THESE!

feet

This is to certify that the

thesis entitled

Joining of Ceramic Materials Using Spin-on Interlayer

presented by

JongeGi Lee

has been accepted towards fulfillment
of the requirements for

I
Master 3 _degree in Materials Science and
Engineering

 

6%03er

Major professor

DateCDCEE ' '7’! $300

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

LlahAnv

Miehlgan State

University

"‘ -._, '—

 

 

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

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

11m mus-p.14

 

 

 

 

 

 

JOHN} ()l‘ (TR

up

JOINING OF CERAMIC MATERIALS USING SPIN-ON INTERLAYER
By

Jong-Gi Lee

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF SCIENCE

Department of Materials Science and Mechanics

2000

 

 

 

JOINING 0i (i. R \\

 

.n‘ ’ . ‘
in». mm WW7
. .

rm ~-.- .1 “

' \ l n.‘ v 7“."

\-¢..L him-(LIN; slfianu s:
e

U"

. . "P r ‘ ‘
~04..,:u;...ilb(.'ni|1 Ll‘lnup' “:10

r

'1' “IV l!t\ V
Kn... v \‘ ‘ “ r’N‘ ‘ 2 ‘1'
:Vlus_n Leis/the u cc

~‘W '
‘A. I 9 ‘ 7
“MN killing UT U 'r' ‘ "
~ _ .. s
"'1' .7.” .,. ‘
Mains) to DU mlfid'b

Fri-www- “
. amine: in C 4'

:3‘ ‘7'.ng [v
nu. .Tm.‘ cf: Jim ‘5'
“Viv

“L ‘19,. I I .
' ".‘uB WU. {flit Err-C“ r

P121," "3 .- '
k..L0-[Uin.n" Tl“:

--
\

If; ""I I-{lr r I
s“ Mi.) fiés’ti ~, ..
U >E‘K’s i.

a.

"CI ‘
.‘EHVLI' 1:,“ ' H '
l s..pdlllall‘~ 31'““"

u~ c..l

14 I ‘! I
.. ~35 - :1 «L '
“43¢ \lidlrwf" J "l.
-sb 5
~

ABSTRACT
JOINING OF CERAMIC MATERIALS USING SPIN-ON INTERLAYER
By

Jong-Gi Lee

Using various spin-on materials with submicron initial coating thickness, oxide
ceramics including alumina, partially stabilized zirconia, fully stabilized zirconia,
alumina-zirconia composites, hydroxyapatite, and MaCorTM and non-oxide ceramics
including silicon carbide and magnesium fluoride have been joined using either
microwave heating or conventional heating with applied low external pressure for a range
of 40 minutes to 60 minutes. Non-oxide ceramics including Mng and SiC were joined at
a low temperature (700°C and 900°C) using ceramic precusor spin—on interlays
transformed into an amorphous ceramic upon heating. SEM micrographs showed smooth
bond layers with thickness ranging from ~0 to 10 microns.

Prior to joining, notches about 400 microns in depth and 350 microns in width
were cut into selected specimens using a low speed diamond saw for alumina-zirconia
composites, partially stabilized zirconia, fully stabilized zirconia, hydroxyapatite, and
MaCorTM. Microscopic investigation of the joined specimens showed that the notch

dimensions had changed by less than 3 % on average-

   

I. lnlmduction.

J

I‘ I " ’

ll .ivl'jtilli‘n 0’. i I‘

q r t

l. Jig-Tin: ltm'miw
. .

|Q ' I

Tom: Tammie
5 I

11 \l‘.;f;~t\.r~c lurk".

.-
~

[\ptrimrntal Prowl L

..'. Him} Twit: ‘3‘
3.3 Cid Prom: (Ca
1‘ P7347533? 0', R;

‘C R
\ .p.
a)" .V
n—. SHIIR, ”I I

r ‘\
Mi; ‘l\L\\
.

‘, A
f at
Satin-c: P't“ "
I . rbAb\
‘-
. CuttirtaPrwt-ti '
‘.
\ i . .
.k -\A.irl"Ku‘\CJki"‘.'
‘
I
.L Ivy-9““? Ya ~Ltv‘.‘ -
.u..l: \\\.§._.\. 4
‘

-‘ijpv!"'.-. 1 ' '
“IMMJ‘JL \\, u

‘ I\.
‘H

. ll Primate nt Pr-

‘ .

:1‘ “Hr

AKIII\J\C £J‘\

‘1 .
L- 3 ‘ "~

. l.\‘k\“d\a]
:1 \ k"I.“

‘ .
--~. \JV‘thQ
\ ‘1 (2“

Table of Contents

page
1. Introduction. ............................................................................ 1
1.1 Classification of Ceramic Joining ................................................ 2
1.2 Joining Techniques Using interface ............................................. 2
1.3 Joining Techniques Without Using Without bond agent ..................... 7
1.4 Microwave Joining ................................................................ 9
2. Experimental Procedure. ............................................................. 17
2.1 Materials Used in This Study .................................................... 17
2.2 Cold Pressing Ceramic Powder Compacts ..................................... 22
2.3 Preparation of Refractory Insulator Used in This Study ..................... 22
2.4 Microwave Control System ....................................................... 24
2.5 Sintering Process Using Microwave Power .................................... 27
2.6 Specimen Preparation ............................................................. 33
2.7 Coating Procedure Prior to Joining .............................................. 35
2.8 Microwave Joining Procedure ................................................... 38
2.9 Joining technique for Non-Oxide Ceramics ................................... 41
2.10 Producing Notch Mechanically Using A Low Speed Saw. ................. 41
2.11 Procedure of Producing Notch Prior to Sintering ............................. 45
2.12 Mounting and Polishing of Joined Specimen ................................. 49
2.13 Vickers indentation Testing of Joined Specimen ............................. 49
2.14 SEM Observation of Bond-layer ................................................ 49
2.15 Infrared Transmittance Measurement of Mng ............................... 53
2.16 Microwave cavity cleaning ...................................................... 54
3. Results. .................................................................................... 59
3.1 Joining of Diamond Thin Film to Optical AndIR Materials. ................. 59
3.2 Microwave Joining of Particulate Composites. ................................ 80
3.3 Geometrical Stability of Holes and Channels During Joining
of Ceramics and Ceramic Composites. .......................................... 96
3.4 Joining Ceramics To Produce Components With Precise
lntemal Channels. .................................................................. 117
3.5 Joining of Non-Oxide Ceramics Using Conventional
And Microwave Heating. .......................................................... 138
3.6 Protective Coatings for Infrared Materials. ....................................... 151
3.7 Joining Of Polycrystalline Ceramics And Ceramic Composites
Using Microwave Heating. ......................................................... 164
3.8 Joining Dissimilar Ceramic Materials. .......................................... 175
4. Discussion and Conclusion. .......................................................... 196
4.1 Thermal Stress in the Dissimilar Ceramics Joined in This Thesis ............ 196
4.2 Summary of Results .................................................................. 204

iii

 

 

1 mi! \iJui‘i‘ ..

1pptndl\ 1an'i'xc ;

SCkiIl.\n Ila-C

1ppendiiBSt. it‘TlYi .«i't' .

energ\ ”‘1 ill
be

typrndiif \1:.'riig'.;;~‘:. .

11". it“; '.\ \t'.

itiirtnces

 

4.3 Future studies ......................................................................... 207

Appendix A Literature surveyed joining techniques cited in introduction
section. The number was set via references number .................... 208

Appendix B Sintering andjoining heating rate using the Microwave
energy in this thesis. ........................................................ 225

Appendix C Micrographs of joined ceramic/ceramic not included
in results section. ............................................................ 239

References. . ....... ................ . .......................................... 248

iibkl. Alumni 1.13111
liblt’ dezoiiipitzic ti
liblei. Slilcnil intumiit.

oi (
lib“. Infurtiiuiznn o. M.

 

lIka. 111267.41 iniurttii“
111)“. The spin-on 111.113?

11bit Iconiillliins 01' 5""

..nn\

11111725 thesis. ..

t‘ .Iric'

ltbltfii’cishzng 212i
Valid-ed In this 11":
twig piltsiicd. l?
are reduced. ......

We 9. A sum

171m 01 175.
immature. livid

"‘9 m ‘l '\ - l ‘ I
\\T‘1&Illlech-'TI;L

lible 10. Eiptti mental ;

I“
am silicon car's

Wt 11. Type of flit-:5 .:

l .3 .
111161.13p5 or hilt-rs .I

'-
I

1111113. The condim-

.,.\i

hit 14. C

“1. v
4 .1...“ , ‘” -
“1“‘lt11‘n1‘1: 'r

Tab! -
*1: Sinzssnz card»-

rain
6‘ ",3 b
Artur-Spied jt‘itif'
111)! .
1E 17. \‘ii‘

Ir 0 '
.t . .
k ’ \‘V‘Lx'ilhi l.

List of Tables

Page
Table 1. Alumina (TM-DAR, Taimi Co, Japan) powder information .................. 18

Table 2. Hydroxyapatite powder material (Cerac INC, Milwaukee, WI) information. .. 18

Table 3. Material information of two kinds of zirconia ..................................... 19
Table 4. Information of MaCorTM from vendor’s web site. ................................ 19
Table 5. Material information of Mng. ...................................................... 20
Table 6. The spin-on materials used for ceramic/ceramic joining. ........................ 20

Table 7. Conditions of Sintering for each of the commercial powder material used
in this thesis. ............................................................................ 31

Table 8. Polishing time and diamond grit size for each of the ceramic materials
included in this thesis. The time corresponds to seven mounted specimens
being polished. If fewer specimens are polished simultaneously the times
are reduced. ............................................................................. 31

Table 9. A summary of the interlayer materials, joining materials, joining
temperature, hold time, and deadweight loading for all of the

ceramic/ceramic joining included in this thesis. ................................... 44

Table 10. Experimental condition of conventional joining of silicon nitride
and silicon carbide. .................................................................. 44

Table 11. Type of filters and detectors used to measure the IR transmittance

of the Mng specimens. ............................................................. 52
Table 12. Type of filters and detector for various wavelengths. ........................... 68
Table 13. The conditions of joining both microwave and conventional furnace. ...... 68
Table 14. Calculation of the optical absorption factor, a ................................... 75
Table 15. Sintering conditions for powder materials used in this thesis. ................... 125
Table 16. Attempted joining conditions for joining MaCorTM and HAP. ............... 125

Table 17. Notch stability of joined HAP with MaCorTM before and after joining ....... 126

Wt 18. “it nitit‘h (54'

 

111119. Tn: nail: .1:.".':.'

I" ‘9 r
at! tomb». .u.

hill 30. Tris notch 1131i;

.1i'.fll(lil‘:ll‘.£ for '
11bit :1. Wm] exam:

111122. lien! fund...‘

. o

W 23. Young's mica}:

lible21. Sim-mm ‘Oi‘IdZIh

11. b
s

 

Milt 15 PTlX‘LKNII“: 0.1-. 1..

Milk ”T

Titmal prune .2;
l w m, .

1116.1. Emil privpctftz
13ka Inertial expire.-
I‘ . .

01:39. 10qu s mmjsii

l ' '
11130. Pawn 312:?

9
.t‘ _.

libki -‘ '
.1.Ca.t..itcdihcttti.

11bit 31 A167.

“Di-i ,-
3 Lining” \

T
IM‘I.JOIE:IAC "V‘ .
-: mags? '

:Ifi‘r‘fin e
'“‘I:'.- .‘arbi
C ”‘11..”

l L'
Lac; ’3'.
\I '1: “113' \T '\
'3
t '11]?! -
S :17 01 f7-
61'5‘AH \
110. R
F. :?
Table i
.Lkimw,"
I‘“Utr~‘.t. .
‘r ‘ “was .-
r-r _ 'H.
sum-2n, 'L
S

Table 18. The notch change before and afterjoining forjoined Z102 with MaCorTM. 126

Table 19. The notch dimension change before and

after joining for joined ZrOz (8 mol % Y203). .................................. 127
Table 20. The notch dimension change before and

after joining for joined Ale3/ZrOz composites. ................................. 127
Table 21. Thermal expansion coefficients for the Materials Included in this Study. 184
Table 22. Thermal Conductivites of the Materials Included in this Study. ............... 184
Table 23. Young’s moduli for the Materials Included in this Study. ....................... 184
Table 24. Sintering conditions for materials used in this study. .......................... 185

Table 25. Processing conditions for dissimilar materials using microwave power. 185

Table 26. Thermal properties of materials used in reference [59]. ....................... 199
Table 27. Thermal properties of materials used in reference [55]. ....................... 199
Table 28. Thermal expansion coefficients for the materials included in this thesis. ...... 205
Table 29. Young’s moduli for the materials included in this thesis. ..................... 203
Table 30. Poisson’s ratio for the materials included in this thesis. ...................... 203
Table 31. Calculated thermal stress values of joined specimens. ........................ 203
Table 32. Averaged notch stability before and after joining in this thesis. .............. 207

Table A1. Joining materials, dimension, adhesive, thickness of adhesive,
heating method, heating rate cooling rate, Tmax, hold time,
bonding pressure, mechanical test, span length, dimension,
loading rate, cross speed, micrograph, thickness of bondlayer,
strength of material, bonding or shear strength, and atmosphere
were written. Reference number 1 to 18 used bonding agent during joining. 209

Table A2. Joining materials, dimension, adhesive, thickness of adhesive,
heating method, heating rate cooling rate, Tmax, hold time,
bonding pressure, mechanical test, span length, dimension,
loading rate, cross speed, micrograph, thickness of bondlayer,
strength of material, bonding or shear strength, and atmosphere
were written. The number was set via references number.
Reference number 24 to 29 used no bonding agent during joining. ...... 220

vi

figurtl Sthcmiiic 111.16

m pnngipic oi

 

lb: martini ; 1
1C1 complete w.“
|
11211111.. '1\1".-'C ...:"..‘i1-.1.
1101111311? \LIi'?.I..
1116 RBSC. 11h! 1.
across the RSV
111611;}81121]

figure 3. P415 "is 01 nu . 1

ml mitrtiuaxc I,

131121 Imitation of 31:. '

 

1111111101311 iitim 8.11.1]
FEHRD. LOSS. 114112;:8111 15 1,.

an6, 41 PHUtUVy-th U?

l

11‘ St‘htmaiic 0.1".

Pruitt...

i‘emd: 1C 111$
11691 13417.41 m
r
T1116 31mm REIT

19 Mon 01‘6““; .

“TM Ii

\.1\t1d1111_1n ("I rct .

111M ph,

1111:11114ap.‘iffi 11;,
W‘tersim

mudc 01 \"u
l \

F'VUR11-118C1‘iem.m

uid1b41gh111

\lv-Q

111111611 P111

. it Q}
10161111159“ 1
. ““X‘t'

List of Figures

Figure 1. Schematic illustration of sessile drop formation:
(a) principle of wetting denoted contact angle and
(b) interfacial energy, (b) non-wetting drop,

(c) complete wetting [21] . ..............................................

Figure 2. A schematic illustrating diffusion bonding of free silicon onto
a butting surface. (a) Migration of free silicon onto surface of
the RBSC, (b) As force (pressure) is applied, the silicon spreads
across the RBSC surface, (c) Free silicon forms a continuous

interlayer [24]. ................................................................

Figure 3. Patterns of heating in (a) conventional furnace and

(a) microwave (adapted from Sutton [31] and Lee [48]). ..............

Figure 4 Interaction of microwaves with materials at ambient temperature

(adopted from Sutton [31] and Lee [48]). ...................................

Figure 5. Loss tangent (8 to 10 GHz) versus temperature [31]. ....................

Figure 6. (a) Photograph of the presser used in this thesis,

(b) Schematic sketch of pressing powder compact in a steel die. .......

Figure 7. Schematic illustration of the technique used to fabricate
the cylindrical portion of the refractory casket.
The starting material was a 30 cm long section of a zirconia

refractory open cylinder. .....................................................

Figure 8. Illustration of refractory cavities used in this thesis. .....................

Figure 9. Photograph of the microwave cavity and the microwave

power supply used in this thesis. ............................................

Figure 10. Schematic of specimen temperature measurement

technique used in this thesis. ................................................

Figure 11. Schematic illustration of refractory casket arrangement

used batch Sintering of ceramic specimens in this thesis. ..............

Figure 12. Photograph of heater and aluminum plate used in
mounting the specimens for polishing.
Note that several specimens have been fixed to

the plate via thermoplastic. .................................................

vii

Page

....... 8

....... 10

ll

12

23

25

26

28

...... 30

....... 32

36

YEW? 13. PML‘STJp'f‘. t“

w

 

I'.

figurell.Schcmuuc1 ..
. ahzghsr'cdv

' ‘ 1v ‘-
szure 1:. A phohg. .19.
”mm“ M C (.1.

figure 1&3: Pha‘nngmp.“
. ffffdih‘r) (mi-u
(b: Phuiugmp‘i

:1ka min 11m-

dead \sczehzs :r

We”. Photograph vi.
the spmmcn h.

 

figurtlfl. Sghemang M in.
cut nuzghw :ntw

figure 19. Schemang

pm oer (mpg;

OI ..\

Fm 31}. Schematic 11iu~tt
after Jmn

' .
WIS 431d

1 J '
Pram. Synsmauc or V;

FM] ‘1 --' -~ ‘ '
1- 1.8.. dcmutlc 0? {fix

the IR trummrth

f‘ 1
W3. A us“ of the V

Snoun are m:- :1

m4 ‘.
“NRC shdzhu s

Fm] ~1
,MCU —

. 1ptuppm
”e flnzer x1
5.. f ..
ppm} €0.5qu
..us.a:\:z‘“»c* '

.A \d

Figure 13. Photograph of the automatic polisher with the aluminum

plate in place for polishing. ..............................................

Figure 14. Schematic illustration of coating procedure using

a high speed spinner. .......................................................

Figure 15. A photograph of the 2.45 GHz single mode

microwave cavity used in this thesis. ....................................

Figure 16. (a) Photograph of the type of zirconia/aluminosilicate
refractory casket used during ceramic/ceramic joining,
(b) Photograph of the bottom SALI plate of the refractory
casket with the specimen, with the specimen setter and

dead weights in place. .....................................................

Figure 17. Photograph of a low speed diamond saw after

the specimen had been in the saw. ......................................

Figure 18. Schematic of the technique used to mechanically

cut notches into the specimens. ..........................................

Figure 19. Schematic of “stamping” notches into the green (unfired)

powder compacts, prior to Sintering. ....................................

Figure 20. Schematic illustration of mounting the specimen

after joining and cutting. ...................................................

Figure 21. Schematic of Vikers indentation near the bond layer. .................

Figure 22. Schematic of the apparatus used for measuring

the IR transmittance of Mng specimens. ...............................

Figure 23. A view of the microwave cavity disassembled for cleaning.

Shown are the main microwave cavity, the lanch probe,

and the sliding short ........................................................

Figure 24. Cavity top plate showing finger stock.
The finger stock makes electrical and
physical contact with the inner cavity wall.
Thus, as the cavity top plate moves, the effective

cavity length changes. .....................................................

Figure 25. For the launch probe part of the probe assembly

includes small finger stack. ...............................................

Figure 26. SEM image of the fracture surface of Mng. ...........................

viii

........ 37

........ 4O

42

....... 43

46

......... 47

......... 48

50

51

........ 52

56

57

58

......... 61

figure 27. St'hcmiu, xl‘.

used In £0.11?

   
 
  

lure 28V Schcmitu wt '5
The dud \tt‘ig'

Figure 29 Schematic u!
the IR Irinxnzzt'

llgunw SElemngr.
the cun\t‘t‘:‘.1n'
l3! MIJIMJ'.
\ch (“gin \x,

name 41 ”t!

“gulf“. SEM lmdgc u: ‘

lhcstdum ~2T..

 

b) migrwuxc .'
dad waist lla:
llgurt32. lransrrgtmntc '
(at Ctmxcmznt.
lbl Xlit‘rtm 4w

130R” Comp-in son Ul
\lgFj From l‘. _'

531W. As Stllt’llisl'c \'

v , -. .
di'lUUF tratcffa,

F .
{Mimi SE.\l muni-
lbl is a a.-. T

IQl_‘\-‘I| I”

FWMMI PL,

I:l_‘ll“§.'.::‘,’; l
casket dun} c
[Kiln-“Tl SKL‘I n:
‘ t
specmcn 3c";
.\&.

Rm”. SEM m;

Figure 27.

Figure 28.

Figure 29.

Figure 30.

Figure 31.

Figure 32.

Figure 33.

Figure 34.

Figure 35.

Figure 36.

Figure 37.

Figure 38.

Figure 39.

Schematic showing the high-speed substrate spinner

used in coating to the Mng substrates. ......................................... 62
Schematic of the refractory casket used for microwave joining.
The dead weight loading also is illustrated [2]. ................................. 65
Schematic of the spectrophotometer used for
the IR transmittance measurements. .............................................. 66
SEM micrographs of both the microwave and

the conventionally joined Mng showing the bond-layer.

(a) Microwave joined Mng (joining at 700°C for 20min

with 60gm weight loading), (b) Conventional joined Mng

(joining at 700°C for 20min with 85gm weight loading). .................... 70

SEM image of a Mngl Mng specimen joined using

the sodium silicate solution. The specimen was joined

by microwave heating at 700°C for 20min with a 60gm

dead weight loading. ............................................................... 71

Transmittance for both microwave and conventionally heated MgFg.
(a) Conventionally-heated Mng and MgFgl Mng joined and
(b) Microwave-heated Mng and Mng/ Mng joins. .......................... 73

Comparison of transmittance for microwave and conventionally joined
MgF2(From Figure 32). ............................................................ 74

As schematic showing the reflections at the
various interfaces of a two-layer specimen (which gives m = 3). ............ 75

(a) SEM micrograph of a cured silica film on a MaCorTM substrate,
(b) as a function of spin rate for cured silica coatings on MaCorTM [after 6]. 85

(a) Photograph of the zirconia/aluminosilicate refractory
casket during specimen joining, (b) Photograph of the
bottom SALI plate of the refractory casket with the specimen,

specimen setter and dead weights in place. ...................................... 86
SEM micrographs of alumina/zirconia composites

joined at 1450°C for 20 minutes showing interface. ........................... 89
SEM micro aphs of zirconia (3 mol % Y203)

with MaCor joined at 1020°C for 20 minutes showing interface. ......... 90

SEM micrograph of HAP with MaCorTM joined
at 1020°C for 10 minutes. The mica platelet rainforcement

In lht‘ MAW

lzllrt 40. SEM midi-u:
huh \lit‘ur ‘

lgurtil. SEM mlt'rng'
the tour mitt“ I

lit alumina.

figurell Procedure for ~

for notih :rwr‘

figural Schema; nt t:
shown: the n'.

Figure“. Schcmihc at It

 

 

FRUIT 45- Microfif‘d‘i urn \

tut

: 'l ~
with. St‘ilt’mallt' n" T"
“ i

F ..
M48 Sthfimdiit‘ n! 71‘

Sllmtmg H“ n.

Fm 49' A schemi'i ~

. 0
\i\ i].

Fm” 50- Nozch

‘ ..
\(‘njjik r

hm :
,re.l. .\l.iCOI'T\1lM.L

lttri 51

‘1‘

Figure 40.

Figure 41.

Figure 42.

Figure 43.

Figure 44.

Figure 45.

Figure 46.
Figure 47.

Figure 48.

Figure 49.
Figure 50.
Figure 51.
Figure 52.
Figure 53.
Figure 54.
Figure 55.
Figure 56.
Figure 57.

Figure 58.

in the MaCorTM is evident in the micrograph. ................................. 91

SEM micrograph of SiC platelet reinforced alumina

with MaCor joined at 1020°C for 10 minutes. .............................. 92
SEM micrographs of the fracture surfaces for
the four materials included in this study, namely
(a) alumina, (b) MaCorTM, (c) zirconia, and (d) hydroxyapatite ........... l 10
Procedure for sectioning the reference specimen

for notch geometry in the “pre-joined” state (after [1]). ..................... lll
Schematic of the microwave processing system,

showing the microwave cavity and the power supply (after [20]). ......... 112
Schematic of the refractory casket used during joining (after [1]). ......... 113
Microstructures of the joint regions of the joined HAP specimens. ........ 114
A schematic for compact powder materials and Sintering. ..................... 128
Schematic of producing notches. ................................................ 129
Schematic of microwave processing system,

showing microwave cavity and microwave power supply. .................. 130

A schematic of showing Vickers indentations placed near the bond layer. 131

Notch configuration in both HAP and MaCorTM specimens. ............... 132
MaCorTM notched specimen before and after joining. ....................... 133
HAP notched specimen before and after joining. ............................. 134
ZrOz notched specimen before and after joining. ............................. 135
2102 notched specimen before and after joining. ............................. 136
A1203/ ZrOz notched specimen before and after joining. ..................... 137
Schematic of the coating and joining technique. .............................. 145
Fracture surface of SiC used in this study. .................................... 146
(a) SiC joined at 900°C for 20 minutes in flowing N2

with a 20 gram deadweight applied during joining and

[bl 51:11:: ‘i‘ '
m :11ch .1

o \ “THIS"
[13111139. SE. 1 1 , '
heelt‘d (0113 L'.r

m mieruwt‘ '

 

llgllfl‘w S lit'm I‘UI ls \ln !
UxCilln ‘4??in

figure 61. ESE.“ C\.ii7.::
not added hex;
101 :0 X’ct‘iith .

Figure 61 The piilieihfixli
seizing ind t
la) the film wig".
intH grim hex.-
spun 2113M!) rp.’
«hi the him wig.
and 6 grim hem:
spun at SUI It.) mrt:

 

m Transmitlir‘ -

5 l

lturett. I

‘un\m1iiur lgc l

“M“ Tidi‘Slflj'th .3 -
dl 501C “1'." d

l
“”65. T'di"\lhlit.1"i“[

41600 C “Ill”: e
l“! .
litre 66. Schematic d: -. "
“56d 10 3],“ [3' .
lditf.’ Lee Cd\"

figurt67. SChDIT
7141;; xii“ 3..

h tilt. ref

lie \‘L\Ik, at
L: ‘7‘“
dis;

X'Sn “PM T?
3‘8

L \ Ernen' {Tn-it‘-

"den"

“31an S *1
mi

“'1: 1L 0‘ '>_
InLnls Stud: :( '

(b) MgF2 specimens joined at 700°C for 20 minutes
in air with a 60 gram deadweight applied during joining. .................. 147

Figure 59. SEM images of the SiC specimens that had been first

heated conventionally at 900°C (Figure 58a), followed by
(a) microwave heating at 1200°C for 20 minutes or by
(b) conventional heating at 1200°C for 20 minutes. ........................... 148

Figure 60. Schematic showing the hi gh-speed substrate spinner

used in applying the coatings to the MgF2 substrates. ........................ 157

Figure 61. ESEM examined polycarbosilane film which is

not added hexane after coating spun at 3000 rpm
for 20 seconds and cured at 100°C for 20 minutes in air. ..................... 158

Figure 62. The polycabosilane film thickness using ESEM after

coating and transmittance of coated specimens.

(a) the film solution of coated specimen was 1 gram PCS

and 4 gram hexane mixed solution. The coated MgF2 was

spun at 3000 rpm and cured at 400°C in N2,

(b) the film solution of coated specimen was 1 gram PCS

and 6 gram hexane mixed solution. The coated MgF2 was

spun at 5000 rpm and cured at 600°C in N 2,

(c) Transmittance for (a), and (d) Transmittance for (b). ..................... 159

Figure 63. Transmittance of both sides polished MgF2 specimen. ....................... 160

Figure 64. Transmittance of MgF2 coated with polycarbosilane

at 500°C with different spin rate. ................................................ 161

Figure 65. Transmittance of MgF2 coated with polycarbosilane

at 600°C with different spin rate. ............................................... 162

Figure 66. Schematic diagram of the microwave apparatus

used to sinter and to join the ceramic specimens
(after Lee, Case, and Asmussen, 1997). ........................................ 169

Figure 67. Schematic showing cross-sectional view of the microwave cavity,

with the refractory casket centered along the cavity axis.

The casket, composed of a hollow zirconia (ZYC) cylinder and

disc-shaped aluminosilicate (SALI) endplates, encloses a ceramic

specimen. The specimen temperature is measured by an optical

pyrometer ............................................................................. 170

Figure 68. Schematic of the insulating casket used

in this study for both microwave Sintering and microwave joining. ........ 186

xi

M" 69. Pht‘lt,tgf,;plj . w
used buth ti.
311d [010111 a:
The PW‘UIt'l'.

[eure'fli SEM mzerttg'
alumni - 137.
at i distem"

ltgtttt 71. A Tlihl 17.1;1.
the same 43...".
an SEM Inter

no Slinli‘

«'1'
i\u-'\

latte 72. A SEM mt. r.

-15 ‘? Imam.

 

flatten. The nntthed 35
defined in 11;:

‘ 1p-
” mkli ( \Yffid‘

l'ettrti-i. .1 SEM mtgtv.‘

»

FEM” "a. A TEX! I i‘lxri" '
for “high the“:
interface h\ :t

f 7
121m .6. A schemata i-t"
tntm dlilt‘?"

.411.“ of
km Sen 1“ t) W
hen
. R78 .
‘\ wmlmtiu‘ g
comp76§\l"e o}.

.1.

Shine

'-
u

d “mm "~

Pet
5 ”31.4 grdph m. '

t ‘1’“
Sim
mired dl I?”

Figure 69. Photograph of the single-mode, 2.45 GHz microwave cavity
used both to sinter the alumina, zirconia, alumina/zirconia powders
and to join all specimens included in this study.
The pyrometer and the insulating cavity (casket) also are shown. ......... 187

Figure 70. SEM micrograph of the bond region of a joined
alumina - zirconia specimen. Note the absence
of a discemable bond layer between the two joined specimens. ............ 188

Figure 71. A TEM micrograph of the bond region of
the same alumina — zirconia specimen depicted by
an SEM micrograph in Figure 70. In the TEM micrograph,
no significant bond layer is visible. ............................................. 189

Figure 72. A SEM micrograph of a microwave-sintered 85 wt % alumina
- 15 % zirconia specimen containing a “pressed-in” channel. ............. 190

Figure 73. The notched 85 wt % alumina — 15 wt% zirconia specimen
depicted in Figure 72, joined with a partially stabilized zirconia
(3 mol% yttria — zirconia) specimen. ........................................... 191

Figure 74. A SEM micrograph of the join between zirconia and MaCorTM. ......... 192

Figure 75. A TEM micrograph of the MaCorTM — zirconia join,
for which there appears to be extensive wettinngf the
interface by the glassy matrix phase of MaCor . ........................... 193

Figure 76. A schematic of the thermal stress induced
in two different joined materials upon cooling. ............................... 197

Figure 77. A schematic of different thermal stress distribution
between two materials that have been joined along a planar interface61].. 201

Figure 78. A continuous crack occurred in alumina due to
compressive thermal stress. The spacing between

alumina and interface was about 500 um and the crack
started from the interface [55]. ................................................ 202

Figure Bl. A graph of sintering TM-DAR (Alumina). TM-DAR was
sintered at 1300°C for 60 minutes with heating rate 10°C/min. ............ 226

Figure B2. A graph of sintering partially stabilized zirconia. Partially stabilized
zirconia was sintered at 1375°C for 60 minutes with heating rate

10°C/min ............................................................................ 227

Figure B3. A graph of sintering fully stabilized zirconia. Fully stabilized

xii

ltrt‘nnti um
l1] C'mm

figure Bi. A graph at \1'
Aluminum

ht‘diln‘.‘ [L135 1

figure BE. A griph (it he
at 'm C lot 2

@1136 A graph t‘lfl hit.
it HUI] C tnr.‘

FM 37. A graph or ‘m- .

Ilka

 

 

ztreomi mm;

_‘ i ‘ .‘I
115.71g 31.1.4 Etit‘.

“$111 38 A $111111 01' i‘tt‘a'

zmoma \xtth J

‘

.1 t
‘ \.. ..
5.22.4 tum “11‘.

1
figure 89. A graph ’01 he;
jomed .11 1'13“

figure Bl0. A

graph 01 he
ZlICOl‘.ld\\:1:*

Stltea ttlm W
“are
. 311. A graph of '5;

£1le:

we «21?:
Silteti ftim (a.

lr
574116 Hi 7.;
C()m:x’\\l[fo D

\\. i3

31 1450' C {1...

harem A

zirconia was sintered at 1350°C for 60 minutes with heating rate
10°C/min ............................................................................ 228

Figure B4. A graph of sintering alumina (85 wt %)-zirconia (15 wt %) composites.
Alumin/zirconia composites was sintered at 1350°C for 60 minutes with
heating rate 10°C/min. ........................................................... 229

Figure B5. A graph of heating rate of microwave joined MgF2/MgF2 joined
at 700°C for 20 minutes using sodium silicate solution spin-on interlayer. 230

Figure B6. A graph of heating rate of microwave joined AKP 30/AKP 30 joined
at 1400°C for 20 minutes using sodium a1 ginate spin-on interlayer ........ 231

Figure B7. A graph of heating rate of microwave joined partially stabilized
zirconia/partially stabilized zirconia joined at 1500°C for 20 minutes
using silica film spin-on interlayer. ............................................. 232

Figure B8. A graph of heating rate of microwave joined partially stabilized
zirconia with alumina joined at 1500°C for 20 minutes using
silica film spin-on interlayer. ................................................... 233

Figure B9. A graph of heating rate of microwave joined HAP with MaCorTM
joined at 1020°C for 20 minutes using silica film spin-on interlayer. ...... 234

Figure B10. A graph of heating rate of microwave joined partially stabilized
zirconia with MaCorTM joined at 1020°C for 20 minutes using
silica film spin-on interlayer. ................................................... 235

Figure B11. A graph of heating rate of microwave joined SiC platelet reinforced
alumina with /MaCorTM joined at 1020°C for 20 minutes using
silica film spin-on interlayer. ................................................... 236

Figure B12. A graph of heating rate of microwave joined alumina/zirconia
composites with alumina/zirconia composites joined
at 1450°C for 20 minutes using silica film spin-on interlayer. ............ 237

Figure B13. A graph of heating rate of microwave joined fully stabilized
zirconia with fully stabilized zirconia joined at 1450°C for 20
minutes using silica film spin-on interlayer. ................................. 238

Figure C1. ESEM micrographs of conventionally-joined AKP 3O (alumina)
joined at 1400°C for 20 minutes using sodium a1 ginate spin-on interlayer
(a) lower magnification of joined A1203 and
(b) higher magnification of joined A1203. .................................... 240

Figure C2. ESEM micrographs of microwave-joined AKP 30 (alumina) joined

xiii

a] 11W t‘?
111’ ltl'sA'Cl m

(bl htght‘t 7T.

llgttrt C3. tilt A mutt
1.: mtli "i l"
mtetlher \l

indeniittnn .

Figure C4. SE.\l mm
at lttltt (‘ tit-

llgttrt CS- SEM mung?
tor l0 mzhtte
lS endertt in t:

Figure C6 SE.\l mterngru.

jt‘lnt‘d it lttjtt

letteC'l. ESEM mien -:
jomed .it 151!)
tat lottcr m2:
ol'jmnetl A3,]

 

 

“MIX SEM mtetm

tr.“
\nh.
b

at ltjttl C 31.3.7.
til letter trim.
Oljt‘tmeu SEC .

at 1400°C for 20 minutes using sodium alginate spin-on interlayer
(a) lower magnification of joined A1203 and
(b) higher magnification of joined A1203. .................................... 241

Figure C3. (a) A micrograph of microwave-joined alumina with zirconia
(3 mol % Y2O3) at 1500°C for 20 minutes using silica spin-on
interlayer showing interface and (b) a crack induced byVikers
indentation was deflected from ZrO2 to A1203 through interface. ......... 242

Figure C4. SEM micrographs of zirconia (3 mol % Y2O3) with MaCorTM joined
at 1020°C for 20 minutes showing interface. ................................. 243

Figure C5. SEM micrograph of HAP with MaCorTM joined at 1020°C
for 10 minutes. The mica platelet rainforcement in the MaCorTM
is evident in the micrograph. .................................................... 244

Figure C6. SEM micrograph of SiC platelet reinforced alumina with MaCorTM
joined at 1020°C for 10 minutes. ................................................ 245

Figure C7. ESEM micrographs of microwave-joined ZrO2 (3 mol % Y203)
joined at 1500°C for 20 minutes using silica spin-on interlayer
(a) lower magnification of joined ZrO2 and (b) higher magnification
of joined A1203. .................................................................. 246

Figure C8. SEM micrographs of conventionally-joined SiC joined
at 1000°C for 20 minutes using Blackglass spin-on interlayer
(a) lower magnification of joined SiC and (b) higher magnification
of joined SiC. ..................................................................... 247

xiv

Ceramic mJiCi'ldi\ 't‘
rennet. ind prote. th "
:1-ireof'rom putt tiff". I
:rrgfrent. lhe hnitle rut
Sigsnents. 0tet the ldxt
Chit} of processing eer

Icteiszeg teehmeue to t .r

 

"“ 31 did-mt» 10min: ts ti.

-i.-.:r.-s stipes 0t ceramic .\

techniques that 1th.

..

