close

Вход

Забыли?

вход по аккаунту

?

Development of near-field scanning microwave and optical dual probe: Application to characterization of high-T(c) superconductors

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 th is reproduction is dependent upon the quality of the
copy subm itted. 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 a t the upper left-hand comer and continuing
from left to right in equal sections with small overlaps.
ProQuest Information and Learning
300 North Zeeb Road, Ann Arbor, Ml 48106-1346 USA
800-521-0600
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Development of Near-field Scanning Microwave
and Optical Dual Probe: Application to
Characterization of High-Tc Superconductors
by
Roberto S. Aga, Jr.
Submitted to the Department o f Physics and Astronomy and the
Faculty o f the Graduate School of the University of Kansas in partial
fulfillment of the requirements for the degree o f Doctor o f Philosophy
Dissertation Committee
r
Chair
?L l uut
-
CiwCiJ
S ftc K '
S U V v v v i - t - t e r t *. I *2- | I " 2 - j 'V o O / 2-
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number 3082642
_____
________
(f t
UMI
UMI Microform 3082642
Copyright 2003 by ProQuest Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
ProQuest Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Abstract
Over the last decade and a half, many scanning probe microscopy (SPM)
techniques have been developed. With their ability to image spatial variations o f
many different physical properties at a resolution ranging from micrometers to sub­
nanometers, SPM has become a powerful tool in many scientific research fields
including Physics, Chemistry, Biology and Material Science. As the spatial resolution
o f the SPM constantly improves and approaches sub-nanometer scale, as driven by
the needs in nanosciences and biophysics, correlation of different physical properties
at the same scale becomes highly desirable to understand the fundamental physics
that governs the overall behaviors of a sample. For example, the overall performance
o f a high-Tc superconductor (HTS) microwave device can be dominated by
impurities, defects and grain boundaries of micrometer to nanometer dimensions.
Correlation of these features with the microwave response at different pixels at
microscopic scale is key to understanding and improving the performance o f a HTS
microwave device.
However, such a correlation is incredibly difficult at the
microscopic scale and presents a major challenge in advancement o f SPM.
In this dissertation, a novel dual-channel near-field scanning microwave and
optical microprobe (NSMM/NSOM) was developed for simultaneous mapping o f
microwave and optical properties of a sample at microscopic scales. This microprobe
is composed o f an open-end coaxial resonator with its center conductor being
replaced by a stainless steel tube terminated by a titanium/silver coated fiber optic
with a tapered tip. The optical fiber serves as the channel for NSOM, while its metal
coating is the channel for NSMM. Using this dual-channel NSMM/NSOM probe, a
spatial resolution o f ~5 pm, that is comparable to the best reported for single-channel
NSMM, has been achieved on metallic samples. This resolution is mainly limited by
ii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the sensitivity of the NSMM channel and may be further improved when the
sensitivity o f NSMM is enhanced.
Characterization o f the microwave properties of the highest-Tc Hg-based
superconductors has been carried out using a traditional resonant cavity technique, as
well as a novel single-channel NSMM and the dual-channel NSMM/NSOM. Using
the traditional technique, the microwave surface resistance (Rs) and power handling
capability (Pc) of HgBaiCaCujOe (Hg-1212 with Tc~ 125 K) films have been
measured for the first time, and the results are superior to the best achieved on other
superconductors. For example, a comparable Rs~0.3 m fi (10 GHz) can be obtained
on Hg-1212 at close to 120 K as opposed to the same Rs for YBaiCuaO? (the most
popular high-Tc superconductor with Tc~ 92 K) at around 77K. This can be attributed
to the large difference in the Tcs between the two materials and has demonstrated the
potential o f Hg-1212 for microwave applications. A comparison o f the microwave
properties o f Hg-1212, Tl-2212 and YBCO films at reduced temperature scale
suggested further room for improvement o f Hg-1212 performance. Using NSMM, the
localized microwave properties, such as Tcs, sheet resistance and power handling
capability have been investigated and nonuniformity, revealed. Attempts to correlate
the observed nonuniformity o f microwave properties with the microstructures of the
sample have been made using the dual-channel NSMM/NSOM probe and interesting
results, obtained. With further refinement and improvement, these scanning probes
can be powerful diagnostic tool for high-Tc superconducting microwave devices,
particularly because they are nondestructive. In addition, these probes have been
employed to study dynamics o f a system, such as degradation of Hg-1212 films in
humid environment, the effect o f photon-doping induced superconductivity (in the
latter case, the dual-channel SPM is at the “pump probe” mode).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Contents
Abstract
ii
List of Figures
vii
1. Introduction
1
1.1 Scanning probe microscopy (SPM)....................................................................1
1.1.1 Brief history of SPM................................................................................ 1
1.1.2 Basic SPM................................................................................................ 3
1.1.3 Examples of SPM techniques.................................................................4
1.1.4 Features of scanning probe microscopes................................................7
1.2 Scanning probe microscopy o f high-Tc superconductors (HTS).................... 9
1.2.1 General properties o f superconductors.................................................10
1.2.2 Characterization of HTS using SPM..................................................... 12
1.3 Motivation.........................................................................................................15
2. Microwave characterization of HTS film: traditional approach
16
2.1 Overview o f superconductivity at microwave frequencies.............................16
2.1.1 Two-fluid model..................................................................................17
2.1.2 Surface impedance o f superconductors..............................................21
2.2 Review of microwave characterization o f HTS films.................................... 23
2.2.1 Introduction to microwave resonator................................................... 23
2.2.2 Common measurement techniques.......................................................25
iv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.3 Nonlinear effects in HTS.................................................................................. 29
2.3.1 Rs power dependence............................................................................ 29
2.3.2 Harmonic generation.............................................................................30
2.3.3 Intermodulation..................................................................................... 31
3. Microwave characterization of HgBa2CaCu:06+s thin films
33
3.1 Microwave surface resistance (Rs) measurement............................................33
3.1.1 Sample preparation...............................................................................33
3.1.2 Sample properties.................................................................................. 34
3.1.3 Experimental technique........................................................................38
3.1.4 Results and discussion.......................................................................... 40
3.2 Microwave power handling measurement....................................................... 43
3.2.1 Device preparation................................................................................ 43
3.2.2 Experimental technique........................................................................44
3.2.3 Results and discussion.......................................................................... 45
3.3 Limitations of traditional techniques................................................................50
4. Near-field scanning microwave microprobe
SI
4.1 Principle of operation........................................................................................ 51
4.1.1 Probe designs........................................................................................ 52
4.1.2 Signal measurement.............................................................................. 54
4.2 Open-end coaxial resonant probe......................................................................55
4.3 Experimental setup for mapping.......................................................................60
4.4 Sensitivity and spatial resolution......................................................................62
5. Applications of near-field scanning microwave probe
65
5.1 Uniformity of large area HTS film...................................................................65
5.2 Local measurement o f superconducting properties.........................................67
5.2.1 Integration of the near-field probe with cryogenic enclosure.............68
5.2.2 Addition o f mutual inductance coils to NSMM probe.........................70
v
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5.2.3 Measurement of local Tc...................................................................... 71
5.2.4 Single-pixel measurement of power handling capability.................... 74
6. Development of a dual-channel near-fleld scanning
79
optical and microwave microprobe
6.1 Overview o f near-field scanning optical microscopy..................................... 80
6.2 Integration of fiber optic NSOM probe to NSMM probe............................... 81
6.3 Tapering of fiber optic tip.................................................................................84
6.4 Metal coating of tapered fiber optic................................................................. 85
6.5 Evaluation of the dual probe.............................................................................88
7. Applications of dual-channel near-field scanning optical
94
and microwave probe
7.1 Degradation study of HgBa2CaCu2 C>6+6 thin film.......................................... 94
7.2 Diagnostic tool for HTS microwave devices...................................................99
7.3 Local observation o f photo-induced effects: A potential
pump and probe application............................................................................102
7.3.1 Introduction......................................................................................... 102
7.3.2 Photo-induced superconductivity (PISC) in
HgBa2CaCu2 C>6+ 6 thin films.................................................................103
8. Conclusion and future direction
107
Bibliography
109
Publications
117
vi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
List of Figures
1.1 Historical progress of microscope spatial resolution[2]........................................ 2
1.2
Schematic diagram of a generalized SPM............................................................. 4
2.1
Equivalent circuit of complex conductivity..........................................................19
2.2 A typical plot of insertion loss (IL) o f a resonator versus
the normalized frequency.......................................................................................24
2.3 Three configurations of patterned resonator using HTS film.
(a) microstrip resonator (b) stripline resonator and
(c) coplanar resonator............................................................................................26
2.4 Configuration of dielectric resonator for measuring the surface
resistance (Rs) of HTS films................................................................................ 28
3.1
Zero-field Jcs as a function of temperature of Hg-1212 made in cationexchange process (open circle), Hg-1212 by conventional thermal
reaction process (solid diamond) and Tl-2212 precursor film
(solid circle), which have dimension o f 2 mm x 3mm. Inset: Plot of
magnetic susceptibility vs temperature................................................................ 35
3.2 Surface morphology of (a) Hg-1212 by cation-exchange,
(b) Hg-1212 by conventional process and (c) Tl-2212 precursor taken
by a scanning electron microscope..................................................................... 37
3.3 Diagram o f the experimental probe for measuring the surface
resistance (Rs) of HTS films using the cavity perturbation technique.
The resonator was excited in the TEoi i mode at 11.2 GHz................................. 38
vii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.4 Surface resistance (Rs) scaled at 10 GHz against temperature o f
the Hg-1212 samples from cation-exchange (open shapes) and
a Tl-precursor (solid circle). Inset: RS(T) o f Hg-1212 made
from conventional process..................................................................................... 40
3.5 Surface resistance (Rs) scaled at 10 GHz of two different HTS
films namely: (a)YBCO and (b) Tl-2212 as reported by I.S. Gergis
et al. [47] and Holstein et al. [48] respectively.................................................... 42
3.6 Experimental setup for power handling measurement of microstrip
lines. In this configuration, the reflection coefficient (T) and insertion
loss (IL) are determined simultaneously. The RF amplifier is only
inserted when power level of network analyzer is not adequate. Inset:
layout o f the Hg-1212 microstrip line on LaAlC>3 substrate as
prepared for this measurement...............................................................................44
3.7 Insertion loss (IL) of two Hg-1212 samples (Dl and D2) at 1 GHz is
plotted as function of temperature. It shows Tcs o f 121 K and 119 K for
samples Dl and D2 respectively............................................................................46
3.8 Critical input power o f two Hg-1212 samples (Dl and D2) at 1 GHz is
plotted as a function of temperature. Inset: typical behavior of Pout
measured at the end o f the pulse vs P®................................................................. 49
3.9 Comparison of Pc for Hg-1212, Tl-2212 and YBCO films when
plotted against reduced temperature. Inset: Pc of three superconductors
plotted against real temperature.............................................................................49
4.1 Different techniques o f implementing a metallic tip in scanning near­
field microwave probe, (a) Microstrip resonator terminated by a wire tip.
