Theory and application of microwave joining

Material Information

Theory and application of microwave joining
Cozzi, Alex Douglas, 1963-
Publication Date:
Physical Description:
xx, 263 leaves : ill. ; 29 cm.


Subjects / Keywords:
Ceramic materials ( jstor )
Dielectric materials ( jstor )
Electric fields ( jstor )
Gels ( jstor )
Heating ( jstor )
Iron oxides ( jstor )
Microwaves ( jstor )
Nickel oxides ( jstor )
Oxides ( jstor )
Silicon carbides ( jstor )
Dissertations, Academic -- Materials Science and Engineering -- UF
Materials Science and Engineering thesis, Ph. D
bibliography ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
non-fiction ( marcgt )


Thesis (Ph. D.)--University of Florida, 1996.
Includes bibliographical references (leaves 255-262).
General Note:
General Note:
Statement of Responsibility:
by Alex Douglas Cozzi.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
023798013 ( ALEPH )
35771605 ( OCLC )

Full Text








Copyright 1996


Alex Douglas Cozzi

I would like to dedicate this dissertation to my father, Thomas. My only regret is that you
could not see me graduate.


The author would like to thank a number of people for their professional

suggestions and personal support towards this study. First I would like to thank Dr. David

Clark who not only showed me how to be a better scientist, but how to be a better person.

His guidance and constant reassurance allowed me to complete the graduate program. I

would like to express special thanks to Dr. Matt Ferber of Oak Ridge National Labs for

taking the time to provide assistance and guidance even though he is extremely busy with

many other projects.

I would like to also thank Drs. J.J. Mecholsky, B.V. Sankar, R.J. Singh, E. D.

Whitney and V.J. Tennery for their comments and suggestions that increased the technical

value of this study. I wish to thank Dr. Ron Hutcheon of the Atomic Energy of Canada

Ltd., Chalk River Laboratories for providing essential dielectric data and insight into the

meaning of that data.

The author appreciates the assistance of the administrative and technical staff of the

High Temperature Materials Laboratory at Oak Ridge National Labs, especially Arvid Pasto

and Billie Russell, the ringleaders of the HTML Fellowship program. I would like to thank

the members of Dr. Ferber's research group, Kristin Breder, Jason Canon, Dana Green,

Allen Haynes, Tim Kirkland, Edgar Lara-Curzio, Ron Ott, Laura Riester and Andy

Wereszczak. I would like to also thank Mark Janney and Jim Kiggans for providing me

with all the materials and assistance they could spare.

I would like to convey special thanks to Diane Folz and Rebecca Lynn Shulz for

both their professional help and friendship throughout my stay at the University of Florida.

I would like to also thank all of the past and present members in Dr. Clark's research

group, my peers in the Department of Materials Science and Engineering, Iftikhar

Ahmad,Salwan Al-Assafi, Attapon Boonyapiwat, Robert Dalton, Arindam Dd, Robert

DiFiore, Zak Fathi, Carl Jones, Don Jones, Michelle Lococo, Edmund Moore, Lorie

Stapler and Bruce Zoitos. I would like to convey special thanks to my friends, especially

Philip McCluskey, Andrew Duncan, Mark Weaver, Chris O'Gara, Lisa Demmy and

George Demmy.

I would like to express particular thanks to my family, my parent's Thomas and

Lillian, my sister Dianne, my brothers Tommy and Alan as well as their families for their

emotional and financial support through eight long years of graduate school. Finally, I

would like to acknowledge my wife, Sunday, who along with my family believed in me

when even I wasn't so sure. She watched me through the highs and lows of this study,

tried tirelessly to keep me focused on the goal and provided love and support without

which this study could not have been performed.

This research was sponsored by the Assistant Secretary for Energy Efficiency and

Renewable Energy, Office of Transportation Technologies, as part of the High

Temperature Materials Laboratory Fellowship Program, Oak Ridge National Laboratory,

managed by Lockheed Martin Energy Research Corp. for the U.S. Department of Energy

under contract number DE-AC05-960R22464.


ACKNOWLEDGEMENTS ....................................... iv

LIST OF TABLES ............................................. ix

LIST OF FIGURES ............................................ xi

ABSTRACT ............................................... xix


1 INTRODUCTION .................................. 1

2 SURVEY OF THE LITERATURE ........................ 7
M icrow aves .................................. 7
Background ............................. 7
Microwave Applicators ..................... 12
Microwave Processing of Materials ............. 16
Numerical Modeling of Microwave/Material
Interactions ........................ 17
Evidence of Microwave Enhancements ........... 26
Joining ..................................... 28
Joining M ethods .......................... 28
Joining processes .................... 32
Criteria for Joining ........................ 35
Effect of Process Variables on Joining ........... 38
Temperature ........................ 38
Pressure ......................... 41
Time ............................. 41
Reactions ......................... 44
Scales of Joining .......................... 44
Atom ic ........................... 44
N ano ............................ 46
M icro ............................ 49
Macro ............................ 50
M materials Joining .......................... 50
Metal/ceramic joining .................. 51
Ceramic/ceramic ..................... 52
Microwave Joining ........................ 53
Sum m ary ........... ........................ 60

3 MATERIALS AND METHODS .......................... 62
Phase O ne ................................... 63

Sol-G el ................................ 64
Basegel .......................... 64
Doped gel ......................... 65
Sol-Gel Characterization ..................... 68
Stand alone heating ................... 68
Differential scanning calorimetry .......... 70
X-ray diffraction analysis ............... 70
Dielectric properties ................... 71
Alum ina ............................... 71
Nickel Oxide ....................... 72
Surface roughness ................... 73
Phase Two ................................... 74
Microwave Joining Apparatus ................. 75
Microwave Susceptors ...................... 77
Specimen Preparation ...................... 86
Mechanical Testing ........................ 87
Phase Three .................................. 90
M microwave Joining ........................ 91
Conventional Joining ....................... 93
Optical Inspection ......................... 93
Statistical Analysis ........................ 94
Phase Four ................................... 95

4 NUMERICAL MODELING ............................ 96
Heat Transfer ................................. 97
Transient Heat Conduction ............... ........ 101
Boundary Conditions ........................... 103
The Physical M odel ............................ 105
Case Title ............................. 105
General Problem Parameters ................. 106
Region Data ............................ 106
M materials .............................. 107
Initial Temperature ....................... 109
Heat Generation ......................... 109
Boundary Conditions ...................... 110
G rids ... ... ... ... ... ... ... ... ... ... ... 112
Analytical Functions ........................... 113
Specific heat ............................ 114
Thermal conductivity ...................... 114
Power generation ........................ 114
Boundary condition ....................... 123
Tabular Functions ............................. 123
Thermal conductivity ...................... 126
Printout Tim es ............................... 127
Nodes M monitored ............................. 128
Transient Data ................................ 128

5 RESULTS ....................................... 137
M achined Surfaces ............................ 137
Sol-Gel Interlayers ............................ 137
Stand Alone Heating ...................... 137

Differential Scanning Calorimetry ............. 143
X-ray Diffraction Analysis .................. 145
Dielectric Measurements ................... 147
Microwave Susceptors .......................... 153
Joining .................................... 160
Joining Conditions ....................... 160
Visual Inspection ........................ 160
Flexure Testing .......................... 163
Statistical Analysis ....................... 184
Numerical Modeling ........................... 203

6 DISCUSSION .................................... 217
Phase O ne .................................. 217
Machined Surfaces ....................... 217
Sol-Gel Interlayers ....................... 218
Summary of Phase One .................... 221
Phase Two .................................. 223
Microwave Susceptors ..................... 223
Microwave Joining Apparatus ................ 224
Summary of Phase Two .................... 224
Phase Three ................................. 225
Experimental Design ...................... 225
Joining Experiments and Flexure
Tests ........................... 226
Statistical Analysis ....................... 229
Summary of Phase Three ................... 231
Phase Four .................................. 231
Computer Simulations ..................... 232
Summary of Phase Four .................... 234

7 SUMMARY AND CONCLUSIONS ..................... 235

8 FUTURE WORK .................................. 241


THICKNESS FOR Al-66 CEMENT ................ 243

.. .. . 245

ANALYSIS ................................. 251

REFERENCES .............................................. 255

BIOGRAPHICAL SKETCH ..................................... 263


Table 2-1. Error Associated with Simplified Equation for Penetration Depth for
Several e"eff/ E' Ratios. .............................. 11

Table 3-1. Manufacturer Reported Properties of Coors Alumina Used in Study.
. 7 2

Table 3-2. Levels of Varied Processing Conditions. ................... 91

Table 3-3. Experimental Design used to Aid in Determining the Effect of
Processing Conditions on the Strength of the Joint. ............. 92

Table 4-1. Summary of the Parameter Cards Used to Describe the Nine Model
Regions. ....................................... 108

Table 4-2. Input Parameters for the Four Materials Represented in the Model. .. 109

Table 4-3. Heat Generation Cards for the Four Materials in the Model. ...... 110

Table 4-4. Table of the Gross and Fine Grid Lines for Both the r and z Axes. 113

Table 4-5. Thermal Conductivity of Coors AD995 Alumina at Different
Temperatures ..................................... 126

Table 4-6. Thermal Conductivity of the Nickel Oxide Used in the Model at
Different Temperatures. ............................. 126

Table 4-7. Calculated Values for the Thermal Conductivity of a 30 wt.%
SiC/Cement Susceptor. .............................. 127

Table 4-8. Nodes of Interest and Their Placement in the Model. ........... 130

Table 5-1. Flexure Strength of AD995 Alumina Joined Under Various
Conditions Using Several Different Interlayer Materials. ......... 163

Table 5-2. Mean Strength and Standard Deviation of Bars Joined by either
Microwave or Conventional Heating. ................... .. 169

Table 5-3. Data Used to Perform ANOVA Analysis. ................... 192

Table 5-4. Output of Statistical Analysis of Data from Table 5-3. .......... 193

Table 5-5. Statistical Analysis of the Effect of Bar Position on the Flexure
Strength for both Microwave and Conventional Heating. ........ 199

Table 5-6. Effect of Position/Heating Method Interaction on the Flexure
Strength ........................................ 203


Figure 1-1. Schematic depicting the intersection of two fields of modeling to
form a basis for research. ............................. 5

Figure 2-1. Representation of the lag in the current that determines the loss
tangent ......................................... 10

Figure 2-2. Ratio of transmitted to incident field for aperture diameter,
0.05< d/X
Figure 2-3. Radial temperature profiles in a cylindrical roast beef. ............ 20

Figure 2-4. Temperature of bread heated with two different field strengths. ..... 21

Figure 2-5. Effect of variable (-) versus constant (---) dielectric properties on
temperature in a raw beef sample. Depth in sample refers to fraction
of whole ......................................... 22

Figure 2-6. Measured and computed temperatures for the specimen and the
insulation during a microwave sintering run. ................. 24

Figure 2-7. The activation energy for 180 diffusion in sapphire measured using
both conventional and microwave heating. ................... 27

Figure 2-8. Diffusion of potassium in sodium aluminosilicate glass using
conventional heating and two different microwave power levels at a
temperature of4500C for 30 minutes. ..................... 29

Figure 2-9. Sketches of a) Typical setup used for joining ceramics in a hot press
and b) ceramic shapes commonly joined in a hot press. .......... 34

Figure 2-10. Schematic showing the six mechanisms of mass transfer that can
lead to joining; a) surface sources, b) bonding interface sources c)
bulk deformation after yield or during creep. ................. 36

Figure 2-11. Contact area of a) textured surface and b) smooth surface. Contact
area in a) is greater than in b) when the external dimension is the
same. ...................... .................... 37

Figure 2-12. Effect of surface roughness (as described in equation 3-1 and figure
3-4) on sample strength .............................. 39

Figure 2-13. Effect of joining temperature on the strength of a joint. ........... 40

Figure 2-14. The effect of pressure applied during joining on the strength of the
joint ................. ......................... 42

Figure 2-15. The effect of joining time on the strength of the joint. ............ 43

Figure 2-16. Graphic representation of the four scales of joining discussed in the
text ........................................... 45

Figure 2-17. Energy versus separation for two ions in proximity to each other. ... 47

Figure 2-18. Force versus separation for the two ions in figure 2-17. .......... 47

Figure 2-19. Computer derived, relaxed structure of the BaO/NiO interface. ..... 48

Figure 2-20. Effect of joining temperature on the strength of 92% alumina. ...... 55

Figure 2-21. The effect of joining time on bend strength of silicon nitride. ....... 56

Figure 2-22. Knoop hardness across joint for microwave joined alumina. ....... 57

Figure 2-23. Acoustic emission traces for a mullite joining experiment. ......... 58

Figure 3-1. Binary Cr203-Al203 phase diagram. ...................... 67

Figure 3-2. Binary Fe203-Al203 phase diagram. ...................... 67

Figure 3-3. Setup used to measure temperature of gel compositions heated solely
by microwaves (setup was placed in a microwave for heating). ..... 69

Figure 3-4. Simulated material profile depicting the arithmetic mean. .......... 73

Figure 3-5. Schematic of the microwave joining apparatus. ................ 78

Figure 3-6. Photograph of the microwave joining apparatus used in all of the
microwave joining experiments. ......................... 79

Figure 3-7. Flowchart used to produce susceptors for microwave joining. ...... 81

Figure 3-8. Cutaway view of the setup used to evaluate the heating ability of
susceptors ........................................ 82

Figure 3-9. Schematic of the specimen/susceptor setup used in the microwave
joining experiments .................................. 84

Figure 3-10. Photograph of susceptor set up inside the microwave cavity. ....... 85

Figure 3-11. Flow diagram of the microwave joining process and specimen
preparation ....................................... 88

Figure 3-12. Flexure fixture for room temperature testing of bars using four-point
bending, a) schematic of flexure rig and b) labels of relevant
dimensions. ....................................... 89

