close

Вход

Забыли?

вход по аккаунту

?

Processing and characterization of functionally graded titanium/titanium boride/titanium diboride composites by combustion synthesis/compaction and microwaves

код для вставкиСкачать
INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI films
the text directly from the original or copy submitted. Thus, some 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 manuscript
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 comer 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.
ProQuest 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.
PROCESSING AND CHARACTERIZATION OF FUNCTIONALLY GRADED
TI/TiB/TIBz COMPOSITES BY COMBUSTION SYNTHESIS/COMPACTION AND
MICROWAVES
A Dissertation
Presented in Partial Fulfillment of the Requirements for the
Degree o f Doctoral of Philosophy
with a
Major in Materials and Metallurgical Engineering
in the
College o f Graduate Studies
University of Idaho
by
Menderes Cirakoglu
July 2001
M ajor Professor: Sarit B. Bhaduri, Ph.D.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number: 3022334
___
®
UMI
UMI Microform 3022334
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.
AUTHORIZATION TO SUBMIT THESIS
This thesis o f Menderes Cirakoglu, submitted for the degree of Doctoral o f Philosophy with
major in Materials and Metallurgical Engineering and titled "Processing and Characterization
o f Ti/TiB/TiB^ Composites by Combustion Synthesis/Compaction and Microwaves" has now
been reviewed in final form, as indicated by the signatures and dates given below. Permission
is now granted to submit final copies to College o f Graduate studies for approval.
Major Professor
Sarit B. Bhaduri
Committee Members
V,
^
Date
Patrick R. Taylor
^m LA
Date
Keith A Prisbrev
I/O<Q1
Date % / 1 / / o f
0
Edwin M. Octom
Department
Administrator
Date
Discipline’s
College Dean
Hate
Harley Johansen
Final Approval and acceptance by the collene o f Graduate Studies
ft
Date
Charles R. Hatch
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
0
/
Dedicated to my parents.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ABSTRACT
The use o f functionally graded materials (FGMs) in aircraft, armor, medical and electronics
industries is becoming more common as their performances exceed the capabilities o f
homogeneous materials. The concept of functionally graded materials (FGMs) was initially
proposed to minimize problems such as poor mechanical integrity and interfacial adhesion
associated with the presence o f abrupt interfaces in metal-ceramic bonded structures.
In this thesis, the combustion synthesis (CS) method was explored in fabricating FGMs in the
Ti-B binary system. Among other methods CS has shown potential in terms o f process
economics and simplicity. Compositional ly graded trilayered and five layered composites
were produced by using three combustion methods: conventional combustion synthesis,
combustion synthesis/compaction and microwave activated combustion synthesis. The
porosity formation during CS has precluded the widespread use of this method. Therefore, a
strategy was proposed to reduce the porosity, first, through control o f vigorous combustion
reactions and second, by applying pressure on the ignited samples. The control of the
reactions was done by selecting compositions away from the stoichiometry without resorting
to adding a third element into the system. These initiatives resulted in net shaped graded
composites with improved density and hardness. As another processing method, microwave
activated combustion synthesis was utilized by using SiC as a susceptor. When compared to
conventional CS, microwave processed FGMs exhibited better microstructurai homogeneity.
All three methods resulted in Ti-TiB-TiBi graded composite materials with continuous and
Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission.
crack free interfaces. Moire Interferometry tests were conducted to compare the deformation
behavior of conventional combustion synthesized FGMs with that o f microwave activated
FGMs. The comparison was based on their in-plane displacements under compression
loading. It was found that microwave produced FGMs exhibit a more compliant behavior.
Under loads as low as 0.37 kN, the composites strained plastically. Also higher boron content
layers exhibited a more compliant behavior compared to pure titanium layers.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ACKNOWLEDGEMENTS
This doctoral work is culmination of several years o f hard work and it is truly a
shared experience. First, I would like to take this opportunity to express my deepest gratitude
to my advisor, Dr. Sarit B. Bhaduri. He has provided me with the freedom to pursue my own
ideas with his unconditional support. His encouragement when research progressed well and
his patience and kindness when it did not were extremely helpful.
I also would like to thank to my other committee members. Thanks to Dr. Patrick R.
Taylor for giving me the opportunity to be a part of his research group and the trust he
showed in my abilities. Dr. Keith A. Prisbrey and Dr. Edwin M. Odom for their helpful
suggestions and insights.
There are a few colleagues past and present, in the Nanomaterials and Advanced
Ceramics Laboratory that I must thank. First, Dr. Sutapa Bhaduri for her constant support,
advice and constructive criticism over the course o f this thesis. She taught me to expect more
from my work and o f myself. Zhixue (Henry) Peng and Jianguo Huang, who laid the ground
work for my research. Muralithran Kutty and Sreerangasai Kesepragada, thanks for all the
discussions on research and life in general -can’t wait to read yours. I have had the privilege
to work with several undergraduate researchers who contributed to the development and
quality of this thesis: Jake Jokisaari, Bill Pribrey, Benjamin Shores and Joanie Loertscher.
Thanks for all your help. Mr. Kenneth G. Gordon, electronic instrument specialist for the
College of Mines, contributed this thesis by his expert technical skills and advice.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
I would like to acknowledge the support and resources made available to me through
the collaborations with Washington State University (WSU) and Idaho National Engineering
and Environmental Laboratory (ENEEL). I would like to thank Dr. W. J. Thomson and his
research group at WSU for providing access to the high temperature X-ray diffraction
facilities. The mercury porosimetry tests were performed at Dr. Amit Bandyopadhyay’s
laboratory at WSU, I gratefully acknowledge his and his students’ help. I would like to thank
to Dr. David F. Bahr and undergraduate student Christy Woodcock at WSU for
nanoindentation tests and for valuable discussions. Moire Interferometry studies were carried
out at the ENEEL with the help o f Sr. Engineer William Windes and Dr. Eric Steffler. Will
deserves special thanks. Not only did he help me on the Moire interferometry tests but also
kept the correspondence to clarify issues and uncertainties. His dedication to his work and his
professionalism is something I hope to emulate in my career.
All my respect and gratitude goes to my parents, to whom I owe everything that I am
and will become. To them I dedicate this thesis.
During my extended stay at the UI, many friends and classmates have made these
years a memorable experience. My heartfelt thanks to Lorraine Mallett and her family for
opening their home for me and supporting me in so many ways. I want to thank my friends
Ana Maria, Shelley, Sherri, Mutlu, Angel, Marie Elaine, Daniel, Yildiz and Bill Kochman,
Barbara, Aki, Miso, Deepa, Wiola, Peter, Ellen and all o f those I may have forgotten in the
rush, for all the support and fun times together.
Finally I would like to thank the Army Research Office (ARO) for supporting this
work under grant #DAAG-55-9810281 with Dr. W.M. Mullins as the monitor.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
TABLE OF CONTENTS
Page No
AUTHORIZATION TO SUBMIT THESIS................................................................................ ii
DEDICATIO N.................................................................................................................................. iii
ABSTRACT........................................................................................................................................ iv
ACKNOWLEDGMENTS............................................................................................................... vi
LIST OF FIGURES........................................................................................................................ xii
LIST OF TABLES...........................................................................................................................xx
DISSERTATION O UTLINE.......................................................................................................xxi
Manuscript One: Functionally Graded Materials, A Review................................................... 1
Cover P a g e ........................................................................................................................................ 1
A bstract............................................................................................................................................. 2
1. Introduction................................................................................................................................ 3
2. Functionally Graded M aterials................................................................................................ 4
2.1. Potential Applications.......................................................................................................5
2.2. Processing M ethods.......................................................................................................... 9
2.2.1. Powder Metallurgy Processes...................................................................... 10
2.2.2. Infiltration...................................................................................................... 12
2.2.3. Casting............................................................................................................ 13
2.2.4. Plasma Spraying............................................................................................ 14
2.2.5. Laser C ladding.............................................................................................. 16
2.2.6. Vapor Deposition Processes........................................................................ 16
2.2.7. Reactive Processes......................................................................................... IS
2.3. Overview on F G M s...................................................................................................... 18
3. R eferences................................................................................................................................ 20
M anuscript Two: Functionally Graded Ti-TiB-TiB 2 Composites Produced By
Combustion Synthesis —I: Process Analysis............................................................................... 25
Cover P a g e ......................................................................................................................................25
A bstract........................................................................................................................................... 26
1. Introduction............................................................................................................................. 27
1.1. Combustion Synthesis-A Review................................................................................ 27
1.1.1. Thermodynamic Considerations...................................................................29
1.1.2. Comparison to Other Constructive Processing Techniques......................32
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.
3.
4.
5.
1.2. Ti-B Binary System ....................................................................................................... 34
1.3. Objectives........................................................................................................................38
Experimental.............................................................................................................................. 39
2.1. Materials..........................................................................................................................39
2.2. Phase Formation Studies.....................................................................................
40
2.3. Preparation o f Single Composition Pellets and FG M s.............................................. 40
2.4. Experimental System ..................................................................................................... 43
2.5 Combustion Experiments.............................................................................................. 45
Results and Discussions...........................................................................................................49
3.1. Thermodynamic Considerations.................................................................................. 49
3.1.1. Adiabatic Temperatures for Stoichiometric Compositions (x=0) ..........51
3.1.2. Adiabatic Temperatures for Non-Stoichiometric Compositions (x>0).. 57
3.2. High Temperature X-Ray Diffraction Studies...........................................................62
3.3. Combustion Experiments with Single Compositions............................................... 64
3.3.1. Combustion Characteristics......................................................................... 64
3.3.2. Time-Temperature P rofiles......................................................................... 66
3.3.3. X-Ray Diffraction Studies............................................................................70
3.4. Combustion Experiments with Graded Compositions............................................. 74
3.4.1. Combustion Characteristics......................................................................... 74
3.4.2. Time-Temperature Profiles......................................................................... 75
3.4.3. Effect o f Atmosphere....................................................................................80
3.4.4. X-Ray Diffraction Studies............................................................................83
Conclusions............................................................................................................................... 85
R eferences..................................................................................................................................87
Manuscript Three: Functionally Graded Ti-TiB-TiB 2 Composites Produced By
Combustion Synthesis —LI: The Effect of Compaction on Properties.................................. 90
Cover P a g e ......................................................................................................................................90
A bstract........................................................................................................................................... 91
1. Introduction............................................................................................................................... 92
1.1. The Origin o f Porosity in Combustion Synthesized M aterials................................92
1.2. Densification M ethods o f Combustion Synthesized M aterials................................93
1.3. Densification o f Combustion Synthesized FG M s..................................................... 95
1.4. Motivation and Objectives........................................................................................... 97
2. Experimental Procedure........................................................................................................... 99
2.1. Combustion Synthesis/Compaction Experim ents.....................................................99
2.2. Characterization...........................................................................................................102
3. Results and Discussions......................................................................................................... 106
3.1. Density and Porosity M easurements........................................................................ 106
3.1.1. Combustion Synthesis o f Single Compositions...................................... 106
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.1.2. Combustion Synthesis/Compaction o f FG M s........................................ 116
3.2. Microstructure............................................................................................................. 119
3.3. Mechanical Properties................................................................................................124
3.3.1. Vickers Hardness Tests.............................................................................. 124
3.3.2. Evaluation o f Interlayer Properties with Nanoindentation..................... 127
4. Conclusions.............................................................................................................................134
5. Recommended Future W ork..................................................................................................136
6. References...............................................................................................................................138
Manuscript Four: Processing and Characterization o f Functionally Graded Materials
Produced By Microwaves in Ti-B Binary System.................................................................. 141
Cover P a g e ................................................................................................................................... 141
A bstract.........................................................................................................................................142
1. Introduction.............................................................................................................................143
1.1. Microwave H eating.................................................................................................... 143
1.2. FGM Processing by Using Microwaves................................................................... 145
1.3. Microwave Activated Combustion Synthesis..........................................................146
1.4. Objectives....................................................................................................................147
2. Experimental Procedure.........................................................................................................149
2.1. Preparation of Layered Structures............................................................................. 149
2.2. Microwave Activated Combustion Synthesis Experiments................................... 152
2.3. Characterization........................................................................................................... 157
3. Results and Discussions........................................................................................................ 159
3.1. Combustion Characteristics.......................................................................................159
3.2. X-Ray Diffraction Studies.........................................................................................169
3.3. Density and Porosity Measurements.........................................................................172
3.4. Microstructural Analysis............................................................................................ 175
3.5. Mechanical Properties............................................................................................... 179
4. Conclusions.............................................................................................................................181
5. Recommended Future W ork..................................................................................................182
6. References...............................................................................................................................183
Manuscript Five: Deformation Analysis of Ti-B Based FGMs by Phase Shifted Moire
Interferometry................................................................................................................................185
Cover P a g e ................................................................................................................................... 185
A bstract.........................................................................................................................................186
1. Introduction............................................................................................................................187
1.1. Moire Interferometry..................................................................................................188
1.2. The Formation o f Fringe Patterns............................................................................. 190
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
XI
2. Experimental Procedure......................................................................................................... 195
2.1. Sample Preparation......................................................................................................195
2.2. Moire Interferometry Experiments............................................................................196
3. Results and Discussions.........................................................................................................203
3.1. Compression Test Results o f CS Produced FGMs.................................................. 203
3.2. Compression Test Results o f MW Produced FG M s...............................................213
3.2.1. Re-loading Under Compression............................................................... 221
4. Conclusions and Recommended Future W ork................................................................... 227
Appendix 1.................................................................................................................................... 229
Appendix II................................................................................................................................... 232
5. References................................................................................................................................ 237
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
XI i
LIST OF FIGURES
Manuscript One
Figure
Page No
1
Schematic of microstructure o f a two component FGM
4
2
The concept o f functionally graded armor composite
7
Manuscript Two
1
Combustion wave propagation.
28
2
Ti —B binary phase diagram.
37
3
Flow chart of experimental procedure.
42
4
Schematic drawing (a) and a picture (b) o f a five-layered green sample.
43
5
A schematic o f the experimental set-up.
44
6
Photograph o f experimental set-up showing the reactor, data acquisition
44
system and pyrometer arrangement.
7
Die assembly used in CS/DC experiments.
46
8
Video image taken during the passage o f combustion wave.
47
9
Free energy versus temperature for the reactions.
51
10
Enthalpy-temperature plot for reactants and products for reaction 1 where
53
reaction stoichiometry o f x=0.
11
Enthalpy-temperature plot for reactants and products for reaction 2 where
53
reaction stoichiometry o f x=0.
12
Enthalpy-temperature plot for reactants and products for reaction 3 where
54
reaction stoichiometry o f x=0.
13
Enthalpy-temperature plot for reactants and products for reaction 4 where
reaction stoichiometry o f x=0.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
54
xiii
Figure
14
Page No
Enthalpy-temperature plot for reactants and products for reaction 1 with
59
reaction stoichiometries ofx=0, x=0.5, x=l and x=1.5.
15
Enthalpy-temperature plot for reactants and products for reaction 2 with
59
reaction stoichiometries of x=0, x=0.5, x=l and x=1.5.
16
Enthalpy-temperature plot for reactants and products for reaction 3 with
60
reaction stoichiometries of x=0, x=0.5, x=l and x=1.5.
17
Enthalpy-temperature plot for reactants and products for reaction 4 with
60
reaction stoichiometries of x=0, x=0.5, x=l and x=1.5.
18
HTXRD pattern o f 82Ti-18B composition (CoKa radiation A.=1.7889A).
62
19
HTXRD pattern o f 90Ti-10B composition (CoKa radiation X=1.7889A).
63
20
Combustion wave propagation images from combustion o f 85Ti-15B.
64
21
Combustion wave propagation images from combustion of 90Ti-10B.
65
22
82Ti-18B single composition (without degassing) after combustion
65
reaction. Sample exploded after the combustion wave propagated after half
way down the sample.
23
Schematic drawing o f the thermocouple arrangement in the compact and the
67
temperature profile measured with two W-5%Re/W-26%Re and a
pyrometer for combustion reaction in single composition o f 82Ti-18B.
24
Schematic drawing of the thermocouple arrangement in the compact and the
69
temperature profile measured with two W-5%Re/W-26%Re and a
pyrometer for combustion reaction in single composition o f 95Ti-5B.
25
X-ray diffraction patterns o f single compositions 95Ti-5B and 90Ti-10B
72
after combustion synthesis.
26
X-ray diffraction patterns o f single compositions 85Ti-15B and 82Ti-18B
73
after combustion synthesis.
27
The propagation o f combustion wave through the graded layer compacts.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
74
Figure
28
Page No
Schematic drawing o f the thermocouple arrangement in the compact and the
78
temperature profile measured with two W-5%Re/W-26%Re and a
pyrometer for combustion reaction in a five layered FGM.
29
Schematic drawing o f the thermocouple arrangement in the compact and the
79
temperature profile measured with three W-5%Re/W-26%Re for
combustion reaction in a five layered functionally graded material.
30
Schematic drawing o f the thermocouple arrangement in the compact and the
81
temperature profile measured with W-5%Re/W-26%Re for combustion
reaction in argon atmosphere and in vacuum in a five layered functionally
graded materials.
31
Schematic drawing o f the thermocouple arrangement in the compact and the
82
temperature profile measured with W-5%Re/W-26%Re for combustion
reaction in argon atmosphere and in vacuum in a five layered functionally
graded materials.
32
X- ray diffraction patterns from two layers (95Ti-5B and 90Ti-10B) of
83
FGM after combustion synthesis.
33
X- ray diffraction patterns from two layers (85Ti-15B and 82Ti-18B) of
84
FGM after combustion synthesis.
Manuscript Three
1
Flow chart of the experimental procedure.
101
2
Bulk density change before and after combustion experiments.
107
3
Mercury intrusion plot for composition 82Ti-18B after degassing.
112
4
Mercury intrusion plot for composition 95Ti-5B after degassing.
112
5
Cumulative pore size distributions for composition 82Ti-18B after
113
degassing
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
XV
Figure
6
Page No
Cumulative pore size distributions for composition 95Ti-5B after
113
degassing
7
Mercury intrusion plot for composition 82Ti-18B after CS.
114
8
Mercury intrusion plot for composition 95Ti-5B after CS.
114
9
Cumulative pore size distributions for composition 82Ti-18B after CS.
115
10
Cumulative pore size distributions for composition 95Ti-5B after CS.
115
11
Bulk density change in different layers o f FGM after CS with and without
117
pressure application.
12
Optical micrographs o f the interfaces o f the FGMs produced by (a)
119
conventional combustion synthesis (b) combustion synthesis/compaction.
13
SEM micrograph o f a cross section o f a five layered FGM.
120
14
SEM of polished sample produced by conventional combustion synthesis
121
(a) interface o f the first and second layers (b) interface second and third
layers.
15
SEM micrographs ofC S samples (a) Ti+5wt%B (b) 10wt% (c) 15wt%.
121
16
SEM micrographs o f CS/compaction samples (a) Ti-Ti+5wt%B interface
122
(b) Ti+5wt%B - Ti+10wt%B interface.
17
Macrograph o f lateral macro cracks on the outside surface o f an FGM.
123
18
Microhardness values o f combustion synthesized FGM layers
124
19
Microhardness values o f FGM layers after the application o f pressure of
126
137 kPa
20
Microhardness values o f FGM layers after the application o f pressure of
126
220 kPa
21
Schematic drawing showing different parameters used in Oliver-Pharr
method.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
128
Figure
22
Page
The results o f a nanoindentation test obtained from interface 1. Figure
shows a high resolution SPM image o f the interface region with the image
o f the indentation and the load -displacement history involved in making
such an indentation.
23
The results o f a nano indentation test obtained from interface 2. Figure
shows a high resolution SPM image o f the interface region with the image
o f the indentation and load -displacement history in making such an
indentation.
Manuscript Four
1
Schematic drawing (a) and a picture (b) o f a 5-Iayered FGM green
compact.
2
Flowchart showing the steps of microwave processing o f FGMs.
3
A schematic o f the experimental set-up. 1) Magnetron 2) Main Controller
3) Forwarded Power Indicator 4) Reflected Power Indicator 5) Wave guide
6) Computer Control 7) Optical Pyrometer 8) Sample Insulation 9) Argon
Tank 10) Vacuum Pump.
4
Photograph of the microwave furnace used in our experiments.
5
Insulation and indirect heating arrangement used in our experiments.
6
Photograph of the sample exploded during MW processing
7
Combustion wave propagation of sample 82Ti-18B during microwave
processing.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
xvii
Figure
8
Page No
A schematic o f the single composition sample used in microwave
164
experiments. The cross indicates the spot where the pyrometer was
focused. Time temperature profiles obtained from single compositions.
9
Schematic o f a five layered FGM and video images from the combustion
166
experiments. The black dot in the schematic indicates the hole where the
pyrometer was focused.
10
Schematic drawing o f a five layered FGM and time temperature profile
167
recorded. The black dot in the schematic indicates the hole where the
pyrometer was focused.
11
The cross section o f a 5 layered FGM after MW processing.
168
12
X-ray diffraction patterns o f two layers o f FGMs ignited in argon.
170
13
X-ray diffraction patterns o f two layers o f FGMs ignited in air.
171
14
Bulk density change before and after combustion synthesis
173
15
Bulk density change in different layers o f FGM after microwave
174
combustion experiments.
16
SEM image showing the dense inner section o f FGM. The outer shell
175
contains more porosity.
17
Optical micrograph showing the interface between Ti layer and Ti-5wt% B
176
layer (magnification: x 200).
18
SEM image o f Ti layer o f FGM ignited in air (a) and in argon (b).
177
19
SEM images o f the 82Ti-18B layers o f the FGMs (a) ignited in air (b)
178
ignited in argon (c) EDS spectra.
20
Vickers hardness data o f the individual layers in FGM ignited in air.
179
21
Vickers hardness data o f the individual layers in FGM ignited in argon.
180
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
XV111
Manuscript Five
Figure
Page No
1
Schematic diagram o f Moire Interferometry system.
2
Schematic drawing o f a four-beam moire interferometry used to record U
191
193
(x) and V (y) displacements.
3
Steps in producing the specimen grating by replication process.
197
4
Photograph o f the Moire Interferometer used in our tests.
199
5
Schematic o f the fixture used for compression testing.
201
6
Specimen geometry and loading configuration under compression.
202
7
A schematic o f the test sample produced by combustion synthesis after
204
machining.
8
Wrapped images of displacement under a null load o f 0.01 kN (a) and
206
under a maximum load o f 0.65 kN (b)
9
Unwrapped images o f displacement under a null load o f 0.01 kN (a) and
207
under a maximum load o f 0.65 kN (b)
10
Displacement (V) as a function o f distance along the left, right and middle
210
o f the CS produced FGM sample under a compression load o f 0.65 kN
(Null load subtracted out). Vertical dashed lines indicate approximate
locations o f the interfaces.
11
Compression data of displacement (V) and strain as a function of sample
211
length along the right side o f the FGM. Vertical dashes indicate
approximate locations o f the interfaces.
12
Stress—strain curves for three-layered FGM produced by combustion
212
synthesis.
13
A schematic o f the test sample produced by microwave activated
combustion synthesis after machining.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
214
xix
Figure
14
Page No
Wrapped images o f displacement under a null load o f 0.01 kN (a) and
216
under a maximum load o f 0.37 kN (b)
15
Unwrapped images o f displacement under a null load o f 0.02 kN (a) and
217
under a maximum load o f 0.37 kN (b)
16
Displacement (V) as a function o f distance along the left, right and middle
218
of the MW produced FGM sample under a compression load o f 0.37 kN
(Null load subtracted out). Vertical dashed lines indicate approximate
locations o f the interfaces.
17
Compression data o f displacement (V) and strain as a function o f sample
219
length along the right side o f the FGM. Vertical dashes indicate
approximate locations of the interfaces.
18
Stress-strain curves obtained during the first loading stage for three layers
220
of FGMs produced by microwaves.
19
Wrapped images o f displacement after unloaded from 0.37 kN (a) and
223
under a maximum load o f 0.60 kN (b)
20
Displacement (V) as a function o f distance along the left, right and middle
224
of the MW produced FGM sample under a compression load o f 0.60 kN
(Null load subtracted out). Vertical dashed lines indicate approximate
locations o f the interfaces.
21
Compression data o f displacement (V) and strain as a function o f sample
225
length along the right side o f the FGM. Vertical dashes indicate
approximate locations of the interfaces.
22
Stress-strain curves o f a three layered FGM produced by microwaves.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
226
XX
Appendix I
Figure
1
Page No
Compression test wrapped fringe patterns, sample FGM #10 1-21-01A,
230
(V field, 1200 line grating)
2
Compression test wrapped fringe patterns, sample FGM #10 1-21-01A
231
(V field, 1200 line grating)
Appendix II
1
Compression test wrapped fringe patterns, sample FGM #9A 1-23-01
234
2
Compression test wrapped fringe patterns during re-loading
sample FGM #9A 1-23-01
236
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
XXI
LIST OF TABLES
Manuscript Two
Table
Page No
1
The adiabatic temperatures o f reactions for x=0
55
2
The change in adiabatic temperature with composition
61
Manuscript Three
1
Mercury porosimetry results o f the single composition green compacts
108
2
Mercury porosimetry results o f the ignited single compositions
110
3
Mercury porosimetry results o f two layers o f FGMs produced by
118
conventional combustion synthesis
4
Mercury porosimetry results o f two layers o f FGMs after the application o f
118
a pressure o f 220 lcPa.
Manuscript Four
1
Mercury porosimetry results o f two FGM layers after microwave
174
combustion in air.
Manuscript Five
1
Force-pressure conversions for tested combustion synthesized FGM
205
2
Force-pressure conversions for tested microwave activated FGM
215
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
DISSERTATION OUTLINE
This thesis is structured into 5 chapters. The outline o f the chapters is as follows:
A review o f the current literature on functionally graded materials is given Chapter I .
It is a summary o f functionally graded materials (FGMs) in terms of applications, fabrication
methods, and primary challenges facing its design and processing. This chapter provides the
necessary background for the development of the thesis in the manuscripts that follow.
In spite o f all the extensive study on combustion synthesis, the utilization o f this
method in FGM processing is relatively new and there are some serious challenges needs to
be solved. In this thesis, we address some o f these problems. Due to the extent o f the work,
the entire combustion synthesis processing section is divided into three main parts. Chapter 2
presents an investigation o f thermodynamic calculations that illustrates the change o f
adiabatic temperatures with different compositions in Ti-B system. Starting with composition
selection and single composition experiments, the design and processing steps o f functionally
graded materials are explained. The phase formation studies are mentioned and explained on
the basis o f thermodynamic calculations. This chapter also includes, time-temperature
profiles o f the FGMs along the length o f the sample during combustion reactions. Chapter 3
particularly addresses the primary challenge of porosity and densification o f combustion
synthesized products. It presents an accompanied compaction method, which has the
potential o f producing denser products. Specifically, those issues that impact the performance
o f graded materials such as density, porosity, microstructure and hardness, are examined. The
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
results o f the combustion synthesis/compaction method are compared to that of conventional
combustion synthesis. At the end o f this chapter, the contributions made are summarized.
Current challenges faced are presented and future research directions are included. In
Chapter 4, microwave activated combustion synthesis is investigated in producing
functionally graded materials in Ti-B system. A brief review o f the fundamentals of
microwave processing is given. Process planning and experimental procedure is explained.
Specifically, the effect of thermal insulation and the effect o f atmosphere on the process are
discussed. In addition, combustion characteristics, time- temperature profiles, microstructure
and hardness is studied. Also process characteristics and results are compared to those of
combustion synthesis.
Finally, in Chapter 5, the elastic/plastic mechanical response o f FGMs under
compression loading is analyzed. The concept of Phase Shifted Moire Interferometry (PSMI)
is introduced. The capabilities o f the system, measurement technique, sample preparation and
data analysis are explained. The chapter then explains the differences between conventional
combustion synthesized FGMs and microwave assisted combustion synthesized FGMs.
Additionally, a complete series o f wrapped fringe patterns obtained from all load stages are
given in Appendix I and Appendix n.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1
MANUSCRIPT ONE
FUNCTIONALLY GRADED MATERIALS- A REVIEW
Menderes Cirakoglu
Department o f Materials and Metallurgical Engineering
University o f Idaho
Moscow-Idaho 83844-3024
will be submitted for publication
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2
ABSTRACT
Further developments in science and technology rely upon the development o f new material
systems. The research on alternative materials that can withstand severe engineering
conditions is motivated by the need for properties that are unavailable in any single material.
In this area, functionally graded materials (FGMs) emerge as an attractive research platform.
FGMs promise increased design flexibility as well as improved properties over the
conventional materials that are currently in production and in use. The FGM process offers a
potential way that a part can be designed to meet the required specifications in a certain
technological application. They hold the key for applications requiring ultra-high material
performance such as thermal barrier coatings, bone and dental implants, piezoelectric and
thermoelectric devices, and optical materials with graded refractive indices and new'
applications are continuously being discovered. To understand why such features are
desirable and why FGMs can be an attractive research field, the current state o f the FGMs
must be analyzed. In this chapter, a literature review on FGMs is given in order to provide a
broad perspective, which will be useful in reaching the goals o f this thesis.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1. Introduction
New technological advances are imposing stronger demands on materials performances. To
meet these new demands, new material systems and processing methods are being introduced
everyday. In many engineering applications, materials encounter service conditions, which
vary within the component. For example, the body of a gear must be tough, whereas its
surface must be hard and wear resistant, and the body o f a turbine blade must be strong,
tough and creep resistant, while its outer surface must be refractory and oxidation resistant or
in a spacecraft one side may be exposed to extremely high temperatures and the other side
may be exposed to extremely low temperatures. Therefore, it is important to be able to design
materials to meet these varying requirements. The solution to these demanding current
applications may lie in the use of spatially inhomogeneous materials [1-4].
Functionally graded materials, like other composites, are designed to achieve superior
performance compared to that of homogeneous materials by combining the desirable
properties o f each constituent phase. This new concept marks the beginning of a paradigm
shift in the way we think about materials and structures because it allows to integrate
material and structural design considerations. Therefore, FGMs are receiving considerable
attention from the materials science community and industry, particularly in Japan, where the
concept originated [1-2].
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4
2. Functionally Graded Materials
Unlike conventional coated materials and composites, FGMs have a continuous grade in
composition as shown in Figure 1. In these materials the volume fractions o f the constituents
continuously change such that its composition and microstructure varies in one direction.
Therefore they take the advantage o f the properties of two different materials within the same
body [1-4].
Figure 1.
Schematic o f microstructure o f a two component FGM [5],
Joining ceramics to metals or ceramics to ceramics to combine inherent advantages o f these
two materials has been pursued to meet the materials requirements in many applications. As
is well known, ceramics possess low density, good high temperature strength and creep
resistance but their fracture toughness and thermal shock resistances are poor. On the other
hand, metallic materials are associated with high fracture toughness and excellent thermal
shock resistance with low strength at high temperatures. A composite of these two may meet
the material requirements in many engineering applications [6,7], However, joining two
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5
dissimilar materials like metals and ceramics pose some serious problems such as the
generation o f thermal residual stresses. Thermal stresses localize especially at the vicinity o f
the sharp interfaces and promote debonding o f the interface or cracking [8-11], In an attempt
to overcome these problems, multi-layered structures were fabricated with gradients in
composition, microstructure and properties. By placing a compositionally graded layer
between two dissimilar solids with different thermal properties, smoothens the distribution o f
thermal stresses and diminishes the magnitude of thermal stresses at the critical locations.