5km! ‘ ° - - ‘
:. .thts. 1111817an tr

..,"7;J _
h

INTRODUCTION

Ceramic materials have many attractive properties in terms of high temperature
application and protection against chemical attack. Ceramic materials are typically
produced from powders, thus, a subsequent sintering process is required to form a dense
component. The brittle nature of ceramics leads to difficulties in obtaining net-shape
components. Over the last decade, many researchers have attempted to reduce the
difficulty of processing ceramic materials. For this reason, joining has been studied as a
processing technique to form a desired shape without additional processing. The main
idea of ceramic joining is to produce a complex shape from simpler subcomponents using
various shapes of ceramics such as bars, discs, tubes, and rods. There are many ceramic
joining techniques that reduce processing difficulty, save processing time, and reduce

energy costs. In terms of the joining processing techniques, microwave energy has been

utilized as the main heating source in this thesis.

thlol This Stud.‘

11! this thesis. .1 pw

 

tenses spmen llliCTid}t -
3751:; including .il'ir:

5.2) Slzf‘ililt‘d Itrt‘timi. ‘~
llf; ind SIC here _ltl'.l‘tt.
ire raked tor .1 range 0

without heitme.

 

ll Classification of (era
[15 he 181 th 0 demo.

‘3‘
.\

tex Ceramic mot:

l

*3 "'f: . a
L 31.4. .Ot pretexsmv q.

“37“” '-
7---i..1101[\“|“ lhc d" '
s. ‘1
Chi. tie/ceramic him

:1"?le
\ ,_. MES 01 . - .

TIE-ha]: "'4 I ~
““01“ ascnt. \t

it?

.3 V

1llttini
n .
g ec"mttuest

lili 44‘
. utiltul: {0 TNT”

:15...“

lrta—i. .
ltk! ldk‘c IT;

.‘h U . w
- LJT‘M

”‘1!

Goal of This Study

In this thesis, 3 possible ceramic/ceramic joining technique was proposed using
various spin-on interlayers to join oxide and non-oxide ceramic materials. Oxide
ceramics including alumina, alumina-zirconia composites, partially stabilized zirconia,
fully stabilized zirconia, MaCorTM and hydroxyapatite, and non-oxide ceramics including
MgF2 and SiC were joined to produce a smooth and thin bond layer. Low external loads
were applied for a range of 40 minutes to 60 minutes via either microwave heating or

conventional heating.

1.1 Classification of Ceramic Joining

In the last two decades, many researchers have developed various ceramic joining
techniques. Ceramic materials are mainly produced in a powder form, thus, sintering is
essential for processing ceramic materials. The ceramics joining process is very
important to form the desired various shapes from simpler components.

Ceramic/ceramic joining techniques are classified as using bonding agents including
metal brazes, glass powder, slurry, or spin-on materials. Diffusion bonding is performed
without a bond agent. Microwave joining is another joining technique using microwave

energy to join ceramic/ceramic materials as a heating method.

1-2 Joining Techniques Using Interface
It is difficult to form ceramics into complex shapes without making simpler sub-
components. Interface materials, which act as “glue’.’ to join ceramic materials, have

been widely used by many researchers [l-18]. Various types of interfaces have been

"iii ill CCIJmlt‘ {Chill};

1‘.
Ll“

3,1373%“. 13. ll].
112711516). 3”]
F151. htuzng the;
151-7'3 prtress. \"lf‘tt't’ .1 ]
:i’pi'ee' to the tern, ‘t.-.'
referral properties he:

:iiio'ttizei of thine hrs

I

 

 

9"?

33.55.0111 l‘tiulls 1n 1 .'

«A ' ' '
..»..~ biL 81.x; ‘1.\1\P\c:
”fl 1, l .,

.11 stresses LilOU'lLi

131:3; 17161.1“

'3‘:- t'flav r‘
“1...... 61.116 \\ .15 UFC

A

o
1

..

:sze and

\

if: v.3 ,3
-51 hrperiture [I].

V

- Is utthtn the t.

1"
. _..‘\_

‘rfir‘l. “91.» ‘
“m-Ttlk malt“

‘.
MAT

used in ceramic/ceramic joining such as metal brazes [1, 4, 8. 10, 11, 16-18], glass
materials [7, 12, 14], glass slurry [2, 3, 5, 6, 9, 13], glass powder [15]. or spin-on
materials [19, 20].

First, brazing (liquid-solid state joining) technique is currently the most typical
joining process, since a lower temperature (melting temperature of brazes) is required
compared to the temperatures required by other joining techniques. However, different
mechanical‘properties between the substrate and interface are one of the considerable
disadvantages of using brazes as an interface. A different CTE (Coefficient of Thermal
Expansion) results in thermal stress between the substrate and bond-layer. For example,
nickel filler braze was used in joining of ZrO2 with Si3N4. The different CTE between
nickel and Si3N4 (Nickel: 13.6 x10'°K" and 313m; 3 x10’°K") induced the highest
residual stresses around the Ni-Si3N4 interface [1]. Thus, thermal expansion differences
between metal braze and ceramic substrate induced thermal stress upon cooling from the
bonding temperature [2]. The brazing technique has disadvantages for joining
ceramic/ceramic materials; (1) high costs, since a large-scale vacuum furnace is needed,
(2) oxides within the bond-layer can be a problem during heating, (3) it is difficult to
produce a hole or channel during brazing, and (4) the bond-layer is much thicker and
could induce a thermal mismatch through the interface causing poor mechanical
properties [16, 23].

The brazing is performed by filling braze filler metal between two ceramic
substrates and then heating to the melting temperature with or without external pressure.
A braze filler metal alloy is chosen depending upon wettability (Figure 1) between the

substrate material and metal braze. Additional components can improve the wettability

‘13 11.271 rescaleho h-

:lri‘}. 1nd Tl-Sn .tliut
than. Also. 4111111; ..
11:211. ll],

To: 5:111:11. tenth".
13.1713). \Vt‘liéi‘iilih .'

"i . g? .0 , .l ‘ I
-[ r..i..\1&\ ‘1?th mt“ ccr‘.

"f'r. ifuwhv- w H " '
.1. we: 1 as C‘t‘t'."
&

 

 

..
I

9:51 : t
t... :1. h the intertdt'1c.‘

leteen $01111 .tild \dT“

wi- .,
"f5i1n1451‘1r8

y

1
1A
A\.

i '1
..I

.I'
I
3111:».
W:
iihu
1. '—. '
{Intt't .
\..
Lh'l
1.4" ’1'
H.1L' l‘t
., . .
— 116531
liki\,\~

[4]. Many researches have developed brazes depending upon the wettability between
substrate material and braze filler metal such as Ag-Cu alloy brazes [4, 18], aluminum
[16,17], and Ti-Sn alloy braze [10] since bond strength improves with improved wetting
behavior. Also, adding a third element can reduce the melting temperature of braze filler
metal [10, 11].

For brazing, contact angle and interfacial energy states are directly connected with
wettability. Wettability has been evaluated from the sessile drop technique [21,22] where
the contact angle between liquid and solid phase is measured (Figure 1). A spreading
theory of wetting has been developed by Harkins [21].

7.. = rev —m 0089 (l)
where 75L is the interfacial energy between solid and liquid, ygv is the interfacial energy
between solid and vapour, fly is the interfacial energy between liquid and vapour, and 0
is the contact angle.

When ygL is larger than st the value of the contact angle 0 is greater than 90°, no
spreading occurs (Figure 1(b)). On the other hand, “when YSL is lower than st (contact
angle 0 < 90°), the solid surface will be covered by liquid (Figure l(c))” [21], since liquid
in contact with a solid reduces the surface energy of the solid [21, 22]. Thus, in order to
obtain wetting and spreading, the following equation applies:

YSL < st - YLV C050 (2)
Equation (1) can also be written for the case of brazing a metal to a ceramic
y, = ym c039 + yr," (3)
where yo," is the interfacial energy between ceramic and metal, ym is the surface energy of

metal, and ye is the surface energy of the ceramic.

~- '. . l .,

lt the lntemtli. t!

I i 6 .
[iii llltl not \j‘Ttdu 11.
”M“ ‘1'“ ‘l 1"!" \l'
wafsklnuitk J.-.\ .

5

9:5 sitiier illul‘. W .m:

3‘ "w l ‘
...\..tethiten.m After

a...

I
1 b.
. 2
kn. HiSl

 

 

 

 

 

 

. A", o, 9'.

.11. .td?\;1ll'1'.‘:t‘tii
t'sC
. firfipna '
*KU-u IE ICLllnikl'Jc i‘ ‘2'
R71?“

1w.“ ~
4.1 -151. {ml‘chlure

\ .

1£,,__ ,
tyfi"...1 "‘5
5k Lnrfl‘ ya x

9.11

If the interfacial energy van is larger than the ceramic surface energy 7... then the

liquid will not spread freely or flow into capillary gaps. In this case c050 will be negative
and the contact angle will be greater than 90°. On the other hand, when the contact angle
0 is smaller than 90° and y,.,,, is smaller than ye, then wetting is possible on the ceramic
surface.

Using a glass powder or slurry are other techniques for ceramic/ceramic joining.
The bond agent is selected upon the basis of having nearly the same properties of the
substrate materials. After heating at a desirable temperature, interface materials pyrolize
and then are transformed into amorphous ceramics. For high temperature applications,
the brazing technique is limited due to using low melting temperature metal brazes.
Brazing can be used in high temperature applications using glass ceramic as a bond agent
due to high temperature endurance of the bond agent. Using glass ceramic has certain
advantages over using metal brazes since (1) chemical compatibility with substrate
material is generally assured and (2) some physical properties such as viscosity, flow
properties, and melting characteristics of glasses can be controlled over a wide range [3].
For example, sialon was joined using a Y-sialon slurry [2] and Si3N4 was joined using a
Y2O3-SiO2—Al2O3-Si3N4 slurry [5]. Using a glass powder or slurry as a bonding agent
reduced differences in CTE and elastic modulus at higher temperatures [2, 5]. Also, a
MAS filler (MgO (99%, 0.5 pm), A1203 (99.99%, 0.2 um), SiO2 (99.9%, 0.8 pm» has

been calcined to produce a slurry with almost the same CTE as SiC substrate [12].

 

 

 

F.» .2
n: l_ 3;.
'I a

I) _ I-
«mutt tt' .-

.fflrp...
\‘ “1“] a1“?!
mm” M 5163171] If}?

Vapour Vapour

 

 

 

[Aqukl Liqukl
My
0
21 DP
Mn 'hv
Sohd Sohd
(a) (b)
IAqukl
*
Sofid
(C)

Figure 1. Schematic illustration of sessile drop formation: (a) principle of wetting, where
the contact angle and interfacial energy are noted, (b) non-wetting drop, and (c) complete
wetting [21].

13 luinlng Techniquo
Another cert-ml. 1.

9.4: lending technlqit‘

'1': the 21m factors for ,.

33.111011“? ‘1‘“

 

foe.“ belt een lnter'. .1.

11:; temperature e'
13.22607 01' the \uhc-
itspie 051 tiltiusrt'lr. h.

213% p.» w -
.u '.Ld\.tlL\n‘bl°nLiL'Li \

 

l at. m sttficttt‘d 1'“ '

:23: {24].

lt-w n "
tinné 03 dtxslmtid’

an»? ... . .
3:: grid n1:- hontilr "

l

f

A
u‘\
x

(I

.151“ "'
....\-§mrn\tere lt‘

‘.:

"rfl" u r '-
1..)511I60i I] (11

I l

11.5 about (I) '} tt

&.
1'.

‘13P;

L . ..
“9&3an 136701“
. \1
”his-«t
«AC-k ‘f‘erh
. .eitmn -
- 5 41 1
:CLifl . 1-
“OT-Al and \l .3-

x‘“ '

[inhu'fl \
.w._l-
:- . LT? '
“til 1"?-
“l
'Qf: in ..
eleme-
b ‘ 1hrcc p 1"-

tth.

. H ,

‘3

1.3 Joining Techniques Without Using a Bond Agent

Another ceramic/ceramic joining technique is diffusion bonding which is a solid
state bonding technique. A bonding agent is not applied in this case. Heat and pressure
are the main factors for using diffusion bonding. The mechanism of diffusion bonding
can depend on the viscous flow of the glassy grain-boundary phase, which can be
diffused between interface [24].

High temperature and pressure are required to join ceramic/ceramic materials since
at least 60 % of the substrate melting temperature is needed for significant diffusion. An
example of a diffusion bonding mechanism is shown in Figure 2. The substrate was
RBSC (reaction-bonded silicon carbide). As pressure and temperature were applied, the
interface was softened by the glassy phase due to free silicon migration across the
interface [24].

Joining of dissimilar ceramics was reported by Sato et al. [26] using microwave
energy and no bonding agents. Alumina and magnesia rods with the dimension of 5 mm
X 5 mm X 25 mm were joined for temperature ranges of 1577°C to 1877°C using an
applied pressure of 0.03 - 0.5 MPa for 2-10 minutes. The 90 MPa of joint strength,
which was about 60 % of the MgO bending strength (150 MPa) and 70 % for that of
MgO specimen bending strength (130 MPa) heat treated at 1877°C for 4 minutes, was
observed after heating at 1877°C for 4 minutes under 5 MPa applied pressure. Due to
diffusion of Al3+ and Mg2+ across the joining plane, MgA12O4 formed at the interface
after joining. A crack ran along the interface between the MgAl2O4 phase and the MgO

phase during the three point bend test due to lower bending strength than alumina (340

MPa).

Force and

 

1 Heating
Free /‘:\0066
silicon /E
or—Si

(a)
Force and
1 Heating

£3365

 

0))
Force and
1 Heating

Decent

(C)

Figure 2. A schematic illustrating diffusion bonding of free silicon onto a butting
surface. (a) Migration of free silicon onto surface of the RBSC, (b) As force (pressure) is
applied, the silicon spreads across the RBSC surface, (0) Free silicon forms a continuous
interlayer [24].

 

57.),- lltlnt title 10 the 11‘.
}: height liter hertz;-

:itirence h nut ldCdl a.

ll llicmllale Joining
Alumnae ener '

tires ihsorh interntl ._

 

131““) F77“ ‘ .
.vu‘.‘.\' liurc.‘i.1lll(\l
s

 

 

7‘ a r
..-'=.~ lnll. At rwn‘. :e

617') g. - .r \ u
with... .15le {HUSH-m l]
S .

la. ‘
.‘Hao,
‘.

6‘
a: .nrn the nude

him [656‘]? ‘11,.“ in

Aki!\

Lhflsvl‘ t.‘ .

. .stghhelm 1.17"”

ltethetelittl ti ‘
iL'x

Applied pressures (20 MPa and 21 MPa) induced a large creep deformation in the
Si3N4 joint due to the high pressure [27, 29]. The specimen dimension changed about 1
% in height after heating at 1500°C with applied 21 MPa for 1 hour. More than a 1 %

difference is not ideal accuracy to form net-shape joining [27].

1.4 Microwave Joining

Microwave energy has been used to process ceramic materials since ceramic
materials absorb microwave energy and heating can be generated from the inside to the
outside (Figure 3 (a)). Conventional heating is generated from the outside to the inside
(Figure 3 (b)). At room temperature, alumina is nearly transparent to microwave energy,
however, above roughly 1000°C, alumina absorbs microwave energy (Figure 4) and starts
heating from the inside.

Many researchers have demonstrated the advantages of using microwave energy in
ceramic processing. A key consideration when using microwave energy in ceramic
processing is the loss tangent (tan 5), defined as the ratio of the effective relative loss

factor to the relative dielectric constant, as given in equation (4) [31].

E. e" or
tandz—flf—=—.+ .
E 8 we 8

r r f 0

 

(4)

where 8” = dielectric loss factor, 80 = permittivity of free space, 8’fo = effective relative
dielectric loss factor, 6 = total effective conductivity caused by conduction

and displacement currents, (u = frequency (219‘), and 8’, 2 relative dielectric constant.

External Heating Source

 

 

 

Specimen
- ~ 1 e ..
I 27> <
i A i
Furnace
Insulation
(a)

Intern a1 H eating

Microwave port

 

 

 

/ \\
/ V \ \\
¥/' Specimen \
\wa/ / \
, . t e . \‘
KLV i i .12,“ //
A Insulation /

/-\
":i/

\

 

 

/

l” t

’\

A/

 

 

/¥""
/

 

 

 

 

 

 

 

Metal shell

(b)
Figure 3. Patterns of heating in (a) conventional furnace and (b) microwave (adapted
from Sutton [31] and Lee [48]).

10

 

 

R

”‘5‘

.4324
5 ~ .

Po.

‘llcraCtll-ln - I

..W 'a _, U] n-

i-mxl', 1 ill “
‘T l

Material Type Penetration

 

TRANSPARENT Total

 

 

 

(Low loss insulator)

 

 

 

 

 

 

 

 

 

 

OPAQL'E None
(Conductor)
M ABSORBER Partial to Total
(Lossy insulator)
°l . \ o; 4. 3 ‘- ABSORBER Partial to Total

2

°\ °"%‘- 0’ 50 (Mixed

Matrix = low loss insulator
Fiber/particles/additives = absorbing
materials

 

 

 

Figure 4. Interaction of microwaves with materials at ambient temperature (adopted from
Sutton [31] and Lee [48]).

ll

0.010

5)

0.008

ngcnt ( t

_O
C)
CD
(35

F F F

I

"i 0.004

0.002

hens -_
i 1.0331

 

 

/\

‘fi-w-o‘; ,
d“;~\nt ’\

Temperature (° C)

 

200 g 600 1000 1400
I I l ot presse 97 A) pure
0.010 — BN A1203
5; Pyrolytic Si,N4 ’E .
g 0.008 —— f, 310,
9". .
8’0 0 006 _9606 (Glass - Ceramic) ’ ‘ 00 pure
g ' ’ 2 3
3 0.004 4

0.002

 

 

 

 

 

Figure 5. Loss tangent (8 to 10 GHz) versus temperature [31].

12

Comm: mater:

 

e:per.:ure mes .zbm -.
rediyrex‘uizmg In J
Leas: to more effect: a ' '
;-":?_’.e::ng at a frequent
from 1 mm In 1-
irastna! apphwtmn. “3.

In general. {he loss

 

:1? the temperature h.-

"‘ 74301} Mb the ad.

'I

“.524 tan 0 Ms 1N

Tne potential .10.:

.A I I
A}: f‘3 .7 ‘

{f‘fi'fl
- t ‘Mthm the m

3.

. Jud unce hedt-m

’tn.
. ":4
~41...“ [0 '
jOm term

”10

;.4':."~L

"~k”. p,
45431613011“

"1 «1.

C10"!-
.u.'.\ ind Sum",1 [a
l l“-

5‘

....t..“~f,‘ R hen [ha .
‘ \

.‘K
‘5.
V‘ . TV
t-‘Ae
yerarl ..
l t ' 0
rTm-z.
3““
“ ' 1
“551.1?
. .vq'a

Ceramic materials are transparent to microwave radiation. However, as the
temperature rises above a certain temperature (critical temperature, Tm), tan 5 rises
rapidly, resulting in a greater absorption of the microwave energy by the materials, which
leads to more effective heating. Generally. microwaves are electromagnetic waves
oscillating at a frequency between 0.3 and 300 GHz which correspond to wavelengths
ranging from 1 mm to 1m. Magnetrons (microwave power supplies) commonly used in
industrial applications operate at 2.45 GHz [32].

In general, the loss tangent (tan 8) initially rises slowly with increasing temperature
until the temperature has reached a critical temperature (Tm). The loss tangent rises
more rapidly with the addition of compositional additives and impurites. Thus, 99% pure
alumina tan 5 was less susceptable to heating than that of 92% or 97% pure alumina
(Figure 5)[31, 32].

The potential advantages of using microwave energy are; (l) a cost reduction is
possible since less energy is required, (2) uniform heating is possible since heating is
generated within the material, (3) a material’s microstructure and properties can be
improved since heating is more efficient within ceramic material, and (4) less time is
required to join ceramic materials using microwave energy since less energy is required
to reach a same temperature compared to conventional joining technique [31, 32, 33].

Clark and Sutton [31] described the microwave joining process as “joining is most
effective when the ceramic sufficiently absorbs microwaves and the grain boundaries
contain phases that soften and ‘glue’ the surfaces of adjacent materials”. As mentioned
in Figure 4, ceramic materials can absorb microwave energy efficiently as the

temperature increases.

13

 

Alumina ~pet‘zme
tget: Lil temperatures :
in; .1 single mode re.‘
trim; HS done lflx l0:~
wed more et't'tgzertt

B:r:rteret 41. [W2 ‘.'
5.: '7 dammit. The jm‘

1.3:: when heited .0 tlx'

\.

 

 

.-:.*.;:r.son.alumtn tux".
h", 4: ~.‘
etc Alulfl Ol .06 sneetmer

gee.) dliiflbultd. ll“ ‘

an} to heat the ma'e‘

Selecttte and is»?

it: ..v
Fe. alumina'Ztreur

T‘jrig‘ .
4...; §\l. ‘
~ -.n.e the mate"
.41.

ex-
rz.‘x_ ‘
, Dtmg micro“ a‘

Fae ° '
lltt reduction or

it» -
{ATOW d\ 63 hd\ ;
\
-‘- ,titned qt

LI:
\ . ..
“BSCHU‘EK
‘ ~' U?-
\
5.,1‘de,‘

Alumina specimens with purity of 92 % to96 % were joined without a bonding
agent at temperatures above 1750°C with 0.6 MPa applied pressure for 3 minutes in air
using a single mode rectangular microwave cavity [33]. The most successful microwave
joining was done by lower purity alumina, indicating that the microwave energy was
absorbed more efficiently for lower purity alumina.

Binner et al. [19] investigated microwave joining via diffusion bonding of 94 % and
85 % alumina. The joined specimens showed the almost invisible layer (very thin bond
layer) when heated at the highest temperature (16000C) and pressure (0.25 MPa). In
comparison, alumina joined at 1400°C alumina clearly showed a visible bond layer across
the width of the specimen and regions of contact between the opposing faces were
sparsely distributed. High purity alumina 99.8 ‘70 was not successfully joined due to the
inability to heat the material to a sufficiently high temperature.

Selective and fast heating using microwave energy suggested by Aravindan et al.
[13]. For alumina/zirconia composites, temperature could reach at 1000°C within 2
minutes since the material absorbed microwave energy very quickly via the amorphous
phase. During microwave joining zirconia transformed from tetragonal to the monoclinic
phase. The reduction of hardness at the interface was due to Al-Zr inter-diffusion [l3].

Microwaves have been used for other various joining techniques. Katz et al. [34,
35] joined silicon carbide using three different microwave methods: ( 1) Reaction bonded
SiC (RBSC) tube specimens with an outer diameter of 3.5 cm, an inner diameter of 2.54
Cm, and a length of 2.54 cm were joined with temperatures range of 1410°C, l420°C,
1465°C, 1515°C, and 1565°C using an applied load of 0.23 MPa in a 2.45 GHz

rectangular multimode cavity. No adhesive was used between the specimens, making

14

its an eumple 0t mt;
handing. '.tht.‘h regain.
travail) done it lil‘
lt‘t7zdeEC SiC spe.'trr.. :
tee arterial tn the 4 p
:Frpistta dais. 3.54 .
Man-l ABA” Jill‘x e ~

firsttmtelt W." C M

 

, .9, . .
13.. . _._~.t mttrtm .1‘~ e

T” ’\

...r.;}e.'0l braze allux t
art. Hemmer. the tr;

:iiftrted SiC 731C Ct tt‘: ”

the: amt3 a

perhmerp

141:3 :0 SlC p0“ do

iittl

,"7 l .
l-Féftti‘l’ l“

hand [34]. T

zine“ t
,.he mleTCM a‘\ 6

wheel-
.tLq I]

erred in a try-

this an example of microwave diffusion bonding. Compared to conventional diffusion
bonding, which requires 2000°C and a long time (1 hour), SiC/SiC joining was
successfully done within 5 minutes at temperature between l420°C and 15150C. The
joined SiC/SiC specimens showed equal or greater fracture toughness than that of the
base material in the 4 point bend test. (2) Continuous fiber-reinforced SiC/SiC
composites disks, 2.54 cm in diameter and approximately 0.32 cm thick, were joined with
Cusin-l ABATM active brazing alloy (2.54 cm diameter and about 0.005 cm thick) at
approximately 900°C for 30 minutes in 95% N2 and 25% H2 atmosphere using a single
mode (TE103) microwave cavity. The specimens showed a homogenous metallic
interlayer of braze alloy that wet the ceramic composites and spread into pores in the SiC
matrix. However, the mechanical behavior was not reported. (3) Continuous fiber-
reinforced SiC/SiC composites plates of dimension 1.27 cm X 1.27 cm X 0.0.32 cm were
joined using a polymer precursor (allyhydridropolycarbosilane (AHPCS) which was
added to SiC powder to form a slurry). Prior to joining, the specimens were squeezed
together by hand [34]. The refractory insulator was heated to lOOOOC for one to 1.5 hours
and then the microwave cavity was evacuated to 6X103 psi with argon. The specimen
was then heated in a microwave cavity at 1100°C for 30 minutes. Occasional porosity
was observed throughout the interface. However, the interface was invisible and smooth
but mechanical testing was not reported [34].

Many researchers have used microwave power as a heating source due to its
selective heating of particular materials since each ceramic material has different

microwave absorption ability. Microwave energy offers the potential of a new

15

:r;;.lt;'.‘erimt; ttttrtt'

.t '1 H
‘ ’7'sal‘tctvll‘lt'ai'

.14.». rut

 

 

ceramic/ceramic joining processes because of absorption characteristics in ceramic

materials and fast heating ability.

16

3 [XPERNENTi

2.1 Materials 1x111 in
Cttmmeretal p.
3:14:13} Sial‘llllt‘d In.

terms for this. thes'

.‘sed in this thesis.
11}ditl\}:palllt‘ t

fiftfisifilt‘llt'tn of dam;

 

’1.-.~Q\ ‘ .. .
.Jalvé 92.9588 due In?

"1|ng 4"
"h :‘Uk‘\\dpd'11€ hub;
. 5 . 1

“A? '1 ’1. 1’1 I V
....t_..p...te 15Cldxxffi

111"" "l ‘ |
-605} thermcal ."w

111 0

1, «I .-
klfius 0i [1'

3r. . ~
.0. ltrettmat and . ‘

-

-\1-JC 01’TM . a N

,1.
it'uft‘tz
I

r4 t
“\l_& Ill, ‘ 1"
eMCnL'm :7.fo

Iv”
-\1_
W.

5111!“
' , ~ 1‘11 ”Hr
- ‘uii‘p 11,1 .12“
aft...
‘§t\ .‘ {j ‘
'thfk‘lSltm 1" .—

31:3”;
H‘kd,‘
‘ I it \

PG

“34:5ch 1
r U "r.
Hi, and]
UL \‘-"\‘
s,” r
.‘sch'1‘e1
TWICE?“

“3 ”l the rr~

2 EXPERIMENTAL PROCEDURE

2.1 Materials Used in This Thesis

Commercial powders of alumina (Table 1), fully stabilized zirconia (Table 3),
partially stabilized zirconia (Table 3), and hyroxyapitite (Table 2) were used to fabricate
specimens for this thesis. As—received MaCorTM (Table 4) and MgF2 (Table 5) were also
used in this thesis.

Hydroxyapatite (HAP) is a known bio-ceramic used for replacement, repair, or
reconstruction of damaged human bodies. Hydroxyapatite and HAP composites are good
bioactive glasses due to their resemblance to human bone. When human blood flows
over hydroxyapatite, hydroxyapatite acts like human bone. Polycrystalline
hydroxyapatite is classified in bio-ceramics as bioactive fixation meaning that it can be
attached by chemical bonding directly to bone [36].

Two kinds of zirconia powder were used: (1) partially stabilized zirconia (3 mol
% Y203 zirconia) and (2) fully stabilized zirconia (8 mol % Y203 zirconia).

MaCorTM, a flurophlogopite mica platelet reinforced glass ceramic in the system -
SiOz-Ale3-MgO-KZO-F was also used for this thesis. The properties of MaCor TM
include (1) excellent physical properties, (2) high dielectric strength, (3) good electrical
resistivity, (4) non-porous and non-shrinking, and (5) machinable into complicated
shapes and precision parts using ordinary metal working tools. MaCorTM is used as an
electrical or thermal insulator, for structural components, in semi-conductor processing,
and in electrical and opto-electronic equipment in industrial fields. Randomly oriented

mica platelets in the microstructure of MaCorTM are the key to its machinability [43].

17

1.03:1 Alumina t'l.\f

:tef't'tetent [38]. and \.
-~——".

(1'.

L01

(‘09..

SLYa.1

 

 

Aterige ~
Pt.’

117‘“
1.1 .

 

\- i

!‘ A. \
(JYLK n x.

x“

1:"; -‘ ~ I
heettenstt} utter he-
h.
Trier-"trial C\Pdn.\lt in

\

hint

liturtg’x r

 

Easel derottmtttt-
siettzctent [ll]

Dixnptton

PM... «0415030er

8311130’2 1 '

*s.g. hi“ ’ 1
y r“ \ dnda‘\\‘l‘

 

Table l. Alumina (TM-DAR, Taimi Co. Japan) powder information. Purity [37], thermal
coefficient [38], and Young’s moduli and thermal conductivity[39]

 

 

 

 

 

 

 

Grade TM-DAR
Lot No 2831
Crystal form a
Surface area 13.6 mZ/g
Average surface size 0.21 um
Purity 99.99 %
Impurity Na (3 ppm), K (1 ppm), Fe (6 ppm), Ca

(lppm), Mg (1 ppm), and Si (6ppm)

 

 

 

 

Green density 2.32 g/cm3
Fired density after heating at 1350°C for 1 3.96 g/cm3
hour
Thermal expansion coefficient (25°C- 8.8 X 10'6 K'I
lOOO°C)
Young’s moduli 390 GPa

 

 

Thermal conductive

 

30.0—35.0 W/m -°C

 

Table 2. Hydroxyapatite powder material information [40] and thermal expansion

coefficient [41]

 

 

 

 

 

 

Description Calcium hydroxylapatite
Ca10(OH)2(PO4)6, Powder, 99 %
Specific analysis or property Test for Found Theoretical
Ca 37.28 % 37.00 %
Spectrographic analysis Element Result (wt %)
A] < 0.01
Cr < 0.01
Fe < 0.01
Mg 0.08
Mn < 0.01
Na < 0.01
Si 0.08
Sr < 0.01
Thermal expansion coefficient
25°C - 400°C 14.5 x 10°K"
400°C - 800°C 15.5 x 10'6K"

 

 

 

18

 

135133 31416041 tr.'.

/—

llaenrls

 

['(l‘tmol ‘it
[30:11“ (11
Ntlltttt ‘1!

specific surface Are-
\

 

 

3:103: Strenzth R l
Fracture TtlUgllIlC\\ R
111% mUSl
H:Cress«H\' 11h

lla-

19.111.1ch diam-"-

 

5:. 1 ~ '
I...» lntorrttatzon 01 3

"‘yr '1'
194435th

\
lengerature Ltr
Dtezecme Cary.

Deleetne Stren-

   

‘ Pttt

”V '—
"$0115 Rdlitl

Table 3. Material information of two kinds of zirconia [42]

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Materials TZ-3 YS (Partially TZ-8 YS (Fully
stabilized) stabilized)

Y203 (mol %) 3 % 8 %

Y203 (wt %) 5.15 i 0.20 13.3 i 0.60

A1203 (Wt (70) .<. 0.1 S 0.1

SiOz (wt %) S 0.02 S 0.02

FeOz (wt %) S 0.01 S 0.01

NaZO (wt %) S 0.04 0.12

Specific surface Area (mZ/g) 7 i 2 7 1r 2

Density (g/cm‘) 6.05 5.90

Bending Strength R.T. (MPa) 1200 300

Fracture Toughness R.T. 5.0 1.5

(MPa mos)

Hardness (HV 10) 1250 1250

Mean Particle diameter (pm) 0.59 0.54

 

Table 4 Information of MaCorTM [43] (Thermal expansion coefficient and Poisson’s

ratio [56])

 

 

 

Temperature Limit 1000° C
Dielectric Constant 1Khz at 25° C 6.03
Dielectric Strength 40KV/mm

 

DC Volume Resistivity

at 25° C >1016 ohm-cm

 

Thermal Conductivity

at 25° C 1.46 W/m° C

 

Thermal expansion coefficient

 

 

 

 

 

25-300°C 9.3 /K x 10°
25-600°C 11.4/K x 10°
25-800°C 12.6/K x 10°
Density 2.52 g/cm3
Modulus of Elasticity at 25° C 64 GPa
Porosity 0

Poisson's Ratio 0.29

 

 

19

 

 

Lie 5. Material 11111

1‘."‘.‘.-....;

Other r15:
llemiCor-duti '
Hzghttiteiehzfr '
d<t~rrtrtt

High Marni-l
d<hmm

Densttt t\_' .\ ~

lleltm: writ

 

 

Hardness IKTIHID'
l’eml etninxtnn .
, $32111: heat eupaetti
11 ‘
113030111th \

 

 

 

    

 

Table 5. Material information of MgF2 [44].

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Formula MgF2
Other name Itran I, sellaite
Thermal Conduct'y mW/cm- K ~150
High wavelength limit, 11.
8
d<6mm
High wavelength limit, [1.
6.5
d<6mm
Strength (FaEL,FAEL) MPa 49
Density (g/cm3) 3.18
Melting point (°C) 1255
Hardness (Knoop) kgf/mm2 415 (575)
Thermal expansion coefficient 11
Specific heat capacity (J kg-l K— 920
1)
H2O Solubility wt.% 0.0076
Refractive Index 134
n at 5pm

 

Table 6. The spin-on materials used for ceramic/ceramic joining.

 

 

 

 

 

 

 

 

 

Material Spin-on Spin-rate Curing
TM Material
ZrOz, MaCor and . .
A1203Zr02 composites Silica film 3000 rpm Yes
A1203 (AKP 30) Sodium alginate 3000 rpm Yes
SiC BlackglasTIW 3000 rpm No
MgF2 Sodium silicate 3000 rpm No

 

20

 

 

 
  

HgF; 1111.111 1. ;
111511116111 grim sti. -
1151; tnrtsmtts 11111.11.

11 the them. tl
11:30:10.5 511 1111- ft Int
\L'" alginate.

11165115401131
lazfnpttrtt} lC\Cl.\ 1e...

3.3.1! '
15.521.511111111‘11114111‘11. _

4L... .‘.. F
1511:1111) annealtr

(V
1O. "
\

81.11111.” 12‘.

51 After spinning at “

 

N1 7“

tired. Blatltm is p'

.150. .15 it into rel..~t='

1... A
0411?." ‘ 1" V '
\p‘anh £0114!" 51 \k N

W: matnt in CW.

5111mm 119m

\.

13391;“
{h} pi)“ d’CI mt:

kit». -
3471301 Sfldjg

:III’Ii‘i
“jut. f. .
:tlldtt? are utr”/
:..‘1 .C

3099.0.6
‘ 1"“! I“ S. " v
5'41 urn .1..

“Ti-1".“ (
"hi-17,7 . .
I lldgcufq .
.lLt‘

MgF2 (Itran I, Eastman Kodak Company) is a hot pressed polycrystalline material
with a mean grain size of approximately 2.8um determined by line intercept method.
MgF2 transmits infrared ration over the wavelength range of 1.5 pm to 9 pm (Table 5).

In this thesis, the following four materials were used as interlayer materials, (1) an
amorphous silica-film, (2) an amorphous BlackglasTM, (3) sodium silicate, and (4)
sodium alginate.

The silica-filmTM(Emulsitone Company, Whippany, NJ) used as spin-on material
has impurity levels less than one part per million of metallic ions from Emulsitone
President information. The silica film is an alcohol based liquid material and can be
densified by annealing at 200°C for 20 minutes.

BlackglasTM (2vol %) mixed with a catalyst (98 vol %) was utilized for joining of
SiC. After spinning at 3000 rpm with BlackglasTM, the coated SiC specimen was not
cured. BlackglasTM is preceramic polymer and pyrolizing it the 800-10000C range
transforms it into relatively high yielding in system of Si-O-C ceramics. BlackglasTM
mixed with catalyst was heated and then formed into ceramics. BlackglasTM was used for
a fiber matrix in CMC’s (ceramic matrix composites) [45].

Sodium alginate (Alginic Acid Sodium Salt, Sigma Chemical Co, St. Louis, MO)
is a glass powder material. Alginic acid is water soluble, thus, adding 100 ml of water
per 3 grams of sodium alginate controlled the viscosity of glass powder. The properties
of alginate are utilized in the thickening, stabilizing, gel-forming, and film-forming
properties. Sodium alginate is used for application in textiles, welding electrodes,

ceramics, pharmaceuticals, cosmetics, etc [47].

St‘dtum silicate ,1

 

e111 :rliter to [0111 \
urination 01 111110
team; $0111qu

merete [46].

 

12(01d1’ressing Ur

Commercial pm $1
cehteszte. and HAP 1*
lie has of the pets 0.
Miller Instrument C
Met compacts 11 e1
erotmtel) 32 Ml

. g 1 ‘
F1213]:

011511 liar-1:

,
Jl’reparation of l

a- .. _
RL Meter“ ‘1

rf

 

 

Sodium silicate (Colmbus Chemical Industries INC, Coumbus, WI) was used as spin-
on interlayer to join MgF2. Sodium silicate is called ‘water glass’ and its system is a
combination of almost equal proportions of silica and soda. Adding water can control

viscosity. Sodium silicate is used as a deflocculent in ceramics and as a hardener in

concrete [46].

2.2 Cold Pressing Ceramic Powder Compacts

Commercial powder ceramic materials, alumina, zirconia, alumina-zirconia
composite, and HAP were pressed into powder compacts prior to microwave sintering.
The mass of the powder to be pressed was first measured using an electronic balance
(Mettler Instrument Corporation) with an accuracy of i 0.001 gm. The green unfired
powder compacts were about 2.2 cm in diameter and 0.2 cm thick after cold pressing at
approximately 32 MPa (Figure 6 (a)). Up to seven powder compacts were sintered

simultaneously using a 10.3 cm diameter refractory casket (Figure 6 (b)).