Figure 4-1. Grid structure surrounding node o, the node of interest. .......... 99

Figure 4-2. Two-dimensional grid structure using the r-z coordinate axes. ...... 99

Figure 4-3. Model used for the simulation of microwave hybrid heating. ...... 100

Figure 4-4. Time-temperature coordinates involved in the Crank-Nicolson
equation. The trapezoid connects the time-position values used to
calculate temperature values at the point indicated,0 ............ 102

Figure 4-5. Logarithmic fit to data for the specific heat of alumina. .......... 115

Figure 4-6. Linear fit for the specific heat of nickel oxide. ................ 116

Figure 4-7. Polynomial fit to thermal conductivity of insulating fiber board. .... 117

Figure 4-8. Temperature of water as a function of time. ................ 119

Figure 4-9. Power absorbed (calculated) as a function of temperature for AD995
alumina. ......................................... 121

Figure 4-10. Power absorbed (calculated) as a function of temperature for nickel
oxide. .......................................... 122

Figure 4-11. Power absorbed (calculated) as a function of temperature for the 30
vol. % SiC susceptor. ............................... 124

Figure 4-12. Curve fit for the emissivity of alumina as a function of temperature. 125

Figure 4-13. Sample input to calculate a heating profile of the model in figure 4-4.
..................... ..... ..................... 129

Figure 4-14. Map of node numbers for the input file in figure 4-13. Geometry
corresponds to the model in figure 4-3. The nodes that represent
thermocouple locations are 328 interlayerr) and 423 (bulk). The
pyrometer is represented by node 328. ................... 131

Figure 4-15. Temperature profiles for simulated microwave hybrid heating
generated using the input in figure 4-13. a) 10 minutes; b) 20
minutes; c) 30 minutes; d) 40 minutes; e) 50 minutes. .......... 132

Figure 5-1. Area of the surface of alumina end member used for roughness
analysis ........................................ 138

Figure 5-2. Profile of 5 mm x 5 mm area of alumina from figure 5-1 analyzed
using a laser profilometer. Light regions correspond to the area
above mean peak height and darker regions correspond to area
below mean peak height. ............................. 139

Figure 5-3. Area of the surface of nickel oxide interlayer used for roughness
analysis. ....................................... 140

Figure 5-4. Profile of 5 mm x 5 mm area of nickel oxide from figure 5-3
analyzed using a laser profilometer. Light regions correspond to the
area above mean peak height and darker regions correspond to area
below mean peak height. .............................. 141

Figure 5-5. Temperature of the potential interlayer materials heated using stand
alone microwave heating. .............................. 142

Figure 5-6. Differential Scanning Calorimetry (DSC) of the base gel, alone and
with additives. .................................... 144

Figure 5-7. X-ray Diffraction Analysis (XRD) of powders synthesized during
DSC analysis. Specimens containing iron compounds exhibited
spectra with reduced intensities. This is most likely due to the
fluorescence of iron from Cu k-alpha radiation. Fluorescence tends
to reduce peak/background ratios. ....................... 146

Figure 5-8. Dielectric constant versus temperature of potential interlayer
materials. Measured at 2.46 GHz ........................ 148

Figure 5-9. Dielectric constants, measured at 2.46 GHz, plotted without the iron
(11) oxide composition .............................. 149

Figure 5-10. Dielectric loss factor, measured at 2.46 GHz, versus temperature for
potential interlayer materials. .......................... 150

Figure 5-11. Dielectric loss factors plotted without the iron (III) oxide and nickel
oxide compositions. Measured at 2.46 GHz. ................ 151

Figure 5-12. Loss tangent, measured at 2.46 GHz, versus temperature for
potential interlayer materials. .......................... 152

Figure 5-13. Loss tangents plotted without iron (III) oxide and nickel oxide
compositions. Measured at 2.46 GHz. .................... 154

Figure 5-14. Temperature of alumina load in silicon carbide/alumina cement
susceptors heated using 3.2 KW with a 75% duty cycle (% time
on) ............................................. 155

Figure 5-15. Dielectric constant, measured at 2.46 GHz, versus temperature for
several compositions of susceptors. ..................... 156

Figure 5-16. Dielectric loss factor, measured at 2.46 GHz, versus temperature for
several compositions of susceptors. The minimum in the loss
factor, associated with some types of silicon carbide, has yet to be
explained. ....................................... 157

Figure 5-17. Loss tangent for several susceptor compositions at 2.46 GHz. ..... 158

Figure 5-18. Temperature of susceptors as a function of silicon carbide content. .. 159

Figure 5-19. Microwave processing conditions for trial #6 (14500C; 45 min; 3
MPa; nickel oxide interlayer). ......................... 161

Figure 5-20. Conventional processing conditions for trial #11 (1550C; 15 min; 1
MPa; nickel oxide interlayer). ......................... 162

Figure 5-21. Fracture surfaces of specimens from trial #4 (45 min; 1550'C; 1
MPa; nickel oxide interlayer; microwave heated). ............ 164

Figure 5-22. Fracture surfaces of specimens from trial #6 (45 min; 1450C; 3
MPa; nickel oxide interlayer; microwave heated). ............ 165

Figure 5-23. Fracture surfaces of a flexure bar joined using a gel-derived (4
mol% Cr203) interlayer .............................. 166

Figure 5-24. Mean flexure strength of bars for each of the trials in the
experimental design. Error bars correspond to one standard
deviation .............. .......................... 168

Figure 5-25. Fracture surface of specimen from trial #12, representative of all the
specimens joined using AD94 alumina as the interlayer material. ... 170

Figure 5-26. Joint line between AD94 alumina interlayer and AD995 alumina end
member (250x). Trial #14 (45 min/1450C/3 MPa/C/AD94). ..... 171

Figure 5-27. Micrograph of joint region in figure 5-26 showing porosity and
grain boundary material (20,000x). ..................... 172

Figure 5-28. Intact interlayer material in bar that fractured away from the joint
area. Trial #15 (15 min/15500C/3 MPa/C/AD94). ............. 173

Figure 5-29. Joint region of bar from trial #15, flexure strength 220 MPa (250x).
Trial #8 (45 min/1550C/3 MPa/M/AD94). .................. 175

Figure 5-30. Magnified micrograph of the joint line in figure 5-29 (20,000x). 176

Figure 5-31. Joint area of a bar microwave joined in trial #3 (15 min/15500C/l
MPa/M/AD94) (6,000x). a) micrograph and b) WDS x-ray map of
silicon ......................................... 177

Figure 5-32. Joint area of a bar conventionally joined in trial #15 (15
min/1550C/3 MPa/C/AD94) (6,000x). a) micrograph and b) WDS
x-ray map of silicon. ................................ 178

Figure 5-33. Joint area of a bar microwave joined in trial #2 (45 min/14500C/l
MPa/M/AD94) (6,000x). a) micrograph and b) WDS x-ray map of
silicon ......................................... 179

Figure 5-34. Joint area of a bar microwave joined in trial #8 (45 min/15500C/3
MPa/M/AD94) (6,000x). a) micrograph and b) WDS x-ray map of
silicon ......................................... 180

Figure 5-35. Matching fracture surface of bar joined with a nickel oxide interlayer
that failed before flexure testing. Compare to Figures 5-36 and 5-
37. Trial #11 (15 min/15500C/l MPa/C/NiO). ............... 181

Figure 5-36. Matching fracture surfaces of bar joined with a nickel oxide
interlayer with a flexure strength of 38 MPa. Trial #11 (15
min/1550OC/l MPa/C/NiO). ............................ 182

Figure 5-37. Matching fracture surfaces of bar joined with a nickel oxide
interlayer with a flexure strength of 45 MPa. Texture difference is
due to slow crack growth from controlled loading (slow rate). Trial
#11 (15 min/15500C/1 MPa/C/NiO). ..................... 183

Figure 5-38. Micrograph of joint region of bar from figure 5-35 joined with a
nickel oxide interlayer (2,000x). Flexure strength 0 MPa. ....... 185

Figure 5-39. Micrograph of joint region of bar from figure 5-35 joined with a
nickel oxide interlayer (5,000x). Flexure strength 0 MPa. ....... 186

Figure 5-40. Micrograph of joint region of bar from figure 5-37 joined with a
nickel oxide interlayer (1,000x). Flexure strength 45 MPa. ...... 187

Figure 5-41. Micrograph of joint region of bar from figure 5-37 joined with a
nickel oxide interlayer (5,000x). Flexure strength 45 MPa. ...... 188

Figure 5-42. Roughness of joined surface (trial #8) calculated from micrograph
taken at 500x. ..................................... 189

Figure 5-43. Roughness of joined surface (trial #12) calculated from micrograph
taken at 500x. ..................................... 190

Figure 5-44. Roughness of as-machined surface calculated from micrograph
taken at 500x. ..................................... 191

Figure 5-45. Effect of time on the flexure strength of joined bars. ........... 194

Figure 5-46. Effect of temperature on the flexure strength of joined bars. ...... 195

Figure 5-47. Effect of pressure on the flexure strength of joined bars. ......... 196

Figure 5-48. Effect of the heating method on the flexure strength of joined bars. .. 197

Figure 5-49. Effect of the interlayer material on the flexure strength of joined
bars. ......................................... 198

Figure 5-50. Effect of the heating method on the flexure strength of joined bars
plotted with error bars at the 90% confidence interval. .......... 200

Figure 5-51. Map of original bar positions in the joined cylinder before
machining ....................................... 201

Figure 5-52. Interaction chart showing the combined effect of the position of the
bar and the heating method on the flexure strength. ............ 202

Figure 5-53. Temperature of the workpiece measured using a thermocouple and
an optical pyrometer. ................................. 205

Figure 5-54. Variation in temperature of the workpiece between the thermocouple
and the optical pyrometer. ............................. 206

Figure 5-55. Temperature of the susceptor measured using both a thermocouple
and an optical pyrometer. .............................. 207

Figure 5-56. Variation in temperature of the susceptor between the thermocouple
and the optical pyrometer. ............................. 208

Figure 5-57. Temperature (calculated and measured) at the surface of the AD94
interlayer. ........................................ 209

Figure 5-58. Temperature (calculated and measured) at the surface of the NiO
interlayer. ........................................ 210

Figure 5-59. Temperature (calculated) at the surface of the joint region for various
heating conditions. .................................. 211

Figure 5-60. Temperature (calculated) at the surface of the end member for
various heating conditions. ........................... 213

Figure 5-61. Temperature (calculated) at the surface of the joint region (AD94)
and end member (AD995). ............................ 214

Figure 5-62. Temperature (calculated) at the surface of the joint region (NiO) and
end member (AD995). ................................ 215

Figure 5-63. Comparison of calculated temperatures for specimens heated with
nickel oxide interlayers of different thickness. ................ 216

Figure 6-1. Phase diagram of the FeO-A1203 system. ................... 220

Figure 6-2. Qualitative representation of the effect of temperature on the
dielectric loss factor of a typical dielectric material. Tc refers to the
critical temperature at which the efficiency of the absorption of
microwaves increases dramatically. ...................... 222

Abstract of Dissertation Presented to the Graduate School of the University of Florida in
Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy




August 1996

Chairperson: Dr. David E. Clark
Major Department: Materials Science and Engineering

Microwave processing of materials is advancing from a novel processing technique

into a viable method for processing industrial materials. It has been adopted by several

industries for the following reasons. Microwave processing can afford savings in energy

or space or, more importantly, it can provide a product with qualities that cannot be

achieved through conventional means. Even as the use of microwave processing increases

in industry, it is vigorously investigated by researchers attempting to better understand and

control the mechanisms that are responsible for microwave heating of materials. These

investigators are seeking evidence that can either demonstrate or disprove the concept of

enhanced microwave processing. Simultaneously, researchers also are looking for

innovative techniques that can provide other industrial applications for microwaves.

Joining has always been an integral component of the manufacturing industry. The

joint is often the weakest link of any manufactured item and researchers are continually

pursuing means of increasing the strength of the joint. Several ceramics have been joined

successfully using microwave energy. However, there are no processing guidelines

available so the data that have been generated cannot be easily compared. Furthermore,

much of the previous work reports the condition of the joint qualitatively, making

comparisons of joining procedures nearly impossible.

The research performed in this study attempts to address many of the issues

discussed above. The main goal of this research is to join alumina using microwave energy

and to develop an understanding of the microwave interactions with materials and its

relation to the joining process.

An apparatus for the microwave joining of materials was designed and pu into use.

A procedure for producing susceptors for microwave hybrid heating was developed.

Using the microwave joining apparatus and hybrid heating, high-purity alumina was joined

using several materials as the interlayer between the alumina end members. An

experimental design was used to provide data that then could be analyzed statistically to

determine the significance of the processing parameters varied during joining. A numerical

model simulated the microwave heating of the joining process. Simulated temperature

profiles were compared to experimentally measured temperatures to assess the usefulness

of the model.



Joining is an important step in the construction of engineering structures. This may

seem obvious when one thinks about the erection of a building or the assembly of an

automobile. However, joining is also a vital process in the formation of many of the

components that are later joined to produce these buildings and automobiles. The joining

process also extends beyond mechanical joints formed through the use of fasteners such as

bolts, nails and clamps. A remarkable number of products contain components that are

joined through chemical reactions.

While it is often preferable to fabricate the components that are used to assemble

products in one piece, it is not always possible or even the prescribed method for

manufacturing the desired shape. This is most apparent when the final product consists of

more than one material. Some of the newer types of cookware consist of a piece of

stainless steel sandwiched between two aluminum sheets. The stainless steel provides a

high heat capacity base needed for stove top cooking, while the aluminum contributes a

high thermal conductivity for even heating. Another example where joining is the preferred

method of manufacture is when the piece consists of a complex shape that would require

either a complicated mold or extensive machining. Joining can also be useful when a

compressive stress is added to increase the strength of the component. This effect is

observed when glaze is fired on a ceramic piece with a lower coefficient of thermal

expansion than the glaze. When the product cools, the glaze will contract more and be put

under a compressive stress.