This distinctive feature o f FGMs provides properties that are not offered by monolithic
materials [12-15].
2.1. Potential Applications
FGMs are ideal candidates for a wide range o f applications. The areas involve advanced
aircraft and aerospace engines, computer circuit boards, medical implants, armor, electric or
dielectric devices and optical devices [15,16],
Structural components used in pressure vessels and pipes in nuclear reactors, high-speed
aircrafts are subjected to thermal loading due to high temperature, high temperature gradient
and cyclical changes o f temperature. In these applications, metal/ceramic composites are
widely used because it is unlikely to find a single material that will satisfy the optimum
properties. Because o f the mismatch in the material properties o f ceramics and metals, the
temperature changes during processing and operation would create residual and thermal
stresses at the sharp interfaces. However, by using FGM concept, the integration o f these
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6
dissimilar materials can be achieved successfully. In these applications, a ceramic rich region
o f an FGM is exposed to hot temperature while a metallic rich region is exposed to cold
temperature, with a microstructural transition in the direction o f the temperature gradient.
Therefore, normally incompatible properties such as high hardness and corrosion resistance
are incorporated in the same material [8],
C/C composites are candidates for applications such as engines and airframes o f space planes
due to their high thermal resistance and high specific strength. One problem with C/C
composites is their low oxidation resistance and therefore they are usually coated with SiC by
chemical vapor deposition (CVD) technique. However, the difference in thermal expansion
between the SiC layer and the C/C base material causes cracking and consequently oxygen
diffuses to the base material through these cracks. To improve the oxidation resistance by
reducing the mismatch and increasing the bonding strength between the coating layer and the
base material, C/C composites are coated with SiC/C functionally graded material by CVD
method where the change from carbon substrate to SiC is compositionally graded [2],
In fusion experimental reactors, the piping system o f actively cooled diverter plates must be
protected against the strong electromagnetic force from the induced current generated by a
transient phenomenon called “plasma disruption”. The induced current can be reduced by
integrating an electrical insulation breaks in the current circuit. Ceramic insulation joints are
widely used for this purpose by joining them directly to metallic structures. However, these
joints are suspected to produce excessive residual stress and leakage through the
ceramic/metal joint interface. FGM joints with a continuous graded chemical composition
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7
(stainless steel/PSZ-AkCVstainless steel) promise to solve these problems by eliminating the
discrete ceramic/metal interface [16].
From a ballistic standpoint; FGMs offer significant advantages to armor designers since the
ceramic/metal composites are capable o f supporting a significant structural load with no
discrete material interfaces. It is well known that hard materials make the best armors.
However, they are also brittle. This limits the ability o f the armor to sustain multiple hits to
the same area. The idea behind using FGMs is to disrupt the shockwave to limit damage and
improve multi hit capability. A typical modem armor consist a hard outer surface backed by
a ductile metal as shown in Figure 2. The purpose of this concept is to expend the kinetic
energy o f the projectile without being penetrated [18-19].
Projectile
Ceramic
Hard but brittle ceramic
material shatters during
the first impact
Metal
Ductile metallic material
absorbs th e re st o f the
kinetic energy
Graded Layer
Discrete interlaces replaced
with graded layer provides
better jo in t between ceramic
and metal
Figure 2. The concept of functionally graded armor composite
In the field of biomaterials, metal-ceramic system has potential applications especially in
orthopedics and dentistry. For example; hydroxyapatite (HA)-Ti FGMs. HA is bioactive and
reacts chemically with living bone to give a strong bond that can resist high stresses and by
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
8
producing an FGM with Ti, this goal can be achieved without a significant degradation in
mechanical properties [20],
In the field o f optical materials, it was found that ZnGa 2 C> 4 solid solution exhibits white
emitting phosphor characteristics if the solid solution consists o f a suitable graded
composition. This may yield a new method o f designing new type o f luminescence materials.
Another novel approach was refractive index gradation in optical filter design. By
introducing layers with graded refractive index into the conventional TiCh/SiCb multilayer
films, the sidelobs, which causes the transport efficiency, can be greatly reduced [21].
Sintered AIN in which W particles are dispersed in graded manner is a highly radiative
material. These materials use AIN as the matrix for transmittance and metallic W particles as
emissivity particles and have a structure in which the dispersion o f tungsten particles is
changed in stages. In the graded structure, the amount o f dispersed metallic tungsten particles
is greatest near the heat source and is gradually decreased [22].
Zhu et al [23 ], prepared functionally graded piezoelectric actuators in PbNit/3Nb2/3C>3PbZrTi0 3 -PbTi0
3
(PNN-PZ-PT) system which consists o f three layers, namely, a
piezoelectric layer, a dielectric layer and a sandwich layer whose dielectric and piezoelectric
properties change gradually.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
9
2.2. Processing Methods
Many industrial applications can benefit from FGMs consequently there is substantial
interest in devising an inexpensive and versatile process for their fabrication.
Various production methods are available for manufacturing graded structures. Among those,
some o f them are suitable for large and/or complex shapes while others can only be used for
small and/or simple shapes [1,2].
In some o f these fabrication methods, the gradients o f FGMs are constructed as a preform.
The preforms consist o f a series o f mixtures from the constituent materials producing the
composition gradient. After mixing, the graded compact is initially formed by die pressing.
The pressure should be enough to prevent sample damage during subsequent handling
therefore cold isostatic compaction is usually used [1,3].
There are other novel preform preparation methods reported in the literature. Lamas et al
[24], developed a slip-casting method to produce alumina bodies with differential porosities
so that molten metal impregnation can be carried out. Darcovich et al [25] again utilized a
casting method to produce graded ceramic membranes with a continuously increasing mean
particle size from top to bottom. They achieved to optimize the porous structure o f green
compacted powders by choosing appropriate sintering temperatures and controlling by pH o f
the slurries. Takao et al [26] utilized an aerosol filtration method and studied the processing
o f multilayer composite o f an oxide superconductor and a nickel zinc ferrite chip
components. They prepared the compositionally graded composites by regulating the particle
flow time under a constant flow rate and particle size. Lee et al [27] employed centrifugal
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
10
force to obtain a compositional gradient of epoxy/carbon polymeric composite. The gradient
structure o f the composite was controlled by varying the rotation time and the material
properties such as fiber length, fiber content, and matrix viscosity.
A variety o f processes have been developed and reported in literature. Some of these
methods have been commercialized while some only exist as research projects. The
following describe and summarize the most common reported fabrication methods.
2.2.1. Powder Metallurgy (P/M ) Processes: Powder metallurgy which can create a wide
range o f variable compositions and locate them in discreet regions, are an important
technology for fabricating FGMs. P/M process offers an attractive and versatile method for
the fabrication of FGMs. The method involves cold pressing followed by sintering heat
treatment or hot pressing, hot isostatic pressing (HIP) [28-29],
Rabin et al [30] studied fabrication o f disk shaped Ni-Al2 0 3 FGMs by using three different
consolidation methods: pressureless sintering (at 1400°C for 3 hours under flowing argon),
hot pressing (1300°C for 1 hour at < lOMPa), and hot isostatic pressing (1350°C and at 100
MPa). Although hot pressing resulted in densified graded compacts, the reactions between
the green compact and the graphite dies caused problems. They minimized the reactions by
coating the dies with boron nitride and by limiting the pressure to < 20MPa. In this case, the
products contained with significant amount o f porosity. Hot isostatic pressing proved to be a
successful method in producing dense, defect-free graded materials. However, the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
11
differential shrinkage caused by the green density variations resulted in non-uniform final
product shapes, although no cracks are formed.
Kawasaki et al [31], prepared disk shaped FGMs o f partially stabilized zirconia and stainless
steel by hot pressing at 1473K under Ar-5% H 2 atmosphere at a pressure o f 30 MPa. In
another study, they achieved to produce defect free SiC-AlN/Mo FGMs by hot pressing at
2123K at a pressure o f 200 MPa under argon atmosphere. Mo was chosen to be the metallic
material for the bottom surface because its sintering temperature is close to SiC-AIN and it is
a heat resistant material. To avoid the reaction between SiC and Mo, AIN was used as the
intermediate layer [32]. Bishop et al [33] produced FGMs by using titanium and
hydroxyapatite (HA). The titanium powder and the powder mixtures with different
proportions o f HA were layered in the following arrangement: Ti, Ti-10% HA, Ti-20% HA,
Ti-30% HA, Ti. The layered powders were cold compacted and hot pressed at 500°C at 1630
MPa pressure. This work demonstrates that an FGM might be produced at low temperatures,
thus reducing the problems due to the differences in properties. Jung et al [5], produced
tetragonal zirconia polycrystal (TZP)/Ni and TZP/stainless steel 304 with 10 vol %
compositional gradient by hot pressing. When compared to directly joined TZP/metal
systems, functionally graded materials concept provided a better interfacial stability and
thermal protection characteristics. Zhu et al [34] produced ZrCb-Ni FGMs. After being
blended with different Zr 0 2 /Ni ratio, the powders were stacked sequentially in a steel die.
Then the green compact was hot pressed at 1350°C under a pressure o f 25 MPa for 1-2 hours
in nitrogen atmosphere.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
12
2.2.2. Infiltration: In this method, the powder preforms are heated to a temperature high
enough for a liquid phase to form in the preform. This method results in dense graded
metal/ceramic composites with interpenetrating network microstructure. Infiltration
technique has been utilized in the automotive industry components (pistons) featuring graded
transitions from aluminum to a ceramic reinforced aluminum at surfaces forming the
combustion chamber. Compared to liquid phase sintering, this technique has several
advantages such as little or no bulk shrinkage, complex shape capability and more rapid
reaction kinetics [35-37].
The production of graded WC-Co and W-Cu structures was studied with this technique.
However, during the process, the liquid phase migrates across the component and yields to a
homogenous product rather than a graded structure [38].
Marple et al [39], fabricated compositionally graded mullite/alumina composites by fully
immersing porous alumina preforms in a pre-hydrolyzed ethyl silicate solution. The preforms
then heat-treated at 1200°C.
Corbin et al [40] developed a process combining tape-casting lamination with molten metal
infiltration. Porous ceramic preforms were (ZrC>2 ) prepared by tape casting/lamination
technique. A pyrolyzable pore-forming agent (graphite) added to the suspension, which
yields a porous structure after binder removing and sintering treatments. Infiltration
experiments were carried out under nitrogen atmosphere at temperatures between 7501000°C by using Al-10%Mg alloy as an infiltrating alloy.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
13
Low et al [41], studied the processing o f functionally graded AI2 T 1O 5 -AI2 O3 composites by
infiltration. Layered alumina preforms prepared by infiltrating the preform by immersing
them in a solution o f TEOT (tetra-ethyl-ortho-titanate). Infiltrated preforms then heat-treated
and sintered. The product provides high thermal shock resistance and good mechanical
strength. A similar approach was used by Pratapa et al [42] to produce functionally graded
aluminum titanate mullite-ZTA (zirconia toughened alumina).
The methods explained so far require an initial preform preparation. The following provides
a summary o f some o f the methods, which do not require such an initial step.
2.2.3. Casting: The graded casting process is based on the slip casting technique used for
shaping ceramics. Graded casting offers the possibility o f manufacturing relatively complex
parts having continuous or stepwise structural or compositional gradients [43],
Baosheng et al [44] produced TiC-2020Al FGMs by in-situ reaction melting and subsequent
centrifugal casting. They reported that TiC particles distributed evenly in the aluminum alloy
and improved mechanical and physical properties were obtained. Yeo et al [45] produced
ZrC>2 and stainless steel 316 (SU S316) FGMs by tape casting method in an aqueous system.
Tape casting method offered to make thin sheets of uniform thickness. And it was suggested
that as the number of compositional gradient increases the residual stresses induced on FGMs
were relaxed. After tape casting samples were sintered at 1350°C in Ar/FL atmosphere for 2
hours. Sintering defects due to the different shrinkage rates o f the starting materials was
controlled by adjusting the particle size and phase type ofZrC>2 .
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
14
2.2.3. Plasma Spraying: Plasma spraying is an attractive process method for functionally
graded materials, because it can simultaneously melt highly refractory phases and a metal,
blending the two in ratios that can be preset by control of the relative feeding rates o f the two
powdered materials. Different processing atmospheres can be used, e.g. air, low pressure gas,
and vacuum [46],
Plasma spraying is the most attractive method o f forming thermal barrier coatings (TBCs).
These coatings reduce the surface temperatures o f metallic components. A typical plasma
sprayed TBC system consists o f an oxidation resistant metallic bond coat overlaid on a
thermally insulating ceramic coating. The metallic surface o f the bond coat produced by
plasma spraying is usually rough and irregular. This provides a good bonding along the
ceramic/metal interface [47,48], The conventional TBC is usually in the form o f a two layer
(non-FGM) coating. This design presents a problem that the coating spalls from the substrate.
A solution to this problem is to produce coatings in graded fashion. Ceramics/metal graded
coating is a new type o f thermal barrier coatings (TBCs) which compositions varied
gradually from metal bond layer to ceramics surface layer and eliminate the macro interface
between ceramics and metal that lies in common duplex TBCs , so the thermal stress could
be relaxed and the thermal shock property could be largely improved. (TBCs) are
commonly used to protect air-cooled superalloy hardware in hot sections o f gas turbine
engines.
Due to Y2 O3 partially stabilized Zr 0 2 has a relatively low thermal conductivity and large
coefficient thermal expansion compared with other structural ceramics and MCrAlY
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
15
(M=Ni,Co,Fe etc.) alloy has excellent oxidation resistance at elevated temperatures, plasma
sprayed ZrC>2 -Y2 0 3 /McrAlY coatings has become the most important thermal barrier
coatings. Yin et al achieved Zr02/NiCoCrAlY graded coating with excellent gradually
compositional distribution and displays typical laminar structure. The relative density o f the
graded coating is 90% or so, and contains a few pores [50],
Coating ZrC>2 on metals using NiCrAlY interlayer is a well-established method, used for high
temperature protection o f metals, such as aircraft turbine engine components. Nippon Steel
Corp. has investigated the gradient composition coating process in ZrC>2 -NiCr system by
using a single gun low-pressure plasma system with four ports. Both, ceramic and metal
powders are fed to the plasma arc at the same time [1,21].
Kim et al [51], used plasma spraying technique for coating Cr 2 0 3 on mild steel substrate. To
improve the bond strength, a composite interlayer (C^Cb/NiCr) is incorporated between
NiCr and Cr2 0 3 layers.
Porosity graded titanium implant materials are produced by varying the powder size during
spraying. Wang et al [52], utilized plasma spraying technique to produce bioactive
functionally graded coatings of hydroxyapatite on Ti-6A1-4V plates.
Fukushima et al [53] utilized plasma twin torch spraying method for producing TiC/Mo
functionally graded coatings on Mo substrate. Graded coatings (coating thickness ~ 320-330
pm) were obtained by stepwise control o f both TiC and Mo powder feeding rates. A smooth
and linear composition gradient was achieved. After coating the graded material is heat
treated at 1470K for 16 hours in vacuum.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
16
2.2.5. Laser Cladding: Laser alloying/cladding has been used to deposit layers o f different
compositions to produce functionally graded materials o f corrosion resistant and hard facing
alloys like NiCrBSi and composite materials like WC/Co. In laser beam cladding, the surface
o f the substrate is melted by using a laser beam and at the same tim e a consumable powder
(metal, ceramic or metal/ceramic mixture) is blown into the melted zone. The low melting
point material becomes completely molten and the other species (usually a ceramic) remains
as solid particles or partially dissolves. Thus, a surface layer consisting o f a particlereinforced composite is obtained [54-56].
Jasim et al [54], produced compositionally graded 3 layers of Al-SiC on nickel alloy
substrate by laser processing. In another study Abboud et al [52], fabricated a functionally
graded Ti-Al/TiB2 composite on a pure Ti substrate. A powder mixture o f Ti, Al and TLEL
was fed into a laser generated melt pool on a titanium substrate and layers with progressively
increasing Al content were produced.
2.2.6. Vapor Deposition Processes: Vapor deposition processes have the advantage that
very thin layers can be produced. Ti-TiC and Ti-TiN multilayer films were deposited onto
various substrates by using hollow cathode argon plasma assisted reactive physical vapor
deposition. Carbon or nitrogen was introduced in the coatings using controlled additions o f
N 2 and C2 H 2 gas to plasma gas. Layers are composed of Ti, Ti+Ti2N, Ti2 N+TiN, and TiN.
Schulz et al [57], achieved compositional grading of NiCoCrAlY coatings by electron beam
physical vapor deposition (EB-PVD). The grading was obtained by variation of processing
parameters, such as power, focus, and electron beam pattern, thereby changing the local heat
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
17
input into the melt. Because vapor pressures o f constituent elements differ widely, the actual
evaporation rate can be influenced.
Leushake et al [58], produced Al2 0 3 -ZrC> 2 graded thermal barrier coatings by electron beam
physical vapor deposition technique. This concept combines the thermal insulation potential
o f zirconia with the low oxygen diffusivity o f alumina. Alumina and zirconia are co­
evaporated from aluminum and zirconium ingots. Due to the difference o f melting points and
evaporation behavior, a composition gradient is achieved. Seifried et al [59], used a
combined CVD/CVS (chemical vapor synthesis) to produce gradient films consisting o f
nanograined material in Si-C-B system. Concentration gradients are obtained simply by
variation o f precursor mass flows.
To avoid the oxidation of C/C composites Sato et al [60] studied coating of oxidation
resistant SiC. C/C composites are applied to high temperature structures. The SiC coating
layer consists of two layers: a thin SiC conversion layers (several pm) and a thick coating
layers. The conversion layer was formed by chemical reaction o f gas phase Si with carbon in
the substrate. The aim of the conversion layer is to make bonding strength of the coating
higher. The CVD layer was thermally deposited at three different temperatures 1200°C,
1600°C and 1800°C with nominal thickness o f 200pm, 100pm, and 60pm.
Salito et al [61] have studied graded copper-tungsten coatings onto a copper substrate by
plasma spraying. Clyne et al [62] studied deposition of zirconia (ZrC>2 - 8 %Y 2 0 3 ) and
NiCrAlY (Ni-22wt%Cr-l0wtAl-lwt%Y) onto Nimonic superalloy substrates by plasma
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
18
spray technique. The function of the interlayer NiCrAlY was suggested to be two fold. First
to improve the adhesion between substrate and zirconia and second, to protect the substrate
from oxidation by gases penetrating through the Z 1O 2 layer.
2.2.7. Reactive Processes: Functionally graded materials can be prepared by reactive
processes by exploiting a rapid exothermic chemical reactions. A multilayered compact with
a composition gradient is usually prepared by stacking layers o f reactant powder mixtures in
appropriate amounts. Each of these mixtures contains a slightly different percentage o f
reactants. FGM fabrication by reactive processes is the main theme of this dissertation. The
details o f the method will be given in the following sections.
2.3. Overview on Functionally Graded Materials
The term o f FGM is now widely accepted by the materials community, which was originated
in Japan in the late 80s. However, the graded material is not something new, humans have
extensively utilized materials containing microstructural gradients long ago [63].
The research on graded materials has gained a remarkable momentum after the first
international symposium on functionally gradient materials, which was held in Japan in 1990.
Unlike Japan, in the United States, there is no centralized national program on Functionally
Graded Materials. However, there is an expanding interest in areas such as thermal barrier
coatings for propulsion systems, wear resistant coatings for ground vehicle components,
structural components for space and military applications. Some o f the centers where
research activities on FGMs are currently in progress are: Alfred University, Caterpillar
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
19
Corporation, Colorado School o f Mines, University o f Idaho, Idaho National Engineering
and Environmental Laboratory, Lawrence Livermore National Laboratory, Los Alamos
National Laboratory, MIT, United Technologies Research Center, Westinghouse R&D
Center [63-65],
From the literature review given in this chapter, it can be concluded that there is a
considerable activity in the field o f functionally graded materials. The key issue in
developing ceramic/metal FGMs is to identify an optimum compositional gradation or
microstructure according to the response to service loads or environments. An important
aspect at the present stage seems to be related to the question of reliability and durability of
FGMs. The methods to ascertain the quality, reproducibility, and reliability by accepted
standards and procedures are insufficient. Considerable basic research is required to fully
understand and predict the behavior along with applied research on processing and
performance evaluation. Considering these factor and ever-increasing need for improved
materials, it is evident that there is a need to further investigate. It is obvious that the interest
in gradient materials will continue to increase as more is learned about their potential [6465].
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
20
3. References
1.
S. Suresh and A. Mortensen “Fundamentals o f Functionally GradedMaterials”
The
University Press, Cambridge, (1998).
2.
M. Koizumi, Ceram. Trans. Func. Grad. Mat., 34, ed. by J.B. Holt, M. Koizumi, T.
Hirai and Z. A. Munir, The American Ceramic Society, (1993).
3.
A. Mortensen and S. Suresh, Int. Mat. Rev., 40, 6, (1995), 239-265.
4.
R. Rawlings, Mat. World, 3, 10, (1995), 474-475.
5.
A.J. Markworth, K.S. Ramesh and W.P. Parks Jr., J. o f Mat. Sci., 30, (1995),21832193.
6.
M. Finot, S. Suresh, C. Bull and S. Sampath, Mat. Sci. & Eng. A205, (1996), 59-71.
7.
M. Dao, P. Gu, A. Maewal and R.J. Asaro, Acta Mater., 45, 8, (1997), 3265-3276.
8.
B.L. Wang, J.C. Han, S.Y. Du, J. of Therm. Stres., 23, (2000), 143-168.
9.
N. Noda, J. of Therm. Stres., 22, (1999), 477-512.
10. K.S. Ravichandran, Mat. Sci. & Eng. A201, (1995), 269-276.
11. A.S. Kim, S. Suresh and C.F. Shih, Int. J. Solids Struct., 34, 26, (1997), 3415-3432.
12. M. Finot and S. Suresh, J. Mech. Phys. Solids., 44, 5, (1996), 683-721.
13. P.R. Marur and H.V. Tippur, J. of Test. & Eval., 26, 6, (1998), 539-545.
14. F. Erdogan, MRS Bull, Jan 1995, 43-44
15.
D.P. Miller, J.J. Lannutti and R.D. Noebe, J. Mater. Res., 8 , 8, (1993), 2004-2013.
16. Y. G. Jung, S.C. Choi, C.S. Oh and U.G. Paik, J. of Mat. Sci., 32 (1997), 3841-3850.
17. M. J. Pindera, J. Aboudi, and S. M. Arnold, Functionally Graded Materials (1996) eds:
I. Shiota and M.Y. Miyamoto, 113-121
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
21
18. W.A. Gooch, B.H.C. Chen, M.S. Burkins, R. Palicka, J. Rubin and R. Ravichandran,
Mat. Sci. Forum, Trans. Tech. Publications, 308-311, (1999), 277-282.
19. E.S.C. Chin, Mat. Sci. and Eng. A259, (1999), 155-161
20. S. Maruno, S. Ban, Y. F. Wang, H. Iwata, and H. Itoh, J. o f Ceram. Soc.o f Japan, 100,
4, (1992), 362-367.
21. M. Niino and S. Maeda, ISIJ Int., 30,9, (1990), 699-703.
22. K. Sogabe, M. Tanaka. T. Miura, and M. Tobioka, Functionally Graded Materials
(1996) eds: I. Shiota and M.Y. Miyamoto page: 155-158
23. X. Zhu, W. Wang, Z. Meng, J. ofM at.. Sci. Let., 14, (1995), 516-518.
24. A. G. Lamas, M Almeida, and H.M.M. Diz, Ceram. Int., 19, (1993), 121-124.
25. K. Darcovich and M.E. Price, J. o f Can. Ceram. Soc., 66,2, (1997), 146-151.
26. Y. Takao, M Awano, Y. Kuwahara, and Y. Murase, J. Mater. Res., 9, 8, (1994), 21282132.
27. N.J. Lee, J. Jang, M. Park, C.R. Choe, J. Mat.Sci., 32, (1997), pp:2013-2020.
28. R. Watanabe, MRS Bull., Jan 1995, 32-34.
29. J.S. Gam, K.S.Han, S.S. Park, and H.C. Park, Mat. and Man. Proc.,14, 4, (1999), 537456.
30. B.H. Rabin, R.L. Williamson, H.A. Bruck, X. Wang, T.R. Watkins, Y. Zhong and D R.
Clarke, J. Am. Ceram. Soc., 81, 6, (1998), 1541-1549.
31. A.Kawasaki, and R. Watanabe, Ceram. Trans. Func. Grad. Mat., 34, ed. By J. B. Holt,
M. Koizumi, T. Hirai and Z. A. Munir, (1993), 157-164
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
22
32.
A. Kawasaki, J.F.Li, R. Watanabe, and T. Matsubara, 3rd Int. Symposium on Structural
and Functional Gradient Materials, 10-12 October 1994, Lausanne Switzerland, eds: B.
Ilscher and N. Cherradi, 433-439.
33.
A. Bishop, C.Y. Lin, M. Navaratnam, R. D. Rawlings, H.B. McShane, J. of Mat. Sci.
Let., 12, (1993), 1516-1518.
34.
J.C. Zhu, S.Y. Lee, Z.D. Yin and Z.H. Lai, Functionally Graded Materials (1996) eds:
I. Shiota and M.Y. Miyamoto, 203-208
35.
J J . Hilden and K.P. Trumble, Mat. Sci. Forum, Trans. Tech. Publications, Switzerland,
308-311, (1999), 157-162
36.
M. Joensson, U. Birth and B. Kieback, Functionally Graded Materials (1996) eds: I.
Shiota and M.Y. Miyamoto, 167-172
37.
F. Stromberger, R. Jedamzik, A. Neubrand and J. Rodel, Mat. Sci. Forum, Trans. Tech.
Publications, Switzerland, 308-311, (1999), 163-168
38.
M. M. Gasik, 3rd Int. Symposium on Structural and Functional Gradient Materials, 1012 October 1994, Lausanne Switzerland, eds: B. Ilscher and N. Cherradi, 575-579.
39. B. R. Marple, D. J. Green, J. o f Mat. Sci. 28, (1993), 4637-4643.
40. S.F. Corbin, X. Zhao-Jie, H. Henein and P.S. Apte, Mat. Sci. & Eng. A262, (1999),
192-203.
41.
I. M. Low, R.D. Skala, and D. Zhou, J. ofM at. Sci. Let., 15, (1996), 345-347.
42. S. Pratapa and I. M. Low, J. ofM at. Sci. and Sci. Let., 15, (1996), 800-802
43. J. Chu, H. Ishibashi, K. Hayashi, H. Takebe, and K.Morinaga, J. of Ceram. Soc. o f Jap.,
Int. Edition, 101,7, (1993), 818-820.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
23
44.
Z. Baosheng, Z. Jingchuan, Z. Yongjun, Y. Zhongda, C. Hongsheng, A. Geyin,
Functionally Graded Materials (1996) eds: I. Shiota and M.Y. Miyamoto. 179-184
45.
J. G. Yeo, Y.G. Jung, and S. C. Choi, YeO Functionally Graded Materials (1996) eds: I.
Shiota and M.Y. Miyamoto, 185-190.
46.
S. Eroglu, N.C. Birla, M. Demirci, and T. Baykara, J. Mat. Sci. Let., 12 (1993), 10991102
47.
.
C.H. Hsueh, P.F. Becher, E.R. Fuller, S.A. Langer and W.C. Carter, Mat. Sci. Forum,
Trans. Tech. Publications, Switzerland, 308-311, (1999), 442-449.
48.
M. Alaya, G. Grathwahl, and J. Musil, 3rd Int. Symp. on Struc. And Functional Gradient
Materials, 10-12 Oct. 1994, ed by B. Ilschner and N. Cherradi, 405-411.
49.
C.J. Jian, T. Hashida, H. Takahashi, N. Shimoda, and M. Saito, 3rd Int. Symp. On Struc.
And Functional Gradient Materials, 10-12 Oct. 1994, ed by B. Ilschner and N.
Cherradi, 419-424.
50.
Z. Yin, X. Xiang, J. Zhu and Z. Lai, Functionally Graded Materials (1996) eds: I.
Shiota and M.Y. Miyamoto, 269-274.
51.
H.S. Kim, H.H. Yang, I.H. Oh, and S.S. Kim, Mat. Sci. Forum, Trans. Tech.
Publications, Switzerland, 308-311, (1999), 226-231.
52.
M. Wang, X.Y. Yang, K.A. Khor, and Y. Wang, J. ofM at. Sci: Mat. In Medicine, 10,
(1999), 269-273
53.
T. Fukushima, S. Kuroda, S. Kitahara, K. Ishida and M. Sano, Functionally Graded
Materials (1996) eds: I. Shiota and M.Y. Miyamoto, 245-250.
54.
K.M. Jasim, R.D. Rawlings, and D.R.F. West, J. Mat. Sci., 28, (1993), 2820-2826.
55.
J.H. Abboud, D.R.F. West, and R.D. Rawlings, J. ofM at. Sci., 29, (1994), 3393-3398.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
24
56. T. Seefeld, C. Theiler, E. Schubert and G. Sepold, Mat. Sci. Forum, Trans. Tech.
Publications, Switzerland, 308-311, (1999), 459-466.
57. U. Schulz, K. Fritzcher, M. Peters, and W. A. Kaysser, 3rd Int. Symp. On Struc. And
Functional Gradient Materials, 10-12 Oct. 1994, ed by B. Ilschner and N. Cherradi,
441-446.
58. U. Leushake, U. Schulz, T. Krell, M. Peters and W. A. Kaysser, Proc. o f the 4th Int.
Symp. On Functionally Graded Materials, Tsukuba, Japan, (1996), edited by I. Shiota
and Y. Miyamoto, Elsevier Publishing, 263-268.
59.
S. Seifried, M. Winterer, and H. Hahn, Mat. Sci. Forum, Trans. Tech. Publications,
Switzerland, 308-311, (1999), 277-282.
60.
N. Sato, I. Shiota, H. Hatta, T. Aoki, and H. Fukuda, Functionally Graded Materials,
1996, Edited by I. Shiota and M. Y. Miyamoto, 257-262.
61. A. Salito and G. B arbezat, 3rd Int. Symp. On Struc. And Functional Gradient Materials,
10-12 Oct. 1994, ed by B. Ilschner and N. Cherradi, 123-128.
62. T.W. Clyne and Y.C. Tsui, 3rd Int. Symp. On Struc. And Functional Gradient Materials,
10-12 Oct. 1994, ed by B. Ilschner and N. Cherradi, 129-136.
63. B.H. Rabin and I. Shiota, MRS Bull., 20,1, 14-19.
64. S. Suresh, A. Mortensen, and H. McManus, MRS Bull., 20, 1, 54-55.
65. S. Suresh, 3rd Int. Symp. On Struc. And Functional Gradient Materials, 10-12 Oct.
1994, ed by B. Ilschner and N. Cherradi, 709.
Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission.
