close

Вход

Забыли?

вход по аккаунту

?

Thallium-based high temperature superconducting thin films for microwave electronics applications

код для вставкиСкачать
INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI
films the text directly from the original or copy submitted. Thus, some
thesis and dissertation copies are in typewriter face, while others may
be from any type of computer printer.
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 bleedthrough, substandard margins,
and improper alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI 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.
Oversize materials (e.g., maps, drawings, charts) are reproduced by
sectioning the original, beginning at the upper left-hand comer and
continuing from left to right in equal sections with small overlaps. Each
original is also photographed in one exposure and is included in
reduced form at the back of the book.
Photographs included in the original manuscript have been reproduced
xerographically in this copy. Higher quality 6" x 9" black and white
photographic prints are available for any photographs or illustrations
appearing in this copy for an additional charge. Contact UMI directly
to order.
University Microfilms International
A Beil & Howell Information Com pany
3 0 0 North Z eeb Road. Ann Arbor. Ml 48106-1346 USA
313/76 1-4 700
80 0/5 2 1 -0 6 0 0
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
O rd e r N u m b e r 9 407819
T h alliu m -b ased high tem p era tu re su p ercon d u ctin g th in film s for
m icrow ave electronics a p p lication s
Subramanyam, G urunathan, P h.D .
University of Cincinnati, 1993
UMI
300 N. ZeebRd.
Ann Arbor, MI 48106
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UNIVERSITY OF CINCINNATI
, 19.93__
1:
n in -n n a th a n
SuhramanvaTii
.
hereby submit this as part o f the
requirements for the degree of:
Doctor of Philosophy
i n Electrical Engineering
It iS
Thallium Based High Temperature
Superconducting Thin Films for Microwave Electronics
Applications
mroved by:
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
THALLIUM BASED HIGH TEMPERATURE SUPERCONDUCTING THIN FILMS
FOR MICRO WA VE ELECTRONICS APPLICA TIONS
A Dissertation submitted to the
Division of Graduate Studies and Research
of the University of Cincinnati
In Partial Fulfillment of the Requirements for the Degree of
DOCTOR OF PHILOSOPHY
Department of Electrical and Computer Engineering
College of Engineering
Division of Graduate Studies
University of Cincinnati
July 1993
by
GURUNATHAN SUBRAMANYAM
B.E., University of Madras, India, December 1984
M.S., University of Cincinnati, December 1988
Committee Chair: Professor Vik J. Kapoor
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
MEMBERS OF THE Ph.D. DISSERTATION COMMITTEE
PROF. VIK J. KAPOOR (CHAIRMAN)
DEPARTMENT OF ELECTRICAL AND COMPUTER ENGINEERING
PROF. PUNIT BOOLCHAND
DEPARTMENT OF ELECTRICAL AND COMPUTER ENGINEERING
PROF. ALTAN M. FERENDECI
DEPARTMENT OF ELECTRICAL AND COMPUTER ENGINEERING
PROF. MARC M.CAHAY
DEPARTMENT OF ELECTRICAL AND COMPUTER ENGINEERING
PROF. DARL H. McDANIEL
DEPARTMENT OF CHEMISTRY
Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission.
DEDICATED
TO THE MEMORY OF
M Y BELOVED MOTHER
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ABSTRACT
The main objective of the dissertation research is to investigate thallium
based high temperature superconducting thin films for microwave electronics
applications.
The preparation, structural properties, electrical transport
properties, and microwave properties of the superconducting thin films were
investigated in this research.
The superconducting thin films were also
investigated for superconductor semiconductor hybrid phase shifter circuits.
TI-Ca-Ba-Cu-0 high temperature superconducting thin films were deposited on
lanthanum aluminate substrates, by rf magnetron sputtering and post-annealing
methods.
A reproducible fabrication process with low resistance metal
contacts has been established for high critical temperature(Tc) and high critical
current density(Jc) TICaBaCuO thin films after patterning using standard
microelectronic photolithography and w et chemical etching techniques. Low
resistance gold contacts on TICaBaCuO thin films were obtained by annealing
in an oxygen flow of 1 liter/minute followed by a slow furnace cooling.
Specific contact resistivity was approximately 10'8 Ohm-cm2 below 77 K. Tc
as high as 100 K, and Jc at zero magnetic field greater than 105 A/cm 2 at 77
K are routinely obtained in 0.3 -0 .5 //m TICaBaCuO thin films. The microwave
properties of TICaBaCuO thin films were investigated by designing, fabricating
and characterizing microstrip ring resonators with a fundamental resonance
frequency of 12 GHz on 10 mil thick lanthanum aluminate (LaAI03) substrates.
i
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Ring resonators with superconducting ground plane of 0 .3 pm thickness and
gold ground plane of 1 pm thickness were fabricated and characterized in the
temperature range of 60-95 K. Typical unloaded quality factor (Q) for the ring
resonators at 12 GHz were above 1500 at 65 K, compared to an unloaded Q
of 370 for a gold ring resonator. The surface resistance as low as 1.5 mQ at
12 GHz and 77 K was obtained in 0 .3 pm TICaBaCuO thin films using the ring
resonator Q measurements. Typical values of penetration depth at 0 K in the
TICaBaCuO thin films were determined to be between 7000 and 8 0 0 0 A using
the temperature variation of resonance frequency measurements.
A
hybrid phase shifter circuit based on high T c TI-Ca-Ba-Cu-0
superconducting thin films and GaAs MESFETs was investigated.
A 180
degrees phase bit designed for a 10 degrees phase error in the bandwidth 3 .5 4 .5 GHz showed a phase shift of 180 degrees near the center frequency of 4
GHz. The insertion loss in the on-state of the devices was as small as 1.76 dB
at 4 GHz and 7 0 K for the superconducting phase shifter circuit compared to
the lowest insertion loss of 3.1 dB obtained in a gold based hybrid circuit. The
improvement in insertion loss is mainly due to the lower surface resistance and
lower conductor losses in the superconductor based microstrip circuits. The
insertion loss in the off state of the superconducting circuit was as small as 2 .9
dB in the off state of the MESFETs as compared to 3 .2 7 dB for a gold based
circuit in the bandwidth between 3.5 and 4.5 GHz. The improvement is again
due to lower conductor losses in the superconducting thin films.
ii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
A CKNO WLED GEMENTS
I sincerely thank my advisor, Professor Vik J. Kapoor for suggesting this
dissertation topic and providing constant guidance, inspiration and support
during the course of this program.
His assistance in publications and
presentations of the research work is gratefully appreciated. Special thanks to
Professor Boolchand for his constant guidance and inspiration during our
collaborative work. Thanks are also extended to Professors Dr. Marc Cahay,
Dr. Altan Ferendeci, and Dr. Dari McDaniel for serving on my dissertation
committee and for their support in this program. I also thank Professor Dr.
Farhad Radpourfor his useful suggestions and discussions during the research
work.
I would like to thank the support and constant encouragement of Dr.
Regis Leonard, Dr. Kul Bhasin and Dr. John Pouch of NASA Lewis Research
Center.
Their guidance throughout the course of this project is gratefully
acknowledged. I also thank Mr. Chris Chorey, Dr. Felix Miranda, Dr. Mark Stan
and Dr. Sam Alterovitz of NASA Lewis Research Center for their guidance in
measurements and useful discussions. I sincerely thank Dr. David Ginley and
his colleagues at Sandia National Laboratory for their assistance during the
initial stages of this project.
I would like to thank my colleagues, past and present, Mr. George
Lemon, Dr. Mike Beidenbender, Dr. Gregory Johnson, Dr. Mohsen Shokrani, Mr.
iii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Alan Hoofring, Dr. Chung Kun Song, Mr. Gregory De Brabender, Dr. Kevin
Walsh, Dr. Chuck Blue, Mr. Homyar Mogul, Mr. Shi Lin Lu, Mr. H. Chang, Dr.
Min Zhang, Mr. Dan Dotson, Mr. Ron Flenniken, Ms. Leslie Ries and Ms. Peggy
Simpson for their help in various parts of this research program. Special thanks
to Ms. Teresa Hamad for her assistance and cooperation throughout this
research program.
Finally I would like to thank my beloved parents, my sisters, my brother
and sister-in-law, my brother-in-law, my favorite nephew Amrith, everybody in
our family, and friends in Cincinnati (especially the UC cricket team) for their
love, constant encouragement and support.
iv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
TABLE OF CONTENTS
PAGE
Abstract
i
Acknowledgements
iii
List of figures
viii
List of tables
xi
CHAPTER 1. INTRODUCTION AND LITERATURE SEARCH
1
1.1
High Temperature superconductors
1
1.2 Chemistry of TICaBaCuO superconductors
4
1.3 TICaBaCuO superconducting electronics
7
1.4 Objectives of the research program
17
CHAPTER 2. RF SPUTTERING OF TI-Ca-Ba-Cu-0 THIN FILMS
2.1
20
Sputtering system
20
2.2 Sputtering targets
22
2.3 Sputtering gases
26
2.4 Substrates
29
2.5 Sputtering conditions
29
2.6 Post-deposition processing of thin films
30
v
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Page
CHAPTER 3. STRUCTURAL CHARACTERIZATION OF THIN FILMS
35
3.1 Scanning Electron Microscopy
35
3.2 X-ray Diffraction
35
3.3 Auger Electron Spectroscopy
41
3 .4 High Resolution X-ray Photoelectron
Spectroscopy
45
CHAPTER 4. ELECTRICAL TRANSPORT PROPERTIES OF TICaBaCuO
SUPERCONDUCTING THIN FILMS
4.1
CHAPTER 5.
Patterning of TICaBaCuO thin films
57
57
4.2 Electrical contacts on superconducting thin films
59
4.3 Resistance vs temperature measurements
60
4 .4 Electrical transport measurements
67
TICaBaCuO SUPERCONDUCTING MICROSTRIP RING
RESONATORS
5.1
72
Fabrication and Testing
74
5.2 Unloaded Q vs Temperature
78
5.3 Resonance frequency vs Temperature
80
5.4 Analysis and discussions
82
vi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Page
CHAPTER 6. SUPERCONDUCTOR-SEMICONDUCTOR HYBRID
PHASE SHIFTER CIRCUITS
90
6.1
90
Reflection type hybrid phase shifter circuits
6.2 Design
94
6.3 Experimental
105
6.4 Results and discussions
107
CHAPTER 7. SUMMARY AND CONCLUSIONS
120
BIBLIOGRAPHY
124
APPENDIX A
140
APPENDIX B
143
vii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
LIST OF FIGURES
Page
2.1
Schematic diagram for the sputter-up configuration
21
2 .2
Photograph of the sputtering system
22
2 .3
Magnetic susceptibility vs Temperature for a bulk 2 1 2 2 Tl
superconductor
25
2 .4
Powder x-ray diffraction spectra for a bulk 2122 sample
2.5
Post-deposition methods (a) Free surface configuration and
(b) Confined surface configuration
3.1
32
X-ray diffraction spectrum of an annealed TICaBaCuO thin
film on a SrTiOs substrate
3 .2
27
36
X-ray diffraction spectrum of an annealed TICaBaCuO thin
film on a LaAI03 substrate
37
3 .3
SEM Micrograph of a TICaBaCuO film on a SrTi03 substrate
39
3 .4
SEM Micrograph of a TICaBaCuO film on a LaAI03 substrate
40
3 .5
AES survey of a TICaBaCuO thin film on a SrTi03 substrate
42
3 .6
AES depth profiling obtained on the TICaBaCuO film
44
3 .7
AES depth profile of a TICaBaCuO film on a LaAI03 substrate
46
3 .8
High resolution XPS for Tl obtained from a TICaBaCuO thin
3 .9
film on a SrTi03 substrate
48
High resolution XPS for Ca obtained from the TICaBaCuO film
50
viii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Page
3 .1 0 High resolution XPS for Ba obtained from the TICaBaCuO film
51
3.11 High resolution XPS for Cu obtained from the TICaBaCuO film
53
3 .1 2
High resolution XPS for 0
54
4.1
Geometry of four probe test devices for electrical transport
measurements
58
4 .2
Photograph of a fabricated four probe device with Au contacts
61
4 .3
Schematic diagram for the closed cycle cryogenic system
62
4 .4
Measurement setup for dc electrical transport measurements
64
4 .5
Resistivity vs Temperature measurements on a four probe
66
test device
4 .6
Measurement set-up for pulsed current technique
4 .7
Zero-field critical current density vs Temperature
68
measurements
70
5.1
Geometry of a microstrip ring resonator for 12 GHz
73
5.2
Photograph of a superconducting microstrip ring resonator
75
5.3
Swept frequency reflected power measurement performed
on an all-superconducting ring resonator
5 .4
77
Resonator unloaded Q vs Temperature characteristics of a
TICaBaCuO superconducting ring resonator
ix
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
79
Page
5 .5
Resonance frequency vs Temperature characteristics of a
TICaBaCuO superconducting ring resonator
81
5 .6
Attenuation vs Temperature characteristics for >1(0) = 6 0 0 0 A
86
5 .7
Attenuation vs Temperature characterisitics for A(0) = 7 0 0 0 A
87
5 .8
Surface resistance vs Temperature characteristics
88
6.1
Schematic of a Lange Coupler for 3 dB coupling
92
6 .2
(a) GaAs MESFET switching characteristics (i) on state
insertion loss (ii) off-state isolation and (b) equivalent circuit
for the on and off states.
6 .3
97
Effective characteristic impedance vs Temperature with
thickness as a parameter
100
6 .4
Schematic of a hybrid 18 0° phase bit using a Lange Coupler
102
6.5
Design page from HPMDS (a) off state of the MESFETs
103
and (b) on state of the MESFETs
104
6 .6
Theoretical results from HPMDS simulation for the hybrid
circuit
106
6 .7
Photograph of a superconductor semiconductor hybrid circuit
108
6.8
Photograph of the superconducting circuit
109
6 .9
Measured Phase shift vs frequency characteristics for a super­
conducting 180 degrees phase bit
x
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
110
6 .1 0
Phase shift vs frequency characteristics fo ra gold based 1 8 0°
phase bit after bonding alternate Lange coupler fingers
6.11
Temperature dependence of center frequency for the
superconducting phase bit.
6 .1 2
Insertion loss vs frequency characteristics for a super­
conductor based hybrid circuit in the on state of the GaAs
MESFETs
6.1 3
Insertion loss vs frequency characteristics for a super­
conductor based hybrid circuit in the off state of the GaAs
MESFETs
6 .1 4
Insertion loss vs temperature characteristics for the super­
conducting hybrid 180 ° phase bit
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
List of Tables
Page
Table 2.1
Optimum conditions for sputtering of TICaBaCuO films
31
Table 3.1
Summary of high resolution XPS studies
56
Table 6.1
Experimental results summarized for gold based 180
degrees hybrid phase bits
Table 6.2
118
Experimental results summarized for superconductor
based 180 degrees hybrid phase bits
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
119
1. INTRODUCTION AND LITERATURE SEARCH
1.1 High Temperature Superconductors:
Since the discovery of a copper-oxide high transition temperature(Tc)
superconductor by K.A.Muller and J.G.Bernorz in January 1986[1], there has
been substantial progress in superconducting electronics.
Several new
compounds such as YBaCuO [2], BiSrCaCuO [3], and TICaBaCuO [4] have been
found to be superconducting above 90 K, thus making it feasible for electronic
applications operating at liquid nitrogen temperature (77 K).
An advantage of
using liquid N2 as compared to liquid helium in superconductor electronics is the
slower boil-off rate of liquid N2, so that one can design cryostats of modest
size. Also, liquid nitrogen is at least ten times cheaper and hence economically
feasible for numerous applications. Operation of electronic systems at 77 K is
advantageous in reducing the noise, improving overall system performance
especially for communication systems.
Since semiconductor circuits and
systems offer improved performance at 77 K, hybrid circuits and systems using
the high Tc superconductors and the existing semiconductors are feasible.
The high T c superconducting materials are a new class of materials
which posess entirely new electrical, structural and metallurgical properties
compared to conventional low T c superconductors. The new high Tc materials
are superior to conventional low temperature superconductors in two ways.
One is their high T c, and second is their high critical magnetic field. There are
several problems in the use of high Tc superconductors, especially in the
Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission.
development of these materials for electronic applications, mainly because of
their complex nature. The synthesis of high T c compounds to obtain the right
stoichiometry and phase is very difficult.
Presence of secondary phases
hindered most of the high Tc materials. Another fundamental problem in high
T c materials is their anisotropy, ie., they possess different structural and
electrical properties in different directions.
The superconducting properties
such as critical current density (Jc) and critical magnetic field (Hc) along the a-b
planes are superior compared to those along the c-axis. A major challenge for
researchers was to develop a fabrication technology imposing the condition that
electrical conduction will be along the a-b planes. The high Tc materials also
exhibit higher penetration depths compared to conventional low T c materials.
The coherence length, a length scale that characterizes superconducting
electron pair coupling, is very short in high Tc materials, increasing the difficulty
of making Josephson Junctions[5j.
The attractiveness of high Tc materials overshadows the demerits and
currently, worldwide research is underway for developing suitable thin film
technologies for possible electronic applications of high T c materials. Already
rapid progress has been made for applications of high Tc materials in areas such
as
Superconducting
microwave
devices,
Quantum
infrared(IR)
Interference
detectors,
Devices(SQUIDs),
and
passive
interconnections
in
microelectronics[6-9].
Among the high T c materials, the TICaBaCuO compound has proven to
2
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
possess the highest T c, 125 K [5]. Because of the highest T c, this material
offers a wide margin of operational range for electronic applications at 77 K.
The TICaBaCuO thin films are very attractive for electronic applications, as they
have shown T c as high as 120 K[10], and Jc greater than 105 A/cm2 at 77
K[7].
The TICaBaCuO thin films are chemically very stable compared to
YBaCuO and BiSrCaCuO[12].
Another advantage is that multiple phases,
especially TI2Ca2Ba2 Cu3Ox (2223) and TI2Ca.,Ba2Cu2Ox (2122) phases, can
co-exist, without degrading the superconducting properties in thin films[13].
Multiphases in polycrystalline thin films of TICaBaCuO have aided in improving
the Jc to greater than 105 A/cm2[14J.
The major disadvantage of TICaBaCuO compound is the toxicity of Tl
which needs very careful processing and handling procedures. The volatility of
Tl gives rise to Tl loss during the high temperature heat treatments. The high
volatility of Tl also causes the non-homogeneous nature of the TICaBaCuO
phases. Most of the Tl based superconducting phases contain a large number
of structural defects which affect the superconducting properties of the Tl
superconducting compoundsd 5].
Thin film fabrication of TICaBaCuO superconductors has been reported
using several techniques such as electron beam evaporation, sputtering, pulsed
laser ablation, and metal organic chemical vapor deposition (MOCVD)[10-17].
Electron beam evaporation is easily adaptable to new materials, but lacks the
precise control of stoichiometry over a large area[17]. The dc and rf sputtering
3
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
techniques are widely used for deposition of Tl films[12-14].
Single target
sputtering is especially promising since target composition can be adjusted to
compensate for the losses that occur during deposition and post-annealing
processes. The pulsed laser ablation technique has produced high quality films,
but is limited by the laser spot size.
The highest Tc reported to date in
TICaBaCuO thin films (120 K) was obtained using a multiple target sputtering
technique[10]. Tl based superconducting thin films are usually fabricated by
post-processing at high temperatures between 830 and 900 °C, for several
minutes[18]. Recently, the Du Pont research group has demonstrated in-situ
growth of Tl based superconducting thin films[19]. The TIBa2CaCu20 7 thin
films have been grown epitaxially on singie crystal LaAI03, NdGa03 and C e02
buffered saphire substrates by in-situ off-axis magnetron sputtering in the
presence of Tl20 vapor. A Tc as high as 97 K has been obtained for in-situ
annealed Tl films[19].