All types of materials--ceramics, metals and polymers--have been joined to one

another for a multitude of applications. The applications that entail the joining of ceramics

are extensive. The most common joint involving ceramics is a ceramic to metal joint. These

types of joints are used in light bulbs to connect the metal base to the glass tube, and also in

spark plugs, to attach the metal core to the ceramic sheath. Ceramic/metal joints are created

in the enameling of household appliances such as washing machines and dishwashers. The

enamel coating on the steel housing not only increases the aesthetic value of the appliance,

but also serves as a rust-resistant coating on the steel. While the ceramic/ceramic joint is

somewhat less prevalent than the ceramic to metal joint, it plays an important role in the

traditional ceramic industries. Two of the most widely used examples of ceramic/ceramic

joints are joining bricks with mortar and attaching a handle to a ceramic mug. The use of

mortar to join bricks relies on the room temperature cementitious reaction. This reaction

causes the mortar to harden, thus joining the two bricks. The practice of joining a handle to

a ceramic mug is similar to joining bricks with mortar in that the interlayer material contains

an appreciable amount of water. When attaching handles to mugs, the interlayer material is

usually the same composition as the end members (cup and handle). The joint is formed

during the sintering of the interlayer material with the end members at elevated

temperatures. While these applications are less glamorous than the heat resistant coating on

space shuttle tiles, they make up a much greater share of the ceramic market. In the field of

structural ceramics, the greatest motivation to apply joining techniques is the size limitations

placed upon the part by the densification apparatus. Large parts such as heat exchanger

tubes can exceed fourteen feet in length. This requires the joining of smaller sections to

complete the part.

The formation of a ceramic to ceramic joint is not as straightforward as the typical

metallic joint. Most metallic joining operations include the formation of a liquid phase.

The liquid phase is usually made up of material from both end members and produces a

joint with properties that gradually change from those of one end member to those of the

other. The properties of ceramic joints made without liquid phase can change abruptly

from one end member to the other. Ceramic to ceramic joints are pursued because they

provide an opportunity to improve the final properties of the product. Ceramic to ceramic

joints can furnish an increased resistance to operating temperature over metallic alloys.

They also are more resistant to oxidation than most metal alloys. The enhanced resistance

to degradation and oxidation of ceramics over metals in high-temperature, corrosive

applications can be used to increase the efficiency of heat exchangers and turbine engines.

The joint area in this dissertation is considered to be the area that encompasses the

interlayer and the portion of the end members that has been altered during the joining

process. The joint can be described in three ways: according to its chemical,

microstructural or physical properties. In the first description, joining can be described

chemically by the mixing of the atoms of the components being joined. The mixing of the

atoms of the two components can result in the formation of a third component. When the

mixing of atoms is limited to materials that have mutual solubility, the concentration of

atoms gradually changes from 100% of the first material to 100% of the second material.

Secondly, the joint region may possess a microstructure similar to one or both of the pieces

being joined or it may be unique. A third classification would be to define the joint

physically. This is the manner in which the end members are macroscopically connected or

somehow attached. Although basic in description, this is the definition of "joined" that

matters most to consumers.

The definition of "joint" depends on the desired use of the end product. For the

ceramic to metal joint of the light bulb described above, hermeticity of the joint is the most

important feature with the strength of the joint only having to meet some minimum

requirement. For the purpose of this dissertation, "joint" will be referred to as the region

between two bulk materials that has a chemical composition or mechanical properties

different from either of the end members or the interlayers used in the joining process.

Materials are determined to be "joined" when they do not separate during normal handling.

Until recently, most of the work performed in the joining of materials has used

conventional heating. That is, the specimens to be joined were heated in either a resistance

or radio frequency (RF) heated furnace. Radio frequency heating is limited to materials that

are electrically conductive and capable of supporting an induced current. There has also

been some work done using lasers to heat the workpieces. Lately, microwaves have been

used to provide the energy necessary to heat the specimens to temperatures suitable for

joining. This has brought about a discussion of whether microwave heating can contribute

to the joining process in some way besides its use of an alternate heat source. There are

many reports of enhanced diffusion and lowered activation energies of transport that will be

discussed later. It is certain that if either of the above mentioned phenomena can be shown

to occur during the joining process in a microwave field, there could be a definite advantage

in the use of microwaves to join ceramics.

The goal of this research then, is to join alumina using microwave energy and to

develop an understanding of the joining process. In the literature, there are many models to

predict joining behavior between two materials of identical, similar or differing

compositions using conventional heating. There is also an abundance of experimental data

to support these models. Several models of microwave interactions with materials also can

be found in recent publications. These models and the experimental data to support them

are not as prevalent. In attaining the goal of this research, it is necessary to identify, and

begin to establish, the intersection of these two separate fields of study, figure 1-1.

Models of Joining My Work Models of Microwave/
Materials Interactions

*equilibrium at interface *evaluate properties of -microwave most important
*surface texture ceramics joined using >communications
*mass transfer mechanisms microwave energy >probe
*residual stress elimination -material most important
>effect on properties

Figure 1-1. Schematic depicting the intersection of two fields of modeling to form a
basis for research.

While working toward the stated goal, the following objectives were identified and

* Review the fundamentals of joining.
* Review the principles of microwave interactions with materials that are relevant to
* Design and construct an apparatus for microwave joining.
* Prepare joined specimens that have strength approaching that of the original material.
* Evaluate the integrity and uniformity of joints based on strength.
* Demonstrate enhancements, if any, achieved during microwave joining.
* Use statistical tools to qualitatively assess the joining parameters used in this study.
* Employ numerical modeling to approximate heat flow during joining.
These objectives form the framework of the research performed. They also acted as

guidelines throughout the course of the study.


Several significant achievements have been realized during the course of this study.

These achievements either extend beyond the anticipated success of the associated objective

or, were accomplished in addition to the objectives. 1) A new generation of microwave

susceptors has been developed and characterized. 2) A microwave joining apparatus was

designed and constructed using closed loop control of the applied force and temperature.

3) An experimental design to provide the data for a statistical analysis of the effects of

joining parameters on flexure strength was constructed. 4) A heat transfer program was

manipulated to generate temperature profiles in a model that compares calculated

temperatures with experimentally measured temperatures. The calculated values provided

results comparable with other microwave heating models available in the current literature.





Like visible light, radio waves and x-rays, microwaves are a part of the electromagnetic

spectrum. Microwaves extend between the frequencies of 300 megahertz (MHz) and 300

gigahertz (GHz) This corresponds to wavelengths ranging from about one meter to less

than one millimeter in length.

In the past, microwaves were used almost exclusively in the field of communications2.

Any electromagnetic losses in the materials used were considered detrimental. The

relatively recent introduction of microwave ovens (1960s) to the general public was a result

of the realization that electromagnetic losses could be used for the purpose of cooking.

This was possible because the interactions of microwaves at 2.45 GHz with the water in

foods provided efficient and rapid heating. In 1985, Meek3 exploited the generation of heat

due to the electromagnetic losses in materials that do not contain water. He was able to heat

a ceramic material to a high enough temperature to induce melting. Shortly thereafter,

Meek4 proposed a model for the microwave sintering of powders of dielectric materials.

He describes the powder compact as a material/porosity (air) composite. Meek then

suggests that the higher electric field in the porosity induces densification in nearby grains.

Metaxas and Meredith5 furnish a thorough explanation of the theory behind the

microwave heating of materials. Heat generated due to losses in a material is related to the

electric permittivity (E*) and magnetic permeability (I*) of the material. Most ceramic

materials are not significantly influenced by the magnetic portion of the wave, therefore the

contribution to heating produced by the magnetic losses are generally ignored. Fathi6

provides an excellent review (from a materials science perspective) of the theory behind the

interactions of microwaves with materials. He discusses the different types of polarization

and the losses associated with them. Space charge polarization and orientation polarization

are the two active mechanisms in the microwave frequency range. Space charge

polarization refers to the charge build up at the interfaces of a non-homogeneous material in

an alternating electric field. Orientation polarization occurs in materials that are comprised

of permanent dipoles that have asymmetric charge distributions. The dipoles tend to

reorient in an alternating electric field. Fathi also describes the historical development of

the theory and equations pertaining to the dielectric heating of materials. The electrical

permittivity of a material is defined by the equation

E* = e' je" = e( E J"E) (2-1)
e' : dielectric constant (real)
e" : dielectric loss factor (imaginary)
E'r : relative dielectric constant (E'/Eo)
eeff: effective relative loss factor = E"r + o/Eo(0
"r : relative loss factor
S : electrical conductivity (ac + dc)
oE : permittivity of free space
S : angular frequency = 2cf,f in Hz.

The dielectric constant (real) and the dielectric loss factor (imaginary) of a material can be

measured experimentally. Using cavity perturbation theory, Hutcheon et al.7 were able to

measure the complex permittivity of several materials. They then were able to

mathematically isolate the real and imaginary portions of the complex permittivity. These

values describe the polarization and conductivity losses generated in a material. The loss

tangent, tan 8, relates the real and imaginary terms:

tan 8=- (2-2)

The value of tan 8 determines the amount of energy dissipated in a material exposed to an

alternating electric field. In an ideal dielectric, the induced current will lead the applied

voltage by an angle of 900 and 8 = 0. In a real material, the time required for polarization

causes a phase shift in the induced current in the dielectric, represented by the angle 8 in

figure 2-1. These losses are temperature dependent and can change rapidly over small

temperature regimes. Hutcheon et al.8 were able to measure the complex permittivity in a

material up to 1400'C. This made it possible to predict heating behavior of the material in a

microwave field. Another equation developed by Metaxas and Meredith5 that is necessary

to understand microwave heating identifies the average power absorbed in a unit volume,

Pay as

Pav= oEo EE.2 (W/m3) (2-3a)

in terms of tan 6

Pv = 27fee' tan E2 (W/m3) (2-3b)
o : angular frequency = 2nf
f : frequency of microwaves, in GHz
Erms :internal electric field (root mean square) (V/m).

The complexity of this equation arises in the determination of the internal electric field,

which can vary with position and temperature. The authors also derive and simplify other

I, current in
ideal dielectric

,L -loss current in
dielectric material

'I induced current in
dielectric material

Vapp applied voltage

Figure 2-1. Representation of the lag in the current that determines the loss tangent. The
dielectric conductivity is a real number so the loss current is in phase with
the applied voltage,Vapp.

pertinent equations such as the penetration depth of the microwaves in the material, Dp.

The penetration depth is defined as the distance from the material's surface to the point at

which the power drops to e-1 of its surface value. This equation is also affected by

temperature due to the temperature dependence of "eff and E',

S" (2-4a)
Dp =---- [(1+( )
27c(2e')' E

For low loss dielectrics ("eff/ e'l 1) the equation is

'0(o') 2
Dp (2-4b)

where 0'o is the wavelength of the electric field in free space. The errors associated with

the simplification of the penetration depth increase as the ratio E"eff/ approaches one.

Table 2-1 contains some representative values of ratios and errors calculated for a

frequency of 2.45 GHz.

Table 2-1. Error Associated with Simplified Equation for Penetration Depth for Several
e"eff/ E' Ratios.

eff/ e' % error
0.009 0.1
0.3 1

0.7 5

1 9

Microwave Applicators

The difference between the two most common types of microwave ovens is in the

microwave cavity where the material is processed. A multimode cavity is a metal box with

an opening in which microwaves are introduced. The microwaves, once introduced into

the cavity, reflect off the walls setting up a number of standing waves. The electric field in

multimode cavities is not very uniform and requires either a turntable or a mode stirrer

(rotating metal fan) to improve the uniformity of the electric field. A single mode cavity has

dimensions that are more precise than the multimode cavity. The length of the single mode

cavity is approximately the size of the wavelength of the microwave frequency being used.

One wall is adjustable so that the cavity can be tuned to the precise length to produce a

single standing wave. The advantage of multimode microwave ovens is that they can be

fabricated in almost any dimension and are relatively inexpensive. Single mode

microwaves have a calculable electric field that permit the identification of the position of

the high field strength. As opposed to single mode microwaves, multimode microwave

ovens are not capable of developing the power densities necessary to produce efficient

heating. Power densities of 107 kW/m3 can be achieved in a tuned single mode cavity5.

Typical power densities of 3 kW/m3 are produced in multimode microwave ovens for home

uses. In a tuned single mode cavity, incident and reflected waves are superimposed and the

position of maximum electric field can be determined. The specimen to be heated can be

positioned to provide the optimum transfer of energy. However, in a multimode cavity the

microwaves are constantly being perturbed and scattered, minimizing the chance that a

specimen is positioned at the point of maximum electric field for more than a few seconds.

In order to heat most ceramic materials in a multimode microwave oven, it is

beneficial, if not essential to use microwave hybrid heating (MHH). Microwave hybrid

heating incorporates a strong microwave absorbing material into a surrounding piece of


insulation. This material absorbs more microwaves than the material being processed.

Silicon carbide is the most common material of choice as the microwave absorbing material

in MHH. Silicon carbide is an excellent and inexpensive microwave absorber that can be

used in air at elevated temperatures. Above 9000C, the silicon carbide will oxidize to form

a protective silica layer. The weight gain associated with the oxidation increases as the

temperature increases9. The combination of the absorbing material and the insulation is

commonly called a microwave susceptor. The purpose of the susceptor is to assist the

material of interest in heating to the processing temperature. The development of

microwave susceptors has empowered researchers to investigate low loss materials such as

high-purity alumina using multimode microwave ovens.

One of the common modifications made to a microwave cavity is the addition of

holes to provide access for either a thermocouple, optical pyrometer or other probes. Holes

also may be necessary for the application of pressure to a workpiece in a microwave cavity.