25
MANUSCRIPT TWO
FUNCTIONALLY GRADED Ti-TiB-TiB2 COMPOSITES PRODUCED BY
COMBUSTION SYNTHESIS -I: PROCESS ANALYSIS
Menderes Cirakoglu
Department o f Materials and Metallurgical Engineering
University o f Idaho
Moscow-Idaho 83844-3024
will be submitted for publication
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
26
ABSTRACT
Combustion synthesis (CS) process employs exothermic reaction between the constituents
and reduces difficulties associated with conventional methods e.g. time and energy. In this
study, CS process was utilized in producing functionally graded materials in Ti-B binary
system. This system was chosen mainly due to the desirable properties of titanium borides in
this system. Depending on the composition, the reaction between Ti and B can be vigorous
and may result in explosion. This study describes the approach to control the reactions by
selecting compositions away from highly exothermic TLB2 stoichiometric composition
instead of adding a third element into the system. Thermochemical calculations were carried
out to provide information about the adiabatic temperature reached by the products for
selected compositions. The influence o f composition on the combustion process, temperature
profiles and phase formation was investigated experimentally. It is shown that, due to high
exothermicity, TiB and TiB2 formations are possible even at off-stoichiometric compositions.
As expected, higher boron content compositions yielded higher combustion temperatures and
combustion wave velocities. X-ray diffraction studies showed the formation o f TiB and TiB2
along with unreacted Ti metal as matrix. Due to the pre-heating occurring in FGM
compacts, the measured reaction temperatures were higher than corresponding single
compositions. These facts corresponded well with the thermodynamic analysis.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
27
1. Introduction
Combustion synthesis is an attractive technique to synthesize a wide range o f materials for
engineering applications. A variety o f process constraints, material selection and fabrication
aspects are involved with this process and needs to be described. Therefore, in this
dissertation, the broad topic was divided into two sections. The basic concepts o f combustion
synthesis method in Ti-B binary system are discussed in this chapter. The specifics o f
process development and product characterization issues are kept for the following chapter.
In this chapter, starting with the general information on combustion synthesis, associated
issues such as thermodynamic concepts and fundamentals are addressed. Part o f the
introduction presents a comparison o f combustion synthesis technique with other FGM
manufacturing techniques. A detailed review o f processing methods is provided in the
previous chapter. As a consequence o f wide and diverse methods, we narrowed our
comparison to oniy on constructive powder densification methods where FGMs are produced
layer by layer stacking. The Ti-B binary system is explained and finally the objectives o f the
present study are mentioned.
1.1. Combustion Synthesis - A Brief Overview
The combustion synthesis (CS) technique has various advantages, including the simplicity o f
the process, higher purity of the products and relatively low energy requirements. Currently
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
28
the method is utilized to prepare many advanced ceramics, intermetallic and composite
materials [1-6].
In a typical combustion synthesis process, a powdered mixture (which is capable o f
undergoing an exothermic reaction) is cold pressed into a cylinder then placed into an
evacuated chamber to prevent oxidation. Surface o f the sample is exposed to an external heat
source (e.g. a heated coil, electric match or laser beam) for a short time. Once the mixture is
ignited, a strong exothermic reaction liberates enough heat to the adjacent layer o f reactants
and the reaction becomes self-sustaining and propagates in the form o f a combustion wave
[7-8]. As the combustion wave advances, the reactants are converted to the product(s) as
shown in Figure 1.
Reacted portion o f the sample
unreacted portion o f th e sample
Figure 1. Combustion wave propagation
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
29
1.1.1. Thermodynamic Considerations
A fundamental understanding o f the combustion synthesis process can be obtained from a
thermodynamic analysis of the process. The theoretical background of thermodynamics o f a
combustion system is well established and a number of different systems were studied and
reported in the literature [1-8]. Studies in this category are aimed to provide a set o f criteria
that can be used to predict the formation o f self-sustaining combustion wave phenomenon.
The adiabatic temperature (Tad) can be used as a semi-quantitative evaluation for whether a
reaction is self-sustaining or not. It can be defined as the maximum temperature to which the
product is raised under adiabatic conditions (thermally insulated) as a consequence o f
evolution o f heat due to the chemical reaction [1-2],
The adiabatic temperature can be calculated from thermodynamic principles. The
temperature can be calculated from the thermodynamic functions of heat capacities (Cp) and
enthalpies o f formation (AHf) and transformation (A H t) and fusion (AHf). Merzhanov et al
[3] suggested that systems with Tad < 1800K will not become self-sustaining. For a given
reaction system, calculation o f the adiabatic temperature can be done graphically by plotting
both the reactant and product enthalpy vs temperature [1-3].
If one assumes that all of the energy evolved from the reactants to form the product goes into
heating that product up to some temperature, then the point at which the product enthalpy
equals that o f the reactants (at the ignition temperature) will represent the maximum
temperature o f the reaction (T ad). For a reaction, taking place in a thermally isolated
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
30
exothermic system, the sum o f the initial reactant enthalpies is equal to the sum o f the
product enthalpies and therefore the heat balance can be written as follows
H ? ( T 0 ) = H 0p (Tad)
(1 )
where T0 is the initial temperature o f the reactants and Tad the final temperature o f the
products.
The adiabatic temperature can be calculated from the enthalpy o f reaction in other words the
heat released during the reaction by using the equation given below
AH° = A /7 ° 298 +
f A C p ( product) d T
(2)
298
where AH°f 2Si is the standard enthalpy o f formation of the product at 298K, ACp is the
change in heat capacity for the formation o f the product.
Due to the fast kinetics o f combustion reactions, it is reasonable to assume that the reactions
are pseudo-adiabatic if not adiabatic. For the sake of simplicity, adiabatic system assumption
is usually made where AH °
~
/.2 9 8
=
0. Hence equation (2) becomes
= \ ACP(product ) dT
if
Tad
<Tm
(3)
298
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
31
On the other hand, if adiabatic temperature equals to melting point o f the reactants (Tad=
T mp)
T,
= JACp{product) d T +v&Hm
(4)
298
where v is the fraction o f the product that is in the liquid state, and AHm is the heat o f
fusion o f the product. Complete solid-state combustion occurs when the melting points are
less than the adiabatic temperature. In this case (Tad > Trap), the relationship will be
- A H 7j??!Sg =
^m p
Tod
298
Tmp
J ACp {product,solid) dT + AH m + j*ACp {product,liquid)dT
(5)
where, ACp (product, solid) and ACp (product, liquid) are the heat capacities o f the product
in its solid and liquid phases, respectively [1-3].
In general, the experimentally observed combustion temperature is lower than the calculated
temperature. Nevertheless, the theoretical Tad may be used as a semi-quantitative evaluation.
1.1.2.
Comparison to other Constructive Processing Techniques
FGM processing methods can be divided into two principal classes: constructive processes
and transport based processes. In the first class, FGMs are constructed layer by layer (e.g.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
32
powder densification, coating, lamination etc.). In the second class, the gradients are created
by transport phenomena based processes (e.g. mass transport, thermal processes, settling and
centrifugal separation, macrosegregation etc.) [10-12].
Powder densification processes are an important sub category under the constructive
processes. They are widely used in producing bulk FGMs and can be studied in 4 groups:
conventional solid-state powder consolidation, liquid phase sintering, infiltration, and
reactive powder processes [10-11].
FGM processing with these methods starts with the fabrication of a preform compact
containing the desired gradient. Usually the compact is stacked and densified in a closed die
system. Depending on the method utilized, one or more o f the following conventional
procedures are utilized: uniaxial pressing, cold isostatic pressing, hot isostatic pressing or by
hot pressing. All o f these methods have been demonstrated to be successful [10-11],
The utilities o f uniaxial pressing and cold isostatic pressing are limited. Hot pressing offers a
one step process wherein an applied pressure drives macroscopic flow and enhances
densification. The die preserves the shape o f the component by avoiding deformation and
failure of the powder compact. However, the main drawback with these techniques is that
they are energy intensive processes and require expensive equipments. Liquid phase sintering
is an attractive process and widely studied in WC-Co FGMs. Although this method provides
a much more rapid densification kinetics compared to solid phase sintering, there is an
important drawback in processing FGMs. During liquid phase sintering, the liquid phase
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
formation and motion o f the phase across the component diffuses the sharp gradients initially
placed in the structure. The degradation o f layered structure is significantly accelerated when
pores are present. This problem may be overcome with infiltration method [10-11],
Reactive processes i.e. combustion synthesis, offers a fast, cheap, and a reliable method in
producing FGMs. These aspects are quite beneficial in overcoming the shortcomings of
conventional processes for manufacturing graded materials. Compared to techniques
mentioned above the most essential feature of this approach is that synthesis and
densification may be achieved simultaneously. The energy required is provided by the
strongly exothermic reaction between the constituents making it a cheaper process. Further,
expensive equipment are not necessary as in the case o f hot pressing and hot isostatic
pressing. Reactions are usually accompanied by the formation of at least one liquid phase,
which aids densification. Due to the fast reactions, the duration of the process and the
residence time o f the liquid phase is shorter compared to other techniques. Compared to
liquid phase sintering, FGM fabrication by combustion synthesis is less prone to the
elimination o f the gradient. Compared to the green compact, after combustion reactions, the
sharp boundaries separating layers have been smoothened which indicates some local
homogenization due to transport between the components. A common problem with
combustion synthesis is the formation o f porosity. This issue will be addressed in the
following chapter in detail. Here it is suffice enough to indicate that this problem is closely
related to the combustion temperature. When the temperature is too high, evaporation of a
phase present in the structure may occur [13-16]. Therefore, in order to lower the
temperature, a low melting point metal is usually added to the system as a diluent. This issue
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
has been addressed by Sata [17], In his study on SiC-TLE^/Cu, they added the product, TiB2
to the system. They concluded that this method lowered the reaction temperature and
avoided the vaporization o f Cu in the system. However, it has been mentioned that the
amount o f diluent added is important. Excessive dilution can cause insufficient heat release,
which is critical for densification and reaction propagation. This parameter is also important
in terms o f the stability of the combustion front. An unstable combustion front results in
microstructural heterogeneity. Nevertheless, combustion synthesis is an efficient and
developing area in the production o f FGMs. The process provides high temperatures and high
velocity reactions, which result in high rates of mass and heat transfer phenomena. The heat
released during the process is utilized which reduces the cost of energy and simplifies the
equipment. The main drawback o f porosity can be overcome by physical treatment o f the
final product [15].
1.2. Ti-B Binary System
Titanium- boron binary system was chosen for several reasons. The products in the system
(TiB and TiB 2 ) exhibit extreme hardness, high electrical conductivity, good thermal shock
resistance, high melting point, chemical inertness and durability. These materials are
candidates for a variety of applications, including refractory, abrasive and electrochemical
processes. TiB 2 is a covalently bonded ceramic with a hexagonal crystal structure typified
AIB2 . In this crystal structure, layers o f titanium and boron atoms alternate in hexagonal
coordination. TiB 2 , which has a great potential not only as a refractory material, but also as
an electronic material because of its high melting point, hardness, electrical conductivity and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
35
thermal conductivity. They have been used as vacuum metallizers and as cathodes in HallHeroult cells for aluminum production. TiB2 is also considered to be an important component
in the fabrication of functionally graded materials for use as heat shielding/structural material
in future space applications. TiB is used as ballistic armor, coating for cutting tools and also
in the electrolytic production o f aluminum instead o f carbon electrodes due to its superior
cathode quality [18-22], A functionally graded material with composition changing from
pure titanium on one side to Ti/TiB/TiB2 composite on the other side can find several
potential application areas. Considering the ballistic properties of titanium and titanium
borides, they can be used in military applications such as armor, armaments, and military
vehicle structures. Graded composite armor structures consist o f a hard component for
destroying the projectile tip and a backing structure for maximizing the strength and energy
absorption [23,24],
The binary phase diagram for Ti-B system is shown in Figure 2. The system contains three
intermediate phases: TiB and Ti3 B 4 which decompose incongruently (peritectic
decomposition) at 2350 K and TiB2 which melts congruently at 3193 K [25,26], During the
first stages o f combustion synthesis, thermal energy (i.e. heat) is provided to the sample at
the sample’s ignition surface. This results in an increase in the enthalpy o f the reactants (Ti
and B). The increase in enthalpy, in turn, raises the temperature of the sample. When the
local temperature reaches to 1166 K, a solid-state phase transformation occurs in Ti. The
crystal structure of Ti changes from oc-Ti (hep, a=0.295nm, c=0.468nm) to (3-Ti (bcc,
a=0.332nm). When the temperature reaches 1936K, titanium starts melting. Since both phase
transformation and melting are endothermic events, local heat within the sample is affected
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
36
by this phenomena. As titanium continues to melt, solid boron dissolves into the liquid
titanium. When titanium becomes saturated with boron, boride formation starts. Titanium
monoboride has a narrow homogeneity range o f about 49 to 50 at% B. On the other hand,
the homogeneity range o f titanium diboride is 65.5 to 67 at%B.
The formations of borides are exothermic process (e.g. 323.8 kJ/mol for TiB2 ) and contribute
most o f the enthalpy to the system, which manifests itself a sudden increase in the local
temperature. When the reaction between constituent elements is extremely violent, the
control o f the process becomes necessary. The control of such reactions might be sought
through either by dilution, or by stoichiometric changes. The diluent acts, as a moderator to
the violent elemental reactions. Diluent phase is similar to a heat sink, and distributes the
heat o f reaction. The lowering o f the adiabatic reaction temperature with dilution results in a
continuous decrease in the velocity of the CS reaction. It was also reported that previously
reacted products can also be added as a diluent to the unreacted powder mixtures in order to
slow down the reaction [27-32].
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
37
Atom ic Ftercent Boron
o
*Q 30^0 90
y>
ro
3000-
ZBOO
1BOO -
900 ..
lO
20
40
HO
66
Wfejght Percen t B oron
70
too
Figure 2. Ti-B binary phase diagram [22].
Previously in our laboratory, Peng et al [22] carried out a systematic research on combustion
behavior of single compositions in Ti-B binary system. They achieved controlled combustion
reactions by carefully selecting appropriate compositions by shifting away from the
stoichiometric titanium borides. They pointed out that compositions containing Ti-30 and 25wt% B, the combustion reactions were uncontrolled yielding explosion into small pieces. The
composition containing 20 wt% B yields a sponge like layered structure. On the other hand,
controlled combustion reactions were observed with compositions Ti- 18,15 and 12 wt% B.
Although, these selected compositions are away from the stoichiometric boride compositions,
TiB and TiB 2 phases were observed in the end products.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
38
1.3 Objectives
In this research, our primary goal is to introduce combustion synthesis as an alternative
method for the fabrication of functionally graded materials in Ti-B binary system. Despite
the fact that there is a large number o f research on combustion synthesis o f titanium borides
available in the literature, to the best o f our knowledge no published work analyzes the
processing parameters o f functionally graded materials in Ti-B binary system. As mentioned
above, several groups added a diluent phase to the reactants in order to control the reaction.
However, this usually results in the formation o f undesirable intermetallics. In this study, we
have taken a different approach to the problem by shifting stoichiometry away from the
diboride composition. Although the basic concepts o f combustion synthesis method are well
established and easy to apply, with this new approach a number o f questions such as
combustion characteristics and phase formations yet need to be investigated. Therefore, this
first part o f the research can be considered as a proof-of-feasibility study.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
39
2. Experimental
In this section, the procedure o f fabrication of graded layers, experimental set-up and
combustion experiments are explained. Since knowledge of high temperature phase
formation is o f importance, we utilized high temperature X-ray diffraction (HTXRD)
technique. The results obtained from this technique was compared with and explained by
means o f thermodynamic evaluations. The combustion wave velocities were measured as
they provide an understanding o f the mechanisms involved during combustion. Combustion
temperature profiles were measured by using thermocouples and a two-color pyrometer in
the zones parallel to the reaction front. Accurate determination o f temperature o f the reacting
mixture, and its variation with time and position is a key point in our study. Such analysis
provides information on heat transfer and reaction kinetics, which are difficult to obtain by
other means. In an attempt to elucidate the effect o f atmosphere on the combustion, several
series o f experiments were carried out under Ar atmosphere.
2.1. Materials
Elemental Ti (99% purity, -325 mesh, Johnson Mathey, Ward Hill, MI) and crystalline B (>
99% purity, average 15 pm or less, Cerac Inc., Milwaukee, WI) were used as starting
materials. Titanium and boron powders were mixed in 82/18, 85/15, 90/10 and 95/5 wt %
ratios. Dry mixing was carried out in polyethylene bottles with alumina as grinding medium.
The process involves several steps. The flow chart shown in Figure 3 depicts the basic steps.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
40
2.2. Phase Formation Studies
High temperature phase formation studies were carried out by using high temperature X-ray
diffraction (HTXRD) technique. HTXRD (Phillips X ’Pert MPD System) analyses are carried
out on several compositions (CoKa radiation, 35kV and 35mA). The system consists o f an Xray diffractometer (0-20 geometry equipped with an incident beam monochrometer), a high
temperature furnace with a Be window, a programmable temperature controller, and an I-N2
cooled single crystal Ge detector. The furnace has internal radiation baffles in order to
decrease the effect o f radiated heat. Isothermal measurements and variable temperature
programs can be carried out between room temperature and 1050°C (the maximum rated
temperature is 1250°C). The powder mixtures of different compositions are placed on an
alumina holder and heated up with a heating rate o f 50° min'1. Argon gas is passed through
the system to prevent oxidation. The X-ray scans are recorded in-situ, after soaking for 10
minutes, upon reaching the programmed temperatures (namely, 300,600,900 and 1050°C).
2.3. Preparation of Single Composition Pellets and FGMs
Samples with homogeneous compositions (fixed boron content) were studied to elucidate the
combustion behavior o f each composition. Samples o f 5, 10, 15, and 18-wt % boron
prepared. The consolidation o f the powder mixtures was accomplished first by uniaxially
pressing cylindrical compacts. A stainless steel die with double acting rams was used and an
axial load of 15,000 lbs was applied. The resulting compacts have typical dimensions o f 12.7
mm and about 14 mm long. The compacts are further densified using a cold isostatic press
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
41
(ISI Inc., Columbus OH) at a pressure o f 100 MPa. The goal was to achieve consistent
compaction procedure.
The graded layers were prepared by stacking up the different compositions in a steel die in a
layered fashion. The required amount o f powder for each layer was loaded into the
cylindrical steel die. Each layer was vibrated and tapped gently before the next layer was
added to obtain flat interfaces and uniform compaction. The geometry and a picture o f the
green compact is shown in Figure 4 (a) and (b), respectively. Powder stacking method offers
microstructure and geometry control. This method is general enough to allow preparing even
larger samples.
Prior to combustion experiments, the as-prepared samples are de-gassed at 600°C for 3 hours
under flowing argon to remove the adsorbed moisture and gases. A vacuum furnace (Centorr
Assoc. Inc., Suncook, NH) was used for degassing. In the absence o f such a treatment, the
samples exploded during the reaction because o f intensive gas evolution.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
42
^
Dry Mixing
Q
Uniaxial Pressing
j in g l e Compositions)
c
^
^
^
1
Graded Layers
)
HTXRD
Q
Cold Isostatic Pressing
c
)
I
Degassing
j
l
^ Combustion Synthesis
^
Characterization
*)
Figure 3. Flow chart of experimental procedure
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
43
12.7 mm
(a)
(b)
Figure 4. Schematic drawing (a) and a picture (b) of a 5 layered green sample.
2.4. Experimental System
A schematic drawing of the experimental set-up used in combustion experiments is shown in
Figure 5. This reactor was designed and built in our laboratory. The photograph of the
experimental set up and data acquisition system is shown in Figure 6. The set-up consists of
a die cavity (~ 24 mm in diameter and 12.20 mm in height), a mechanical roughing pump, Ar
gas, and view port, pressurized air activated punch, tungsten coil.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
44
LM -
C u n tiu ln
Heatmg c o ls
VUeo
Cam era
T S a l —I
m
f T
-P = g
^
r
/ y t t
O j art**viirxto»r
Data Acquastidnv.w
.w.v.w.-,
Systetn
Power
Supply
o ^
LJ
U
Figure 5. A schematic drawing o f the experimental set-up [98],
Figure 6. Photograph o f experimental set-up showing the reactor, data acquisition system
and pyrometer arrangement.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
45
The die is made out o f steel and the cavity is about 1 inch in diameter. The main function of
this cavity is to hold the sample in place during experiments. The gap between the sample
and the walls o f the die was about 5 mm. This gap was filled with thermal insulation to
insulate the sample from the metal fixture. For this purpose we used alumina felt at the
bottom and around the sample (Type AB, Zircar Fibrous Ceramics, NY). The thermal
insulation also provided lateral confinement o f the pellet during compaction.
2.5. Combustion Experiments
The green compact was placed into the die cavity and the chamber was flushed with argon
gas a few times and then evacuated again by using a mechanical pump attached to the
system. Our preliminary experiments with functionally graded samples showed that ignition
from higher boron content layer resulted in uncontrollable ignition. Therefore these samples
were ignited from the Ti side. A tungsten coil with a wire diameter of 2 mm (Model
H2.040W, R.D. Mathis Co., Long Beach, CA) was used as an external heat source by passing
an electric current through it. The coil was placed at 2-3 mm from the top o f the sample as
shown in Figure 7. An AC power supply capable of 25 V and 120 A output was used.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
46
W-coil
Sample
Die cavity
Figure 7. Die assembly used in CS/DC experiments
The sample first heated up by radiation from the tungsten coil and after some elapsed time
the surface o f the sample reached to a certain ignition temperature and the sample ignited. A
combustion front formed and started to propagate downwards as shown in Figure 8. The
view port facilitated the observation o f the combustion wave and recording the combustion
process. The rate of propagation o f the combustion wave was measured by a video camera
through the view port at a speed o f 60 frames per second. The time required to initiate a selfpropagating combustion wave depends upon the heat source as well as thermal properties of
the sample. It is usually difficult to measure the exact ignition temperature.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
47
W coil
Sample
Alumina Blanket
Thermocouple
Figure 8. Video image taken during the passage of combustion wave.
When a non-contact measurement technique (e.g. a pyrometer) was used, a high ignition
temperature was usually measured due to the interference o f the heat source. In order to
monitor the temperature distribution in the samples, thermocouples were inserted into
different location in the samples. The temperature measurement configuration consists of a
data acquisition system to record and process thermocouple output voltage data and two Ctype thermocouples (0.24-mm diameter, W-5%Re/W-26%Re). The junction o f the
thermocouples was inserted through preexisting holes drilled in the green pellets.
Thermocouples were placed in different axial and radial positions in order to get an accurate
temperature distribution within the sample. The average bead size o f the thermocouples was
0.63 mm and the maximum service temperature was rated as 2316°C with short-term
exposure up to 2760°C. The thermocouples, output the current (in milliamps) readings in real
time during the experiment to a host computer. A data acquisition board (WB-Dynares,
Omega Engineering, Inc. Stamford, CT) together with a graphical interface application
software (QuickLog PC,Omega Engineering, Inc., Stamford, CT) was used to collect data at
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
48
every 0.5 s intervals. The DAQ system has eight differential analog input channels. The
thermocouple wires were connected to the specified channel’s screw terminals inside a
terminal panel. The terminal panel is connected to the data acquisition board inside the host
computer with a 50-pin ribbon cable.
After the experiments samples were left to cool inside the chamber. The samples were cross­
sectioned and the phase constituents were identified at room temperature by X-ray diffraction
(Siemens D5000, Ni filtered CuKa radiation) employing a Cu tube (wavelength: 0.15406
nm), operating at 40 kV and 30 mA. The step scanning employed was 0.05° per second.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
49
3. Results and Discussions
This section describes the process issues and results o f the experimental work. It begins
with a general analysis on the effect o f composition on the adiabatic temperatures. The
results o f high temperature phase formation studies are presented and explained
according to the calculated thermodynamic data. Subsequently, the combustion
characteristics o f single compositions and FGMs are presented and explained. The issues
cover the combustion wave velocity studies, time-temperature profiles and phase analysis
in the products.
3.1. Thermodynamic Considerations
Most o f the experimental facts can be explained by use o f thermodynamic calculations,
which can also predict combustion mechanisms. Therefore, in this section, our objective
is to present the results of thermodynamic analysis for the preparation o f Ti-TiB-TiB 2
composites. The results are expected to predict the adiabatic temperatures.
As explained previously, in our experimental system, we investigated different
compositions (Ti - 5, 10,15 and 18 wt %B) in the Ti-B binary system. Based on these
different compositions, four general reactions can be written as follows. In these
reactions, “x” refers to the excess molar amount of titanium where x = 0 refers to the
stoichiometry. We assumed that the reactions proceed to form the desired boride phases
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
50
and depending on the amount o f titanium present in the reactants, there may be unreacted
Ti present in the product phase.
reaction 1:
(1+x) Ti +2B
= TLB2 + xTi
reaction 2:
(1+x) Ti+ B
= T iB + x T i
reaction 3:
(2+x)Ti + 3.5B
= 1.5TLB2 + 0.5TiB + xTi
reaction 4:
(2+x)Ti + 2.5B
= 0.5TiB2 + 1.5TiB + xTi
First concern is the thermodynamic feasibility o f these reactions. The feasibility o f the
reactions is largely determined by their free energies o f formation. These calculations
were carried out by using reported Gibbs free energies o f formation at temperatures
ranging from 298 K to 4000 K as shown in Figure 9. In our studies, all thermodynamic
data were taken from JANAF Thermochemical Tables [33]. In this temperature range, the
Gibbs free energy o f the reactions (AG™) is negative which suggest that they are
thermodynamically possible throughout this temperature range. Among the competing
reactions, reaction 3 is thermodynamically the most favored, while the reaction 2 is the
least favored.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
51
-100
reaction 2
reaction 1
^ -3 0 0
u.
reactio l 4
-4 0 0
reaction 3
-5 0 0
-6 0 0
0
1000
200 0
3000
4000
Temperature (K)
Figure 9. Free energy versus temperature for the reactions.
3.1.1. Adiabatic Temperatures for Stoichiometric Compositions (x=0)
The second important aspect o f the thermodynamic analysis is to calculate the adiabatic
temperatures of the reactions from the energy balances. As explained in the introduction
section, the adiabatic temperature is an empirical criterion for determining the feasibility
o f combustion synthesis. The energy balance is given by the enthalpy curves (Figures 1013) from which the maximum temperature of the product for an adiabatic process can be
estimated.
These calculations were first carried out for the reactions in their stoichiometric
compositions where x = 0. During the chemical reaction, if any o f the components
undergoes a change o f state or order (it may be a transformation, fusion or evaporation)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
52
within the temperature range under consideration, additional heat effects must be taken
into account. The heat o f transformation o f a reactant must be subtracted from the total
while that o f a product must be added. In the above reactions, Ti undergoes a —> P solidstate phase transformation at 1166K accompanied by a heat o f transformation o f 0.997
kj/mole, and melts at 1963K accompanied by a heat of fusion o f 14.146 kJ/mole. On the
other hand, B melts at 2350K with a heat o f fusion o f 50.2 kJ/mole. In the products, TiB 2
melts at 3 193K accompanied by a heat o f fusion o f 100.4 kj/mole. The results o f the
thermodynamic calculations o f reactions are schematically shown in Figures 12-15 and
the estimated adiabatic temperatures and melting point of reactants and end products are
listed in Table I.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
53
1000
Ti b.p..
Enthalpy (kJ/m ol)
750
B m.p
5 00
Ti m
a -p
250
iHr29S
Tad
-2 5 0
PRODUCTS
-5 0 0
0
500
1000
2000
1500
2500
30 0 0
3500
4000
g
Temperature (K)
10. Enthalpy-temperature plot for reactants and products for reaction 1, where
reaction stoichiometry
o f x = 0.
10 0 0 T
TiB
Enthalpy (kJ)/mole
750
500
250
REACTAI
PRODUCTS
-2 5 0
-5 0 0
0
500
1000
1500
2000
25 0 0
30 0 0
3500
4000
Temperature (K)
Figure 11. Enthalpy-temperature plot for reactants and products for reaction 2, where
reaction stoichiometry o f x = 0.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
54
2000
1750
1500
1250
2
1000
fC
750
£
500
UJ
250
EAi
'T ad
-250
PRODUCTS,
-500
0
500
1000
1500
2000
2500
3000
3500
4000
Temperature (K)
Figure 12. Enthalpy-temperature plot for reactants and products for reaction 3, where
reaction stoichiometry
o f x=0.
1750
1500 H
a)
1250
|
1000
2
750
a!
500
ID
c"
250
REACTANTS
UJ
Tad
-250
CT
-500
0
500
1000
150 0
2000
2500
3000
3500
4000
Temperature (K)
Figure 13. Enthalpy-temperature plot for reactants and products for reaction 4, where
reaction stoichiometry
o f x=0.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
55
Table I. The adiabatic temperatures o f reactions for x=0
Reaction
No
Melting Temperature (K)
Ti
B
TiB2
3193
Reaction 1
1936
2350
Reaction 2
1936
2350
Reaction 3
1936
2350
Reaction 4
1936
2350
Estimated
TiB’
—
Tad (K)
3200
2350
3350
3193
2350
3200
3193
2350
3200
—
Decomposition temperature
The above thermodynamic analysis indicates that, the estimated Tad is higher than the
melting temperature o f both o f the reactants. For reactions 1,3 and 4, adiabatic
temperatures were estimated to be 3200K. Since it is very close to the melting
temperature o f TiB2(3250K), we can assume that a melting might be occurring partially
in a narrow reaction zone. However, in reaction 2, a slightly different adiabatic
temperature was calculated (3350K). This result suggests that the formation o f TiB is
energetically more likely to occur, leading to higher exothermicity. Further evidence for
this conclusion will be presented in the high temperature X-ray diffraction results. In
reaction 2, the calculated adiabatic temperature exceeds the melting temperature o f
reactants and decomposition temperature of the product phase. The product phase, TiB,
decomposes into liquid and TiB 2 at 23 5OK.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
56
During the pre-ignition stages, thermal energy is added to the sample at the sample’s
ignition surface, thereby increasing the enthalpy o f the sample in the vicinity o f ignition
surface. The increase in enthalpy raises the temperature o f the sample originating at the
ignition surface. The local temperature increases until the melting temperature o f titanium
is reached. Liquid formation within the porous compact causes a redistribution of
particles and increases the contact area (intermaterial area) between liquid titanium and
solid. As titanium continues to melt, boron dissolves into the liquid, which results in the
formation of borides (an exothermic event). Since the melting o f titanium is an
endothermic event and expected to inhibit sudden temperature rise due to the exothermic
reaction, an energy redistribution may occur within the sample. Furthermore, some of the
generated heat will be lost from the compact to the surroundings by convection and
radiation. The sums o f all these contributions dictate whether or not the reaction will be
self-sustaining. The involvement o f the liquid phase strongly intensifies combustion,
enhances the contact between reactants resulting in greater homogeneity. As indicated
previously, while calculating adiabatic temperature, the system was assumed to be a
closed system. However, in reality it is obvious that, heat is lost to the surroundings and
therefore the reaction temperature will be lower than the calculated temperature.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
57
3.1.2. Adiabatic Temperatures for Mon-Stoichiometric Compositions (i> 0 )
One o f the main problems encountered in CS process, when applied specifically to
borides and carbides, is the process control. Sometimes the heat generated is too high and
results in an uncontrolled reaction with explosion o f the sample. In situations, where the
reaction between the constituent elements is extremely violent, the control o f the process
becomes necessary. In the literature, it was reported that the control o f CS reaction is
possible by means of dilution, or by stoichiometric changes [6,18]. The diluent acts, as a
moderator to the violent elemental reactions. The diluent phase is similar to a heat sink,
and distributes the heat o f reaction. The lowering of the adiabatic reaction temperature
with dilution results in a continuous decrease in the velocity o f the CS reaction. It was
also reported that previously reacted products can also be added as a diluent to the
unreacted powder mixtures in order to slow down the reaction [34],
Our goal is to control the vigorous exothermic reactions without adding a third element as
a diluent into the system. Shifting away from the stoichiometric compositions o f titanium
borides allowed us to do so. The deviations from the initial stoichiometric ratios o f
reactants (Ti and B, in our system) can also decrease the adiabatic temperature as in the
case for the dilution. In the latter case, the decrease in the Tad is mainly due to the
lowering o f the heat o f reaction caused by the presence of unreacted reactant Ti within
the final product phase. To confirm this, we carried out further thermodynamic studies.
