This is to certify that the
dissertation entitled
Fabrication of Functionally Gradient Materials with

Internal Channels in Ceramics and Ceramic Composites

presented by

Hyea-Weon Shin

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in Materials Science
and Engineering

 

 

WW

/ Major professor

Date £22]; £44m;

MSU is an Affirmative Action/Equal Opportunity Institution 0-12771

 

 

.UBRARY
Michigan State
University

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:lCIRC/DateDue.p65-p.15

 

FABRICATION OF FUNCTIONALLY GRADIENT MATERIALS WITH

INTERNAL CHANNELS IN CERAMICS AND CERAMIC COMPOSITES

Hyea-Weon Shin

A DISSERTATION

Submitted to
Michigan State University
In partial fulfillment of the requirements

for the degree of

DOCTOR OF PHILOSOPHY

Department of Chemical Engineering and Materials Science

2002

ABSTRACT

FABRICATION 0F FUNCTIONALLY GRADIENT MATERIALS WITH
INTERNAL CHANNELS IN CERAMICS AND CERAMIC COMPOSITES
By

HYEA-WEON SHIN

Functionally Gradient Materials (FGMs) are inhomogeneous materials whose
compositions vary from one phase to another. By tailoring the inhomogeneous properties,
FGMS can be used to reduce the stresses that are caused by severe thermal gradients.
Thermal gradient loading can further be compensated by heat transfer into a cooling fluid
circulating in a network of channels and manifolds. In an envisioned application, heat
from a localized source is transferred to the cooling fluid, easing sharp thermal loads
while minimizing the unwanted spread of heat energy to the ambient surroundings.

This study reports on the fabrication of functionally gradient ceramics and the
embedding of simple internal channels within these ceramics. Functional gradiency
(variation of composition) is built in via the layering of different components across the
thickness of a plate sample. Traditional powder processing techniques are applied to
fabricate the test pieces, and recently developed methods of joining are used to build
assemblies from individually sintered plate layers.

For a well-formed FGM to be made, materials parameters need to be selected
based on mechanical, thermal and chemical properties. As a class, ceramics are hard,
wear-resistant, refractory, electrically and thermally insulative, nonmagnetic, chemically

stable, and oxidation-resistant. However, because of their brittleness, ceramics with

minute channels are diflicult to machine. Instead, for this study, a graphite fugitive phase
is used as a spacer to support channel volumes within a ceramic powder compact; during
pre—sintering, the graphite burns out to expose a network of channels. Full sintering fixes
the final shape. At the operating temperatures of the ovens used in our fabrication study,
sintering of alumina, partially stabilized zirconia, fully stabilized zirconia and
hydroxyapatite have been successful, and these ceramic powders form the basis of the
present fabrication studies.

Inhomogeneities inherent in the composition of layered FGMS and non-
uniformities introduced in the fabrication process (possibly during powder compaction)
give rise to non-uniform densification behavior, which leads to warping, cracking, and
the development of residual stress both within the layers and along layer boundaries. As
a remedy, powders of different particle size are mixed within component layers, and the
effect of particle mixture on final specimen curvature is shown. The optimal mixture ratio
gives a flat layer, wherein the effects of thermal expansion and densification are balanced.
Different pressing procedures and layer-stacking orders have also been used to minimize
warping; these are viewed as optional, complementary steps to particle mixing.

A study of the difference between conventional and microwave heating is also
presented. The main interest has been to quantify the different sintering behavior
associated with conventional heating and microwave heating.

Finally, scanning electron microscopy and energy dispersive spectroscopy are
used to illustrate the gradual changes in element composition across the thickness of the

produced functionally gradient ceramic plates.

DEDICATION

To

MY PARENTS

WHO LOVED ME AND SUPPORTED ME DURING THE PAST

YEARS

iv

ACKNOWLEDGEMENTS

I would like to express my deep gratitude to my advisor, Dr. Patrick Kwon,
and co-advisor, Dr. Eldon Case for their instruction, guidance and encouragement

to complete this research.

I also wish to thank the members of my committee, Dr. Dashin Liu,
Department of mechanical Engineering, and Stanley L. Flegler for their advice,

assistance and valuable comments on this research.

Special thanks are extended to my parents for their patient support and

encouragement during entire graduate study.

TABLE OF CONTENTS

LIST OF TABLES ........................................................................... ix
LIST OF FIGURES ........................................................................ xi
CHAPTER 1
INTRODUCTION ............................................................................ 1
CHAPTER 2
BACKGROUND ............................................................................. 7
2.1 FGM. ............................................................................... 7
2.2 Joining. ............................................................................. 18
2.2.1 Joining Techniques. ................................................... 18
221-1 Mechanical Attachment. ........................... 18
2.2. 1-2 Direct Bonding Process. ................................. 19
CHAPTER 3
EXPERIMENTAL PROCEDURE ......................................................... 33
3.1 Materials. ......................................................................... 33
3 .2 Powder preparation. .................................... . ......................... 43
3.3 Specimens preparation. ....................................................... 45
3.3.1 Specimen Pressing. ................................................... 45
3.3.2 Presintering. ............................................................ 46
3.3.3 Conventional Sintering for Final Sintering. ....................... 46
3.3.4 Microwave Sintering Processing. ................................... 51
3.3.5 Sectioning the Sintered Specimen. ................................ 57
3.3.6 Specimen Polishing. .................................................. 57
3.3.7 Coating procedure. ................................................... 59
3.3.8 Joining procedure. .................................................... 59
3.3.9 Specimen Characterization. .......................................... 63
3.3.9-l To Compare the Degree of Warping. ................... 63
3.3.9-2 SEM Observation. ......................................... 64
CHAPTER 4

FABRICATION OF INTERNAL CHANNELS IN CERAMICS AND
CERAMIC COMPOSITES

vi

4.1 Introduction. ....................................................................... 70

4.2 Experimental Procedure. ........................................................ 73
4.3 Results and Discussion. ...................................................... 78
4.4 Summary and Conclusion. ...................................................... 94

CHAPTER 5

NOVEL POWDER PROCESSING TECHNIQUES TO F ABRICATE

EFFICIENT MESO-SCALE HEAT EXCHANGER ..................................... 99
5.1 Introduction. ....................................................................... 100
5.2 Background. ....................................................................... 102
5.3 Conclusion. ........................................................................ 120

CHAPTER 6

JOINING OF BIOACTIVE AND BIOINERT CERAMICS ........................... 124
6.1 Materials and preparation. ...................................................... 126
6.2 Results and Discussion. ......................................................... 130

CHAPTER 7

COMPARISON OF MICROWAVE SINTERING AND CONVENTIONAL

SINTERING .................................................................................. 142
CHAPTER 8
MINIMIZING WARPING , CRACKING AND RESIDUAL STRESS ............... 151
8.1 Warping and Cracking. .......................................................... 151
8.1.1 Initial Observation. ................................................... 151
8.1.2 Variation of Pre-Sintering Temperatures to Minimizing
Cracking. .............................................................. 152
8.1.3 Use of FGM/Variation of Layer Composition to
Minimize Warping. ................................................ 153
8.2 Residual Stress. .................................................................... 163
CHAPTER 9
SCANNING ELECTRON MICROSCOPY (SEM) ..................................... 165
9.1 SEM images of Topography Based on Secondary Electrons. .............. 165
9.2 EDS Images of Element Composition Based on Released X-Rays. ...... 166
CHAPTER 10
CONCLUSIONS ............................................................................. 175

vii

APPENDICES
APPENDIX 1. Micrographs of Fracture Surface of Materials Used in This

Study. . 179
APPENDIX II. Micrographs of Various Shapes of Manifold Channels. ............. 185
APPENDIX III. Micrographs of Various Channels in Homogeneous

Materials and Composite Materials. ......................................................... 190
APPENDIX IV. Micrographs of Various Functionally Gradient Materials. ........ 195

APPENDIX V. Element of Functionally Gradient Materials. ......................... 199

viii

LIST OF TABLES

Table 3.1. Alumina (TM-DAR, Taimi Co, Japan) powder information. .............. 37
Table 3.2. Hydroxyapatite powder information. ....................................... 37
Table 3.3. TOSOH Zirconia powder information. ....................................... 38
Table 3.4. CERAC Zirconia powder information. .................................... 38
Table 3.5. MaCor Information. ............................................................. 39
Table 3.6. Thermal expansion coefficients for the materials included in

this thesis. ...................................................................................... 39
Table 3.7. Graphite Information. ....................................................... 41
Table 3.8. Thermal Conductivities of the Materials included in this thesis. .......... 42
Table 3.9. Young’s Modulus for the Materials Included in this Thesis. ............ 42
Table 3.10. Powder size used in this thesis. ............................................... 42
Table 3.11. Summary of mixing ratios of powders. ..................................... 44
Table 3.12. The conditions of conventional sintering. .................................. 53
Table 3.13. The different conditions of microwave sintering. .......................... 60
Table 3.14. The conditions of polishing which is done for this study (time,

diamond grit Size and materials used). ...................................................... 60
Table 3.15. The condition of silicafilm coating. .......................................... 61
Table 3.16. The comparison of joining comparison of joining conditions

between microwave and conventional furnace. .......................................... 63
Table 4.1: List of Powders used. ............................................................ 74
Table 4.2. Initial Powder Mass for each specimen. ...................................... 76

ix

Table 4.3. Dimension differences of the Specimen (in cm). ............................. 80

Table 5.1. Dimensional changes during processing (Units are in cm). ............... 1 10
Table 5.2. Height difference of the five samples on figure 5.5. ...................... 111
Table 6.1. Summary of Powder Average Size. ........................................... 128
Table 6.2. For the specimens included in this study, the furnace type,
sintering temperatures, sintering times and channel descriptions for
channels fabricated without machining. ................................................... 128
Table 7.1. Summary of comparison of sintering using conventional and
microwave means. ............................................................................ 147
Table 8.1(a). Size Comparison of Zirconia Layers of Different Mixtures
after pre-sintering at 900°C for 4 hours (cm). ............................................. 159
Table 8.1(b). Size Comparison of Zirconia Layers of Different Mixtures
after final sintering at 1430°C for 4 hours (left column lists temperatures
at which the specimens were pre-sintered) (cm). ......................................... 160
Table 8.2. Minimizing Warping by Mixing Two ZrOz Powders in a Top
Layer. ........................................................................................... 161

LIST OF FIGURES

Figure 2.1. Application of FGM. ........................................................... 12
Figure 2.2. Schematic figure of Functionally Gradient Material. ..................... 12

Figure 2.3. The difference between homogeneous material and
fimctionally gradient material. ............................................................... 13

Figure 2.4. Schematic figures for sintering faults (a) ideal FGM, (b)
stepwise figures of FGM, (c) warping, (d) frustum formation, (e) splitting

and (f) cracking. .............................................................................. 16
Figure 2.5. Procedure of P /M. .............................................................. 17
Figure 2.6. Schematic drawing for active brazing process [32]. .i ...................... 23
Figure 2.7. Diffusion bonding Techniques [43]. ......................................... 24
Figure 2.8. Schematic figure of illustrating the multiple interlayers to join

the ceramic to metal [31]. ............................................................... 25
Figure 2.9. Schematic drawing of the reaction of diffusion bonding on

bond interface. (a) contact, (b)deformation and interfacial boundary

formation, (c)grain boundary migration and elimination and ((1) volume

diffusion and pore elimination. .............................................................. 26
Figure 3.1. Schematic drawing of “Carver Laboratory press”. ........................ 47
Figure 3.2. SEM picture of Pencil lead (a) one end of pencil lead, (b) the

surface of pencil lead. ........................................................................ 48
Figure 3.3. Overview of specimen processing (a) one layer embedded

fugitive phases and fugitive phases on the top surface and (b) two layers

embedded fugitive phases and on the top surface (0) dimension of the

specimen. ....................................................................................... 49

Figure 3.4. Configuration of FGMS specimens with internal channels and
embedded channels (a) one internal channels and channels on surface and
(b) two different levels of embedded channels and channels on surface in

xi

one homogeneous material (c) two levels of embedded channels and
channels on surface in FGMS Specimens.

Figure 3.5. Schematic figure of specimens on the setter. ............................... 52
Figure 3.6. Conventional high temperature furnace. ..................................... 52

Figure 3.7. Photograph of microwave cavity and power supply used in
this study.

Figure 3.8. (a) Microwave cavity and (b) the schematics of microwave
processing unit.

Figure 3.9. Schematic showing casket with the position of specimens. .............. 56
Figure 3.10. The schematic drawing for process of spin-coating. ..................... 62
Figure 3.11. Picture of Magnetic Base with Dial Indicator. ........................... 65

Figure 4.1. A SEM micrograph of an alumina specimen that was pre-
sintered at 900°C for one hour in air in a conventional furnace, then
sintered at 1300°C for 1 hour via microwave heating. The cylinder channel

was formed by pressing a 500 micron pencil lead along with the alumina
powder.

Figure 4.2. A particulate ceramic specimen (25% alumina/75% HAP)
with four cylindrical channels, using 500 micron-diameter pencil lead as
the fugitive elements. The specimen was pre-sintered in air at 900°C for
one hour in a conventional fiIrnace, then sintered at 1300°C for 1 hour via
microwave heating.

Figure 4.3. A high-magnification view of a single channel (formed fi'om a
500 micron diameter pencil lead) in a particulate ceramic specimen (25%
3mol% Y203-partially stabilized zirconia/75% HAP). The pre-sintered in
air at 900°C for one hour in a conventional furnace, then sintered at
1300°C for 1 hour via microwave heating. The specimen was cut such that
the "bottom" of the channel is visible in the micrograph. ............................... 85

Figure 4.4. A fiIlly-stabilized zirconia (FSZ) specimen pre-sintered
conventionally at 900°C for one hour in air, then sintered for four hours at
1430°C in air in a conventional fiJmace. Both the four internal channels
and two surface channels were formed using 900 micron diameter pencil
lead ................................................................................................. 86

Figure 4.5. A close-up view of one of the channels in the F82 specimen
shown in Figure 4.4. .......................................................................... 87

xii

Figure 4.6. A FSZ specimen pre-sintered conventionally at 900°C for one
hour in air, then sintered for four hours at 1430°C in air in a conventional
furnace. The four internal channels were formed using 300 micron
diameter pencil lead. ......................................................................... 88

Figure 4.7. Two channels in a FSZ specimen pre-sintered conventionally
at 1200°C for two hours in air, then sintered for four hours at 1430°C in air
in a conventional fumace. The larger channel was formed using 700
micron diameter pencil lead, while the smaller channel was formed using a
300 micron diameter pencil lead. The cylinder axes of the two channels
are oriented in different directions.

Figure 4.8. A FSZ specimen pre-sintered at 1200°C for two hours in air,
then sintered for four hours at 1430°C in air in a conventional furnace.
Although the specimen included four internal channels, only two internal
channels (formed using 700 micron-diameter pencil lead) are in the field
of view, along with one surface channel (formed using 500 micron
diameter pencil lead). ......................................................................... 90

Figure 4.9. A FSZ specimen pre-sintered at 1200°C for one hour in air,
then sintered for one hour at 1430°C in air in a conventional furnace. Both
the surface channel and the internal channels were formed using 900
micron-diameter pencil lead. Note that two of the internal channels are
within less than one channel diameter of the specimen surface. ........................ 91

Figure 4.10. Two rows of closely spaced channels in a FSZ specimen pre-
sintered at 1200°C for two hours in air, then sintered for four hours at
1430°C in air in a conventional furnace. The larger channels were formed
using 700 micron diameter pencil lead, while the smaller channels were
formed using 300 micron diameter pencil lead. .......................................... 92

Figure 4.11. For a FSZ specimen (pre-sintered and sintered in air at
1200°C for two hours and for four hours at 1430°C, respectively), where
microstructure (a) at a distance of about mm away form an internal
channel is perhaps slightly less porous that the microstructure (b) at a
distance of roughly 200 - 300 microns from the outer wall of an internal
channel. ......................................................................................... 93

Figure 5.1. The micro-textured and configured body. ..................................... 105
Figure 5.2. Three processing routes to attain FGMS. ...................................... 105

Figure 5.3. A multiple-layer sample with a rectangular cross-section
channel affected by residual stress. ...................................................... 106

xiii

Figure 5.4. Shrinkage control using different powder size distribution. ............... 110

Figure 5.5. Five samples with two layers (alumina and zirconia mixed in
five different ratios) showing varying degrees of warpage. .............................. 111

Figure 5.6. Multiple channels on a composite of 75%TMDAR and
25%TZ-8YS. .................................................................................. l 14

Figure 5.7. A channel showing the micro-ridges (the surface feature of a
pencil lead) on alumina sample. ............................................................ 115

Figure 5.8. Channel formed by joining an alumina-zirconia composite
specimen to a partially stabilized zirconia specimen [Lee, 2001]. ..................... 116

Figure 5.9. Manifold-shaped surface channel created by the graphite sheet
cut out with a wire EDM. ..................................................................... 117

Figure 5.10. A spoke-shaped surface manifold created by rapid
prototyping unit .................................................................................. 118

Figure 6.1. Overview of Processing Route. ................................................ 129

Figure 6.2. A SEM micrograph of the fracture of an alumina specimen
sintered from commercial powders (TM-DAR, Tamaii Corp, Japan). The
specimen was pre-sintered for one hour at 9000C in a conventional
fumace in order to burn out the fiJgitive phase. Then, the specimen was
sintered at 1300°C in a microwave cavity.

Figure 6.3. An SEM micrograph of an internal channel in a microwave
sintered alumina specimen. A fugitive-phase rod 0.5 millimeters in
diameter was used to initially form the channel in the powder compact. .............. 133

Figure 6.4. For a alumina/hydroxyapatite particulate composite (25
weight percent alumina, 75 weight percent hydroxyapatite), a low speed
diamond saw cut reveals the four cylindrical channels formed internally in
the specimen. ................................................................................. 134

Figure 6.5. A SEM micrograph of a sectioned hydroxyapatite specimen
showing one of the internal channels formed without machining using the
fugitive-phase rod elements. After sectioning on the low speed diamond
saw, the specimen surface was polished using diamond paste. ........................ 135

Figure 6.6. For the joined hydroxyapatite/lVlaCorTM specimen, the region

near the interface between the two materials. Note that a portion of one of
the channels is also included in the micrograph. ....................................... 136

xiv

Figure 6.7. The interface of the joined sample made of TZ-8YS and 25%
TMDAR / 75% TZ-3YS joined at 1475°C for 20 min. in our microwave
unit.

Figure 7.1. Heating patterns in (a) conventional and (b) microwave
furnaces [1].

Figure 7.2. Temperature gradients generated during (a) conventional and
(b) microwave heating of materials [2].

Figure 7.3. Schematic figure of density versus temperature of microwave
(ZSGHz) and conventionally sintered A1203 [4]. ........................................ 146

Figure 7.4. Fracture surface of alumina (a) sintered by conventional at
1430°C for 4 hours, (b) by microwave at 1300°C for 1 hour. ......................... 148

Figure 8.1a. Shrinkage difference in the diameter of zirconia and alumina
specimens after pre~sintering at different temperatures, and their
corresponding diameter and thickness afier full-sintering at 143 0°C. ................. 157

Figure 8.1b. Shrinkage difference in the thickness of zirconia and alumina
specimens after pre-sintering at different temperatures, and their
corresponding diameter and thickness after full-sintering at 1430°C.

Figure 8.2. Comparison of the out-of-plane height at the center of
zirconia/alumina specimens, based on a variation of particle-size mixture
in the zirconia component. .................................................................. 162

Figure 9.1. Element mapping based on energy dispersive spectroscopy
(EDS) of the SEM sample. X-ray beams characteristic of aluminum,
yttrium and zirconium are collected from the sample surface and the signal
is magnified proportionally as brightness. The different colors are
arbitrarily chosen. Line Scan of the x-ray and EDS element scan shown in
figure 9.2, 9.3.

Figure 9.2. Line scan of the x-ray count based on energy dispersive
spectroscopy. The amount of an element (A1, Zr, or Y) along the line
(figure 9.3 a) is shown by the number of x-ray particles detected over a
dwell time of 100 milliseconds. Element mapping and EDS element scan
shown in figure 9.1, 9.4. ...................................................................... 170

Figure 9.3. (a) Line across the samples for line scan of the x-ray count

based on energy dispersive spectroscopy (Figure 9.2.), (b) square region
(For figure 9.4) of the sample surface. ..................................................... 171

XV

Figure 9.4. EDS elemental scan over the indicated, square region (figure
9.3 b) of the sample surface. X-ray beams (created by shining an imaging,
electron beam on the target) of different energies are detected are
correlated to x-ray signatures of different elements. The JEOL 6400
microscope is capable of detecting all elements from boron to uranium.
Element mapping and line scan of the x-ray shown in figure 9.1, 9.2. ............... 172

Figure I-l. SEM image of fracture surface of Alumina (a) Sintered at
1430°C for 4 hours in conventional furnace, (b) Sintered at 1300°C for 1
hour in microwave.

Figure I-2. SEM images of Fracture Surface of TZSYS (fully stabilized 8
mol% Yttria-Zirconia) (a) Sintered at 1430°C for 4 hours in conventional
furnace (b) Sintered at 1300°C for 1 hour in microwave. .......................... 181

Figure I-3. SEM image of Fracture Surface of TZ3YS (partially stabilized
3mol% Yttria-Zirconia) Sintered at 1430°C for 4 hours in Conventional
Furnace ((a) and (b) represent images taken at different locations). ................... 182

Figure I-4. SEM image of fracture surface of HAP (Hydroxyapatite) (a)
Sintered at 1300°C for 4 hours in conventional fumace, (b) Sintered at
1100°C for 1 hour in microwave.

Figure I-5. SEM image of fracture surface of (a) 25% TMDAR /
75%HAP sintered at 1300°C for 1 hour in microwave and (b) 50%
TMDAR / 50% TZ3YS sintered at 1430°C for 4 hours in conventional
furnace.

Figure II-l TZSYS specimens pre sintered at 900°C for 2 hours and
sintered 1430°C for 4 hours in conventional furnace with various shapes of
manifold channels of (a) ABS which is a thermoplastic (b) ABS and pencil
lead in center and (0) paper. ......................................................... 186

Figure Il-2 TZ8YS specimens pre sintered at 900°C for 2 hours and
sintered 1430°C for 4 hours in conventional furnace (a) surface channels
created by paper as fiIgitive phase, (b) embed channel. ............................... 187

Figure II-3 TZ8YS specimens pre sintered at 900°C for 2 hours and
sintered 1430°C for 4 hours in conventional furnace channels created by
paper as fugitive phase (a) surface channel, (b) embed channel. ....................... 188

Figure Il-4 TZ8YS specimen pre sintered at 900°C for 2 hours and
sintered 1430°C for 4 hours in conventional furnace with manifold shaped
channels created by pure graphite as fugitive phase, (a) Surface channel,
(b) Embed channel (c) Embed channel in position marked in (b). ..................... 189

xvi

Figure III-1 Various channels in homogeneous materials and composite
materials. .......................................................................................
(a) A SEM micrograph of an alumina specimen that was pre-sintered at

900°C for one hour in air in a conventional fumace, then sintered at

1300 °C for 1 hour via microwave heating. Two rows of closely

spaced channels were formed by pressing 500microns pencil leads.

(b) A particular ceramic specimen (25% alumina/ 75% TZ3YS) with two
layered cylindrical channels, using 500 micron-diameter pencil leads
as fugitive elements. The specimen was pre-sintered in air at 900°C
for one hour in a conventional fumace, then sintered at 1300°C for
one hour via microwave heating.

(c) A fully - stabilized sirconia (TZ8YS) specimen pre-sintered
conventionally at 900 °C for 2 hours in air, then sintered for four
hours at 1430 °C in air in a conventional furnace. The channels were
formed using 900 microns pencil lead.

(d) A FSZ specimen pre-sintered conventionally at 950 °C for two hours
in air, then sintered for four hours at 1430 °C in air in a conventional
furnace. Both four internal channels and three surface channels were
formed using 500 micron diameter pencil leads.

(e) A FSZ specimen pre-sintered at 900 °C for two hours in air, then
sintered for four hours at 1430 °C in air in a conventional furnace.
The larger channels were formed using 900 micron diameter pencil
lead, while the smaller channels were formed using 500 micron
diameter pencil lead for internal channels. And, 300 micron diameter
pencil lead were used for surface channels.

(f) A FSZ speciemn (pre-sinterd and sintered in air at 1200 °C for two
hours and for four hours at 1430 °C, respectively) was shown with
900 micron and 700 micron pencil lead as fugitive phase.

Figure III-2 A high-magnification view of a single channel formed (a)
from a 500 micron diameter pencil lead (b) from 900 micron diameter
pencil lead in a FSZ specimen. The specimen was pre-sintered in air at
900 °C for two hours and sintered at 1430 °C for four hours in a
conventional furnace. ........................................................................

xvii

Figure III-3 A high-magnification view of a single channel (a)formed
fi'om a 300 micron diameter pencil lead (b) from 700 micron diameter
pencil lead in a FSZ specimen. The specimen was pre-sintered in air at
950 °C for two hours and sintered at 1430 °C for four hours in a
conventional furnace.

Figure IV-l Pre-sintered at 900 °C for four hours and sintered at 1430 °C
for four hours in air in a conventional fumace. (a) with 700 micron
diameter pencil lead as fugitive phase, (b) with 500 micron diameter
pencil lead as fugitive phase. materials were shown in configuration
respectively. ................................................................................. 1 96

Figure IV-2 Pre-sintered at 900 0C for four hours and sintered at 1430 °C
for four hours in air in a conventional fiimace. (a) with 700 micron
diameter pencil lead as fugitive phase, (b) with 700 micron diameter
pencil lead as fugitive phase. materials were shown in configuration
respectively. .................................................................................... 197

Figure IV-3 Pre-sintered at 900 °C for four hours and sintered at 1430 °C
for four hours in air in a conventional fumace. (a) with 900 micron
diameter pencil leads, (b) with 900 micron diameter pencil leads. materials
were shown in configuration respectively.

Figure V-l. Element mapping based on energy dispersive spectroscopy
(EDS) of the SEM sample. X-ray beams characteristic of aluminum,
yttrium and zirconium are collected from the sample surface and the signal
is magnified proportionally as brightness. The different colors are
arbitrarily chosen.

Figure V-2. Square region (for the figure V-3) of the sample surface. ............... 201

Figure V-3. EDS elemental scan over the indicated, square region (figure
V-2) Of the sample surface. X-ray beams (created by shining an imaging,
electron beam on the target) of different energies are detected are
correlated to x-ray signatures of different elements. The JEOL 6400
microscope is capable of detecting all elements from boron to uranium. ............. 202

Figure V-4. EDS elemental scan over the indicated, square region (figure
V-2) of the sample surface. X-ray beams (created by shining an imaging,
electron beam on the target) of different energies are detected are
correlated to x-ray signatures of different elements. The JEOL 6400
microscope is capable of detecting all elements from boron to uranium. .............. 203

Figure V-5. Line across the samples for line scan of the x-ray count based
on energy dispersive spectroscopy (For figure V-6.). ..................................... 204

xviii

Figure V-6. Line scan of the x-ray count based on energy dispersive
spectroscopy. The amount of an element (Al, Zr, or Y) along the line
(figure V-5) is shown by the number of x-ray particles detected over a
dwell time of 100 milliseconds. Element mapping and EDS element scan

shown in figure V~l, figureV-3, 4. ......................................................

Figure V-7. Element mapping based on energy dispersive spectroscopy
(EDS) of the SEM sample. X-ray beams characteristic of aluminum,
yttrium and zirconium are collected from the sample surface and the signal
is magnified proportionally as brightness. The different colors are
arbitrarily chosen.

Figure V-8. Line across the samples for line scan of the x-ray count based

on energy dispersive spectroscopy (For figure V-9.). ................................

