemoto and T.G. Langdon, Scripta Mater. 35,
143 (1996).
91. V.M. Segal, Mater. Sci. Eng. A197, 157 (1995).
92. Y. Wu and I. Baker, Scripta Mater. 37, 437 (1997).
93. A. Shan, I. Moon, H. Ko and J. Park, Scripta Mater. 41, 353 (1999).
94. P.B. Prangnell, C. Harris and S.M. Roberts, Scripta Mater. 37, 983 (1997).
95. D.P. DeLo and S.L. Semiatin, Metall. Mater. Trans. 30A, 1391 (1999).
96. N. Hansen, Mater. Sci. Tech. 6, 1039 (1990).
97. D.A. Hughes and N. Hansen, Acta Mater. 45, 3871 (1997).
98. K. Nakashima, Z. Horita, M. Nemoto and T.G. Langdon, Acta Mater. 46, 1589
(1998).
99. Y. Iwahashi, Z. Horita, M. Nemoto and T.G. Langdon, Acta Mater. 45, 4733
(1997).
100. Y. Iwahashi, Z. Horita, M. Nemoto and T.G. Langdon, Acta Mater. 46, 3317
(1998).
101. Y. Iwahashi, Z. Horita, M. Nemoto and T.G. Langdon, Metall. Mater. Trans. 29A,
2503 (1998).
102. K. Oh-ishi, Z. Horita, M. Furukawa, M. Nemoto and T.G. Langdon, Metall. Mater.
Trans. 29A, 2011 (1998).
103. V.M. Segal, Mater. Sci. Eng. A271, 322 (1999).
104. S. Ferrasse, V.M. Segal, K.T. Hartwig and R.E. Goforth, Metall. Mater. Trans. 28A,
1047 (1997).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
105. W.H. Huang, M.F. Li, P.W. Kao and C.P. Chang, in The Third Pacific Rim
International Conference on Advanced M aterials and Processing (PRICM 3),
Edited by M.A. Imam, R. DeNale, S. Hanada, Z. Zhong and D.N. Lee, The
Minerals, Metals & Materials Society, 1943 (1998).
106. M. Furukawa, Y. Iwahashi, Z. Horita, M. Nemoto and T.G. Langdon, Mater. Sci.
Eng. A257, 328 (1998).
107. Y. Iwahashi, M. Furukawa, Z. Horita, M. Nemota and T.G. Langdon, Metall.
Mater. Trans. 29A, 2245 (1998).
108. P.B. Berbon, M. Furukawa, Z. Horita, M. Nemoto and T.G. Langdon, Metall.
Mater. Trans. 30A, 1989 (1999).
109. S.L. Semiatin, V.M. Segal, R.E. Goforth, N.D. Frey and D.P. DeLo, Metall. Mater.
Trans. 30A, 1425 (1999).
110. D. Yamaguchi, Z. Horita, M. Nemoto and T.G. Langdon, Scripta Mater. 41, 791
(1999).
111. M. Mabuchi and K. Higashi, J. Mater. Sci. Lett. 17, 215 (1998).
112. S. Ferrasse, V.M. Segal, K.T. Hartwig and R.E. Goforth, J. Mater. Res. 12, 1253
(1997).
113. D. Shin, W. Kim and W. Choo, Scripta Mater. 41, 259 (1999).
114. M. Mabuchi, K. Ameyama, H. Iwasaki and K. Higashi, Acta Mater. 47, 2047
(1999).
115. D.P. DeLo and S.L. Semiation, Metall. Mater. Trans. 30A, 2473 (1999).
116. M. Mabuchi, H. Iwasaki and K. Higashi, Nanostruc. Mater. 8, 1105 (1997).
117. J.E. Hatch, Aluminum: Properties and Physical M etallurgy, Am. Soc. Met, Metals
Park, Ohio, (1984).
118. I.J. Polmear, Light Alloys-M etallurgy o f the Light M etals, Arnold, a division of
Hodder Headline PLC, 3rd ed., London, p.24-194 (1995).
119. Metals Handbook, in M etallography and M icrostructures, Amer. Soc. Met, 9th ed.,
Metals Park, Ohio, 9, p.23-388 (1985).
194
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
120. S. Xiang, K. Matsuki, N. Takatsuji, M. Tokizawa, T. Yokote, J. Kusui and K.
Yokoe, J. Mater. Sci. Lett. 16, 1725 (1997).
121. E.O. Hall, Proc. Phys. Soc. B64, 747 (1951).
122. N.J. Petch, J. Iron Steel Inst. 174, 25 (1953).
123. M.F. Ashby and D.R.H. Jones, in Engineering M aterials I, Pergamon Press,
Oxford, p. 105 (1980).
124. M. Furukawa, Z. Horita, M. Nemoto and T.G. Langdon, Ann. Chim. Fr. 21, 493
(1996).
125. M. Furukawa, Z. Horita, M. Nemoto, R.Z. Valiev and T.G. Langdon, Acta Mater.
44, 4619 (1996).
126. M. Furukawa, Y. Iwahashi, Z. Horita, M. Nemoto, N.K. Tsenev, R.Z. Valiev and
T.G. Langdon, Acta Mater. 45, 4751 (1997).
127. M. Furukawa, Z. Horita, M. Nemoto, R.Z. Valiev and T.G. Langdon, Phil. Mag.
78A, 203 (1998).