So far, the discussion has focused on stoichiometric compositions where x = 0. Further
calculations were conducted in order to see how the amount o f Ti in the reactants,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
58
changes the thermal effect and corresponding adiabatic temperature. The ratio o f the
constituents was changed by shifting the composition away from the stoichiometric TiB
and TiB 2 compositions. In these calculations, the amount o f excess Ti (x was varied from
0 to 1.5) was considered in the energy balance and the results are presented in Figures 1417.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
59
2000
1500
o
2
1000
>-
a. 500
«i
zz
c
UJ
Tad
d3
Ta 12
-500
0
Figure 14.
500
1500
20 0 0
2500
Temperature (K)
1000
3000
T ad l
4000
3500
Enthalpy-temperature plot for reactants and products for reaction 1 with
reaction stoichiometries of x = 0, x = 0.5, x = 1 and x = 1.5.
2000
EM EU
1500
>-
D
i
i
i
1000
■Y -
CL
X —
500
1
-
0
-500
l, 5j
BBBH
1 1 Tad3
Tad4
1
- ■■■;—-i — i— v ■■ . . . T. . .. — ,—
500
Figure 15.
1000
SB
S!
1
Tad2
I----!---- r —r - -1— :
1500
2000
2500
Temperature (K)
1-j
—-
3000
l
T
U -- .— i
3500
4000
Enthalpy-temperature plot for reactants and products for reaction 2 with
reaction stoichiometries of x = 0, x = 0.5, x = 1 and x = 1.5.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
60
3000
2500 H
5 - 2000
o
.E
2
to
1500
1000
500
c
lU
-5 0 0
-1 0 0 0
0
500
1000
1500
2000
3000
2500
4000
3500
T e m p e r a tu r e (K)
Figure 16.
Enthalpy-temperature plot for reactants and products for reaction 3 with
reaction stoichiometries o f x = 0, x = 0.5, x = 1 and x = 1.5.
3000
2500
2000
1500
.c 1 0 0 0
C
UJ
500
Jtsflu
J2JL
T ac1
-5 0 0
0
500
1000
1500
2000
Tad2
2500
Tad3 Tad4
3000
3500
4000
T e m p e r a tu r e (K)
Figure 17.
Enthalpy-temperature plot for reactants and products for reaction 4 with
reaction stoichiometries of x = 0, x = 0.5, x = 1 and x = 1.5.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
61
Examination of the above analysis shows that, an increase in the amount o f excess Ti
decreased the temperature o f the combustion reaction and the corresponding adiabatic
temperatures also decrease. Increasing the amount o f titanium in the reactants, results in
the presence of a large amount molten phase during reaction. The results are tabulated in
Table II.
Table EL The change in adiabatic temperature with composition.
Adiabatic Temperatures (K)
Reaction
Number
off-stoichiometric
stoichiometric
x=0
x = 0.5
x= 1
Reaction I
3193
3000
2500
1939
Reaction 2
3350
2500
2000
1939
Reaction 3
3193
3100
2800
2300
Reaction 4
3193
3000
2600
2000
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
x = 1.5
62
3.2. High Temperature X-Ray Diffraction Studies
The HTXRD patterns o f compositions 82Ti-18B and 90Ti-10B are shown in Figure 1819, respectively.
counts/s
14
12
1050C
10
8
900C
6
4
2
* iwi
300C
0
°2Theta
Figure 18. HTXRD pattern o f 82Ti-18B composition (Co K« radiation X= 1.7889A)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
63
counts/s
15-
• Ti
A TiB2
♦ TiB
1050C
■**KOO,’!
900C
1
:
600C
. . . .
<
5-
m ■_ ■
300C
0
15
~l~l~l ■•’ »I
20
25
30
35
4-0
45
50
60
°2Theta
Figure 19. HTXRD pattern of 90Ti-10B composition (CoKa radiation A.= 1.7889A)
From the high temperature X-ray diffraction (HTXRD) results, some conclusions can be
drawn regarding to the mechanisms involved in the reactions. It seems evident that the
higher boron content results in more boride phase formation. Titanium monoboride forms
at low temperature around 900°C prior to the formation of titanium diboride. It is
speculated that there are two main reasons for this observation. First, the compositions
selected are closer to or equal to stoichiometric TiB stability region. Second, the
estimated adiabatic temperature for TiB (3355K) is higher than that of TiB 2 (3200K) as
predicted in the thermodynamic analysis. It is important to mention here that due to the
heat losses, adiabatic temperatures are never reached, as it will be proven in the following
sections. This is the explanation why the conversion to TiB was not completed in
composition 82Ti-18B, which is stoichiometric composition of TiB.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
64
3.3. Combustion Experiments with Single Compositions
3.3.1. Combustion Characteristics
The combustion reactions began from the top surface and propagated nearly uniformly
towards the bottom as shown by the sequence o f photographs in Figures 20 and 21. The
figures indicate that the reaction became self-sustaining after ignition with the
propagation velocity strongly depending on the initial composition. The decrease in the
combustion wave velocity is mainly due to two reasons: reduced volumetric heat release
and reduced mass diffusivity. The first is related to the thermodynamics o f the system.
The thermodynamic studies suggested that, any deviation from stoichiometry results in a
reduction in Tad- Further, the experimental data show that such compositions also result in
a decrease in the combustion rate. A combustion wave velocity o f 2.7 mm/sec was
determined for 85-15B single compositions and 1 mm/sec for 90Ti-10B. The lower
boron content compositions resulted in is less exothermic and more sluggish reactions in
nature as compared to the higher boron content.
t=0 seconds
t=1 seconds
t=2 seconds
t=3 seconds
t=5 seconds
Figure 20. Combustion wave propagation images from combustion o f 85-15B
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
65
t=0 seconds
t=2 seconds
t=4 seconds
t=6 seconds
t=12.8 seconds
Figure 21. Combustion wave propagation images from combustion o f 90-1 OB
It is important to note that the bright cloud surrounding the compact is due to the
oxidation of tungsten coil and expulsion o f adsorbed gases. No significant shape change
was observed. Uncontrolled vigorous combustion reactions were observed with single
compositions 85-15B and 82-18B, when samples ignited without a degassing treatment.
The samples suddenly exploded when the combustion wave reached half way down the
sample as shown in Figure 22. Adsorbed gases on the surface o f the boron are believed to
be the cause of this type of behavior. The final product no longer resembled a pellet but
consisted o f crumbled pieces.
Figure 22. 82Ti-18B composition (without degassing) after combustion reaction. Sample
exploded after the combustion wave propagated after half way down the sample.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
66
3.3.2 Time-Temperature Profiles
C-type thermocouples were embedded into a small hole drilled at different locations. A
typical time- temperature profile along the sample is shown in Figure 23 for a single
composition o f 82Ti-18B. The temperature reached 200°C in about 10 seconds and then
thermocouples and pyrometer recorded an abrupt rise in temperature followed by
relatively slow cooling rate to below 100°C. The temperature suddenly increased to
1850°C fo rtc l and 1450°C fortc2. Since tel was embedded halfway inside the compact,
it exhibited a higher temperature compared to tc2. This sharp increase (at about 12
seconds) in the profile marks the arrival o f the combustion front at the region o f the
thermocouple. Both o f these measured temperatures are much lower than predicted Tad,
presumably due to the heat losses. The measured combustion temperatures by
thermocouples are likely to be lower than the actual combustion temperature within the
sample, mainly due to the fast kinetics characteristic associated with the combustion
reactions. It is noted here that during the combustion process, the reaction rate is very fast
while the thermocouples response rate is limited. The temperatures consistently dropped
within 80-90 seconds and the resultant cooling curves exhibited an exponential decay.
For this particular sample, pyrometer measurements were carried out by focusing onto
the surface o f the sample just below the heating coil. The maximum temperature recorded
by the pyrometer was around 3300°C. This composition corresponds to the
stoichiometric TiB composition and in the thermodynamic section an adiabatic
temperature o f 3350K was estimated (Table m , reaction 2). However, it should be noted
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
67
here that since the point was too close to the tungsten coil, it is highly possible that some
o f this temperature is actually from the tungsten coil.
py ro m eter
•7.5 m
•11.5 mn
•14 mm
tc1
tc 2
3 mm
12 mm
3500
pyrometer
3000 2500 -
3
2000
tc1
9- 1500 -
I—
1000
500
- ________
L-
0
10
20
30
40
50
60
70
80
90
time (seconds)
Figure 23. Schematic drawing o f the thermocouple arrangement in the compact and
temperature profile measured with two W-5%Re/W-26%Re and a pyrometer for
combustion reaction in single composition o f 82Ti-18B.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
100
In Figure 24, the time-temperature profile obtained during the ignition o f single
composition 95Ti-5B is shown. It is informative to compare the thermal profiles o f
compositions 82Ti-18B and 95Ti-5B. It can be seen that both o f the plots essentially
showed a steady ramp up to around 200°C and then temperature suddenly increased. This
general profile was observed for other single compositions. The temperature suddenly
increased to 1375°C for tel and 950°C for tc2. The binary phase diagram shown in the
introduction section indicates that at 95Ti-5B composition, both of these temperatures lie
below the eutectic temperature (1540°C). Thus the reactions are mostly in the solid state.
Experimentally measured temperatures were comparatively lower than estimated values,
which are, as explained before, attributable to the heat losses. It is, however, clear that,
the maximum temperature attained decreased as the relative ratio of Ti in the compact
increased. This result is in good agreement with our thermodynamic results and shows
the effect o f initial stoichiometric ratios. More importantly this observation confirms that
the formation o f the titanium borides was the primary heat generation process which
gives rise to the self-propagating behavior o f the combustion wave.
After the combustion wave had passed, the temperature decreased by dissipating the
generated heat to the surroundings by radiation and conduction through the sample’s
outer surface. The conduction heat loss also takes place internally (internal conduction)
within the sample by transferring the heat to the adjacent layers. The heat loss by
radiation is a negative factor for combustion synthesis type processing methods.
However, it is inevitable. As summarized before, if the heat losses were too high, the
reaction would be localized and as a result the combustion wave would not propagate.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
69
ii
-7.5 m n
*12 mm
td
*15 mm
tc2
3 mm
12 mm
1600
combustion
duration
1400
1200
pre-combustion
duration
1000
800
cooling
tc1
600
'V
tc2
400
200
0
0
50
100
150
200
time (sec)
Figure 24. Schematic drawing o f the thermocouple arrangement in the compact and
temperature profile measured with two W-5%Re/W-26%Re and a pyrometer for
combustion reaction in single composition o f 95Ti-5B.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
250
70
As thermodynamic calculations indicated, 82Ti-18B composition is more exothermic
than 95Ti-5B. Moreover, this composition involves a liquid phase (titanium) and hence as
Figures 20 and 21 shows it is relatively rapid.
3.3.3. X-ray Diffraction Studies
The room temperature X-ray diffraction patterns from the cross sections o f ignited single
compositions are shown in Figures 25 and 26. The X-ray diffraction patterns of the
ignited compacts reveal that, although some boron reacted with titanium, the conversion
o f titanium and boron is not fully completed, even with the composition o f Ti-18wt% B
which corresponds to stoichiometric TiB composition. In this composition, we observed
peaks o f TiB and TLB2 in significant quantities besides unreacted elemental Ti. The XRD
results indicated that as the boron content increased, the formation o f borides was also
promoted. The single composition o f Ti-5wt%B only showed the presence o f titanium
phase. This result is expected due to the fact that the amount of boron was very low.
When the boron content increased to 10-wt %, we observed the formation o f TiB (JCPDS
Card # 5-700) and TiB2 (JCPDS card # 35-741) in minor amounts. It is important to note
here that, although the composition Ti-10wt% B is away from the stoichiometric TiB (Ti18wt%B) and TiB2 (Ti-30wt%B) due to the high activation energy for the Ti-B reaction it
is possible to see the formation o f these phases as proposed by Holt et al [35].
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
71
Figure 26 shows that, the relative amount o f borides increased as the boron content
increased from 15wt% to 18wt% B. It can be seen that the presence o f TiB is significant
but not to the extent expected from the composition as 82Ti-18wt%B, which corresponds
to TiB stoichiometric composition. Unreacted Ti was still present in the reacted compact.
This might be explained by involving time-temperature profile data. The profiles clearly
indicated that, the real temperatures are much less than the estimated adiabatic
temperatures.
It is to be noted that the TiB ( t i l ) and Ti (002) peaks are overlapped; therefore the
existence o f a peak in this 20 position (~ 38.4°) may suggest the presence of both Ti and
TiB phase. In composition 85Ti-15B, although in the x-ray diffraction pattern the
maximum relative intensity for TiB corresponds to TiB (201), it is not consistent with
JCPDS standard data. The JCPDS standard have maximum intensity for TiB (102). A
similar phenomenon was also observed in composition 82Ti-18B, the maximum intensity
correspond to TiB (210) however, the JCPDS standard data predicts TiB (102) as the
maxima. A possible explanation for this could be related with preferred orientation in the
Ti matrix.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
72
2-theta (degrees)
Figure 25. X-ray diffraction patterns o f single compositions (95Ti-5B) and (90Ti-10B)
after combustion synthesis.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
73
o
CM
CM
O
m m
P P
82Ti-18B
CM
ZZ
w
CM
ca
P
cm '
O
a O
P h
W
mm
nil
C/3
C
<D
85Ti-15B
ni^H %■* my
30
35
40
45
50
2-theta (degrees)
Figure 26. X-ray diffraction patterns o f single compositions (85Ti-15B) and (82Ti-18B)
after combustion synthesis.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
55
74
3.4. Combustion Experiments with Graded Compositions
3.4.1. Combustion Characteristics
As explained previously, functionally graded samples were prepared by stacking up
different compositions. Similar to single composition experiments, a brightly glowing
zone, which moves along the specimen from the ignition region, was observed. The
combustion was recorded by a video camera through the view port.
t=0
t=12sec
t=14sec
Figure 27. The propagation of combustion wave through the graded layered compacts.
When the samples were ignited from the boron rich layers, they deformed and in some
cases exploded therefore, we chose to ignite from the Ti layer.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Combustion wave velocities are obtained by noting the time involved for the reaction
front to travel from the top to the bottom of the sample. The cylindrical compacts o f 3
layered FGMs ignited from the Ti side after a sufficient amount o f energy transferred
from the tungsten coil to the sample. The reaction front propagated along the length o f
the sample in a controlled fashion as shown in Figure 27. The bright section on the upper
side corresponds to the reacted portion of the sample. According to Figure 27, as the
boron content increases the corresponding combustion velocity becomes higher as well.
This is an expected result knowing that as the B content increases the exothermicity o f
the reaction also increases.
3.4.2. Time-Temperature Profiles
The position of tel roughly corresponds to 90Ti-10B layers, and that o f tc2 corresponds
to 82Ti-18B layers. The ignition time, also known as pre-combustion duration can be
defined as the length of time required initiating a self-propagating combustion wave. It is
usually difficult, if not possible, to identify this length o f time precisely. From Figure 28,
an approximate pre-combustion time was determined as 110 seconds. Compared to single
composition samples the pre-combustion time is longer with FGMs. This is an expected
result mainly because unlike FGMs, single compositions do not have a titanium layer.
The titanium layer in the FGMs delays combustion. Prior to the arrival o f the combustion
wave at the exact thermocouple positions, the local temperature at these thermocouple
positions was raised gradually essentially by conduction. During this time, the
conduction heat transfer was primarily responsible for the preheating o f the reactants.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
76
The effect o f preheating is an important parameter in combustion synthesis as explained
in process parameters in the introduction section. By preheating, the internal heat losses
(heat losses to the layers ahead as well as behind of the combustion wave) are reduced.
The result is, as seen in the time-temperature profiles, a higher local temperature. Once
the combustion wave reached to the thermocouple positions, a high rate o f heat
generation occurred, which manifested itself as a sudden rise in the temperature. As can
be seen from Figure 28, a maximum temperature of 1900°C was recorded from tel
position and of 2150°C from tc2 position. After the ignition started, the temperature
started increasing rapidly at tc2 location mainly due to its location being close to the heat
source.
Similar to the time-temperature profiles obtained from the single compositions, the
profiles have two distinct stages; a slower first stage in the beginning and a fast second
stage. However, the profile obtained from te l position exhibited a plateau at around
1000°C. This plateau formation delays the temperature evolution. This plateau formation
might be indicating an endothermic heat exchange since the tungsten coil continued to
heat the sample. It may be possible that fine titanium particles start melting locally at this
time. Further, since the thermocouple tips were insulated by alumina cement, the reai
temperatures within the sample are possibly higher than the measured temperatures. After
adequate amount o f liquid had formed, the exothermic reaction begins and results in a
sharp increase in the temperature. On the other hand, the occurrence o f the temperature
plateau is not as distinct at tc2 profile, however a reduction in the temperature ramp is
clearly visible. A sudden rise in temperature had occurred simultaneously at both
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
77
positions at around 110 seconds from the beginning o f the experiment. This increase is
more remarkable in tc2 position, which corresponds to the highest boron content layer in
the FGM. The temperature at tc2 position reaches to 2150°C and at te l position it is
around 1900°C. It is evident that higher boron contents yield to reactions with higher
exothermicity. After this sharp rise, the temperature gradually decreased room
temperature. Another important observation was related to the cooling rates. From the
initial portion o f the cooling curves cooling rates of 41°C/sec and 33°C/sec were
measured fortc2 and te l, respectively.
In Figure 29, the variation o f measured temperature with respect to time is shown for 3
positions along the center o f the sample. The profiles indicate that the temperature before
ignition is relatively low. Therefore, the influence of heat losses by radiation and
convection is smaller than conduction. During pre-ignition stage, the net rate o f energy
accumulation o f the layers increase and when it exceeds the ignition temperature, or the
accumulated energy in the heated layer can overcome the barrier o f the activation energy,
the ignition o f the layer is achieved. Once the layer ignites, the high temperature o f the
combustion helps to sustain the combustion wave. Although it takes longer ignition time,
when compared to the temperature profiles of the single compositions, it is clear that
maximum temperatures attained in FGM arrangement are higher than corresponding
single compositions. The relatively longer ignition times of FGMs is due to the fact that
the samples were heated from the Ti side instead of highly exothermic, high boron
content side. The explanation for this phenomenon can be attributed to preheating effect.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
78
90-1 OB
2500
combustion
duration
2000
O
3
C
3
k->
1500
pre-combustion
duration
cooling
tc1
5 - 1000
E
o
E-1
:c2
500
0
50
100
150
200
time (secon d s)
Figure 28. Schematic drawing o f the thermocouple arrangement in the compact and
temperature profile measured with two W-5%Re/W-26%Re and a pyrometer for
combustion reaction in a five layered functionally graded material.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
250
79
- 6 mrr
-3.5 mn i
-tc1
-tc2
- 1 6 mm
-14mm
■tc3
12 mm
2500
2000
tc3
tc2
1500
ca
w
2E
iooo
tel
a
E-
500
0
50
100
150
200
250
time (seconds)
Figure 29. Schematic drawing of the thermocouple arrangement in the compact and
temperature profile measured with three W-5%Re/W-26%Re for combustion reaction in
a five layered functionally graded material.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
80
3.4.3. Effect o f atmosphere
The combustion chamber has several ports that provide attachments for a gas purge. After
flushing a few times with argon the chamber was filled with argon gas and sealed.
Similarly, thermocouples were located away from the ignition source. The samples were
thermally insulated from the stainless steel die cavity by using alumina fiber blanket at
the bottom and around the sample. Temperature profiles were recorded in real time by
using C-type thermocouples. The time temperature profiles obtained from thermocouples
at certain locations are shown in Figures 30 and 31.
The time-temperature profiles showed that the heating rates are higher under argon
atmosphere as compared to vacuum. The maximum temperature recorded in argon
atmosphere was 1200°C while the maximum temperature recorded in vacuum was
around 2190°C, lower by around 1000°C. Part o f the reactants did not participate in the
reaction, which in turn reduced the heat of reaction. No sudden temperature increase was
observed in the experiments carried out in argon atmosphere, which indicates that the
ignition did not take place. Keeping the ignition coil on longer times (up to around 150
seconds) did not help substantially to ignition. Only a very small temperature increase at
around 240 second was observed. A severe oxidation o f the tungsten coil had occurred
and it was difficult to see through the quartz window. This is believed to cause
inefficiency in the energy input to the sample and eventually inhibited the ignition take
place. After the experiment, the outer surface of the sample was covered with tungsten
oxide layer.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
81
90-1 OB
2500
in vacuum
2000
O
23
1500
42)
in argon
I* 1000
500
0
50
100
150
200
250
300
350
time (seconds)
Figure 30. Schematic drawing o f the thermocouple arrangement in the compact and
temperature profile measured with W-5%Re/W-26%Re for combustion reaction in argon
atmosphere and in vacuum in a five layered functionally graded material.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
400
82
iL
- 8 mr i
95-5B
-1 6 mm
90-1 OB
85-15B
82-18B
12 mm
2500
2000
in vacuum
1500
in argon
2- 1000
500
0
0
50
100
150
200
250
300
350
400
tim e (seconds)
Figure 31. Schematic drawing o f the thermocouple arrangement in the compact and
temperature profile measured with W-5%Re/W-26%Re for combustion reaction in argon
atmosphere and in vacuum in a five layered functionally graded material.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
83
3.4.4. X-Ray Diffraction Studies
Typical XRD patterns after combustion experiments o f FGMs are shown in Figures 32
and 33. X-ray diffractograms obtained from individual layers are very similar to that o f
the corresponding single compositions indicating TiB and TLB2 formation and remaining
unreacted titanium as a matrix phase.
350
O
300
o
250
1/1
c
<v
4->
c
o o
CH
200
CD
CD
CD
150
- 90Ti-10B *T
100
t
95Ti-5B
30
35
40
45
50
55
2-Theta (degrees)
Figure 32. X-ray diffraction patterns from two layers (95Ti-5B and 90Ti-10B) o f FGM
after combustion synthesis.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
84
350
300
250 -I
I I
»
a>
i/i 200
c01
l (
p n
o■
CXoit
*-F
p
■g 150
i e
100
85Ti-15B
50
r*
30
35
40
45
i ■ • < * « 1(1.1
50
55
2-Theta (degrees)
Figure 33. X-ray diffraction patterns from 2 layers (85Ti-15B and 82Ti-18B) o f FGM
after combustion synthesis.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
85
4. Conclusions
Our work on processing of functionally graded materials in Ti-B binary system by using
combustion synthesis have revealed the following conclusions:
1. The results obtained have demonstrated that the fabrication o f Ti/TiB/TiB2 net shape
FGMs using a method based on combustion synthesis is possible. The method
suggests an attractive and low cost route in synthesis and densification of this type of
materials.
2.
The FGMs were fabricated from titanium and boron powders with nonstoichiometric ratios. Stoichiometric mixtures o f TiB 2 react almost explosively with
reaction temperatures rising almost instantaneously above 2200°C. Shifting the
composition away from the stoichiometry created excess titanium phase, which acted
both as reactant and diluent. Although the excess amount o f titanium decreased the
adiabatic temperature, it provided a control on exothermic reactions.
3. High temperature X-ray diffraction studies have revealed some important results
regarding to the reaction mechanisms. As expected, the higher boron content
compacts showed more boride phase formation. Probably the most important
information we gained is that although the selected compositions are away from the
stoichiometric TiB2 composition, due to the high exothermicity of the reaction, this
phase was observed as a second phase along with TiB.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
86
4. The maximum reaction temperatures recorded for single compositions 82Ti-18B and
95Ti-5 B were 1850°C and 1375°C, respectively. The results are in good agreement
with our thermodynamic calculations in the sense that as the boron content increases,
the reactions become more exothermic. However, in both cases, the measured
temperatures were lower than predicted theoretical calculations due to the heat losses.
In the case o f FGMs higher local temperatures were measured. This is attributed to
pre-heating.
5. The desired products, i.e., TiB and TiB2 in a titanium matrix, were produced.
6. In argon atmosphere, no clear ignition and combustion wave propagation was
observed.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
87
§. References
1.
H.C. Yi and J.J. Moore, J. o f Mat. Sci., 25, (1990), 1159-1168.
2.
Z.A. Munir, Ceram. Bull., 67,2, (1988),342-349.
3.
A.G. Merzhanov and B.I. Khaikin, Energy Combust. Sci., 14, (1988), 1-98.
4.
J. B. Holt and Z. A. Munir, J. Mat. Sci. (1986), 251-259.
5.
C. He and G.C. Stangle, J. Mater. Res., 13,1, 135-145, (1998).
6.
B. Moss, Combust. Sci. and Tech., 98, 337-340, (1994).
7.
J.B. Holt, Ceramics and Glasses, Engineered Materials Handbook, 4, ASM
International, (1991), 227-231
8.
B.I. Khaikin and A.G. Merzhanov, Comb. Expl. and Shock Waves, 2,3, (1966), 3646.
9.
G.K. Dey, A. Arya and J.A. Sekhar, J. Mat. Res., 15,1, (2000), 63-75.
10. S.Suresh and A. Mortensen “Fundamentals of Functionally Graded Materials” The
University Press, Cambridge, (1998).
11. A. Mortensen and S. Suresh, Int. Mat. Rev., 40, 6, (1995), 239-265.
12. W. Lai, Z.A. Munir, B.J. McCoy and S.H. Risbud, Proc. o f the 4th Int. Symp. on
Functionally Graded Materials, Tsukuba, Japan, (1996), edited by I. Shiota and Y.
Miyamoto, Elsevier Publishing, 275-281.
13. H.C. Yi and J.J. Moore, J. o f Mat. Sci., 25, 1159-1168, (1990).
14. R.Z. Yuan, Int. J. o f SHS, 6, 3, (1997), 265-270.
15. G.C. Stangle and Y. Miyamoto, MRS Bull., (1995), 52-53.
16. H.J. Feng and J.J. Moore, J. o f Mat. Sci. and Perform., 2, 5, (1993), 645-650.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
88
17.
N. Sata, Ceramic Trans. Functionally Gradient Materials, 34 edited by J.B. Holt, M.
Koizumi, T. Hirai andZ. A. Munir, The American Ceramic Society, (1993), 109116.
18.
P.Yih, and D D L . Chung, J. o f Mat. Sci., 32, (1997), 1703-1709.
19.
S. Deb, B.K. Sarkar, and T.K. Dan, J. of Mat. Sci. Lett., 13, (1994), 597-599.
20.
S.K. Roy, and A. Biswas, J. o f Mat. Sci. Lett., 13, (1994), 371-373.
21.
M.K. Ferber, P. F. Becher, and C. B Finch, J.o f Am. Ceram. Soc. Comm. C-2-C-4,
(1983).
22.
Z. Peng, Ph.D. Thesis, (1998), University o f Idaho.
23.
W.A. Gooch, B.H.C. Chen, M.S. Burkins, R. Palicka, J. Rubin, and R.
Ravichandran Mat. Sci. Forum, Trans. Tech. Publications, 308-311, (1999), 277282.
24.
E.S.C. Chin, Mat. Sci. and Eng., A259, (1999), 155-161.
25.
K.E. Spear, P. McDowell, and F. McMahon, J. of Am. Ceram. Soc., (1986),C-4C-5.
26.
D.C. Halverson, Z.A. Munir, Ceram. Sci. Proc., 1001-1010.
27.
V.A. Neronov, M.A. Korchagin, V.V. Aleksandrov, and S.N. Gusenko, J. o f the
Less Common Metals, 82, 125-129, (1981).
28.
L. Brewer, D.L. Sawyer, D.H. Templeton, and C.H. Dauben, J. o f the Amer. Ceram.
Soc., 34,6, (1951), 173-179.
29.
E.J. Huber Jr., J. Chem. Eng. Data, 11,3, (1966), 430-431.
30.
B. Post, F.W. Glaser, and D. Moskowitz, Acta Metall., 2, (1953), 20-25.
31.
H.R. Ogden and R.I. Jaffee, Trans. AIME, (1951), 335.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
89
32.
A.E. Palty, H. Margolin, and J.P. Nielsen, Trans, of ASM, 46, (1954), 312-328.
33.
NIST-JANAF Thermochemical Tables, 4th edition, Am. Chem. Soc. (1998).
34.
A.G. Merzhanov, A.M. Stolin and V.V. Podlesov, J. of European Ceram. Soc., 17,
(1997), 447-451.
35.
J.B. Holt, D.D. Kingman and G.M. Bianchini, Mat. Sci. and Eng., 71, (1985), 321327.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
90
MANUSCRIPT THREE
FUNCTIONALLY GRADED Ti-TiB-TiB2 COMPOSITES PRODUCED BY
COMBUSTION SYNTHESIS -H : THE EFFECT OF COMPACTION ON
PROPERTIES
Menderes Cirakoglu
Department of Materials and Metallurgical Engineering
University o f Idaho
Moscow-Idaho 83844-3024
will be submitted for publication
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
91
ABSTRACT
In this study, we utilized CS method along with compaction, as a processing method for
Ti-B based FGMs. Application o f a mechanical load to the sample during the CS process
is usually necessary in order to produce dense FGMs. After combustion synthesis
reactions, higher boron content layers exhibited higher porosity due to the high
exothermicity o f these compositions. The pore sizes were smaller in FGM samples
compared to single compositions. The application o f pressure with combustion synthesis
helped to eliminate the pores to some extent and resulted in a denser product. A
noticeable feature o f the microstructure in the cross-section is that the particles were
irregular in shape and elongated perpendicular to the direction o f the application o f
pressure during processing. The SEM studies revealed a well-bonded composite
structure. As the boron content and the applied pressure increased, the hardness values
also increased indicating a denser product. A nanoindentation test was utilized for testing
two interlayers with somewhat smooth surfaces. The results showed that porosity within
the interface region has an effect on the mechanical properties causing microstructural
irregularities and scatter o f the data. The increase of hardness and elastic modulus was
attributed to the presence o f higher amount o f ceramic phases TiB and TiB2 .
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
92
1. Introduction
A commonly recognized characteristic o f combustion synthesis is the porous nature of the
products. This chapter highlights the significant influence o f porosity on the final properties
o f functionally graded materials.
This section contains a brief overview on the porosity aspects o f the combustion synthesis
process. First, the origin of the porosity formation during a CS process is explained.
Subsequently, a literature survey on the methods utilized to reduce the porosity is presented.
In the last part o f this section, the motivation and objectives o f this w ork are outlined.
1.1. The Origin o f Porosity in Combustion Synthesized Materials
Several factors contribute to the formation o f porosity during a combustion synthesis process.