Figure V—9. Line scan of the x-ray count based on energy dispersive
spectroscopy. The amount of an element (Al, Zr, or Y) along the
line(figure V-5) is shown by the number of x-ray particles detected over a
dwell time of 100 milliseconds. Element mapping and EDS element scan

shown in figure V-7, figureV-l 1. .......................................................

Figure V-l0. Line across the samples for line scan of the x-ray count

based on energy dispersive spectroscopy (For figure V-l 1.). .......................

Figure V-l 1. EDS elemental scan over the indicated, square region (figure
V-10) of the sample surface. X-ray beams (created by shining an imaging,
electron beam on the target) of different energies are detected are
correlated to x-ray signatures of different elements. The JEOL 6400

microscope is capable of detecting all elements from boron to uranium. .........

xix

oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo

205

207

..... 208

209

Chapter 1

INTRODUCTION

Recently developed techniques in ceramic joining and processing are applied to
the development and fabrication of thermal management materials. The intended
application is a mesoscale, ceramic heat exchanger in which coolants flow through a
network of interior channels and manifolds. Although this technology can be applied in
aerospace, manufacturing, and medical industries, the current discussion will focus on the
issue of thermal management in microelectronics. In the past, common methods of
removing heat fi'om microelectronic devices have included the conduction of heat
through highly conductive materials such as diamond, silicon nitride, and molybdenum
[1,2] and cooling by air convection [3]. However, by circulating coolants through a
closed system of channels, we can achieve not only better conduction by design, but we
can also minimize the unwanted spread of heat to other peripheral devices.

The current research work combines existing techniques in ceramics processing
and an economical technique to introduce channels within the interior of ceramics; it is
viewed as a step towards the eventual implementation of mesoscale heat exchangers with
an active cooling system.

As a class of materials, ceramics are wear-resistant, hard but brittle and prone to
thermal shock, electrically and thermally insulative, nonmagnetic, chemically stable, and

oxidation-resistant [4 ]. Although the hardness and fracture toughness of ceramic

composites are adequate for microstructural applications, ceramics having complex
shapes are difficult to manufacture.

While the production of channels inside ceramics bodies has previously been
achieved, for instance, by the etching of silicon wafers [5], the current method of creating
channels uses of a graphite fugitive phase that offers more control over channel geometry
at a relatively low cost. The graphite fugitive phase in the shape of the channels is
embedded inside the ceramic powder before compaction. During pre-sintering, the
graphite is dissolved to expose a system of internal channels. Crumm et a1. [6] used the
term “fugitive material” to describe a space-filling buffer that is extruded fi'om an initial
mixture of carbon black powder in a thermoplastic/mineral-oil melt. The buffer supports
channel volumes in the green state, but oxidizes at high temperatures to leave the channel
volumes empty. In the current work, the choice of fugitive material is readily available
and does not require the pre-processing steps of thermoplastic-melt mixing and extrusion.
Also, the work of Crumm [6] et a1. is primarily concerned with a process of multiple-pass
co-extrusion (for size reduction) of repeating units of ceramic channels separated by
metal electrodes, which is a different process from the one envisioned in this dissertation.

Because ceramics are brittle, a main concern for their use in building a heat
exchanger is the presence of residual stresses, warping and interface cracking caused by
thermal gradients. During fabrication, thermal gradients can result from the joining of
two or more simpler parts with different coefficients of thermal expansion, or different
temperatures of sintering. Also, during the operation of the heat exchanger, thermal
gradients can result from the circulation of coolants around the embedded channels, or

from the desired removal of heat from localized, external sources.

The following technique is used to address the problems of residual stresses and
warping which arise during the fabrication process. Namely, ceramic particles of
different sizes are mixed in the composite layers to ensure a flat specimen surface (e. g.
for joining). The Optimal mixture ratio is obtained by varying powder characteristics so
that compatibility between layers (i.e. flatness) is assured in the presence of thermal
expansion and shrinkage from sintering. Different pressing procedures and layer stacking
orders have also been used as complementary steps to minimize warping. Details of our
study on warping are given in Chapter 4.

Joining techniques are applied to combine component layers. Hence, instead of
sintering an entire FGM stack at one time, we join different layers that have been
separately sintered to control warping and cracking. A background of joining techniques
is given in Chapter 2, and an elaboration of the current results on the use of joining to
control warping and residual stresses can be found in Chapter 4.

The proposed media belong to the class of functionally gradient materials (FGMS),
which consist of component phases endowed with spatially varying properties,
compositions, or microstructures for the purpose of achieving desired material response
[7,8,9]. By tailoring their inhomogeneous properties, FGMs can be designed to minimize
the stress during operation while reducing detrimental, process-induced residual stress. A
background of FGM’s is given in Chapter 2, fabrication procedures in Chapter 3, and
results related to the control of warping, SEM examination, etc., in Chapter 4.

Two different types of sintering processing—conventional sintering and
microwave sintering—were applied. Fundamentally, there are many differences between

the two heating methods. Conventional electrical furnace for sintering materials is

composed of heating elements and their insulation. By contrast, the heating chamber of a
microwave is composed of a port through which electromagnetic waves are guided into
the cavity from a power supply. By using these two methods of heating, different
sintering mechanisms can be studied. This is discussed in more detail in Chapter 4.

To check the effectiveness of the current fabrication procedure, the sintered FGM
specimens are observed using scanning electron microscopy (SEM) and energy
dispersive spectroscopy (EDS). This work is found towards the end of Chapter 4.

This dissertation is divided into five parts. Chapter 2 contains a background to
FGM and the different types of joining techniques. The salient features of FGM and a
discussion on the challenges of using ceramics as a base material are presented. This is
followed by a discussion on the different types of joining techniques. Glaze bonding is
chosen for this work because of its reliability in bonding materials having different
coefficients of thermal expansion. The processing parameters such as fusing temperatures
and hold times are also well known to us from previous work.

Chapter 3 contains general experimental procedures. The composition and
properties Of the ceramic powders used—alumina, partially stabilized zirconia, fiIlly
stabilized zirconia and hydroxyapatite—are given. Powder preparation and sintering
procedures, including powder pressing, pre-sintering and fiIll sintering are discussed
along with data on processing temperature, pressure, and hold times. Specifics on
specimen polishing, coating, and joining are also described. The final section touches
briefly on the sectioning of specimens for SEM observation.

The beginning sections of Chapter 4 contain previously published results by the

author on the fabrication of ceramic FGMS with internal channels, joining techniques

applied to alumina, zirconia, and bioinert ceramics, and the control of warping and
residual stress. This is followed by section 4.5 on the difference between conventional
and microwave sintering, section 4.6 on some updated results regarding the control of
warping, cracking and residual stress, section 4.7 on SEM/EDS imaging of FGM

specimens, and finally, a summary and conclusion section.

REFERNCES

 

1 K. J. Gray, Diamond and Related Materials, Vol. 9, No. 2, 2000, pp. 201—204. K.
Jagannadham, “Model of Interfacial Thermal Resistance of Diamond Composites,”
Journal of Vacuum Science and Technology, Vol. A17, No. 2, 1999, pp. 373—379.

2 P. Hui, H. S. Tan, and Y. S. Lye, IEEE Transactions on Components and Packaging
Technologies, Part A, Vol. 20, No. 4, 1997, pp. 452—457

3 R. Hall, Electronic Design, Vol. 47, No. 7, 1999, pp. 29—30

4 Takashi Kawai, Shun-ichi Miyazaki, Muneki Araragi, “A New method for Forming a
Piezo-Electric FGM Using a Dual Dispenser System, ” The first International
Symposium FGM, 1990, pp191-196

5 D. B. Tuckerman and R. F. W. Pease, “High Performance Heat Sink for VLSI,” IEEE
Electron Devices Letters, Vol. EDL-2, May 1981, pp. 126—129

6 A. T. Crumm and J. W. Halloran, “Fabrication of Microconfigured Multicomponent
Ceramics,” Journal of the American Ceramic Society, Vol. 81, No. 4, 1998, pp. 1053—
1056

7 D. P. H. Hasselman and G. E. Youngblood, “Enhanced Thermal Stress Resistance of
Structural Ceramics with Thermal Conductivity Gradient,” Journal of the American
Ceramic Society, Vol. 61, No. 1—2, 1978, pp. 49—52.

8 P. Kwon, Macroscopic Design and fabrication of Functionally Gradient Materials, PhD
Dissertation, U. C. Berkeley, 1994.

9 P. Kwon, C. K. H. Dharan, and M. Ferrari, Journal of Energy Resources Technology,

Vol. 116, 1994, pp. 115—120

Chapter 2

BACKGROUND

2.1 FGMS

In recent years many studies have been undertaken to develop new materials.
Consequently they are continually being introduced as the demand from materials
became more critical. Ceramic composite material is made fi'om ceramic matrix and a
dispersion phase of either ceramics or metal and is fabricated to use excellent properties
of each of these materials. Usually, the size of the dispersion phase of ceramic
composites is in the order of micrometers; however, to improve the properties of
composite materials, the size in the order of nanometer (ceramic nanocomposite) have
been developed [1]. These nanocomposite materials have evenly spread dispersion phase
thus the property can be much more homogeneous as called homogeneous functional
material. Considering aerospace applications these composite materials cannot withstand
the severe thermal stress condition due to severe environment, thus, the new technologies
often force to have that composites perform multiple fiInctions instead of having only
single characteristic. In order to meet this demand, a new class of composite materials
called Functionally Gradient Materials (FGM) have been emerged in two recent
international conferences in the mid-19803 [2, 3, 4]. The term FGM, now widely used by
the materials society, as a new material having spatially inhomogeneous microstructures

and properties. FGM, a novel engineered composite material, was arisen to seek a high

temperature material for a recently proposed space-plane project by Nino in Japan [5].
The resistant material for use in the space plane must endure a ultrahigh temperature
oxidizing environment At the same time it must be able to withstand a high degree of
thermal stress related with its own ability to be a thermal barrier [6, 7]. The space
technologists, estimate that nose cone of the space-plane will be heated to about 2000K
and the air intake of the multi-propulsion engine up to 1900K with a temperature
difference of over 1000K in a thin layer toward the inner portion of the material. Such a
high temperature heating is mainly caused by the aero-dynamic friction with air during
long time cruising in the atmosphere in comparison with the present vertical launching
type rocket. NO industrial materials could withstand such severe thermomechanical
loading. Three features are important in the designing of the material used under the
above mentioned environment; heat-resistant and anti-oxidation properties in the high
temperature surface layer of the material, mechanical toughness in the other side and an
effective thermal stress relaxation throughout the material. Also, FGMS provide low cost
ceramic surfaces displaying grace failure at moderate to high temperatures and high
thermal heat flux. F GM is applicable not only to space-plane thermal-protection systems
but also to the electronic, chemical, nuclear power and biomedical such as dental
implants fields as well as many others (Figure 2.1) [8, 9]. FGM is a novel engineered
composite which is designed to optimize material properties for use in real environments
by place-to-place control of composition and microstructure. In this simplest structure, it
consists of one material on one side, a second material on the other, and an intermediate
layer whose structure composition and morphology vary smoothly from one material to

the other at the micron level. Similar to composite materials, FGM consists of at least two

distinct phases. However, the concentration level of each phase is spatially varying to
attain inhomogeneous properties in macroscopic and microscopic level (Figure 2.2).
Therefore, FGM is the novel material providing the innovative properties and / or
functions that cannot be performed by homogeneous materials, emphasizing the
following two essential features. The first essential feature includes the tailoring of
chemical composition and microstructure due to the gradient change of properties. The
second essential feature includes the availability of good reproducibility. Thus, FGM may
undoubtedly be classified into a distinct category separated from conventional
homogeneous composites. Figure 2.3 explains the differences between the homogeneous
materials and FGMS.

It is now well known that sudden transitions in materials composition and
properties within component can be one of reasons of sharp local concentrations of stress.
In other words, if the transition in materials gradually changes from one side to the other
these residual and thermal stress concentrations can be much more reduced and enhanced
the bonding strength. From the view point of fracture mechanics, the distinguishing
feature of FGMS is in their strength-related properties such as yield and ultimate strength,
fracture toughness, and creep, fatigue, and corrosion crack growth parameters [10, 11, 12,
13 , 14 , 15 ]. Continuously or discontinuously varying compositions and / or
microstructures over definable geometrical distances porosity ratio between the low and
high — temperature can control the thermal stress. Similar to the service condition of a
tool, one practical application of FGMS is to compensate for very stiff thermal gradient
loadings. In many applications, FGMS may be exposed to constant or alternating thermal

loading with high temperature gradient from one side to the other. Therefore, the thermal

stability of the FGM structure must be evaluated, because of microstructural change
caused by Ostwald ripening and because of unstable compositional distribution due to the
diffusion of structural elements [16]. Also thermal shock resistance, thermal fatigue
characteristics, temperature profile and overall heat flow must be evaluated. The main
problem with this kind of FGM is the development of residual thermal stresses at the
interface between the different layers as a consequence of the thermal expansion
mismatch.

In these works, a steep gradient temperature field was compensated by the
variation of the coefficient of thermal expansion (CTE). Designing FGMS for thermal
gradient loading conditions was shown by Kwon [17] and Kwon and Dharan [18,19]. In a
simple elastic analysis, the thermoelastic properties (which are fiJnctions of the volume
fraction of the reinforcement phase) as well as the type of applied loads, such as thermal
gradients, pressure, force and moments, and the geometry of the structure have to be
considered simultaneously in defining the type of gradient, and the required properties.
The gradients can be continuous on a microscopic level, or they can be laminates
comprised of gradients of metals, ceramics, polymers, or variations of porosity/ density.
Brittle structural ceramics for high-temperature applications are highly susceptible to
thermal stress fracture. The incidence of such fracture can be reduced by the selection of
materials with values as low as possible of the coefficient of thermal expansion, Young’s
modulus of elasticity, poison’s ratio, and emissivity in combination with high values of
tensile strength, thermal conductivity, and thermal diffiIsivity. Rapid crack arrests after
thermal stress fracture can be promoted by selecting materials with moderate strength,

high young’s modulus, and high values of fracture energy and which exhibit stable crack

10

propagation. From the point of view of engineering design, thermal stresses can be
reduced by decreasing the structure size and developing highly compliant structures with
a minimum of external mechanical constraints. Thus, the introduction of compositional
gradients materials can be beneficial such as (i) the “smoothing” of the transitions in
thermal with vast thermal expansion mismatch (ii) the flexibility to tailor the magnitude
of thermal stresses and hence the onset and spread of plastic flow and cracking by
tailoring the geometry of the graded layer as well as the gradient in composition within
that layer ; (iii) the possibility to minimize or even fully eliminate the deleterious effects
of stress concentrations and singularities at free edges ; and (iv) making possible the
deposition of thick protective ceramic layers of 1mm thickness or higher by thermal
spraying on to metallic substrates. For the design of FGMS, necessarity of material
properties, such as thermal-expansion-coefficient and Young’s modulus in the specific
region, is Optimized by controlling the distribution profiles of composition and
microstructures, as well as micropores in the materials [20,21]. Various methods are used
in FGM technology to control the composition and structure of a composite. The simplest
one is step by step stacking of mixed powders of different compositions such as Chemical
Vapor Deposition (CVD), Physical Vapor Deposition (PVD), ion plating, plasma
spraying, diffusion treatment, sedimentation, sintering or self-propagating high
temperature synthesis. Another methods which enable to make continuous stacking with
changing composition, include wet filtration process, vibrating stacking process,
centrifugal process, wet powder spray forming process, sequential slip casting and slurry

dipping process [22, 23].

ll

 

<ENGINEERING>
cutting tool, shaft, roller, engine components, turbine blade,
etc.

 

 

<AEROSPACE>
rocket engine components, space plane body, etc.

 

 

<NUCLEAR ENERGY>
nuclear reactor components, first wall of fiIsion reactor, the]
pellet, etc.

 

 

Applica
-tion of
FGM

 

 

 

 

<BIOMATERIALS>
implant, artificial skin, drug delivery system, etc.

 

 

<CHEMICAL PLANT>
heat exchanger, heat pipe, slurry pump, reaction vessel, etc.

 

 

<ELECTRONICS>
graded band semiconductor substrate, sensor, etc.

 

 

<ENERGY CONVERSION>
thermoelectric generator, thermionic convertor, fuel cell,
solar cell, etc.

 

 

<COMMODITIES>
building material, sport goods, car body, window glass, etc.

 

 

 

 

<OPTICS>
optical fiber lens, etc.

 

Figure 2.1. Application of FGM.

 

 

 

 

Figure 2.2. Schematic figure of Functionally Gradient Material

 

 

 

HOMOGENEOUS
MATERIAL

 

FUNCTIONALLY
RADIENT MATERIAL

 

 

CTE

 

 

 

 

 

 

 

 

FRACTURE
TOUGHNES S

 

 

 

 

 

 

 

 

STRESS

 

 

 

 

 

 

 

OXIDATION
RESISTANCE

 

 

 

 

 

 

 

 

 

 

HEAT RESISTANCE

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 2.3. The difference between homogeneous material and functionally gradient

material

13

 

This study involves sintering powder mixtures with spatially varying
concentrations of two or more constituent powders. However, due to the apparent
inhomogeneous nature of FGMS, residual stress occurs in the structural level. This causes
non-uniform shrinkage [24,25]. Most solids shrink when the temperature decreases after
the final phase of consolidation processing. This phenomenon is best described by
examining the Coefficient of Thermal Expansion (CTE). During consolidation, a
mechanism for shrinkage exists to produce densification. Densification occurs in a
powder compact because the interstitial spaces are filled by the mass transport
mechanisms of densification. While the shrinkage difference due to CTE mismatch
cannot be altered for a given combination of materials, shrinkage occurring during
densification can be mixing the powder and optimizing the sintering conditions (time,
temperature, heating/cooling cycles, etc) [26, 27]. During cooling from the sintering
temperature, thermal stress will be generated in the graded compacts, causing crack
formation within the compacts. We must incorporate a composition profile into a graded
compact for sufficient thermal stress reduction to avoid crack formation. The retarding
influence of inclusions on matrix densification can be significant, although the spread in
sintering rates with varying inclusion volume fraction in the non-percolating regime is
not dramatic. Discrepancies in sintering rate are none of the less undesirable, as they
cause warping and, potentially, cracking in the FGM during densification. These effects
may therefore have to be taken into account in FGM densification. Thermal stress
generated during fabrication shows a maximum at the ceramics rich side of the surface of
the FGM sample, which contributes to crack formation. Thus to fabricate crack-flee FGM

it is necessary to improve the fracture strength and toughness of the material in the

14

composition range where maximum thermal stress arises. For the fabrication of the PM
(powder metallurgy) FGM, the thermal stress generated during cooling from the sintering
temperature for the first, is to be minimized. For the actual condition, the thermal stress
generated in a temperature gradient in a thermal barrier situation is to be minimized. In
powder processing of homogeneous materials, the starting powder is often characterized
by an average particle size and particle size distribution. As shown in Reed [28] with
alumina powder, the different combinations of these characteristics result in different
packing fraction. Because the packing fraction after consolidation should approach 1.0,
the overall shrinkage will be different even using the same material with different powder
characteristics. This study address the solution to the most difficult problems associated
with PM processing of F GMs. The processing behavior of PM materials depends on the
starting powder characteristics such as size, shape, particle size distribution, etc. and
sintering temperature, pressure and time. Shrinkage difference between wide and narrow
size distribution powders was studied by Yeh and Sacks [29]. To quantify the nonuniform
deformation in the PM during sintering, the measurement of two-and three-dimensional
and local sintering shrinkage is required by all means for their effective dimensional
control. The powder metallurgical process is a very viable and suitable route for FGM
fabrication. The key points are essentially two-fold: how to obtain the graded green
compact and how to temperature without defect formation. The defects, such as large
residual porosity or small cracks in the compacts, are healed, if necessary, by HIP.
Powder metallurgical (P/M) processing of FGMS provides a wide range of compositional
and microstructural control, along with shape-forming capability [30]. Apart from the

usual advantages and disadvantages of powder metallurgy, there are several technical

15

challenges to be met for successful fabrication of FGM samples: (1) stacking powders
with different mixing ratios according to the predesigned composition profile with the
desired sample shape and size, and (2) compacting and sintering the stacked powder
heap, which has variable compacting and sintering behaviors layer by layer. Powder
compacts with different mixing ratios usually exhibit a different sintering behavior,
which will cause various sintering faults in the FGM samples (such as warping, fi'ustum
formation, splitting, and crack formation) (Figure 2.4). The following figure 2.5 shows

the PM process.

 

 

 

 

 

 

 

 

 

(e) (d) (f)
Figure 2.4. Schematic figures for sintering faults (a) ideal FGM, (b) stepwise figures of

FGM, (c) warping, (d) fiustum formation, (e) splitting and (f) cracking.

l6

 

Materials
(A1203 & ZrOz)

 

 

 

l

 

 

Mixing powder by Ball Milling

 

 

l

 

 

Stepwise stacking in a steel die

 

 

i

Compact

i

 

 

 

 

 

Pre-Sintering

 

 

 

/\

 

 

Final Sintering in
Conventional Oven

 

 

 

 

Final Sintering in
Microwave

 

Figure 2.5. Procedure of P /M

17

 

2.2 Joining

Ceramic materials (A1203, ZrOz, SiC, and Si3N4, etc.) have been used extensively
due to their structural and firnctional properties for the last two decades. These properties
make them attractive for the industrial application. Generally, ceramics are strong
(particularly in compression), excellent oxidation — resistant, wear- resistant, brittle, high
temperature properties, corrosion resistance, electrically and thermally insulative,
nonmagnetic, chemically stable, low densities, the lack of ductility, sensitivity to defects
and very difficult to fabricate. Most ceramic materials are produced in a powder form,
thus, making the certain shapes as a green state and then sintering to consolidate them..
Hence, it is necessary to depend on joining techniques to fabricate the desired complex
shapes from easily produced simple components. The need for joining technologies has
arisen recently due to the frequent use of ceramics. Accordingly, the joining techniques
must be considered in dependence on the partners to be joined.
2.2.1 Joining techniques

There are many processes to join ceramic materials [31, 32, 33, 34, 35, 36, 37, 38].
The main techniques can be divided to two categories such as mechanical attachment and
direct bonding processes.
2.2.1-l Mechanical Attachment

Mechanical attachment is widely used to join ceramics to other materials by
avoiding stress concentrations due to uneven contact. Crimp, clamp, shrink fit, bolt,
screw thread and other variants applications are used as this technique. The techniques,
which are performed without bonding medium and creating interface can be quite simple

to apply and cheap. Usually, screw thread includes expensive machining and shrink

l8

fitting uses the difference in thermal expansion coefficient between ceramics and metals.
Generally, metals contract more than ceramics on cooling and expand more on heating.
The joint is performed at the temperature where the metal expanded enough to surround
the ceramic, and contract and clamp it on cooling.

2.2.1-2 Direct Bonding processes

Direct bonding processes such as Glaze bonding, Metallized and Brazed joints,
Diffusion bonding, Adhesive bonding, Cast around includes the formation of an intimate
and continuous interface between the different materials [39, 40, 41, 42, 43, 44, 45]. The
interface area describes a discontinuity. This discontinuity is a essential key to make a
decision of whether the joining is successful or not.

Glaze bonding is a good process for most types of ceramic materials to join. To
join, the mating surfaces should be polished to be flat then coated with a suitable glaze
slip and the fired to fuse the glaze and then bonding the two pieces together. Many of the
techniques used to join oxide ceramics can be used to join non-oxide ceramics. However,
it is important to distinguish between the two classes of materials; due to differences in
the chemical nature of non-oxide ceramics, the interfacial ceramic/metal reactions,
bonding mechanisms and resulting joint properties are often significantly different.
Especially, most oxide ceramics can be joined by silicate glasses or glazes. Also joints
can be made with borate and phosphate based glasses or glazes. But non-oxide ceramics
are more difficult to join. They tend to react with the glazes and release gas bubbles,
which weaken the bond.

Brazing (liquid phase process) has been most often used for joining ceramic to

ceramic and ceramic to metal due to the requirement of lower temperature than the

19

requirement of other techniques. There are two different ways to join by brazing [46].
First, indirect brazing, where the surface of the ceramic material has been metallized with
braze filler metal before brazing second, direct brazing, where the filler alloys include
active elements like titanium. One of the advantages of the brazing compared with the
conventional welding is that the base materials do not melt. Brazing can be applied for
dissimilar materials that are impossible to be joined by fusion process because the
metallurgical properties are incompatibility. Even though, different mechanical property
such as Coefficient of thermal expansion results in thermal stress between the substrate
and the bond-layer generally brazing produces less thermal stress and distortion since the
whole component is subjected to heat treatment, therefore the distortion can be avoided
from localized heating in welding. Moreover, it is a process that can be easily adopted for
mass production. The process is relatively simple but depend on complex chemistry,
which is occurring at the interface to promote wettability between the substrate material
and metal braze such as Ag-Cu alloy braze [47,48], aluminum [49,50] and Ti-Sn alloy
braze[51] due to improving the bond strength with improving wettability. For example
72Ag-28Cu alloy is often used because its low melting temperature (780°C) makes work
easy and its good ductility make to overcome the mismatches of thermal expansions.
Also, CuNiTiB paste are used as the braze filler metals to join Si3N4 to Si3N4 [52]. A
blaze filler metal has been melted up to the melting temperature between two ceramic
substrates to perform the brazing with or without external pressure but most advanced
ceramics such as alumina, zirconia, silicon carbide and silicon nitride are not wetted by
silver, copper or gold that is basic constituents of many brazing alloys [31]. The steps of

brazing process have been depicted in Figure 2.6 [32].

20

Diffusion bonding can be used to join ceramic to metal than ceramic to ceramic
(Figure 2.7) [43]. Diffusion boned joints have higher strength than brazed or mechanical
joints [53]. It is important than direct diffusion bonding must be used only when the
thermal expansions of the two components are matched. With the materials of different
thermal expansions, interlayers must be used. This is call as the indirect diffusion
bonding. Interlayers are widely used in the form of foils or coatings; electroplated,
evaporated or sputtered to reduce values for the bonding parameters such as temperature,
pressure and time. The high strains introduced during the cooling of the bond from a
relatively high temperature can be accommodated by a combination of elastic and plastic
deformation within the interlayer. Thus, interlayers are sometimes used to reduce this
effect. However, to avoid the changes in microstructure or composition, the interlayers
must be carefully chosen. These changes can be affected in mechanical properties and
corrosion behavior. Schematic examples are shown in Figure 2.8. The two components,
pressure and high temperature for a certain period of time are needed to apply this
method. Generally, small grain size helps for fast grain boundary diffusion, however,
only time and temperature are essential factors for many metals in experiment. Even
though diffusion bonding requires high temperatures to enhance diffusion but little
macroscopic deformation is involved. The mechanism of diffusion bonding illustrates in
Figure 2.9. As pressure were applied, plastic deformation and creep deformation have
been arisen in the first step then diffusion, recrystallization and grain boundary migration
have be placed as following step. Finally, the low pressure (much less than the
macroscopic yield stress) and the high temperatures (> 0.5m Tm where Tm is the

absolute melting point) have been used to eliminate the pore. Isothermal anneals without

21

pressure can be help to remove the small voids. The deformation, which has been

effected in the beginning of the process involves the power law of creep:

e = AO'"exp — (Qc/RT)

where e = creep rate, 0 = stress, Qc= activation energy for diffusion and A and n are

constant with n z 3-4 for Ti6Al-4V.

The interface consists of bonded areas separated by area containing small voids at the end
of this step. Hydrostatic pressure can accelerate void closure by diffusion and plastic
deformation for voids greater than 20pm, but below the size diffusion alnoe controls the
elimination of voids by surface, grain boundary, and volume diffusion mechanism.

One of the most exciting examples of this process has been observed between alumina
and yttria stabilized zirconia since good interface contact can be achieved under
superplastic conditions. The diffusion bonding joints were produced at 1475°C under
pressure of 12.5MPa to give a bonding strength of 1360MNm'2 [54].