The presence of holes in the cavity necessitates the use of some type of choking system to

prevent microwaves from escaping and presenting a potential health hazard. Holes of a

diameter less than one-quarter of the wavelength (3.06 cm for microwaves at 2.45 GHz)

theoretically emit no microwave energy. This originates from the equation

Z = Ztan(2ntd/X) (2-5)

Z = input impedance of hole
Zo = characteristic impedance of hole
d = diameter of hole

, = wavelength of microwaves in cavity.

If d = V/4, the input impedance goes to infinity and the hole acts as a open circuit. The

microwaves are reflected back into the cavity as if the hole did not exist. At aperture sizes

from 0 < d J1/2, there is a small impedance associated with the aperture than can lead to a

small emitted electric field. In reality, the wall material (usually aluminum or steel) has a

finite impedance and can emit a portion of the incident microwaves. Cathey to developed

two equations to calculate the ratio of the electric field scattered through an aperture to the

incident electric field,

E, d ] z d
Sx exp 1 I (2-6a)


it 2d n1 2d
E, 4 x 8 1 z d (2-6b)
Ei z 4A


Es = scattered electric field, V/m
Ei = incident electric field, V/m
d = aperture diameter, m
X = wavelength of microwave, m
z = distance from aperture, m.

This is the case for the viewing port for an optical pyrometer and the thermocouple

feedthrough. At an aperture diameter to wavelength ratio of 0.5, the intensity of

microwaves transmitted through the aperture becomes significant. Proximity to an aperture

greater than one-half the wavelength of the microwaves can pose a health hazard. Cathey

then calculated the ratio of transmitted to incident field for a range of aperture diameters,

figure 2-2. It can be ascertained from figure 2-2 that for an aperture of X/4 or less, the

fraction of transmitted power is small and attenuates quickly. As the aperture size is

increased, the intensity of the transmitted field increases greatly. In the situation where a

load is to be applied within the cavity, the hole, or aperture, is partially filled with a

0 0.25 0.5 0.75 1.0 1.25 1.5 1.75

Figure 2-2. Ratio of transmitted to incident field for aperture diameter,
0.05 < d/< 110


dielectric material. This material, typically alumina, is used to transfer the load from the

external load cell to the workpiece. Here, the wavelength of the microwaves is decreased

by the square root of the dielectric constant of the filling material5. The decrease in

wavelength makes it possible for the incident microwaves to pass through a smaller

aperture than expected. This condition requires additional modifications to effectively

attenuate the microwaves. Metaxas and Meredith5 describe the use of a circular tube of

conducting material as a choke. If the diameter of the tube is less than one-half of the

incident wavelength, the microwaves are attenuated and the transmitted power decreases

rapidly. Under these conditions (d < 1/2), the attenuation of the power is proportional to

e-z, where X is the wavelength of microwaves in the dielectric material and z is the length of

the choke.

Microwave Processine of Materials

In the past fifteen years, numerous efforts to take advantage of the unique capabilities

of microwave processing have been undertaken. Suttonl outlines the potential advantages

of several characteristics of microwave processing over conventional heating processes.

These advantages include 1) uniform heating of large sections of material, 2) heating above

2000"C in air, vacuum or controlled atmospheres, 3) rapid processing with reduced grain

growth, 4) accelerated sintering and diffusion due to high electric fields, 5) selective

heating of particular materials or phases and 6) improvement of mechanical properties.

Katz and Blakell demonstrated that uniformity could be achieved over a large area by

simultaneously sintering twenty alumina cylinders. They reported that dimensional

tolerances were maintained from experiment to experiment as well as among specimens

within each experiment. Holcombel2 reported attaining a temperature of 2150"C when

sintering boron carbide and Tian et al.13 reached a temperature of 2200'C when sintering

silicon carbide in a positive pressure argon atmosphere. Other non-oxides have been

processed under flowing argon with the resulting x-ray diffraction pattern primarily

indicating peaks of the desired phase 14. Using microwave hybrid heating, D615 rapidly

sintered high-purity alumina to a nearly fully dense state. In D6's experiments, specimens

were held at the sintering temperature for thirty-five minutes. The lack of significant grain

growth reported was analogous to that exhibited in conventional rapid sintering. The

author also reported uniform hardness measured across a 25 mm diameter specimen.

Moore et al. 16 used the concept of selective coupling to demonstrate the effectiveness of

microwave heating. The authors used a lossy binder in the forming step of alumina

compacts. When heated using microwaves, the binder burned out cleanly leaving the

alumina unaffected. In an attempt to organize their approach to microwave processing,

Janney et al. 17 constructed guidelines for microwave processing at the Oak Ridge National

Laboratory. They describe the specimen/insulation configurations used and the temperature

measurement methods that provided consistent results. The progress demonstrated by

researchers worldwide in the microwave processing of materials gained the attention of the

U.S. government. Stein et al. 18 summarized a report commissioned by the Department of

Defense (DoD) and the National Aeronautics and Space Administration (NASA). This

report assessed the current status of the technology, ascertained which applications of the

technology show promise to provide energy or space savings and recommended future

activities in the field of microwave processing. Two industries in which microwave

processing has progressed from an emerging technology to a mature process are the rubber

industryl9 and the food industry20.

Numerical Modeling of Microwave/Material Interactions

Until now, all of the discussion concerning microwave/materials interactions has

been phenomenological. That is, only the behavior of materials after exposure to


microwaves has been considered. Due to the complexity and multiple interactions of the

components in equation 2-3, many researchers do not consider the theoretical aspects of

microwave heating. Another more practical reason that theoretical approaches are avoided

is the difficulty in measuring the conditions within the microwave during processing. More

specifically, it is very difficult to determine the electric field in a material inside a

microwave oven. Properties such as the dielectric constant and the loss tangent are

temperature dependent. The interaction of the electric field with the dielectric material in the

cavity alters the field through absorption or reflection of the waves.

To calculate the electric field in a microwave cavity, it is necessary to solve

Maxwell's equations simultaneously. Yee21 was the first to develop a numerical solution

for these equations in an isotropic medium by replacing them with a set of finite difference

equations. Previous to Yee's work, scattering of the electric field was calculated by a

laborious integration method that required numerous subdivisions of the material being

evaluated22. Taflove and Brodwin23 extended Yee's findings to determine the field within

a dielectric cylinder.

One of the early applications of the finite difference modeling techniques was used in

response to cataracts that developed in the eyes of microwave generator operators24. The

model was implemented to predict the electromagnetic fields and temperature induced in a

microwave irradiated human eye25. Recently, a similar model was used to determine the

specific absorption rate of various body components26. The authors suggest that it is

possible to take advantage of the potential for selective heating of various types of tissue in

surgical procedures.

The bulk of the research in microwave/material interactions has been performed in the

food industry. Datta27 judged the determination of the fields within a loaded microwave

environment to be too complex to predict accurately. The author concentrated on predicting


the heat and mass transfer associated with microwave processing of foods. Figure 2-3

compares the experimentally and predicted temperatures in a cylindrical roast beef heated

using microwaves. In a rigorous evaluation of microwave induced heat and mass transfer

in baked dough products, Tong and Lund 28 considered the temperature dependence of the

dielectric properties of the materials. The authors were able to measure the field strength

near the subject's surface and use these values for the model. Figure 2-4 compares their

model to experimental results obtained in two different microwave ovens. Pangrle et al.29

introduced the use of finite element analysis to model the microwave thawing of pure water

cylinders. Subsequently, Pangrle et al.30 extended the model to multilayer beef/water

composites. Many authors choose to neglect the temperature dependent properties of the

materials being heated. Ayappa et al.31 used the finite element method to determine the

effect of the temperature dependent properties on the temperature of various materials.

Ayappa et al.32 then extended their finite element model to a 2-dimensional situation.

Figure 2-5 is the temperature distribution of microwave heated raw beef using both

temperature dependent and constant dielectric properties.

In ceramics, Watters et al.33 used difference equations to calculate transient

temperature profiles in a ceramic using a simplified geometry to represent the ceramic

specimen. Gupta and Evans34 developed a 1-dimensional numerical model for chemical

vapor infiltration (CVI) using microwave heating. Using an integral method35 for

microwave material interactions, Gupta and Evans36 calculated the temperature in a

microwave heated 2-dimensional body. In the complicated system of hydrated porous

concrete with variable dielectric properties, Li et al.37 developed a numerical 1-dimensional

model to determine the temperature and steam pressure developed during microwave

heating. Ferber et al.38 were able to calculate stress and temperature distributions for a

simplified geometrical model of a ceramic window subjected to microwave heating.


o 70

S Iexperimental

E 60 computed

6 5 4 3 2 1 0 1 2 3 4 5 6
Radial Distance at Mid Height (cm)

Figure 2-3. Radial temperature profiles in a cylindrical roast beef26.

Time (s)

Figure 2-4. Temperature of bread heated with two different field strengths28.

15 min

10 min


E 5 min
I 67

1 min

0 0.2 0.4 0.6 0.8

Depth in Sample
Figure 2-5. Effect of variable (-) versus constant (---) dielectric properties on
temperature in a raw beef sample. Depth in sample refers to fraction of


The joining of ceramics using microwave energy has been considered

theoretically39. It was suggested that the joint area be positioned in the area of a hot spot to

facilitate joining. However, no experimental data is presented in conjunction with the

numerical model. Dibben et al.40 presented a finite element and a finite difference method

for the modeling of the sintering of ceramics in a multi-mode microwave cavity. Their

effort included solving Maxwell's equations and producing 3-dimensional plots of electric

field and power density as a function of position. The finite difference method was used to

calculate temperatures in a 3-dimensional specimen41. To calibrate their model with the

microwave used for the experiments, water loads were heated and the electric field was

calculated from the rate of rise in temperature of the water. In a 1-dimensional model,

Skamser and Johnson42 developed a system to simulate microwave hybrid heating. The

microwave field was assumed to be uniform throughout the microwave. Temperature

dependent dielectric properties of the alumina fiber specimens were varied; those of the

susceptor were kept constant. Microwave hybrid heating experiments were carried out by

Thomas et al.43 using zirconia as the suscepting material. Measured temperatures of the

workpiece and the insulation were compared to those calculated using a 2-dimensional

numerical model. Differences between the measured and calculated temperatures of the

workpiece were as much as 120C during heating. In the insulation these temperature

differences exceeded 200C. When the temperature of the workpiece reached the sintering

temperature, the calculated and measured temperatures converged. Figure 2-6 is the

computed and measured temperatures of the workpiece and the insulation.

Kriegsmann44 proposed a 1-dimensional model to explain the phenomenon of

thermal runaway that is occasionally experienced during the heating of ceramics using

microwave energy. The author later expanded the model to encompass ceramic fibers and

proposed that there exist stable temperatures for given power levels in systems with highly

1600 -
O --- --- Sample
1400 -

S1000 r -- etA----:-A Insulation
' experiment


400 I
10 15 20 25 30 35 40 45 50

Time (min)

Figure 2-6. Measured and computed temperatures for the specimen and the insulation
during a microwave sintering run43.

specified boundary conditions45. Spotz et al.46 use the principles put forth by Kriegsmann

to compare experimentally determined thermal stability conditions. They concluded that

microwave hybrid heating is an effective procedure for limiting the formation of hot spots

that can lead to thermal runaway. Beale and Arteaga47 proposed the use of a closed loop

control system to use feedback from a thermocouple to prevent thermal runaway during the

microwave joining of ceramics. Numerical simulations by the authors indicate that this

system demonstrates potential for experimental applications.

One element of the modeling that has been a constant source of difficulty is resolving

the electric field strengths through Maxwell's equations. Ayappa et al.48 investigated the

applicability of using Lambert's law to predict electric field strengths. Lambert's law

describes the exponential decay of the power flux from an initial intensity and is expressed


I(z) = Ie-2Z (2-7)


Io = transmitted power flux (W/m2),

B = penetration depth (cm) and

z = distance from the sample surface (cm).

The authors have determined a critical slab thickness (Lerit) above which Lambert's law is

valid. The critical thickness is dependent upon the penetration depth, equation 2-4a as

L, ,=2.7P 0.08 (2-8)

The validity of Lambert's law can be quickly determined for any materials system by

calculating the critical thickness at the lowest temperature of interest. That is, the

temperature with the largest penetration depth. This will ensure that Lambert's law will be

valid over the entire temperature range examined. The critical thickness for several ceramic

materials at a microwave frequency of 2.45 GHz are listed in Table 2-2. A sample

calculation of the penetration depth and critical thickness is in Appendix A. When the

critical thickness exceeds the sample thickness, it becomes necessary to solve Maxwell's

equations to determine the power within a material.

Table 2-2. Critical Thickness and Penetration Depth at Different Temperatures for Several
Ceramic Materials at a Frequency of 2.45 GHz.

Material Temperature Penetration Depth Critical Thickness
(C) (cm) (cm)
25 9553 25,793
1000 205 553.4

Al-66 cement 25 127 342.8
1000 28 75.5

nickel aluminate 12 148 4989
1000 5.4 14.5

Evidence of Microwave Enhancements

In the first reported evidence of enhanced diffusion, Janney and Kimrey49 used 180

tracer in a single crystal sapphire substrate in side by side annealing experiments using

microwave and conventional heating. Oxygen was selected instead of a more easily traced

species to avoid differential microwave coupling during the test. Figure 2-7 compares the

activation energies of diffusion for the two processes over a range of temperatures. A

single crystalline sample was used to eliminate microwave interactions with the grain

boundary phase. These interactions contribute to the decrease in activation energy. The

authors do not offer any explanation for the perceived effect. Fathi et al.50 performed ion

exchange in sodium aluminosilicate glasses. The authors exchanged potassium ions from a

salt with the sodium of the glass. The diffusion constants then were calculated from ion

concentrations that were measured using electron microprobe analysis. A variation in the

soda/alumina ratio resulted in differing penetration depths of the potassium ions. These






4.7 4.9 5.1 5.3 5.5 5.7 x 10-4
1/T (1/K)
Figure 2-7. The activation energy for 180 diffusion in sapphire measured using both
conventional and microwave heating49.