These can be outlined as follows:
(i) Porosity in the green compact: The major source o f porosity in combustion-synthesized
samples is the initial value o f the porosity in the sample. The reactant mixtures are usually
pressed in the form o f powder compacts before the combustion reactions. These compacts are
usually 50-70% dense. The initial porosity has significant effect on both the combustion
wave velocity and the maximum reaction temperature. Smaller pores were found to give rise
to a higher contact area between the reactants. When the pores sizes are small, decreased
porosity enhances local conduction process and hence it gives rise to a higher rate of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
93
enthalpy increase and a higher combustion velocity. On the other hand highly porous
reactants may not support a self-sustaining reaction because of inadequate heat transfer ahead
o f the combustion wave [2-6],
(ii) Porosity from the volatilization of impurities: Because of the high temperatures
associated with the combustion synthesis process, contaminants (O, N etc), which may exist
on the starting powders, volatilize, leaving behind escape paths in the synthesized product
[4](iii) Porosity from molar volume change: The products of exothermic reactions are
generally expected to have shorter atomic bonds and hence to have higher densities
compared to the reactants. Therefore, a net decrease in the molar volume occurs during the
process. This decrease in the actual volume o f the reacted compact is accompanied by the
generation o f porosity as high as 20-25 % [4-6],
(iv) Porosity from the thermal migration. The other source of porosity is related to the
thermal migration due to the existence o f steep temperature gradients ahead o f the
combustion wave [4],
1.2. Densiftcation Methods of Combustion Synthesized Materials
The presence of pores in structural materials can affect the material's mechanical properties
from both safety and performance perspectives. The porosity can be reduced if the
combustion synthesis process is combined with the application o f pressure. Pressure is
applied during or shortly after the combustion reaction. The idea behind this is to rapidly
consolidate the reacted CS product while it is still at a relatively high temperature. The brief
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
94
existence o f a liquid phase during the combustion synthesis reactions provides a significant
opportunity for densification as well as simultaneous synthesis. This type of forced
consolidation techniques occur in a small time scale compared to hot pressing and hot
isostatic pressing [3,7-12],
A number o f techniques are being studied or employed to densify the CS produced products.
Zhang et al [3] carried out a series o f theoretical work on the densification o f combustion
synthesized products. Kecskes et al [10] utilized explosive compaction together with
combustion synthesis to produce dense HfC and binary HfC-TiC composites. In their study,
the explosives were detonated remotely, right after the CS experiment completed. Dense
products o f 74-87 % o f the theoretical densities were obtained. In another study, Kecskes et
al [9], fabricated full density TiB 2 and TiC by CS followed by dynamic compaction by using
explosives. The products with 98% of theoretical density with hardness values equal to or
greater than commercially available hot pressed materials were produced. A ceramic-toceramic joining method was introduced by Rabin et al [11]. They studied producing TiCAJ2 O 3 composites by combustion synthesis followed by consolidation. They achieved to
produce full density composites using a shock pressure of 1 GPa. They found that the
samples have a well-bonded composite structure. Song et al [8 ] demonstrated the fabrication
ofT iB 2/TiC composites using a method o f combustion synthesis followed by dynamic
compaction. They utilized titanium and boron carbide powders as precursors and after
compaction composites with more than 98 % o f the theoretical densities were achieved.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
95
1.3. Densification o f Combustion Synthesized FGMs
Research into the use o f combustion synthesis to produce FGMs resulted in development of
innovative techniques together with the traditional reactive process. Compared to
conventional production methods e.g. hot pressing, hot isostatic pressing, reactive processes
utilizes the heat generated by the reactions therefore it is economically competitive.
Furthermore, the other drawback o f pressing methods is that they degrade layers in the
structure [13].
The distribution o f the compositions in these compacts can be symmetric or asymmetric.
Feng et al [14] studied utilized combustion synthesis process along with infiltration. They
added aluminum metal into TiC>2 + B 2 O 3 precursor composition. The graded composites were
prepared by changing the amount o f aluminum in each individual layer. They produced a
porous TLB2 -AI2 O 3 matrix with liquid Al infiltrates the porous structure. During the
combustion, a light compaction pressure (as low as
2 0 0
psi) was also applied simultaneously.
In a similar way Pityulin et al [15] produced symmetric gradient materials o f TiC-Ni-TiC.
In another study, Levashov et al [16] utilized combustion synthesis in order to produce
diamond-containing FGMs. Pellets were produced with a layer to layer changing diamond
concentration (changing from 0 to 12 vol %) in Ni + Al, Ti + B and Ti + C + Co systems.
After the completion o f the reaction, hot product was compacted in a hydraulic press with
pressures not more than 400 MPa for 10 seconds. Otherwise, with conventional methods,
high temperatures (2500°C) and pressures (up to 8000 MPa) are necessary. It was reported
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
96
that FGMs produced by conventional methods have greater abrasive resistance compared to
those produced by conventional methods and more importantly diamond consumption was
reduced to 2-4 times.
Ohyanagi et al [17] investigated processing o f three layered (diamond, diamond/(TiB 2 /Si)
and TiB2 /Si) FGMs. Titanium, boron and silicon elemental powders were used. The sample
was embedded in sand and a pseudo isostatic pressure was applied to sand (a total pressure o f
255 MPa was applied) by a piston after the combustion reaction completed.
Li et al [18] studied processing o f Al2 0 3 /TiC/Ni/TiC/Al2 C>3 and Kang et al [19] studied
Al2 0 3 /Cr 3 C2 /Ni/Cr 3 C 2 /Al2 0 3 symmetric gradient structures by combustion synthesis
accompanied with hot isostatic pressing. In both o f these methods, the strong exothermic
reaction between Si and nitrogen was used to initiate the reactions. The energy released from
this reaction instantaneously melted the green compact and by using hot isostatic press the
compact was consolidated.
Lai et al [20] utilized centrifugal force along with combustion synthesis in order to produce
AI2 O3 -CU FGMs. The product had three regions layers namely a copper region, an alumina
region and a transition region with a graded composition. Copper was added into the initial
mixture o f Al and CuO to control the temperature and combustion reaction. The green
compact was placed inside a specially designed furnace with a centrifuge system. After the
centrifugal force has stabilized at a set value, the ignition was initiated by applying power to
the furnace through tungsten wires. They have concluded that at high rotations, complete
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
97
phase separation occurred and the product was in the form o f a bi layer material on the other
hand at low speed rotations, no phase separation was observed and the product was a macro
uniform material.
1.4. Motivation and Objectives
The main objective o f this study is to apply the FGM architecture to ceramic/metal armor
materials utilizing combustion synthesis/dynamic compaction as a processing method. The
chosen Ti-B system has potential application in defense related applications such as armor,
armaments, and vehicle structures. Earlier studies demonstrated promising ballistic attributes
for MMCs. Improved penetration resistance was observed with high strength ceramic
particulate reinforced MMCs [21]. The concept o f functionally graded armor composites
(FGAC) is intended to further enhance the ballistic space and mass efficiency o f MMCs by
tailoring the through-thickness incorporation and distribution o f various reinforcement
morphology, size and chemistry to mitigate shock damage. The present armor scheme to
defeat light to medium threats typically consists o f a hard frontal surface and softer backing.
The hard frontal materials are usually ceramics. The purpose o f the hard surface is to blunt
and to induce a destructive shock wave on to the projectile upon impact. The softer backing
materials act as a “catcher” for residual broken fragments in preventing target penetration. In
this type of armor scheme, the hardest frontal material will typically provide the best level of
ballistic protection [2 1 - 2 2 ].
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
98
Our previous studies showed that combustion synthesis processing approach is feasible in
fabricating Ti/TiB/TiB2 FGMs, however, the challenge lies in the production o f denser
composites. Several processing parameters and aspects yet to be investigated are
summarized below.
(i) To devise a simple, inexpensive and repeatable way to fabricate dense FGM specimens.
(ii) To evaluate the effect o f porosity and density on the combustion synthesis process and
products; and to correlate these results with the combustion characteristics reported in the
previous chapter.
(iii) To study the microstructure and mechanical properties o f individual layers and
interlayers.
(iv) After establishing the necessary parameters, it is important to investigate the possibility
o f applying the same approach on larger scale products.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
99
2. Experimental Procedure
This section presents the experimental procedure and characterization methods. The material
system, sample preparation and basic experimental procedures are similar to the previous
section. Compaction step is additionally added to the conventional combustion synthesis.
2.1. Combustion Synthesis/Compaction Experiments
The process consists o f several steps, which are shown, schematically in Figure 1. Titanium
(99% purity, -325 mesh, Johnson Mathey, Ward Hill, MI) and crystalline boron (> 99%
purity, average 15 pm or less, Cerac Inc., Milwaukee, WI) powders were mixed in 82/18,
85/15, 90/10 and 95/5 wt % ratios. The graded layers were prepared by simply stacking up
the different compositions in a steel die in a discrete homogeneous layers fashion. The
consolidation o f the powder mixtures was accomplished by first uniaxially pressing into
cylindrical compacts under an axial load o f 15,000 lbs. The compacts are further densified
using a cold isostatic press (ISI Inc., Columbus OH) at a pressure o f 100 MPa. The resulting
compacts have typical dimensions of 12.7 mm and about 14 mm long. Prior to combustion
experiments, the as-prepared samples are de-gassed at 600°C for 3 hours under flowing argon
to remove the adsorbed moisture and gases. A vacuum furnace (Centorr Assoc. Inc.,
Suncook, NH) was used for degassing.
Graded compacts were ignited from Ti side in a reactor designed and built in our laboratory.
A tungsten coil (Model H2.040W, R.D. Mathis Co., Long Beach, CA) placed at 2-3 mm
from the top o f the sample is used as an external heat source for ignition. The compaction
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
100
was performed after the green compact FGM is converted into the final product. This is
practically when the combustion wave reaches to the bottom o f the sample, which was
verified by visual observation. A pressurized air activated steel punch is used for this
purpose, which is located on the upper section o f the reactor assembly.
The potential energy stored in pressurized air is transformed into the kinetic energy o f the
punch and collapses the porous microstructure. Since the diameter of the punch is slightly
smailer than that o f the die, the wall o f the die is protected. The tungsten coil moves out of
the way right before the compaction and coil was saved. The stroke distance was about 4 cm.
The applied pressure was adjusted as desired by simply increasing the air pressure from a
compressed air tank. This technique, based on our present investigations, is promising for
producing dense compacts.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
101
Ti
I
(
C
)
^
Dry Mixing
f
'
Uniaxial Pressing
B
1
)
)
f
(
)
)
Cold Isostatic Pressing
D egassing
Z
,
\
^Conventional Combustion Synthesis)
Q
j
^Combustion Synthesis/Compaction)
I
~
Characterization
^
Figure 1. Flow chart o f experimental procedure.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.2. Characterization
The samples were analyzed for density, porosity, hardness and microstructure. Density was
measured before and after combustion experiments by using Archimedes technique. Ethyl
alcohol (d=0.7893 g/cm3) was used as the immersion medium. The specimen was placed in
ethyl alcohol and vacuumed for at least 30 minutes to ensure that all the pores were filled
with alcohol. The volume occupied by the solid plus the volume of the voids when divided
into the powder mass yields the bulk density. There
Bulk Density =
^
^
Weight) x {Density o f immersion medium)
(,Saturated Weight)—{Suspended Weight)
1
(Dry Weight) x (Density o f immersion medium)
(Dry Weight) —(Suspended Weight)
In order to evaluate the type o f pore and pore volume distribution, porosimetry measurements
were carried out by using Micromeritics AutoPore III mercury porosimetry. The operation is
performed in two stages namely, low pressure (from
0 . 1
to 3100 psia) and high pressure
(3100 psia to 30,000 psia). The test samples were in the range of 0.75 to 1.5 grams. The
samples were oven-dried before testing to remove the adsorbed vapors. The first step in
measurement was evacuation o f all gases from the cell, which holds the sample. Therefore,
the system was evacuated to 50
Hg for 5 minutes. Then, the glass penetrometer was filled
with mercury. Mercury flowed into the narrow end o f the penetrometer under vacuum by
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
103
capillary forces. Then the sample cell was removed from the mercury filling mechanism and
inserted into the high-pressure cavity for high-pressure analysis. This requires that the cell be
moved from the horizontal position to the vertical position, which creates an additional
pressure head of 5-6 psia above ambient on the sample. Pressure (up to 3100 psia) was
applied to force mercury into the interparticle voids and intraparticle pores. The mercury
porosimetry was capable o f producing continuous plots o f both the intrusion and extrusion
curves by monitoring both the applied pressure and the intruded mercury volume. The
volume forced into the pores was usually monitored in a penetrometer, which was calibrated
precision stem of a glass cell, containing the sample and filled with mercury. Extrusion
from the voids and large pores can be continuously monitored by evacuating the housing.
This completed a full low-pressure intrusion-extrusion cycle.
This law, in the case o f a non-wetting liquid like mercury and cylindrical pores, is indicated
by Washburn equation [23],
D=-
Ay cos (p
(3)
where
D: pore diameter
P: the applied pressure
y : Surface tension (485 dynes/cm)
cp : The contact angle (130°)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
104
Pore size distributions are therefore generated by monitoring the volume o f mercury intruded
into pores as a function o f increasing applied pressure. The choice o f cylindrical pore
geometry was one o f the mathematical conveniences in order to avoid the complexities o f
having to deal with mean radii and contact angles in pores of irregular cross sections.
Mercury porosimetry also provided a convenient method for measuring the density o f
powders. The volume o f the sample was calculated by weighing the cell without mercury and
then the cell containing the sample with mercury. After converting the weights o f mercury to
the corresponding volumes, using the density table, the volume o f the sample was determined
as the difference of the two mercury volumes. Using the volume o f mercury intruded at
various pressures, the apparent density was obtained. The calculated apparent densities were
obtained by subtracting the intruded volume from the initial sample volume and dividing the
resulting value into the sample weight.
CS reacted samples were polished sequentially using SiC abrasive paper from grit sizes 180
to 4000 then polished with alumina suspension from 5 p.m to 0.01 p.m. The polished samples
were then lightly etched using a solution o f 10 % HF, 30 % HNO3 and 60 % H 2 O. The
microstructural characterizations o f the combusted samples were made by optical microscopy
(Leco, Olympus PMG3, St. Joseph, MI) and by using scanning electron microscopy (SEM,
AMRAY 1830, Amray Inc., Bedford MA).
Hysitron Triboscope nanoindentor in conjunction with Park Scientific Instrument’s
AutoProbe™ CP Scanning Auto Probe Microscope (SPM) was used for nanoindentation
tests. All indentations were made at room temperature and a three-sided 90°(cube comer)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
105
diamond tip was used as an indenter, which was also used as the imaging probe. The indenter
tip was calibrated prior to tests by using fused silica due to its known isotropic elastic
properties. The area function is deduced from these calibration tests. It is generally described
as a function of the plastic depth (he), which is the depth o f the indenter in contact with the
sample under the tip. Later, it is used in calculation of area under the tip (A). A piezoelectric
driver was used to move the tip into the sample and the load and displacements were
monitored in situ by capacitance probes. The samples were first scanned with the indentation
tip to determine the appropriate region for analysis. After the indentation was performed and
the load -displacement data were recorded, the surface is scanned a second time to capture a
3 dimensional image o f the indentation. This image is a height map o f the surface topography
where the contrast is such that lighter areas correspond to features at more elevated heights.
KaleidaGraph™ data analysis/graphing application software (Synergy Software, Reading,
PA) was used carry out the calculations and to plot the data.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
106
3. Results and Discussion
In this section, we present the results o f compacted combustion synthesized products in
comparison to the conventional combustion synthesis method. The improvements made by
compaction are addressed. Processing conditions are shown to be highly related to the
material properties like porosity, density and hardness.
3.1. Density and Porosity Measurements
In order to investigate the effect o f density and porosity on the combustion process, density
and porosity measurements of single compositions and FGMs were carried out before and
after the CS experiments.
3.1.1. Combustion Synthesis of Single Compositions
Figure 2 shows the density change with boron content in single compositions before and after
CS experiments. It indicates that the bulk density o f the compacts increases as the boron
content decreases. This is an expected result considering the fact that the density o f boron
(2.34 g/cm3) is much lower than that o f Ti (4.51 g/cm3). We calculated theoretical densities
based on the initial components by using rule of mixtures for each composition. The results
indicated that before CS, the green compacts had densities in the range of 68-73 % TD
depending upon the relative amount o f boron. The XRD results indicated, TiB, TiB 2 are the
reaction products along with unreacted Ti. The densities o f TiB and TiB2 are 5.26 g/cm 3 and
4.52 g/cm3, respectively. Therefore, it can be suggested that as the boron content increases,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
107
the density o f the CS product also increases mainly due to the formation o f boride phases.
Experimentally however, it was found that the higher boron content layers have lesser
density. This can be explained by the presence of porosity observed in these layers, which is
a common occurrence in the products o f CS. When the boron content is increased closer to
TLB2 stoichiometric composition the volumetric heat release also increases. As explained in
the introduction section, expelling volatile impurities as well as molar volume changes result
in porosity formation. Although, in our experiments we carried out de-gassing prior to
combustion experiments, the degassing temperature cannot be high enough to cause
extensive volatilization o f the impurities. However, unlike other sources of porosity (from
green compact, molar volume changes etc.) it is difficult to quantify the effect o f this source.
Rice et al [6 ] calculated the volume change for TiB2 formation about -23.3 %. This value
was calculated for complete conversion. In our experiments however, due to excess titanium,
volume change is expected to be lower than this value.
^cn
3 -7
ao 3.5
35 3.3
C/3
1
|
2.9
3
2.7
PQ
a fter CS
2.5
82Ti-18B
85Ti-15B
90Ti-10B
95Ti-5B
Single Compositions
Figure 2. Bulk Density change before and after combustion experiments
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
108
In order to examine the sequence o f the porosity change during the process, it is important to
mention and characterize the porosity in the green compacts. Therefore, in order to quantify
the amount o f porosity and pore size distribution, mercury porosimetry tests were carried out
before and after the combustion experiments. In Table I, the porosimetry results o f the
compacts before combustion synthesis are given.
Table I. Mercury porosimetry results o f the single composition green compacts
Porosity
Bulk Density
Avg. Pore Dia.
(%)
(g/cm3)/ (%TD)
(fim)
82-18
16.52
2.7621/ (67%)
16.41
85-15
24.70
2.9207/ (69%)
22.08
95-5
26
3.255 /(74%)
2 2 .2 2
Composition
As can be seen from the above table, the green densities o f the compacts after degassing
varied from 67% to 74% TD (calculated by using rule o f mixtures based on their initial
compositions) indicating a relatively porous microstructure. The porosity in green compacts
decreases as the boron content increases. The reason for this observation is that: the boron
particle size is smaller than titanium. This leads to a reduction in the green porosity. The
smaller particles of boron tend to fill the interstices in the green compact. As it was
mentioned previously, this is expected to provide a better particle-to-particle contact, which
can enhance the combustibility o f the green compact [24]. However, in our experiments, the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
109
influence of this effect on the overall profile was small because the exothermicity gained by
the increase o f boron outweighs this criterion.
Table I also indicates the average pore diameter of the green compacts is in the range o f 1522 pm. It is well known that green compacts were left from the shaping procedures with
large amounts o f in situ interconnected pores. Another important observation is that the
change in porosity with composition is not large. This is mainly due to the smaller particle
sizes o f both o f the constituents. This therefore prevented us directly correlate the results to
the macroscopic features (combustibility, combustion wave velocity etc.) of the combustion
process. It is evident that as the boron content increases in the green compact, the initial
porosity decreases from 26 % to 16%. However, this change affects the thermal conductivity
and the combustion characteristics o f the compacts.
Figure 3 and 4 show typical mercury intrusion plots for single composition o f 82Ti-18B and
95Ti-5B, respectively. The spectra were plotted as incremental pore volume (mL/g) versus
pore diameter (A) and illustrate a bimodal pore size distribution. In these figures, the slope
between points A and B indicates mercury filling the space in the range of around
1 0 6
A
(1000 pm) mainly the space between the contacting particles o f Ti and B. As the pressure
increases, mercury penetrates into the smaller cavities between particles. Between points C
and D, intrusion commences into the pores with radii smaller than ~
1 0 4
A ( 1 0 pm). It can be
seen that the slope between points A and B is not as discreet with 95Ti-5B compositions as it
is with 82Ti-18B. The probable reason for this is that the green compact of 95Ti-5B is
mainly composed o f Ti (mono sized).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
110
Pore volume distributions o f samples 82-18 and 95-5 are given in Figures 5 and 6 ,
respectively. The peaks in these figures are related to the slopes in Figures 3 and 4 and
indicate the most frequently detected pore sizes. Figure 5 shows 3 peaks; a main peak in the
order o f 1 0 4A ( 1 0 pm) and two sub peaks in orders o f
1 0
6A
(1 0 0 0
pm),
1 0 3
A ( 1 pm). Since
the area under the volume distribution curve is proportional to the total pore volume, it can
be suggested that the pores are mostly in the range o f
1 0 4
A ( 1 0 pm). Similarly, Figure
6
reflects three different pore size distributions. However, the plot indicates that the green
compact consist o f narrowly distributed pore sizes compared to 82Ti-18B. Mainly due to the
fact that green compact of 95Ti-5B is mainly composed o f Ti (mono sized) as intrusion plots
suggested.
Mercury intmsion tests were also performed on the compacts after combustion synthesis. The
results are tabulated in Table
n.
Table EL Mercury porosimetry results o f the ignited single compositions.
Porosity
Bulk Density
Avg. Pore Dia.
(%)
(g/cm3)
(M-m)
82Ti-l8B
31.9
2.8148
48.3
85Ti-15B
28.74
2.9200
22.38
95Ti-5B
22.49
3.2902
10.06
Composition
When compared to the results obtained from green compacts (Table I), it is clear that after
combustion the average pore sizes and porosity decreased for composition 95Ti-5B and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Ill
increased for compositions 82Ti-18B and 85Ti-15B. For composition 95Ti-5B, the timetemperature profiles indicated a maximum temperature of 1400°C. Although this temperature
is lower than the melting temperature o f Ti (1660°C), we believe that the temperature at the
combustion front locally exceeded the melting temperature o f titanium. We suggest this,
based on the fact that after the combustion experiments we observed thermocouples stuck to
the sample indicating a liquid phase formation. Therefore, the decrease in porosity after
combustion reaction can be explained by melted titanium capillary spreading and filling the
pores during the process. It is also important to note here that after combustion synthesis,
higher boron content had the highest porosity. This could be ascribed to the high
exothermicity o f these compositions. In the highly exothermic system, thermal migration due
to the existence of a steep temperature gradient at the vicinity o f the combustion wave causes
more porosity in the product. Figures 7 and
8
show the mercury intrusion spectra obtained for
compositions 82-18 and 95-5, respectively. Corresponding pore size distributions were given
in Figures 9 and 10. The main peak and sub peak positions maintained their positions.
These observations showed that, the final density o f the product was directly related to the
relative amount of liquid phase during the combustion reaction and hence exothermicity o f
the reaction. During combustion synthesis, due to the exothermic nature o f the reactions,
liquid phase forms within the combustion front. Depending on the quantity of the liquid,
mass and heat re-distribution occurs. The presence of a liquid phase can lead to a decrease in
porosity through rearrangement or sintering. Thus, we anticipate that compacts in which the
reactions are primarily in the liquid phase will have lower porosities at the end of the
reactions. Results obtained in this study show agreement with this anticipation.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
112
82Ti-18B
0.07
o>0.06
0.05
£ 0.0 4
1 0 03
1 0.02
o 0.01
1.00E+07
1.00E +06
1.00E +05
1.00E +04
1.00E+03
1.00E+02
1.00E+01
1.00E+00
Cumulative D iam eter (A)
Figure 3. Mercury intrusion plot for composition 82Ti-18B after degassing
95Ti-5B
0.08
o> 0-07
E 0.06
a> 0.05
-(O 0 .0 4
g 0.03
I 0.02
oc
0.01
1.00E+07
1.00E +06
1.00E +05
1.00E +04
1.00E+03
1.00E+02
1.00E+01
1.00E +00
Cumulative D iam eter (A)
Figure 4. Mercury intrusion plot for composition 95Ti-5B after degassing
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
113
82Ti-18B
0.0 3
cn
■g 0 .0 2 5
j|
3
0.02
;> 0 .0 1 5
a»
I
0.01
|
a
0.00 5
3
1 .00E+07
1 .00E+06
1.00E+05
1 .OOE+04
1.00E +03
1.0OE+02
Cumulative D iam eter (A)
Figure 5. Cumulative pore size distributions for composition 82Ti-18B after degassing
95Ti-5B
O)
E
0.0 3 T
0.0 2 5
|
0.02
>
0.01 5
3
a>) __
0.01
as
3£
5 0.00 5
a
1.00E+07
1 .00E+06
1 .00E+05
1 .00E +04
1.00E+03
1.00E+O2
Cumulative D iam eter (A)
Figure 6 . Cumulative pore size distributions for composition 95Ti-5B after degassing
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
114
82TM 8B
0.1 2
-
0.08
—
(O 0.06
E 0 .0 4 ;
o
2
=
0.02
1.00E+07
1.00E+06
1 .OOE+05
1.00E+04
1 .OOE+03
1.0OE+02
Cumulative Diam eter (A)
Figure 7. Mercury intrusion plot for composition 82Ti-18B after CS
9 5 -5
0.12
™
0.1
,E
<u 0 .0 8
I(O 0 .0 6
I
<u
co
0 .0 4
0.02
1.0OE+O5
1.00E+04
1.00E +02
Cumulative Diameter (A)
Figure 8 . Mercury intrusion plot for composition 95Ti-5B after CS
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
115
82T1-18B
0 .0 3 5 T
■S> 0.03
E
a> 0.02 5
E
_i
=
0.02
o> 0.01 5
§
0.01
E
O 0.00 5
1.00E+07
1 .00E +06
1.00E+05
1 .00E + 04
1.00E+03
1.00E+02
Cumulative D iam eter (A)
Figure 9. Cumulative pore size distributions for composition 82Ti-18B after CS
9 5T i-5B
0.04 5 j
-S’ 0 .0 4 E. 0.03 5
|
0.03
-§ 0.02 5
«
I
0.02
0.01 5
I
0.01
o
0.00 5
1.00E+07
1 .00E +06
1 .OOE+05
1.00E + 04
1 .OOE+03
1 .OOE+02
Cumulative D iam eter (A)
Figure 10. Cumulative pore size distributions for composition 95Ti-5B after CS
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
116
3.1.2. Combustion Synthesis/Compaction o f Functionally Graded Materials
As mentioned before, a uniformly dense product can be produced when the mechanical load
is applied at the same time as the ignition and maintained throughout the combustion process.
In our experimental system, we applied the compaction load on FGM samples to collapse the
pores. After the compaction the height o f the sample decreased about 10 % for a load o f 220
kPa and the diameter increased about 1.5 %.
Figure 11 shows the density change o f FGM layers with applied pressure. Similar to
conventional combustion synthesized samples, as the boron content increased, the density o f
the layers decreased mainly due to the presence o f porosity. Although, the effect o f
compaction increased the density in all layers, the change is more prominent in titanium
layer. The scenario envisioned is related to the timing of the pressure. This issue is addressed
in detail in the following paragraphs along with the porosity results.
Each individual layer was first sliced through the interface and porosity measurements were
carried out on these layers by using mercury porosimetry. The results are shown in Table
in
and IV. When compared to the porosity data o f the single compositions after CS (reported in
Table
n), the pore sizes are considerably smaller in FGMs.
This is attributed to the pre­
heating effect during the combustion reactions o f FGMs. Higher local combustion
temperatures were achieved as it was reported in the previous manuscript. Consequently,
more liquid titanium forms. The final porosity o f the product is directly related to the relative
amount of the liquid phase, which facilitates densification.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
117
4.25
cn
*▼
I
«> 3.75
c
<u
2
35
90H-10B
3.25
3
M
2.75
Ti
95Ti-5B
90Ti-10B
85Ti-15B
FGM layers
Figure 11. Bulk density change in different layers o f FGM after CS with and without
pressure application.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
118
Table HI. Mercury porosimetry results o f two layers o f FGMs produced by conventional
combustion synthesis.
Composition
Porosity
Bulk Density
Avg. Pore Dia.
(%)
(g/cm3)
(pm)
82-18
28.75
2.6302
14.24
95-5
16.88
3.2548
17.20
Table IV. Mercury porosimetry results o f two layers o f FGMs after the application o f a
pressure of 220 kPa.
Composition
Porosity
Bulk Density
Avg. Pore Dia.
(%)
(g/cm3)
(pm)
82-18
7
2.8891
9.69
95-5
1.72
3.0593
8.59
However, the application o f an external mechanical load is quite important as a poreelimination method besides local pore filling phenomena due to the redistribution o f the
liquid titanium. At any instant during the combustion reaction, the liquid phase only exists
within a fraction of a second and confined only to a very narrow region in space.
Subsequently it disappears by both solidification and dissolution o f boron and forming a
solid boride product phase. The time temperature profiles reported in the previous manuscript
indicated a sharp decrease in temperature after combustion took place. The temperature
dropped from the maximum temperature of 2300°C to 1500°C in about 5 seconds. This
suggests that, the application o f the mechanical load on time is as important as the magnitude
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
119
o f the load in order to produce uniformly dense products. Once the punch is activated there
was a brief delay in the order o f 3-5 seconds. The difficulty with our approach is this
inherent delay in the application o f the mechanical load very likely leading to incomplete
densification.
3 .2 . Microstructure
In Figure 12 a, optical micrograph o f the first interface of the combustion-synthesized
samples is shown. In Figure 12b, the same interface o f the CS/DC produced sample is shown.
Comparison o f Figure 12a with 12 b provides an interesting feature o f the two types of
experiment i.e. conventional combustion synthesis and combustion synthesis followed by
dynamic compaction. A noticeable feature o f the microstructure in the cross-section is that
the particles are irregular in shape and elongated perpendicular to the direction of the
application o f pressure during processing.
(a)
(b)
Figure 12.
Optical micrographs o f the interfaces of the FGMs produced by
(a) Conventional combustion synthesis
(b) Combustion synthesis/ compaction
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
120
A low magnification scanning electron microscopy (SEM) of the CS/DC sample is shown in
Figure 13. As the boron content increases, the porosity level increases. In the micrograph the
light areas are void free regions and the dark spots are open pores.
lplpli§|i
90Ti*l0B
Figure 13. SEM micrograph o f a cross section o f a five-layered FGM
In Figures 14 (a) and (b), the interfaces o f the samples produced by conventional combustion
synthesis are shown at a higher magnification. The white areas are titanium metal, the gray
areas are ceramic phase and the black spots are voids. As can be seen from the figures, the
cross section consists o f somewhat dense layers with a network o f micro pores and dense
areas. The dense areas are indicative of liquid phase formation.
In Figure 15, the microstructures of 3 different layers are shown. The microstructure is
showing irregular shaped pores in the range o f 16-27 pm. The apparent porosity in these
figures is due to the particle pullout during metallographic preparation. Such phenomena
were also observed in all layers.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
121
mmmmm
(a)
(b)
Figure 14. SEM o f polished sample produced by conventional combustion synthesis
(a) interface o f the first and second layers
(b) interface o f second and third layers
J/
fVA.'■
#v
VV
■W,
s ✓'
&it>
' .'iS.\
'tt?.
