173

MICROWAVE PROCESSING OF CERAMICS:

Guidelines Used at the Oak Ridge National Laboratory

MARK A. JANNEY, HAL D. KIMREY, AND JAMES 0. KIGGANS

Oak Ridge National Laboratory*
P.O. Box 2008
Oak Ridge, Tennessee 37831-6087

ABSTRACT
To make meaningful comparisons between conventional and microwave processing
of materials, one must conduct experiments that are as similar as possible in the two
environments. Particular attention must be given to thermal conditions, sample
parameters, and furnace environment. Under thermal conditions, one must consider
temperature measurement (pyrometer or thermocouple, sheath type, and arcing of thermo-
couples), thermal history (heating and cooling rates, thermal gradients), and exothermic
reactions. Regarding sample parameters, one must. consider sample size, and packing
powders and insulation systems. With respect to furnaces, one must consider differences
in atmosphere, impurities, and uniformity of heating. Examples will be drawn from
diffusion, grain growth, sintering, nitridation, and drying experiments conducted at the
Oak Ridge National Laboratory (ORNL) over the past six years.

INTRODUCTION

There is a critical need to compare the results obtained from microwave processing
with those from conventional processing when conducting an investigation of the effects
of microwave heating on materials. Side-by-side comparisons are required for at least
three reasons. First, literature data from similar investigations are not sufficient. For
example, in the study of diffusion, the scatterband of data from various investigators using
different materials is typically an order of magnitude or greater. Second, there are lot-to-
lot variations in the composition and processing of materials that greatly affect their
properties. If the same material is not used in both the conventional and the microwave
studies, meaningful comparisons might not be made. Finally, the claims in the microwave
processing literature such as lowered sintering temperatures [1,2], accelerated kinetics in
ceramic, polymeric, and organic systems [3-6], and reduced activation energies [1,3-5] are
suspect to many researchers both inside and outside the microwave processing
community.

STANDARD ASSUMPTIONS
The experimental procedures described in this paper are built upon a few standard
assumptions that are common to all of the experiments cited herein. Multimode cavities
were used in all of the work reported here. Thus, there are certain details that might not be
strictly applicable to work in single-mode cavities. Also, most of the experiments reported
here were conducted at 2.45 GHz, the most common microwave processing frequency.
However, the general principles involved are the same irrespective of cavity mode
structure or frequency. It is further assumed that processing parameters such as heating
rate, soak temperature, processing atmosphere and total pressure, and calibration of
instruments such as pyrometers and thermocouples are the same for both conventional and
microwave processing.

*Operated by Martin Marietta Energy Systems, Inc., under contract DE-AC05-

840R21400 for the U.S. Department of Energy.
Mat. Res. Soc. Symp. Proc. Vol. 269. 01992 Materials Research Society
174

THERMAL EFFECTS

The ability to heat a material

There are major differences between microwave and conventional thermal

processing. To heat a part conventionally only requires putting the part into a furnace and
heating the furnace to the required temperature. If enough time is taken to permit heat to
flow into the part, it must, by necessity, heat to the required temperature. Such is not the
case when heating a material in a microwave furnace. Whereas heat is supplied externally
to the part being processed conventionally, it is generated within the part being processed
by microwaves. Depending on the dielectric properties of the part, several different
phenomena may occur: (1) it may heat quickly; (2) it may not heat at all; (3) it may heat
uniformly; or (4) it may generate hot spots. How the part heats depends critically on the
dielectric loss properties of the part, how those properties change with temperature, and on
the microwave field distribution in the furnace. It is important to remember that microwave
furnaces are not like conventional furnaces.
In addition to chemical and thermal considerations, dielectric properties must be
considered in microwave heating. This includes the material being processed, the
insulation surrounding the part, and auxiliary materials such as thermocouples. As an
example, consider the parts shown in Fig. 1. The ZrO2 - 12 mol% CeO 2 part shown in
Fig. la has a conical shape from top to bottom because the temperature was much higher
on the top. The temperature gradient led to severe cracking because of differential
sintering. The insulation that was used around this part was pelletized ZrO 2 - 12 mol%
CeO 2, which heated very nonuniformly. The sample shown in Fig. lb sintered very
uniformly because uniform heating was achieved. In this case, zirconia fiber insulation of
composition ZrO2 - 8 mol% Y20 3 was used, which heated much more uniformly than the
pelletized ZrO2 - 12 mol% CeO 2.