1.2 Chemistry o f TI2Ca1Ba2 Cu20x superconductors:
Processing of single phase TI2Ca.jBa2Cu20x samples is an important
research area to optimize the superconducting properties of each phase. The
volatility of Tl at the high processing temperatures necessitates careful
treatment procedures.
It is important to understand the correlation of high
temperature chemistry with processing variables such as starting composition,
heat treatment temperature, and time.
There has been a large number of
4
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
studies on 2223 samples previously [20-23].
Previous studies on various
superconducting phases indicate reversible reactions between various phases
of the TICaBaCuO system [24-25]. For example, TI2122 phase can be formed
from TI2223 phase by adding additional Tl20 3 source to TI2223 pellets[16].
The reverse reaction is also true, ie., by removal of Tl20 3 from the 2 1 22 pellets
22 2 3 phase can be formed[25].
The 2 1 22 and 2 2 2 3 phases can also be
formed by diffusion of Tl into the CaBaCuO precursor pellets[25]. In all the
cases reported earlier, the main understanding is that the Tl20 content in the
crucible determines the phase formation. Previous experimental results suggest
rapid reaction between Tl20 3, BaCu02 and CaO powders. The reaction speed
is reported to be almost the same as the loss of Tl20 3 by sublimation in the
applied conditions [28]. The experimental results indicate that the optimum
composition for synthesis of TI2223 phase is 2:2:2:3.
In the author's study, T I2122 thin films were fabricated from both 2122
and 2223 sintered superconducting compounds.
In the case of thin films
deposited from TI2223, the post-processing heat treatments had to be
performed in excess Tl20 partial pressure, in the free surface sintering
configuration[11]. The films deposited from 2122 had to be heat treated in the
confined surface sintering configuration[11]. The loss of Tl had to prevented
from the thin films in the later case. In both cases, the heat treatment times
and temperatures were optimized to obtain the smooth morphology, high phase
purity, superior electrical and microwave properties.
5
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Structural studies of TICaBaCuO compounds are very important in
understanding the complex chemistry.
X-ray diffraction (XRD) technique is
useful in determining the crystal structure, unit cell dimensions and also the
presence of multiple phases if any. Other techniques such as Transmission
Electron microscopy(TEM), Scanning Tunneling Microscopy(STM) etc are used
in
determining
the
microstructure
and
defects
in
the
high
Tc
superconductors[20, 29].
Greenblatt et a I [20], have studied the synthesis, structural, and
electronic
properties, and the correlation between
superconducting properties.
the chemistry and
Reduced oxidation state of Tl, and increased
valence of Cu, has been reported by several researchers[21, 23].
This
particular finding could be a possible mechanism for introducing holes in the
CuO layers and hence the origin of superconductivity in these compounds.
Also Ca substitution in Tl sites has been reported by Hiraga et al [29] using
high resolution electron microscopy.
Recently, the author verified the 2 +
valence state of Cu, and a valence state of Tl between + 1 and + 3 in a
TICaBaCuO thin film, using xray photoelectron spectroscopy (XPS)[30]. From
the current studies, the 2 + valence state of Cu and the valence state of Tl
between + 1 and + 3 seem to be the key in hole creation on the Tl based
superconductors.
Defects in the crystal structure are abundant in the
TICaBaCuO compounds, and these do affect the physical and superconducting
properties in these compounds.
The studies in the structural properties of
6
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
TICaBaCuO thin films were performed using XRD, AES and XPS. The XRD
studies confirmed that the major phase in the TICaBaCuO thin films fabricated
in-house, is the 2 1 22 phase. TI2223 phase was present in the thin films. The
AES was used to study the surface chemistry, compositional uniformity, and
also the thin film-substrate interaction. XPS was used to study the chemistry
in the TICaBaCuO thin films, by determining the chemical state of Tl, Ca, Ba,
Cu and 0 .
These studies confirmed the 2 + valence state of Cu and also a
reduction in valence state of Tl, between + 1 and + 3 [3 1 ].
1.3 TICaBaCuO Superconducting Electronics:
An important requirement of high Tc thin films for electronic applications
is the ability to pattern thin films into fine line features without degrading the
superconducting properties. Shih et al[32], reported the fabrication of fine line
TICaBaCuO structures from rf sputtered CaBaCuO thin films. Patterning was
performed by standard photolithography and wet chemical etching in a weak
acid. By heat-treating the BaCaCuO thin films with a Cu protective layer in 0 2
and Tl ambient, superconducting thin films were obtained on cubic zirconia
substrates. A T c of 70 K was reported for a 150 pm wide line, 1.4 pm thick
film. Other etching techniques such as a non-aqueous chemical etch described
by Vasquez et al[33], have been successfully used for patterning high Tc thin
films. The non-aqueous solution consisted of 1 % by volume Br2 in absolute
ethanol. Pulsed laser etching using a pulsed excimer laser (248nm, 30 secs)
7
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
[34], has also been successfully used for patterning high Tc thin films.
Recently, focused ion beam lithography has been successfuly used for
patterning high T c thin films[35].
The author has developed a patterning
process for as-deposited TICaBaCuO thin films, using standard positive resist
photolithography and w et chemical etching techniques[36].
Patterned films
were post-processed to obtain superconductivity. A 1:100 phosphoric acid:DI
H20 solution was used for etching the thin films.
Critical current density (Jc) is one of the most important superconducting
properties determined in thin films.
It is an estimate of how much current
density a superconductor can transport before becoming normal. High critical
currents at the level of magnetic field needed for specific applications is an
important pre-requisite for high Tc superconductors. Jc greater than 10 5 A/cm 2
at 77 K is required for possible electronic applications. Jc of a superconductor
can be estimated in two different ways.
One is the Bean's critical state
mode![36], extracting Jc from magnetization data, and the second is the
transport current measurements using four probe devices.
The transport
current method is very reliable and usually lower than the evaluated Jc from
Bean model.
Although high Jcs are being achieved in thin films of high T c
superconductors, several fundamental problems are yet to be understood
consistently.
These problems are: 1. grain boundaries, and metallurgical
defects, 2. flux pinning mechanisms, and 3.
flux creep[37-38].
All these
problems influence the Jc of high Tc superconductors, due to their higher
8
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
transition temperature and short coherence length.
In general, high T c thin
films with epitaxial growth on substrates, and a microstructure with a smooth
morphology and grains smaller than 0.5 //m have yielded the highest Jc. One
of the important objectives of the author was to obtain such smooth
morphology and microstructure that gives high Jc.
The highest Jc obtained in TICaBaCuO thin films on LaAI03 substrates
is 1 .0 6 * 1 0 6 A/cm2 at 77 K[39j. In that work, a microbridge 20 //m wide by
100 jjm long, was used for the transport measurements using an electric field
criterion of 1 //V/mm.
The author's Jc measurements were performed using
a 50 //m wide by 1 mm long four-probe test devices. DC and pulsed current
methods were used for determining Jc, subject to the electric field criterion of
1 //V/m m.
One of the main problems in TICaBaCuO thin films has been the
fabrication of low resistance contacts for reliable electrical measurements on
thin films.
Low resistance contacts with specific contact resistance in the
range of 10‘8 ohm cm2 and 10 '10 ohm cm2 at 77 K, have been reported on
superconducting thin films[40-42j. The process involves evaporation of gold
or silver on the films, followed by annealing the film in oxygen between 500
and 600 °C for several hours. Tl thin films pose problems for low resistance
contacts mainly because of their non-uniform surface. A reliable method for
the fabrication of low resistance contacts has been established in our
laboratory.
Contact resistance below 10'6 ohm cm2 have been rountinely
9
Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission.
obtained.
The foremost applications of high Tc thin films are expected to be in the
area of 'passive microwave devices' such as resonators, filters and delay
lines[8-9].
High Tc superconductors have a greater impact on selected
microwave devices because of two important properties that differ from normal
metals at high frequencies. One is the lower surface resistance in high Tc thin
films compared to Cu and Au, corresponding to higher Q and improved
performance in passive microwave devices.
The second advantage is the
frequency independent penetration depth as compared to frequency dependent
skin depth in normal conductors.
This means, dispersion introduced in
superconducting components will be negligible up to frequencies as high as 1
THz. Because of lower losses in superconductors, reduction in size is another
advantage using the high Tc thin films. Compact delay lines, chirp filters, and
resonators are possible[9].
One of the most important developments in microwave applications of
high Tc materials, is the fabrication of high quality thin films on low loss
substrates such as LaAIOs and M g 0[43-45].
At microwave frequencies, a
fundamental quantity of importance for superconductors is the microwave
surface resistance (Rs). From two fluid theory[46], the Rs of a superconductor
is proportional to the square of the frequency, as compared to f 1/2 dependence
in normal metals. Superconductors have orders of lower Rs at lower microwave
frequencies[47-48].
However, as frequency increases, the superconductor
10
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
losses exceed that of normal metals. Surface resistance o f superconducting
thin films can only be measured by indirect methods, because of lower values
in superconductors, below Tc. By introducing a superconducting film in a high
Q cavity, the loss in the sample would directly affect the cavity Q, and the
bandwidth[47-49].
Rs can be obtained from the change in Q.
A second
method for obtaining the surface resistance is from the real and imaginary parts
of the complex conductivity[50]. Measurements of complex conductivity have
been reported by several researchers[50-51].
Bardeen Cooper Schrieffer
This method is based on the
(BCS) theory[52],
homogeneous superconducting materials.
which
is valid only for
Since the high Tc materials are
complex and inhomogeneous, the usage of BCS theory has to be examined.
The BCS theory, being the only theory for superconductivity, is being used
widely.
Another indirect method for mesuring Rs, is by fabrication and testing of
microstrip resonators [53]. By measuring the quality factor and bandwidth in
resonators, we can obtain the Rs from the conductor losses in the microstrip.
TICaBaCuO thin film based passive microwave devices, have shown superior
performances. Chang et al [47], have reported a surface resistance at least an
order smaller than Cu at 77 K and 9.5 GHz. Bourne et al [54], reported a one
nano second microstrip delay line using thin films of TICaBaCuO.
Again a
factor of 10 improvement in loss was observed at 3 .2 9 GHz and 77 K.
Hammond et al[55], reported TICaBaCuO microstrip resonator and its power
11
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
handling performance at 77 K. Loaded Q as high as 7 3 0 0 at 2 .6 GHz was
obtained. At effective power levels in the resonator upto 100 W , the Q was
still atleast 3 times higher than a silver resonator at 2 .6 , 5.2, and 7.3 GHz.
Hammond et al[55], reported the power dependence of microwave losses in
epitaxial TICaBaCuO thin films on LaAI03 substrate.
Low power surface
resistance 50 times lower than Cu at 9.55 GHz and 77 K, was measured. At
microwave fields greater than 30 G, the surface resistance was more than ten
times lower compared to Cu. Millimeter-wave ring resonators with an unloaded
Q as high as 26 50 at 77 K and 35 GHz, have been obtained at 0 dBm input
power[56]. This is approximately 5 times higher compared to a gold resonator
at the same frequency and temperature.
Another major impact for applications of TICaBaCuO thin films is in the
area of high T c SQUIDs, for sensitive detection of magnetic fields[57-59].
TICaBaCuO SQUIDs have proven to have very low flux noise, compared to
SQUIDs made from YBCO and bismuth superconductors[58], mainly due to
differing responsivities of natural grain boundaries. TICaBaCuO SQUIDs have
sensitivity and low flux noise comparable to commercial rf and dc SQUIDs[59].
Recently, new developments have been made in the area of switching devices
based on TICaBaCuO thin films. J.Martens and D.S.Ginley have developed a
superconducting flux flow transistor(SFT)[60].
The SFT consists of two
regions of TICaBaCuO superconductors, joined by a dozen or so weak-links.
The weak-links, 0.1 //m thick, are arranged like the rungs in a ladder.
12
Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission.
A
magnetic field from a superconducting control line can pin or unpin the flux
lines in the weak-links, and so shift them between a superconductive and a
non-superconductive state. Such devices could find applications in microwave
circuits for switching app!ications[61-62].
Other applications such as thermal switches based on TICaBaCuO thin
films have been investigated by Martens et. al.[63]. The thermal switch is
based on the principle of turning a superconducting microbridge normal using
a control current passed through a normal conductor. The two distinct states
of the switch could be used in such applications as superconducting phase
shifters.
An important area of research in high Tc superconductors is the
modelling and simulation of high Tc thin film based quasi-TEM tranmission lines.
The high temperature superconducting thin films are different from conventional
superconductors because of higher penetration depth and lower coherence
length.
The penetration depth in high Tc thin films is as high as the film
thickness or higher due to the quality of the films.
High penetration depth
results in field penetration through the entire thin film. The superconductor is
entirely penetrated by the field and introduces an additional internal inductance
called the kinetic inductance[64]. For modelling of high T c thin films the kinetic
inductance effects should be included when the penetration depth is in the
order of film thickness. Several models have been reported in literature which
in general are based on the calculation of internal impedance using complex
13
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
conductivity approximation for the superconductor[65-67]. The author's model
is based on the phenomenological loss equivalence (PEM)
model [65]
approximation and kinetic inductance calcualtion based on the internal
inductance[67J.
One of the objectives of the author was to design
superconducting microstrips based on the characteristic impedance changes
due to the kinetic inductance. By knowing the superconducting properties of
the films, the kinetic inductance and hence the wave slowing factor were
calculated. The superconducting microstrips were designed for a predetermined
effective characteristic impedance so that when the sample is superconducting
the characteristic impedance of the microstrip will be the desired characteristic
impedance. The microstrip width and lengths were designed based on the
standard design formu!ae[68-71].
Superconductor-semiconductor hybrid circuits
are being currently
investigated for cryoelectronics applications. The advantages of using hybrid
circuits combine the advantages of both superconductors and semiconductors
\
[72-73].
Hybrid amplifier circuits[74], and oscillators[75] and phase shifters
[76] have been reported using high Tc thin films and compound semiconductor
devices. Narrow band-smplifiers and oscillators with improved performance
have been reported utilizing the lower conductor losses in the superconducting
thin films. The main goal for the investigations of hybrid circuits is to develop
microwave subsystems for communication applications.
One of the objectives of the author was to design, fabricate and test a
14
Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission.
hybrid phase shifter circuit using TICaBaCuO superconducting thin films and
GaAs MESFETs. Phase shifters are essential components for electronic steering
of phased array antennas in communication systems. The function of a phase
shifter is steady control of relative phase between the input and output. The
phase of each antenna element is controlled such that a radiated beam of any
desired shape can be formed. Digital phase shifters obtain phase control in
steps of a least significant bit, generally 11.5 degrees.
The analog phase
shifters offer a continuous control of phase. Problems such as size, drift with
time, large phase errors and large power supply requirements complicate the
analog phase shifter designs. The digital phase shifters offer the advantages
of lower insertion loss, high speed switching, small drive power, small size, and
low cost per bit. The fundamental theory of electronic steering of phased array
antennas and the phase shifter designs can be found in several text books[7778].
Superconducting phase shifters have been investigated by several
researchers, including the use of HTS and semiconductor based hybrid
circuits[76,79-80].
The use of distributed Josephson inductance (DJI)
approach for HTS based monolithic phase shifter has been reported recentiy[7980].
Phase swings greater than 60 ° were reported at 65 K and lower
temperatures. The main principle is to control the wave velocity and hence the
phase shift
by providing a variable inductance to
a superconducting
transmission line[79].
15
Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission.
In this research, the phase shifter bits were designed using reflection
type configuration for 1 8 0 ° phase shift using HTS passive components and
GaAs MESFET switches.
The circuit consists of a 3 dB coupler that divides
the input power equally between the direct and coupled ports.
By careful
design o f the impedance transforming network, and proper terminations in the
direct and coupled ports, any desired phase shift can be obtained.
The
switching device acts as the reflection plane. The desired phase shift could be
obtained by changing the switching states of the device. The bandwidth of the
circuit is limited by the type of the coupler employed. An inter-digitated Lange
Coupler was chosen for the 3 dB coupler, to obtain a large bandwidth of about
25% .
Also, the inter-digitated Lange coupler could be easily realized in the
microstrip form[69].
The phase shifters were designed for 4 GHz center frequency, and
operation at 77 K.
The 1 8 0 °
phase bits were designed for: (a) minimum
insertion loss, (b) a minimum bandwidth of 25 % , (c) the phase error less than
10 ° , and (d) input and output return losses greater than 15 dB.
A fabrication process was established for superconductorsemiconductor
hybrid circuits using a two level mask set.
A reflection type hybrid phase
shifter circuit for 180 degrees phase shift was designed, fabricated and tested
by the author[81 -82]. The hybrid circuit showed more than 1 dB improvement
in insertion loss over a gold based hybrid circuit.
16
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1.4. Objectives o f the Research Program
The first major objective in this research program was to develop a
fabrication process for high quality (high Tc and high Jc) TICaBaCuO thin films
mainly on LaAlOs substrates.
A fabrication process was established by
optimizing the sputter deposition parameters such as the chamber pressure, rf
power density, substrate to target distance, and substrate temperature during
deposition. The post-deposition parameters such as the sintering and annealing
temperatures, Tl and 0 partial pressures were optimized for obtaining high
quality thin films, mainly on LaAI03 substrates.
The second major objective of the research program was the fabrication
of device structures, without degradation of superconducting properties. A
patterning process for fabricating fine line structures in TICaBaCuO thin films
was developed. The fabrication of useful device structures would enhance the
applications of TICaBaCuO thin films for microelectronic applications.
A
photolithography mask for four-probe test devices was designed and fabricated.
A w et chemical etching technique using a dilute phosphoric acid solution was
investigated for etching the TICaBaCuO thin films. The four-probe test devices
were used for critical current measurements. The author's objective was to
obtain critical current density greater than 5x10 5 A/cm 2 at 7 7 K.
correlation of microstructure and Jc were examined.
The
Also, a method for
obtaining low resistance gold and silver contacts on Tl films was developed.
17
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Contact resistances as low as 10‘8 ohm-cm2 at 77 K were routinely obtained.
The third major objective in the author's research program was to
fabricate high quality factor ring resonator based on TICaBaCuO thin films, for
microwave and millimeterwave applications. Ring resonators were designed for
a fundamental resonance frequency at 12 GHz.
The all-superconducting ring
resonators were fabricated and tested. Unloaded Q of the ring resonators were
obtained from swept frequency reflected power measurements. The unloaded
Q of the superconducting resonators were compared to that of a gold resonator
fabricated for the same design.
The superconducting resonators were
fabricated on 10 mil thick both sides polished LaAI03 substrates. A process
was developed in this lab for obtaining superconducting thin films on both sides
of the substrates. From the unloaded Q measurements, the effective surface
resistance of the superconducting thin films was extracted. The Rs at 12 GHz
and 77 K was typically 1.5-2.75 m£l almost an order of magnitude lower than
Cu at the same temperature and frequency.
Our fourth objective in this research programme was to design, fabricate
and test superconductor-semiconductor hybrid reflection type phase shifter
circuits using TICaBaCuO thin films and GaAs MESFETs. A 180 degrees phase
bit was designed using a superconducting Lange coupler and impedance
transformation networks, for a 1 GHz bandwidth, 10 degrees phase error, and
low insertion loss during the on and off state of the MESFETs. The measured
results indicate an improvement of approximately 1 dB in the insertion loss over
18
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the gold based phase shifter circuit.
Chapter 2 describes the processing of TICaBaCuO thin films by rf
sputtering and post-annealing methods.
Chapter 3 describes the structural
studies on the superconducting thin films including SEM, XRD, AES and XPS.
Chapter 4 describes the electrical transport measurements on the patterned
superconducting thin films. Chapter 5 explains the microstrip ring resonator
results, and chapter 6 the design, fabrication and testing of hybrid phase shifter
circuits.
19
Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission.
CHAPTER 2. RF SPUTTERING OF TI-Ca-Ba-Cu-0 THIN FILMS
2.1 Sputtering System:
TICaBaCuO thin films were prepared by sputtering from a single
composite powder target in a CVC model 601 rf sputtering system.