409 kJ/mol



results were compared to identical ion exchange experiments carried out using conventional

heating. At soda/alumina ratios near 1:1, the measured diffusion distance of the potassium

ions is much greater in the microwave processed glasses than in the conventionally

processed specimens. As the ratio is varied from 1:1 in either direction, the differences in

the diffusion distance of the potassium ions were less discernible. In another variation of

the experiment, a smaller susceptor was used and more microwave power was utilized to

attain the required temperature. Figure 2-8 compares the use of two different microwave

powers with conventional heating for the ion exchange of K+<=>Na+ at 4500C.

To examine the effect of defect concentrations on diffusion in a microwave field,

Ahmad5s diffused zinc oxide into single crystals of sapphire with differing amounts of

lattice defects. He correlated the concentration of defects with the rate of formation of zinc

aluminate spinel. Similar experiments were executed by Fathi et al.52 using chromium

oxide as the diffusing species. No microwave effect was reported, but the authors

considered this evidence that thermometry was an acceptable measurement technique in a

microwave field. Tiegs et al.53 employed the premise of selective coupling described

earlier to couple microwave energy with the grain boundary phase of a low loss silicon

nitride. A post sintering anneal of the sintered specimens was carried out using both

microwave and conventional heating54. It was reported that in the specimens with a

significant volume of grain boundary phase, there was an improved resistance to creep



Joining Methods

The term "joined" has several meanings in the materials world. It refers to

components that are held together by one of several methods. The two main types of

joining are mechanical joining and chemical joining. Mechanical joining uses fasteners or

0.10 I i

0.08 8 K (800 W)
S"0. (1600 W)
0 0.06 "*

0.04 -

0.02 -


0.00 10 i !!"" I I *Ell.u.
0 10 20 30 40 50 60 70 80 90


Figure 2-8. Diffusion of potassium in sodium aluminosilicate glass using conventional
heating and two different microwave power levels at a temperature of 450C
for 30 minutes50.


clamps to join two or more distinct pieces together. It is probably the most preferred

method of joining because the parts can often be easily separated at a later time for

repositioning or for re-use. Another type of mechanical joining involves the use of

interlocking pieces, such as a tongue-in-groove floor or a dovetail joint on a birdhouse


The other type of joining -- chemical joining -- can be further subdivided into direct

joining and indirect joining. Direct joining is the adherence of two components directly to

each other. These components can be either identical or different. The case in which the

two materials are identical is an attractive process because it minimizes the number of

components involved. It also eliminates the chance of mismatches in the thermal expansion

of the materials, which can lead to a weak joint. Direct joining of similar materials can be

accomplished by heating the area near the joint to a temperature at which a liquid will form.

However, there are not many materials that can be effectively joined in this manner55. In

ceramics, low thermal conductivity can result in material failure due to stresses that develop

during heating. When the two materials to be joined directly are different, the presence of a

compositional gradient provides a driving force for joining. Here also, care must be taken

to minimize the thermal stresses that can be induced during the joining process. Thermal

stresses can lead to low joint strengthss6. These stresses are induced by a mismatch in the

thermomechanical properties of the dissimilar materials.

Indirect joining employs a third component between the end members that facilitates

the bonding of the two parts. This is an excellent alternative to the direct bonding method.

Here, the end members can be either identical or dissimilar in composition. When the

materials are identical, the interlayer provides a compositional gradient to assist in the

joining process. In the case of dissimilar materials, the interlayer material is often chosen

to have thermomechanical properties between those of the end members so that thermal


stresses are minimized57. These interlayers are more commonly referred to as functionally

graded interfaces and will be discussed later in this chapter.

In addition to the various methods of joining, there are many ways to determine

whether components are joined. The final application of the joined materials is instrumental

in defining what is considered to be an acceptable joint. To say that pieces are joined may

mean welding two parts together so there is a gradient of chemical composition or

mechanical properties from one material to the next. It can also indicate the joining of a

film to a substrate where the purpose of joining the pieces is for physical support of the

film. Little or no chemical reaction is desired between the two materials. This joint may be

mechanically weak but still acceptable because it met the minimum requirements for use.

An example of this type of joint would be superconducting films on an alumina substrate.

The superconducting films form a Josephson junction and can detect small magnetic fields.

There is also the case of the ceramic/metal joint in a light bulb where the hermeticity of the

joint is the determining criterion for suitability. In this case it is not so much the strength or

the amount of reaction between the end members that's important, but the ability of the joint

to maintain a vacuum at elevated temperatures.

For the purpose of this work, a joined specimen will be defined as two or more

components that are physically and chemically attached after exposure to a high temperature

and an applied load. Additionally, the components may not be easily separated during

normal handling.

Currently, there are many processes in which the previously described methods of

joining can be accomplished. Joining by mechanical means requires not only knowledge of

the materials being joined, but also requires extensive design skill55. In chemical joining,

some processes involve only the heating of the area to be joined, whereas others require the

entire workpiece to be heated to the joining temperature. The application of pressure on the

joint area is usually required during the joining process to form a suitable bond.

Joining processes

Bolting. Bolting materials together is a popular method of joining metals. With

ceramic/metal joints care must be taken to avoid stress concentrations. This can be done

with the use of a soft washer (either lead or copper) between the metal bolt and the ceramic


Clamping. As with bolting, clamping requires the spreading of the load placed on

the ceramic material. Similarly, bending loads should be avoided and gaskets should be

used at all ceramic/metal interfaces55.

Welding. Oxide ceramics can be joined by welding. A laser or an electron beam is

used to heat the edges of the end members above their liquidus temperature. The liquid fills

the gap between the materials. This can produce intense local stresses as described

previously that can produce cracking during processing or cooling58. There is also the

possibility of the formation of an undesirable microstructure during the recrystallization of

the molten material55.

Brazing. Brazing entails the use of a low melting filler between the end members.

This is very common in the joining of metals because of the ease of the procedure. When

working with advanced ceramics, the use of a low temperature braze material decreases the

usefulness of the joined pieces by lowering their maximum use temperature. Other

undesirable aspects of brazing are a possible mismatch of thermal expansion coefficients or

a chemical reaction between the brazing material and the end member55. In the joining

process called active metal brazing, these phenomena are controllable and therefore


Adhesive bonding. This process incorporates the use of an organic adhesive

between the ceramic and metal pieces. The organic adhesive is typically flexible compared

to the end members. The relative flexibility of the adhesive material can accommodate

stresses that develop between materials with differing thermomechanical properties. Joints

made with organic adhesives are usually limited by the strength of the adhesive and the

temperature at which the adhesive loses its mechanical integrity. An exceptional example of

adhesive bonding is the application of the ceramic insulating tiles to the aluminum skin of

the space shuttle.

Hot pressing. Hot pressing is used for both direct and indirect joining. Hot

pressing consists of heating the entire assembly to be joined and applying a load at some

time during the heating cycle. Direct and indirect joints are both possible using this

process. Pressure can be applied either uniaxially, where the end members are usually

cylinders, plates or blocks or the pressure can be applied isostatically where any shape can

be accommodated. The main limitation of uniaxial hot pressing is the simplicity of the

shapes that may be joined. Figure 2-9 represents a typical uniaxial hot pressing setup and

the shapes that are most often used. Hot pressing lends itself to the application of diffusion

bonding. Diffusion bonding is the formation of a joint through the interdiffusion of two

materials with a minimal amount of material deformation. The key requirements to

diffusion bonding are 1) intimate contact between the surfaces to be joined and 2) sufficient

interdiffusion between the materials being joined to produce a joint in a reasonable time

frame. The application of pressure aids in fulfilling both requirements for diffusion

bonding. It is a relatively simple matter to hold an assembly as shown in figure 2-9 at the

desired temperature and pressure for the extended periods of time that are sometimes

necessary for significant diffusion bonding to occur. The isostatic application of pressure

is a special case of hot pressing called hot isostatic pressing (HIPing). HIPing allows for

application of large pressures and high temperatures. The capital and operational cost of

HIPing limits its use to necessary applications.



parallelepipeds cylinders

-- -

a) b)


Figure 2-9. Sketches of a) Typical setup used for joining ceramics in a hot press and b)
ceramic shapes commonly joined in a hot press.

Criteria for Joining

The joining method of interest in this study is indirect joining through the formation

of a diffusion bond. It is now accepted that the mechanisms of mass transfer responsible

for joining are the same as those that lead to pressure sintering59,60,61,62. Diffusion and

creep mechanisms are responsible for the pore closure that makes up the joining process.

These can be grouped most easily by sources and sinks and are illustrated in Figure 2-10.

Derby and Wallach62, describe the driving forces for each of the mechanisms in Figure 2-

10. The driving force for 1 (lattice diffusion) and 2 (vapor transport) is surface curvature

and is halted when uniform curvature is achieved. Mechanisms 3 (boundary diffusion) and

4 (surface diffusion) are driven by the chemical potential gradient; 5 (viscous creep) and 6

(dislocation motion) are driven by the applied pressure. The relative importance of each of

these mechanisms is determined by the material properties and the time interval required by

the joining process of interest.

In diffusion bonding, there are two distinct factors that must be considered for the

formation of a good joint. The first factor is the preparation of the surfaces to be joined.

For a diffusion bond, it seems reasonable to assume that a mirror smooth surface would be

desirable to provide intimate contact between the two surfaces63. Conversely, it also

makes sense that a larger contact area provided by a heavily textured surface, figure 2-11,

would be conducive to the joining process. The amount of surface texture can be described

quantitatively in terms of the ratio of the contact area to the unit area of the surface. The

smooth surface in figure 2-1 lb would have a ratio of one, whereas the textured surface in

figure 2-11a would have a ratio greater than one. Villagio64 contends that increasing the

area of the surfaces to be joined by enlarging the surface profile from a mirror flat finish

will increase the contact area and yield the strongest joint. He suggested that sinusoidal

shaped surfaces matched together would yield the optimum conditions. The final result is

material a


a) 2

b) jint

5 and 6

c) int

S- material source
Figure 2-10. Schematic showing the six mechanisms of mass transfer that can lead to
joining; a) surface sources, b) bonding interface sources c) bulk deformation
after yield or during creep.

a) textured surface

textured surface

b) smooth surface

smooth surface

external dimension

Figure 2-11. Contact area of a) textured surface and b) smooth surface. Contact area in a)
is greater than in b) when the external dimension is the same.


a compromise between the two theories that includes a textured surface but not enough to

introduce notch stresses at the asperities caused by a reduced area supporting the applied

load. The amplitude of the curve suggested by Villagio is small relative to the period of the

surface. The ratio of contact area to unit area for the ideal surface discussed by Villagio is

only slightly greater than one. In figure 2-12, Suganama et al.65. show that an increase in

surface roughness of the ceramic decreased the strength of a Si3N4-Al brazed joint. In a

review paper, Akselson66 suggests that improved joint strengths could have been attained

using a higher pressure during processing.

The second factor that must be considered in joint formation is the driving forces

responsible for the processes that hold the two pieces together. The impetus for joining can

be in the form of van der Waals forces or with an electronic structure across the interface55.

The goal of all of these mechanisms working together is to reduce the free energy of the

system by eliminating the free surface.

Effect of Process Variables on Joining


The mechanisms that control joining involve either diffusion or plastic deformation.

For the mechanisms that are diffusional, temperature is the most influential variable in the

joining process. This is explained by the presence of temperature in the exponent of the

diffusion equation. However, the maximum temperature for diffusion bonding is limited to

a fraction of the melting temperature of the material to be joined. Akselson66 reports the

temperature range for sufficient joining to be 0.5-0.8 of the lowest melting point of the

materials to be joined. Higher temperatures would lead to excessive creep deformation of

the specimen. Crispin and Nicholas67 calculated the bond strength of a steel/alumina joint

with an aluminum interlayer as a function of temperature. Figure 2-13 plots these values

and compares them with values measured between 0.75-0.95 of the melting temperature of

600 3
base material
500 -........ ...........................................--


c 300
m 200
Ln 100

0 0.1 0.2 0.3 0.4

Surface Roughness (pm)

Figure 2-12. Effect of surface roughness (as described in equation 3-1 and figure 3-4) on
sample strength65.

0 60



S30 calculated values

L5. 2 0
- 20

O9 10[
0 -----
400 450 500 550 600 650 700

Bonding Temperature (oC)

Figure 2-13. Effect of joining temperature on the strength of a joint67.

aluminum. Derby and Wallach62 review the role of temperature in each of the mechanisms

in figure 2-10.


The application of pressure during diffusion bonding has the effect of increasing the

intimate contact between the surfaces to be joined. When the joining temperature has been

attained, the application of pressure assists in deformation near the joint surface that results

in the closure of voids at the interface62. As reported by Allen and Borbidge68 in figure

2-14, at low applied pressures an increase in pressure greatly increases the bond strength.

This effect is diminished as the applied pressure is increased. This indicates that there may

be, for each joint, an optimum pressure where the greatest benefit of increased bond

strength can be realized without significant deleterious deformation of the original



The solid-state joining process of interest is diffusion controlled. The effect of time

on the strength of a joint is similar to that of time on atomic diffusion66. The well-known

equation for diffusion as a function of time is

X = k(Dt)n (2-9)

where X is the distance traveled by the diffusing species, k is a constant, D is the diffusion

coefficient and the exponent, n, is usually 0.5. The analogous expression


where BS is the bond strength, and Bo is a constant was developed by Crispin and

Nicholas67. The equation relates the joining time to the bond strength. Figure 2-15 shows

the effect of bonding time on the strength of an A1203-Al joint. This equation and plot




0 2 4 6 8 10 12
Bonding Pressure (MPa)

Figure 2-14. The effect of pressure applied during joining on the strength of the joint68.