ZrCe 92006 ZrCe 92027

Sample Hotter on Top Uniform Heating
Differential Sintering Good Sample
Cracking ZrO2 (8% Y203) Bulk Fiber Insulation
Conical Shape
Pelletized TZ1 2CE Matched Insulation

Figure 1. ZrO2 - 12 mol % CeO2 was challenging to process at 2.45 GHz.

 175

Some materials cannot be processed at all. For example, carbon fiber preforms arc
continuously upon exposure to microwave energy. The arcing is a result of the high
electrical conductivity of the fibers combined with their small radii, which tend to
concentrate the electrical field around them.
The dielectric loss factor of materials being microwave processed must be within
certain ranges to be heated successfully. The microwave conductivity of bulk metals is
much too high to permit their heating by microwaves; they simply reflect the energy back
into the cavity. However, finely divided metals have been heated in many cases [7-9].
The increased surface area provided by powders accentuates the role of the "skin depth" of
the metal. The skin depth is that near-surface region (on the order of 1 to 100 gm) which
absorbs a small amount of the microwave energy incident on the metal surface owing to its
finite conductivity. In a metal powder, a significant fraction of the material is in the near-
surface region, and hence can play a role in microwave absorption. At the other end of the
spectrum are low loss insulators, which are difficult to heat because they absorb little of
the microwave energy incident upon them. Materials such as high purity alumina, high
purity magnesia, and high purity silica can be heated at 2.45 GHz only at high field
strengths (e.g., in a tuned cavity at high power density), at high frequencies (e.g., above
-20 GHz), or with auxiliary preheating. Table 1 lists several materials that have been
successfully processed in multimode cavities at 2.45 GHz at ORNL.
The use of indirect heating may be required (1) to preheat a material to a temperature
at which it will couple to 2.45 GHz microwaves or (2) to mitigate problems associated
with thermal runaway. Examples of indirect heating techniques include the "picket fence"
[2,14,15] coatings [16], and lossy second phases [8,17]. All of these have been used
successfully in materials processing.

Temperature measurement
The ability to measure temperature accurately is crucial to conducting meaningful
microwave and conventional experiments. In fact, the first question asked by observers
from outside the microwave community is often - "how do you know that you are
measuring a meaningful temperature in the microwave furnace?" Even after standard
temperature measurement procedures have been established in a laboratory, it is important
that the need for accurate temperature measurement remain in the forefront of the
researcher's concerns.

Table 1. Various ceramics have been processed at 2.45 GHz at ORNL.

Maximum Auxiliary
Material Temperature Heating Comments
Achieved ('C) Required
SiC 1600 No Parts as large as 2.5 kg have been heated to
1600*C at rates up to 2000'C/h.
LaCrO 3 1700 No Arcs to bare thermocouple [10].
NiO 900 No Could not be heated above 900'C due to
increase in reflectivity of NiO at high
temperature
Fe3 0 4 600 No Oxidation to Fe20 3 prevented further
heating.
B 4C 2500 No Heating must be done quickly [11].
TiB 2 2200 No Yttria insulation [ 12].
Silicon 1400 No Melting prevents further heating. [13].
A12 0 3 1600 Yes Conducting particles added. [8].
A12 0 3 1600 Yes Picket fence used [14].
ZrO2 1600 Yes Picket fence used [2].
A12 0 3-ZrO2 1500 Yes Picket fence used [ 15].
176