The
system consisted of three isolated sputtering stations. Each station contained
a water cooled cathode assembly coupled to a 1 KW rf power supply operating
at 13.56 MHz. The system was operated in the rf magnetron mode to improve
the sputtering yields at low working pressures.
The depositions were
performed in a 'sputter up' configuration on substrates placed face down on
the anode. Figure 2.1 shows a schematic for the sputtering configuration. The
sputter-up configuration is useful especially for powder targets.
An infrared
radiant quartz heater assembly is provided for substrate heating. The substrate
to target distance was fixed at 3 inches to minimize the negative ion backsputtering effect. The gas flow rate was controlled by a mass flow controller,
and the chamber pressure was adjusted using a butterfly valve placed in the
vacuum line. The chamber is pumped down by a diffusion pump which has a
liquid nitrogen baffle for recycling the used oil. The diffusion pump has the
capability to pump down to 10 '7 Torr. The high vacuum valve is opened only
after pumping down the chamber below 50 mTorr using the roughing pump.
The chamber pressure is measured by an ionization gauge in the sub-milli Torr
range. Figure 2.2 shows a photograph of the CVC 601 sputtering system.
20
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Substrate
Digital
Powder
Flow
Target
Controller
Cathode
Ar gas
RF Power
Diffusion Pump
Supply
Fig 2.1 Schematic diagram for the sputtering configuration
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. 2 .2 Photograph of the CVC 601 Sputtering System
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2 .2 Sputtering Targets:
The sputtering targets were powder targets made in house using three
different methods. The TI2Ca2Ba2 Cu3Ox powder target was prepared from high
purity BaO, CaO, CuO and Tl20 3 powders.
Stoichiometric amounts of BaO,
CaO, Tl20 3 and CuO (in the ratio 2:2:2:3) were mixed and ground in an agate
mortar using a pestle. The target was enriched with about 2 0 % excess Tl20 3
to compensate for the loss of Tl during the post-processing of the thin films
and to maintain sufficient composition for several deposition runs. The powder
target was spread over an 8 inch diameter copper plate which was part of the
cathode assembly, and pressed to obtain a uniform target of 2-3 mm
thickness.
The second method consisted of making a BaCu02 precursor by heat
treating (850 °C for approximately 24 hours) appropriate amounts of Ba02 and
CuO. The advantage of using Ba02 is the lower melting point (820 °C), which
helps in obtaining a homogeneous BaCu02 through liquid phase reaction. After
forming BaCu02 precursor, appropriate amounts of Tl20 3 and BaCaCu02 were
mixed and ground to form TICaBaCuO compund. An excess of approximately
7% Tl20 3 was added to compensate for the loss during the sintering process.
Pellets of approximately 12 gm each were made by pressing in a pelletizer at
15000 psi load for 10 minutes. The pellets were sintered at 850 °C for 10
minutes to obtain superconductivity.
From resistance vs temperature
characteristics, the critical temperature of the pellets fabricated was determined
23
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
to be above 120 K, indicating that the samples were of the 2223 major phase.
The pellets were ground again to a very fine powder, which was spread
uniformly on the Cu backing plate and pressed to form the target.
Such a
sintered powder target gave very little contamination to the system vacuum
vessel during sputtering. The target could be used for multiple runs.
The third target studied was the sintered 2122 powder target prepared
similar to the above case, but the amount of Tl20 3 was adjusted to form 2122
phase. Again 7% excess Tl20 3 was added to compensate for the loss of Tl
during the sintering process. Figure 2 .3 shows the magnetic susceptibility vs.
temperature in a sintered 2122 pellet. The pellet was sintered at 8 5 0 °C for
10 minutes followed by another heat treatment at 860 °C for 10 minutes. The
critical temperature of the superconducting samples, the superconducting
volume fraction for the samples were determined from the slope of the
magnetic susceptibility vs magnetization measurements. The Tc of the sample
was approximately 97 K. The sharpness of transition from the onset of T c to
the zero resistance T c is very high. The volume fraction of superconductivity
was as high as 54% for the sample. A few samples of 2122 made separately
showed consistently higher volume fraction of superconductivity proving that
this gives the optimized heat treatment for highest volume fraction. The liquid
phase Tl20 formation above 8 2 0 °C is the key for the rapid reaction between
Tl20 , BaCu02 and CaO. The Tc of the sample is lower than 110 K which is the
T c of the 2122 phase, may due to oxygen deficiency in the samples, or
24
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
a. - 4
-6
80
100
TEMPERATURE (K )
Fig. 2.3
Magnetic susceptibility vs Temperature for a bulk 2122 Tl
superconductor
stoichiometrically off from the 2122 phase, or a combination of both.
The
higher volume fraction indicates that the samples are of high quality.
The
samples with higher Tc closer to 110 K, have very low volume fraction of
superconductivity again indicating mixed phase creation in the samples.
The samples used for magnetization measurements were ground into a
fine powder and mixed with acetone to form a film on a glass plate which is
mounted on an aluminum sample holderfor the X-ray diffraction measurements.
The samples were scanned from a diffraction angle of 4 ° to 6 0 °. The 2122
phase has a low angle (002) peak at 6 ° and the 2223 phase has a peak at 5 °,
the 1212 at 7 °, distinguishing each phase[20]. Figure 2 .4 shows the X-ray
diffraction spectra obtained for the same 2 1 22 sample. The low angle peak
around 6 ° confirmed the 2122 phase in the sample. Also, by comparing the
spectra with other references, the sample was determined to be single
phase[20].
The optimum two step heat treatment for single phase 2122 formation
obtained was: first heat treatment at 8 5 0 °C for 10 minutes in air, and second
at 8 6 0 °C for additional 10 minutes in air with both steps performed in a
platinum crucible closed by a platinum lid[83].
2 .3 Substrates:
(100) SrTiOs substrates have a very close a-axis lattice match with the
TICaBaCuO compound, and were used in this study.
The substrates were
26
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2 -T h e ta - S cale
Department o f Geology a t the U n iv e r s ity o f C in c in n a ti ll-M a y -1 9 9 2 15: 25
o
HI
0)
go
rs
«H
O
*
V
(0
°
|
i
i
r
i
f vf l i ^i
S
10
15
20
B:\MZ003.RAWTLi3L0NG (CT:
9 .0 a .
°
Fig. 2.4
Powder
O
'
y\
2fe
30
SS:0.050dg.
x-ray
superconductor
diffraction
II
WL:
i
i
35
40
i.540BAo)
spectra
for
a
bO
bulk
2122
Tl
degreased in acetone and methanol, rinsed in Dl water and blown dry before
loading into the vacuum chamber.
Microwave transmission properties of
superconducting passive components are governed by the losses in the
superconducting thin films as well as the dielectric losses associated with the
substrate. The best quality of high T c TICaBaCuO thin films grown until the
intiation of this research work were obtained on substrates which have
dielectric losses too high to be useful for microwave applications.
SrTi03
substrates have a static dielectric constant above 300 at room temperature,
and over 1000 at liquid nitrogen temperature [84]. The loss tangent of SrTi03
is extremely high at microwave frequencies compared to other useful
microwave substrates[85], leading to poor microwave transmission properties.
The YSZ and MgO substrates have low dielectric constants, but due to the
large mismatch in the a-axis lattice constants, epitaxial TICaBaCuO thin films
have not been grown.
Single crystal LaA!03 substrates offer the advantages of low loss
tangent[43], low dielectric constant, and good lattice match to the TICaBaCuO
compound. Thin films of YBaCuO and ErBaCuO have been studied on LaAI03
substrates[43]. Excellent superconducting properties comparable to the films
on S rTi03 were obtained.
The dielectric constant of LaAI03 is 24.5[44] and
the loss tangent is approximately 8 .3 *1 O'5 at 77 K[44]. The LaAI03 substrates
were degreased in boiling acetone followed by rinsing in methanol and Dl water
before loading the substrates into the chamber.
28
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2 .4 Sputtering Gases:
The sputtering of TICaBaCuO thin films from a sintered powder target
was performed in a pure argon partial pressure.
The reason for not using
oxygen and reactive sputtering is due to the volatility of Tl from the target. Tl
readily combines with oxygen and forms Tl20 3 gaseous phase which can be
easily pumped out from the chamber. The target gets depleted of Tl much
faster when oxygen is used as part of the reactive gases. Hence the author
has used only pure argon sputtering for TICaBaCuO thin film deposition from
both TI2122 and 2223 targets.
2 .5 Sputtering Conditions:
The chamber is pumped down to a background pressure of 2x1 O'6 Torr
before introducing the working gas. Substrates were subjected to an initial insitu bake at 125 °C for 10 minutes.
After establishing a constant flow of
argon, the chamber pressure was raised to 5 mTorr using the butterfly valve.
Prior to the sputtering on the substrates, pre-sputtering of the target was
performed to clean and equilibrate the target surface.
The shutter located
between the target and the substrates was closed during this pre-sputtering
process. The rf power was slowly increased to the power level desired for
sputter deposition. Typically, a 3-4 hours pre-sputtering was needed for new
targets and a half hour for used targets to outgas the powder target.
TICaBaCuO thin films were deposited to a thickness of 0.3-0.7 pxr\ using the
29
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
conditions summarized in table 2.1.
The deposition rate obtained using a
dektak profilometer, was approximately 30 A/min.
2 .6 Post-processing o f thin films:
2.6.1 Thin films fabricated from 2223 target:
The sputter deposited thin films were post-processed in two steps: first,
sintering in air at 850 °C for 12-15 minutes in an excess Tl20 partial pressure
and second, annealing in an oxygen flow of 500 see at 750 °C for 15 to 30
minutes. An excess Tl partial pressure was maintained during the annealing
process. Sintering was performed in a small box furnace in the free surface
configuration described by Ginley et al[18].
shown in figure 2.5a.
The sintering configuration is
The free surface configuration in excess Tl20 partial
pressure is required for sintering the thin films sputtered from the TI2223
powder targets, since the additional Tl20 reacts to the 2223 phase and forms
T I212 2 phase.
The thin films were placed on a sintered pellet of TI2Ca2Ba2Cu3Ox
(2223) with the film side facing the free surface in a small platinum crucible
covered by a lid. A second pellet was placed above the sample in a platinum
wire mesh support. The pellets provided the excess Tl partial pressure in the
crucible, to minimize the loss of Tl from the thin films. Sintering of TICaBaCuO
thin films at 850 °C in air for 12-15 minutes established the superconducting
phase and morphology. At 850 °C, Tl20 3 decomposed into a Tl20 liquid phase
30
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
TABLE 2.1
TYPICAL DEPOSITION PARAMETERS
RF POWER
220-240 W
CHAMBER PRESSURE
5 m-Torr
TARGET-SUBSTRATE DISTANCE
3"
DEPOSITION RATE
30 A/min
THICKNESS OF THE FILMS
3000-7000 A
Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission.
Fig. 2.5a. Free Surface Configuration
TI2122 pellets
Fig. 2.5b. Confined Surface Configuration
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
and a rapid liquid phase sintering occured. After sintering, the crucible was
removed from the furnace and cooled rapidly.
Oxygen annealing of the air
sintered TICaBaCuO thin films was carried out in the same configuration as the
sintering, with an oxygen flow of 50 0 see to about 1 liter/min. During oxygen
annealing, the grains in the thin films grew into large platelets.
Oxygen
annealing improved the electrical properties of the thin films by including
oxygen into the compound.
For reproducible processing of TICaBaCuO thin films, it was necessary
to provide the optimum Tl partial pressure during post-deposition processes.
A simple technique was used to monitor the reduction in Tl content in the asdeposited thin films, after each sputtering run.
Percentage reduction in Tl
content from run to run was obtained through AES surface analysis on the asdeposited samples.
The percentage reduction in Tl content compared to a
standard reference 2223 pellet gives an approximate estimate of additional
Tl20
partial pressure needed during the post-deposition processes.
The
estimated additional Tl20 partial pressure was provided for the post-deposition
processes by adding additional 2223 pellets in the crucible. This technique has
yielded reproducible high Tc and high Jc films[86].
2 .6 .2 Thin films sputtered from 2 1 2 2 target:
Thin films sputtered from T I2122 powder target were required to be heat
treated in a confined surface configuration shown in figure 2.5b. The confined
33
Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission.
surface sintering prevents the loss of Tl20 3 from the films.
The confined
surface sintering of 2122 thin films was performed in the presence of a Tl
partial pressure introduced by sintered 2122 pellets. An additional annealing
step in oxygen flow of 1 liter/min for approximately 10 minutes was required
for obtaining high quality thin films.
34
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 3. STRUCTURAL CHARA CTERIZA TION OF THIN FILMS
3 .1 SEM ANALYSIS:
The surface morphology of the annealed thin films was examined using
an ISI model SX-30 SEM.
The effect of annealing on the thin films was
studied by SEM. During the oxygen annealing, the grains in the thin films grew
into large platelets greater than 10 jjm in diameter. The oxygen annealing and
slow furnace cooling improved the grain boundaries in terms of defects and
irregularities. Figure 3.1 shows the SEM micrograph of an annealed thin film
on a SrTi03 substrate. The thin film contained large platelets oriented parallel
to the surface of the substrate. The as-deposited thin films were smooth with
a thickness uniformity greater than 95% , while the annealed thin films had a
thickness uniformity of
86%
as determined from dektak profilometer
measurements.
Figure 3.2 shows the surface morphology of an annealed film on LaAI03
substrate. The micrograph shows platelets oriented parallel to the surface of
the substrate. The platelets were smooth and varying in size from 2 to 5 //m .
Some randomly distributed defects in the form of pin-holes on the surface of
the thin film were observed.
3 .2 X -ray diffraction fXRDJ:
The TICaBaCuO thin films were analyzed by XRD to determine the crystal
35
Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission.
5.0kx
15kv
032
Fig 3.1 SEM Micrograph of a TICaBaCuO thin film on a SrTi03 substrate
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig 3 .2 SEM Micrograph of a TICaBaCuO thin film on a LaAI03 substrate
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
structure and the phases present in the films. A Siemens x-ray diffractometer
Model 3 0 7 0 with a Cu Ka radiation was used to obtain a diffraction pattern
from 4 to 60 degrees Bragg angles {28).
Figure 3 .3 shows the XRD spectrum obtained on an annealed thin film.
The characteristic peaks corresponding to 2 1 22 and 2223 phases and the
SrTi03 substrate were present. No other impurity phases were observed. From
the (0 0 I) peaks, the highly c-axis oriented nature of the thin films was
revealed.
The c axis lattice constant of the 2 1 22 and 2223 phases were
calculated as c = 30 .4 A and c = 35.6 A respectively. The 2122 phase was the
dominant phase in these thin films, as determined from the intensities of the
XRD peaks.
Figure 3 .4 shows the XRD spectrum of an annealed thin film on LaAI03
substrate.
The figure shows the characteristic peaks of TI2Ca1Ba2Cu20x
(2122), TI2Ca2Ba2Cu3Ox (2223) and the LaAI03 substrate. No other impurity
phases were present in the film. From the ( 0 0 I) peaks present in the XRD
spectrum, the highly c-axis oriented growth is evident. The 2 1 22 phase is the
dominant phase in the thin film, as determined from the intensities of the XRD
peaks. The c-axis lattice constant calculated from the XRD spectrum is 29 .2
A which is close to the previously reported values[20]. From the XRD data, the
a-axis lattice constant of LaAI03 is 3.7801 A. The a-axis lattice constant of
a unit cell of 2122 phase is 3.8503 A. The lattice mismatch from these values
is less than 2% .
38
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
03
2
D
OQ
OC
<
>•
H
TOJQ
03
2
-Q TO
oocvT
CD (O
oo
oo
t -C M
O O
O O
111
f—
2
20
28
36
44
52
60
TWO - THETA (DEGREES)
Fig. 3 .3
X-ray diffraction spectrum of an annealed TICaBaCuO thin film on
a SrTi03 substrate
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
X R D PATTERN
•
2122
y 2223
D
<
Z
>H
00
Z
Ui
I-
z
CM O
T- T-
o o
o o
20
28
36
44
TWO-THETA IN DEGREES
Fig. 3 .4
52
60
X-ray diffraction spectrum of an annealed TICaBaCuO thin film on
a LaAI03 substrate
3 .3 Auger Electron Spectrocopy:
The superconducting TICaBaCuO thin films were analyzed by AES,
and argon ion sputtering, to determine the compositional uniformity and
impurity contents.
Measurements were carried out on as-deposited and
annealed samples in a Perkin Elmer Physical Electronics Model 56 0 ESCA/SAM
system.
The excitation was provided by a primary electron beam, with an
energy of 5 KeV and 3.0 pA beam current. Sputtering was performed with a
2 KeV argon ion beam rastered over an area of 6 mm x 6 mm. This ion beam
energy was chosen to minimize the knock-on mixing and sputtering non­
uniformities. The sputter rate was determined to be approximately 1.7 A/min.
The depth profiles of Tl, Ca, Ba, Cu and 0 in the samples were obtained by
recording the peak to peak heights of the derivative AES signals for TI(MNN),
Ca(LMM), Ba(MNN), Cu(LMM) and O(KLL) respectively. Since the instrumental
sensitivity to metals in the form of oxides are usually different from those of
pure metals, only the elemental depth profiles were obtained. AES survey and
depth profile of a T I 2 Ca2Ba2 Cu3 0 x pellet were used as a reference in order to
estimate the composition of the thin films from the intensities of the AES
peaks.
Figure 3.5 shows the AES survey in the bulk of (a) a sintered
TI2 Ca2 Ba2 Cu3 0 x pellet, (b) an as-deposited thin film and (c) thin film after
sintering and annealing.
The pellet was used as a reference for the AES
measurements. The figure shows the Tl(242 eV), Ca(292 eV), Ba(584 eV),
41
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Ca
Ba
Cu
Ba
Cu
Ba
Cu
>
DC
<
DC
H
CD
DC
<
>
H
Ca
C/3
0
200
400
600
800 1000 1200
K IN E T IC ENERGY CeV)
Fig. 3 .5
AES survey of a TICaBaCuO thin film on a S rTi03 substrate
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Cu(920 eV) and 0 (5 1 2 eV) peaks of the pellet, as-deposited and annealed
films. The as-deposited thin films were rich in Tl at the surface and the amount
of Tl decreased with depth into the sample.
This was attributed to the
tendency of Tl to segregate at the surfaces. After sintering and annealing, the
Tl content decreased in the sample confirming the loss of Tl during the sintering
and annealing processes. The intensities of the AES peaks measured in the
pellet were used as the reference to determine the approximate composition in
the thin films.
Since the target had been used for several times, the exact
composition of the target was not known. The as-deposited thin films had a
chemical composition of Tln 8Ca3 1Ba1 9Cu2 _5Ox in the bulk. After sintering
and annealing the composition changed to Tl0.gCa1 3 4Ba2.4 Cu2 .i Ox, indicating
that the Tl deficient 2 1 22 phase dominates in the bulk of the thin film. Figure
3 .6 shows the AES depth profiles of (a) the sintered TI 2 Ca 2 Ba2 Cu3 0 x pellet,
(b) an as-deposited thin film, and (c) annealed thin film.
The depth profiles
show the variation of Tl, Ca, Ba, Cu and 0 with depth. The depth profiles
indicate good compositional uniformity in the thin film through the depth
analyzed. This also indicates the homogeneity of the annealed thin films. A
small detectable amount of carbon contamination was found at the surface of
some of the annealed samples, which was attributed to the exposure of thin
films to the laboratory environment.
A TICaBaCuO thin film of approximately 0.1 //m thcikness on a LaAI03
substrate was analyzed by AES depth profiling, to determine the compositional
43
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
PELLET
Ca
Ba
co
b-
Cu
z
13
m
cc
<
<
in
Ca
AS DEPOSITED
F IL M
CL
Ba
Cu
<
Ui
Q.