>'•
"■'-A
S*f
'^
f
'
'
^
^ -r > '
' ' -> ? f
(a)
(b)
"
"J
" ''''A
*
(c)
Figure 15. SEM micrographs o f CS samples.
(a) Ti+5wt%B layer
(b) 10wt%B
(c) 15wt%B
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
,
122
In Figure 16, the SEM images of the samples produced by combustion synthesis followed by
compaction (220 kPa) are shown. As is shown the pore sizes are in the range o f 7-18 pm.
These results are in good agreement with the porosimetry data reported earlier. This might be
indicating the compaction energy was not sufficient enough to collapse these voids.
(a)
(b)
Figure 16. SEM micrographs CS/compaction samples.
(a) Ti - Ti+5wt%B interface
(b) Ti +5wt%B - Ti+10wt%B interface
The examination o f the cross section o f the layers showed randomly distributed pore
distribution. LaSalvia et al [7] reported tearing type o f cracks starting in the lateral surface
and propagates towards the center in their combustion synthesized/compacted TiC samples.
This type o f extended cracks was not observed in our samples under low compaction
pressures (up to 220 kPa). When the applied compaction pressure was increased (to = 317
kPa) in order to reduce the residual porosity, lateral macro cracks were observed as shown in
Figure 17. These cracks were found to be within a few millimeters in range and extend 1-2
mm towards the thickness o f the sample. Because o f the axisymmetry o f the combustion
reaction and compaction, the heat losses from the system is largest in the radial direction i.e.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
123
the outer surface is cooler than the interior. As explained previously, to avoid heat losses we
used alumina blankets as thermal insulation. However, due to the geometry o f the die cavity
the thickness o f the lateral insulation used is much higher than that o f radial insulation
(insulation at the bottom o f the sample). Therefore, there is a difference between the heat
transfer in axial and radial directions. Due to this difference, fast cooling outer surface will
be stronger than slowly cooling interior. During compaction, the hardened surface becomes
fragmented.
Figure 17. Macrograph of lateral macro cracks on the outside surface o f an FGM
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
124
3.3. Mechanical Properties
3.3.1. Vickers Hardness Tests
In a typical FGM, the volume fractions of the constituents are varied in layers gradually. In
that sense, FGMs can be visualized as a sandwich o f thin layers o f varying material
properties. Each layer corresponds to a composite material. Therefore, the mechanical
properties are expected to change from one layer to another [13,25], In order to compare the
effect o f compaction on the mechanical properties, hardness tests were carried out on
samples produced by both CS and CS/DC. Figures 18 shows the hardness data o f the sample
produced by conventional combustion synthesis. As the boron content increased, the
hardness value is also slightly increased. A maximum hardness o f 126 HVN was measured
on the 5th layer.
Figure 18. Microhardness values o f combustion synthesized FGM layers.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
125
It was noticed that microhardness values were relatively low. These low values may be
attributed to the presence o f porosity. As was mentioned in the previous sections, due the
nature o f the CS, the product is usually porous. Therefore, a pressure application is necessary
to form more dense products. Figure 19 shows the hardness profile o f the FGM layers
produced by CS/DC under a compaction pressure o f 137 kPa. Similar to CS produced
samples, the hardness increases with increasing boron content due to the formation o f hard
boride phases. It was evident that microhardness o f the FGM layers is increased by the
application o f pressure after the CS indicating a more dense microstructure. Figure 20 shows
the hardness data after the application o f a pressure o f 220 kPa. Increasing the pressure
resulted in higher hardness values and a denser product. Although, there is considerable
scatter in the hardness data, significant improvements in hardness are clear. The fluctuation
in the error bars, especially in the higher boron content layers, is directly related to the ability
o f placing the indenter in a void-free region o f the sample without collapse o f adjacent
grains. The improvement in density and hardness even with compaction loads of 220 kPa is a
promising result for the potential production o f tough, dense, near net shaped FGMs.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
126
200
150
100
>
50
0
Figure 19. Microhardness values o f FGM layers after the application o f pressure o f 137 kPa.
1500
1000
z
>
X
500
0
Figure 20. Microhardness values o f FGM layers after the application o f pressure o f 220 kPa.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
127
3.3.2. Evaluation o f Interlayer Properties with Nanoindentation
In FGMs, failure was typically observed to initiate at the interfaces. Therefore, interface is an
important factor controlling the properties o f the overall FGMs. Traditional hardness tests are
not suitable when testing the properties o f areas in the range o f a few micrometers as in the
case o f FGM interfaces. For this type o f applications, nanoindentation has been widely used.
This method provides a mechanical signature o f a material’s response to deformation and
enables us to calculate Young’s modulus (E) and hardness (H) [25-29],
Nanoindentation tests require high quality finished surfaces. Any roughness on the surface
will cause incorrect estimation o f the effective contact area under the tip and therefore it may
lead to uncertainties and scatter in data. Towards the higher boron content layers, the surfaces
had more porous structure due to the exothermicity of the reaction. Therefore, w e had to limit
our nanoindentation tests with only two interfaces. A typical indentation cycle was
(i)
approach the tip to locate the interface surface as accurately as possibly
(ii)
load with a rate of lON/sec up to the set force
(iii)
hold for 10 seconds (to minimize the inelastic effects by allowing tim e dependent
plastic effects (e.g. creep) to diminish
(iv)
unload at the same rate. Hardness and elastic modulus are determined from the
unloading part of the load-depth curve by using Oliver-Pharr method [30]. Three
measurements were taken from each layer and the averaged value was reported as
hardness. Figure 21 shows the schematic, which identifies the parameters used in
this method.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
128
Load (P)
Surface profile
after load
removal
indenter
Surface profile
under load
I
h : total displacement at any time during the loading.
hc: contact depth (vertical distance).
hs: displacement o f the surface at the perimeter of the contact.
hf: final residual depth.
Figure 21. Schematic drawing showing different parameters used in Oliver-Pharr method.
Load (P) versus depth of penetration (h) curves are continuously recorded in situ. The
analysis was carried out by fitting the unloading segment o f the curve to the power law
equation given in equation 4.
P =B{h-hy)m
(4)
where P is the indentation load, h is the displacement, B and m are fitting parameters
determined by empirically, and hf is the final displacement after complete unloading. The
hardness (H) can be calculated by using the following formula.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
where A is the projected area o f the indentation [30-3 I], The elastoplastic properties of the
material are reflected on their indentation hystereses as shown in Figure 22 and 23. In the
second interface, a maximum load o f lOOOpN was applied. In both situations, it was noted
from the P-h curves that once the maximum load is reached some creep occurred during the
holding step. This is expected due to the presence o f metallic phase (Ti). There are two main
reasons for the lower hardness values. First, as X ray diffraction studies indicated that there
still unreacted titanium remains in the layers and the conversion to TiB and TiB2 is low.
Second, the materials hardness not only depends on the composition but also sensitive to the
microstructure including porosity and porosity distribution. The existence of pores weakens
the materials strength. The elastic modulus can also be determined from P-h curves based on
the slope o f the unloading curve can be used as a measure o f the elastic properties of the
sample [27], As shown in Figures, the estimated E values for interface I and interface 2 are
147.56 GPa and 152.1 GPa, respectively. Room temperature elastic moduli of titanium
borides are reported to be in the range o f 500-550 GPa, and for titanium it is around 115 GPa
[25]. The low elastic modulus in our samples can be related to the lower amounts of boride
phases (as X-ray diffraction studies indicated) and the presence o f voids within the interface
region.
The depth o f indenter can be determined from the P-h curves by taking the tangent to the
unloading segment at the maximum load and extrapolating it to zero. The intercept at the xaxis gives the plastic depth. For interface 1, this depth was about 220 nm compared to 65.3
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
130
nm for interface 2. Although, the lower value obtained from interface2 indicates a higher
hardness for this interface, it should be reminded here that a lower maximum load (lOOOpN)
was applied at interface 2. However, when carefully examined, it can be seen that at lOOOpN
load, the displacement in the interface 1 was 140pN compared to 62.3 p.N for interface 2. As
can be seen from Figure 22, some o f the on-load is recovered elastically as the indenter loads
is removed.
When compared the P-h curves obtained from two interfaces, a greater curvature is
noticeable in the unloading segment o f interface 2 than that o f interface 1. This is related
with the hardness to elastic modulus ratio (H/E). This ratio represents the mechanical
behavior o f material with H/E < 0.01 indicates a ductile material. The H/E ratio for interface
1 was calculated as 3.21xl0'2 on the other hand for interface 2 it was 5.43x1 O'2. Interface I
represents a somewhat more ductile behavior than interface 2. Therefore, the amount of
deformation recovered (elastic deformation) during unloading is less than interface 2 [28],
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
131
InterfaceI
Ti/95
[..-T- T_ r ...f....l—T1400
—
1200
--
1000
■ -*
Hardness: 4.74 GPa
Er: 147.56 GPa
Contact Stiffness: 97.3 pN/nm)
800
eoo
400
200
0
50
100
150
D i s p l a c e m e n t (rtm )
w m m M
.
Figure 22. The results o f a nanoindentation test obtained from interface 1. Figure shows a
high resolution SPM image o f the interface region with the image of the indentation and also
the load -displacem ent history involved in making such an indentation.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
132
Interface 2
9 5 T i - 90 T i In terface
H ardness: 8.26 GPa
Er : 152.1 GPa
C ontact Stiffness: 59.45 pN/nm )
H a rd n e ss Im p ressio n
Displacement (nm)
Figure 23. The resuits o f a nanoindentation test obtained from interface 2. Figure shows a
high resolution SPM image o f the interface region with the image o f the indentation and also
the load -displacement history involved in making such an indentation.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
133
The residual imprints after the indentation show a “pile-up” phenomenon, which is usually, a
material response, observed with metals. During the indentation the displaced material
(mostly consists o f Ti) tends to flow up the faces of the indenter. One notable feature is that
the pile-up is non-uniformly distributed around the tip. The material under the load is more
constrained at the edges than comers therefore as it can be seen from Figures 22 and 23, the
pile up is localized at the edges o f the tip rather than comers [31].
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
134
4. Conclusions
This research has resulted in the following conclusions
1. The method pursued herein is an attractive and low cost route in synthesis and
densification o f functionally graded materials.
2. Externally applied pressure exploits the brief appearance o f liquid titanium during
combustion reaction and reduces the porosity drastically. The external pressure assists
in the filtration o f molten titanium into pores.
3. Porosity was reduced from 28.75 % to 7 % for 82Ti-18B layer and from 16.88 % to
1.72 % for 95Ti-5B after compacting under a pressure o f 220 kPa. For further
densification the inherent delay in the application o f the mechanical load must be
compensated.
4. The microstructure shows elongation perpendicular to the direction o f the application
o f pressure during processing.
5. When the applied compaction pressure was increased (to = 317 kPa), lateral macro
cracks were observed. The difference between the heat transfer in axial and radial
directions causes fast cooling outer surface and slowly cooling interior. During
compaction, the hardened outer surface becomes fragmented. It is suggested to that
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
135
by tailoring the ratio o f lateral insulation thickness to radial insulation thickness, it is
possible to apply higher compaction pressure without cracks.
6. The microhardness o f the FGM layers is increased by the application o f pressure after
the CS indicating a denser microstructure. The scatter in the hardness data, especially
in the higher boron content layers, is directly related to the ability o f placing the
indenter in a void-free region o f the sample without collapse o f adjacent grains.
Compared to monolithic TiB 2 (3200-3400 kg/mm2), the measured values are smaller.
However, as X-ray diffraction studies indicated the relative amount o f TiB 2 within the
layers are also small. With further development the method described could be an
economical way compared to hot pressing.
7. The load-penetration depth plots obtained from the nanoindentation tests indicated
deformation occurring at the maximum loads during holding time. This is expected
due to the presence o f metallic phase (Ti). Elastic modulus values were estimated as
147.56 GPa for interface 1 and 152.1 GPa for interface 2.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
136
5. Recommended Future Work
One o f the major problems encountered in this work have been lateral cracks observed at
higher compaction loads. In order to be able to produce denser products this problem
should be solved first. As an initial step the thickness o f the insulation ratio used at the
bottom and around the compact can be changed. This will allow minimizing the
difference in the heat transfer in axial and radial directions and consequently might allow
decreasing and/or completely eliminating the lateral cracks. Likewise decreasing the
annular gap between the sample and the die cavity can also be tried. At the moment, this
cavity is around 1 inch in diameter (slightly bigger than the diameter o f the punch). This
leaves an annular space o f lA inch when sample is placed. In our experiments, this space
was filled with alumina insulation. Preparing green compacts with diameter larger than lA
inch but smaller than 1inch can be tried as a first attempt. This will not also decrease the
annular space but also help to preheat the sample effectively by using the coiled cable
heaters (1/8” diameter, 120 V, 375W) surrounding the outside o f the die cavity. The cable
heater is capable o f heating the steel die up to 200°C. Heating o f the compact surrounding
will somewhat provide uniform temperature distribution within the compact during both
combustion and more importantly during cooling.
The inherent delay in the application of compaction is also considered to be an important
parameter causing insufficient densification. As was mentioned in the preceding sections,
the compact must have some plasticity at the time o f compaction. The formation o f liquid
titanium provides this. Based on the fact that the plasticity o f the material is temperature
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
137
dependent, a sharp decrease in temperature during post combustion period (a
characteristic o f the method employed), reduces the plasticity drastically. Heating the
compact surrounding as explained in the previous paragraph can also help to slow down
cooling rate and increase the plasticity.
We had to limit our nanoindentation tests with only two interfaces mainly due to the
difficulty in polishing the composites and porous structure at the higher boron content
layers. Further work on the remaining layers is believed to provide an overall
understanding of the properties o f interlayers.
Further detailed work concerning the microstructural features o f the process is also
advised. Especially when the process is conducted in conjunction with the application of
a mechanical load. An improved understanding o f the process can lead to several benefits
including, improved products and process technique, predict and prevent possible failures
related to the micro structure.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
138
6. References
1.
A. K. Bhattacharya, J. Am. Ceram. Soc., 74, 9, (1991), 2113-2116.
2.
C. He, C. Blanchetiere, and G.C. Stangle, J. Mater. Sci., 13, 8, 2269-2280.
3.
Y. Zhang and G.C. Stangle, J. Mater. Res., 10, 7, (1995), 1828-1845.
4.
Z.A. Munir, J.ofM at. Synt. And Proc., 1,6, (1993), 387-394.
5.
J.J. Moore and H.J. Feng, Prog. In Mat. Sci., 39, (1995) 243-273.
6.
R.W. Rice, and W.J. McDonough, J. Am. Ceram. Soc., 68, 5, (1985), C-122-C-123.
7.
J. C. LaSalvia, L.W. Meyer, and M.A. Meyers, J.Am. Ceram. Soc., 7 5 , 3, (1992),
592-602.
8.
I. Song, L.Wang, M.Wixom, and L.T. Thompson, J. o f Mat. Sci., 35, (2000), 26112617.
9.
L. J. Kecskes, T. Kottke, and A. Niiler, J. Am. Ceram. Soc., 73, 5, (1990), 12741282.
10. L. J. Kecskes, R.F. Benck, and P.H. Netherwood Jr., J. Am. Ceram. Soc., 73, 2,
(1990), 383-387.
11. B.H. Rabin, G.E. Korth, and R.L. Williamson, J. Am. Ceram.Soc.,73 ,7, (1990),
2156-2157.
12. B. H. Rabin, J. Am. Ceram. Soc., 75,1, (1992), 131-135.
13. S.Suresh and A. Mortensen “Fundamentals of Functionally Graded Materials” The
University Press, Cambridge, (1998).
14. H.J. Feng and J.J. Moore, J. o f Mat. Eng. and Perform., 2, 5, (1993), 645-650.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
139
15.
A.N. Pityulin, Z.Y. Fu, M.J. Jin, R.Z. Yuan and A G . Merzhanov, 4th International
Symposium on Structural and Functional Gradient Materials, (1996), Elsevier
Science, ed. by I. Shiota and M.Y. Miyamoto, 295-300
16.
E.A. Levashov, I.P. Borovinskaya, A.V. Yatsenko, M. Ohyanagi, S. Hosomi and M.
Koizumi, 4th International Symposium on Structural and Functional Gradient
Materials, (1996) Elsevier Science, ed. by I. Shiota and M.Y. Miyamoto, 283-288.
17.
Ohyanagi, T.Tsujikami, M. Koizumi, S. Hosomi, E.A. Levashov, and I.P.
Borovinskaya, 4th International Symposium on Structural and Functional Gradient
Materials, (1996),Elsevier Science, ed. by I. Shiota and M.Y. Miyamoto, 289-294.
18.
Z. Li, K. Tanihata and Y. Miyamoto, 3rd International Symposium on Structural and
Functional Gradient Materials, 10-12 Oct. 1994, ed. by B. Ilschner, and N.
Cherradi, Pres, polytechniques et universitaires romandes, 109-114.
19. Y.S. Kang and Y. Miyamoto, 3rd International Symposium on Structural and
Functional Gradient Materials, 10-12 Oct. 1994, ed. by B. Ilschner, and N.
Cherradi, Presses polytechniques et universitaires romandes, 115-120
20.
W. Lai, Z. A. Munir, B.J. McCoy and S. H. Risbud, 4th International Symposium on
Structural and Functional Gradient Materials, (1996) Elsevier Science, ed. by I.
Shiota and M.Y. Miyamoto, 275-282
21. W.A. Gooch, B.H.C. Chen, M.S. Burkins, R. Palicka, J. Rubin and R. Ravichandran,
Mat. Sci. Forum, Trans. Tech. Publications, 308-311, (1999), 277-282.
22. E.S.C. Chin, Mat.Sci.and Eng. A259, (1999), 155-161.
23. T. Allen, Particle Size Measurements, 4th edition, Chapman and Hill, (1990).
24. G.K. Dey, A. Arya and J.A. Sekhar, J. Mater. Res., 15,1, (2000), 63-75.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
140
25.
A.J. Markworth, K.S. Ramesh and W.P. Parks Jr., J. o f Mat. Sci., 30, (1995), 21832193.
26.
S.F. Corbin, X. Zhao-Jie, H. Henein and P.S. Apte, Mat. Sci. & Eng. A262, (1999),
192-203.
27.
G.M. Pharr, W.C. Oliver and F.R. Brotzen, J. Mater. Resi5 7, 3, (1992), 613-617.
28.
S.V. Hainsworth, W.W. Chandler, and T.F. Page, J. Mater. Res., 11, 8 , (1996),
1987-1995.
29.
D.F. Bahr, J.W. Hoehn, N.R. Moody and W.W. Gerberich, Acta Mater., 45, 12,
(1997), 5163-5175.
30.
T.Y. Tsui and G.M. Pharr, J. Mater.Res., 14, 1, (1999), 292-301.
31.
K.W. McElhaney, J.J. Vlassak and W.D. Nix, J. Mater. Res., 13, 5, (1998), 13001306.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
141
MANUSCRIPT FOUR
PROCESSING AND CHARACTERIZATION OF FUNCTIONALLY GRADED
MATERIALS PRODUCED BY MICROWAVES IN Ti-B BINARY SYSTEM
Menderes Cirakoglu
Department o f Materials and Metallurgical Engineering
University o f Idaho
Moscow-Idaho 83844-3024
will be submitted for publication
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
142
ABSTRACT
The feasibility o f processing Ti-B based functionally graded materials with microwave
activated combustion synthesis was studied. The use o f microwaves has been found to offer
advantages as compared to conventional combustion methods. When combined with
microwaves, combustion synthesis offers great potential for the fabrication of dense ceramics
and composites. Samples were ignited in a microwave furnace at a power of 1.5 kW and a
two-color pyrometer was used for measuring temperatures. The effects of hybrid heating,
thermal insulation, and atmosphere on the process were investigated. When the thermal
insulation surrounding the sample was reduced, more controlled combustion reactions and
net shape products were obtained. X-ray diffraction studies indicated titanium monoboride
and diboride along with unreacted titanium. When compared with conventional combustion
synthesized products, microwave processing resulted in smaller pores. However, the total
amount o f porosity remained almost the same. Scanning electron microscope and optical
microscope were used to examine the interfaces and graded layers. Over the entire crosssection, the interfaces were continuous and crack-free. Hardness was determined by Vickers
indentation technique. As expected by increasing the amount of boron in the layers, increases
hardness. The samples ignited in air have slightly higher hardness values.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
143
1. Introduction
The background information on microwave heating and associated issues are explained in
this section. This brief but necessary background includes material-microwave interactions,
comparison to traditional heating methods and microwave hybrid heating. The microwave
activated combustion synthesis is also described and then FGM systems previously studied
by this process are discussed. Finally, the objectives of this research are defined.
1.1. Microwave Heating
Although the research on heating and processing of ceramics and composites with
microwaves have started in 70’s, most o f the effort was carried out during the 80’s and is
continued into today. Early studies on microwave heating o f solid materials were mostly
concerned with the removal of moisture or organic liquids. In recent years, the versatility and
full potential o f this process have gained widespread attention. This is evident from the
growth in the number o f papers presented at the technical meetings and symposia dedicated
to microwave processing [1-4].
Materials differ in their response to microwave fields. Some materials (e.g. glasses) are
transparent to microwaves. Polar, ionic and conductive materials absorb microwaves and
therefore can be processed with microwaves. Bulk metals tend to reflect microwaves back
into the microwave cavity. On the other hand, fine metal powders can be heated with
microwaves. The increased surface area in the powder form accentuates the role o f the skin
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
144
depth o f the metal. The skin depth, which is on the order o f 1 to 100 pm, absorbs a small
amount o f microwave incident on the metal surface [3,4]. The microwave processing of
ceramics is limited by the fact that many ceramics do not absorb microwaves efficiently. A
broad range o f ceramic materials (e.g., AI2 O 3 , MgO, SiC>2 , most glasses etc) is transparent to
microwaves at room temperature. They are difficult to heat initially with microwaves and
also they are prone to localized heating and eventually cracking.
When the part consisted of materials with different microwave interactions, as in the case of
composites, local overheating may occur. For this type o f situations, hybrid (or indirect)
heating method was developed. This method involves the use o f a microwave absorbing
material to thermally heat the poorly absorbing material until it reaches to a temperature
where it will start absorbing microwaves. This strong microwave coupling materials are
called as “susceptor”. Usually SiC is used as a microwave absorbing material [1,5,6].
Microwave energy offers significant advantages over conventional thermal techniques. In
conventional heating the heat is supplied to the part externally. On the other hand, with
microwaves, the energy is directly coupled to the ceramic body causing the sample itself to
generate the heat instead of receiving it from the surrounding heat source. This provides a
volumetric heating which is more uniform than surface heating. Therefore, the normal
temperature gradient in a conventionally heated sample is the opposite o f that in a
microwave-heated sample. Volumetric heating produces less thermal stresses in samples.
This will ultimately allows fabrication o f more complex parts and shapes than are possible by
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
145
conventional processing. It also enables very high heating rates, which can be used to limit
grain growth and produce materials with superior and tailored properties [1-5].
1.2. FGM Processing by Using Microwaves
Volumetric heating achieved by microwaves reduces thermal stresses, which are critical
when joining different materials. Most o f the reported studies on microwave processing o f
FGMs have dealt with microwave sintering in less exothermic systems. The Borchert et al [7]
fabricated FGMs by using microwaves in the systems such as AI2 O 3 /M 0 , AkC^/steel and
ZrO2-Ni80Cr20. The graded layers were prepared by sequential filling o f different
compositions in a silicone mould and then cold isostatic pressing. They investigated the
influence o f the FGM composition on the formation o f thermal gradient during microwave
sintering. Pressureless microwave sintering was performed at a power level o f 2 kW under
argon-hydrogen mixture atmosphere.
In another study Porada et al [8 ] reported processing o f ZrCVmetal FGMs by infiltration and
microwave hybrid heating using SiC as a susceptor. Infiltration procedure was performed
by pouring the metallorganic compound (zirconiumtetrapropylate) onto the porous metal
matrix (copper and iron based alloys and nickel) through a nozzle and heating under argon or
vacuum. They reported that by repeated infiltration and decomposition o f the metallorganic
compound, complete filling o f the pores with a ceramic phase was achieved.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
146
Our study differs from these earlier ones in that it examines the feasibility o f microwave
hybrid heating to fabricate FGMs in Ti-B binary elemental system. As is well known
depending on the composition chosen, the reaction between titanium and boron can be highly
exothermic. This method makes use o f both the energy released from the combustion
reaction and microwave material interaction. It holds the promise to be commercially viable
as it is rapid and low cost. However, as a manufacturing process it poses some challenges as
described in the following section.
1.3. Microwave Activated Combustion Synthesis
In recent years, microwave heating has also been used as an alternative method to ignite
materials during combustion synthesis. In conventional combustion synthesis, the heat is
provided by current carrying coils however microwaves penetrate the material to produce
volumetric heating. Thus, microwave heating provides rapid and uniform heating. Since the
heating is generated internally, the heating process is not heat transfer limited. Clark et al [9]
pointed out that by utilizing microwaves, the dependence of combustion reaction on the
thermal conductivity and density o f the compact is greatly reduced and further densification
can be achieved simply by leaving the microwave power on after the reaction. This is not
possible in conventional combustion processes because once the combustion has taken place
the temperature decreases rapidly. For materials, which are difficult to heat by microwave
energy, hybrid-heating technique is utilized to initiate the reaction. A susceptor is used to
conduct the heat to the surface o f the sample and materials are pre heated to a temperature at
which they react exothermically. The microwave energy is absorbed within the material and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
147
after the ignition, microwave energy and combustion reactions assist each other in sustaining
the reaction [9,10].
Microwave activated combustion synthesis can overcome the shape complexity limitations of
the traditional combustion processes and allow rapid production in large amounts. Therefore,
they are considered as an efficient process compared to conventional combustion processes
[11,12]. However, in order to implement this process into large-scale production, several
issues such as the control o f the ignition and propagation o f reaction front have to be well
established. The control o f the microwave heating is an important issue especially when
highly exothermic reaction are involved. A rapid increase in the absorbed power with
increasing temperature along with high exothermic reactions often result in sudden,
uncontrolled temperature surges in the specimen. Having control on the combustion wave
front allows for a gradual release o f volatiles and result in denser products. Besides the
density and thermal conductivity o f the compact, the control o f the wave propagation can be
controlled by the incident microwave power. It has been reported that the wave propagation
can be terminated by simply turning o ff the power. Additionally by pulsing the incident
power would even give more precise control on the velocity o f the propagation [11-15],
1.4. Objectives
This work reported in this chapter has been undertaken as a part o f our FGMs processing
studies reported in the previous chapter. Due to the obvious advantages o f microwave
heating, a great deal of research was carried out on the processing o f a variety o f engineering
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
148
materials including FGMs. However, to the best o f our knowledge, we are not aware o f any
effort that utilizes microwave activated combustion synthesis for the fabrication o f FGMs.
Given the difficulty as outlined above, the question is whether net shaped FGMs can be
produced with microwave activated combustion synthesis. It is our goal in this chapter to
seek an answer to this question. Several processing parameters including the effect of thermal
insulation and atmosphere are studied. The comparison o f results obtained from microwave
assisted combustion synthesis will be compared to those from conventional combustion
synthesis. Based on this knowledge, expanded research can open up the possibility o f the
processing o f functionally graded materials using microwave energy.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
149
2. Experimental Procedure
This section describes the experimental procedures followed to fabricate FGMs by
microwaves. The materials, composition selection and sample preparation processes are the
same as the one used in conventional combustion synthesis as described in the preceding
chapters. Therefore, these issues are briefly explained. Being a new manufacturing process,
microwave activated combustion synthesis poses new challenges in process planning.
Therefore, a variety o f experimental techniques are investigated in order to optimize the
process. The descriptions o f these experimental techniques are also described in this section.
The section ends with the description o f the characterization methods employed to evaluate
the products.
2.1. Preparation of Layered Structures
Elemental Ti (-325 mesh, purity: 99%, Johnson Mathey, Ward Hill, MI) and crystalline B
(average 15 pm or less, purity > 99%, Cerac Inc. Milwaukee, WI) powders were used as raw
materials. Powder premixes o f different mixing ratios were prepared by ball milling for 3
hours in polyethylene bottles with alumina balls. Different compositions were stacked layer
by layer in a steel die having a cavity diameter o f 12.75 mm as shown in Figure la. The
gradient consisted usually o f 5 layers of approximate the same thickness with 0, 5, 10,15 and
18-wt % boron. Cylindrical samples with 12.75 mm diameter and 16.30-16.50 mm height
were obtained after uniaxial pressing under a load o f 15,000 lbs subsequently cold
isostatically pressed at 100 MPa. The green compacts were placed in a tube furnace at 550°C
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
150
for 3 hours under flowing argon gas for degassing. Microwave experiments were conducted
in an industrial microwave furnace (Microwave Materials Technology (MMT), Knoxville,
TN). Figure 2 shows the experimental flow chart.
Figure 1. Schematic drawing (a) and a picture (b) of a 5-layered FGM green compact
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
151
C
B
Ti
1r
^
Dry Mixing
^
r
^
Uniaxial Pressing
^
r
Cold Isostatic Pressing
r
^
Degassing
mMicrowave
i
Activated Combustion A
)
(in air'i
^
^
f Microwave Activated Combustion
Characterization
(in arson)
J
Figure 2. Flowchart showing the steps o f microwave processing o f FGMs
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
152
2.2. Microwave Activated Combustion Synthesis Experiments
The experimental set-up is shown in Figure 3 and a photo o f the microwave furnace is shown
in Figure 4. The system has variable power output magnetron source capable o f operating
from 0 to 3 KW at 2.45 GHz. The microwave processing system is a cylindrical multi-mode
cavity with an internal volume o f 500 liters (4 ft. in length, 2-1/2 ft. in diameter). The cavity
is equipped with a mode stirrer and multiple microwave inputs to enhance the uniformity of
the microwave field. The microwave sources generate up to 3 kW o f microwave power at a
frequency o f 2.45 GHz.
Samples were placed into a box made out of low-density alumina fiberboard and alumina
blanket was used for thermal insulation. The purpose o f insulation in a microwave energy
heating process is to keep the heat from escaping quickly. The low density alumina
fiberboard (Type ZAL-15, Zircar Fibrous Ceramics, NY) was used as the outer envelope.
The box was made by joining the fiberboards by using an alumina cement (Type AL-CEM,
Zircar Fibrous Ceramics, NY). The cement contains 98.5 wt% AI2 O 3 and 1.5 wt% SiC>2 and it
was applied by using brushes. The boards are high strength; rigid, low heat storage and
excellent thermal shock resistance and contains high purity inorganic binders. Since it is
microwave transparent it allows penetration o f the microwaves into the sample. As inner
layer o f insulation, alumina blanket (Type AB, Zircar Fibrous Ceramics, NY) was used. An
input power o f 1.5 kW was applied to the cavity through waveguides. The experiment was
observed through the quartz window and as soon as the ignition started to propagate the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
153
power was shut down. The experiments were recorded by a Sony TRV900 digital video
camera. Samples were left to cool inside the chamber.
Figure 3. A schematic of the experimental set-up.
1) Magnetron 2) Main Controller 3) Forwarded Power Indicator 4) Reflected Power Indicator
5) Wave guide 6) Computer Control 7) Optical Pyrometer 8) Sample Insulation 9) Argon
Tank 10) Vacuum Pump
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
154
Figure 4. Photograph o f the microwave furnace used in our experiments.