At ORNL, both thermocouples and radiation pyrometers are used extensively for
temperature measurement. Several guidelines can be formulated regarding the preferred
techniques at ORNL, based on our experience. Thermocouples are most often used for
direct measurement of sample temperature and control of the heating program for materials
processing. Thermocouples must be shielded from the microwave field to prevent arcing
of the fine thermocouple wires. Molybdenum, Inconel, stainless steel, and platinum
sheaths have been used successfully on Type C (tungsten-rhenium), Type K (chromel-
alumel), and Type S (platinum-rhodium) thermocouples. The sheath diameter is typically
1/8 inch (3mm). This diameter represents a reasonable compromise between response
time of the thermocouple (one prefers a small diameter sheath to reduce response time) and
the tendency of the sheath to arc in the furnace (arcing is reduced by using a large
diameter sheath). Figure 2 shows the variation in time of both the electric field and the
sample temperature (as indicated by a Type-C thermocouple embedded in an alumina part)
in a 28-GHz microwave furnace 3. As the microwave electric field is cycled off and on,
the thermocouple signal responds with decreasing, then increasing, temperature.
However, there are no discontinuities in the temperature measurement when the power is
turned off or on; this shows that the microwave field has no effect on the temperature
measurement as made by the thermocouple. Only normal heating and cooling effects are
observed, which are similar to those that would be observed in a conventional furnace if
the power were turned off and on. Additional support that thermocouple temperature
measurements are correct is given by the following example. For the highest microwave
annealing temperatures that were used in the alumina heating study, i.e., 1600"C, no
degradation, densification, or embrittlement of the alumina bulk fiber insulation used was
observed. When the bulk alumina fiber insulation is heat treated conventionally, it
degrades rapidly at temperatures above 1600 0 C. If the temperature measurement in the
microwave furnace had been significantly incorrect, say that 1500'C was recorded when
the temperature was actually 1700"C, then densification and embrittlement of the insulation
would have been observed. Since no such observations were made, the indicated
temperature is likely the correct temperature. As a further example, it has been shown
[10] that there is excellent agreement (-5'C) between the temperature measured by an
embedded thermocouple and the temperature measured by a two-color infrared pyrometer
when a sample of LaCrO3 was heated in the microwave furnace. LaCrO 3 approximates a
black body and represents an excellent material for temperature measurement using a two-
color IR pyrometer.

1500 VIIVII..1 30

1400 25
o 20
S1 3 0 0
15 .

S1
2 00
C. 10.
E wl
I- 1100 5

1000 0
0 2 4 6 8 10 12
Time (min)

Figure 2. Turning the electric field on and off had no effect on the thermocouple
readings during annealing of alumina at 28 GHz [3].
 177

In conventional furnaces, accurate measurement of temperature is often obtained by

using a sensing thermocoup!e in addition to the control thermocouple. The sensing
thermocouple is placed directly on, or within a few millimeters of, the sample. In the
procedures normally used at ORNL, the sensing thermocouple normally is the control
thermocouple. To assure that good temperature measurements are made, particular
attention must be made to placement of the thermocouple. Figure 3 shows two examples
of thermocouple placement that have been found to give reproducible results. For a
cylindrical specimen, a thermocouple well is drilled about 3/4 of the way through the
sample. The thermocouple is then inserted until it abuts the back wall of the well. For
plate specimens, the thermocouple is inserted between two pla . This ensures that the
thermocouple sees the same thermal environment in both directions. Simply placing the
thermocouple in contact with the surface of a cylinder, or on the top or bottom of a single
plate, guarantees that the temperature that is measured will be too low. Instead of
measuring the temperature of the part, the thermocouple will measure some average of the
part temperature and the temperature of the surrounding insulation.

Temperature uniformity

Temperature uniformity within a sample is a concern on several scales. Johnson

[18] has shown that there can be essentially no microscopic gradient within a sample
caused by preferential heating of grain boundaries. His calculations were based on a
model in which all of the microwave energy is deposited in a region close to the grain
boundary. Even in grains as large as 100 p.m, the worst case temperature differential was
only -0.1 *C. In contrast, significant gradients can exist on a macroscopic scale in both
microwave and conventional processing. Conventional heating deposits energy on the
surface of a sample and depends upon thermal diffusion to transfer the heat to the center of
the sample. Microwave heating deposits energy throughout the volume of the sample
(assuming the penetration depth of the microwaves is larger than the thickness of the
sample) with cooling of the sample from the surface. Therefore, the normal temperature
gradient in a conventionally heated sample is opposite to that in a microwave heated
sample, Fig. 4. The difference in surface and centerline temperature can be rather large.
Figure 5 shows the surface and centerline temperatures for microwave and conventional
drying of a castable insulating refractory, K3300 (Thermal Ceramics, Augusta, Ga.) The
centerline temperature of the microwave dried sample is 30 to 40'C above the surface
temperature. The centerline temperature of the conventionally dried sample is 20 to 30'C
below the surface temperature. In this case, a one-to-one comparison of the microwave
and conventional processes is difficult because the physics of the situation are so different
in the two cases.
Thermal gradients often arise from improper insulating systems. During microwave
processing, such gradients often lead to differential sintering as is shown in Fig. la.
Development of a workable insulating system may require extensive effort. For example,
the sample shown in Fig. l b was the first "good" sample processed in the ZrO2-12 CeO 2
system at ORNL. This was the 27th sample of the series. Numerous other insulating
schemes were tested to try to alleviate the thermal gradient problems. The first 26 attempts
were unsuccessful. Unfortunately, such efforts are not unusual. Development of a
workable, repeatable insulating system remains one of the most challenging tasks in the
high temperature microwave processing of ceramics. To determine whether the insulating
system is working appropriately normally requires investigation of the microstructure of
the sample after firing. If the microstructure is uniform, one can be assured that the
thermal history of all parts of the sample was the same. Microstructural development is
the most sensitive probe of thermal history available.
Sometimes the gradients developed by microwave processing can be used to
advantage. Clark, et al [19] have shown advantages in solid state high temperature
synthesis (SHS) reactions from microwave heating. The reaction front is initiated at the
center of the pellet, and then proceeds toward the periphery, which results in more
complete reaction of the pellet than conventional heating. Likewise, Kiggans, et al [13]
showed that nitridation of silicon proceeds from the inside of a powder compact to the
surface under microwave processing. In this way, the material was converted from silicon
to Si 3N4 more completely than in conventional processing.
178