70 -11

50 50

4T 40


m 10
" 20

0 20 40 60 80 100 120 140

Bonding Time (min)

Figure 2-15. The effect of joining time on the strength of the joint67.


assume that there is no chemical reaction in the form of a phase change associated with the

joining process.


While not strictly process variables, reactions in and between the end members and

the interlayer may have a significant effect on the quality of the joint and the bond

strength. In metal/ceramic joining, there is the possibility of controlling the interfacial

reactions. Loehman and Tomsia60 compared results from two different investigations of

the strength of Nb-A1203 joined in a vacuum. They discovered that the formation of a

reaction layer reported by one set of authors agreed with theoretical prediction based on

thermodynamics, while reports by other researches yielded contradictory results. Loehman

and Tomsia suggest60 that these contradictions can be attributed to a difference in the

oxygen partial pressure during processing.

Scales of Joining

Joining can be described on a variety of scales. When two materials are joined

chemically, the interactions are typically discussed considering only the scale of interest.

That is, when the matching of crystal lattices is the topic, it would be discussed on a nano

or atomic scale and the macro and micro scales would not be considered. Figure 2-16

illustrates the four scales of joining that are referred to most often.


On the atomic scale, joining can be thought of as interatomic bonding. This can be

illustrated by looking at adjoining atoms from the two surfaces. In the case of most metal

oxides, the interatomic bond is largely ionic. Kingery et al.70 state that as two oppositely

charged ions approach one another, the molecule becomes more stable. The energy of

attraction is due to the coulombic forces that arise between the oppositely charged ions.

interatomic forces interdiffusion

0000 O 0

atomic / nano

Figure 2-16. Graphic representation of the four scales of joining discussed in the text.

The increase in stability continues until the electron shells associated with the two ions


As described in the Pauli exclusion principle, only one electron is allowed for each

quantum state. As two ions approach each other, a strong repulsive force arises to work

against the coulombic attractive force. The combination of these two forces leads to an

energy minimum in an energy versus separation diagram, Figure 2-17. Ferrante et al.71

identify this energy minimum as the binding energy. The authors also state that the

derivative of figure 2-17 is the force versus separation curve, figure 2-18, and that the

maximum can be considered in simplest terms as the tensile stress (Fma/Asuface) necessary

to separate the two ions.


In the next largest scale examined, the first few atomic planes of each surface are

considered. Ferrante et al.71 discuss the difficulties in increasing the scale in the theoretical

evaluations. They note deviations from symmetry that muddles the theoretical approach.

Some of these deviations are caused by the use of two different planes of the same material

or the existence of two different materials in proximity. To simplify this problem, Clarke

and Wolf72 reduce the interface to a Van der Waals pair potential and sum the potentials

over the interface. This can be used for ceramics when the bonding is localized. Metals

require another approach that is more complicated and will not be discussed in this

dissertation. Stoneham and Tasker73 employed computer codes to model the interfaces

between materials. In the evaluation of the BaO/NiO interface, the authors matched

surfaces by rotating the 100 planes by 45. The resulting relaxed structure, figure 2-19,

placed the Ni3+ ion and vacancies from the disordered NiO lattice near the BaO surface.

This is depicted as the lowest energy structure. Unfortunately, the authors do not consider

any interdiffusion of the cations in their model.

Inflection point

Figure 2-17. Energy versus separation for two ions in proximity to each other71.

FA !

Figure 2-18. Force versus separation for the two ions in figure 2-1771.



)) axis

0) axis

Bulk NiO below
this plane

Figure 2-19. Computer derived, relaxed structure of the BaO/NiO interface73.


Joining on a micro scale can be reviewed in two ways. The most common way to

consider joining is by bringing together two distinct pieces of material. If the pieces are

determined to be joined, at some position in both materials there will be a deviation in

composition and/or structure. If the joining is done properly, the joint may be

indistinguishable from the bulk to the naked eye. Optical or electron microscopes may be

required to locate the position of the joint. Usually, the area where the initial surfaces

contact can be easily identified. Binner et al.74 recently reported grain growth across the

initial interface, making it difficult to recognize. Al-Assafi75 joined alumina of two

different purities using an alumina gel as an interlayer. The original interface was

indistinguishable from either end member. Backscatter electron microscopy was used to

determine the original interface by identifying the silica impurities in the two types of

alumina. Another feature associated with the micro scale of joining is the possible

observation of porosity. When joining fully dense end members, porosity can appear at the

original interface. This usually has a negative effect on the strength of the joint. The micro

scale is also where interdiffusion across the interface can be detected and measured. This

can be done with backscatter electron spectroscopy (BES), Secondary Ion Mass

Spectroscopy (SIMS)51 or with electron microprobe analysis (EMPA)76.

A second way to envision joining on a micro scale is to consider sintering.

Sintering is the joining of many smaller particles to each other to form a large mass. Solid

state sintering is analogous to joining via diffusion bonding. As discussed earlier, the

mechanisms for joining are based upon those developed for sintering. The end members

could be considered as grains and the joint line would represent a grain boundary.


In this paper, the macro scale is considered to encompass those aspects of the joint

that can be seen with the unaided eye. The features of a joint that can be discussed on a

macro scale include poorly formed joints, interlayers, reaction products and abrupt

discontinuities when materials of differing colors are used. When a joint is not fully

developed, the original interface is apparent because the edges of the end members still

contain porosity. Interlayers are almost always a different material from the end members.

Even when the joint is well made, the presence of an interlayer makes joint identification

fairly routine. In the situation where there is a reaction product, such as the joining of

alumina and nickel oxide to form a spinel, the color change provides a good indication of

the extent of the reaction.

Materials Joining

The joining of materials has attracted the interest of a growing number of

researchers. There is vast potential in the joining industry to improve the properties of

joined components. However, there are many theoretical and experimental issues

regarding the joining process that must be addressed if the industry is to fulfill its potential.

Some of the theoretical aspects of the joining process have been discussed previously in

this dissertation. A recent focal point has been the effect of the use of an interlayer on the

stresses developed in the final piece during joining77. Experimental work is being

performed on several elements of the joining process. This includes the development of

new methods for evaluating joint strength and new ways of joining materials. The joining

field includes laminar composites that consist of polymer/metal or ceramic/metal bonding,

and the more common butt joints of metal/metal, metal/ceramic, ceramic/metal/ceramic and

ceramic/ceramic. The most recent addition to the joining field is the functionally graded

interface (FGI). The functionally graded interface can be applied to any of the above

mentioned methods of joining. The joint consists of a gradual compositional variation from


one surface to the other (e.g.concentration and/or porosity). These changes result in

property gradients (e.g. strength, mechanical toughness or thermal expansion coefficients),

which can be adjusted by controlling the composition and/or the microstructure of the


Metal/ceramic joining

The metal/ceramic joint is present in two different situations. In one instance, it is

an element of the design of the component. This enables the exploitation of the properties

of both metals and ceramics. One example, mentioned earlier, is the case of a light bulb.

The metal base is electrically conductive and permits current flow to the filament with a

minimum of resistance. The ceramic component, a glass envelope, is transparent to the

visible light emitted by the filament. Joined together to form a vacuum seal, they provide a

package that would be difficult to imitate using only a metal or a ceramic. The other

instance of metal/ceramic joining occurs when it is necessary to join two ceramic pieces

using a metal interlayer. A common example of this is in high-temperature structural

applications. Under these circumstances, the high-temperature strength and refractory

properties of ceramics are desired. High melting point metals such as titanium are used as

interlayers when brazing ceramics for high-temperature applications. The brazing process

is used because it requires a minimum amount of metal. Increased metal content may

detract from the desired behavior of the ceramics.

Alumina is the most common ceramic material to be studied in metal/ceramic

joining. Perhaps this can be attributed to its popularity as a structural material, as well as

the wealth of information available on its thermodynamic and mechanical properties.

Routine polishing of the surface is the only treatment the ceramic piece receives prior to

joining. Conversely, the metals are carefully prepared from high purity materials and

meticulously cleaned and polished before these are used. When the joint is to be formed

via diffusion bonding, several different metals of varying thickness are used78,79. Usually


the added metallic layers are very thin and have a low melting temperature and elastic

modulus compared to either the ceramic or the primary metal. This correlates with several

thermal stress reduction methods in metal/ceramic joining discussed by Suga et also. Ishida

et al.81 relieved thermal stresses via the introduction of structural defects. This was

accomplished by orienting the lattice when joining single crystals of niobium and alumina.

Serier and Treheux82 increased the strength of a silver/alumina joint by work hardening the

silver. They offer several hypotheses for the increase in strength, including the inducement

of structural defects. Frequently, the methods described above are not practical for service

operations, and the formation of a liquid phase at the joining temperature is necessary to

form a good joint.


In 1979, Hauth83 noted that several emerging technologies have high-temperature

use specifications for materials that will require ceramic/ceramic joining without the aid of a

metal interlayer. The example referred to was a high purity alumina pumpout port for a

thermonuclear fusion containment system. The pumpout port is a complex shape that is

comprised of several regular shapes (cylinders and plates) fitted together. The joining

method consisted of making a slurry of the unsintered material and buttering (as with toast)

the surfaces to be joined. This is the same procedure for attaching cup handles and other

extremities in the clay based ceramic industry. In the years since, numerous advances have

been made in ceramic/ceramic joining but there are still many hurdles to be overcome.

In the direct joining of similar ceramic materials, some chemical or physical

variation must be present to assist the driving forces that make up the joining process.

Scott and Brewer84 diffusion bonded alumina at 1500'C and 69 MPa. The joined pieces

were then heated to 1875'C to promote grain growth across the joint. They reported a

maximum joining strength of 59% of the starting material. This was achieved with an axial

deformation of 18% during joining. In similar experiments85, small amounts of MgO were

added as a sintering aid to the alumina cylinders to be joined. The joining temperatures

ranged between 1300"C and 1500'C with a pressure of 69 MPa. Deformation increased to

over 20% and no mechanical testing results were reported. Zdaniewski and Kirchner86

inserted thin polystyrene sheets between 96% alumina plates. The decomposition of the

polystyrene during heating created reducing conditions that enhanced interfacial bonding

and diffusion. In specimens joined at 1350'C and 17.5 MPa, the authors reported flexural

strengths of 194+94 MPa. Direct joining has also been investigated for SiC and Si3N4

with modest success87.88.

When the need for an indistinguishable joint is not imperative, there are more

straightforward ways of joining ceramics for high temperature use. Zimmer89 describes

the development of several ceramic cements that can withstand high temperatures

(>1450'C). Glass or glass-ceramics are also popular choices as interlayer materials.

Whereas some authorso9,91 chose glasses for their mechanical properties and crystallization

abilities, other researchers92.93 chose an interlayer glass composition that approximated the

composition of the grain boundary phase of the crystalline end members. When the glass

was chosen for its crystallization ability, joints of modest flexural strength and fracture

toughness were produced. Both groups of investigators who joined Si3N4 with a glass

similar to the grain boundary phase reported room temperature strengths of 450 MPa.

Microwave Joinine

Microwave joining of materials appears to be one of the most promising among the

areas of materials processing that have been included in microwave processing research.

Researchers have successfully joined alumina, silicon carbide, silicon nitride and mullite94-

100. With the exception of a few studies, most of the work has been performed using

single mode microwave applicators. The first report of joining ceramics using microwaves

was by Meek and Blake3. They successfully joined two alumina plates with a glass

interlayer in a home model microwave oven. Palaith et al.94 provide an overview of the

fundamentals of microwave heating and how it applies to joining. Using a tunable, single

mode microwave oven modified to allow for an application of pressure, Fukishima et al.95

joined several different purities of alumina and silicon nitride. Specimens consisted of

cylindrical rods measuring three millimeters in diameter. The alumina purities varied from

92% to 99%. Joining experiments were conducted at temperatures ranging from 1400'C to

1850'C and pressures up to 2.4 MPa. The authors were able to produce direct joints using

92% and 96% alumina that approached the original strength of the material. Figure 2-20 is

the effect of joining temperature on the bending strength of the 92% alumina. The 99%

alumina was successfully joined indirectly using both the 92% and the 96% alumina as

interlayers. Bending strengths of 90% of the original material were reported. No

information was provided regarding creep deformation. However, even at the low joining

pressures used, the temperature was high enough for creep to occur in high-purity

alumina. The same authors joined silicon nitride in a flowing nitrogen atmosphere using

silicon nitride with a greater dielectric loss than the end member material. The effect of

joining time on the bending strength of the joined silicon nitride is shown in figure 2-21.

Al-Assafi and Clark96 successfully joined 94% alumina to 99.5% alumina using a sol-gel-

derived alumina as the interlayer. The researchers used a multimode cavity and microwave

hybrid heating. Figure 2-22 is the Knoop hardness across the joint area for joints made

with and without the gel interlayer. The purpose of the gel interlayer is to provide a high-

purity powder with a large surface area to join the two surfaces. In an effort to evaluate

the joint non-destructively during processing, Palaith and Silberglitt97 introduced an

apparatus that provides an acoustic pulse echo axially through the sample during joining.

This provides information as to the extent of the joint formation. Figure 2-23 is the result


Before joining

S S *
I /1

I /

Joining time: 3 min
Pressure: 0.6 MPa

1400 1600 1800 2000
Joining Temperature (C)
Figure 2-20. Effect of joining temperature on the strength of 92% alumina95.


Base: SN220
Adhesive: SN501

a. 500 Base

C Adhesive


m 300

(1720C; 6.2 MPa; 0.6 mm)

100 1
0 5 10
Joining Time (min)
Figure 2-21. The effect of joining time on bend strength of silicon nitride95.


1800 99.5% Alumina 94% A


1400 -



800 mr

600 -

400 -

200 -

-300 -150 0 150

Distance from Joint (lim)

Figure 2-22. Knoop hardness across joint for microwave joined alumina96.