Combustion experiments were divided into four categories:
(i) without SiC susceptor: In this case, samples were placed inside an alumina fiber board
enclosure and surrounded with alumina blanket. Alumina fiberboard was chosen for its
transparency to microwaves and its stability to withstand relatively high temperatures. The
enclosure served for the purpose o f containing the heat during the process. The front o f the
enclosure was also covered with alumina blanket. This assemblage was heated in air. A
small hole (0.5" x 0.5") was opened on the blanket in order to measure the temperature.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
155
(ii) with SiC Susceptor: Samples were placed on a SiC block, which acts as a susceptor
material. The specimens were inserted into the box and placed on a strip o f dense P-SiC,
which was used as a susceptor. SiC. As described earlier, in order to heat many technical
ceramics, susceptors are required to boost the temperature up to a point where the ceramic
itself will couple with the microwave energy. The front opening o f the box was also covered
with alumina blanket (thickness Vz in.) and a square slit was opened in order to see through
the sample. The insulation and indirect heating arrangement used in our experiments is
shown in Figure 5.
alumina blanket
sample
^
“ 10 cm—
►
Figure 5. Insulation and indirect heating arrangement used in our experiments.
(iii) with SiC susceptor and reduced thermal insulation: We kept outer thermal insulation
(alumina fiberboard box) and reduced the amount o f inner insulation (alumina blanket) so
that a space o f 1 inch was left around the sample. SiC susceptor was placed on alumina
blanket and the front opening in the enclosure was covered with a thin layer o f insulation
with an enlarged hole ( 1 " x 1 ") to measure the temperature and to record the combustion
reactions.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
156
(iv) in Argon: The effect o f process environment was studied under argon atmosphere. The
samples were placed in a Pyrex glass flask, which was first evacuated and then filled with Ar
gas. The gap between the sample and glass flask was filled with an alumina fiber blanket.
A two-color pyrometer (Mikron L77, temperature range o f 900-3000°C, Mikron Inc.
Oakland, NJ) was used for temperature measurements viewing through a quartz window. A
pyrometer is a non-contact heat-sensing device that measures inferred wavelengths.
Pyrometers have the advantage o f measuring very high temperatures without having in
contact with the material to be heated. The temperature o f the sample was monitored by
using a two color pyrometer (Mikron Instruments,Oakland, NJ, model M77LS, temperature
range: 1000-3000°C). Main problem associated with this temperature measuring system is its
inability to sense low temperature and the necessity to visibly see the material to be heated.
This means that a sighting hole must be available through the insulation surrounding
material. The pyrometer was set about 35 cm away from the sample and aligned with the
quartz view port with a diameter o f about 2 cm which is located on the front door o f the MW
chamber. The measured spot on the surface o f the sample was approximately 2 mm in the
vicinity o f the center o f the cylindrical sample. The pyrometer output the current (in
miliamps) readings in real time during the experiment to a host computer. A data acquisition
board (WB-Dynares, Omega Engineering,Inc. Stamford, CT) together with a graphical
interface application software (QuickLog PC,Omega Engineering, Inc., Stamford, CT) was
used to collect data at every 0.5 s intervals. The DAQ system has eight differential analog
input channels. The pyrometer wires were connected to one o f the specified channel’s screw
terminals inside a terminal panel. The terminal panel is connected to the data acquisition
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
157
board inside the host computer with a 50-pin ribbon cable. This specified channel was set to
measure the current in the range o f 4-20m A The temperature profiles were then calculated
off-line at the end o f the data acquisition process by using the calibration data for the
pyrometer with an accuracy o f 0.5 % o f reading.
2.3. Characterization
The as-prepared samples were sectioned and mounted in thermosetting resin. Samples were
initially polished using SiC papers from grade size 180 to 4000 and finally with alumina
suspension from 5 pm to 0.01 pm. A solution of 10 % HF, 40 % HNO 3 and 60 % H 2 O was
used to etch the surface of the samples. The composition o f each layer was characterized by
X-ray diffraction technique (Siemens D5000, CuKa radiation).
The samples were analyzed for density, porosity, phase constituents and microstructure.
Density was measured before and after combustion experiments by using Archimedes
technique. Ethyl alcohol (d=0.7893 g/cm3) was used as the immersion medium. The
specimen was placed in ethyl alcohol and vacuumed for at least 30 minutes to ensure that all
the pores were filled with alcohol. In order to evaluate the type o f pore and pore volume
distribution, porosimetry measurements were carried out by using Micromeritics AutoPore
i n mercury porosimetry.
The room temperature hardness o f each layer was determined by a Vickers testing machine.
A load o f 1 kgf was applied for 30 seconds. The microstructures of graded layers were
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
158
examined by using an optical microscope (Leco,Olympus PMG3, St. Joseph, MI) and a
scanning electron microscope (AMRAY 1830, Amray Inc., Bedford MA). Elemental analysis
was performed using an energy dispersive spectroscopy (EDS) attached to the SEM.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
159
3. Results and Discussions
In this section, the results o f microwave experiments are presented. First, combustion
characteristics o f single and graded compositions including time- temperature profiles are
explained. The effect of hybrid heating and reaction atmosphere are discussed. The results o f
density and porosity measurements are summarized and compared to conventional
combustion synthesis products. Finally microstructure characteristics and hardness test
resuits are discussed.
3.1. Combustion Characteristics
Combustion experiments with microwaves were divided into four categories. The results are
summarized below.
(i)
Samples placed without a SiC susceptor:
No significant temperature rise was observed and hence no meaningful temperature signal
was recorded mainly due to the fact that our pyrometer has a lower limit o f 900°C. Therefore,
we can conclude that the temperature never reached to 900°C.
(ii)
Samples placed on a SiC susceptor with thermal insulation:
As previous experiments indicated, no significant microwave heating took place without a
susceptor. Therefore, we utilized hybrid heating concept and used a SiC susceptor to pre heat
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
160
the compacts of Ti and B to temperatures at which they will readily absorb microwave
energy.
Approximately, 240-300 seconds were required for samples o f mass 6-7 grams to reach the
combustion temperature with this arrangement. A maximum surface temperature o f 1900°C
was recorded by pyrometer. Packed insulation around the sample combined with the high
exothermicity of the reaction caused high in-situ temperatures and uncontrolled reaction,
which resulted in deformed products as shown in Figure 6 . At the moment o f ignition,
sample reached intense white incandescence within the cavity.
Figure 6 . Photograph of the sample exploded during MW processing
(iii)
Samples placed on a SiC susceptor with reduced thermal insidation:
Figure 7 shows the video images from combustion o f a single composition o f 82Ti-18B. The
susceptor started glowing at about 450 seconds and the sample started to heat up. At about
540 seconds, the whole sample was glowing and then combustion wave started to form at the
top o f the sample and propagated towards bottom. The heat loss from the reaction front to
the environment reduced the heat transfer to the adjacent layers and therefore all single
compositions except 82Ti-18B retained their original shapes.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
161
The measurement o f the heat generated in a combustion synthesis reaction is important. The
heat generated is directly related to kinetics and thermodynamics o f the process. Under
normal conditions, thermocouples would be adequate measuring device and are used in many
applications. However, in a microwave environment, electromagnetic field within the cavity
can damage the metallic thermocouple especially at the tip [3]. For this reason, a non-contact
temperature measurement technique was utilized. Heating profiles were recorded by using a
two-color pyrometer focused on a small hole drilled in the middle o f the green pellet after
degassing. The depth o f the hole was around
6
mm. As mentioned previously, the pyrometer
only records temperatures above 900°C. There is a discontinuity in our temperature
measurements mainly due to the lack o f low temperature measurement means. Figure
8
shows the time temperature profiles obtained from three different single compositions, which
are used to produce FGMs. As expected, 82Ti-18B composition resulted in a highest
combustion temperature o f about 2700°C. On the other hand composition 95Ti-5B showed a
much lower combustion temperature o f about 1300°C. Although a sharp increase in
temperature was observed with this composition, no clear combustion wave formation was
observed.
In conventional combustion synthesized samples, maximum temperatures were reached in
about 10 seconds. However, this study also showed that in the microwave activated
combustion process, the pre combustion time is much longer. We recorded a pre combustion
time o f 240-300 seconds with intensive thermal insulation and 550-680 seconds with reduced
thermal insulation as shown in Figure 8 . The possible reason for this is related to the heating
technique. In conventional combustion method, the sample is directly heated by the external
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
162
heat source; however, with microwaves, the sample is heated only after the susceptor is hot.
Similar to conventionally ignited samples, the pre-ignition time is directly related to the
composition. The ignition time decreased as the amount o f boron increased. Another
important point is related to the maximum combustion temperatures. In microwave activated
combustion synthesis, the maximum temperatures are higher compared to conventional
combustion synthesized samples. The difference is quite discernible for composition 82Ti18B. Because, in the case o f conventional combustion synthesis, a maximum temperature o f
1850°C was recorded, on the other hand in the microwave heating the temperature was
2700°C.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
163
t « 555 sec
t * 570 sec
Figure 7. Combustion wave propagation o f sample 82Ti-18B during microwave processing.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
164
4
-6.5 mm
*-13 mm
Sin«it
Composition
12mm
3000
82Ti-18B
2500
90Ti-10B
O 2000
oa.
M
1500
5Ti-5B
io o o
500
500
700
600
800
time (seconds)
Figure 8. A schematic o f the single composition sample used in microwave experiments.
The cross indicates the spot where the pyrometer was focused. Time temperature profiles
obtained from single compositions
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
165
In FGM experiments, Ti side was in contact with the SiC block. By visual examination we
observed that, the ignition started from bottom (Ti side) and propagated towards the top (high
ceramic content side) as shown in Figure 9. When the combustion wave reached to the top,
another ignition took place propagated towards the bottom similar to single composition
82Ti-18B. On the basis o f these observations, the following event sequence can be
envisaged. Initially the sample is preheated during the pre-ignition stage due to the
conduction from the SiC susceptor. When the ignition temperature is reached, combustion
reaction started similar to conventional combustion synthesis. Combustion wave propagated
towards the top layers. As the temperature is increased and top 82Ti-18B layer is reached,
considerable amount o f heat generated and another ignition took place. Initiation o f this
reaction is likely to influence the product formation. Schematic drawing of the compact and
temperature profile for five-layered functionally graded material is shown in Figure 10.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
166
-12 mm
82Ti -18B
85Ti
V 15B
I-
4 mm
90Ti -10B
-1 6 mm
95Ti -5B
Combustion wave direction
t=561 sec
t=565 sec
Figure 9. Schematic of a five layered FGM and video images from the combustion
experiments. The black dot in the schematic indicates the hole where the pyrometer was
focused.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
167
- 1 2 mm
i I
82T i -1 8B
85T i
•
15B
- 4 mm
'r
90Ti -10B
- 1 6 mm
95T i -5B
Ti
ir
1600
1400
P . 1200
1000
_
<
L> 800
I*
600
H
400
200
600
700
800
900
1000
1100
1200
time (seconds)
Figure 10. Schematic drawing o f a five layered FGM and time temperature profile recorded.
The black dot in the schematic indicates the hole where the pyrometer was focused.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
168
This method of controlling the combustion is not an optimal method, however the product
FGM kept its original shape. By reducing the heat insulation, sample retained its original
shape as shown in Figure 11 as opposed to Figure 6.
Figure 11. The cross section o f a 5 layered FGM after MW processing.
(iv)
Samples placed into an Ar filled flask with SiC susceptor and thermal insulation:
To determine the effect o f atmosphere on combustion, we carried out several experiments in
argon filled flask. Similar to the experiments carried out in air, a strip o f SiC block was used
as susceptor and alumina fiber was used for thermal insulation. In these experiments, the
main problem was arcing during processing. As is well known, the processing atmosphere,
argon interacts with the microwave field. If a non-oxidizing atmosphere is required, in order
to prevent problems associated with arcing nitrogen is often used instead o f argon [10]. But
in our case, in order to prevent TiN formation, we tried to avoid using nitrogen gas. During
our experiments, we observed intense arcing and tiny sparks appeared on the sample surface,
which can be attributed to local electric field concentration around surface defects and rough
spots. Some of the earlier works [6,11], showed that arcing and the breakdown o f gas
disturbs the electric field and degrades coupling and as a result o f this the sample temperature
drops. This condition also caused difficulty in measuring the proper temperature. Under these
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
169
conditions approximately 11-12 minutes were required for samples o f mass 6-7 grams to
reach the ignition temperature. The temperature was measured through a small hole on the
alumina blanket and a maximum temperature of 1160°C was recorded. It must be noted here
that, this temperature corresponds to the outside temperature of the samples. As it was
pointed out by Clark et al [14] unlike conventional heating microwave heating provides an
inverted temperature profile with the highest temperature in the center o f the sample.
3.2. X-Ray Diffraction Studies
X-ray diffraction studies were carried out on the individual FGM layers. The patterns of the
samples ignited in argon atmosphere and in air are shown in Figure 12 and 13, respectively.
The first (Ti), second (95Ti-5B) and third (90Ti-10B) layers contained mainly Ti. No boride
phase was detected by X ray diffraction. Fourth (85Ti-15B) and fifth (82Ti-18B) layers
consisted o f TiB and TiB2 along with unreacted Ti. The fifth layer contained an increased
amount o f TiB and TiB2 as expected. The presence o f unreacted titanium indicates that
despite the combustion reaction, the conversion to titanium boride was not fully completed.
This suggests that the temperature did not reach adiabatic temperature due to heat losses. The
adiabatic temperature for the formation of TiB2 phase is 3 190K. However, the temperature of
the specimen would never rise to such a high value and therefore the amount o f TiB2 phase
forming was very small.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
170
The entire outer surface o f the samples, which were ignited in air, was covered by a thin
layer, identified as TiC>2 by X-ray diffraction. On the other hand, the cross section shows a
considerable amount o f TiN as a reaction product along with titanium borides. This might be
due to the reaction between titanium and air trapped in the pores o f the sample. The
formation o f Ti0 2 coating on the surface prevented further oxidation.
+ Ti
• TiB
#TiB2
>»
*35
c
<u
c
82T1-18B
*
MkwhrtfcAtnLilwAy
8ST1-1SB
25
30
35
40
2-Theta
45
50
55
Figure 12. X-ray diffraction patterns o f two layers o f FGMs ignited in argon.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
171
+ Ti
•TIB
#TiB2
ATiN
Intensity
82TI-18B
85TI-15B
"'I
25
30
35
40
45
50
2-Theta
Figure 13. X-ray diffraction patterns of two layers of FGMs ignited in air.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
55
172
3.3. Density and Porosity Measurements
Figure 14 shows the density change in the single compositions before and after MW
processing. The bulk density o f the single composition compacts increases as the boron
content decreases for both before and after microwave experiments. This is an expected
result as the density of boron (2.34 g/cm3) is much lower than that o f Ti (4.51 g/cm3). After
microwave experiments, the change in density is prominent in lower boron content layers. It
might be expected to have an important increase in density for composition 82Ti-18B after
the microwave combustion, however as found in conventional combustion experiments, the
formation o f pores suppresses the increase. When compared to the results obtained from
conventional combustion synthesis experiments, no much of difference was observed.
After microwave experiments, each individual FGM layers was sliced through their interface
and density measurements were carried out. The results are shown in Figure 15.
As mentioned earlier, porosity in composite materials is an important physical parameter.
Porosity is related to density and mechanical integrity o f these materials. It also can be an
indication of delamination, local voids and microstress concentrations in ceramics. Porosity
and pore size distributions o f the FGMs ignited in air were measured by mercury
porosimetry. The values of the porosity o f FGMs after MW processing are shown in Table I.
When compared to the pore sizes o f the FGMs produced by conventional combustion
synthesis, the pores are smaller. However, there is not much difference in the total amount o f
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
173
porosity. Microwave heating provides low heating and cooling rates compared to
conventional combustion synthesis, which possibly led to an increase in the capillary
spreading o f pores with liquid titanium.
3.4
after MW
g 3.3
o
^ 3.2
3.1
e
~
£
3
=* 2.9
®
iefore MW
2.8
82-18
85-15
90-10
95-5
Composition (wt%)
Figure 14. Bulk density change before and after combustion synthesis
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
174
3.4
i
3-3
95Ti- 5B
lb 3.2
*c /i 3 1^
c
3
Q 2.9
3
09
82TM8B,
2.8
2.7
Ti
95Ti-5B
90Ti-10B
85Ti-15B
82Ti-18B
FGM Layers (wt%)
Figure 15. Bulk density change in different layers o f FGM after microwave combustion
experiments.
Table I. Mercury porosimetry results of two FGM layers after microwave combustion in air.
FGM layer
Porosity (%)
Ti
82-18
19.6
31.3
Bulk Density
(g/cm3)
3.2805
2.9157
Avg. Pore Dia.
Gun)
9.03
9.28
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
175
3.4. Microstructural Analysis
Both SEM and optical microscopy analysis were carried out on the polished and etched
samples. Examination o f the products by a scanning electron microscope indicated a denser
core and porous outside as shown in Figure 16.
i'-' t . ' - - . '
, hi.
O
KV
‘ ** ...........
f 1 J ’>
Figure 16. SEM image showing the dense inner section of FGM. The outer shell contains
pores
Figure 17 is an optical microscope picture showing the interface between the first and the
second layers. Over the entire cross-section, the interfaces were continuous and crack-free.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
176
Q STi-SR
Figure 17. Optical micrograph showing the interface between Ti layer and Ti-5wt% B
layer (magnification: x200)
In Figure 18, surface structures of Ti layer of the FGMs ignited in air (a) and in argon (b) are
shown. As the figures indicate, the surface consists of dense, locally sintered areas with areas
o f small pores.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(a)
(b)
Figure 18. SEM image of Ti layer o f FGM ignited in air (a) and in argon (b)
The apparent porosity evident in these figures is due to the particle pull-out during
metallographic preparation. A similar phenomenon was also observed for the other layers.
The microstructure o f the FGM produced in argon 18 b, indicates a considerable contact
between Ti particles.
In Figure 19, the microstructures o f fifth layer o f the FGMs ignited in air (a) and in argon (b)
are shown. A typical EDS spectra obtained from the marked particles is given in Figure 19 c.
The layer exhibits a microstructure consisting essentially o f titanium boride particles
surrounded by Ti matrix for both cases. However, in argon atmosphere, the layer contains a
higher volume fraction of titanium boride phase.
The EDS analysis has identified the gray particles as Ti and darker particles (indicated with
a cross) as Ti and B, as shown in Figure 19 c. This indicates the formation o f titanium boride.
Quantitative identification by EDS has been attempted but was not successful because B
gives low levels o f X-rays. Therefore, we could not determine the exact chemical formula o f
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
178
this boride phase. As minor elements O, A1 and Si were also detected. These are mainly from
the polishing stage since we used alumina suspension and SiC abrasive papers.
(a)
(b)
(c)
Figure 19. SEM images of the 82Ti-18B layers o f the FGMs (a) ignited in air (b) ignited in
argon (c) EDS spectra
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
179
3.5. Mechanical Properties
Hardness testing can be a relative measure o f the degree of reaction since the hardness
increases as the density increases. Therefore, a number of hardness measurements were
conducted on the individual layers. In Figure 20 and 21, the hardness test results were shown
for the FGM samples ignited in air and argon atmosphere, respectively. Increasing the
content o f boron in the mixture slightly increases hardness. The increased hardness resulted
from the formation of borides as revealed by the X-ray diffraction analysis. However, the
influence of the boride phases on the hardness values was small because o f the small volume
fraction o f these phases. The samples ignited in air have slightly higher hardness values.
200
_
150
“
100
O)
50
1 st la y e r
2 n d la y e r
3rd la y e r
4 th l a y e r
5th l a y e r
FGM L ayers
Figure 20. Vickers hardness data o f the individual layers in FG M ignited in air.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
180
200
_
cn
o
o
15 0
2
100
Z
>
X
50
1st lay er
2 n d lay er
3 rd l a y e r
4 th la y e r
5 th l a y e r
FGM L ay ers
Figure 21. Vickers hardness data of the individual layers in FGM ignited in argon.
The variation in hardness data may be due to the type of porosity present and to the irregular
shape of the samples. The present setup o f microwave cavity does not allow the application
o f pressure during the process. The hardness values o f microwave activated combustion
synthesis carried out in air is slightly higher than conventional combustion synthesized
samples. The formation o f TiN phase along with titanium borides is a possible explanation
for this observation. Thomas et al [16] reported a hardness change along the diameter o f their
reaction bonded silicon nitride samples. They found that the samples were significantly
harder in the center compared to the edges. In our samples we have not observed a significant
change in hardness profile. However, as Figure 16 indicates that the interior o f the samples
seems to be fully reacted than outside.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
181
4. Conclusions
The main conclusion from this work is the fact that it is possible to use microwave hybrid
heating for the production o f functionally graded materials. Microwave activated combustion
processes and conventional combustion processes are inherently different in terms o f
processing, thermal history o f the samples and final properties. Using a SiC susceptor
provided an effective mean in having a uniform temperature distribution and pre heating.
Combustion experiments carried out under argon atmosphere resulted in TiB and TiB2 in a
titanium matrix formation on the other hand under air TiN formation was observed along
with borides and titanium. Arcing was the main problem when operating under argon
atmosphere due to its low breakdown potential. This also caused difficulty in measuring the
temperature.
Microstructural examination revealed that over the entire cross section, the interfaces were
continuous and crack free. The hardness tests showed no considerable variations in hardness
values between air and argon atmosphere. But, the hardness values o f FGMs produced in air
are slightly higher than conventional combustion synthesized samples. The formation of TiN
phase along with titanium borides is a possible explanation for this observation.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
182
S. Recommended Future Work
At the present time, several problems have still to be solved because o f the experimental
applications o f microwaves. The problems demand essentially the control o f the heating rate
by measuring the temperature and optimize the power consumption throughout the process at
any time. In our experiments, the major obstacle was the temperature measurements at low
temperature avoided us to control the process. A low temperature measurement technique
should be implemented to the present system. An insulated thermocouple or a low
temperature pyrometer can be used. This will provide a feedback to the controller and it will
be possible to continuously vary the microwave power in response to the difference in
temperature between the measured and set point. However, it should be noted here that
although the use o f susceptors permits preheat the samples, it has its own problems. The
temperature control and balance the temperature difference between the susceptor and the
component seems will remain a difficult task. This is mainly because the dielectric properties
o f susceptor and sample are different and it changes with temperature. Accordingly it is also
important to study the material properties i.e. the high temperature microwave properties of
both titanium and boron must be studied. This can provide an ability to modify and control
them. The processing of FGMs also requires detailed knowledge o f the actual thermal
gradient developing during the microwave heating and its dependence on the local phase
composition. Mainly because they are inherently heterogeneous and different components
may preferentially absorb microwaves. The knowledge o f thermal history and temperature
profile within the specimen is crucial for the scale up o f this method.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
183
6. References
1.
W.L. Sutton, Ceram. Bull., 68, 2, (1989), 376-386.
2.
A.C. Metaxas and J.G.P. Binner, Adv. Ceram., 1, (1990), 285-367.
3.
M.A. Janney, H.D. Kimrey and J.O. Kiggans, Mat. Res. Symp. Proc., eds. R.L. Beatty,
W.H. Sutton, and M.F. Iskander, 269, (1992), 173-185.
4.
I. E.Campisi, L.K. Summers, K.E. Finger and A M . Johnson, Mat. Res. Symp. Proc.,
eds. R.L. Beatty, W.H. Sutton, and M.F. Iskander, 269, (1992), 157-162.
5.
M.C.L. Patterson, P.S. Aste, R.M. Kimber and R.Roy, Mat. Res. Symp. Proc., eds. R.L.
Beatty, W.H. Sutton, M.F. Iskander, 269, (1988), 291-300.
6.
H. Fukushima, T. Yamanaka, and M. Matsui, Mat. Res. Symp. Proc., eds. W.H. Sutton,
M.H. Brooks, and I.J. Chabinsky, 124, (1988), 267-272.
7.
R. Borchert, M. Willert-Porada, Ceram. Trans., 80, (1997), 491-498.
8.
M.Willert-Porada, T. Gerdes, S. Vodegel, Mat. Res. Soc. Symp. Proc, 269, (1992), 2052210 .
9.
D.E. Clark, I. Ahmad, and R.C. Dalton, Mat. Sci. and Eng., A144, (1991) 91-97.
10. D.Atong and D.E. Clark, Ceram. Eng. & Sci. Proc., 20, (1999), 111-118.
11.
W.H. Sutton, Mat. Res. Symp. Proc., eds. R.L. Beatty, W.H. Sutton, and M.F. Iskander,
269, (1992), 1-20.
12. D.E. Clark, D.C. Folz, R.L. Schulz, Z. Fathi, A.D. Cozzi, A. Boonyapiwat, P.
Komarenko, R. DiFiore and C.B. Jones Mat. Res. Symp. Proc., eds. M.F. Iskander, R.J.
Lauf, and W.H. Sutton, 347, (1994), 489-500.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
184
13.
M. Cirakoglu, S. Bhaduri and S.B. Bhaduri, 6th International Symposium on
Functionally Graded Materials, September 10-14, 2000, Estes Park, Colorado.
14.
J. Jokisaari, B.E. Shores, W.A. Prisbrey, M. Cirakoglu, S. Bhaduri and S.B. Bhaduri,
2nd World Congress on Microwave and Radio Frequency, Processing o f Bridging
Science, Technology and Applications, April 2-6, 2000, Orlando, Florida.
15.
S. Bhaduri, J. Jokisaari, M. Cirakoglu and S.B. Bhaduri, 2nd World Congress on
Microwave and Radio Frequency, Processing o f Bridging Science, Technology and
Applications, April 2- 6 2000, Orlando, Florida.
16.
J. J. Thomas, T.R. Jesse, D.L. Johnson, and H. M. Jennings, Mat. Res. Symp. Proc.,
eds. R.L. Beatty, W.H. Sutton, M.F. Iskander, 269, (1988), 277-283.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
185
MANUSCRIPT FIVE
DEFORMATION ANALYSIS OF Ti-B BASED FGMs BY PHASE SHIFTED
MOIRE LNTERFEROMETRY
Menderes Cirakoglu
Department o f Materials and Metallurgical Engineering
University o f Idaho
Moscow-Idaho 83844-3024
will be submitted for publication
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
186
ABSTRACT
Phase Shifted Moire Interferometry (PSMI) was used to measure the deformation of
functionally graded materials produced by conventional combustion synthesis and
microwave activated combustion synthesis. The PSMI uncovered the influence of
compositional gradient and heterogeneity on the deformation. The three-layered FGM
samples were loaded in compression and surface displacements were measured. All samples
exhibited smooth uniform deformation despite their anisotropic architecture. Also, in all
samples, higher boron containing layers indicated a more compliant behavior as opposed to
titanium layer. The microwave-processed samples exhibited a more compliant behavior
compared to those produced by conventional combustion synthesis. In these samples, some
residual fringes remained at under lower loads (0.37 kN) indicating a plastic strain. Also after
re loading they exhibited an increased elastic modulus, which is attributed to a strain
hardening behavior. A complete series o f fringe patterns obtained from all load stages are
given in Appendix I and IT.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
187
1. Introduction
In the previous manuscripts, we have discussed the processing aspects o f FGMs. During their
service, they are expected to perform successfully under high loads, impacts and large
temperature gradients. The engineering performance o f functionally graded materials
depends on the transfer o f load from matrix to the reinforcement and changes from layer to
layer. Therefore, it is important to have an understanding of the influence o f graded layers
and processing technique on the overall properties o f the composite.
In this study, our objective is to utilize phase shifted moire interferometry technique to
evaluate the mechanical response of our FGMs. Although, the mechanical behavior o f graded
structures were widely studied and reported in the literature, the majority o f these works are
theoretical. This method allowed us to measure very fine spatial variations o f deformations in
each individual layer. To the best of our knowledge, Winter et al first utilized phase shifted
moire interferometry to study full field surface displacements in graded Ni-Al2 C>3
composites. Current work was conducted to measure measure surface displacements o f our
unsymmetric Ti-B based FGM specimens loaded in compression.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
188
1.1. Moire Interferometry
Moire interferometry is capable o f measuring in-plane displacements with very high
sensitivity and excellent clarity by providing whole-field patterns. In this technique, the
interference between sample grating and reference grating is used to magnify the surface
deformations. The result is a contour map or Moire fringe patterns, which consist o f broad
dark and light bands. The fringes are created by constructive and destructive interference o f
between two mutually coherent beams o f laser light. The details of the fringe formation will
be discussed in the following section.
Moire Interferometry offers several advantages. It is real time technique therefore the
displacement fields can be viewed as loads are applied. The in-plane displacements i.e. the
displacements parallel to the surface o f the specimen can be measured with high sensitivity
(typically 2.4-fringes/|j.m displacement) and measurements can be taken from small regions.
The method is compatible with a large range of displacements, strains, and strain gradients.
Although, Moire interferometry is a relatively new technique, it has been used on a variety o f
materials and stress-analysis problems. The method developed mostly in 1980s and materials
such as steel, aluminum, titanium, concrete, composites (both fibrous composites and metal
matrix type composites) have been investigated with this technique.
This technique has been used for residual stress measurements in welded joints. Welded
joints show premature failures when they are used in corrosive environments. This type o f
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
189
failure is attributed to the service induced cyclic stresses and residual stresses. The residual
stresses in welded joints are usually measured by resistance strain gages in combination with
stress relaxation methods. These methods deduce the stress information from the elastic
strain release that occurs when either the specimen is cut into pieces or material is locally
removed like hole drilling. However, the measurements with strain gages or extensometers
are prone to error because the measurement is highly sensitive to the size and the location of
the instrumentation. Nicoletto showed that moire interferometric hole drilling method brings
out distinctive features o f the welded joints compared to conventional counterparts.
Similarly, residual stresses in composites and in metals can also be measured by Moire
Interferometry. Especially high performance composite materials are often highly
inhomogeneous and the physical properties exhibit directional dependence.
Mechanical properties o f graded layers have been studied theoretically by several
investigators. Winter et al first applied phase shifted moire interferometry to measure fine
variations in surface displacements for graded Ni-AbCb composites. To the best o f our
knowledge, no other study has been reported in the open literature on the micromechanical
behavior o f graded composites by Moire approach.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
190
1.2. The Formation of Fringe Patterns
Figure I represents the basic principle o f Moire Interferometry. As the figure shows, a double
beam optical system is used for measuring the in-plane displacements in real time. A high
frequency cross-line diffraction grating is bonded to the specimen. The specimen grating
diffracts the incident two coherent beams with certain incident angle, and in the direction
normal to the specimen two strong diffracted beams are obtained. When the specimen surface
deforms under a certain type o f loading, the optical diffraction grating also deforms with the
specimen.
The two incident beams are symmetrical with respect to the normal to the specimen gratings.
From the optics, it is well known that, overlapping beams o f coherent light gives rise to a
grating structure. Laser light is usually employed as a light source because when discharge
lamps were used, the useful area o f fringes is quite small. When the grating is illuminated by
a collimated beam o f light with a wavelength of A. at angles + a and -a, the two beams
generate a virtual grating (a constructive and destructive interference) in the zone of their
intersection as shown in Figure 1 where a is the angle o f the beams a n d /is the frequency of
the grating (usually lines/mm).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
191
Figure 1. Schematic diagram o f Moire Interferometry system.