Adhesive bonding such as gluing and cementing (for example, Portland cement)
give stress free joint when high temperature and vacuum tightness are not presented. This
process depends on temperature, the duration and the rate of stressing and may be
possible to reduce the strength due to long service. Epoxide, phenolic, acrylic and
polyurethane adhesive can be used up to 200°C then, the temperature above the 200°C

polymides or other thermally stable polymers can be used.

22

 

CERAMIC POWDER PROCESSING

 

Powder press

 

l

 

 

Sintering

 

 

/\

 

 

As-sintered

 

Grinding

 

 

 

 

 

 

/\

 

 

 

 

Lapping Resintering

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

V BRAZING PROCESS
Active ._, Assembly 4— Metal
brazing member
filler
metal Brazing
Testing

 

Figure 2.6. Schematic drawing for active brazing process [3 2]

23

 

 

DIFFUSION BONDING TECHNIQUES

/\

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

PROCESS WITH THE SOLID PROCESS WITH THE LIQUID
STATE PHASE
/\
Parent/pa Interf Interface Low melting Liquid phase
rent -ace coating point Produced by
material foil elements diffusion
interface and
alloy
| | I

Roll Electro- CVD CVD
Bond- plated Interf- Inter Inter Inter.
ed ace faceing Face Facerng

foil coating foil coating

 

 

 

 

 

 

 

 

 

 

 

 

Figure 2.7. Diffusion bonding Techniques [43]

24

METAL.

         

INTERLAYER SEC ONDINTERLAYER METAL-CERAMIC
GRADED
INTERLAYER

 

    

FIRST INTERLAYER

 

    

 

 

 

 

 

 

 

Figure 2.8. Schematic figure of illustrating the multiple interlayers to join the ceramic to
metal [31]

25

FORCE AND HEATING

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 2.9. Schematic drawing of the reaction of diffusion bonding on bond interface.
(a)contact, (b)deformation and interfacial boundary formation, (c)grain boundary
migration and elimination and ((1) volume diffusion and pore elimination.

26

REFERNCES

 

1 Makoto Sasaki and Toshio Hirai, “Fabrication and properties of Functionally Gradient
Materials”, The Chemical Material Issue of the Ceramic Society of Japan, 9Vol.99,
No.10, 1991, pp1002-1013

2 Holt, J. B., Koizumi, M., Hirai, T. and Munir, Z., 1993, Ceramic Transactions:
Functionally Gradient Materials, The American Ceramic Society, Westerville, OH.

3 Akira Kawasaki and Ryuzo watanabe, “Concept and Fabrication of Functionally
Gradient Materials”, Ceramics: charting the future: proceedings of the world Ceramics,
l994,ppl715-l726

4 Kawasaki, A, Watanabe, R, “Concept and PM Fabrication of Functionally Gradient
Materials”, Ceram. Int., Vol. 23, No. 1, 1997, pp.73-83

5 HI Feng and J.J.Moore, [“The application Of Combustion Synthesis in the In-Situ
Production of Functionally Graded Materials (FGM)”, The Minerals, Metals and
materials Society, 1994, pp81-95

6 M. Koizumi and M.Niino, “Overview of FGM Research in Japan”, MRS Bulletin, Jan.
1995, pp19-21

7 Shabana, Y M, Noda, N and Tohgo, K, “Combined Macroscopic and Microscopic
Thermo-Elasto-Plastic Stresses of Functionally Graded Plate Considering Fabrication
Process”, J SME International Journal Series A: Solid Mechanics and Materials
Engineering (Japan), Vol. 44, No. 3, July 2001, pp.362-369

8 Takashi Kawai, Shun-ichi Miyazaki, Muneki Araragi, “A New method for Forming a
Piezo-Electric FGM Using a Dual Dispenser System”, The first International Symposium

FGM, 1990, ppl91-l96

27

 

9 Chu, C, Zhu, J, Yin, Z and Lin, P, “Structure Optimization and Properties of
Hydroxyapatite-Ti Symmetrical Functionally Graded Biomaterial”, materials Science and
Engineering A(Switzerland), Vol.316, No. 1-2, 15 Oct. 2001, pp.205-210

10 John J. Lannutti, “ FGMS in High-Heat —Flux Environments: Cost/ Performance
Issues”, MRS Bulletin, Jan. 1995, ppSO-Sl.

11 Uzun, H, “Effect of Different Factors on Crack Grth Behavior of SiC(P)/A12124
Single Core Bulk Functionally Graded Materials”, Proc. Turkish Ceram. Soc. IV.
Ceramic Congress Proceedings Book, Eskisehir, Turkey, 22-25 September 1998, 4, 1999,
pp.481-487

12 Huang, P], Hubbard, C , Swab, JJ, Cole,MW, Sampath, S, Depalo,S, Gutleber, J,
Kulkarni, A and Margoloes, J, “Functionally Graded Materials for Gun Barrels”, Ceram.
Eng. Sci. proc., Vol.20, No.4, 1999, pp.33-38

13 Lee, C S, De Jonghe, L C and Thomas, G, “Mechanical properties of polytypoidally
Joined Si sub 3 N sub 4— Al sub 2 0 sub 3”, Acta materialia (USA), Vol. 49, NO. 18, 26
Oct. 2001, pp.3767—3773

l4 Yokoyama, A, Watari, F, Miyao, R, matsuno, H, Do, M, Kawasaki, T, Kohgo, T,
Omori, M and Hirai, T, “Mechanical Properties and Biocompatibility of Titanium-
Hydroxyapatite Implant Material Prepared by Spark Plasma Sintering Method”, Key
Engineering Materils (Switzerland), Vol.192-195, 2001, pp.445-448

15 Ma, J, Tan, GEB, “Processing and Characterization of Metal-Ceramics Functionally
Gradient Materials”, Journal of Materials Processing Technilogy (Netherlands), Vol. 113,
No. 1-3 15 June 2001, pp.446-449

16 Akira “Concept and PM Fabrication of FGMS, p73

28

 

17 Kwon, P., 1994, Macroscopic Design and Fabrication of Functionally Gradient
Materials, Ph. D. Dissertation, U.C. Berkeley

18 Kwon, P., Dharan, C.K.H. and M. Ferrari, 1994, J. Energy Res. Tech., 116, pp.115-
120

19 A. J. Sanchez-Herencia, K. Morinaga & J. S. Moya, “A1203/Y-TZP Continuous
Functionally Graded Ceramics by Filtration-Sedimentation”, Journal of European
Ceramic Society, 17, 1997, pp1551-1554

20 D.P.H.Hasselman and G.E.Youngblood, “Enhanced Thermal Stress Resistance of
Structural Ceramics with Thermal Conductivity Gradient”, Journal of The American
Ceramic Society, Vol. 61, No.1-2, 1978, pp.49-52

21 Tohru I-Iirano and Kenji Wakashima, “Mathematical modeling and design”, MRS
Bulletin, Jan. 1995, pp40-42

22 Makoto Sasaki and Toshio Hirai, “ Fabrication and Properties of Functionally
Gradient Materials”, Journal of the Ceramic society of Japan international edition, vol99,
No. 10, 1991 pp970-980.

23 Biesheuvel, PM and Verweij, H, “Calculation of the Composition profile of a
Functionally Graded Material Produced by Centrifugal Casting”, J. Am. Ceram. Soc,
Vol.83, No.4, 2000, pp.743-749

24 Zang, L.M., Yuan, R.Z., Oomori,M. and hirai, T., 1995, J.Mat.Sci.Let., 14, 22,
pp1620-1623

25 Mizno, Y., Kawasaki, A., and Watanabe, R., 1995, Metall. Trans, Vol. 26B, pp75-79
26 German, RM. 1994 Powder Metallurgy Science, metal Powder Industries Federation,

Princeton, New Jersey

29

 

27 Hur, S K, Yoo, S H, Groza, J R , Doh, J M, Yamazaki, K and Shoda, K, “Graded
Coatings by Gradient Temperature Densification”, J. Mater. Res., Vol. 13, No. 5, 1998,
pp.1255-1259

28 Reed, J., 1989, Introduction to the principles of Ceramic processing, john Wiley &
Sons, New York, p99 &p191

29 Yeh, T. and Sacks. M.D., 1988, J. Am. Ceram. Soc, Vol. 71, no. 12, ppC484-487

30 Zhu, J, Lai, Z, Yin, Z, Jeon, J and Lee, S, “Fabrication onrO sub 2- NiCr
Functionally Graded Material by Powder Metallurgy”, Material Chemistry and Physics
(Switzerland), Vol. 68, No. 1-3, 15 Feb. 2001, pp.130-135

31 J. Femie, A. Sturgeon, “Joining Ceramic Materials”, Ceramic Te°hnology
International, 1993, pp183-186

32 AS. Bahranic, “The Bonding of Ceramic to Ceramic and Ceramic to Metal”, The
international journal for the joining of metals : the journal of the European Institute for
the joining of materials (JOM), vol.2(4), Dec. 1990, pp. 104-107

33 “Ceramics and Glasses’, Engineered Materials handbook, vol4, ASM International

pp.477-53 1, section 7 Joining

34 A.Meier, D.A. Javernick, and G.R.Edwards, “Ceramic-Metal Interfaces and the
Spreading of Reactive Liquids,” JOM Feb. 1999, pp.44-47

35 Servet Turan, Ian A. Bucklow, and Robert Wallach, “Capacitor-Discharge Joining of
Oxide Ceramics,” J. Am. Ceram. Soc., vol.82 No.5, 1999, pp.1242-1248

36 A.G.Foley, “Ceramic Joining Applications in Gas Turbines,” Industrial Ceramics,

vol.l9 No.3, 1999, 193-195

30

 

37 A.Kara-Slimane, D.Juve, E.Leblond and D.Treheux, “Joining of A1N with Metals and
Alloys,” Journal of European Ceramic Society, vol.20, 2000, pp. 1829-1836

38 N.D.Tinskey, J .Huddleston, and M.R.Lacey, “Reduction of Residual Stress Generated
in Metal-Ceramic Joining,” materials and manufacturing processes, vol.13, No. 4, 1998,
pp.491-504

39 E.Butler, M.H. Lewis, A.Hey and RV. Sharples, Ceramic Joining in Japan. Report,
The welding Institute, 1986

40 RE. Loehmanm, “Joining of Ceramics”, American Ceramic Society Bulletin, 67(2),
pp375-380, 1988

41M.G. Nicholas and R]. Lee, “Joining dissimilar Materials”, Metals and Materials, 5(6),
1989, pp.348-351

42 MG. Nicholas and DA. Mortimer, “Ceramic / Metal Joining for Structural
Applications”, Material Science and Technology, vol. 1, pp.657—665, 1985

43 PG. Partridge and CM. Ward-Close, “Diffusion Bonding of Advanced materials”,
Metals and Materials, vol.5 (6), pp334-339, 1989

44 HE. Pattee, “Joining Ceramics to metals and other materials”. Welding Research
Council Bulletin 178, pp1-43, 1972

45 Lee, C S, Zhang, X F and Thomas, G, “Novel Joining ofDissimilar Ceramics in the Si
sub 3 N 4 — Al sub 2 0 sub 3 System Using polytypoid Functional Gradients”, Acta
Materialia (USA), Vol. 49, No. 18, 26 Oct. 2001, pp.3775-3780

46 O.M.Akselsen, “Review Advances in Brazing of Ceramics”, Journal of Materials

Science vol. 27, 1992, pp1989-2000

31

 

47 T.Kuzumaki, T. Ariga and Y.Miyamoto, “Effect of Additional Elements in Ag-Cu
Based Filler Metal on Brazing of Aluminum Nitride to metal”, ISIJ International, Vol. 30,
No. 12, 1990, pp.1135-1141

48 D.Sherman, “The Mechanical Behavior of Layered Brazed Metal/Ceramic
Composites”, Materials Letteres, Vol. 33, 1998, pp.255-260

49 K.Suganuma, ‘New Process for Brazing Ceramics Utilizing Squeeze Casting”, Journal
of Materials Science, Vol. 26, 1991, pp.6144-6150

50 K.Suganuma, T. Okamoto and M. Koizumi, “Joining of Silicon Nitride to Silicon
Nitride and to invar Alloy Using An Aluminum Interlayer”, Journal of Material Science,
Vol. 22, 1987, ppl359-1364

51 CG. Wang, H.P.Xiong and Z.F. Zhou, “Joining of Si3N4/Si3N4 with CuNiTiB paste
Brazing Filler Mtals and Interfacial Reactions of the Joints”, Journal of Materials Science,
Vol. 34, No. 12, 1999, pp.3013-3019

52 Chuahgeng Wan, Huaping Xiong and Zhenfeng Zhou, “joining of Si3N4/Si3N4 with
CuNiTiB paste brazing filler metals and interfacial reactions of the joints”, Journal of
materials science, Vol. 34, 1999, pp.3013-3019

53 Yamada T, et al, high Temp Tech, Vol. 5, 1987, pp. 193

54 Nagano T, hidezumi K and Fumihiro W, MRS int. meeting on advanced materials,

Tokyo, 1988

32

Chapetr 3

EXPERIMENATAL PROCEDURE

3.1 Materials

Materials used to fabricate the specimens for this thesis are commercial powders
of alumina (Table 3.1), hydroxyapatite (Table 3.2), two kinds of partially stabilized
zirconia (Table 3.3, Table 3.4), fiIlly stabilized stabilized zirconia (Table 3.3) and
MaCorTM (Table 3.5) were used.

Alumina is chemically so stable, because of their strong ionic and covalent
bonding between the Al and 0 ion, as expected from the value of heat formation (~400
Kcal / mol), A1203 has also outstanding physical stability, such as a high melting point
(2050°C, or 3720°F), the highest hardness among oxides, and high mechanical strength
that is high at room temperature but is lowered largely above 1100°C for wear application.
The superb tribological properties (friction and wear) of A1203 occur only when the
grains very small (<4 um) and have a very narrow size distribution. If large grais are
present, they can pull out and lead to very rapid wear of bearing surfaces due to local dry
friction. And, high density, high purity (> 99.5%) alumina is used in load-bearing hip
prostheses and dental implants because of its combination of excellent corrosion
resistance, good biocompatibility, wear resistance. Moreover, A1203 is a typical
electrically insulating material. The most demanding use is for A1203 as a high-voltage,

high temperature insulator with thermodynamic stability in a reducing environment. In

33

these reasons, a 99.99% pure alumina powder (TM-DAR Taimi chemical Co. LTD) has
been successfully used.

Two kinds of partially stabilized zirconia powder and one kind of fully stabilized
zirconia powder were used: (1) A 99.9% pure partially stabilized 3 mol% yttria — zirconia
(PSZ) powder (TZ — 3YS, Tosoh Ceramic Division, Bound Brook, New Jersey), (2) a
typically 99.95% pure stabilized 5 mol% yttria powder ( Zr02, Cerac INC. Milwaukee,
WI) and (3) a 99.9 % pure fully-stabilized 8 mol % yttria- zirconia (FSZ) powder (TZ-
8YS, Tosho Ceramic Division, Bound Brook, New Jersey ). In this study, two difference
kind zirconia powders with almost similar chemical composition had the difference in
average particle size. Yttria stabilized zirconia powder has attracted considerable
attention for their superior mechanical properties such as bending strength and fracture
toughness [1,2]. The high temperature stable form of zirconia has a cubic fluorite
structure. Because the ionic radius of Zr4+ is too small (0.084nm, or 0.003 um.) to satisfy
the required ratio of cation and anion radi for an 8-coordinated polyhedron Of the fluorite
structure, pure Zl'Oz shows a polymorphic transformation [3]:

- - [000°C 37 ° . . 0 _ . .
M0n0clm1c<————> T etragonal 6i) Cubrc< 268° C >quurd

 

Stabilized zirconia shows high oxygen ion conductivity, because of the high
concentration of oxygen vacancies. It has been widely applied as the solid electrolyte in
oxygen concentration detectors or fuel cells. Generally, multiphase ceramics with
tetragonal and cubic forms are called partially stabilized zirconia (PSZ), and monophase
ceramics in tetragonal form are called tetragonal zirconia polycrystal (TZP). Those
ceramics show very high mechanical strength and fracture toughness. Fully stabilized

zirconia has flexure strength values from 360 to 660 MPa and fracture toughness values

34

of 2.4 MPa‘Jm, which are much lower than those of TZP. Moreover, because stabilized
ZIOz has a high melting point (~2680°C or 4855°F), low thermal conductivity, low vapor
pressure and good corrosion resistance, it gives good performance as a refractory and a
heat insulator for high temperature use.

Hydroxyapatite, known as a bio-ceramicm, is a compound that is similar to bone
mineral, sharing the formula Cam (PO4)6OH2. Crystal lattice of hydroxyapatite is
hexagonal or pseudohexagonal [4]. As shown by radiographic diffraction analysis,
calcium and phosphate mineralize in bone in the form of an apatite, which is similar to
hydroxyapatite [Ca10(PO4)6(OI-I)] - other configurations may include brushite (CaHPO4
2HzO) or octacalcium phosphate [CagH2(PO4)65HzO]. Crystalline structure can be
changed by manipulation of the calcium-to-phosphate ratio, as well as changing the
content of carbonate and fluorine. When human blood meets the hydroxyapatite,
hydroxyapatite acts like human bone. Due to the chemical similarity between synthetic
HAP and bone, no inflammatory reaction of tissue encapsulation is occurring during the
implantation, but chemical interaction takes place. HAP with optimized microstructure
supports the cell adhesion and -minera|ization strongly. However, the biofirnctionality of
synthetic HAP is not comparable to bone. Low strength, insufficient fracture toughness
and reliability exclude its application as a bone substitute material with load bearing
function at the moment [5, 6,7].

MaCorTM, polycrystalline alumina, is a machinable glass ceramic reinforced by a
flurophlogopite mica plates (Coring code 9658) [8]. MaCorTM gives the performance of a
technical ceramic with the versatility of a high performance plastic. MaCorTM glass

ceramic is an outstanding engineering material and is machinable with ordinary

35

metalworking tools. It has a high maximum use temperature of 800°C continuously,
1000°C at peaks and low thermal conductivity and is a useful high temperature insulator.
It is strong and rigid. MaCorm, unlike high temperature plastics, will not creep or deform.
Moreover, it is an excellent electrical insulator and it has zero porosity. When properly
baked out, it will not outgas in ultra high vacuum applications. Table 3.6 gives
coefficients of thermal ecpansion used in this study.

Pencil lead was used as fugitive phase to make the internal channels and surface
channels in this study. Pencil lead is made of nontoxic mineral called graphite. Graphite
is one of the three forms of crystalline carbon; the other two are diamond and firllerenes.
The layered crystal structure of graphite gives it some of its most important
characteristics and one of the lowest coefficients of friction of any material. Over a wide
temperature range, graphite is a sectile, flexible material exhibiting low coefficient of
thermal expansion, high thermal and electrical conductivity, and is almost chemically
inert (Table 3.7). Consequently, graphite is commonly used for the following attributes:
lubrcity, electrical conductivity, refractoriness, thermal conductivity, chemical inertness,
low thermal expansion, and as a carbon source.

Because of graphite's unique structure, it is highly anisotropic or directional in properties.
This is due to the differences in separation distances of atoms between the layer planes
(0.335mm). Hence, electrical conductivity is good along the direction of the lattice planes,
and poorer perpendicular to the lattice planes. A similar result is seen with strength. The
good lubricity of graphite products also results from the hexagonal arrangement of layer
planes. Because each plane is weakly bonded to those above and below it, each is

relatively free to slip or side past one another.

36

Table 3.1. Alumina (TM-DAR, Taimi Co, Japan) powder information [9].

 

 

 

 

 

 

 

 

 

 

Grade TM-DAR
Lot No. 31 16
Crystal Form or
Surface Area (mz/g) 13.8
Average particle Size 1pm) 0.20
Purity A1203 (%) 99.99
Impurity Na (ppm) 1
K (ppm) 1
Fe (ppm) 3
Ca (ppm) 1
Mg(ppm) 1
Si (ppm) 4
Green Density (g/cm3) 2.31
Fired Density (g/cm3) 3.96

 

 

Table 3.2. Hydroxyapatite powder information [10].

 

Description

Calcium Hydroxyapatite
Cal 0(OH)2(PO4)6, Powder,99%

 

Lot Number

X21949

 

Specific analysis or property

Test for Found Theoretical
Ca 36.24% 39.90%

 

 

Spectrographic analysis

Element Result (%)
Al <0.01
Cr <0.01
Fe <0.01
Mg 0.1
Si 0.1
Sr <0.01

 

X-Ray Diffraction Analysis

 

Chart Number: D34565
Major Ca5(PO4)3(OH) matches PDF 9-432
Hexagonal

 

37

 

 

 

Table 3.3. Tosoh zirconia powder information [1 l].

 

 

 

 

 

 

 

 

 

 

 

 

 

Materials TZ-3YS (partially TZ-8YS (fiIlly stabilized)
stabilized)
Y203 (mol %) 3% 8%
Y203 (wt %) 5.15i0.20 13.335060
A1203 (wt %) $0.1 $0.1
SiOz (wt %) $0.02 $0.02
F602 (wt %) $0.01 $0.01
NazO (wt %) $0.04 0.12
Specific surface Area 7i2 7:2
(mz/g)
Density (g/cm3) 6.05 5.90
Bending Strength RT. 1200 300
(MPa)
Fracture Toughness R.T. 5.0 1.5
(Mpa mos)
Hardness (HV 10) 1250 1250
Mean particle diameter 0.59 0.54
(11m)

 

 

 

Table 3.4. Cerac zirconia powder information [12].

 

Lot Number

X17183-l

 

Description

Zirconium Oxide-Yttria Stab.
5.2Mo/o Y203 100%) < 3 mic.
Typically 99.95% pure

 

Specific analysis or property

Test for Found Theoretical
Y203 5.2% 5.2%

 

Spectrographic analysis

Element Result(%)
Ca 0.004
Hf 1.54
Mg <0.001
Si 0.002
Ti 0.03

 

 

X-Ray Diffraction Analysis

 

Major Yo,15Zro,3501,93matches PDF30-l468
Cubic
Minor ZrOz matches PDF37-1484
Monoclinic

 

38

 

 

Table 3.5. MaCor information [13].

 

 

 

Temperature Limit 1000°C
Dielectric Constant l Khz at 25°C 6.03
Dielectric Strength 40KV/mm

 

DC Volume Resistivity

At 25 °C. 1016 ohm-cm

 

Thermal Conductivity

At 25°C 1.46 W/m°C

 

 

 

 

 

Density (g/cm3) 2.52
Modulus of Elasticity At 25°C 64 GPa
Porosity 0

 

Table 3.6. Thermal expansion coefficients for the materials included in this thesis.

 

 

 

 

 

 

Materials Temp. Interval (°C) Thermal Expansion Reference
(Io‘K‘l)
A1203 20 8 14
25- 8.1 15
0-1000 8.8 16
ZrOz Room temp 8.5 17
25- 10.0 10
1000 ll 18
Monoclinic 7
Tetragonal 12
HAP 25-400 14.5 19
400-800 17.5
MaCor 25-300 9.3 20
25-600 11.4
25-800 12.6

 

 

 

 

39

 

 

Thermal conductivities, young’s modulus and powder sizes which are used in this
study are given in table 3.8, 3.9, and 3.10.

The Silica-FilmTM (Emulsitone Company, Whippany,NJ) used as an interlayer
material (spin-on material) to join has impurity less than one part per million of metallic

ions in alcohol based liquid material. It can be densified by annealing at 200°C for

20minutes.

40

Table 3.7. Graphite information.

 

 

 

 

 

 

 

 

 

property Reference

Specific Gravity 2.2

Electrical Conductivity (ohm.m) l6

Perpendicular to c-axis 9.8 x 10"

Parallel to c-axis 4,1 x 10'5

Natural 1.2 x 10*6

Thermal Conductivity 16

(Watts/meter.Kelvin at 273K)

Perpendicular to c-axis 250

Parallel to c-axis 80

Natural 160

Coefficient of Thermal Expansion 17

(IO‘K'l)

Overall 7.8(at 293K), 8.9(500K)

a-axis* -1.2(at293K), -0.7(at500K)

c-axis 25.9(at 293K), 28.2(500K)

Bulk Modulus (GPa) 34 13, 19
7.3 -10.7(non-irradiated, uncoated)
2.5-7.3 (non-irradiated, coated)

1 4.0- 16.9(irradiated, uncoated)
7. 8-8.4(irradiated, coated)

Mohs Hardness 1-2

Young’s Modulus (GPa) 20

a-axis 1060

c-axis 36.5

 

 

 

*z“a” direction is parallel to the basal planes

“c” direction is perpendicular to the basal planes

41

 

Table 3.8. Thermal conductivities of the materials included in this thesis.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Materials Thermal Conductivity Reference
(W/m°K)
A1203 21 14
16-39 21
ZIOz 1.9 14
1.675 (at 100°C) 17
2.094 (at 1300°C) 17
HAP 0.013 22
MaCor 1.46
Table 3.9. Young’s modulus for the materials included in this thesis.
Material Young’s modulus (GPa) Reference
A1203 322 2
366 23
390 22
ZrOz 131 10
190 (partially stabilized) 24
220 (cubic)
HAP 80-100 22

 

 

 

Table 3.10. Powder size used in this thesis.

 

 

 

 

 

 

 

Materials Average particle size (um)
Alumina (TMDAR) 0.2
3 mol % Zirconia (TZ3YS) 0.59
8 mol % Zirconia (TZ8YS) 0.54
5.2 mol % Zirconia (CERAC) 90% < 1.79
50 % < 1.23
10 % < 0.77
Hydroxyapatite (HAP) 2~3

 

 

42

 

 

 

3.2 Powder Preparation

Monolithic ceramic specimens were prepared using the zirconia, alumina, or
hydroxyapatite powders discussed above. In this study, not only each powder by itself
was processed but also a variety of composite powders were processed. To fabricate the
FGMS specimens, alumina and zriconia powders were balled in variety compositions.
Table 3.11 presents the variety of the compositions. The mixing of powders with two or
more components was performed by dry ball milling in a plastic ball mill for 24 hours
with alumina grinding media at a rotation rate of 110 revolutions per minute. The ball
mill charge consisted of a total of 50 grams of powder and about 250 grams of alumina
media and zirconia media, which gave a volumetric ratio of roughly 1 part powder to 2
parts of the alumina media. The kind of media used was dependent on the kind of
powders. For example, to mix the zirconia powders, the zirconia media was used, and to
mix the two different powders together the media was dependent on the ratio of powders
so the media was chosen by depending on the powder with a larger amount. The mixing
was done for approximately 24 hours, with the procedure of interruptions at about 8 hour
intervals to shake the ball mill. The reason for shaking is to evenly distribute the powders

in order to get microscopically homogeneous materials.

43

Table 3.11. Summary of mixing ratios of powders.

 

Group

Composition

 

TMDAR/HAP

75% TMDAR / 25%HAP

 

25% TMDAR / 75%HAP

 

TMDAR / TZ8YS

75% TMDAR / 25%T28YS

 

25% TMDAR/ 75%TZSYS

 

TMDAR / TZ3YS

90% TMDAR/ 10% TZ3YS

 

80% TMDAR / 20% TZ3YS

 

75% TMDAR / 25% TZ3YS

 

70% TMDAR / 30% TZ3YS

 

60% TMDAR / 40% TZ3YS

 

50% TMDAR / 50% TZ3YS

 

40% TMDAR / 60% TZ3YS

 

30% TMDAR / 70% TZ3YS

 

25% TMDAR / 75% TZ3YS

 

20% TMDAR / 80% TZ3YS

 

10% TMDAR / 90% TZ3YS

 

TZ3YS / CERAC

85%TZ3YS / 15% CERAC

 

75%TZ3YS /25% CERAC

 

72.5%TZ3YS / 27.5% CERAC

 

70%TZ3YS / 30% CERAC

 

65%TZ3YS / 35% CERAC

 

60%TZ3YS / 40% CERAC

 

55%TZ3YS / 45% CERAC

 

50%TZ3YS / 50% CERAC

 

25%TZ3YS / 75% CERAC

 

 

TMDAR / TZ3YS / CERAC

70% TMDAR/30%(70%TZ3YS / 30%

CERAC)

 

50% TMDAR / 50%(70%TZ3YS / 30%

CERAC)

 

 

30% TMDAR / 70%(70%TZ3YS / 30%

CERAC)

 

44

 

3.3 Specimens preparation
Thin and round shape specimens were produced by using the powders which were
described above and sintered. And then, the specimens were polished, joined and

sectioned to see via SEM.