(b) Quarter-wavelength resonator shielded by a cavity.
(c) Resonant coaxial transmission line..................................................................53
4.2 Microwave probe design employing a slit to confine emitted fields
at sub-wavelength resolution................................................................................. 53
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.3 Design and implementation of the microwave probe based
on a quarter wavelength open-end coaxial resonator............................................56
4.4 Plot of the real (black) and imaginary (red) terms in the complex
input impedance (Zm) o f the open-end resonator as a function of
frequency as measured by a vector network analyzer..........................................58
4.5 (a) Electric field configuration at the tip o f the probe when a sample
is placed very close to i t (b) Interaction between tip and sample via
effective capacitance, Cs ........................................................................................ 59
4.6 Change in reflection property o f the resonator when a conducting
sample is placed under the probe tip with distance o f 0.5 mm.
This measurement was performed at room temperature and
the corresponding change in unloaded Q was from 8400
(no sample) to 1800 (with sample)........................................................................60
4.7 Configuration o f the three motorized stages to move the top most
stage in three dimensions....................................................................................... 61
4.8 Linescan o f a tapered microwave probe with tip diameter
around 15 pm on a patterned superconducting Tl-2212 line
whose width is 150 pm .......................................................................................... 63
4.9 (a) Photograph o f the sample with “KU” letters made of copper. The
letters were patterned on a printed circuit board (PCB).
(c) Microwave image o f the same letters as generated by
the near-field scanning probe................................................................................. 64
5.1 Image scanned by the near-field scanning microwave probe showing
the non-uniformity o f microwave property o f a large-area Tl-2212 HTS
film on LaAlCb substrate. Resolution is 0.5 mm.................................................. 66
5.2 Sketch o f the experimental cryogenic vacuum enclosure used for single­
pixel characterization o f microwave properties of HTS films..............................69
5.3 Sketch illustrating the addition of mutual inductance coil to the
NSMM probe and stage..........................................................................................70
ix
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5.4 Plot o f the pick up voltage (open circle) and microwave reflected
power expressed as Sn versus temperature. These data were taken
from a Hg-1212 HTS film...................................................................................... 73
5.5 Data obtained from local Tc measurement of three different
superconductors. Solid circles correspond to Hg-1212. The solid
squares and circles in the inset correspond to Tl-2212 and
YBCO respectively................................................................................................ 73
5.6 Behavior of the resonance curve as input power to the NSMM
probe is increased. The sample used in this experiment was Hg-1212 at
80 K. and tip-to-sample distance was 0.4 mm..................................................... 75
5.7 Summary o f the local power handling measurements performed
on three different superconductors namely: Hg-1212 (black circles),
Tl-2212 (blue squares) and YBCO (red circles). In this experiment,
the tip-to-sample distance was approximately 0.3 mm........................................77
6.1 Major configurations in near-field scanning optical microscopy
(NSOM). They are named as follows: (a) illumination; (b) collection;
(c) oblique illumination; and (d) oblique collection [62]....................................80
6.2 Design o f the dual-channel near-field scanning optical and microwave
probe. Inset: exploded view of the tip and the collection
mode configuration and cross-sectional drawing of the dual probe
tip showing the multi-concentric structures.......................................................... 83
6.3 Photograph o f the actual dual-channel near-field scanning optical
and microwave probe developed in this work...................................................... 83
6.4 (a) Diagram o f the designed Teflon container for etching fiber optic.
(b) Qualitative description of the tip diameter during etching. The critical time
for getting the sharpest tip is around 38 minutes.................................................. 84
6.5 Schematic diagram o f the setup for metal coating o f the optical.
fiber using thermal evaporation............................................................................. 86
x
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6.6 (a) Scanning electron microscope (SEM) picture of tapered tip coated
with single layer of gold, (b) Magnified SEM picture showing the
bottom part o f (a), (c) SEM picture of another tip coated with a
single layer of silver............................................................................................... 87
6.7 (a) Schematic diagram and (b) actual experimental scanning setup for
the newly-developed dual probe............................................................................89
6.8 Behavior of resonant frequency as conducting sample
approaches the tip.................................................................................................. 91
6.9 Dual line scan of the probe on a patterned conducting strip
with 20 pm line width............................................................................................ 91
7.1 Spatial variation of microwave reflection before (a) and after 6
minutes of exposure to humid environment (b). Lower images correspond
to the spatial variation of optical transmission before (c) and after 6
minutes of exposure (d). Scan area is 1.5 x 1.5 mm2.......................................... 96
7.2 AFM images showing the process o f surface deformation occurring
as sample degrades. Scan area is 5 x 5 pm2 and height scale is from
0 nm to 300 run...................................................................................................... 98
7.3 (a) Microwave image obtained by the dual probe on a segment of a
microwave resonator R l. (b) Corresponding optical image of the
same segment mapped by the dual probe........................................................... 100
7.4 (a) Microwave image obtained by the dual probe on a segment of a
microwave resonator R2. (b) Corresponding optical image of the
same segment mapped by the dual probe........................................................... 101
7.5 (a) Effect of 12-hr photodoping at 295 K to the Tc of Hg-1212
thin film, (b) Diagram of transport measurement setup for studying
photo-induced effects........................................................................................... 104
7.6 Observation of PISC using (a) transport and
(b)magnetization measurements..........................................................................105
xi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 1
Introduction
Scanning probe microscopy has become an emerging field in nano-sciences and
materials research. Scanning probe microscopy refers to a variety o f high-resolution
surface microscopy techniques that use a tip o f a probe maintained very close to or in
contact with a sample surface. The probe is scanned in the x-y plane over a sample
by utilizing a combination o f piezoelectric transducers and stepping motors. Various
types of scanning probes have been developed to create an image o f certain physical
properties o f a sample at up to a sub-nanometer resolution, providing powerful tools
for researches in many different disciplines including material sciences, biology,
medicine, etc. Nevertheless, functions o f the SPM in the current status are limited.
Sometimes there are two different physical properties such as microwave and optical
properties that need to be imaged simultaneously for correlation studies at
microscopic scale. In this chapter, an overview of the current SPM techniques, its
applications as well as limitations that motivate this thesis work will be presented.
1.1
Scanning probe microscopy (SPM)
1.1.1 Brief history of SPM
In 1981, when Binnig and Rohrer invented scanning tunneling microscope
(STM), they had superconductors on their mind [I]. The main motivation was not
I
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
really to image surface morphology because at that time electron and field ion
microscopy were already mature technologies achieving spatial resolution better than
one nanometer (or 10'9 m) as shown in Fig. 1.1. However, with the newly invented
STM the electronic states of conducting samples could be probed aside from the
surface morphology. In addition, the STM (also all the SPM measurements)
measurement can be performed in air as opposed to the required vacuum environment
o f the electron and ion microscopes. The most important impact o f the invention of
STM was the development of other SPM probes with different capabilities.
For
example, due to the limitation o f STM to conducting samples, the atomic force
microscope (AFM) was developed in 1986. This instrument could measure surface
morphology of practically any samples, conductive or nonconductive. It was not a
complete substitute to STM because the STM measures electronic state in addition to
the morphology. During the past decade,
•
to-7 j .
0 Optical microscopy
— to-®
D
TGSA
O
4*
* io-’ _L ■
♦
Electron microscopy
o
□
Field ion microscopy
Scanning probe microscopy
o
Near-fieid optical microscopy
o□ □
t o 10-!.
1850
♦
♦
♦
♦
1950
1900
2000
YEAR
Fig. 1.1: Historical progress of microscope spatial resolution [2].
2
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
many different probes with unique imaging capabilities have been developed at an
impressively fast rate. Many o f them are now commercially available and serve as the
major tools in nano-sciences related research. SPM now represents a large number of
techniques for surface characterization using scanning probes and it has been applied
for characterization of a wide range o f materials including metals, semiconductors,
superconductors, dielectric, magnetic, ferroelectric, biological, etc. The focus o f this
thesis is on the high-temperature superconductors (HTS) that was discovered about
16 years ago and has attracted much attention from many groups in the world due to
their potential applications.
1.1.2 Basic SPM
In SPM, the primary objective is to be able to measure the desired physical
property o f a sample at a local spot (or pixel) with adequate sensitivity. Repeating this
measurement on many other pixels will then result in a two-dimensional mapping of
certain physical properties at resolutions defined by the pixel dimension. In many
existing SPM, the single-pixel measurement is performed with a microscopic sensor
placed at the proximity of the sample. Another way is to use a focused beam such as
electrons, ions and photons to locally excite the sample and measure their interaction
with separate sensors. The size of this pixel that can be characterized determines the
spatial resolution o f the microscope. Figure 1.2 is a schematic representation of
generalized SPM. The microscopic sensor is depicted with a probe tip. Scanning this
sensor (or focused beam) over the sample and measuring the synchronous signal from
the sensor records the spatial variation o f the physical property on each pixel. Thus,
the scanned region can be represented by the measured physical property. The
microscopic sensor used for probing can be practically anything and the choice is
mainly determined by the kind of sample to be studied and what physical property is
to be imaged. This makes SPM a versatile and flexible technique.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Probe tip vertical
coarse positioning I
Meter for measuring
tip to sample distance
IProbe tip
Meter for measuring
(microscopic physicai property
sensor)
SAMPLE
Feedback
Fig. 1.2: Schematic diagram of a generalized SPM.