2.3 Preparation of Refractory Insulator Used in This Thesis

Refractory insulators were used during both microwave sintering and joining.
Two types of cylindrical caskets were used, one 10.3 cm in diameter and the other 7.5 cm
in diameter. The refractory caskets both (1) reduced heat loss and (2) provided
microwave susceptors to allow heating at low temperatures. The wall thicknesses of the
Cylindrical refractories ZYC(Zirconia) were 1.3 cm for all caskets. The cylindrical

refractories were composed of zirconia ZYC about 40 % densified.

5’ t
I r.
I'-

 

U
I:
i?"r:6 -
‘503. '1‘, L
‘ ‘1 1.41:1

 

Press

7 ./' Powder
HardDie . . ,

Pressing 7 '
between 2000

to 3000 lbs

‘\ A //

‘ powdfir Green dimensions’
Compact 2.2 cm diameter
~ 0.2 cm thick

 

“" (b)

Figure 6. (a) Photograph of the presser used in this thesis, (b) Schematic sketch of
pressing powder compact in a steel die.

23

As-ret‘etted .
monolthemlet .
recened refractory e-
1:51: 115mg 11 penal
: I
115555 e: then 1115
‘Iiglltfl. Attereutt '
fennel paper (Ft-gm

‘sttiied rheut 0.5 cm tr

 

 

 

figure :1. The hole .11
0301118101.
Retraeton end
teeter . ' ‘
.1 equal to the 0

1:15.21 blade. A s'X“
4 I \

1.‘

tin 51.

7 -
.1.\licr0\sare Contr

In this thesis .1

0.171111 Inc. Plvr
_. it".

f!~el ‘Efi P“
g 1:6,‘11350 \
‘ .tj‘gw 4 .
d 6 L311“
. an
All (1w .
weal
5'3...“
w,"\a\.a

As-received cylindrical refractory was about 30 cm in long. The cylindrical
portion of the casket used for microwave heating was 3 cm in height after cutting the as-
received refractory casket. The 30cm long cylindrical casket was marked 3 cm on the
surface using a pencil to suitable size to cut of refractory casket. The cylindrical part of
the casket then was cut following the pencil mark on the surface using a hacksaw blade
(Figure 7). After cutting, the rough surface of the casket was briefly polished using clean
notebook paper (Figure 7). After cutting and polishing, a hole was made in the casket
located about 0.5 cm from bottom of the cut cylindrical casket using a screw driver
(Figure 7). The hole allows measurement of the specimen temperature using an optical
pyrometer.

Refractory end plates composed of aluminosilicate 2 cm thick disks, with a disk
diameter equal to the outer diameter of the cylindrical section of the casket using the

hacksaw blade. A specimen setter was also cut using the hacksaw blade into 0.5 cm thick

(Figure 8).

2.4 Microwave Control System

In this thesis, a 2.45 GHz single mode microwave cavity (Model CMPR-250,
Wavemat Inc., Plymouth, MI) was the main heating source for sintering and joining
ceramics (Figure 9). The microwave power supply (SAIREM, Model 24, Rue Louis,
France) generates 0 watt to 2000 watts. For all sintering and joining work, the
microwave cavity was operated using TM] 11 electro-magnetic mode.

An optical pyrometer (ACCUFIBER, Model 10) setting was performed prior to

microwave heating. The optical pyrometer was set as following steps to measure

24

13 011E—

 

 

Hacksaw

 

 

 

ado
Marked by a
pencil
1.3
\
1
/
1361* 4
30 cm

rl In
Paper -
—~.—~

Making hole Hole made

 

using a 0.5 cm above

screw driver bottom of
the refractory
casket

Figure 7. Schematic illustration of the technique used to fabricate the cylindrical portion
of the refractory casket. The starting material was a 30 cm long section of a zirconia
refractory open cylinder.

25

1501 for .\

 

 

mm m,

fluff; '
b y. ‘... ’-
11103[fur.1(.m '

Used for Sintering Used for Joining

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

It 10.3 cm pl '4 7.3 cm bl
> Alumino-
2 cm silicate end
plate
C D T Q
Cylindrical
3cm Zirconia ZYC
Casket
Vietz) hole L View hole
Alumino-

 

silicate

Alumino-

'2 cm silicateend
plate

Figure 8. Illustration of refractory cavities used in this thesis.

 

 

 

 

26

nsrde micron .11 e . ..
stelt. first. the = ‘1 -
11111111141111) ..
one :11. ° and shut 1-1;
petite the enter 1e;
1131121158356le 1

1131111 on the screen.

 

51212116} 11115 pu~hetl

in.
.1

siting nttmher 51W
Afterplaetne .
aten the Opitt‘al par

Tn \ \ . '
.te .mputer Lt.llliIt‘l.

1,

"l

g heating. The n

[1261mm “.1: set to (1p;

it't‘rl v." ‘
1.11.1.0 kitten “.18 lit

1‘ . .
4 Smtenng PFOCW

inside microwave cavity. After power on the pyrometer, it was automatically calibrated
itself. First, the # 9 key was pushed and then the number 1 key was pushed. The screen
showed black body after pushing the down arrow key and the right arrow key was pushed
one time and showed pyrometer on the screen. The enter key was then pushed. After
pushing the enter key, the right arrow key was pushed one time and then the screen
showed sense type. The down arrow key was pushed one time and sense factor was
shown on the screen. Sense factor was pushed following the # keys 1572 and than the
enter key was pushed. After setting sense factor, the upper arrow key was pushed until
showing number 500.00.

After placing specimen inside the microwave cavity, the cooling water was turned
on then the optical pyrometer, the microwave power, and the computer were turned on.
The computer controlled the short and the probe position to minimize reflected power
during heating. The reflected power meters were turned on. The initial value of the short
position was set to approximately between from 10 cm to 10.05 cm while the probe’s

initial position was from about 0.2 cm to 0.3 cm.

2.5 Sintering Process Using Microwave Power

A maximum of seven compact powder specimens can be sintered simultaneously
using a 10.3 cm in diameter refractory casket (Figure 11). First, the microwave input
power was set to 100 watts and the input power was increased in 100 watt increments
was performed every 3 minutes before coupling. The coupling was heat-starting point
that material inside the microwave casket started to absorb the microwave energy then

started heating after increment of power about 300 watts.

Monitor
(Control Short
& Probe) S“Pp'y . - .44
' ' M Reflected power
Detector

Microwave

 

Figure 9. Photograph of the microwave cavity and the microwave power supply used in this thesis.

28

One prehle."
253:: :he mtertm 1x 1'
{reeling 111111.111) 1.
newline heating :.

V. 1- - .
its). 111115. 1n\r'\lc1‘

- l . ’ .
mm 01 dt‘t‘ICJsll‘ .

 

 

charging the short Jl‘t
gt'rertndteatnr in mi.
.‘t’t‘lttmtghl cause at.
511: off immediately
1:3“de ustem.

1:01 tempcm L.‘

30571031116 11 prese‘

"1.): .

1111 C. 100 watt the
EM; - ~.

.....r1ture n as aha:

1116 tempera:

HT.“ .
sup“ 4 » a, .

Allfi‘.’ read.“ 11

x ‘0?“
~.:r'\‘ '~‘
“er ,

is... -
““r‘id hold 11 1..

"”53”,? the

3110“ d‘

91».

”34.41... —
“411110. w. ,

t 1&1.
)2)"
"'sfi’l‘.‘ ..
“*d\a‘_

* WI‘

One problem encountered when heating the 7.5 cm diameter refractory casket
using the microwave was that the temperature tended to decrease after coupling.
Coupling typically occurred in the temperature range of 900°C to 1000°C. The
microwave heating rate cannot be controlled at temperatures less than 1000°C in this
thesis. Thus, increasing input power in 100 watt increments every 3 minutes solves the
problem of decreasing temperature after coupling. Every time the input power increased,
changing the short and probe positions minimized the reflected power. If the reflected
power indicator in microwave power controller reached 45, the reflected microwave
power might cause arcing in the microwave cavity. In this case, the microwave must be
shut off immediately. The reflected power above 45 could cause damage to the
microwave system.

For temperatures above lOOOOC, increases in microwave power were controlled in
order to give a preselected heating rate. For example, at 9 minutes temperature was
1100°C, 100 watt increments causes increasing temperature. After 19 minutes, if
temperature was about 1200°C, heating rate was 10°C/min.

The temperature was measured by an optical pyrometer though a view port in
microwave cavity (Figure 10).

After reaching a desired sintering temperature, the temperature was considered to
be constant if the temperature changed by less than i 2°C for 5 minutes. After a pre-
determined hold time at the sintering temperature, decreasing the input power without
changing the short and probe position did cooling. Table 7 shows sintering temperature,
time, and heating rate for of the each powder material used in this thesis. The

microwave-sintered specimens were fractured using a hammer. The sintered specimens

29

112:1
C311

 

 

ESure [Q S;
11ers.

Microwave
Cavity w all

'/

 

 

 

 

A

Refractory
Caseket
A A A A A View
hole

JYJYV

I>

Optical
Pyrometer

 

 

 

 

 

 

 

 

 

 

 

=+f

Figure 10. Schematic of specimen temperature measurement technique used in this

thesis.

30

- E
111337. Condition:

tens.

31.116.11.11

 

Alumtna
.‘lllt’lllr 1:11

llTCOI‘sld
511101” 11:1)

7.1 It‘tlll l d

 

 

 

 

Alumina —11re0

CflIIiH’1\':IL‘\

1
1". f’lg‘ - ~
“131116515. The 11‘
5m.- ..; -

«“u tfiglmfnb LIT:

\

1111161101

Tn] 3“ I -
131.1. POllslllllL’ t'

Alumina
Smell \'-0-.

 

Table 7. Conditions of sintering for each of the commercial powder material used in this
thesis.

 

 

 

 

 

Material Mass Sinter Temp Hold time Heating rate
Alumina 2 grams 1300°C 1 hour 10°C/min

3m°'% I203 3 grams 1375 °C 1 hour 10°C/min
zrrconra

8m°l% I203 3 grams 1350°C 1 hour 10°C/min
zrrconra

Alumna ‘Z?r°°“'° 3 grams 1350 °C 1 hour 10°C/min
composrtes

 

 

 

 

 

Table 8. Polishing time and diamond grit size for each of the ceramic materials included
in this thesis. The time corresponds to seven mounted specimens being polished. If
fewer specimens are polished simultaneously the times are reduced.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Material 35 um 17 pm 15 um 10 pm 6 ptm 1 pm
Alumina 20 min 30 min 30 min 30 min 30 min 30 min
3mol% Y.203 20 min 30 min 30 min 30 min 30 min 30 min
zrrconra
8mol% Y.203 20 min 30 min 30 min 30 min 30 min 30 min
zrrconra
A'“"“"° “z.”°°“‘° 20 min 30 min 30 min 30 min 30 min 30 min
composrtes
Silicon Nitride 5 min 10 min 10 min 10 min 10 min 10 min
Silicon Carbide 5 min 10 min 10 min 10 min 10 min 10 min
MaCorTM 2 min 10 min 10 min 10 min 10 min 10 min
MgF2 N/A 10 min 10 min 10 min 10 min 10 min

 

00m;

 

 

111:6 312cm
< 3--

 

 

 

 

Cylindrical 3 cm
View Zirconia ZYC
[1019\E’ 038k“

 

 

Seven batch \e

compact powder<e§02=1=05 cm
Alumlno-
sllicate E 3 I 2 cm
and plate

Figure 11. Schematic illustration of refractory casket arrangement used batch sintering of
ceramic specimens in this thesis.

 

 

 

 

32

 

 

 

were 1111 h) the hi."
I

 

preees hit the suite
The grain me..
fracture surfaces 11.

Friar 0.1 1.5.

2.6 Specimen Prep

The specr .re
1523:: an automate ‘
1111. The grit si'

41 IL'\

Pimp-1‘ .
uxuiorpqht

‘1“ F‘
-1. or 10011517117;

‘
4.? .
“1.11111: ’, an Cm
' iii

were hit by the hammer to break into several pieces. The hammer to break into several
pieces hit the sintered specimens.

The grain mean size of the specimens was calculated from SEM micrographs of the
fracture surfaces using the liner line intercept method with a stero-graphic projection

factor of 1.5.

2.6 Specimen Preparation

The specimens sintered and the as-received commercial specimens were polished
using an automatic polisher (Model Vari/PoLTM VP-SO, LECO Corporation, St. Joeph,
MI). The grit sizes of the diamond paste (Warren Diamond powder Co. Inc, Olyphant,
PA) used for polishing included 35 um, 17 um, 15 um, 10 pm, 6 um, and 1 pm (Figure
12). For polishing, the specimens were first mounted onto an aluminum plate disk about
2 cm thick, 20 cm in diameter, with 0.5 cm diameter center hole. The plate was heated
for about 10 minutes on a heater to allow a thermoplastic (Thermoplastic Cement,
BUEHER, Lake Bluff, IL) to melt which was the adhesive between the specimens and
the surface of the aluminum plate (Figure 13). After mounting, the plate and specimens
were cooled for about 20 minutes to allow the thermoplastic to densify.

Polishing oil (Microid Diamond, LECO Corporation, St. Joeph, MI) and diamond
pastes were applied to the polishing cloth after placing a polishing plate in the polisher
each time. The polishing oil was applied each time diamond grit changed. For newly
mounted polishing cloths about 1 gram of diamond paste was applied to the polishing
cloth. The polishing oil gave a good contact between the polishing cloth and the polished

specimens. When the polishing for each diamond gn't size was complete, the specimens

33

acre cleaned Using

 

   
 
 

utzcrand 1 "t ““1“"
timind gnt. The '
lubricated during p
C'Wfilt‘dllng of the ‘
polishing plates \\ d‘
itecimens' surfaces
3m HS indicated h
satires “ere \er}

snared good pg! 1. _\

4??ij L116 p011) hi

After poli’s

Minted from the

n, -
tr: specrmens \Kt

5:574”
1...-L5 [0 rEmU\

litm-

Met U1 Sm

:1.i-""T" ’
“on bx

237$ rel-h Q 1' f‘
\ 110\ f
\L‘ L (1

were cleaned using liquid soap, which was strong resolve liquid mixed with water 99 ‘7c
water and l % liquid soap (Liquid-Nox, Alconox, Inc, New York, NY), to remove oil and
diamond grit. The polishing oil was applied when the cloth was not sufficiently
lubricated during polishing since not enough oil caused tearing of the polishing cloth and
over heating of the polisher. For all polishing done in this thesis, the spin rate of
polishing plates was fixed at 120 rpm. After finishing with the 35 um polishing, the
specimens’ surfaces were the same color. For example, insufficient polishing after 35
um was indicated by an apparently dark color through the specimens. The specimens’
surfaces were very shinny after finishing with the 17 um. Thus, a shiny specimen surface
ensured good polishing after polishing with the 17 um grit. Table 8 shows the
appropriate polishing time and diamond grits size for each of the ceramic materials used
in this thesis.

After polishing with the l ptm gn't was complete. The specimens were dis-
mounted from the aluminum disk by heating the disk for 5 minutes to melt thermOplastic.
The specimens were then taken from the aluminum plate and put into acetone for 1 or 2
minutes to remove residual thermoplastic in a beaker. An ultra—sonic cleaner (Model
Ultra-Met III Sonic Cleamer, Burhler LTD, Evanston, IL) was used to remove
contamination by ultras-sonification for 20 minutes using de-ionized water. Excess water
was removed from the ultra-sonically cleaned specimens using a clean tissue.

As-received MgF2 billets were cut into 1 cm X 1 cm X 0.2 cm specimens using a
low speed saw (Buehler IsoMetTM, low speed saw, Lake Bluff, IL). MgF2 is much more
brittle than the oxide ceramic materials, thus, a mass of about 1.5 kg was applied to the

saw arm and the cutting speed was set at between 5 rpm to 6 rpm. During cutting, the

34

“if; Spt‘t‘lmt'm ‘v
duetocontat‘ttng
tut specimens it as
tlstmum plate vi a

L34.

....ettt.isusu.1ll\ E

 

regimen could he L
timed to cool ire:

)lgF; Vt as p
wishing. Thus. at?
tlutztnum plate is .1:
that: rammed from
“taremmed usmg
“fitter w remm ed
SE3.5115'Jnttitlishetl 5'
3353- Aiitt both 51'

liter emlldlnjno G

Cg.

Tl'fituP
4:.1l"“~‘
- [0 re mitt C

MgF2 specimens were wrapped with a small piece of paper towel to reduce scratching
due to contacting with metal specimen holder during cutting. After cutting, each of the
cut specimens was mounted onto the aluminum polishing plate using thermoplastic. The
aluminum plate was first placed on a heater to melt the thermoplastic. The set point of
heater was usually 8 and after 10 minutes heating. melted thermoplastic so that the cut
specimen could be attached to onto the aluminum plate. The specimens were then
allowed to cool freely for 20 minutes to avoid failure of the thermoplastic bond.

MgF2 was polished on both sides to examine the infrared transmittance after
polishing. Thus, after one side was polished, specimens were dis-mounted. The
aluminum plate was heated to melt the thermoplastic. After 10 minutes, the specimens
were removed from the aluminum plate and the residual thermoplastic on the specimens
was removed using acetone. The specimens were cleaned using water and any excess
water was removed using a clean tissue. The specimens were then remounted with the
single unpolished side up. Polishing was done using the same series of diamond grit
sizes. After both sides were polished, the specimens were dismounted and placed in
beaker containing de-ionized water. The beaker was placed in the middle of an ultra-
sonic cleaner where the height of the water in the beaker matched the height of the water
in the DI beaker (about 5 cm). The specimens were cleaned ultrasonically for about 20

minutes to remove contamination during cutting and polishing.

2.7 Coating Procedure Prior to Joining.

Prior to joining, the ceramic materials used in this thesis were coated with various

spin-on materials (Table 6). Silica film were used most often as the interlayer for the

35

 

m

Pgtte 13.
Dism—

r [Q

Polishing
Oil

Pressure

I".
.1

 

Figure 12. Photograph of the automatic polisher with the aluminum plate in place for
polishing.

Set as
8 During Heating

 

 

Figure 13. Photograph of heater and aluminum plate used in mounting the
specimens for polishing. Note that several specimens have been fixed to the plate via
thermoplastic.

ceramic-ceramic .1
spinning at 500 rr‘
It etontis.

First. douhle ~;
mi‘iunted on the p."
hen the TOUghmg \.
lie spin rate and m

Eyre 14. the spin-t

 

 

 

 

 

 

iii) soured the spe-

After spinntn
sntteth. E‘lCCpI tor
tired after spinning
sadtum alginate to l

SC-Clle alginate and

these.
For 31ng “ l
nasal. but it

lm \k'.

1 ~ ltdic On
- p1

 

 

ceramic/ceramic joining done in this thesis. A high speed spinner was capable of
spinning at 500 rpm to 6000 rpm. For this thesis, the spin rate was fixed at 3000 rpm for
20 seconds.

First, double sided adhesive tap was applied to the petri dish, and then specimen was
mounted on the petri dish. The petridish was placed in high speed spinner holder and
then the roughing vacuum pump (which held petri dish during spinning) was turned on.
The spin rate and the spin time were set to the desired value (Figure 14). As shown in
Figure 14, the spin—on material was liquid. Two or three pipette drops of spin-on material
fully covered the specimen surface.

After spinning, the spin-on liquid material that had been applied was thin and
smooth. Except for the silica film and sodium alginate, the spin-on materials were not
cured after spinning. Water was added to the sodium alginate, at a ratio of 3 grams of
sodium alginate to 100 ml water to control the viscosity. Following spinning, both the
sodium alginate and silica film was cured at 200°C for 20 minutes in a conventional
furnace.

For MgF2 with MgF2 joining, sodium silicate solution was used for the spin-on
material, but film was not cured after spraying. MgF2 specimens were coated with

sodium silicate on polished side using a high speed spinner at 3000 rpm for 20 seconds.

2.8 Microwave Joining Procedure
Ceramic/ceramic joining was performed in a 2.45 GHz single mode microwave
cavity (Figure 15). The particular resonant cavity mode used for this thesis was TM] H

mode. Microwave power from 0 watts to 2000 watts was generated by a Sariem (Sairem

38

 

:4. RM“ [All

1,193! WM!

0" Witt:

asset \\ is :
stared at
:titrgiit .v. c,
Sniping he.
(with; at
less; iihgiu;

Arte

': .
iii 4 r\

. ..
~ filial)

 

th

Fttt }

'7
Ht
:2. it“

24, Rue Louis, France) power supply. The Microwave power controller controlled the
input power. Cylindrical refractory caskets, 7.3 in diameter and 1.3 cm thick, were
prepared and placed in the microwave cavity (Figure 16 (a)). Alumina (AKP 30)
deadweight disks were placed of the specimens to provide low pressure during joining
(Figure 16 (b)). The microwave power control is discussed in section 2.4. A heating rate
of approximately 45°C per minute was used in this thesis. Duringjoining, the reflected
power was minimized by adjusting sliding the short and probe positions. Coupling
occurred at input powers between 200 watts and 250 watts within 7 or 8 minutes for most
microwaves heating. The probe position was changed from 0.3 cm to about 1.7 cm after
coupling occurred, while short position was changed between 10.2 cm to 9.3 cm.
Coupling occurred when the specimen absorbed microwave energy at critical energy
level (about 250 or 300 watts).

After coupling the temperature quickly rose to 800°C. Thus, it was very difficult
to control the heating rate for the temperatures less then 1000°C.

For MgF2 joining, the maximum joining temperature was 700°C. The pyrometer
attached the microwave system in this thesis can measure temperatures from 500°C.
Holding 700°C was not easy operation since after coupling, specimen temperature was
reached about at 900°C. However, it was possible to hold temperature at 700°C, after
step by step power decrement (one time per 5 or 6 watts decrement) approximately 30
seconds per one power decrement.

Table 9 summarizes the interlayer materials, joining materials, joining

temperature, joining time, and deadweight loading for the ceramic/ceramic materials

39

 

Pol

—*
Polished Dmpthespin-onmm'ial
specimen onfliesubstrate

 

 

 

 

 

IEQPJ i

Spinningcontrol

 

Figure 14. Schematic illustration of coating procedure using a high speed spinner.

40

- L
joined in [his ti.
Defltiih Siui‘tlt.

pifitiih Stahtl‘.

2.9 Joining 1H

Nztnde
initiate tTherr
fleeing nitrog
sire-on carbide
:eqemis but t’r
short earhic}
hide: to join
tag The

a ’
.lU* rdLE \\ d.

5

3.35427va ‘
£15 I)!
P...“

joined in this thesis including (MgFgl MgFg, AKP 30/AKP 30, partially stabilized ZrOgl
partially stabilized Zl'Oz, fully stabilized Zr02/ fully stabilized ZrOz, HAP/MaCorTM,
partially stabilized ZrOleaCorTM, SiC/AlgOg/MaCorTM, and partially stabilized

ZrOzlAlZO3, and alumina-zirconia composites alumina zirconia composites).

2-9 Joining technique for Non-Oxide Ceramics
Nitride and carbide ceramics joining attempts were made in a conventional

furnace (ThermoCo Products Corporations, Model Mini Brute, Orange, CA) with a
flowing nitrogen atmosphere and a flow rate of about 120 sccm. Only one polished
Silicon carbide or nitride specimen was coated with BlackglasTM spun at 300 rpm for 20
Seconds but the other polished silicon nitride specimen was not coated at all. Pair of
silicon carbide or nitride specimen was first squeezed by hand then placed onto specimen
holder to join using the conventional furnace without BlackglassTM curing prior to
joining. The heating rate for conventional joining was 20°C per minutes and the nitrogen
flow rate was 120 sccm during heating. The conventional furnace used for joining was
limited to temperatures of 1000°C or less. During joining, a deadweight of 20 to 60

grams was applied.

2.10 Producing Notch Mechanically Using A Low Speed Saw

As part of the ceramic/ceramic joining work, notches were cut into selected
specimens prior to joining.

Polished and silica-coated specimens were mounted in a low speed diamond saw

specimen holder, with the silica-coated side faced toward the saw blade (Figure 17). The

41

MW.“ *1:

 

Cylindrical
Cavity

Refractory
Casket

 

Figure 15. A photograph of the 2.45 GHz single mode microwave cavity used in this
thesis.

42

 

 

FL

6".
/ I.

‘0.
.1

ftr‘r‘jw I '6' If! ‘-

i _

— H u.
i F

L11

79616- lul P“

"V

‘1.

I

in: .. ,

C. \.
”drug“ .,
. ku'c'

fl

Alurrlnpa ;
‘ SALI
Zirc‘oflg
ZYC

.— " Dead {iv-«eight
‘— Speciments-

.. speimen ' 71:5. 1

“1"."

 

Figure 16. (a) Photograph of the type of zirconia/aluminosilicate refractory casket used
during ceramic/ceramic joining, (b) Photograph of the bottom SALI plate of the
refractory casket with the specimen, with the specimen setter and dead weights in place.

43

idi‘if 9. .~\ '
hddtmex
LitOlS.

 

 

y‘-

XKPFM-

leO: Vs
x

SIC.-

f/
a-

Table 9. A summary of the interlayer materials, joining materials, joining temperature,
hold time, and deadweight loading for all of the ceramic/ceramic joining included in this

 

 

 

 

 

 

 

 

 

 

thesis.
Attempted joining S in-on in terla er Joining Hold Dead
materials p y Temp(°C) time Weight
MgF2 with MgF2 Sodium silicate solution 700°C 20 min 60 grams
AKP 30 with AKP 30 Sodium alginate 1400°C 20 min 60 grams
*ZrOz with *ZrOz Silica film 1500°C 20 min 60 grams
*ZrOz with A1203 Silica film 1500°C 20 min 50 grams
HAP with MaCorTM Silica film 1020°C 20 min 20 grams
*ZrOz with MaCorTM Silica film 1020°C 20 min 20 grams
SIC/A1203 With . . . 0 .
MaCorTM Srllca film 1020 C 20 min 20 grams
A1203/Zr02 with . . o .
A120 3/Z r02 Silica film 1450 C 20 min 50 grams
**Zr02 with **ZI'02 Silica film 1500°C 20 min 50 grams

 

 

 

 

 

 

*ZrOz is partially stabilized zirconia and **ZrOz is fully stabilized zirconia in this thesis.
HAP is hydriteappitite and SiC/A1203 is a SiC platelet reinforced alumina

Table 10. Experimental condition of conventional joining of silicon nitride and silicon

 

 

 

 

 

 

 

 

 

 

carbide.
Material Spin-on Spin-rate Max Temp Hold time Atmosphere
material (°C)
SiC BlackglasTM 5000 rpm 900°C 20 minutes N2
Si3N4 BlackglasTM 3000 rpm 900°C 20 minutes N2
Si3N4 BlackglasTM 5000 rpm 1000°C 20 minutes N2

 

 

 

legit ipplled t-
it specimens.

hit. about l T;
ii Sh‘NS it uh:
itch me that
teatime the ~

new the he.

i

.1] Procth
Notch

legitimate

:51 Tie the pa:

weight applied to the saw during notching was much less than weight used during cutting
the specimens. Often 1.5 kg of deadweight was used to produce notch, but a mass greater
than about 1 kg was detrimental to the life of the low speed diamond saw blades. Figure
18 shows a schematic of the notch processing technique. After specimens containing a
notch were joined with another specimens, the specimen was cut in half in order to
determine the stability of the notch dimensions. The notched specimen was then

microwave heated in an attempt to join the notched specimen.

2.11 Procedure of Producing Notch Prior to Sintering

Notched were “stamped” prior to sintering. Monofilament fishing line
(approximately 250 mm in diameter) was cut into 1.5 cm lengths and then glued onto one
of the die punch faces. Approximately 2.5 gm of alumina powder was uniaxially pressed
at 1.75 MPa using a uniaxial press.

The powder was put into the hard die then smoothed using the second die punch
which had no glued fishing line on it. Upper punch was replaced and then pressed at 17.5
MPa (Figure 19).

Alumina (TM-DAR) enclosed notches was sintered at 1300°C for 1 hour with
heating rate 10°C/min using the microwave energy. Also, alumina-zirconia composites
enclosed notches was sintered at 1300°C for 1 hour with a heating rate 10°C/min.
Alumina—zirconia composites was then joined with zirconia (3 mol % YzO3 sintered at
1375°C for 1 hour using the microwave energy), which was without notches at 1400°C

for 20 minutes.

45

 

Figure 17. Photograph of a low speed diamond saw after the specimen had been in the
saw.

46

Ԥ
5

ll)

Coated . Notches cut

 

 

 

specimen into specimen
, QED
\\_____.
Specimen sectioned Joining of Notched
through notches specimen
/"i
m > t

 

 

 

 

I
W 1r \ m p\
U U U

 

Figure 18. Schematic of the technique used to mechanically cut notches into the
specimens.

47

 

 

 

€36 19 S;

Filter 10 Smift‘.

        

 

 

 

 

@

 

Monofilam -
fishing line
K

Die
Punch

 

 

 

 

 

 

Compact Powder
> After Making Notches

Figure [9. Schematic of “stamping” notches into the green (unfired) powder compacts,

prior to sintering.

48

112 )ltluntit

101nm

{transmit
l

.1, Ld
i1 ll\ll iMl.|

lflCJEfllT‘lfl at ,

Silfilflifls \\

2.12 Mounting and Polishing of Joined Specimen

Joined specimens were cut into two or more pieces using a low speed diamond
saw. Joined and cut specimens were then mounted in diayllyl-phthalate powder
(Compression Mounting Powders, LECO Corporation, St. Joeph, MI) using a press with
a 1 inch diameter rum (Figure 20).

At least 3 of the joined and mounted specimens were required to polish using the
automatic polisher. During polishing, 3 mounted specimens were balanced without
inclining of a polishing holder. As mentioned in section 2.6, joined and mounted
specimens were polished as using diamond grit size from 35 pm to 1 pm. After
polishing, cleaning was performed in DI water using the ultra-sonic cleaner for 20

minutes.

2.13 Vickers indentation Testing of Joined Specimen

The fracture toughness of joined specimens was estimated from the crack
deflection at an interface using Vikers indentation (Figure 21). Indentation was applied
away from the interface to near the indentation using Vikers indentation then applied
more indentations to near interface as shown in Figure 21. Same load and load speed
were applied onto joined specimen to observe the deflection of bond layer after joining

compared with mother material.

2.14 SEM Observation of Bond-layer

Joined and mounted specimens were coated with gold using a sputter coater,

(Emscope SC 500) coating for 1 minute produced 7 nm gold coating. Carbon paint,

49

r-it ‘

-0. 3

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Joined A cross-sectional Sectioned
specimen cut through the specimen
specimen

©E
Diayll

Thlate

mounted
material
Mounted specimen

 

 

Figure 20. Schematic illustration of mounting the specimen after joining and cutting.

50

 

l’lclel
llldetltl

VleCl‘S l -y - i
are} . ,l
lndentor (y

 

EL Mounted
specimen

Figure 21. Schematic of Vikers indentation near the bond layer.

51

 

lable l l.

specimens

Fl»

Ԥ- '
T l
~‘e
s“ '

3??»

Ml}:

Table 11. Type of filters and detectors used to measure the IR transmittance of the MgF2

 

 

 

 

 

 

 

 

 
   
    

 

specimens.

Wavelength range Types of filters Types of detectors
400-650nm Transparent filter 813-SL
650-840nm Red filter 813-SL
850-1650nm Black filter 800-IR

{:3 33.3 .33 Optical power meter
MEAL Dill] Eli] (NewportCo., Model 835)
F” 33 333 3 ,
Wavelength
Specimen selector Filter Condenser
lens

Detector

   

Holder \

 

/ \.
'1? ~33 to

 

 

 

 

 

 

 

Light source

\

\ /'
Condenser Monochromator
]ens (OrielCo.,Model77200)

Figure 22. Schematic of the apparatus used for measuring the IR transmittance of MgF2

specimens.

 

£5in

The s

f‘ i
“-‘l sh ;

~i‘\.-

. -ltt

Y‘s.

«11 T71
a”

‘r‘

which was carbon based black liquid (Ted Pella, Inc. Redding, CA), was applied before
gold coating from edge of the joined specimen to end of the mounted side to the
specimen to prevent charging during SEM observation. A JEOL 6400 SEM was used to
examine bond layer. The best image resolution was obtained using accelerating voltage
of 20 kV, a working distance 15 cm, and condenser lens intensity setting of 9.

A set of notched and joined specimens was mounted to an aluminum stub using
carbon tape in order to observe changes in notch dimensions before and after joining.

The specimens mounted onto aluminum stub also were coated with gold for 1 minute.

2.15 Infrared Transmittance Measurement of MgF2

The transmittance of polished, coated, and heated Mngspecimens was measured
for wavelength from 400 nm to 1650 nm using a spectrophotometer (Oriel Co, Model
77200) (Figure 22).

Transmittance was first measured without the specimen in the spectraphotometer
and then measured with the specimen in the spectrophotometer. Transmittance was
calculated after measuring both without specimen and with specimen. Transmittance was
calculated from dividing the measurement of without specimen data into measurement of
with specimen data. Which was

Transmittance measurement from with specimen.
Transmittance measurement from without specimen

Transparent filter transmitted UV light only in range of wavelength 400-650nm,
red filter transmitted infrared only in range of wavelength 650-840nm, and black filter
transmitted only infrared in range of 850-1650nm shown in Table 11. Two types of

detectors were used to get correct data at each range.

53

 

ll

”5.“. ‘
Lnk 5

55

ill
'I'J

I.

hrs a

.3 ;\L
3

..‘-I. _ ’A
r Cl: 15
~l

2.16 Microwave cavity cleaning

The microwave cavity was cleaned every each month. Volatile components of
the spin-on coatings deposit on the water-cooled microwave cavity walls. The
microwave cavity is cleaned using metal polish (No. 22 Metal and Chrome Polish, Blue
Ribbon Products Division, Indianapolis, IN) and commercially pure alcohol.

The microwave cavity is composed of the main cavity, the sliding short assembly,
and the probe (Figure 23). Inside the microwave cavity, the short assembly and the probe
are gold coated.

Before removing either the short assembly or the probe assembly from
microwave cavity, the short was positioned at 17 cm and the probe positioned at 0.0 cm
to prevent possible damage while removing the short and probe parts. After setting both
the short and the probe to the proper position, they were removed from the main
microwave cavity using proper screwdriver. The short and the probe assembly each have
finger stock (Figure 23, 24, and 25). Finger stock distributes microwave energy inside
the microwave cavity during heating are very brittle. Thus, separating the short and the
probe parts must be done carefully to avoid breaking the finger stock.

The inside of the main microwave cavity is coated with gold and the bottom of
the microwave cavity (which does not move) has finger stock (Figure 23). The cavity top
plate is the sliding short, which has also finger stock, and is goal coated (Figure 24).

Each of the separated parts (short, probe, and main microwave cavity) were
cleaned by first applying metal polish to metal part and then cleaned by alcohol. Cotton
swaps and cotton balls were used to clean the microwave cavity. Cotton swaps were

often used inside the microwave cavity, the short assembly, and the probe part that can be

54

 

 

 

reach

tings.

reached with one’s hands. The cotton balls were used to clean the finger stock since the
finger stock is very fragile.

After cleaning the main microwave cavity, the short assembly, and the probe
assembly, careful reassembly is required. In this case, the short was positioned at about
17 cm and the probe was positioned at 0.0 cm. After replacement, the short was
positioned to about 10.2 cm and the probe was positioned at about 0.6 cm to prepare the

cavity for its next use.

'Jl
LII

3 Microwave
- Cavity

   
   

Figure 23. A view of the microwave cavity disassembled for cleaning. Shown are the
main microwave cavity, the lanch probe, and the sliding short.

56

F343 34
Hither] Co

 

Figure 24. Cavity top plate showing finger stock. The finger stock makes electrical and
physical contact with the inner cavity wall. Thus, as the cavity top plate moves, the
effective cavity length changes.

57

 

 

Eire 25.

 

Figure 25. For the launch probe part of the probe assembly includes small fingers stack.

58

jMwM

3,1 Joint“?~ 0

JOHN

.tBSTRif
Mall) 1
spectrum (St
prc'tect such
difficult. sin
exertion.

amend tiln

l. lNTROD

Cerdn

A

Jllt‘ mate:

‘3w
«.1

Based slumln

3. Results

3.1 Joining of Diamond Thin Film to Optical and IR Materials

JOINING OF DIAMOND THIN FILM TO OPTICAL AND IR MATERIALS
J. G. Lee, K. Y. Lee and E. D. Case

Materials Science and Mechanics Department,

Michigan State University
East Lansing, MI 48824

ABSTRACT

Many materials that are transparent in both the visible and infrared regions of the
spectrum (such as ZnS and MgF2) are easily scratched. Hard coatings may help to
protect such materials, but direct deposition of diamond films onto such substrates is
difficult, since the materials tend to degrade at the temperatures required for diamond
deposition. We discuss attempts to make MgF2/MgF2 joins, as well as efforts to bond

diamond films to MgF2.

1. INTRODUCTION
Ceramic/ceramic joining has been done successfully for a variety of structural
ceramic materials, including alumina, SiC, and zirconia. For example, Case et al. has

joined alumina, MacorR, and zirconia using spin-on layers [1-2]. In addition to structural

 

J. G. Lee, K. Y. Lee and E. D. Case, “Joining of Diamond Thin Films to Optical and IR Materials,"
Ceramic Transactions, Volume 94, American Ceramic Society, Inc., Westerville, OH. pp. 509-520 (1998)

59

ennuinnnil
penne.ienc
thnnerhem
hnnllinnh
mhmmnd

Thspn
hunnndthln l
35.53th as a

‘- NERD
ll'lhfieri

Asre
ififiedpnh

Steed d1 amt.

ceramics, optically-transmitting ceramics have been joined, usually at high joining
pressures. Yen et al. [3] joined MgF2 specimens using a direct diffusion bonding
technique at temperatures of 800°C and greater, using pressures of 17 MPa to 25 MPa.
Yen et al. attributed a drop in the joined MgF2 specimens’ transmittance to grain growth
in the polycrystalline MgF2 specimens.