Ca
ANNEALED
F IL M
Ba
Cu
24
Fig. 3.6
48
72
96
120 144
S P U T T E R T IM E (M IN .)
168
192
AES depth profile obtained on a TICaBaCuO thin film on a SrTiO
substrate
Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission.
uniformity through the thickness of the film, and the substrate-thin film
interaction at the interface. The excitation was provided by a primary electron
beam with an energy of 5 KeV and 3.0 pA beam current. The sputter etching
was performed with a 5 KeV argon ion beam rastered over an area of 6 mm x
6 mm. The depth profile obtained is shown in figure 3.7. The depth profile
shows good compositional uniformity through the depth for Ca, Ba, and Cu
elements. The energy window used for Tl and Al are very close to each other
resulting in uncertainty in the distribution of Tl in the substrate. The interaction
between the substrate and the thin film was estimated to be less than 2 5 0 A
[87].
3 .4 High Resolution XPS:
High resolution XPS spectra were obtained using the same ESCA/SAM
system. The chemical states of Tl, Ca, Ba, Cu and 0 were investigated at the
surface, in the bulk of the thin film and near the interface with the SrTi03
substrate.
Excitation for the XPS analysis was provided by a Mg Ka x-ray
source. The photoelectron energy was measured with a spherical retarding grid
in conjunction with a double pass cylindrical mirror analyzer (CMA). For the
measurement in the bulk of the thin film, the AES depth profiling was
temporarily suspended to obtain the XPS data. Since these films were about
0.5 pm thick, and the argon sputter rate of 1.7 A/min was very slow, most of
the film was etched away using a 1:10 phosphoric acidrDI H20 solution. The
45
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
_J
CO
C
o
O
3
o
co
CD
CO
O
LL
H
O
cc
-2
ca
©
c
c
CO
c
CO
c
CL
X
H
Q.
LU
o
•o
©
c
’ co
a
C/5
3
LLl
o
<
o
a.
X>
o
a)
0)
s
Q.
JZ
®
-o
s
3
w
Q
<
C/D
LU
CO
—1
<
r^
co
03
'l l
CD
>lV3d 0 1 »V 3d
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CO
co
XPS survey close to the interface with the substrate was then obtained. High
resolution XPS spectra were obtained for the Tl(4f), Ca(2p), Ba(3d), Cu(2p) and
0(1 s) peaks in order to study the chemistry of the various elements in the thin
films.
The chemical states of Tl, Ca, Ba, Cu and 0 in the annealed TICaBaCuO
thin films were studied using XPS[31 ]. The XPS spectra were referenced to the
C 1s internal standard at a binding energy of 284.6 eV[88]. Figure 3 .8 shows
the Tl 4 f core level spectra of the TICaBaCuO thin films (a) at the surface, (b)
in the bulk and (c) near the interface with the substrate. The chemistry of Tl
at the surface and in the bulk of the thin film are very much the same. There
is a shift to a higher binding energy of Tl 4f near the interface with the
substrate. The binding energies of the main Tl 4 f7/2 peaks in figures 3 .8 .a,
3 .8 .b and 3.8.C are 118.0, 118.1 and 119.3 eV respectively. The binding
energy of Tl 4 f7/2 is 117.7 eV for T l3+ in TI20 3 and is between 118.7 and
1 1 9 .4
eV for Tl1 + depending upon the ligands[88].
In the TICaBaCuO thin
films analyzed, the binding energy of Tl 4 f7/2 was found to be higher than that
o f Tl20 3 and lower than Tl20 except near the interface with the substrate,
where the binding energy of Tl 4 f 7/2 matched that of Tl20 . This indicates that
the valence of Tl in the TICaBaCuO thin films is between + 3 and + 1 .
The
exact determination of the Tl valence is difficult based on the XPS spectra. Our
results agree with the previously reported chemical state of Tl in a
T I2Ca2Ba2Cu3O10 bulk sample studied by XPS[21].
47
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7/2
INTENSITY
(ARB,UNITS)
5 /2
127
124.2
121.4
B IN D IN G
Fig. 3.8
118.6
115.8
ENERGY (eV )
High resolution XPS spectrum for Tl
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
113
The Ca 2p core level spectra of the TICaBaCuO thin films are shown in
the figure 3 .9 .
The binding energies of the main Ca 2p3/2 peaks in figures
3 .9 .a, 3.9.b and 3.9.c are 34 5.0 7, 3 4 5 .7 2 and 3 4 7.2 eV respectively. The
XPS high resolution spectra of Ca 2p at the
surface shows a complicated
mixed state. The binding energy of Ca 2p3/2 in CaO is 346.2 eV[88]. From
figure 3.9.b, the binding energy of Ca 2p3/2 peak is closer to that of CaO.
There could be a mixed state of Ca and CaO in the bulk of the thin film. Tl
deficiency
and
excess
in
Ca
are
evident from
the
composition
^’,0 .9 ^ ai. 34 Ba2 .4 ^ u2 .i^ x ' obtained for the annealed thin film.
of
The thallium
deficient and calcium rich composition of the 2122 phase may be necessary for
stabilization of the superconducting phase similar to the TI2Ca2Ba2Cu3Ox
compound[21-22]. The calcium rich phase implies that there may be a partial
substitution of Ca2+ in place of Tl3+ resulting in hole creation, as reported
previously, based on electron probe micro-analysis[21].
Figure 3.10 shows the Ba 3d XPS spectra of the TICaBaCuO thin films.
The binding energies of the main Ba 3d5/2 peaks in 3.10. a, 3.10.b and 3.10.C
are 77 9 .6 5 , 7 7 9.7 and 7 7 9.7 eV respectively.
From these spectra we
conclude that the chemistry of Ba is unchanged from the surface up to the
interface with the substrate. Since the binding energy of Ba 3d5/2 in BaO is
7 7 9 .6 5 eV[88], the chemical state of Ba in the TICaBaCuO thin film is BaO, as
expected from the structure. Since the chemistry of Ba has not changed near
the interface with the substrate, we could conclude that Ba has not reacted
49
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
INTENSITY
(ARB. UNITS)
Ca
354
351.6
349.2
B IN D IN G
Fig. 3 .9
346.8
344.4
ENERGY (eV)
High resolution XPS spectrum for Ca
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
342
3d
Ba
INTENSITY
( ARB. UNITS )
3d
800
794
788
BINDING
Fig. 3 .1 0
782
776
ENERGY ( e V j
High resolution XPS spectrum for Ba
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
770
with the substrate.
The chemical state of Cu in the TICaBaCuO is complicated as evidenced
by the Cu 2p core level spectra in figure 3 .1 1 .
The binding energies of the
main Cu 2p3/2 peaks in 3 .1 1 .a, 3 .1 1.b and 3 .1 1.c are 93 1.7 5, 9 3 2.0 and
9 3 5 .8 eV respectively.
For Cu metal and the monovalent Cu20 , only the
primary line at 932 eV is observed[88]. For the divalent CuO, the main peak
Cu 2p3/2 is at 9 3 3 .6 eV with satellite peaks nearby[88].
The presence of
satellite peaks, near the main peak in our spectra, implies the CuO state of Cu.
But the binding energies of the Cu 2p main peaks on the surface and in the bulk
of the thin film lie near the Cu20 state.
By examining the 0 1s peak, the
chemical state of Cu can also be determined. If the 0 1s peak is at 530.1 eV,
the CuO state will be dominant, and if it lies at 530.8 eV, the Cu20 state will
be dominant[89]. Figure 3 .1 2 .a, 3 .1 2 .b, and 3 .1 2 .c show the XPS spectra of
0 1s peaks in the TICaBaCuO thin film on the surface, in the bulk and near the
interface with the substrate respectively. The binding energies of the main 0
1s peaks in figures 3 .1 2 .a, 3 .1 2 .b and 3.12.C are 530.8, 5 2 9 .8 and 531.5 eV
respectively.
Binding energy of the main O 1s component at 530.8 eV
indicates that Cu20 state is predominant on the surface of the thin film. The
binding energy of 0 1s peak at 5 2 9.8 eV indicates the CuO state of Cu in the
bulk of the thin film. The shift in binding energy of the main Cu 2p peak to a
lower energy could be due to a charge transfer from Tl ions to the CuO layers.
Charge transfer between the Tl3+ and CuO layers, would lead to the observed
52
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
INTENSITY CA.U)
975
965.2
955-4
945.6
935.8
BINDING ENERGY IN eV
Fig. 3.1 1
High resolution XPS spectrum for Cu
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
926
(ARB. UN ITS)
INTENSITY
536
534
532
B IN D IN G
Fig. 3 .1 2
530
528
ENERGY (eV)
High resolution XPS spectrum for 0
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
526
valence of Tl between + 3 and + 1 , and also to the hole creation in the CuO
layers. Since superconductivity is obtained in the TICaBaCuO compound only
by the creation of holes, two of the possibilities for hole creation are: (1) partial
substitution of Ca2+ in place of Tl3 + , and (2) charge transfer from Tl3+ to the
CuO layers resulting in the valence of Tl between + 3 and + 1 .
Table 3.1
summarizes the results obtained from the high resolution XPS measurements.
55
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
TABLE 3.1
SUMMARY OF HIGH RESOLUTION XPS RESULTS
ELEMENT
PEAK
BINDING
CHEMICAL
ENERGY (eV)
STATE
Tl
4 f7/2
118.1
Ca
2P3/2
346 .0
CaO
Ba
3d5/2
779.7
BaO
Cu
2P3/2
9 3 2 .0
CuO
t i 2o 3
- t i 2o
CHAPTER 4. ELECTRICAL TRANSPORT PROPERTIES OF TICaBaCuO
SUPERCONDUCTING THIN FILMS
4 .1 Patterning TICaBaCuO thin films:
For the electrical transport measurements, four-probe test devices were
designed with line-widths of 10 pm, 25 pm, and 50 pm. The geometry of the
test devices is shown in figure 4.1. The voltage sense lines were 1 mm apart,
and the width of the sense lines was less than the line-widths in order to
approximate a point contact as closely as possible.
patterned
on
as-deposited
TICaBaCuO
thin
The test devices were
films
using
standard
photolithography and wet chemical etching techniques. Positive photoresist AZ
1421 was used for the lithography. The as-deposited TICaBaCuO thin films on
LaAI03 substrates were pre-baked at 180 °C for 20 minutes before the
photoresist was spun. The photoresist AZ 1421 was spun on to a thickness
of about 1 pm.
The samples were soft-baked at 90 °C for 20 minutes,
followed by exposure to UV light in a mask aligner.
The photoresist was
developed in a 1:5 developer:DI H20 solution for 45 seconds. The samples
were post-baked at 85 °C for 15 minutes to complete the photolithography
process. A 1:100 phosphoric acid : Dl H20 solution was used for chemically
etching the films. The solution was kept at a constant temperature of 75 °C.
The etch rate was approximately 40 A/min. After the etching process was
completed, the photoresist was removed by immersing the samples in acetone
57
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1mm
1mnr
1mm
1m m — »
1mm
1mm
a =10, 25,50Aim
b< a
Fig. 4.1
Geometry of four probe test devices for electrical transport
measurements
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
for 25 seconds, followed by a 30 second rinse in Dl H20 . The patterned
samples were post-processed using our standard methods described in section
2 .6 .
4 .2 Low Resistance Contacts:
For electrical measurements on the test devices, metal bonding pads
were formed by thermally evaporating 6000 A thick gold film on the
superconducting pads, through a shadow mask,
in order to obtain low
resistance contacts, the samples were annealed in an oxygen flow of 1
liter/min, for about 15 minutes at 600 °C, followed by a slow furnace cooling
for 30 minutes after the furnace was switched off. The samples were removed
when the furnace temperature was approximately 300 °C. Gold wires of 1 mil
diameter were bonded to the gold pads using a Kulicke and Soffa Model 4 1 23
ultrasonic wedge bonder. The bonding process did not require sample heating.
For redundancy, multiple bonds were attached on the contact pads.
The
contact
resistance
obtained
from
four-probe
resistance
measurements is typically a few milli-Ohms, at temperatures below the Tc. The
specific
contact
resistivity
calculated
from
the
four-probe
resistance
measurements range from 3 .6 5 * 1 0‘5 Ohm-cm2 at 90 K, to 10'8 Ohm-cm2
below 77 K. These results were reproducible from sample to sample and are
comparable with the results for Au contacts on YBaCuO high temperature
superconducting thin films[40]. A photograph of a four probe test device with
59
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
gold contacts is shown In figure 4.2.
4 .3 Resistance vs Temperature Measurements:
4.3.1 Closed cycle cryogenic system:
The closed cycle cryogenic refrigeration system consists of a CTI
cryogenics model 8 3 00 helium gas compressor controlled by model 8001
controller, model 22 cold head and associated electrical and mechanical
interconnections.
The schematic for the installation of the closed cycle
cryogenic system is shown in the figure 4 .3 .
The model 8300 compressor
assembly consists of a water cooled compressor, the 8001 controller and a
helium filtration cartridge.
The cold head consists of a two stage cold head cylinder and drive unit
assembly, that produce refrigeration that range from 60 to 120 K for the first
stage station and 10 to 20 K for the second stage station. During operation,
compressed helium gas from the compressor enters the cold head at the helium
supply connector and flows through the displacer-regenerator assembly before
exiting through the helium gas return connector and back to the compressor.
A differential pressure relief valve in the compressor controls the pressure
between the supply and return lines.
Helium expansion in the displacer-
regenerator assembly provides cooling at the cold head stations. Typically, the
system is evacuated to below 20 mTorr before the compressed helium is
supplied to the cold head. The typical refrigeration capacity for the model 22
60
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig 4 .2
Photograph of a fabricated four-probe test device for
electrical transport measurements
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Thermocouple
Vacuurr
Gauge
Roughing
Roughing
Pump
Valve
Pressure
Eq. valve
CTI model22
Refrigeration
CTI Cryogenics
system
8001 Controller
Supply
Return
Line
CTI Cryogenics
Cartrdige
8300 Compressor
Fig 4.3 Schematic diagram for the cryogenic system
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
refrigeration system is approximately 8 Watts for the first stage heat load at 77
K, and 2.0 Watts for the second stage heat load at 20 K. More details of the
cryogenic system operation can be found in the operating manual provided by
CTI Cryogenics.
4 .3 .2 DC electrical transport measurements:
Since superconductors have negligible resistance to dc and at low
frequencies, measurement of resistance with temperature shows a very rapid
transition below the onset temperature for superconductivity. The accurate
measurement of low resistance is crucial to determine the superconducting
properties such as the zero resistance T c. A computer automated dc technique
with noise control from electrical and magnetic fields was used for electrical
transport measurements.
The measurement setup for the resistance vs
temperature measurements and dc critical current density measurements is
shown in figure 4.4.
The setup consists of a Keithley model 181 nano­
voltmeter, a Keithley model 2 2 4 current source, and a Lakeshore model 805
temperature controller all of them connected to a control computer through an
IEEE 488 interface. The BASIC computer software called "SC" obtained from
Keithley
Instruments,
controls the
sequencing of the
instrumentation
measurements, records data and computes the sample resistance.
The
technique used for resistance measurements on superconductors is a four wire
measurement that eliminates lead resistance. The four wire technique consists
63
Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission.
HP7475A
Plotter
Keithley 2 2 4
Keithley 181
Program
Current
source
Nano
IBM PC
Voltmeter
Lakeshore 805
Temperature
Controller
Refrig n
CTI Cryogenics
Model 22
Fig. 4.4 Measurement setup for dc electrical transport
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
of tw o wires for passing a current through the sample and tw o wires to
measure the voltage drop across it.
To eliminate the thermal voltages,
measurements were made by applying current in both forward and reverse
directions. It is very important to complete the voltage meaurements in both
directions of current flow before the thermal gradients change causing errors
in the voltage [90].
The specimen temperature was controlled using the
temperature controller, connected to a closed cycle helium gas refrigeration
system. The error in temperature measurement was less than 0 .2 5 K. Once
the resistance of a superconducting sample is obtained, the resistivity in Ohmcm can be calculated from standard formulae. An overall resistance accuracy
of 0.1 %
with
a resolution of 1 pQ. was obtained in the
resistance
measurements[90].
The resistivity vs temperature characteristics of a 50 jjm wide fourprobe test device is shown in figure 4 .5 . A constant dc current of 10 pA was
applied between the current leads and the corresponding voltage drop was
measured across the voltage sense lines. From the voltage measurements the
resistance was computed at every temperature by the control software and the
measurements were displayed in real time on the control computer display.
From the measured resistance vs temperature characteristics, the resistivity vs
temperature characteristics were obtained. The area of cross-section of the
device for the current flow is the thickness of the film multiplied by the
linewidth.
The thickness of the thin film was measured by DEKTAK
65
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
RESISTIVITY VS TEMPERATURE AFTER PATTERNING
AND ANNEALING
1600-
1200
-
800
400-
50
100
150
200
250
T E M P E R A T U R E IK)
Fig. 4 .5
Resistance vs Temperature measurements on a 5 0 /vm four probe
test device
300
profilometer thickness measurements.
From the figure 4.5, the onset temperature for superconductivity occured
at 106 K, and the device showed zero resistance at 9 8 .5 K.
The room
temperature resistivity was 1.5x1 O'3 Ohm-cm. Zero resistance T c between 97
and 100 K is routinely obtained. The larger transition width is probably due the
mixed phase nature of the thin film.
4 .4 Jc vs Temperature Measurements:
The critical transport Jc was measured using dc and pulsed current
techniques, using a 1 /c/V/mm electric field criterion.
A Keithley model 181
nano-voltmeter and a Keithley model 224 current source were used for the dc
transport measurements. The specimen temperature was controlled using a
Lakeshore model 805 temperature controller, connected to a closed cycle
helium gas refrigeration system. The error in temperature measurement was
less than 0.25 K. Thermal equilibrium was established before measurements
at each temperature below T c. The dc current method was not used above a
current density of 104 A/cm2, since sample heating at higher currents could
cause the films to crack before measurements could be completed.
The pulsed current measurements were performed using tw o EG&G
PARC 5 2 10 lock-in amplifiers, a HP 214B pulse generator, and an adjustable
current source capable of supplying 1 A.
The measurement setup for the
pulsed current measurements is shown in figure 4.6.
The pulse generator
supplies a 10 V, 1 KHz pulse train with a 10% duty cycle. This signal is the
67
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Signal
Generator
Lock-in Amp
Current
Current pulse
Forming Network
Lock-in Amp
Voltage
Four Probe Device
Fig 4.6 Schematic diagram for pulsed Jc measurements
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
reference signal for the lock-in amplifiers. This technique is very sensitive to
the reference frequency and rejects all other frequencies. The pulsed current
is applied to the test device, and the corresponding voltage pulse is measured
across the sample using the lock-in amplifiers. The amplitude of the current
pulse at which the voltage across the sample exceeds 1 //V , yields the critical
current at a particular temperature.
The main advantage of the pulsed current technique is the reduction in
heating at the contacts due to short duration pulsed current applied. Also,
problems associated with AC ripple problems associated with current sources
could be eliminated using the pulsed current technique[91-92].
The pulsed
current measurements were compared to the dc values at as many
temperatures as possible to insure the accuracy and compatibility of the two
methods. The morphology of the finished TICaBaCuO devices was examined
in an ISI SX-30 scanning electron microscope(SEM).
The morphology was
evaluated in order to study the correlation between Jc and the microstructure
of the films.
Figure 4 .7
shows the typical zero field current density (Jc) vs
temperature measurements obtained on the four probe test devices. Current
density at zero magnetic field as high as 5x10 5 A/cm 2 at 77 K, and
approximately 1x106 A/cm2 at 60 K were obtained.
The resistivity of the
sample calculated from the l-V measurements is approximately 2 .3 8 x 1 0‘11
Ohm-cm at 77 K, much lower than any normal conductors at this temperature.