3.3-1 Specimen Pressing
‘ All specimens included in this study were pressed from commercially available
powders in a uniaxial steel die at a pressure of approximately 35Mpa (Figure 3.1). From
1.8 grams to 3 grams of powders were used to form the specimens. The mass of the
powders was measured with accuracy of i 0.003 grams by using an electronic balance
(Mettler Instriment Corporation). The green state specimens were about 2.2 cm in
diameter after cold pressing. The fugitive phase used in this study to form channels and
cavities within the specimens were incorporated during the powder compacting stage.
Two types of fugitive phase were used in this study. One type was a pencil lead,
with nominal diameters of 300, 500, 700 and 900 microns. Figure 3.2 shows the
microstructure of pencil lead. The pencil lead was cut into lengths approximately 0.8 cm
to 1.3 cm and embedded in the ceramic powder. The second type of fiigitive phase used
was graphite, which was machined to a range of geometric shapes. Initially, about one-
half to one-third of the powder mass to be used in pressing a given specimen was loaded
into the cylindrical steel die. Then, the powder bed was smoothed and the fugitive phases
were placed on the top of the smoothed powder bed. Next, the remaining powder was
added to the die, and the specimen was pressed at a press of approximately 35 MPa using

a hydraulic press. Several fugitive phases from one to twelve pencil leads were placed

45

into the powder. Two types of configurations were used for the fugitive phase placement
within the powder compact. In one configuration, a single row of aligned fugitive phase
rods was placed within the powder compact. In a second configuration, aligned fugitive

rods were placed at two different levels within the specimen (figure 3.3 and figure3.4).

3.3-2 Presintering

After the fugitive phase rods had been incorporated into the powder compact, the
presintering to bum-out the fugitive phase was done at a temperature of from 400°C to
1200°C in air for from one hour to four hours, with the average heating rate of from 5°C
to 10°C per minute and cooling rate of from 5°C to 10°C. To figure out the successful
holding time and the peak temperature, variety experimental conditions were done in the
conventional electrical resistance furnace. During the processing the specimens were

placed on a thin alumina powder bed, which was spread on an alumina setter (Figure 3.5).

3.3-3 Conventional sintering for Final Sintering

The pre-sintered specimens were subsequently fully sintered using either
microwave processing or conventional processing. The conventional sintering was done
with an electric resistance-heated firmace (Carbilite, Inc.) with an Eurotherrn temperature
controller (Figure 3.6). The heating and cooling rate was approximately 10°C per minute,
with two different holding times, which are two hours and four hours at two different

peak times that were1300°C and 1430°C (Table 3.12). The conventional sintering was

46

 

K ((((((¢1

 

 

 

 

 

 

“Carver Laboratory Press”

1. /

 

 

 

 

 

((((((((((((((((((((((((((

 

 

 

 

 

 

 

 

 

 

Figure 3.1. Schematic drawing of “Carver Laboratory press”.

47

 

Figure 3.2. SEM picture of Pencil lead (a) one end of pencil lead, (b) the surface of pencil
lead.

48

press

 

Fugitive phase

 

 

 

 

 

 

 

 

  

Figure 3.3. Overview of specimen processing (a) one layer embedded fugitive phases and
fugitive phases on the top surface and (b) two layers embedded fiIgitive phases and on the
top surface (c) dimension of the specimen.

49

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

O O O O
O O O O
(a)
O O O O
O O O O
..... O O O O
(b)
O O
O O
O O
(C)

Figure 3.4. Configuration of FGMS specimens with internal channels and embedded
channels (a) one internal channels and channels on surface and (b) two different levels of
embedded channels and channels on surface in one homogeneous material (0) two levels
of embedded channels and channels on surface in FGMS Specimens.

50

done at higher temperatures and longer times than the microwave sintering, since in order
to get a comparable density and microstructure as obtained via microwave sintering, the
conventional firrnace’s processing temperature and time need to be increased. During the

conventional sintering, from six to eight specimens were sintered simultaneously.

3.3-4 Microwave Sintering Processing

The presintered specimens were sintered in a single mode resonant microwave
cavity at a frequency of 2.45GHz using a 0-2 kW microwave power supply (Model
CMPR-250, Wavemat Inc., Plymouth, MI) (Figure 3.7, 3.8). The specimens were placed
in a refractory casket (specimen enclosure, Zircar Fibrous Ceramics, Zircar Products,
Florida, NY) that served both to thermally insulate the specimens and to provide a
microwave suspector material. The refractory casket was made from the low-density
yitria stabilized zirconia (ZYC, Zircar Fibrous Ceramics). Approximately 5mm diameter
view hole, which is at the 10mm above from bottom of the zirconia casket was bored
through the refractory casket to measure the temperature. Then, the specimens on the
aluminosilicate were placed on the bottom stage of the microwave unit. From five to six
specimens were microwave sintered in a given batch (Figure 3.9).

Microwave input power was set to 100 watts and manually controlled with a ramp
rate of 20 — 100 watts every 3 minutes. The temperature was measured by an optical
thermometer (Accufiber Model 10 Optical Fiber Temperature Control System, Accufiber
Division, Luxtron Corporation, Beaverton, OR) after starting the coupling. Before
coupling, which is around 500°C, the temperature was unknown. Microwave forward

power was

51

Compact

/— powder

u: .6--

  
 

  

Figure 3.5. Schematic figure of specimens on the setter.

 

 

High Temperature
Furnace

 

 

 

 

Figure 3.6. Conventional high temperature firrnace.

52

 

Table 3.12. The conditions of conventional sintering

 

 

 

 

 

 

 

 

Material Sintering Holding Time Heating/cooling rate
Temperature (°C) (hour) (°C)
TMDAR 1430 1 or 4 10
TZ3YS 1300 or 1430 1, 20r4 10
TZ8YS 1300 or 1430 1,20r 4 10
HAP 1375 2 10
Alumina / Zirconia 1300 or 1430 1, 3 or 4 5 or 10
composite
Alumina / HAP 1300 1 10
Composite
Two kinds of 1430 4 5 or 10

zirconia mixture

 

 

 

 

53

 

:er
a;

f
2' .6
i

 

Figure 3.7. Photograph of microwave cavity and power supply used in this study.
(A) reflected power detector, (B) microwave power controller, (C) water cooler, (D)
microwave cavity, and (E) optical pyrometer.

54

 

   

Cylindrical Pmb‘"

cavity

 

 

 

 

(b)

Figure 3.8. (a) Microwave cavity and (b) the schematics of microwave processing unit.

55

 

 

1 cm: Aluminosilicate

 

 

 

 

 

 

 

 

10.20m
< >
A V Low
3 cm Density
Zirconia
Cylinder
i O ‘ View hole
Aluminosilicate

 

 

 

Figure 3.9. Schematic showing casket with the position of specimens.

56

increased to 300 watts to bring the specimens to the desired temperature. The microwave
sintering was done at temperature of 1000°C to 1375°C for a time of one hour. During the
process, the short and probe positions were adjusted to controlled the reflected power.
After holding the desired time, the input power was decreased with the cooling rate of
10°C per minute to 500°C. Table 3.13 explained the sintering temperature and holding

time which is used in this study for the microwave processing.

3.3-5 Sectioning the Sintered Specimen

The sintered specimens were cut by using the low speed saw (Leco Corporation,
St. Joesph, MI) with a 10.16 cm diameter 0.3mm thick diamond saw (Superabrasive
grinding wheel, Norton Company, Worcester,MA), with an oil based lubricant (VC-SO,

Leco Corporation).

3.3-6 Specimen Polishing

The automatic polisher (Model Vari / POLTM VP-SO, LECO Corpoeration, St.
Joeph, MI) were used to polish the sintered specimens and specimens which got
commercially made. The machine moved in two rotations and a swinging motion with
random orbit to polish the specimens. The polishing paste (Warren Diamond Powder CO.
Inc, Olyphant, PA) containing the diamond grit was graded for 35, 25, 17, 15, 10, 6, and
1 micron sizes. The specimens were first mounted to a flat 2 cm thick, 20cm diameter
aluminum disk with thermoplastic (Thermoplastic cement, number 40-8100, Buehler
LTD, Lake Bluff, IL) by using the hot plate (Cole Parmer Model 4658) to melt the

thermoplastic up to a 1-3 mm thick layer. The specimens of almost same height are

57

mounted together to keep the balance during the polishing otherwise, it would take so
long time to polish or it would be impossible to get the polished specimens. To densify
the thermoplastic between the aluminum plate and the specimens, it should be cooled
down. Size graded diamond grit was applied to grit size specific polishing cloth (LECO
Co. and Buehler catalog No. 40-7122, ultra pad with adhesive backing 12”, Lake Bluff,
m) affixed to a rigid metal disk. The cloth covered disk plate and the aluminum disk
where the specimens are mounted were load into the polishing machine. The polishing
cloth was lubricated with Ethylene Glycol solution (Polishing oil, Microid diamond
compound extender, LECO Corporation). The polishing oil gave the smooth contact
movement between the polishing plate and the specimens so it would be applied each
time after diamond grit changed. For newly changed polishing cloth, about 10 drops (1
gram) of the diamond paste were applied otherwise several drops were used. For this
study, the cloth covered metal disk rotation rate was set to 120-150 rpm. Initial polishing
at grit size of 35 or 25 microns were took for 30 minutes to 1 hour after this stage, there
was no change on surface of the specimens, then polishing were done in 17, 15, 10, 6 and
1 microns. After taking the 17 microns diamond grit, the specimens were shinny. Before
going to next grit size and after final stage of polishing, the specimens were cleaned with
liquinox soap (Liquid-Nox, Alconox, Inc, New york, NY) and soft bristle brush to
remove the polishing oil and the diamond grit. Table 14 shows the approximate polishing
time on each different stage.

After the polishing was done up to 1 micron grit size the specimens were
demounted from the aluminum disk by heating up on the hot heater. The demounted

specimens were put into acetone for 5 to 10 minutes to removed the thermoplastic from

58

the specimen then an ultra sonic cleaner (Ultra-Met 111 Sonic Cleaner, Burhler LTD,
Evanston,IL) was used with de-ionized water to clean up for final stage and dried with

paper towels to removed the excess water.

3.3-7 Coating Procedure

The pre-ceramic organic — based silica precursor, spin-on material, (Silica Flimm,
Emulistone Co., Whippany, New Jersey) was used as an interlayer material for ceramic /
ceramic joining. The specimens were coated with the silica film (Table 3.15).

To perform this procedure, the specimens were affixed to the petridish by using
the double-sided adhesive tape then a few drops of the silica film to cover the whole
surface were applied to the polished surfaces of the specimens that were to be joined. The
specimens were then spun for 20 seconds at 3000 rpm on a commercial substrate vacuum
spinner (Figure 3.10.). The coating was then cured by heating the specimens in air for
twenty minutes at 200°C in conventional furnace. The resulting amorphous silica coating

was a high purity silica coating approximately 200 nm thick.

3.3-8 Joining procedure

Ceramic / Ceramic joining were done in both conventional furnace (Carbilite,
Inc.) and microwave cavity (2.45 GHz single mode that included a 0-2kW microwave
power supply-Saiem, MWPS 2000, Wavemat Inc., Mi). The specimens were placed
inside of the 7.3 cm in diameter and 1.3 cm thick cylindrical refiactory casket with the 20

grams of dead

59

Table 3.13. The different conditions of microwave sintering.

 

 

 

 

 

 

Material Sintering time (°C) Holding Time(hour) Heating/ cooling
rate
Alumina (TMDAR) 1300 1 10°C/min.
3 mol% yttria 1375 1 10°C/min.
stabilized Zl‘Oz
(TZ3YS)
8 mol% yttria 1375 1 10°C/min.
stabilized ZIOz
(TZ8YS)
Hydroxyapatite 1 100 1 l 0°C/ min.
(HAP)

 

 

 

 

Table 3.14. The conditions of polishing which is done for this study (time, diamond grit
size and materials used).

 

 

 

 

 

 

 

 

 

 

Material Polishing time (minutes) on each diamond grit size
35pm 25 um 17pm 1 5pm 10pm 611m 111m
Alumina 20 20 40 40 30 30 30
(TMDAR)
3mol%YzO3 20 40 40 30 30 3O
Zirconia
(TZ3YS)
8mOI%Y203 20 40 40 30 30 30
Zirconia
(TZSYS)
Hydroxyapatite 20 30 30 30 30
(HAP)
Alumina/ 20 20 30 30 30 30 30
Zirconia
composite
Alumina/ 20 40 30 30 30 30
Hydroxyapatite
composite

 

 

 

 

 

 

 

 

60

 

Table 3.15. The condition of silicafilm coating.

 

 

 

Spin-on Spin rate Spin Time Curing Curing holding
Material (rpm) (Sec) temperature Time (minutes)
(°C)
Silica Film 3000 20 200 20

 

 

 

 

 

61

 

. Apply silica
P011§h6d film on desired /
Specrmen surface . .
6 ————> 6
V v
Q o o
g g

——>

 

 

 

 

 

/ / "1152; "3:6;
Spinning Z
contrOl

 

 

 

 

 

. Vacuum
pump

 

Figure 3.10 The schematic drawing for process of spin-coating.

62

weights (Alumina specimens) to give a low pressure for microwave joining and put into
the conventional furnace with dead weight too. The heating and cooling rate was about
20°C per minute. During the microwave joining process, the reflected power was
controlled by adjusting the short and probe positions. After coupling, which is occurred at
input power between 200 and 250 watts the temperature quickly reached to 800°C. The

comparison of joining conditions between microwave and conventional firrnace are

shown in Table 3.16.

Table 3.16. The comparison of joining comparison of joining conditions between

microwave and conventional fumace

 

 

 

 

Joining Temp. Joining Time Dead Weight Heating Rate
(°C) (min) (gram) (°C/min)
Conventional 1475 240 20 10
Heating
Microwave 1 3 00 60 20 1 0
Heating

 

 

 

 

 

3.3-9 Specimen Characterization
3.3.9-1 To compare the degree of warping

To fabricate the FGMS specimens, the most difficult part is to control the residual
stress to minimize the cracking or warping. In this study, two different size of zirconia
powders were mixed together to reduce cracking or warping. Magnetic Base with Dial
Indicator was used to measure the heights of warped the specimens to indicate the how

much it was not warped (Figure 3.11.)

63

 

3.3.9-2 SEM Observation

The polished specimens were mounted on aluminum stub using carbon tape
(Structure Probe Inc., West Chester, PA) with the specimen surface of interest facing
upwards. Also, to prevent charging during SEM observation, carbon paint (Ted Pella,
Inc., Redding, CA) was applied to the edges of the sample before gold coating. The
specimens were coated with about 21 nm (1 minute produce approximately 7 nm gold
coating) of gold by a sputter coater (Emscope SC 500). To proper coating, the timer was
set for 3 minutes with a current of 20 mA.
A JEOL 6400 scanning electron microscope was used to look the around the channel of

FGMS and the joined surface and fracture of materials. The accelerating voltage was 20

kV.

64

 

Figure 3.11. Picture of magnetic base with dial indicator.

65

REFERENCES

 

1 T. Sakuma, Y. Yoshizawa, H. Suto, “The microstructure and mechanical properties of
yttria-stabilized zirconia prepared by arc-melting”, Journal of Material Science 20, 1985,
pp2399-2407

2 T. Tsukuma, Y. Kubota, T. Tsukitate, in: N. Claussen, M.Ruhle, AH. Heuer (Eds),
Advances in Ceramics, Vol. 12, Science and Technology of Zirconia II, American
Ceramic Society, Columbus, OH, 1984, pp.382-390

3 EC. Subbarao, “Science and Technology of Zirconia”, Vol.3, Advances in Ceramics,
AH. Heuer and L.W. Hobbs, Ed, American Ceramic Society, 1981, pp1-24

4 J .C. Elliott, Structure and Chemistry of the Apatite and Other Calcium Orthophosphates,
Elsevier, 1994

5 Ralph W.G. Wyckoff: Crystal Structures Volume 3 Inorganic Compounds Rx(MX4)y,
Rx(MnXp)y, hydrates and ammoniates, second edition, Interscience publisher

6 J .C. Elliot, P.E. Mackie, R.A. Young: “Monoclinic Hydroxyapatite”, Acience, Vol. 180,
1995, pp.1055-1057

7 Jiming Zhou, Xingdong Zhang and Jiyoung Chen, “High Temparature Characteristics
of Synthetic Hydroxyapatite”, Journal of Material Science: Materials in Medicine, 4,
1993, pp.83-85

8 http://wwwferroceramiccom

9 TM-DAR materal data sheet from Taimi Co

10 HAP material data sheet from Cerac INC, Milwaukee, WI.

11 Zirconia powder data sheet from Tosoh Ceramics Division, Bound Brook, NJ

66

 

12 Zirconia powder data sheet from Cerac Cerac INC, Milwaukee, WI.

13 http://www.coming.com/search_main.html

14 “Technical ceramics”, Advanced materials technology international, KM Nutter and
M P Davidson, 1990, p. 13 1-133

15 O.M.Akselsen, “Review Advances in Brazing of Ceramics”, Journal of Material
Science, Vol. 27, 1992, pp.1989-2000

16 “Introduction to Ceramics”, 2nd edition, W.D.Kingery, HK. Bowen and
D.R.Uhlmann, pp 595

17 Zirconia powder data sheet from Cerac INC, Milwaukee, WI.

18 Engineered Materials handbook Vol.4, Ceramics and Glasses, ASM international
pp776

19 Joining of Ceramic Materials Using Spin-on Interlayer Master thesis of Jong-Gi Lee
Michigan State University

20 ED. Case, K.Y. Lee and J.G.Lee, “joining of Polycrystalline Ceramics and Ceramic
Composites Using Microwave Heating”, in proceedings of the 33rd International
Microwave Power Symposium, International power Institute, Manassas, VA, pp. 17-20
(1998)

21 Engineered Materials handbook Vol.4, Ceramics and Glasses, ASM international
pp974

22 Engineered Materials handbook Vol.4, Ceramics and Glasses, ASM international
pp776

23 Engineered Materials handbook Vol.4, Ceramics and Glasses, ASM international pp

759

67

 

24 ‘Fundamentals of Ceramics’, michel Barsoum, McGraw-Hill series in Materials

Science and Engineering, pp401

68

Chapter 4

FABRICATION OF INTERNAL CHANNELS IN CERAMICS AND CERAMIC

COMPOSITESl

ABSTRACT

In engineering applications, ceramics possess important characteristics such as
chemical inertness and high stiffness and hardness, which cannot be surpassed by most
metals. However, difficulties in the processing of ceramic materials have limited the
utilization Of these excellent properties. In particular, it is difficult to fabricate ceramics
with complex geometry, especially internal geometric features such as channels or
cavities in the bulk of a ceramic component. This paper presents a unique and economical
processing technique that utilizes a fugitive phase to produce cavities in the bulk of
ceramics and ceramic composites. The fugitive phase in the Shape of the desired channels
is introduced into a die during powder compaction. During pre-sintering of the green
body, the fugitive phase burns away, leaving the cavities and channels. A final sintering
step via either conventional or microwave heating produces a densified ceramic
component having bulk-penetrating channels or cavities. Ceramic/ceramic joining

technology, developed by the authors and other researchers, together with the presented

 

l H.W. Shin E.D. Case and P. Kwon, “Fabrication of Internal Channels in Ceramics and Ceramic
Composite”. Journal of Advanced Material accepted for publication

69

processing technique provide the potential for fabricating very complex parts by joining

subcomponents with other such subcomponents, which all have complex cavities.
4.1 INTRODUCTION

Although ceramics Offer many attractive material properties, including a good
resistance to chemical attack and erosion, biocompatability, and good high temperature
stability, the use of ceramics has been limited, in part by the brittle nature of ceramics and
in part by the difficulty in fabricating complex shapes with ceramics. The limitation of
fabricating ceramics with complex shapes impacts their potential use in many
engineering applications, since it is very difficult to fabricate ceramic components having
internal channels. Thus, although a particular ceramic or ceramic composite may have a
superior resistance to chemical attack and erosion or a remarkable biocompatability,
another material will be used in place of the ceramic if the application demands internal
channels that are difficult to produce in the ceramic material.

Even though the presented processing protocol can be applied in many practical
engineering applications including aerospace, manufacturing and medical, we will focus
this introduction with the applications to microelectronics. The development of highly
efficient temperature management materials (TMMs) is becoming essential to allow the
continued progresses in microelectronics such as miniaturization and multichip modules.
Such devices generate higher heat output, which must be managed to avoid deleterious
effects on the basic functions of the device. Advanced TMMs resolve the thermal
management problem by efficiently dissipating the heat being generated. A variety of

high conductivity materials such as diamond [l] and aluminum nitride have been utilized

70

to spread the heat generated by electronic devices. Without such materials, the generated
heat locally increases the temperature, which directly affects the integrity of the devices.

In addition to spreading out the heat, TMMS can extract the heat from the site
where the heat is generated by circulating cooling fluids. Internal channels, which serve
as conduits for cooling fluids, can be included in the design of TMMs. The mechanical
stresses and the thermal and chemical environments to which the components are
subjected also help guide the selection of materials to be used in such components.

Because of the difficulties in producing ceramic components with internal channels,
silicon wafer has been used in some recent work to address this challenge. For example,
Tuckmerman and Pease [2] produced high flux electronic cooling channels on lcm2
silicon integrated circuit chips with cooling channel widths Of 57 um and heights of 365
pm. These rectangular-cross section passages were etched onto <110> silicon wafers
using KOH, where deionized water serves as the coolant through the passage. The
spacing between channels was about equal to the channel width. The etched channels in
the silicon wafers were converted into closed channels by anodically bonding a Pyrex
plate over the channels. Tuckmerman and Pease's TMMS heat dissipation rate was 40
times higher than that for conventional IC-heat-sink cooling schemes [2].

Tuckerman and Pease’s concept has extended by many researchers. For example,
by adding manifolds on multiple microchannels Harpole and Eninger [3] fabricated high
aspect ratio channels (250m width, 200nm depth and 40 um distance apart between
channels) on silicon wafer diffusion bonded to diamond face sheet. Bower and Mudawar
[4] fabricated mini- (2.54mm diameter) and micro- (0.51mm diameter) channeled heat

sinks made of copper. The heat sink geometry such as channel spacing and overall

71

thickness was optimized through the experiments. By using fluorocarbon liquid as the
coolant, channel length and wall thickness was thinner than when water was used as the
coolant. The new design reduces the maximum component temperature and drastically
reduces variation in component temperature [5].

This paper represents the part of our current study focusing on the development of
powder processing protocols for highly efficient TMMS made of ceramics and ceramic
composites. One of the important processing aspects related to ceramics and ceramic
composites is being able to produce channels internally in the bulk of ceramic materials.
Our processing protocols enable us to produce channels economically. Low temperature
presintering burned out the cylindrical fugitive phases elements from ceramic powder
compacts. (In this study, the fugitive phase elements consisted of pencil lead with
diameters between roughly 300 and 900 microns). As a second processing step, both
conventional and microwave heating were used to sinter the ceramics and ceramic
composite materials, resulting in dense ceramics having circular internal channels.

The size scale of the channels included in this study is appropriate to the active
cooling of electronic materials [6 - 8]. For example, some recently developed cooling
systems for computer hardware include channels that are 0.78 cm across, while a variety
of "microchannels" for active cooling applications range from several hundred microns
across to as small as about 150 microns across.

To date, the authors and co-workers have formed internal channels in ceramics by
joining specimens with "open" channels either machined into densified specimens [9 -
11] or "pressed" into the surface of powder compacts prior to sintering [12]. The internal

channels formed by joining specimens with surface channels are of course constrained to

in the plane of the interface between the joined components. In contrast, in this study, we
form internal channels by the burnout of fugitive phases pressed into the powder compact.

Although this study focuses on fabricating internal channels in ceramics and in
particulate ceramic composites, ceramic/ceramic joining techniques are available that
allow the joining of two or more subcomponents to build up increasing complex ceramic
parts. For example, the authors and co-workers have successfully joined a variety of
ceramics, including alumina/alumina [9], zirconia/zirconia [l4], SiC/SiC, HAP/HAP [14],
and MaCorTM/MaCorTM [9]. (HAP is the bioceramic hydroxyapatite and MaCorTM is a
machineable glass ceramic composite composed Of an aluminosilicate glass matrix
reinforced by mica platelets). In addition, the authors and co-workers have joined a
variety of dissimilar ceramics, including alumina/zirconia [15], HAP/MaCorTM [16],
MaCorTM/zirconia and a variety of other ceramic/ceramic combinations. Thus, specimens
with internal channels fabricated by the technique discussed in this study might be used
as either stand-alone components or joined with other ceramic components to produce

components with quite complicated channel or cavity structures.

4.2 EXPERIMENTAL PROCEDURE

Materials: For this study, the following commercial powders were used: (1) A 99.99%
pure alumina powder (TM - DAR, Taimi chemicals Co. LTD), (2) A 99.9% pure
partially-stabilized 3 mol% yttria — zirconia (PSZ) powder (TZ — 3YS, Tosoh Ceramic
Division, Bound Brook, New Jersey), (3) A 99.9 % pure fully-stabilized 8 mol% yttria —
zirconia (FSZ) powder (TZ — 8YS, Tosoh ceramic division, Bound Brook, New Jersey)

and (4) 99% pure Hydroxyapatite (HAP: Cam (PO4)(, (OH)2, Cerac Inc., Specialty

73

Inorganic, Milwaukee, WI). The average initial particle size of each powder is given in
Table 4.1. Both single-phase specimens (alumina and FSZ) and particulate composite
specimens (25% PSZ/ 75% HAP and 25% alumina! 75% HAP) were included in this
study.

The particular materials included in this study (alumina, FSZ, and the particulate
composites of PSZ, alumina and HAP) were selected on the basis of their characteristics
and a wide range of applications. For example, alumina is a relatively tough ceramic that
is also very hard and very resistant to chemical attack. Many applications of alumina
include electronic substrates and a bioinert material in implants. FSZ (fully stabilized
zirconia) is an ionic conductor that is used in oxygen sensors and is a candidate material
for use in fuel cells. While the coefficient of thermal expansion (CTE) of alumina (7.2-
8.8x10'°K") is close to that of FSZ (7.5xlO'6 K'l), their thermal conductivities are quite
different (35W/mK for alumina vs. l9W/mK for FSZ). HAP is the major mineral
constituent of bones and teeth and is a bioactive material that is used in a variety of
implants. However, HAP has relatively low fracture toughness and strength, thus the
composites HAP with ceramics of higher fracture toughness such as alumina and PSZ,
are of interest [13].

Table 4.1: List of powders used

Powder 1 Average particle size (um)

 

Alumina (TMDAR ) 0.2

3 mol % Zirconia (TZ — 3YS ) 0.5
8 mol % Zirconia (TZ — 8Y8 ) 0.54
Hydroxyapatite (HAP) 2 ~ 3

 

 

 

 

 

74

Thus, the materials included in this study represent a range of current and potential
application areas of ceramics and ceramic composites.