This paper discusses attempts to bond (1) MgF2 to MgF2 and (2) polycrystalline
diamond thin films to MgF2. For successfully joined specimens, the transmittance is
measured as a function of wavelength and the optical absorption coefficient, 0t, in order

to compare the optical qualities of joined and unjoined specimens.

2. EXPERIMENTAL PROCEDURE
2. 1. Materials

As received commercial MgF2 (Itran I, Eastman Kodak Company) is a hot-
pressed polycrystalline material with a mean grain size of approximately 2.8um (Figure
26). Specimens were sectioned into roughly lcm X 1cm and 2mm thickness using a low
speed diamond saw.

After cutting, the specimens were polished by an automatic polisher (LECO
Corporation) using the following series of diamond grit sizes 17, 10, 6, and 1pm. The
specimens were polished on both sides to facilitate IR transmittance measurements before
and after joining. The polished specimens were coated with a sodium silicate solution
(Columbus Chemical Industries Inc.) using a high-speed spinner (Figure 27). After
spinning at a rate of 2000 rpm to 5000 rpm, coating was uniform in thickness and

smooth.

60

 

Figure 26. SEM image of the fracture surface of MgF2.

61

ll
We 77

his
33'
ubst

 

 

 

 

 

 

 

IEQPJ l

Spinningoontml

Figure 27. Schematic showing the high-speed substrate spinner used in coating to the
MgF2 substrates.

62

 

 

5355

“(a

mean

at 41!.

hit it
5:063."
Uhfdsr
0310 3
“(re p

[56 ggd

In this study, two silicate solutions were used as bonding agents. A commercial
sodium silicate solution (Columbus Chemical Industries, Inc.) was used as one of the joining
media. Although the sodium silicate was very viscous, after spinning on the hi gh-speed
substrate spinner, the coatings were quite smooth. A second bonding agent that was used
was an organically-based silica solution, which pyrolyses to an amorphous silica film upon
curing at 200°C. However, the silica film was not successful in either the MgF2/MgF2
bonding or the diamond film/Mng bonding.

The diamond films used in this study were between 3 to 4 microns thick, with a
mean grain size of approximately one micron. The films were microwave plasma-deposited
at 400°C to 500°C under 28 torr pressure onto (100) oriented single crystal silicon wafers
that were 0.048 to 0.053 cm thick and 5.08 cm in diameter. The diamond-coated silicon
specimens were fractured into 1 cm X 1 cm sections using a razor blade. Following
ultrasonic cleaning in deionized water, the diamond-coated silicon specimens were placed
onto MgF2 substrates that had been coated with sodium silicate solution. The specimens
were placed such that the diamond film coating on the diamond-silicon specimens contacted
the sodium silicate coating on the MgF2 substrates.

Prior to joining, the bonding agents were spun onto the MgF2 substrates. Using a
pipette, a few drops of either sodium silicate or the silica film was placed onto a polished
MgF2 substrate and the film was spun between 2000 and 5000 rpm using a
hi gh-speed substrate spinner (Figure 27).

2.2. Joining Procedure
Coated specimens were heated in either a single mode microwave cavity (Sairem,

Model MWPS 2000) or in a conventional electrical resistance furnace. For joining the

63

 

.llgf: to.)
lint it
using the 5
heights 01‘
13. lR It
All lR
Co. \lttlel
Prior tt
stretiment
it or tier ti
hut not jun
til-rte on the
lll .‘iS‘POllI
13' Ptnilshed
tIlPohshed

311’.
In alu'ttlo

14» '
“33C abate.

1

3333335 it e

Eittlmeng l'elt

MgF2 to MgF2 using the silica film, the annealing temperatures ranged from 500°C to
1200°C with applied dead weight loads of 20 gm to 85 gm. The MgF2/MgF2 joining
using the sodium silicate solution utilized temperatures of 500°C to 800°C and dead
weights of 20 gm to 85 gm.

2.3. IR Transmittance Measurements

All 1R transmittance measurements were performed using a monochromator (Oriel
Co, Model 77200) over wavelength range of 800 nm to 1600 nm (Table 12).

Prior to the transmittance measurements on the joined specimens, a series of
experiments was done to determine the effect of (1) heating the polished but uncoated MgF2
at or near the temperature used for joining and (2) coating and heating the MgF2 substrates
(but not joining them). In order to examine these effects, transmittance measurements were
done on the MgF2 substrates having the following surface treatments:

(1) As-polished specimens,

(2) Polished specimens which were subsequently heated at 700°C in air,

(2) Polished specimens that were coated (without joining) and then heated at 700°C in
air.

In addition to the three “comparison” surface treatment/heat treatment conditions
listed above, polished specimens were coated, and then the coated surfaces of the two
specimens were placed in contact. The specimens were then heated in an effort to join the

specimens (either MgF2/MgF2 joining or diamond/Mng joining).

figure 28.

“fight l0,“

 

(l
U

.___Top Alumina SALI
end plate

a____ Zirconia ZYC
Dead weights

Joined
Specimens

Specimen setter

 

 

   

 

 

 

.__ Bottom Alumina SALI
end plate

Figure 28. Schematic of the refractory casket used for microwave joining. The dead
weight loading also is illustrated [2].

65

 

 

 

 

“gun I!
Tiltiiftttetni

 

 

{- F 7‘ ‘ , 'C’
‘ lllll u'l'l'l lLJLJ

7‘ {'1 WT F

[iii

 

 

Optical power meter
(Newport Co., Model 835)

Wavelength
Specimen selector

Holder \ l
/ \ t

l
l
i
3 i ip<— - 3
1

Condenser
lens

   
 

Filter
Detector

 

 

 

 

 

 

 

\ /
Condenser Monochromator Light source
lens (Oriel Co., Model 77200)

Figure 29. Schematic of the spectrophotometer used for the IR transmittance
measurements.

66

 

jpmt’
oi the WW
polisher. tht‘
17.10.03”

SE.\l
Sfigrmt‘nS. -
sectioned “in

SEM eiumtl

3. RESl'lJ
3.1. Joining

For 1
TM usrng
contentions
MMmh
mlfitnttal'ej
contentions!
”SPHERE
30). hlterost

turned ilgP/

The coutllltl0

Th:
uit lfmpéfdiUl

Joined specimens were cross—sectioned using a low speed diamond saw. A portion
of the sectioned specimen was mounted in thermosetting powder. Using an automatic
polisher, the mounted specimens were polished using a series of diamond grit, including
17, 10, 6, and lum.

SEM examination determined the bond-layer thickness in the sectioned
specimens. Since ceramic materials are not electrically conductive materials. the
sectioned and mounted specimens were coated with a 0.7 nm thick layer of gold prior to

SEM examination.

3. RESULTS AND DISCUSSION
3.1. Joining MgF2 using sodium silicate solution

For the sodium silicate bonding agent, the MgF2/MgF2 joining was successful at
700°C using both microwave and conventional heating. For both the microwave and
conventional heating, dead weight loading was used during joining (Table 13).
The mean bond layer thickness for the MgF2/MgF2 specimens was 10 microns for the
microwave-joined specimens and 15 microns for the specimens joined in the
conventional furnace. The microwave and the conventionally-joined
MgF2/MgF25pecimens showed very different microstructures near the bond layer (Figure
30). Microstructural differences between the microwave—joined and the conventionally-
joined MgF2/MgF2 specimens may be related to the nature of microwave heating itself.
The coupling of a material with microwave energy is a function of both the material and
the temperature of the material, but if the material does absorb microwave power, heat

can be generated within the specimen [4,5]. In this case, the microstructural differences

67

Table l

Table 12. Type of filters and detector for various wavelengths.
Wavelength range Types of filters Types of detectors

 

 

400-650nm Transparent filter 813—SL
650-840nm Red filter 813-SL
850-1650nm Black filter 800-IR

 

Table 13. The conditions of joining both microwave and conventional furnace.

 

 

Heating method Tmax(°C) Hold time(min) Dead Heating rate
weight(gm) (°C/min)
Conventional 700°C 20min 85 gm 10°C/min
Microwave 700°C 20min 60gm 45°C/min

 

68

til-gills 30
C01“ Clilll‘l
reaction it
tilt'lerence

Future stu

\lticrs ll
Sitiill‘ltll
Shipped d
material,
bond by;

The I
R L'L‘zflsn

‘0 the [R

i 1
”an int
Trr‘ a"

Lei-A53”,
Lb GI\£U

3‘1 Cal

MgF2 Us

(Figure 30) may be a function of differing chemical reaction kinetics between the
conventional and microwave heating (hence forming differing sodium silicate/MgF2
reaction layers). Altematively, the microstructural differences may be linked to
differences in microwave and conventional heating at and near the interfacial region.
Future studies should address these points.

A rough estimate of the bond-layer toughness was obtained by placing a 98N
Vickers indentation crack near the bond-layer of a sodium silicate bonded MgF2/ MgF2
specimen (Figure 31). For the microwave joined MgF2, the Vickers indentation crack
stopped at bond-layer, indicating the bond-layer material is not as strong as the matrix
material. Similar results were obtained for Vickers indentation cracks placed near the
bond layer for conventionally-joined specimens.

The IR transmittance of the polished and heated (but not coated) specimens and the
IR transmittance of the polished coated, and heated (but not joined) were nearly identical
to the IR transmittance of the polished specimens (Figure 32). Thus, (1) heating without
coating, (2) coating, heating and not joining gave an IR transmittance that was very
similar to the as-polished specimens, and therefore heating and coating alone do not seem
to significantly affect the IR transmittance of the MgF2 specimens.

The microwave joined MgF2/MgF2 shows a considerably higher IR transmittance
than the conventionally joined MgF2/MgF2 specimens (Figure 33). The differences in
transmittance lead to differences in the calculated optical absorption coefficients, as will
be discussed in Section 3.2 of this paper.

3.2. Calculation of Absorption Factor for microwave and conventionally joined

MgF2 using sodium silicate solution

69

ll) rim

; Mng
E
F

Bond
layer

4‘

 

(b)
Figure 30. SEM micrographs of both the microwave and the conventionally joined
MgF2 showing the bond-layer. (a) Microwave joined MgF2 (joining at 700°C for 20min
with 60gm weight loading), (b) Conventional joined MgF2 (joining at 700°C for 20min

with 85gm weight loading).

70

 

“Sure
SQlill‘il’l

\
\

Vicker's
indentation
crack

 

Figure 31. SEM image of a Mngl MgF2 specimen joined using the sodium silicate
solution. The specimen was joined by microwave heating at 700°C for 20min with a
60gm dead weight loading.

71

TU tll

ltltt‘tttlt the t

1113‘ used. it
liltll‘dlltlll. (t

ll ll

“heft R l\
Gptlt’il ind,
lite ratio I

it normal

 

To determine the optical absorption factor, or, for both the conventionally and the
microwave joined MgFgl MgF2 specimens, the Lambert-Bouger law [6]

17- = 1, em (1)
was used, where IT is intensity of transmitted radiation, his the intensity of the incident
radiation, (1 is absorption factor, and x is the specimen thickness.

If the optical reflection at each interface is taken into account, 17/11 is given by
17/1, = (1-R)"'e'“‘=T (2)
where R is the reflectivity for normal incidence angles. R is a function of the
optical index of reflection, n, and m is the number of interfaces for the specimen.
The ratio IT/I, is the transmittance, T. For a planar slab, m = 2.

At normal incidence, R, is given by

("—1)2
R: 3
(n+1)2 ( )

 

Solving for optical absorption factor, or, gives

1 r
=——l ————— 4
a xn(l-R)m U

The optical properties of bond-layer are unknown, but we estimated a by assuming a
zero thickness bond-layer in equation (2) for the joined MgF2. For Mng/ MgF2
specimens, m is set equal to 3 (Figure 34).

An optical absorption factor, a, was calculated from equation (4) using an optical
index of refraction, n, of 1.3749 at a wavelength 1500nm [7]. The calculated optical
absorption coefficients, a, were similar for the polished (unjoined) and for the

microwave-joined

44“ ”1.4:.

r-‘

F’illlre 3
(0131‘ sntl.

 

  

 

 

 

 

  

 

 

 

1.0 - i .
* EOllfhsd Conventional *
.__.__._. a .

8 0-3 e 6 Heating
:: . ------ Coated
g 0.6 ,. —-—- Jomed
0-
E t
m ..
= 0.4 .1"
a i
L:
F-t 0.2: /

0.0 A I L ' I l i l i i l i l I l L

800 900 1000 1100 1200 1300 1400 1500 1600 1700
Wavelength (nm)
(a)
1.0 ' l l T. l T i i ' l T l l 1
~ _ POhShed Microwave

8 0-8 r —_ Heated Heating
= . ------ Coated
E
m _
r: 0.4
cu
:-

0.0 L 1 1 i 1 I 1 l J A 1 I l l i i

800 900 1000 1100 1200 1300 1400 1500 1600 1700

Wavelength (nm)
(b)

Figure 32. Transmittance for both microwave and conventionally heated MgF2. (a)
Conventionally-heated MgF2 and Mngl MgF2 joined and (b) Microwave-heated MgF2
and Mngl MgF2 joins.

73

33. Cut
it are ”

~WO~V~L
A:
3.1-v

a .

F-

09

 

 

 

 

 

 

 

 

 

 

1.0 . l . . l . l ' 1 1 - . r a T f
/-_-____ ”—l—_ _ o \

Q 031' —9— MW—Jorned ‘ -

E 1] —<—>~— CV—Joined j

§ 0.6~ 7”“ ’ “ ”‘ r”

.-

E

2 0.4~ -1

a

h

E" 0.2:» _
0.0‘L1L11124L1111111
800 900 1000 1100 1200 1300 1400 1500 1600 1700

Wavelength (nm) Figure

33. Comparison of transmittance for microwave and conventionally joined MgF2(From
Figure'32).

74

Figure .
Sptt‘tme

 

 

 

 

 

 

 

 

 

 

 

V
111

Figure 34. As schematic showing the reflections at the various interfaces of a two-layer

specimen (which gives m = 3).

Table 14. Calculation of the optical absorption factor, a.

 

 

Processing Measured T R x(mm) a(mm")
Polished 0.703 0.025 2.029i‘0.002 0.149

MW-joined 0.486 0.025 4.1341‘0009 0.156

CV-joined 0.229 0.025 4209:0012 0.342

 

75

and iht
{ON ‘3"
Indugc
mqiddl
the mi}
N EEC
center a
£3 < H
ranem
mile 1
mic-mm
mrrons
themes;

micron r

specimens (0.149mm'l and 0.156mm", respectively Table 14). The a values for the
polished (unjoined) and the microwave-joined specimens implied that the polished MgF2
and the microwave-joined MgF2 had a similar optical quality. The higher a for
conventionally-joined MgF2 indicated optical losses resulting from IR scattering [8,9]
induced by crystallographic phases formed in and near the bond region (Figure 30).

The scattering of electromagnetic radiation depends on a number of factors,
including (1) the mismatch in optical indices of refraction between the scattering center and
the matrix material, and the (2) the relative size of the scattering center, compared to the
wavelength of the incident radiation [9, 10]. If we let a = characteristic size of the scattering
center and A = the wavelength of the incident light, then for scattering is significant for 0.1 <
Na < 10, with a maximum in scattering at about Ma = 1. As an example of how effectively
scattering can reduce the transmitted light intensity in a ceramic, consider the scattering of
visible light by pores in a ceramic [10]. At the wavelength of red light (700 nm or 0.7
micron), a 3 percent volume-fraction porosity consisting of pores with a diameter of 2
microns reduces the transmittance to about 0.01 % compared to the transmittance for a
theoretically-dense material [10]. When the pores have a diameter of 0.7 micron, then 0.7
micron radiation will be reduced by a similar amount for a volume fraction porosity of only
one-percent [10]. From Figure 30, we see that the characteristic dimension of the scattering
centers (second phase particles) is roughly a few microns. Thus, second phases that are
several micron diameter should scatter the incident radiation used in this study (a
wavelength range from 880 nm to 1600 nm, or equivalently 0.88 micron to 1.6 micron),
since Na should be very roughly in the range from about 0.1 to 0.5. However, a detailed

scattering study for non-spherical particles that are not uniform in size is extremely complex

76

 

{5.101, '

themes

33. Di;

the 51hr
rpm ant
adtddt
0f fiou;
maniC§

“Will

111 flow

[8-10]. Therefore, attempts to extract further information from the transmittance curves

(Figures 32 and 33) will be a topic of future study.

3.3. Diamond joining on MgF2 using sodium silicate solution

Diamond/MgF2 joining was attempted for both the sodium silicate solution and for
the silica film. For the sodium silicate bonding agent, the films were spun between 3000
rpm and 5000 rpm from for 20 seconds. Then, using a heating rate of 100 C per minute and
a dead weight loading of 85 gm, the diamond/MgF2 specimens were heated in a atmosphere
of flowing nitrogen at maximum temperatures of 5000C to 8000C for hold times of 20
minutes. However, the joining was not successful at any of these conditions. In separate
experiments involving single-slab specimens of magnesium fluoride, it was found that even
in flowing nitrogen, the specimens were discolored at temperatures above 800°C,
presumably due to point defects that evolve during heating. Due to the degradation of the
magnesium fluoride's optical properties, no joining attempts were made at temperatures
higher than 800°C.

In addition to the sodium silicate solution, the silica film was used to attempt both
MgF2/MgF2 bonding and the diamond film/MgF2 bonding. The specimens were heated by
a conventional furnace at maximum temperatures between 5000C and IZOOOC in an
atmosphere of flowing nitrogen. Dead weight loads ranged from 20 gm to 85 gm. The
MgF2/MgF2 bonding occurred only at for a maximum anneal temperature of lZOOOC, but at
that temperature (despite the flowing nitrogen atmosphere) the surface of the MgF2
specimen was milky-colored (likely due to oxidation). The resulting transmittance of the

MgF2 specimen was severely degraded by the hi gh-temperature reaction.

77

ha bet
aROH

'\ 41*
t0mul

4. C0

smut:
¥?hfl
mECl'O‘t
the IR
Ptcz'm
acorn
bfilll €€

fikkd

Rm
1. rr

Had the joining been successful, the silicon substrate (on which the diamond film
has been microwave-plasma deposited) would have been etched away using a nitric acid or
a KOH etching solution. The transmittance of the specimen, without the silicon substrate,

could then have been measured.

4. CONCLUSIONS

The MgF2/MgF2 specimens were successfully joined using: (1) spun-on sodium
silicate interlayers, (2) both conventional and microwave heating and (3) low externally
applied pressures. The microstructure of the bonded region was quite different for the
microwave and the conventionally bonded MgF2/MgF2 materials (Figure 30). In addition,
the IR transmittance of the microwave-bonded specimens, as determined from the optical
absorption factor, on, was similar in optical quality to the polished but unjoined MgF2
specimens (Table 14). In contrast, the on for the conventionally-joined MgF2/MgF2 indicates
a considerably degraded optical transmittance (Table 14). The difference in optical quality
between the microwave-joined specimens and the conventionally—joined specimens may be

related to IR scattering induced by second phases [8,9] in the bond region (Figure 30).

REFERENCES

1. KY. Lee and ED. Case, “Microwave Joining and Repair of Ceramics And
Ceramic Composites,” Ceramic Engineering and Science Proceedings, 21‘"
Annual Cocoa Beach Conference and Exposition on Composites, Advanced
Ceramics, Materials and Structure, Cocoa Beach, Florida, V18 543-550, 1997.

2. KN. Seiber, K.Y. Lee, and ED. Case, “Microwave and Conventional Joining of
Ceramic Composites Using Spin-On Materials,” Proceeding of the 12:}: Annual

Technical Conference Dearbom, MI, 941-949, 1997.

3. T.F. Yen, Y.H. Chang, D.L. Yu and ES. Yen, “Diffusion Bonding of MgF2
Optical Ceramics,” Material Science and Engineering, A147 309-321, 1991.

78

4.llL
lllr‘ :ll

5.1..1!
.llu'lr‘i

6 ill

lmm

l. “(If
Hugh

8 P.[k

9. CF.
PM»

erV.[
3nd!

 

 

10.

ML. Santella, “A review of Techniques for Joining Advanced Ceramics,” Journal of
the American Ceramic Society Bulletin, 71[6] 947-953 (1992).

I. Ahmad and R. Silberglitt, “Joining Ceramics Using Microwave Energy,”
Materials Research Society Proceedings, 314, 119-130, 1993.

I.W. Donald and P.W. Mcmillan, “Review Infrared Transmitting Materials,”
Journal ofMaterials Science, 13, 1151-1176 (1978).

. Moses “Refractive Index of Optical Materials in the Infrared Region”; pp7-16,

Hughes Aircraft Company, Culver City, CA 1970.

. P. Debye, HR. Anderson, and H. Brumberger, Journal of Apply Physics., 28: 679, 1957.

CF. Boren and DR. Huffman, Chaper 3 in AbsoLption and Scattering Light by Small
Particles, John Wiley and Sons, New York, 1983.

W. D. Kingery, H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics,
2nd Edition, pp 674-677, John Wiley & Sons, New York, NY (1976)

79

ll}.

.le

can

CWT:

LA

“(15;
0} ll]:

\Cfill'.’

MR1

3.2 Microwave Joining of Particulate Composites

MICROWAVE JOINH\1G OF PARTICULATE COMPOSITES
J. G. Lee and E. D. Case
Materials Science and Mechanics

Michigan State University
East Lansing, MI 48824

ABSTRACT

Using various spin—on materials with submicron initial coating thickness, oxide
ceramics have been joined using microwave heating. Four different ceramic material
combinations were joined in this study, where either one or both of the materials in a
given pair was a particulate ceramic composite. The microstructure of the bond region
was examined using Scanning Electron Microscopy. This study builds upon earlier work
by the authors and co—workers, in which a variety of non—composite polycrystalline

ceramics were joined.

INTRODUCTION
Recently, there has been considerable activity in bonding ceramics with metals [1-
5], bonding ceramics with other ceramics [6 - 10] and bonding ceramics with polymers [ll
-12]. A number of researchers have used brazing techniques to join metals with ceramics
such as silicon nitride [l], sialon [2], molybdenum disilicide [3], and aluminum nitride [4].
In addition, ceramics have been joined to ceramics using brazing, where a metallic

interlayer is used to join the ceramic components. However, if ceramics are joined via

 

J. G. Lee and E. D. Case, “Microwave Joining of Particulate Composites”, Advances in Ceramic Matrix
Composites V, Ceramic Transactions Volume 103. American Ceramic Society, Inc., Westerville, OH, pp.
57 1-581 (2000)

80

 

hazing

iron.

comm:

BACK'

liming
lump
313mm
militia

3011; ll

brazing the final bond thickness can be 40 to 50 microns thick or more [5].

In contrast, work done by the current authors and co-workers [6 - 10] has focused on
minimizing the bond area thickness using spin-on interlayers. The joining work has
included joining of the following particulate ceramics; (1) alumina/zirconia- 3 mol% yttria,
(2) SiC platelet reinforced alumina/glass ceramic, (3) hydroxyapatite/glass ceramic, and (4)
zirconia - 3 mol% yttria/glass ceramic, where the glass ceramic used was MaCorTM, a

commercial glass ceramic reinforced by mica platelets.

BACKGROUND
Microwave heating of ceramic materials gives efficient, volumetric, and rapid
heating [13-15]. The authors and co-workers have used microwave heating to join a
number of polycrystalline ceramic materials, including the following pairs: alumina with
alumina, zirconia with zirconia, MaCorTM with MaCorTM, hydroxyapatite with
hydroxyapatite, and magnesium fluorite with magnesium fluorite [6 - 10]. In ceramic-
cerarnic joining in these previous studies [6 - 10] produced high-quality joins. For example,
the interfaces of the alumina-alurnina joins were relatively tough, as indicated by the fact by
Vickers indentation cracks placed near the joined interface did not deflect as they transited
the interface [6]. For the joined IR transmitting material magnesium fluoride, the infrared
transmittance was especially unchanged by the presence of the interface for microwave
joining [8]. However, the joining in the earlier studies [6 - 10] was limited to like
ceramics, and none of the joined ceramics were composites. This study extends the

previous work to include the joining of particulate-reinforced ceramic materials.

81

 

EXP!

ml
[filer
mm]
mdlt‘
Pll‘ll
using

mitts

in a 3

'ili‘hl;

‘ 3
3 ”)0
P's

UCU

Wags,

EXPERIMENTAL PROCEDURE

The ceramics joined in this study included two polycrystalline materials, namely
(a) ZrOz -3 mol % YZO3 and (b) hydroxyapatite (HAP) and three particulate-reinforced
ceramic composites, including (a) Ale3/ZrOz particulate composites, (b) a mica platelet
reinforced glass ceramic (MaCorTM) and (c) SiC platelet reinforced alumina. Of these
materials, the ZrOz -3 mol % Y203, the hydroxyapatite (HAP), and the A1203/Zr02
particulate composites were sintered as part of this study, with all sintering being done
using commercial powders. The sintering, as described below, was done either using a
microwave cavity or a conventional tube furnace. The mica platelet reinforced glass
ceramic (MaCorTM) and SiC platelet reinforced alumina were commercial materials which
were received in billet form.

Prior to the sintering that was performed as part of this study, all specimens were
hard die pressed at 32 MPa in a uniaxial die, resulting in disc—shaped specimens with
green dimensions of approximately 2.0 cm in diameter and 2.0 mm thick.

Seven powder compact specimens of the zirconia - 3 mol % yttria powders (TZ-
3YS, Tosoh Ceramic Division, Bound Brook, New Jersey ) were sintered simultaneously
in a 2.45 GHz single mode microwave cavity at 1375°C for 1 hour. The densities of
sintered zirconia were about 97 % of the theoretical density, as measured using
Archimedes method.

For the 85 wt % alumina (TMDAR, Taimi Chemicals Co. LTD)/ 15 wt % zirconia
- 3 mol % yttria powders (TZ-3YS, Tosoh Ceramic Division, Bound Brook, New Jersey )

particulate composites, the powders were prepared by milling the alumina and zirconia

powders together for 24 hours using alumina grinding media in a plastic ball mill

to produce uniform alumina/zirconia powder mixture. The alumina/zirconia compact
powders were microwave sintered at 1350°C for 1 hour using a heating rate of
approximately 10°C/min.

Unlike the alumina/zirconia composites and the zirconia specimens included in
this study, the calcium hydroxyapatite (Ca10(OH)3(PO4)6) powder (CeracTM Incorporated
Specialty Inorganics) was sintered in a conventional furnace at 1300°C for 11 hours. The
heating and cooling rate was approximately IOOC/minute.

The as-received billets of MaCorTM (Corning Code 9658, a flurophlogopite mica
platelet reinforced glass ceramic) and billets of the as-received 80 vol% A1203 / 20 vol %
SiC platelet composites (Max Tech, Inc, Lansing MI) were cut using a low speed
diamond saw. The final specimen dimensions were approximately 1 cm x 1 cm x 0.1 cm
for the MaCorTM and about 1 cm x 1 cm x 0.5 cm for the 80 vol% A1203/ 20 vol % SiC
platelet composites.

All specimens used in this study were polished using an automatic polisher at
1200 rpm with a series of diamond paste grit of size ranging from 35 pm to 1 pm. After
polishing, each of the specimens were cleaned first with acetone and then cleaned again
in deionized water using an ultrasonic cleaner.

After polishing and cleaning the specimens, the silica coating was applied to the
specimens. An organic liquid (SilicaFilm, Emulsitone Company, Whippany, New Jersey)
was utilized for the spin-on interlayer. Several drops of silica film on the specimens were
deposited on the specimen surface using a pipette. The coatings were spun at 3000 rpm
for 20 seconds, then coated specimens were cured in a conventional furnace at 200°C for

20 minutes. When heated, the SilicaFilm is pyrolyzed to an amorphous silica coating.

83

 

lit t‘ur
will [ht

For
nm [hr
.lerOt'

For
letting
elector
“Std d't
“AR 0‘
Smirllg
ill to

and Q

‘ACCU'

‘m (11;

. ll ‘
dame

The cured silica coating is thin and uniform (Figure 35a), where the film thickness varies
with the spin rate (Figure 35b).

For the coated materials in this study, the cured silica coating was about 100 to 200
nm thick. Figure 35 a is a SEM micrograph of a 130 nm thick silica coating on one of the
MaCorTM specimens used in this study.

For all of the joining work included in this paper, the specimens were joined by
heating in a 2.45 GHz single-mode microwave resonant cavity using the TM] H
electromagnetic mode. A cylindrical zirconia refractory specimens enclosure (casket) was
used during joining (Figure 36a). The refractory casket consisted of a top and bottom
plate of an aluminosilicate fiber board (SALI, Zircoa) and a hollow yttria-partially
stabilized zirconia cylinder (ZYC, Zircoa). The refractory casket serves two purposes:
(1) to serve as a microwave susceptor to aid in heating the specimens at low temperatures
and (2) to provide thermal insulation. Within the casket, the specimens were placed on a
aluminosilicate setter (Figure 36b) and alumina dead weights of mass ranging between 20
gram to 50 gram were used to apply a normal load during joining (Figure 36b).

The specimen temperature was measured via an optical pyrometer system
(Accufrber Optical Fiber Thermometer, Model 10, Luxtron Co., Beaverton, OR). A 0.5
cm diameter circular hole in the zirconia-cylinder section of the casket (Figure 36b)
allowed the optical pyrometer to be sighted on the specimen. The joining temperature
was controlled by input power where the maximum input power ranged from about 400
watts to 900 watts. The heating rate was approximately 40°C/min for each experimental

I'UI'I.

After joining, the joined specimens were sectioned using a low speed diamond

84

 

 

Silica film —>

 

 

.=> .6
Q \l

.
l

Film
Thickness(micron)

9
u

9
N

 

 

 

1000 2000 siloo
Spin Rate (rpm)

c
u
i—

(b)
Figure 35. (a) SEM micrograph of a cured silica film on a MaCorTM substrate, (b) as a
function of spin rate for cured silica coatings on MaCorTM [after 6]

85

Alum_irra ‘
“ SALI

Zirco_n_ia
ZYC

p66

. i " . Dead weight
<_— Specimen'ts

 

(b) i

Figure 36. (a) Photograph of the zirconia/aluminosilicate refractory casket during
specimen joining, (b) Photograph of the bottom SALI plate of the refractory casket with
the specimen, specimen setter and dead weights in place.

86

it‘d

Silt

lei

t4;

Cm

mitt

saw to allow SEM observation of the bond region. The specimens were mounted in
dially pthlate and then polished with a series of diamond grit. The surfaces were made
sufficiently conductive for SEM examination by sputtering a 7 nm thick coating of gold
onto the specimen surfaces. In conjunction with the SEM analysis, x-ray line scans and

elemental mapping was performed.

RESULTS AND DISCUSSION

For the pairs of ceramic materials joined, either one or both of the materials joined
was a particulate reinforced ceramic material. The ceramic materials that were
successfully joined include: (1) alumina-zirconia composites joined with alumina-
zirconia composites, (2) MaCorTM joined with ZrOz, (3) HAP joined with MaCorTM, and
(4) SiC platelet reinforced alumina with MaCorTM. Of these materials, the particulate
composites include the alumina-zirconia composites, MaCorTM, and the SiC platelet
reinforced alumina.

The alumina/zirconia particulate composite specimens were joined using
microwave heating at 1400°C, 1450°C, and 1500°C for 20 minutes. During joining, a
deadweight of approximately 50 grams was applied via sintered alumina disks placed on
top of the specimen. As viewed in the SEM micrographs, no bond phase is apparent, and
if a bond phase exists, it must be less and one urn thick (Figure 37). Quasi-elliptical
pores are occasionally observed at the interface (Figure 37).

MaCorTM was joined with zirconia - 3 mol % yttria upon heating at 1020°C for
20 minutes using a 20 gram deadweight loading. In contrast to the specimens of

alumina/zirconia composite joined with alumina/zirconia composite (Figure 37), a bond

87

 

liter

trill:

mat
1050

liaise

layer is visible at the MaCorTM /zirconia interface, where the bond region is on the order
of 0.5 - 1 micron thick (Figure 38). Numerous micron-scale pores were visible along the
interface region between the MaCorTM /zirconia interface.

Hydroxyapatite was successfully joined with MaCorTM at 1020°C for 10 minutes
using a 20 gram deadweight (Figure 39). As was the case with the alumina/zirconia
composites specimens joined with alumina/zirconia composites (Figure 37), a bond
phase at the interface was not apparent in the SEM (Figure 39). In Figure 39, the mica
platelet reinforcement in the MaCorTM is evident to the right of the indicated interface.

MaCorTM has a relatively low use temperature material, with a vendor-specified

maximum use temperature of IOOOOC. The attempted joining of HAP with MaCorTM at
1050°C and 1070°C using a 50 gram deadweight failed since the MaCorTM melted under
those conditions, when in contact with the HAP.

Polycrystalline alumina reinforced with 20 vol% SiC platelets (Max Tech Inc.,
Lansing, MI) was joined with MaCorTM by heating at 1020°C for 20 minutes using a
deadweight loading of about 20 grams. As was the case with joining HAP with MaCorTM,
attempts at joining the alumina/SiC platelet composite with MaCorTM at 1050°C failed
since the MaCorTM partially melted. After joining, a 10 micron-thick bond layer was
observed at MaCorTM-alumina/SiC platelet interfaces (Figure 40). Since the thickness of
the original silica coating layer (before joining) was roughly 130 nm, the 10 micron thick
bond layer thickness likely represents a reaction zone between the SiC platelet reinforced
alumina and the MaCorTM. Thus, only one of the four ceramic/ceramic material

combinations included in this study formed a significant reaction zone during the joining

process.

88

 

Figure

millUI

/

— Interface

 

Figure 37. SEM micrographs of alumina/zirconia composites joined at 1450°C for 20
minutes showing interface.

89

 

M 3C or

Bond
g layer

1 run

 

Figure 38. SEM micrographs of zirconia (3 mol % Y203) with MaCorTM joined at
1020°C for 20 minutes showing interface.

90

M aC‘or

<— Interface

 

Figure 39. SEM micrograph of HAP with MaCorTM joined at 1020°C for 10 minutes.
The mica platelet rainforcement in the MaCorTM is evident in the micrograph.

9|

Alest/SiC ;

‘ Bond
layer

 

Figure 40. SEM micrograph of SiC platelet reinforced alumina with MaCorTM joined at
1020°C for 10 minutes.

92

 

she
Cllllt

the 1

C0.)

.9413,
:l ,- ..
:40.)
lle.
. 11,5de
Wt“; m,
if"
V’llfd \

inl’r
. “5 lit l.

X-ray line profiling of the composition of the bond region and the near-bond
region was performed for each of the four ceramic material pairs joined in this study.
However, for each of the joins (except the MaCorTM-alumina/SiC platelet joined
interface, Figure 40) no “bond layer” could be identified, which was to be expected since
the radius of the interaction volume [16] is typically large enough that the compositional
information from the x-ray line profile is averaged over subsurface volumes that are
several microns in radius. Thus the composition differences on the size scale of the sub-
micron bond region (if present) would be impossible to detect with the x-ray line
profiling. However, the preliminary x-ray line scan measurements of the bond region in
the specimens consisting of alumina/SiC platelet composite joined with MaCorTM did
show evidence of the reaction zone (for example, the concentration of aluminum ions was
enhanced near the MaCorTM/reaction zone interface). Also, an elemental map indicated

the possible presence of Mg and K rich precipitates in the reaction zone.

CONCLUSIONS

In this study, we joined the following particulate ceramics: (1) alumina/zirconia
particulate composites with alumina/zirconia composites (2) a mica platelet reinforced
glass ceramic (MaCorTM) with zirconia -3mol % yttria, (3) MaCorTM with polycrystalline
hydroxyapatite and (4) SiC platelet reinforced alumina with MaCorTM. Except for the
inter‘face was relatively free of porosity. For the alumina/zirconia particulate composites
S13>eCimens in which the alumina/SiC platelet was bonded to MaCorTM, the region near the
Joined with alumina/zirconia composites and the MaCorTM joined with polycrystalline

hydroxyapatite, no bond layer was apparent on the SEM micrographs.

93

ill:

En

clc

S

In future work, Vickers indentation will be used to aid in estimating the nature of the
residual stresses near the join. In addition, Transmission Electron Microscopy (TEM) will
be performed in order to further analyze the interface region of the joined specimens. For
the reaction layer alumina/SiC platelet composite - MaCorTM specimens, additional x-ray

profiling of the bond layer compositions will be performed in order to better characterize

the nature of the interface regions.

ACKNOWLEDGMENTS

The authors acknowledge the financial support of the Composite Materials and
Structures Center and the Electronic Surface and Properties of Materials Center, College of
Engineering, Michigan State University. The authors also acknowledge the use of the

electron microscopy facilities at Michigan State University’s Center for Electron Optics.

REFERENCES
l . G. Chaumat, B. Drevet, and L. Vernier, "Reactive brazing study of silicon nitride to a

metal joining", J. European Ceram. Society, 17[15-16]: 1925 - 1927 (1997).

2. A. P. Xian, "Joining of sialon ceramics by Sn-5 at% Ti based ternary active soders", J.
Materials Science, 32[23]: 6387 - 6393 (1997).

3. S. D. Conzonne, D. P. Butt, A. H. Bartlett, "Joining of MoSiz to 316L stainless steel",
J. Mater. SCI, 32[13]: 3369-3374 (1997).