128. M. Nemoto, Z. Horita, M. Furukawa and T.G. Langdon, Met. Mater. 4, 1181
(1998).
129. Z. Horita, S. Komura, P.B. Berbon, A. Utsunomiya, M. Furukawa, M. Nemoto and
T.G. Langdon, Mater. Sci. Forum, 304-306, 91 (1999).
130. M. Furukawa, K. Oh-ishi, S. Komura, A. Yamashita, Z. Horita, M. Nemoto and
T.G. Langdon, Mater. Sci. Forum, 304-306, 97 (1999).
131. S. Komura, A. Utsunomiya and Z. Horita, unpublished data, Department of
Materials Science, Kyushu University, Fukuoka, Japan (1999).
132. J.S. Vetrano, C.H. Henager, Jr., S.M. Bruemmer, Y. Ge and C.H. Hamilton, in
Superplasticity and Superplastic Forming 1998, Edited by A.K. Ghosh and T.R.
Bieler, The Minerals, Metals & Materials Society, 89 (1998).
133. J.E. Bird, A.K. Mukheijee and J.E. Dorn, in Quantitative Relation Between
Properties and M icrostructure, Edited by D.G. Brandon and A. Rosen, Israel
Universities Press, Jerusalem, Israel, 255 (1969).
134. F.A. Mohamed and T.G. Langdon, Phy. Stat. Sol. 33, 375 (1976).
195
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
135. F.A. Mohamed and T.G. Langdon, Metall. Trans. 5, 2339 (1974).
136. J. Wang, Y. Iwahashi, Z. Horita, M. Furukawa, M. Nemoto, R.Z. Valiev and T.G.
Langdon, Acta Mater. 44, 2973 (1996).
137. T. Mukai, T.G. Nieh, H. Watanabe and K. Higashi, Mater. Sci. Forum, 304-306,
109 (1999).
138. S.J. Rothman, N.L. Peterson, L.J. Nowicki and L.C. Robinson, Phy. Stat. Sol. (b),
63, K29 (1974).
139. X.P. Li, L. Coyelli, V. Tagliaferri and Y.W. Liu, J. Mater. Sci. Lett. 10, 585 (1991).
140. R.Z. Valiev, N.A. Krasilnikov and N.K. Tsenev, Mater. Sci. Eng. A137, 35 (1991).
141. C. Xu, M S thesis, Department of Materials Science and Engineering, University of
Southern California, Los Angeles, CA, (1999).
142. K. Matsuki, T. Aida and J. Kusui, Mater. Sci. Forum, 304-306, 255 (1999).
143. K. Matsuki, H. Kawakami, M. Tokizawa, Y. Murakami and S. Murakami, Mater.
Sci. Forum, 217-222, 1163 (1996).
144. K. Matsuki, H. Kawakami, M. Tokizawa and S. Murakami, Mater. Sci. Tech. 13,
1039(1997).
145. B.M. Watts, M.J. Stowell, B.L. Baikie and D.G.E. Owen, Met. Sci. 10, 189 (1976).
146. B.M. Watts, M.J. Stowell, B.L. Baikie and D.G.E. Owen, Met. Sci. 10, 198 (1976).
147. R.Z. Valiev, A.V. Korznikov and R.R. Mulyukov, Mater. Sci. Eng. A168, 141
(1993).
148. P.B. Berbon, M. Furukawa, Z. Horita, M. Nemoto, N.K. Tsenev, R.Z. Valiev and
T.G. Langdon, Phil. Mag. Lett. 78, 313 (1998).
149. H.C. Kim, T.H. Ahn, C.H. So, Y. Ma, X. Zhao and T.G. Langdon, Scripta Metall.
Mater. 26, 423 (1992).
150. N. Ridley, E. Gullen and F.J. Humphreys, in Superplasticity and Superplastic
Form ing 1998, Edited by A.K. Ghosh and T.R. Bieler, The Minerals, Metals &
Materials Society, 65 (1998).
151. M. Mabuchi and K. Higashi, JOM, 50, June, 34 (1998).
196
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
152. M. Furukawa, P.B. Berbon, Z. Horita, M. Nemoto, N.K. Tsenev, R.Z. Valiev and
T.G. Langdon, Mater. Sci. Forum, 233-234, 177 (1997).
153. P.B. Berbon, N.K. Tsenev, R.Z. Valiev, M. Furukawa, Z. Horita, M. Nemoto and
T.G. Langdon, Metall. Mater. Trans. 29A, 2237 (1998).
154. S.X. McFadden, R.S. Mishra, R.Z. Valiev, A.P. Zhilyaev and A.K. Mukheijee,
NATURE, 398, 684 (1999).
155. R.S. Mishra, S.X. McFadden and A.K. Mukheijee, Mater. Sci. Forum, 304-306, 3 1
(1999).
156. P.B. Berbon, M. Furukawa, Z. Horita, M. Nemoto, N.K. Tsenev, R.Z. Valiev and
T .G Langdon, in M odeling the M echanical Response o f Structural M aterials,
Edited by E.M. TalefF and R.K. Mahidhara, The Minerals, Metals & Materials
Society, 173 (1997).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Deformation of bulk nanocrystalline aluminum alloys