PRIMARY
INSULATION

Figure 3. Examples of thermocouple placement for accurate temperature

measurement in a microwave furnace for cylinder and plate sample geometries.

Conventional Microwave
External Heating Causes the Intemal Heating Causes the
Centerline Temperature to Lag Centerline Temperature to Lead
the Surface Temperature the Surface Temperature
Figure 4. Temperature gradients generated during microwave and
conventional heating of samples tend to be different.
 179

6'120
0 90

960
C. " -- I MW Surface
E 3--I Conv Center
30 ---I Conv Surface

0 20 40 60 80
Time (min)

Figure 5. Microwave drying of a castable refractory (Thermal Ceramics K3300)

raised the internal temperature well above the surface temperature.

12 1600
*._• 1500 -'
S 9
1400
0. -- 0- MW Power
5.0
3:4 6 -0-T(SiC) 1300
o- • 1200
CR30 Alumina 1100 •O
2.45 GHz
Isopressed, 30 kpsl
0 1 a 1000
0 100 200 300 400 500
Alumina Sample Weight (g)

Figure 6. Increasing the size of an alumina sample reduced the SiC picket
temperature and raised the total power indicating that more of the power in the furnace was
absorbed by the alumina sample.

SAMPLE CONSIDERATIONS
size
In general, samples used in microwave processing studies at ORNL are larger than
those used in conventional studies. For example, conventional sintering studies are
typically performed using 1 to 2g samples. Microwave sintering studies at Oak Ridge
have typically used 30 to 150g samples [1,2,4,10,11,14,15]. The use of large samples in
the microwave furnace is related to coupling power into the sample and to temperature
measurement. Very small samples do not couple well to the microwave field. Larger
samples couple better. (A demonstration of this is easily done with popcorn in a home
microwave oven. A single popcorn kernel cannot be heated; a cupful of popcorn is easily
popped.) Temperature measurement using small samples is difficult if thermocouples are
used. The thermal mass of the sample relative to that of the thermocouple is small; hence,
the thermocouple reads a temperature that is indicative of both the sample and the
surrounding insulation, which is normally at a lower temperature than the sample, rather
than just the sample temperature. If a radiation pyrometer is used (IR or optical), a small
sample may experience significant radiative cooling. The pyrometer sight hole can
represent a large fraction of the total solid angle through which the sample radiates.
Again, a temperature lower than the true temperature will be recorded.
Another effect of size has been observed for the firing of samples using the "picket
fence" indirect heating arrangement. The picket fence [2,14] consists of a ring of silicon
180