1.1.3 Examples of SPM techniques
STM is perhaps the most well known scanning microscopy technique and it is
regarded as the parent of other techniques. It is based on the quantum mechanical
phenomenon called single electron tunneling. In STM experiment, a sharp metallic tip
is brought close to the conducting surface o f a sample. When the two surfaces are
close enough (~ 1 nm), their wave functions overlap resulting to a finite probability
that electrons will cross the barrier between the surfaces when a bias voltage is
applied between the sample and the tip [3]. The bias voltage ranges from a few
millivolts to a few volts, depending on the conductivity of the sample. The surfaces
where this electron tunnels, takes place in atomic scale. The tunneling current (It),
which typically varies between 10 pA and 1 nA, decreases exponentially as the tip to
4
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
sample distance increases. This relationship can be used to image topography and
electronic states of the sample with an atomic resolution. By scanning the tip over the
sample at constant height, variation of It can be recorded and that corresponds to the
topography of the sample surface assuming that the effects o f electronic density to It
can be sorted out. If It is kept constant during scanning by creating a feedback to the
scanning stage, the signal required to alter the vertical tip position represents the
image of the spatial variation o f electron density.
AFM probes the surface o f a sample with a sharp tip, which is placed at the end
of a cantilever. This tip is few microns long and often less than 10 nm in diameter
near the end of the tip. Forces between the tip and the sample surface cause the
cantilever to bend or deflect. As the tip is scanned over the sample, cantilever
deflections can be measured using optical techniques and recorded into a computer to
generate a map of surface topography. There are several forces that cause deflection
in the AFM cantilever. The first one is the inter-atomic van der Waals force. As the
atoms are gradually brought together, they exhibit attraction, which increases until
their separation is so small that their electron clouds begin to repel each other electro­
statically. Further decrease in the inter-atomic separation, causes the attractive force
to weaken. The force goes to zero when the separation is a couple o f angstroms
(about a length o f a chemical bond). When the atoms are in contact, the van der
Waals force becomes positive or repulsive. For operations in ambient environment,
another force arises when tip is in contact with the sample. It is called capillary force
and it is due to a thin layer o f water between the tip and the sample. This water wicks
its way around the tip creating an attractive force that holds it in contact with the
sample surface. The cantilever itself also exerts force on the sample. This force is like
a compressed spring and it depends on the spring constant of the cantilever.
MFM stands for magnetic force microscopy. This microscope uses a sharp
magnetized tip attached at the end o f the cantilever, which is very similar to the AFM
tips. When a sample is brought near the magnetized tip, localized magnetic fields
5
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
interact with the sample surface just beneath the tip. The magnetic force between the
tip and the sample deflects the cantilever. The direction and magnitude of deflection
are mainly dependent on the magnetic property o f the sample. Without scanning, this
magnetized tip can be used to investigate locally the ferromagnetic and diamagnetic
behavior of a material. Ferromagnetic, diamagnetic, paramagnetic and anti­
ferromagnetic properties are directly observed by studying the cantilever deflections.
The spatial resolution at which the local measurement can be performed is determined
by the cross sectional area o f the magnetized tip.
The addition of a scanning
mechanism enables the tip to scan over the sample surface, thus providing a contrast
based on magnetic properties.
SSM, which stands for scanning SQUID (Superconducting Quantum Interference
Device) microscopy, is also a technique for imaging magnetic field. It utilizes the
ability of the SQUID to detect very small fields. It has the highest flux sensitivity but
it also has the lowest spatial resolution (~ 10 (am), which is limited by the input coil
for the SQUID.
SHPM, or scanning Hall probe microscopy is also a useful technique for studying
magnetic properties of a sample. It is based on a microscopic Hall bar microfabricated
near the corner of a gallium arsenide chip [4], Resolution is limited by the size of the
sensor and it is typically around 0.3 um. The advantage of SHPM is that it is the least
invasive because it has no magnetic or superconducting material. It can be calibrated
to measure the absolute flux or magnetic field perpendicular to the sensor with
sensitivity between that of MFM and SQUID microscopes.
NSMM stands for near-field scanning microwave microscopy. This technique
can probe locally the microwave properties such as conductivity and permitivitty at
microwave frequencies, o f conductors and dielectrics by utilizing the near-fields
emanating from a sharp tip or small aperture. The resolution is determined by the tip
or aperture diameter and not by the wavelength o f the signal used.
6
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
NSOM, or near-field scanning optical microscopy, is an extension of NSMM in
the visible regime. It also utilizes the near-fields to break the diffraction barrier and
obtain resolution independent of wavelength of the signal being used. In this
technique, light is allowed to come out or go into a very small aperture, which
determines the spatial resolution. The interaction o f a sample with light gives rise to
several observable phenomena such as fluorescence, energy absorption, light
scattering and polarization. As a result, NSOM has different mode of operations.
1.1.4 Features of scanning probe microscopes
The scanning probe microscope is a versatile tool for characterization of a wide
range of materials such as dielectrics, conductors, superconductors, polymers and
even biological specimens. One o f the reasons for its versatility is that the microscope
can be designed to exchange scanning probes for imaging different physical
properties. By just changing the proximity sensor or probing signal and modifying the
detecting electronics, an SPM system can have diverse imaging capability without
changing the main system components like the scanning stage, controller or scan
head. For example, there are commercial SPM systems, which can perform as STM,
AFM and MFM. For each technique, there is a corresponding tip that needs to be
mounted.
A scanning probe microscope is not limited to imaging. Some probes can
perform local spectroscopy. Conventional surface-sensitive spectroscopic techniques
such as X-ray photoelectron spectroscopy, Auger spectroscopy and vibrational
spectroscopy are global in nature [5]. The data they provide represent an average over
the entire surface. These techniques are not appropriate for studying surface
phenomena, which are strictly local in nature such as those associated with impurities,
steps and defects. With SPM probes, a specific point on the sample can be selected
for spectroscopic measurements. The image itself generated from the scan can be
used for choosing which spot to study. In the case o f STM, the basis for spectroscopic
operation is the dependence o f the tunneling current on the bias voltage.
7
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Measurement o f I-V curve at a fixed location and a fixed tip to sample distance can
give the local electronic structure. Hamers et al. were able to map out the electronic
structure of the Si(l 1l)-(7x7) unit cell with a lateral resolution o f 0.3 nm [6].
Another SPM technique, which can perform spectroscopy, is the NSOM. The
basic idea in this technique is that light is allowed to pass through a sub-wavelength
aperture, which is usually a tapered fiber optic tip coated with a thin layer of metal.
When a sample is brought very close to the tip, it interacts with the tip via the near
field. This interaction is localized at a spot with its size approximately equal to the
tip. Thus, the resolution of the NSOM is determined by the size of the tip rather than
the wavelength o f the light used. By scanning this tip over the sample, light is applied
to the sample (or collected from the sample) in a spot-by-spot manner. Spatial
resolution o f 12 nm has already been reported [7]. The NSOM technique is capable of
detecting single molecules [8] while the probe is focused on a single molecule to
gather its spectral information. It is important to characterize molecular properties and
chemical changes on the scale of single molecule. Single molecule spectroscopy has
also been utilized at low temperature to study dynamics o f local environments in
crystalline and amorphous materials, local field effects, vibrational modes and
magnetic resonance o f a single molecular spin [9]. Low temperature spectroscopy
allows a more accurate probe of the natural line widths because thermal broadening is
minimized. Moreover, variations in the excitation spectra due to local and induced
effects can be probed more easily.
Another
significant
capability
of
scanning
probe
microscope
is
micromanipulation and lithography. It can be used to alter surface properties such as
conductivity, construct surface patterns and manipulate atoms. Manipulation of
materials at nanometer scale in a controlled fashion is the basis for some nano­
fabrication techniques. For example, Mamin et al. have demonstrated that a gold
STM tip can be used as a miniature solid-state emission source for directly depositing
nanometer-size gold structures at atmospheric pressures [10]. By applying short
voltage pulses (few hundred nanoseconds or less), field evaporation from the tip was
8
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
produced. It created small mounds on Au ( I I I ) surface. Arrays of mounds were
deposited in ordered rows on the sample surface.
1.2
Scanning probe microscopy of high-Tc
superconductors (HTS)
The discovery o f superconductivity in Ba-La-Ca-Cu-O at 38 K. is considered as
one o f the greatest discoveries in the scientific history [11]. It surpassed the maximum
theoretical predicted value o f 30 K. for superconductors from the Nobel price winning
Bardeen-Cooper-Schriefer (BCS) theory for conventional superconductors. The
discovery stimulated intensive activities aimed at searching for new materials
exhibiting superconductivity even at higher temperatures. Not for long, Wu et al.
discovered that YBa2 Cu 3C>7 (YBCO) system exhibits superconductivity around 90 K
[12]. It opened the new era of high-Tc superconductivity. “High-Tc” stands for high
transition temperature and Tc of a superconductor is the critical temperature at which
it undergoes a normal to superconducting state phase transition. Below Tc, it has zero
DC resistance and expels external magnetic field. Above Tc, it has electrical
resistance and it can no longer expel external magnetic field. After YBCO, several
other superconducting compounds were discovered. Bi-Sr-Ca-Cu-0 and Tl-Ba-CaCu-O systems were reported to have Tc’s up to 105 K and 125 K. respectively
[13,14].* Five years later (1993), Hg-Ba-Ca-Cu-O systems achieved the highest Tcto
date o f 138 K [15]. Most HTSs that are being developed into practical applications
have Tc’s above 77 K. They are very appealing for certain applications, which are
already utilizing the conventional low temperature superconductors (LTS) because
they would only require liquid nitrogen (boiling point = 77 K) as a refrigerant. Liquid
nitrogen is relatively inexpensive and easier to handle than liquid helium (boiling
point = 4.2 K), which is the standard refrigerant for LTS. Applications of HTS are not
limited to replacing present LTS technology. Other applications include passive
microwave devices for telecommunications and coated conductors for high power
applications. Another advancement in the application o f HTS is the availability of
commercial cryo-coolers for cryogen-free operation of superconducting devices. For
9
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
example, in mobile communications, HTS microwave devices are now integrated
with these newly developed cooling machines. The rapid advancement in cryo-cooler
technology has really brought more interest in selecting HTS over LTS or normal
metallic conductors (i.e. gold, copper and silver) for applications. The main reason is
that implementation of a cryo-cooler for system cooling is simpler than the use of
cryogen. With liquid nitrogen or helium, more space is taken in a system, creating a
bulky setup. It also requires maintenance. On the other hand, the current-generation
cryo-coolers can operate continuously for more than ten years. They are compact,
low-noise and clean. Lastly, they are well suited for HTS application because they
become more energy efficient as operating temperature increases.
1.2.1. General properties of superconductors
In this section, the most important properties of superconductors are
summarized:
• Superconductivity occurs below a critical temperature (Tc) wherein the material
exhibits zero DC resistance.