69
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
■Samplel
-j-Sample2
10
10
E
o
10
o
10
10
10
50
60
70
80
100 110
TEMPERATURE I K)
Fig. 4 .7
Zero-field critical current density vs Temperature measurements
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The surface morphology of the thin films was essentially featureless, and very
smooth, typical of high quality films.
exceeded 1 0 5 A/cm2 at 77 K.
The current density of such films
Films with numerous inter-grain boundaries
showed lower current densities below 104 A/cm2 at 77 K.
The data from the zero field current density (Jc) measurements shown
in figure 4 .7 was analyzed to investigate the dependence of Jc with the
temperature ratio T/Tc. The slope of the log Jc vs log (1-T/TC) characteristics
is an indication of the type of the superconductor.
The slope of the line
obtained from our measurements was approximately 1.5, for temperatures
between 50 and 80 K. The (1-T/TC)3/2 dependence of Jc is consistent with
earlier reports in high T c thin films[93]. This indicates that the thin films may
contain grain boundaries which are either insulating or behave like a normal
metal, or the thin films may be poly-crystalline in nature. The presence of grain
boundaries and weak flux pinning in TICaBaCuO thin films may be the main
reasons for the lower Jc in TICaBaCuO thin films compared to epitaxial insitugrown YBaCuO thin films. However, among the poly-crystalline high T c thin
films, TICaBaCuO thin films have shown superior electrical properties and hence
very attractive for electronic applications.
71
Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission.
5. TICaBaCuO SUPERCONDUCTING MICROSTRIP RESONA TORS
A microstrip resonator is a useful device for measurement of
dispersion, phase velocity and effective dielectric constants of dielectric
substrates.
Ring resonators are being widely used for realizing filters, and
stabilization of oscillators. A microstrip ring structure resonates if its electrical
length is an integral multiple of the guide wavelength. A simple ring resonator
device was designed which consisted of a ring structure separated from the
feed line by a small coupling gap. The size of the coupling gap determines the
coupling between the feed line and the ring resonator.
Loose coupling is
desired to minimize excessive loading effects. A ring resonator designed for 10
mil thick LaAIOs substrates (er = 24.5), for a fundamental resonance at 12 GHz
is shown in figure 5.1.
In the figure, the linewidth of the ring and the
microstrip feed line is W = 5.6 mils, the coupling gap G = 1.75 mils, and the
mean radius of the ring R = (R 1+R 2)/2 = 77 mils.
The characteristics
impedance of the microstrip is 41 Ohms at 12 GHz. The details of the design
of the ring resonator have been described by Chorey et. al.f[56J.
In theory, the ring resonator allows only the resonance frequency signal
to pass by, and hence acts as a narrow band filter. The high selectivity of the
ring resonator is useful in stabilization of oscillators in communication systems.
Superconducting microstrip ring resonators offer the advantages of higher Q
compared to gold based microstrip ring resonators.
72
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
W
=
R
=
G
=
Fig. 5.1
5 . 6 M IL S
7 7 M IL S
1 .7 5 M IL S
Geometry
of
a
microstrip
ring
resonator designed
fundamental resonance frequency of 12 GHz
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
for
5. 1 Fabrication and testing:
TICaBaCuO ring resonators were fabricated by patterning 0 .3 pm
thin films using AZ 1421 positive photoresist photolithography and wet
chemical etching techniques similar to the process used for fabricating four
probe test devices described above[35,94]. The ring resonators were annealed
using the same annealing procedure described above.
The samples were
divided into two groups: one set of samples with 1 pm gold film on the bottom
side of the LaAI03 substrate for the ground plane formation and the second set
with 0 .3 pm TICaBaCuO superconducting thin film ground plane. The ground
plane side superconductor was deposited and post-processed using our routine
post-deposition methods described in chapter 2, after the microstrip ring
resonator was fabricated on the top side. Figure 5.2 shows a photograph of
a fabricated superconducting ring resonator.
A ring resonator was mounted in a gold plated Copper test fixture of 1"
wide, 2" long and 1" thickness. The test fixture was placed on the cold head
of the helium gas closed cycle cryogenic system. Electrical connection to the
feed line was obtained by mechanical contact of a launcher at the input side of
the test fixture. Connections to the HP 8720 network analyzer were made
using a 0.1 41 " semi-rigid co-axial cable of 50 ohms characteristic impedance.
Before measurements were performed on ring resonators, standard one port
calibration was performed at room temperature. The calibration was performed
using an open, a short and a broadband load to effectively remove the test
74
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. 5.2
Photograph of a fabricated ail superconducting microstrip
ring resonator
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
system imperfections introduced by the interconnecting cables, adapters etc.
The calibration was also valid at lower temperatures.
The resonator quality factor(Q), the ratio of the energy stored in the
resonator to the energy dissipated in the resonator was obtained from the
swept frequency reflection measurements[95]. The Q value is a figure of merit
for a resonator, and is inversely proportional to the total losses in the circuit.
The measured Q called the loaded Q is a measure of the circuit losses including
the coupling loss and the loss through the feed line. The actual Q of the ring
resonator called the unloaded Q is a measure of the losses only in the
resonator. The unloaded Q is obtained by separating the external losses in the
feed line and due to coupling. The loaded Q and the unloaded Q are related
through the reflection coefficients at resonance and far from the resonance[91].
The derivation for the relationship between the loaded Q and the unloaded Q
is described by Aitken[95].
A computer programme was written to compute
the unloaded Q values from the measured loaded Q and the magnitude of
reflection coefficients at resonance and far from resonance. The determination
of whether the resonator was overcoupled or undercoupled was made from the
Smith chart and also the phase response of the resonator. Typically, the ring
resonators were overcoupled. Measurements for the superconducting resonator
were performed at the fundamental resonance frequency of 12 GHz, and an
input power level of -30 dBm. Figure 5.3 shows the frequency response of the
input reflected powerlS-j-]). T h eS n is approximately-50 dB at 12.1 GHz. The
76
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
i
log MAG
S-\-\
1 0 d B /d lv
Cor
OdB >
1 1 .8 6 GHz
Fig. 5 .3
1 2 .5 GHz
Measured input reflected p o w e rlS ^ ) vs frequency for an all­
superconducting ring resonator
swept frequency reflected power shown in the figure exhibits dual resonance
mainly due to a damaged portion of the ring causing an additional resonance
peak.
The very low value of reflected power indicates almost 100%
transmission to the output if there is an output port.
5 .2 Unloaded Q vs Temperature:
The unloaded Q versus temperature characteristics for a high T c thin film
ring resonator with a superconducting ground plane is shown in figure 5.4. For
comparison, unloaded Q for a gold resonator with a gold ground plane is also
shown in curve B. The unloaded Q of the ring resonator with superconducting
ground plane is approximately four times higher than the goid resonator at 65
K. In addition, the unloaded Q of the superconducting ring resonator shows an
increasing
trend
in
Q
with
decreasing
temperature,
whereas
the
superconducting ring resonators with gold ground plane showed a saturation
of Q at low temperatures due to the dominance of ground plane conductor
losses.
The superconducting ring resonators offer an indirect method for
measuring the effective surface resistance (Rs) of the superconducting thin
films.
This microwave surface resistance is the fundamental quantity
responsible for the conductor losses at high frequencies. The Rs of sputtered
thin films were obtained from ring resonator quality factor (Q) measurements.
By separating the conductor and dielectric losses, the surface resistance of the
78
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UNLOADED Q V S TEMPERATURE
2000
1500
Q
• A
-
10 00 -
500
-
70
80
TEMPERATURE in K
Fig. 5 .4
90
100
Resonator unloaded Q vs Temperature characteristics of a
TICaBaCuO superconducting ring resonator(A) and for a gold
resonator(B)
TICaBaCuO thin films was calculated using the standard microstrip loss
equations described by Pucel et al[68]. The effective Rs at 12 GHz, and 77 K
was determined to be typically between 1.5 and 2.75 mQ, almost an order of
magnitude lower than Rs of Cu at the same temperature and frequency. The
lowest surface resistance reported in TI2122 thin films to date is 0 .1 3 0 mQ at
77 K and 10 GHz[96].
5 .3 Resonance frequency vs temperature:
The swept frequency reflection measurements performed at several
temperatures, is also used in determining the penetration depth of the
TICaBaCuO superconducting thin films.
The resonance frequency is the
frequency at which the magnitude of the reflection coefficient is at the
minimum. The resonance frequency was measured at each temperature for
ring resonators. A typical measured resonance frequency shift with respect to
temperature for a superconducting ring resonator with approximately 1 pm
thick gold ground plane is shown in figure 5.5.
The shift in resonance
frequency with temperature is mainly due to the temperature dependence of the
penetration depth of the superconductor. Thus, the resonance frequency shift
is an indirect method of determining the penetration depth. From the figure,
the change in resonance frequency below 70 K is almost negligible.
The
superconducting resonators with a 0.3 pm thick superconducting ground plane
showed a slightly higher dependence of resonance frequency with temperature
80
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
RESONANCE FREQUENCY VS TEMPERATURE
GHz
12.4
12.3
FREQUENCY
in
-
12.2
-
RESONANCE
12.1
12.0
*
11.9
60
70
80
TEM PE RA TU R E
Fig. 5.5
90
in
K
Resonance frequency vs Temperature characteristics of a
TICaBaCuO superconducting ring resonator
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
100
due to the temperature dependence of penetration depths of the top and the
ground plane superconductors. The detailed analysis of this figure to determine
the penetration depth of the superconducting thin films, is given in the next
section.
5 .4 ANAL YSIS AND D/SCUSSfONS
The penetration depth of TICaBaCuO thin films can be determined from
the resonance frequency vs temperature measurements, by comparing the
experimental data shown in figure 5.5 with theoretical calculations.
The
resonance frequency shift in the ring resonators is assumed to be due to the
change in penetration depth with temperature. Neglecting the effects due to
the substrate contraction at lower temperatures, the effective penetration depth
was extracted from the resonance frequency shift as discussed below.
The phase velocity of a superconducting microstrip transmission line with
a superconducting ground plane is given by[97],
vph = c h /e e ii* ^ +2*M /h*coth(t//l)}'0'5
— > (1 )
where c is the velocity of light, eeff is the effective dielectric constant, h is the
substrate thickness, t is the thickness of the microstrip, A the penetration depth
of the superconducting microstrip.
The penetration depth is temperature
dependent based on the Gorter-Casimir relationship[98] ie.,
AIT) = ^{0)*[1-O7Tc)4]'0'5
— > (2)
for temperature T less than Tc. A(0) is the penetration depth at T = 0 K. The.
82
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
resonance frequency of the ring resonator is given by the equation
f = n *v ph/(2*L)
-->
(3)
where f is in GHz, L is the mean circumference of the ring in mm, and n is the
integer order of resonance. From the temperature dependence of resonance
frequency measurements and the above equations, the best value of >1(0) was
determined to be 6890 A. The typical value ranges between 7 0 00 A and 8 0 00
A. This is an effective value for the penetration depth in the TICaBaCuO thin
films.
Since the thin films are only 0 .3 -0 .4 fjm thick, the penetration depth
depends upon the properties of the superconductor through the entire film.
This may be a reason for the high value of penetration depth.
Also, the
patterned thin films have rough edges, and hence the penetration depth
obtained using the above technique is an averaged value over the whole film
area.
Typical values of penetration depth reported in literature are between
4 0 0 0 and 8000 A in TI2122 thin film s[99-100].
The surface resistance of the TICaBaCuO superconducting thin films
determined from the ring resonator Q measurements was compared with the
theoretical surface resistance vs temperature characteristics for a given
penetration depth. A theoretical model based on the Phenomenological loss
Equivalence Method (PEM) approximation [65-66] was employed to determine
the theoretical variation of conductor losses and the surface resistance with
temperature for the cases of superconducting microstrip/gold ground plane, and
superconducting microstrip/superconducting ground plane. Both these cases
83
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
were compared to the attenuation constant of a gold microstrip on LaAIOs
substrate.
The attenuation constant for a superconducting microstrip is calculated
from the formula[66],
a = (T/Tc)4/[1-(T/Tc)4]3/2 * G-]/4 * an/Z * w 2* p 2 * M 0)3* coth(X)
+ X cosec2(X) Np/m
— > (4)
where X = A * G^/KO) * [1-(T/Tc)4] 1/2.
G t is the geometric factor given by the equation
Gn = 1 /(/7h) *[1 -(W e/(4h))2] *[1 /2 + h/W e + h/(/7We) *ln(2h/t) — > (5)
W e is the effective width of the microstrip, and A is the area of cross-section
of the microstrip, T is the measurement temperature below Tc, and A(0) the
penetration depth at 0 K of the superconductor.
The parameters assumed for the calculations are the relative dielectric
constant (er) of LaAI03 to be 24.5, the loss tangent (tan 6) of LaAI03 to be
8 .3 *1 O'5 below 100 K, the substrate thickness (h) of 10 mil, the width of the
microstrip (W) of 142 pm, corresponding to a characteristic impedance of 41
ohms at 12 GHz, the thickness of the superconducting microstrip(t) to be 0.3
pm, the ground plane thickness of 1 pm for gold ground plane and 0 .3 pm for
superconducting ground plane, the zero resistance T c of the TICaBaCuO thin
films was to be 100 K, and the normal conductivity at T c (<7n) of 1 .5 * 1 0 6 S/m.
The ground plane conductor losses can be calculated by the same
method, using the geometric factor G2 instead of G1 in the equation 4.
84
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
G2 = 1/(2/rh) * [1-(W e/4h)2]
— > (6)
Figure 5.6 and 5.7 show temperature variation of the attenuation due to
conductor losses for a gold microstrip, a superconducting microstrip with a gold
ground plane, and a superconducting microstrip with a superconducting ground
plane as determined using equations 4-6 for the penetration depth at 0 K of
6 0 00 A and 7000 A respectively. The figures show the lower attenuation for
the microstrip with superconducting ground plane compared to the one with
gold ground-plane below 77 K.
The surface resistance of the superconducting thin film is obtained from
the equation
Rs = 2 Z0 a! G,
— > (7)
where Z0 is the characteristic impedance of the microstrip. The theoretical
temperature variation of surface resistance of the superconducting microstrip
with a superconducting ground-plane determined using the equation 7 is shown
in figure 5.8, for A{0) of 6000 A and 7000 A.
For comparison, the surface
resistance of a 1 jum thick gold microstrip on LaAI03 substrate, is plotted for
the same microstrip geometry. The Rs calculated from the measured Q values
of an all-superconducting ring resonator on LaAI03 substrate is also plotted in
figure 5.8. The Rs obtained from the ring resonator Q measurements (curve D)
deviates from the theoretical temperature dependence as seen in the figure.
Since the TICaBaCuO thin films do react with the LaAI03 substrate, it is
possible that the region of interaction contributes to additional losses.
85
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Since
Attenuation vs Temperature
Attenuation in dB/cm
10
0.1
0.01
0.001
40
50
60
70
80
90
Temperature in K
—■— gold
Fig. 5 .6
I
supr/gold
supr/supr
Effect of superconducting ground plane on microstrip
attenuation for/l(0) = 6000 A
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
100
Attenuation vs Temperature
Attenuation in dB/cm
10
1
0.1
0.01
0.001
40
50
60
70
80
90
Temperature in K
—*
Fig. 5 .7
gold
1
supr/gold
~supr/supr
Effect of superconducting ground plane on microstrip
attenuation for yl (0) = 7000 A
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
100
Rs vs Temperature
Surface Resistance in ohms
0.1
0.01
1 .000E-03
1 .00QE-04
40
50
60
70
80
100
90
Temperature in K
—■
gold
f— supr(L = 6 0 0 0 A)
supr(L = 7 0 0 0 A)
Fig. 5.8
Theoretical
and
“ S ” supr/supr(expt.)
experimental
surface
resistance
temperature characteristics
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
vs
BCS theory is valid only for homogenous superconducting materials, the
deviation may not be modelled accurately by the two fluid theory.
For the superconducting ring resonators with gold ground-plane, the
effective surface resistance for the superconducting microstrip was obtained
from the formula
Rs = Rsg - 2nZ0l(Ag G ,) {1/Q g - 1 /Q J
-->
(8)
where Rsg is the surface resistance of the gold microstrip, Qg is the conductor
Q for the gold microstrip, and Qs is the conductor Q for the superconducting
microstrip.
The ring resonators fabricated and tested were not designed for optimum
coupling to maximize the Q of the resonator. The optimum coupling gap to
minimize the loading effects and improve the Q is approximately 5.5 mils as
determined by HP Microwave Design System (HPMDS) CAE software.
89
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6. SUPERCONDUCTOR SEMICONDUCTOR HYBRID PHASE SHIFTER
CIRCUITS
A high T c superconductor and compound semiconductor based hybrid
digital phase shifter was investigated in this research.
TI-Ca-Ba-Cu-0
superconducting thin films and GaAs MESFETs were used in the circuit. The
phase shifter circuit was designed for 180 degrees phase shift using the
reflection type configuration[101-103]. The superconducting circuit consists
of input and output feed lines, 3 dB Lange Coupler which divides the input
power by half into the direct and coupled arms, and impedance transforming
networks for matching the impedance of the switching devices.
GaAs
MESFETs were used for switching application in the hybrid circuit. The phase
shifters were designed for 4 GHz center frequency, and operation at 77 K. The
180 degrees phase bit was designed for a minimum insertion loss in the on and
off states of the switching devices, a minimum bandwidth of 25% , the phase
error less than 10 degrees, and input and output return losses greater than 15
dB.
The lower conductor losses in superconductors would lead to lower
insertion loss of the phase shifter circuit.
6.1 Reflection type phase shifters:
In general, the digital phase shifters could be realized using three
different configurations. 1. Loaded line phase shifters, 2. Reflection type phase
shifters, and 3. Switched line phase shifters. Among the three, the reflection
90
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
type phase shifters are most suitable for large phase bits, due to their low
insertion loss, wide bandwidth, and the use of only tw o switching devices per
bit. Typically, in a multi-bit phase shifter circuit, the bits below 45 degrees are
realized using the loaded line type, and the higher bits, 9 0 and 180 degrees, are
realized using the reflection type.
In this research, a 180 degrees phase
shifting bit is designed using the reflection type configuration.
The circuit
consists of a 3 dB coupler that divides the input power equally between the
coupled and direct ports.
The Lange coupler is a special microstrip 3 dB coupler which provides
octave bandwidth.
The schematic diagram for a Lange coupler is shown in
figure 6.1. The device is a four port device consisting of interdigitated fingers.
The interdigital coupling section compensates for the odd and even mode phase
velocity dispersion over a wide frequency range[69]. If the width of the inter­
digitated fingers and the spacing between the fingers is designed carefully, the
input signal power can be divided equally between the direct and coupled ports.
To improve the coupling over a wide frequency range, bonding short gold wires
between alternate fingers is necessary. The bond wires should be as short as
possible to minimize inductance.
The advantages of this structure is the
increase in spacing compared to simple microstrip 3 dB coupler, and improved
bandwidth. The total length of the coupler should be close to a quarter of a
wavelength[69].
For phase shifter applications, careful design of impedance transforming
91
Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission.
Output
Direct
Input
Coupled
Output
CMPl
MSLMGE
Direct
1 >
Input
Fig. 6.1
Coupled
Schematic diagram for a Lange Coupler
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
networks and proper terminations using semiconductor switches in the coupled
and direct ports could lead to any desired phase shift. In this research, GaAs
MESFETs were investigated for switching application. The desired phase shift
could be obtained by changing the switching states of the device.
For
example, a phase shift of 180 degrees can be obtained by terminating the
direct and coupled ports between an ideal open and an ideal short circuit. The
bandwidth of the circuit is limited by the type of the coupler employed. The
inter-digitated Lange Coupler was chosen to obtain a large bandwidth of at
least 25% . Also, the inter-digitated Lange coupler could be easily realized in
the microstrip form.