carbide rods surrounding the sample and its insulation and enclosed by an outer insulating
cylinder. The picket fence is used to preheat the sample to a temperature high enough that
the sample will couple directly to the microwaves. This arrangement has worked well for
microwave sintering of zirconia [2], alumina [14], and alumina-zirconia composites [ 15].
Figure 6 shows the total microwave power used and the temperature of the silicon carbide
pickets for firing various alumina samples at 1600'C. As the alumina sample size
increases, the temperature of the silicon carbide rods decreases. This indicates that the
amount of power absorbed by the silicon carbide rods decreases as the alumina sample
size increases. This observation is consistent with Kimrey's model [14] of microwave
absorption in multimode cavities. It also indicates that the microwave power level in the
alumina increases as the alumina sample size increases.
In the study of some materials phenomena, small samples must be used. Examples
include grain growth studies [3] and diffusion experiments [4]. In both of these cases, a
ballast/crucible [3] was constructed to permit the processing of the small samples. The
basic premise behind the ballast/crucible is to increase the dielectric load in the microwave
furnace with a material having dielectric properties either identical or very similar to the
sample being studied. Figure 7 shows a schematic of a ballast/crucible used for the
alumina grain growth and diffusion experiments. In the case of the grain growth
experiments, 40 test coupons 0.5 cm 3 in volume were cut from a single hot-pressed
sample for annealing in both conventional and microwave furnaces. A single hot-pressed
sample was used to assure that the initial materials conditions for grain growth were
identical for all of the microwave and conventional annealing studies. The ballast/crucible
was made by isopressing and sintering powder from the same processing batch as the hot-
pressed samples. In the case of the diffusion experiments, single crystal sapphire coupons
1 x 1 x 0.1 cm were used. The ballast/crucible for these experiments was made by
isopressing and sintering ultrahigh purity (undoped) alumina powder, which gave as close
a match as possible to the dielectric properties of the sapphire. To confirm that the
ballast/crucible concept was a reasonable approach, it was tested by sintering small (3g)
samples of alumina. The sintering behavior of the small alumina samples in the
ballast/crucible was identical to that of large (150g) samples.

Insulation and packing powders

As was mentioned above, development of an insulating system that gives

reproducible results is perhaps the most difficult task in the high temperature processing of
ceramics using microwaves. A few guidelines are in order regarding how one proceeds.
As is shown in Fig. 3, the insulation packages used for microwave processing at ORNL
are typically multilayered. The primary thermal insulation, i.e., the outer envelope, should
be microwave transparent to allow penetration of the microwaves into the sample. The
inner layer of insulation should~be matched both chemically and dielectrically to the sample
as closely as possible. If done properly, this approach results in essentially isothermal
conditions within the sample. This is observed experimentally as a uniform microstructure
in the final sintered part [1]. All of the thermal gradient is produced in the insulation. In
some cases, the inner insulation will actually be a packing powder of the same
composition as the sample. This is typical for silicon nitride, for example [20].

FIELD EFFECTS
SMicrowave field effects are important to consider in the processing of materials in a
microwave furnace. Three phenomena that are important are uniformity of the field in the
furnace, concentration of the field around the tip of a thermocouple, and arcing.
Nonuniformity of the microwave field is a serious concern because it leads to the creation
of hot-spots in parts that are being processed. Figure 8 shows a representation of the
variation of the electric field in a typical small microwave cavity (similar to a home
microwave oven) [21]. Field concentrations of this type are ubiquitous in cavities
operating at 2.45 GHz [2,21]. Figure 9 shows the effect of field concentration around the
tip of a thermocouple in the sintering of ZrO 2 - 12 mol% CeO 2. The microwave field was
higher at the tip of the thermocouple than along its length. This produced more heating at
the right side of the sample than at the left side. Differential sintering is evidenced by the
"banana" shape of the sample and the crack on the left-hand side that developed during
sintering. Figure 10 shows an example of arcing at 2.45 GHz between a thermocouple
 181

Alumina Bulk
Fiber Insulation Dense Alumina
- Crucible
Boron Nitride T•
Crucible

!S4-- Alumina
Sample

Type C Thermocouple with

Molybdenum Sheath
Figure 7. A ballast-crucible arrangement was developed for processing alumina
samples for grain growth and diffusion experiments [3].