• Below Tc, magnetic field does not penetrate the material (the Meissner effect) and
to be exact, the field does exist inside a surface region of a thickness comparable to
the penetration depth A, where persistent screening currents flow. This A. has an
empirical temperature dependence given by:
A(D-/l(0) ,
‘-----
Vi-(r/n)
where A(0) is calculated to be I O'8 to 10'7 m [16].
• The penetration depth continuously changes from the finite value to infinity at Tc.
It means that the properties o f electron system also change continuously. At the same
time, resistivity changes abruptly. One could imagine that electrons do not interact
10
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
with the lattice in the superconducting state. In this case one may expect a large
increase in the thermal conductivity. Nevertheless, it is continuous at the transition.
• Superconductivity can.be destroyed by an external magnetic field H > He, where
the critical field He has an empirical temperature dependence given by:
He(T) = //e(0)[i-(r/rc)2]
If the superconductivity is destroyed by a current I > Ic, the critical current Ic is the
threshold current which produces He at the surface.
• Superconductors can be classified as either type I or type II. A type I
superconductor has only one critical He. When H is below He the magnetic flux is
expelled from the bulk (Meissner state) and above it the material is in normal state. A
type II superconductor has two critical fields Hci and H<o. For H < Hci all the flux is
also expelled from the bulk (as in type I), for Hd < H < Hc2 the material is in the
mixed state. In this so-called mixed state, magnetic field can partially penetrate into
the superconductor without completely destroying the superconductivity. One may
consider a superconductor in the mixed state is composed o f a mixture o f
superconducting (no magnetic penetration) and normal (with magnetic penetration)
materials. When H > H<o, the sample is in the normal state.
• The electron contribution to the specific heat at low temperatures behaves as
exp(-A/kaT), where kg is the Boltzman’s constant. That means that there is a energy
gap in the elementary excitation spectrum. But this gap is strongly temperature
dependent at T close to Tc contrary to behavior in semiconductors. Indeed, the gap
vanishes at the transition temperature.
• A characteristic length called coherence length (^,) exists in superconductors. 2*,
is analogous to an electron’s mean free path in a normal metal. It is basically a
measure o f how likely it is that a pair of electrons will interact with each other. It sets
the scale for spatial variations in the density o f superconducting charge carriers.
•
HTSs have layered structures that give rise to anisotropy. This implies that
electrons move easily in the planes, and with difficulty along the normal o f the
11
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
planes. Within a crystal, a constant force applied along different axes produces
different amounts of electron motion. Thus an effective mass o f the electron can be
defined which changes with crystal direction. Anisotropy is defined by the ratio of the
effective mass of the electron in the various directions.
1.2.2 Characterization of HTS using SPM
The mechanism of high-Tc superconductivity remains to be a mystery despite an
extensive effort worldwide in HTS research. Since the macroscopic behavior of
charges is governed by their microscopic natures, an investigation of HTSs using
various tools that can probe the microscopic mechanisms becomes the focus o f many
groups recently. In this section, the use o f SPM techniques for studying properties of
HTS will be reviewed. SPM has brought new and valuable information about HTS,
which could not be possible with other techniques. It has revealed striking features
and has visualized many interesting properties o f these novel materials.
For STM and AFM, a good demonstration of their power is the discovery of
growth spirals in YBCO thin films [17-19].“ These defect structures would be
impossible to detect using x-ray diffraction techniques because of their broken
symmetry and/or their sporadic distribution. The ability to see such flaws is useful for
finding the substrate and growth conditions that can prevent their occurrence. Another
success of STM was the discovery of charge modulations along the Cu-0 chains o f
YBCO [20]. By performing reversed-bias STM at 20 K. on the Cu-O chain layer o f
cold-cleaved single crystals o f YBCO, a 1.3 nm corrugation was discovered, which
change sign under bias polarity reversal. This is consistent with the hypothesis that
the one-dimensional Cu-O chains undergo a charge density wave transition. This
experimental evidence is
very valuable to the understanding o f high-Tc
superconductivity.
Vortex dynamics in HTS has always been a major interest both for theorists and
experimentalists. It contains rich phenomena and understanding them is important for
practical applications as well as for fundamental physics. In this subject, the greatest
12
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
contribution o f SPM techniques is vortex imaging. Vortices are commonly studied
with MFM. As discussed earlier, MFM is based on the localized magnetic interaction
between the magnetized tip and the superconducting sample. If the tip encounters a
non-superconducting region (defect) in the sample, its levitation force goes away.
Moreover, the interaction force between the tip and a vortex could be attractive or
repulsive depending on whether its magnetic dipole and that o f the vortex is aligned
or anti-parallel. Thus, in a constant force image, a vortex is either a dip or a bump.
The tricky part for this imaging technique is that the vortices must be pinned (kept
from moving) when the field from the tip interacts with them. To do this, the sample
must be cooled in external magnetic field. With MFM, Moser et al. has already
demonstrated single vortex imaging on YBCO at 77 K without the need of special
surface preparation [21]. They observed that a specific number of vortices is
produced by cooling the sample in different fields and they have disordered
arrangement. They also found that the smallest inter-vortex distance is much larger
than the effective penetration depth. Another significant result showed that small
topographic defects are good pinning sites. In this study, Yuan et al. used the MFM to
obtain topographic and magnetic images o f the same region o f YBCO sample [22].
They observed that the location of the vortices coincide with the defects on the
topographic image. Vortex pinning by natural defects or those caused by processing
is important for many applications because when these vortices move, they dissipate
energy resulting to a lossy superconductor.
SSM has been used to study in great detail the flux distribution in tri-crystal
Josephson junctions. Tsuei et al. were able to observe junctions with half a flux
quantum, which is consistent with d-wave symmetry o f the order parameter in several
HTSs [23]. This experiment is o f fundamental importance to the understanding of
HTS mechanism. In another report, Oral et al. has employed SHPM to study vortex
phase diagram in HTS [24]. They observed the first order melting transition o f the
vortex lattice in BSCCO single crystals in situ and in real time. This behavior is also
observable with macroscopic measurements, but with SHPM, detailed behavior of
vortices near transition can be obtained. For example, when the field is suddenly
13
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
raised, the vortices reconfigure themselves in a minimum energy state at the given
field and temperature, but in all the temporary configurations most vortices keep six
nearest neighbors.
Mapping the spatial variation of microwave surface resistance, Rs, o f HTS is
desirable for high frequency applications. Microscopic non-uniformity of microwave
properties is the most common cause of nonlinear effects in HTS such as
intermodulation distortion and harmonic generation. Thus, it is important to
investigate such critical issues in order to circumvent the limitations o f HTS in
microwave applications. The appropriate technique for mapping Rs is NSMM. One
group has employed NSMM with S pm spatial resolution to image the edges of a
patterned YBCO at 80 K. [25]. They found that there are regions which are not
superconducting. This result is very useful for applications because the edges of a
microwave device have the highest microwave current density.
NSOM has been used for a high-resolution investigation o f microstructural
defects on grain-boundary Josephson junctions. For example, McDaniel et al. have
observed that an inhomogeneous distribution of submicron-sized structural defects at
the fusion boundary o f polished SrTi0 3 bicrystal substrates can cause the grain
boundary of an YBCO thin film grown on the bicrystal to wander up to I pm in the
film [26]. These structural defects were found to correlate qualitatively with the
electrical characteristics o f grain-boundary Josephson junctions patterned on the
YBCO film. Aside from characterization, the NSOM technique has already been used
for lithography. An NSOM tip was employed to create a superconducting nano-line
by utilizing the photo-induced superconductivity (PISC) effect. In PISC, an oxygendeficient HTS such as YBCO and RBCO can increase their carrier density by
illuminating them with visible lights [27]. As a result, there is an enhancement in
superconducting properties such as an increase in Tc. Based on this observed
phenomena, Decca et al. used an optical near-field probe to locally excite a
GdBa2 Cu3 0 6 . 5 thin film and create a superconducting nano wire [28]. They were able
to induce photo-carriers and confined them on a scale o f 250 nm. Thus, NSOM
writing of HTS nanowires opens up a new technique for micro-device fabrication.
14
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1.3 Motivation
Development of SPM probes with ability to map two or more physical properties
o f a sample simultaneously is highly demanded in order to correlate different physical
properties at microscopic scales. Theoretically, it is possible to scan the same region
o f a sample successively using different probes in order to map different properties.
However it is extremely difficult, if not impossible, to match the same pixel from
different SPM scans when the pixel dimension falls into a microscopic scale.
In high-Tc superconductors, for example, microwave characterization has both
technological and fundamental importance. The traditional measurement techniques
have been very successful in studying HTS at high frequencies, but it has limitation
because the measurement is a weighted average over large areas o f the sample. This
implies that the intrinsic behavior of the material can be masked by the response o f a
small fraction of impurities and defects. To overcome this limitation, NSMM can be
employed to permit local quantitative measurements of electromagnetic response. It
can provide information on the locations of this small fraction of impurities and how
their microwave properties compare with the rest of the material. However, the
single-channel NSMM cannot provide any other information that might explain what
is responsible for having such spatial variations in the microwave response o f the
sample. Such a problem encountered in HTS microwave devices represents a generic
issue in many other SPM studies of many other materials and justifies the need to
develop multi-channel SPM. For example, integrating another SPM technique such as
NSOM with NSMM will make a more powerful tool for characterization o f HTS
microwave devices. Motivated by such needs, this thesis has intended to pioneer
multi-channel SPM.
15
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 2
Microwave characterization of HTS
film: traditional approach
High-Tc superconductors (HTS) are very promising for microwave applications.
Due to their low power dissipation, they can be made into high-performance devices
such as resonators, filters, phase shifters, etc. Technologically, microwave
characterization is important to understand the fundamental physics that governs the
behaviors of HTS at microwave frequencies, to identify factors that affect the
performance o f HTS microwave devices, to correlate with material processing
conditions that can yield the optimized performance. It presents a generally accepted
measure o f film quality. The complex conductivity o f superconductors gives insights
into the physics o f quasiparticle excitations and collective charge properties of the
material. Such properties include superconductivity, spin and charge density waves,
Josephson plasmons, etc. [28].
2.1 Overview of superconductivity at microwave frequencies
In order to understand the behavior o f HTS in the presence o f microwave fields,
it is necessary to discuss the fundamental physics related to the electromagnetic
interaction with the material. So far, the theory behind high-Tc superconductivity is
not yet well understood, but some classical models are found to give satisfactory
approximation and agree qualitatively with experiments.