4- D. Huh and D. H. Kim, "Joining AlN to Cu using In-base active brazing fillers", J.
Materials Res., 12[4]: 1048 - 1055 (1997).

5- I... Esposito, A. Bellosi, S. Guicciardi, G. de Portu, "Solid state bonding of A1203 with
Cu, Ni, and Fe: characteristics and properties", J. Mater. Sci., 33[7]: 1827 - 1836
(1998).

6- K. N. Seiber, K. Y. Lee, and E. D. Case, "Microwave and Conventional Joining of
Ceramics using Spin-on Materials,“ pp. 941-949 in Proceedings of the 12th Annual
Advanced Composites Conference, Technomic Publishing Co., Lancaster, PA (1997).

94

7.

10.

ll.

12.

13.

14.

15.

16.

E. D. Case, K. Y. Lee, and J. G. Lee, “Joining of Polycrystalline Ceramics and
Ceramic Composites Using Microwave Heating,” pp. 17 - 20 in Proceedings of the
33rd lntemational Microwave Power Symposium, lntemational Power Institute,
Manassas, VA (1998).

E. D. Case, J. G. Lee, and K. Y. Lee, “Joining of Optical and Infrared Materials
Using Spin-On Layers”, pp. 17 - 26 in Joining of Advanced and Specialty Materials,
M. Singh, J. E. Indacochea, and D. Hauser, eds., ASM lntemational, Materials Park,
OH (1998).

E. D. Case, K. Y. Lee, J. G. Lee, and T. Hoepfner, “Geometrical Stability of Holes
and Channels During Joining of Ceramics and Ceramic Composites,” pp. 27 - 34 in
Joining of Advanced and Specialty Materials, M. Singh, J. E. Indacochea, and D.
Hauser, eds., ASM lntemational, Materials Park, OH (1998).

J. G. Lee, K. Y. Lee and E. D. Case, “Joining of Diamond Thin Films to Optical and
IR Materials,” pp. 509 - 520 in Ceramic Transactions, Volume 94, American
Ceramic Society, Inc., Westerville, OH (1998).

C. Mukherjee, E. D. Case and A. Y. Lee, “Thin, Protective Silica Layers on
Polymeric Materials”, Accepted for publication, Ceramic Eng. and Sci. Proc., Volume
20(1999).

C. Mukherjee, E. D. Case, K. Y. Lee, and A. Lee, “Silica Coatings on BMI Polymeric
Substrates,” submitted, Journal of Materials Science.

K. Y. Lee and E. D. Case, “Steady-State Temperature of Microwave-Heated
Refractories as a Function of Microwave Power and Refractory Geometry,” Accepted
for publication, Materials Science and Engineering.

K. Y. Lee and E. D. Case, “Microwave Sintering of Alumina Matrix Zirconia
Composites Using a Single-Mode Microwave Cavity,” J. Mater. Sci. Lett., 201-203
(1999).

K. Y. Lee, P. H. Dearhouse, and E. D. Case, “Microwave Sintering of Alumina Using
Four Different Single-Cavity Modes,” Journal of Materials Synthesis and Processing,
7[3] Page numbers not yet determined (1999).

P. J. Grundy and G. A. Jones, pp. 26 - 36 in Electron Microscopy in the Study of
Materials, Crane Russak, Publishers, New York, NY (1976).

95

3.3 Geometrical Stability of Holes and Channels During Joining of Ceramics and
Ceramic Composites

Geometrical Stability of Holes and Channels During Joining of Ceramics and Ceramic
Composites
E. D. Case, K. Y. Lee, J. G. Lee, and T. Hoepfner
Materials Science and Mechanics Department,
Michigan State University

East Lansing, MI 48824

Abstract

An ultrasonic mill has been used to machine precise channels and holes in
ceramic components that are subsequently joined via microwave heating. Microscopic
inspection of the channels then indicates the stability of the channels during the
microwave joining process. The information gained from joining such specimens can be
crucial in the development of techniques for joining ceramics with complex geometries,

including channels for fluid flow in intricate components that are joined from simpler

ceramic subcomponents.

 

E. D. Case, K. Y. Lee, J. G. Lee, and T. Hoepfner, “Geometrical Stability of Holes and Channels During
Joining of Ceramics and Ceramic Composites,” Joining of Advanced and Specialty Materials, M. Singh, J.
E. Indacochea, and D. Hauser, eds., ASM International, Materials Park, OH. pp. 27 — 34 (1998)

96

Introduction

This paper deals with joining densified ceramic bodies that include channels or
holes cut into one or more of the subcomponents. The need for such a technique is
related to the nature of ceramic processing, which often involves (1) a difficulty in
processing components of complex geometry and (2) the considerable shrinkage that is
typical during densification.

This paper discusses and compares work done by Case and co-workers on joining
polycrystalline alumina and MaCorTM [1, 2] with more recent work in joining

polycrystalline zirconia and hydroxyapatite.

Background
Ceramic Processing: Difficulty in Forming Complex Shapes

Ceramic components are typically processed using a powder processing
technique, that is, ceramic powders are formed into a shape, then the powders are
densified at temperatures corresponding to homologous temperatures, TH, of 0.6 or
greater. TH may be defined as the ratio TAMBIENT/T Mp, where the ambient temperature,
TAMB 15m, and the melting point temperature, TMp, are both expressed in units of degrees
Kelvin. In some processing methods, the shape forming and densification steps are
combined (such as in hot-pressing or hot-isostatic pressing), but typically the shape-
forming and densification steps are performed in sequential fashion [3].

The shape-forrning techniques often used in industry often involve either a
pressing operation or slip casting. Pressing powders in a die usually required that the

processed shaped be relatively simple disk or plates [3]. Rods of constant cross-section

97

may be extruded from a die [3]. Slip casting allows a greater range of geometry, but still

the shapes produced are limited and the introduction of precise channels or holes within

the part is problematic.

Components With Complex Geometry: Fabricated From Ceramic “tapes”

A technique that does offer more flexibility in terms of component geometry is
that of building-up a complex shape for tape-casting ribbons. The unfired tape-cast
materials are typically quite thin (on the order of one millimeter or less in thickness), and
when cast with a high fraction of a plastic binder phase, the unfired tape-cast elements
can be cut and compacted into a shape of relatively complex outer geometry.

Organic binders are widely used in the ceramics industry. Organic binders, typically
at the level of several volume percent, are added to increase the green (unfired strength) of a
variety of components pressed from ceramic powders [3]. In the electronics industry, large
volumes of organic binders and solvents are added to ceramic powders resulting in "tapes"
that are flexible in the green state. Cutting and hole punching operations then are relatively
straightforward in the flexible tapes, enabling the fabrication of multilayer ceramic
capacitors and piezoelectrics. In addition, 50 volume percent or more of binder and
lubricants are added to ceramic powders for injection molding of ceramic components.
Difficulties with binder burnout in ceramic tapes.

The benefits of adding organic binders to ceramic powders are at least partially
offset by several disadvantages. The problems inherent to having a large volume fraction
of binder in a ceramic compact focus on binder burnout, where the burnout process (1)

can be very time consuming, (2) binder burnout can leave chemical residues that lead to

98

flaws and defects in the final sintered ceramic, and (3) the high shrinkage associated with
binder burnout can lead to mechanical stresses that cause cracking.

When the volume fraction of the binder phase is high and/or the ceramic
components are thick (such as is the case for electronic components from multilayers of
binder—laden ceramics) then the time required for binder burnout can be very long. For
example, in multilayer ceramic piezoelectrics which may consist of 150 layers or more
of ceramic “tape” [4] or mutilayer ceramic capacitors [5], delaminations (cracks)
between the layers can occur at extremely low heating rates. Heating rates on the order
of 0.015°C per minute, corresponding to a total binder burnout time of 250 hours
(approximately 1.5 weeks) may still be rapid enough to induce delamination between
layers of a multilayer ceramic piezoelectric component [4]. Binder burnout treatments
can require up to several days for injection-molded ceramic components [6].

In addition to the long times required for burnout of multi-layered, binder-laden
ceramic powders, binders also can leave a carbon or sulfur residue [7, 8], which has been
shown to persist even after elevated (1500°C) heat treatments. Many researchers have
linked binder residues with defects in the final, sintered ceramic bodies [9-11].

The high shrinkage inherent to the binder burnout process induces mechanical
stresses which in turn can lead to flaws and cracking in the processed ceramics [12, 13].
It is worth noting that densification for a ceramic compact containing binders is
essentially a two step process in which (1) the binder is first removed from the compact
and (2) the densification of the ceramic powders themselves takes place. That is, after
binder burnout is complete, the sintered powders must still densify. At the point that the

binder is removed (typically at temperatures of a few hundred degrees Celsius), the

99

remaining ceramic powders have a volume fraction porosity of roughly 50 percent.
During densification of the ceramic powders, the component thus must undergo
considerable addition shrinkage upon sintering.

When the fraction of binder is high enough to allow the green (unfired) tape cast
material to plastically deform and therefore form a complex part by building it up layer
by layer, then binder burnout can be a problem. As discussed above, the binder burnout
problems can include: (I) the binder burnout step can be very lengthy for parts
composed of many layers, (2) chemical residues from binder burnout can lead to defects
in the final sintered part, and (3) cracks can form in the part as a result of the very high
shrinkages that are inherent to the binder burnout process. Thus, the plasticity afforded to
green (unfired) ceramics by virtue of their high binder concentrations can come at with a
high cost in terms of processing difficulties. In addition, the act of compacting layer
upon layer to build up the component from tape cast materials might deform precise
channels or holes within the ceramic body, if such channels or holes were desired.

One alternative way to fabricate ceramics of complex geometric shape is to joined

simpler subcomponents of densified ceramics, in order to more the final component.
This can have benefits in terms of quality control (the detection of flaws) and in terms of

allowing one to include precisely machined channels and holes into the ceramic

component.
Ceramic/Ceramic Joining

The joining of ceramic materials can help circumvent some of the difficulties

associated with processing components of complex geometry. The fabrication of

100

complex parts can be accomplished by the joining of simpler subcomponents. In addition
to avoiding the difficulties associated with the burnout of large volume fractions of a
binder phase, a potential benefit of joining simple ceramic subcomponents is that the
detection of flaws in simple subcomponents is easier than the detection of flawed
components of complex shape. Thus, if a flaw is discovered in a subcomponent, it may
be discarded before it is included in the final part.

Often, joining is performed with the assistance of a flux that is placed between the
ceramic pieces that are to be joined. Different joining techniques are distinguished in
part by the (l) magnitude of the externally applied pressure and (2) the joining
temperatures used. Diffusion bonding, for example, uses very high stresses and
temperatures to join ceramics, such that the joining is accomplished under conditions
conducive to creep in that material [14, 15]. In this paper, we use relatively low
externally applied loads to join a variety of specimens, but in each instance we use a spin-

on interlayer to assist in joining.

Experimental Procedure
Materials Used

The materials joined in this study include polycrystalline alumina, hydroxyapatite,
zirconia, and a mica-platelet reinforced glass ceramic. The alumina, hydroxyapatite, and
the zirconia were sintered from commercial powders, while the glass ceramic was a
commercial material, MaCorTM (Corning Code 9658).

The alumina specimens were microwave-sintered from Sumotomo AKP 50

powders using a single-mode microwave cavity, described in the next section. The

101

AKPSO powders have a vendor-specified initial particle size of 0.23 microns and a purity
of 99.95 percent. Alumina is used in a wide variety of electronic and structural
applications of ceramics, due to its high hardness, good resistance to chemical attack, and
low electrical resistivity.

The calcium hydroxyapatite powder (HAP) used in this study (Cerac Inc.,
Specialty Inorganics, Milwaukee, WI, Item #C-2071-1, Lot#Xl6907) had a nominal
purity (as reported by Cerac) of “typically 99% pure”. Calcium hydroxyapatite is an
“bioactive” ceramic material, such that when exposed to blood or to a simulated
biological fluid (SBF), HAP promotes the formation of bone on its surfaces [16]. HAP
coatings have been used to coat biological implant materials. The HAP coatings can act
to reduce corrosion of the metallic substrate materials, since the corrosion can lead to the
release of toxic ions into the body [16, 17].

The MaCorTM specimens were cut from as-received billets of the material. Macor
is a machineable glass ceramic material of the composition SiOz-A12O3-MgO-KzO-F.
The machinability of the material stems from the flurophogopite mica platelets that are
randomly dispersed in the material. Macor is used in a variety of electronic substrate
applications, due to its good dielectric properties and its ability to be precisely machined
by metal-working tools.

The zirconia specimens were formed from 3-mol percent yttria partially-
stabilized zirconia powders (TZ-3Y Grade zirconia powders, Tosoh Ceramics Division,
Bound Brook, New Jersey) with a vendor-specified specific surface area of
approximately 16 mz/gram. The standard vendor chemical assay shows that the major

impurities in the zirconia powders are alumina, silica, iron oxide, and soda, with weight

102

percentages of less than 0.1, 0.02, 0.01, and 0.04, respectively. The initial particle size is
roughly in the 0.1 micron range. Zirconia is used in a wide variety of ceramic
applications. including as grinding media, high-temperature insulation, and in electronic
components.

Two materials were used as bonding agents in this study. “Silicafilm”
(Emulsitone Company, Whippany, New Jersey) is an organically-based silicate, which
upon heating at relatively low temperatures (say, 2000C) yields an amorphous silica film.

The second bonding agent used was a sodium silicate solution

Specimen Preparation
Microwave and Conventional Sintering of the Specimens

The alumina, hydroxyapatite, and zirconia powders were each dry pressed in a
uniaxial, double-acting die at approximately 32 MPa. For the alumina, the nominal mass
of each specimen was about 2 grams, with an as-pressed diameter of about 2 cm and an
as-pressed thickness of about 2 mm. The alumina and zirconia specimens were then
sintered in a single-mode microwave cavity (described below) while the HAP specimens
were sintered in a conventional, radiant energy furnace (also described below).

The HAP specimens were fired at maximum 1300°C for 30-60 minutes in a
conventional tube furnace (MRL Thermtec, with an Eurotherm Controller and SiC
heating elements), at a ramp rate of 10C/ minute. For the alumina and zirconia, the
maximum temperatures ranged from about 1500 to 16000C for the alumina, while the
zirconia and the HAP powders were sintered at about 1450 and 13000C, respectively.

For both the microwave sintering and the microwave joining, the temperature was

measured by an optical pyrometer (Accufiber Optical Fiber Thermometer, Model 10,

Luxtron Co., Beaverton, OR).

Cutting and coating the specimens

The as-received MacorTM billets were approximately 7.8 cm X 7.8 cm X 0.18 cm.
These billets were cut into 1 cm X 1 cm X 0.18 cm specimens using a low-speed
diamond saw. Following sintering, the sintered HAP specimens were sectioned by a low-
speed diamond saw into 12-15 mm square billets which were about 3-5 mm thick. The
alumina and zirconia specimens were disk-shaped in the as-sintered form, and were not
sectioned prior to coating, notching, and joining.

The microstructures of the as-sintered specimens of alumina, zirconia, and the
hydroxyapatite were characterized by Scanning Electron Microscope (SEM) micrographs
of fracture surfaces (Figure 41). The fracture surfaces of the alumina and zirconia show
relatively equiaxed grains with low porosity (Figure 41). However, the HAP material has
a volume fraction porosity of about 0.10 and exhibits considerable transgranular failure
on the fracture surface.(Figure 41) In addition, the fracture section of the MaCorTM
material clearly shows the randomly oriented mica phase (Figure 41).

All specimens were polished prior to joining using an automated polishing
machine (Leco Corporation, St. Joseph, MI). Prior to beginning the polishing procedure,
the edges of the specimens were slightly beveled to reduce wear on the polishing cloth.
After beveling, three or more specimens were simultaneously mounted onto an aluminum
disk, using a thermal plastic to hold the specimens in place. Polishing was done using a

series of diamond grit ranging from 25 microns to 1 micron.

104

After polishing the specimens were coated with one of the bonding agents
described above. A few drops of the coating was applied to the specimen using a pipette,
and then the specimen was spun on a substrate spinnner to distribute the coating material
over the surface of the material to be joined. The spinning rates varied from about 500 to
5000 rpm. After spinning, the Silicafilm coatings were cured in air, in a conventional
furnace, using curing temperatures of 2000C and curing times of 20 minutes. For the

PCS and the sodium silicate bonding agents, the film was not cured prior to bonding.

Notching the Specimens

Following the polishing, coating, (and in some cases, curing) steps, the specimens
were then notched using a ultrasonic vibratory mill (Sonic Mill. Albuquerque, NM). The
principal of operation of the vibratory mill is that a tool is mechanically vibrated at
ultrasonic frequencies (say between about 1 - 10 kHz), and the vibrating tool in turn
causes vibration in a boron nitride slurry which is placed at the tip of the tool. The
cutting is then accomplished by the boron nitride itself, rather than the tool. The cutting
tool used for many of the notches was a single-edge razor blade that had been mounted
via silver soder on a “stub” supplied by the vendor (Sonic Mill) that allows one to
fabricate cutting tools of a variety of shapes. In addition to the “razor-blade tool”, some
wider notches were placed in polycrystalline alumina specimens using a tool made from a
cold-rolled steel piece which also had been attached by silver soder to a stub. (The razor
blade tool was capable of cutting notches of widths ranging from roughly 200 to 330

microns, while notches cut using the cold-rolled steel tool were on the order of 900

microns across).

105

Following the notching procedure using the sonic mill, a 3 to 5 mm section of
each specimen was cut from the notched end of the specimen using a low-speed diamond
saw (Figure 42). The section of the notched specimen was used as the reference for the
dimensions of the notch prior to the joining procedure. Both the notched specimen and
the reference piece were examined in the SEM to determine the geometrical stability of

the notch duringjoining.

Microwave and Conventional Processing
A 2000 Watt, 2.45 GHz power supply (Sairem, Model MW PS 2000, Wavemat
Inc., Plymouth, MI) was used to heat a 17.78 cm diameter cylindrical single-mode cavity
(CMPR-250, Wavemat Inc., Plymouth, MI, Figure 43) [18, 19]. As the specimen’s
temperature increased, the dielectric constants (both the real and the imaginary parts) of
the heated ceramics changed, and thus the microwave system was tuned continuously to
re-establish the resonant condition in the cavity. Cavity tuning entailed separate but
coordinated movements of the cavity short position and the launch probe position via
stepper motors that were interfaced with a computer [18, 19]. The specimens were joined
using a TM111 cavity mode, with an initial input power of 100 Watts. The power was
incremented by 50 Watts approximately every three minutes until the processing
temperature was reached.
During joining, the specimens were placed in an refractory enclosure (called a
“Casket”). The casket (Figure 44) consisted of a hollow zirconia cylinder (ZYC, Zircar,
Inc.) and two disc-shaped end caps made of an aluminosilicate refractory board (SALI,

Zircar, Inc.) Before the specimens were loaded into the zirconia casket, from zero to 60

106

grams of dead weight were placed on the specimens. (The dead weights consisted of disc
of microwave-sintered Sumitomo alumina, which each weighed about 20 grams and were
about 4.1 cm in diameter and 4 mm thick). The specimens were loaded into the casket
near its cylinder axis, and the casket was in turn centered along the cylinder axis of the
microwave cavity.

The conventionally joined specimens were joined in one of two radiant-heat
furnaces. The MaCorTM specimens were joined in an electrical box-type furnace (CM,
Inc.) which employed MoSi2 heating elements. The specimens were placed on an

alumina setter, and an alumina crucible was inverted and placed over the specimen.

Results

The alumina and the MaCorTM were joined well using the silica spin-on layer, and
notches cut into the specimens were quite stable during the joining process. Micrographs
of the specimens obtained via the Scanning Electron Microscope (SEM) indicated that
the width and depth of the channels machined into the alumina and MaCoRTM The
alumina and the MaCorTM were joined well using the silica spin-on layer, and notches cut
into the specimens were quite stable during the joining process. For notches widths in the
range of about 200 to 330 microns, and for large notches about 900 microns in width,
the “after joining dimensions” were within about 2 to 5 percent of the “before joining”
dimensions.

For the zirconia specimens, the geometrical stability of the notches has not yet
been quantified, since the zirconia specimens (microwave sintered from 3 mol% Tosoh

TZ-3Y Grade zirconia powders), however the authors believe that using 8 mol% zirconia

107

powders and/or a different bond phase might assist in the joining of the zirconia. While
the polycrystalline HAP ceramics did not join using the silica spin-on interlayers, the
HAP did join when the sodium silicate was used.

The zirconia specimens did not join well at 14500C using microwave heating of
the spin-on silica film, thus the stability of the notches can not be assessed that this time.
Although the zirconia specimens were joined , the specimens broke apart upon cutting.

Using the silica spin-on interlayers, the polycrystalline HAP ceramics did not join
using either microwave or conventional heating at 12000C. After the joining attempt
using the silica solution and both heating modes (conventional and microwave) the HAP
specimens showed no indication of bonding whatsoever. However, when the sodium
silicate solution was used as the bonding agent in place of the silica film, the HAP did
bond at 12500C in air in the microwave cavity (Figure 45). The silica spin-on interlayers,
the HAP did join when the sodium silicate was used. For the notch shown in Figure 45,
for example, the notch width and depth before joining was about 166 microns and 322
microns, receptively. After joining, the width and depth was 169 and 319 microns,
respectively. Thus, the joining process induced only a about a two-percent change in
width and a one-percent change in depth, compared to the “before-joining state.
Additional notches in the joined HAP showed similar results, with changes in individual

notch dimensions uniformly within about 4-percent or less.
Summary and Conclusions

The alumina and the MaCorTM were joined well using the silica spin-on layer, and

notches cut into the specimens were quite stable during the joining process. For the

108

zirconia specimens, the geometrical stability of the notches has not yet been quantified,
since the zirconia specimens (microwave sintered from 3 mol% Tosoh TZ-3Y Grade
zirconia powders), however the authors believe that using 8 mol% zirconia powders
and/or a different bond phase might assist in the joining of the zirconia. While the
polycrystalline HAP ceramics did not join using the silica spin-on interlayers, the HAP
did join when the sodium silicate was used. In addition, the stability of the notches
during the HAP was excellent, as was seen for alumina and MaCorTM specimens also.
These results mean that precise channels can be cut into polycrystalline ceramics, and the

dimensions of the channels can be maintained very well during joining.

Acknowledgments

The authors acknowledge the financial support of the Michigan Research
Excellence Fund provided through the Electronic and Surface Properties of Materials
Center, Michigan State University and by the Composite Structures and Materials Center.
The authors also acknowledge the use of the Scanning Electron Microscope facilities at

the Center for Electron Optics, Michigan State University.

109

 

(C) (d)

Figure 41. SEM micrographs of the fracture surfaces for the four materials included in
this study, namely (a) alumina, (b) MaCorTM, (c) zirconia, and (d) hydroxyapatite.

110

Top ,
co ted specrm en

'/ >

 

 

 

 

 

 

Joined specimen
Razor blade with notch

tool attached
to sonic-millTM

 

BOIIOéIl .

coate specrmen Specimen

' with notch
/—> — —} -’

— —-

 

Specimen cut
along this line

Figure 42. Procedure for sectioning the reference specimen for notch geometry in the
“pre-joined” state (after [1]).

111

Directional Power meter Cylindrical Single-mode
coupler l microwave cavrty

m 0
o o o Specimen

 

 

 

 

 

Magnetron Circulator

\.
- + -

 

 

 

 

 

 

 

 

 

 

 

 

 

 

— \ / —
Dummy load fit? 0
. O ofo = Casket
L" E E E: .5- Optical
_ pyrometer

Power supply controller
Motor controller

 

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

———-——--= _ for probe
L—J “ 1 "ti——
1 I l. »
. —- T”: —
lllllllllg I :41; accufiber model 10
[33 33232322333223] , l l” Optical Fiber
00 0052:3000 3 j -_. __‘E Thermomete
LET Motor controller
for short

Figure 43. Schematic of the microwave processing system, showing the microwave
cavity and the power supply (after [20]).

7.6 cm

 

Alumina SALI

Al
'I

2cm
t-——-t

Zirconia ZYC

Dead weights

7cm

Specimen

Alumina SALI
specimen setter

Alumina SALI

2.5 cm

 

 

1.3 cm 5.1 cm 1.3 cm

Figure 44. Schematic of the refractory casket used during joining (after [1]).

113

 

(a) (b)

Figure 45. Microstructures of the joint regions of the joined HAP specimens.

114

References

l.

10.

11.

12.

13.

14.

15.

16.

K. N. Seiber, K. Y. Lee. E. D. Case, pp. 941-949 in Proceedings of the 12th Annual
Advanced Composites Conference, Technomic Publishing Co., Lancaster, PA (1997)

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

D. W. Richerson, Modern Ceramic Engineering, 2nd edition, p. 496-497, Marcel
Dekker, Inc., New York (1992)

M. Kahn and M. Chase, J. Am. Ceram. Soc., 75[3]: 649-656 (1992)

S. Masia, P. D. Calvert, W. E. Rhine, and H. K. Bowen, J. Mater. Sci, 24: 1907-
1912 (1989)

M. J. Edirisinghe and J. R. C. Evans, Int. J. High Technol. Ceram., 2: 249-258 (1986)

H. Yan, W. R. Cannon, and D. J. Shanefield, J. Am. Ceram. Soc., 76[1]: 167-172
(1993)

F. J. Klug, W. D. Pasco, and M. P. Borom, J. Am. Ceram. Soc, 65: 619- 626 (1982)
S. J. Bennison and M. P. Harmer, J. Am. Ceram. Soc., 68, 591-597 (1985)

F. F. Lange, B. 1. Davis, and E. Wright, J. Am Ceram. Soc., 69: 66-69 (1986)

C. H. Hseuh, A. G. Evans, and R. C Coble, Acta. Metall., 30, 1269-1279 (1982)

G. Bandyopadhyay and K. W. French, J. Europ. Ceram. Soc., 11[l]: 23-24 (1993)

R. M. German, K. F. Hens, S.T.P. Lin, Am. Ceram. Soc. Bull., 70[8]: 1294-1302
(1991)

T. F. Yen, Y. H. Chang, D. S. Tsai, S. L. Duh, and S. J. Yang, Mat. Sci. and
Engineering, A154, 215-221 (1992)

G. Elssner, W. Diem, and J. S. Wallace, pp. 629-639 in Surfaces and Interfaces in
Ceramic and Ceramic-Metal Systems, Edited by J. Pask and A. G. Evans, Plenum Press,
New York (1981)

Ravaglioli and A. Krajewshi, Bioceramics, pp. 5 — 7, 413—415, Chapman and Hall.
New York (1992)

115

17. A.Krajewshi, A. Ravaglioli and V. Biasin, “Plasma spray coating of prevalently
titanium supports with various ceramics”, paper presented at the lntemational
Conference of ‘Bioceramics and the Human Body’, Faenza, Italy (1992)

18. K. Y. Lee, E. D. Case, and J. Asmussen, Jr. , Ceramic Transactions. 59, 473-480,
American Ceramic Society, Inc., Columbus, OH. (1995)

19. K. Y. Lee, E. D. Case, J. Asmussen Jr., and M. Siegel, Scripta Materialia,
35[l]:107-111 (1996)

20. K. Y. Lee, E. D. Case, J. Asmussen Jr., and M. Siegel, Binder Burnout in a
Controlled Single-Mode Microwave Cavity, Scripta Materialia, 35[l]:107-111 (1996)

116

3.4 Joining Ceramics to Produce Components with Precise Internal

Channels

JOINING CERAMICS TO PRODUCE COMPONENTS WITH PRECISE INTERNAL

CHANNELS

J. G. Lee* and E. D. Case

Materials Science and Mechanics Department,

Michigan State University, East Lansing, MI 48823

ABSTRACT

Geometrically complex ceramic parts are difficult to achieve, especially when one
deals with an “internal” geometric complexity such as a component with internal
channels, holes, or notches. Such components are potentially important for the flow of
cooling fluids, fuel, or biological fluids through a ceramic component. This paper
discusses microwave joining of ceramics (including alumina/zirconia composites,
zirconia, MaCorTM, and hydroxyapatite) to form components with precise internal

notches. The microstructure near the joined/notched regions also will be characterized.

 

J. G. Lee and E. D. Case, “Joining Ceramics to Produce Components with Precise lntemal Channels”. in
Innovative Processing and Synthesis of Glass, Composite, and Ceramic Materials 111, Volume 108,
Ceramic Transactions, American Ceramic Society. pp. 433 — 442 (2000)

117

INTRODUCTION

The difficulty of fabricating near net-shape components, due to the brittle
characteristic of ceramic materials, has been a limitation for machining ceramic [1-4].
However, joining allows one to make complex shape from simpler subcomponents [1-4].
Features such as of interest channels, holes, or notches, are interesting in high
temperature heat exchanger and engine application as well as biological fluids.

Direct ceramic—ceramic joining utilizing microwave energy has been successfully
done in recent year. The ability of rapid and volumetric heating is attractive advantages
of microwave heating [3]. Desirable temperature can be reached at lower power
compared to conventional heating. Case and Lee have been joined alumina, zirconia,
MaCorTM and MgF2 also including notch stability before and after joining using spin-on
materials [6-10].

In this study, notches were induced in alumina/zirconia composites, zirconia,
MaCorTM, and hydroxyapatite. In order to observe the dimension difference before and
after joining, notched and notched/joined specimens were compared by SEM micrograph.
The joining was attempted in 2.45 GHz single mode microwave cavity with a low
external pressure, provided by dead weight loading in air. The joined regions were

characterized by Vikers indentation near the interface to perceive the toughness.

EXPERIMENTAL PROCEDURE
Materials used in this study
Zriconia

In this study, a 3 mol % yttria-ziconia and an 8 mol ‘70 yttria-zirconia powder (Tosoh

118

Ceramics Division, Bound Brook, New Jersey) were sintered in a 2.45 GHz single mode
microwave. Microwave heating could achieve higher density after sintering when compared
to that of conventionally annealed specimen [5]. Table 15 showed the sintering conditions
for both powers including alumina/zirconia composites.
Alumina/zirconia composites

Alumina/zirconia composies composed of 85 wt % alumina powder (99.9 % pure
alumina, TMDAR, TAIMI chemicals Co. LTD) and 15 wt % zirconia powder (3 mol %
yttria-zirconia powder, Tosoh Ceramics Division, Bound Brook, New Jersey) were
mechanically mixed for 24 hours using alumina grinding media in a plastic ball mill. After
ball milling, the powder was sintered at 1350°C for 1 hour in the microwave cavity.
MaCorTM

MaCorTM (Corning Code 9658) is commercial machinable glass ceramic with a good
tolerance for thermal shock that is used as an electrical or thermal insulator. The
composition of MaCorTM is SiOz-Aleg-MgO-KzO-F. As received MaCorTM billets were
cut into 1 cm X 1 cm using a low speed saw.
Hydroxyapatite

Nomially 99 % pure hydroxyapatite powder (Cerac Inc., Specialty Inorganics,

Milwaukee, WI, Item#C-2071-1, Lot#X 16907) is an “bioactive” ceramic material which
used for human bone and had a nominal purity. Unlike other power materials used in this
study, hydroxyapatite(HAP) was sintered in an electrical resistance conventional furnace
at 1300°C for 11 hours to a density about 2.974 g/cm3.

Silicafilm

119

Silica film (Emulsitone Company, Whippany, New Jersey) was used for the bond phase.
The film was cured at 200°C in a conventional furnace.
Specimen Preparation

The specimens of MaCorTM and sintered ZrOz, Ale3/Zr02 composites, and HAP
specimens were polished followed a series of diamond grits size 35 um, 17 um, 15 um,
10 um, 6 um, and 1 pm using an automatic polisher (LECO Corporation, St. Joseph, MI).
After polishing, the specimens were cleaned in DI water using an ultrasonic cleaner for 20
minutes.

The silica film was dropped onto a substrate spun at 3000 rpm for 20 seconds using a
high speed substrate spinner. Curing at 200°C for 20 minutes in a conventional furnace
produce uniform and smooth films approximately, 0.13 um thick.

After the applied silica films were cured, the specimens were notched on coated side
using a low speed diamond saw (Figure 47). The coated and notched specimens were
sectioned though the notches in order to observed the notch prior to joining.

A pair of half sectioned specimens were attempted joining in a single mode microwave
cavity 17.78 cm in diameter (Model CMPR-250, Wavemat Inc) using microwave power
supply (Sariem, Model MWPS 2000) generates from zero to 2000 watts of microwave
power at frequency of 2.45 GHz (Figure 48). During microwave heating, the specimens
were placed in a cylindrical refractory insulator to reduce heat loss during joining. During

joining, alumina deadweights in range of 20 grams to 50 grams were placed on top of the
specimen being joined to provide low external loads.

After joining, the specimens were sectioned using a low speed diamond saw, which

Produced one specimen for SEM examination and one specimen for Vickers indentations.

120

Sectioned specimen enclosed notches was mount on stub with just notched specimen not
heated for observing the notch dimension difference before and after joining using
scanning electron microscope (SEM). The other piece was mounted using hot pressing
presser and then polished finished I urn. To examine the bond layer or interface
toughness, Vickers indentation was introduced near the bond layer or interface (Figure

49).

Results and Discussion
HAP with MaCorTM

Notches were cut into both HAP and MaCorTM specimens (Figure 50). One set of
MaCorTM and HAP were joined at 1020°C with 20 grams applied dead weights. Joining
was failed at temperature range 1150°C to 1050°C with 50 grams dead weights. Table 16
was showed conditions of attempted joining of MaCorTM and HAP.

For MaCorTM joined at 1020°C for 20 minutes, using a 20 grams deadweights, the
difference between before and after joining was less than 6 %, which shows the
geometrical stability of notches was quite stable after joining (Table 17 and Figure 51).
For HAP, also three were introduced and measure (Figure 52) before and after joining
(Table 17). The geometrical stability of HAP was in the range of l % to 4 %.

Zro2 (3mol% YZO3) with MaCorTM

ZrOz was joined with MaCorTM at 1020°C for 20 minutes with applied 20 grams
dead weights. Four notches were made in ZrOz but the MaCorTM was not notched.
Notch stability was determined from SEM mesurements of the notch dimensions, before

and after joining. (Table 18 and Figure 53). The differences between before and after

121

joining values of depth range from 1 ‘70 to 5 % while the width changes ranged from I %
to 2 %. Thus, the notch dimension was quiet stable afterjoining.
ZrOz (8 mol ‘70 Y303) with Zl'Og (8 mol ‘70 Y203)

Zirconia (8 mol ‘72: Y303) was joined to zirconia (8 mol % YZO3) via microwave
heating at 1500°C for 20 minutes with 50 grams dead weights. Notches were at into both
the 8 mol % yttria-zirconia specimens. The differences in notch width and depth before
and after joining was less than 3 % (Table 19). After heating at 1500°C using microwave
energy, notches made in ZrOz (8 mol % Y203) were not effected in terms of difference
dimension changing.

A1203 85 wt %/ 2102 (3 mol ‘70 Y203) 15 wt % composites joining.

Ale3/ZrOz composites were successfully joined using microwave heating at 1500°C,
1450°C, and 1400°C for 20 minutes. In each case, a silica film spin-on interlayer was used
as the bonding phase and a 50 gram dead weight was placed on the top of the specimens
during joining. Three notches were observed the joined specimen at 1450°C. The notches
were changed the dimension in depth of range 1 % to 2 % and width of range 1 % to 4 %

(Table 20).

CONCLUSIONS

A number of different ceramic materials were notched and then successfully joined.
The cerarnic-creamic material pairs that were joined include MaCorTM with HAP, MaCorTM
with ZrOz (3 mol % YzO3), ZrOz (8 mol % YzO3), and Ale3/Zr02 composites. In each
case, a spin-on silica coating was used as the bond phase. The dimensions of the notches

induced in joined specimens were quite stable (less than 6 % for every attempted joining

122

specimen) during microwave joining process. These results indicate that channels and

notches cut into in polycrystalline ceramics can maintain their dimensions during the joining

process.

REFERNCES

1.

D.Palaith and R.Silberglitt, “ Microwave Joining of Ceramics,” Journal of American
Ceramic Society Bulletin, 68[9] 1601-1606 (1989).

RE. Loehman and AP. Tomsia, “Joining of Cermics,” Journal of American Ceramic
Society Bulletin, 67[2] 375-380 (1988).

ML. Santella, “A Review of Techniques for Joining Advanced Ceramics,” Journal of
American Ceramic Society Bulletin, 71[6] 947-953 (1989).

R.Silberglitt, I. Ahmad, W.M. Black, and JD. Katz, “Recent Developments in
Microwave Joining,” MRS Bulletin, 18[7] 47-50 (1993).

SA. Nightingale, D.P. Dunne, and H.K. Nomer, “Sintering and Grain growth of 3 mol
% Yttria Zirconia in A Microwave Filed,” Journal of Materials Science, 31 5039-5043
(1996).

ED. Case, K.Y, Lee, J,G, Lee, and T. Hoepfner, “Geometrical Stability of Holes and
Channels During Joining of Ceramics and Composites,” Proceedings from Materials
Conference ’98 on Joining of Advanced and Specialty Materials, 12-15 October 1998,
27-34.

K.Y. Lee and ED. Case, Microwave Joining and Repair of Ceramics And Ceramic
Composites,” Ceramic Engineering and Science Proceedings, 21‘" Annual Cocoa Beach
Conference and Exposition on Composites, Advanced Ceramics, Materials and
Structure, Cocoa Beach, Florida, 18 543-550, 1997.

K.N. Seiber, K.Y. Lee, and ED. Case, “ Microwave and Conventional Joining of
Ceramic Composites Using Spin-On Materials,” Proceeding of the 12'hAnnual
Technical Conference Dearbom, MI, 941-949, 1997.