100

80
A

60
C 40

20

0
0 20 40 60 80 100
Relative Position
Figure 8. Microwave field variation is sizable in a 40x40x50 cm microwave furnace
operating at 2.45 GHz [21]

ZrCe 92012
Banana shape
Crack from Differential Sintering
TZ12CE/ALBF Insulation
Figure 9. Concentration of the microwave field around the thermocouple sheath
caused differential heating of ZrO 2 - 12 mol % CeO2.
182

and a LaCrO3 sample. Arcing caused the tip of the thermocouple to heat excessively,
which produced a series of upward spikes in the temperature profile. Because the rapid
heating pushed the thermocouple temperature well above the controller setpoint, the
furnace controller shut off the power, the arcing stopped, and the temperature of the
thermocouple tip dropped rapidly. Since the thermocouple temperature was now below
the setpoint temperature of the furnace controller, power was reapplied and arcing
resumed. Thus a rapid periodic temperature fluctuation was established, Fig. 10. To
prevent arcing, a dielectric barrier of calcium aluminate cement (Alundum, Norton Co.,
Worcester, MA) was placed between the thermocouple sheath and the sample, which
reduced the electric field strength in that region, and hence the propensity to arcing. The
resulting furnace run using the Alundum-coated Pt-sheathed thermocouple is shown as the
top trace in Fig. 10, which shows no evidence of arcing. The microwave power-time
traces (not shown) mimicked the temperature-time curves in both cases; i. e., the trace was
irregular for the arcing case and smooth for the non-arcing case. For other applications,
alumina and boron nitride thermocouple protection tubes have been used to accomplish the
same task of mitigating electric field concentration.

0 600
.• 500 No Arcing

0 400 (Alundum Coating)

4. 300
"0Arcing
o 200 (Bare Sheath)
E 100 LaCrO Sample
0-
0 20 40 60 80 100
Time (min)
Figure 10. Arcing from a platinum thermocouple sheath to a LaCrO 3 sample at 2.45
GHz was eliminated by coating the sheath with calcium aluminate (Alundum®) cement
[10].

ATMOSPHERE EFFECTS

Gas composition in conventional processing of ceramics is normally used to fix the

partial pressure of oxygen in the furnace. Therefore, neutral atmospheres of argon or
nitrogen or vacuum are often considered to be equivalent. In microwave processing, an
additional factor must be considered; that is the breakdown potential of the gases under the
influence of the microwave field [22]. Figure 11 shows the effect of pressure, gas
composition, and frequency on thebreakdown field strength. At all pressures, air (or
nitrogen) has a much higher breakdown potential than does argon. This difference
becomes pronounced at high pressures approaching atmospheric pressure. Nitrogen is
often used instead of argon if a non-oxidizing atmosphere is required to prevent problems
associated with arcing. At intermediate pressures, say between 0.01 and 10 torr, plasma
formation is a real possibility. Unless specifically desired, plasmas are best avoided
during microwave processing. At very low pressure, e.g. below 10-4 torr, neither arcing
nor plasma formation is a problem; there just are not enough gas molecules around to
cause either phenomenon to occur. Figure 11 also shows that the higher the frequency,
the higher the breakdown field strength required to cause arcing or plasma formation.
Therefore, operation at 28 GHz is less sensitive to plasma formation than is operation at
2.45 GHz.
 183

10000 1
Data taken from - A.D. Ma~cDonald
1
__ Microwave Breakdown In Gases
j. Wiley and Son@, 1000

760
O5. 1000 torr
"oE

> 100
"0
- -0- Air - 0.992 GHz
-0- Air -9.4 GHz
m 10 --A--Argon - 0.992 GHz

0.01 0.1 1 10 100 1000

Pressure (torr)

Figure 11. Dielectric breakdown in a microwave field is sensitive to total pressure,

gas composition, and frequency. 22

SUMMARY AND CONCLUSIONS

Guidelines have been developed at ORNL for conducting meaningful side-by-side

conventional and microwave thermal processing experiments on ceramics. These
guidelines may be summarized as follows:

1. Microwave and conventional furnaces are inherently different in terms of materials

processing.
2. Dielectric properties must be explicitly considered in planning microwave
experiments, including both temperature and frequency effects.
3. Temperature measurement in a microwave furnace is always an issue. In particular,
each new material or experimental setup brings with it new problems in temperature
measurement.
4. Thermal gradients may be inherently different for conventional and microwave
processing. Conventional samples tend to be cooler in the middle than on the
surface; microwave samples tend to be warmer in the middle.
5. Sample size may have to be adapted to microwave processing conditions; in general,
a minimum sample size of 50 to 100g is required in the microwave furnace unless
indirect heating or a ballast/crucible is used.
6. Field effects must be managed in a microwave furnace including arcing of
thermocouples to samples and field concentration around thermocouples. Also, the
field uniformity in the microwave cavity (or lack thereof) can have a decided effect
on how materials can be processed
7. It must be recognized that the processing atmosphere interacts with the microwave
field. Plasma formation and arcing can be a problem, especially when operating in
argon or other gases with low breakdown potential.