Specimen Preparation: The composite powders included in this study were mixed using
dry ball milling in a plastic ball mill with alumina media at a rotation rate of 110
revolutions per minute. The ball mill charge consisted of a total of 50 grams of powder
and about 250 grams of alumina media (which gave a volumetric ratio of roughly 1 part
powder to 2 parts of the alumina media). The ball milling was carried out approximately
for 24 hours. All specimens included in this study were pressed in a uniaxial steel die at a
pressure of approximately 35 MPa using a Carver hydraulic press. The internal diameter
of the die was 22.2 mm. Between 1.5 and 3 grams of powder were used to form the
specimens (Table 4.2).

During the course of our investigation, a variety of materials including cotton
thread, polymer fibers, pencil lead and shrink tape has been tested as fugitive phases.
Surface channels were successfully produced by each of these fugitive phase materials
[12]. However, the pencil lead was the only fugitive phase medium to consistently
produced internal channels without any damage in the specimen. We speculate the
damage in the other samples were due to the discrepancy between the CTE of the fugitive
phase and the shrinkage during consolidation and the differences between the compacting
behavior of powder and the compression behavior of the fugitive phases.

For the internal channels fabricated in this study, the pencil lead was incorporated
with the powder compact. About one—third to one-half of the measured powder was

initially poured into the die cavity, then the pencil lead in lengths of 1.3 cm to 1.7 cm was

75

Table 4.2: Initial powder mass for each specimen

7 Powder

 

A1203

 

Zl’Oz-3m01%Y203

 

ZrOz-8mol%YzO3

 

HAP

 

25 % A1203 / 75 % HAP

 

25 % ZrOz-3mol%YzO3 / 75 % HAP

 

 

76

placed on top Of the powder in the die, with a lateral spacing of roughly 1.4 mm to 1.6
mm between the lead. More powder was then added to the die cavity before pressing.

To produce two rows of internal channels, two sets Of pencil leads were
sandwiched between three layers of powder in the die cavity while a single row of
internal channels was produced by centering a single row of pencil lead within the
powder in the die cavity. Surface channels were formed by placing pencil lead on top of
the powder bed immediately before pressing.

The pencil lead used had nominal diameters of 900, 700, 500, and 300 microns.
Typically, four pencil leads were used in a given row of channels, with up to four surface
channels per specimen. Both surface and internal channels were produced on several
specimens.

After pressing, the pencil leads were burned out by a pre-sintering in air in an
electrical resistance furnace at either 900°C or 1200°C for between two and four hours.
Pre-sintering was done in a conventional furnace for all specimens included in this study,
however the final sintering of the specimens was done either in a microwave cavity or in
a conventional furnace.

The conventional sintering was done in an electrical resistance furnace in air with
a heating rate of approximately 10°C per minute; with a 4-hour hold time at 1430°C. The
specimens were then cooled to room temperature at a rate of roughly 10°C per minute.
From six to eight specimens were sintered simultaneously.

For the microwave processing, the presintered specimens were sintered in a
single-mode resonant microwave cavity at a frequency of 2.45 GHz using a 0 - 2 kW

microwave power supply. The specimens were sintered in a refractory casket (specimen

77

enclosure) that served to thermally insulate the specimens and to provide a microwave
susceptor material [17]. The microwave sintering in air was done at in the temperature
range between 1300°C to 1310°C for one hour. For each sintering or joining run, the
temperature was measured using an optical pyrometer. The details of the microwave
apparatus and the microwave sintering procedure are given elsewhere [18]. As was the
case for the conventional sintering, multiple specimens (from five to six specimens) were
microwave sintered in a given batch.

Using a low-speed diamond saw, the sintered Specimens were sectioned. The
specimens were then mounted and polished using a series of diamond paste having grit
sizes from 25 micron to 1 micron. Both the structure of the cavities and the specimen

microstructure was examined using optical and scanning electron microscopy.

4.3 RESULTS AND DISCUSSION

The specimen dimensions were measured with a caliper after pressing, after pre—
sintering and after final sintering. The specimen mass was measured with an electronic
balance. The mass density of the specimens in the green (as pressed) state and in the
sintered state was then calculated from the measured Specimen mass and dimensions
along with the specimens' volumetric and linear shrinkages due to sintering, which are
shown in Table 4.3.

The shrinkage of the powder compacts is an important aspect of the processing.
As pressed, the density of the powder compacts ranged between 40% and 62% for the
powder with a narrow initial particle size distribution, where the remaining "volume" of

the powder compacts consists of void spaces among the particles of the powder compact.

78

If a specimen achieved theoretical density, the specimen volume would be reduced by
about 43% - 54%, which corresponds to a linear shrinkage of roughly 15% - 30%. During
fabrication, one would need to take account of the volumetric and linear shrinkage values,
in order to determine the final specimen dimensions and the final dimensions of the
channels. However, the shrinkage of the powder compact is potentially a benefit, since
one could obtain a channel diameter that is about up to 18% smaller than the diameter of
the fugitive phase elements that are introduced into the powder compact. At the present
time, we do not know the exact reason for a large shrinkage with 900micron fugitive
phase and a small shrinkage with 700micron fugitive phase. Thus, the shrinkage provides
a “reduction factor” for the final, processed channel diameters relative to the size of the
fugitive phase elements themselves. For a given type of ceramic powder, if the key
processing parameters (such as the heating rate, the sintering temperature and the
sintering time) remain the same from run to run, then the final density of the component
and the final diameter of the channels should be quite repeatable.

The channels and cavities formed by the elimination of the fugitive phases were
examined by both optical microscopy and scanning electron microscopy (Figures 4.1-11).

The channels formed are generally free of cracks for both the microwave
processed (Figures 4.1-3) and the conventionally processed (Figures 4.4-11) materials.

SEM micrographs showing the details of individual channels in the sinte
redalumina (Figure 4.1), the 25% P82 / 75% HAP (Figure 4.3), and the FSZ (Figures 4.5

and 4.7) reveals that the channels are quite circular.

79

      
 

   
    

 

Acxcmfivv

 

csvdfi

had

 
 

mmmd

mmmd H H

 

.632?me 803 macEficqm :3 805.25 586ch u H was 56:86 5.500% M Q ._.

@cfiwwv

E

    

 

   
    

 

 

 

     
    

 

       
       

 

 

  
 

  

 

 

 

   

  

 

 

 

 

scam SEN SE was as u a as we SE S mm
is 32 ,
S95 scams s2 .8 $3 N2 moo u C.
see. c . . . . u .75 s E.
33 83 E 8 as: as N 2 N a RON? s 8
scene Sada sax. and :2 mac n s
c .
S a. 3m 9% smam we; 2a as u a NE
38.5 Assess sis mono Ed Ed u .9
sm mm 3m 34 an: a: as as u *9 SN?
Own 5.. WNW—WW MMHHWV owe 5.5m “5.82.? wfiufifim wage...— m 2.53
a. . :m . a . a a. v. =3 .32 -8.— csZ .55. _ . 2

A86 .5 SEER; 2: E moocouotfi 56585 ”me 03m...

 

      

80

While no obvious distortion of the channels is visible from the micrographs
(Figures 4.1, 4.3, 4.5 and 4.7), repeated measurements of the channel diameter were
preformed using the SEM software. For example, for 10 measurements on the diameter
of the channel in FSZ material, the mean and standard deviation of the channel diameter
was 729nm and 611m, respectively. The relatively small coefficient of variation (the
standard deviation divided by the mean) of 0.8% indicates the channels are indeed quite
circular.

All of the specimens included in this study were fabricated with multiple internal
channels (Figures 4.2, 4.4, 4.6, 4.8, 4.9 and 4.10) and most included both surface
channels and internal channels (Figures 4.4, 4.6, 4.8 and 4.9). A key feature apparent in
these micrographs is that no cracks are apparent between the channels and between the
channels and the specimen surface. The fact that no cracks were evident on the channels
near the specimen surface is significant, since in some cases the internal channels are
located within approximately one channel diameter (or less) of the specimen surface
(Figures 4.4, 4.6, 4.7, 4.8 and 4.9). Also, the spacing between the internal channels was
as small as roughly two channel diameters or less, but despite their proximity, no
cracking was apparent (Figures 4.6, 4.8. 4.9 and 4.10). The good circularity of the
channels and the lack of cracks in the specimen are important in our application to TM.
The four different diameters of pencil leads used (900 microns, 700 microns, 500 microns
and 300 microns, as specified by the vendor) produced four different channel sizes, from
about 730 microns (from the 900 micron diameter pencil lead) down to about 280

microns (using the 300 micron pencil lead). The submillimeter size range (in this case

81

730 - 280 microns) includes the range Of sizes that are Of interest for fluid transport via
microchannels, including the transport of cooling fluids in TMMS [2-6].

It is important to note that the nominal diameter of the pencil lead (900 microns,
700 microns, 500 microns and 300 microns) is not their actual diameter. SEM
measurements of the pencil lead shown that in fact the diameter Of individual pencil lead
specimens agrees well with the diameter data provided by the suppliers. However, a
careful study of the variation in diameter from one pencil lead to another in a given batch
of lead has not yet been conducted, so that the shrinkage of an internal channel, from the
powder compact (with the pencil lead in place) to the final sintering is probably best
inferred from the overall shrinkage measurements on the specimen.

The size of the pore was measured (by using SEM at high magnification) to be
much smaller than one micron. The purpose of this measurement is to find the effect of
the fugitive phase in the consolidation of the powder and to confirm if the fugitive phase
has any detrimental effect on the integrity of the sintered materials near the channels. Our
finding shows that slightly more pores exist near the channel than away from the channel
as evident in Figures 4.11a and 4.11b. However, all of the pores we observed were small,
isolated pores roughly one micron or less in size. The local variations in the sintering

process must be caused by the presence of the fugitive phase.

82

 

Figure 4.1. A SEM micrograph of an alumina specimen that was pre-sintered at 900°C for
one hour in air in a conventional furnace, then sintered at 1300°C for 1 hour via
microwave heating. The cylinder channel was formed by pressing a 500 micron pencil

lead along with the alumina powder.

83

 

Figure 4.2. A particulate ceramic specimen (25% alumina/75% HAP) with four
cylindrical channels, using 500 micron-diameter pencil lead as the fugitive elements. The
specimen was pre-sintered in air at 900°C for one hour in a conventional furnace, then

sintered at 1300°C for 1 hour via microwave heating.

84

 

Figure 4.3. A high-magnification view of a single channel (formed from a 500 micron
diameter pencil lead) in a particulate ceramic specimen (25% 3mol% Y203-partially
stabilized zirconia/75% HAP). The pre-sintered in air at 900°C for one hour in a
conventional furnace, then sintered at 1300°C for 1 hour via microwave heating. The

specimen was cut such that the "bottom" of the channel is visible in the micrograph.

85

 

Figure 4.4. A fully-stabilized zirconia (FSZ) specimen pre-sintered conventionally at
900°C for one hour in air, then sintered for four hours at 1430°C in air in a conventional
furnace. Both the four internal channels and the two surface channels were formed using

900 micron diameter pencil lead.

86

 

Figure 4.5. A close-up view of one of the channels in the FSZ specimen shown in Figure

4.4.

87

 

Figure 4.6. A FSZ specimen pre-sintered conventionally at 900°C for one hour in air,
then sintered for four hours at 1430°C in air in a conventional furnace. The four internal

channels were formed using 300 micron diameter pencil lead.

88

 

Figure 4.7. Two channels in a FSZ specimen pre-sintered conventionally at 1200°C for
two hours in air. then sintered for four hours at 1430°C in air in a conventional furnace.
The larger channel was formed using 700 micron diameter pencil lead, while the smaller
channel was formed using a 300 micron diameter pencil lead. The cylinder axes of the

two channels are oriented in different directions.

89

 

Figure 4.8. A FSZ specimen pre-sintered at 1200°C for two hours in air, then sintered for
four hours at 1430°C in air in a conventional furnace. Although the specimen included
four internal channels, only two internal channels (formed using 700 micron-diameter
pencil lead) are in the field of view, along with one surface channel (formed using 500

micron diameter pencil lead).

90

 

Figure 4.9. A FSZ specimen pre-sintered at 1200°C for one hour in air, then sintered for
one hour at 1430°C in air in a conventional furnace. Both the surface channel and the
internal channels were formed using 900 micron-diameter pencil lead. Note that two of

the internal channels are within less than one channel diameter of the specimen surface.

91

 

Figure 4.10. Two rows of closely spaced channels in a FSZ specimen pre-sintered at
1200°C for two hours in air, then sintered for four hours at 1430°C in air in a
conventional furnace. The larger channels were formed using 700 micron diameter pencil

lead, while the smaller channels were formed using 300 micron diameter pencil lead.

92

 

Figure 4.11. For a FSZ specimen (pre-sintered and sintered in air at 1200°C for two hours
and for four hours at 1430°C, respectively), where microstructure (a) at a distance of
about mm away form an internal channel is perhaps slightly less porous that the
microstructure (b) at a distance of roughly 200 — 300 microns from the outer wall of an

internal channel.

93

4.4 SUMMARY AND CONCLUSIONS

One processing technique that was developed recently enables to have the channels
in a ceramic material. Crumm et al. [19] fabricated a fugitive compound formed from
ethylene ethyl acrylate resin, acryloid resin, carbon black, and mineral oil. However, they
used in mutliple-pass coextrusion of multicomponent ceramics. Feedstock consisting of
rectangular strips of the ceramic and the fugitive phase were repeatedly co-extruded such
that when firing, the fugitive phase was burned out, Open channels approximately 40
microns wide were formed.

The present study does not require the co-extrusion, thus an economical method to
introduce precise channels and holes into the bulk of the ceramics and ceramic
composites, combining a fugitive phase procedure with a simple powder-pressing
technique. Multiple circular channels with final dimensions of between about 700 and
300 microns were produced in both one and two layers with the specimens. Also, unlike
the work of Crumm et a1. [19], the channels will have diameters of several hundred
microns to allow the unrestricted flow of the cooling fluids. A thennogravimetric analysis
of the fugitive phase removal showed the burnout of the fugitive phase began at a
temperature of about 300°C and was nearly complete at a temperature of 800°C.

As shown in other ceramic processing work by the authors and co-workers [14], the
most significant difference between microwave processed and conventional processed
ceramics is the maximum temperature and time needed for processing. In addition, the
grain size of the microwave-processed specimens tends to be somewhat lower than that

for the conventionally processed ceramics/ceramic composites. However, for this study,

94

the circularity of the channels and the lack of cracking near the channels are very similar
regardless of whether microwave or conventional processing was used.

For metal component fabrication, internal features such as channels can be either cast
or compacted in a mold and then sintered. For ceramic and ceramic composites,
machining and casting are not viable options. Work is presently underway to extend the
techniques presented in this study to accommodate more complicated internal shapes,
such as manifolds in single components. In addition, ceramic/ceramic joining techniques
discussed elsewhere [9; 14-16] can potentially allow a number of individual components

to be bonded into a single, much more complicated assembly.
ACKNOWLEDGMENTS

The authors acknowledge the financial support of the Composite Materials and Structures
Center, College of Engineering, Michigan State University (MSU). The authors also
acknowledge the use of the scanning electron microscopy facilities at the Center for

Advanced Microscopy. Michigan State University.

REFERENCES

l. Jagannadham, K., “Model of Interfacial Thermal Resistance of Diamond
Composites,” J. Vac. Sci. Technol, A17: (2), pp. 373-379, (1999).

2. D. B. Tuckerman and R. F. W. Pease, “High Performance Heat Sinking for VLSI,”
IEEE Electron Device Lett., 2, pp126-129, (1981).

3. G. M. Harpole and J. E. Eninger, “Micro-Channel Heat Exchanger Optimization”
Proc. 7‘“ IEEE semi-Therm. Symp., pp. 59-63, (1991).

4. MB. Bowers and I. Mudawar, “Two-Phase Electronic Cooling using Mini-channel
and Micro Channel Heat Sinks: Part l-Design Criteria and Heat Diffusion
Constraints,” Journal of Electronic Packaging, 116, pp. 290-297 (1994).

5. D. Copeland, “Manifold Microchannel Heat Sink: Isothermal Analysis” IEEE
Transactions on components, packaging and manufacturing technology — part A, 20,
2, pp96-102 (June 1997).

6. S. J. Kim and D. Kim, “Forced Convection in Microstructures for Electronic Equipment
Cooling,” J. Heat Transfer, 121, p. 639, (1999).

7. R. Hopkins, A. Faghri and D. Khrustalev, “Flat Miniture Heat Pipe with Micro Capillary
Grooves," Journal of Heat Transfer, 121,1, pp. 102-109, (1999).

8. K. Take, Y. Furukawa and S. Ushioda, “Fundamental Investigation of Roll Bond Heat
Pipe as Heat Spreader plate for Notebook Computers,” IEEE Trans. Components and

Packaging, 23 80 (2000).

96

9.

10.

11.

12.

13.

14.

15.

K. N. Seiber, K. Y. Lee, and E. D. Case, "Microwave and Conventional Joining of
Ceramics using Spin-on Materials,“ pp. 941-949 in Proceedings of the 12th Annual
Advanced Composites Conference, Technomic Publishing Co., Lancaster, PA, 1997.
E. D. Case, K. Y. Lee, J. G. Lee, and T. Hoepfner, “Geometrical Stability of Holes
and Channels During Joining of Ceramics and Ceramic Composites,” pp. 27 - 34 in
Joining of Advanced and Specialty Materials, M. Singh, J. E. Indacochea, and D.
Hauser, eds., ASM International, Materials Park, OH, 1998.

E. D. Case and J. G. Lee, , "Joining Dissimilar Ceramics and Ceramic Composites for
Biomedical and Structural Applications", pp. 318 - 324, Technology Convergence in
Composites Applications, Proceedings of the ACUN-3 International Composites
Conference, University of New South Wales, Sydney, Australia (2001).

J. G. Lee, H. W. Shin, E. D. Case, P. Kwon, "The fabrication of smooth, submilliter-
diameter channels in polycrystalline ceramics without machining", J. Mater. Sci. Lett.,
20[2]: 107 - 109 (2001).

Y. M. Kong, S. Kim, H. E. Kim, and I. S. Lee, "Reinforcement of hydorxyapatite
bioceramic by addition of Zr02 coated with A1203", J. Am. Cer. Soc., 82: 2963 -
2968, (1999).

J. G. Lee and E. D. Case, “Joining Ceramics to Produce Components with Precise
Internal Channels”, pp. 433 - 442 in Innovative Processing and Synthesis Of Glass,
Composite, and Ceramic Materials HI, Volume 108, Ceramic Transactions, American
Ceramic Society (2000).

L. Zeng, M. A. Crimp and E. D. Case, "The interfacial microstructure Of Zirconia and

MaCorTM joined using spin-on interlayers", accepted, Materials Science and Eng.

97

16. H. W. Shin, E. D. Case, and P. Kwon, "Joining bioactive and bioinert ceramics",
accepted, ASM Fall 2000 Proceedings.

17. K. Y. Lee, E. D. Case, and J. Asmussen, Jr., “The Steady-State Temperature as a
Function of Casket Geometry for Microwave-Heated Refractory Caskets,“ Materials
Research Innovations, 1[2]: 101-116 (1997).

18. K. Y. Lee, L. Cropsey, B. Tyszka, and E. D. Case, "Grain size, density and
Mechanical Properties of alumina batch-processed in a single mode microwave
cavity," Materials Research Bulletin, 32[3]: 287-295 (1997).

19. Crumm, A. T. and Halloran, J. W., “Fabrication of Microconfigured Multicomponent

Ceramics,” J. Am. Ceram. Soc., 81 [4], pp. 1053 — 1056 (1998).

98

Chapter 5

NOVEL POWDER PROCESSING TECHNIQUES TO FABRICATE

EFFICIENT MESO-SCALE HEAT EXCHANGER2

ABSTRACT

This paper reports on novel ceramic processing techniques developed at Michigan
State University (MSU) during the last few years in order to fabricate an active-cooled,
efficient meso-scale heat exchanger. Its unique feature is a network of channels and
manifolds in the micro—textured medium called Functionally Gradient Material (FGM).
FGM is achieved by spatially varying the compositional ratio between two materials. The
gradiency is necessary to compensate thermal gradient loading as a result of heat transfer
into a cooling fluid circulating in the network of channels and manifolds. The gradient
medium also enhances heat transfer from a localized heat source to the cooling fluid,
minimizing the spread of heat energy to other peripheral devices. In order to reach the
goal of making such a heat exchanger, several processing challenges must be resolved.
Using conventional and/or microwave powder processing, the components needed to
achieve the proposed heat exchanger were fabricated. During this process, the residual
stresses caused by spatial gradients in thermomechanical properties are important for the
final integrity of the component. A processing technique to minimize the residual stress is

developed, which uses different powder characteristics to modulate the densification

 

2 H.W. Shin, E.D. Case and P.Kwon “Novel Powder Processing Techniques to Fabricate Efficient Meso-
Scale Heat Exchanger” accepted in North American Manufacturing Research Conference

99

behavior. In addition, when the powders are compacted, a fugitive phase in the shape of
desired channels can be embedded. During sintering, the fugitive phase decomposes
leaving cooling channels in the gradient medium. We will extend this technique to
achieve a complex network of channels and manifolds. When either the CTE differentials
between two materials or the difference in the sintering temperatures between two
materials become too severe, the development of joining technique was necessary to join
these individual components. In addition, because of the geometry complexity of the
network of channels and manifolds as well as the integrity of the material after processing,
the parts or components of the heat exchanger had to be made independently, which
subsequently have to be joined. The incorporation of joining technology was necessary to

achieve our goal.

5.1 INTRODUCTION

Many researchers have attempted to remove heat from microelectronic devices.
The most common method to dissipate the heat [Gray, 2000; J agannadham, 1999 and Hui
et al., 1997] is to spread rapidly by employing highly conductive and high heat capacity
materials such as diamond, silicon nitride, molybdenum, etc. Heat is dissipated to the
environment either by forcing air through pin arrays or fins or by cooling naturally.
Recently MEMS-sized heat exchangers [Hal], 1999] have emerged, which rely on
convective heat transfer via air. In such cases, heat inevitably spreads to its surrounding
and other peripheral devices.

More efficient cooling results when coolants circulated through networked

channels and manifolds in a closed system. Tuckerman and Pease [1981] made a

l 00

unidirectional micro-channeled heat sink whose channels were produced by etching
<110> silicon wafers using KOH on a 1cm2 silicon wafer. The distance between the
channels was the same as the width of the channels. The channels circulating deionized
water were covered using a Pyrex plate by anodically bonding over the channels. These
channels were slightly modified to form pins [Yin and Bau, 1997] to enhance the heat
transfer into the coolants. Zhong and Bau [2001] cut low-temperature, cofired ceramic
tapes into shapes to make the channels before stacking the tapes and firing. Harpole and
Eninger [1991] and Copeland et al. [1997] added a simple manifold by covering only the
middle section of the channel, where one side of the uncovered section is used as inlet
and the other side as outlet. Bower and Mudawar [1994] and Gillot et al. [1999]
developed a two-phase heat exchanger with a higher efficiency. Bower and Mudawar
[1994] made mini- and microchannels with cross-sectional diameters of 2.54mm on a
block of oxygen-free copper and 510Cm on the block made of copper and nickel plate,
respectively. Gillot et al. [1999] produced channels with a circular cross-section, a
diameter of 2mm and length of 40mm. This approach is similar to a Heat Pipe [Nguyen et
al., 2000], where coolants such as fluorocarbon FC72 and water undergo a cycle of
evaporation and condensation to efficiently remove the heat generated by electric
components.

This paper reports on the essential processing techniques for the development of
an efficient, meso-scale heat exchanger. To achieve this, we propose the body of the heat
exchanger to be: (1) micro-textured and (2) micro-configured. Micro-textured or FGM
material can be processed by spatially modulating the compositional ratio between two

powders. Micro—configured medium represents the network of micro-scale channels and

101

manifolds. The network will be eventually employed to transport coolants for our heat
exchangers. Applications for the heat exchanger include a variety of other industrial or
medical uses, such as conduits for fluids in engines or to carry medicine or biological
fluids.

Achieving the aforementioned features hinges on the economical, appropriate
processing technologies especially for hard-to-shape material like ceramics and small-to-
make devices. The developed techniques must facilitate net-shape processing and
minimal post-processing of the material. Compared to selective laser sintering process,
the presented processing methods require conventional powder processing equipment
such as a furnace and a press. Selective laser sintering also requires the sophisticated
equipment that has laser source, computer control and several motors to control powder

feeding.

5.2 BACKGROUND

In the micro-textured and micro-configured body, heat is conducted and
dissipated via the coolant circulating through the holes, channels and manifolds (Figure.
5.1). We will report on the novel processing techniques that will enable us to fabricate the
heat exchanger; (1) processing techniques to fabricate internal channels into an FGM in
order to produce micro-configured body, (2) powder processing techniques to produce
micro—textured Thermal Management Materials (TMMS) undamaged by process-induced
residual stress, and (3) joining techniques to laminate the consolidated FGM layers.

A key issue in producing micro-textured medium (called FGMS) is to avoid

damage due to process-induced residual stress. Residual stresses in FGM’s [Zhang et al.,

102

1995 and Mizuno et al., 1995] occur due to spatial variations in Coefficient of Thermal
Expansion (CTE) and densification behavior during processing. A powder processing
protocol to minimize process—induced residual stress in FGMS has been developed, which
can produce a mild gradiency in FGMS. For severe gradiency, two or more FGMs, each
with mild gradiency, must be joined. The development of a joining protocol for two or
more layers is presented. In addition to the FGM processing, another novel processing
technique developed is the introduction of holes, channels and complex manifolds,
without machining. Such channels can be straight or curved, with constant or variable
cross section. Channels were fabricated on the specimen surface as well as in the interior
of the ceramic. Two or more ceramic sub-components with the surface channels were
joined subsequently to form internal channels in the final component. Ceramic/ceramic
joining could be used, for example, to laminate the thin densified FGM layers.

An FGM ideally has continuous gradiency as shown in Figure 5. 2 (a). One can
approximate such gradient properties by preparing subcomponents in the form of layers.
Each layer is prepared by homogeneously mixing two powders in various ratios, which
can be stacked to form the ‘discrete’ gradiency in thermomechanical properties as shown
in Figure 5.2(b). However, if the process-induced residual stresses can be controlled, each
layer can sustain some gradiency in properties. A smaller number of layers with some
gradiency are then needed to form an FGM as depicted in Figure 5. 2 (c). The techniques
presented in this paper allow us to control the residual stress and to fabricate components
with less number of layers to form FGMS. This paper presents the development leading
up to the arrangement similar to Figure 5. 2 (b) and the developments are currently

underway leading to the arrangement similar to Figure 5. 2 (c).

103

As stated in the previous section, controlling residual stress is a key technological
problem in making multi-layer media as well as FGM components shown in Figure 5.2
(c). Without controlling the residual stress, the processed media would be useless as
shown in Figure 5.3. In order to resolve this problem, the difference in two shrinkage
mechanisms existing in powder processing must be utilized; the effect of CTE and
densification behavior. During densification, the interstitial spaces among the powder are
filled due tO mass transport mechanisms. The shrinkage differences due to CTE mismatch
are intrinsic for a given combination of materials after consolidation. However, the
shrinkage from the densification of powder can be altered by designing the powder
characteristics and optimizing the sintering conditions such as time, temperature,

heating/cooling cycles, etc. [German, 1984].

104

 

Figure 5.1: the micro-textured and configured body.

 

(a)ldeal FGMs (b)Discrete FGMs (c)FGMs

Figure 5.2: Three processing routes to attain FGMS

105

500011111

 

Figure 5.3: A multiple-layer sample with a rectangular cross-section channel affected by

residual stress.

106

The starting powder characteristics and sintering conditions can compensate for the
CTE difference by exploiting the differences in densification behavior among the layers.
Powder shrinkage during densification must be understood as a function of various
processing variables such as the powder characteristics as well as sintering conditions.
The powder characteristics influence the initial powder packing density, which in turn
affects the densification behavior during sintering. As shown in Reed [1989] with
alumina powder, different combinations of the powder characteristics result in different
packing densities.