J .G. Lee, K.Y. Lee and ED. Case, “Joining of Diamond Thin Film to Optical And IR

Materials,” Innovative Processing/Synthesis: Ceramics, Glasses, Composites II, 509-
520, 1999.

123

10. ED. Case, J.G. Lee, and KY. Lee, “Joining of Optical and Infrared Materials Using
Spin-On Layers,” Proceedings from Materials Conference '98 on Joining of Advanced
and Specialty Materials, 12-15 October 1998, 17-26.

124

Table 15. Sintering conditions for powder materials used in this study.

 

 

 

 

. . Pre- Heating Sintering Hold Cooling
Mama” Mass Pressing Rate Temp (°C) Time(min) Rate
2102 (3 mol 0 0 . O , 0 _
% Y203) 3 cm 32 Mpa 10 C/rmn 1375 C 60 min 10 C/mrn
2102 (8 mol 0 o . 0 . 0 .
% Y203) 3 cm 32 Mpa 10 C/mrn 1350 C 60 mm 10 C/mrn
“203/2.02 2 gm 32 Mpa 10°C/min 1350°C 60 min 10°C/min
composrtes

 

Table 16. Attempted joining conditions for joining MaCorTM and HAP.

 

 

 

 

 

 

Tmax (0C) Dead weights Hold time Heating rate Result Comments

“50°C 50 grams 20 min ~ 40°C/min Not joined Macor
melted

1 100°C 50 grams 20 min ~ 40°C/min Not joined Macor
melted

1050°C 50 grams 20 min ~ 40°C/min Not joined Macor
melted

1020°C 20 grams 10 min ~ 40°C/min Joined

1000°C No 20 min ~ 40°C/min Not joined

 

125

Table 17. Notch stability of joined HAP with MaCorTM before and after joining.

 

 

 

 

 

 

 

Notch Notch Notch Notch
Material # of Width, Width, Difference depth, depth, Difference
Notched notches before after (%) before after (%)
joining joining joining joining
HAP 1 364 361 um I % 632 um 636 um I %
um
HAP 2 343 355 pm -4 % 643 pm 614 um 4 %
um
HAP 3 314 321 um -2 % 518 um 496 pm 4 %
um
MacorTM 1 343 335 um 2 % 371 ttm 385 pm -4 %
um
MatCorTM 2 321 335 11m -4 % 514 Mm 528 pm -3 %
um
MaCotTM 3 341 321 pm 6 % 407 pm 421 pm -4 %
um

 

Table 18. The notch change before and after joining for joined ZrOz with MaCorTM

 

 

 

 

 

Notch Notch Notch Notch
Material # of Width, Width, Difference depth, depth, Difference
Notched notches before after (%) before after (%)
joining joining joining joining
ZrOz l 374 um 367 pm 2 % 246 pm 240 pm 3 %
ZrOz 2 375 pm 379 pm -1 % 238 pm 225 u 5 %
m
Zr02 3 379 um 375 um I % 267 um 265 um I %
k ZrOz 4 385 um 379 um 2 % 304 pm 296 um 3 %

 

126

Table 19. The notch dimension change before and afterjoining for joined Zr02 (8 mol %

 

 

 

 

 

 

Y203)
Notch Notch Notch Notch
Material # of Width, Width, Difference depth, depth, Difference
Notched notches before after (%) before after (%)
joining Eining Joining joining
ZrOz 1 314 pm 319 pm 2 % 421 um 407 um 3 %
ZrOz 2 319nm 310 um 3% 381 um 386 pm -1 %
ZrOz 3 319 pm 312 um 2% 402 um 410 pm -2%
ZrOz 4 329 um 319 pm 2 % 460 um 452 um 2 %
ZrOz 5 310 um 314 um -I% 433 pm 419 um 3%

 

Table 20. . The notch dimension change before and after joining for joined A1203/ZrOz

 

 

 

 

composites.
Notch Notch Notch Notch
Material # of Width, Width, Difference depth, depth, Difference
Notched notches before after (%) before after (%)
joining Joining joining joining
A1203/Zr02 1 321 um 307 pm 4 % 321 pm 325 rim -1 %
Ale3/Zr02 2 309 um 307 um I % 313 um 315 um -I %
Ale3/Zr02 3 321 um 318 um I % 352 pm 346 um 2 %

 

l
g
A ./’ Powder

45:25:)
1

l

l

'. 1

l

l

,r

b

 

xy

1

Q Compact
‘ powder

h

 

Seven powder
compact specimens

Figure 46. A schematic for compact powder materials and sintering.

128

Coated Notched

 

 

 

 

5116011113“ specimen
E3 - 9
Cut through Notched and
Notch Joined specimen
m p
1‘\2\L1_Ll/-/i
“\\\~_~/

 

 

Figure 47. Schematic of producing notches.

129

Monitor
(Control Short
& Probe)

Microwave

Reflected power
Detector

Optical
Pyrometer

 

Figure 48. Schematic of microwave processing system, showing microwave cavity and
microwave power supply.

130

 

 

Vickers
lndentor

 

.__ Mounted
specimen

 

 

Figure 49. A schematic of showing Vickers indentations placed near the bond layer.

131

MaCorn"

 

I

Figure 50. Notch configuration in both HAP and MaCorTM specimens.

132

200 um

 

(a) MaCorTM before joining (b) MaCorTM after joining

Figure 51. MaCorTM notched specimen before and after joining.

133

200 um

 

(a) HAP before joining (b) HAP after joining

Figure 52. HAP notched specimen before and after joining.

134

 

(a) ZrOz before joining (b) ZrOz after joining

Figure 53. ZrOz notched specimen before and after joining.

135

200 u m

 

(a) ZrOz before joining (b) ZrOz after joining

Figure 54. ZrOz notched specimen before and after joining.

136

20011111

 

(a) A1203/ ZrOg before joining (b) A1203/2102 after joining

Figure 55. A1203/ ZrOz notched specimen before and after joining.

137

3.5 Joining of Non-Oxide Ceramics Using Conventional And

Microwave Heating

JOINHVG OF NON-OXIDE CERAMICS USING CONVENTIONAL AND
MICROWAVE HEATING

Jong-Gi Lee and E. D. Case
Materials Science and Mechanics Department
Michigan State University
East Lansing, MI 48824
ABSTRACT
The nonoxide ceramic materials silicon carbide (a structural ceramic) and magnesium
fluoride (an optical/IR ceramic) have been joined using conventional and microwave

heating. Joining of both the SiC and MgF2 utilized a spin-on interfacial layer, which

allowed joining with low external applied pressures.

INTRODUCTION

Non-oxide ceramics include many useful materials, including structural ceramics such
as silicon carbide and silicon nitride, with their combination of relatively high toughness
and low mass density. However, a second important category of non-oxide ceramics is
infrared transmitting materials. Particular nonoxide ceramics are much better infrared
transmitting materials than oxide ceramics due to basic physical nature of the metal-oxygen

bond. The cutoff in infrared transmission as a function of wavelength is due to coupling of

 

J. G. Lee and E. D. Case, “Joining of Non-Oxide Ceramics Using Conventional and Microwave Heating”.
Ceramic Engineering and Science Proceedings 21(4]. American Ceramic Society. pp. 589-597 (2000)

138

the infrared photons with the lattice normal modes of vibration for the material [I]. High
atomic mass and weak bonds tend to increase the effective cutoff wavelength such that in
oxide ceramics, the metal-oxygen bond strongly absorbs infrared wavelengths longer than
about 5 to 8 microns [2,3], but for nonoxides such as magnesium fluoride and zinc sulfide,
the relatively low mass and weak bonds enhance the infrared transmission by pushing the
IR cutoff to higher wavelengths. This paper addresses the joining of both optical/infrared
nonoxide ceramics as well as structural ceramics.

As is the case for oxide ceramics, non-oxide ceramics also are difficult to process in
complex shapes and one method of fabricating ceramic components with complex
morphologies is to “build them up” by joining sub-components of relatively simple
geometry. Thus, there have been a number of efforts to join nonoxides via either
conventional or microwave heating [4 - l4].

Nonoxide ceramics have been joined by conventional heating using a variety of
metallic interlayers or braze filler materials. Young et al. fabricated AlN—Ni-Cu-Ni-AIN
“sandwiches” by vacuum hot pressing at 600°C to 700°C at a pressure of 6.5 MPa for 30
minutes [4]. The AIN plates were coated with nickel, then a copper foil was inserted
between the Ni-coated AIN plates prior to joining, resulting in a low-temperature bond
about 100 microns thick [4]. Wang et al. used a CuNiTiB paste brazing filler to join
Si3N4/Si3N4. Before joining, an interfacial layer consisting of a slurry of the metal
powders was applied to the specimens [5]. Joining was done at temperatures between
about 1000°C and 1120°C in a conventional vacuum furnace, with final bond layer

thicknesses ranging between about 48 microns and 58 microns thick [5].

139

A number of researchers used ceramic interlayers to join nonxoxide ceramics via
either conventional or microwave heating. Xie et al. joined sintered Si3N4 using a slurry
of a—S13N4, Y203, S102, and A1203 powders in a conventional furnace with a nitriding
atmosphere [6]. The Optimum joining conditions were 1650°C with a 5 MPa applied load
and a 30 minute hold time, giving bond layers approximately 15 microns thick [6]. Lee et
al. joined sintered silicon carbide using a 1 mm thick MgO-A1203-SiOz (MAS) layer
inserted between the two SiC plates [7]. Using zero applied external load, the SiC
specimens were heated in a conventional graphite-element furnace from room
temperature to 1000°C in vacuum, then from 1000°C to either 1500°C or 1600°C in an Ar
atmosphere, yielding final bond thicknesses of 5 microns and 10 microns, respectively
[7].

Silberglitt et al. used Si interlayers in the microwave joining of SiC at 1450°C for 5 -
10 minutes under an external load of 2 - 5 MPa [8]. The final bond widths were a
function of the interlayer processing technique, with Si powders, a Si slurry, and a
plasma-sprayed Si coating yielding final (as—joined) bond layer thicknesses of 50 microns,
5 microns, and approximately 3 microns, respectively.

In addition to the joining of nonoxides using interlayers, Binner used a 2.45 GHz
microwave cavity to directly join RBSN to RBSN [9]. However, although Binner et al.
did not deposit an interlayer on the RBSN materials, the joining process itself generated a
free-silicon interlayer between the joined RBSN specimens [9], where the free silicon was
thought to “bleed” out the RBSN microstructure into the bond region due to the
application of high temperature and pressure [9]. For the RBSN, a gap between the

joined specimens diminished with increasing temperature and pressure, until at a joining

140

temperature of 1350°C, a joining time of about 15 minutes, and a joining pressure of 1.5
MPa, the silicon-filled bond layer had decreased to about 3 - 5 microns (whereas at
1190°C and 0.5 MPa, the “gap” was about 10 - 20 microns wide and only partially filled
with Si). [9]

Previously, the authors and coworkers successfully joined several ceramic oxide
materials [10-14], including joining alumina, partially stabilized zirconia (PSZ),
MaCorTM, and hydroxyapatite. Also, a number of dissimilar ceramic oxide pairs have
been joined [12, 14]. This paper focuses on the joining of two nonoxides, SiC (a

structural material) and MgF2 (an IR transmitting material).

EXPERIMENTAL PROCEDURE

The as-received billets of silicon carbide and magnesium fluoride were cut using a
low speed diamond saw, with specimen dimensions of 1.6 cm X 0.4 cm X 0.3 cm for the
SiC (Hexoloy) specimens and 1.0 cm X 1.0 cm X 0.2 cm for the magnesium fluoride
specimens. After cutting, the specimens were polished using a series of diamond grit
sizes, with a minimum 1 micron grit size.

The spin-on layers used were applied to the polished specimen surfaces in a liquid
form. Four to five drops of the spin-on liquid (BlackglasTM for the SiC and sodium
silicate for the MgF2) were applied near the center of the polished specimen using a
pipette. The specimen was spun for 20 seconds at speeds ranging from 3000 RPM to
5000 RPM using a substrate spinner. Following spinning, the coated surface of the
specimen was placed in contact with another polished, uncoated specimen of the same

material and was immediately loaded into the furnace (Figure 56).

141

The joining was done using both a conventional furnace and a microwave cavity.
Conventional heating was done with a nitrogen flow rate of approximately 120 sccm and
a heating/cooling rate of approximately 10°C per minute. The microwave heating was
done in air using a 2.45 GHz single-mode microwave resonant cavity. The specimen
temperature was monitored via an optical pyrometer system. A cylindrical zirconia
refractory specimen enclosure (casket) was used during joining [10-14].

The SiC specimens were heated first in the conventional furnace at 900°C for 20
minutes, followed by microwave heating of selected specimens at 1200°C for 20 minutes.
The MgF2 specimens were heated for 20 minutes either in the conventional furnace (with
flowing nitrogen) at 700°C or in air in the microwave at 700°C. After joining, the
specimens were sectioned on a low speed diamond saw and examined in a Scanning
Electron Microscope (JEOL Model JSM-6400V). Vickers indentation testing was also

performed on the joined and sectioned specimens.

RESULTS AND DISCUSSION

The microstructure of the SiC (Hexoloy) materials is shown in Figure 57 (a
micrograph of the fracture surface), while for MgF2 a comparable fracture surface
micrograph is given in reference 11.

Joining the SiC at 900°C in flowing N2 in a conventional furnace results in a dense
bond layer that was roughly 2 to 3 microns thick (Figure 58a). For the MgF2 specimens
that were joined at 700°C both in flowing nitrogen in a conventional furnace and in a
microwave cavity in air at 700°C, the resulting bond thickness was about 10 microns

(Figure 58b shows the MgF2 bond due to microwave heating). For the SiC specimens

142

that were heated conventionally at 900°C, reheating the specimens at 1200°C in the
microwave cavity did not appreciably change either the appearance or mean thickness of
the bond layer (Figure 59). Previously, the authors found that the IR transmittance of the
joined MgF2 was essentially the same as the unjoined MgF2, if microwave heating was
used. However, the IR transmittance was degraded if the specimens were joined by
conventional heating, in part due to IR scattering from the second phases formed in MgF2
bond layer in conventionally heated specimens [11, 14].

In order to investigate the relative interfacial toughness of the bonds in the joined
specimens, a series of 49 N Vickers indentation cracks were placed near the interface of
the conventionally and the microwave joined MgF2 and SiC specimens, such that the
radial cracks were oriented approximately normal to the joined interface. In general,
cracks approaching an interface can deflect at the interface (forming “T-shaped” cracks
along the interface) if the interfacial fracture energy is less than about 60 percent of the
matrix fracture energy [15]. Otherwise, the crack propagates across the interface, without
deflection. For both the SiC and the MgF2 specimens included in this study, the
indentation cracks did deflect at the interface, indicating that the relative interfacial

fracture energy was less than 60 percent of the matrix fracture energy in both cases.

SUMMARY
Two nonoxide materials, SiC and MgF2 were joined using a spin-on interlayer. The
bond phases produced were relatively thin and dense, but the interfacial fracture energy

was low.

143

 

ACKNOWLEDGMENTS
The authors acknowledge financial support of the Composite Materials and Structures
Center of Michigan State University and the use of the electron microscope facilities at

the Center for Electron Optics, Michigan State University.

144

Polished
specimen

9
9" 9" @

Polished Coated with

Coated
specimen spin-on material

specimen

Microwave or

conventional
heating
= /
Joined
specimen

Figure 56. Schematic of the coating and joining technique.

I45

 

Figure 57. Fracture surface of SiC used in this study.

146

 

 

Figure 58. (a) SiC joined at 900°C for 20 minutes in flowing N2 with a 20 gram
deadweight applied during joining and (b) MgF2 specimens joined at 700°C for 20
minutes in air with a 60 gram deadweight applied during joining.

147

‘hwklw‘h-u.~“'

2::

 

Figure 59. SEM images of the SiC specimens that had been first heated conventionally at
900°C (Figure 58a), followed by (a) microwave heating at 1200°C for 20 minutes or by
(b) conventional heating at 1200°C for 20 minutes.

148

REFERENCES

l.

10.

11.

12.

M. W. Barsoum, pp. 618 - 635 in Fundamentals of Ceramics, McGraw-Hill, New
York, NY, 1997.

C. Kittel, pp. 107-116 in Introduction to Solid State Physics, 5th ed., John Wiley and
Sons, New York, 1976.

L. L. Hench and J. K. West, Principles of Electronic Ceramics, pp 351-360, John
Wiley & Sons, New York, NY, 1990.

CD. Young and J.G. Duh, “Bonding Mechanism of Electrodes Ni-P Film with AIN
substrate and Cu Foil,” IEEE Transactions on Components Packaging and
Manufacturing Technology Part A, 21 [2] 330-344 (1998).

CG. Wang, H.P. Xiong and Z.F. Zhou, “Joining of Si3N4/Si3N4 with CuNiTib Paste
Brazing Filler Metals and Interfacial Reactions of the Joints,” Journal of Materials
Science, 34 [12] 3013-3019 (1999).

R.J. Xie, L.P. Huang, Y. Chen and X.R. Fu, “Evaluation of Si3N4 Joints: Bond
Strength and Microstructure,” J. Mat. Sci., 34 [8] 1783-1790 (1999).

H.L. Lee, S.W. Nam, B.S. Hahn, B.H. Park and D. Han, “Joining of Silicon Carbide
Using MgO-AIZO3-SiOz Filler,” J. Mat.Sci., 33 [20] 5007-5014 (1998).

R. Silberglitt, D. Palaith, H.S. Sa’alaldin, W.M. Black, J.D. Katz, and RD. Blake,
“Microwave Theory and Application in Materials Processing,” pp. 487 - 495 in
Ceramic Transactions Vol 21, D. E. Clark, F. D. Gac, and W. H. Sutton, editors,
American Ceramic Soc., Westerville, OH (1991).

J.G.P. Binner, J.A. Femie, and PA. Whitaker, “An Investigation into Microwave
Bonding Mechanisms via a Study of Silicon Carbide and Zirconia,” Journal of
Materials Science, 33 [12] 3009-3015 (1998).

KY. Lee, E.D. Case, D. Reinhard, “Microwave Joining and Repair of Ceramics and
Ceramic Composites,” Ceramic EngSci. Proc., 18 543-550 (1997).

ED. Case, J.G. Lee, and KY. Lee, “Joining Optical and Infrared Materials Using
Spin-On Layers”,pp.l7-26 in Joining of Advanced and Specialty Materials, M. Singh,
J .E. Indacochea, and D.Hauser,eds.,ASM lntemational, Materials Park, OH, 1998.

K.N. Seiber, K.Y. Lee, and ED. Case, "Microwave and Conventional Joining of

Ceramics using Spin-on Materials,“ pp. 941-949 in Proc. 12th Annual Advanced
Composites Conference, Technomic Pub. Co., Lancaster, PA, 1997.

149

13. ED. Case, K.Y. Lee, J .G. Lee, and T. Hoepfner, “Geometrical Stability of Holes and
Channels During Joining of Ceramics and Ceramic Composites,” pp. 27 - 34 in
Joining of Advanced and Specialty Materials, M. Singh, J.E. Indacochea, and D.
Hauser, eds., ASM lntemational, Materials Park, OH, 1998.

14. ED. Case, KY. Lee, and J.G. Lee, “Joining of Polycrystalline Ceramics and Ceramic
Composites Using Microwave Heating,” pp. 17 - 20 in Proc 33rd Int. Microwave

Power Symp., lntemational Power Institute, Manassas, VA., 1998.

15. W. Lee, S. J. Howard, and W. J. Clegg, Acta Mater. 44 3905-3922 (1996).

150

3.6 Protective Coatings for Infrared Materials

PROTECTIVE COATINGS FOR INFRARED MATERIALS

Jong-Gi Lee, E. D. Case and M. A. Crimp
Materials Science and Mechanics Department
J. Malik and D. K. Reinhard,

Electrical Engineering Department

Michigan State University
East Lansing, MI 48824

ABSTRACT

Solutions of polycarbosilane and hexane have been spun onto magnesium fluoride
substrates, producing continuous and adherent coatings upon curing in flowing nitrogen.
The study focused on the infrared properties of the films, which were maintained upon

coating.

INTRODUCTION
The interaction of electromagnetic waves with ceramic materials involves reflectance,
absorptance, transmittance and scattering, such that [1]

S+R+A+T=l (D

where S, R, A, and T are coefficients associated with scattering, reflectance, absorptance

and transmittance, respectively. Each quantity in equation 1 is a function of a number of

 

E. D. Case, J. G. Lee, M. A. Crimp, D. K. Reinhard, and J. Malik, “Protective Coatings for Infrared Materials”,
Ceramic Eng. and Sci. Proc., Volume 20, American Ceramic Society, Inc., Westerville, OH, pp. 145-152
(1999)

151

variables, including the incident wavelength. In addition to the wavelength dependence,
the scattering, S, is a function of the number density and dimensions of scattering
centers, such as pores, rough surfaces, etc. Reflectance is a function of the index of
refraction of the material and the angle of incidence. For the absorptance, A, in the
infrared-visible-ultraviolet regime, there are two major absorption mechanisms that
produce wavelength-dependent “cutoffs”. In the ultraviolet, the cutoff is due to electronic
processes [1], while in the infrared, the cutoff is due to coupling of the infrared photons
with the lattice normal modes of vibration [1,2]. The frequency of the lattice modes in
turn depends on the bond strength and the atomic mass of the atoms that make up the
lattice [3]. High atomic mass and weak bonds tend to increase the effective cutoff
wavelength. For example, in oxide materials, the metal-oxygen bond tends to strongly
absorb infrared wavelengths longer than about 5 to 8 microns [1,2], but if one selects
ceramics with weaker bonds, such as magnesium fluoride, zinc sulfide, and zinc selenide,
then one enhances the infrared transmission by pushing the IR cutoff to higher
wavelengths. However, the weaker bonds tend to give fluorides, sulfides, and selenides
relatively poor mechanical properties, such as low strength and hardness, which in turn
makes infrared transmitting materials susceptible to abrasion and erosion damage. To
reduce abrasion, erosion, and handling damage, one can add a protective coating to the
surface of these materials. In this paper we discuss a method for fabricating a thin,
adherent silicon-based coating on polycrystalline magnesium fluoride.
EXPERIMENTAL PROCEDURE

A commercial hot-pressed polycrystalline magnesium fluoride (Irtran 1, Eastman

Kodak) was used as the infrared transmitting substrate for this study. The magnesium

152

fluoride had an equiaxed microstructure with a mean grain size of approximately 3 microns.
The polycarbosilane (PCS) solution used as a spin-on coating in this study was either
applied in a undiluted form to the MgF2 or it was diluted with hexane, using 1 gram of
polycarbosilane to n grams of hexane, where the values of 11 used in this study were 4
grams, 6 grams, and 10 grams.

Using a low-speed diamond saw, magnesium fluoride specimens roughly 1.0 cm X 1.0
X 0.2 cm were sectioned from as-received billets. Both of the 1.0 cm X 1.0 cm specimen
faces were polished for each specimen, using an automated polishing machine (Leco
Corporation, St. Joseph, MI) and a series of diamond grit sizes ranging from 17 microns to
one micron. Using a pipette, several drops of the PCS solution were placed near the center
of a 1 cm X 1 cm specimen face. The PCS coated specimens were then spun at rates
between 500 and 6000 rpm for 20 seconds using a commercial substrate spinner (Figure
60).

Curing of the coated specimens was done in both air and flowing nitrogen. For the
curing in air, the coatings were heated at about 10°C per minute to maximum
temperatures of up to 265°C with hold times of 20 minutes in a laboratory oven. The
nitrogen-atmosphere annealing was done with a nitrogen flow rate of approximately 120
sccm to 130 sccm, with a maximum temperature ranging from 400°C to 600°C and a
hold time of 20 minutes. For the nitrogen anneals, both the heating and cooling rate was
approximately 20°C per minute.

The infrared transmittance for both the coated and uncoated specimens was
determined using a Beckman Spectraphotometer IR 4420, over the wavelength range

from 2.5 microns to 9 microns (a wavenumber range from 4000 cm'l to 1100 cm"). A

153

scan speed of 150 cm'1 per minute was used for all of the transmittance measurements
reported in this paper. The coated specimens were indented using a Vickers indenter with a

49 N load, a load time of 10 seconds, and a loading speed of 70 microns/second.

RESULTS AND DISCUSSION

The varying dilution of PCS (Starfire Systems, Watervliet, NY) with hexane produced
coatings with differing thickness. Undiluted PCS spun at a 3000 rpm for 20 seconds
produced a continuous coating about 11 microns thick upon curing in air at 100°C for 20
rrrinutes (Figure 61). Dilution of I gram PCS to 4 grams hexane (Figure 62a) and 1 gram
PCS to 6 grams hexane (Figure 62b), both produced coatings about 2 to 3 microns thick
when cured in flowing N2 at 400°C and 500°C, respectively. Specimens sectioned with a
low speed diamond saw were mounted in thermosetting plastic, sputter coated with gold,
and observed in an environmental scanning electron microscope (Figures 62a and 62b).

A dilution of 1 gram PCS to 10 grams hexane produced films roughly 1 micron thick
upon curing in flowing N2. The coatings annealed in air showed numerous strong and
very broad IR absorptions, so the air-annealing of the specimens was not pursued past an
initial measurement of the IR transmittance for several specimens.

The polished, uncoated MgF2 (Figure 63) has transmittance comparable to that of the
PCS coated specimens cured in flowing N2 (Figures 64 and 65). MgF2’s sharp absorption
at a 2.763 micron wavelength is consistent with the literature for Irtran 1 [4]. Nitrogen
atmosphere annealing at 400°C gave a broad absorption at about 4.737 microns or 2112
cm'I which corresponds well to a Si—H group absorption reported at 2120 cm'1 for PCS

[5]. For N2 anneals at 5000C and 600°C (Figures 64 and 65) this absorption was absent,

154

which is consistent with temperature-induced changes in PCS chemistry, since anneals at
400°C in Ar yield considerable weight loss due to the evolution of C6H5 (phenyl) groups
from PCS [6], but by 500°C the weight is stable to above 1200°C [5]. Other volatile
species including H2 and CH4 may be evolved during annealing.

In general, a coating on a substrate can either increase or decrease the transmittance
relative to the uncoated substrate, depending on the optical indices of refraction of the
two media and the coating thickness [6]. An increase in transmission is favored if the
index of refraction of the coating is lower than that of the substrate and the coating
thickness is on the order of 1/4 the wavelength of the incident radiation [6]. Thus, slight
increases or decreases in the transmittance with respect to the uncoated specimens may be
due to reflectance effects [6], and decreases in transmittance may be due to scattering or
absorption effects.

For the thicker (11 micron) coatings, 49N indentations produced coating
delaminations about 300 — 400 microns across, in addition to the associated radial cracks,
while the l to 3 micron-thick films showed a similar extent of delamination. Spalling of
both the thick and the thinner films occurred relatively infrequently upon Vickers
indentation at 49 N. However, Spalling radial coating cracks, and delamination for
ceramic coatings is a function of coating thickness, indentor load, as well as the elastic
and fracture properties of the coating and substrate [7]. Additional indentation work is

underway for the coatings included in this study.

155

SUMMARY

Thin, adherent coatings were successfully applied to polished MgF2 substrates using
spin-on PCS and hexane solutions. Curing in flowing nitrogen at 500°C and 600°C
produced continuous films about 1 to 3 microns thick, with IR transmittances that very
similar to polished and uncoated MgF2 specimens. Thus, such coatings do not disminish the
original IR transmittance of the MgF2 substrates. The coating thickness decreased as the
hexane dilution increased, and as the coating thickness decreased from about 11 microns to
about 1 micron, the tendency for coating delamination remained essentially unchanged.
The indentation work done to date in this study is only a preliminary indication of the
response of the coatings to point contact loading. Future work will include addition

characterization of the mechanical integrity of the coatings.

ACKNOWLEDGMENTS
The authors acknowledge financial support of Composite Materials and Structures
Center, Michigan State University, East Lansing, MI. The authors also thank W.

Sherwood of Starfire Systems, Watervliet, NY for the PCS.

156

Figure 60. Schematic showing the high-speed substrate spinner used in applying the
coatings to the MgF2 substrates.

0 Spin-on material

0
a Polished specimen

Ki 6 D Coated specimen
\ 43 3 after spinning

fig

 

 

High speed substrate spinner
(500 rpm ~ 6000 rpm)

 

 

\ /

157

Figure 61. ESEM examined polycarbosilane film which is not added hexane after coating
spun at 3000 rpm for 20 seconds and cured at 100°C for 20 minutes in air.

3
E
g
.5

 

158

Figure 62.

The polycabosilane film thickness using ESEM after coating and

transmittance of coated specimens. (3) the film solution of coated specimen was 1 gram
PCS and 4 gram hexane mixed solution. The coated MgF2 was spun at 3000 rpm and
cured at 400°C in N2. (b) the film solution of coated specimen was 1 gram PCS and 6
gram hexane mixed solution. The coated MgF2 was spun at 5000 rpm and cured at 600°C
in N2, (c) Transmittance for (a), and (d) Transmittance for (b)

'Fhermoset

Transmittance

 

 

 

 

0.8 ~ _
0.6 r _
0.4 - 3
M PCS:HX=I :4, Cured at 400'C
Spun at 3000 rpm
0.2 2 ~
0.0 1 4 4L 4.; l 1 l

 

Wavelength (micron meter)

(C)

159

 

Transmittance

5 11 In

 

 

 

 

 

 

1.0 Tfi—fi l m —r
0.8 2
0.6 '
0.4 i
PCS:HX=1:6, Cured at 600'C
Spun at 5000 rpm
0.2 P i
0.0 L I l 1 L l l L '
2 3 4 5 6 7 8 9

Wavelength (micron meter)

((0

Figure 63. Transmittance of both sides polished MgF2 specimen.

1.0

0.8

0.6

0.4

Transmittance

0.2

0.0

 

T

l

 

A;

1 J.

l

1

 

Wavelength (micron meter)

3

4

5

160

6

7

8

9

 

Figure 64. Transmittance of MgF2 coated with polycarbosilane at 500°C with different
spin rate.

 

 

 

 

 

l l
1.0 ~ 2
8
a 0.8 P —4
N t
33 ..
"-1 0.6 e 1, a
E T 500 rpm \
U) '2
G 0.4 - 20011 rpm in, -
CB . ,.
h . \‘3 1
E 401111 rpm ‘\.‘,;'
0.2 ” [JR 4
' "' 6000 rpm
0.0 ‘ P i ' 1 g 1 . t . 1 . .
2 3 4 5 6 7 8 9

Wavelength (micron meter)

161

Figure 65. Transmittance of MgF2 coated with polycarbosilane at 600°C with different
spin rate.

 

 

 

 

 

1.0 - f ,
o ” Wilt it“ “1er it" -.
g 0.8 - l [.12. —
CB 1: 1,,
s-r t\
1‘: 0.6 -
E l ‘M—‘“ 500 rpm 11‘1“
g 0.4 * 2000 rpm 1’14. .
In » . 11 1;“, 2, l

41" it rpm 11
H 0.2 - ‘3
j ’ ' ‘ 6000 rpm
0.0 1 ' 1 = 1 .1 1 1 1 1 L

2 3 4 5 6 7 8 9

Wavelength (micron meter)

REFERENCES

l.

[\J

L. L. Hench and J. K. West, Principles of Electronic Ceramics, pp 351-360, John
Wiley & Sons, New York, NY (1990).

M. W. Barsoum, pp. 618 - 635 in Fundamentals of Ceramics, McGraw-Hill, New York,
NY (1997).

C. Kittel, pp. 107-116 in Introduction to Solid State Physics, 5th ed., John Wiley and
Sons, New York (1976).

D. W. Gray, page 6-62 in American Institute of Physics Handbook, Second Edition,
McGraw-Hill Book Co., Inc., New York, NY (1963).

M. Eber and L. E. Jones, pp. 485 - 492 in Ceramic Eng. and Sci. Proc., Volume
19[3], American Ceramic Society, Westerville, OH (1998).

M. V. Klein, pp. 201 - 209, Optics, Wiley and Sons, New York, NY (1970).

D. F. Diao, K. Kato, and K. Hokkirigawa, “Fracture Mechanics of Ceramics Coatings
in Indentation”, J. of Tribology, 116: 860 - 869 (1994).

163

3.7 Joining of Polycrystalline Ceramics And Ceramic Composites Using Microwave

Heating

JOINING OF POLYCRYSTALLINE CERAMICS AND CERAMIC COMPOSITES

USING MICROWAVE HEATING

ELDON D. CASE, KI-YONG LEE, AND JONG—GI LEE
Materials Science and Mechanics Department
Michigan State University, East Lansing, MI 48824

ABSTRACT

Microwave energy can be used to join a variety of ceramics, including both
structural ceramics and optical (infrared) transmitting ceramics. A thin, spin-on interlayer
can be successful in joining such materials. For optical materials such as MgF2, microwave

joining can produce joins that do not appreciably degrade the optical quality of the material.

INTRODUCTION

Microwave processing offers the potential for the rapid fabrication of
ceramic/ceramic composite components of complex geometry. Spin-on layers are widely
used for low-cost fabrication of electronic devices including integrated circuits and solar

cells. Such layers provide protective layers, dielectric layers, and dopant sources. The

 

E. D. Case. K. Y. Lee, and J. G. Lee, “Joining of Polycrystalline Ceramics and Ceramic Composites Using
Microwave Heating,” Proceedings of the 33rd International Microwave Power Symposium, lntemational
Power Institute. Manassas. VA, pp. 17-20 (1998)

164

spin-on material in liquid form is dispensed onto the surface to be coated. The substrate is
then spun at high speeds, leaving a thin layer of the spin-on material.

Techniques for ceramic joining can generally be divided among methods employing
a high extemally-applied stress (such as diffusion bonding) and methods involving low
external stress (such as brazing). Diffusion bonding joins materials at high temperatures
and pressures, where solid state diffusion across the joined interface may be assisted by
fluxes that enhance the mass diffusion rates (Sandhage, Schmutzler, Wheeler, and. Fraser,
1996; Santella, 1992). Diffusion bonding (and creep) can be a sensitive function of grain
size, with the tendency for both creep and successful bonding increasing as the grain size
decreases. For example, Elssner et al. showed that alumina plates with a mean grain size of
18 microns did not join, while similar alumina plates with a one micron average grain size
did join when processed under nominally identical conditions (Elssner, Diem, and Wallace,

1981). In addition to structural materials such as alumina, optical ceramics, such as
magnesium fluoride, have been joined using diffusion bonding (Yen ,. Chang Yu, Yen,
Tsai, and Lin, 1991; Yen,. Chang, Tsai., Duh., and Yang, 1992).

Brazing is another method that has been used to join ceramics. In general, "brazing"
may be defined as "a joining process in which a filler metal and a flux are sandwiched
between the workpieces" (Schaffer, Saxena, Antolovich,. Sanders, and. Waner 1995). At
elevated temperatures, the filler /flux system wets the surfaces to be joined. For ceramics,
brazing has utilized oxide glasses, oxynitride glasses, metal particles slurries, and metal
layers to join ceramic components (Sandhage, Schmutzler, Wheeler, and. Fraser, 1996). In

this study, joining is done via thin, spin-on interlayers of silica or sodium silicate.

165

MATERIALS AND SPECIMEN PREPARATION

A number of polycrystalline ceramic materials have been joined by the authors and
co-workers, including structural ceramics such as alumina and zirconia, electronic ceramics,
such as alumina and MaCoRTM and optical ceramics, such as magnesium fluoride (Lee,
Case, Reinhard , 1997; Seiber, Lee and Case, 1997). The magnesium fluoride
specimens were a polycrystalline commercial, hot-pressed material (Irtran-l, Eastman
Kodak Company) with a mean grain size of approximately 3 microns. MaCoRTM is a
machineable glass ceramic (Corning Code 9658) that is reinforced with randomly-oriented
platelets of fluromica. Rectangular specimens of the both MaCoRTM and the magnesium
fluoride were cut from as-received billets using a low-speed diamond saw.

For the polycrystalline alumina, disk-shaped specimens were pressed from as-
received Sumitomo AKPSO powders having an average particle size of 0.23 microns. The
smaller of the alumina specimens approximately 22 mm in diameter and two mm thick,
when pressed uniaxially at pressures of approximately 32 MPa. Larger disks, having
masses of up to 20 grams and as-pressed diameters of approximately 50 mm also were
pressed and sintered. The polycrystalline zirconia specimens pressed in a similar manner to
the alumina specimens, using Tsoh zirconia, 3-mol-percent yttria powders with an initial
particle size of roughly 0.1 micron.

The bonding agents (that were used as spin-on materials in this study) include a
sodium silicate solution (Columbus Chemical Company) and Silicafilm (Emulsitone
Company, Whippany, NJ). “Silicafilm” is an organic liquid, which when pyrolyzed

produces a film of amorphous silica.

166

Using an automated polishing system, all specimens were polished using a series of
diamond abrasive grit sizes ranging from about 25 microns to 1 micron. For the magnesium
fluoride specimens for which optical and infrared transmittance was measured, the
specimens were polished on both faces, while the remaining specimens were polished on
only one face. The silica film and the sodium silicate solution were applied to the polished
specimen using a pipette, then the specimens were spun at speeds between about 500 rpm to
5000 rpm for approximately 20 seconds. For the silica film, the coatings were typically
cured in air at 200°C for 20 minutes. After coating and curing, selected alumina, zirconia
and MaCorTM specimens, submillimeter-width notches were cut into the specimen surface

using a stationary sonic mill (Sonic-Mill, Albuquerque, NM).