REFERENCES

1. M. A. Janney and H. D. Kimrey, "Microwave Sintering of Alumina at 28 GHz,"

Ceramic Powder Science, II, pp 919-924, G. L. Messing, E. R. Fuller, and H.
Hausner, eds., American Ceramic Society, Westerville, Ohio, 1988.
184

2. M. A. Janney, C. D. Calhoun, and H. D. Kimrey, "Microwave Sintering of Solid

Oxide Fuel Cell Materials, I: Zirconia-8 mol% Yttria," J. Am. Ceram. Soc., 75
[21341-6 (1992).
3. M. A. Janney, H. D. Kimrey, M. A. Schmidt, and J. 0. Kiggans, "Alumina Grain
Growth in a Microwave Field," J. Am. Ceram. Soc., 74 [7] 1675-81 (1991).

4. M. A. Janney and H. D. Kimrey, "Diffusion-Controlled Processes in Microwave-

fired Ceramics," Microwave Processing of Materials, 11, Proc. Mater. Res.
Soc., Vol. 189, 215-27 (1991).

5. D. Lewis, "Review of Polymer Processing with Microwaves," this proceedings.

6. R. Giguerre, "Review of Organic Synthesis Using Microwaves," this proceedings.

7. P. Marquardt, G. Nimitz, G. Heite, and H. Peters, "Microwave Evidence for a Size-

Induced Metal-Insulator Transition in Mesoscopic Conductors," Microwave
Processing of Materials, I, Proc. Mater. Res. Soc., Vol. 124, pp. 155-160
(1989)
8. T. Nishitani, Method for Sintering Refractories and an Apparatus
Therefore, U. S. Patent 4,147,911; filed Aug. 10, 1976; issued Apr. 3, 1979.

9. C. E. Holcombe; personal communication.

10. M. A. Janney and H. D. Kimrey, "Microwave Sintering of Solid Oxide Fuel Cell
Materials, II: Lanthanum Chromite," submitted to J. Am. Ceram. Soc.
11. C. E. Holcombe and N. L. Dykes, "Importance of 'Casketing' for Microwave
Sintering of Materials," J. Matl. Sci. Lett., 9, 425-8 (1990).
12. C. E. Holcombe and N. L. Dykes, ""Ultra" High-Temperature Microwave
Sintering," Ceramic Trans., Vol. 21, 375-85, 1991.
13. J. 0. Kiggans and T. N. Tiegs, "Characterization of Silicon Nitride Synthesized by
Microwave Heating," Ceramic Trans., Vol 21, 403-410, 1991.

14. H. D. Kimrey, "A Quasi-Optical Model for Microwave Processing," this

proceedings.

15. H. D. Kimrey, J. 0. Kiggans, M. A.Janney, and R. L. Beatty, "Microwave

Sintering of Zirconia-Toughened Alumina Composites," Microwave Processing
of Materials, II, Proc. Mater. Res. Soc., Vol. 189, 243-55 (1991).
16. R. A. Peterson, Microwave Heating Temperature Control, U. S. Patent
3,859,493; filed Nov. 21, 1973; issued Jan. 7, 1975.
17. W. H. Sutton and W. E. Johnson, Method of Improving the Susceptibility
of a Material to Microwave Energy Heating, U. S. Patent 4,219,361; filed
Jun. 9, 1978; issued Aug. 26, 1980.
18. D. L. Johnson, "Microwave Heating of Grain Boundaries in Ceramics," J. Am.
Ceram. Soc., 74[41849-50(1991).
19. D. E. Clark and D.C. Foltz, "Microwave Processing Activities at the University of
Florida," Ceramic Trans., Vol. 21, 29-34, 1991.

20. J. 0. Kiggans and T.N. Tiegs, "Microwave Processing of RBSN and SRBSN
Silicon Nitrides," this proceedings.
 185

21. C.R. James, W.R. Tinga, and W.A.G. Voss, "Energy Conversion in Closed
Microwave Cavities," in E. C. Okress, ed., Microwave Power Engineering,
Vol. 2, Academic Press, N.Y., 1973

22. A.D. McDonald, Microwave Breakdown in Gases, J. Wiley and Sons, 1966.