The main steps in the design, fabrication, and characterization of the
phase shifters are:
1. Microwave characterization of a GaAs MESFET for the switching device.
Two port s-parameter measurements are required during the 'on' and 'off'
states of the device, to determine the equivalent circuits in both states.
2. Design of impedance transforming networks for realization of 90 and 180
degree phase bits.
3. Phase shifter circuit design using HPMDS CAD software, for 20 mil thick
LaAI03 substrates.
4. Layout and mask generation for circuit fabrication.
5. Fabrication of the circuits using gold and TICaBaCuO thin films.
6. Packaging and testing the phase shifter circuits at cryogenic temperatures.
93
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The phase shifters were designed for 4 GHz center frequency, and
operation at 77 K. The 180 degrees phase bit was designed for: (a) minimum
insertion loss, (b) a minimum bandwidth of 25 % , (c) the phase error less than
10 degrees, and (d) input and output return losses greater than 15 dB.
6.2.
DESIGN
The first step in the design of the phase shifter was determining the
switching parameters such as the insertion loss and isolation for the GaAs
MESFETs used.
Switching parameters for the MESFETs were obtained by
packaging and testing the MESFETs for scattering parameters.
From the
switching parameters, an equivalent circuit approximation was determined to
facilitate the design of impedance transformation networks and subsequently
the whole phase shifter circuit.
6 .2 .1 S parameter measurements o f GaAs MESFET switches:
The GaAs MESFETs could be used for switching applications at the lower
end of the microwave frequencies since they have advantages such as low
insertion loss, single power supply requirements and high switching speed[106108].
The gate control of the drain-source current gives two distinct
impedance states for a MESFET.
In a normally 'on' depletion type MESFET
94
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
with n-type doped channel, the MESFET will be in the low impedance state for
zero gate bias. The impedance of the device in the 'on' state is mainly due to
the channel resistance.
When a negative bias is applied to the gate, the
channel is depleted of carriers and hence the channel resistance increases, and
the capacitance between the source and drain decreases. This gives rise to the
high impedance state ('off' state). A GEC Marconi general purpose amplifier
GaAs MESFET (P35-110) was chosen for switching, which has a high gm
typically 45 mS, low gate leakage current typically 10 nA at -5 V, a low gatesource capacitance (Cgs) typically 0 .4 pF at low frequencies, and large pads for
ease of bonding. The width of the device is 2x150 pm. The MESFETs were
characterized for switching parameters.
A special chip carrier was made in-
house for testing the MESFETs [Appendix A]. The carrier consists of input and
output microstrip lines on alumina substrates, attached to the sample block
using a conductive silver epoxy. The microstrip lines were designed for 50
Ohms characteristic impedance. The device was mounted in the gap between
the input and output microstrips. The device was packaged with a 5.1 KQ
resistor for gate bias, to provide isolation from the rf path. A standard two port
thru-reflect-line (TRL) calibration was performed between 2.5 and 5.5 GHz
[Appendix A].
The insertion loss, S21 in the on state (VGS = 0.5 V) of the MESFET,
and the isolation S21 in the off state of the MESFET (VGS = -1.25 V) were
measured.
The insertion loss is approximately 1.5 dB at 4 GHz and room
95
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
temperature. The isolation for the device was approximately 16.5 dB at 4 GHz.
Since the superconducting circuits are designed for temperatures below 1.00 K,
the MESFETs were also tested at cryogenic temperatures. The TRL calibration
was performed at room temperature and the calibration was assumed to be
valid at lower temperatures. Figure 6 .2 .a shows the insertion loss and isolation
of the MESFET switch obtained at 90 K. The insertion loss of the device was
approximately 1.2 dB at 4 GHz, when a positive bias of 0.5 V was applied.
The isolation of the device was approximately 15 dB. The equivalent circuits
for the on and off states of the MESFET switch were extracted following a
simplified model[106j.
The on state impedance was assumed to be purely
resistive and the off-state impedance was modelled by a parallel R and C
combination.
The equivalent circuit obtained for both states at cryogenic
temperatures is shown in figure 6.2.b.
6 .2 .2 Design o f hybrid reflection type phase shifter bit for 780 degrees phase
shift:
For modelling of the HTS thin films the kinetic inductance effects should
be included when the penetration depth is in the order of film thickness.
Several models have been reported in literature which in general are based on
the calculation of internal impedance using complex conductivity approximation
for the superconductor[65-67].
The author's main objective was to design
96
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CrT STATE= -15.281 i+00
lM .0 1 8 3 b [)9
on stat :
C9 STATE= -1.2330E
11=4.0183 ;+09
fre q
5 .5
GHz A
fre q
5 .5
GHz C
CMF73
R
CHP'2
R
R=1.471 KOH
R=19.26 OH
C=0.06869 pF
on state
Fig. 6 .2
off state
(a ) GaAs MESFET switching on state insertion loss and off-state
isolation and (b) equivalent circuits for on and o ff states of the
MESFETs
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
superconducting microstrips based on the characteristic impedance changes
due to the kinetic inductance. By knowing the superconducting properties of
the films, the kinetic inductance and hence the wave slowing factor were
calculated. An effective characteristic impedance was determined for which
the superconducting microstrips were designed so that when the sample is
superconducting the impedance of the microstrip would be the desired
characteristic impedance. The lossy superconductor can be treated as a normal
conductor with complex conductivity, as in the PEM model[65]. In this model,
the distributed internal impedance of the superconductor is obtained from an
equivalent single strip model for the planar quasi TEM line. The attenuation
coefficient in Np/m can be obtained from the approximation Ri/(2Z) where Ri
is the real part of internal impedance and Z is the characteristic impedance of
the microstrip line.
The propagation constant of the microstrip line can be
obtained from the approximation Xi/(2Z) where Xi is the imaginary part of the
internal impedance. From the distributed internal impedance, the propagation
characteristics of the superconducting transmission line can be obtained. The
following procedure was followed:
1.
Calculate the internal inductance(Li) of the superconducting microstrip,
based on the PEM model[65].
2. Calculate the effective wave slowing factor n from the additional inductance
introduced due to the field penetration[67].
3.
Calculate the effective characteristic impedance (Ze) of the microstrip
98
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
following the model proposed by Antsos[67].
4. Design the microstrip for this effective characteristic impedance, so that the
impedances of the normal and superconducting microstrips will be the same at
a particular temperature.
The following assumptions were considered in the model:
1.
Width of the microstrip line: 6 .5 07 9 mil for 50 Ohms characteristic
impedance.
2. Substrate: LaAI03 with relative dielectric constant of 24.5 and loss tangent
of 1e-4.
The temperature variations of the substrate parameters were
neglected.
3. The T c of the superconducting thin films: 100 K
4. The normal conductivity of the superconductor at Tc: 1e6 S/m
Based on these above procedure and the assumptions, a program was written
in Quick Basic, to
compute the
inductance vs penetration depth
in
superconducting microstrips at a particular temperature. The thickness of the
film was used as a variable. The inductance increases as the penetration depth
increases, and this change in inductance is mainly due to the field penetration
into the film. The change in inductance causes the change in characteristic
impedance.
The effective characteristic impedance vs penetration depth
characteristics is shown in the figure 6.3, with the film thickness as a
parameter. The measurement temperature was assumed to be 77 K. Again Ze
differs from Z by a large percentage at higher penetration depths and smaller
99
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Characteristic Impedance in Ohms
50
48
46
44
42
40
0.2
0
0.6
0 .4
1
0.8
Penetration depth in um
"■—
t = 0 .5 um
t = 0 .2 um
Fig. 6.3
I
*
t = 0 .4 um
t = 0 .3 um
t = 0 .1 um
Characteristic impedance vs penetration depth at 0 K for
various film thickness
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1.2
thickness of the microstrip.
The design of the phase shifter circuit was performed using HP
Microwave Development Software (HPMDS). The schematic diagram for the
hybrid phase shifter circuit is shown in figure 6.4. The circuit consists of a
Lange coupler for 3 dB coupling, impedance transforming networks and the
transistor switches, similar to the design by Andricos et al[101].
The
impedance transforming network was designed using open circuited stubs to
match the impedance of the switching device to the microstrip line.
The
optimization variables were the lengths of the open circuit stubs, input and
output feed lines, and the mid-section of the matching network. The widths
of the lines were optimized for 50 D characteristic impedance.
The main
reason to fix the characteristic impedance to 50 Ohms is to improve the validity
of the comparison of gold and superconducting circuits for the same
characteristic impedance. The circuit pages for on and off state conditions
were created in the same design file, the off state circuit as circuit 1 and
onstate circuit as circuit 2. The optimization process for the on and off state
conditions were performed simultaneously for this circuit. Figure 6.5a shows
the circuit page for the on state of the MESFETs and figure 6.5b shows the
circuit page for the off state of the MESFETs.
Figure 6.5b contains the
optimization goals such as low insertion loss and phase shift close to 180
degrees.
After the optimization process, the layout for the superconducting
circuit could be generated using HPMDS in an EGS Archive file for mask
101
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
)
(
Output
feed
line
Input
feed
line
t
Lange
— .<
Coupler
Impedance
tranfrmn
network
GaAs
GaAs
j
5 .IK
i ; nd
W
CND
"\7
Vgs
Fig
6.4
S c h e m a t i c diagram for the
phase shifter circuit
hybrid
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
P0RT11UH--2
R=50.0 Oil
JX=0.0 Oil
PORT SPM
5UHST=L<iAl03
l.=93.65 m il
H=N50
HSM
SO
CITIM
KSIL
AGROUND
win
K
SIXW
.t
H3=W50
H3--H50
drilo
HIRES
'yW I =W50
SUlli'rHjM03
SUBSMaAIOJ
Hl=Hb0c«.MV2=H30
I,=1.1
W2=H50
H---HS0
HI=H50
L=i,o
R=30 m il
R=0.'j m il
31.1 um
COIIII 1.0C.II53
UBST-UM03
,= 9 3 .l m il
KSIL
SUHST=,aA 03
WF=25.
H'I0W
77777777777777
l,=L6
11=30 m il
11=0.5 m il
drm C OtH)=1.0F.U53
SUHST=I.j A103
SUDST=l.aA 03
H-H50 \ \
AMG=90 (Icq \ V
H=°.5
\ J L
SUHST=l,oA103
^
EQUATION H50=6.08022 m il
EQUATION Ll = (10<10.99K 120) m il
F.QUATI0H l,2=(20.01<93.028<90) m il
EOUATIOII 1.1= (1 0 0 (2 1 3 .360<250) m il
cirl
KSSUIiSTRAIC
P0R7tlUH=t
R=50.0 Oil
JX=0.0 Oil
W IL L
PORI SPUR
V'
AGROUND
_ --------------
SUH5T=loA103
HU=2000 m il
!;!!^2} ,5
11=20 m il
C01IU=4.1g 199
ROUGH=0 um
| .T=o.oooooP|2i
J_____
TAHU 0.0001
EQUATION I,5=(15<31.676<120) m il
cirill
M NS.<ljtasot=htslB0
* S-PARAMF.TER *
SIMULATION
SHFPT VAR-FRK0
STIHGROUP=s t i kgroup
FRF.U B
OUTI'UI VARS=B
CKP1I5
STP.SIPLOG
STIMULUS
5TIMGR0UP=STIHGR0UP
START=3.5 GII2,
ST0I’ =4.5 GHZ
LOG PTS=201
RKVURSF 111)
Fig. 6.5,1. Circuit, dosiqn |>.iqp from lll’HDS foi llio Off-stale of the GaAs MRSFK.Ts
AGROUND
C=0.069 [>r
CIVR108
0
0
Aiini=o
AHG2=0
SUI)ST=l,aAI03
I.=91 .65 m il
^7
mmmmm.
^,^=UM03
simsT=i,8H(y»
HI =W50 lriT(R=H50
CMP 167
R=1411 Oil
SUI1ST-I,JAI03
^7
AGROUND
0=0.069 pF
of tbo HaAs HFSFFTa
i ic<n IlltfJS foi I bo onstalo
•- °
g-r.
Fitj. 6.5b.
Circuit. <icsiqn ptjo
rp«yy:
ss
|55
111
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
generation.
Both gold and superconducting microstrip phase shifter circuits were
optimized using the HPMDS.
The modelled phase shift of the circuits were
between 176 and 184 degrees, less than 1 0 ° phase error in the frequency
range 3 .5 -4.5 GHz. The insertion loss of the hybrid circuits were approximately
1.0 dB in the off state of the switching devices and 3 .1 6 dB in the on state of
the switching devices.
Figure 6.6 shows the theoretical simulation results
obtained from the HPMDS.
6.3. EXPERIMENTAL
TI2Ca1Ba2Cu2Ox (2122) superconducting thin films were fabricated by
rf magnetron sputtering from a sintered TI2122 powder target in a pure argon
plasma, and post-annealing methods [81-82]].
Thin films of approximately
35 0 0 A were deposited on the double side polished 2" LaAI03 substrates. The
film side was coated with a photoresist layer before the substrate was cut into
desired dimensions. Thin films were sintered in air in an optimum Tl partial
pressure at 8 5 0 °C for 10 minutes in the confined surface configuration to
obtain superconductivity [81].
The sintered TICaBaCuO thin films were
patterned into the phase shifter circuit using standard positive photoresist
photolithography. The films had to be coated with a HMDS adhesion layer
before the photoresist AZ-1421 was spun. After the photolithography, the
films were etched in a 1 : 100 phosphoric acid : Dl H20 solution. The samples
105
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
72
Fig. 6.6 Theoretical simulation results for the 180 degrees
phase bit using HPMDS
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
were annealed in oxygen at 750
°C for 10 minutes.
An additional
photolithography step was needed for lift-off metallization for gold contact
pads. The contact pads of 4 0 00 A were annealed at 4 5 0 °C for 10 minutes
in a flowing oxygen (1 lit/min), and then rapidly cooled on a fire-brick. The
ground plane silver was evaporated to a thickness of approximately 2 pm to
form the phase shifter circuit. The gold based circuits were fabricated by gold
plating, patterning and etching methods described in Appendix B. The GaAs
MESFETs were mounted on the support block along with the gate isolation
resistor (5.1 KQ).
Gold wires were bonded to package the circuit using an
ultrasonic wire bonder. A photograph of the superconducting circuit is shown
figure 6.7. The photograph of a packaged hybrid phase shifter circuit is shown
in figure 6.8. The packaged circuit was mounted in the cryogenic system after
a tw o port TRL calibration was performed. The circuit was characterized by
measuring s parameters at cryogenic temperatures.
6.4. RESULTS AND DISCUSSIONS
The TICaBaCuO
based
hybrid circuit showed a phase shift of
approximately 18 0° at 4.1 GHz below 90 K. The measured phase shift vs
frequency characteristics for a hybrid circuit is shown in figure 6.9.
The
measurement was performed at a cold head temperature of approximately 50
K. The phase of S21 in the on state of the MESFETs was stored in the memory
107
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. 6.7
Photograph of the superconducting circuit part of the hybrid
phase shifter
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. 6.8.
Photograph of a hybrid 180 degrees phase shifter after
packaging
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHI
S2 i -M
phase
90
°/
REF
0
°
3;
O U T P P L )T ; tr ir t
C2
4 . 07E 7
START
Fig. 6 .9
3 .5 0 5 0
°
GHz
J -1 5 E
3 . 9 0 6 1 GHZ
MARI E R
. 07E
15B .S
Hz
GHz
STO P
4 .4 8 5 0
GHz
Phase shift vs frequency characteristics for a superconductor
based hybrid 180 degrees phase bit
of the automatic network anaiyser(ANA). The device was turned off, and then
the phase shift was obtained using the command "data-memory" in the ANA,
in the phase format. This gives the phase shift in degrees, as shown in the
figure 6 .9 . The bandwidth for 10 degrees phase error was typically less than
200 MHz for the superconductor based phase shifter without the bond wires
across the Lange coupler fingers. The lower bandwidth is due to poor coupling
over the bandwidth. Improved coupling has been obtained by wire bonding
alternate fingers of the Lange coupler.
The phase shift vs frequency
characteristics obtained for a gold based hybrid circuit with Lange coupler
fingers bonded is shown in figure 6.10. The figure shows improved bandwidth
of 550 MHz for 10 degrees phase error.
The temperature dependence of center frequency was also measured
experimentally. The center frequency was defined as the frequency at which
the phase shift was closest to 180 °. The figure 6.11 shows the temperature
dependence of the center frequency measured for a HTS hybrid phase shifter
circuit.
The shift in center frequency is probably due to the temperature
dependence of superconducting properties of the microstrip. The shift in center
frequency is very small for temperatures below 60 K.
The insertion loss vs frequency characteristics for the same hybrid phase
shifter circuit is shown in figure 6.12. The measurement was performed at 70
K, with gate bias at 0.5 V. The lowest insertion loss obtained during the on
state of the MESFETs was approximately 1.79 dB at 4 GHz, compared to the
111
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHI
S2 1 -M
phase
90
»/
REF
0
°
a
0 U T P P L 3T: i t i t i
1 71.07
°
A.OBE A G H z
J
17E . 2 2 °
3 . 7 1 C5 G H z
MARI <ER
3
4 . O B E i4
G Hz
?: - 1 7 C . 0 4 °
3 . 9 3 1 4 GHz
n
.X
l
-----------
—A
A
2
START
Fig. 6 .1 0
3 .5 0 5 0
GHz
STO P
A . 4B50
GHz
Phase shift vs frequency characteristics for a gold based 180
degrees phase bit after bonding alternate Lange coupler fingers
Center Frequency vs Temperature
HTS hybrid phase shifter
Center frequency in GHz
4 .2
4 .1 5
4.1
4 .0 5
40
50
60
70
80
90
Temperature in K
Fig. 6.11
Temperature dependence of center frequency for the 1 8 0 °
superconducting phase bit
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
100
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHI s21
log MAG
10 dB/
REF 0 dB
2; “ 1.7993 dB
o u r p p L o i ; n n ttO J T P P L i )T ; it h i
4 . OOC 1 GHz
jL - 9 . 7C 3 B d B
3 . 5 0 E 0 GHZ
3: - 3 . 4 E
4 . /B E 6
D
17
dB
OHz
___
A
A ^
•
START
Fig. 6.12
3 .5 0 5 0
GHz
S TO P
4 . 4 850
GHz
Insertion loss vs frequency characteristics for a superconductor
based hybrid circuit in the on state of the GaAs MESFETs
insertion loss of approximately 3.1 dB for a gold based 180 degrees hybrid
phase bit. The improvement in the superconductor based circuit is mainly due
to the lower conductor losses in the superconducting microstrip transmission
lines.
The insertion loss obtained in the off state of the devices (ILoff) is shown
in figure 6.13.
The minimum insertion loss measured in the bandwidth of 3.5-
4.5 GHz was 2 .9 4 dB at 70 K for the superconductor based hybrid phase
shifter. The best value obtained for a gold based phase shifter was
approximately 3 .2 7 dB in the same bandwidth.
The input and output return losses were also measured in both on and
off states of the MESFETs. The measurement results for a gold based hybrid
circuit are summarized in table 1. The meausrement results for a HTS based
hybrid circuit are summarized in table 2.
The temperature dependence of insertion loss of the circuit is shown in
the figure 6.14. The temperature correction was included in this measurement
to account for the difference between the temperatures at the cold head and
the sample. From our calibration, the sample temperature was atleast 10 K
higher than the cold head temperature. The figure shows the insertion loss vs
the sample temperature. From the figure, the insertion loss is very high above
100 K and decreases to below 5 dB at 90 K.
The insertion loss does not
change appreciably below 70 K.
115
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
C H I Sa l
lo g
MAG
10 d B /
REF 0 dB
3 J -2 .9 3 B 9
0 J T P P L )T ; f t m
A . 4BE 0
i: - 1 1 . * 2B
dB
GHz
dB
3 .5 0 E 0 GHz
MARI ; e r
3
4 . 4BE i
GH
2: - 5 .9 E 50 dB
A . doc 0 G H z
7
______
.