16
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.1.1 Two-fluid model
Superconductivity arises when paired electrons move without any resistive loss
due to the presence o f an electric field E. However, above absolute zero temperature,
there is always a probability that thermal energy can split some electron pairs. As a
result, some unpaired or normal electrons are always present in the material. In
general, current transport in a superconducting medium involves the motion of paired
and unpaired electrons within the lattice of the solid. This is known as the two-fluid
model. It is therefore reasonable to model the superconductor in terms o f a complex
conductivity ct = (cri - jcri), where G\ represents the normal electrons and CT2 represents
a mixture of paired and unpaired electrons. A physical interpretation of this complex
conductivity can be obtained by simple classical approach. First, consider the
equation of motion o f an electron pair under the influence o f E.
dv
2m- J- = -2eE
dt
(2.1)
In this equation, vs is the electron velocity, e is the electron charge, m is the electron
mass and E is the applied electric field. Similarly, a normal electron has the following
equation of motion.
dv
v
m — - + m ————eE
dt
r
o
(2-2)
t is the momentum relaxation time due to the scattering of normal electrons with the
lattice. The motion o f electrons induced by the applied electric field gives rise to a
current density J, which is related to the velocity by: J = -nev, where n is the electron
density. From this relationship, the current densities due to superconducting and
normal electrons can be written as follows:
17
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
J , = -n ,ev , (paired)
(2.3)
J „ = -n,,ev(, (single)
(2.4)
where ns and n„ are the paired and normal electron densities respectively. The total
current density will just be:
Jr =J,+J
(2.5)
It is related to E by:
J r = crE
(2.6)
For a sinusoidal E field, the J and v will also be sinusoidal. Suppose
E (0 = EQe‘a* then J ,( f ) = J
and J„ = J„eJa* . From these definitions, a time
dependent expression for vs and vn can be obtained explicitly using equations (2.3)
and (2.4). Then they can be plugged into the differential equations (2.1) and (2.2) to
give:
(2.7)
1
jm c o
(2.8)
The expressions for J s and J„ can be inserted into Eq. (2.6) to give:
a = <x, - jcr, =
n„e2r
n
.* n,e
j 2
t
e o nft ^ r 2
18
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(2-9)
Eq. (2.9) defines oi and
02
as a function of frequency (to) of the applied field. Note
that for most practical applications, relaxation effects can be ignored because to2 t 2 «
I . With this approximation, CTi does not change with frequency.
02
can be simplified
by factoring out nse2 /mco. The terms left inside the parenthesis will just reduce to
unity because of the assumption that
<o 2t 2
«
1. Thus, as frequency increases,
oi
decreases as l/co. This behavior is similar to that in a circuit composed o f an inductor
and a resistor connected in parallel (Fig. 2.1). When the current is constant, the
inductor completely shorts the resistance, which is equivalent to superconducting
CT-,
Fig. 2.1: Equivalent circuit of complex conductivity.
state. But as frequency of the current increases, the inductive reactance increases
creating an obstacle for the current flow. This is equivalent to the increase of
dissipation in the superconductor at very high frequency. From classical formulation,
a physical interpretation of complex conductivity became clear and it agrees
qualitatively with the actual observation o f HTS at microwave frequencies.
From Maxwell’s equations, the general form o f wave equation can be derived:
V2E = n — + / if
dt
(2-10)
dt~
19
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
here (i and s are the magnetic permeability and dielectric permittivity, respectively. If
the normal and paired electrons are treated separately,
J„ = cr,E
m Al
- 5L
^
(2-U)
= e
n,e* at
The first relation in Eq. (2.11) is the ordinary Ohm’s law and the second one is
derived from Eqs. (2.1) and (2.3). The time derivative o f J in Eq. (2.10) can be
evaluated using the assumption that J = J n + J s and that the normal and
superconducting current densities J n and J s are related to E by Eq. (2.11). The wave
equation then becomes:
at
A
at
where A = m/nse". For time independent E, this differential equation reduces to:
V2E = ^-E
A
E and B are related by Faraday’s law so that the equation above can be converted to:
V2B = ^ B = —
A
XL
(2.13)
Considering a simple one-dimensional case, the solution is:
B(r) = B(0)el' ^ t)
This indicates that the magnetic field decays exponentially inside the superconductor
from the surface (z=0) with a decay length equal to Xl, which is called the London
20
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
penetration depth. For a time varying field E(7) = E0e’ed . Eq. (2.12) becomes:
(2.14)
From the above equation, it can be recognized that
02
= 1/coA = 1/ cop^L2 = nse2 /o)m.
This is the same as the one obtained in Eq. (2.9) when the relaxation effect is ignored.
2.1.2 Surface impedance of superconductors
If a plane electromagnetic wave is traveling in the z-direction in a material filling
the half-space z > 0 , then the solution to the wave equation for a wave propagating
forward is:
E = Eze~r:
H = Hye~rz
(2.15)
Here, y is the propagation constant, which is a complex quantity. The real and
imaginary parts represent the attenuation and phase respectively of the traveling
wave. E and H are related by Faraday's law to give:
(2.16)
dz
The above equation shows that E and H are perpendicular. The surface impedance of
the medium is defined as the ratio between Ex and Hy.
(2.17)
21
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In the above equation, ct can be replaced by the complex conductivity cti- j g z if a
superconducting material is considered. It is reasonable to assume that the
displacement current is negligible (o » coe) and the transport current is dominated by
paired electrons (cri «
02
). From these assumptions, it can be shown that:
(2.18)
Again, Eq. (2.18) is a complex quantity, which can be separated into its real and
imaginary parts. Just like the complex impedance of lumped circuits, the real part is a
resistance, which represents real dissipation and the imaginary part is the reactance,
which represents phase. Since Zs is surface impedance, its real part is called surface
resistance and it is denoted by Rs. The imaginary part is called surface reactance and
it is denoted by Xs.
(2.19)
(2.20)
Using the expression for London penetration depth obtained in Eq. (2.13),
02
could be
replaced by l/copA./.2 to yield:
2
R
2
i3
_ P & o A
(2.21)
2
X, =ficoAL
(2.22)
These equations, although obtained from phenomenological model, can apply to most
practical microwave applications. In experiments, Rs and Xs are measured and they
can be used to find Oi and cto. Investigation of these quantities can give insights into
22
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the behaviors of quasi-particle excitations and collective charge properties of the
material. In practical applications, both resistive and reactive components play
important roles in determining performance of a HTS device. The Rs determines the
quality factor and the insertion loss whereas Xs determines the sensitivity to
temperature variations. Long-term stability o f a device requires low values o f Rs and
stability of Xs.
2.2 Review of microwave characterization of HTS films
A superconducting film is the main constituent of many o f the device
applications.
Traditional microwave characterization of HTS
film
involves
measurement o f Rs and Xs. In this dissertation, the emphasis is on the real microwave
dissipation. Many techniques for measuring Rs on films have already been established
and three of them will be discussed here. Note that these measurement setups are also
capable of determining Xs. Sensitive measurement of Rs can be achieved using
resonant methods. It is therefore necessary to discuss the fundamentals of the
microwave resonator before going into the details o f the techniques.
2.2.1 Introduction to microwave resonator
A microwave resonator is an apparatus, which is capable of maintaining
oscillating electromagnetic field at microwave frequency. Generally, it can have
many distinct resonant frequencies depending on its geometry. In the presence of an
oscillating field, a resonator stores and at the same time dissipates energy. The
dissipation arises because o f the finite losses in the resonator. For example, for a
resonator composed of conductors and dielectrics, dissipation is due to the finite
resistance and loss tangent (tan5) o f the conductor and dielectric respectively. Loss
tangent is defined as the ratio between the imaginary and the real part o f the dielectric
constant o f a material. Radiation loss may also add to the total energy loss. The
performance o f a resonator is quantified by its quality factor, which is defined as:
23
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
^
Energy stored in the resonato
Q = tu
—----------------------------Average power lost
(2.23)
With advances in microwave testing electronics, experimental measurement of Q is
now straightforward using commercial vector network analyzers. In the transmission
mode, a resonator has an input and output feed lines. The sample to be measured is
inserted between the input and the output terminals of the analyzer, which sends a
frequency sweep signal, centered at one of the resonant frequencies (f0) of the
resonator. With a constant input power, the analyzer measures the power output as a
function o f the frequency. Figure 2.2 is a plot showing a typical transmission
characteristic of a resonator. S21 is called the insertion loss (IL). It is expressed as 10
log (Pout/Pin), where Pin is the power input to the resonator and P0Ut is the measured
3dB
. 3..
HPBW
T5
-6 --
-
12 -
-1 5 -0.04
-0.02
0
0.02
0.04
( f-fo )S fo
Fig. 2.2: A typical plot of insertion loss (IL) of a resonator versus the normalized frequency.
24
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
power output. At resonance, Pout is maximum and almost equal to Pjn so IL is nearly
zero. The half-power bandwidth (H P B W ) is a frequency band where Pout is above a
half of Pout(/o). To measure Q, the ratio /o /H P B W is calculated. Thus, a high-Q
resonator is featured with a sharp resonance peak because the corresponding HPBW
is very narrow. Resonators are classified into three categories based on their structure.
They could be one-dimensional (i.e. patterned line resonator), two-dimensional (i.e.
patterned ring resonator) or three-dimensional (i.e. cavity resonator). They will be
discussed in detail in the next section.
2.2.2 Common measurement techniques
Patterned line resonator technique employs a patterned narrow planar line on a
film. The line is open at both ends so that reflections can occur at those points. As a
result, standing waves are set up in the section o f the line and hence resonance is
achieved. The fundamental mode o f the resonator is when the line is half-wavelength
(A/2) long. Other modes occur at nA/2, where n = 2,3,4.... The line resonators can be
configured into three different ways namely: microstrip [29], stripline [30,31]’ and
coplanar [32], These configurations only differ in the implementation o f the ground
plane. Figure 2.3 shows the diagram of the three configurations. In the microstrip
resonator (Fig. 2.3a), only one ground plane exists on the back o f the patterned strip.