An array o f very closely spaced bright and dark bars appears on the specimen surface when
the specimen cuts these beams. This is a closely spaced horizontal band o f constructive and
destructive interference is also called virtual gratings. The frequency o f the virtual gratings
(f) is defined by the formula given below.
f = —sin a
X
(1)
where A. is the wavelength o f light employed and a is the angle o f incidence as indicated
previously. Virtual grating is a convenient way to visualize the meaning o f fringe pattern
actually created by optical interference.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
192
When the two vertical coherent beams are used, the virtual reference grating interacts with
the horizontal set o f lines o f the deformed specimen grating and forms a moire fringe pattern.
In the absence o f specimen deformation, fringes are straight uniformly spaced and extended
throughout the field o f view. Under an applied load, the grating deforms together with the
underlying specimen and the straight fringes become curvilinear. This pattern is a contour
map and represents the displacement field in the vertical direction (x). Similarly, the virtual
reference grating created by the two horizontal coherent beams combines with the vertical set
o f lines of the deformed specimen grating and forms fringe pattern which depicts the
displacement field in the horizontal direction (y). The displacement is usually described by
scalars, U, V and W in the x, y, and z directions, respectively. O f these, U and V lie in the
original plane of the surface therefore they are called in-plane displacements. Generally
displacements in U (x) and V (y) directions fully define in-plane deformation o f the surface.
This can be achieved by using a four-beam moire system, where two orthogonal sets o f
beams are arranged. Figure 2 depicts the optical arrangement for a four beam.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
193
fin
Specimen
<3rating
N x
o r
N y
F r in g e s
\
Figure 2. Schematic drawing o f a four-beam moire interferometry used to record U (x) and
V (y) displacements.
The patterns of contour maps o f displacements governed by the relationships.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
194
Ux = — N x
f
"
,
=
J
N
y
where Ux and Uy are in plane x and y components of displacement at any point. Nx and Ny
are the fringe orders at that point in the corresponding fringe patterns. Strains can be
calculated from the displacement fields by the relationships given below.
d x
d U
f
v
d x
d N
=
d y
v
0)
a y
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
195
2. Experimental Procedure
There are two relatively independent parts in the entire testing procedure. The first part is
related to the sample preparation and the other is optical measurement. In the following
sections, these processes will be explained.
2.1. Sample Preparation
Ti-B based functionally graded composites were produced by conventional combustion
synthesis and microwave processing. The details o f the processes were explained in previous
manuscripts. FGMs consisted o f three ~ 3.75mm thick layers. The surface o f the FGMs
polished flat, cleaned with acetone and dried thoroughly before the application o f the grating.
The process used to transfer the reflective film to the specimen grating is illustrated in Figure
3. The procedure consists o f a specimen, the mold with a cross-line grating and a liquid
adhesive (PC=10C, Measurements Group, Inc., Raleigh, NC USA). The adhesive was poured
between the specimen surface and grating to replicate every detail of the grating. The
gratings were reported to be prepared by exposing a photographic plate to a virtual grating,
coating with a releasing agent (Photoflo) and aluminizing by physical vapor deposition
(PVD). In this study, we used previously prepared gratings o f 1200 lines/mm. For this
grating, the accuracy o f displacement was reported to be less than ± 80 nm. Before grating is
replicated on the specimen, the specimen should be aligned with respect to the grating lines
o f the mold. This was done by using a flat alignment bar by first aligning the bar with respect
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
196
to an alignment fixture. The specimen was pressed against the mold by placing small weights
to avoid slippage o f the sample relative to the mold surface. Also a slight weight helps the
adhesive polymerize uniformly. Otherwise, when the liquid adhesive solidifies in one
section, the remaining liquid polymerizes and shrinks later causing severe defects on the
replica surface. Excessive liquid adhesive was cleaned away by using cotton swabs. The
specimen was left overnight to let the adhesive to be cured by polymerization. The curing
was carried out at room temperature due to the fact that the adhesives tend to form channels
when the curing rate is high which eventually results in a non-uniform coating. After
polymerization, the mold was pried off carefully by using sharp razor blades. A high prying
force usually leads to delamination o f the grating.
2.2. Moire Interferometry Experiments
The test system used in this study was a four-beam argon ion laser (A.=514.5nm)
interferometer with a manually operated load frame. A photograph o f the interferometer is
shown in Figure 4. The whole testing system is placed on an isolated floating table to
eliminate any external vibrations. The samples were placed in between the parallel flat
surfaces o f a servohydraulic testing machine. The sample loading arrangements and the
fixture are shown in Figure 5. The ball bearing joint in the top fixture allowed us to align the
sample.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
197
Specimen
Metallic film
Uncured
adhesive
Uncured
adhesive
Specimen _
and grating
Cured adhesive
Metallic film
Figure 3. Steps in producing the specimen grating by replication process.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
198
An initial fringe pattern (displacement field) was first obtained under a compressive load of
0.01 kN (assumed as a null load). This recording o f undeformed (or slightly deformed)
gratings themselves is used as a reference. Then, the specimen was taken through a loading
sequence up to a pre selected maximum load then the load was returned to null load, in order
to check if the elastic limit had been exceeded or not. The exact loading sequences were
explained in the Results and Discussion part. When the specimen surface deformed, the
previously coated optical grating also deformed with the specimen. Under each constant load
stage, the two coherent laser beams diffracted from the grating in the normal direction and
generated feature interferometry pattern that represents the in-plane displacement
distribution. The concurrent four-beam laser system applies to both horizontal and vertical
direction by using two-laser beam for each direction. Therefore, deformation information in
two perpendicular directions was obtained. The fringe pattern generated by the two vertical
laser beams represents the vertical deformation field and the fringe pattern generated by two
horizontal laser beams represents the horizontal deformation field (V field and U field,
respectively in our notation). For each set o f experiment, the null field values were subtracted
from a specific load field values. The results gave the net displacement from the applied load.
Phase shifting was accomplished by applying a series o f increasing voltages to a
piezoelectric transducer coupled to an optical fiber. The data extraction was carried out by
analyzing both the fringes o f undeformed and deformed gratings individually. These fringes
yield to the maps of the displacements.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 4. Photograph of the Moire Interferometer used in our tests.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
200
The fringe images were captured by a CCD camera. The camera was located behind the
sample and the distance between the camera and the sample was adjusted to ensure proper
field o f view. The images from the camera were acquired by a frame grabber, which digitized
the field o f view into 640 x 480 pixel array. These images are grayscale representations of
matrices, from 0 to 255.
The analysis o f the reference gratings and deformed gratings was
carried out by using a phase unwrapping algorithm and converted to a useful displacement
format. The software to unwrap the phase maps and construct a pixel-by-pixel representation
o f specimen surface displacements was developed at INEEL. The unwrapping program also
reduces surface defects and optical noise. The details of the unwrapping algorithm can be
found elsewhere [26]. In simple terms, the fringes are generated by the algorithm and stored
as a digital array o f numbers. Each element o f the array given either a binary 0 (a dark point)
or a binary 1 (a bright point). A fringe can therefore be built up o f a regular pattern o f zeroes
and ones. An array o f continuous zeros would produce the dark line o f the Singe pattern and
in a similar way an array o f continuous ones give the bright line. One point worth noting here
that, the method requires a good fringe contrast. Therefore, the images are stored in HDF
(hierarchical data format) file format that allows for information to be associated and retained
with the image. In the unwrapped image, origin and length scales were specified. Length
scales were based on our initial fiducials made on the gratings indicating the approximate
locations o f the interlayers. A java-based tool (Java HDF Viewer JHV) was used for easily
viewing the contents o f HDF files. This software is developed by the National Center for
Supercomputing Applications (NCSA) and freely available to public from the NCSA - HDF
homepage (http://hdf.ncsa.uiuc.edu). Data extraction was performed with the data analysis
package Noesys™ (version 2.0, Research Systems Inc., Boulder, CO).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Ball Bearing
Top Punch
Sample
Bottom Punch
Figure 5. Schematic o f the fixture used for compression testing.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
202
A typical FGM sample geometry used in the measurements is shown in Figure 4. The data
were extracted in the direction o f the load (V field) and at 3 positions along the width o f the
sample for all samples. These locations are; left (L), middle (M) and right (R). Displacement
field under a constant load were computed by measuring the distance between the two
neighboring fringe centers and plotted as a function o f the sample length. The analysis was
based on these plots. One position (either left, middle or right) was chosen based on the
linearity o f the curves. The strain measurements were computed by measuring the derivative
o f the displacement field.
Load
I
IV
R
95Ti-5B
-f
T
*
h
?
♦
+ U (x)
Fixed end
Figure 6. Specimen geometry and loading configuration under compression.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
203
3. Results and Discussions
The results of Moire Interferometry tests are presented in this section. Data extraction and
manipulation were repeated for each group of sample (both conventional combustion
synthesized and microwave activated) as explained in the experimental section. In the
presentation o f the results, first, displacement versus sample length plot at the maximum load
will be given. This data were traced from the Moire fringes. These plots represent how the
displacement varies on different sections of the sample cross section depending upon loading
conditions. Subsequently, the changes in the displacements under different load stages are
shown. This is followed by the strain plots analysis. Elastic modulus o f each layer was
calculated from stress-strain plots.
3.1. Compression Test Results of CS Produced FGMs
The FGM samples produced by conventional combustion synthesis machined to the
dimensions and geometry shown in Figure 7. The load sequence applied to the sample is
shown in Table I. After each load stage, the moire fringes were captured. In Figure 8 and
Figure 9, the images obtained from null state (0.01 kN) and maximum load (0.65 kN) are
presented in wrapped and unwrapped forms, respectively. Figure 8a is considered as a
reference pattern, which represents undeformed state o f the sample. On the other hand,
Figure 8b is the pattern o f the deformed sample under the maximum applied load. The
complete series of representative wrapped fringe patterns obtained from all load stages are
given in Appendix I. These contour plots show a number o f fringes with a good fringe
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
204
contrast. These grayscale images are a representation o f the generated displacement due to
the mechanical deformation.
8.58 mm
A
4.9 mm
S S 85Ti-15B
12.3 mm
A
llllllllllllll
95Ti-5B
mmiiiimi
4.3 mm
A
3.1 mm
Ti
A 3.44mm
Figure 7. A schematic o f the test sample produced by combustion synthesis after machining.
The system is very sensitive to contact conditions between the punch and sample. Usually the
applied load was not evenly distributed across the samples. Usually, it was biased towards
one side o f the sample. As shown in Figure 8b, the right bottom side o f the sample exhibits
slightly more fringes indicating that this side o f the sample experienced larger stresses.
Figures 9 a and b show the corresponding unwrapped images under the null load and the
maximum applied load, respectively. These images can be considered a translation o f the
moire fringe patterns into displacement plots. The origin and length scales were also
indicated on the images. On a careful examination, it can be identified that the displacement
bars underneath the images displays the deformation (compression) in the form of a contour
bars. The bars indicate that darker regions correspond to the highly deformed sections. It is
clear that the right bottom side of the sample has undergone a higher compression
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
205
deformation. As indicated, this was also observed in Figure 8b that there is a slight
unsymmetry in the fringe order distribution on this section of the sample.
In Table I., the computed stress values were given assuming that a uniform distribution o f the
load had occurred. In these calculations, surface area for the test sample was calculated to be
2.952x10*5 mm2.
Table I. Force-pressure conversions for tested combustion synthesized FGM
Force (kN)
M Pa
0.01
0.338
0.15
5.08
0.25
8.46
0.35
11.85
0.40
13.55
0.50
16.93
0.55
18.63
0.65
22.01
Under compression loading, displacements were plotted as a function o f sample length based
on the fringe patterns. In Figure 10, the V(y) field displacements as a function of distance
across the sample are shown. The displacements are shown along the left, middle and right of
the sample at a load o f 0.65 kN. Attention was focused on the right side o f the sample
because the largest displacements occurred on this section. The figure clearly indicates that, a
change in slope of the curves occurs through the graded layers. This is an expected result due
to the fact that the composition changes along V(y) direction.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
206
(a)
(b)
Figure 8. Wrapped images of displacement under a null load o f 0.01 kN (a) and under a
maximum load o f 0.65 kN (b)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
-s.a
- 2 .s
o .o
2 .5
s.a
-5.0
-2 .5
0 .0
2.5
5 .0
11T
t
T
-0 .0 0 0 2
0 ,0 0 0 0
Displacement
(a)
0.0C 02
- 0 . 0Q 37S -0 . 0 0 2 S G -0.0 0 1 2 5 0 . OOQQO
Displacement
(b)
Figure 9. Unwrapped images o f displacement under a null load o f 0.01 kN (a) and under a
maximum load of 0.65 kN (b)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
208
The displacement —sample length curve was divided into three sections. Each section with a
constant slope represents different layers (Ti, 95Ti-5B, and 85Ti-15B). The interface
locations were estimated from the fiducials marked on the gratings and recorded in wrapped
images.
A trend line was fitted to the displacement data for each individual layer. The slope o f the
linear fit (first order derivative o f the polynomial that defines the line) yields strain data
along the right side at a particular single load stage. For each section, a linear fit was
appropriate. By examining the displacement and strain plots, an understanding o f the
mechanical properties can be gained for each individual layer. It should be noted here that
regardless of the method used to calculate the derivatives, the accuracy o f the calculated
strains is less than that o f the measured displacements. The reason for this is that the
differentiation of experimental data always results in some uncertainty. However, in our
tests, the displacement fields provided abundance of data points, therefore, the strain data
was determined satisfactorily. Figure 11 shows the results o f this calculation. It can be seen
that as the compression load increases the displacements and corresponding strain values also
increases. The transition from titanium layer to titanium matrix boride layers showed some
difference in displacements. The change is more pronounced in titanium layer. It changed
from about -5x1 O'4 mm to —4xl0'3 mm as the pressure increased from 0.15 kN to 0.65 kN.
On the other hand same increment in pressure created a less change in 85Ti-15B. The higher
boron containing layers were expected to show a more stiff behavior than titanium layer.
However, the influence o f adjacent layers on the strain values seems insignificant. For all
loading stages, titanium and 85Ti-15B layers exhibited approximately equal strain values and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
209
layer 95Ti-5B showed a slightly less strain compared to those adjacent layers. There might
be two reasons related to this phenomenon. First, as X ray studies indicated, the relative
amount o f boride phases within the layers is small. Titanium remains as a continuous matrix
along the interfaces. This indicates the effect o f matrix phase on the micromechanical
behavior o f the composite. Second, the porosity created in the higher boron content layers
during the processing might also be an additional factor contributing this compliant behavior.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
210
95Ti-5B
\
► U(x)
Fixed end
1.0000E-03
left
middle
-1.0000E-03
right
-2.0000E-03
-3.0000E-03
-4.0000E-03
85Ti-15B
95Ti-5B
-5.0000E-03
0
1
2
3
4
5
6
7
8
9
10
Sam ple L ength (m m)
Figure 10. Displacement (V) as a function of distance along the left, right and middle o f the
CS produced FGM sample under a compression load o f 0.65 kN (Null load subtracted out).
Vertical dashed lines indicate approximate locations o f the interfaces.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
211
1 .0 0 0 0 E -0 3
£
£ - 5 .0 0 0 0 E -0 4
so
£
0.15 kN
0.25 kN
0.40 kN
0.55 kN
0.65 kN
- 2 .0 0 0 0 E -0 3 -
oo
J2 - 3 .5 0 0 0 E -0 3 D.
.a
Q
85T -15B
95Ti-5B
- 5 .0 0 0 0 E -0 3
Sampie Length (mm)
O.OOOE+OO
_
-1 .0 0 0 E -0 4
jj
-2 .0 0 0 E -0 4
£
-3 .0 0 0 E -0 4
•I
- 4 .0 0 0 E -0 4
0.15 kN
0.25 kN
0.4p kN
0.5b kN
0.65 kN
is
™
-5 .0 0 0 E -0 4
85Ti-15B
-6 .0 0 0 E -0 4
0.0
2.5
9 5 T i-5 B
5.0
7.5
10.0
Sample Length (mm)
Figure 11. Compression data o f displacement (V) and strain as a function of sample length
along the right side o f the FGM. Vertical dashes indicate approximate locations of the
interfaces.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
212
25 i
85Ti-15B
20
Stress (M Pa)
*
95Ti-5B
15
10
E-modulus:
Ti layer : 65.9 GPa (y int: 0.383 MPa)
95Ti-5B : 50.5 GPa (y int: 0.163 MPa)
85Ti-15B: 55.6 GPa (y-int: 1.338 MPa)
0
0.0E+00
-1.0E-04
-2.0E-04
-3.0E-04
-4.0E-04
-5.0E-04
Strain (mm/mm)
Figure 12. Stress -strain curves for three-layered FGM produced by combustion synthesis.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
213
3.2. Compression Test Results of MW Produced FGMs
The FGM samples produced by microwave processing machined to a rectangular prism and
the dimensions are shown in Figure 13. The same test procedure was repeated. In the forcepressure conversions the surface area o f the FGM was calculated 1.653x10'5 m2 . Due to a
rather peculiar phenomenon occurred during the tests o f M W produced FGMs, we will
present the results in two sections. First sets of tests were carried out up to a pressure o f 0.37
kN. Then the sample unloaded and re loaded up to 0.60 kN. The complete loading sequence
is given in Table n.
In Figure 14 and Figure 15, the images obtained from null state (0.01 kN) and maximum load
(0.37 kN) are presented in wrapped and unwrapped forms, respectively. Figure 14a is the
null pattern showing undeformed state o f the sample; Figure 14b is the pattern o f the
deformed sample under a pressure o f 0.37 kN. It is noticeable that MW produced FGMs
produced a large number o f deformation fringes compared to CS produced FGMs even under
low pressure. This indicates their more compliant behavior. In fact when the sample
unloaded to null state, some residual fringes remained indicating a plastic deformation. The
complete series of representative wrapped fringe patterns obtained from all load stages are
given in Appendix H.
In order to avoid redundancy a detailed explanation o f the behavior of this material will be
explained in the following reloading section. Here we only present the results obtained in this
low pressure loading stage.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
214
|^Tv"
11.27 mm
8.79 mm
^
s \\\\\V
S
w S 90Ti-10B
k \\ \ \
\V^
llllllllllllll
95Ti-5B
4.3 mm
3 .65 mm
linn
Ti
3.32 mm
-4*1.88mm
Figure 13. A schematic of the test sample produced by microwave activated combustion
synthesis after machining.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
215
Table EL Force-pressure conversions for tested microwave activated FGM
Force (kN)
MPa
0.01
0.605
0.10
6.05
0.20
12.09
0.25
15.12
0.35
21.17
0.37
22.38
Re-loading
0.20
12.09
0.30
18.15
0.40
24.19
0.50
30.24
0.60
36.29
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(a)
(b)
Figure 14. Wrapped images o f displacement under a null load o f 0.01 kN (a) and under a
maximum load of 0.37 kN (b)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
-4
-2
0
2
4
6
m ______
- 0 . 0 0 S « a OO&OS 0 0 0 0 0 0 0 0 X 5 0 0 0 -5 0 COO 7 5
Disp iacerssent
(a)
i-----------1-------- t
i
i
- C . 0 0 7 5 - 0 ,0 0 S 6 -0 .0 0 2 5 0 .0 0 0 0 0 .0 0 2 5
Displacewent;
(b)
Figure 15. Unwrapped images of displacement under a null load o f 0.02 kN (a) and under a
maximum load of 0.37 kN (b)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
218
Load
90Ti-10B
V(y)
j j s jj
I'
95Ti-5B
JL
'
41
Ti
I '
XL_
-► U(x)
T
Fixed end
5 .0 0 0 0 E -0 3
^
cr-«
2 .5 0 0 0 E -0 3
left
middle
'T ' O.OOOOE+OO
>
§
£o
right
-2 .5 0 0 0 E -0 3
J -5 .0 0 0 0 E -0 3
a.
.as
O -7 .5 0 0 0 E -0 3
95Ti-5B
90Ti-10B
-1 .0 0 0 0 E -0 2
0
1
2
3
4
5
6
7
8
9
10
Sample Length (mm)
Figure 16. Displacement (V) as a function o f distance along the left, right and middle of the
MW produced FGM sample under a compression load o f 0.37 kN (Null load subtracted out).
Vertical dashed lines indicate approximate locations o f the interfaces.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
219
B
B
c
2.500E-03
O.OOOE+OO
0.10 kN
-2.500E-03
<D
Ba -5.000E-03
o
•S
n. -7.500E-03 in
Q -l.OOOE-02
0.20 kN
0.25 kN
9 0 T -1 0 B
2
0.37 kN
95Ti-5B
3
4
5
6
10
7
Sample Length (mm)
5.00E-05
|
-2.00E-04
0.10 kN
-4.50E-04
0.20 kN
0.25 kN
0.37 kN
B
'^rc ' -7.00E-04
2
C/3
-9.50E-04
9 0 T i-1 0 B
1
9 5 T -5 B
-1.20E-03
0
2.5
5
7.5
10
Sample Length (mm)
Figure 17. Compression data o f displacement (V) and strain as a function o f sample length
along the right side o f the FGM. Vertical dashes indicate approximate locations of the
interfaces.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
220
25
95Ti-5 B
20
-
Stress (MPa)
90Ti-10B
E-modulus:
Ti layer : 37.5 GPa (y-int: 1.88)
95Ti-5B : 29.5 GPa (y-int: 0.884)
90Ti-10B : 26.1 GPa (y-int: 1.411)
0.0E+00
-2.0E-04
-4.0E-04
-6.0E-04
-8.0E-04
-1.0E-03
Strain (mm/mm)
Figure 18. Stress-strain curves obtained during the first loading stage for three layers of
FGMs produced by microwaves.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
22 L
3.2.1. Re-loading Under Compression
Figure 19 a depicts the remnant fringe pattern after the first test indicating the plastic
deformation. Figure 19 b shows the fringe pattern at a load of 0.60 kN. As observed in
previous tests, the fringe patterns skewed right side o f the sample as a result o f rotation of the
sample. This was a dominant feature in all our tests. This problem is difficult to control and
often encountered in this type of loading regime. A minor rotation o f the sample introduces
more fringes o f that particular side o f the sample. However, this does not significantly
influence the results.
When compared to CS processed FGMs (Figure 8b), microwave processed FGMs showed a
higher density o f fringes which indicates a more compliant behavior. Similar to previous
tests, displacement- sample length profiles indicated three distinct slopes as shown in Figure
20. Each slope corresponds to a different layer in FGM sample. It is evident that as the load
increases the displacements also increases as Figure 21 indicates. Deformation
(displacement) was observed to be linear as the composition changes from one layer to
another. In terms o f deformation, microwave processed FGMs exhibited a different
displacement behavior than conventional CS produced FGMs. With a careful examination it
can be seen that, with CS produced FGMs (Figure 10) as the composition changes from
95Ti-5B to 85Ti-15B the displacement curve deviates from the linearity. This is due to the
wide change in composition. However, in microwave processed FGMs since the composition
change is less drastic the displacement curves stayed linear.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
222
The average strains across individual layers were determined for each loading condition by
overlaying a linear fit on each section. The layers showed linearly changing strains. As
expected the strain values increased as the applied pressure increased. However similar to CS
produced FGMs, the effect o f adjacent layers is not significant. Higher boron content layers
showed a more compliant behavior. It is also important to note that for all compositions and
load stages, the strain values obtained from microwave produced FGMs are higher compared
to conventional CS produced FGMs. This can also be seen in the fringe patterns shown in
Appendix II. This occurred even at low loads and resulted in plastic deformation. As a
consequence of this, average elastic moduli values are lower. The average elastic moduli
values for each layer are shown in Figure 22. These values are smaller compared to the
corresponding layers o f combustion synthesized FGMs.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
223
Figure 19. Wrapped images of displacement after unloaded from 0.37 lcN (a) and under a
maximum load of 0.60 kN (b)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
224
8.0000E -03
4.0000E -03
left
C
middle
right
>
eo
e
.OOOOE-03
.0000E-03
.2000E -02
90Ti-10B
95Ti-5B
.6000E -02
0
1
2
3
4
6
7
8
9
10
S am p le L e n g th (m m )
Figure 20. Displacement (V) as a function o f distance along the left, right and middle o f the
MW produced FGM sample under a compression load o f 0.60 kN (Null load subtracted out).
Vertical dashed lines indicate approximate locations o f the interfaces.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
225
4.00E-03
O.OOE+OO
sO
g
eg
o.
q
-4.00E-03 4
0.30 kN
-8.00E-03 -|
0.40 kN
-1.20E-02
0.50 kN
90T1-10B
•1.60E-02
1
2
0.60 kN
95T i-5B
3
4
5
6
7
8
9
10
Sample Length (mm)
0
2.5
5
7.5
10
Sample Length (mm)
Figure 21. Compression data of displacement (V) and strain as a function of sample length
along the right side o f the FGM. Vertical dashes indicate approximate locations o f the
interfaces.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
226
40
95Ti-:
Stress (M Pa)
30 -
20
10
0
E-modulus;
Ti layer 41.79 GPa (y-int: 2.85 MPa)
95Ti-5B 33.57 GPa (y-int: 2.74 MPa)
90Ti-10B 35.21 GPa (y-int: 6.86 MPa)
-
t
-2.0E-04
i
l l
i' —i—■■ .
-4.0E-04
.
i
|
-i
i
-6.0E-04
-8.0E-04
--- 1--- r-
-1.0E-03
i
1----1---- 1----r-
-1.2E-03
-1.4E-03
Strain (m m /m m )
Figure 22. Stress-strain curves o f a three layered FGM produced by microwaves.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
227
4. Conclusions and Recommended Future Work
Evaluation o f elastic/plastic deformation o f Ti-B binary-based functionally graded
composites with phase shifted moire interferometiy has resulted in several important
conclusions. The technique has been demonstrated for examining the deformation o f this
type o f composites successfully. The Moire technique uncovered the influence o f a
compositional gradient on deformation. All samples showed smooth uniform deformation
behavior with biased loading which is a common observation for compression loading.
Some differences between combustion synthesized FGMs and microwave activated
combustion synthesized FGMs were detected. Comparison o f the results indicated that
microwave processed samples exhibited a more compliant behavior and plastically
strained under lower loads. Hence, they exhibited lower elastic modulus values. Also
after re loading they exhibited an increased elastic modulus indicating a strain hardening
behavior. While further studies are needed to address this issue, we believe that the
measured strain and elastic modulus depends on the microstructure. Pore collapse during
the uniaxial compressive loading was first thought as a reason o f higher strains.
Although, volume percentage of porosity in the layers remains approximately the same,
the pore sizes were smaller in microwave-produced samples. Interestingly, they exhibited
higher displacements. Although, we believe that pores are undoubtedly play an important
role in the strain response o f test specimens, in our tests the results contradicted with our
hypothesis. The previously described microstructure o f microwave activated samples
showed a uniform second phase distribution in titanium matrix. Due to lack o f exact
temperature profiles with the microwave experiments, we do not know the exact
temperature history o f these samples. However, we are certain that the samples were
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
228
exposed to high temperatures extended times compared to conventional combustion
synthesized samples. This might very likely resulted in a refined microstructure. This
suggests to us that the effect o f a uniform metal matrix phase on the displacements might
actually be responsible for higher displacements. The research can be extended a step
forward and combustion synthesized/compacted samples can be tested with this
technique as well. This will provide an additional understanding on the effect o f
microstructure.
The displacement fields from the perpendicular to the loading directions (U field) should
be measured. This, combined with the primary direction (V) measurements will provide
biaxial strain and Poisson’s ration can be calculated.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
229
APPENDIX I
MOIRE FRINGE PATTERNS
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
230
0.25 kN
0.35 kN
Figure 1. Compression test wrapped fringe patterns
sample FGM #10 1-21-01A
(V field, 1200 line grating)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
231
0.4 kN
0.65 kN
Figure 2. Compression test wrapped fringe patterns
sample FGM #10 1-21-01A
(V field, 1200 line grating)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
232
APPENDIX II
MOIRE FRINGE PATTERNS
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
233
MM
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
234
0.35 kN
0.02 kN
Figure 1. Compression test wrapped
fringe patterns
sample FGM #9 A 1-23-01
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
235
0.4 kN
0.5 kN
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
0 .6 0 k N
U
0 .0 2 k N
Figure 2. Compression test wrapped
fringe patterns during re-loading
sample FGM #9 A 1-23-01
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
237
5. References
1.
D. Post, B. Han, and P. Ifju , High Sensitivity Moire-Experimental Analysis for
Mechanics and Materials, Springer-Verlag, (1994).
2.
F. Dai, J. McKelvie and D. Post, Optics and Lasers in Engineering, 12, 101-118, (1990)
3.
A. Asundi, Exp. Mech., 9, (1994), 230-241.
4.
C.A. Walker, Exp. Mech., 12, (1994), 281-299.
5.
J. McKelvie, P.M. MacKenzie, A. McDonach, and C.A. Walker, Exp. Mech., 12,
(1993), 320-325,
6.
Y.Z. Dai and F.P.Chiang, Exp. Mech., 3, (1991), 76-81.
7.
D. Joh, K.Y. Byun and J. Ha, Exp.Mech. 3, (1993), 70-76.
8.
D. Post, R. Czamek, D. Joh, J. Jo and Y. Guo, Exp. Mech., 6,(1987),90-194.
9.
W.G. Gottenberg, Exp. Mech., 9, (1968), 405-410.
10. W. N. Sharpe Jr., Exp. Mech., 3, 62-67, (1992).
11. B. Han, P. Ifju and D. Post, Exp. Mech., 9, (1993), 195-200.
12. D. Post, R. Czamek and D. Joh, Exp. Mech., 9, (1985), 282-287.
13. B. Han, D. Columbus, Z. Wu and J. Lu, Exp. Tech., 1-2, (1999), 16-19
14.
C. Y. Poon, M. Kujawinska and C. Ruiz, Exp. Mech., 9, (1993), 234-241.
15.
E.D. Steffler and A.N. Winter, Proc.ofthe Soc. of Exp. Mech., (1999), 665-670.
16. F. P. Chiang, V.J. Parks, and A.J. Durelli, Exp. Mech., 9, (1968), 554-560.
17. T. W. Shield, and K.S. Kim, Exp. Mech., 6, (1991), 126-134.
18.
H.M. Hsiao, I.M. Daniel, and R. D. Cordes, Exp. Mech., 9, (1998), 172-179.
19. J. L. Sullivan, Exp. Mech., 12, (1991), 373-381.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
238
20.
K. Arakawa, R.H. Drinnon Jr., M. Kosai, and A.S. Kobayashi, Exp. Mech., 12, (1991),
306-309.
21.
G. Cloud, and S. Paleebut, Exp. Mech., 9, (1992), 273-281.
22.
D. Post, Exp. Mech., 9, (1991), 276-280.
23. B. Han, and D. Post, Exp. Mech., 3, (1992), 38-41.
24. K.E. Perry Jr., and J. McKelvie, Exp Mech., 3, (1996), 55-63.
25. C.J. Jenkins, Exp. Mech., 7, (1968), 331-332.
26. A.N. Winter, PhD Thesis, Colorado School o f Mines,(1999).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Документ
Категория
Без категории
Просмотров
0
Размер файла
9 694 Кб
Теги
sdewsdweddes
1/--страниц
Пожаловаться на содержимое документа