SINTERING AND JOINING

The microwave sintering of the alumina and zirconia powders, along with the
subsequent joining of all of the specimens, was performed in a cylindrical single-mode
microwave cavity, 17.78 cm in diameter (Model CMPR-250, Wavemat Inc., Plymouth,
MI). The microwave power was generated by a 0 - 2000 Watt, 2.45 GHz continuous-wave
power supply (Sairem Model MWPS 2000, Wavemat Inc., Plymouth, MI). The microwave
system is equipped with automated sliding short and launch probe position controls (Figure
66). Details of the microwave cavity and power supply are given detail elsewhere (Lee,
Case, Asmussen, and Siegel 1996).

Prior to heating in the microwave, the specimens to be sintered or to be joined were
placed in a cylindrical refractory enclosure, which was in turn centered along the axis of the

microwave cavity. The refractory enclosure (typically called a "casket") consisted of a

167

hollow cylinder made from a partially-stabilized zirconia (ZYC, Zircar, Inc) with disc-
shaped end plates made of an aluminosilicate refractory board (SALI, Zircar, Inc). The
zirconia cylinder was 7 cm high with an outer diameter of 10.2 cm and the two
aluminosilicate end plates were each 10.2 cm in diameter and 2 cm thick (Figure 67).
Processing temperatures ranged from about 1050°C for the MarcorTM to about 1625°C for
the alumina. The specimen temperatures were measured using an optical pyrometer
(Acufiber Optical Fiber Themometer, Model 10, Luxtron Co., Beaverton, OR), which
was aligned with a viewport in the microwave cavity and a 5 mm diameter circular hole

in the casket wall (Figure 67).

CHARACTERIZATION OF THE J OINED CERAMICS

Using the silica film interlayer, fifteen pairs of MaCorTM specimens were
successfully joined at temperatures between 1050°C and 1075°C. Five of the 15 specimens
were joined with no extemally-applied loading, while the 10 remaining specimens were
joined using a 20 gram dead-weight load applied during joining using a sintered alumina
disk placed on top to the specimen to be joined. Notches cut normal to the joined surfaces
(using the Sonic-Mill) retained their shape very well during the joining process. For
example, for a notch 331 microns depth and 228 microns wide prior to joining, the notch
dimensions after joining decreased by only about 5-percent, to a width of 314 microns and a

depth of 215 microns.

168

Directional Power meter Cylindricalsingle-mode
microwave cavrty

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

coupler ”l”. O
o 010 Specimen
Magnetron Circulator
/ \.
Z \ / :
Dumm load {127
y o 01: Casket
§ © [2:]
—[§° "E EOptical
Power supply controller pyrometer
Motor controller
for probe
1‘“ ‘
l

 

 

“III ‘1 III—l

 

 

 

000130 000 ,
00090 000

 

 

 

 

 

 

 

 

 

 

 

 

l
i—L E accufiber model 10
ll 2 Optical Fiber
“HE Therm omete

Motor controller

for short

Figure 66. Schematic diagram of the microwave apparatus used to sinter and to join the
ceramic specimens (after Lee, Case, and Asmussen, 1997).

169

Cavity wall

/ -\ View port A

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Casket Spejcrmen
View SALI /
PortB\? zvc\ :11111éLJ E ,/ 1) )
' : ------ 11111111
”’1 1:1—1 Optical pyrometer
SALI I
11 L '\ j Silica window

 

 

 

 

 

 

Cavity bottom plate

Figure 67. Schematic showing cross-sectional view of the microwave cavity, with the
refractory casket centered along the cavity axis. The casket, composed of a hollow
zirconia (ZYC) cylinder and disc-shaped aluminosilicate (SALI) endplates, encloses a
ceramic specimen. The specimen temperature is measured by an optical pyrometer.

170

The microwave-sintered alumina specimens also were joined using a spin-on silica
layer. Joining was accomplished at temperatures between about 1550°C and 1640°C.
Microstructural examination in both a conventional Scanning Electron Microscope (SEM)
and a Environmental Scaning Electron Microscope (ESEM) showed no apparent “bonding
layer”, the microstructure seemed quite continuous across what was the plane of the join.
(Unlike the SEM, no conductive metal specimen coating is required to examine ceramic
specimens in the ESEM, thus the joint region in the specimens could be examined in the
ESEM without possible interference from an applied conductive surface coating. The
joined alumina specimens were fractured in bend, and the fracture plane did not deviate at
the join. In addition, Vickers indentation cracks were induced into the joined alumina
specimens (near the interface), and the subsequent Vickers cracks propagated undeflected
across the interface. The macrocrack behavior and the indentation crack behavior both
suggest that the alumina/alumina bond has very high mechanical integrity. In addition to
the alumina specimens, zirconia specimens were joined at temperatures from about 1450°C
to 1520°C, but the mechanical properties of the bonds have not yet been fully evaluated.

The MgF2 specimens did not join using the spin-on silica film, but the MgF2
specimens did join upon heating at 700°C for 20 rrrinutes, using the sodium silicate solution
as a bonding agent. The IR transmittance of the microwave-joined MgF2 specimens was
measured over the wavelength range of 880 nm to 1600 nm using a spectraphotometer. For
specimens of equivalent thickness, the transmittance of the microwave-joined specimens
was nearly identical to the transmittance for un joined specimens over the measured
wavelength range. MgF2 specimens also were joined at under the same conditions (sodium

silicate solution bonding agent, heated to 700°C for 20 minutes), but using the conventional

171

radiant heating rather than microwave heating. Due to scattering at second phases observed
only in the bonded region of the conventionally-heated MgF2, the transmittance of the
conventionally-heated MgF2 was considerably lower than the microwave-joined MgF2 for
specimens of equal thickness. At 1600 nm the transmittance of the microwave-joined
specimens were about a factor of two higher than that for the conventionally-joined

material.

CONCLUSIONS

Joins made using microwave heating of spin-on interlayers can be made both quite
thin and quite strong. For the joined alumina specimens, the bond layer after joining could
not be detected using a Scanning Electron Microscope. In addition, a Vickers indentation
crack placed near the interface propagated across the interface without deflection.

The transmittance of optical and infrared materials can be a function of the joining
technique, as was observed for magnesium fluoride. Presumably, the degradation of optical
properties is due to optical scattering by second phases that form in the bond region during
firing. The genesis of the differences in microstructure between the conventionally-fired
and the microwave heated specimens needs further study.

Using spin-on interlayers and microwave joining, ceramic specimens can be joined
without significantly perturbing either the dimensions or the geometry of submillimeter

notches cut into the specimen (for the MaCorTM, alumina, and zirconia specimens).

172

ACKNOWLEDGMENTS

The author acknowledges the financial support of the Michigan Research

Excellence Fund provided through the Electronic and Surface Properties of Materials

Center, Michigan State University

REFERENCES

1.

Elssner, G., Diem, W. and Wallace, J. S., 1981, pp. 629-639 in “Surfaces and Interfaces
in Ceramic and Ceramic-Metal Systems”, Edited by J. Pask and A. G. Evans, Plenum
Press, New York.

Lee K. Y, Case E. D., Asmussen, J. Jr., and Siegel M., 1996, "Binder Burnout in a
Controlled Single-Mode Microwave Cavity," Scripta Materialia, 35[l]:107-111.

Lee K. Y, Case E. D., and Reinhard D., 1997, “Microwave Joining and Repair of
Ceramics and Ceramic Composites,” Ceramic Eng. and Sci. Proc., 18: 543-550.

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

Sandhage K. H., Schmutzler H. J., Wheeler R., and. Fraser H. L, , 1996, "Mullite
Joining by Oxidation of Malleable, Alkaline-Earth-Bearing Bonding Agents", J. Am.
Ceram. Soc., 79[7]: 1839-1850.

Santella M. L., "A Review of Techniques for Joining Advanced Ceramics", 1992,
Amer. Ceram. Soc. Bull., 71: 947-954.

Schaffer J. P., Saxena A., Antolovich S. D.,. Sanders T. H, Jr., and. Waner S. B, 1995,
pp. 718 - 720 in The Science and DCSIQ of Engineering Materials, Irwin Press,
Chicago.

Seiber K. N ., Lee K. Y and Case E. D., 1997, "Microwave and Conventional Joining
of Ceramics using Spin-on Materials,” pp. 941-949 in Proceedings of the 12th Annual
Advanced Composites Conference, Technomic Publishing Co., Lancaster, PA.

Yen T. F.,. Chang Y. H, Yu D. L., Yen F. S., Tsai D. S., and Lin I-Nan, 1991,

“Diffusion bonding of MgF2 optical ceramics,” Mat. Sci. and Engineering, A147 121-
128.

173

10. Yen T. F.,. Chang Y. H, Tsai D. S., Duh S. L., and Yang S. J., 1992, “LiF-film-assisted
diffusion bonding of MgF2 ceramics,” Mat. Sci. and Engineering, A154, 215-221.

174

3.8 Joining Dissimilar Ceramic Materials

JOINING DISSIMILAR CERAMIC MATERIALS

J. G. Lee, E. D. Case, L. Zeng, and M. A. Crimp
Materials Science and Mechanics Department

Michigan State University
East Lansing, MI 48824
Abstract
Bi-material ceramic joins will be discussed, where the joining was accomplished by
heating in a single—mode microwave resonant cavity. The joined ceramics include the
following combinations of polycrystalline ceramics: (1) alumina and a partially stabilized
zirconia (2) SiC platelet reinforced alumina and a glass ceramic, (3) hydroxyapatite and a
glass ceramic, (4) zirconia - 3 mol% yttria and a glass ceramic and (5) an 85wt% alumina
- 15 wt% partially stabilized zirconia particle composite and partially stabilized zirconia.

The glass ceramic used was a commercial glass ceramic reinforced by mica platelets.

Introduction
The past few years, researchers have bonded ceramics with a variety of metals, other
ceramics and with polymeric materials [1 -13]. A number of researchers have used brazing
techniques to join metals with ceramics such as silicon nitride [l], sialon [2], molybdenum
disilicide [3], and aluminum nitride [4]. In addition, ceramics have been joined to ceramics
using brazing, where a metallic interlayer is used to join the ceramic components. However,

the final bond thickness can be 40 to 50 microns thick or more [5].

 

E, D. Case, J. G. Lee, L. Zeng, and M. A. Crimp, “Joining of Dissimilar Ceramic Materials”, Joining of
Advanced and Speciality Materials 11, ASM International. Materials Park, OH. pp. 10-17 (2000)

175

In contrast with the studies that join ceramics using thick bond layers, work done by
the current authors and co-workers [6 - 10] has focused on minimizing the bond thickness.
While we have joined like materials (such as alumina to alumina, or silicon carbide to
silicon carbide), dissimilar ceramics have been joined, including: (1) alumina/zirconia- 3
mol% yttria, (2) SiC platelet reinforced alumina/glass ceramic, (3) hydroxyapatite/glass
ceramic, and (4) zirconia - 3 mol% yttria/glass ceramic, where the glass ceramic used was
MaCorTM, a commercial glass ceramic reinforced by mica platelets and (5) an 85wt%
alumina — 15 wt% partially stabilized zirconia particle composite and partially stabilized
zirconia.

As used in this paper, the term “dissimilar” refers to a number of material
properties. For example, among the ceramic materials included in this study ( alumina,
zirconia- 3 mol% yttria, SiC platelet reinforced alumina, hydroxyapatite, and a mica
reinforced glass ceramic, MaCorTM) the linear coefficient of thermal expansion ranges
from a 9 X 10'° K’1 for zirconia to 17.5 X 10“5 K" for HAP (Table 21). In addition, the
thermal conductivites and the Young’s modulus for the materials included in this study

also span a broad range of values (Tables 22 and 23).

Experimental Procedure
Materials
Two of the materials used in this study were commercially obtained as densified
billets, while the remaining materials were microwave sintered from commercially
TM

available powders. One of the commercial materials included in this study was MaCor

(Corning Code 9658), which is a flurophlogopite mica platelet reinforced glass ceramic

176

in the system SiO2-AI2O3-MgO-K2O-F. In addition to MaCorTM , the other commercial
material included was an A1203/SiC platelet composite material having 80 volume %
alumina and 20 volume % SiC platelets (Max Tech, Inc, Lansing MI).

The three commercial ceramic powders used in this study include a partially
stabilized zirconia powder, an alumina powder, and a hydroxyapatite (HAP) powder
(Table 24). The partially stabilized zirconia powder was a 3 mol % yttria-zirconia
powder (TZ-3YS, Tosoh Ceramic Division, Bound Brook, New Jersey) powder having a
vendor-specified initial particle size of 350 nm and 7 mZ/g specific surface area, as
determined by BET measurements [14]. A 99.99 % pure alumina power (TMDAR, Taimi
Chemicals Co. LTD) was also used, which has a vendor-specified specific surface area of
13.6 mzlg specific surface area and a 0.21 micron average initial particle size, and 2.32
g/cm3 green density. The calcium hydroxyapatite (HAP, Calo(OH)2(PO4)6) specimens
were prepared using a nominally 99% pure powder (CeracTM Incorporated Specialty
Inorganics) having an initial particle size of roughly one micron as determined by direct
SEM measurements on the HAP powders.

Specimen preparation

The alumina, zirconia, and the alumina/zirconia composite powder specimens were
first made into powder compacts using uniaxial pressing in a hard die at approximately
32 MPa. The as—pressed dimensions of the disk-shaped powder compacts were
approximately 2 cm in diameter and 2 mm in thickness. All of the specimens sintered in
this study were processed inside an aluminosilicate casket (Figure 68) using a microwave

cavity (Figure 69).

177

For the alumina and zirconia powders, four to seven compact powder specimens
were then sintered in a 2.45 GHz single mode microwave cavity (Model CMPR-250,
Wavemat, Inc., Plymouth, MI, Figure 69). The microwave power supply was a Sairem,
Model MW PS 2000, which generates from zero to 2000 Watts of microwave power.
During sintering, the sintering temperature was controlled by adjusting an input power
with automatic sliding short and probe [6 - 10].

Before sintering, the specimens were placed in a ceramic specimen enclosure
(casket) composed of a hollow, porous partially stabilized zirconia cylinder capped on the
top and bottom with disk-shaped aluminosilicate end plates (Figure 69). The casket was
then placed in the microwave cavity, which the center of the casket aligned coaxially
with the microwave cavity. The specimen temperature was monitored using an Optical
pyrometer system (Accufiber Optical Fiber Thermometer, Model 10, Luxtron Co.,
Beaverton, OR), which was sighted through a 0.5 mm circular hole that was drilled
through the porous zirconia cylinder that comprised the body of the casket.

In contrast to the microwave sintering of the alumina and zirconia compacts, all the
HAP specimens included in this study were sintered in air in a conventional furnace at
1300°C for 11 hours, giving relative densities of up to approximately 95 percent of
theoretical.

For the commercially densified materials, the as-received billets of both the
MaCorTM and the Al2O3/SiC platelet composites (Max Tech, Inc, Lansing MI) were cut
into specimens approximately 1 cm X 1cm using a low speed diamond saw.

All of the ceramic materials included in this study were polished using a series of

diamond paste, with grit sizes from 35 microns to 1 micron. After polishing, the

178

specimens were cleaned for 20 minutes in an ultrasonic cleaner using de-ionize water and
detergent.

After polishing and cleaning, the specimens were coated with the organic silica
precursor liquid (Emulsitone Company, Whippany, New Jersey). In order to coat the
specimens with a thin, uniform film of silica, the specimens were placed on a commercial
high speed substrate spinner. Four to five drops of the silica precursor were pipetted onto
the specimen surface. The specimens were spun at 3000 rpm for 20 seconds, followed by
curing in air in an electrically heated furnace for 20 minutes at 200°C.

Joining Procedure

Joining was performed via microwave heating. The coated Specimens were placed in
the same 2.45 GHz single-mode microwave cavity (Model CMPR-250, Wavemat Inc.,
Plymouth, MI) used in sintering the powder compacts. Also, the same cylindrical
zirconia refractory casket insulated the specimens during joining. Alumina dead weights
ranging from 20 to 50 grams were placed on the specimens prior to joining. The joining
was performed at temperatures between 1020°C to 1500°C (Table 25). A summary of the
joining conditions is given in Table 25.

The joined specimens were then sectioned using a low speed diamond saw. The
cut sections of the specimens were mounted with in a thermal setting polymer. The cut
surfaces were polished with a series of diamond paste. The joined specimens were then
examined in a Scanning Electron Microscope (SEM).

TEM specimens were prepared by sectioning joined specimens in a low speed
diamond saw. Disks 3 mm in diameter were then cut from the sections using an ultrasonic

disk cutter. The disks were glued into Molybdenum rings and dimpled using a VCR

179

Model D5001 dimpler. The final thinning was done using a Gatan PIPS with single-sided

milling at an angle of 4 degrees. TEM observations were made at 200 kV using a Hitachi

H-800 microscope.

Results and Discussion

Dissimilar ceramics may also be taken to mean dissimilar in terms of their use. The
ceramics included in this study represent a broad spectrum of current and potential uses,
as will be briefly discussed here.

MaCorTM is used as electrical and/or thermal insulators and structural components,
as well as in electrical and opto-electronic devices [15]. MaCorTM is a very interesting
material since it is machinable using ordinary metal working tools. The physical basis
for the machinability of MaCorTM is the mica platelets that are randomly distributed in its
glass ceramic matrix. Cracks that form during the machining of MaCorTM interact with
the mica platelets such that when the MaCorTM is machined, cracks deflect around the
platelets rather than propagating long cracks. Ceramic — ceramic joining that provides a
machinable ceramic base for a component or component assembly has many potential
advantages. For example, such machinable bases may aid in interfacing ceramic
components in a variety of environments, such as allowing convenient attachments for
engine or biological applications.

HAP is a bioacitve ceramic material, in that when it is in contact with either blood or
simulated biological fluid, bone grows on its surface. In fact, the major mineral

constituent of both bones and teeth is HAP, although living bones and teeth have a

180

significant non-mineral content. Consequently, HAP is used for hard tissue implants
because of its close resemblance to human bone and teeth.

Alumina and zirconia have widespread use as ceramic materials. The 3 mol% yttria-
partially stabilized zirconia included in this material has excellent fracture toughness.
Both alumina and zirconia have important uses as bioinert ceramics [l9], ceramics that
unlike HAP do not induce bone growth and their surfaces when subjected to biological
fluids, and hence are use materials for replacement joint surfaces, where one does not
wish to induce bone growth [19].

The alumina — zirconia joins (Figures 70 and 71) were quite interesting. The SEM
micrograph (Figure 70) shows are very interface that lacks a discemable interface. In
fact, the interface is slightly uneven (on the scale of the grain size within the specimen).
In greater detail, the TEM micrograph (Figure 71) shows that after joining, the interface
between the two materials lack an apparent interface layer, although a silica layer
(approximately 150 nm thick) had been applied to both the alumina and the zirconia prior
to joining. These types of microstructures are similar to those observed during Transient
Liquid Metal Phase bonding in alloy materials [21].

Figure 72 shows a channel, approximately 400 microns wide at the “top” of the
channel and about 175 microns deep that was formed in a 85 wt% alumina - 15 wt%
zirconia specimen. Figure 72 depicts the channel following sintering of the 13500 C for
one hour in the microwave cavity described in the experimental procedure section. The
specimen was then polished and joined with a sintered and polished zirconia disk (Figure
73), where the joining was done at 1450°C, with a hold time of 20 minutes and a dead

weight loading of 20 grams (Table 26). As was the case with the other joins, a silica

181

interlayer was used, where the silica film was applied to only the zirconia before joining.
Such channels could be of interest in to distribute fluids (such as cooling fluids, fuels, or
biological fluids) in a component.

For the MaCorTM — 3 mol% yttria-zirconia bonds, SEM and TEM micrographs are
shown in Figures 74 and 75, respectively. Figure 75 shows considerable grooving of
grains at the MaCorTM — zirconia interface, which may be indicative of considerable local
dissolution of the zirconia.

It should be emphasized that although silica interlayers were applied to all specimens
prior to bonding, preliminary TEM observations of the joints has revealed direct bonding
of the base materials, with no evidence of interfacial bond layers. For example, Figure 75
shows the interface between MaCorTM and 3-mole% Y stabilized ZrO2, which shows
extensive wetting of the ZrO2 by the glassy matrix of the MaCorTM. This results in
significant interlocking along the interface. In comparison, the TEM micrographs of the
A1203—Zr02 bonds (Figure 71), show some wetting, but the wetting that is observed is
more limited than in the MaCorTM-Zr02 case. The A1203 and ZrO2 appear to have
bonded directly and possesses sharp interfaces. These results indicate that there may be a
number of joining mechanisms operative in this specimens as one spans the spectrum of
materials with a considerable glassy phase (such as MaCorTM) to those materials with

very limited grain boundary phases, such as the alumina and zirconia materials included

in this study.

Acknowledgments
The authors acknowledge the financial support of the Composite Materials and
Structures Center and the Electronic Surface and Properties of Materials Center, College of

Engineering. Michigan State University.

183

Table 21. Thermal expansion coefficients for the Materials Included in this Study

 

 

 

 

 

 

 

Materials Temp. Interval Thermal expansion (10'6 K' References

MaCorw 25°C-300°C 9.3 15
A1203 25°C-1000°C 8.8 17
ZrO2 Room Temp 9 14
ZrO2 25°C-1000°C 1 l 14
HAP 25°C400°C 14.5 16
HAP 400°C-800°C 17.5 16

 

Table 22. Thermal Conductivites of the Materials Included in this Study

 

 

 

 

 

Materials Thermal conductivity References
MaCorm 1.46 W/m -°C

A1203 30.0-35.0 W/m -K

ZrO2 N/A

HAP 0.013 J s'Tm" K"

 

Table 23. Young’s moduli for the Materials Included in this Study

 

 

 

 

 

Materials Young’s moduli References
MaCorTM 64 GPa

A1203 390 GPa

ZrO2 190 GPa

HAP 80 —100 GPa

 

184

Table 24. Sintering conditions for materials used in this study.

 

 

 

 

 

Materials Initial Mass Sintering Sintering Final
particle Temp Time Density
size

Alumina 021 um 2 grams 1300°C 1 hour 96 %

Zirconia 0.59pm 3 grams 1375°C 1 hour 97 %

HAP 10 grams 1300°C 11 hours 2.974 g/crn3

 

 

 

 

 

 

 

Table 25. Processing conditions for dissimilar materials using microwave power.

 

 

 

 

 

 

Materials Joined Joining Hold Dead
Spin-on interlayer Temp(°C) Time Weight
(min) (gm)
ZrO2 with A1203 Silica film 1500 20 50
HAP with MaCorTM Silica film 1020 20 20
ZrO2 with MaCorTM Silica film 1020 20 20
sick/(1121523113 “h Silica film 1020 20 20

 

 

 

 

 

185

 

10.3cmforlnrpeubt
L 7.3cmtorsmallennui-t4I

VI

1'
Alumlno- —,
sum 1
and plate '—

 

 

 

 

 

72cm
glylnlonlam 3cm
View
l‘l010\ m

V

W ’—

3“ 4:05 cm
NUI'I'III'IO- - fir—
..... .,
and plate

Figure 68. Schematic of the insulating casket used in this study for both microwave
sintering and microwave joining.

 

 

 

 

186

Insulation
cuv it)”

 

Figure 69.Photograph of the single-mode, 2.45 GHz microwave cavity used both to Sinter
the alumina, zirconia, alumina/zirconia powders and to join all specimens included in this
study. The pyrometer and the insulating cavity (casket) also are shown.

187

A1203

4-Interface

5 um

 

Figure 70. SEM micrograph of the bond region of a joined alumina — zirconia specimen.
Note the absence of a discemable bond layer between the two joined specimens.

188

      

3 .-

 

A1.“

Figure 71. A TEM micrograph of the bond region of the same alumina — zirconia
specimen depicted by an SEM micrograph in Figure 70. In the TEM micrograph, no
significant bond layer is visible.

 

189

 

Figure 72. A SEM micrograph of a microwave-sintered 85 wt % alumina — 15 %
zirconia specimen containing a “pressed-in” channel.

190

A1203—Zr02

composite 106-Em

 

Figure 73. The notched 85 wt % alumina — 15 wt% zirconia specimen depicted in Figure
72, joined with a partially stabilized zirconia (3 mol% yttria — zirconia) specimen.

191

 

Figure 74. A SEM micrograph of the join between zirconia and MaCorTM.

192

 

Figure 75. A TEM micrograph of the MaCorTM - zirconia join, for which there appears
to be extensive wetting of the interface by the glassy matrix phase of MaCorTM.

193

10.

ll.

12.

13.

I4.

15.

References

. G. Chaumat, B. Drevet, and L. Vernier, J. European Ceram Society, 17 [15-16], 1925-

27 (1997).

A. P. Xian, J. Materials Science, 32 [23], 6387-93 (1997)

S. D. Conzonne, D. P. Butt, A. H. Bartlett, J. Mater. Sci, 32 [13], 3369-74 (1997)
D. Huh and D. H. Kim, J. Materials Res, 12 [4], 1048-55 (1997)

L. Esposito, A. Bellosi, S. Guicciardi, G. de Portu, J. Mater. Sci., 33 [7] 1827- 36
(1998)

K. N. Seiber, K. Y. Lee, and E. D. Case, pp. 941-949 in Proceedings of the 12th
Annual Advanced Composites Conference, Technomic Publishing Co., Lancaster, PA.
(1997)

E. D. Case, K. Y. Lee, and J. G. Lee, pp. 17 - 20 in Proceedings ofthe 33rd
International Microwave Power Symposium, lntemational Power Institute, Manassas,
VA. (1998)

E. D. Case, J. G. Lee, and K. Y. Lee, pp. 17 - 26 in Joining ofAdvanced and
Specialty Materials, M. Singh, J. E. Indacochea, and D. Hauser, eds., ASM
lntemational, Materials Park, OH. (1998)

E. D. Case, K. Y. Lee, J. G. Lee, and T. Hoepfner, pp. 27 - 34 in Joining of Advanced
and Specialty Materials, M. Singh, J. E. Indacochea, and D. Hauser, eds., ASM
lntemational, Materials Park, OH. (1998)

J. G. Lee, K. Y. Lee and E. D. Case, pp. 509 - 520 in Ceramic Transactions,
Volume 94, American Ceramic Society, Inc., Westerville, OH. (1998)

C. Mukherjee and Eldon D. Case, submitted, J. European C eram. Soc. (1999)

C. Mukherjee, E. D. Case, K. Y. Lee, and A. Lee, accepted for publication, J. ofMat.
Sci. (1999)

C. Mukherjee, E. D. Case and A. Y. Lee, Accepted for publication, Ceramic Eng. and
Sci. Proc., Volume 20. (1999)

For Tosoh TZ-3YS partially stabilized zirconia powders, data from vender web site
(www.tosoh.com)

MaCorTM data from vender (Corning) web site

194

16. Je-Won Choi, Young—Min Kong, and Hyoun- Ee Kim, J. Am. Ceram. Soc., 81 [7]
1743-48 (1998)

17. W.D. Kingery, H.K. Bowen, and D.R.U. Uhlmann, p 595, Introduction to Ceramics,
Second Edition, New York, NY. (1976)

18. M. Barsoum Fundamental of Ceramics, p 401 and 505, First edition, McGraw—Hill,
New York, NY. (1997)

19. L. L. Hench, J. ofAm. Cer. Soc., 81, [7] 1711 (I998)
20. Kijima T, Tsutsumi M, J. ofAm. Cer. Soc., 62, 455-60 (1979)

21. W. F. Gale, JOM, 51[2], 49 — 52, (1999).

195

Discussion and Conclusion

4.1 Thermal Stress in the Dissimilar Ceramics Joined in This Thesis

In this thesis, the dissimilar ceramics were joined using a silica spin-on interlayer
(section 3.8 in this thesis). The different thermal expansion between two dissimilar
ceramics caused thermal stress in the joint. We shall use research of Ning et a1. [61] and
Lee et al. [57] to help in our analysis of stresses developed in the materials joined in this
thesis.

Ning et al. [61] reported joining of Si3N4(13 mm square and 20 mm long) with a
99.992 % pure aluminum foil (no dimension was reported) to produce a 6 mm thick braze.
The bond layer formed a silica —a1umina noncrystalline layer after the specimen was heated
at 800°C between the Si 3N4 and Al braze. The thermal expansion mismatch caused thermal
stress in the joint upon cooling. According to Ning et al. [61], “If two dissimilar materials
can deform only elastically with stress and the stresses developing on both sides of the bond
layer do not interfere with each other” [59]. Ceramic materials used in this thesis deform
elastically to fracture. Ning et al. [61] evaluated the thermal stress of dissimilar joined
specimens using the mismatch in coefficient of thermal expansion by [57, 61]

E‘ El ( )AT (1)

0° =—0' - =—— at—a-
' 1 E,- + E j ’ l
where E is Young’s modulus, or is the thermal expansion coefficient, ATis (Tfinal'Tinitial) and

i and j are the materials being joined. Assuming elastic deformation, 0',- = E its,- and

O'j = Ejej , where the thermal strain is given by 81: or iATand ej = (IjAT. In figure 76,

material i, which has a lower thermal expansion coefficient, experiences a compressive

stress and material j, which has a higher thermal expansion coefficient, experience a

196

Elastic deformation
0' = EaAT

 

material 1'

 

material j

 

 

 

Upon cooling

 

 

 

 

 

—8;'
-> -— —- <-
' - ' - AT
_, L; I materralz I :‘J ‘_
4— 1 material j 1 81 -> alzajAT
4- ->
-> 4- Compressive
-> <-
-> 4- stress
4- -> °
‘_ _, Tensrle
" + stress
(I. = -6

Figure 76. A schematic of the thermal stress induced in two different joined materials upon
coohng.

197

tensile stress. The tensile and compressive stresses are equal in magnitude, thus, 0',- = —0 j .

Equation (1) may be derived as follows:

O’iinS[=—O'j=EjEj (2)
5total = 8i “Ej (3)
o- '0'
812—1. and£j=—Jf (4)
1 J
0i “j 0i 0i
e :—--—-=——+—— o-z—o') 5)
total Ei Ej Ei Ej( t J (
1 l
Etotal =01 (3+?) (6)
l J

solving for 0',- gives

EiEj (7)
0" = ———-—-——8
t Ei + Ej total
since Etotal = 8,- — 81° and Etotal = OtiAT- (IjAT: (01,-- Olj)AT (8)

since ATthe same in both cases. Substituting equation (8) into (7) yields

EiEj
—Ei+EJ‘

O',‘ (Oti—aj)AT=—0'j

Figure 77 showed different thermal stress distribution between two materials being
joined. Compressive stress (lower thermal expansion) was denoted with minus Sign while
tensile stress was denoted with plus sign.

Ning et al. [61] assumed the silica —alumina noncrystalline layer to be mullite. The

thermal stresses calculated according to equation (1) were (for the following dissimilar

198

Table 26. Thermal properties of materials used in reference [61].

 

 

 

 

 

 

 

Materials Thermal expansion (10’6 K") Young’s moduli
Al 23 70 GPa
Si3N4 2 300 GPa
Mullite 5 100 GPa

 

Table 27. Thermal properties of materials used in reference [57].

 

 

 

 

 

 

 

Materials Thermal expansion (10° 0C‘]) Young’s moduli
Inconel
600 12.1 214 GPa
Ti braze 8.9 106 GPa

 

199

 

 

material combinations) Si3N4/Al = 476 MPa, Si3N4/mullite = 114 MPa, and mullite/Al =
296 MPa, the data from Table 26. Ning et a1. [61] explained that the lower thermal stress
in the 513N4 and silica-alumina noncrystalline layer due to smaller CTE difference between
the Si 3N4 and silica-alumina noncrystalline layer than that of CTE difference between the
813N4 and Al. However, no cracks were reported for Si3N4/Si3N4 joint using the Al braze
with a joint strength of 146 MPa. Thus, the strength of the joint was not strong enough to
overcome the thermal stress. Following Ning et al. [61], Lee et al. [57] joined 10 mm x 10
mm x 5 mm bar of 99 % alumina that was joined with a 10 mm diameter and 5 mm height
Inconel 600 disk. The Inconel 600/alumina specimen showed a crack in the alumina joint.
Using table 27 and AT = 1000°C, the thermal stress was about 560 MPa. Due to the
thermal stress, a crack was produced continuously at the alumina joint (Figure 78). As
expected from Ning et al’s work [61], the joint strength was poor due to high thermal stress.
Lower thermal stress was calculated 90 MPa (using table 27). When a Ti braze was used as
a bond agent, no cracks were found.

Table 31 shows calculated thermal stress values of joined alumina with zirconia (3
mol % yttria), zirconia (3 mol % yttria) and MaCorTM with HAP used data from table 28
and 29. These thermal stresses may exceed the strength of materials being joined and thus
could produce a crack in the bond layer. However, unlike the results from Ning and Lee, no
cracks were observed despite the difference in CTE (Coefficient of Thermal Expansion)
between two substrates. Silica-film after coating and curing was about 130 nm (very thin),
thus, the bond layer was ignored to calculate thermal stresses of joined A1203 with ZrO2

(Figure 70). HAP with MaCorTM (Figure 38), and ZrO2 with MaCorTM (Figure 74).

200

interface

1

 
  
   

 

 

   

lfigher
g thennal
,: expansunr
CID
x

lower _

thennal

expansunr

 

Figure 77. A schematic of different thermal stress distribution between two materials that
have been joined along a planar interface [61].

201

 

Alunnna

A Him

+ lun
Inconel600

 

SOOlun

 

 

 

Figure 78. A continuous crack occurred in alumina due to compressive thermal stress. The
spacing between alumina and interface was about 500 um and the crack started from the
interface [57]

202

Table 28. Thermal expansion coefficients for the materials included in this thesis.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Materials Temp. Interval Thermal expansion (10'6 K”) References
25°C-300°C 9.4
MaCorTM 25°C-600°C 11.4 58
25°C-800°C 12.6
A1203 25°C-1000°C 8.8 38
ZrO; Room Temp 9 60
ZrO2* 25°C-1000°C l l 59
HAP 25°C-400°C 14.5 41
HAP 400°C-800°C l 7.5
2102* : 3 mol % ytrri — zirconia (Tosoh InC)
Table 29. Young’s moduli for the materials included in this thesis.
Materials Young’s moduli References
MaCorTM 64 GPa 43
A1203 390 GPa 39
ZrO2* 200 GPa 60
HAP 80 -100 GPa 36

 

 

 

 

ZrO2* : 3 mol % ytrri - zirconia (Tosoh InC)

Table 30. Poisson’s ratio for the materials included in this thesis.

 

 

 

 

 

 

 

 

Materials Poisson’s ratio References

MnCorTM 0.29 58
A1203 0.25 39
ZrO2 0.30 60
HAP 0.28 62

 

Table 31. Calculated thermal stress values of joined specimens

 

 

 

 

 

 

 

Joined Materials Calculated Thermal stress
A12O3/ZrO2 501 MPa

ZrO2/MaCorTM 100 MPa

HAP/MaCorTM 247 MPa

 

 

 

 

203

Alumina with zirconia joined specimen showed an almost invisible bond layer and
no cracks although occassional quasi-elliptical pores were observed in the bond layer.
Alumina has a lower thermal expansion coefficient. From equation (1), the calculated
thermal stress was 501 MPa which is a compressive in the alumina. The partially
stabilized zirconia tensile strength was 520 MPa [63]. The calculated thermal stress in the
partially stabilized zirconia was 501 MPa. Thus, the partially Stabilized zirconia tensile
strength was higher than the thermal stress induced due to joining PSZ with alumina. The
MaCorTM tensile strength was 90 MPa [64] and the calculated thermal stress in joined HAP
with MaCorTM was 247 MPa. Due to the thermal expansion coefficients (equation 1) the
stress in the MaCorTM should be tensile. Since the tensile stresses in MaCorTM far exceed its
tensile strength, the MACorTM should fracture according to equation (1). However, no

cracks were found with MaCorTM.

4.2 Summary of Results

Thin and strong bond layers can be produced using spin-on interlayers as bonding
agents in ceramic/ceramic joining. Using various spin—on interlayers, ceramic/ceramic
joining was successfully done in this thesis. The joined ceramics include MgF2/MgF2 ([49],
section 3.1 of this thesis), alumina/zirconia composites with alumina/zirconia composites, a
mica platelet reinforced glass ceramic (MaCorTM) with zirconia -3mol % yttria, MaCorTM
with polycrystalline hydroxyapatite and SiC platelet reinforced alumina with MaCorTM
([50], section 3.2 in this thesis), alumina with zirconia —3 mol % yttria ([56], section 3.8 in

this thesis), and SiC with SiC ([53], section 3.5 in this thesis).

204

The optical transmittance of joined MgF2 [49] was higher for the microwave joined
specimens than for the conventionally joined specimens. Microwave joined infrared
material (MgF2) showed optical absorption factor or (0.156) which was similar to the
or value measured for a polished MgF2 specimen (0.149). Conventionally joined MgF2
showed higher or (0.342) due to second phases in the bond-layer caused by optical

scattering, where the higher or value indicates the conventionally joined specimens
transmitted much less of the IR radiation than did the microwave joined specimens.
Bond-layers of joined dissimilar ceramics were in range of 0 to 3 microns thick
except SiC platelet reinforced alumina with MaCorTM joined specimen bond layer (10
microns, Figure 40) due to a reaction bond-layer. The “zero” thickness bond layer refers to
those cases in which a bond layer was not observed by either SEM or TEM observation.
Observation by TEM (work done by L. Zeng and M. A. Crimp in section 3.8 in this thesis)
showed no apparent bond layers in joined alumina with partially stabilized zirconia.
Ceramic/ceramic joining was successfully done without significantly changing the
dimensions in alumina/zirconia composites, fully stabilized zirconia, partially stabilized
zirconia, MaCorTM, and hydroxyapatite using a silica spin-on interlayer. The averaged
notch dimension of about 400 microns depth and 350 microns width were measured and
calculated from table 17, 18, 19, and 20 in this thesis. Table 32 shows each averaged notch
dimension in depth and width. Each measured notch depth and width produced in one
specimen was added, then divided by the number of notches. The dimension change notch

was calculated by the following equation.

Averaged notch depth before joining - Averaged width afterjoining X 100
Averaged depth before joining

 

205

Table 32. Averaged notch stability before and after joining in this thesis.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. # of Notch Notch . Notch Notch .
Material Width, Width, Difference depth, depth, Difference
Notched “(2:11 before after (%) before after (%)

joining joining joining Joining
Ale3/Zr02 3 317nm 311 11m 2 % 329 pm 327 um I %
**ZrOz 4 318 um 315 11m 1% 419 um 415 11m 1%
*ZrO2 5 372 11m 375 um I % 264 11m 247 11m 3 %
MaCorTM 3 340 11m 346 pm 2 % 431 run 444 um 3 %
HAP 3 335 um 330 pm 2 % 431 pm 444 um 3 %

 

*ZrO2: 3 mol % yttria-zirconia

**ZrO2: 8 mol % yttria-zirconia

*ZrO2 was joined with MaCorTM (no notches were made in MaCorTM)

206

 

The differences of the notch dimension before and after joining were 3 % in depth
and 2 We in width after average difference of each single notch from table 17, 18, 19, and 20
([52], 3.4 in this thesis).
Using BlackglasTM spin-on interlayer, one conventionally joined SiC (Hexoloy) with
SiC (Hexoloy) at 900°C showed smooth and thin (in range of 2 to 3 microns thick) bond

layer ([53], section 3.5 in this thesis).

4.3 Future Studies
Conventionally joined SiC with SiC followed heat-treatment by the microwave at
1200°C for 20 minutes showed a decreasing bond layer. Additional heating made a thinner

bond layer. However, the nature of additional heating should be studied further.

207

Appendices

APPENDIX A

Literature surveyed joining techniques cited in introduction section. The number was set

via references number.

Table A1. Joining materials, dimension, adhesive, thickness of adhesive, heating method,
heating rate cooling rate, Tum, hold time, bonding pressure, mechanical test, span length,
dimension, loading rate, cross speed, micrograph, thickness of bondlayer, strength of
material, bonding or shear strength, and atmosphere were written. Reference number 1 to
18 used bonding agent during joining.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Refere Materials Dimension Adhesive Thickness of Heating
-nces adhesive method
1 Si3N4 with 10 mm in diameter Pure nickel 0.2-0.8 mm Diffusio
Zirconia 5 mm thick foil n
bonding
furnace
Hot pressed Y-sialon
sialon
Hot pressed 1 5x1 5x 5mm bar or Y-sralon
sralon hot ressed dic of
Hot pressed p . Y-sialon Unknow
2 . 20 mm diameter Slurry
sralon -n
and 5 mm .
Pressureless . Y-sralo
. . thickness
srntered sralon
Pressureless Y-sialon
sintered sialon
Si3N4 with Not
3 Si3N4 20x20x8 mm glass Slurry Slurry reported
Ag-Cu-Ti No
AIN with Ag-Cu-Ti- No Convent
4 copper Not reported Co ional
or tungsten Ag-Cu-Ti- furnace
No
Nb
Sl3N4 With Y203-SIOz-
5 Si3N4 20x20x8 mm A12 03-Si3N4 Slurry Furnace
6 Sialon with 25x20x5 mm Y2O3-SiO2- Slurry Furnace
sialon Ale3-SI3N4
. . methyl-
7 RBSIC.Wlth 76x7x6 mm3 hydroxyl- 18i4ltm Furnace
RBSlC .
sfloxane
ZrO2/T1/ZrO 10 mm in diameter
ZrO /Pt/T'/ZrO 5 mm height f°r Convent
2 ' TZP Ti-6wt%Al- .
8 2 lelel mm for 4wt%-V mm)“ mm mm“
ZrO2/Pt/ Ti- Ti furnace
6wt%Al-
4wt%-V

 

209

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Refere Heating Cooling Tmax Hold time Bonding Mechani
-nces rate rate pressure -cal test
1 5- lOoC/min 1000- 90 min 20-37 MPa Shear
25°C/min 1 100°C strength
1130°C 1 hour 5 MPa
Not Not 1400:C 1 hour 5 MPa Microha
2 reported reported 1600 C 1 hour 5 MPa rdness
1600°C 1 hour 0 MPa test (Hv)
1600°C 1 hour 0 MPa
Three
3 10°C/min NOI 145(1),- 10-20 min No pressure point
reported 1750 C
bend
820-920°C 0-30min Not reported
4 12:30 0.2 K/sec 820-920°C 0-30min 8:23:11
820-9200C 0-30min
Three
5 regfied regcr’ied 1600°C 30 min 5 MPa point
bend
6 100°C/mi 100°C/min 1600°C 10 min 2 MPa no
11
7 1°C/min 1°C/min 800-1200°C 1 hour No pressure no
[328°C and 1-180 min 5 MPa Vickers
1494°C indentati
N N t °"
8 repofied repoited 1162°C and 15-180 min 5 MPa Not
1245°C reported
1162°C and 90 min 5 MPa Not
1245°C reported

 

 

 

 

 

 

 

 

210

 

 

 

 

 

 

 

 

 

 

 

 

 

Refere Span Dimension loading cross Micrograp Thickness
-nces length rate speed h of bond
layer
1 no no no no Yes No
Yes 5-10 um
Yes 1-2 um
Yes less than
0.5 urn
2 no no no no Yes less than
0.5 um
Yes almost
invisible
3 no 3x4x36mm no no Yes 20-35 1.1m
Yes ~3-10 pm
4 no no no no Yes ~ 140} m
Yes ~ 0.8—3.5 p.
m
5 no no no 0.5mm/ Yes 12 um
min
6 no no no no Yes 1.5 :05
um
inner Yes 6.5, 4.5,
7 6mm 1.5x3x15m 0.2mm/ and 2.5um
outer 12 m3 no min
mm
no no no no Yes Unknown
8 no no no no Yes Unknown
no no no no Yes Unknown

 

 

 

 

 

 

 

211

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

References Strength of materials Bonding or Atmosphere
shear strength
1 No No 4x 10’4 Pa (Vacuum)
1549 Hv 1371 Hv
1572 Hv 1550 Hv
2 1535 Hv 1553 Hv Flowing Nitrogen
1440 Hv 1431 Hv
1440 Hv No
3 700 MPa 400 MPa Flowing argon
118 MPa
4 Not reported 116 MPa Not reported
147 MPa
5 700 MPa (7) 550 MPa Flowing Nitrogen
6 No No Flowing Nitrogen
7 250:10 MPa 70-220 MPa Flowing argon
736 i 71 Hv Just annealed
8 metal 60 Hv Flowing argon
Not reported Not reported
Not reported Not reported

 

IQ
_
ls)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Refere Materials Dimension Adhesive Thickness of Heating
-nces adhesive method
9 Pressurelessly 20x20x8 slurry in the Slurry furnace
sintered Si3N4 mm3 system of Y-
with Si3N4 Si-Al-O-N
10 Sintered Sialon Bar Sn-5at% Ti 0.15 mm Cold
(3x4x20mm alloy wall
) type of
Vacuum
furnace
11 Si3N4 with Si3N4 bar CuNiTiB No Vacuum
(3x4x20mm paste Brazing furnace
) filler metal
12 Pressureless 15x25x20 MgO-Aleg- 1 mm Graphite
Sintered SiC min S102 (MAS) vacuum
filler furnace
and
Alumina
tube
furnace
l3 A12O3-ZrO2 Block sodium Slurry Microw
(30%) (15x4x4 silicate glass (powder ave
composites mm) powder mixed with (2.45
glycerol) GHz,
700 W)
Sintered
A1203 ceramic 92 % 0.6 mm
3 mm 10 96 ‘70 A1203 MICI'OW-
14 A1203 diameter No No ave
SN220 10 mm long Sheet of 0.6 mm
SN501
SN501 No No
A1203 (99%) disc( 10 mm sodium Slurry Microw-
15 x 6 mm) silicate glass (powder ave
powder mixed with
glycerol)
Refractory Not reported sodium Slurry Microw-
15 alumina (low silicate glass (powder ave
purity powder mixed with
glycerol)

 

213

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Refere Heating Cooling rate Tmax Hold time Bonding
-nces rate pressure
9 10°C/min Not reported 1450- 10-20 min 0-5 MPa
1650°C
10 Not Not reported 827°C 20 min 8mPa
reported (vacuum)
11 15 K/min 15 K/min 1007- 10 min Vacuum
1127°C
12 15°C/min 100°C/min 1500- 30 min No load
1600°C
13 within 2 Not reported 10000C 9 min No
min at
1000°C
40°C/min 15°C/min 1601°C and 3 min 0-2.4MPa
1720°C
[750°C or 3 min 0.6MPa
14
above
No No 1720°C 3 min 6.2 MPa
No
reached at cool down ~1000°C ~ 5 min Not
about fast after de reported
15 1000°C coupling
within
about10
min
reached at cool down
about fast after de
1000°C coupling
within
about10
min

 

 

 

 

 

 

214

 

 

 

 

 

 

 

 

 

 

References Mechanical Span length Dimension loading rate cross Speed
test
9 Three point No 3x4x36mm 0.5mm/mi no
bend 3 m
10 Four Point 27 mm No 0.2 18 mm
mm/min
l 1 Three Point No No 0.5 No
mm/min
12 Three point No 3x4x35mm No 0.5
mm/min
13 Three point 30 mm no no no
and Hv
Hardness
14 Four point No No No No
Hv No No No No
15 Hardness

 

 

 

 

 

 

215

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Refere Microgr Thickness of Strength of Bonding or shear Atmosphere
-nces -aph bond layer materials strength
9 Yes ~10 pm 700 MPa 650 MPa Flowing
Nitrogen
10 Yes ~0.l mm 400 MPa ~100 MPa (Sn- Vacuum
5Ti-Ag active
solder)
l 1 Yes Vary with Not reported 224.8 (highest Vacuum
braze (about joint) at 1353K (3.0-6.5x10'
48 — 58 3) Pa
microns)
12 Yes 10 pm at 370 MPa 380 MPa Up to Ar flowing
1500°C 800°C and 80
5 11m at MPa at 9000C
1600°C
No (but
joint
13 Yes about 500 strength was 28MPa Not reported
um stronger
than glass
phase)
No No from 370 MPa Air
1 4 Yes ? reference 420 MPa
No No 525 MPa ~390MPa N
No No 344 MPa ~520 MPa 2
15 Yes About 300 Glass: 515 582-1100 Hv Air
urn kg/mmz)

 

 

216

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Refere Materials Dimension Adhesive Thickness Heating
-nces of adhesive method
A1203 with Convention
A1203 al brazing
A1203 with Convention
A1203 15 mm al brazing
A1203 with square face SQ Brazing
A1203 20 mm (tightly
length fixed)
A1203 with 99.5 wt% SQ brazin
16 A1203 pure alumina NOt reported (loosely g
contact)
As received SQ brazing
Si3N4
Pre- 15x 10x 10 SQ brazing
Oxidation min
(900°C)
treated Si3N4
l7 Invar- disc 1mm Aluminum 0.5 mm fumace
alumium- high and 13 (99.9 wt%) thick and 7
silicon mm in sheet mm in
nitride- diameter diameter
alumium- (SiC)
Invar Invar Rod
(7mm in
diameter
and 7 mm
height )
18 Ti-6Al-4V lmmx0.63 63 wt%Ag, 50 microns Furnace
alloy sheets mm for 1.75 Ti , bal (840°C (brazing)
with both brazing alloy eutectic T)
alumina thin alumina and
plates Ti alloy (7
layers)

 

 

 

 

 

217

 

 

 

 

 

 

 

 

 

 

 

Refere Heating Cooling Tmax Hold Bonding Mechan Span loading
-nces rate rate time pressure ical test length rate
800°C 10min 0.01 MPa Three- 30 mm No
point
800°C 0.01 MPa
Not 6000& 50 MPa
16 reported No 800 C
600 & 50 MPa
800°C
600°C 50 MPa 25 mm
600°C 50 MPa
17 Not no 700°C- 2 to 0 to 0.15 Four 10 mm No
reported 950°C 30min MPa point
18 15°C/min ~15°C/ 850°C 10 min 0.15 MPa Three 40 mm
min point
bending
test

 

 

 

 

 

 

 

 

 

218

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Refere Dimension cross Microg Thickness Strength Bonding Atmosp
-nces speed -raph of bond of or shear -here
layer materials strength
3x3x40 0.5 Yes 0 350 MPa Not Air
mm mm/ joined
min
10 49 MPa Flowin
microns g
Argon
l 6 10 325 Air
microns MP3
300 228 Air
microns MPa
3x3x30 No 200 900 MPa 57 MPa Air
mm microns
200 404 Air
microns MPa
17 2x2x15 No Yes 10 to 20 400 to Ar, air,
mm3 microns 500 and N2
MPa
18 50 mm 0.3 Yes Not A1203 263 Vacuu
long, 5 mm mm/ reported :280 MPa MPa m (10‘
wide and min Ti alloy : 5torr)
4.4 and 1100
5.75 mm MPa
thick for
mutilayers

 

219

 

Table A2. Joining materials, dimension, adhesive, thickness of adhesive, heating
method, heating rate cooling rate, Tmax, hold time, bonding pressure, mechanical test,
span length, dimension, loading rate, cross speed, micrograph, thickness of bondlayer,
strength of material, bonding or shear strength, and atmosphere were written. The
number was set via references number. Reference number 24 to 29 used no bonding
agent during joining.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Refere Materials Dimension Adhesive Thicknes Heating Heating Cooling
-nces s of method rate rate
adhesive
HPSC rod 0.6 °Cs' 4.6 °CS’
(5x5x50mm) ' ‘
RBSC disc . 2.0 °Cs' 3.0 °Cs’
(15x9mm) Microwave 1 1
24 Y-PSZ rod (7x40 “° "° 23°12”, 2'45 0.5 °Cs' 3.5 °cs
Z, TE102 1 1
mm)
Y-PSZ tube 2.8 °Cs‘ 4.0 °Cs‘
(I.8x6.4x45) ' ‘
Clay- Tube (1/2” &
bonded 1”)
SiC SALD
25 HiNicalo N o (Selective
11 Fiber in Area Laser
SiC Deposition)
Hexoloy Tube (1/2” &
3/8”)
26 A1203 5 mm X 5 No No Microwave 120°C/ 120°C/
with mm X 25 min min
MgO mm
27 Si3N4 15x15x20m No (but No Diffusion No No
with m heated in bonding
Si 3N4 powder and then
(A1203 bed) heat treated
and Y203 at 19500C
additives for 3.6Ks
for
sinterin g)

 

 

 

 

 

 

 

220

 

 

Refere
-nces

Materials

Dimension

Adhesive

Thickness
of
adhesive

Heating
method

Heating
rate

Cooling
rate

 

Zirconia
with
zirconia/
Hydroxy
apatite
composit
es

 

Alumina
with
alumina/
Hydroxy
apatite
composit
es

Not reported

No

No

First CIP
at 300

MPa and
then HIP

Not
reported

Not
reported

 

 

 

Si3N4
with
Si3N4

 

15x15x20 min3

 

No

 

No

 

Diffusion
bonding

 

Not
reported

 

Not
reported

 

221

 

Jr

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

References Trudi Hold time Bonding Mechanical test
pressure
1310°C - 0.5 MPa No
74 1190-1375°C up to 2 0.5-2.0 MPa No
" 1400-1565°C 10 min 2,7 i 20 MPa Viker’s indent
[230°C 10 min 3.2 i 2.0 MPa N0
. Vicker indenter
7 . ‘7
-5 Usrng Laser Beam (.) (500g load)
26 1577°C to 2- 10min 0.03-0.5 MPa Three point
1877°C
27 1500°C 3.6Ks 20 MPa Bending test
1225°C 1 hr 200 MPa .
7
"8 Fractured durin, CIP Three bending
29 1500°C 3.6Ks ] 21 MPa Three bending

 

 

222

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

References Span length Dimension loading rate cross speed
No No No No
24
No No No No
25
26 30 mm no no 0.5
mm/min
27 outer 30 3x4x40mm No No
mm
inner 10
mm
No 30x3.5x3.5 No No
28 mm3
29 No 3x4x40 No No
mm3

 

223

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Refer- Micro Thickness of bond layer Strength of Bonding or Atmosphere
ences graph materials shear
strength
No No
Yes from ((10-20 microns at
“90°C, 0.5 MPa) to (3-5
24 microns at 1375°C, 2 No No Not reported
MPa)
Yes Invisible
No No
25 Yes Hard to see Not reported Vacuum
26 Yes Almost impossible to MgO: 150, 70MPa Air
fine out A1203I340
Joined
:~0.3GPa
27 Yes Almost invisible about 0.8 GPa Joined and N2
annealed
:~0.5GPa
Zirconia : 916 845 MPa
MPa
Yes Not identical Z/HAP
composites :
28 860MPa Not reported
A1203: 577 MPa
No Fractured Al/HAP. Fractured
composrtes: 601
MPa
29 Yes occasionally porous about 0.8 GPa 148 MPa N2 gas (0.1
interface and 1 MPa
pressure)

 

 

 

224

 

APEPPENDIX B

Sintering and joining heating rate using the Microwave energy in this thesis.

225

 

1400 . . n

 

1300*
1200*-
1100"

l

1

10001-
900
800:.
700

l
111

l
1

Temperature (0C)

7

. r
1

 

 

 

J

 

500 1 1 - l i l u l L l .
0 20 40 60 80 100 120 140

Time (minute)

Figure B1. A graph of sintering TM-DAR (Alumina). TM—DAR was sintered at
1300°C for 60 minutes with heating rate 10°C/min.

226

 

1400 . - . t . r
1300;
1200’—
1100-

Temperature (0C)

 

 

 

 

500 . 1 - l 1 1 m 1 Q l i 1 r
0 20 4O 60 80 100 120 140

Time (minute)

Figure B2. A graph of sintering partially stabilized zirconia. Partially stabilized zirconia
was Sintered at 1375°C for 60 minutes with heating rate 10°C/min.

227

1400 -

 

 

1300~

l

1200
1100~

Temperature (0C)

 

 

 

 

500 L 1 l 1 1 r l 1 1 m l 1 1 1 1
0 20 40 60 80 100 120 140 160 180

Time (minute)

 

Figure B3. A graph of sintering fully stabilized zirconia. Fully stabilized zirconia was
sintered at 1350°C for 60 minutes with heating rate 10°C/min.

228

Temperature (DC)

1400

 

1300:
1200!-
1100;
10001
900;
800;
7001
600;

 

500

 

 

H4 1

 

 

1 u l 1 1 1 1 l 1 l

 

0

20

40160180 100 120 140 100 180 200

Time (minute)

Figure B4. A graph of sintering alumina (85 wt %)-zirconia (15 wt %) composites.
Alumin/zirconia composites was sintered at 1350°C for 60 minutes with heating rate

10°C/ min.

229

 

8001- -

 

 

Temperature (0C)
8

 

 

 

500 ' 1 l 1 1 . l 1
0 10 20 30 40 50

Time (minute)

Figure BS. A graph of heating rate of microwave joined MgF2/MgF2 joined at 700°C for
20 minutes using sodium silicate solution spin-on interlayer.

230

 

1500

 

 

 

 

 

1400-
A 1300-
U
O 1200
V
113 1100-
3 3
C6 1000”
33 goof
a. .
E 800-
O.) .
1“ 700-
600-
500

40 60 180‘

Time (minute)

Figure B6. A graph of heating rate of microwave joined AKP 30/AKP 30 joined at
1400°C for 20 minutes using sodium alginate spin-on interlayer.

231

 

I

 

1500
1400-
1300:-
1200-
1100-
1000’-
900-
800»-
700’-
600:-
5000’10’210’3040’50’60’70’80
Time (minute)

Temperature (0C)

 

 

 

 

 

Figure B7. A graph of heating rate of microwave joined partially stabilized
zirconia/partially stabilized zirconia joined at 1500°C for 20 minutes using silica film
spin-on interlayer.

232

 

 

1500
1400—
13001
1200-
11001
1000-
900’-
800:
700’-
600-
5000 ’ 110°20’310’4’0’50 60 70 8019101100
Time (minute)

_T_T

Temperature (0C)

 

 

 

1 J 1 L 1 L 1

 

Figure B8. A graph of heating rate of microwave joined partially stabilized zirconia with
alumina joined at 1500°C for 20 minutes using silica film spin-on interlayer.

233

 

 

 

 

 

1100 - * l T l 1 l
A 1000» -
Q i
O
v 900- -
O.)
H
:5
g 800- -
H
8 1 .
E 700- -
Q) 1
I-1

600- -

500 1 1 I 1 l 1 l

0 20 30 40 50

 

Time (minute)

Figure B9. A graph of heating rate of microwave joined HAP with MaCorTM joined at
1020°C for 20 minutes using silica film spin-on interlayer.

234

 

 

 

 

 

 

l 100 . l e l T l r 1
l
A _
D
0
v —l
O.)
3.1
D
4—r _
CU
i-t
8.
E .1
PO") 1
1 1 1 . l r 1
20 30 40 50

Time (minute)

Figure B 10. A graph of heating rate of microwave joined partially stabilized zirconia with
MaCorTM joined at 1020°C for 20 minutes using silica film Spin-on interlayer.

235

1100 l

 

 

1000*-

900

Temperature (0C)
2

 

 

 

 

700 — -
600 .- 2
500 1 l 1 l n l . 1 i L

0 10 20 30 40 50

Time (minute)

Figure Bl l. A graph of heating rate of microwave joined SiC platelet reinforced alumina
with IMaCorTM joined at 1020°C for 20 minutes using silica film spin—on interlayer.

236

 

 

1600 _ie.i..tiirn

1500- i
1400- 1
1300
1200
1100
1000-

T

 

lfil'

T I T

800~
700*-

Temperature (0C)

 

 

 

 

500.111111111L11111L11
010 20 30 40 50 60 70 80 90100

Time (minute)

Figure B12. A graph of heating rate of microwave joined alumina/zirconia composites
with alumina/zirconia composites joined at 1450°C for 20 minutes using silica film spin-
on interlayer.

237

 

 

1500-
1400:
1300
1200-
1100

l

1

Temperature (0C)
8

 

 

 

 

 

010’20’30 40 50
Time (minute)

Figure B 13. A graph of heating rate of microwave joined fully stabilized zirconia with
fully stabilized zirconia joined at 1450°C for 20 minutes using silica film spin—on
interlayer.

238

APPENDIX C

Micrographs of joined ceramic/ceramic not included in results section.

239

1 til-1«W‘htlmkifltanrttnltttlt

 

Bond
layer

 

(b)
Figure C 1. ESEM micrographs of conventionally-joined AKP 30 (alumina) joined at

1400°C for 20 minutes using sodium alginate spin-on interlayer (a) lower magnification
of joined A1203 and (b) higher magnification of joined A1203.

240

 

1' W.MhM. to...“
[\J

 

(b)

Figure C2. ESEM micrographs of microwave-joined AKP 30 (alumina) joined at 1400°C
for 20 minutes using sodium a1 ginate spin-on interlayer (a) lower magnification of joined

A1203 and (b) higher magnification of joined A1203.

241

A1203

4—lnterface

Interface

Vickers
Indentation

 

(b)
Figure C3. (a) A micrograph of microwave-joined alumina with zirconia (3 mol % Y203)

at 1500°C for 20 minutes using silica spin-on interlayer showing interface and (b) a crack
induced byVikers indentation was deflected from ZrO2 to A1203 through interface.

242

 

Figure C4. SEM micrographs of zirconia (3 mol % Y203) with MaCorTM joined at
1020°C for 20 minutes Showing interface.

243 /

Interface

 

Figure C5. SEM micrograph of HAP with MaCorTM joined at 1020°C for 10 minutes.
The mica platelet rainforcement in the MaCorTM is evident in the micrograph.

244

”I 5%
it?
i

SiC/ MaCorTM
AlZQlReaction
-- . layer

1100 um

 

Figure C6. SEM micrograph of SiC platelet reinforced alumina with MaCorTM joined at
1020°C for 10 minutes.

245

 

 

 

 

(b)

Figure C7. ESEM micrographs of microwave-joined Z102 (3 mol % Y2O3) joined at
1500°C for 20 minutes using silica spin-on interlayer (a) lower magnification of joined
ZrO2 and (b) higher magnification of joined A1203.

246

«Interface

s'fit'n

Ln -\

 

(b)
Figure C8. SEM micrographs of conventionally-joined SiC joined at 1000°C for 20

minutes using Blackglass spin-on interlayer (a) lower magnification of joined SiC and (b)
higher magnification ofjoined SiC.

247

References

l.

10.

11.

R. H. Vegter and GD Ouden, “Diffusion Bonding of Zirconia to Silicon Nitride
Using NiKel Interlayers”, Journal of materials Science, Vol. 33. pp. 4525-4530
(1998)

L. M. Weldon, S. Hampshire and M. J. Pomeroy, “Joining of Ceramic Using Oxide
and Oxynitride Glasses in the Y-Sialon System”, Jouranl of the European Ceramic
Society, Vol. 17, pp. 1941-1947 (1997)

R. Xie, L.Huang, Y. Chen, X. Fu, “Effect of Chemical Compositions of adhesive and
Joining Processes on Bond Strength of Si3N4/Si3N4 joints”, Ceramic Intematioanl,
Vol.25, pp. 101-105 (1999)

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

R.Xie, L.Huang, and X.Fu, “Bond Strength and Microstructual investigation on
813N4/S13N4 Joint bonded with Glass-Ceramic”, Journal of Materials Science Letters,
Vol. 17, pp. 761-763 (1998)

P.A.Walls and M.Ueki, “Joining of SiAION Ceramics Using Composite B-SiAION-
Glass Adhesives”, Journal of American Ceramic Society, Vol. 75, No. 5, pp. 2491-
2497 (1992)

P.Colombo, V.Sglavo, and E. Pippel, and J. Woltersdorf, “Joining of Reaction-
Bonded Silicon Carbide Using A Preceramic Polymer”, Journal of Materials Science,
Vol. 33, pp. 2405-2412 (1998)

R. N. Corriea, J. V. Emiliano, and P. Moretto, “Microstructure of Diffusional
Zirconia-titanium and zirconia-(Ti-6wt% Al-4 wt% V) Ally Joints”, Jouranl of
Materials Science, Vol. 33, pp. 215-221 (1998)

R.Xie, L.Huang, Y.Chen, X.Fu, “Evaluation of S13N4 Joints: Bond Strength and
Microstructure”, Journal of Materials Science, Vol. 34, PP. 1783-1790 (1999)

A.Xian, “Joining of Sialon Ceramics by Sn-5 at % Ti Based Ternary Active Solders”,
Journal of Materials Science, Vol. 32, PP. 6387-6393 (1997)

C.G.Wang, H.P.Xiong, Z.F.Zhou, “Joining of Si3N4/S13N4 with CuNiTiB Paste

Brazing Filler Metals and Interfacial Reactions of the Joints”, Journal of Materials
Science, Vol. 34 [12], PP. 3013-3019 (1999)

248

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

H.L.Lee, S.W.Nam, B.S.Hahn, B.H.Park, and D.Han, “Joining of Silicon Carbide
Using MgO-Al2O3-SiO2 Filler”, Journal of Materials Science, Vol. 33, pp. 5007-5014
(1998)

S. Aravindan and R. Krishnamurthy, “Joining of Ceramic Composites by Microwave
Heating”, Materials Letters, Vol. 38, PP. 245-249 (1999)

H. Fukushima, T. Yamanka, and M. Matsui, “Microwave Heating of Ceramics and Its
Application to Joining”, Microwave Processing of Materials, vol. 124, Edited by
W.H. Sutton, M. H. Brooks, and l. J. Chabinsky. Materials Research Society,
Pittsburgh, PA, pp 267-272 (1988)

S. Aravindan and R. Krishnamurthy, “Microwave Joining of Alumina Using Sodium
Silicate Glass As a Bonding Agent”, Industrial Ceramic, Vol. 15[3], PP. 173-176
(1998)

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

K.Suganuma, T.0kamoto, and M.Koizumi, “Joining of Silicon Nitride to Silicon
Nitride and to Invar Alloy Using An Aluminum Interlayer”, Journal of Materials
Science, Vol. 22, pp. 1359-1364 (1987)

D.Sherman, “The mechanical Behavior of Layered Brazed Metal/Ceramic
Composites’, Materials Letters, Vol. 33, pp. 255-260 (1998)

KY. Lee, E.D. Case, and D. Reinhard, “Microwave Joining and Repair of Ceramics
and Ceramic Composites,” Ceramic Engineering and Science Proceedings, vol. 18,
pp. 543-550 (1997).

K.N. Seiber, K.Y. Lee, and ED. Case, “Microwave and Conventional Joining of
Ceramic Composites Using Spin-On Materials,” Proceedings of the 12th Annual
ASC Technical Conference on Composite Materials, Dearbom, MI, pp. 941-949
(1997).

O.M. Akselsen, “Review Advances in Brazing of Ceramics,” Journal of Materials
Science, vol. 27, pp. 1989-2000 (1992)

MG. Nicholas in “Joining of Ceramics”, edited by M.G. Nicholas, (Chapman and
Hall, London, 1990) pp. 15-18 and 71-93

J .D.Cawley, “Joining of Ceramic-Matrix Composites,” Journal of American Ceramic
Bulletin, vol. 68, no. 9, pp. 1619-1623 (1989)

J .G.P.Binner, J .A.Femie, and P.A.Whitaker, “An Investigation into Microwave

Bonding Mechanisms Via A Study Of Silicon Carbide And Zirconia”, Journal of
Materials Science, Vol. 33, PP. 3009-3015 (1998)

249

 

27.

29.

30.

31.

32.

33.

34.

35.

36.

37.

S.Harrison and H.L. Marcus, “Gas-Phase Selective Area Laser Deposition (SALD)
Joining of SiC”, Materials and Design, Vol.20, pp. 147-152 (1999)

. T.Sato, N.Takahashi and K.Shimakage, “Microwave Joining of Alumina to

Magnegia”, Journal of Ceramic Society of Japan, vol. 104, no. 10, pp. 905-907 (1996)

M. Nakamura, I. Shigematatsu, M. Nakanishi, K, Shimojima, Y. Yamada, and Y.
Shigegaki, “Microstructure of Joining Interfaces of Si 3N4 Ceramics Formed by

Diffusion Bonding”, Journal of Materials Science Letter, vol.l6, no. 20, pp. 1654-
1655 (1997)

. J .Li, L.Hermansson, and R.Soremark, “Joining of Ceramics of Different Biofuction

by Hot Isostatic Pressing”, Journal of American Ceramic Society, Vol. 76, pp. 2685-
2686 (1993)

M.Nakamura, I.Shigematsu, “Diffusion Joining of Si3N4 Ceramics by Hot pressing
Under High Nitrogen Gas Pressure”, Journal of Materials Science Letters, Vol. 16,
pp. 1030-1032 (1997)

J .G.PBinner, J .A.Femie, P.A.Whitaker, and TE. Cross, “The effect of composition
on the Microwave Bonding of Alumina Ceramics”, Journal of Materials Science, Vol.
33, pp. 3017-3029 (1998)

W.H. Sutton, “Microwave Processing of Ceramic Materials,” Journal of American.
Ceramic Society Bulltin. 68[2], pp. 376-386 (1989).

D.E.Clark and W.H.Sutton, “Microwave Processing of Materials”, Annual Review of
Materials Science, vol. 26, pp. 299-331 (1996)

H. Fukushima, T. Yamanka, and M. Matsui, “Microwave Heating of Ceramics and Its
Application to Joining”, Microwave Processing of Materials, vol. 124, Edited by
W.H. Sutton, M. H. Brooks, and I. J. Chabinsky. Materials Research Society,
Pittsburgh, PA, pp 267-272 (1988)

I.Ahmad, R.Silberglitt, Y.L.Tian, and J D.Katz, “Microwave Joining of SiC Ceramics
And Composites”, Ceramic Transactions, Microwaves: Theory and Application in

Materials Processing IV, vol. 80, pp. 455-463 (1997)

R.Silberglitt, W.M.Black, and J. D.Katz, “Recent Developments in Microwave
Joining”, MRS Bulletin, vol. 18, pp 47-50 (1993)

LL. Hench, “Bioceramics,” Journal of American Ceramic Soceity, vol. 81 [7], pp
1705-1728 (1998)

TM-DAR material data sheet from Taimi Co

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

. W.D. Kingery, H.K. Bowen, and D.R.U. Uhlmann “Introduction to Ceramics,”

 

Second Edition, New York, NY. pp 595 (1976))

M.Barsoum“Fundamenta1 of Ceramics” First edition, McGRAW-HILL, New York,
NY. pp. 401 and 505 (1997))

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

J .W.Choi, Y.M.Kong, and H.E.Kim, Reinforcement of

Hydroxyapatite Bioceramic by addition of Ni3Al and A1203, J. Am. Ceram. Soc 81:
[7] pp1743-l748 (1998)

Two kinds of zirconia data sheet from Tosoh Ceramics Division, Bound Brook, NJ
materials data sheet)

http://www.corning.com/search main.html

http://www.vidrinc.cornlvidrine/iropmat.htm

 

S.Rangarajan, R.Belardinelli, and P.B.Aswath, “Processing, Physical and Thermal
properties of BlackglasTM matrix composites reinforced with NextelTM Fabric”,
Journal of Materials Science 34: (3) 515-133 (1999)

http://wwwglassonline.com/dictionary/375.html

htt ://members.rediff.com/manua0arwal/RAM2.html

 

K.Y.Lee, “Microwave Processing of Ceramics and Ceramic Composites Using a
Single-Mode Microwave Cavity,” Ph. D Thesis, Michigan State University (1998)

J. G. Lee, K. Y. Lee and E. D. Case, “Joining of Diamond Thin Films to
Optical and IR Materials,” Ceramic Transactions, Volume 94, American Ceramic
Society, Inc., Westerville, OH. pp. 509 -— 520 (1998)

J. G. Lee and E. D. Case, “Microwave Joining of Particulate Composites”, Advances
in Ceramic Matrix Composites V, Ceramic Transactions Volume 103, American

Ceramic Society, Inc., Westerville, OH, pp. 571-581 (2000)

E. D. Case, K. Y. Lee, J. G. Lee, and T. Hoepfner, “Geometrical Stability of

Holes and Channels During Joining of Ceramics and Ceramic Composites,” in Joining
of Advanced and Specialty Materials, M. Singh, J. E. Indacochea, and
D. Hauser, eds., ASM lntemational, Materials Park, OH, pp 27-34 (1998)

J. G. Lee and E. D. Case, “Joining Ceramics to Produce Components with Precise

lntemal Channels”, in Innovative Processing and Synthesis of Glass, Composite, and
Ceramic Materials III, Volume 108, Ceramic Transactions, American Ceramic

Society. pp. 433 — 442 (2000).

251

 

 

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

J. G. Lee and E. D. Case, “Joining of Non-Oxide Ceramics Using Conventional and
Microwave Heating”, Ceramic Engineering and Science Proceedings 21[4], American
Ceramic Society. pp. 589-597 (2000)

E. D. Case, J. G. Lee, M. A. Crimp, D. K. Reinhard, and J. Malik, "Protective
Coatings for Infrared Materials", 23rd Cocoa Beach lntemational Conference on

Engineering Ceramics and Structures, American Ceramic Society, Cocoa Beach, FL,
January, pp. 145-152 (1999)

E. D. Case, K. Y. Lee, and J. G. Lee, “Joining of Polycrystalline Ceramics and
Ceramic Composites Using Microwave Heating,” in Proceedings of the

33rd lntemational Microwave Power Symposium, lntemational Power Institute,
Manassas, VA, pp. 17-20 (1998)

E. D. Case, J. G. Lee, L. Zeng, and M. A. Crimp, “Joining of Dissimilar Ceramic
Materials”, Joining of Advanced and Speciality Materials H, ASM lntemational,
Materials Park, OH. pp. 10-17 (2000)

H.R.Hwang, R.Y.Lee, “The Effect of Metal Coating on the Diffusion Bonding in
A1203,/Inconel 600 and the Modulus of Rupture Strength of Alumina,” Journal of
Materials Science, vol. 31, pp. 2429-2435 (1996)

http://www.accuratus.com/MACOR.htm

 

http://wwwtosoh.com

 

N.Yoshikawa, A.Kikuchi, S.Taniguchi, and T.Takahashi, “Tetragonal to Monoclinic
Transformation in Y-TZP Joined with Metallic Materials,” Journal of Materials
Science, vol. 34, pp. 5586 (1999)

X.S.Ning, T.0kamoto, Y.Miyamoto, A.Koreeda, and K. Suganuma, “Effect of Oxide
Additive in Silicon Nitride on Interfacial Structure and Strength of Silicon nitride

Joints Brazed with Alumina,” Journal of Materials Science, vol. 24, pp. 2865-2870
(1989)

D.E.Grenoble, J .L.Katz, K.L. Dunn, R.S.Gilmore, and K.L.Murty, “The Elastic
Properties of Hard Tissues and Patites,” Journal of. Biomed Materials. Research, vol.
6 (3), PP. 221-223(1972)

K.Noguchi, M.Fujita, and T.Makaki, and M.Mizushina, “Tensile Strength of Yttria-
Stabilized Tetragonal Zirconia Polycrystals,” Journal of American Ceramic Society,

vol. 72 [7], pp. 130—1307 (1989)

www.ferroceramic.com/tables/t_07.htm