1400 700

1200) -600

1_000 500

800 400 "

0- 600 -300

200 100

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Time (min)

Figure 2-23. Acoustic emission traces for a mullite joining experiment97.


of an experiment involving the joining of mullite. Yu et al.98 employed this method in the

simultaneous sintering and joining of alumina powders. At a temperature of 1400'C and a

pressure of 0.238 MPa, they achieved a density of 99% theoretical.

Silicon carbide is an excellent microwave absorber and a ceramic material that has

many potential uses. Therefore, it has attracted the bulk of attention for microwave joining.

The processing of the silicon carbide, whether it is reaction bonded silicon carbide or hot

pressed silicon carbide, can significantly affect its ability to absorb microwaves. Reaction

bonded silicon carbide contains residual silicon metal throughout the microstructure.

Ahmad et al.99 direct joined hot pressed silicon carbide to hot pressed silicon carbide and to

reaction bonded silicon carbide. They accomplished this using both a single mode

microwave and a multimode microwave oven. Yin et al.loo joined the high dielectric loss

siliconized silicon carbide using a thin aluminum foil as an interlayer. The authors achieved

fracture strengths comparable to the original material. Other attempts to join silicon carbide

include the use of combustion synthesis. It has been demonstrated that microwaves are an

excellent source of energy for initiating and controlling the combustion synthesis

reaction0io. Silberglitt et al. 102 made successful SiC/SiC joints by initiating a combustion

reaction between titanium, silicon and carbon at the interface.

In an overview paper discussing the problems and opportunities in microwave

joining, Loehman 103 states that there are applications that hold promise but much work

needs to be done to prove the value of microwave joining. He discusses potential

advantages that could be realized by applying microwave processing to the joining of

ceramics. The author also identifies problems that may be encountered and offers solutions

through rigid process control.


Several authors 104,105 have discussed the potential for energy savings through the

use of microwave processing. It still is not clear if there is a significant energy

conservation to be realized. Katz 106, regardless of his involvement as a researcher in the

field of microwaves, displayed a skepticism towards many of the positive results being

published regarding microwave processing. Despite a wary community of researchers,

exploration of microwave processing has continued with success in specific areas. One

promising niche involves the recovery of common and precious metals from electronic

circuit boards107. Microwaves also have proved successful in several industrial

applicationso0. They are currently being used in some areas of food processing, rubber

vulcanization and wood curing.

Research in the microwave joining of ceramics is developing into a field that will

have viable approaches to engineering problems. Microwave joining has proved feasible

for several oxide and non-oxide ceramics. These materials can be joined with strengths at

or above those attained on conventional joining processes. Some of the advantages that

have been demonstrated by researchers include the ability to join materials with unfinished

surfaces (large surface roughness) as well as non-destructive evaluation of the joint during

the joining process. The in-situ joint evaluation can be performed because it is possible for

the end members to extend out of the heating zone. This provides access to the end

members for the acoustical equipment. In the brief period of time that microwave joining

has been investigated, many discoveries have been made that have led to further research.

Much of the work reported has been presented in one of two formats. In the first method,

the strength of a joint formed under specific conditions is presented with reference to the

strength of the original material. The second method is somewhat more informative in that

the results are compared to conventional joining of similar materials under comparable


conditions. The consequences of changing the processing parameters have not yet been

investigated. Results need to be evaluated in a more scientific manner than plotting strength

versus the processing parameter (temperature, pressure, etc.). These parameters need to be

evaluated to determine their relative importance in microwave joining. The intent of this

research was to provide reproducible, microwave joining data based upon

an experimental design coupled with numerical modeling of the experiment.

The data generated will contribute to the advancement of microwave joining

of materials by assisting in clarifying the effects of some of the processing

parameters on the joined materials.



There are several objectives encompassed by the experimental procedure. The first

objective was to choose a system of materials, consisting of end members and an

interlayer, that was of technical interest and could be joined successfully using microwave

energy. The second objective was to develop a device and procedure that could fabricate

joined specimens with reproducible properties. The first two objectives were incorporated

into the third objective to manufacture joined specimens that were tested for flexure

strength. These results then were analyzed to evaluate the relative importance of the

processing parameters on the strength of the joined specimens. A fourth objective of this

chapter was to monitor the temperature of the specimen and the susceptor during joining so

comparisons could be made to the numerically derived temperature profiles discussed in the

next chapter. The experimental work performed in this study can be separated into four


The first phase of the work was the preliminary testing of potential interlayer

materials. Various properties of several materials were tested to determine which materials

possess the characteristics required to produce a strong joint. The material properties of

greatest interest were the mechanical strength, refractoriness and microwave absorbing

capabilities. The results reported by Al-Assafi75 indicate that an alumina based sol-gel was

an excellent candidate to provide the desired characteristics.

The second phase consisted of designing and implementing a microwave joining

device to control and record the processing parameters. A microwave susceptor was


developed to heat the end members to a temperature at which they become more efficient

microwave absorbers. Using the materials that were classified as tenable interlayers,

experiments to join two viable end member materials were initiated. Flexure bars were

machined from the joined specimens and tested using four-point bending to evaluate the

interlayer and end member materials based on strength.

Phase three was the development of an experimental design that would use the

chosen materials system to evaluate the significance of each of the processing parameters

on the flexure strength of the joined material. Two values were selected for each

processing parameter and an analysis of variances was performed on the results of the

flexure tests.

In phase four, a heat transfer program was used to simulate microwave hybrid

heating. The results were compared to measured values collected in experiments run to

measure the temperature of the specimen and of the susceptor. Two methods of

temperature measurement were used to test the validity of the measuring process.

Phase One
The materials that were identified as potential interlayers were a series of alumina

based gels, 96% pure alumina and nickel oxide. The alumina gel used by Al-Assafi was

investigated as an interlayer material. It also served as a base gel that was combined with

other materials. The other materials were introduced to the base gel either by substituting

for the aluminum salt in the gel during formation or by admixing them to the stable sol

before gelation. All of the materials added to the base gel were considered to be strong

microwave absorbers. The intent of this approach was to increase the absorption of

microwave energy in the joint area. All of the joining experiments in this study were

performed in a constantly changing microwave field. Therefore, unlike the single mode

applicator used by Fukishima et al.95, there were no fixed positions where a high electric


field was present. Due to these processing conditions, there was no technique available to

determine the position of a stable hot spot as described by Booty and Kriegsmann39.

Five materials (chromium oxide, iron (II) oxide, iron (III) oxide, nickel oxide and

silicon carbide) were used in conjunction with the base gel to make ceramic powders for

subsequent evaluation. These materials were selected from a myriad of strong microwave

absorbers because of the way they combine with alumina. Chromium oxide is mutually

soluble in alumina for an entire range of compositions. Iron (III) oxide also forms a solid

solution with alumina, but each material is capable of dissolving around 10% of the other.

Iron (I) oxide and nickel oxide both form a spinel when combined with alumina. Silicon

carbide does not form any intermediate compounds with alumina.

Thin monolithic disks of 96% pure alumina and of nickel oxide also were examined

as potential interlayer materials. Alumina was considered because it is very similar to the

materials being used as end members: 99.5% and 96% alumina. Using a similar material

as the interlayer reduces the risk of stresses from differential expansion and maintains a

high strength material throughout the joint region. However, the similarity in composition

presents a challenge to joining because it reduces the presence of a compositional gradient

to act as a driving force. Nickel oxide was chosen over iron (II) oxide as the second solid

material because it is less sensitive than iron (II) oxide to oxygen partial pressures.


Base gel

The alumina based sol-gel interlayer was produced in accordance with a patent by

Clark et al. 109 for the creation of boehmite sols. The materials used to produce the base

gel are water*, aluminum sec-butoxide*, and aluminum nitrate nonahydrate***. The

* Deionized water resistance > 10 Mohms
* Alpha Chemicals # 11140 Al(OC4H9)3 F.W. 246.33
** Fisher Chemicals # A586-500 AI(NO3)3-9H20 F.W. 375.14


purpose of the aluminum sec-butoxide is to provide the solution with metal ions. The

water hydrolyzes the aluminum sec-butoxide. Aluminum nitrate nonahydrate is a peptizing

agent that breaks up the precursor, permitting the hydrolysis to occur. This facilitates the

formation of colloidal aluminum hydroxide particles. Throughout the remainder of this

thesis, aluminum sec-butoxide will be referred to as ASB, and aluminum nitrate

nonahydrate will be referred to as aluminum nitrate. Working within the parameters

established by the patent, Dalzelllo reported a ratio of ASB:aluminum nitrate of 25:1 as

providing a stable sol. The ratio of 25:1 ASB:peptizer was used throughout the course of

this study. The ratios of the three components for the initial sol were 100:1:0.04

corresponding to water:ASB:aluminum nitrate.

Under reflux conditions, 675 milliliters of deionized water was heated to 880C in a

1000 milliliter Pyrex flask that had been washed in a 6M nitric acid/0.2M hydrofluoric acid

solution and rinsed in deionized water. During heating the water was stirred continuously

by a Teflon coated magnet. Meanwhile, six grams of aluminum nitrate was dissolved in 45

milliliters of water. Rapidly, 104 milliliters of ASB was measured to minimize hydrolysis

of the ASB from the air. The ASB then was added to the water using two syringes*. The

aluminum nitrate solution then was immediately added to the flask. The result was a

cloudy, milky-white solution. This was stirred continuously until it became translucent.

The clearing process took several hours. After the sol cleared, it was cooled to room

temperature and stored in a polypropylene container until used.

Doped gel

According to the patent by Clark et al.109. several other metal nitrate salts may be

substituted for aluminum nitrate during the sol-gel process. Both chromium nitrate** and

*Fisher Chemicals # 14-823-2D 60 ml syringes
**Fisher Chemicals # C331-500 Cr(N03)3-9H20 F.W. 400.15

iron (1II) nitrate* can replace aluminum nitrate in the sol. Chromium oxide forms a solid

solution with alumina at every composition. Figure 3-1 is the binary A1203-Cr203 phase

diagram at elevated temperatures. The A1203-Fe203 phase diagram, figure 3-2, is more

complex. Below 1300C, there is a mixture of the Fe203 (hematite) phase solid solution

with A1203 and A1203 corundumm) phase solid solution with Fe203. Near 1400C,the

presence of FeO becomes significant as the spinel, FeAI204, is formed. McGill et al. mI

reported that both chromium oxide and iron (II) oxide are excellent microwave absorbers

in the temperature range investigated at the frequency of 2.45 GHz. The other additives,

silicon carbide**, nickel oxide*** and iron (II) oxide****, were admixed as powders into an

alumina sol and reduced to a solid. The yield of the sol was determined in two ways. In

the direct method, 200 milliliters of sol was dried and then heated to 12000C. This

temperature was above the crystallization temperature of the stable a-alumina phase. X-ray

diffraction analysis confirmed that the alpha phase was the only phase present in detectable

amounts. The resulting powder was weighed and had a mass of 5.08 grams. The yield of

the sols was also determined mathematically. Using the ratios of materials stated earlier,

calculations were made based on using 100 moles of water, one mol of ASB and 0.04

moles of aluminum nitrate. This would provide 1.04 moles of aluminum available for

oxide formation. This, in turn, would produce 0.52 moles of alumina which represents a

mass of 53 grams of alumina. By calculating the volume of the resulting sol, a yield of

0.026 grams of alumina per milliliter of sol was determined. For a 200 milliliter

*Fisher Chemicals # 1110-500 Fe(N03)3-9H20 F.W.
**Norton Company SiC E-85 Crystolon 220 grit F.W. 40.1
***Fisher Chemicals # N69-500 NiO F.W. 74.71
****Fisher Chemicals # 1119-500 FeO F.W. 231.54



s 2100
h 2045W5
2000 -

0 20 40 60 80 100
A1203 mol% Cr203 Cr203

Figure 3-1. Binary Cr203-A1203 phase diagram 12.
Spinel ss Spinel s s
Hemotite ss Corundum s s
SpineSpine s

Fe20/ AI203
1400 -
0 ( Fe203-AI,20
Corundum s s
a) ff/ Fe2O03A1203

E Hematite ss + Fe20zA1203z ss

Hematite s s + Corundum s s

Fe203 20 Fez03-A2g03 60 80 AI203
wt.% A1203
Figure 3-2. Binary Fe203-A1203 phase diagram']3.

sol, this renders 5.12 grams of alumina -- a difference of less than one percent from the

experimentally determined value.

The powders were added to sols in a 1:1 molar ratio with the expected yield of

alumina. The mixture then was ultrasonicated for ten minutes to break up agglomerates.

The sol/powder mixture then was gently heated on a hot plate while stirring until much of

the liquid was removed. The resulting gel was heated overnight at 150 oC to remove most

of the remaining liquid. The end result was a solid mass of opaque material. This was

crushed in a mortar and pestle and sieved through a 325 mesh (44gm) screen. Using a 1.2

centimeter diameter steel die, 1.5 grams of each powder were pressed into pellets using a

13.3 kN force.

Sol-Gel Characterization

The gels produced were characterized using several diverse methods. The different

compositions of pellets were heated using microwave energy at 2.45 GHz to evaluate each

as a microwave absorber. Differential scanning calorimetry (DSC) was performed on the

various compositions with x-ray diffraction analysis conducted on the resulting material.

Dielectric properties of the different compositions were measured at 2.45 GHz as a function

of temperature.

Stand alone heating

A home-style microwave oven* was fitted with an on/off temperature controller.