One method to control the initial powder packing is by mixing a few batches of
powders with different average particle sizes as demonstrated in Figure 5. 4. Yeh and
Sacks [1988] and Ting and Lin [1994 and 1995] studied shrinkage difference between
wide and narrow size distribution (WSD and NSD) powders. They Show that WSD and
NSD powder exhibit not only different initial packing of powder but also different
densification rates. In the experiments [Kwon, 1998] conducted in our laboratory, two
different powders, AA-S and AKP-15 (manufactured by Sumitomo Chemical) were
compacted with a uniaxial press. Packing density of 59.1% was attained with all AA-5
powder (an average particle size of 5.5-5.8’. m). The AA-5 powder mixed in 1:3 ratio
with the AKP-IS powder (whose average particle size is 0.6-0.8Em) yielded the packing
ratio of 71.2%. McGeary’s [1961] experiments with spherical particles show higher
packing by mixing ‘designed’ size particles because smaller particles can nest
interstitially among bigger particles. Irregular shaped powder can also be used to achieve

low packing and different densification behavior.

107

Powder Preparation

Four different powders, alumina, Partially-Stabilized Zirconia (PSZ) and two
kinds of Fully-Stablized Zirconia (FSZ), were used. The alumina powder designated TM-
DAR is manufactured by Taimai Co., Japan, whose average size is about 0.2 micron. The
PSZ powders manufactured by Tosoh designated TZ-3YS, whose average size was about
0.6 microns. Two FSZ powders were designated TZ-8YS manufactured by Tosoh Corp.,
Japan and 5.2 weight percentage (w/o) Y203 zirconia manufactured by CERAC,
respectively. The average size of the TZ-8YS powder was about 0.58 microns. The

average size of the CERAC powder was about 1.23 microns (90% <1.79 um, 50% <l.23
11m and 10% <0.77um). The CTES of alumina and zirconia are close at 8.8x106and

7.6x10-6/°C, respectively. In a more practical sense, the CTE data as a function of
temperature for these materials [Purdue University, 1970] must be used. The CTE-
induced shrinkage can be compensated with the shrinkage due to densification in making
multi-layers and FGMS. In addition, the composite layers are needed to make the discrete
FGM shown in Figure 5. 2(b). The mixing of powders with two or more components was
performed by dry ball milling for approximately 24 hours in a plastic ball mill with
alumina media. The ball mill charge consists of approximately 50 grams of powder.
Pressing

A uniaxial steel die was used to press all specimens in this study at a pressure of
approximately 35 MPa. Typically the thickness of pressed disk in a green (unfired) state
ranges from 2 to 3 mm. In preparing for subcomponents of FGM, various ‘composite’

powder mixtures of TM-DAR and P82 were prepared. A sub-layer of one mixture, '

108

another sub-layer of another mixture and so on were deposited layer by layer in the die
and pressed to form the discrete FGM.
C ontrolling Residual Stress

For a given set of two materials, the extreme condition for residual stress is
expected when two powders are processed together. Initially they are processed
separately to understand their shrinkage behavior individually without modulating the
starting powder. Table 1 shows the dimensional changes after each processing step. TZ-
3YS and TZ-8YS powders shrink substantially more than alumina TM-DAR powder
after final sintering. However, after pre-sintering (an essential step to burn out fugitive
phase to make channels), the dimensional differences are minute.

The sintered samples are subject to warping since the TM-DAR layer is in a state of
compressive residual stress while the TZ-3YS is in a state of residual tensile stress based
on the shrinkage rate reported in Table 5.1. In McGeary’s [1963] experiments with
spherical particles, the diameter ratio between larger and smaller particles was 6.5:1.
However, the powders we used had relatively wide size distribution and were non-
spherical in shape.

To understand and control the residual stress, two-layer samples, one layer with
TM-DAR and the other with a various mixture of TZ-3YS and CERAC powders, are
pressed together and sintered. All of the samples had some degree of warping as shown in
Figure 5.5 with the mixture ratios. The minimum warpage can be observed in C
(65T/35C) and D (SOT/50C) showing the total height of the warped samples. Based on
the observation, the ideal mixture of 45% CERAC and 55% TZ-3YZ is chosen to make

an ideal two-layer (alumina and P52) sample.

109

Table 5.1: Dimensional changes during processing (Units are in cm).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. After pre After final Linear Shrinkage
Material After Pressing 31 nt e n n 9 51 nt e ri n 9 Rate
TMDAR 0:219 D=2.16 o: 1.85 15.5%

T = 0.273 T = 0.272 T = 0.205 24.9%
T2-3YS (p32) 0 = 2.19 o = 2.19 o = 1.685 23.0%
T = 0.33 T = 0.318 T = 0.244 26%
TZ-8YS by Tosoh D = 2.19 D = 2.19 D = 1.68 23.3%
(FSZ) T = 0.33 T = 0.317 T = 0.24 27.3%
Narrow Size _
Distribution H1811
(NSD) Pow r Shnnkage
I
Low CTE Material
Wide Size Low
Distribution Shrinkage

(WSD) Powder

 

 

Figure 5.4: Shrinkage control using different powder size distribution.

110

C :65'I‘/35C E:25T/75C

~.

 

Figure 5.5: Five samples with two layers (alumina and zirconia mixed in five different

ratios) showing varying degrees of warpage.

Table 5.2: Height difference of the five samples on figure 5.5

 

A l B | C [ D | E
1 Height (cm) I 0.05 T 0.03 l 0.03 I 0025* J 0.09 I

 

 

*Specimen destroyed in a fracture test and value extrapolated from Figure 8.2

111

Channels and Manifolds

 

Channels that penetrate the bulk Of the specimen were formed using a “fugitive
compound” added to the ceramic powders before firing. Among many fugitive phases
used, polymers, papers and graphite work well. For circular channels, pencil leads with
nominal diameters of 700, 500 and 300 microns were sectioned into 1 cm to 1.5 cm
lengths prior to embedding them in the ceramic power. Prior to pressing, placing the
fugitive phase on top of the ceramic powder produces surface or external channels while
embedding them in the ceramic powder produces internal channels. To make internal
channels, part of the powder was placed in the die before placing the fugitive phase and
then the rest of the powder is placed into the die. The same process was used for other

fugitive phases. The presintering was conducted in air for three hours in an electric
furnace to bum out the fugitive phase, with an average heating rate of roughly 5°C per

minute to 900°C, and then a cooling rate of 10°C per minute. During heating, the
specimens were placed on a thin powder bed of the same material as the specimen being
sintered.

We have fabricated internal channels in a variety of ceramics and ceramic
composites, including alumina, zirconia, and hydroxyapatite [Shin et al. 2000 and Lee et
al., 2000]. The alumina/zirconia composite samples with internal channels are shown in
Figure 5.6, 5.7 and 5.8. These channels are completely internal, not exposed to any
surfaces. The samples shown in Figure 5.6, 5.7 and 5.8 were cut after processing using
low-speed diamond. Closed internal channels were formed when two specimens with
surface channels were joined [Lee and Case, 1999b]. Virtually all the samples with the

surface channels made to date were successful without any process-induced defects even

112

very complex shapes as shown in Figure 5.9 and 5.10. However, only pencil leads show
consistent success as a fugitive phase for all the samples with internal channels. These
samples can be joined to form the channels penetrating the bulk of ceramic.

Interesting points can be made with Figure 5.7 and 5.8. Figure 5.7 shows the
ridges along the channel, which are the exact ridges on pencil lead like a tree trunk. The
surface texture of the fugitive phase is exactly reprinted on the channel. On Figure 5.8,
two channels at the top right hand comer are half of the diameter apart. The manifold-
shaped channel shown in Figure 5.9 was achieved by cutting a paper into the shape of
manifold and embedding on the powder before pressing. The spoke-shape manifold was
achieved by making the spoke-shaped fugitive phase using rapid prototyping machine.
Sintering

Both microwave processing and conventional sintering were used to fabricate our
specimens. For microwave processing, a refractory casket (specimen enclosure) serves
both to thermally insulate the specimens and to provide a microwave susceptor material.
Microwave processing

The specimens that were sintered using microwave energy were processed at
temperatures between 1300°C and 1450°C for up to one hour. The Specimen temperature
was measured by an optical pyrometer that was sighted on the specimen being sintered.
The microwave power supply generated up to 2 kW of power at a frequency of 2.45 GHz.
The specimens were heated in a single-mode microwave resonant cavity having both an
adjustable short and an adjustable launch probe position. Details of the sintering

procedure are given elsewhere [Lee et al., 1997, Lee and Case, 1999a].

 

Figure 5.6: Multiple channels on a composite of 75%TMDAR and 25%TZ-8YS.

114

 

Figure 5.7: A channel Showing the micro-ridges (the surface feature of a pencil lead) on

alumina sample.

115

- can.
'-. .n»

[11203—21402 ‘
composite 10mm

 

Figure 5.8: Channel formed by joining an alumina-zirconia composite specimen to a

partially stabilized zirconia specimen [Lee, 2001].

116

 

Figure 5.9: Manifold—shaped surface channel created by the graphite sheet cut out with a

wire edm.

ll7

5000pm

 

Figure 5.10: A spoke-shaped surface manifold created by rapid prototyping unit.

118

Conventional sintering

An electrical furnace with SiC heating elements (Carbolite, Corporation) was
used to conventionally process the specimens. A furnace with a controlled atmosphere
was used to sinter some of the oxide materials, while in a future controlled atmosphere
furnaces will be used for the nitride materials. A number of researchers have observed
that the temperatures required for “equivalent sintering” via microwave and conventional

heating can be substantially different [Lee et al., 1999a]. TM-DAR powders were fired
conventionally at a heating rate of approximately 5°C per minute and a 3-hour hold time

at 1430°C. As with the microwave sintering, during conventional sintering multiple-
specimen batches can be sintered, with from at least six to eight specimens in a given
batch.
Jaimg

The authors and co-workers have joined like ceramics such as alumina/alumina
[Lee et al., 1997], zirconia/zirconia [Lee and Case, 2000], SiC/SiC, HAP/HAP [Lee and
Case, 2000], and MaCorTM/MaCorTM [Lee et al., 1997] as well as dissimilar ceramics
such as alumina/zirconia [Zeng et al., 2001], HAP/MaCorTM[Shin et al., 2000],
MaCorTM/zirconia and a variety of other ceramic/ceramic combinations. Here HAP is a
bioceramic hydroxyapatite and MaCorTM is a machineable glass ceramic composite
composed of an aluminosilicate glass matrix reinforced by mica platelets.

In order to prepare the sintered specimen surfaces for joining, the surfaces to be
joined were first polished and cleaned. Then a coating of amorphous silica was applied to
the polished surfaces. The polishing was done with diamond paste having an abrasive

size as small as one micron. The polished specimens were then ultrasonicated in a

119

deionized water bath to help remove debris that accumulated during polishing. A coating
of an organically-based silica precursor liquid was then spun onto the surfaces that were
to be joined. The coatings were then converted into a thin (roughly 200 nm thick) layer of
amorphous silica by heating the coated specimen in air for approximately 20 minutes at
200°C. Joining was done at temperatures of roughly 1300°C - 1450°C using the same
microwave cavity and power supply that had been used to sinter the powder compacts.
Sintered disks with surface channels can be joined with other disks to produce channels at
the interface between the two disks.

5.3 CONCLUSIONS

Being able to micro-configure a ceramic material will impact the way we design and
manufacture ceramic components. In addition, with the current research emphasis on
micro—devices, we recognize that the immediate applications of the proposed processing
techniques are in micro—heat exchangers for microelectronics and cryogenic surgery.
Because of the inherent size of such a device, the effective use of materials is necessary.
Texturing a material optimally based on a loading condition is the key element that
triggered the development of FGMS. The most difficult problem with FGM is to control
process-induced residual stress such that residual stresses would not damage the details

(like channels) as well as the material itself.

This work is applicable to high temperature applications where the cooling is essential
such as advanced semiconductor device and cutting tools. The coolant circulating through
the channels and manifolds creates thermal gradient in the medium. And thermal gradient

can be compensated by the micro-textured medium such as FGMS.

120

ACKNOLWDGEIVHENTS
The authors acknowledge the assistance of Michigan State University graduate student

Jong-Ge Lee with the micrographs.

121

REFERENCES

Bower, M. B. and Mudwar, I., 1994, J. Electron Package, 116, pp. 290-297.

Copeland, D., 1997, IEEE Trans. Comp. Pack. Manufi Tech. -— part A, 20, 2, pp96—102.
Gillot, C., Meysenc L., Schaeffer, C. and Bricard, A., 1999, IEEE Trans. Comp. Pack.
Tech, 22: (3), pp. 384-389.

Gray, K. J ., 2000, Diamond and Related Materials, 9: (2), pp. 201-204

German, R. M., 1984, Powder Metallurgy Science, Metal Powder Industries Federation,
Princeton, New Jersey.

Hall, R., 1999, Electronic Design, 47: (7), pp. 29-30.

Harpole, G. M. and Eniger, J. E., 1991, Proc. 7th IEEE Semi-Therm. Symp., pp. 59-63.
Hui, P., Tan, H. S. and Lye, Y. S., 1997, IEEE Tran. Comp. Pack. & Manuf. Tech. Part
A, 20: (4); pp. 452-457.

J agannadham, K., 1999, J. Vac. Sci. Technol, A17: (2), pp. 373-379

Kwon, P., 1996, Unpublished Results.

Lee, K. Y. and Case, E. D., 1999a, Materials Science and Engineering, A269: 8 - 20.
Ice, J. G. and Case, E. D., 1999b, Ceramic Transactions, American Ceramic Society,
103: 571-581.

Lee, J. 0., Case, E. D., Shin, H. and Kwon, P., 2000, Ceramic Engineering and Science
Proceedings, 21[4]: 103 - 110, American Ceramic Society.

Lee, J. 6., Shin, H. W., Case, E. D., Kwon, P., 2001, Journal of Materials Science

Letters, 20, 2, pp. 107-109.

Lee, K. Y., Case, E. D., and Reinhard, D., 1997, Ceramic Eng. and Sci. Proc., 18: 543-
550.

McGeary, R. K., 1961, J. Am. Ceram. Soc., 44, 10, pp. 613-522.

Mizuno, Y., Kawasaki, A., and Watanabe, K, 1995, Metall. Mater. Trans, 26B, p. 75.
Nguyen T, Mochizuki M, Mashiko K, et al., 2000, IEEE Tran. Compon. Pact. T., 23: (1),
pp. 86-90.

Purdue University, Thermophvsical Properties of Matter, V.13, Ed. Y.S. Touloukian,
R.K. Kirby, RE. Taylor and T.Y.R. Lee, [Fl/Plenum, New York, 1970.

Reed, J. S., 1989, Introduction to the Principles of Ceramic Processing, Wiley & Sons,
New York, NY.

Shin, H. W., Kwon, P. and Case, E. D., ASM Fall 2000 Proceedings.

Ting, T. -M. and Lin, R. Y., 1994, J. Mat. Sci, pp. 1867-1872.

Ting, T. -M., and Lin, R. Y., 1995, J. Mat. Sci., pp. 2382-2389.

Tuckerman, D. B. and Pease, R. F. W., 1981 IEEE Electron. Devices Lett., vol. EDL-2,
pp. 126-129.

Yeh, T. and Sacks, M. D., 1988, J. Am. Ceram. Soc., 71, 12, pp. C484-V487.

Yin, X. and Bau, H. H., 1997, Journal of Electronic Packaging, 119, pp. 51-57.

Zhang, L. M., Yuan, R. Z., Oomori, M. and Hirai, T., 1995, J. Mat. Sci. Let, 14, 22, p.

1620.

Zeng, L., Crimp, M. A. and Case, E. D., 2001, Materials Science and Engineering,

A307: 74 - 79.

Zhong, J. and Bau, H. H., 2001, American Ceramic Society Bulletin, 80, 10, pp. 39-42.

123

Chapter 6

JOINING OF BIOACTIVE AND BIOINERT CERAMICS3

For a number of years, ceramic materials have been of interest as permanent hard
tissue implant materials [1] such as bone screws, dental implants [2], bone replacement
[3], and as coatings on metallic implant devices [4]. For example, a number of
commercial orthopedic manufacturers produce implant materials that include plasma-
sprayed hydroxyapatite (HA, Ca10(PO4)6(OH)2) coatings. Typically these ceramic
coatings have been placed on metallic alloys, such as surgical grade Ti-6Al-4V [4—6].

Hip replacement operations are a relatively frequent procedure, with more than
200,000 Total Hip Arthroplasties (THA) performed annually in the United States.
Ceramic-coated implants have the prospect of improving the long-terrn outcomes of the
hip replacement surgeries [4]. For example, bone ingrowth that occurs in the presence of
bioactive ceramics such as hydroxyapatite can overcome the fixation problems that are
associated with the cracking of cemented implants [7] (implants that are cemented into
place). However, bioceramics are now being used in medical applications that go far
beyond their service as a coating material for metals such as Ti-6Al-4V.

A bioinert ceramic does not induce HA to grow on its surface when it is exposed to

blood or to a simulated biological fluid. In addition, a bioinert ceramic does not generate

 

3 Summarized from H. W. Shin, E. D. Case and P. Kwon, “Joining of Bioactive and Bioinert Ceramics”,
Proceedings from Joining of Advanced and Specialty Materials, ASM International 2000, pp.15-22, and
H. W. Shin, P. Kwon, and E. D. Case, “Fabrication of Internal Channels without Machining in Joined
Alumina and Zirconia Ceramics”, Proceedings from Joining of Advanced and Specialty Materials, ASM
International 2000, pp.23-30.

124

corrosion products when exposed to blood or to a simulated biological fluid (although for
the typical surgical alloys such as Ti-6Al-4V, corrosion products can accumulate in the
tissues near the implant site).

Alumina and zirconia are two bioinert, polycrystalline oxide ceramics included as
materials that go into hip replacement units, for example. Either alumina or zirconia is
used in the joint region of the hip replacement, where the good wear properties and
chemical inertness of these ceramics are very beneficial. In fact, bioactivity would be
counterproductive for implant materials that are included in joints. Because of their
superior hardness and fracture toughness, alumina and zirconia also find use in
engineering applications, e. g., thermal management materials.

In addition to implant materials, bioinert ceramics are being examined as potential
materials for dental restorations and tooth implants [8—10]. Glass ceramics are one
particular class of ceramics that is being considered for the exterior portions of dental
restorations [8]. Glass ceramics are formed by the controlled nucleation and growth of
crystalline phases from a glass melt [11, 12]. If the crystalline phase that precipitates in
the glass ceramic is a micaceous phase, then the mica can promote the formation of small,
localized microcracks during machining, rather than forming large, catastrophic cracks.
MaCorTM and other glass ceramics that contain randomly oriented mica platelets may be
machined using metal-working tools, thus making such micaceous glass ceramics a
potentially versatile element in the fabrication of biomedical materials.

In this chapter, we present our results on the joining of the bioactive ceramic
hydroxyapatite and the bioinert ceramics alumina, zirconia, and MaCorTM. The challenge

is to join two or more layers of materials with distinct coefficient of thermal expansion

125

(CTE). Joining techniques are necessary when an assembly from simpler components is
necessary and when severe warping and cracks are developed in simultaneously sintered
composite layers.

6.1 Materials and Preparation

With the exception of the MaCorTM, all specimens included in this section were
sintered using commercial hydroxyapatite, zirconia, and alumina powders. The alumina
(A1203) powder, designated by TM—DAR, is from Taimei Chemicals Co. The two types
of zirconia powders used in this study include a partially stabilized zirconia (PSZ,
zirconia powder with 3 mol % yttria), designated by TZ-3YS, from the Ceramic Division
of Tosoh Corp., Japan, and a fully stabilized zirconia (FSZ, zirconia powder with 8 mol%
yttria), designated by TZ-8YS, also from Tosoh’s Ceramic Division. The specific surface
areas of both the PSZ and the FSZ were approximately 7 mzlg. The chemical purity was
about 99.9 percent for both the partially-stabilized zirconia (three mol % yttria-zirconia)
powder and for the fully-stabilized zirconia (eight mol % yttria-zirconia powder). The
hydroxyapatite (Cerac Inc., Specialty Inorganic, Milwaukee, WI) was nominally 99%
pure. The average particle size for each of the powders and its abbreviated name used in
this paper are listed in Table 6.1.

In addition to the specimens that were sintered from the commercial powders,
specimens were cut from billets of the machineable glass ceramic, MaCorTM (Corning
Code 9658, Corning NY). MaCorTM consists of a borosilicate glass matrix phase with a
reinforcing phase of fluorophlogopite mica, where the mica is in the form of randomly
oriented platelets. By volume, the mica platelets make up about 55 volume percent of

MaCorTM, with the borosilicate glass comprising the remaining 45 volume percent.

126

In this set of experiments (prior to October, 2000), the disc-shaped powder
specimens are pressure-compacted to approximately 2.2 cm in diameter in the green
(unfired) state (as compared to 2.25 cm diameter in tests conducted after October, 2000).

The powder compacts were then sintered using either a 2.45 GHz single-mode
microwave cavity or a conventional, electrically-heated furnace (see Chapter 3). During
microwave heating, the specimens were placed in a refractory specimen enclosure (a
refractory casket) and then sintered in air for the temperatures and times given in Table
6.2.

During processing, the specimen temperature was monitored by an optical pyrometer.
The conventional furnace (Carbolite Inc., with a Eurotherrn temperature controller)
needed longer sintering temperatures and times to accomplish the same degree of
sintering as obtained with the microwave furnace (Table 6.2). This difference in
sintering behavior is typical for microwave and conventionally sintered ceramics [13—14].

After sintering, specimens were polished using diamond paste, with the series Of
grit sizes 35, 25, 17, 10, 6, and 1 microns. In order to inspect the channels, the specimens
were cut with a low speed diamond saw. The channels and the specimen microstructure
were then examined using optical and scanning electron microscopy.

The reader is referred to Chapter 3 for the types and composition of ceramic
powders used, mixture ratios used for composite specimens, procedures for powder

mixing of composites, specimen pressing, pencil lead as the choice of fugitive phase and

127

Table 6.1: Summary of powder average size

 

 

 

 

 

 

Powder Material Average Particle
size (um)
TMDAR A1203 0.2
TZ-3YS 3 mol% PSZ 0.5
TZ-8YS 8 mol% PSZ 0.54
HAP Hydroxyapatite 2-3

 

 

 

 

Table 6.2. For the specimens included in this study, the furnace type, sintering
temperatures, sintering times and channel descriptions for channels fabricated without

machining.

 

 

 

 

 

 

 

 

 

Material Furnace Type Sintering Sintering Time Channels
Temperature (Hours)
(°C)
100% HAP Microwave 1300 1 4 Internal
75%Alumina+ Microwave 1300 1 4 Internal
25%HAP
25%Alumina+ Microwave 1300 l 4 Internal
75%HAP
75% Zirconia+ Microwave 1300 1 4 Internal
25%HAP
25%Zirconia+ Microwave 1300 1 4 Internal
75%HAP
100%HAP Microwave 1300 1 4 Internal
+
4 External
100%HAP Conventional 1375 4 3 Internal

 

 

 

 

 

 

its embedding configurations, temperatures and hold times for pre-sintering and full
sintering, method used for polishing and joining, as well as information on the pressing
and sintering equipment. It is worth noting that a polymer line (fishing thread) has also
been tried as a fugitive phase. However, under the high heat of pre-sintering, the polymer
phase melts before it evaporates, causing the channel shape to lose its circular cross-
sectional shape. In addition, the difference in elastic rebound properties between
polymers and ceramics leads to internal fractures in the both the pre-sintered and fully
sintered specimens. Even shrinkage tapes used in electric work have been considered as a
fugitive phase at one point. However, they exhibit anisotropy in Shrinkage; e.g. one of
the dimensions shrinks while another expands, which leads to fracture. Graphite/pencil
lead was finally chosen as the fugitive phase.

The overview of the process route associated with the joining of ceramic layers

with internal channels is depicted in Figure 6.1.

'63..

Loose Compacted Presintering Fully Joining

   

Powde Powder (Burnout Sinterin g
Fugitive
Phase)
Powder
Pressing

Figure 6.1: Overview of processing route

129

6.2 Results and Discussion

The average grain size of the sintered specimens was determined using the linear
intercept method on SEM micrographs of fracture sections of the sintered specimens
(Figure 6.2). The mean grain size of the alumina and zirconia specimens ranged
approximately from one to two microns.

The 0.5-mm diameter and the 0.3-mm diameter fugitive-phase rods produced
cylindrical channels with a circular cross section (Figure 6.3). Using a low speed
diamond saw, the sintered specimens were sectioned along planes perpendicular to the
cylinder axes of the internal channels (Figure 6.3). Following sectioning, selected
specimens were mounted and polished with diamond paste with grit sizes down to 100
microns.

For the 0.5-mm diameter fugitive-phase rods, the final, as sintered channels were
425 microns in diameter. Using these fugitive-phase rods, internal channels were
successfully formed in pure alumina, 3 mol% yttria-zirconia, and 8 mol% zirconia, using
both conventional and microwave heating. In addition, pure hydroxyapatite specimens
were fabricated, along with various hydroxyapatite/alumina and hydroxyapatite/zirconia
particulate composites. The sintering temperature, sintering time, and chemical
composition of each of the hydroxyapatite-containing specimens is listed in Table 6.2. It
should be noted that hydroxyapatite/alumina and hydroxyapatite/zirconia particulate
composites could be bioactive while having superior mechanical properties compared to
single-phase hydroxyapatite materials [15—17]. Low magnification SEM and optical

microscope examination of the sectioned specimens (Figure 6.4) showed the relative

130

placement of the multiple internal channels. Figures 6.5 and 6.6 are SEM micrographs of
two different regions of a joined hydroxyapatite/h/IaCorTM specimen.

Fig. 6.7 shows the joined interface between a homogeneous layer of zirconia (TZ-
8YS) and a composite layer made of alumina and zirconia (25% TMDAR/75% TZ-3YS).
Prior to joining, one side of each specimen was polished using diamond paste of
gradually reducing, mean abrasive size from 35 to 111m. These specimens were spin-
coated with silica film at a rotational speed of 3000 rpm for 20 seconds. Then, they were
cured in a furnace at 200°C for 20 min. Finally, they were joined in our microwave unit at
1475°C for 20 minutes with a heating and cooling rate 10°C/min. A dead weight of 11.75
g was used to press the two specimens together. After they were joined, the specimen was
cut at a low speed using diamond saw.

For the sectioned specimens containing the multiple internal channels, no
cracking around or near the channels was noted upon either optical or electron
microscopic examination of the sectioned surfaces, for both surface preparations (namely,
the as-diamond saw cut specimen surfaces or the sectioned and then diamond—polished
specimen surfaces).

Thus, pressing the fugitive-phase rods into the powder compacts, and then
burning out the fugitive elements prior to sintering was a successful method of
fabricating internal circular channels in both bioactive and bioinert ceramics.

To extend the potential for the use of these techniques, one can envision uses in
which bioactive materials and bioinert materials would be joined to form a single
structural elements. In this study, we successfully joined MaCorTM specimens with pure

hydroxyapatite specimens having three internal channels.

131

 

Figure 6.2. A SEM micrograph of the fracture of an alumina specimen sintered from
commercial powders (TM-DAR, Tamaii Corp., Japan). The specimen was pre-sintered
for one hour at 900°C in a conventional furnace in order to burn out the fugitive phase.
Then, the specimen was sintered at 1300°C in a microwave cavity

 

Figure 6.3. An SEM micrograph of a internal channel in a microwave sintered alumina
specimen. A fugitive—phase rod 0.5 millimeters in diameter was used to initially form the
channel in the powder compact.

133

 

Figure 6.4. For a alumina/hydroxyapatite particulate composite (25 weight percent
alumina, 75 weight percent hydroxyapatite), a low speed diamond saw cut reveals the
four cylindrical channels formed internally in the specimen.