------------
......... V
\
r — -
S T A F tT
Fig. 6.13
3 .5 0 5 0
GHz
STOP
A. 4 B 50
GHz
Insertion loss vs frequency characteristics for a superconductor
based hybrid circuit in the off state of the GaAs MESFETs
Insertion Loss vs Temperature
Superconducting hybrid phase shifter
Insertion Loss in dB
30
25
20
15
10
40
50
60
70
90
80
100
Temperature
—
Fig. 6 .1 4
IL(ON)
—
IL(OFF)
Insertion loss vs temperature characteristics for the
superconducting hybrid 1 8 0 ° phase bit
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
110
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
TABLE 6.1
SUMMARY OF RESULTS FOR GOLD BASED HYBRID PHASE SHIFTERS
PARAMETER
@ON STATE*
@ OFF STATE*
Insertion Loss (S21) dB
-3 .1 2
-3 .2 7 2 7
Input Reflection Loss
-8
-1 6 .3 4
-3 .4
-5 .0 6 7
(S-j <j) dB
Output Reflection Loss
(S22) dB
* On and OFF states of the GaAs MESFETs
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
TABLE 6.2
SUMMARY OF RESULTS FOR SUPERCONDUCTOR BASED HYBRID PHASE SHIFTERS
PARAMETER
@ON STATE*
@ OFF STATE*
Insertion Loss (S21) dB
-1.76
-2.9389
Input Reflection Loss
-5.62
-5.44
(S11} dB
Output Reflection Loss
-23.098
(S22) dB
* On and OFF states of the GaAs MESFETs
-16.35
7. SUMMARY AND CONCLUSIONS
The first major objective in this research program was to develop a
fabrication process for high quality (high T c and high Jc) TICaBaCuO thin films
mainly on LaAIOs substrates.
TICaBaCuO superconducting thin films were
fabricated on LaAI03 substrates by rf magnetron sputter deposition in a pure
argon plasma, followed by post-annealing techniques.
A reproducible
fabrication process has been established for TICaBaCuO thin films on LaAI03
substrates, for high T c and high Jc characteristics. The TICaBaCuO thin films
were patterned into four-probe test devices using standard microelectronic
lithography and wet etching techniques.
Low resistance gold contacts on
TICaBaCuO thin films were obtained by annealing at 600 °C in an oxygen flow
of 1 liter/minute followed by a slow furnace cooling for about 3 0 minutes. The
critical current density measurements were performed using dc and pulsed
current techniques under the electric field criterion of 1 //V/m m .
The zero
resistance T c between 97 and 100 K were routinely obtained for the patterned
TICaBaCuO thin films. Zero field current density (Jc) as high as 5 * 1 0 5 A/cm2
were obtained in four-probe test devices.
The specific contact resistivity
measured when the sample is superconducting, ranges from 3 .6 5 *1 O'5 Ohmcm2 at 9 0 K, to 10 '8 Ohm-cm2 below 77 K.
The microwave properties of TICaBaCuO thin films were investigated by
designing, fabricating and characterizing a microstrip ring resonator.
120
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The
resonator was designed for a fundamental resonance frequency of 12 GHz, and
for fabrication on 10 mil thick LaAI03 substrates. Ring resonators with gold
ground plane of 1 pm thickness and TICaBaCuO superconducting ground plane
of 0 .3
pm
thickness were fabricated and characterized at cryogenic
temperatures. The unloaded Q for the superconducting resonators were above
1500 at 65 K, compared to 370 for a gold resonator. The surface resistance
of the TICaBaCuO thin films obtained by separating conductor losses from the
Q measurements is typically between 1.5 and 2.75 m-Ohms at 12 GHz and 77
K, almost an order lower than Cu and Au at the same temperature and
frequency. The penetration depth at 0 K, was calculated from the resonance
frequency shift with temperature measurements. The typical values for the
penetration depth at 0 K is approximately between 7000 and 8000 A.
The
conductor
losses
in the
superconducting
microstrips
with
superconducting ground plane were compared to the ones with gold ground
plane using a theoretical model called the Phenomenological loss Equivalence
Method (PEM). This model predicted lower conductor losses for the microstrip
with superconducting ground plane, below 77 K.
Theoretical temperature
variation of surface resistance (Rs) for different penetration depths was
obtained for the all-superconducting microstrip (with superconducting ground
plane).
The Rs obtained from the Q measurements in the ring resonators
deviates from the theoretical temperature dependence.
This is possibly
because of additional losses introduced in the devices due to interaction
121
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
between the TICaBaCuO thin films and the LaAI03 substrates. Nevertheless,
the poly-crystalline TICaBaCuO thin films have almost an order lower surface
resistance compared to gold at 80 K. The design of the ring resonator was not
optimized for the highest Q, but the results of our investigations show that
TICaBaCuO ring resonator devices fabricated with superconducting ground
plane do show higher Q compared to a gold resonator below 9 0 K, proving their
usefulness for all-superconducting microwave circuit applications.
A high Tc superconductor and compound semiconductor based hybrid
digital phase shifter was investigated in this research.
TI-Ca-Ba-Cu-0
superconducting thin films and GaAs MESFETs were used in the circuit. The
phase shifter circuit was desiged for 180 degrees phase shift using the
reflection type configuration. The superconducting circuit consists of input and
output feed lines, 3 dB Lange Coupler which divides the input power by half
into the direct and coupled arms, and impedance transforming networks for
matching the impedance of the switching devices. GaAs MESFETs were used
for switching application in the hybrid circuit.
The phase shifters were
designed for 4 GHz center frequency, and operation at 77 K. The 180 degrees
phase bit was designed for a minimum insertion loss in the on and off states
of the switching devices, a minimum bandwidth of 25% , the phase error less
than 10 degrees, and input and output return losses greater than 15 dB.
Experimental results showed the insertion loss in the on-state of the devices as
small as 1.76 dB at 4 GHz and 70 K for the superconducting phase shifter
122
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
circuit compared to the lowest insertion loss of 3.1 dB obtained in a gold based
hybrid circuit. The improvement in insertion loss is mainly due to the lower
surface resistance and lower conductor losses in the superconductor based
microstrip circuits. The center frequency for the phase shift of 180 degrees
varied from 3.9 6 GHz below 40 K to approximately 4.11 GHz at 70 K. The
bandwidth of the phase shift for 10 degrees phase error is only about 20% of
the design mainly due to poor coupling over the bandwidth in the Lange
coupler. The Lange coupler finger has to be bonded by short gold wires to
improve the coupling.
The progress made in the recent years in TICaBaCuO superconducting
thin films improves the possibility for superconducting electronics at 77 K,
especially for superconductor-semiconductor hybrid communication systems
and
microelectronic interconnections,
bringing
us one step
closer to
commercialization of high temperature superconducting electronics in the near
future.
123
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
BIBLIOGRAPHY
1.
J.G.Bednorz and K.A.Muller, 'Possible high Tc superconductivity in the
Ba-La-Cu-0 system', Z.Phys.B 64, 1986, pp 189-193.
2.
M.K.W u, J.R.Ashburn,C.J.Torng,P.H.Hor, R.L.Meng, L.Gao,Z.J.Huang,
Y.Q.W ang, and C.W.Chu,'Superconductivity at 93 K in a new mixed
phase
Y-Ba-Cu-0
compound
system
at
ambient
pressure',
Phys.Rev.Lett., vol.58, no.9, March 2, 1987, pp 908-910.
3.
H.Maeda, Y.Tanaka, M.Fukitomi, and T.Asano,'A new high Tc oxide
superconductor without a rare earth element', Jap. J.Appl.Phys., vol.27,
no.2, Feb.1988, pp L209-210.
4.
Z.Z.Sheng and A.M.Hermann,'Bulk superconductivity at 120 K in the TlCa-Ba-Cu-0 system', Mature, vol.332, no.6160, March 1988, pp 138139.
5.
S.S.P.Parkin, V.Y.Lee, E.M.Engler, A.i.Nazzal, T.C.Huang, G.Gorman,
R.Savoy,
and
R.Beyers,'Bulk
superconductivity
at
125
K
in
Tl2 Ca2 Ba2 Cu3 0 x', Phys.Rev.Lett., vol.60, no.24, 13 June 1988, pp
2539-42.
6.
V. O. Heinen, K. B. Bhasin, and K. J. Long,"Emerging applications of
high temperature superconductors for space communications", NASA
Technical Memorandum 103629, 1990.
7.
Special issue on superconductivity, Cornell Quarterly, vol. 24, no.2,
124
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Winter 1990.
8.
W.G.Lyons, and R.S.Withers/Passive microwave device applications of
high Tc superconducting thin films'. Microwave Journal, Nov. 1990, pp
85-102.
9.
R.S. Withers, and R.W.Ralston/Superconductiveanalogsignal processing
devices', Proc.lEEE, vol.77, no.8, Aug.1989, pp 1247-62.
10.
W .Y.Lee, V.Y.Lee, J.Salem, T.C.Huang, R.Savoy, D.C.Bullock, and
S.S.P.Parkin/SuperconductingTICaBaCuO thin films with zero resistance
at temperatures of upto 120 K', Appl.Phys.Iett., vol.53, no.4, 25 July
1988, pp 329-331.
W. Y. Lee, S. M. Garrison, M. Kawasaki, E. L. Venturini, B. T. Ahn, R.
Boyers, J. Salem, R. Savoy, and J. Vasquez, 'Low temperature formation
of epitaxial TI2 Ca2 Ba2 Cu3 0 10 thin films in reduced 0 2 pressure', Appl.
Phys. Lett., vol.60, pp 772-774, 1992.
11.
D.S.Ginley, J.F.Kwak, R.P.Hellmer, R.J.Baughman, E.L.Venturini, and
B.Morosin,'Sequential electron beam evaporated films of TI2CaBa2Cu2Oy
with zero resistance at 97
K', Appl.Phys.Lett., vol.53,
no.5,
1
Aug.1988, pp 406-408.
12.
M.Hong, S.H.Liou, D.D.Bacon, G.S.Grader, J.Kwo, A.R.Kortan, and
B.A.Davidson, 'Superconducting TI-Ca-Ba-Cu-0 films by sputtering',
Appl.Phys.Lett., vol.53, no.21, 21 Nov. 1988, pp 2102-04.
13.
H.L.Chang, C.Wang, M.L.Chu, T.M .Ven, and Y.S.Gou,'Preparation of
125
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
T l 2 Ca2 Ba2 Cu 3 0 x superconducting
thin
films
by
dc
sputtering',
Jap.J.Appl.Phys., vol.28, no.4, April 1989, pp L631-634.
14.
S.I.Shah, N.Herron, C.R.Fincher, and W.L.Holstein, 'Fabrication of Tl-CaBa-Cu-0 films by annealing rf sputtered Ba-Ca-Cu-0 films in thallium
oxide vapors', Appl.Phys.Lett.,vol.56, no.8, 19 Feb.1990, pp 782-784.
15.
S.H.Liou, K.D.Aylesworth, N.J.Ianno, B.Johs, D.Thompson, D.Meyer,
J.A.Woollam, and C.Barry,' Highly oriented TI2Ca2Ba2Cu3O10 thin films
by
pulsed
laser evaporation',
Appl.Phys.Lett.,
vol.54,
no.8,
20
Feb.1989, pp 760-762.
16.
K.Zhang, E.P.Boyd, B.S.Kwak, A.C.Wright, and A.Erbil,' Metalorganic
chemical vapor deposition of TI2CaBa2Cu2Oy superconducting thin films
on sapphire', Appl.Phys.Lett., vol.55, no.12, 18 Sep.1989, pp 1258-60.
17.
R.W.Simon, ' High Tc superconducting thin films', Superconductor
Industry, Spring 1990, pp 22.
18.
D.S.Ginley, J.F.Kwak, E.L.Venturini, B.Morosin, and R.J.Baughman, '
Morphology control and high critical currents in superconducting thin
films in theTI-Ca-Ba-Cu-0 system', Physica C, vol. 160, 1989, pp 42-48.
19.
D. W. Face, and J. P. Nestlerode, "In-situ growth and properties of
epitaxial TIBa2CaCu20 7 and TIBa2(Ca1.xYx)Cu20 7 thin films", presented
at the 1992 Applied Superconductivity Conference, Chicago, August
1992.
20.
M.Greenblatt, S.Li, E.H.McMills, and K.V.Ramanujachry,' Chemistry and
126
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
superconductivity of thallium based cuprates', Studies of the high Tc
superconductors, Nova Scientific publishers, NY, vol.8, 1992.
21.
T.Suzuki, M.Nagshi, Y.Fukuda, Y.Syono, M.Kikuchi, N.Kobayashi, and
M.Tachiki,'Chemical state ofTI in the superconductorTl2 Ca2 Ba2 Cu3 0 10
studied by XPS', Phy.Rev.B, vol.40, no.7, 1 Sep.1989, pp 5184-86.
22.
R.B.Beyers, S.S.P.Parkin, V.Y.Lee, A.I.Nazzal, R.J.Savoy, G.L.Gorman,
T.C.Huang, and S.J.LaPlaca,' TI-Ca-Ba-Cu-0 superconducting oxides',
IBM J.Res.Develop., vol.33, no.3, May 1989, pp 228-237.
23.
B.Morosin, R.J.Baughman, D.S.Ginley, J.E.Schirber, and E.L.Venturini,
'Structure studies on TI-Ca-Ba-Cu-0 high Tc superconductors: effects of
cation disorder and oxygen vacancies', Physica C, vol.161, 1990, pp
115-124.
24.
R. Sugise, M. Hirabayashi, N. Terada, M. Jo, M. Tokumoto, T.
Shimomura, and H. Ihara, 'Reversible reactions in TI-Ca-Ba-Cu-0
system', Japanese J. App. Phys., vol. 27, no. 12, December 1988, pp
L2310-13.
25.
C. T. Cheung and E. Ruckenstein, 'Formation route of higher member
phases from lower ones in the TI-Ca-Ba-Cu-0 system', J. Mater. Res.,
vol. 5, no.9, Sept. 1990, pp 1860-70.
26.
M. Kikuchi, T. Kajitani, T. Suzuki, S. Nakajima, K. Hiraga, N. Kobayashi,
H.
Iwasaki, Y. Syono, and Y.
Muto,
'Preparation and chemical
composition of superconducting oxide TI2 Ba2 Can. 1Cun0 2 n + 4 with
127
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
n = 1,2, and 3 ', Japanese J. Appl. Phys., vol.28, no.3, March 1989, pp
L382- L385.
27.
B. T. Ahn, W . Y. Lee, and R. Beyers, 'Effect of oxygen pressure on
TI2 Ca2 Ba2 Cu3 0 x formation', Appl. Phys. Lett., vol. 60 (17), 27 April
1992, pp 2150-2152.
28.
M. Itoh, R. Liang, K. Urabe, and T. Nakamura, 'Reproducible processing
for high T c superconducting TI-Ba-Ca-Cu-0 phases', Japanese J. Appl.
Phys., vol. 27, no. 9, Sept. 1988, pp L1672-75.
29.
K.Hiraga,
D.Shinde,
M.Hirabayashi,
M.Kikuchi,
N.Kobayashi,
and
Y.Syono, 'Crystal structures of TI-Ca-Ba-Cu-0 superconducting phases
by high resolution electron microscopy', Jap.J.Appl.Phys., vol.27,
no.10, O ct.1988, PP L1848-51.
30.
M.DiStasio, K.A.Muller, and L.Pietronero,' Non-homogeneous charge
distribution in layered high T c superconductors', vol64, Phys.Rev.Lett.,
4 June 1990, pp 2827-30.
31.
G.Subramanyam, F.Radpour, V.J.Kapoor, and G.H.Lemon, 'Fabrication
and chemical composition of rf magnetron sputtered TICaBaCuO high T c
thin films', J.Appl.Phys., vol.68, 1 Aug.1990, pp 1157-63.
32.
I.Shih, and C.X.Qiu,'Creation of TICaBaCuO fine features by a w et
process', J.Appl.Phys., vo.65, no.12, 15 June 1989, pp 5215-5217.
33.
R.P.Vasquesz, B.D.Hunt, and M.C.Foote,'Non-aqueous chemical etch for
YBaCuO', Appl.Phys.Lett., vol.53, no.26, 26 Dec. 1988, pp 2692-94.
128
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
34.
A.lnam,
X.D.W u,
T.Venkatesan,
S.B.Ogale,
C.C.Chang,
and
D.Dijkkamp/Pulsed laser etching of high Tc superconducting thin films',
Appl.Phys.lett., vol.51, no.14, 5 O ct.1987, pp 1112-14.
35.
G.Subramanyam, V.J.Kapoor, C. M. Chorey, and K. B. Bhasin,'Electrical
transport properties and microwave device performance of sputtered
TICaBaCuO superconducting thin films', J. Appl.Phys., vol.72, no.6, pp
2 3 9 6 -2 4 0 3 , 1992.
36.
C. P. Bean,'Magnetization of high field superconductors', Rev. Mod.
Phys. vol. 36, pp 3 1 -39 , 1964.
37.
J.F.Kwak,
D.S.Ginley,
J.C.Barbour,
and
E.L.Venturini,
M.O.Eatough,
B.Morosin,
Studies
of
R.J.Baughman,
high
temperature
superconductors edited by A.Narlikar, vol.7, pp 4 5 -73 , Nova Science
Publishers, New York, 1991.
38.
D.Shi, and M.S.Boley,'Superconducting transition broadening and flux
creep activation energy in TI2 Ca2 Ba2 Cu3 0 x', Supercond.Sci.Technology,
vol.3, 1990, pp 289-292.
39.
R.B.Hammond, G.V.Negrete, L.C.Bourne, D.D.Strother, A.H.Cardona,
and M.M.Edddy,'Epitaxial Ti2CaBa2Cu20s thin films with low 9 .6 GHz
surface resistance at high power and above 77 K', Appl.Phys.lett.,
vol.57, no.8, 20 A ug.1990, pp 825-827.
40.
J.W.Ekin, A.J.Panson, and B.A.Blankenship,' Method for making low
resistivity contacts to highTc superconductors', Appl.Phys.lett.,vol.52,
129
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
no.4, 25 Jan.1988, pp 331-333.
41.
K.Mizushima, M.Sagoi, T.Miura, and J.Yoshida,' Electric properties of
the YBaCuO/Au interface', Appl.Phys.lett., vol.52, no.13, 28 March
1988, pp 1101-1103.
42.
J.W.Ekin, T.M.Larson, N.F.Bergren, A.J.Nelson, A.B.Swartzlander,
L.L.Kazmerski,
A.J.Panson,
and
B.A.Blankenship,'
High
Tc
superconductor/noble metal contacts with surface resistivities in the
1 0 '10 ohm cm2 range', Appl.Phys.Lett., vol.52, no.21, 23 May 1988,
pp 1819-21.
43.
R.W.Simon, C.E.Platt, A.E.Lee, G.S.Iee, K.P.Daly, M.S.Wire, J.A.Luine,
and M.Urbanik,'Low loss substrate for epitaxial growth of high
temperature superconducting thin films', Appl.Phys.Lett., vol.53, no.26,
26 Dec. 1988, pp 2677-79.
44.
G.A.Samara,
'Low temperature dielectric properties of candidate
substrates for high temperature superconductors, LaAI03 and Z r0 2:9.5
mo!% Y20 3', J.Appl.Phys., vol.68, no.8, 15 Oct. 1990, pp 42 1 4 -1 9 .
45.
G.Subramanyam, F.Radpour,and V.J.Kapoor,'FabricationofTICaBaCuO
superconducting thin films on LaAI03 substrates', Appl.Phys.Lett.,
vol.56, 30 April 1990, pp 799-802.
46.
F. London, and H. London, 'The electromagnetic equations of the
superconductor', Proc. of Royal Soc. A, vol.A 149, pp 71-88, 1935.