An inconel-shielded, Type-K thermocouple was placed in contact with the specimen for

temperature measurement within the microwave. Refractory brick** was sculpted to

accommodate a specimen and a thermocouple feedthrough. The brick provides insulation

to reduce heat loss from the specimen during heating. The setup in Figure 3-3 has a

*General Electric 750 Watt Dual Wave II Microwave System
**Thermal Ceramics K-3000 3000F refractory brick


Figure 3-3. Setup used to measure temperature of gel compositions heated solely by
microwaves (setup was placed in a microwave for heating).


maximum test temperature of 1175 C. Specimens were heated at full power (800 W)

using the setup in Figure 3-3. Temperature readings were taken periodically. Initial results

may have been influenced by arcing between the thermocouple and the specimen. Arcing is

the release of a charge build up between two points which can cause localized heating and

even melting of the thermocouple and the specimen. Another problem associated with

arcing is inaccurate, widely varied temperature readings. One way in which arcing was

reduced was by increasing the surface area of the tip of the thermocouple. In this case, a

thin nickel disk was placed between the thermocouple tip and the specimen. This

significantly reduced the amount of arcing encountered during the study.

Differential scanning calorimetry

Differential scanning calorimetry was performed on a Stanton Redcroft DSC

1500S. The gels were heated from room temperature to 14000C at ten degrees per minute

in flowing air. A second experiment was conducted on the sample containing FeO. This

time, the maximum temperature was 12500C in order to avoid any reactions that occur

above 12500C. A piece of sapphire, similar in mass to the sample powder, was used as a

reference material to maintain a level baseline.

X-ray diffraction analysis

The ceramic powders that resulted from heating in the DSC were subsequently

analyzed using a Scintag X-Ray Diffraction Analyzer. The x-rays were copper k-alpha at

45 kilovolts and 40 milliamps at a wavelength of 1.54060 angstroms. The powders were

scanned at one degree 20 per minute from twenty to sixty degrees 20. This 20 range was

sufficient to identify any of the compounds formed during heating that involved any of the

materials used.

Dielectric properties

The relative dielectric constant and loss factor, e', and e"eff, were measured for

several compositions of the dried powder at 2.45 GHz as a function of temperature.

Before analysis, the powders were heated to 4500C to remove all of the bound and

structural water. This was done to provide the instrument with a relatively constant mass

of material. The dielectric properties were measured from room temperature up to 14000C.

These measurements were performed at Chalk River Laboratories/Atomic Energy of

Canada, Ltd. using the cavity perturbation method*.


The two purities of alumina used as an interlayer material were Coors AD94, 94%

alumina and Coors AD96, 96% alumina. The major impurities in all of the Coors alumina

products are silica, magnesia and calcia. These impurities act as fluxes for densification

and as grain-growth inhibitors. The manufacturer reports a room temperature flexure

strength for the AD94 of 350 MPa. The grain size ranges from 2-25 microns with an

average grain size of 12 microns. Other relevant properties of the alumina are in Table 3-1.

The interlayers were produced by cutting disks approximately 2 millimeters thick from a

cylindrical rod 2.54 0.0025 centimeters in diameter and 20.32 0.0025 centimeters long.

The disks were mounted on a glass plate and the surfaces were machined using a high

speed diamond grinding wheel*. The machining procedure used a downfeed of five

microns per pass and a crossfeed of 2.5 millimeters. The disks had a final thickness of 1.5

0.15 millimeters.

*R. Hutcheon AECL Chalk River Laboratories, Chalk River, Ontario, CA
**Norton ASD320-R75B99-1/4 320 grit wheel

Table 3-1. Manufacturer Reported Properties of Coors Alumina Used in Study.

+Coors Alumina and Beryllia Properties Handbook, Bulletin 952, Coors Ceramics, (1969)
Nickel Oxide

High-purity nickel oxide disks, hot pressed to >80% dense, were obtained from

Super Conductor Materials, Inc. The as-received disks were 2.5 centimeters in diameter

and 1.2 centimeters thick. The surface of the disks were impregnated with a graphite

Property Units AD-94+ AD-96+ AD-995+

Specific Gravity 3.62 3.72 3.84
Crystal range crowns 2-25 2-20 10-50
Size average 12 11 20
Flexural 250C 350 359 310
Strength 10000C 140 172 230
Modulus of Elasticity GPa 282 303 358
Shear Modulus GPa 117 124 152
Bulk Modulus GPa 166 173 207
Poisson's Ratio 0.21 0.21 0.21
Maximum Use (no load)
TC 1700 1700 1,750
Thermal 25C 18.0 18.0 31.4
Conductivity 100C 14.6 14.6 27.2
4000C 7.1 7.1 11.7
8000C 4.2 4.2 7.1
Specific Heat 100C J/K(kg) 880 880 880

250C 8.9 8.9 9.4
Dielectric Constant 500C ratio 10.4
1 GHz 800C -- 11.0

250C 0.0008 0.0001 0.0001
Dielectric Loss 500C ratio 0.0002
Factor 1 GHz 8000C 0.0003

250C 0.007 0.001 0.001
Loss Tangent 5000C g"/g 0.002
(tan 8) 1 GHz 800C 0.003

residue from the manufacturing process. A low speed watering saw was used to section

the disks into thinner pieces. As with the alumina interlayer specimens, the nickel oxide

disks were mounted on a glass plate and machined to a thickness of 0.33 0.02

millimeters. All of the graphite was removed during the grinding step.

Surface roughness

The machined surfaces were characterized using a laser profilometer*. A 3-

dimensional scan was performed on a five millimeter by five millimeter square section of

the surface. A total of 250 line scans were performed on each material. The table speed

was 10 mm/minute. The average surface roughness was calculated from a single scan

using the same distance and table speed as the 3-dimensional scans. The roughness

reported is the deviation from the arithmetic mean and was calculated automatically from the

profiles using the equation

R, = l(x)dx. (3-1)

Figure 3-4 illustrates the arithmetic mean in a simulated profile and identifies the

components of equation 3-1 that are used to compute the surface roughness.

mean line

Ay m / a i R-

Figure 3-4. Simulated material profile depicting the arithmetic mean.

*Rodenstock RM-600 Laser Profilometer

Phase Two

Currently, there are no commercially available systems to perform the microwave

joining of ceramics. Therefore, an apparatus had to be designed and constructed to

accomplish this task. The modification of a microwave oven to accommodate the

application of pressure has many challenges associated with it. Turntables or fans used to

stir the microwave field may interfere with the application of pressure and must be removed

or repositioned. The external holes needed to apply a load on a specimen in the microwave

cavity are a breach of the microwave containment system and present a potential health

problem (see equation 2-5). Extreme care must be taken in the design and fabrication of the

joining apparatus to address these challenges and prevent them from hindering the joining


There are several reasons why alumina was chosen as the material of interest for

this study. Alumina is one of the most widely used technical ceramics. It is prominent as a

substrate for electronic materials, as a refractory material, and as a structural ceramic.

Although alumina is not always the ideal material for the intended use, it is often chosen

because of its low cost. The thermal or mechanical properties of alumina are also

considered acceptable for many applications when the ideal material is either unavailable or

too costly. Alumina can be obtained in a multitude of purities, grades and sizes designed to

meet most of the needs of the technical community. Another predominant reason for

selecting alumina for this study is that it is a common material for testing and analysis.

There is an abundance of documentation available detailing its properties and performance

under widespread conditions. This includes the thermal, electrical, optical and mechanical

behavior upon exposure to elevated temperatures.

The end members considered for the joining experiments, 99.5% and 96% pure

alumina, are both poor microwave absorbers compared with the materials being surveyed


as potential interlayers. Without assistance, the end members are unable to heat to the

temperature required for joining using the microwave oven designated for this study.

Microwave hybrid heating, discussed in Chapter 2, is a logical approach to provide the

thermal energy for joining these end members. A microwave susceptor was used to

generate the radiant heat necessary to raise the temperature of the end members being joined

to a temperature where they could then more efficiently absorb microwaves. The susceptor

can be made from a number of materials, including those being examined as potential

interlayers. Design parameters that should be considered include the shape of the

workpiece being heated, the air gap desired between the workpiece and the susceptor, the

method in which the temperature of the workpiece will be measured, and most importantly,

the ability of the workpiece to absorb microwaves. This last factor will determine the

amount of microwave absorber needed in the susceptor to attain the desired temperature.

There are at least two methods of making susceptors. The first consists of fixing a

microwave absorbing material to a piece of low density insulation that has been shaped into

the desired configuration. This was simple to construct and works well for low

temperature applications. The second method involves casting a susceptor using a

refractory cement with granules of microwave absorbing material dispersed in it. This

method works very well at high temperatures. These susceptors tend to be more massive

than the first type and can shield very small specimens from the microwave field.

Microwave Joining Apparatus

The joining apparatus consisted of several components, each with a specific

function. To physically support the remaining components, a load frame, capable of

sustaining the maximum applied load, was machined from an aluminum alloy. A home

model microwave oven* was modified to facilitate the application of pressure and the

*Goldstar MA-1172M 1000 Watt microwave oven

monitoring and control of temperature. An air cylinder* capable of supplying a load of 2.2

kN was mounted on the top of the load frame. A load cell**, mounted in line with the air

cylinder, measured the load being applied. The feedback provided by the load cell in the

form of a voltage, was used to control the load applied by the air cylinder. An 18

millimeter diameter high-purity alumina rod was used to transfer the load into the

microwave cavity. A 20 millimeter feedthrough was machined in the top of the microwave

to provide the ram access to the cavity. The diameter of the hole was less than one-sixth

the wavelength of the microwaves used in this study. In Chapter 2, the critical hole

diameter for significant transmittal of microwaves was found to be one-half the wavelength

of the microwave. However, in this setup, there was a dielectric material (alumina push

rod, 'r = 9) partially filling the aperture in the wall of the microwave cavity. In this setup,

the dielectric material acts as an antenna for microwave propagation and the ability to

contain the microwaves was brought into question. To ensure a minimum transmittal of

microwave energy via the feedthrough, a 7.5 centimeter long copper tube of the same

diameter as the aperture was inserted into the feedthrough. The placement of an electrical

conductor tube prevents microwave leakage. Metallic tape was used to cover any other

gaps where microwaves might leak. The microwave's turntable was discarded and the

turntable motor was relocated to the side of the microwave opposite the wave guide. An

aluminum mode stirrer was attached to the motor to improve the uniformity of the field

within the microwave. Most of the operations performed during the joining process were

controlled by a computer program*** running on a Macintosh IIC. This included storing

the values of applied pressure and using this information to adjust the air cylinder. The

*Bellofram model # 900-009 500# air cylinder
**Revere Transducers model # 63HC-D3-500-10pl 500# load cell
**National Instruments Labview Version 2.0


computer was also used to store the temperature readings from two Type-R

thermocouples*. The two thermocouples were fed through an eight millimeter hole in the

back wall of the microwave cavity. A setpoint controller** was used to regulate the

maximum temperature during the process. The temperature controller was spliced into the

magnetron power supply so that the temperature could be regulated by turning the

magnetron on and off as necessary. Figure 3-5 is a schematic and figure 3-6 is a

photograph of the microwave joining setup.

Microwave Susceptors

In preliminary microwave hybrid heating experiments, D6114 heated alumina in a

microwave furnace using a commercially available susceptor***. These microwave

furnaces were simply a silicon carbide coating on a low density alumino-silicate refractory.

At temperatures near 15000C, both the silicon carbide coating and the refractory lining

experienced significant degradation. This led to a reduction in the furnace's ability to

absorb microwaves. Al-Assafi75 produced susceptors using a refractory with a lower

dielectric loss factor than the commercially available furnace. This enabled the maintenance

of temperatures of 15000C for one hour without destruction of the susceptor. These

susceptors proved sufficient for one experiment, but assembling a susceptor with a uniform

coating that maintained its microwave absorbing ability throughout a course of high-

temperature experiments was not feasible. It became necessary to seek an alternative

approach to microwave hybrid heating to provide a high-temperature, extended-use

susceptor. For this research, a mold of the desired susceptor was formed out of concentric

*ARi Inc. model # HIT-99N-4DR9AA-36 Type-R thermocouple
**Omega model # CN9000A Miniature Autotune Temperature Controller
**National Superconductor Microwave Glass Melter

air cylinder

HKtl I

temperature controller
Figure 3-5. Schematic of the microwave joining apparatus.


Figure 3-6. Photograph of the microwave joining apparatus used in all of the microwave
joining experiments.

polypropylene cylinders. High alumina content cement* was mixed with silicon carbide

granules and enough water to make a slurry. This was cast into the mold and allowed to

set. The process for forming the susceptor is outlined in figure 3-7. The polypropylene

rings were removed and the cement/SiC cylinder was cured by slowly heating it in a

conventional furnace to 1000 'C for one hour. After cooling, the susceptor was ready to


The amount of cement and silicon carbide was varied to determine the amount of

silicon carbide that would absorb the appropriate amount of microwaves to provide the

desired thermal environment for the joining process. This was done systematically by

mixing the cement and silicon carbide in different ratios. Silicon carbide granules were

added in amounts of 5, 10, 20, and 30 volume percent of the cement. The volumes were

determined using the density of alumina to approximate the density of the cement (96%

alumina) and the density of silicon carbide. Enough water was added to make the mixture

castable. The cylinders were all approximately 300 grams. All of the cylinders were cured

as described earlier.

To determine the ability of the cylinders to heat, they were insulated as shown in

figure 3-8, placed in a microwave oven and heated using 3200 watts of power. The

temperature was measured in both the center of the suscepting cylinder and outside of the

refractory wall using Type-K thermocouples shielded from the microwaves by a inconel

sheath. The dielectric properties of the cylinders as a function of temperature were

measured at AECL using the cavity perturbation method.

The susceptor used for the joining process was similar to those used in the

susceptor evaluation experiments. A 36 volume percent loading of silicon carbide was

used in the susceptor. Two 4 millimeter in diameter hollow alumina tubes were fixed in

*Alcoa Al-66 97% A1203 3% CaO

Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EQ0DP8YAS_4TPZUT INGEST_TIME 2013-01-23T13:56:17Z PACKAGE AA00012928_00001