134

 

Figure 6.5. An SEM micrograph of a sectioned hydroxyapatite specimen showing one of
the internal channels formed without machining using the fugitive—phase rod elements.
After sectioning on the low speed diamond saw, the specimen surface was polished using
diamond paste.

135

0.2mm

 

Figure 6.6. For the joined hydroxyapatite/MaCorTM specimen, the region near the
interface between the two materials. Note that a portion of one of the channels is also
included in the micrograph.

136

 

Figure 6.7: The Interface of the Joined Sample made of TZ-8YS and 25% TMDAR / 75%
TZ-3YS Joined at 1475°C for 20 min. in our microwave unit.

137

ACKNOWLEDGMENTS

The authors acknowledge the financial support of the Composite Materials and
Structures Center, College of Engineering, Michigan State University, East Lansing, MI.
The authors also acknowledge the use of the scanning electron microscopy facilities at
the Center for Advanced Microscopy, Michigan State University. In addition, the authors
acknowledge the assistance of Mr. Jong-Gi Lee, a graduate student in the Materials
Science and Mechanics Department at Michigan State University, for his assistance in

preparing some of the specimens included in this study.

138

REFERENCES

I.

T. L. Seitz, K. Noonan, L. L. Hench, N. E. Noonan, “Effect of Fibronectin on the
Adhesion of an Established Cell Line to a Surface Reactive Biomaterial,” Journal of
Biomedical Materials Research, 1982, 16: 195-207

H. Kawahara and M. Hirabayashi, “Single Crystal Alumina for Dental Implants and
Bone Screws”, Journal of Biomedical Materials Research, 1980, 14: 597-605

A. Uchida, S. M. L. Nade, E. R. McCartney, and W. Ching, “The use of ceramics for
bone replacement”, Journal of Bone and Joint Surgery, 1984, 66-B: 269-275

S. H. Maxian, J. P. Zawadsky, and M. G. Dunn, “In Vitro Evaluation of Amorphous
Calcium Phosphate and Poorly Crystallized Hydroxyapatite Coatings on Titanium
Implants,” Journal of Biomedical Materials Research, 1993, 27: 111-117

Y. Fujita, T. Yamamuro, T. Nakamura, T. Kitsugi, and S. Kotani, “Mechanism and
Strength of Bonding Between Two Bioactive Ceramics in Vivo,” Journal of
Biomedical Materials Research, 1992, 26: 1311-1324

E. P. Paschalis, Q. Zhao, B. E. Tucker, S. Mukhopadhayay, J. A. Bearcroft, N. B.
Beals, M. Spector, and G. H. Nancollas, “Degradation Potential of Plasma-sprayed
Hydroxyapatite-coated Titanium Implants,” Journal of Biomedical Materials
Research, 1995, 29:1499-1505

N. C. Blumenthal and V. Cosma, “Inhibition of Apatite Formation by Titanium and
Vanadium Ions,” Journal of Biomedical Materials Research: Applied Biomaterials,

1989, 24, Al: 13-22

. Y. O. Jung, 1. M. Peterson, D. K. Kim and B. R. Lawn, “Lifetime-limiting strength

139

degradation from contact fatigue in dental ceramics,” J. Dent. Research, 2000, 79:

722-731

9. G. J. P. Fleming, R. M. Shelton, P. M. Marquis, “The influence of clinically induced

10.

ll.

13.

I4.

15.

variability on the bi-axial fracture strength of cemented aluminous core porcelain
discs”, Dent. Materials, 1999, 15: 62-70

P. Magne, K. R. Kwon, U. C. Belser, J. S. Hodges and W. H. Douglas, “Crack
propensity of porcelain laminate veneers: A simulated operatory evaluation, J.
Prosthet. Dentistry, 1999, 81: 327-334

W. D. Kingery, H. K. Bowen and D. R. Uhlmann, Introduction to ceramics, Second

Edition, John Wiley and Sons, New York, 1976, pp. 368-375

. Y. M. Chiang, D. P. Bimie, III, and W. D. Kingery, Physical Ceramics: Principles for

Ceramic Science and Engineering, John Wiley and Sons, New York, 1997, pp. 446-
464

K. Y. Lee and E. D. Case, “Microwave Sintering of Alumina Matrix Zirconia
Composites Using a Single-Mode Microwave Cavity,” Journal Materials Science
Letters, 1999, l8[3]: 201-203

K. Y. Lee, P. H. Dearhouse, and E. D. Case, “Microwave Sintering of Alumina Using
Four Different Single-Cavity Modes,” Journal of Materials Synthesis and Processing,
1999, 7[3]: 159-166

J. A. Delgado, L. Morejon. S. Martinez, M. P. Ginebra, N. Carlsson, E. Fernandez, J.
A. planell, M. T. Clavaguera-Mora, and J. Rodriguez-Viejo, “Zirconia-toughened
hydroxyapatite obtained by wet sintering”, J. Materials Sci. — Materials in Medicine,

1999, 10: 715-719

140

16. Y. M. Kong, S. Kim, H. E. Kim and I. S. Lee, “Reinforcement of hydroxyapatite
bioceramic by addition of ZrOz coated with A1203”, J. Amer. Ceram. Soc., 1999, 82:
2963-2968

17. L. Fu, K. A. Khor and J. P. Lim, “Yttria—stabilized zirconia reinforced hydroxyapatite

coatings”, Surf. Coating Tech., 2000, 127: 66-75

141

Chapter 7

COMPARISON OF MICROWAVE SINTERING AND CONVENTIONAL

SINTERING

The principle of microwave heating is fundamentally different compared to
conventional heating (Figure 7.1) [1]. In the microwave processing, heat is generated
within the material; meanwhile, the heat is created by an external heating source in the
conventional processing. Because of the result of this internal and volumetric heating, the
thermal gradients and the flow of heat in microwave processed materials are opposite of
those in materials processed by conventional heating (Figure 7.2) [2]. Due to these
reasons many advantages in microwave processing over conventional processing includes
(1) reduction in manufacturing costs because of saving of energy and processing time.
For example, Sheppard [3] showed microwave drying and sintering require less energy
than conventional processing by a factor of about two and ten respectively and lower
sintering temperatures and decreasing in processing time [4,5]. Temperature differentials
as high as 400°C between microwave and conventional processing have been reported
[6]; (2) improved product uniformity and yields; and (3) improved microstructure and
properties [7,8,9].

Janny and Kimrey [4] sintered A1203 doped with 0.1 wt% MgO under vacuum by
using a 28-GHz microwave [10]. For a given temperature, a much higher degree of
densification can be achieved with the use of microwave as compared to conventional

heating (Figure 7.3).

142

Using 3.96 g/cm3 for the theoretical density of alumina, 6.05 g/cm3 for partially
stabilized zirconia (TZ3YS), 5.90 g/cm3 for fully stabilized zirconia and 1.64~7.44 g/cm3
for hydroxyapatite (HAP), relative densities over 90% of theoretical were calculated for
specimens which were made by using each powder. Relative mass densities of each
specimen are given in Table 6.1. To get the similar densification of specimen sintered in
the microwave and the specimen sintered by conventional means, the sintering
temperature was increased from 1300°C to 1430 °C for alumina and zirconia specimens
and increased from 1100 °C to 1300 °C for HAP. Also, the sintering time is 1 hour for the
microwave as opposed to 4 hours for conventional processing.

Microwave sintering can yield smaller grain size than sintering with conventional
furnaces due to fast heating and enhanced densification at lower temperature which is so
called the ‘microwave effect”[ll,12,l3]. The microwave effect results from the direct
interaction of the processed material with the microwave field, resulting in internal
volumetric heating. The rrricrostructures for Specimens of alumina were shown in figure4.
Figure 7.4a indicates that the fracture of alumina that sintered at 1430°C for 4 hours in
conventional furnace and figure 7.4b shows that the fracture mainly occurred by

intergranular fracture.

143

 

 

 

 

M g a"; $ W Insulation

 

Specrmen

MZH‘W"

 

 

 

 

 

 

 

 

 

Microwave port

\ \‘l/ , /

 

 

 

 

 

 

   
 

 

Insulation

 

 

 

 

 

 

N .
\ Specrmen

 

 

 

 

 

 

Figure 7.1. Heating patterns in (a) conventional and (b) microwave furnaces [l]

144

 

Temperature

 

 

 

 

Temperature

 

 

 

Figure 7.2. Temperature gradients generated during (a) conventional and (b) microwave
heating of materials [2]

145

 

100

  
    

90 — Microwave

80 —
Density (%)

70 — .
Conventional

 

 

 

60 —-
50 1 L 1 1 1
800 1000 l 200 1400

Temperature

Figure 7.3. Schematic figure of density versus temperature of microwave (280Hz) and
conventionally Sintered A1203 [4]

I46

.338 395828 23 3:263:06 was: $525 .8 53358 .6 x3885 4.5 658.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ado 8: fl v.3 82 fl om _ .m min n23
wco cog fl Eco 83 v 8% $6 m>wNH
odm oofl _ 9mm omi v 3.0 and m>mNH
NS 00m _ ~ wco 0.0.3 v cod Nd M<QEH
cocmocacon Ocv C305 cosmocacon— 6% 2:05
cs @th SEC. ca .383. 0:5. $8on 3:: SB
.3855 209mm
wctocfim 0%.,»82—2 wctoEG 3:233:00 Rocococnh cwmco>< €532

 

 

147

 

Figure 7.4. Fracture surface of alumina (a) sintered by conventional at 1430 °C for 4
hours, (b) by microwave at 1300 °C for 1 hour.

148

REFERENCES

1 W.H.Sutton, “Microwave Processing of Ceramic Materials”, Am.Ceram.Soc.Bull.68[2],
1989, pp376-386

2 MA. Janney, H.D. Kimrey, and J.O.Kiggans, “Microwave Processing of Ceramics:
Guidelines used at the Oak Ridge National Laboratory”, in Microwave Processing of
Materials, 111, R.L. Beatty, W.H.Sutton, and M.F. Iskander, eds., MRS Symp. Proc.,
Vol.269, 1992, 173-185

3 L.M. Sheppard, “Manufacturing Ceramics with Microwaves: The Potential for
Economical Production”, Am Ceram. Soc. Bull., 67[10], 1988, pp.1656—1661

4 MA. Janney and H.D. Kimrey, “Microwave Sintering of Alumina at 28GHz”, Ceramic
Powder Science, 11, B, Ceramic Transactions, vol.1, 1988, pp.919-924

5 MA. Janney and H.D. Kimrey, “Diffusion-Controlled Processes in Microwave-Fired
Oxide Ceramics”, Mater.Res.Soc.Symp. Proc., vol.189, 1991, pp215-227

6 MA. Janney, H.D. Kimrey, W.R.Allen and J0. Kiggans, “Enhanced Diffusion in
Sapphire during Microwave Heating,” Journal of Materials Science vol.32, No.5,
1997,pp.1347-1355

7 A. De, I.Ahmad, E.D. Whitney, and D.E.Clark, “Microwave (Hybrid) heating of
Alumina at 2.45 GHz: 1. Microstructural Uniformity and Homogeneity,” Microwaves:
Theory and Application in Materials Processing, Ceramic Transactions, vol. 21, 1991,
pp.3l9-328

8 CE. Holcombe, T.T.Meek, and N.L. Dykes, “Enhanced Thermal Shock Properties of
Y203 —2wt.%Zr02 heated Using 2.45 GHz Radiation,” Mat.Res.Soc.symp. proc.vol. 124,

1988, pp.227-234

I49

9 M.C.L.Patterson, P.S.Apte, R.M. Kimber, and R.Roy, “Mechanical and Physical
Properties of Microwave Sintered Si3N4,” Mat. Res. Soc. Symp. Proc.v01269 1992,
pp.301-310

10 Yang Bin, Li Jun, Wang Jinxiang, and Zhong Xiangchong, “Influence of ZrOz on the
Sintering Microstructure and Properties of MgO Materials,” proceedings of the
International Symposium on the Science and Technology of Sintering, Intering and
Materials, Proc., 6‘“, vol. 6, 1995, pp.197-202

11 M.A. Janney and H.D. Kimrey, in Mat.Res. Soc. Symp. Proc., ed. W.B. Snyder, Jr.,
W.H. Sutton, M.F. Iskander and D.L.Johnson, vol.189, Materials Research Society,
Pittsburgh, Pensnsylvania, 1990, pp.215-227

12 K.Y.Lee, E.D.Case, J .Asmussen, Jr., and M.Siegel, in Proceedings of the 11th Annual
ESD Advanced Composites Conference, Ann Arbor, Michigan, 1995, pp.491-503

l3 K.Y.Lee, E.D.Case, J.Asmussen, Jr., and M.Siegel, in Cer. Trans, vol.59, The

American Ceramic Society Inc., Westerville, Ohio, 1995, pp.473-480

150

Chapter 8

MINIMIZING WARPING, CRACKING, AND RESIDUAL STRESS

8.1 Warping and Cracking
8.1.1 Initial Observations

Pre-sintering refers to partially sintering of the powder compact during which the
graphite fugitive phase from the ceramic substrate evaporates. Heated in an oven by itself,
the graphite fugitive phase evaporated near 700°C after 2 hours. However, when the same
temperature and heating time were used to pre-sinter a ceramic substrate having graphite
rods embedded on its surface, un-evaporated graphite residue remained. As a remedy,
pre-sintering e was performed at 900°C. In a homogeneous specimen this high-
temperature process presented no problem. However, warping and cracking were evident
after both pre-sintering and final sintering of layered zirconia/alumina samples (viz.,
layers were compacted together and sintered without any joining of layers). The cause of
warping and cracks was identified by visual inspection to be the difference in shrinkage
behavior of the individual layers. In separate tests, the amount of shrinkage of
homogeneous samples of zirconia and alumina specimens after pre-sintering and full
sintering have been measured and are given in Figure 8.1 ; an interpretation of the results
will be given shortly.

Few ideas are available for minimizing warping in the finally sintered specimens.
Among these are (1) finding the total shrinkage (after final sintering) of different layers

as a function of the pre-sintering temperature, in hopes that a range of optimal pre-

151

sintering temperatures can be found that would ensure shrinkage of the different layers to
the same structural dimensions; (2) varying the composition of the individual layers to
achieve flat, sintered composites, which can then be joined to other layers, and (3)
adjusting the initial sizes of the green specimens to ensure a compatible fit between
different layers after sintering.
8.1.2 Variation of Pre-sintering Temperatures to Minimize Cracking

In Figure 8.1, the presintering temperature was varied to study its effect on
shrinkage. The initial, green-state geometries of the specimens are nearly the same: their
diameters are controlled by the inside diameter of the die (2.25 cm), while their thickness
are controlled by making sure that equal amounts of powder (3 grams) are compacted. In
both plots (a) and (b), the fine lines indicate the amount of shrinkage after a set of
specimens have been pre-sintered at different temperatures. The heavier lines indicate the
final specimen dimensions after final sintering at 1430°C; the seeming variation with
temperature is merely a reference to the respective temperatures to which the specimens
have been exposed in the pre-sintering step.

Two observations can be inferred from Figure 8.1. First, the amount of total shrinkage
after final sintering is independent of the choice of pre-sintering temperature. Hence, in
making a two-layered, alumina/zirconia composite, we can expect the final, sintered
specimen to warp regardless of the choice of pre-sintering temperature.

The second observation is with regards to cracking. It is based on the variation in
the amount of shrinkage in the diameters of the alumina and zirconia specimens with
respect to pre-sintering temperature. In the extreme case presented in Figure 8.1 (a), it is

seen that at a processing temperature of 1200°C, there is a flip in Shrinkage behavior

between alumina and zirconia in going from pre-sintering to full sintering. In particular,
after 4 hours of pre-sintering at 1200°C, alumina (TMDAR) shrinks more than zirconia
(TZ3YS); however. after an additional 4 hours at 1430°C to achieve final sintering, the
situation is reversed—the diameters of the zirconia specimens end up smaller than those
of the alumina specimens. We conclude that during the processing of two-layered,
alumina/zirconia composites, a concavity will first develop along the alumina side after
pre-sintering, but after final sintering, the curvature is flipped to surround the zirconia
side instead. This reversal in curvature imposes cyclic loading that is the suspected cause
of macroscopic cracking and layer separation along the interface, as well as spalling on
the exposed concave face undergoing compression. By contrast, while the warping of a
two-layered, alumina/zirconia specimen is observed even for pre-sintering temperatures
of less than 900°C, there is no reversal in curvature in the sintering process (Figure
8.1(a)). Correspondingly, the severity of cracking is not nearly as severe; in fact, only
hairline cracks parallel to the interface and wrinkle cracks on the compressed face are
evident.

In summary, unlike the case of warping, the amount of interfacial cracking found
in the alumina/zirconia composite depends greatly on the pre-sintering temperature.
Based on the results in Figure 8.1, pre-sintering temperature of 900°C is chosen not only
for the full evaporation of the graphite phase embedded in ceramic powders, but also for
the minimization of cracking.

8.1.3 Use of FGM/Variation of Layer Composition to Minimize Warping
The most severe cracks found in the alumina/zirconia specimens are caused by

warping; in turn, warping is caused by a mismatch in the densification behavior of the

153

different layers. The coefficient of thermal expansion of alumina and zirconia are similar.
While pre-sintering temperature of 900°C minimizes cracking, it did not eliminate
cracking, nor did it control warping. To control warping and cracking, we introduce an
independent physical mechanism to counteract mismatches in densification. The idea is
to use differences in particle-mixin g ratios to control densification.

When the composition of a material is varied spatially in order to achieve a
desired effect, we speak of it as a functionally gradient material (FGM). In the current
application, the desired effect is the flatness of the composite layers after sintering. After
sintering, such layers can be joined to other layers to produce more complex assemblies.

In what follows, the proper mixture of powders must be chosen to give the desired
densification behavior. The goal is to produce a warp-free composite layer. A composite
layer of mixed zirconia and alumina is being considered (see the ‘Configuration’ figure
of Table 8.1). Mixing zirconia particles of different sizes allows us to vary the
densification behavior.

Tables 8.1 (a) and (b) show the results of a preliminary study on the shrinkage of
zirconia layers of different mixtures (average particle sizes are listed under the tables).
Being an initial control step, this study does not include alumina. As mentioned before,
the diameters of all specimens in the green state are the same, whereas the initial
thickness is constant within each composition. Due to the different sizes of particles used,
two powder piles of equal mass are compacted to different sizes in general; this gives rise
to different, initial specimen thickness across different mixture ratios. .

It can be inferred from Table 8.1 (b) that the total shrinkage after final sintering is

independent of the choice of pre-sintering temperature, and this is true for all mixtures of

154

zirconia considered. Also, the two compositions labeled ‘60T/40C’ and ‘55T/45C’ shrink
differently from specimens of other compositions, in an unexpected way. Namely, after
pre-sintering (Table 8.1 (a)), the specimens labeled 60T/40C and SST/45C shrink to
diameters that are smaller than those of the other mixtures, while after final sintering
(Table 8.1 (b)), these specimens have diameters that are bigger than those of other
mixtures. While no physical mechanisms are available to explain this shrinkage behavior
across different compositions, the effect does highlight the 60T/40C and SST/45C
compositions as possible candidates for further study.

Since the publication of chapter 5. we have added more samples to refine the exact
compostion where the residual stress be minimized. The results were represented in
figure 8.2. and the composition is 55%T/45%C. Having conducted experiments to isolate
the shrinkage behavior of zirconia mixtures, we moved on to study the warping behavior
of the zirconia mixture/alumina composite layer. A schematic of the composite layer is
given as part of Table 8.2. Densification behavior is controlled by varying the mixture of
zirconia (Zl'Og) powders in the zirconia component layer. This mixture is composed of
zirconia powder from Tosoh Corp., designated TZ-3YS, with an average particle Size of
0.6 microns (histogram: 92%<2 microns, 50%<0.59 microns and 10.12%<0.3 microns),
and zirconia powder from CERAC, Inc., with an average particle size of 1.23 microns
(histogram: 90%<1.79 microns, 50%<1.23 microns, and 10%<0.77 microns).

The amount of warping after final sintering of different zirconia mixture/alumina
composite layers is given in Table 8.2. Warping is quantified by the amount of the
displacement of the specimen’s center away from true flatness, as shown in the last figure

of Table 8.2. A plot of this center displacement versus composition is given in Figure 8.2,

155

where it is seen that a composition of 55% TZ-3YS and 45% CERAC in the zirconia

component can be used to minimize the warping in the zirconia/alumina composite layer.

156

Diameter Shrinkage after Pre-Sintering

 

 

 

 

 

 

 

 

 

2.35
2.25 r
E 2.15 r .
- 2.05 - -, ~-—- -.
.3 --c--’11\11)AR1 °
g 1.95 ~ —o—_____'Izsvs_]
.6
1.85 ~
1.75 r
1.65 p .
500 600 700 800 900 1000 1100 1200
temp °C
Diameter Shrinkage after Final Sintering
2.35
2.25 ~
E 2.15 ~
8
s. 2 . --—‘i ,-,____
g 05 --c--'li\ll>ARl
g 1.95 — +.-,,-_.T43Y51
.5
1.85 ~ 0 O 4, ...... c ------ o ------ o o- - . .0
1.75 -
WWW
1.65
A B C D E F G H
sample label

Figure 8.1a. Shrinkage difference in the diameter of zirconia and alumina specimens after

pre-sintering at different temperatures, and their corresponding diameter after final-

sintering at 1430°C.

157

Thickness Shrinkage after Pre-Sintering

 

 

 

 

 

 

 

 

 

 

 

 

0.35
0.33 5
0.31 1 W
E 0.29 —
v A B C D E F G H
§ 0.27 ~
: ..... o...
f.) 0.25 b o-.--"'° ------ 0 """"" O ..... '0 ----- 0
5 0.23 1 ..
O
0.21 1 1
inc-VIMDAR
0'” * :9—__:r_z_3:§__
0.17
500 600 700 800 900 1000 1100 1200
temp °C
Thickness Shrinkage after Final Sintering
0.35
0.33 ~
1--<>--'IMDAR
A 0'31 * 1+_,_2_£z_3.Y_s_
g 0.29 1
3 0.27 ~
(D
E 0.25 ~
.9 W
5 0.23
021 "'0'.
1 'A ....... o ........... °-._. ' o."
0.19 ~ 0 o """"" °"' -....-° °
0.17
A B C D E F G H

 

sample label

Figure 8. lb. Shrinkage difference in the thickness of zirconia and alumina specimens

after pre-sintering at different temperatures, and their corresponding thickness after

final-sintering at 1430°C.

Table 8.1(a). Size comparison of Zirconia layers of different mixtures after pre-
sintering at 900°C for 4 hours (cm)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

‘ omposit 85T 75T 72.5T 70T 60T SST 25T

" " /15C® /15C /27.5C /30C /40C /45C /75C

Pre-

Sinteri

Temp.

600°C D 2.235 2.235 2.24 2.24 2.235 2.235 2.23
Q)
T 0.25 0.26 0.26 0.26 0.24 0.235 0.26
(SD

700°C D 2.23 2.23 2.235 2.235 2.23 2.235 2.24
T 0.255 0.255 0.255 0.26 0.24 0.24 0.255

800°C D 2.225 2.225 2.23 2.225 2.215 2.22 2.216
T 0.255 0.265 0.26 0.26 0.23 0.24 0.255

900°C D 2.215 2.225 2.22 2.22 2.21 2.22 2.21
T 0.25 0.265 0.255 0.27 0.24 0.24 0.25

1000°C D 2.21 2.215 2.215 2.215 2.195 2.195 2.195
T 0.25 0.255 0.26 0.26 0.225 0.24 0.265

1100°C D 2.165 2.17 2.17 2.165 2.11 2.125 2.14
T 0.24 0.248 0.246 0.25 0.22 0.225 0.245

1200°C D 2.055 2.03 2.045 2.02 1.92 1.975 1.985
T 0.235 0.24 0.24 0.235 0.22 0.21 0.23

 

 

G) T: Zirconia from Tosoh Company (TZ-3YS; average powder particle size = 0.6 pm)
C: Zirconia from Cerac Company (5 wt% Y203; average powder particle size = 1.23
Mm)

® D: Diameter of the specimen

6) T: Thickness of the specimen

159

Table 8.1(b). Size Comparison of Zirconia Layers of Different Mixtures after final
sintering at 1430°C for 4 hours (left column lists temperatures at which the

specimens were pre—sintered) (cm)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

omposition 8ST 7ST 72.5T 70T 60T SST 25T
/15C® /1 SC /27.5C /30C /40C /45C /75C

Sinterin

Temp.

600°C D 1.77 1.755 1.755 1.755 1.805 1.8 1.75
G)
T 0.2 0.2 0.205 0.2 0.19 0.2 0.2
(3

700°C D 1.77 1.755 1.755 1.74 1.8 1.805 1.755
T 0.2 0.2 0.2 0.205 0.195 0.195 0.2

800°C D 1.77 1.76 1.775 1.745 1.815 1.81 1.75
T 0.2 0.2 0.2 0.2 0.19 0.195 0.2 '

900°C D 1.77 1.755 1.755 1.75 1.81 1.81 1.76
T 0.195 0.2 0.2 0.2 0.19 0.19 0.2

1000°C D 1.77 1.775 1.75 1.75 1.815 1.8 1.76
T 0.2 0.2 0.2 0.205 0.195 0.195 0.2

1100°C D 1.775 1.76 1.755 1.755 1.815 1.81 1.76
T 0.2 0.205 0.21 0.205 0.195 0.195 0.2

1200°C D 1.775 1.75 1.76 1.75 1.8 1.805 1.765
T 0.195 0.205 0.205 0.21 0.2 0.2 0.2

 

 

(D T: Zirconia from Tosoh Company (TZ-3YS; average powder particle size = 0.6 pm)
C: Zirconia from Cerac Company (5 wt% Y203; average powder particle size = 1.23
1m)

(2) D: Diameter of the specimen

(3 T: Thickness of the specimen

160

Table 8.2. Minimizing warping by mixing two 2102 powders in a top layer
1. Initial configuration of powder compact

 

ZrOz Mixture (Tosoh/Cerac)

 

 

A1203 (TMDAR)

 

 

2. Height Difference as Measured by Dial Indicator for sintered specimens (Cm)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Material Height at Height at Height (HC — HE)
Edge center differece T
(HE)* (HC)* (HQ-(HE)
85%(T)/ 15%(C) 0.4 0.45 0.05 0.111
75%(T) / 25%(C) 0.4 0.43 0.03 0.0698
72.5%(T) / 0.4 0.435 0.035 0.0804
17.5%(C)
70%(T) / 30%(C) 0.405 0.44 0.035 0.0795
65%(T) / 35%(C) 0.4 0.43 0.03 0.0698
60%(T) / 40%(C) 0.4 0.425 0.025 0.0588
55%(T) / 45%(C) 0.405 0.425 0.02 0.047
50%(T) / 50%(C) 0.405 0.44 0.035 0.08
40%(T) / 60%(C) 0.405 0.459 0.054 0.118
30%(T) / 70%(C) 0.41 0.495 0.085 0.172
25%(T) / 75%(C) 0.41 0.5 0.09 0.18

 

*Position for height measurement

 

 

 

 

161

 

 

0.2 1
0.18 —
0.16 —
0.14 —
0.12 ~»

 

 

 

 

 

 

 

 

 

 

 

 

 

0.08 — -

0.06 —
0.04 —_—»-~——~— - ~—

0.02 -——~~ -—

0 -~—-*~~ - 1 +— - 1

0 0.2 0.4 0.6 0.8 1

Weight % of Cerac zirconia powder

 

 

 

 

 

Normalized Height Difference
(HC-HE)/HE
o

 

 

 

 

 

 

 

 

Legend
a = 85% (T) / 15% (C) Q = 55% (1) / 45% (C)
b = 75% (T) / 25% (C) n = 50% (T) / 50% (C)

c = 72.5% (T) / 27.5% (C)
d = 70% (T) / 30% (C)
e = 65% (T) / 35% (C)
1 = 60% (T) / 40% (C) k = 25% (T) / 75% (C)

i = 40% (T) / 50% (C)
j= 30% (T) / 70% (C)

 

 

 

Figure 8.2. Comparison of the out-of—plane height at the center of zirconia/alumina
specimens, based on a variation of particle-size mixture in the zirconia component.

162

8.2 Residual Stress

When a multi-layer specimen is warped due to a mismatch in thermal expansion
and/or densification of the different layers, residual stresses are stored inside. Assuming
spherical symmetry in the warping and small deformations, and taking into account that
zirconia densifies more than alumina after full sintering, a formula for the final curvature

can be derived:

 

 

l _ [(aA -aZ)AT-(6A—52 )1 ~ [(aA '"az )AT—(5A_82)]

Kzr_d(l+l/2(01A+012)AT—1/2(8A+82))~ d

 

9

where [(18 the mean curvature, r is the radius of curvature, ais the coefficient of thermal
expansion (subscript A for alumina, and subscript Z for zirconia), AT is the change in
temperature in going from pre-sintering to final sintering (assuming zero curvature after
pre-sintering), 6 is the densification per unit original length (defined positive for
shrinkage), and d is the distance between the center planes of the two layers. In the above
formula; the curvature is positive when the concavity points downward towards the
zirconia side. A higher thermal expansion on the alumina side (aA > 612) or a higher
shrinkage on the zirconia side (5; > 6}.) will both lead to a positive curvature. In this case,
a stress distribution is set up in the composite layer that varies linearly from maximum

tension on the alumina top surface to maximum compression on the zirconia bottom face.

For convenience, this stress distribution will be referred to as the curvature stress. Also, it
is noted that on z a2, so thermal effects do not act to increase the curvature.

Because of the difference in shrinkage behavior of the two layers, axial stresses are
also present across the specimen. In particular, the alumina layer, which does not shrink
as much as the zirconia layer, is under compression, while the zirconia layer is under
tension. The total residual stress that is felt by the composite layer is the sum of the axial
and curvature stresses. The presence of such a sum of stresses account for the
experimental observation that vertical cracks can develop in various positions along the
thickness.

Residual stresses are minimized by choosing two materials of similar thermal
expansion and densification behavior. In the case of the alumina/zirconia composite,
since a1 = a2, our approach has been to match the densification behavior by varying the
particle-size mixture within the zirconia layer. Within the limits of what we can achieve
in the experiments, a composition of 55% TZ-3YS and 45% CERAC in the zirconia

component has been found to minimize the residual stresses.

164

Chapter 9

SCANNING ELECTRON MICROSCOPY (SEM)

The functionally gradient ceramic being proposed in this work is composed of a
spatially varying distribution of different constituent materials. To verify its successful
fabrication, topographical and compositional images of the ceramic samples have been
taken using a JEOL (Japan Electron Optics Laboratory) 6400 scanning electron
microscope, located at the Center for Electron Optics at Michigan State University. At
working distances of 15—25 mm (depending on specimen size), the microscope has an
approximate range of S—7 nm resolution for secondary electron images; this is based on
an interpolation of the following two given points: {working distance = 8 mm, attainable
resolution = 3.5 nm} and {working distance = 39 mm, attainable resolution = 10 nm}.
The magnification range is 10x—300,000x. In addition, an energy dispersive spectroscopy
(EDS) system is attached to the SEM. An energy range of 0—20 KeV enables the EDS
module to detect the presence of all elements from boron to uranium.

Details of SEM/EDS imaging can be found in [S. L. Flegler, J. W. Heckman, Jr., and
K. L. Klomparens, Scanning and Transmission Electron Microscopy, An Introduction,
Oxford University Press, 1993].

9.1 SEM Images of Topography Based on Secondary Electrons

When the electron beam (imaging beam) generated the SEM machine is directed at

the specimen surface, many electrons within the specimen’s atoms are given enough

energy to escape their orbitals. These escaped electrons, called secondary electrons, have

165

low kinetic energy. An electron detector having positive potential collects these easily
picked electrons, and the signal is amplified as an image. Cavities on the specimen
surface tend to trap secondary electrons and appear as dark regions on the image, while
convex bumps expose greater amounts of secondary electrons for picking and hence
appear bright. SEM images based on secondary electrons primarily show the topography
of specimens rather than composition. This is so because of two reasons: (1) the
secondary electrons are formed closed to the specimen surface, and (2) the availability of
secondary electrons does not depend significantly on the atomic number of the
specimen’s constituent atoms.

Another mechanism for image generation is based on the reflection of the imaging,
electron beam from the nuclei of the target. These reflected electrons are called primary,
or back-scattered, electrons, and they have high kinetic energy. Because the quantity of
back-scattered electrons depends on the size of the nuclei within the specimen’s
constituent atoms, SEM images based on back-scattered electrons not only measure the
surface topography of the specimen, but also its weight density. Back-scattered electron
images have not been used in this work.

On the JEOL 6400 microscope, we have chosen a pixel resolution of 512x512 for
secondary electron images, with an acquisition dwell time of 1e-S seconds per pixel.
From the point of view of element mapping, secondary electron images of specimen
samples are merely used as reference images of the mapped regions, and examples will
be included in the next subsection.

9.2 EDS Images of Element Composition Based on Released X-Rays

166

 

If the escaped electrons being forced out by the imaging beam come from the inner
shells of the target atoms, electrons from the outer shells move to fill the vacancies. In the
process, x-rays characteristic of the target atoms are dispersed. An energy detector can be
used to generate a spectral energy plot indicating the presence of different elements in the
target sample. The detector range in the EDS system on the JEOL 6400 allows all
elements from boron to uranium to be detected.

In the x-ray, mapping mode of the EDS system, a detection window is chosen around
a particular element’s signature, x-ray peak(s). Under a two-dimensional scan of the
target surface using the JEOL microscope, dots on the observation screen are generated in
proportion to the x—ray count of that element (Figure 9.1). The same spectral detection
window around an element’s x-ray peak(s) can be retained for an element x-ray line-scan.
This time an x-ray count of element composition can be plotted along the chosen line
(Figure 9.2, 9.3, 9.4).

In Figure 9.2, which shows the x-ray line-scan, errors in the element count are noted
in the depleted regions corresponding to the dark regions in the 2-D, x-ray map. A
possible explanation is system noise. Here, we suggest a relationship between noise and
the size of the imaging beam. When the resolution of the image is high, the diameter of
the imaging beam should be low; this means that the resultant signal (and output) is
weak, leading to a higher sensitivity to noise. Hence, to test the nature of the error in
element count, one option may be to lower the resolution in the x-ray line—scan to see if
the noise is reduced. If not, the observed noise may well be a characteristic system noise

that needs to be reported.

167

To generate the element x-ray map on the JEOL microscope, a 2-D resolution of
128x128 pixels was used with an acquisition dwell time of 100 milliseconds per pixel.
For the x-ray line-scan, a linear resolution of 128 pixels was used with an acquisition

dwell time of 100 milliseconds per pixel.

168

 

‘ 111-2111'- . .

1 00011111

   

 

 

' 100011111
1003 um ,m
1, TZ3YS
2, 25 % TMDAR / 75% TZ3YS
1 2 3 4 3, 75 % TMDAR/25% TZ3YS
4, TMDAR

 

 

 

 

 

 

 

 

 

Figure 9.1. Element mapping based on energy dispersive spectroscopy (EDS) of the SEM
sample. X—ray beams characteristic of aluminum. yttrium and zirconium are collected
from the sample surface and the signal is magnified proportionally as brightness. The
different colors are arbitrarily chosen. Line Scan of the x—ray and EDS element scan
shown in figure 9.2, 9.3.

169

Counts

Counts

Counts

45

32M W

 

 

 

0 111m n. “n
1400 2100 28 0
Microns
Zr
36
27
18
9 “(WM
0' l '1
1400 2100 2800
Microns
Y
8
6
4
21111601111411“!
0 111., .
1400 2100 2800
Microns

Figure 9.2. Line scan of the x-ray count based on energy dispersive spectroscopy. The
amount of an element (Al, Zr, or Y) along the line(figure 9.3a) is shown by the number of
x-ray particles detected over a dwell time of 100 milliseconds. Element mapping and
EDS element scan shown in figure 9.1, 9.4.

170

1‘
i
i
i

 

(b)

Figure 9.3. (a) Line across the samples for line scan of the x-ray count based on energy
dispersive spectroscopy(Figure 9.2.), (b) square region (For figure 9.4) of the sample

surface

171

2000 1 A'

1500 _
Zr
3 . ,
g 1000 ..
o
O
500 _

 

 

 

 

 

 

 

 

LIE-Wow - 7 1'5 ' 20

Energy (keV)

Figure 9.4. EDS elemental scan over the indicated, square region (figure 9.3b) of the
sample surface. X-ray beams (created by shining an imaging, electron beam on the target)
of different energies are detected are correlated to x-ray signatures of different elements.
The JEOL 6400 microscope is capable of detecting all elements from boron to uranium.
Element mapping and line scan of the x-ray shown in figure 9.], 9.2.

172

REFERENCES

 

1 W.H.Sutton, “Microwave Processing of Ceramic Materials”, Am.Ceram.Soc.Bull.68[2],
1989, pp376-386

2 M.A. Janney, H.D. Kimrey, and J.O.Kiggans, “Microwave Processing of Ceramics:
Guidelines used at the Oak Ridge National Laboratory”, in Microwave Processing of
Materials, 111, R.L. Beatty, W.H.Sutton, and M.F. Iskander, eds., MRS Symp. Proc.,
Vol.269, 1992, 173-185

3 L.M. Sheppard, “Manufacturing Ceramics with Microwaves: The Potential for
Economical Production”, Am Ceram. Soc. Bull., 67[10], 1988, pp.16S6-166l

4 M.A. Janney and H.D. Kimrey, “Microwave Sintering of Alumina at 28GHz”, Ceramic
Powder Science, 11, B, Ceramic Transactions, vol.l, 1988, pp.919-924

S M.A. Janney and H.D. Kimrey, “Diffusion-Controlled Processes in Microwave-Fired
Oxide Ceramics”, Mater.Res.Soc.Symp. Proc., vol.189, 1991, pp215-227

6 M.A. Janney, H.D. Kimrey, W.R.Allen and J0. Kiggans, “Enhanced Diffusion in
Sapphire during Microwave Heating,” Ioumal of Materials Science vol.32, No.5,
1997,pp.1347-1355

7 A. De, I.Ahmad, E.D. Whitney, and D.E.Clark, “Microwave (Hybrid) heating of
Alumina at 2.45 GHz: 1. Microstructural Uniformity and Homogeneity,” Microwaves:
Theory and Application in Materials Processing, Ceramic Transactions, vol. 21, 1991,
pp.3l9-328

8 CE. Holcombe, T.T.Meek, and N.L. Dykes, “Enhanced Thermal Shock Properties of
Y203 —2wt.%ZrOz heated Using 2.45 GHz Radiation,” Mat.Res.Soc.symp. proc.vol. 124,

1988, pp.227—234

173

 

9 M.C.L.Patterson, P.S.Apte, R.M. Kimber, and R.Roy, “Mechanical and Physical
Properties of Microwave Sintered Si3N4,” Mat. Res. Soc. Symp. Proc.v01269 1992,
pp.301-310

10 Yang Bin, Li Jun, Wang Jinxiang, and Zhong Xiangchong, “Influence of ZrOz on the
Sintering Microstructure and Properties of MgO Materials,” proceedings of the
International Symposium on the Science and Technology of Sintering, Intering and
Materials, Proc., 6'“, vol. 6, 1995, pp. 197—202

11 M.A. Janney and H.D. Kimrey, in Mat.Res. Soc. Symp. Proc., ed. W.B. Snyder, Jr.,
W.H. Sutton, M.F. Iskander and D.L.Johnson, vol.189, Materials Research Society,
Pittsburgh, Pensnsylvania, 1990, pp.215-227

12 K.Y.Lee, E.D.Case, J.Asmussen, Jr., and M.Siegel, in Proceedings of the 11th Annual
ESD Advanced Composites Conference, Ann Arbor, Michigan, 1995, pp.491-503

l3 K.Y.Lee, E.D.Case, J.Asmussen, Jr., and M.Siegel, in Cer. Trans, vol.59, The

American Ceramic Society Inc., Westerville, Ohio, 1995, pp.473-480

174

Chapter 10

CONCLUSION

In this study, to fabricate the functionally gradient materials, commercially
available alumina, zirconia and hydroxyapatite powders were sintered. The specimens
were densified to over about 90% of theoretical density.

To densify the specimens, two different heating systems—a conventional furnace
and a cylindrical single-mode microwave cavity unit—were used to compare each
behavior. Microwave-sintered alumina specimens reached about 97% theoretical density
at a sintering temperature of 1300°C after 1 hour, while a specimen that is sintered in
conventional heating process reached about 97% at 1430°C after 4 hours. This verified
that by increasing the sintering temperature and time in conventional sintering process,
similar densifications comparing to microwave heating were achieved. In addition, the
fracture toughness of alumina between microwave and conventional heating was
examined and gave the same result. This sintering study revealed that the microwave
sintering could be done rapidly and uniformly comparing conventional sintering.

For a given set of materials, the extreme condition for residual stress is expected
when two or more powders having significantly different thermal expansion and/or
densification behavior are processed together. The sintered samples are subject to
warping since the alumina layer is in a state of compressive residual stress while the

zirconia layer is in a state of tensile residual stress based on shrinkage differences.

175

Initially, homogeneous powder layers are processed separately to expose their individual
shrinkage behavior. Then, to understand and control the residual stress, two-layered
samples, one layer with alumina the other with various mixtures of two different kinds of
zirconia powders, are pressed sintered. While all of the samples had some degree of
warping, the mixture ratio 55% TZBYS (from Tosoh) / 4S % zirconia (from Cerac) yields
the minimum amount of warping observed.

Among the many fugitive phases used to make channels (polymer, paper, pure
graphite and pencil lead), pencil lead worked well. For making circular channels, pencil
leads varying in lengths from 1 to 1.5 cm with nominal diameters of 300, 500, 700 and
900 microns were sintered. To make internal channels, part of the powder was placed in
the die before placement of the fugitive phase, followed by the rest of the powder.

Samples with various internal channels such as manifold-shaped surface channels
and spoke-shaped surface channels were fabricated in ceramics and ceramic composites
including alumina, zirconia and hydroxyapatite. In addition, the FGMS specimens were
fabricated. FGMS with variable composition were characterized by a process called
element mapping. The basic idea is the detection of characteristic x-rays emanating from
the surface of the target material due to impact from an imaging, electron beam. This
element mapping showed well the composition of each layer of FGMS.

From the past experiments and results, one of the difficult problems in fabricating
FGMS is the determination of process variables so that the residual stresses would not
damage the details (like channels) as well as materials themselves. Because of the
presence of warping and cracks, instead of sintering an entire FGM stack at one time,

specimen layers are sintered separately and then joined. With 20-gram dead weights,

176

HAP and Macor layers have been successfully joined at a temperature of 1475°C with 4
hours of holding time. SEM examination of the joined region showed no evidence of
cracking or microcracking around channels or across the interlayer boundaries. Also, the
grain structure near the joint does not differ from the grain structure in the bulk of the
specimen, indicating that the process of generating the joint does not greatly perturb the

specimen’s microstructure, at least on a size scale of a few microns.

177

APPENDICES

178

APPENDIX I

Micrographs of Fracture Surface of Materials Used in This Study

179

 

 

Figure I-l. SEM image of fracture surface of alumina (a) Sintered at
1430°C for 4 hours in conventional furnace, (b) Sintered at 1300°C for
1 hour in microwave.

180

 

(b)
Figure I-2. SEM image of fracture surface of TZSYS (fully stabilized 8
mol% Yttria-Zirconia) (a) sintered at 1430°C for 4 hours in
conventional furnace (b) sintered at 1300°C for 1 hours in microwave

181

 

Figure I-3. SEM image of fracture surface of TZ3YS (partially
stabilized 3mol% Yttria-Zirconia) sintered at 1430°C for 4 hours in

conventional furnace ((a) and (b) represent images taken at different
locations)

182

 

(b) 8

Figure 14. SEM image of fracture surface of HAP (Hydroxyapatite) (a) sintered at
1300°C for 4 hours in conventional furenace, (b) sintered at 1100°C for 1 hour in
microwave.

183

 

 

Figure I-5. SEM image of fracture surface of (a) 25% TMDAR /
75%HAP sintered at 1300°C for 1 hour in microwave and (b) 50%
TMDAR / 50% TZ3YS sintered at 1430°C for 4 hours in conventional
furnace.

 

APPENDIX 11

Micrographs of Various Shapes of Manifold Channels

 

185

 

(a)

 

(b)

 

(C)

Figure II-I TZ8YS specimens pre sintered at 900°C for 2 hours and sintered
1430°C for 4 hours in conventional furnace with various shapes of manifold
channels of (a) ABS which is a thermoplastic (b) ABS and pencil lead in center
and (c) paper

186

1250um

 

Figure II-2 TZ8YS specimens pre sintered at 900°C for 2 hours and sintered
1430°C for 4 hours in conventional furnace (a) surface channels created by paper
as fugitive phase, (b) embed channel

187

 

(b)

Figure II-3 TZ8YS specimens pre sintered at 900°C for 2 hours and sintered
1430°C for 4 hours in conventional furnace channels created by paper as
fugitive phase (a) surface channel, (b) embed channel

188

Position

1,1111: picture
\_

2500um

 

Figure II—4 TZ8YSspecimen pre sintered at 900°C for 2 hours and sintered 1430°C for
4 hours in conventional furnace with manifold shaped channels created by pure
graphite as fugitive phase, (a) Surface channel, (b) Embed channel (c) Embed channel
in position marked in (b)

189

 

APPENDIX III

Micrographs of Various Channels in Homogeneous Materials and
Composite Materials

 

190

2001mm

250011111 ‘ 7 250011111

2500um

 

(e) (0

Figure HI-l Various channels in homogeneous materials and composite materials

191

Figure 111-1.

(3)

A SEM micrograph of an alumina specimen that was pre-sintered at 900°C for one
hour in air in a conventional furnace, then sintered at 1300 0C for 1 hour via
microwave heating. Two rows of closely spaced channels were formed by pressing
500microns pencil leads.

(b) A particular ceramic specimen (25% alumina / 75% TZ3YS) with with two layered

(C)

cylindrical channels, using 500 micron-diameter pencil leads as fugitive elements.

The specimen was pre-sintered in air at 900 °C for one hour in a conventional

furnace, then sintered at 1300 °C for one hour via microwave heating.

A fully- stabilized sirconia (T Z8YS) specimen pre-sintered conventionally at 900°C
for 2 hours in air, then sintered for four hours at 1430 0C in air in a conventional
furnace. The channels were formed using 900 microns pencil lead.

(d) A FSZ specimen pre-sintered conventionally at 950 0C for two hours in air, then

(6)

(f)

sintered for four hours at 1430 °C in air in a conventional furnace. Both four internal
channels and three surface channels were formed using 500 micron diameter pencil
leads.

A FSZ specimen pre-sintered at 900 0C for two hours in air, then sintered for four
hours at 1430 °C in air in a conventional furnace. The larger channels were formed
using 900 micron diameter pencil lead, while the smaller channels were formed
using 500 micron diameter pencil lead for internal channels. And, 300 micron
diameter pencil lead were used for surface channels.

A FSZ speciemn (pre-sinterd and sintered in air at 1200 °C for two hours and for

four hours at 1430 °C, respectively) was shown with 900 micron and 700 micron
pencil lead as fugitive phase.

192

 

(b)

Figure III-2 A hi gh-magnification view of a single channel formed (a) from a 500
micron diameter pencil lead (b) from 900 micron diameter pencil lead in a FSZ specimen.
The specimen was pre-sintered in air at 900 °C for two hours and sintered at 1430 0C for
four hours in a conventional furnace.

193

 

b
Figure III-3 A high-magnification view (Sf 21 single channel (a)formed from a 300 micron
diameter pencil lead (b) from 700 micron diameter pencil lead in a FSZ specimen. The
specimen was pre-sintered in air at 950°C for two hours and sintered at 1430 °C for four
hours in a conventional furnace.

194

APPENDIX IV

Micrographs of Various Functionally Gradient Materials

195

TWO LAYERED SPECIMENS

 

60%T / 40%C

 

0 TMDAR O

 

 

 

 

55%T 745 %C

 

O TMDAR O

 

 

 

2501mm

 

(b)

Figure IV-l Pre-sintered at 900 °C for four hours and sintered at 1430 °C for four hours
in air in a conventional fumace.(a) with 700 micron diameter pencil lead as fugitive
phase, (b) with 500 micron diameter pencil lead as fugitive phase. Materials were shown
in configuration respectively.

196

FOUR LAYERED SPECIMENS

250011m

 

65 %TZ3YS / 35%CERAC

 

70 %TZ3YS / 30%CERAC

 

30%TMDAR / 70%(70%T/30%C)

 

 

 

TMDAR

 

 

65 %TZ3YS / 35%CERAC

 

70 %TZSYS I 30%CERAC

 

70%TMDAR / 30%(70%T/30%C)

 

TMDAR

 

 

 

250011111

 

(b)

Figure IV-2 Pre-sintered at 900 0C for four hours and sintered at 1430 °C for four hours
in air in a conventional fumace.(a) with 700 micron diameter pencil lead as fugitive
phase, (b) with 700 micron diameter pencil lead as fugitive phase. Materials were shown
in configuration respectively

197

ll LAYERED SPECIMENS

 

 

 

2501mm 1 l

 

1,TMDAR
2,90%TMDAR+10%TZ3YS
3,80%TMDAR+20%TZ3YS
4,70%TMDAR+30%TZ3YS
5,60%TMDAR+40%TZ3YS
' 6,50%TMDAR+SO%TZ3YS
7,40%TMDAR+60%TZ3YS
8,30%TMDAR+70%TZ3YS
. 9,20%TMDAR+80%TZ3YS
250011m 10,10%TMDAR+90%TZ3YS
— “ 1 1,TZ3YS

 

(b)
Figure IV-3 Pre-sintered at 900 °C for four hours and sintered at 1430 °C for four hours
in air in a conventional fumace.(a) with 900 micron diameter pencil leads, (b) with 900
micron diameter pencil leads. Materials were shown in configuration respectively

198

APPENDIX V

Element Mapping of Functionally Gradient Materials

199

"from.“ 1 10011111
U i 110011111-

J.Hum

 

 

65% TZ3YS/35% CERAC
70% TZ3YS/30% CERAC
30%TMDAR/ 70% (70%T/30%C)
TMDAR

Figure V-l. Element mapping based on energy dispersive spectroscopy (EDS) of the
SEM sample. X-ray beams characteristic of aluminum, yttrium and zirconium are
collected from the sample surface and the signal is magnified proportionally as brightness.
The different colors are arbitrarily chosen.

 

 

 

 

 

 

200

J

1100 um

 

Figure V-2 Square region (for the figure V-3) of the sample surface

201

 

Al

 

 

 

 

 

 

 

 

1000‘ Point1
800-
U)
E 600-
8
O 400‘ AU
ZOO-COM. Au AU
0 I A? 'L ' 1' ' T
5 1O 15 20
Zr Energy (keV)
‘ Point2
6004
B _
g 400- A1
0 _ 1
O
200-
. BOVLE A“ AAuA
0- 1 i ' ' 1 1‘
5 10 15 20
z Ener keV
800‘ r QY( ) .
- Po1nt3
600-
.9
5 400-
8 .
' A
200- 1
k - 4 441.114
0" l I l l
5 10 15 20
Energy (keV)

Figure V-3. EDS elemental scan over the indicated, square region (figure V-2) of the
sample surface. X—ray beams (created by shining an imaging, electron beam on the
target) of different energies are detected are correlated to x-ray signatures of different
elements. The JEOL 6400 microscope is capable of detecting all elements from boron to
uranium.

202

Zr

 

 

 

 

 

1000 .
_ Point4
800 ~
0) '4
E 600 -
8 -
200-B Si l Au
0‘ . A. -
5 10 15 20
_ Zr Enerov (keV)
600 - Point 5
49 ‘ AI
g 400 -
O - 1
0
200- 1
-BC J AU Au
0 T r ‘ 3.1.11 .L.‘*.
5 10 15 20
Energy (keV)

Figure V-4. EDS elemental scan over the indicated, square region (figure V-2) of the
sample surface. X-ray beams (created by shining an imaging, electron beam on the
target) of different energies are detected are correlated to x-ray signatures of different
elements. The JEOL 6400 microscope is capable of detecting all elements from boron to
uranium.

203

 

1100 um

Figure V-S. Line across the samples for line scan of the x-ray count based on energy
dispersive spectroscopy(For figure V-6.)

204

 

 

 

 

 

 

 

   

181
15-
Al
,9 12-
C
8 9-
O 6-
3-
0 .
1600 2400 3200
Microns
20‘ Zr I
a) 15" .. 1
E , l:
g 10- W I 1 11 i
o ‘ l , '
5- 1
0, 11.11.11.» f v. “.3
800 1 600 2400 3200
Microns
6.1
5- Y
.9 4-
5 3- J , , .
O 1 i, .
O 2‘ . l , ~
1- 1111 111
0_ 1111111! 11% , I 1 11 1| I
800 1 600 2400 3200
M1crons

Figure V-6. Line scan of the x-ray count based on energy dispersive spectroscopy. The
amount of an element (A1, Zr, or Y) along the line(figure V-S) is shown by the number of
x-ray particles detected over a dwell time of 100 milliseconds. Element mapping and
EDS element scan shown in figure V-1,figureV-3,4.

205

100011m 1 ()()()11m
"1fifi1='_

1000 um

1 00011111 100011 m
m m

 

 

55%TZ3YS / 45 %CERAC

 

TMDAR

 

 

 

Figure V-7. Element mapping based on energy dispersive spectroscopy (EDS) of the
SEM sample. X-ray beams characteristic of aluminum, yttrium and zirconium are
collected from the sample surface and the signal is magnified proportionally as brightness.
The different colors are arbitrarily chosen.

206

 

Figure V-8. Line across the samples for line scan of the x-ray count based on energy

dispersive spectroscopy(For figure V-9.)

207

30- . , , AI
25- f l 1 1'
20- , 1 11 1
15- R 1 i 1
10- l
5 _
0 .1111 .JLignrnflr..11m 11',

400 800 ' 1200 1500 77 72000
Microns

Counts

 

25 _ Zr

20- ,
15-
10-
5-

Counts

 

 

. . .111.111 11

400 800 1 200 1 600 2000
Microns

7 " 1

6 -

5 -

Y

Counts
A

2-

3,‘ . . _ [11111111

' l
400 800 1 200 1600 2000
Microns

 

Figure V-9. Line scan of the x-ray count based on energy dispersive spectroscopy. The
amount of an element (Al, Zr, or Y) along the line (figure V-8) is shown by the number
of x-ray particles detected over a dwell time of 100 milliseconds. Element mapping and
EDS element scan shown in figure V-7, figure V-ll

208

1000 um

 

Figure V-10. Line across the samples for line scan of the x-ray count based on energy
dispersive spectroscopy(For figure V-l 1.)

209

1400 -1 Zr

1200 -1 Point 1

1 000 -
800 -

600 -
400 -
Hf

200‘0 Au
01.21.22,? 712 ”7111'

Energy (keV)

Counts

 

 

 

 

20

18001 N
1 500 q Pomt 2

1200 -
900 -

500~
3004 0 AU

Au
0 41.4.1.5, - -- - - --
4 8 12 15 20
Energy (keV)
Figure V-ll. EDS elemental scan over the indicated, square region (figure V-10) of the
sample surface. X-ray beams (created by shining an imaging, electron beam on the

target) of different energies are detected are correlated to x-ray signatures of different

elements. The JEOL 6400 microscope is capable of detecting all elements from boron to
uranium.

Counts

 

 

210

 

   

111'1111111111111