PDF

Creep deformation in superplastic yttria -stabilized zirconia ceramics and yttria-stabilized zirconia /alumina composites

PDF

Development of steel foam processing methods and characterization of metal foam

PDF

Application of electrochemical methods in corrosion and battery research

PDF

Electrochemical test methods for advanced battery and semiconductor technology

PDF

Corrosion protection of anodized aluminum alloys and aluminum metal matrix composites and stainless steels

PDF

Chemisorption studies of hydrocarbons on clean and hydrogen precovered platinum(111) and tin/platinum(111) surface alloys

PDF

Developments of chromium-free conversion coatings on galvanized steels

PDF

Corrosion protection of high-copper aluminum alloys and stainless steels by surface modification

PDF

EFAB: A novel, high-aspect-ratio true three-dimensional microfabrication process for rapid, low-cost desktop micromachining of MEMS

PDF

Flow of fluids with yield stress in porous media

PDF

Ferroelectric-Paraelectric Phase Transitions In The Lead(X) Calcium(1-X) Titanium Trioxide And The Lead(X) Barium(Y) Strontium(1-X-Y) Titanium Oxide Perovskite Systems

PDF

Mechanical processing of two -phase alloys by severe plastic deformation

PDF

High-Temperature Creep In Aluminum Alloy Matrix Composites

PDF

Structural And Phase Stability Of Aluminum(3) Titanium With Modifying Transition Metal Elements

$Fractal theory and groundwater flow in fractured media$

PDF

Fractal theory and groundwater flow in fractured media

PDF

Investigation of sacrificial layer and building block for layered nanofabrication (LNF)

PDF

Processing by equal-channel angular pressing: an investigation of high temperature mechanical properties

PDF

Design and analysis of ultra-wide bandwidth impulse radio receiver

PDF

An evaluation of ultra -wideband propagation channels

Asset Metadata

Creator Lee, Sungwon (author)

Core Title Evaluation of the superplastic potential in commercial aluminum alloys through equal -channel angular pressing

School Graduate School

Degree Doctor of Philosophy

Degree Program Materials Science

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag engineering, materials science,Engineering, Metallurgy,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Langdon, Terrence G. (committee chair), Goo, Edward (committee member), Sammis, Charles G. (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c16-63532

Unique identifier UC11326998

Identifier 3018014.pdf (filename),usctheses-c16-63532 (legacy record id)

Legacy Identifier 3018014.pdf

Dmrecord 63532

Document Type Dissertation

Rights Lee, Sungwon

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA