close

Вход

Забыли?

вход по аккаунту

?

Growth and characterization of barium tantalate-based microwave ceramics and barium and strontium titanate ferroelectrics

код для вставкиСкачать
GROW TH AND CHARACTERIZATION OF BARIUM TANTALATE-BASED
MICROWAVE CERAMICS AND BARIUM AND STRONTIUM TITANATE
FERROELECTRICS
by
Shaojun Liu
A Dissertation Presented in Partial Fulfillment
of the Requirements for the Degree
Doctor of Philosophy
ARIZONA STATE UNIVERSITY
December 2005
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
UMI Number: 3194934
Copyright 2006 by
Liu, Shaojun
All rights reserved.
INFORMATION TO USERS
The quality of this reproduction is dependent upon the quality of the copy
submitted. Broken or indistinct print, colored or poor quality illustrations and
photographs, print bleed-through, substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.
®
UMI
UMI Microform 3194934
Copyright 2006 by ProQuest Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
ProQuest Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346
R e p r o d u c e d w ith p e r m issio n o f th e co p y rig h t o w n er . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
GROW TH AND CHARACTERIZATION OF BARIUM TANTALATE-BASED
MICROWAVE CERAMICS AND BARIUM AND STRONTIUM TITANATE
FERROELECTRICS
by
Shaojun Liu
has been approved
November 2005
APPROVED:
. Chair
SupervisoryXjoimnittee
ACCEPTED:
Director of th e Program
Dean, Division of G raduate Studies
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
ABSTRACT
In this thesis, we explore the growth and properties of electronic ceramics used in mi­
crowave resonator and energy storage. Their applications require high dielectric constants,
which do not vary significantly with tem perature and loss dissipation loss.
Single-phase B a ( C d i/ 3 T a 2 / 3 )Os powder is produced using conventional solid state
reaction methods. Ab-initio electronic structure calculations show th a t the covalent nature
of the directional d-electron bonding in these high-Z oxides plays an im portant role in
producing a more rigid lattice with higher melting points and enhanced phonon energies
and consequently resulting materials with a high dielectric constant and a low microwave
loss for B a (C d i/z T a 2 /z)Oz and B a ( Z n i/z T a 2 /z)Oz ceramics.
B a {C d iizT a 2 / 3 )Oz samples with high sintering density and excellent microwave
properties are made with boron oxide as sintering aid at 1200 — 1350°C, corresponding
to tem peratures 300°C lower than samples prepared without a sintering aid. XRD com­
bined with High Resolution Electron Microscopy (HREM) indicates th a t B a (C d \/z T a 2 /z)Oz
ceramics prepared with boron oxide have a well-ordered hexagonal structure. Transmission
Electron Microscope (TEM ) results indicate th a t the improvement in densification con­
tributes to the liquid sintering mechanism for boron concentrations exceeding 0.5wt%. An­
nealing treatm ent and high boron concentrations are also found to improve the microwave
properties. For example, B a (C d i/z T a 2 /z)Oz doped with 0.5wt% B 2Oz ceramics annealed
at 1250°C for 40 hours has a dielectric constant (er ) and tem perature coefficient of resonant
frequency (ry) of 32 and 80 ± 15ppm/°C respectively and a loss tangent (Q) of < 2 x 10-5
at 2 GHz.
Ceramic injection molding methods were subsequently developed to fabricate the mi­
crowave devices. A high sintering density (~ 94%) sample with er (~ 30), tj (0.1ppm/°C),
iii
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
and loss tangent (< 1.7 x 10~5) at 2 GHz was achieved using a high tem perature 1680°C
and 48h sintering process.
Doping Sc into Bao^SrQ.zTiOs ceramics changes its crystal structure from tetrag­
onal to rhombohedral structure and significantly reduces the dielectric constant of
B a o^SrosT iO s.
In contrast, BaT iO s and B a o jS ro ^ T iO ^ with V-doping m aintain the
tetragonal crystal structures. Leakage current in these m aterials can be reduced by doping
with vanadium. The leakage current is also strongly affected by point defects induced by
neutron damage or annealing treatm ent for undoped B a o^Sro^T iO ^ ceramics.
iv
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Dedicated to My Parents, Wife and little Sabrina
In memory of my father, Chenke Liu
v
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
ACKNOWLEDGMENTS
At the beginning, I wish to thank my father, Chenke Liu, a great father with two
sons, a daughter, two grandsons and two granddaughters, who has fight with cancer for
three years, passed away at the moment I finished my dissertation. However, I can not sit
with him in his last life moment. This is my lifetime regret. W hat I can do is to beg his
forgiveness.
I would like to feel my deepest gratitude towards my dissertation advisor, Professor
N athan Newman. Throughout the years, without his invaluable guidance, numerous en­
couragements, and financial and spiritual supports he has provided, it would not have been
possible to finish my research and Ph.D dissertation.
I am grateful for Professor David J. Smith, Professor W illiam Petuskey and Profes­
sor Mark Van Schilfgaarde for their valuable comments while correcting this thesis. Prom
bottom of my heart, I want to thank Professor David J. Smith, Professor William Petuskey
and Professor M ark Van Schilfgaarde for their insightful comments, valuable suggestions,
and strong encouragements throughout the course of my thesis research. I am also thankful
to Professor David J. Smith, Professor William Petuskey and Professor M ark Van Schilf­
gaarde for serving on my thesis supervisory committee. I like to thank Professor Jian Sun,
a visiting scholar Professor from Shanghai Jiaotong University, for his beautiful TEM and
HREM work. I also would like to thank David W hite for their help and encouragements
during my PhD study in ASU.
I am forever in debt to all other colleagues in the Newman research group for their
assistance throughout the years. Dr. Rakesh Singh helped me with many measurements
and suggests th a t provided me with extremely im portant insight to this research. Especially
I like to thank my ofhcemates, Dr. Zhaoyang Fan, Mr. Stephen Wu, Dr. Jihoon Kim, and
vi
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Dr. Lei Yu. We spent a large amount work and studying tim e together. I want to thank all
other members of the group for the m utual support, sharing knowledge, and enlightening
discussion. These members include: Dr. R. Singh, Mr. H.X Liu, Mr. R. Nandivada, Mr.
M. Espinasse, Mr. C. Beach, Mr. P. Bandaru, Mr. S. Bandyopadhyay, Mr. B. Strawbridge,
Mr. Y. Chen, Mr. L. Hao, Mr. Zhizhong Tong, and Mr. S. M ohapatra.
I am very grateful to my friends at Arizona State University, Dr. Xiaolin Gao, Mr.
Lin Gu, Dr. Xiaolong Fang, Dr. Zhian He, Mr. Yinghui Na, Mr. Liqing Ke, Ms. Lin Zhou,
Ms. Hua Wang, Mr. Diafeng Gu, and Mr. Wei Cao for their encouragement and friendship.
At last, Xian Liu, my dearest and a great wife, gave me unselfish encouragement,
support and love even when I struggled to go through a tough time. I am really indebted
to her. I would also like to thank my daughter, little Sabrina Jiayi, for all the happiness
she has brought me. She is my source of motivation . Many thanks to my parents, my
parents-in-law, my brother and sister for their support and love.
vii
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
TABLE OF CONTENTS
Page
LIST OF T A B L E S ......................................................................................................................
xiii
LIST OF F I G U R E S ..................................................................................................................
xiv
CHAPTER 1
IN T R O D U C T IO N ........................................................................................
1
1.1. Research background for microwave c e r a m ic s ....................................................
1
1.1.1. M aterials requirem ents..................................................................................
3
1.1.2.
M a t e r i a l s .........................................................................................................
6
1.1.3.
M easurement of dielectric p r o p e r tie s ........................................................
16
1.2. Research background for fe rro e le c tric s ..................................................................
17
1.2.1.
Barium titan ate (B a T iO s )
........................................................................
18
1.2.2.
Breakdown m e c h a n is m s ...............................................................................
23
1.3. M aterials p r o c e s s in g ..................................................................................................
24
1.3.1.
Powder s y n th e s is ............................................................................................
24
1.3.2.
C a lc in a tio n .....................................................................................................
25
1.3.3.
P r e s s in g ............................................................................................................
25
1.3.4.
Sintering
.........................................................................................................
25
1.4. Research motivation and thesis o rg an izatio n ........................................................
29
1.4.1.
Research motivation and g o a l .....................................................................
29
1.5. Thesis o rg an iz atio n .....................................................................................................
30
CH A PTER 2
EXPERIM ENTAL AND THEORETICAL INVESTIGATION OF
BARIUM TANTALATE BASED CERAMICS
...........................................................
viii
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
32
Page
2.1. In tro d u c tio n ..................................................................................................................
32
2.2. Experim ental p ro ced u res...........................................................................................
33
2.3. Theoretical a p p ro a c h ..................................................................................................
36
2.4. Results and D iscu ssio n ..............................................................................................
37
B a ( C d i/ 3 T a 2 /s ) 0 3 therm ogravim etry e x p e rim e n ts ..............................
37
2.4.2. X-Ray d iffrac tio n ............................................................................................
40
2.4.3. Densities and M ic ro s tru c tu re .....................................................................
42
2.4.1.
2.4.4.
Ordered
domain
B a {C d ii^ T a 2 /z)Oz
and
its
boundaries
in
undoped
.....................................................................................
42
2.4.5. Dielectric P r o p e r tie s .....................................................................................
49
2.4.6.
2.5.
structures
Local Density-Functional Calculation for B a (C d 1 / zT a 2 /s)O z and
B a { Z n l i ZT a 2 / z ) O z ........................................................................................
51
C o n clu sio n s..................................................................................................................
56
CH APTER 3
MICROSTRUCTURAL AND DIELECTRIC PR O PER TIES OF BAR­
IUM CADMIUM TANTALATE CERAMICS W ITH BORON OXIDE AS SINTER­
ING AID
...............................................................................................................................
58
3.1.
In tro d u c tio n ..................................................................................................................
58
3.2.
Experim ental p r o c e d u r e ...........................................................................................
59
3.3.
Results and D iscussion..............................................................................................
61
3.3.1. X-ray d if f r a c tio n ............................................................................................
61
3.3.2.
Sintering density and m icrostructure of boron-doped B a ( C d i/ 3 T a 2 / z)Oz 64
3.3.3.
TEM o b s e rv a tio n s ........................................................................................
ix
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
67
Page
3.3.4.
Sintering m e ch an ism ....................................................................................
73
3.3.5.
Dielectric p r o p e r tie s ....................................................................................
75
3.4. Conclusion
CH APTER 4
..................................................................................................................
77
DIELECTRIC-LOADED MOBIUS RESONATOR AND STRUC­
TURAL, CHEMICAL AND DIELECTRIC PRO PERTIES OF CERAMIC IN JEC­
TION MOLDING BARIUM ZINC TANTALATE MICROWAVE CERAMICS
.
79
4.1. Research b ack g ro u n d ..................................................................................................
79
4.1.1.
Concept of Mobius r e s o n a to r ....................................................................
79
4.1.2.
Dielectrically loaded Mobius r e s o n a to r ....................................................
80
4.1.3.
Ceramic injection molding (C IM )..............................................................
80
4.2. Experim ental p r o c e d u r e ............................................................................................
84
4.2.1.
Dielectric-loaded Mobius r e s o n a to r ..........................................................
84
4.2.2.
Synthesis and characterization of BZT ceram ics...................................
85
4.3. Results and d is c u s s io n ...............................................................................................
86
4.3.1.
Dielectric Mobius r e s o n a to r .......................................................................
86
4.3.2.
XRD analysis and m icro stru ctu re.............................................................
87
4.3.3.
Dielectric properties of ceramic injection molded B a { Z n i/z T a 2 /z)Oz
93
4.3.4.
Dielectric-loaded Mobius bandpass f i l t e r ................................................
96
4.4. C o n c lu sio n s..................................................................................................................
98
CH APTER 5
D IELECTRIC PROPERTIES AND I-V CHARACTERISTICS OF
VANADIUM AND SCANDIUM DOPED BARIUM AND STRONTIUM T I­
TANATE UNDER HIGH DC BIAS V O L TA G E...........................................................
x
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
99
Page
5.1.
In tro d u c tio n ..................................................................................................................
5.2.
Experim ental p r o c e d u r e ................................................................
102
5.3.
Results and d is c u s s io n ..............................................................................................
103
5.3.1.
X-Ray patterns and crystal s tru c tu re .......................................................
103
5.3.2.
M icrostructure and sintering d e n s it y ........................................................
106
5.3.3.
Dielectric p r o p e r tie s .....................................................................................
106
5.3.4.
I-V characteristics for V -dop ed B a o .iS r o .z T iO z ................................
109
5.3.5.
Dependence of dielectric constant on the freq u en cy .............................
115
C o n c lu sio n s..................................................................................................................
117
5.4.
CH APTER 6
99
INVESTIGATION OF PO IN T D EFECTS INDUCED BY NEUTRON
DAMAGE AND ANNEALING AND TH EIR E FFE C T ON LEAKAGE CURRENT
IN BARIUM AND STRONTIUM TITANATE F E R R O E L E C T R IC S ....................
120
6.1.
In tro d u c tio n ..................................................................................................................
120
6.2.
Experim ental p r o c e d u r e ...........................................................................................
122
6.3.
Point defects induced by neutron irradiation and its effect on the leakage
6.4.
c u r r e n t............................................................................................................................
123
6.3.1.
XRD p a t t e r n .................................................................................................
123
6.3.2.
M agnetic measurement
..............................................................................
123
6.3.3.
I-V c h a ra c te riz a tio n ....................................................................................
125
Point defects induced by annealing treatm ent and its effect on the leakage
current of B a T i O z .....................................................................................................
127
6.4.1.
127
Leakage current under different annealing condition
..........................
xi
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Page
6.5.
6.4.2.
Activation e n e rg ie s...................................................................................
130
6.4.3.
M agnetic measurement and E P R .........................................................
134
Conclusion
137
CONCLUSION AND FUTURE W O R K ............................................
139
Conclusions for barium tantalate-based microwave c e r a m ic s ....................
139
CH APTER 7
7.1.
.................................................................................................................
7.1.1.
M icrostructure, structural, chemical, electronic and high frequency
dielectric properties of barium cadmium tantalate-based ceramics . .
7.1.2.
139
M icrostructure and dielectric properties of B a ( C d i/3T a 2/ 3) 0 3 mi­
crowave ceramics synthesized with a Boron oxide sintering aid
. . .
140
7.1.3. Dielectric-loaded Mobius resonator and structural, chemical and di­
electric properties of ceramic injection molded B a { Z n i / 3T a 2/ 3)Oz mi­
7.2.
crowave c e ra m ic s ......................................................................................
141
Conclusions for barium and strontium titan ate ferro electrics....................
141
7.2.1. Dielectric properties and I-V characteristics of Vanadium and Scan­
dium doped B a i - xS r xT i 0 3 under high dc bias v o ltag e..................
7.2.2.
7.3.
141
Effect of neutron irradiation and annealing treatm ent on leakage current 142
Future w o r k ............................................................................................................
143
7.3.1.
Microwave dielectrics................................................................................
143
7.3.2.
F e rro e le c tric s.............................................................................................
144
R E F E R E N C E S .......................................................................................................................
145
xii
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
LIST OF TABLES
Table
Page
1.
Summary of the dielectric properties of ceramics dielectrics, source [22]
. .
2.
A lternate paths for m atter transport during the initial stages of sintering.
Source: [ 3 9 ] ...................................................................................................................
3.
27
The lattice constants of B a ( Z n x C d (i/z-x)Ta 2 /z)Oz powder samples fit to the
cubic and hexagonal structure...................................................................................
4.
7
41
Table of lattice positions of B a ( Z n i/z T a 2 /z)Oz and B a {C d 1 /z T a 2 /i)O z in
Cartesian coordinates. Note: the dimensions are scaled to the cubic unit cell
dimensions. Note: Done by Dr.M ark V. Schilfgaarde..........................................
53
5.
Particle size analysis of samples annealed for different durations......................
67
6.
Resonant frequency and Q of the dielectric-loaded Mobius wire resonator.
Note th a t the Q is smaller than anticipated from the dielectric alone as a
result of losses in the gold wire..................................................................................
96
7.
Calculated to tal trap density for V-doped BaQ.tSra.zTiOz.................................
113
8.
Concentration of param agnetic point defects irradiated by different neutron
dose...................................................................................................................................
xiii
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
125
LIST OF FIGURES
Figure
Page
1.
The tem perature dependence of the T C F of polycrystalline T i 0 2. Source [22]
2.
A comparison of the tanS versus tem perature at 3 GHz of YSZ-doped and
8
undoped fully dense T iO 2 - Source [22] ..................................................................
9
3.
The tem perature dependence of the tanS of ZTS at 7.8 GHz. Source: [22] .
10
4.
Two crystal structures for B a ( B i / 3 T a 2 /s)Os complex perovskites, where B =
M g ,o r Z n . (a) disordered ; (b) ordered, source: [26]
5.
.......................................
Tem perature dependence of tand for the B a ( M g 1 / 3 T a 2 /s ) 0 3 at 8.3 GHZ.
Source [22]
6.
13
14
The quality factor, grain size and density of BZT as a function of sintering
tem perature. Source: [ 3 0 ] ........................................................................................
15
7.
Diagram of the configuration for cavity technique.................................................
15
8.
Illustration of the magnetic field distribution for the T E qi$ mode. Source [22]
16
9.
Crystal structure of B a T iO s . Source: [ 4 3 ] ............................................................
18
10. Phase diagram of the B aO — T i 0 2 system (> M m o l.% T i 0 2 )- Source: [42] .
19
11. Tem perature dependence of dielectric constant as a function of dc-biasing
field for B a x S r i - xTiOs system. Measuring frequency is 10 kHz. Source: [115] 22
12. T he neck between particles forms a miniscus which exerts capillary pressure
drawing particles together. Source [3 9 ]..................................................................
26
13. A lternate pathes for mass transport during the initial stages of sintering.
Mechanism number is explained in Table 2. Source: [39]...................................
27
14. Therm ogravim etry measurements of the mass loss from B a {C d i/^T a 2 /^)0^
as a function of time and tem perature B a (C d i/ 3 T a 2 / 3 ) 0 3 ................................
xiv
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
37
Page
Figure
15.
X-ray diffraction spectra of decomposed B a (C d i/ 3 T a 2 / 3 ) 0
3
powder after ex­
posure to 1050°C during the therm ogravim etry measurements.........................
38
16.
Ellingham diagram for B a{C d i/ 3 T a 2 / 3 )Oz............................................................
39
17.
X-Ray diffraction spectra of B a ( Z n xC d i/ 3 _xT a 2 /z ) 0 3 powder........................
40
18.
X-Ray diffraction spectra of sintered B a {C d i/ 3 T 0 ,2 / 3 ) 0 3 synthesized with
2wt% ZnO as sintering aid..........................................................................................
19.
Dependence
of
relative
density
on
sintering
tem perature
41
for
B a ( C d i/ 3 T a 2 / 3 ) 0 3 ceramics synthesized with 2wt%ZnO sintering agent.
Note: The theoretical density of B a (C d iiz T a 2 /z) 0
20.
3
is 7.94g /c m 3 ...................
Scanning electron micrograph of Ba{Cdi/zTa, 2 /z)Oz with 2wt%ZnO sintered
at 1550°C for 48h..........................................................................................................
21.
43
43
Selected area electron diffraction pattern for B a (C d i/ 3 T 0,2/ 3)O3 viewed along
the [110] zone axis. Source (J. Sun, S. J. Liu, N. Newman, and D. J. Smith,
M at. Res. Soc. Symp. Proc. Vol. 783, 2004 M aterials Research Society,
B 5 .12.1.).........................................................................................................................
22.
44
Ordered domain structures for B a (Cd 1/3T a 2/ 3 )O3 ceramics w ithout boron
additive sintered at relatively high tem perature. Source (J. Sun, S. J Liu, N.
Newman, and D. J. Smith, Appl. Phy. Lett, 84, 3918 ( 2 0 0 4 ) ) .......................
xv
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
45
Figure
Page
23. High
resolution
electron
micrographs
of
twin
B a {C d ii^ T a 2 /^) 0 ^ viewed along the [110] zone axis:
boundaries in
(a) boundary on
(001) plane; (b) boundary on (110) plane, as indicated by the white arrows.
Localside steps are also observed at boundaries. The superstructure unit
meshes are indicated by white rectangles.
Source (J. Sun, S. J Liu, N.
Newman, and D. J. Smith, Appl. Phy. Lett, 84, 3918 ( 2 0 0 4 ) ) .......................
24. High
resolution
electron
micrographs
of
antiphase
B a (C d i/^ T a 2 / 3 ) 0 3 viewed along the [110] zone axis:
46
boundaries in
(a) boundary in­
clined to (111) plane; (b) boundary parallel to (111) plane. The projected
displacement vectors of the type [001] for antiphase boundaries are indicated
by small white arrows. Source (J. Sun, S. J Liu, N. Newman, and D. J.
Smith, Appl. Phy. Lett, 84, 3918 (2004))
25.
47
Dependence of Q x f on the sintering tem perature of B a (C d 1 / 3 T a 2 / 3 )Os sam­
ples synthesized w ith 2wt%ZnO sintering agent....................................................
26.
Dependence
of
B a ( C d i/ 3 T a 2 /z ) 0
27.
dielectric
3
constant
on
sintering
tem perature
49
for
samples synthesized with 2wt% ZnO sintering agent. . .
50
Ball and stick model of (a) jBa(Zn1/ 3T a 2/ 3)03 and (b) H a (C d i/3T a 2/ 3)0 3 .
Solid black balls are Z n in (a) and Cd in (b). The distortion relative to the
bond-centered configuration has been amplified by a factor of five to more
clearly show the distortion, in particular the buckling of the T a — O — Cd
bond. Note: Done by Dr.M ark V. S c h ilfg a a rd e .................................................
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
52
Page
Figure
28.
Energy as a function of the generalized coordinate Q. Q parameterizes the
collective displacement of the O atoms; Z = 0 corresponds to the highsymmetry position of the oxygen atom between the Cd and T a atoms; Q =
1 to the minimum energy configuration given in Table 2. Note: Done by
D r.M ark V. Schilfgaarde..............................................................................................
29.
Electronic
band
B a iZ n x j^ T a ^ i^ O z
structure
of
(a)
B a (C d i/^ T a 2 / 3 )Os
and
as calculated by the Linear Muffin Tin
54
(b)
O rbital
m ethod w ithin the Local Density Functional approximation. Note: Done by
Dr.M ark V. Schilfgaarde..............................................................................................
30.
X-ray
diffraction
B a ( C d i/ 3 Ta,2 / 3 ) 0 3
p attern
of
Ba(C'd1/ 3T a 2/3)03
powder
and
ceramics with 0.05 wt% B 2 O 3 as sintering aid an­
nealed at 1050°C and varying tim es.........................................................................
31.
X-ray
diffraction
B a (C d 1 / 3 T a 2 / 3 ) 0 3
55
p attern
of
B a (C d 1 / 3 Ta,2 / 3 ) 0 3
powder
62
and
ceramics with 0.5 wt% B 2 O 3 as sintering aid an­
nealed at 1250°C for varying times.
Note th a t the powder sample is
indexed w ith the random-alloy pseudocubic structure and the ceramic
samples are indexed with the ordered hexagonal structure. Asterisks denote
superstructure peaks characteristic of B-site ordering.........................................
32.
62
Dependence of the lattice constants of B a(C d 1/ 3T a 2/ 3)03 containing a
0.05wt% boron sintering aid on the annealing time at 1050°C..........................
xvii
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
63
Page
Figure
33.
Sintering density versus boron concentration for B a {C d i/^ T a 2 /z)Oz sin­
tered at 1350°C for 2 h and (b) sintering density versus annealing time for
B a (C d i/z T a 2 / 3 )Oz with O.liut% boron oxide as sintering aid annealed at
1250°C..............................................................................................................................
34.
SEM images of B a {C d i/zT a 2 /z)Oz specimen prepared with 0.5wt% boron
oxide sintered at 1350°C and annealed for (a) 40 h and (b) 120 h ...................
35.
65
66
Electron diffraction p attern for the boron-doped B a (C d i/ 3 T a 2 / 3 )Oz ceramics
along the < l l 0 > direction. Source: 1) J. Sun, S. Liu, N. Newman, M.R.
McCartney, and D.J. Smith, J. M ater. Res. 19, 1387 (2004); 2) S. Liu, J.
Sun, R. Taylor, D. J. Smith, and N. Newman, J. M ater. Res., 19, 3526 (2004). 68
36.
(a) Annular dark-field image of triple junction and (b) electron energy loss
spectroscopy elementary profile along the line indicated in (a) showing pres­
ence of boron at triple junction. Source: 1) J. Sun, S. Liu, N. Newman, M.R.
McCartney, and D.J. Smith, J. M ater. Res. 19, 1387 (2004) 2) S. Liu, J.
Sun, R. Taylor, D. J. Smith, and N. Newman, J. M ater. Res., 19, 3526 (2004). 69
37.
(a) High-resolution electron micrograph of grain boundary showing absence of
amorphous phase and (b) lattice image from region in triple junction showing
presence of amorphous phase. Source: 1)J. Sun, S. Liu, N. Newman, M.R.
McCartney, and D.J. Smith, J. M ater. Res. 19, 1387 (2004); 2) S. Liu, J.
Sun, R. Taylor, D. J. Smith, and N. Newman, J. M ater. Res., 19, 3526 (2004). 70
38.
Lattice image for B a (C d i/ 3 T a 2 ^3 ) 0 3 ceramics with boron additive sintered
at 1350°C Source (J. Sun, S. Liu, N. Newman, M.R. M cCartney, and D.J.
Smith, J. M ater. Res. 19, 1387 (2 0 0 4 ))..................................................................
xviii
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
72
Figure
Page
39. Effect of the annealing duration at 1050°C on the microwave properties. . .
40. Effect of annealing duration at 1250°C on the microwave properties.....
41.
75
76
W hen A ’ —> A and B ’ —> B, the rectangles become (a) and (b) cylinders and
(c) a Mobius strip. The length of either of the two edges (A-A’ or B-B’ )
of the cylinder formed in (a) is equal to the length of the single edge (A-B’
-B-A’ ) of the Mobius strip formed in (c). As representations of transmission
lines, the two orthogonal modes (solid and dashed lines) are (a) at resonance
for a ring-type resonator, (b) at antiresonance for a half-length ring, and (c)
at resonance for a half-length ring with a 180° twist (Mobius resonator), from
Ref. [85]...................................................................................................................
81
42.
Several key processes for ceramic injection mold technique......................
83
43.
(top) Perspective drawing of the dielectric-loaded Mobius wire resonator
showing the Macor puck and the embedded gold wire with the cross over
accomplished w ith two vias. (bottom) Edge view of a two-pole filter con­
sisting of a dual-mode dielectric-loaded Mobius wire resonator in a copper
c a v i ty .............................................................................................................................
44.
Weakly coupled response of the dielectric-loaded Mobius wire resonator
shown in Fig. 43............................................................................................................
45.
X-Ray diffraction pattern of ceramic injection molded B a ( Z n i / 3 T a 2 / 3 ) 0
89
3
doping Zr and Ni sintered at: (a) 1580°C and (b)1680°C..................................
46.
88
89
Dependence of relative density on the sintering tem perature for ceramic injec­
tion molded Z r and AT-doped B a ( Z n ij 3 Ta<2 / 3 ) 0 3. The theoretical density
is 7.94g/cm - 3 .................................................................................................................
xix
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
90
Page
Figure
47.
SEM photom icrograph of a (a) ceramic injection molded Z r and N idoped B a ( Z n i/ ^ T a 2 /s ) 0 3
sintered at 1520°C; (b) Z r and Ni-doped
B a ( Z n i / 3 T a 2 / 3 )Oz prepared using conventional powder processing meth­
ods sintered at 1520°C; (c) ceramic injection molded Z r and iVz-doped
B a ( Z n i/ z T a 2 /z ) 0 3 sintered at 1680°C....................................................................
48.
Dependence of the Q x f product on the sintering tem perature for ceramic
injection molded Z r and IVz-doped B a ( Z n 1 /z T a 2 /z)O z......................................
49.
92
94
Dependence of the the tem perature coefficient of resonant frequency on
the sintering tem perature for ceramic injection molded Z r and jVi-doped
B a { Z n ijz T a 2 /z)O z........................................................................................................
50.
94
Black and white photograph of a Mobius resonator synthesized from Z r and
IVi-doped B a,{Zn\/zTa 2 /z)Oz using ceramic injection molding. The sample
itself has a light yellow color......................................................................................
51.
97
Schematic drawing of the dielectric-loaded Mobius wire resonator illustrating
the geometry of the dual mode Mobius resonator.................................................
97
52.
X-Ray Diffraction of (a) F-doped BaT iO z and (b) F -doped Bao.rSro.zTiOz■ 104
53.
X-Ray
Diffraction
of
(a)
S'c-doped
B aT iO z
and
(b)
F-doped
Bao. 7 SrQ.zTiQS8 Sco. 0 2 Oz................................................................................................
54.
SEM
pictures
(a) BaTiOz]
for BTO
doped
with
different
F2O5 concentrations:
(b) 0.1wt% F2 0 s-doped BaTiOz](c) 0.5wt% FiOs-doped
BaTiOz](d) lw t% V^Os-doped B a T iO z ....................................................................
55.
105
107
Dependence of B a T iO z (a) sintering density;(b) dielectric constant on V2 O 5
concentrations...................................................................................................................
xx
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
108
56.
Dependence of dielectric constant on the tem perature
(a)
F-doped
BagjSro.zTiO z and (b) S'c-doped B a ojS ro .zT iO z...............................................
57.
Dependence of dielectric constant on the applied dc biasing for undoped and
F -doped BaojSro.zTiO z sample...............................................................................
58.
110
Ill
I-V characteristics for undoped and F-doped BaojSro.zTiO z sample at room
tem perature....................................................................................................................
112
59.
I-V characteristics for Sc-doped B ao^Sro^T iO z sample at room tem perature. 112
60.
Tem perature dependence of I-V curves for 0.5 at% Sc-doped B a 0 .7 S r 0 .zTi.Oz- 115
61.
Typical log (V) log(I) curve for 0.5 at% Sc-doped BaojSro.zTiO z sample
measured at 140°C showing Ohmic and space- limited-charge current.
62.
. . .
Dependence of dielectric constant on the frequency at room tem perature (a)
F-doped B a o jS ro .zT iO z; (b) Sc-doped B a ojS ro .zT iO z.....................................
63.
-
119
XRD for non-irradiated and neutron irradiated B aojSro.zTiO z sample. Neu­
tron irradiation shows the same patterns as th a t of non-irradiated.................
65.
118
Dependence of dielectric constant on the frequency at room tem perature(a)
Sc doped B a 0 .7 S r 0 .zTi 0 .g8 V0 .0 2 Oz; (b) F-doped B a 0 .7 S r 0 .zTi 0 .Q8 Sc 0 .0 2 Oz.
64.
116
Magnetic
susceptibility
for
non-irradiated
and
neutron
123
irradiated
Bao. 7 Sro.zTiOz samples. The noise for sample irradiated by 4.24xl013/c m 2
result from environmental artifact in the measuring room during experiment. 124
66.
Logarithmic plots of the leakage current as a function of applied field strength
for the undam aged and damaged sample with neutron dose 2.12 x 1013/c m 2. 125
67.
I-V characteristics for non-damaged and neutron irradiation Bao. 7 Sro.zTiOz
with dose 2.12 x 1013/c m 2, 4.24 x 1013/c m 2, and 6.36 x 1013/c m 2.................
xxi
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
126
Figure
68.
Page
I-V curves for unannealed, annealed B aT iO z in pure 0
2
and 5vol%H2 +
95vol%Ar forming gas at 500°C' respectively. Samples are measured at 100°C'.129
69.
Dependence of leakage current of annealed B a T iO z in oxidizing gas at 500°C
on the annealing tim e...................................................................................................
70.
129
Dependence of leakage current of annealed BaT iO z in reducing atmosphere
on the annealing tim e...................................................................................................
130
71.
Typical log (V)-log (I) curves for B aT iO z annealed at 500°C in O 2 for 48 hrs.131
72.
Typical Arrhenius curve at low voltage range for the calculation of the acti­
vation energy of annealed B aT iO z in O 2 for 48 hrs at 500° C ...........................
73.
Typical Arrhenius curve at low voltage range for the calculation of the acti­
vation energy of annealed B aT iO z in #2 for 12 hrs at 500°C...........................
74.
132
132
Activation Energy of BTO annealed in H 2 and O 2 Measured at low dc voltage
(51V)................................................................................................................................
133
75.
M agnetic susceptibility for samples annealed in varying annealing atmosphere. 134
76.
Calculated concentration of paramagnetic point defects from Fig.75..............
77.
E P R spectra of unannealed B aTiO z and B a T iO z annealed in oxidizing and
reducing gas for 48 hrs at 500°(7. Spectra are obtained at 120K.....................
xxii
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
136
136
CH A PTER 1
INTRODUCTION
1.1. R esearch background for m icrow ave ceram ics
M iniaturization of satellite communication and cellular systems requires low-loss
tem perature-com pensated microwave ceramics with enhanced dielectric constants[1, 2,
3].
Since the 1970’s, there have been significant advances in the performance of
several microwave dielectric ceramics, including, B a 2 T i g 0 2, B a T i 4 Og, Z r \ - xS n xT i 0 4,
B a ( Z n i/ 3 T a 2 / s ) 0 3 and BaQ-^xRE%^ 2 x T i\ 8 0 ^ [4]. The performance of dielectric filters
has significantly improved as a result of these advances, as well as from improvements in
device and system design [5]. B a ( B y 3 Ta-2 /g)Og based perovskite compounds, where B ’ =
Mg, Zn, show trem endous potential for widespread use in microwave systems owing to their
excellent high frequency properties. B a (Z n i/g T 0.2/ 3)0 3 , for example, has a large dielectric
constant (~30) and u ltra low loss tangent (tanS < 2 x ICE5) at 2 GHz. Furthermore, when
doped w ith Ni, its tem perature coefficient of resonant frequency,
t j
,
can be tuned to near
zero [6]. Zr is also commonly added since it has been found th a t high quality factors (Q)
can be obtained in much shorter annealing times [7]. As yet we do not have a fundamental
understanding of why this class of materials can have both a high dielectric constant and
low loss, although a number of experimental and theoretical investigations have proposed
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
2
th a t the perm ittivity is dom inated by polar phonon mode contributions described by a clas­
sical oscillator mode and the predominant microwave loss mechanism in practical dielectrics
arises from the low-energy tail of anharmonic lattice vibrations [8, 9, 10].
A basic understanding of the mechanisms of microwave loss in practical materials has
eluded researchers to date. Nevertheless a number of material properties have shown to be
strongly correlated w ith loss. For example, an early paper reported an inverse relationship
between Zn-Ta site ordering in B a ( Z n i / 3 T a 2 jz)Oz (typically referred to as B-site ordering
in A ( B i / ^ B y 3 )Os) perovskites and the loss tangent at microwave frequencies [11, 12]. It
was reported th a t high-quality factors and B-site ordering could be attained through high
tem perature annealing at 1350°C for extended time (120 hours). Microwave loss in other
materials, including Z r T i O 4 doping with Sn, has also been correlated w ith cation ordering
[113]. Later reports found th a t the addition of B a Z rO z to B a { Z n i/z T a 2 /z)Oz resulted in
low loss even w ithout extended annealing times and significant B-site ordering [7]. A num­
ber of reports attrib u ted this to various factors, including specific atomic configurations at
grain boundaries [14]. The am ount of impurities in the material has also been correlated to
the microwave loss [15]. A recent report found a direct correlation between microwave loss
and the concentration of point defects, as quantified by optical spectroscopy [2]. Results
show th a t the point defects could play an im portant role in the dielectric loss of the mi­
crowave ceramics. However, it is not clear how the point defects are related to the dielectric
properties. Clearly the role of intrinsic and extrinsic factors in the microwave loss process
in these materials is ju st beginning to be uncovered.
On the other side, a growing number of high-frequency passive devices require intri­
cately shaped microwave m aterials [16]. This has placed increasing dem and on the ceramic
research community to develop a low-cost process th a t can manufacture high performance
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
3
microwave m aterials in complex-shapes. The use of dielectric ceramics w ith low loss, high
dielectric constants, and near-zero tem perature coefficient of resonant frequency is also es­
sential in achieving the required microwave performance for the devices of interest [1, 4],
Several traditional ceramics have been successfully m anufactured in high volume and with
great dimensional precision using ceramic injection molding, including silicon nitride [17],
zirconia [18, 19] and alum ina [20, 21]. Most research to-date has focused on investigating
the mechanical properties of ceramics parts, while little attention has been paid to the elec­
tronic and dielectric properties. Thus, for microwave devices, it is im portant th a t ceramic
injection molded ceramics have a sufficiently high dielectric constant (tan<5), low loss tan­
gent (Q = 1/tan<5) and a near zero tem perature coefficient of resonant frequency (ry) to be
useful for practical applications.
1.1.1. M a te r ia ls re q u ire m e n ts . Useful microwave dielectrics should have high
dielectric constant, low dielectric loss, and near-zero tem perature coefficient of resonant
frequency. er , tan<5 or the quality factor (Q = l/tan<5), and r / are these three im portant
dielectric properties for the microwave dielectrics. These properties are discussed in the
following subsections.
1.1.1.1. Dielectric constant. The relationship between /o (resonant frequency) and
er is given by the equation:
/ o oc ^
Er
(L 1)
Increasing er can reduce the size of ceramics without altering the resonant frequency. Thus,
a high dielectric constant is an im portant param eter when size is a factor in its actual
application.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
1.1.1.2.
Dielectric loss. A low dielectric loss (tand) is m andatory for high perfor­
mance microwave devices because the low dielectric loss enables the development of high-Q
devices with the low noise characteristics. The Q of a resonator in the absence of other
losses is l/tam fi In general, dielectrics materials with enhanced dielectric constant are found
to usually have higher dielectric losses. Among all of the three properties, tan<5 is the most
sensitive to m aterial preparation and processing conditions. Because of the dependence of
dielectric loss on frequency, results are often quoted in term s of Q x f. To a first approxima­
tion, Q x f is a constant in the microwave region.
As mentioned above, there has been dram atic progress in m aterial development of
ceramic dielectrics, but the responsible basic physics for the dielectric properties has not
been investigated in detail. Thus, it is necessary to b etter understand the physical origin of
these m aterials’ properties first, specifically of the dielectric loss. The dielectric losses can
be divided into two classes: intrinsic and extrinsic.
• Intrinsic losses: Intrinsic losses represent the lower limit of loss th a t would be found
in a perfect single crystal and are believed to result from the anharmonic lattice
forces th a t control the phonon-phonon interactions within a crystal. Since microwave
frequencies are much smaller than the transverse optical modes, the microwaves do
not directly interact with individual phonons. Losses are created by three phonon
processes. The dominant process is the decay of one transverse optical phonon into
two acoustic (thermal) phonons. The theory of intrinsic losses has been investigated
in a number of papers [9, 11]. From classical dispersion theory, the dielectric loss is
given by:
( 1 .2 )
where 7j are the dam pening constants and A er are the dielectric strengths of the j t h
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
5
transverse optical mode. Since ceramic dielectrics are far from perfect single crystals,
their dielectric loss is not expected to be dominated by intrinsic loss.
• Extrinsic Losses: Extrinsic losses are additional losses resulting from imperfections in
crystal structure. These imperfections include raw m aterials impurities, grain bound­
aries, and im purities from the raw materials and from sintering additives th a t are
used to achieve high density, and charge carriers and structural defects during oxida­
tion/reducing annealing processing. These can be divided into two groups:
1. Losses due to point defects such as dopant/im purity atoms, vacancies, or defect
pairs th a t lead to quasi-bonded states.
2. Losses due to extended dislocations, grain boundaries, pores, inclusions, and
secondary phases. W ith the second type of defect, either dipole relaxations of
im purities concentrated at interfaces or relaxations of space-charge polarizations
present at interfaces are found. For both effects, the dielectric loss tangent is
[22 ]:
tan 5 <x
UTs 9
1+
(1.3)
where the relaxation time is
n oc
(L4)
and E q is the activation energy of the process. The first type of defect leads
to dielectric loss due to the dampening of the optical lattice vibrations through
scattering. These losses can be considered as modifying the dampening con­
stan t t j of the transverse optical modes. The degree th a t the defect affects the
dam pening constants will depend on the type of defect and the structure of the
material.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
6
1.1.1.3.
Temperature coefficient of resonant frequency. For a microwave device to
be used practically, its resonant frequency ( /0) must be extremely stable in the changing
environment tem perature. The tem perature coefficient of frequency (TCF) is defined as:
(1.5)
and should be very close to zero for most applications. The shift in
with tem perature re­
sults from therm al expansion dimensions and tem perature dependence of er . Thus the TCF
can be related to the m aterial’s linear therm al expansion coefficient (cq) and tem perature
coefficient of er (T C £r) by
—T C
T C F = — —^ - ax
z
(1.6)
Adjusting the T C Sr so th a t the TC F is zero (no frequency drift) is probably the most
difficult aspect in the development of a microwave dielectric ceramic. Research has shown a
distinct correlation between the m agnitude of T C £r and the structure of the oxygen network
(tilted/untilted) in complex perovskites [33]. Further investigation [24] shows th a t tilting of
the hexagonal structure is the most likely explanation of the controllability of the TC F or
T C e in Ca and S r titanate-based compounds if dilution effects, associated w ith the addition
of low-dielectric-constant perovskites, such as NdAIOz and LaAlO 3, are taken into account.
1.1.2.
M a teria ls. A wide range of dielectric materials have been investigated. Ta­
ble 1 summarizes the dielectric properties of some classical microwave ceramics. We will
describe the advantages and limitations of some of the most im portant microwave materials.
1.1.2.1.
Alumina. Alumina (AZ2O3) is not a practical m aterial for most applications
due to its low dielectric constant, er = 10, and high T C F ~ —60 p pm /K , but it is still
interesting for several reasons. First, sapphire (single crystal AI 2 O 3 ) has the lowest known
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
7
M aterial
Alumina(Al203)
Ba{M gii^Ta2/z)Oz
(BMT)
B a ( Z n 1 / 3 Ta2iz)03
(BZT)
(Zno.8Sno.2)Ti04
(ZTS)
T itania (T ^ )
BaTi^Og
Sf
10
24
Q
420009 GHz
2000010 GHz
Q x f(T H z)
380
200
T C F ( p p m /K )
-60
29
98007 GHz
69
« 0
37
70007 GHz
49
« 0
100
38
51
58.5
450
14
Ba2Tig02o
39
17000 3 GHz
130004.5
GHz
120004.5
GHz
35003 GHz
20003 GHz
54
4
10.5
6
93
» 0
120004 GHz
48
± 20
81
B aN d2 T i3 0 u
BaO - PbO 88
Nd,203
NdzOz — AZ2O3 — « 45
CaO —T i 2
Table 1. Summary of the dielectric properties of ceramics dielectrics, source [22]
loss of any microwave materials. Alumina of a very high purity (99.999%) is available.
T hat makes it a very good m aterial to study the fundamental physics of dielectric loss.
Also because of its very intrinsic loss, AI 2 O 3 is very sensitive to the materials preparation
and processing. W ith high purity powder, Alumina produced using high purity powder has
been found to have a Q of 42,000 at 9 GHz [25]. This makes it an excellent candidate for
low noise oscillators. There are several factors th a t can affect the dielectric properties of
alumina, such as, purity, porosity, and grain size [22],
1.1.2.2.
[1].
Titania. TiO<i is very cheap and has a er = 104 and Q — 14,000 at 3GHz
Unfortunately, near room tem perature, T iO 2 has a T C F of +427 ppm/TT-1 [26].
The properties of titan ia are also sensitive to oxygen stoichiometry. For example, oxygendeficient T i 0
2
is a semiconductor and thus has conduction loss. However, titan ia is still
of interest because of the combination of high dielectric constant and high Q not found in
other materials, and it can still be used in dielectric resonators. On the other side, recent
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
800
700
800
X 500
3 400
Uw
P 300
200
100
0
50
150
200
250
3Q0
Temperature (K)
Figure 1. The tem perature dependence of the T C F of polycrystalline TiOz- Source [22]
research found th a t near 25K, the TC F of T iO i is near zero, see Fig. 1. Thus, with the
recent developments into cryogenic technology, there has been renewed interest in TiO^An oxygen deficiency with the accompanying T i 3+ ions is found to significantly
reduce the Q in TiO<i- Thus, doping TiO^ with an im purity having a valence of 2+ or
3+ ions might compensate for the oxygen loss and could potentially produce a high Q and
high density T iO 2 - As an example, Fig. 2 shows the effect of doping w ith y ttria stabilized
zirconia on the dielectric loss of TiO i- Doping TiOg w ith pure zirconia with a valence of
4+ does not improve the Q as it is the y ttria th a t provides the needed 3+ ions.
1.1.2.3.
Barium titanate and its derivatives. There are two phases of barium titanate
th a t are useful for microwave applications, BaTi^Og and BagTigOgg- These components
are referred to as B T 4 and BgTg respectively. The properties of these m aterials are very
sensitive to processing param eter. The typical properties of these two materials are listed as
followings: BaTi^Og'. er = 38, TC F = 14 ppm /K , and Q — 15,000 at 4.5 GHz; B a 2 Tig 0 2 oer = 49, T C F = 4 ppm /K , and Q = 12,000 at 4.5 GHz.
The low Q has been attributed to two main causes: the presence of lossy impurity
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
9
7E-4
6E-4
Undoped
5E-4
26-4
Doped with 0.25 weight % YSZ
1E-4
0
50
100
150
200
250
300
350
Temperature (K)
Figure 2. A comparison of the tanS versus tem perature at 3 GHz of YSZ-doped and undoped
fully dense TiO i- Source [22]
phases and of T i 3+ as a result of the reduction of T i 4+. Im purity phases are introduced
from low purity raw m aterials (BaC O s and T iO 2‘ ), the use of binders, and the use of
alumina grinding media. Similar to T iO 2, the reduction of T i A+ to T i 3+ is mainly due to
the presence of structural defects th a t are compensated by the formation of T i 3+ or due to
incomplete reoxidation during the cool down after the sintering process.
The main disadvantage of these m aterials is the TCF. Many applications require
TC F between —4 and 2 ppm /K . Composites of HT4/H 2T9 can produce high Q materials
with a T C F between 4 and 14 ppm /K . Reducing the T C F requires the addition of dopants.
Substituting S n 0 2 for TiO<i in R 2T9 or B T 4 / B 2 TQ composites can reduce the TC F to
acceptable levels, but this also causes a reduction in Q.
1.1.2.4.
( Z r i - x ) S n xT iO i. Although Zirconium titanate-based ceramics have been
used as dielectrics in capacitors for many years, Z r xT i yS n z 0 4 ceramics were not incorpo­
rated into microwave filters until the late 1970s. Compositions in the range 0.15 < Z < 0.3
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
10
2E-4
1E-4
s
8E-5
8E-5
2E-5
0E+0
0
50
100
150
200
250
300
Temperature (K)
Figure 3. The tem perature dependence of the tand of ZTS at 7.8 GHz. Source: [22]
and (1 — z/2) < y < 1.05 were found to be of particular interest for dielectric resonator
application because they have a high Q, a low tem perature coefficient, and a high dielectric
constant (~ 35). One of the advantages of zirconium titan ate stannate (ZTS) is th a t its
TC F can be controlled by varying the Sn content w ithout drastically affecting the other
properties. This is im portant for applications because a TC F of precisely zero is not always
required. In fact, it can be used to compensate amplifier and other microwave devices th a t
have a non-zero TC F. In general increasing the TiO^ content at the expense of Z rO i or
S n O i increases Q and er , whereas increasing S n O -2 content at the expense of Z r O i in­
creases Q with little effect on er . Z r^^Srio^T iO i th a t has received particular attention has
a TC F = 0 ppm /K , er = 38, and Q — 7000 at 7 GHz [1]. However, ZTS is very difficult to
sinter to high density by solid-state diffusion. M aterials with high porosity have reduced Q.
The densification of ZTS can be enhanced by the use of additives through the formation of
liquid phases at grain boundaries during sintering, although secondary phases th a t might
result can degrade the dielectric properties. A typical additive m ixture of 1 wt% Z n O is
used. 0.5 wt% Nb^O^ is also sometimes added to this formulation. High density (>96%)
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
11
Zro&SriQzTiOii can be produced using a m etal alkoxide route [28]. This m ethod produced a
fine (0.3pm) monosized amorphous ZTS powder th a t was pressed and sintered without any
precalcination. These materials were found to have good microwave properties. Materials
sintered at 1650°C for 3 h had er = 40, Q = 5000 and TC F = 3 p p m /K at 10 GHz. The
Q was further increased to 5300 at 10 GHz by annealing the samples in O2 at 1450°C for
15 h. Tem perature dependence of the tanS of undoped Zro^Sno/zTiO^ is shown in Fig. 3.
The tanS decrease roughly linearly down to the lowest tem perature measured (15K).
1.1.2.5. B a ( B i / 3 T a 2 /s ) 0
3
, B = Z n ,M g . Several oxides based on the perovskite
structure with Q factors exceeding 10,000 at 10 GHz were developed for high frequency
microwave applications by careful control of their chemistry, processing m ethods and de­
gree of structural order [29] - [32]. Especially, these complex perovskites with a general
formula A (B [/ 3 B%/ 3 ) 0 3, where A = B a 2+, B ' = M g 2+, Z n 2+ or N i 2+ and B " = T a 5+ or
N b 5+, show very interesting properties at microwave frequencies. These ceramics have high
permittivity, low dielectric loss and almost near zero tem perature coefficient of resonance
frequency. Thus, these m aterials are commonly used as resonators in practical microwave
devices. Among these compound, B a ( B y 3 T a 2 / 3 ) 0 3, where B ’ = Zn or Mg exhibit the
highest Q values. It is well known th a t many perovskites with a 1 : 2 m ixture of diva­
lent and pentavalent ions on the B ’-site adopt a structure in which the two cations order
onto individual (111) crystallographic plane (Fig.4) [14]. The long range ordering, which
is often accompanied by a small lattice distortion, yields a hexagonal structure with a
,.Bi2+ — T a 5+ —T a 5+.. repeat sequence along the < 111 > direction of the parent cubic
perovskite cell. Studies of the structures and properties of the tan talate systems show that
the degree of ordering between the B ,2+ and T a5+ significantly affect on the dielectric loss
at microwave frequencies. By inducing long-range cation order through extended high tem ­
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
12
perature annealing (> 1400°C), the Q values of B a fB y ^ T a ^ /Q O z, B = Z n , M g ceramics
can be increased from ~ 500 to > 35000 at 10 GHz. Several investigation have clearly
dem onstrated th a t the changes in dielectric loss and tem perature coefficient of resonant
frequency arise from alterations in the degree of intrinsic cation order [2, 14, 33]. However,
the expensive raw material, To^Os, presents materials scientists w ith a pressing challenge
to understand the role of tantalum and replace it with a less costly element [3].
• B a ( M g 1 / 3 T a 2 / 3 )O s: B a { M g ij 3 T a 2 / 3 ) 0 3 sintered at 1650°C for 2 h possesses a high
dielectric constant (er ~ 23), a high quality factor ( Q x f value ~ 86000 GHz) and
a small tem perature coefficient of resonant frequency ( t j ~ 7.5p p m /°C ) [104]. Bet­
ter microwave dielectric properties (er ~ 23.7) and Q x f value ~ 124000 GHz) can be
achieved by extending the sintering time to 10 h [35]. It is reported th a t its microwave
dielectric properties (er ~ 25.7, Q x f value ~ 176000 GHz, r / ~ 2.7p p m /°C ) could be
improved with 1 mol% Mn addition [36]. However, B a (M g x/ 3 T a 2 / 3 ) 0 3 is not easily
sintered to high density because of the slow atomic diffusion rate during solid state sin­
tering. Fig. [?] shows the tem perature dependence of tanS for the B a ( M g ij 3 T a 2 / 3 ) 0
3
at 8.3 GHZ. The tanS decreases with the decreasing tem perature.
• B a ( Z n i / 3 T a 2 / 3 ) 0 3-. In the search for tem perature stable, high Q, higher dielectric
constant, many m aterials have been studied. B a ( Z n i/ 3 T a 2 / 3 ) 0 3 (BZT) ceramics a t­
trac t more attention and interests because this ceramics have a er = 30, Q = 84,000
at 2 GHz, and a T C F of ~ 0.1 —0.5 ppm /K [30]. Samples sintered above 1450°C were
found to have a high Q. M aterials sintered above 1600°C were found to have a lower
density, as small as 80% of theoretical value, but still m aintain a high Q. The effect of
grain size on the Q was found to be small. The Q was found to be very dependent on
sintering time. Even though the density and grain size changed little, samples were
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
13
Q
rB‘ /B*
O
(a)
#«•
O
Figure 4. Two crystal structures for Bci(B 1 / 3 T (1 2 / 3 )O3 complex perovskites, where B
M g ,o r Z n . (a) disordered ; (b) ordered, source: [26]
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
14
5.0E-5
4.5E-5 • ■
4.GE-5 ■■
3.5E-5 - •
G
2.SE-5
2.0E-5 . -
1.5E-5-1.QE-5-5.0E-6 - O.GE+O
0
50
100
150
200
250
300
350
Temperature (K)
Figure 5. Tem perature dependence of tanS for the B a ( M g i/ 3 T a 2 / 3 ) 0 3 at 8.3 GHZ. Source
[22 ]
found to exhibit high Q when sintered at 1350°C for greater th an 32 h (see Fig. 6).
X-ray diffraction shows th a t this was accompanied by ordering of the Z n and Ta.
Initially during processing a disordered cubic perovskite is formed with the Zn and
Ta ions on random sites. At tem peratures greater than 1350°C, the structure orders
into a hexagonal perovskite with the Zn and Ta showing 1 : 2 order in the B site.
The high Q could arise as a result of ordering reducing the number of degeneracies
of the phonon modes and the reduction of point defects. A nother factor th a t might
affect the Q values is the loss of volatile Z n O during sintering. The loss of Z n O can
distort the lattice through the creation of zinc an d /o r oxygen vacancies. However, the
lattice distortion and ordering do not occur simultaneously as reported in reference
[30]. Thus, it has been suggested th a t the presence of lattice distortion after ordering
is complete can be attrib u ted to the loss of ZnO [67].
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
15
J 5000
10000
5000
8.0
m
o>
UJ
M
1.5 •
"’ c
t.0-
5 1
o
as2
8
32
120
LOG SINTERING TIME ( h)
Figure 6. The quality factor, grain size and density of BZT as a function of sintering
tem perature. Source: [30]
Top Plate
Dielectrics Sample
Insulator Spacer
Bottom Plate
Figure 7. Diagram of the configuration for cavity technique.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
16
Figure 8. Illustration of the magnetic field distribution for the T E q\$ mode. Source [22]
1.1.3.
M e a s u r e m e n t o f d ie le c tric p r o p e r tie s . Resonant techniques are the best
methods for measuring the dielectric properties [37]. There are several advantages of res­
onant techniques. First, the materials being measured have similar geometry to the way
they will be used in a device. Second, because the microwave energy is mainly contained
within the material, they are less susceptible to influence from external losses. Among the
resonant techniques, the Haki-Coleman m ethod [22] or postresonator technique is one of
the most common techniques for measuring dielectric properties in the microwave region.
It consists of a cylinder of dielectric sandwiched between two conducting planes. This is
just a basic shielded dielectric resonator configuration. Microwave energy can be coupled
to the resonator using coupling loops. The Q and resonant frequency can be determined
in transmission using a vector network analyzer. This technique is good for measuring
the dielectric constant and TC F. However, this m ethod has limited accuracy when used to
measure Q (or tand). A b etter technique for measuring Q is to use a cavity technique with
dielectric suspended above the bottom plate on a low loss spacer, as shown in Fig. 8. The
spacer, such as, teflon, m ust have a low dielectric constant and low dielectric loss so th a t
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
17
it does not interfere with the measurement. The Q is usually measured from the T E m$
mode. This mode is used because it is a fundamental mode and the field is symmetric
with the geometry of the resonator and the cavity, resulting in a reduction of radiation at
near-field losses. The field p attern of this mode is shown in Fig. 7. In this case, any current
generated w ithin the shielding cavity does not have to cross from one surface to another
which would increase losses. The cavity technique is more accurate at measuring Q than
the Haki-Coleman technique because the contribution from the end plates is lower and the
(cylindrical) cavity wall reduces radiation losses. Usually the losses from the (cylindrical)
cavity wall can be minimal if the diameter of the cavity is between two and three times the
diameter of the dielectric. A diam eter greater th an three times th a t of the dielectric can
make it difficult to couple to the dielectric resonator modes. The contribution of the plates
to loss, 1/ Q c, cannot be determined analytically, but requires numerical techniques. Several
techniques have been developed for calculating the modes. Kajfez and Guillon [55], who
also provide software code for calculating the frequency of E modes, have detailed some of
these.
1.2. R esearch background for ferroelectrics
Ferroelectric ceramics have been used in a wide variety of applications and constitute
the materials base for a very large number of commercial components and devices used
in electronics. The most experimentally and theoretically studied system is barium and
strontium titan ate because it is the prototype ferroelectric as well as an im portant material
in the capacitor industry. In this section, simple concepts and factors affecting the properties
of barium titan ate were introduced. A more complete description for the structure and
dielectric properties of this material can be found in Ref. [22, 42, 43]. As shown in Fig. 9,
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
18
• Ti*+
(a)
(b)
T>TC
T<TC
#
Ba2+ O
0*
Tc : Curie temperature
Figure 9. Crystal structure of B a T iO z. Source: [43]
BaTiOz has a perovskite crystal structure. In the high-tem perature paraelectric phase there
is no spontaneous polarization and it is non-polar phase. Below the transition tem perature
Tc, called the Curie tem perature (about 120°C), spontaneous polarization occurs, and the
crystal structure becomes slightly elongated, th a t is tetragonal. The dielectric characteristic
of barium titan ate ceramics w ith respect to tem perature, electric strength, frequency and
time (ageing) are very dependent on the substitution of minor am ounts of other ions for
Ba or Ti. These factors affecting the properties of some commonly used ferroelectrics are
discussed below.
1.2.1. B ariu m tita n a te (B aT iO z)•
1.2.1.1. A O /B O 2 ration. The A O /B O 2 ratio is the ratio of the to tal number of ions
on Ba sites to the number on T i sites. The partial phase diagram for the BaO —T iO 2
system (Fig. 10) shows th a t there is only very slight solubility for excesses of either BaO or
T 1 O 2 in B aT iO z■ Excess T i02 (A O /B O 2 < 1) results in the formation of a secondary phase
of Ba§TinOio, and this forms a eutectic with BaTiOz th a t melts at about 1320°C. This
facilitates the liquid phase sintering mechanism to occur at tem peratures above 1320°C. An
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Figure 10. Phase diagram of the BaO —TiO<i system (> 34mol.%Ti02). Source: [42]
excess of BaO results in the formation of Ba^^TiO^ which forms a eutectic with BaTiOz that
melts at about 1563°C. As is often the case with solid insoluble phases, B a z T iO 4 inhibits
the grain growth of BaTiOz sintered at tem peratures up to 1450°C, giving rise to grain sizes
in the 1 —5/am range. Excess BaO also lowers the cubic-hexagonal transition from 1570°C
to about 1470°C in pure B a T iO z■ Hexagonal m aterial seldom occurs in sintered ceramics of
technical purity because many common substituent, such as strontium for barium, stabilize
the cubic form.
1.2.1.2.
Substituent. Uniformly distributed isovalent substituents do not greatly af­
fect the shape of the er —T curve and other characteristics. Their main effect is to alter
the Curie point and the lower tem perature of BaTiOz phase transition. Alloying with lead,
strontium and calcium enables the transition tem peratures to be shifted to suit particular
requirements. The Ba, Pb and Sr ions can be alloyed in any proportions to produce a single­
phase perovskite, while the solubility of CaTiOz is limited to about 20mol.%. T iA+ can be
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
20
replaced by isovalent ions with radii between 60 and 75 pm. Zirconium, hafnium and tin
have similar effects on the three transitions, although the solubility of tin may be limited to
about 10 mol.%. They reduce Tc but raise the tem perature of the other two transitions to
such an extent th a t in the range 10 —16mol.% they almost coincide in the neighbourhood of
50°C. Particularly high values of perm ittivity are found for such compositions. Aliovalent
ions are usually limited in their solubility which may depend on the A O / B O 2 ratio. For
example, K + can replace Bo?+ to which it is very similar in radius.
A number of trivalent ions with radii between 110 and 133 pm, e.g. B i and La, can
substitute on the A site. La?+ confers a low resistivity at low concentrations (< 0.5mol.%).
Higher-valency ions on B sites with radii between 58 and 70pm have similar effects to La
on the A site at both high and low concentrations. N b 5+ at the 5 mol.% level has been
found to improve resistance to degradation. In sufficient concentration these higher-charge
substituent both suppress oxygen vacancies and promote the formation of cation vacancies
th a t act as acceptors. The resulting dielectrics have a high resistivity and are resistant to
degradation.
T i4+ can be replaced by a number of trivalent ions with radius in the range 60 - 70
pm (C r , Ga, M n , F e and Co) up to about 2mol.%. They lower Tc.A bout 0.5 mol.%
M 71O 2
is frequently added to all classes of dielectric and results in a reduction in the dissipation
factor. This may be due to its presence as M n A+ in the sintered bodies w ith the possibility
of trapping carriers by the reactions:
M n A++ e ' -> M n 3+
(1.7)
M n 3+ + e' -» M n 2+
(1.8)
and
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
21
M
n and M n^{ also act as acceptors. T i 4+ can also be replaced by about 2mol.% of
divalent ions with radius in the range 60 —70 pm, such as N i 2+ and Z n 2+ , with similar
results to substitution by trivalent ions. Larger divalent ions such as M n 2+ may be soluble
to a lesser extent.
1.2.1.3. Grain size. It was generally recognized th a t as the grain size was reduced to
the micron level, the dielectric constant at room tem perature increased, and th a t the tem­
perature dependence of the dielectric constant was modified significantly below the Curie
tem perature (Tc). It was observed th a t polycrystalline B a T iO z exhibited an enhanced di­
electric response for ceramic specimens prepared with a grain size as small as I p m [44, 45].
The increase in dielectric constant is now understood in term s of the twinning behavior of
polycrystals and the m icrostructure stress with decreasing grain size [46]. Further relation­
ship for grain size reduction to the nano scale becomes more difficult because there exist a
large number of defects at such a small size. Although this idea of a critical size for ferroelectricity has long existed [47, 48], unfortunately, few reports have been p u t forth which
lead to a solid understanding of size-dependent phenomena for polycrystalline BaTiO z at
this scale level.
1.2.1.4. Applied voltage. The magnitude of the applied voltage has a very significant
effect on dielectric properties. The domain width motion under dc biasing is believed to
be the factor influencing the dielectric behavior [115]. As shown in Fig. 11, the dc-biasing
effect usually is stronger in the paraelectric state than in the ferroelectric state, but the
dc-biasing effect on the dielectric constant occurs in both the ferroelectric and paraelectric
states. The change of dielectric constant under dc voltage is characterized by dielectric
tunability D t th a t is defined at a specific working tem perature T and dc-biasing field E as,
D t( T ,E ) = (1 -
j j )
k0
x 100%
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
(1.9)
22
12000
OV/cm -
1250 V/cm
3750 V/cm -
5000 V/cm
2500 V/cm
10000
8000
6000
4000
P
x = 0.5
x = 0.75
2000
x = 0.25
X =
0
x = 1.0
-200
-150
-100
-50
0
50
100
150
Temperature (°C)
Figure 11. Tem perature dependence of dielectric constant as a function of dc-biasing field
for B a xSr\-.xT iO i system. Measuring frequency is 10 kHz. Source: [115]
where k' and k '0 are the real part of the relative perm ittivity (i.e., dielectric constant) under
zero bias field and under bias field E, respectively. According to Devonshire’s theory [115],
in the perovskite cubic structure, titanium ions oscillate in an anharm onic potential of the
form ar2 + fer4, where r is the position of the titanium ion. The Helmholtz free energy
P (P ,T ) of the titanium ion can be expanded in even powers of the polarization P with
coefficients th a t are a function of the tem perature only; th a t is,
F ( P ,T ) = F ( P , T ) + A ( T - 6 ) P 2 + B P 4 + C P e
where
6
(1.10)
is the Curie-Weiss tem perature, and A, B, and C are the expansion coefficients.
In the paraelectric state, the free energy increases as polarization, and there is only one
minimum, at polarization P — 0. If a small field E is considered to apply on the materials,
th a t is, approximately the minimum, the P 6 term can be neglected.
It is appropriate
to neglect the P 6 term in the expansion of free energy because the polarization in the
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
23
paraelectric state is much smaller th an th a t in the ferroelectric state.
The dielectric constant k' can be obtained by taking the second derivative of free
energy (equation 1.10) with respect to polarization:
4tt = 2A(T - 6) + 12BP2
(1.11)
In the case of a small field E, it is assumed th a t k' « Anp can be substituted into the second
term in Eq. (1.11). Thus, a representation of the dielectric constant is obtained as
4tr = 2A(T - d ) +
Hence, the dielectric constant varies with dc-biasing E. Equation
(1.12)
(1.12) reduces to the
Curie-Weiss law when the anharmonic term is negligible. The lattice anharmonic interaction
of titanium ions is responsible for the field dependence of the dielectric constant of the
BST system. A phenomenological equation describing the relationship between the applied
voltage and loss tangant th a t is valid in the paraelectric state is listed as below [115]:
k'
1
k" = (1 + ak>03E 2)V 3
(L13^
where a — 12B /(4 tt)3 is the phenomenological coefficient (or anharmonic coefficient),
which is derived from the anharmonic term in the Helmholtz free energy. By these phe­
nomenological equations the field dependence of the dielectric properties for the paraelectric
B a \ - xS rxTiOz system can explain satisfactorily.
1.2.2.
B re a k d o w n m e c h a n ism s. Dielectric breakdown is defined as the voltage
gradient or electric field sufficient to cause large current to flow. It depends on many factors,
such as sample thickness, tem perature, electrode composition and shape, and porosity. In
ceramics, there are two basic types of breakdown; intrinsic and therm al.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
24
• Intrinsic: In this mechanism, often referred to as avalanche breakdown, electrons in
the conduction band are accelerated to such a point th a t they sta rt to ionize lattice
ions. As more ions are ionized and the number of free electrons increases, an avalanche
effect is created. Clearly, the higher the electric field applied, the faster the electrons
will be accelerated and the more likely this breakdown mechanism will be.
• Therm al breakdown: The criterion for therm al breakdown is th a t the rate of heat
generation in the dielectric, as a result of losses, m ust be greater th an the rate of heat
removal from the sample. Whenever this condition occurs the dielectric will heat up,
which in tu rn will increase its conductivity, which causes further heating, etc. This is
term ed therm al breakdown or therm al runaway.
1.3. M aterials p ro cessin g
This section will discuss the processing of bulk microwave dielectrics. There are four
main processes for the preparation of ceramics: powder preparation, calcinations, pressing,
and sintering.
1.3.1.
P ow d er sy n th esis. The most common route is the mixed oxide route,
where oxides of the constituent cations are mixed in stoichiometric proportions. However,
B aO powder is not suitable because it is either unstable or highly hygroscopic in the air.
Thus, usually BaCO% th a t they will decompose to oxides on the elevated tem perature is
used. The purity of raw powder is im portant since the impurities can affect the microwave
properties of dielectrics. However it would be not common to use extremely pure powder
due to the expense. The particle size of a powder is the secondary consideration. The finer
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
25
powder size produces the less uniform material resulting from the increase of the inter­
particle Van de Waals attraction forces. Howerver it can reduce significantly the sintering
tem peratures. The finer powder has its limited application. For example, agglomerates can
result in nonuniform sintering, porosity, and irregular grain growth. After the dry powders
have been weighed out in their stoichiometric proportion they must be mixed. This is usu­
ally performed in some form of ball mill. By this procedure, particle size and agglomerates
can be reduced. Zirconia milling media and water with the w ater are usually used in a
polyethylene bottle for this procedure.
1.3.2. C a lcin ation . Calcination is a process th a t involves heating the mixed pow­
der to a tem perature th a t is below the sintering tem perature. This process can decompose
any existing carbonates or nitrates. It also may form interm ediate compounds of the con­
stitutes prior to sintering or fully react the constitute powder to form a single phase final
materials. Finally during the calcination process there is solid-state diffusion th a t promotes
the coarsening of the powder.
1.3.3. P ressin g . The next step is to produce the green body to be sintered. The
preparation of powders and their formation into green is extremely im portant. Although
the addition of a binder is necessary to reduce the pressing faults of the green body, usually
green bodies up to about 50 mm diameter can be formed by uniaxial pressing in a stainless
steel dies. If a binder is added, the heating rate must be slow enough to prevent producing
a large am ount of trapped gas from the binder decomposing during the sintering process.
Larger samples can be produced by isostatic pressing.
1.3.4. S in tering.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
26
Material
transfer
Figure 12. The neck between particles forms a miniscus which exerts capillary pressure
drawing particles together. Source [39]
1.3.4.1.
Solid state sintering. Sintering refers to the process of firing and consoli­
dating a green body from powder particles. The step is very im portant to achieve the
excellent microwave properties. However, these materials with complex crystal structure do
not sinter easily because of the narrow tem perature window available for sintering due to
the existence of the volatility of some components of undesirable secondary phases. Thus,
it is common to use additives th a t can facilitate the liquid phase sintering. This process
will be discussed later in this section. The particle size of a powder strongly affects the
conditions needed for optimizing the sintering process. Smaller particles have a greater
curvature (i.e., smaller radius of curvature). The surface energy of a curved surface causes
a pressure difference across the surface. It is the same surface energy effect th a t it causes a
liquid to rise in a capillary. The driving force for sintering is the reduction in surface energy.
In a powder compact this means the reduction in free surface area. An im portant effect of
the pressure difference across a curved surface is th a t increases the vapor pressure. In fine
powders, where the radius of curvature is large, these effects can become very significant.
At the microscopic level, the driving force for sintering is the capillary pressure associated
with surface curvature where particles come into contact (see figure 12). The neck formed
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
27
‘Grain b o undary
Figure 13. A lternate pathes for mass transport during the initial stages of sintering. Mech­
anism number is explained in Table 2. Source: [39].
between particles has a saddle curvature characterized by two radii of curvature, x, the
diameter of the neck, and p, the characteristic diam eter of the inter granular region (see
Fig. 12, which is small and negative. The capillary pressure is:
AP
Mechanism number
1
2
3
4
5
6
= 7(I + I)
Transport path
Surface diffusion
Lattice diffusion
Vapor transport
Boundary diffusion
lattice diffusion
lattice diffusion
p
(1.14)
x
Source m atter
Surface
Surface
Grain boundary
Grain boundary
Grain boundary
Dislocations
Sink of m atter
Neck
Neck
Neck
Neck
Neck
Neck
Table 2. A lternate paths for m atter transport during the initial stages of sintering. Source:
[39]
Fig. 13 describes several m atter transport paths during the initial stages of sintering:
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
evaporation-condensation, viscous flow, surface diffusion, grain boundary diffusion, and
lattice diffusion. Some of these lead to densification, which refers to the shrinkage process
and requires th a t the centers of particles approach one another. O ther transport mechanisms
lead to coarsening, which is a growth of the neck between particles leading to reduction
of the specific surface area w ithout shrinkage.
Generally densification is the dominant
process involving in achieving the high density sample.
Only m atter transfer from the
particle volume or grain boundary results in densification. For ceramics it is the solidstate processes, th a t is, viscous flow and diffusion th a t are of the most im portance for
densification. Evaporation-condensation requires th a t m aterials be heated to a tem perature
high enough for the constituent’s vapor pressure to become appreciable. This is a condition
th a t does not usually occur in the sintering of oxide ceramics. More complete descriptions
of the sintering process can be found in Ref.[39, 40, 41]
1.3.4.2.
Liquid phase sintering. Complex perovskite compounds are especially diffi­
cult to sinter. For this reason, liquid phase sintering is a very im portant sintering mech­
anism. The main advantage of using a liquid phase is th a t densification is more rapid.
Examples include B a T iO %, in which a slight excess of titan ia in the overall composition
forms a Ti-rich eutectic liquid at above 1320°C; see Fig. 10. The essential requirements
for a successful liquid additive include having an appreciable solubility for the principle
constituent, a reasonably low viscosity for rapid diffusion kinetics, and a contact angle th a t
allows it to wet and penetrate between the principal constituent particles. The driving force
for densification arises from the capillary pressure of the liquid between the solid particles.
Several processes assist the densification processing. W hen the liquid forms, particle re­
arrangement, which can increase the packing, may occur. Second, at the particle contact
points, creep can occur due to the high stresses. The solution-precipitation processes may
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
29
take place. Solution-reprecipitation from smaller particles and highly curved surfaces leads
to flattened boundaries and core densely packed grains. Such effects can greatly assist the
sintering of highly refractory ceramics, but there are serious disadvantages. The main dis­
advantage is th a t the liquid phase can solidify into as a very lossy component. In the case
of peritectic reactions, the reaction product is often small in volume, but still sufficiently
lossy to degrade density performance. The presence of liquid phases during sintering can
also cause microcracking.
1.4. R esearch m o tiv a tio n and th esis organ ization
1.4.1.
R esearch m o tiv a tio n and goal. In the previous sections, the research
background for microwave dielectrics ceramics and ferroelectrics was briefly discussed .
Key steps in m aterials processing th a t can produce bulk ceramic samples were described.
The motivation and goal of this thesis research are summarized as followings:
• High performance microwave ceramics play an im portant role in microwave resonators
and filters, but we do not have a fundamental understanding of why this class of m ate­
rials can have both a high dielectric constant and low loss. On the other side, a growing
number of these devices require intricately shaped microwave materials. Theoretical
investigation of R a (Z n xC'd1/ 3_x)T a2/303 using ab initio calculation suggests th a t the
relative contribution of d-electron bonding can provide enhanced directional covalent
bonding. This directionality can strengthen the soft anharmonic lattice modes th a t
are believed to be involved in the microwave loss mechanism of practical dielectrics.
Thus, B a { Z n x C di/^_x)T a 2 /zOz is the system th a t was chosen to study the structural,
chemical, electronic and high frequency dielectric properties of this class of materials.
While initial attem pts to produce the structures with conventional ceramic forming
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
30
technology had limited success as a result of the intrinsic hardness and brittleness of
B a ( Z n i / 3 Ta,2 / 3 ) 0
3
, the ceramic injection molding m ethod is proposed to fabricate
microwave devices w ith complex structure.
• Ferroelectric ceramics have been used in a wide variety of applications in electron­
ics. For example, barium and strontium titan ate are used extensively as high-voltage
capacitors due to their high dielectric constant, low dispersion, and wide frequency
range of response. However, besides the strong tem perature dependence of its di­
electric constant near the Curie tem perature, these m aterials have large dissipative
losses and uncontrolled electrical breakdown at m oderate electric fields (3kV/mm).
Previous investigations have shown a strong correlation between the presence of point
defects and breakdown voltage. Thus, another goal of this thesis is to further identify
and investigate the effect of the point defects on the leakage current and breakdown
voltage of B a xS r \ - xT i 0
3
.
1.5. T h esis organ ization
There are two parts in this thesis. In the first part, the structural, chemical, elec­
tronic and high frequency dielectric properties of B a ( Z n x C d i ^ _ xT a 2 /s ) 0 3 are discussed
in detail. The synthesis of microwave resonators with complex shape using ceramic injec­
tion molding are also described. In the second part, the effect of Vanadium and Scandium
doping on the dielectric properties and I-V characteristics of B a x S r i - xTiO s under high dc
voltage is discussed first. Then, how the point defects induced by neutron irradiation and
annealing treatm ent affect the leakage current is explored.
In chapter 1, the research background for microwave dielectric and ferroelectric ce-
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
31
ramies is discussed briefly. The m ethods to produce bulk ceramic samples is described next.
The motivation and goal of this thesis research also is discussed after that.
In chapter 2 and 3, the B a { Z n xCdij-z-xT a 2 j^)Oz solid solution powder is synthe­
sized. Ba{Cdii^,Ta 2 /z)Oz ceramics with high densities are attained by adding 2wt% ZnO.
Further reduction in sintering tem perature and time can be achieved by adding boron oxide.
Their m icrostructure and properties are characterized using a wide range of characteriza­
tion techniques, including therm ogravim etry technique, SEM, X-Ray diffraction, immersion
pycnometry, TEM , HREM, EELS, and microwave measurements. Local density functional
calculations are carried out to investigate the unusual nature of this class of material.The
results of the ordered domain structures and its boundaries in B a ( C d i/z T a 2 /z)Oz ceramics
produced with and w ithout boron oxide as a sintering aid was summarized.
In chapter 4, the development of a ceramic injection molding (CIM) process to
produce complex-shaped structures using high-performance microwave ceramic materials is
reported. In particular, the synthesis methods and the structural, chemical and dielectric
properties of B a ( Z n i/ z T a 2 /z)Oz doped with Z r and N i (BZT) ceramics produced using
ceramic injection molding are described.
The aim of chapter 5 is to investigate the influence of high d.c field on the dielectric
properties and I-V characteristics of Scandium (acceptor) an d /o r Vanadium (donor) doped
B aTiO z and BaajSro.zTiO z as well as the effect of Vanadium an d /o r Scandium on the
crystal structure and dielectric properties of B aTiO z and B a o jS ro ^T iO z- In chapter 6, the
effect of point defects induced by neutron irradiation and annealing treatm ent in different
atmosphere ( 0 2, 95vol%Ar+5vol%H2) at different time and tem perature is investigate,
using I-V measurement, param agnetic measurement, EPR , and XRD. All the conclusions
and significant research progress and proposed future work are listed in chapter 7.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
C H APTER 2
EXPERIMENTAL AND THEORETICAL
INVESTIGATION OF BARIUM TANTALATE BASED
CERAMICS
2.1. In tro d u ctio n
In this chapter, the structural, microstructural, chemical, electronic and high fre­
quency dielectric properties of barium cadmium tantalate-based ceramics are reported.
B a {C di/ 2,T a 2 /z)Oz based ceramics was chosen and studied since initial theoretical calcu­
lations predict th a t the relative contribution of d-electron bonding will be stronger in Cdcontaining compounds th an in other A ( B 1 / 3 T a 2 / 3 ) 0 3 perovskites [69]. On the other side,
previous investigations show clearly the ordering can affect significantly the microwave prop­
erties of perovskite compound [14, 81]. The tendency for ordering in B a {C d i/^T a 2 / 3 ) 0 3 is
much higher compared with th a t for B a ( Z n i / 3 T a 2 / 3 ) 0 3 and other similar ceramics because
of the large difference in ionic size between the B ' and B" cations [52, 53], Prior work that
measured low frequency properties of B a { Z n xC d \j 3 _xT a 2 j 3 ) 0 3 alloys reported a maximum
dielectric constant of ~30 at 1 kHz.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
33
2.2. E x p erim en ta l p roced u res
B a (C d i/ 2 T a 2 /z)Os powder were made from reagent grade BaCO s, T a 2 0
3
, and CdO.
The raw powder was blended by using distilled water and Z r 0 2 ball milling media (diam­
eter ~1 cm) with a 20 : 1 milling ball to powder weight ratio for 16 hours in a milling
machine. This step served to deagglomerate the powders and provide a homogeneous dis­
tribution of raw powder. The slurry was subsequently dried. The dried powder was filtered
through a 14-mesh screen. Then, the powder was heated to 1350°C for 10 h with an ini­
tial ramp of 100°C/h in air in a box furnace (CM furnaces Model 1700) to bring about
reaction of the raw powder to form single-phase powder.
B a (C d 1 / 3 T a 2 / 3 ) 0
3
After the reaction step, the
powder was milled in a poly(vinylacohol)-poly(ethylene glycol) aqueous
slurry in order to reduce the particle size to th a t which will facilitate densification dur­
ing sintering. B a ( Z n x C d i/ 3 _xT a 2 / 3 ) 0 3 alloy powders were also produced with the same
procedure using reagent grade B a C 0 3, T a 2 0 3, and Z n O and CdO.
Undoped B a ( C d i/ 3 T a 2/ 3)Os ceramics were sintered at 1520°C for 48 hours in air,
with an initial heating rate of 300°C/h.
These sintered samples were porous with low
density. To produce high density B a (C d i/ 3 T a 2 / 3 ) 0 3 compounds, the addition of 2wt% Z nO
powder as a sintering aid was found to be essential since high density ceramics could not be
successfully produced w ithout a sintering aid. B a (C d i/ 3 T a 2/ 3)Os samples with 2wt% Z n O
were pressed to ~ 60% of theoretical density and then sintered at 1450°C,1520°C, 1550°C
and 1580°C for 48h in a P t crucible with an initial ram p rate of 100°C/h in air. During
sintering, the crucible was sealed by P t foil and the samples were covered with powder
enhanced with additional CdO to reduce and offset the loss of this volatile component. The
samples were slowly cooled and no additional post-firing heat treatm ents were used in this
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
34
investigation.
Phase stability of powder was studied using a thermogravimetric measurement sys­
tem. In each run, the mass loss of powder was monitored at constant tem perature ram p as
a function of time. The powders were allowed to decompose to completion.
The structure of the powder and ceramics was characterized using a Rigaku D/MAXIIB diffractometer. A single crystal graphite monochromator was used to attain Cu K a
radiation. Polysilicon powder with a size less than 75mm was used as a reference standard.
The lattice constants were determined by fitting at least 6 of the dom inant diffraction peaks
in the spectra using a least square error minimization fit in the M DI-Jade program with
the pattern-indexing feature. Simulations of X-Ray diffraction spectra were performed with
the same software using the P attern calculation feature.
Immersion pycnometry was used to quantitatively determine the density for wellsintered samples. The bulk density of other samples was evaluated by measuring the di­
mensions and weight of the specimen.
Samples for transm ission electron microscopy were prepared by a standard technique
involving mechanical grinding and dimpling to a thickness of about 10 mm, and final argonion milling to electron transparency at 3.5 keV. Samples were coated lightly with carbon
before TEM observations to avoid surface charging. Electron diffractions and high resolution
transmission electron microscopy were performed at 400 KV (JEM 4000EX).
The microwave quality factor (Q ), dielectric constant and tem perature coefficient
of resonant frequency, r / , were measured with the T E qis mode of a dielectric resonator.
The unloaded quality factor (Q q) was determined using techniques th a t are a variation on
Ginzton [54], Kajfez [55, 56] has extended Ginzton’s measuring and graphical techniques of
using VSW R to measure Qo and coupling of the dielectric sample in a resonant cavity. The
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
35
procedure consists of placing the cylindrical DR sample to be measured in an Au-coated
aluminum cylindrical cavity of dimensions approximately three times greater than the mea­
sured sample. The T E qi$ mode is measured using
S 'n
reflection d ata at the terminals of
the one port cavity. S ll d ata is collected spanning the resonant frequency and the imme­
diate lower and upper frequency bounds. This d ata is processed [55, 56] using a fractional
linear curve fitting routine and then graphically displayed as a Q circle on a Smith chart.
Values of Q o, loaded Q and the coupling coefficient are also displayed, the routine using
the over-determined system of equations derived from the S u data. This method, which
we will refer to as the dielectric resonator method, does not produce precise measurements
of the dielectric constant (error ~ ±20%) as a result of the near-field coupling between
the dielectric resonator and the metal resonant cavity. However, it can give reasonable
trends between materials when similar size dielectric samples are used. The tem perature
coefficient of resonant frequency was measured over a tem perature range of 25 —60° C. To
accurately determine the dielectric constant (ey), a different technique was used which uti­
lizes an open sided, parallel plate holder. This concept, proposed originally by Haki and
Coleman, was further developed by Courtney, and has subsequently retained the name of
the Courtney M ethod. In our measuring technique, we have adopted the additional refine­
ments outlined by Wheless and Kajfez [57] th a t allow for the identification of the T E qis
mode as well as other modes th a t arise in this measuring configuration. The DR sample
of known dimensions is placed between the plates of the two ports Courtney holder and
connected to a vector network analyzer in transmission mode. The coupling probes of the
two ports are either magnetic loops or electric probes th a t can be oriented both horizontally
and vertically. Using this flexibility of probe type and orientation, the resonant frequency
of the desired T E qu mode to be measured can be isolated. Mode charts [57, 58] are also of
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
36
assistance in this mode identification process. Once the resonant frequency is identified, the
transcendental formulations [59, 60] are utilized and compared to provide the D R sample
er . Root finding and graphing routines in M athem atica [61] was used to extract the value
for sr for each DR sample.
2.3. T h eo retica l approach
Phonon frequencies and eigenmodes at the P-point were calculated within the
local-density approximation, using a full-potential, generalized linear muffin tin orbitals
method [63]. Initially the lattice was relaxed assuming the P 3 m l sym m etry associated
with B a ( Z n 1 / 3 T a 2 / 3 ) 0 3 . To compute phonon frequencies a frozen phonon approach was
adopted by computing forces Fj for each of a sequence of small, finite displacement Sxi in
coordinate i of the unit cell (there are three Cartesian components i for each site). The
dynamical m atrix was constructed by Vij = Fj / 5XI and the phonon eigenvalues determined
from Vij. Local orbitals [64] were used to include the B a and T a 5p states in the valence
simultaneously w ith the usual Qp states. Local orbitals were found to be necessary for ac­
curate total energies, as is often the case in transition-m etal oxides w ith their short bond
lengths. A rather large L M T O basis set (about 15 orbitals/atom on average) was used,
resulting in a well-converged L D A calculation. It was found th a t for B a {C d i/ 3 T a 2 / 3 ) 0
3
an imaginary phonon eigenvalue was found, indicating a symmetry-lowering distortion to
P321 symmetry, bending the T a — 0 — Cd bond from 180 degrees as described below. The
relaxation was traced until a zero-force condition was found. The P-point phonon eigen­
values were recalculated at the new positions and found to be positive, indicating a stable
geometry.
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
37
(0
V)
o
_l
</>
TO
2
-
2-
-
6-
1200°C
1250°C
1300°C
1350°C
1400°C
0
-
10 -
-
12-
0
10000
30000
20000
40000
Time (s)
Figure 14. Therm ogravim etry measurements of the mass loss from B a (C d 1 i 3 T a 2 / 3 )O3 as a
function of time and tem perature B a (C d 1 / 3 T a 2 / 3 ) 0 3 .
2.4. R e s u lts a n d D iscu ssio n
B a ( C d i/ 3 T a 2 / 3 ) 0 3 th e rm o g r a v im e tr y e x p e r im e n ts . Figure 14 shows
2 .4 .1 .
therm ogravim etry experiments th a t measure the mass loss of exposure to 1500°C during
the therm ogravim etry measurements. Evidence of the presence of secondary Ba^Ta^O^
and Ba^Ta^Oi^ phases can be clearly seen B a (C d i/ 3 T a 2 / 3 ) 0
3
as a function of time and
tem perature.
X-ray diffraction d ata of the resulting product, as shown in Figure 15, indicate th at
B a (C d 1 / 3 T a 2 / 3 ) 0
3
decomposes according to the following reaction:
3B a 3 (C dTa 2 )Og(s) —> B a 3 T a^O \ 3 {s) + B a ^ T a ^ O ^ s ) + 3CdO(g)
(2.1)
The evaporation flux under equilibrium conditions is determ ined by the HertzKnudsen-Langmuir equation: Jv = Pe(2pm kT ) ~
1/ 2
, where m is the molecular mass, k
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
38
700-
* Ba.Ta.O,
5
4 1
# BaTa„0,
600500400-
c
300200-
100-
20
30
40
50
60
70
20
Figure 15. X-ray diffraction spectra of decomposed Ba{ Cdi /^T a 2 /z) 0 ^ powder after expo­
sure to 1050°C during the therm ogravim etry measurements.
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
39
BaCd^Taa/sOs (s)
Q.
■o
cr
Ba5Ta4 0 15 (s)+Ba4Ta20 9(s)+CdO(g)
0.01
7.2
7.4
7.6
7.8
8
8.2
10000/T (1/K)
Figure 16. Ellingham diagram for B a ( C d 1 / z T a 2 /z)Oz-
is the Boltzmm an constant, and T is the tem perature. Under conditions in which kinetic
barriers are small, as would be expected for this system, the therm odynam ic param eters for
a therm ally activated process can be deduced from: A G = —R T ln k and A G = A H —A S T ,
where k is the reaction equilibrium constant and equals the partial pressure of CdO in this
chemical reaction, and AG, A H and A S correspond to Gibbs free energy, enthalpy and
entropy changes, respectively. The evaporation rate, as inferred from the Hertz-KnudsenLangmuir equation and the mass loss rate at maximum slope for each isothermal run, is
plotted in Figure 16.
From this analysis, A H and A S of this decomposition reaction
were determined to be 170.82K J / m o l and 113.07J / m o l respectively. This plot represents
a pressure-tem perature phase diagram th a t is divided into two areas by the critical stabil­
ity line. Above the line, B a ( C d i /z T a 2 /z)Oz is stable. Below the line, B a ( C d i /z T a 2 /z)Oz
decomposes into BG^To^Og and Ba$TaiO\z.
The results show th a t B a{ C di jz T a 2 /z)Oz
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
40
(3) B a(Zni/4C d1/12)Ta2/30 3
(b) Ba(Zn2/9Cd1/9)Ta2/30 3
(c) B a f Z n ^ C d J T a ^
(d) Ba(Zn1,12C d i,4) T a 2,30 3
( 110 )
( 101 )
(201)
(200)
^
(211)
>»
4->
'55
c
o
+->
c
20
30
40
50
60
70
20
Figure 17. X-Ray diffraction spectra of B a ( Z n x C di/ 3 _xT a 2 /z)Oz powder,
decomposes at a m oderate rate at the tem perature and pressures required for sintering.
2 .4 .2 .
X -R a y diffraction. X-Ray diffraction spectra of B a { Z n xC d \ j z - xT a 2 /z)Oz
powders, as shown in Figure 17, indicate th a t this system forms solid solutions over the
entire range of alloy compositions. The observed trend in the increase in lattice constant
with increasing Cd content, summarized in Table 3, is expected since the ionic radius of
Z n 2+ is smaller than th a t of Cd2+. In this case, we labeled the XRD p attern in figure
17 using the notation of the cubic structure.
Fig. 18 shows X-ray diffraction spectra
of Ba ( C d i /z T a 2 /z)Oz ceramics with 2wt% Z n O as a sintering additive as a function of
sintering tem perature. Evidence for Cd-Ta ordering, as indicated by the presence of weak
(100) superstructure peak in the X-ray diffraction spectra (indicated by an asterisk) near
~ 18° can been found for all the sintered samples. It should be noted th a t X-ray diffrac-
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
41
a (A) (cubic)
4.094
4.104
4.105
4.146
4.148
4.163
B a ( Z n i / 3T a 2 / 3 ) 0 3
Ba {Z n i/ ±C di /i 2 )T a 2 /?,0 3
B a ( Z n 2 /gCdi/g)T a 2 / 3 0 3
B a( Z n i/ gC d 2 /g)Ta 2 / 3 0 z
B a ( Z n i / i 2 C d i/ 4 )Ta, 2 / 3 0 3
B a( C di /3T d2 /z) 0 3
a (A) (hexagonal)
5.788
5.802
5.804
5.868
5.871
5.880
Table 3. The lattice constants of B a ( Z n xCd^i/ 3 _x^Ta 2 / 3 ) 0
and hexagonal structure.
3
c (A) (hexagonal)
7.099
7.109
7.111
7.193
7.195
7.210
powder samples fit to the cubic
second phase
(110)
superstructure phase
(202)
*
</>
(220)
(214)
. ( 101)
(104)
(100)
1450PC
C
_A.
1550PC
1580PC
20
30
40
50
60
70
26
Figure 18. X-Ray diffraction spectra of sintered B a{ Cd ij 3 T a 2 j 3 ) 0 3 synthesized with 2wt%
ZnO as sintering aid.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
42
tion simulations of ordered structures indicate th a t the superstructure peak intensity of
B a( C di /z T a 2 /z)Oz will be ~ 3 times smaller than th a t of B a ( Z n i / z T a 2/ 3)Oz for the same
amount of ordering due to the smaller difference in scattering am plitude (i.e., smaller AZ)
between Cd and Ta than between Zn and Ta. B a ( C d i /z T a 2 /z)Oz samples prepared with a
boron oxide sintering aid found high intensity superlattice peaks using selected area elec­
tronic diffraction (SAED), although the XRD superlattice peaks were also very weak [65].
Earlier work by A. J. Jacobson [66] did not detect superlattice lines resulting from cation
ordering for the B a — T a — Cd —O system.
2.4.3. D en sities and M icrostru ctu re. The density of B a {C d i/ z T a 2 /z)Oz is less
than 80% when prepared without a sintering additive. Significantly improved densification
is attained with the addition of Z n O as a sintering additive. Figure 19 shows the depen­
dence of sample density on sintering tem perature. The small density of samples sintered
above 1450°C is presumably due to the evaporation of CdO , as would be expected from
our therm ogravim etry results. Typical SEM of B a ( C d i / 3 T a 2 /z)Oz with 2wt% Z n O after
1550°C sintering for 48h are shown in figure 20. There is significant grain growth compared
with the starting powder (~l-2/um).
2 .4 .4 . O rdered
d om ain
stru ctu res
and
its
b o u n d a ries
in
undop ed
Ba{Cdi/zTa2/z)Oz> 1
2.4.4.1.
Selected area electron diffraction. 2 Figure 21 shows the selected area elec­
tron diffraction p attern of the ordered domain structure for B a ( C d i /z T a 2 /z)Oz taken along
1Work done by Prof.Smith group: 1) J. Sun, S. J Liu, N. Newman, and David J. Smith, Appl. Phy.
Lett, 84, 3918 (2004); 2) J. Sun, S. J. Liu, N. Newman, and David. J. Smith, Ordered Structures in
Ba(Cdi/ 3 Ta 2 / 3 ) 0 3 Microwave Ceramics: A Transmission Electron Microscopy Study, Mat. Res. Soc.
Symp. Proc. Vol. 783, 2004 Materials Research Society, B5.12.1.
2The indices are indexed by reference to the reciprocal lattice of the cubic perovskite structure with
lattice constant ~4.1 A
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
43
97
o^.96
>.
CO 95
♦
C
CD
Q 94
<D
— 93
CD
♦
91
♦
♦
90
1460 1480 1500 1520 1540 1560 1580 1600
Temperature (°C)
Figure 19. Dependence of relative density on sintering tem perature for B a ( C d i / 3 T a 2 / 3 ) 0 3
ceramics synthesized with 2wt%ZnO sintering agent. Note: The theoretical density of
Ba (C d i / 3 T a 2 / 3 ) 0 3 is 7.94p/cm 3.
Figure 20. Scanning electron micrograph of Ba{C di / 3 T a 2 / 3 ) 0
1550°C for 48h.
3
with 2wt%ZnO sintered at
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
44
the zone axis [110]. Strong superlattice reflections are observed at positions of (h ± 1/3,
k ± 1/3, I ± 1/3) away from the fundamental reflections for the cubic perovskite cell (lattice
constant of ~ 0.41 nm), along both of the allowed [111] and [111] directions in the [llO] zone
diffraction pattern. Thus, the ordering is present in the sequence of one Cd2+ layer and
two T a 5+ layers along [111] or [111] directions, which implies a twinned crystallographic
relationship for the ordered domains in Ba{Cdti/^Ta 2 /z) 0 ^.
^B
^B
^B
^B
■■BB
111
B■
I
E9■S11^9
19IIIBB
19B||■B
■H B
19■
B1■BB
1HB^9
B
1B
B 1H
B
BB11B^9
■B
■ iH
Figure 21. Selected area electron diffraction p attern for S a(C 'd1/ 3T a 2/ 3)03 viewed along
the [110] zone axis. Source (J. Sun, S. J. Liu, N. Newman, and D. J. Smith, M at. Res. Soc.
Symp. Proc. Vol. 783, 2004 M aterials Research Society, B5.12.1.)
2.4.4.2.
H R E M observation. The direct evidence for the presence of ordered domains
with a twinned relationship is shown in Fig. 22, where the ordered domain structure are
visible. The average domain size was estim ated to be about 18 nm. Along both [111] and
[111] directions triple-period lattice fringes with a periodicity of ~ 0.71nm, corresponding to
1/3(111) superlattice reflections in the diffraction pattern, are observed. Single domains are
well-ordered with hexagonal symmetry with lattice constants a ~ 0.58 nm and c ~ 0.71nm
respectively.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
ditive sintered at relatively
• J- Smith, Appl. Phy. Lett, 84, 3“
o46) , ^
ceramics without boron ad
<J' SU" ' & J Li“ > N ' N e w m l°"^ d
Reproduced with permission of ,he copyright owner. Further reproduc
p ro d u ctio n p ro h ib ited w ith o u t p erm ission .
46
Figure 23. High resolution electron micrographs of twin boundaries in Ba {Cd i/ ^Ta 2 /^) 0 ^
viewed along the [110] zone axis: (a) boundary on (001) plane; (b) boundary on (110) plane,
as indicated by the white arrows. Local side steps are also observed a t boundaries. The
superstructure unit meshes are indicated by white rectangles. Source (J. Sun, S. J Liu, N.
Newman, and D. J. Smith, Appl. Phy. Lett, 84, 3918 (2004))
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
47
Figure 24.
High resolution electron micrographs of antiphase boundaries in
B a( C d i / 2,Ta 2 /^)Oz viewed along the [110] zone axis: (a) boundary inclined to (111) plane;
(b) boundary parallel to (111) plane. The projected displacement vectors of the type [001]
for antiphase boundaries are indicated by small white arrows. Source (J. Sun, S. J Liu, N.
Newman, and D. J. Smith, Appl. Phy. Lett, 84, 3918 (2004))
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
48
Domain boundaries result from the growth and contact of the different ordered
domains during the ordering transition. These domain boundaries show as either twins
or antiphase boundaries w ith the different orientational and translational variants. As an
example, Fig. 23 shows an orientational boundary between two domains with a twinned
relationship taken along the [110] zone axis. As shown in Fig. 23(a), the triple-period lattice
fringes extending into the twin boundary along the (111) and (111) planes in each domain
are clearly visible. The twin boundary parallels to the (001) plane of the cubic perovskite
structure which shows twin boundaries along the (110) plane. Similar to Fig. 23(a), the
abrupt stop of triple-period lattice along (111) and (110) planes in each domain is observed,
as shown Fig. 23(b).
As shown in Fig. 24, antiphase boundaries are observed w ithin a single orientational
domain of the ordered structure.
Fig.
24(a) shows an antiphase boundary inclined to
the (111) plane viewed along the [110] zone axis. The lattice fringes show a projected
displacement vector of the type [001] for the antiphase boundary, as indicated by the small
white arrow. A nother example of a geometrically different antiphase boundary parallel to
the (111) plane is shown in Fig. 24. While twin boundaries appear well-ordered structure,
antiphase boundaries have a disordered region with a width of several atomic distances.
In
summary,
electron
diffraction
and
HREM
results
show
the
undoped
Ba{C di/^ Ta 2 /z)Oz ceramics sintered at 1520°C for 48 hrs have a well-ordered structure with
hexagonal symmetry. These results are analogous to those formed in R a (Z n 1/ 3T a 2/ 3)03
and other similar perovskite ceramics, with a ..B'2+ — T a5+ —T a5+.. repeat sequence along
the < 111 > direction of the parent cubic perovskite cell. High density of domain bound­
aries also are observed. The appearance of ordering structure is expected to significantly
affect the dielectric properties of B a { Z n i / ^ T a 2 /z)Oz-
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
49
4.0
BCT with 2wt%ZnO
3.5
o
3.0
M—
x
a
25
2.0
CT
1440 1460 1480 1500 1520 1540 1560 1580 1600
Temperature (°C)
Figure 25. Dependence of Q x f on the sintering tem perature of B a { C d i jz T a 2 /z)Oz samples
synthesized with 2wt%ZnO sintering agent.
2.4 .5 .
D ie le c tric P r o p e r tie s . Figure 25 illustrates the Q x f product (i.e. mi­
crowave quality factor times resonant frequency) of Ba(Cdi/z'^'a2 /z)Oz samples sintered
over a range of tem peratures.
It is clear th a t the Q x f product of Ba ( C d 1 / 3 T(i 2 /z)Oz
is improved substantially when a Z n O sintering agent is added. The Q x f product ex­
hibits a maximum at a sintering tem perature of ~1550°C. It is interesting to note th a t
the highest Q x f are attained in samples th a t contain a significant fraction of secondary
phases. It is particularly surprising given th a t the decomposition products would not be
expected to necessarily produce low loss phases. The same phenomenon has been observed
for B a ( Z n i / 3 T a 2 /z)Oz when the loss of Z n O is correlated with reduced a microwave loss
tangent [67]. The simultaneous appearance of ordering and significant secondary phases
does not allow us to isolate the role of each of these param eters on the microwave dielectric
properties using our current data.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
BCT without sintering aid
26 -
/
■
241440
1460
1480
1500
1520
1540
1560
1580
1600
Temperature (°C)
Figure 26.
Dependence of dielectric constant on sintering tem perature
B a ( C d 1 / 3 T a 2 / 3 ) 0 3 samples synthesized with 2wt% ZnO sintering agent.
for
Dielectric constants of B a ( C d i /z T a 2 jz)Oz with and w ithout 2wt%ZnO, as measured
by the dielectric resonator method, are illustrated in figure 26. The Courtney m ethod was
used to verify the results in figure 26 determined from the following equation:
/o =
34 ,a
( 2 .2 )
where a is the sample radius in mm , t is the sample thickness in m m , er is the relative
dielectric constant, and /o is the resonance frequency in GHz.
The dielectric constant
of Ba{Cdi / 3 Ta, 2 /z)Oz doped with a 2wt%ZnO sintering aid measured with the Courtney
method is ~ 32.5. This value is close to 33.2, the dielectric constant (er ) measured by
the dielectric resonator method. The tem perature coefficient of resonant frequency r / is
measured to be 80 ppm /oC for sample sintered at 1580°C. The presence of a significant
fraction of secondary phases does not allow us to confidently associate this value with the
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
51
intrinsic param eter for this material.
2.4.6. L ocal D en sity -F u n ctio n a l C alcu lation for B a( C d 1/zTa 2 /z)Oz and
Ba{ Znx/^Ta 2 /z) 0 ^. Ab initio calculations within the local density approximation (L D A )
of Ba(Zji] / 3T ^2/ 3)0.3 and Ba(Cd] / 3T 02/ 3)0.3 predict equilibrium lattice constants of
a = 0.574nm, c = 0.700nm and a = 0.583nm, c = 0.717nm, respectively. The full-potential
variant of the m ethod of linear muffin tin orbitals employed was described in ref. [64]. Note
th a t the c /a ratio is very near (3/2)0 5, as is characteristic of an undistorted pseudocubic
crystal. The predicted lattice constants are slightly smaller (0.01 ~ 1%) th an the experi­
ment, as is typically found in the LDA. The Ba {C di /zT a 2 /z)Oz bulk modulus was calculated
to be 1.91M6ar th a t is slightly less than the value for B a ( Z n i / z T a 2 /z)Oz (1.99Mbar) as a
result of the dilated lattice. The most im portant difference between B a ( C d i /z T a 2 /z)Oz and
B a ( Z n i / 3 Ta, 2 / 3 ) 0 3 is the additional distortion found in B a ( C d i / 3 T a 2 /z)Oz th a t results in
a bond angle of 172° for the T a —O —Cd bond. B a { Z n \ j z T a2 jz)Oz has a threefold rotation
about the c-axis, a twofold rotation about y and finally an inversion symmetry, making
12 group operations in all. The relaxed positions for all atoms as calculated within the
LDA are shown in Table 4, The crystal structure for B a ( Z n i / :iT 0,2/ 3)03 has a hexagonal
Bravais lattice and is in the P 3m l space group, see Figure 27. The Bravais lattice for
B a ( C d i /z T a 2 /z)Oz is also hexagonal , but a distortion of the T a and Cd lowers the energy
and breaks the inversion symmetry as discussed below. U nfortunately it is not possible to
detect experimentally whether this distortion actually exists, because the intensity differ­
ence of the simulated X-ray spectra resulting from the distortion of the oxygen atoms is
*only 0.01%. This change is too small to be distinguished experimentally since other factors
including the presence of strain and varying degrees of order in the B-site sublattice can
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
52
180deg
Figure 27. Ball and stick model of (a) B a { Z n i / ^ T a 2 /z)Oz and (b) B a ( C d i / 2,Ta 2 /s)Oz.
Solid black balls are Z n in (a) and Cd in (b). The distortion relative to the bond-centered
configuration has been amplified by a factor of five to more clearly show the distortion, in
particular the buckling of the T a — O —Cd bond. Note: Done by Dr.M ark V. Schilfgaarde
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
53
Ba2
Ba
Ba
Ta
Ta
Zn(Cd)
01
01
01
02
02
02
02
02
02
BZT
0.0000
0.3333
-0.3333
0.3333
-0.3333
0.0000
0.0000
-1.5000
0.5000
0.1714
-0.1714
-0.1714
0.3428
-0.3428
0.1714
0.0000
-0.3333
0.3333
-0.3333
0.3333
0.0000
-0.5000
-0.5000
0.0000
0.3428
-0.3428
0.1714
0.1714
-0.1714
-0.1714
0.0000
0.3386
-0.3386
-0.1754
0.1754
-0.5000
0.0000
0.0000
0.0000
-0.3245
0.3245
0.3245
0.3245
-0.3245
-0.3245
BCT
0.0000
0.3333
-0.3333
0.3333
-0.3333
0.0000
0.0000
0.4836
-0.4836
0.1539
0.1969
-0.1539
0.3507
-0.3507
0.1969
0.0000
-0.3333
0.3333
-0.3333
0.3333
0.0000
-0.4836
0.4836
0.0000
0.3507
-0.3507
0.1969
0.1539
-0.1969
-0.1539
0.0000
0.3398
-0.3398
-0.1670
0.1670
0.5000
0.0000
0.0000
0.0000
-0.3153
0.3153
0.3153
0.3153
-0.3153
-0.3153
Table 4. Table of lattice positions of B a ( Z n i j 3 T a 2 /s)Oz and B a ( C d i / 3T a 2/z)Oz in Carte­
sian coordinates. Note: the dimensions are scaled to the cubic unit cell dimensions. Note:
Done by Dr.M ark V. Schilfgaarde.
result in similar modifications to the spectra.
The energy of the crystal as a function of the oxygen distortion is illustrated in
Fig. 28, with the minimum in energy representing the relaxed position shown in Table
4.
Note th a t the oxygen atom would be expected to oscillate between these positions
at room tem perature since the energy barrier separating the minima (6 MeV) is small
compared to therm al energies (26 MeV). Interestingly, even though B a ( C d i / 3 T a 2 / 3 ) 0 3 was
found to have this distortion while B a { Z n i / 3 T a 2 /z)Oz did not, the phonon modes were
found to be fairly similar, with the Cd-bearing case having slightly softer modes at lower
energy and stiffer modes at high energy. Nevertheless the phonon mode associated with
the O distortion would be anticipated to have a strong anharmonic component in the Cdbearing case.
It is interesting th a t both B a ( Z n i / z T a 2 /z)Oz and B a ( C d i /z T a 2 /z)Oz are
expected to have atypical physical properties due to its unusual d-electron bonding. The
presence of significant charge transfer between the cation d-orbitals is predicted to provide
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
54
I
m
1 .6 -1 .2 -0 .8 -0 ,4 0
Q
Figure 28. Energy as a function of the generalized coordinate Q. Q parameterizes the
collective displacement of the O atoms; Z — 0 corresponds to the high-symmetry position of
the oxygen atom between the Cd and T a atoms; Q = 1 to the minimum energy configuration
given in Table 2. Note: Done by Dr.M ark V. Schilfgaarde.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
k
Figure 29. Electronic band structure of (a) B a ( C d i / 3 T a 2 / 3 )Os and (b) B a ( Z n i / zT a 2 / 3 ) 0 3
as calculated by the Linear Muffin Tin O rbital m ethod within the Local Density Functional
approximation. Note: Done by Dr.M ark V. Schilfgaarde.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
56
a degree of covalent directional bonding between atoms th a t resist angular distortions, a
property absent in conventional ionic compounds. This may strengthen ’’soft lattice modes”
correlated with microwave loss, as might occur in both defective and high-quality ionic
compounds. The influence of the d-electron type bonding, as compared to conventional
non-directional metal-oxygen ionic bonding, may play an im portant role in achieving high
melt tem peratures and enhanced phonon energies. The latter may presumably play a role
in the ultra-low microwave loss (loss tangent) th a t is observed in this class of materials.
In the case of B a { C d i / 3 T a 2 /z)Oz and B a ( Z n i / 3 T a 2 iz) 0 3, theoretical work using the LDA
approximation indicates th a t charge is transferred from Ta 5d-levels in the conduction
band (empty states near the conduction band minimum) to Cd —4d and Z n — 3d levels
(full states near th e valence band maximum), see figure 29. Typically phonon energies
scale inversely with the bond distance. Therefore, high Z materials tend to have reduced
phonon energies and enhanced loss tangents. The high dielectric constant of high Z material
is a result of the large polarizability of the core electrons.
Thus, it is speculated th a t
the presence of a significant amount of d-electron covalent bonding in a compound with
many high-Z components such as B a ( Z n xC di/ 3 _xT a 2 / 3 ) 0 3 can result in enhanced the
phonon energies, possibly resulting in reduced microwave loss, while still maintaining a
large dielectric constant.
2.5. C onclusions
Therm ogravim etry experiment show th a t B a ( C d 1 / 3 T a 2 / 3 ) 0 3 decomposes at a mod­
erate rate at the tem perature and pressure required for sintering, so it is m andatory
to choose carefully the processing param eters to achieve high quality samples.
While
B a { Z n xC d ii 3 _xT a 2 /z)Oz ceramics with high densities could not be attained without the
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
57
aid of a sintering agent, the addition of 2wt% ZnO was needed to achieve over 97% of
the theoretical density for pure BaiCd-y/^T0-2/ 3)03 ceramics. Evidence for Cd-Ta ordering,
as indicated by the presence of superstructure peaks in the X-ray diffraction spectra, was
found. For a sample sintered at 1550°C for 48h, the dielectric constant and microwave
loss tangent were measured to be ~ 32 and 5 x l0 ~5 at 2 GHz. Local density functional
calculations of B a ( C d i /^ T a 2 /s ) 0 3 and B a { Z n i / ^ T a 2 /z) 0 ^ give insight into the unusual na­
ture of this class of material. The conduction band maximum and valence band minimum
are strongly composed of weakly itinerant Ta 5d-and Z n —3d/C d —4d levels, respectively.
This is believed to play an im portant role in having a high melt tem perature and enhanced
phonon energies, as well as the unusual property of having both a high dielectric constant
and low loss tangent.
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
CH A PTER 3
MICROSTRUCTURAL AND DIELECTRIC PROPERTIES
OF BARIUM CADMIUM TANTALATE CERAMICS
WITH BORON OXIDE AS SINTERING AID
3.1. In tro d u ctio n
In chapter 2, theoretical investigation of B a { Z n xCdii^_xT a 2 /z) 0 ^ suggests th a t the
relative contribution of d-electron bonding can provide enhanced directional covalent bond­
ing, a property absent in classical ionic compounds. This directionality can strengthen
the soft anharmonic lattice modes th a t are believed to be involved in the microwave loss
mechanism of practical dielectrics. Methods to produce high-density ceramics at reduced
sintering tem peratures and shorter durations are needed to obtain high quality materials
due to the large Cd vapor pressures th a t occur during high tem perature sintering. Several
procedures can be used to minimize this effect. For example, the sintered m aterials can be
covered with a powder of similar composition or enriched with Cd during sintering. Z r is
also commonly added to B a { Z n i / 2,Ta 2 /z) 0 3 since it has been found th a t high quality fac­
tors (Q) can be obtained using much shorter annealing times [122, 14]. Unfortunately, the
use of Z r does not appear to have the same beneficial effect in B a ( C d i ^ T a 2 /s) 0 3 . Because
significant volatilization of cadmium occurs during sintering, reduced sintering tem perature
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
59
and duration are considered desirable for the synthesis of highly densified Cd-containing
ceramics. However, B a ( C d i / 3 T 0.2/ 3)03 has exhibited poor sintering properties even for
tem peratures as high as 1550°C for 48 hours [68]. Recent experiments have shown th a t ad­
dition of boron as a sintering aid can enhance the sintering properties of Ba iC dxj ^T a^j ^O z
significantly [69]. Highly densified ceramics have been achieved by sintering at tem pera­
tures as low as ~1200 - 1300° for short durations. In this chapter, a comprehensive study is
reported using a wide range of characterization techniques, including SEM, X-Ray diffrac­
tion, immersion pycnometry, TEM, HREM and microwave measurements to characterize
the influence of the boron oxide sintering aid on the m icrostructure and dielectric properties
of Ba(Cdii%Ta 2 /z)Oz.
3.2. E x p erim en ta l proced u re
B a i C d i / ^ T 0.2/ 3)03 was made from reagent grade BaCOs, cadmium oxide (CdO)
and tantalum oxide (Ta^Os).
The raw m aterials were blended using Z r O '2 milling ball
media with a diam eter of ~ 1 cm and distilled water for 16 hours in a milling machine to
mill and deagglomerate the powders. The weight ratio between milling ball and powder
was 20 : 1. This procedure was used to successfully produce a homogeneous distribution
of particle w ith a grain size of about 1 to 2 fim. The powder was heated to 1050°C for 6h
with an initial ram p rate of 100°C/h in air in a box furnace (CM furnaces Model 1700) to
initiate reaction of the raw materials to form single-phase materials.
After the reaction, the powders of B a (C d i / 3 T a 2 /s ) 0
3
with varying am ounts of boron
oxide powder were milled in aqueous slurry with poly(vinylacohol)-poly(ethylene glycol) to
reduce the particle size to th a t which will facilitate densification during sintering. The
resulting slurry was dried in an oven at 60° C, and then filtered through a 14-mesh screen
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
60
to produce agglomerates suitable for pressing.
B a ( C d i / 3 T a 2 / 3 )0
3
samples were heated
in an AfaOs crucible at 1350°C for 2h with an initial ramp rate of 300°C/h in air. The
samples were covered w ith powders of the same composition to avoid excessive cadmium
loss.
The samples were cooled in the furnace.
B a ( Z n i / 3 T a 2 / 3 )0
3
It has been established th a t undoped
exhibits low dielectric loss when annealed for a long time as a result
of 1 : 2 ordering on the B site [14]. Thus, different post-sintering annealing tem peratures
and times were used to explore the effect of this process on ordering and microwave loss in
Ba{Cdi/3Ta2/3) 0 3.
The structure of ceramic samples was characterized with a Rigaku D/M AX-IIB Xray diffractometer (XRD) using Cu K a radiation produced w ith a graphite single crystal
monochromator. Polysilicon powder with size less th an 75 /rm was used as a reference
standard. The X-Rays spectra were measured twice; once w ith silicon powder for lattice
constant determ ination and once without to produce a clean spectrum for phase identifica­
tion. The lattice constants were determined by fitting at least 6 of the dominant diffraction
peaks in the spectra using a least squares error minimization fit in the M DI-Jade program
with the pattern indexing feature. Immersion pycnometry was performed to determine the
sample density.
Samples suitable for observation by transmission electron microscopy (TEM) were
prepared in three steps. The sample was first polished to a thickness of approximately
lOOnm, followed by mechanical dimpling of the central part to about 10/jm. The final step
involved the use of argon-ion milling to attain a central portion of the sample th a t was
transparent to the electron beam. Samples were lightly coated w ith carbon before TEM
examination. Samples for examination by scanning electron microscopy (SEM) were coated
with gold to minimize the effect of charging when exposed to the electron beam.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
61
The microwave quality factor (Q) and tem perature coefficient of resonant frequency
(tj)
were measured using the T E qis mode of the dielectric resonator. The resonator was
enclosed in a Au-coated test cavity with dimensions about three times larger than the sample
size. Low-loss supports were used to suspend the specimen in the center of the cavity. The
quality factor was measured in reflection at room tem perature. Microwave coupling was
adjusted to achieve ~40 dB loss by spatially adjusting the electric-coupling probes.
3.3. R e s u lts a n d D iscu ssio n
3 .3 .1 .
X -ra y
B a ( C d i ^ T a 2 /s ) 0
3
d iffra c tio n . Fig.
powder without
boron
oxide
30
and
and
31
show
XRD
B a ( C d i ^ T a 2 / 3 )Os
patterns
ceramics
prepared w ith boron oxide as sintering aid. The B a. (CVi:l/ 3T a 2/ ,3) O 3 powder synthesized
without boron oxide does not exhibit the features characteristic of the B-site ordered
hexagonal compound [i.e., presence of superstructure diffraction peaks and significant
splitting of the (104) and (220) peaks]. Thus, this XRD pattern has been indexed using
the pseudocubic structure.
As can be seen in Fig. 30 and 31, all of the as-prepared ceramic samples, as well
as those exposed to subsequent annealing, have XRD d ata th a t exhibit significant splitting
of the (104) and (220) diffraction peaks, but there is only weak evidence for superstruc­
ture peaks.
Because of the similarity in the atomic number of Cd and Ta, the super­
structure peaks are expected to be smaller by a factor of approxim ately 3 than those in
B a ( Z n i / 3 T a 2 / 3 ) 0 3 with the same degree of ordering. As mentioned before, we see strong
evidence for ordering in the splitting of the (104) and (220) peaks and give additional ev­
idence using selected-area diffraction in the TEM section [65, 71]. These results indicate
th a t Ba {C d i/ 3 T a 2 / 3 ) 0 3 has substantial ordering. Thus, we indexed the XRD p attern using
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
of
62
(1 1 0 )
( 202 )
(024)(220)
( 101 )
(210)
( 201 )
140h
100h
20h
( 100 )
( 110)
■A
Powdei
n ^ ^r
20
30
40
60
50
70
2 T h elta
Figure 30. X-ray diffraction p attern of B a( C di / 3 T a 2 / z ) 0 3 powder and B a ( C d i / 3 T a 2 / 3 ) 0 3
ceramics with 0.05 wt% B 2 O 3 as sintering aid annealed at 1050°C and varying times.
(110)
A nnealed a 1 1250°C w ith 0.5wt% B2O 3
(202)
BaSTa40 15 ■
(104|(122)
T ^\
(100)'
J
(201)‘ (103)'
h
i.........V.
\
(024) (220)
(210) i
J i 8° h
U40 h
(100)'
— ji_.,
p »h
........... ** W“
JO h
.
(100)
i
(211
(110) (201 p ° °
(220)* P ow dei
—1—1—1—1—1—1—1—1—1—1—'—1—1—1—1—1—1—1—1—1—f—
15
20
25
30
35
40
45
50
55
60
65
70
2 theta
Figure 31. X-ray diffraction pattern of B a (C d i / 3 T a 2 / 3 ) 0 3 powder and B a ( C d i / 3 T a 2 / 3 ) 0 3
ceramics w ith 0.5 wt% B 2Os as sintering aid annealed at 1250°C for varying times. Note
th a t the powder sample is indexed with the random-alloy pseudocubic structure and the
ceramic samples are indexed with the ordered hexagonal structure. Asterisks denote super­
structure peaks characteristic of B-site ordering.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
63
5.882 -
7.210
7.208
5.880 -
c axis- 7.206o<
7.204 E
°< 5.878+c■»
S
c
5.876-
7.202
7.200
7.198
7.196
a axis
» 5.874o
« 5.872-
w
c
5
8
I
7.194
7.192
5.870 0
20
40
60
80
100
Annealing Time (h)
Figure 32. Dependence of the lattice constants of B a{ Cd ij zT a 2 /z)Oz containing a 0.05wt%
boron sintering aid on the annealing time at 1050°C.
the ordered hexagonal perovskite structure. In contrast, annealing Ba (Z rii /zT a 2 /z)Oz for
over 50 h is necessary to obtain well-ordered structures and outstanding dielectric properties
[30, 72].
Evidence for a second phase composed of Ba^Ta^Oi^, can be seen at a 29 angle of
approximately 28° in Fig. 31. Fig. 32 is a plot of lattice constant as a function of annealing
time. Careful cell refinements have been used to eliminate any significant effect on the lattice
constant determ ination from the secondary phases. Previous experimental results indicate
th a t B a ( C d i /z T a 2 /z)Oz decomposes rapidly at the elevated tem peratures typically used for
sintering [68]. The observed decrease with increasing annealing time may therefore result
from the formation of Cd and oxygen vacancies and possibly the homogeneous incorporation
of boron ions (radii of only approximately 0.26(A) into the B a ( C d i / 3 T a 2 /z)Oz lattice.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
64
3 .3 .2 . S in terin g
d en sity
and
m icro stru ctu re
of
b oron -d oped
B a (C d i^ T a 2 /s)0 3 .
3.3.2.1. Sintering density. Fig. 33(a) shows the dependence of sintering density on
the concentration of the boron sintering aid. It is clear th a t the use of as little as 0.01
wt% boron additive can produce high density (~ 95%) B ai Cd ^i ^T 0 .2 / 3 ) 0 3 samples after
sintering at only 1350°C and 2 h. The relative sintering density reaches 97.5% when 0.5wt%
boron oxide was added. Further enrichment of the boron oxide concentration was found
to decrease the sintering density significantly. Fig. 33(b) illustrates the dependence of
sintering density of B a{ C di / 3 T a 2 / 3 ) 0
3
samples prepared with 0.1wt% boron oxide annealed
at 1250°C for different times. It shows th a t the extended annealing decreases the sample’s
sintering density. The density loss is believed to be a consequence of CdO evaporation from
B a ( C d i / 3T a 2 / s ) 0 3 .
The sintering and annealing tem peratures used in this study are significantly lower
due to the effectiveness of the boron oxide sintering aid.
Our earlier work found that
the addition of a small am ount of Z n O improves the densification significantly, but high
sintering tem peratures (>1520°C) and long exposure times (>48 h) are still required to
achieve a density of over 95% [68].
3.3.2.2. Scanning electron microscopy. Fig. 34(a) and 34(b) show typical scanning
electron micrographs of sintered samples after 40 h and 120 h anneals. The white spots in
Fig. 34(a) are believed to be associated with contamination produced during the polishing
procedure.
We used an improved cleaning procedure to eliminate these white spots in
subsequent samples. M easurements of the average grain size are summarized in Table 5
using the linear intercept method. This SEM analysis indicates th a t a significant difference
in the average grain size (of ~ 1.7pm) is not observed between as-made samples and those
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
65
S in te re d a t 1350°C fo r 2h
1 ,00 -
D.B0-
0.01w t% B oron
O'"
0.0
0.2
0.3
0.6
0.7
Composition (wt%}
(a)
A n n ealed a t 12S0°C w ith 0.1w t% B, 0
96-
>.
to
c
a>
Q
^ 94-
ro
a)
tr
0
(b)
20
40
60
80
100
120
A n n ealin g tim e (hrs)
Figure 33. Sintering density versus boron concentration for B a ( C d 1 / 3 T a 2 / 3 ) 0 3 sintered at
1350°C for 2 h and (b) sintering density versus annealing time for B a{ C di / 3 T a 2 / 3 ) 0 3 with
0.1'u;i% boron oxide as sintering aid annealed at 1250°C.
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
66
(b)
Figure 34. SEM images of B a ( C d i /z T a 2 /z) 0 s specimen prepared w ith 0.5wt% boron oxide
sintered at 1350°C and annealed for (a) 40 h and (b) 120 h.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
67
Annealing duration(h)
0
20
40
80
120
Sample number
1
2
3
1
1
Average particle size(pm)
1.7
1.7
1.7
3.2
4.7
Table 5. Particle size analysis of samples annealed for different durations.
annealed up to 40 h. A considerable increase in the average grain size is, however, found
after approxim ately 80 h of annealing. During this period, the grain growth results from
the disappearance of smaller grains with accompanying grain growth of the larger grains.
This can be observed by comparing the grain size in Figures 34(a) and 34(b). When the
annealing time reaches 120 h, the average particle size is 4.7/rm, which is almost three times
th a t of samples annealed for 40 h.
3 .3 .3 .
T E M o b s e rv a tio n s . 1 Fig. 35 shows the selected-area electron diffraction
(SAED) p attern recorded along the [110] direction of the Ba {C di /^T a 2 /^) 0 ^ ceramics with
0.5 wt% boron oxide. The SAED p attern can be indexed using a hexagonal structure. The
superlattice reflections are visible at positions of (h ± 1/3, k ± 1/3, I ± 1/3) away from the
fundamental reflections for the cubic perovskite when the patterns are indexed using the
cubic structure. This periodicity indicates th a t cation ordering is present in the sequence of
one Cd2+ layer and two T a 5+ layers, with tripling of the unit cell constant along the [111]
direction for Ba {C di /^T a 2 /z)Oz. B a ( Z n xC d i / s - xT a 2 / 3)O3 ceramics prepared by Ganguii
and coworkers showed superlattice reflections in electron-diffraction patterns indexed by
pseudocubic structure [73]. In the present work, SAED patterns revealed well-ordered ma­
terial for B a ( C d i /^ T a 2 /z ) 0
3
ceramics with 0.5 wt% boron oxide sintered at >1350°C for 2 h.
1EELS, TEM and TREM were done by Prof.Smith group, see: J. Sun, S. Liu, N. Newman, M.R.
McCartney, and D.J. Smith, J. Mater. Res. 19, 1387 (2004).
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
68
Figure 35. Electron diffraction pattern for the boron-doped B a ( C d 1 / 3 Ta, 2 /z ) 0 3 ceramics
along the < l l 0 > direction. Source: 1) J. Sun, S. Liu, N. Newman, M.R. McCartney, and
D.J. Smith, J. M ater. Res. 19, 1387 (2004); 2) S. Liu, J. Sun, R. Taylor, D. J. Smith, and
N. Newman, J. M ater. Res., 19, 3526 (2004).
Evidence for ordering is also found in XRD data, as discussed earlier. The ordered domain
structures will be discussed in more detail below. Analytical electron microscopy and HREM
elucidated the beneficial effect of boron on the sintering behavior of B a ( C d i /z T a 2 / 3 )Oz at
grain boundaries and at multiple junctions of the grain boundaries. A low-magnification
annular-dark-field (ADF) image involving a triple junction of B a { C d i /z T a 2 /z)Oz is shown
in Fig. 36(a).
The multiple junctions along the grain boundaries showed markedly different con­
trast from w ithin the individual grains. Analysis by electronic energy loss spectroscopy,
as shown in Fig. 36(b), indicated th a t the triple junctions were rich in boron. However,
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Ba
s
S3
+■>
d
l-H
-10
(b )
0
10
20
30
40
distance (nm)
50
60
70
Figure 36. (a) Annular dark-field image of triple junction and (b) electron energy loss
spectroscopy elementary profile along the line indicated in (a) showing presence of boron at
triple junction. Source: 1) J. Sun, S. Liu, N. Newman, M.R. McCartney, and D. J. Smith,
J. Mater. Res. 19, 1387 (2004) 2) S. Liu, J. Sun, R. Taylor, D. J. Smith, and N. Newman,
J. Mater. Res., 19, 3526 (2004).
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
70
MBMHM
i^H
su
gj
H H H
(b)
Figure 37. (a) High-resolution electron micrograph of grain boundary showing absence of
amorphous phase and (b) lattice image from region in triple junction showing presence of
amorphous phase. Source: 1) J. Sun, S. Liu, N. Newman, M.R. McCartney, and D.J. Smith,
J. M ater. Res. 19, 1387 (2004); 2) S. Liu, J. Sun, R. Taylor, D. J. Smith, and N. Newman,
J. Mater. Res., 19, 3526 (2004).
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
71
segregation of boron to the grain boundaries was not detected. The lattice image of a grainboundary region in B a ( C d i /z T a 2 /z)Oz with 0.5 wt% boron oxide is shown in figure 37(a).
The lattice fringes in each grain stop abruptly at the grain boundary plane, and neither
secondary phases nor amorphous materials are observed along the grain boundary. HREM
confirmed the presence of an amorphous phase at the triple junctions for B a ( C d 1 / 3 T a 2 /z ) 0 3
with 0.5 wt% boron oxide, as shown by the example in Fig. 37(b). From these observa­
tions, we conclude th a t the boron-oxide-rich amorphous phase located at the grain boundary
junctions was a liquid phase formed during sintering.
Fig. 38 further shows the lattice image of Ba{C di j 3 T a 2 /z)Oz ceramics with boron
additive sintered at 1350°C for 4 hrs. Contrast modulated structures along the [111] di­
rection were observed and lattice fringes along both [110] and [001] directions are also dis­
cernible. The wavelength of the contrast m odulated structures was found to be ~ 0.71
nm for the 1/3(111) superlattice reflection from the corresponding diffraction pattern,
which is three times the (111) interplanar spacing for B a ( C d 1 / 3 T a 2 / 3 ) 0 3. Although weak
diffraction reflections and contrast modulations were also found occasionally in some areas,
the Ba{C di / 3 Ta, 2 /z)Oz ceramics with boron additive exhibit a well-ordered structure with
hexagonal symmetry (a ~ 0.58 nm and c ~ 0.71nm). Lattice images of B a ( C d i / 3 T (1 2 / 3 ) 0 3
with boron additive annealed at 1250°C are the same as shown in Fig. 38. Fully ordered
structures were observed w ithin the interior of grains of the sample.
The electron diffraction and HREM results showed a presence of a well-ordered
structure with 1 : 2 ratio for B a (C d i / 3 Ta, 2 /z) 0
3
ceramics with boron additive sintered at
1350°C with 4 hours. No significant change in ordered structures was observed by electron
microscopy for th e sample subjected to a long-period annealing at 1250°C. In contrast,
earlier research indicated th a t B a ( Z n i / 3 T a 2 /z)'Oz prepared at tem peratures below 1377°C
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
72
Figure 38. Lattice image for B a ( C d i / 3 T a 2 / 3 )Os ceramics with boron additive sintered at
1350°C Source (J. Sun, S. Liu, N. Newman, M.R. McCartney, and D.J. Smith, J. Mater.
Res. 19, 1387 (2004))
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
73
forms a disordered cubic structure. The ordered structure is obtained only when a sintering
tem perature of 1400°C is used [30, 66]. The normally sintered ceramics are only partially
ordered except if high tem peratures and long annealing times are used [30, 72], In th a t case,
the degree of ordering is generally kinetically determined. The present results indicate th a t
the B a( C d i / 3 T 0 *2 / 3 ) 0 3 ceramics are ordered much more easily because of the large difference
in size between B' and B " atoms (0.29 A), which is consistent with the conclusion made
by Galasso and coworkers th a t the ordering increases as the difference in valence and size
between B' and B " atoms increased in A ( B y 3 B y 3)Oz perovskite ceramics [52, 53].
3.3 .4 .
S in te rin g m e c h a n ism . Boron has previously been used as a sintering aid
for preparation of B aT iO z and B o q^ S tq ^ T iO z perovskite ceramics [74, 111, 125]. Densification starting tem peratures of those ceramics was greatly reduced when a boroncontaining sintering aid was used to induce liquid phase sintering. The current results for
Ba {C di /zT a 2 /z)Oz are clearly different from those for Ba TiO z ceramics, where boron seg­
regation on grain boundaries and a film of B a B O amorphous phase along grain boundaries
were reported for B aT iO z ceramics fabricated with boron as a sintering aid [74, 111],When
solid and liquid phases come into contact in an equilibrium condition, the energy relation
for the grain boundary and interface can be expressed by the following equation:
Igb = 2')iscos{<p/2)
(3.1)
where 7gb is the grain boundary energy, q;s the interfacial energy between solid and liq­
uid phases, and ip is the dihedral angle. The morphology and distribution of the liquid
phase in the solid phase depends on the ratio of 'jgb/'Yis, th a t is, dihedral angle. W hen <p
equals zero (7^ > 27/s), grain boundaries of the solid phase are fully wetted by the liquid
phase. More th a n 10 dihedral angles were measured from several ADF micrographs for
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
74
B a( Cd 1 / 3 T a 2 / 3 ) 0 3 ceramics with boron additive, and dihedral angles were found to fall
into the range of 30 to 60°C. These results indicate th a t the liquid phase does not appear
as a film along grain boundaries in Ba {C di j 3 T a 2 / 3 ) 0 3 ceramics, b u t partially penetrates
along the multiple grain junctions, which is consistent w ith the current electron microscopy
observations. In the early stages of sintering of B a ( C d i / 3 T a 2 / 3 ) 0 3 ceramics with boron
additive, B 2 0 3 is expected to quickly form elemental boron and is present as a liquid phase
as the tem perature increases (melting point about 450°C. The B 2 0 3 liquid phase volatilizes
during the subsequent processing because the formation tem perature of B a 0 - n B 2 0 3 and
CdO-nB^Oz liquid phases (close to 1000°C) is much higher than the melting point of B 2 0 3
[77, 78]. Densification of ceramics can arise from particle rearrangem ent under the influ­
ence of capillary forces and the filling of pores by the liquid phase. The remaining B 2 0 3
liquid phase transform s into B a 0 - C d 0 - n B 2 0 3 liquid phases located at multiple junctions
during sintering. This process is responsible for the improvement of sintering behavior for
the Ba{ Cdi / 3 T a 2 / 3 ) 0 3 ceramics with boron additive.
It is interesting to note th a t samples doped with as small as 0.01 wt% boron have the
same high density as the samples doped with 0.5 wt% boron oxide. A high concentration of
mobile vacancies formed due to the evaporation of CdO at sintering tem perature could be
responsible for this behavior. Evidence th a t mobile vacancies can play an im portant role
in the densification of other ceramics has been shown by a number of other studies [79].
It has been reported th a t compensated oxygen vacancies in B a T i 0 3 doped with trivalent
acceptors improved the densification processing [80]. Research is in progress to identify the
mechanism responsible for densification in Ba{C di / 3 T a 2 / 3 ) 0 3 with such low boron oxide
concentration.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
75
BCT w ith 0.5w t% B
50,000 45,000 N 40,000 -
35,000 30,000 -
BCT w ith 0.05w t% B
25,000 -
20,0 0 0
-
0
20
40
60
60
100
120
140
A n n e a lin g T im e {Hour)
Figure 39. Effect of the annealing duration at 1050°C on the microwave properties.
3.3 .5 . D ielectric p ro p erties. We reported in C hapter 3 th a t the properties of
zinc-doped Ba {C di /^ Ta 2 /%)0 2 , dielectric prepared at high sintering tem peratures (1500°C)
and long sintering time (48 h)[68]. In this chapter, the use of boron as a sintering aid has
facilitated the production of ceramics with 95% of the theoretical density at relative low
sintering tem perature (1350°C) and short sintering times (<2 h). Figure 39 shows the effect
of the annealing time and the boron content on the microwave properties at 1050°C. The
results show th a t annealing results in significant improvements to the microwave properties.
The product of the quality factor and frequency (Q x f ) increases from approximately 21,000
to ~ 3 4 ,000 after 20 h annealing of the sample with 0.05 wt% boron at 1050°C. For the
sample with 0.5 wt% boron at the same condition, the Q x f product is significantly improved
to approximately 45,000. Figure 40 shows the influence of the duration of annealing on the
microwave properties at 1250°C. The Q x f product reaches >50,000 for samples annealed
at 1250°C for 40 h.
Further increases in the annealing tem peratures and times result
in degradation of the microwave properties. We note th a t a number of the physical and
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
76
50000 -
BCT with 0.5wt%B
45000 -
N
40000-
O
75
•JJ 35000 -
O
BCT with 0.1wt%B
30000 -
26000 -
0
20
40
60
80
100
120
Annealing time (hour)
Figure 40. Effect of annealing duration at 1250°C on the microwave properties.
chemical properties can be correlated with the microwave performance. For example, it
is interesting th a t the maximum Q x f product produced using 1250°C anneals for 40 h
corresponds to the conditions in which the secondary phases are found. The result shows
th a t the presence of the secondary phase can be correlated to improved microwave properties
of B a ( C d i / 3 T a 2 / 3 ) 0
3
. The correlation, as well as the anti-correlation, of the presence of
secondary phases and the magnitude of microwave loss has been observed in other ceramic
systems.
Of course this depends on the detailed chemical and microwave properties of
the secondary phases and the host material. In the system th a t we study, we note the
correlation but do not feel th a t we have sufficient information to reach strong conclusions.
The results in figure 39 and 40 also indicate th a t the use of larger am ounts of the boron
sintering aid significantly improves the Q x f product, compared w ith samples prepared with
lower boron oxide concentration. It is clear from figure 40 th a t the improvement might then
be attributed to the enhanced density of samples with a large boron oxide concentration.
It is well established th a t porosity in ceramics often leads to a degradation of microwave
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
77
properties [22], At this point the influence of ordering on the microwave properties is not
clear. Since all of the annealed samples exhibit ordering in the SAED results, it is difficult
to reach a strong conclusion about the influence of ordering on microwave loss in this
material system.
samples.
Work is underway to better quantify the degree of ordering in these
The dielectric constant and the tem perature coefficient of resonant frequency
for B a ( C d i / 3 T a 2 / 3 ) 0 3 doped with 0.5 wt% B 2 0 3 ceramics annealed at 1250°C for 40 h
were measured to be 32 and 80±15 ppm /°C \ respectively. No significant differences were
observed between the dielectric constant of samples annealed for different durations.
3.4. C onclu sion
The use of boron as a sintering aid reduces the sintering tem perature, enhances
the sintering density, and improves the microwave properties of B a ( C d i / 3 T a 2 / 3 ) 0 3 ceramic
dielectrics. Observations by transmission electron microscopy indicate th a t the liquid sinter­
ing mechanism contributes to the improvement in sintering density for boron concentrations
exceeding 0.5 wt%. No boron segregation and amorphous phase were observed along grain
boundaries. An amorphous phase rich in boron-oxide forms pockets partially penetrating
along multiple grain junctions. The enhancement of sintering behavior results from the
formation of boron-containing liquid phase in Ba{ Cdi / 3 T a 2 / 3 ) 0 3 ceramics with boron ad­
ditive. The introduction of as small as 0.01 wt% boron also results in the production of
high-density samples (~ 95%), presumably indicating th a t a point defect mechanism may
also play an im portant role in the sintering process. XRD combined with high-resolution
transmission electron microscopy images show th a t boron-doped B a ( C d i / 3 T a 2 / 3 ) 0 3 ceramic
material has a well-ordered hexagonal structure. Annealing treatm ent is found to improve
the microwave properties. The best sample has a dielectric constant of ~ 32, a tem perature
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
78
coefficient of resonant frequency of 80±15 ppm/°C', and a quality factor of >25,000 at 2
GHz.
Electron diffraction and high resolution transmission electron microscopy studies
showed also a well ordered structure of 1:2 with hexagonal symmetry for B ^ C d i / ^ T a ^ / ^ O ^
with boron additive sintered at 1300°C. No significant changes in ordered structures were
observed for B a ( C d i / :iT a 2 /:i)O3 subjected to a long-period annealing subsequently.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
C H APTER 4
DIELECTRIC-LOADED MOBIUS RESONATOR AND
STRUCTURAL, CHEMICAL AND DIELECTRIC
PROPERTIES OF CERAMIC INJECTION MOLDING
BARIUM ZINC TANTALATE MICROWAVE CERAMICS
4.1. R esearch background
4 .1 .1 .
C o n cep t o f M ob iu s reson ator. Bandpass filters can be realized from dual­
mode Mobius resonators with intrinsic transmission zeros h The Mobius resonator utilizes
a geometrical deformation of a transmission line to obtain a four-fold reduction in volume
[85, 5]. Mobius resonators are the result of projecting a transm ission line onto a Mobius
strip, which is the prototypical nonorientable surface. Although traditionally referred to as
one-sided surfaces, nonorientable surfaces are those for which the concept of left and right
are globally nonsensical [85].
A practical definition [86] of a Mobius strip is a one-sided surface th a t is constructed
from a rectangle by holding one end fixed, rotating the opposite end through 180°, and
applying it to the first end. The Mobius strip is probably the most common geometric
^ h e concept of Mobius resonator was proposed by J.M. Pond [85]
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
80
shape cited to illustrate topology [86].
Topologically, the Mobius strip traditionally has been referred to as a one-sided
surface [85] th a t possesses one edge.
Ignoring the sinusoidal patterns, Fig.
41(a) and
(b)illustrates the construction of cylinders from long rectangles, while Fig. 41(c) illustrates
the realization of a Mobius strip from a long rectangle. The cylinders resulting from Fig.
41(a) and (b) when connects to and connects to each contain two edges, A ’ and B ’. In
contrast, the Mobius strip created from Fig. 41(c) when A ’ connects to A and B ’ connects
to B possesses only one edge, A-B’-B-A. Although the length of the rectangles used to
construct the cylinder in Fig.
41(a) is twice th a t of the edge of the rectangle used to
construct the Mobius strip in Fig. 41(c), the length of the edge A-B’-B-A of the Mobius
strip is equal to the length of the edges of the cylinder.
4 .1 .2 . D ie le c tric a lly lo a d e d M o b iu s re s o n a to r . More compact bandpass filters
are realized using dielectric loading.
The use of dielectric ceramics with low loss, high
dielectric constants, and near-zero tem perature coefficient of resonant frequency is also
essential in achieving the required microwave performance for the devices of interest [1, 4],
Since the Mobius modes involve counter flowing currents in the wires, the electric field is
concentrated largely between the wires with little field interaction w ith the cavity wall.
Thus, by embedding the Mobius wire resonators in a dielectric, the circumference can, to
first order, be reduced by the square root of the relative dielectric constant. Hence, the
resonator volume is reduced by a factor equal to the relative dielectric constant.
4 .1 .3 . C e ra m ic in je c tio n m o ld in g (C IM ).
4.1.3.1.
The advantages of CIM technique. Rapid progress in electromagnetic simu­
lation methods has enabled a significant improvement in the performance of state-of-the-art
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
180
a)
B'->B
360
Phase (degrees)
A->A
b)
Phase (degrees)
B'->B
0 '- ) 0 (180° twist)
A'->A
180
c)
Phase (degrees)
Figure 41. W hen A ’ —+ A and B ’ —> B, the rectangles become (a) and (b) cylinders and (c)
a Mobius strip. The length of either of the two edges (A-A’ or B-B’ ) of the cylinder formed
in (a) is equal to the length of the single edge (A-B’ -B-A’ ) of the Mobius strip formed in
(c). As representations of transmission lines, the two orthogonal modes (solid and dashed
lines) are (a) at resonance for a ring-type resonator, (b) at antiresonance for a half-length
ring, and (c) at resonance for a half-length ring with a 180° twist (Mobius resonator), from
Ref. [85].
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
82
microwave devices. A growing number of these devices require intricately shaped microwave
materials. This has placed increasing demand on the ceramic research community to de­
velop a low-cost process th a t can manufacture high performance microwave materials in
complex-shapes. CIM provides special advantages. First, complex shapes are achievable in
ultra-hard materials th a t are difficult, costly, or impossible to make through traditional pro­
cesses. Features such as internal and external threads, knurls, organic curves, intersecting
and odd-shaped holes are possible. CIM is also able to combine multiple components that
previously required assembly. Second, CIM offers a high degree of reproducibility. Once
tooling has been established and the process fine-tuned, the CIM process is very repeatable.
The use of robotic p art handling and careful process monitoring furthers these capabilities.
Further, CIM is an additive process rather than subtractive.
Therefore in many cases,
adding complexity has minimal impact on cost. Finally, combining components also takes
advantage of the flexibility of CIM, reducing part count and assembly costs.
4.1.3.2.
CIM processing. As shown in figure 42, there are several key processes for
CIM technique: feedstock preparation, molding, debinding, sintering and densification.
Ceramic powder is selected along with careful attention to particle size, shape, and distri­
bution. Under carefully monitored conditions, these powders are blended with binders and
additives to enable the m ixture to flow. Specially tailored injection molding machines are
used to inject the feedstock into a mold, which is similar to plastic injection molding design.
However, the tooling requires very tight tolerances and wear-resistant components to stand
up to the abrasive powders being molded.
Then, the binders are removed via evaporation and exothermic reaction, leaving a
fraction of binder m aterial behind. Binder removal represents a difficult step in the ceramic
injection molding process. W hen using fine ceramic powders and preparing large bodies,
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
83
Feedstock
Preparation
Molding
Debinding
Sintering
Densification
Figure 42. Several key processes for ceramic injection mold technique.
the removal of binder is even more difficult. In most cases, binder removal puts limits on
the size and complexity of the shape of injection molded parts since w ith large body cross
section it is not possible to remove the binder in an acceptable period of time without
defects. Although a number of advanced methods for binder removal have been developed
(supercritical debinding [87], catalytic debinding [88], solvent debinding [89]) th a t speed up
this process and make it more effective, therm al removal of binder is still the most widely
used due to its universality and simplicity.
Finally, depending on the material, parts are next sintered in either an oxidizing
atmosphere, a reducing atmosphere, or in vacuum. Complete processing description and
discussion can be found in Ref. [90].
4.1.3.3.
Application. Several traditional ceramics have been successfully manufac­
tured in high volume and w ith great dimensional precision using ceramic injection molding
(CIM), including silicon nitride [17], zirconia [18, 19] and alumina [20, 21]. However, most
research to-date has focused on investigating the mechanical properties of ceramics parts,
with little attention paid to the electronic and dielectric properties. For example, Ni-doped
B a ( Z n i / s T a 2 /z)Oz is commonly used for many microwave m aterials as it exhibits excel­
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
84
lent high frequency properties with a high dielectric constant (er ~30), a low loss tangent
(< 2x10-5 at 2 GHz) and has a near-zero tem perature coefficient of resonant frequency,
tj
-
[6, 7]. For practical microwave devices, it is im portant th a t ceramic injection molded ceram­
ics have a sufficiently high dielectric constant, low loss tangent and a near zero tem perature
coefficient of resonant frequency. Our initial attem pts to produce parts w ith the complex­
shaped structure using conventional ceramic forming and machining technology failed as
a result of the intrinsic hardness and brittleness of B a ( Z n 1 / 3 T a 2 / 3 ) 0
3
. In this chapter,
we report (a) the initiate development of dielectric-loaded Mobius bandpass filters using
alumina and macor (b) the development of a ceramic injection moulding process th a t can
produce complex geometric structures from high-performance microwave dielectrics th a t
can further reduce the device size [i.e. Ba,(Zni/ 3 T 0-2/ 3)
t hr ough a collaborative effort
with Jeff Pond at NRL and (c) the microwave, structural and sintering properties of the
materials produced. Further, we explored the ceramic injection moulding capability by
synthesizing a recently proposed microwave high power filter th a t uses a dual-mode Mobius
geometry resonator to achieve a four-fold reduction in volume over conventional designs [5].
4.2. E x p erim en ta l p roced u re
4.2 .1 .
D ielectric-lo a d ed M ob iu s reson ator. A number of dielectric Mobius res­
onator test structures were initially fabricated using alumina ceramics, in conjunction with
small diameter gold or platinum metal wire and the machine milling and laser ablation
machining methods. A parallel loop geometry was employed since it can be accurately
fabricated and modeled. Alumina was chosen for initial testing because of its excellent
microwave performance, high therm al conductivity, and relatively large dielectric constant.
Further, for ease of fabrication during development, Macor ceramics was used for the di­
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
85
electric. Macor is an easily machinable glass ceramic dielectrics with a relative dielectric
constant of 5.6 at 8.5 GHz. The dielectric loss tangent has been reported to be ~ 7.1 to 10
at 8.5 GHz.
4 .2 .2 .
S y n th esis and ch aracterization o f B Z T ceram ics. B a ( Z n i/ 3 T a 2 / 3 ) 0 3
doped with N i and Z r was made from reagent grade BaCOs, Z n O , T a 2 0§, N i O and Z r 0 2
powders. The raw materials were blended by using Z r O <2 ball milling media and distilled
water for 16 hours to deagglomerate the powder and provide a homogeneous distribution
of raw powder.
The slurry was dried and then filtered through 14-mesh screen.
It is
subsequently heated to 1350°C for 10 h with an initial ram p of 100°C/h in air to form
single-phase B a ( Z n i / 3 T a 2/ s)Oz powder by solid state reaction. For the ceramic injection
molding process, low viscosity molded feedstock comprised of a wax-based binder, ceramic
powder and a lubricant is used. The binder is composed of paraffin wax, carnauba wax,
polyethylene, and stearic acid in a weight ratio of 70 : 10 : 19 : 1. Stearic acid serves as
the lubricant in this process. The B a { Z n i j 2,Ta2/^)Oz powder was loaded to 50vol% and
the ceramic feedstock was stirred at 130°C for 4 h to homogenize the m ixture of powder
and binder. The feedstock was subsequently fed into a plastic injection-moulding machine
(Gluco, type LP20, Jenison, MI, USA) to form the Mobius structure with 100MP pressure
at 180°C. The molded samples were therm ally debound in air w ith a ram p rate of 20°C/h up
to 700°C. Note th a t the organic binder is removed at m oderate-tem peratures (150 —250°C),
and it is im portant th a t the ram p rate be sufficiently slow to ensure th a t this process is
complete before exposing the sample to tem peratures at which the polymers decompose.
The debound samples were then successively sintered either at 1520°C, 1550°C, 1580°C,
1650°C, and 1680°C using a ram p rate of 300°C/h. In order to minimize Z n O evaporation
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
86
when sintering Bai^Znxj^Ta^j^O^ at very high tem perature (1680°C), the ceramic injection
molded samples were packed by powder of the same composition and the P t crucible was
covered with P t foil.
The structural properties of the samples were characterized using a Rigaku D/MAXIIB diffractometer equipped with single crystal graphite Cu K a monochromator. The mi­
crostructure of the samples was characterized using a Hitachi S —4200 scanning electron
microscope. SEM specimens were prepared for examination in a two-step process in which
they were first polished and then a thin gold film sputter deposited. The sintered density
of high quality samples was measured using the Archimedes method. The bulk density of
low-density samples was evaluated by measuring the dimension and weight of the specimen.
The microwave quality factor (Q ) and tem perature coefficient of resonant frequency
(t/ )
were measured using the TEqis mode of the dielectric resonator. The TEqis mode was
measured using S ii reflection d ata at the terminals of the one p ort cavity th a t is connected
to a vector network analyzer in transmission mode. The resonator was enclosed in an Aucoated test cavity with dimensions about three times larger than the sample size. Low-loss
supports were used to suspend the specimen in the center of the cavity. The quality factor
was measured in reflection at room tem perature. Microwave coupling was tuned to ~40 dB
loss by carefully adjusting the position of the electric probes. The tem perature coefficient
of resonant frequency r / at microwave frequencies was measured between 25 and 60°C.
4.3. R e s u lts a n d d isc u ssio n
4 .3 .1 .
D ie le c tric M o b iu s r e s o n a to r . Dielectric-loaded Mobius wire resonators
were fabricated using Macor, in conjunction with 0.25-mm-diameter gold wire. The struc­
tures consisted of two parallel Au loops recessed into 0.38-mm-deep circular grooves in the
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Macor ceramic. The loops are 0.74 cm in diameter and are separated by a distance of 0.31
cm. The Au wire was configured into the Mobius geometry by connecting the loops through
0.4-mm-diameter via holes. The cylindrical resonators have an overall diameter of 1.27 cm
and a thickness of 0.38 cm. A 0.22-cm-diameter hole was drilled through the cylindrical axis
of the Macor resonators through which a metallic rod could be press fit. Figure 43 shows a
cross-sectional representation of a Macor dual-mode Mobius resonator in a cylindrical cop­
per cavity. W ith the Macor-loaded Mobius wire resonator inserted into the copper cavity,
as shown in figure 43, the weakly coupled response was measured and is shown in figure
44. The dual modes and two transmission zeros are evident and are centered at 3.5 GHz.
Further discussion is beyond the topics of this thesis. However, complete analysis can be
read in Ref. [5] and [16].
4.3 .2 ,
X R D a n a ly sis a n d m ic r o s tr u c tu r e . Fig. 45(a) and (b) compares the X-
ray diffraction (XRD) patterns of B a { Z n i / ^ T a 2 /z) 0 ^ samples prepared with conventional
powder processing and ceramic injection molding methods. There is not significant struc­
tural difference between the two methods for ceramics sintered at 1520°C and 1680°C. These
two samples do not exhibit ordering, as indicated by absence of additional superstructure
peaks in the XRD patterns. Also, there is not evidence of any second phases th a t would
be produced from decomposition as a result of preferential Z n O evaporative loss during
sintering. This is im portant as significant levels of secondary phases can lead to enhanced
microwave loss.
Fig. 46 shows the dependence of density on the tem perature. The rate of heating
during sintering for all the samples was the same (5°C/m in). The low sintering density
th a t was observed after sintering at 1520°C for 48 hrs (72%) dem onstrated th a t ceramics
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
copper
cavity
plastic
spacer
coax
in
/
Mobius
wire
coax
Macor®
puck
is s
tuning screw
Figure 43. (top) Perspective drawing of the dielectric-loaded Mobius wire resonator showing
the Macor puck and the embedded gold wire with the cross over accomplished with two
vias. (bottom ) Edge view of a two-pole filter consisting of a dual-mode dielectric-loaded
Mobius wire resonator in a copper cavity.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
89
-20
-40
-60
-80
-100
2.5
2.0
3.0
3.5
4.0
5.0
4.5
Frequency (GHz)
Figure 44. Weakly coupled response of the dielectric-loaded Mobius wire resonator shown
in Fig. 43.
(a) Prrsssed samp le sintered at 1520°C
(110) <b) Ce ramie Injecl ion Molded sample
siiitered at 16 80°C
(2(
(101)
2)
(104) (22°)
I
(201)
<2: 1 ) , _
. 1
(21f>
L
J
L
(a)
1
,
■
15
■ I
20
,
25
i
I
30
i
i
35
i
,
40
A
(b)
i
i
45
i
i
50
i
i
55
i
i
60
i
i
65
i
i
70
i
75
20
Figure 45. X-Ray diffraction p attern of ceramic injection molded B a ( Z n 1 / 3 T a 2 /s ) 0 3 doping
Zr and Ni sintered at: (a) 1580°C and (b)1680°C.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
90
100
c o n v e n tio n a l p o w d e r p r o c e s s e d s a m p le
90
0s
'co
c
80
<D
Q
<D
■M
to
cu
01
70
c e ra m ic injection m o ld e d s a m p le
60
1520
1560
1600
1640
1680
Temperature (C)
Figure 46. Dependence of relative density on the sintering tem perature for ceramic injection
molded Z r and IVf-doped B a ( Z n i ^ T a 2 fs) 0 3 . The theoretical density is 7.94g/cm ~3.
injection moulded B a ( Z n i / ^ T a 2 / ^ ) 0 3 does not undergo significant densification under the
same conditions th a t conventional powder processed B a { Z n i / ^ T a 2 /^) 0 ^ samples do (i.e.
~96%). Surprisingly, a sintered density of only ~82% was found after th e ceramic injection
molded samples were sintered at 1650°C. Only when the sample was sintered at 1680°C does
the density achieve a respectable value of ~94%. Thus, high tem peratures and extended
sintering times had to be used to achieve high density with our B a{ Z n xj ^T a 2 j^) 0 'i ceramic
injection molding process.
Densification usually accompanies the elimination of porosity and grain growth dur­
ing high tem perature sintering. The structure and homogeneity of pores in the powder
compact strongly affects the sintering behavior [91].
Agglomerates, in particular, limit
the attainable green density, interfere with the development of m icrostructure, and impede
initial-stage sintering kinetics. This is especially characteristic of powder injection molded
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
91
samples, which have poorly dispersed powders in the powder and binder-mixing step [91, 92].
Previous investigations have clearly shown th a t large pores and pore clusters in the sample,
as a result of viscous binder flow at the debinding tem perature; result in microstructure
inhomogeneity, e.g the presence of agglomerate th a t subsequently affects the diffusion on
sintering, in the powder compact, which is often preserved in the final structure even after
completion of the sintering process [92, 93]. Such m icrostructure inhomogeneity can cause a
significant reduction in the driving force for densification [91, 94], We suggest th a t this is the
mechanism responsible for requiring a significantly higher sintering tem perature in ceramic
injection molding of B a ( Z n i ^ T a 2 /s)Oz than in conventional ceramic powder processing.
Fig. 47(a) is SEM micrograph of a ceramic injection molded sample sintered at
1520°C th a t shows th a t there exist a large number of pores with sizes ranging from ~2/_/m to
4/j.m. There is no evidence for densification outside of the small areas of attachm ent of neigh­
boring particles and a minor amount of grain coarsening (i.e. an increase in size from ~lyum
in the raw powder to ~5/rm ). In contrast, pores are not observed in 5 a (Z n 1/ 3T a 2/ 3)03
samples prepared using conventional powder processing techniques when sintered under the
same conditions (figure 47(b). Furthermore, the sample prepared using conventional pow­
der processing techniques sintered at lower tem peratures (1520°C) exhibits significant grain
growth (i.e. a grain size twice than th a t (~16/mi) of the ceramic injection molded sample
processed under the same conditions). Thus, the significant grain growth is an im portant
factor in the densification of pressed sample [95]. Figure 47(c) shows an SEM micrograph
of a ceramic injection molded sample with an average grain size of ~8^xm th a t was pre­
pared by sintering at 1680°C. Similar to the pressed samples, no obvious large pores are
visible, as might be expected given its high density. Thus, to achieve enough high density
ceramic injection molded samples; a very high sintering tem perature is required. Factors
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
92
Figure 47. SEM photom icrograph of a (a) ceramic injection molded Z r and iVi-doped
Ba(Zni/zTcL 2 /z) 0 z sintered at 1520°C; (b) Z r and iVi-doped Ba{Z rii /z T 0 ,2 / 3 ) 0 3 prepared
using conventional powder processing methods sintered at 1520°C; (c) ceramic injection
molded Z r and Ni-doped Ba (Zrii/?JT a2/3)0.3 sintered at 1680°C.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
93
th a t affect the kinetics of processing such as the sintering tim e and atmosphere must be
carefully controlled to minimize the loss of volatile Z n and its compounds. It might also be
possible to choose additives th a t facilitate densification and reduce the required sintering
tem perature as a result of the liquid sintering mechanism, as we have accomplished for
B a ( Z n 1 / 3 T a 2 /z)Oz, a similar perovskite compound [69].
D ielectric p ro p erties o f ceram ic in jectio n m o ld ed B a ( Z n i / 3 T a 2 /z)Oz-
4 .3 .3 .
Fig.
48 shows Q x f (quality factor times frequency) of ceramic injection molded
B a ( Z n i / z T a 2 /z ) 0
3
ceramics as a function of sintering tem peratures. It is noteworthy th a t
relatively high quality factors are achieved even in the absence of any direct evidence for
Z n / T a B-site ordering in the XRD patterns. The realization of high Q x f in the absence of
ordering is similar to th a t found for Zr-doped Ba(Zni/zTa, 2 /z ) 0
3
samples prepared by con­
ventional powder processing methods [7]. As noticed, the relationship between Q x f values
and sintered tem peratures reveals the same trend as th a t between density and tem perature.
The highest Q x f value corresponds to the highest sintering density.
The Q x f values of the sintered Ba{Zni/zTa, 2 /z)Oz ceramics ranged from ~ 1 6 ,000
to ~63,350. In contrast, the Z r and A i-doped Ba(Zni/zTa, 2 /z)Oz ceramics prepared by
conventional powder processing methods th a t were sintered at 1520°C for 48h shows higher
Q x f p ro d u cts(~ 9 5 ,000). Further investigation is necessary to clarify the difference in the
microwave performance between ceramic injection molded and pressed Ba(Zni/zTa, 2 /z)Oz
samples.
There is no obvious difference in the dielectric constant (~30) of Z r and iVi-doped
B a ( Z n i / z T a 2 /z)Oz samples prepared with conventional processing methods sintered at
1520°C and ceramic injection molded samples sintered at 1650°C for the same 48h.
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
94
70000
60000500004—
X
40000-
o
3000020000
-
10000
1520
1560
1600
1640
1680
Temperature (C)
Figure 48. Dependence of the Q x f product on the sintering tem perature for ceramic injec­
tion molded Z r and iVi-doped B a (Z n 1(/3T a 2/ 3)0 3 .
a.
CL
0 .8 0 .6 0.40 .2 -
Q.
0.0 -
0.2
1520
1560
1600
1640
1680
Temperature C
Figure 49. Dependence of the the tem perature coefficient of resonant frequency on the
sintering tem perature for ceramic injection molded Z r and Ni-doped S a ( Z n 1/ 3T a 2/ 3)0 3 .
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
95
The dependence of the tem perature coefficient of resonant frequency on the sintering
tem perature for ceramic injection molded Z r and Ni-doped B a ( Z n 1 / z T a 2 /z)Oz samples is
shown in Figure 49. A near zero tem perature coefficient of resonant frequency was achieved
for samples sintered at 1680°C for 48h. A maximum tem perature coefficient of resonant
frequency value of ~1.4 ppm /°C is observed after sintering at 1580°C. In contrast, a near­
zero (0.1ppm /°C) tem perature coefficient of resonant frequency for a B a ( Z n 1 /z T a 2 /z)Oz
sample with the same composition prepared using conventional powder processing could
be achieved. As noticed, the dependence of Tf on sintering tem perature did not show the
same trends as th a t of Q x f and density. Earlier experimental results have also found th a t
Tf can depend on synthesis param eters [96, 97]. For example in A B O z perovskites, r / can
be directly correlated to the nature and extent of B-site ordering, as reported in a study
of B a xS r i - x { Z n i/ z N b 2 /z)Oz [20]. As the order-disorder transition in B a { Z n 1 /z T a 2 /z)Oz
(1500°C-1625°C) [98] falls in the range of the sintering tem peratures used in our study,
the changes in t / with sintering tem perature may be attrib u ted to variation in the type
and extent of ordering. Note th a t large changes in the degree of ordering as monitored
by X-ray diffraction in our study were not found. This technique is, however, known to
have limited sensitivity for detecting this effect, especially since the extent of ordering and
the total volume of the ordered phases may be small. Earlier work by Davies et. al. had
reported ordering in regions at and near the grain boundary in B a( Z r iii zT a 2 /z) 0 z which
would be difficult to detect with X-ray diffraction [14]. The presence of secondary phases
can also significantly influence
Tf .
We did not find evidence for secondary phases in XRD
resulting from the evaporation of zinc during the elevated sintered tem perature (1680°C)
used in this study. However, we do need to note th a t the secondary phases th a t constitute
a small volume fraction, are randomly oriented an d /o r are possibly defective in nature can
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
96
also be very difficult to detect.
Although the physical mechanism(s) responsible for changes in r / are just being
uncovered, a number of workers have shown th a t variations can be correlated with structural
param eters [33, 99, 24]. For example, Colla reported an empirical correlation between t/
and the extent of octahedral tilting in perovskites [33]. It is clear th a t to uncover the
fundamental issues related to r / , further investigation is needed.
Sample number
1
2
Resonant frequency (GHz)
1.251
1.258
Q
928
954
Table 6. Resonant frequency and Q of the dielectric-loaded Mobius wire resonator. Note
th a t the Q is smaller than anticipated from the dielectric alone as a result of losses in the
gold wire.
4 .3 .4 .
D ie le c tric -lo a d e d
M o b iu s b a n d p a s s filte r. A number of dielectric
Mobius resonator test structures were fabricated using injection molded B a ( Z n i / 3 Tci 2 /z)Oz
ceramics, in conjunction w ith small diameter gold metal wire. A parallel loop geometry was
employed since it can be accurately fabricated and modeled. The structures consisted of
two parallel Au loops recessed into circular grooves in the B a ( Z n \ / z T a 2 /z)Oz ceramics.
The Au wire was configured into the Mobius geometry by connecting the loops via holes. A
hole was drilled through the cylindrical axis of the B a ( Z n i / 3 T a 2 /z)Oz resonators through
which an alum ina ceramics rod could be press fit. Figure 50 shows a dual-mode Mobius res­
onators fabricated from Z r and ./Vi-doped B a ( Z n l/'i T a 2 /z)Oz prepared by ceramic injection
moulding. Figure 51 shows a schematic drawing of the Mobius wire resonator structure. The
structures consisted of two parallel A u loops recessed into grooves in the Ba(Z rii/ 3 T 02/ 3)0 3 .
W ith the R a (Z n 1/ 3T a 2/ 3)0 3 -loaded Mobius wire resonator inserted into the copper cavity,
the weakly coupled microwave response was measured. Two m atched Ba(Zni/zTa, 2 /z)Oz-
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
97
Figure 50. Black and white photograph of a Mobius resonator synthesized from Z r and
N i -doped B a ( Z n i / ^ T <3.2/ 3)03 using ceramic injection molding. The sample itself has a light
yellow color.
Figure 51. Schematic drawing of the dielectric-loaded Mobius wire resonator illustrating
the geometry of the dual mode Mobius resonator.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
98
loaded Mobius wire resonators were fabricated during the same sintering run.
Table 6
shows th a t they have very similar microwave properties. A small difference in center fre­
quency is found and probably arises from differences in the air gaps between the wire and
the groove due to the imperfect fit. This is not surprising since the recessed gold wires
were inserted into the grooves manually. A four-pole bandpass filter was also realized using
two cascaded dual-mode dielectric-loaded Mobius resonators. The dielectric-loaded Mobius
wire resonators were nominally identical to the one described previously. The separation
between resonators was held fixed by the alumina spacer.
4.4. C onclu sion s
In this chapter, the development of a CIM process to produce complex-shaped struc­
tures using high-performance microwave ceramic materials was reported. In particular, we
describe the synthesis methods and the structural, chemical and dielectric properties of
B a { Z n i/^ T a 2 /z)Oz doped with Z r and N i (BZT) ceramics produced using CIM. Sintering
the Ceramic Injection Molded Ba(Zn-i/ 3 Ta,2 /z)Oz to a relative density of ~94% was possible
at a tem perature of 1680°C and a time of 48 hrs. The best samples to date exhibit a er of
~30, a Q value of ~31250 at 2 GHz, and a r / of 0.1 ppm /°C .
The incorporation of dielectric loading of Mobius wire resonators was successfully
demonstrated.
W ith this approach, a very compact resonator volume can be achieved
since the size reduction due to dielectric loading and the Mobius resonator concept are
synergistic. Besides the resonators fabricated by machinable Macor, a number of dielectric
Mobius resonator test structures were fabricated using injection molded B a ( Z n 1 /z T a 2 /z)Oz
ceramics. The weakly coupled microwave response was measured. These results indicate
th a t these technologies can be used to develop compact high-performance filters.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
CH A PTER 5
DIELECTRIC PROPERTIES AND I-V
CHARACTERISTICS OF VANADIUM AND SCANDIUM
DOPED BARIUM AND STRONTIUM TITANATE
UNDER HIGH DC BIAS VOLTAGE
5.1. In tro d u ctio n
Ferroelectric ceramics have been used in a wide variety of applications in the elec­
tronics industry. The most experimentally and theoretically studied systems are barium and
strontium titan ate and their alloys because they are the prototype ferroelectric and very im­
portant materials in th e capacitor industry. For example, B a T iO s based ferroelectrics are
used to produce dielectric capacitor, particular in the form of MCCs (Multilayer Ceramic
Capacitor) and P T C R (positive tem perature coefficient resistors).
These m aterial’s main disadvantage is the strong tem perature dependence of its
dielectric constant near the Curie tem perature. They must be modified chemically and
physically to produce the required properties. As a result of the inherently strong tem pera­
ture dependence of ferroelectrics, dopants are added to modify this characteristic to achieve
improved stability. These desired characteristics can be achieved through cation alloying
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
100
to shift the Curie tem perature, control of grain size or m icrostructural modification of the
grain core/shell structure through doping with impurities such as cerium. The resulting
chemical inhomogeneity features a ferroelectric grain core and a paraelectric shell. The
flat high-dielectric-constant tem perature characteristics, as achieved in the XR7 ceramics,
can be related to the high dielectric constant of the grain core on the high-tem perature
regions and the grain shell in the low-temperature region. X7R specifications demand no
change in dielectric constant of more than +15% or -15% over tem perature range —55°C to
125°C. The dopants used to achieve the reduced tem perature dependence of the properties
can result in undesired leakage currents, particularly during high field bias conditions. The
donor-doped B a T iO z leads to the formation of semiconducting behavior when prepared by
sintered under a low oxygen partial pressure since electron com pensation is favored under
these conditions. Several studies show th a t the oxygen pressure also has a pronounced influ­
ence on grain growth for donor-doped B a T iO z . An im portant application of donor doped
B aTiO z in industry is to make positive tem perature coefficient resistors (PTCR). W ithin
the materials used in the devices, ceramic grain boundaries are formed in a specific way so
th a t the state of polarization controls the large transport across Schottky barriers.
B aT iO z and S rT iO z also exhibit a number of other undesirable characteristics in­
cluding large dissipative loss and uncontrolled electrical breakdown at m oderate electric
fields (3kV/ m m ). Especially, degradation of the insulation nature is found to occur when
this material is exposed to prolonged dc field and tem perature stress [100] - [103]. BaTiO z
also has another lim itation as its perm ittivity decreases and dielectric loss increases under
high dc fields. Thus, the high field dielectric properties of barium and strontium titanate
and its I-V characteristics are im portant factors for evaluating the utility of this material
for many applications.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
101
There are several factors th a t affect the dielectric properties and I-V characteristics
of barium and strontium titan ate ferroelectrics. For example, the dielectric polarization is
not a linear function of the applied electric field strength. A phenomenological equation
based on the Devonshire’s theory can be used to describe the reduction of dielectric constant
under applied dc bias [50, 115].
While many studies have focused on the tem perature-dependence of dielectric prop­
erties and the effect of dopants on the dielectric properties and grain growth for barium and
strontium titan ate [109, 112, 114, 123], only a few studies have investigated the relationship
between the dielectric properties, I-V characteristics and dc field [100] - [103].
All of these studies mentioned above focus on the undoped or acceptor doped
B a T iO z . To our knowledge, there have been very few investigation of the influence of
large dc voltage on the dielectric properties of donor doped B a T iO z . Despite the fact th a t
work has been done on Sc and V doping in B aT iO z and SrT iO z, these dopants might be
considered as prototype defects since they differ by one atomic number from Ti. A recent
study has simulated the electronic structure and properties of ScTi and VTi defects and
concludes the substitution of V for Ti is n-type doping and substitution of Sc for Ti is
p-type doping. The hole doping yields a larger density of states at the Fermi level than
the electron doping ??. The goal of this chapter is to investigate the influence of high dc
field on the dielectric properties and I-V characteristics of scandium (acceptor) an d /o r vana­
dium (donor) doped B a T iO z and Bao^Sro ^TiOz. We also study the influence of vanadium
and/or scandium on the crystal structure of B aT iO z and Bao^Sro.zTiOz-
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
102
5.2. E x p erim en ta l p roced u re
The starting raw powders are B a T i O z , S r T i O z , V2O5, and Sc^Oz- The composition
prepared is B a o j S r o . z T i O z . To produce different net donors, Sc^Oz was mixed with V2O5,
B a T i O z and S r T i O z to get the compound with composition BaQjSr0.zTi0.98-xV0.02Oz,
where x is 0, 0.01, 0.015, 0.0175, and 0.02.
To produce different net acceptors,
V2O5 is mixed with ScjOz and B a o j S r o .z T iO z to get the compound w ith composition
Bao. 7Sro.zTio.98-xSco.o2VxOz, where x is 0, 0.01, 0.015, 0.0175, and 0.02. Then, the ob­
tained powders were milled by using Z r 0
2
ball milling media and distilled water for 16 hrs
in a milling machine to provide a homogeneous distribution of raw powders. The slurry was
subsequently dried. Samples were pressed to 60% of theoretical density. These samples
were sintered in alum ina crucible at 1350°C for 2 hrs with an initial ram p rate of 300°C/h
in air. After sintering, the samples were cooled slowly. No additional annealing treatm ents
were used in this investigation.
The crystal structure of the ceramic samples were characterized using a Rigaku
D/M AX-IIB diffractometer. A single crystal graphite monochromator was used to attain
Cu K a radiation. The I-V (current-voltage) and C-V (capacitance-voltage) measurement
was performed using a Q uadtech 7400 Precision RCL and Quadtech 1865 M egohm m eter/IR
tester. Gold thin film electrodes were deposited on the surfaces of the as-sintered samples.
The following equation:
- S
(51)
was used to infer the relative dielectric constant (er), where C is the measured capacitance,
d is the sample thickness, A is the contact area, and e 0 is the perm ittivity of free space
(8.8542xl0“ 12F /m ). Scanning electron microscopy was used to examine the microstructure
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
103
of samples.
5.3. R esu lts and d iscu ssion
5.3 .1 .
X -R a y p a ttern s and cry sta l stru ctu re. Fig. 52 (a) and (b) shows XRD
patterns of V-doped B a T iO z and F-doped Bao, 7 Sro,zTiOz sintered at 1350°C for 2 hrs
respectively. Fig. 53 (a) and (b) shows XRD patterns of 5c-doped B a T iO z and F-doped
Bao. 7 Sro.zTio.Q8 -xSco,o 2 0 z sintered at 1350°C for 2 hrs respectively. The XRD spectra of
B aTiO z and B a o ^ S ro zT iO z doped with F2O5 indicate th a t these ceramics have the same
tetragonal structure as their undoped counterparts. There is evidence of a small amount of
secondary phases present in samples doped with 0.5 and 1 wt% V2 O 5 . This indicates th a t
there is a vanadium solid solubility limit of less than 0.5% in B a T iO z and Baa. 7 Sro,zTiOzIn contrast, a small am ount Sc doping (0.5 at%) changes the regular structure of
B a ().7 S r 0,zTiOz from tetragonal to rhombohedral structure (see Fig. 53). The Bragg reflec­
tions are indexed to the tetragonal crystal structure for below 0.5% F-doped Bao jS r u z T iO z
and the others are indexed to rhombohedral structure in Fig. 53(a).
The difference in ionic
radius between Sc and Ti is 30% larger th an th a t between Va and Ti (The Ahrens ionic
radii for F 5+, T i i+ and S c3+ with VI coordinate number are 0.59A,
0 .6 8 A ,
and
0 .8 lA
[17]).
Thus, the significant difference in crystal structure results from the ionic radius difference.
The electron configuration ({Arr}d(y) of Sc3+ and V 5+ is the same when these two atoms are
fully ionized . In this case, there does not exist Jahn-Teller effect (JTE) proposed in Ref.
[119] and described in Ref. [120] to explain the cubic-hexagonal phase transform ation for
BaTiO z resulting from the lattice structure distortion.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
104
S in te r e d a t 1 3 0 0 ° C fo r 2 h rs
(1 1 0 )
(1 1 1 )
(2 1 1 )
( 2 20 )
<f)iwt%v2o5
( J(2 2 1 )(3 1 0 )
(1 0 0 )
A.
i
:.
1
. f i *
...............J
.
A ....i ..... ,
,
,
L
i__
. . A ...JL ....
A .
.- .
(e) 0.5wt%V2O5
JV
-A ..
Jw
Vtt
1
i
“ ," T «
|
|
L
(c) 0.15wt%V2O
.... i \
A . .ill
k
(b)0.1wj%V2O5
k
(a) 0,05wt%V2C
A .
A ...................
k
j
i
. . . ........
A ...... ....
20
30
j .
L
F
T
40
j
L
50
60
70
80
2 th e ta
(a)
(a) Ba,7Sr03TiO3
(c) Ba07Sr03TiO3-0.2%V2O5
(d) Ba07Sr03TiO3-0.5%V2O5
(e>Ba».7Sro,Ti03-1%V2°5
(f) Ba„Sr0.,TiO3-2%V2O5
(111) (200)
2 1 1)
(220 )
$
<
C/)
o
c
A .
A
j\
xM
T
70
80
T h e ta (d e g r e e )
Figure 52. X-Ray Diffraction of (a) F-doped B aT iO z and (b) F -doped Bao.jSro.zTiOz■
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
105
“
1
20
I
'
30
I
'
I
40
50
'
1“
60
Theta (Degree)
(a)
(a)Ba0.7Sr0.3TIO3-2%SC2O3
(b) Ba0.7S''0.3TiO3-2%SC3O3-1%V3O5
( c ) Ba07Sr03TIO3-2%Sc2O3-1.5%V2O5
(d) Ba„7Sr03TI03-2%Sc20 3-1 .75%V20„
(e) Ba„,Sr„ ,TiO-2%Sc.O -2%V O
( 110 )
A(001)
J
(221 ) ( 1 12 )
wA-VvWaM*-/
J l
~r20
30
“I--
—r~
- r~
40
50
60
Theta (Degree)
(b)
Figure 53.
X-Ray Diffraction
BaQ.TSro.zTio.szScamOz-
of
(a)
S’c-doped
B a T iO z
and
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
(b)
F-doped
106
5.3.2. M icro stru ctu re and sin terin g den sity. Fig. 54(a)-(d) show SEM pic­
tures for B aT iO z doped with different V2 O 5 concentrations. A small am ount (0.1wt%) of
V2O5 doping significantly reduces the sample’s grain size. However, significant grain growth
is observed for B a T iO z doped with 0.5wt% V2O5. There is no further grain growth observed
in B aTiO z doped with lw t% V2 O 5 . At this concentration, the grain growth is apparently
restrained by the existence of secondary phases. This appears to be the case in our study as
we observed additional secondary phases in Fig. 52(a).
Fig. 55(a) shows the dependence
of sintering density on the V2 O 5 . Maximum sintering density for B a T iO z is seen at 0.1wt%
vanadium doping.
5 .3 .3 . D ielectric p ro p erties. Fig. 55(b) shows the dependence of the dielectric
constant of B a T iO z on vanadium concentration. The dielectric constant is enhanced with
0.1wt% V2O5 doping. At this doping level, secondary phases are not found. The increase
in dielectric constant might be attributed to the significant reduction in the grain size as
observed from SEM micrographs. As the grain size decreases the domains become smaller;
the domain w idth is roughly proportional to the square root of the grain size. The number
of domains per grain therefore decrease as the square root of grain size. So the smaller grain
results in the larger unrelieved stress within the grain. The increase in stress is accompanied
by an increase in permittivity. In addition to the direct effect of stress, a reduction in the
90° domain w idth can enhance perm ittivity because the domain wall area per unit volume
of ceramic increases [129]. The decrease of dielectric constant of samples doped with higher
V2 O 5 might result from the appearance of secondary phases as indicated in Fig. 52 (b).
Figs. 56(a) and (b) indicate the dependence of dielectric constant of Bao^Sro^T iO z
on the tem perature. The Curie tem peratures of both samples for S c and V decrease mono-
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
107
Spin
5pm
Figure 54. SEM pictures for BTO doped with different V2 O 5 concentrations: (a) B a T iO 3;
(b) 0.1wt% V^Os-doped BaTiOz',{c) 0.5wt% V^Os-doped BaTiOz',(d) lw t% V^Os-doped
B a T iO z .
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
108
5.825.80
5.78
5.76
O) 5.72
c
5.70
5.68
5.66
5.64
5.62
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
V2Os In composition (wt%)
(a)
3500340033005(0 3200? 3100o
O 30002900o 2800“
2700-
|
2600-
a ; 250024002300
0.0
0.2
0.4
0.6
0.8
1.0
V20 5 in Composition(wt%)
(b)
Figure 55. Dependence of B a T iO 3 (a) sintering density;(b) dielectric constant on V2 O 5
concentrations.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
109
tonically with Sc and F . V doping of 0.1% results in maximum Tc and er . In contrast,
for Bao.-jSrQ.zTiOz doped with Sc, both the Curie tem peratures and dielectric constant
decrease monotonically w ith Sc and F concentration. Figs. 57 shows the dependence of
dielectric constant on th e applied dc biasing for undoped and F-doped jBao.7SVo.3T iOz
sample. There exist two regions in Fig. 57. At low electrical field, the dielectric constant of
B a o jS ro ^ T iO s increases with applied electrical field. Then, the dielectric constant reduces
continuously w ith the applied electrical field.
However, as shown in Fig. 57, the dependence of the dielectric constant on the
dc bias is obvious only for B a o ^ S r o ^ T iO z doped with 2 wt% V2O5.
Usually a broad
ferroelectric-to-paraelectric transition occurs for samples containing strontium . The dielec­
tric tem perature measurements for F-doped and S'c-doped samples show th a t the phase
transition tem peratures for B a o ^ S r zz T iO z doped with vanadium decrease monotonically
with vanadium. W hen the doping concentration of V2O5 reaches 2 wt%, the phase tran ­
sition tem perature is about 18°C. Thus, at the measurement tem perature of 22°C, the
crystal structure of B a o ^ S r o .z T iO z doped with 2 wt% V2O5 is close to cubic and exhibits
the paraelectric state. A strong dependence of dielectric constant on applied electric field is
observed, as predicted by Equation 1.12. This equation is based on the Devonshire’s theory
assuming th a t the ferroelectrics is in the paraelectric state. For other Bao,-jSrQ,zTiOz doped
with V2O5 samples, their phase transition tem peratures are higher th an the measurement
tem perature This shows the samples are in the ferroelectric states and the dependence of
dielectric constant on the applied electric field is not expected to be as large.
5 .3 .4 .
I - V c h a ra c te r is tic s fo r F - d o p e d B ao.iS rQ ^T iO z- I-V characteristics for
F-doped Bao. 7 Sro.zTiOz are shown in Fig.
58.
Fig.
59 shows I-V characteristics for
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
110
180 0 0
--B
b\, s'.,ticU>'1°av!0.
a07Sr0aTiOJ_0.2%
VJO(
~Ba>V,Ti°s-1%V!°!
-Ba.,Srn,Ti0.2%V,0,
160001400012000<0
tn
c
10000-
O
o
8000-
<D
<D
o
600040002000-
—I—
— 1—
20
40
T e m p e ra ru e (C°)
(a)
Bao.7SrojTIOJ_0.25%ScJOf
■Ba^Sr,TIO0.5%ScO.
Sr.TiO. 1%
Sc„0
Sr.JIO 2%Sc.O
O
2500
2000-
1500-
T e m p e ra tu re (°C)
(b)
Figure 56.
Dependence of dielectric constant on the tem perature
Bao.iSro.sTiOz and (b) 5c-doped Bao^Sro^TiO z.
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
(a) F-doped
I ll
8200
8000
• B a0 7S r03T i0 3 doping
- Ba07S r03TiO3 doping
- B a07S r03TiO3 doping
- Ba07S r03TiO3 doping
7800
7600
7400
with
with
with
with
0.1 wt%V20 5
0.5wt%V2O 5
1wt%V20 5
2wt%V20 5
7200
7000
6800
*****
6600
6400
6200
6000
5800
\ r
5600
Electric Field (V/cm)
Figure 57. Dependence of dielectric constant on the applied dc biasing for undoped and
F-doped Bao.iSra,zTiOz sample.
Sb-doped BaojSro,zTiOzAs expected, for all undoped, F-doped and Ac-doped B a o jS ro zTiOz samples, the
leakage current increases with applied voltage. For ferroelectric single crystals, assuming
th a t trapping is negligible and th a t the equilibrium carriers are zero, the current and voltage
is related by:
I ~ Va
(5.2)
where a equals 1.2 for strong electrical field and 2 for weak field. This derivation of this
equation can be seen in Ref. [109] - [110]. For polycrystal ferroelectrics, I-V characteristics
are more complicated due to the appearance of deep trap states at the defected grain
boundaries. This influences the behavior of the electrons near the surface. Below an onset
voltage, quasi-ohmic current flows. For higher voltage, the current and voltage are related
by I ~ V m . W hen m is ~ 2, the current is space charge limited. In this case, the carrier
density injected from the electrodes exceeds the native bulk carrier concentration. There
exist three I-V characteristic regions for Bao. 7 Sro,zTiOz doped with V2 O 5 : Ohmic current,
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
3
200 H
Region(lll)
' Region II
Region I
-i—>—i—<
—i—<
—i—•—i—<
— i—■
—i—■
—i—
2000
4000
6000
8000
10000
12000
14000
16000
B ia sin g F ield (V /cm )
Figure 58. I-V characteristics for undoped and V-doped B clqj Sro.zTiOz sample at room
tem perature.
c
<D
i_
3
o
u
3)
re
re
Q)
1.80E-04
1.60E-04
1.40E-04
1.20E-04
1.00E-04
8.00E-05
6.00E-05
4.00E-05
2.00E-05
0.00E+00
♦ B7S3TO
mB7S3TO+O.25at% Sc203
B7S3TO+O.5at% Sc203
XB7S3T0+1at%Sc203
200
400
600
Applied Voltage (V)
Figure 59. I-V characteristics for .Sc-doped B a o^S ro^T iO s sample at room tem perature.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
113
space charge limited current, and trap filled limit region (Fig. 58). As shown, the onsets
of traps-filled-limited current shift to higher electrical field for samples doped with higher
V2 O 5 concentration.
The onset of the trap-filled-limit curve results when there are already unneutralized
charges in the traps th a t prevent the injection of additional electrons from the cathode.
The onset voltage can be expressed as:
v™
=
(53)
where, Nt is the to tal trap density necessary to overcome this repulsion [22], Thus, the
onset voltage can be used to derive the trap density as described below. The calculated
total trap density determ ined using d ata in Fig. 56(a) and Fig. 58 and equation 5.3 is listed
in Table 7. Table 7 indicates th a t the trap density increases w ith vanadium concentration.
Although the onset to the trap-filled-limited current increase monotonically w ith increasing
V concentration, a drop in the calculation of trap density for 0.2 wt% V2O5 doped sample
is found since the dielectric constant is much smaller at this measurement tem perature.
Another feature th a t should be noted is th a t doped vanadium shifts the onset regions of the
Sample
Baz.iSrQ.zTiOz
BaQ -jSra,zTiOz
BaQ.fSrQ,zTiOz
Baz,iSrQ,zTiOz
doped
doped
doped
doped
w ith
w ith
with
w ith
0.1%V2O5
0.2%V2C>5
1%V205
290V2O5
Total
3.4 x
2.9 x
4.5 x
7.2 x
Trap Density (cm 3)
102U
1020
1020
1020
Table 7. Calculated total trap density for V-doped B a ^ jS r o ^ T iO z -
nonlinear curves to higher applied voltage region and reduces the leakage current. This can
be explained viewed from point defect chemistry. It is well known th a t there is a transition
from semiconductor to insulator as the concentration of donor is increased when donordoped B aT iO z is equilibrated with an oxidizing atmosphere. One example is Nb-doped
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
114
BaTiO^- As Nb concentration increases, the material becomes insulator showing the excess
oxygen is retained within the lattice. Thus, doping donor will not attrib u te the electrons
to BaTiOz- This also might happen to V-doped B aT iO z since V has a higher valence than
Ti.
In contrast, a different type of I-V behavior is found for S'c-doped BaQ.iSro.zTiOz
samples. Fig. 59 shows th a t the leakage current increases continuously with scandium in
the range from 0 at% to 0.5 at%. For these doping concentrations, there are no crystal
structure changes as illustrated in Fig. 53. Again, viewed from th e point defect chemistry:
2B a O + S c 2 0 3 -» 2B a Ba + V ? + 50 0 + 2S cTi
(5.4)
Thus, doping Sc increases the leakage current presumably as a result of the enhanced
concentration of oxygen vacancies. However, the leakage current decreases when the doping
concentration is 1 at%. Combining with XRD in Fig. 53(b), the transition in the leakage
current corresponds to the crystal structure change and the associated decrease in the
dielectric constant (see Fig. 56(b)). A significant influence of crystal structure on the leakage
current is found. Fig. 60 shows the dependence of leakage current on the tem perature for
Sc-doped Bao.fSro.zTiOz measured from room tem perature to 200°C. As expected, the
leakage currents increase w ith increasing tem perature. There exits two regions in these
curves: Ohmic at low applied voltage range followed by space charge current at higher
applied voltage region. As an example, Fig. 61 shows the typical log (V) log (A) dependence
for sample measured at 140° C.
It is worthy noting in Fig. 60 th a t the applied voltages corresponding to the onsets of
the nonlinear region in the I-V curves shift to lower values as the tem perature increases and
the trap filled limit region disappears at high measurement tem perature. The disappearance
of trap-filled-limited region shows th a t at the higher measurement tem peratures the density
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
115
9.00E-04
8.00E-04
7.00E-04
c 6.00E-04
£l.
3 5.00E-04
o 4.00E-04
0
o>
0 3.00E-04 j
00 2.00E-04
1.00E-04
O.OOE+OO
♦ 25C
80C
x 100C
x 120C
• 140C
+ 160C
-2 0 0
1000
A pplied V o ltag e (V)
Figure 60. Tem perature dependence of I-V curves for 0.5 at% Sc-doped B ao^SrozT iO z.
of therm al excited electrons is enhanced, increasing the m agnitude of the ohmic current.
Thus, the high level injection condition required to enter the space charge limited region
is moved to higher voltage and current levels and only the ohmic region is observed in
these structures at high tem perature. Presumably, at higher power conditions, the devices
would eventually reach the conditions where the unneutralized defects are filled th a t prevent
injection from the cathode and the SCLC electrical properties would be observable.
5.3 .5 . D ep en d e n c e o f d ielectric co n sta n t on th e frequency. Fig. 62(a) and
(b) show the dependence of dielectric constant on the frequency for V-doped BaQ^SrQ.zTiOz
and Sc-doped B ao ^S ro^T iO z respectively.
Fig.
63(a) and (b) indicate the depen­
dence of dielectric constant on the frequency for Sc doped ilao.7SVo.3Tzo.98Vb.02O3 and
V-doped Tao.7SVo.3Tzo.98Sco.02O3 respectively. No significant dependence of the dielec­
tric constant on the frequency is observed for Sc-doped B a o ^S ra ^T iO z and V-doped
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
116
1.00E+00
1.00E-01
_ 1.00E-02
i
1.00E-03
= 1.00E-04
| 1.00E-05
1.00E-06
g> 1.00E-07
re 1.00E-08
•3 1.00E-09
1.00E-10
1.00E-11
1
10
100
1000
A p p lied V o lta g e (V)
Figure 61. Typical log (V) log(I) curve for 0.5 at% Sc-doped B a o ^S ro ^T iO z sample mea­
sured at 140°C showing Ohmic and space- limited-charge current.
£ao.75Vo.3Tio.98K).0203- In contrast, significant dependence of the dielectric constant on the
frequency is observed for F-doped B ao^SrozT iO z and F-doped BO fnSro^Tio ^ S
samples. There are several factors th a t might influence the dielectric properties of ferroelectrics, such as, crystal structure, domain wall structure, grain size, and grain core/shell
microstructure when present. In our case, we do observe the change of grain size as described
in previous section for V-doped BaT iO z samples, but the similar frequency dependence for
all V-doped samples is seen in Fig. 62(a). The same phenomena happens to Sc-doped
BaTiO z samples where doping Sc can change the grain size of B a T iO z , but the similar fre­
quency dependence for all Sc-doped BaT iO z is observed, as shown in Fig. 62(b). While it
is reported th a t th e cerium-doped regions are closely related to grain size [113], it seems the
grain size and cerium-doped regions will not affect significantly / —e curves except th a t they
might change the value of dielectric constant. Previous simulation results by Monte-Carlo
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
117
model have shown th a t the motion of domain walls proceeds by nucleation and growth of
new nuclei onto an existing wall [136]. The dependence of the capacitance on frequency
might be due to the time needed for new nucleation processes to occur, thereby limiting the
domain wall contributions to the perm ittivity at high frequencies. Further experimental
and theoretical investigation is necessary to clarify the difference.
5.4. C onclusions
Doping Sc into Bao.ySro,zTiOz changes its crystal structure from tetragonal to
rhombohedral and significantly reduces the dielectric constant of Bao^Sro.zTiOz- In con­
trast, doping V into B a T iO z and B a o ^S r^^T iO z does not change their crystal structure.
The leakage current can be reduced by doping V2 O 5 . No significant dependence of the
dielectric constant on the frequency is observed for Sc-doped Baa^Sro^T iO z- In contrast,
significant dependence of the dielectric constant on the frequency is observed for F-doped
Bao,^Sro,zTiOz and V-doped BaQ.iSro.zTio.szScQmOz samples.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
118
16000
c
-Ba.,Sr TIO0.1%V,O.
14000
0.3 y— 5 5
*2
g
12000
O
o
j£ 10000
<x>
N,.
b
CD
•3
8000
6000
-4 -A -
4 -A --A -
a
A - A A.
A -A -A -A -A -A -A -
4000
10
100
1000
10000
100000
Frequency (Hz)
(a)
14000
13000
12000
11000
a
w
0C
10000
8000
1
7000
<D
5000
<D
S
-undopedBa^Si^TlOj
_Ba„,Sr.,TiO. 0.25%Sc,Oe
-B‘>t.7Sr„T|0 _1%Sc!Oi
-bS.^ut,03-2%Sc.°.
9000
°
6000
'• S
4000
a?
cc
3000
' 1
* ------* ------A------a ------A------A------A------A------A------A------A------A------A------A
A
2000
;[ ' 1
1000
0
1 : 1
1000
J r ; . - - ; j 10000
i n i
100000
Frequency (Hz)
(b)
Figure 62. Dependence of dielectric constant on the frequency at room tem perature (a)
F-doped Bao. 7 SrQ,zTiOz; (b) Sc-doped Bao -jSro.zTiOz.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
119
3500-
-Ba07Sr0.3TiO3-2at%VA
-- Ba„
Bai7,SSr0
lT,O^2at%V2lV 1al%SCA
r , TiO 2 at% V o 1 .5 a1 % S c .O ,
3000-
- Ba„ ,S r . ,T iO
2 at% V O
2 a t % S c ,0 ,
2500-
O
2000-
0<D
1
1500-
b
1000-
i
t
1000
t
10000
t tt= t
100000
Frequency (Hz)
(a)
1600-
- Bao,Sr»3Ti(V2al%ScA
- B ao ,,S ''„ T i ° ^ 2 a t % S c I0 ^ 1 a t% v 2° s
- B a0;S r0 3TiO J_ 2 a t% S c aO j_1,5al% V 2O s
- B a 0JS r0JT1OJ_ 2 a t% S c JO3_1 .7 5 at% V 2O !
1400-
- B anTS r ' TiO_ 2 a t% S c O
2 at% V O .
c 1200o
o
o
t5
0)
■55 1 0 0 0 Q
800-
-n-TT....... ,...........
1000
10000
I-100000
Frequency (Hz)
(b)
Figure 63. Dependence of dielectric constant on the frequency at room tem perature(a) Sc
doped B a 0 .7 S r 0 .3 T i 0 .9 8 V0 .0 2 O 3 -, (b) F-doped Bao, 7 Sra. 3 Tio.g$Sco.o2 0 3 .
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
CH A PTER 6
INVESTIGATION OF POINT DEFECTS INDUCED BY
NEUTRON DAMAGE AND ANNEALING AND THEIR
EFFECT ON LEAKAGE CURRENT IN BARIUM AND
STRONTIUM TITANATE FERROELECTRICS
6.1. In tro d u ctio n
High-voltage capacitors th a t are smaller, lower-loss, more reliable and less tem per­
ature dependent th an current components have a wide applications in the electronic and
microelectronic industry [121]-[125]. High dielectric constant materials, such as ferroelec­
tric (Ba, Sr)T iO s, are typically used in this technology to attain large capacitances in
compact structures [126, 127]. Despite significant progress in improving the properties of
high voltage capacitors, the properties of the dielectrics still limit the capacitor’s electrical
performance. It is believed th a t the point defects play a key role in the conductive behavior
and breakdown of ferroelectrics, but the fundamental conduction and high field breakdown
mechanism still needs to be further clarified.
There are several means th a t can introduce point defects into the ceramics samples,
such as doping with impurities, annealing in different atmospheres, and ion an d /o r neutron
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
121
irradiation. Radiation induced defects in semiconductors and ceramics have been exten­
sively studied [128] - [131]. For example, radiation degradation can be caused by point
defects generated by way of ionization and atomic displacement damage [132]. One advan­
tage of using the neutron damage to induce point defects is th a t the concentration of point
defects can be systematically controlled without changing the other param eters, such as,
crystal structure, particle size, and the secondary phase th a t might significantly affect the
dielectric and electrical properties of material. It is well known th a t annealing treatm ent
can significantly affect the dielectric properties and leakage current of (B a , Sr)T iO z ferro­
electrics by change the concentration and type of point defects. This is a main research
field in point defect chemistry and has been well documented, but these point defects are
studied mainly by measuring conductivity at different oxygen pressure at elevated tem pera­
ture combining with the point defect reaction equation, such as, V** and Vga e.g. However,
point defects in perovskite structure may contain unpaired electrons and thus would be
paramagnetic. It is reported th a t B aT iO z samples sintered at different atmosphere pro­
duce several param agnetic point defects. Thus, the influence of param agnetic point defects
on the leakage current needs to be studied and considered.
In chapter 5, the effect of point defects induced by doping impurities has been
investigated. In this chapter, the effect of point defects induced by neutron irradiation
on the electrical conductivity of B a o jS ro ^ T iO z is discussed first. Then, the point defects
induced by annealing treatm ent and its effect on the leakage current is investigated next.
The concentration and type of param agnetic point defects can be directly identified by
electron param agnetic resonance (EPR) techniques. By understanding the fundamental
defect chemistry of B a T iO z , combined with the measurement of leakage current of samples
annealed at different annealing condition, we hope to gain insight on how point defects
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
122
affect the leakage current and electrical breakdown.
6.2. E x p erim en ta l p roced u re
Conventional ceramic fabrication processes were the same as the ones described as
chapter 5. These sintered samples were irradiated with high-energy neutrons at the Idaho
Accelerator Center, Idaho State University. Samples were irradiated w ith three different
neutron dose: 2.12 x 1013/c m 2, 4.24 x 1013/c m 2, and 6.36 x 1013/c m 2 respectively . In
order to investigate the effect of annealing condition on leakage current, sintered BaTiO z
samples were annealed at 300°C and 500°C for 12 hrs, 24 hrs, 48 hrs in either reducing
5%J?2 + 95% Ar forming gas and pure O2 gas.
To carry out the I-V measurements, the surfaces of the as-sintered pellets were
ground and polished, followed by sputtering a thin layer of A u film as electrode on both
side surfaces. The I-V characterization was done using Quad 1865 M egom m eter/IR tester.
The crystal structure was characterized by a Rigaku D /M ax-IIB diffractometer.
Small samples exposed to the same processing treatm ent were used for the magnetic mea­
surement. To determ ine the concentration of param agnetic center, static magnetic suscep­
tibility measurements were done with a vibrating sample m agnetom etery from 2K to room
tem perature range. E P R was used to identify the type of param agnetic point defects in­
duced by annealed treat. E P R studies were performed at 120 K using quartz finger dewar
and CW measurements at X-band (9-10 GHz) using a Bruker Elexsys E580 spectrometer.
Polycrystalline samples were ground into fine powders. The weight of the powder samples
was kept at ±50 mg.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
123
6.3. P o in t d efects in d u ced by n eu tron irrad iation and its effect on th e leakage
current
X R D p a ttern . Fig. 64 shows XRD spectrum for non-irradiated and neu­
6 .3 .1 .
tron irradiated B a o ^S ro ^T iO z ceramic. XRD patterns of neutron irradiated samples have
not been presented since they are the same as th a t of non-irradiated sample. Secondary
phases were not detected for either the as-made or the neutron-irradiated samples. In con­
clusions, XRD show the crystal structure of Bao.jSro.zTiOz does not change by neutron
irradiation for this level of dose.
800
(a) Undamaged Ba07Sr03TiO3
(b) Damaged Ba07Sr03TiO3
( 110 )
700600-
( 220 )
( 111 )
(200 )
500( 100 )
c
400-
( 211 )
300 - •**
200-
100
20
30
40
50
60
2 Theta
Figure 64. XRD for non-irradiated and neutron irradiated BaQ.-jSrz.zTiOz sample. Neutron
irradiation shows th e same patterns as th a t of non-irradiated.
6 .3 .2 .
M a g n etic m easu rem en t. Mass susceptibility x (em u/g) is inferred from
the measured magnetization [M (em u/cm 3)] and sample mass [m (g)] by the following equa-
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
124
3. 00E-08
Dose 4. 24xlOVcin2,p a ra m a g n e tic p o in t
d e f e c ts :6. 0xl018/cm2
2. 50E-0;
Su
1 .5 0 E -0 8
Dose 2 . 12xl013/cm2, paramagnetic point
d e fe c ts :7. 6xl018/cm2
s c
e p
1 .0 0 E -0 8
t i
5. 0 0 E -0 9
Before damage, paramagnetic point d e fe c ts:4 . 7x10 /cm
Temperature (K)
Figure 65. M agnetic susceptibility for non-irradiated and neutron irradiated BaQjSrQ^TiOz
samples. The noise for sample irradiated by 4.24x1013/c to 2 result from environmental
artifact in the measuring room during experiment.
tion [4]:
where H is the external magnetic field (G). As shown in Fig. 65, the magnitude of the
observed param agnetic moment of the BaQ.iSro^TiOz samples is found to increase with
neutron dose. Table 8 lists the referred concentration of param agnetic point defects from
Fig. 65.
It is well known th a t the kinetic energy transferred by an energetic impinging particle
to a lattice atom results in the formation of point and extended defects [132], although the
type of induced defects need to be further identified in this case. While no paramagnetic
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
125
im purity is introduced into the samples, the increase in param agnetic moment is due to
the appearance of param agnetic native defects resulting from the neutron irradiation, as
the neutron irradiation process does not introduce any param agnetic impurities into the
samples.
Sample
Undamaged
Irradiated w ith dose 2.12 x 1013/c m 2
Irradiated w ith dose 4.24 x 1013/c m 2
Irradiated with dose 6.36 x 1013/c m 2
Concentration of param agnetic point de­
fects measured at 10K (cm -3 )
4.2 x 1020
6.0 x 1020
7.6 x 1020
8.3 x 1020
Table 8. Concentration of param agnetic point defects irradiated by different neutron dose.
1000
IM/L8ashowing space change limtedojT^nt
With neutron dose 2.12x10t13/ati2
<
o
100
c
£
15
Trapsfilledlim
it
G
CD
O)
CO
0)
I ^ A 90
10
lu ^Undam aged Ba^Sr^TiOj,
' J8100
1000
10000
Electrical Strenght (V/Crrf
Figure 66. Logarithmic plots of the leakage current as a function of applied field strength
for the undam aged and damaged sample with neutron dose 2.12 x 1013/c m 2.
6 .3 .3 .
I-V c h a r a c te r iz a tio n . Fig. 66 and 67 are current-voltage curves for the
undamaged and neutron damaged Bao.iSrQ.zTiOs samples. The leakage current was found
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
126
Dose 2.12x10 /cm'
Undamaged Ba07Sr03TiO,
Dose 4.24x10 /cm
Dose 6.36x10 /cm'
100
1000
10000
Electrical Field Strength (V/cm)
Figure 67. I-V characteristics for non-damaged and neutron irradiation B a ^ jS r o ^ T iO s
with dose 2.12 x 1013/c m 2, 4.24 x 1013/c m 2, and 6.36 x 1013/c m 2.
to be essentially identical to those presented in the figure when the bias was reversed (not
shown here), indicating th a t non-rectifying contacts are formed between the electrode and
insulator. Fig. 66 shows the logarithmic plots of the leakage current as a function of applied
field strength for the undam aged and damaged sample with neutron dose 2.12 x 1013/c m 2.
For these two samples, in the low and middle field region, log(I) vs log(E) curves are close
to linear, characteristic of Ohmic conduction. At higher field region, log(I) vs log(E) curves
indicated a power relation:
IocVm
(6.2)
where m = 2, showing a characteristics of space limited conduction [134]. Thus, the space
charge limited conduction is dominant in the higher field. We also observe th a t the leakage
current increases in both the Ohmic region and space charge limited current region with
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
127
2.12 x 1013/c m 2 dose neutron irradiation.
In the as-made from, Bao^Srg.zTiOz has a relatively low density of free carriers [133]
and a sufficiently small point defect concentration not to have significant hopping conduction
in a defect band. The introduction of paramagnetic point defects during 2.12 x 1013/c m 2
neutron dose irradiation is found to result in a measurable increase in the leakage current.
For example, oxygen vacancies appear as donors th a t generate free carriers or facilitate
electron hopping w ithin the defect band. It is also evident from Fig. 66 th a t the sharp
increase region labeled as traps-filled-limited. The appearance of these regions results from
the fact th a t there exist unneutralized charges in the discrete traps th a t prevent the injection
of additional electrons form the cathode [135]-[134].
It is interesting th a t a significantly reduced leakage current at all applied bias is
observed (Fig. 67) for samples irradiated by 4.24 x 1013/c m 2 and 6.36 x 1013/c m 2 neutron
dose. This might result from the deep level traps generated by the point defects th a t are
induced by higher neutron dose. The appearance of traps would be expected to trap free
carriers and thereby the m agnitude of Ohmic and space-charge-limited current.
6.4. P oin t d efects in d u ced by an n ealin g trea tm en t and its effect on th e leakage
current o f B a T iO z
6.4 .1 .
Leakage current under different a n n ealin g co n d itio n . Fig. 68 com­
pares I-V characteristics of unannealed B a T iO z, annealed B a T iO z in oxidizing (pure O2)
and reducing (5vol% ii/2+95vol%Ar) atmosphere at 500°C' for 24 hrs. Significant differences
are observed. Annealing B a T iO z in a reducing atmosphere increases the leakage current,
while annealing in oxidizing atmosphere reduces it. According to La chatelier’s principle and
i
basic point defect chemistry, annealing in reducing atmosphere, the donor electrons would
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
128
be expected to compensate dopants and oxygen vacancies at low Po2, whereas at high
Po 2 (annealing in oxidizing atmosphere) the excess positive charge of donor ions would be
expected to com pensated by cation vacancies, as shown in the equations below:
~ 02 + V** — O * + 2h* (oxidation)
(6.3)
^ 0 2 (g) + 4e' = Og + V?" (reduction)
(6.4)
7^ 2(5) + 2e' = Og + Vga(reduction)
(6.5)
or
The point defects act as color centers in ceramic materials. They can absorb light
in the visible spectrum , resulting in a perceivable change in color. B aT iO z annealed in O 2
shows shallow yellow, while the color of sample annealed in reducing atmosphere is much
dark.
Fig. 69 further shows the dependence of the leakage current from BaTiOz annealed
in oxidizing atmosphere as a function of annealing time. As shown, the leakage current
reduces with the tim e for samples annealed in the oxidizing atmosphere. In contrast, the
leakage current increases with annealing time for samples annealed in reducing atmosphere,
see Fig. 70.
Fig. 71 shows a series of typical log (V)-log (I) curves measured from room tem pera­
ture to 200°C for Ba T iO z samples annealed at 500°C for 48 hrs in an oxidizing atmosphere.
As expected, the leakage current continuously increases with the increasing tem perature
since a greater therm al energy is available for hopping an d /o r generating therm ally excited
electrons.
A nother characteristic th a t is worthy of noting is th a t Ohmic region is followed
at higher fields by space charge limited current. The transition from Ohmic conduction
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
129
2.50E-04
Annealed in 02 at 500C for 24 hrs
<
2.00E-04
8
1.50E-04
Annealed in H2 at 500Cfor 24 hrs
Unannealed
1.00E-04
5.00E-05
0.00E+00
fttfvwyrfi
*
200
0
'
400
600
800
Applied Voltage (V)
Figure 68. I-V curves for unannealed, annealed BaTiOs in pure 0% and 5 vol%H 2 +
95vol%Ar forming gas at 500°C respectively. Samples are measured at 100°C.
1.20E1.00E-
0
\
X
^
C
u1— 8.00E
D
o 6.00E0
U) 4.00E03
.id
03
2.00E0
X
■
0
2 4
-
X
w
.... V
■
$
J i
00
/
0.00E+
0
10
20
30
40
50
60
Applied Voltage (V)
Figure 69. Dependence of leakage current of annealed BaT iOz in oxidizing gas at 500° C
on the annealing time.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
130
1.10E-03
9.00E-04
ol.
0 h rs
12 h rs
24 h rs
48 h rs
7.00E-04
k.
3
5.00E-04
0)
a 3.00E-04
re
re
re
1.00E-04
-1.00E-04
TOO"— 200“— 300...... 400-----500— 600----- 700
A p p lied V o lta g e (V)
Figure 70. Dependence of leakage current of annealed BaTiOz in reducing atmosphere on
the annealing time.
to space charge limited current shifts to lower applied voltage region with the increasing
measurement tem perature. Space charge limited current occurs when the carriers density
injected from the electrodes exceeds the native bulk carrier concentration. This current
usually relates to electronic carriers, since the contribution to ionic transport is small in
these materials. At elevated measurement tem peratures, the onset of space charge limited
current results at lower applied voltage since the injected electron density from the electrodes
is significantly higher under these conditions th an the native electron concentration.
6 .4 .2 .
A c tiv a tio n en e rg ie s. Ohmic current can consist of electronic (electron
and/or hole) and ionic contributions. The voltage (V) dependence of current (I) is ex­
pressed as:
/ = ^
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
(6.6)
131
Slope =~2
shoving space-lim ited
1 .00E+00
(A)
1. 00E-03
Leakage Voltage
1. 00E-02
1 .00E-04
1. 00E-05
1 .00E-06
1. 00E-07
1. 00E-08
1. 00E-09
1 .00E-10
1
J
10
Slope = 1 showing
Applied Voltage (V)
Ohimic current
1000
Figure 71. Typical log (V)-log (I) curves for BaTiOz annealed at 500°C in O 2 for 48 hrs.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
132
1. 00E -04
1. 00E-05
<c
g 1 .00E -06
J3h
§T
O 1 . 00E-07
aj
03
1 .00E-08
1. OOE-09
2
1. 5
2 .5
3
3 .5
1000/T (1/K )
Figure 72. Typical Arrhenius curve at low voltage range for the calculation of the activation
energy of annealed BaT iOz in O 2 for 48 hrs at 500°C'.
1.00E-03
1.00E-04
1.00E-05
§ 1.00E-06
Ea=0.40(ev)
g, 1.00E-07
1
1.5
2
2.5
3
3.5
1000/T(1/K)
Figure 73. Typical Arrhenius curve at low voltage range for the calculation of the activation
energy of annealed Ba T iO z in H 2 for 12 hrs at 500° C.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
133
0.65
0 .6 0 > 0 .5 5 o
>*
U i 0 .5 0 vo
= 0 .4 5 -
0 .4 0 -
|
<C
> 0.3 5 -
o
<
0 .3 0 0.25
48 in H2 2 4 in H 2 12inH 2
0
12 in 0 2 48 in 02
Annealed Time (hour)
Figure 74. Activation Energy of BTO annealed in H 2 and O 2 Measured at low dc voltage
(51V).
Where: q electronic charge; A and L cross-sectional area and thickness; n carrier concentra­
tion; /i carrier drift mobility. For transport dominated by hopping mode, the tem perature
dependence of current may be expressed in term s of activation energy E a:
I - I
0
e
x
.
(6-7)
Thus, activation energies of samples annealed under different annealing condition can be
determined by measuring current versus tem perature at fixed voltage. 51 V is chosen to
characterize the Ohmic current since all of the curves are dom inated by this transport be­
havior at this voltage. Two typical examples of samples annealed in oxidizing and reducing
atmosphere are seen in Fig. 72 and 73 respectively. Fig. 74 summaries the activation ener­
gies for samples annealed for different times at 500°C in oxidizing and reducing atmosphere.
It is seen th a t the activation energies of samples annealed in reducing atmosphere are less
than those of samples annealed in oxidizing atmosphere, while the leakage currents show
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
134
8.0E-04
Annealed in Ar/Hj
a 6.0E-04
I
4.0E-04
*
2.0E-04
0
10
20
30
Temperature (K)
40
SO
Figure 75. Magnetic susceptibility for samples annealed in varying annealing atmosphere.
the opposite tendency as described in previous section. Further, samples annealed for the
longer duration in O 2 result in the higher activation energy. In contrast, samples annealed
with the longer duration in
5
vol%iH2 + 95vol%Ar result in the lower activation energy.
While the measured energies equal to neither the intrinsic band gap and nor the ionization
of B a T i O z , the measured activation energies th a t are between 0.30ev to 0.62ev show the
dependence on th e annealing condition. Thus, it is concluded th a t th e dominant conduc­
tion mechanism is by electron hopping. The lower activation energy (barrier) and increased
concentration of point defects for samples annealed in reducing atmosphere facilitate the
electron transport and increase the leakage current.
6 .4 .3 .
M a g n e tic m e a s u re m e n t a n d E P R . It is observed in previous sections
th a t different annealing condition causes a significant change in leakage current. As ex­
plained earlier, these changes might result from the appearance of induced point defect and
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
135
the change of activation energy during annealing. On the other side, in densely packed
perovskite structure, the Frenkel defects usually are ignored and only Schottky disorder are
found to dominate. Depending on the occupancy of these vacancies, they may contains
unpaired electrons and thus would be paramagnetic. Thus, the magnetic measurement and
E P R can be used to further characterize the samples annealed in different atmosphere.
Mass susceptibility x (emu/g) is inferred from E quation.(6.1). Results of magnetic
measurement for unannealed sample and samples annealed at 300° C and 500° C in reducing
and oxidizing atmosphere are shown in Fig. 75 and 76. As shown in Fig.75, the magnitude
of the observed param agnetic moment of BaTiOz annealed in reducing atmosphere is larger
than th a t of sample annealed in oxidizing atmosphere. Fig. 76 shows the concentration of
param agnetic point defect for samples annealed in O 2 calculated from the magnetic sus­
ceptibility has the lowest value, while the concentration for samples annealed in reducing
gas has the higher values. It is also observed th a t the sample annealed at higher annealing
tem perature in reducing atmosphere consequently increases the concentration of param ­
agnetic point defects. In contrast, the sample annealed at higher annealing tem perature
in oxidizing atmosphere consequently decreases the concentration of param agnetic point
defects. The change of param agnetic moment results from the creation or annihilation of
point defects during the different annealing treating as there are not a significant number
of param agnetic im purities introduced into the samples during annealing.
E P R is used to further identify the physical nature of the param agnetic point defects.
The E PR resonance signal arises when the energy of the electromagnetic wave, hu, passing
through the sample is equal to the Zeeman splitting; A E, (i.e. the difference in energy
between the electronic levels with high and low spin configurations) of defects w ith unpaired
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
136
10
Annealed in Ar/H f**^
As sintered""” **!
s«
♦
Annealed in 0 f
% 6
1
s
«
v
2 4
Concentration (cm'
*>
As
sintered
300"C
7.3SX1018
500<’C
6.83X101*
Annealed in
Temperature
A n n e a le d in
Oa
Ar/Ha
?,05*101s
7.74X1018
6.3Sx101*
7.58X101*
;
g 2
X.
0
0
100
200
JL,
300
400
500
000
Tempera tore (**€)
Figure 76. Calculated concentration of param agnetic point defects from Fig.75.
A* sintered
Annealing in Ar/Hj
tg=2.004
0
3200
3250
3300
3350
3400
3450
3500
Magnetic Field (G)
Figure 77. E P R spectra of unannealed BaTiOz and BaTiOz annealed in oxidizing and
reducing gas for 48 hrs at 500°C. Spectra are obtained at 120K.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
137
electrons: [140]:
A E = hv — g/j,0H
(6-8)
where nn is the Bohr magneton, H is a m agnitude of the magnetic field, and g is the
giromagnetic constant. It is this constant th a t is specific for each param agnetic defect.
If the sample contains atoms with non-zero nuclear magnetic spin, then additional energy
splitting may occur due to their hyperfine interaction w ith unpaired electrons.
E P R results for measurement of BaTiOz ceramics are shown in Fig. 77. As shown,
E P R spectra of unannealed and annealed BaTiOz in reducing atmosphere are identical.
The only signal in these samples is the g=2.004 signal. According to the defect chemistry of
BaTiOz and giromagnetic constant values of param agnetic centers in BaT iOz summarized
in [140]-[144], the param agnetic point defects originates from oxygen and barium vacancy
complexes in a charge state w ith unpaired electron. Thus, since their concentration de­
creases in oxidizing atmospheres and increases in reducing atmosphere, they can tentatively
be assigned to be Vo-Vj3a complex.
As mentioned, the oxygen (V**) plays key role in
the resistance degradation and leakage current. Unfortunately, V** is not paramagnetic
point defect because it has paired electrons and it is not detectable by E P R and magnetic
measurement.
However, combining with the measurement of leakage current, magnetic
measurement and E P R results show th a t the param agnetic point defects, such as, Vo-V^a
might make a significant contribution to the leakage current.
6.5. C onclu sion
In conclusion, the leakage current is strongly affected by point defect induced by
neutron irradiation for undoped Bao^Sra.zTiOz ceramics. M agnetic measurement shows
the continuous increase of the concentration of param agnetic point defect with the neutron
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
138
irradiation. The current-voltage (I — V ) characteristics indicate th a t the initial neutron
irradiation causes the slight increase of Ohmic current and space charge limited current.
However, further neutron irradiation can significantly reduce the leakage current.
We also study the effect of annealing treatm ent and induced point defects on the
leakage current of BaTiOz.
The leakage current decreases in oxidizing atmosphere and
increases in reducing atmosphere. Leakage current increases w ith the annealing time for
BaTiOz samples annealed at H 2 . In contrast, the leakage current reduces w ith annealing
time for O2 annealed Ba T iO z samples. The dominant conductive mechanism is hopping
mode. M agnetic measurement and E P R are used to identify the type and concentration
of param agnetic point defects. Combining with the measurement of leakage current and
magnetic measurement, results show th a t the presence of point defects including Vo-Vga
complex might make a significant contribution to the leakage current.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
CH A PTER 7
CONCLUSION AND FUTURE WORK
7.1. C on clu sion s for barium ta n ta la te-b a sed m icrow ave ceram ics
7 .1 .1 .
M icro stru ctu re,
stru ctu ral,
ch em ical, electro n ic and h igh fre­
q u en cy d ielectric p ro p erties o f barium cad m iu m ta n ta la te-b a sed ceram ics. The
nano-sized 1 : 2 ordered domain structures with a twin crystallographic relationship
and high-density domain boundaries induced by ordering were observed for undoped
Ba{C dij^ Ta 2 /z)Oz sintered at 1520°C.
While B a ( Z n xCdi/z_xT a 2 /z)Oz ceramics with high densities could not be attained
without the aid of a sintering agent, the addition of 2wt% ZnO was needed to achieve
over 97% of the theoretical density for pure Ba {C di jzT d 2 /z)Oz ceramics. Evidence for CdTa ordering, as indicated by the presence of superstructure peaks in the X-ray diffraction
spectra, was found. For a sample sintered at 1550°C for 48h, the dielectric constant and
microwave loss tangent were measured to be ~ 32 and 5 x 10-5 at 2 GHz. Local density
functional calculations of B d ( C d i / 3 T d 2 /z)Oz and B d ( Z n i / 3 T d 2 /z) 0
3
give insight into the
unusual nature of this class of material. The conduction band maximum and valence band
minimum are strongly composed of weakly itinerant Ta 5d-and Z n — 3d /C d — 4d levels,
respectively. This is believed to play an im portant role in having a high melt tem perature
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
140
and enhanced phonon energies, as well as the unusual property of having both a high
dielectric constant and low loss tangent.
7.1.2.
M icro stru ctu re and d ielectric p ro p erties o f B a ( C d i/ 3 T a 2 / 3 ) 0
crow ave ceram ics sy n th esized w ith a B oron oxid e sin terin g aid. The use of boron
as a sintering aid reduces the sintering tem perature, enhances the sintering density, and im­
proves the microwave properties of Ba{Cdi / 3 T a 2 /z) 0
3
ceramic dielectrics. Observations by
transmission electron microscopy indicate th a t the liquid sintering mechanism contributes to
the improvement in sintering density for boron concentrations exceeding 0.5 wt%. No boron
segregation and amorphous phase were observed along grain boundaries. An amorphous
phase rich in boron-oxide forms in pockets partially penetrating along multiple grain junc­
tions. The enhancement of sintering behavior results from the formation of boron-containing
liquid phase in B a ( C d i / 3 T a 2 /s)Os ceramics with boron additive. The introduction of as
small as 0.01wt% boron also results in the production of high-density samples (~ 95%), pre­
sumably indicating th a t a point defect mechanism may also play an im portant role in the
sintering process. X-ray diffraction d ata combined with high-resolution transmission elec­
tron microscopy images show th a t boron-doped B a ( C d i / 3 T a 2 / 3)Os ceramic m aterial has a
well-ordered hexagonal structure. Annealing treatm ent is found to improve the microwave
properties. The best sample has a dielectric constant of 32, a tem perature coefficient of
resonant frequency of 80±15 ppm /°C , and a quality factor of >25,000 at 2 GHz.
Electron diffraction and high resolution transmission electron microscopy studies
showed a well ordered structure of 1:2 with hexagonal symmetry for B a ( C d i / 3 T a 2 / 3 ) 0 3
with boron additive sintered at 1350°C. No significant changes in ordered structures were
observed for B a ( C d i / 3 T a 2 / 3 ) 0 3 subjected to a long-period annealing subsequently.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
3
m i­
141
7.1.3.
D ielectric-lo a d ed M ob iu s reson ator and stru ctu ra l, ch em ical and
d ielectric p ro p erties o f ceram ic in jectio n m old ed Ba(Zrii/zTa2iz)Oz m icrow ave
ceram ics. The development of a Ceramic Injection Molding (CIM) process to produce
complex-shaped structures using high-performance microwave ceramic m aterials was re­
ported. In particular, we describe the synthesis methods and the structural, chemical and
dielectric properties of B a ( Z n 1 / 3 T a 2 /z)Oz doped w ith Z r and N i ceramics produced by
CIM. Sintering the CIM-ed B a ( Z n 1 /z Ta 2/ 3)Oz to a relative density of ~94% was possible
at a tem perature of 1680°C and an extended duration of 48 hrs. The best samples to date
exhibit a dielectric constant, ey, of ~30, a Q value of ~31, 250 at 2 GHz, and a tem perature
coefficient of resonance frequency, r / , of 0.1 ppm /°C .
The incorporation of dielectric loading of Mobius wire resonators was successfully
dem onstrated.
W ith this approach, a very compact resonator volume can be achieved
since the size reduction due to dielectric loading and the Mobius resonator concept are
synergistic. Besides the resonators fabricated by machinable Macor, a number of dielectric
Mobius resonator test structures were fabricated using injection molded Ba(Z rii /zT a 2 /z)0z
ceramics, in conjunction with small diameter gold metal wire.
7.2. C on clu sion s for barium and stro n tiu m tita n a te ferroelectrics
7 .2 .1 .
D ielectric p ro p erties and I-V ch a ra cteristics o f V anadium and
Scandium d o p ed B a i ^ xS r xTiOz under high dc bias v o lta g e. Doping Sc into
Bao.’j Sro.zTiOz changes its crystal structure from tetragonal to rhombohedral structure
and significantly reduces the dielectric constant of Bao^Sro^TiOz-
In contrast, doping
V into BaT iOz and Bao. 7 Sro,zTiOz does not change their crystal structure. The leakage
current is reduced by doping with V2 0^. In this case, the oxygen vacancy concentration is
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
142
suppressed when the presence of a positively charged donor reduced the concentration of
the other positively charged defects, including the oxygen vacancies. No significant depen­
dence of the dielectric constant on the frequency is observed for Sc-doped BaojSro.zTiOzIn contrast, significant dependence of the dielectric constant on the frequency is observed
for V-doped Bao^Sro^TiOz- It is believed th a t this difference originates from the crystal
structure changes. The breakdown voltage for 2 wt% V doped Ba^/jS r^z TiOz is above
17, OOOV/cm and its dielectric constant at room tem perature is ~ 13,000.
7.2.2. E ffect o f n eu tron irrad iation and a n n ealin g tr ea tm en t on leakage
current. The leakage current is strongly affected by point defect induced by neutron dam­
age for undoped Bao.-jSro.zTiOz ceramics. Magnetic measurement shows the continuous
increase of the concentration of param agnetic point defect with the neutron irradiation. The
current-voltage ( I —V) characteristics indicate th a t the initial neutron damaged causes the
slight increase of Ohmic leakage current and space charge limited current behavior. How­
ever, increasing the neutron irradiation dose to 4.24 x 1013cm ~2 and 6.36 x 1013cm -2 can
significantly reduce the leakage current. Work is under way to further investigate the type
and state of point defects induced by neutron damage in order to identify their role in
reducing the leakage current.
We also study the effect of annealing treatm ent and induced point defects on the
leakage current of B a T i O z ■ The leakage current decreases in oxidizing atmosphere and
increases in reducing atmosphere. Leakage current increases w ith the annealing time for
BaTiOz samples annealed at Hi- In contrast, the leakage current reduces w ith the annealing
time for O 2 annealed B aT iO z samples. The dominant conductive mechanism is by electron
hopping. The results of the electrical, magnetic and E P R indicate th a t the paramagnetic
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
143
point defects, e.g, Vo-Vj3a complex might make a significant contribution to the leakage
current.
7.3. F uture work
7.3.1.
M icrow ave d ielectrics. Future work on the microwave dielectrics should
focus on the m icrostructure and point defects on the dielectric properties. The following
tasks should be addressed:
• Clarify further the role of point defects in determining the dielectric properties, in­
cluding the dielectric constant, loss tangent, and tem perature coefficient of resonant
frequency. The point defects induced by neutron irradiation will be studies by EPR,
magnetic and optical measurement, ft is expected th a t the role of the concentration
and type of point defect in determining the dielectric properties will be clarified.
• Investigate the effect of annealing treatm ent on the dielectric properties of microwave
dielectrics since initial experiment shows the strong dependence of loss tangent on
the annealing atmosphere. The annealing treatm ent will be carried out in different
atmosphere and for different time. Special attention will be paid to the relationship
between the change of crystal structure and point defects.
• Study the relationship between the processing param eter and the tem perature coeffi­
cient of resonant frequency. Special attention should be paid on the crystal structural
and m icrostructural changes caused by processing param eter.
• Identify the sintering aid to significantly reduce the sintering tem perature for CIM-ed
B a { Z n i / z T a 2 /z)Os samples. Favorite sintering aid should form liquid phases at the
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
144
sintering tem perature or it can facilitate the atom diffusion during the solid state
sintering.
7.3.2.
F e rro e le c tric s . Future work on the microwave dielectrics should focus on
the effect of m icrostructure and point defects on the dielectric properties. The following
tasks should be address:
• Do further experiments, such as, E P R and magnetic measurement, combining with
the point defect chemistry to investigate what types of point defects are induced by
neutron irradiation. The nature of traps in the BSTO and its effect on the leakage
current should be clarified.
• Use a combination of electronic structure calculation to ascertain the defect properties
and TR IM /SR IM to model the damage process to predict the nature and concentra­
tion of defects formed during neutron damage.
• Investigate the point defect induced by neutron damage and annealing treatm ent and
its effect on the breakdown voltage and further clarify the breakdown mechanism.
• Use defect and grain boundary engineering to choose the proper impurities th a t can
significantly increase the barrier height at the grain boundary. This might be a key to
achieve the high breakdown voltage. In addition to the electrical measurement, TEM
and HREM will be widely used to analysis the domain structure, the secondary phases
and structure of grain boundary to determine the possible improvement of breakdown
voltage.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
REFERENCES
[1] K. Wakino, T. Nishikawa, Y. Ishikawa, and H Tamura, Dielectric resonator materials
and their application for mobile communication systems, Br. Ceramic. Trans. J., 89,
1990, 39.
[2] D. Davies, Ceramic transaction: Materials and processes for wireless communication,
Vol.53, 137-149, (1995)
[3] T. Vanderah, Talking Ceramics, Science, 298, 2002, 1182-1184.
[4] G. Rong, N. Newman, B. Shaw, and D. Cronin, Role of Ni and Zr doping on the elec­
trical, optical, magnetic, and structural properties of barium zinc tan talate ceramics,
J. M ater. Res., 14, 1999, 4011.
[5] J. Pond, S. Liu, N. Newman, IEEE Trans. M TT. 49, 2001, 2363.
[6] T. Negas, G. Yeager, S. Bell, and R. Amren, Chemistry of electronic ceramic materials,
Proceedings of the international conference, Jackson, WY, July 17-20, 21 (1990)
[7] H. Tamura, T. Konoike, Y. Sakabe, and K. Wakino, J. Am. Ceram. Soc., 67, C59
(1984).
[8] R. Zurmiihlen, J. Petzelt, S. Kamba, G. Kozlov, A. Volkov, B. Gorshunov,D. Dube,
A. Tagantsev, and N. Setter, J. Appl. Phy., 77, 5351 (1995)
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
146
[9] D. A. Sagala and S. Nambu, J. of Phy. Soc. Japan, 61, 1791 (1992).
[10] S. Kamba, H. Hughes, D. Noujni, S. Surendran, Robert. C. Pullar, P. Samoukhina, J.
Petzelt, R. Freer, N. M. Alford, and D. M. Iddles, J. Phys. D: Appl. Phys., 37, 1980
(2004)
[11] T.
Takahashi,
First-principles
investigation
of
the
phase
stability
for
B a ( B y 3 2 +B y 3 5 +)C>3 microwave dielectrics with the complex perovskite structure,
Jpn. J. Appl. Phys., 39, 2000, 5637-5641.
[12] D. A. Sagala and S. Nambu, Microscopic calculation of dielectric loss at microwave
frequencies for complex perovskite Ba(Zni/zTa, 2 /z)Oz, J- Am. Ceram. Soc., 75, 1992,
2573-75.
[13] Yung P ark and Yoonho Kim, M ater. Res. Bull. 31, 7(1996)
[14] P. Davies, J. Tong and T. Negas, Effect of ordring-induced domain boundaries on lowloss Ba{Zni/%Ta 2 /z)Oz — B a Z r O z perovskite microwave dielectrics, J. Am. Ceram.
Soc., 80, 1727 (1997)
[15] V. Tolmer and G. Desgardin, J. Am. Ceram. Soc, 75, 2573 (1992)
[16] J. M. Pond, S. Liu, N. Newman, Dual-mode and quad-mode Miibius resonators, IEEE
MTT-S Int. Microwave Symp. Dig. 3 (2000) 1771-1774
[17] T.Chartier, E.Delhomme, J.F.B aum ard, G.Veltl,F.Ducloux, Injection moulding of hol­
low silicon nitride parts using fusible alloy cores, Ceram. Int., 27 (2001) 821 -827
[18] S.I.-En.Lin, Near-net-shape forming of zirconia optical sleeves by ceramics injection
moulding, Ceram. Int., 27 (2001) 205 -214
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
147
[19] G. de W ith and N. M. W ithbreuk, Injection moulding of Zirconia (Y-TZP) ceramics,
J. Euro. Ceram. Soc., 12 (1993) 345-351
[20] S. Krug, J. R. G Evans, and J. H. H. ter M aat, Residual stresses and cracking in large
ceramic injection mouldings subjected to different solidification schedules, J. Euro.
Ceram. Soc., 22 (2002) 345-351.
[21] H. M. Shaw and M. J. Edirisinghe, Shrinkage and particle packing during removal
of organic vehicle from ceramic injection mouldings, J. Euro. Ceram. Soc., 15 (1995)
109-116
[22] H. S. Nalwa, Handbook of low and high dielectric constant materials and their appli­
cation, Vol.2, Academic press, San Diago, CA, 1999.
[23] Colla, E. L., Reaney, I. M. and Setter, N., Effect of structural changes in complex
perovskites on the tem perature coefficient of the relative perm ittivity, J. Appl. Phy.,
74, 1993, 3414-3425.
[24] Reaney I. M., Wise P., and R. Ubic R., On the tem perature coefficient of resonant
frequency in microwave dielectrics, Philos. Mag. A, 2001, 81, 501-510.
[25] S. J. Penn, N. McN. Alford, A. Templeton, X. Wang, M. Xu, M. Reece, and K. Schrapel,
J. Am. Ceram. Soc., 80, 1997, 1885.
[26] K. Wakino, K. Minai and H. Tamura, J. Am. Ceram. Soc., 67, 1984, 278.
[27] T. Negas et.c, Am. Ceram. Soc. Bull. 72, (1993) 80.
[28] S. Hirano, T. Takashi, and A. H attori, J. Am. Ceram. Soc., 74, 1991, 1320.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
148
[29] C. L. Huang and R. J. Lin, Liquid Phase Sintering and Microwave Dielectric Properties
of Ba(Mgi/zTa< 2 /z)Oz Ceramics, Jpn. J. Appl. Phys. 41A (2002), pp. 712-716 P art 1.
[30] S. Kawashima, M. Nishida, I. Ueda, and H. Ouchi: B a ( Z n i / 3T 0 ,2 / 3 ) 0 3 ceramics with
low-dielectric loss at microwave frequencies. J. Am. Ceram. Soc. 66, 421 (1983).
[31] K. Endo, K. Fujimoto and K. Murakawa, J. Am. Ceram. Soc. 70 (1987) C215.
[32] I. N. Lin, C. T. Chia, H. L. Liu, H. F. Cheng and C. C. Chi, Dielectric Properties of
xBa(Mg-l/ 3 T d 2 / 3 ) 0 3 -(l — x ) B a { M g i / 3 Nb 2 /z ) 0 3 Complex Perovskite Ceramics, Jpn.
J. Appl. Phys. 41B (2002) pp. 6952-6956 P art 1
[33] E. L. Colla, I. M. Reaney and N. Setter, Effect of structural changes in complex per­
ovskites on the tem perature coefficient of the relative perm ittivity, J. Appl. Phy., 74,
1993, 3414-3425..
[34] K. H. Yoon, D. P. Kim and E. S. Kim, J. Am. Ceram. Soc. 77 (1994) 1062.
[35] H. J. Youn, K. K. Kim and H. Kim, Jpn. J. Appl. Phys. 35 (1996) 3947.
[36] S. Nomura, K. Toyama and K. K aneta, Jpn. J. Appl. Phys. 21 (1982) L624.
[37] Y. Kobayashi and M. K atoh, Microwave measurement of dielectric properties of lowloss materials by the dielectric rod resonator m ethod IEEE Trans. M TT., 33, 1985,
586-592.
[38] D. Kajfez and P. Guillon, Dielectric resonators, Artech House, Dedham, Mass., 1996.
[39] Y. M. Chiang, D. Birnie III, and W. D. Kingery, Physical ceramics, John Wiley &
Sons, Inc., Newyork, 1997.
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
149
[40] W. D. Kingery, H.K Bowen, and D. R. Uhlmann, Introduction to ceramics, 2nd ed.
Wiley, Newyork, 1976.
[41] M. Barsoum, Fundam entals of ceramics, McGraw-Hill Companies, Inc., Newyork, 1997.
[42] A. J. Moulson and J. M. Herbert, Electroceramics, Chapm an and Hall, 1990, Cam­
bridge, U.K
[43] R. W. calm, P. Haasen, and E. J. Kramer, M aterials science and technology, Vol.11,
VCH publisher Inc., Newyork, 1994.
[44] M. H. Prey and D. A. Payne, Grain-size effect on structure and phase transformations
for barium titanate, Phy. Rew. B., 54, 1996, 3158-3168.
[45] A. E. Jacobs, Phys. Rev. B 52, 1995, 6327.
[46] W. R. Buessen, L. E. Cross, and A. K. Goswami, J. Am. Ceram. Soc. 49, 1966, 33.
[47] K. Ishikawa, K. Yoshikawa, and N. Okada, Phys. Rev. B 37, 1988, 5852.
[48] W. L. Zhong, B. Jiang, P. L. Zhang, J. M. Ma, H. M. Cheng, Z. H. Yang, and L. X.
Li, J. Phys. Condens. M atter 5, 1993, 2619.
[49] J. W. Liou and B. S. Chiou, Effect of direct-current biasing on the dielectric properties
of barium strontium titanate, J. Am. Ceram. Soc., 80, 1997, 3093-99.
[50] K. M. Johnson, Variation of dielectric constant w ith voltage in ferroelectrics and its
application to param etric devices, J. Appl. Phys., 33, 1962, 2826-31.
[51] L. C. Tien, C. C Chou, and D. S. Tsai, Ordered structure and dielectric properties of
lanthanum -substituted B a ( M g i / 3 T a 2 /s) 0
3
, J. Am. Ceram. Soc., 83, 2000, 2074-78.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
[52] F. Galasso and J. Pyle, Inorg. Chem. 2, 482 (1963).
[53] F. Galasso and J. Pyle, J. Phys. Chem. 67, 1561 (1963).
[54] E. L. Ginzton, Microwave Measurements, (McGraw Hill Book Co., 1957).
[55] D. Kajfez, Q Factor, Vector Fields, Oxford, MS (1994).
[56] D. Kajfez, IEEE Trans. M TT 42, 1149 (1994).
[57] P. Wheless and D. Kajfez, IEEE MTT-S Symposium Dig. 473, St Louis, (1985).
[58] W.E. Courtney, IEEE Trans. M TT. 18, 476, (1970).
[59] B.W. Hakki and P.D. Coleman, IEEE Trans. M TT, 8, 402 (1960).
[60] C.C. Johnson, Field and Wave Electrodynamics, 173-176 (McGraw-Hill, New York,
1965).
[61] M athem atica, V4.0, (Wolfram Research, 1999).
[62] J. Sun, S. J. Liu, N. Newman, and David. J. Smith, Ordered Structures in
B a( C d i/ ^ Ta 2 /s)Oz Microwave Ceramics: A Transmission Electron Microscopy Study,
Mat. Res. Soc. Symp. Proc. Vol. 783, 2004 Materials Research Society, B5.12.1.
[63] D. Singh, Phys. Rev. B 43, 6388 (1991)
[64] M. Methfessel, M. van Schilfgaarde, and R. A. Casali, “A full-potential LMTO method
based-on sm ooth Hankel functions,” in Electronic Structure and Physical Properties of
-Solids: The Uses of the LMTO M ethod, Lecture Notes in Physics, 535. H. Dreysseed.
(Springer-Verlag, Berlin, 2000).
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
151
[65] J. Sun, Shaojun Liu, N. Newman, and David J. Smith, Appl. Phy. Lett, 84, 3918
(2004).
[66] J. Jacobson, B. M. Collins and B. E. F. Fender, A cta Cryst. B32, 1083 (1976).
[67] S. B. Desu and H. M. ’Bryan, Microwave loss quality of B a (Z n 1/ 3T a 2/ 3)0 3 , J. Am.
Ceram. Soc., 6 8 , 1985, 546-51.
[68] S. Liu, R. Taylor, L. Budd, N. S. Petrovic, M. Van Schilfgaarde, and N.Newman,
Experimental and theoretical investigation of the structural, chemical, electronic and
high frequency dielectric properties of Barium Cadmium Tantalate-based ceramics, J.
Appl. Phy. 97, 014105 (2005).
[69] S. Liu, J. Sun, R. Taylor, D. J. Smith, and N. Newman, M icrostructure and dielectric
properties of B a ( C d i / z T a 2/z)Oz microwave ceramics synthesized w ith a boron oxide
sintering aid J. M ater. Res., 19 (2004) 3526-3533.
[70] Jung-In Yang, Sahn Nham, Chang-Hack choi, Hwack-Joo Lee, and Hyun-Min Park,
M icrostructure and microwave dielectric properties of B a ( C d i /z T a 2/z)Oz ceramics with
Z r O 2 addition. J.Am.Ceram. Soc. 85, 165 (2002).
[71] J. Sun, S. Liu, N. Newman, M.R. McCartney, and D.J. Smith: Electron microscopy
characterization of B a{ Cd i/ zT a 2/z)Oz microwave dielectrics with boron additive. J.
M ater. Res. 19, 1387 (2004).
[72] K.M atsumoto, T. Hiuga, K. Takada, and H. Ichimura: B a (M g 1/ 3T a 2/z)O3 ceramics
with ultra-low loss at microwave frequencies, in Proceedings of the 6th IEEE Interna­
tional Symposium on Application of Ferroelectrics, Bethlehem, PA (IEEE, New York,
NY, June 8-11, 1986), p. 118.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
152
[73] A.K. Ganguii, K.P. Jayadevan, G.N. Subbanna, and K.B.R. Varma:
Prepara­
tion and dilectric properties of zinc and cadamium doped barium tantalates,
B a ( Z n xC d i/ 3 _xT a 2 /z)C>3 . Solid State Commun. 94, 13 (1995)
[74] I.C. Ho, J. Am. Ceram. Soc. 77, 829 (1994).
[75] J.H. Lee, Y.W. Heo, J.A. Lee, Y.D. Ryoo, J.J. Kim, and S.H. Cho, Solid State Ionics,
101-103, 787 (1997).
[76] S.M.Rhim, H. Bak, S. Hong, and O.K. Kim, J. Am. Ceram. Soc. 83, 3009 (2000).
[77] E.M. Levin and H.F. McMurdie, Phase Diagram for Ceramists Supplement (American
Ceramics Society, Inc., 1975), p. 97.
[78] E.C.Subbarao and F.A. Hummel, Phase Diagram for Ceramists (American Ceramics
Society, Inc., 1964), p. 109.
[79] H.M. Chan, M.P. Hamer, and D.M. Smyth: Compensating defects in highly donordoped B aT i03. J. Am. Ceram. Soc. 69, 507 (1986).
[80] G.V. Lewis, C.R.A. Catlow, and R.E.W . Casselton: P T C R effect in B aT i03. J. Am.
Ceram. Soc. 68, 555 (1985).
[81] , L. C. Tien, C. C Chou, and D. S. Tsai, Ordered structure and dielectric properties
of lanthanum -substituted Ba(Mgi/zTa, 2 /z)Oz, J. Am. Ceram. Soc., 83, 2000, 2074-78.
[82] C. C. Lee, C. C. Chou, and D. S. Tsai, Effect of L a/K A-site subsititutions on the
ordering of B a ( Z n i / z T a 2 / 3 )Oz, J. Am. Ceram. Soc., 80, 1997, 2885-90.
[83] I. M. Reaney, I. Qazi, and W. E. Lee, Effect of structural changes in complex perovskites
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
153
on the tem perature coefficient of the relative permittivity, J. Appl. Phys. 88, 6708
( 2000 ).
[84] M. A. Akbas and P. K. Davies, J. Am. Ceram. Soc. 81, 670 (1998).
[85] J. M. Pond, Mobius Dual-mode mesonators and bandpass filters, IEEE Trans. MTT.
48 (2000) 2465-2471.
[86] W ebster’s New Collegiate Dictionary. Springfield, MA: G.&C. M erriam , 1974.
[87] Chartier, T., Ferrato, M. and Baumard, J. F., Supercritical debinding of injection
molded ceramics. J. Am. Ceram. Soc., 1995, 78, 1787-1792.
[88] Ebenho ch, J. and ter M aat, J., Ceramic injection molding with a polyacetal based
binder system. In Advances in Powder M etallurgy and P articulate M aterials, Vol. 5,
Powder Injection Molding. M etal Powder Industries Federation, Princeton, 1993, pp.
73-89.
[89] Tsai, D. S. and Chen, W. W., Solvent debinding kinetics of alum ina green bodies by
powder injection molding. Ceram. Int., 1995, 21, 257-264.
[90] R. German, Powder Injection Molding
[91] D. M. Liu and W. J. Tseng, Influence of debinding rate, solid loading and binder
formulation on the green microstructure and sintering behaviour of ceramic injection
moldings, Ceram. Int., 24 (1998) 471-481.
[92] Yimin Li, Shaojun Liu, Xuanhui Qu, Baiyun Huang, Therm al debinding processing of
316L stainless steel powder injection molding compacts, J. M ater. Proc. Tech., 6487
(2003) 1-5.
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
154
[93] C. L. Huang, K. H. Chiang, S. C. Chuang, Influence of V2 O 5 additions to
B a ( M g i / 3 T a 2 /s ) 0
3
ceramics on sintering behavior and microwave dielectric proper­
ties, M ater.Res.Bulletin, 39 (2004) 629-636.
[94] W. H Rhodes, Agglomerate and particle size effects on sintering yttria-stabilized zirconia. J. Am. Ceram.Soc., 1981, 64, 19-22.
[95] Y. M Chang, D. P Birnie, W. D Kingery, Physical ceramics: principles for ceramic
science and engineering, John Wiley and Sons, New York, 1997, pp. 371-388.
[96] Chen H. C., Weng, M. H., Horng, J. H., Houng, M. P. and Wang, Y. H., Effect of
Bismuth Addition on Sintering Behavior and Microwave Dielectric Properties of Zinc
T itanate Ceramics, J. Electron. Mater., 34, 2005, 119.
[97] Chu, S. Y. and Chen, T. Y., Fabrication of modified lead titan ate piezoceramics with
zero tem perature coefficient and its application on SAW devices, IEEE Trans. Ultrason.
Ferroelectr. Freq. Control., 51, 2004, 663 - 667.
[98] Reaney, I. M., Qazi, I. and Lee, W. E., Order-disorder behavior in B a ( Z n i / z T a 2 /z)Oz,
J. Appl. Phy., 88, 2000, 6708-6714.
[99] Reaney I. M., Colla E. L. and Setter N., Dielectric and structural characteristics of
Ba- and Sr-based complex perovskites as a function of tolerance factor, Jap. J. appl.
Phys., 1994, 33, 3984-3990.
[100] R.Waser, T. Baiatu, and K.H. Hardtl, dc electrical degradation of perovskite-type
titanates: I, J.Am. Ceram. Soc., 73, 1990, 1645-53.
[101] R. M. Waser, Electrochemical boundary conditions for resistance degradation of doped
Alkaline-earth titanates, J.Am. Ceram. Soc., 72, 1989, 2234-40.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
155
[102] S. W. Freiman and R. C. Pohanka, Review of mechanically related of ceramic capac­
itors and capacitor materials, J.Am. Ceram. Soc., 72, 1989, 2258-63.
[103] K. Lehovec and G. A. Shirn, Conductivity injection and extraction in polycrystalline
barium titanate, J. Appl. Phys., 33, 1962, 2036-44.
[104] S. H Yoon, J. H. Lee, D. Y. Kim, and N. M Hwang, Effect of the liquid-phase character­
istic on the m icrostructure and dielectric properties of donor- (Niobium) and acceptor(magnesium) doped barium titanate, J.Am. Ceram. Soc., 86, 2003, 88-92.
[105] U. Weber, G. Greuel, U. Boettger, S. Weber, D. Hennings, and R. Waser, Dielec­
tric properties of B a ( Z r ,T i) O z -based ferroelectrics for capacitor applications, J.Am.
Ceram. Soc., 84, 2001, 759-66.
[106] Y. S. Jung, E. S. Na, U. Y. Paik, J. H Lee, and J. H. Kim, A study on the phase
transition and characteristics of rare earth elements doped Ba TiO z, Mater.Res.Bull.,
37, 2002, 1633-1640.
[107] S. M. Rhim, S. K. Hong, H. J Bak, and O. K. Kim, Effects of B aT iO z addition on the
dielectric and ferroelectric properties of Ba o^Sro^TiOz ceramic, J.Am. Ceram. Soc.,
83, 2000, 1145-48.
[108] B. Su and T. W. B utton, M icrostructure and dielectric properties of Mg-doped barium
strontium titan ate ceramics, J. Appl. Phy., 95, 2004, 1382-1385.
[109] W. Osak and K. Tkacz, Investigation of I-V characteristics in polycrystalline BaTiOz,
J. Phy.D.: Appl. Phys., 22, 1989, 1746-50.
[110] V. M. Fridkin and K. Kreher, Theory of space-charge-limited currents in ferroelectrics,
Phy. Stat. Sol. (a), 2, 1970, 281-285
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
156
[111] Y. S. Lee and H. I. Yoo, Current-voltage characteristic of B a T i O z s in its mixed n /p
regime under oxygen potential gradients, Solid State Ionics, 150, 2002, 373-382.
[112] O. G. Vendik,
S. P. Zubko, and M. A. Nikol’ski, Microwave loss-factor of
B ax S r \ - xTiOz as a function of tem perature, biasing field, barium concentration, and
frequency, J. Appl. Phy., 92, 2002, 7448-7452.
[113] Y. P ark and H. G. Kim, dielectric tem perature characteristics of cerium-modified
barium titan ate based ceramics with core-shell grain structure, J.Am.ceram.Soc., 80,
1997, 106-112.
[114] L. Wu, Y. C Chen, C. L Huang, Y. P Chou, and Y. T Tsai, D irect-Current Field
Dependence of Dielectric Properties in Alumina-Doped Barium Strontium T itanate, J.
Am. Ceram. Soc., 83, 2000, 1713-19.
[115] J. W Liou and B. S Chiou, Effect of direct-current biasing on the dielectric properties
of Barium Strontium T itanate, J. Am. Ceram. Soc., 80, 1997, 3093-99.
[116] J. H Hwang and Y. H Han, Defect chemistry of E r-doped BaTiOz, Solid State Ionics,
140, 2001, 181-186.
[117] D. M Smyth, M. P Harmer, and P. Peng, Defect chemistry of relaxor ferroelectrics and
the implications for dielectrics degradation, J. Am. Ceram. Soc., 72, 1989, 2276-78.
[118] D. M Smyth, Comments on the defects chemistry of undoped and acceptor-doped
BaTiOz, J. Electroceram., 11, 2003, 89-100.
[119] H. A. Jahn and E. Teller, Stability of polyatomic molecules in degenerate electronic
states. I. orbital degeneracy, Pro. Roy. Soc., A 161, 1937, 220-235.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
157
[120] H. T. Langham m er,T Muller, K. H. Feigner,+ and H. P. A bicht,Crystal Structure and
Related Properties of Manganese-Doped Barium T itanate Ceramics, J. Am. Ceram.
Soc. 83, 605 (2000).
[121] T. Nagai, K. Iijima, H. J. Hwang, M. Sando, T. Sekino, and K. Niihara, Effect of MgO
doping on the phase transform ation of B a T iO z , J. Am. Ceram. Soc. 83, 107 (2000).
[122] G. Y. Yang, E. C. Dickey, C. A. Randall, D. E. Barber, P. Pinceloup, M. A. Henderson,
R. A. Hill, J. J. Beeson, and D. J. Skamser, Oxygen nonstoichiometry and dielectric
evolution of B a T i O z ■ P art I: improvement of insulation resistance with reoxidation,
J.Appl.Phy.96, 7492(2004).
[123] H. Y. Lee, K. C. Lee, J. N. Schunke, and L. C. Burton, Leakage currents in multilayer
ceramic capacitor, IEEE Trans.Components, Hybrids, Manuf. Technol. CHMT-7, 443,
(1984).
[124] W. D. Luo, W. H. Duan, S. G. Louie, and M. L. Cohen, Phy. Rev. B. 70, 2004,
p214109
[125] S. M. Rhim, S. Hong, H. Bak, and 0 . K. Kim, Effects of B 2 O 3 addition on the
dielectric and ferroelectric properties of B ao j S ro ^ T iO z ceramics, J. Am. Ceram. Soc.
83, 1145 (2000).
[126] U. Weber, G. Greuel, U. Boettger, S. Weber, D. Hennings, and R. Waser, Dielec­
tric properties of Ba( Zr , T i)0 3 -Based ferroelectrics for capacitor applications, J. Am.
Ceram. Soc. 84, 759 (2001).
[127] Y. S. Jung, E. S. Na, U. Paik, J. Lee, and J. Kim, A study on the phase transition
R e p r o d u c e d with p e r m issio n o f th e co p y r ig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
158
and characteristics of rare earth elements doped B a T i O 3, M ater. Res. Bull. 37, 1633
( 2002 ).
[128] A. Cavallini, B. Fraboni, W. Dusi, and M. Zanarini, Defective states induced in CdTe
and C d Z n T e detectors by high and low energy neutron irradiation, J. Appl. Phy. 94,
3135 (2003).
[129] G. Arlt, D. Hennings, and G. de W ith, Dielectric properties of fine-grained barium
titan ate ceramics, J. Appl. Phy. 58, 1619-25, 1985
[130] M. Satoh, K. Kuriyama, M. Yahagi, K. Iwamura, C. Kim, T. Kawakubo, K. Yoneda,
and I. Kimura, Influence of fast neutrons on electrical properties in neutron transm u­
tation doped GaAs: New annealing stage, Appl. Phys. Lett. 50, 580 (1987)
[131] C. E. Barnes, N eutron damage in SiC light-em itting diodes, Appl. Phys. Lett. 20, 86
(1972).
[132] P. Padmini, M. Pulikkathara, R. Wilkins, and R. K. Pandey, N eutron radiation effects
on the nonlinear current-voltage characteristics of ilm enite-hem atite ceramics, Appl.
Phys. Lett. 82, 586 (2003).
[133] X. Qi, J. Dho, R. Tomov, M. G. Blamire, snd J. L. MacManus-Driscoll, Appl. Phys.
Lett. 86, 062903 (2005).
[134] M. A. Lam pert, Simplified theory of space-charge-limited currents in an insulator with
traps, Phy.Rev. 103, 1648 (1956)
[135] R. W. Smith, and A.Ross, Phy.Rev. 97, 1531 (1955).
[136] D. Bolten, U. Bottger, and R. Waser, Frequency and tem perature dependence of the
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
159
relative perm ittivity in ferroelectrics: Monte-Carlo simulation study, J. Appl. Phy., 93,
2003, 2890-2894.
[137] H. Yang, B. Chen, X. Qiu, and B. Zhao, Tem perature- and field-dependent leakage
current of P t / ( B a o ^ S r o ^ T i O s ) interface, Appl.Phy.Lett. 83, 1611 (2003)
[138] A. Rose, Space-charge-limited currents in solids, Phy.Rev. 97, 1538 (1955)
[139] T. Kolodiazhnyi and A. Petrie, Analysis of point defects in polycrystalline BaTiOz
by electron param agnetic resonance, J. Phy.Chem.Solid, 64(2003) 953-960
[140] T.R.N. Kutty, P. M urugaraj, N.S. Gajbhiye, Activation of trap centers in PTC
BaTiOz, M ater. Lett. 2 (1984) 396-400.
[141] S. Jida, T. Miki, Electron param agnetic resonance of Nbdoped B aT iO z ceramics with
positive tem perature coefficient of resistivity, J. Appl. Phys. 80 (1996) 5234-5239.
[142] T. Miki, A. Fujimoto, S. Jida, An evidence of trap activation for positive tem pera­
ture coefficient of resistivity in BaTiOz ceramics with substitutional Nb and Mn as
impurities, J. Appl. Phys. 83 (1998) 1592-1603
[143] N.S. Hari, P. Padmini, T.R.N. Kutty, Complex impedance analyses of n-BaTiOz
ceramics showing positive tem perature coefficient of resistance, J. M ater. Sci.-Mater.
Electron. 8 (1997) 15-22.
[144] D. Makovec, Z. Samardzija, U. Delalut, D. Kolar, Defect structure and phase relations
of highly lanthanum -doped barium titanate, J. Am. Ceram. Soc. 78 (1995) 2193-2197.
R e p r o d u c e d with p e r m issio n o f th e co p y rig h t o w n e r . F u rth er rep ro d u ctio n p roh ib ited w ith o u t p e r m issio n .
Документ
Категория
Без категории
Просмотров
0
Размер файла
4 754 Кб
Теги
sdewsdweddes
1/--страниц
Пожаловаться на содержимое документа