47.
L.D.Chang, M.J.Moskowitz, R.B.Hammond, M.M.Eddy, W.L.Olson,
130
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
D.D.Casavant,
E.J.Smith,
M.Robinson, L.Drabeck,
and
G.Gruner,
'Microwave surface resistance in Tl based superconducting thin films',
Appl.Phys.Lett., vol 55, no.15, 25 Sept., 1989, pp 1357-59.
48.
Technical Bulletin-'Surface resistance of thin films'-Superconductor
Technologies Inc.,Santa Barbara, CA, 1990.
49.
S.Sridhar, and W.L.Kennedy,'A novel technique to measure microwave
response of high Tc superconductors between 4 .2 K and 200 K', Rev.of
Sci. Instruments, 1988.
50.
J.P.Karini, A.M.Awasthi, W.Beyerman, G.Gruner, T.Hylton, K.Char,
M.R.Beasley,
and
A.Kapitulnik/Millimeterwave
surface
resistance
measurements in highly oriented YBCO thin films', Phys.Rev.B, vol.37,
1988, pp 9726-29.
51.
F.A. Miranda, W. L. Gordon, K. B. Bhasin, V. 0 . Heinen, J. D. Warner,
and G. J. Valeo,'Millimeterwave transmission studies of YBaCuO thin
films in the 26.5 to 40 GHz frequency range', Proc. of Supercond. and
its Applns conf., Buffalo, NY, 1989.
52.
J.
Bardeen,
L.
N.
Cooper,
and
J.
R.
Schrieffer,'Theory
of
superconductivity', Phys. Rev., vol.108, no.5, pp 1175-1204, 1957.
53.
J.H.Takemoto, F.K.Oshita, H.R.Fetterman, P.Kobrin, and E.Sovero,'
Microstrip ring resonator technique for measuring microwave attenuation
in high Tc superconducting thin films', IEEE MTT, vol.37, no.10,
Oct. 1989, pp 1650-52.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
54.
L.C.Bourne, R.B.Hammond, McD.Robinson, M.M.Eddy, W.L.OIson, and
T.W .J am es/A low loss microstrip delay line in 7120383200208',
Appl.Phys.Lett., vol.56, pp 2333-35, 1990.
55.
R.B.Hammond, G.V.Negrete, M.S.Schmidt, M.J.Moskowitz, M.M.Eddy,
D.D.Strother, and D.L.Skoglund,' Superconducting TI-Ca-Ba-Cu-0 thin
film microstrip resonator and its power handling performance at 77 K',
IEEE MTT symposium, June 1990.
56.
C.M.Chorey, K.B.Bhasin, J.D.Warner, J.Y.Josefowicz, D.B.Rensch, and
C.W .Nieh,' An experimental study of high Tc superconducting microstrip
transmission lines at 35 GHz and the effect of film morphology', 1990
Appl.Superconductivity conf., IEEE trans. on Magnetics., March 1991.
C.M.Chorey, K.S.Kong, K.B.Bhasin, J.D.Warner, and T.ltoh, 'YBCO
superconducting ring resonators at millimeterwave frequencies', IEEE
Trans. Microwave Theory Tech. vol M T T 39, pp 1480, 1991.
Robert R. Romanofsky,'Analytical and experimental procedures for
determining
propagation characteristics of millimeterwave gallium
arsenide microstrip lines', NASA Technical paper 2899.
57.
R.H.Koch, W.J.Gallagher, B.Bumble, and W .Y.Lee,' Low noise thin film
TICaBaCuO dc SQUIDs operated at 77 K', Appl.Phys.Lett., vol.54,
no.10, 6 March 1989, pp 951-953.
58.
R.H.Koch,'SQUIDs made from high temperature superconductors', Solid
state technology, May 1990, pp 255-260.
132
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
59.
B.Oh, R.H.Koch, W.J.Gallagher, V.Foglietti, G.Koren, A.Gupta, and
W .Y.Lee, 'YBaCuO input coils with low Tc and high Tc SQUIDs',
Appl.Phys.Lett., vol.56, no.25, 18 June 1990, pp 2575-77.
60.
'Superconducting transistors show their speed'. Scientific American,
Feb.1991, pp 28.
61.
J. S. Martens, V. M. Hietala, T. E. Zipperian, D. S. Ginley, C. P. Tigges
and
J.
M.
Phillips,'S-parameter
measurements
and
applications of superconducting flux flow transistors',
microwave
IEEE Trans, on
MTT., vol. 39, pp 2018-2027, December 1991.
62.
J. S. Martens, V. M. Hietala, T. A. Plut, D. S. Ginley, G. A. Vawter, C.
P. Tigges and M. P. SiegaI,'Flux flow microelectronics', Presented at the
Applied Superconductivity conference, Chicago, Illinois, August 1992.
63.
J. M. Martens, V. M. Hietala, T. E. Zipperian, D.S. Ginley, C. P. Tigges,
and G. K. G. Hohenwater,'A reflective microwave switch made of
TICaBaCuO for signal control applications', IEEE Microwave and Guided
Wave Lett., vol. 1, no. 10, pp 291-293, 1991.
64.
A. B. Pippard,'The surface impedance of superconductor and normal
metals at high frequencies, III. The relationship between impedance and
superconducting penetration depth', Proc. of Royal Soc. London., vol.
A 191, pp 399-415, 1947.
65.
H.- Y Lee and T. Itoh, 'Phenomenological loss equivalence method for
planar quasi-TEM transmission lines with a thin normal conductor or
133
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
superconductor', IEEE Trans. Microwave Theory Tech., vol. MTT 37, pp
1904-190 9, 1989.
66.
0.
R.
Baiocchi, K. -S.
Kong, H. Ling, and T. Itoh,
'Effect of
superconducting losses in pulse propagation on microstrip lines', IEEE
Microwave and Guided Wave Lett, vol 7, pp 2-4, 1991.
67.
D. Antsos, W . Chew, A. L. Riley, B. D. Hunt, M. C. Foote, L. J. Bajuk,
D. L. Rascoe, and T. W . Cooley, 'Modeling of planar quasi-TEM
superconducting transmission lines', IEEE trans. Microwave Theory
Tech., vol. MTT-40, 1128-1132, 1992.
68.
R.A.Pucel, D.J.Masse, and C.P.Hartwig, 'Losses in microstrip', IEEE
Trans. Microwave Theory Tech., 16, 342 (1968). Also 'Correction to
losses in microstrip', IEEE Microwave Theory Tech., 16, 1064 (1968).
69.
T.C.Edwards, 'Foundations for Microstrip circuit design', J.Wiley, 1981,
pp 209-251.
70.
R. K. Hoffman/Handbook of Microwave integrated Circuits', Artech
House, London, 1987.
71.
E. J. Denlinger,'Losses of microstrip lines', IEEE Microwave Theory and
Tech., vol. MTT 28, pp 513, 1990.
72.
H.Kroger,
C.Hilbert,
D.A.Gibson,
U.Ghoshal,
and
L.N.Smith,
'Superconductor-Semiconductor hybrid devices, circuits and systems',
Proc.of IEEE, vol.77, no.8, Aug.1989, pp 1287-1301.
73.
T.Van Duzer, and S.Kumar, 'Semiconductor-Superconductor hybrid
134
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
electronics', Cryogenics, vol.30, 1990, pp 1014-1023.
74.
K. B. Bhasin, S. S. Toncich, C. M. Chorey, R. R. Bonetti, and A. E.
Williams/Performance of a YBaCuO superconducting filter/GaAs Low
noise amplifier hybrid circuit', NASA Technical memorandum 105546,
1992.
75.
A. P. S. Khanna, Mike Schm idt/Low phase noise superconducting
oscillators', IEEE MTT_Symposium digest, pp 1239-1242, 1991.
76.
Guo-chun Liang, Xiaohong Dai, David F. Hebert, Theodore Van Duzer,
Nathan Newman, Brady F. Cole, 'High-temperature superconductor
resonators and phase shifters', IEEE Trans Applied Superconductivity,
vol.1, no.1, pp 58-66, March 1991.
77.
R. C. Hansen and H. Jasik, Eds., Antenna Handbook, McGraw Hill, New
York, 1984.
78.
R. V. Garver, Semiconductor Diode Control devices, Artech House,
1976.
79.
D. A. Durand, J. Carpenter, E. Ladizinsky, L. Lee, C. M. Jackson, A.
Silver, and A. D. Smith, ' The distributed Josephson Inductance Phase
shifter', IEEE Trans on Applied Superconductivity, vol.2, pp 33-38,
March 1992.
80.
J. H. Takemoto-Kobayashi, C. M. Jackson, C. L. Pettiette-Hall, and J. F.
Burch, 'High T c superconducting monolithic phase shifter', IEEE Trans on
Appl Superconductivity, vol. 2, pp 39-44, March 1992.
135
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
81.
G. Subramanyam, and V. J. Kapoor,'TICaBaCuO superconducting thin
films for hybrid circuit applications', presented at the ECS Spring
Meeting, Honoloulu, Hawaii, May 1993.
82.
G. Subramanyam, V. J. Kapoor, and K. B. Bhasin, 'A Hybrid phase
shifter circuit based on TICaBaCuO superconducting thin films and GaAs
MESFETs', submitted to IEEE Trans, on MTT, 1993.
83.
G. Subramanyam, M. Zhang, P. Boolchand, V. J. Kapoor,' Processing of
single phase TI2Ca1Ba2Cu2Ox bulk superconductor', submitted to the
Material Science Letters, 1993.
84.
G. A. Samara, and A. A. Giardini, Phys. Rev. A vol. 140, pp 95 4 (1965).
85.
F. A. Miranda, W . L. Gordon, V. 0 . Heinen, and B. N. Ebihara,
'Measurements of complex permittivity of microwave substrates in the
20 to 300 K range from 26.5 to 4 0 GHz', NASA technical memorandum
102123, 1989.
86.
G.Subramanyam, V.J.Kapoor, and G.H.Lemon, 'Processing of high Tc
and high Jc TICaBaCuO thin films', Bulletin of Am. Phys. Soc. 36 (3),
519 (1991).
87.
G. Subramanyam, V. J. Kapoor, C. M. Chorey, and K. B. Bhasin,'Studies
on sputtered TI-Ca-Ba-Cu-0 high Tc superconducting thin films for
microelectronic applications', to be published in the book on "High
Temperature Superconductors", edited by J. J. Pouch et al., Trans Tech
Publications, Aedermansdorf, Switzerland, 1993.
136
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
88.
C.D.Wagner, W.M.Riggs, L.E.Davis, J.F.Moulder and G.E.Muilenberg
(editor),
Handbook
of
Photoelectron
Spectroscopy,
Perkin-Elmer
Corporation, Minnesota, 1979.
89.
H.W.Willard, L.L.Merritt,Jr., J.A.Dean, and F.A.Settle,Jr., Instrumental
Methods of Analysis, 6th Edn., Wadsworth publ.co., Belmont, CA,
1984.
90.
William Pelster,'Accurate, Automated Measurement of Superconductor
Material Resistance', Research & Development, pp 80-84, June 1988.
91.
Technical Booklet: Low Voltage Measurements, Applications Overview,
Model 182 Sensitive Digital Voltmeter, Keithley Instruments.
92.
L. F. Goodrich and S. L. Bray/High Tc superconductors and critical
current measurement', Cryogenics, vol.30, pp 667-677, 1990.
93.
J. F.Kwak,E.L.Venturini,R.J. Baughman, R.Morosin, and D.S.Ginley,'High
critical currents in polycrystalline TI-Ca-Ba-Cu-0 films', Cryogenics,
vol.29, March 1989, pp 291-295.
94.
G. Subramanyam, V. J. Kapoor, C. M. Chorey, and K. B. Bhasin,'High
Tc superconducting microstrip resonators for 12 GHz', Presented at the
World Congress on Superconductivity, Munich, Germany, August 1992.
95.
J.Aitken, 'Swept frequency microwave Q factor measurements', Proc.of
Inst.Elec.Eng. vof 123, pp 855, 1976.
96.
W.
L.
Holstein,
L.
A.
Parisi,
C.
Wilker,
and
R.
B.
Flippen,
'Tl 2 Ba2Ca1Cu2 0 8 films with very low microwave surface resistance up
137
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
to 95 K', Appl. Phys. Lett., vol.60, pp 2014-2016, 1992.
97.
R.L.Kautz, 'Pulse propagation in superconducting microstrip lines',
J.Appl.Phys. vol.49, 308 (1978).
98.
T.Van Duzer, and C.W.Turner/PrincipIesof Superconductive devices and
circuits', chapter 3, pp 92 -1 3 8 . Elsevier North Holland Inc., New York,
1981.
99.
V . M. Hietala, J. S. Martens, D. S. Ginley, T. E. Zipperian, C. P. Tigges,
M . S. Housel, and T. A. Plut, 'Evaluation of superconducting Tl-Ca-BaCu-0 thin film surface resistance using a microstrip ring resonator', IEEE
Microwave and Guided Wave Lett., vol.1, pp 84-86, 1991.
100.
Private Communications with Dr. Felix A. Miranda of NASA Lewis
Research Center.
101.
Constantine Andricos, Indar J. Bah! and Edward L. Griffin, 'C-band 6 bit
GaAs monolithic phase shifters', IEEE Trans. Microwave Theory Tech.,
vol. 33, no. 12, pp 15 9 1 -1 5 9 6 , 1985.
102.
M . H. Kori and S. Mahapatra,'Integral analysis of hybrid coupled
semiconductor phase shifters', IEE Proceedings, vol. 134, Pt. H, no. 2,
pp 156-162, 1987.
103.
H. Chang, Ed., Handbook of Microwave and optical components, vol. 2,
chapter 5, pp 230-262.
104.
R.C.
Hansen/Superconducting antennas',
IEEE Trans.
Aerospace
Electron. Systems., vol.26, pp 34 5-3 55, Mar 1990.
138
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
105.
R. J. Dinger/Some potential antenna applications of high temperature
superconductors', J. of superconductivity, vol. 3, no. 3, pp 287-296,
1990.
106. Yalsin Ayasli, 'Microwave switching with GaAs FETs', Microwave
Journal, pp 61-74, November 1982.
107.
Ronald J. Guttman and David J. Fryklund, 'Characterization of linear and
non-linear GaAs MESFETs for broad-band control applications', IEEE
Trans. Microwave Theory and Tech., vol MTT 35, no. 5, pp 516-521,
1987.
108.
Anand Gopinath and Bruce Rankin,'GaAs MESFET RF switches', IEEE
Trans. Electron Dev., vol ED 32, no.7, pp 1272-1278, 1985.
109.
Nitin Jain and Ronald J. Guttman, 'Modeling and design of GaAs
MESFET control devices for control applications', IEEE Trans. Microwave
Theory Tech., vol. 38, no.2, pp 109-117, 1990.
110.
A. K. Sharma,'Solid-state control devices: State of the Art', Microwave
Journal, pp 95-112, 1989.
111.
G. A. Johnson, 'InGaAs Power Transistors', PhD Dissertation Thesis,
June 1992.
112.
M. Shokrani,'InGaAs Insulated Gate FETs for power and control
applications', PhD Dissertation Thesis, March 1993.
139
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
APPENDIX A
This section describes the procedure to fabricate calibration standards
used for the two port Thru-Reflect-Line (TRL) calibration. A 1 cm long thru line,
a 90 degrees and 45 degrees delay lines and 50 Ohm microstrip lines for input
and output feeds for the sample block were designed using TOUCHSTONE for
25 mil thick Alumina substrates. The layout was generated in MICAD and then
the pattern was defined on a rubylith.
A 2.5" square emulsion mask was
generated using the first stage reduction camera.
The microstrip standards were fabricated on 25 mil thick Alumina
substrates with
Cr/Au metallization on both sides.
metallization was thicker than the top microstrip.
The ground-plane
The patterns for the
standards were transferred on to the microstrip substrates using positive
photoresist (AZ 1421 )photolithography. The ground plane side was coated
with the photoresist to be protected from the etching solution. The top gold
metallization was first etched in a gold etchant solution prepared as follows:
8 gm of Kl and 2 gm of Iodine were dissolved in 80 ml of Dl H20 . The
solution is stirred constantly during the etching process. The approximate etch
rate at room temperature is 0.8 //m per minute.
The Cr etching solution is
prepared as follows:
A saturated eerie sulphate solution is prepared by heating up Dl H20 and
dissolving eerie sulphate in it until no more dissolves.
Mix 9 parts of eerie
140
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
sulphate solution with 1 part concentrated nitric acid.
After the gold metallization is removed, the microstrip substrate appears
dark in color, the color of Cr. The samples were immersed in the Cr etching
solution to remove the Cr metallization.
It is important to note that the Cr
etching solution is effective when we switch between the Cr and Au etching
solutions with Dl rinse in between. A clean white surface is obtained wherever
the Cr/Au metallization were removed.
After etching the standards on the
alumina substrates, the photoresist is stripped in acetone(30 seconds), followed
with immersion in methanol(30 seconds) and finally a Dl rinse for 2 minutes.
These standards were cut using a diamond saw so that each one can be
inserted separately in the test fixture.
exactly between the launchers.
The standards were designed to fit
The sample block is raised by a screw
assembly. The sample is raised slowly until there is good contact between the
microstrip and the launchers. For reproducibility the contacts were made at the
same marked spot.
Also, the same torque is applied when the sample is in
contact with the launchers.
The two port calibration for the test fixture was performed as follows:
1.
Insert the thru line and measure the magnitude and phase of the s
parameters.
2. Insert the 90 degrees delay line and measure the magnitude and the phase
of the s parameters.
3. Measure the s parameters for the 'open' condition at the launchers.
141
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
These three measurements are sufficient for the 12 term error model used in
extracting the DUT data[111]. The validity of the calibration is always checked
with a 'thru' as a DUT measurement. A reliable calibration had the following
features:
1. The magnitude and the phase of S21 and S 12 are almost equal (symmetrical
condition).
2. The magnitude of S n and S22 are below -40 dB.
3. The magnitude and phase of S21 measured for the thru line as a DUT is
almost close to 0 dB and a few degrees respectively.
142
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
APPENDIX B
Fabrication o f gold based phase shifter circuit:
The gold based 180 degree phase shifter circuit was first fabricated on
20 mil thick LaAI03 substrate.
The substrates were cleaned using the
following procedure.
1. Substrates degreased in boiling acetone for 15 minutes.
2. Immersed in methanol for 5 minutes
3. Dl water rinse for 2 minutes
4. Blow dry the substrates in nitrogen
5. Bake the substrates at 120 °C for about 10 minutes.
After the substrates were cleaned, a thin layer of Ti (300 A) was deposited on
the substrates. The Ti thin film layer acts as a buffer layer, since gold does not
adhere to the LaAI03 substrates. Ti layer evaporation is followed immediately
by gold evaporation to a thickness of approximately 2 0 00 A. The samples
were removed from the evaporation system and then gold was electroplated to
a thickness of approximately 3 pm using the procedure developed in our
laboratory[112].
After gold plating, the samples were patterned by photolithography using
the phase shifter mask.
AZ 1421 positive photoresist was used for the
patterning process. The samples were etched in the gold etchant solution, and
then the Ti layer was etched in a Dl H20 : 30% H 20 2 : NH4OH (5:1:1) base
143
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
solution heated to 60 °C.
The groundplane gold was evaporated with 300 A buffer layer of Ti. The
groundplane side was gold plated to a thickness of 3 //m . This completes the
circuit fabrication. The circuit was mounted on a gold plated carrier using silver
epoxy. Then the GaAs MESFETs were mounted on the chip carrier close to the
circuit. Gold wire was bonded to the device and the microstrip transmission
lines to complete the assembly of the circuit.
144
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Документ
Категория
Без категории
Просмотров
0
Размер файла
5 559 Кб
Теги
sdewsdweddes
1/--страниц
Пожаловаться на содержимое документа