This plane could be deposited on the same substrate as the sample when two-sided
fabrication is achievable. It can also be another HTS film pressed on the back o f the
patterned strip. The stripline resonator (Fig. 2.3b) utilizes two ground planes. The
resonating strip is sandwiched between two superconducting films. The resonator is
separated from the upper and lower ground planes by the dielectric substrate. The
coplanar configuration (Fig. 2.3c) consists o f two ground planes and a central strip of
superconductor deposited on the same surface of a substrate. For all these three
configurations, microwave field is coupled in and out o f the resonator capacitively
through the coupling gaps. The resonating strip could also be meandered or spiraled
25
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
coupling gap
coupling gap
resonator
feed line
feed Jin?—
substrate
t
ground plane
£
___________________________
(a)
ground plane
ground plane
(b)
ground plane
ground
(c)
Fig. 2.3: Three configurations of patterned resonator using HTS film, (a) microstrip resonator
(b) stripline resonator and (c) coplanar resonator
26
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
to increase its length without increasing the dimension o f the substrate. This reduces
the resonant frequency since it is inversely proportional to X. Working at the lower
microwave band is relatively easier as compared to X-band (8-12 GHz) and higher
ones where more microwave engineering is necessary. From all these line resonator
techniques, Rs is determined by measuring the Q of the resonator. Their relationship
is Rs = k/Q, where
k
is a geometrical constant which could be determined by proper
calibration of a particular measurement setup. It should be remembered that the
measured Rs using patterned line resonator technique may differ from those measured
on original films due to the extrinsic effects associated with photolithographic
processing. However, from a circuit designer’s point of view, the Rs values obtained
on patterned samples are more useful because real applications of HTS require
patterning. For routine sample characterization, this method is not suitable because it
is destructive and requires fabrication o f the circuit on a fairly large sample. Although
for power dependence measurements, this has advantage because high current density
can be easily achieved with a low input power due to the narrow line width of the
resonator.
Dielectric resonator [33]. Here, a pair of HTS films, assumed to have equal
properties, sandwich a piece of dielectric (commonly a cylinder) [34]. The dielectric
has a high-dielectric constant and low loss tangent.
Sapphire (AI2 O 3 ) is such a
material. The dielectric constant and loss tangent (10 GHz) perpendicular to c-axis o f
AI2 O3 are 11.5 and
8 .6
x 10' 5 respectively. Figure 2.4 shows a typical measurement
configuration using this technique. Microwave power is coupled in and out of the
resonator using coupling loops and the assembly is typically secured in a Cu
enclosure. For this resonator structure, radiation loss is almost eliminated due to the
confinement of microwave energy inside the dielectric. Thus, loss would only come
from the HTS films and the dielectric. The latter is also negligible due to its low loss
tangent The average Rs o f the two HTS films is related to their Q-value by Rs =
G/Ofnm. G is a geometric factor, which can be calculated theoretically. The
experimentally measured Q-value (Qefr) is the effective Q-value o f the film (Qfiim) and
the dielectric (Q<j) and they are related via:
27
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In this method, calibration is not required.
Measurement is simple and
straightforward. It is also nondestructive since the HTS film does not need to be
patterned. It is thus well suited for routine sample characterization. The only
Copper enclosure
Loopterminated
cable
Loopterminated
cable
dielectric
Fig. 2.4: Configuration of dielectric resonator for measuring the surface resistance (R,) of
HTS films.
disadvantage is that it requires two samples, which should have very close microwave
properties and dimensions. They should be large enough to cover the ends of the
dielectric. A simpler version requiring one sample has also been reported [35]. Here,
only one end of the dielectric is covered by the HTS film and the enclosure is large
enough to minimize the loss due to the enclosure. However, the excessive radiation
loss can lead to uncertainty in the measurement.
Cavity perturbation technique (CPT). CPT has been developed in the late
1940’s [36] and has undergone many variations. However, the general concept
remains the same. CPT uses a conducting cavity to generate standing wave patterns or
28
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
resonance. During the measurement, this resonance is disturbed by introduction of the
sample into the cavity. This disturbance results from the interaction of the sample
with the microwave fields, which changes the original resonant frequency and Q o f
the cavity. Detection of these changes will allow determination of Rs of the sample.
Note that for a particular geometry, there are a number o f different standing wave
configurations, called modes. It is necessary to choose the one that is very sensitive to
Rs changes of the sample. In the next chapter, a more detailed discussion of a
particular CPT setup will be presented.
2.3 Nonlinear effects in HTS
The high performance exhibited by HTS microwave devices is only at low power
levels due to the detrimental nonlinear effects even at moderate power levels. These
nonlinear effects, including power dependence o f Rs, harmonic generation and
intermodulation, have been a subject of considerable interests but so far are not yet
well understood. Systematic investigation of these phenomena is therefore necessary
and could help lead to production of HTS films with better power response, critical to
second and third generation digital wireless communication. Such investigation could
also contribute to the fundamental understanding of the nature of HTS.
2.3.1 R, power dependence
In the classical model of the surface impedance of HTS film (section 2.1), Rs is a
function of frequency and temperature. But it was not shown explicitly that Rs is also
a function of the microwave power applied to the HTS film. Nevertheless, many
groups have observed power dependence of the Rs on HTS films. In a simplistic way,
Rs can be interpreted as a response, in the form of dissipation, of the material to an
applied microwave field. It has been experimentally observed that at very low input
power, R* exhibits very weak or no power dependence. This is known as the linear
regime of the superconductor. When power is continuously increased, a critical power
29
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(Pc) will be reached wherein Rs starts to increase with the input power. In this state,
the superconductor is said to be in the nonlinear regime. Thus, the critical power Pc
separates the linear and nonlinear regimes. HTS devices are normally operated below
their Pc. Determination of Pc, or the microwave power handling measurement, is
therefore desirable in order to find the power limitation of an HTS sample. Most
power handling measurement techniques are simply extensions of Rs measurements.
The major additional equipment is the microwave amplifier, which enables the input
power to be amplified. Rs can then be monitored as a function of the input power
level. For example, Oates et al. [37] used their stripline resonator technique to
measure power dependence of Rs. They converted the input power into the equivalent
RF magnetic field (Hrf) generated by the applied RF current. This eliminates the
dependence of results on circuit geometry. For YBCO, they found that R* can be
approximated by a quadratic dependence on Hrf: Rs = Rso + a Hrf2, where Rso is the
surface resistance at very low fields and a is a fitting parameter. In post-annealed
films and in films sputtered at lower substrate (LaAlCh) temperature, they observed a
linear Hrf-dependence of Rs. In another report, Nguyen et al. observed quadratic Hrfdependence o f Rs even at low and intermediate fields [38]. At high fields, R$
increases faster than Hrf2. The variation in the observed power dependence from
sample to sample could be attributed to the domination of extrinsic effects, besides
intrinsic ones. It is well known that microwave power handling of HTS films is very
sensitive to the defects in the material, making investigation difficult to distinguish
the intrinsic and extrinsic mechanisms.
2.3.2 Harmonic generation
Like any other nonlinear element in an RF circuit, HTS nonlinearity generates
harmonics. With a nonlinear resistance the relationship between I and V can no
30
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
longer be defined by Ohm’s law alone, rather expressed as a Taylor’s series
expansion:
f £J3T
1 \
f dr I
6 V 3+ 0 {S V 4)
I(V) = I ( 0 ) + f ^ l <5V+ i
&V2 + —
3!
KdV >v=o
2. cN 2 v=o
v=o
(225)
where the coefficient in the second term can be expressed as:
d l\
dV Jw
= l_
R
In this expansion, all the terms higher than the linear term represent nonlinearity. It
should be realized that all even order terms vanish because o f symmetry with respect
to polarity. Without imposing a DC bias, I(-V) = -I(V). Hence, the lowest harmonic is
the third-order term. Measuring the third harmonic generation can be performed using
a non-resonant technique with a single-tone input signal. This signal, which should be
free o f harmonics, is fed to a sample (i.e. HTS transmission line) then a spectrum
analyzer is used to measure the output power o f the fundamental and the third
harmonic.
2.3.3 Intermodulation
Aside from the harmonic generation, another observable nonlinear effect is the
intermodulation. At moderate power levels, different frequency components interact
with each other. This implies that modulated signals in different frequency channels
may get mixed up. Intermodulation is a serious concern for applications such as
telecommunications because the signal often has spectrum with different frequency
components. Investigation o f intermodulation is commonly performed on patterned
resonator or narrowband filter using a two-tone input. This input is a mixture of two
signals having different frequencies, f t and f> but having equal amplitude or power.
The tones are separated symmetrically about the center frequency of the resonator or
31
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
narrowband filter by an amount o f A f such that both tones are within the 3 dB
bandwidth of the device. The output power in the third order mixing products 2// - _/}
and 2 f*- f i are then measured in a spectrum analyzer as a function of the input power.
In this measurement, the quantity being determined is the third order intercept (TOI).
This number characterizes the intermodulation distortion (IMD) signal observed in a
sample. TOI is defined by the intercept between the linear extrapolation o f the output
power of fundamental and IMD signals.
The mechanism of the third order harmonic generation and the third order
intermodulation is the same. Moreover, the recent report of Lahl et al. [39] showed
that there is a linear correlation between the power handling capability and IMD
measured in a YBCO coplanar resonator. This indicates that the mechanism of
nonlinear power dependence o f R* is also the same as IMD and harmonic generation.
Also, they found out that IMD signals are more sensitive to input power. In sweeping
the input power from low to high levels, generation of these IMD signals were
observed before the degradation o f Q value. It suggests that the IMD measurement
can be used for characterization o f power handling capability.
32
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 3
Microwave characterization of
HgBa2CaCu20 6+s thin films
The discovery of superconductivity near 135 K on Hg-based HTS makes them
very attractive for microwave applications since higher Tc implies higher operation
temperature. This could significantly ease the cooling requirements that have been
one of the major obstacles in commercialization o f HTS microwave devices. It
becomes more appealing when integration with cryo-cooler technology is considered.
This eliminates the use of conventional liquid cryogen. As operating temperature
increases, efficiency of a cryocooler also increases. For example, the efficiency of a
cryocooler is calculated to be 0.345 at 77 K and it improves to 0.5 to 0.579 if the
device operating temperature is raised to 100 K and 110 K. respectively. It is therefore
necessary to investigate their microwave properties to evaluate their potential for
microwave applications. The results reported in this chapter are more or less the first
study of the microwave properties on Hg-HTSs, while many o f them represent the
best so far achieved on HTSs
3.1 Microwave surface resistance (R,) measurement
3.1.1 Sample preparation
33
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In this experiment, HgBaiCaCuiOe+s (Hg-1212) thin films were fabricated using
two different processes: a cation-exchange process [41] and a conventional thermalreaction process [42]. In the cation-exchange process, Tl-based HTS films (either Tl2212 or Tl-1212) are employed as precursors. They are converted to Hg-1212 films
by Hg-vapor annealing. The second process involves deposition of Ba-Ca-Cu-O
precursor films followed by thermal annealing at high Hg-vapor pressures. The
dimensions of Hg-1212 films, as well as that of Tl-2212 and YBCO films used in this
experiment for comparison, were typically 2 mm x 3 mm. Although many samples
have been measured, the data shown in this dissertation were from: four Hg-1212
films made using cation-exchange process (labeled as samples Ia-Id), a Hg-1212 film
by conventional thermal reaction process (labeled as sample II), and a Tl-2212
precursor film (labeled as sample III). All these films were grown on (100) single
crystal LaAlCh substrates. Several YBCO films, with their Rs values measured at the
Los Alamos National laboratory by cavity wall replacement technique, were used to
calibrate the measurement system.
3.1.2 Sample properties
X-ray diffraction (XRD) was used to determine the orientation of the film, which
shows that all samples used have their c-axis perpendicular to the plane o f the
substrate. XRD pole figure measurements indicate that samples la - Id and III are
highly epitaxial, with their a axis aligned with the (010) axis o f the substrate. To
characterize Tc and Jc o f the films, magnetization (M) was measured as a function o f
temperature (T) and magnetic field (H) in a Quantum Design superconducting
quantum interference device (SQUID) magnetometer. Jc values were estimated using
Bean model [43,44]’, which states that Jc of a cylindrically shaped type II
superconductor is proportional to its remanent magnetic moment, Mmn. Mrem is
typically extracted from an M-H hysteresis loop obtained at constant temperature
where the width o f the magnetic branch AM taken at H=0 is equal to 2Mren,. In
the case o f a cylindrical geometry, the critical current density is expressed as [45]:
34
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Jc
30 M rem/Rcff
The units for Mrem and R^fr are emu/cm3 and cm respectively so that the resulting Jc is
in A/cm2. R«fr represents a length scale where shielding currents can circulate freely.
Most often, it is assumed to be equal to the sample radius or sample side for square
samples. Figure 3.1 shows Jc vs T curves for samples la, II and III at zero field. It
seems that Jcs of the two Hg-1212 samples are comparable while that o f Tl-2212 is
higher at low temperatures. For example, at 5 K, samples la, II and III have Jc values
1 0
r
\
•
0000
•
1 0
0.001
-0002
* - ^ -0003
:3
-0
004
' ® -0005
; 2 -0.006
•0007
•0.006
•0008
1 0
s
1
0
•
.............................................
20 40
B0
80 100 120
T4n y n f t if»(*)
0
i
20
•
I'
40
I
0
'------------------r
60
80
'
1------- 1
100
|
120
Temperature(K)
Fig. 3.1: Zero-field JcS as a function of temperature o f Hg-1212 made in cation-exchange
process (open circle), Hg-1212 by conventional thermal reaction process (solid diamond) and
Tl-2212 precursor film (solid circle), which have dimension of 2 mm x 3mm. Inset: Plot of
magnetic susceptibility vs temperature.
35
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
o f 22, 23 and 33 MA/cm2, respectively. It should be realized that these Hg-1212 films
have not been fully optimized. In several Hg-1212 films, Jc ~ 40 MA/cm2 have been
obtained at 5 K and self-field. Tl-2212 film displays slightly better Jc in the low
temperature regime but it has also the most rapid decrease with increasing
temperature. At 77 K, Jc of the Tl-2212 film drops to ~ 4 MA/cm2, which is in
between that of sample la (~ 5 MA/cm2) and II (~2 MA/cm2). The decrease in Jc of
the Tl-2212 film is significant at 100 K. as the Tc (— 103 K. as shown in the inset o f
Fig. 3.1) of the sample is approached. Both Hg-1212 films still show moderate Jc at
100 K. due to their higher Tc ~ 120 K (inset o f Fig. 3.1). The Jc - 2 MA/cm2 of sample
la, however, is four times higher than that o f sample II, suggesting that poorer quality
epitaxy is a major current limiting factor. This Jc difference is even enlarged at higher
temperatures. For example, sample la still has a Jc of 0.5 MA/cm2 at 115 K while that
o f sample II is not measurable. This result indicates that high-quality epitaxy is the
key to achieve high Jc at temperatures close to Tc as weak-link effect from high-angle
grain boundaries becomes a dominant factor to suppress current flow. By comparison
between samples la and II, it has been noticed that the quality of the film epitaxy has
been greatly improved in the cation-exchange process, although their Tc values are
nearly the same.
Surface features of the three samples are shown in the scanning electron
microscopy (SEM) pictures in Figure 3.2. A comparison between the upper (sample
la) and lower (sample III) panel reveals changes in film surface morphology during
the Tl-Hg cation-exchange process. The number of voids or holes increases on Hg1212 and the surface becomes rougher after conversion. The surface o f sample la
however, is much smoother than that o f sample II. It is interesting to notice the
presence of finely dispersed impurity particles on the surface of sample la, which are
Hg-rich from an analysis using energy dispersive x-ray spectroscopy (EDS) and most
o f them can be wiped away with tissues. Since no impurity phases were visible in
XRD data, we suspect that most of these Hg-rich particles present on the surface o f
the film and their volume portion is negligible. In contrast, massive impurity
particulates can be identified on the surface o f sample II (middle panel o f Fig. 3.2).
36
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
10 KV x2000
M ITV .M M
2a KV t2000
! • HU
10 tun
Fig. 3.2: Surface morphology of (a) Hg-1212 by cation-exchange, (b) Hg-1212 by
conventional process and (c) Tl-2212 precursor taken by scanning electron microscope.
37
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The dimension o f these impurity particulates is much larger than that of Hg-rich
particles on sample la so that the surface of sample II is rougher than that o f sample
la. Consistent with the XRD results, these impurity particulates are either Hg
(possibly HgCaCh) or Ba-Cu-Ca rich phases as suggested by the EDS analysis. The
presence o f impurities could be another reason for the degraded Jc (and Rs, which will
be discussed in the next section) o f sample II. Reducing of impurity phases of HgHTS films is an important success o f the cation-exchange process.
3 .1 J Experimental technique
Rs o f the films were characterized using a modified CPT developed by Ormeno et
al. [46] as shown schematically in Figure 3.3. This setup is one o f the many
variations of CPTs. It is designed for measurement at liquid helium environment
.S. Tube
Teflon Guide
S.S. Rod
Loop Terminated Coax
Thermal Insuladon
Heater
Brass
apphire Rod
in Film Sample
Sapphire Disk Resonator
*
Nb Cavity
Fig. 3.3: Diagram of the experimental probe for measuring the surface resistance (R5) of HTS
films using cavity perturbation technique. The resonator was excited in the TEon mode at
11.2 GHz.
38
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(T~ 4.2 K.). This technique does not require large samples. Typically, measured
samples have dimensions from 2 mm x 2 mm to 4 mm x 4 mm. The superconducting
cylindrical cavity, made of niobium, is loaded with a sapphire disk dielectric
resonator. The disk has a good thermal contact with the bottom end of the cavity. Its
cylindrical axis coincides with cavity axis. The resonator is operated in the
transmission mode, with microwave power coupled in and out of the resonator by
semirigid stainless steel cryogenic coaxial cables terminating in loops. It is excited in
the TEou mode at 11.2 GHz. to induce an ab-plane screening current in the sample.
This mode is the most sensitive to the Rs of the sample. During measurement, the
sample is mounted with a very thin layer of vacuum grease on a vertically movable
sapphire rod that goes through a cutoff hole centered on the top end of the cavity. The
rod connects the sample to an external heater block. The heater block assembly
contains a resistive winding and a thermometer. The sample-heater-thermometer
assembly is thermally isolated from the cavity so that the sample temperature can be
controlled independently without perturbing the superconducting state o f the cavity.
The surface of the film faces the dielectric resonator. By moving the sapphire rod, the
field incident to the sample can be changed. This causes f 0 o f the cavity to shift and
the half-power bandwidth to change. Let 8 / denotes shift in f 0 and 8BW represents
change in the half-power bandwidth for a certain change in the sample position. As
long as the TEon is not polluted by spurious low Q cavity modes that arise due to the
movement o f the sample, the dependence of SBW to 5 / is linear. The ratio between
8BW and 8 /is directly related to the Rs of the sample, that is: Rs = k(SBW/8/), where
k
is the resonator constant and depends on the geometry of the resonator. In principle,
there is a need to calibrate the resonator in order to obtain tc. In this measurement, ic
was determined using samples with known R$. For each sample, a position sweep at a
chosen sample temperature was made and the SBW is plotted as a function of 8 / The
slope was multiplied by
k
to give Rs. Position sweeping, temperature ramping and
data gathering were controlled by a computer.
39
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.1.4 Results and discussion
The results for several Hg-1212 samples are plotted as a function of temperature
in Figure 3.4. Tc o f these films can be determined from the transition of Rs. They all
have Tc values around 120 K, which agree very well with the values obtained from
magnetization measurements. The Rs of most Hg-1212 thin films fabricated from
cation-exchange are less than 1 m fi below 110 K when scaled at 10 GHz. This
scaling is based on the dependence of Rs to co2 obtained from phenomenological
treatment discussed in chapter 2. The scattering o f Rs values from sample to sample is
attributed to the fluctuation in the growth processing parameters. It is expected to be
reduced when fine control of parameters is reached. Sample III is the result for a Tl2212 film showing comparable Rs values below
Tc although Tc is 20 K less
than that o f Hg-1212. Sample lb has the smallest Rs in the temperature range. At 77
6
5
- O — la
- A —lb
Swnpta II: Hg-1212
ic
4
3
2
CD .
1
0
60
80
100
Temperature(K)
120
Fig. 3.4: Surface resistance (R*) scaled at 10 GHz against temperature of the Hg-1212
samples from cation-exchange (open shapes) and a Tl-precursor (solid circle). Inset: RS(T) of
Hg-1212 made from conventional process.
40
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
tC, it is around 0.2 mQ and at 100 K it increases to 0.3 mQ. It has been noticed that
sample la has nearly constant Rs~0.3 mQ below Tc while other samples show
stronger temperature dependence. This indicates that R* is dominated by residual loss.
Residual loss can be understood by considering that RS(T) o f HTS does not really
drop continuously as temperature approaches zero. Most often, it exhibits a plateau
over a wide range o f temperature. Thus, for practical situations a residual surface
resistance is added to the expression for Rs(T). Its value depends on the type and
quality of the superconductor and arises from the effects of impurities, trapped flux or
weak links at grain boundaries. It should be pointed out that Rs o f sample II, which is
made from conventional process, is much higher than that o f samples Ia-Id (inset of
Fig. 3.4). At 77 K, it is up to five times higher than that of sample lb although the two
films have similar Tcs.
Sample III is the result for a Tl-2212 film showing
comparable Rs values below Tc although Tc is 20 K less than that of Hg-1212. Thus,
an increase of 20 K. in operation temperature seems to be feasible in Hg-1212
microwave devices over their Tl-2212 counterpart. Significant improvement o f R* has
been achieved on Hg-1212 thin films made in cation-exchange process. This
improvement may be explained by better phase purity, better epitaxy and smoother
surface morphology of samples processed by this technique.
The Rs data for Hg-1212 thin films from the cation-exchange process is fairly
comparable to the data reported earlier for YBCO and Tl-2212 HTS films. Figure 3.5
shows a plot of Rs(T), which is scaled at 10 GHz, for these two materials. For YBCO
at 77 K, Rs of YBCO ranges from 0.44-0.7 mQ [47]. At temperatures (50 - 70 K.),
which is close to Tc (~- 88 K), Rs exhibits weak temperature dependence. This
behavior, which is also observed in the Hg-1212 films (Fig. 3.4), suggests that in this
temperature range, Rs is dominated by residual loss probably due to extrinsic defects.
Below 50 K, it begins to drop. At 16 K. its value is 37 |oQ. This low temperature
behavior does not agree with BCS theory where Rs is proportional to e '^ 7. They also
calculated Rs(T) using two-fluid model where R*(T) =
(T/Tc)4 / 2(1 —
(T/Tc)4)374. They assumed a normal conductivity just above Tc (crn = 10"4 Q 'cm '1) and
41
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
—
-
•
o
X
□
.
1
—
• 1
•2
•3
V4
2-FLUID
—1--------- 1--------- 1--------- 1--------- 1—
0
*
W
X *» □
X
a
#
□
•
o
s
o
-
o
E
X
■
«n
o □
0.1
0 .01
10
o
•
s
w
-
•
■
-
0 °
X
•
5
■
*
•
-
1
1
l
L
L
20
30
40
50
60
1 ------1
70
80
90
Temperature (K)
(a)
300
a3L
Ui
O
200
100
CO
0
20
40
60
80
100
TE M PE R A T U R E (K)
(b)
Fig. 3.5: Surface resistance (R*) scaled at 10 GHz of two different HTS films namely: (a)
YBCO and (b) Tl-2212 as reported by I.S. Gergis et al. [47] and Holstein et at. [48]
respectively.
42
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
penetration depth at 0 K to be Xo = 150 nm. The calculated Rs values are represented
by dots in Fig. 3.5(a). They are lower compared to the measured values at all
temperatures. The behavior of Rs at low temperatures varies from sample to sample
and in fact, it is not yet resolved. The key problem is to understand the mechanism of
residual loss at low temperatures. This has been a difficult task because o f problems
with sample variation and resolution limit o f measurement techniques. For Tl-2212
(Fig. 3.5(b)), the measured R* are 300, 130 and 23 pQ at 95, 77 and 4.2 K
respectively [48]. The temperature dependence of Rs is approximately linear over a
very wide temperature range extending to 80 K, which indicates a high-quality film.
Since the residual loss is not dominant in this particular sample, Rs at 77 K. is
expected to be lower than for YBCO at the same temperature. Tl-2212 has higher Tc
(110 K.), implying a significantly smaller k at 77 K. relative to YBCO. This translates
to lower Rs(77 K).
3.2 Microwave power handling measurement
3.2.1 Device preparation
In this particular experiment, the non-resonant technique was employed to
investigate the microwave power handling capability of Hg-1212 microstrip
transmission lines. Fabrication o f patterned Hg-1212 structures by standard
photolithography process is usually difficult due to the susceptibility o f Hg-based
compounds to water and water-based chemicals. To overcome this difficulty, device
patterning was performed on the Tl-2212 “precursor” film and this “precursor” device
was then converted into Hg-1212 device in a Tl-Hg cation-exchange process [41].
The centerline has a dimension o f 5 mm x 0.15 mm. The corresponding characteristic
impedance (Zo), which can be calculated from the line width, substrate thickness (0.5
mm) and the dielectric constant o f the substrate (~26) is about 34 Q. The inset o f
Figure 3.6 shows the layout of the microstrip sample already mounted on a Cu stage
with corresponding SMA connectors. The ground plane is spring loaded so that it
43
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
presses the Ag contacts to their corresponding transition pins. The mounted sample is
then placed inside a probe, which is equipped with a heater and temperature sensor.
The probe is pumped to a good vacuum and immersed in liquid nitrogen bath.
3.2.2 Experimental technique
Figure 3.6 is the diagram for the measurement setup. In this measurement, the
critical input power (Pc) was defined as the maximum microwave power the
Ag contact-
Network Analyzer
Temperature Controller
Power Meter
Temperature Controlled
v Environment
Device Under Test
(see inset)
RF Amplifier
Fig. 3.6: Experimental setup for power handling measurement of microstrip lines. In this
configuration, the reflection coefficient (T) and insertion loss (IL) are determined
simultaneously. The RF amplifier is only inserted when power level of network analyzer is
not adequate. Inset: layout of the Hg-1212 microstrip line on LaA103 substrate as prepared
for this measurement.
44
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
superconducting line could carry continuously without showing any performance
degradation. This criterion is the one applied to real HTS devices. Their critical
power rating is based on how much microwave current they can carry without
showing undesirable effects. Their normal operation is at power levels where their
design specifications are met. The measurement o f Pc on Hg-1212 microstrip lines
was performed at several selected temperatures using 1 GHz microwave. To
determine the actual power (Pin) penetrating the sample, it was necessary to know the
reflection coefficient (T), which was measured using a network analyzer (HP8722C)
connected at the input side of the microstrip line. An RF power meter (Boonton 4220)
was connected at the output side (Fig. 3.6) to serve as the load of the sample. With
this setup, T and Insertion Loss (IL) o f the sample were measured simultaneously.
Note that if the power applied to the sample (Pa) is known, then Pu, is approximately
(l-T)*Pa. On the other hand, IL(dB)=Pin(dBm) —P0Ut(dBm), where Pout was monitored
by the power meter.
3.2.3 Results and discussion
Figure 3.7 shows the measured IL(T) o f two Hg-1212 samples (labeled as D-l
and D-2, respectively). Both samples have T around 0.3 below Tc. It can be seen
from Figure 3.7 that the transition from normal state to superconducting state occurs
at 119 K and 121 K for D-l and D-2 respectively, consistent with the magnetic and
microwave measurements on un-pattemed Hg-1212 films. The Tc values of the Hg1212 microstrip structures fabricated using cation-exchange method are typically in
the range o f 110-122 K. The insertion loss below Tc is below I dB for both samples
and they are almost flat at lower temperatures. This suggests that the residual loss at
temperatures below Tc is dominated by the loss from connectors and contact
resistance. For example, the two SMA connectors were found to contribute about 0.2
dB and the contact resistance on the Ag/Hg-1212 and Ag/transition pin interfaces
may contribute even more. The large drop in IL below Tc is an indication that most of
45
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
18
16-
—
D-1
— o — D-2
14 -
CD
5m m
-0.15mm
108
-
6
-
—A g
c o n ta c t
4mm
10m m
80
100
11 0
120
130
140
150
T(K)
Fig. 3.7: Insertion loss (IL) of two Hg-1212 samples (D1 and D2) at 1 GHz is plotted as
function of temperature. It shows TcS of 121 K. and 119 K for samples Dl and D2
respectively.
the microwave power is transmitted through the Hg-1212 microstrip lines. The IL’s
o f D-l and D-2 decrease from 10 dB and 12 dB, respectively, at 125 K to less than 1
dB at 110 K. This means that approximately 75% of the input power is dissipated in
the Hg-1212 microstrip lines in the normal state. In addition, the curves in Figure 3.7
exhibit the real temperature behavior of Rs o f Hg-1212 above Tc suggesting that the
power transmission property of the structure is determined by the quality of
superconducting film. So indeed it is reasonable to assume that at 1 GHz, the power
leakage due to coupling between connector pins, which are separated by about 9 mm,
is negligible. The critical input power measured below Tc will then be a good estimate
o f the actual power penetrating the superconducting microstrip line and the power
output measured at the other end will be almost the power passing through the
sample.
46
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Since the power rating of the network analyzer was not adequate to exceed the Pc
o f the Hg-1212 microstrip lines, an RF amplifier (Amplifier Research 1W1000) was
inserted between the analyzer and the sample to magnify Pa (Fig. 3.6).
The input
power was fed as a long pulse (pulse width = 10 s) to the sample at a fixed
temperature and the Poul was measured at a resolution of 0.05 dBm. Although
microwave devices need to maintain their performance during continuous operation,
the 10-second time interval is generally much longer than the time scale for IL
degradation to occur at P> Pc. The Pout was then recorded as a function of Pa.
Generally P0Ut was very stable during the pulse period at low Pa levels at which the
calculated IL=Pin-P0ut was consistent with what was given in Figure 3.7. This is
expected since the Insertion Loss o f the line is proportional to Rs, which is known to
have very weak power dependence at low microwave power levels [49]. As Pawas
further increased and exceeded a certain level that was defined as the Pc, Pout became
unstable during the pulse period. Although its initial value was the expected Pout for
the particular applied power and normal IL, Pout decreased moderately by 10% to
25% at the end of the pulse period. Inset of Fig. 3.8 is a typical plot of Pout measured
at the end of the pulse versus Pi„. It can be seen that P0Ut behaves almost linearly with
Pin
until Pc is reached. At this critical power level, P0Ut drops significantly during the
pulse interval.
It should be noticed that the Pc measured on these Hg-1212 microstrip lines is
much lower than the intrinsic Pc o f the material at which an instantaneous decrease o f
the Pout is generally expected. The observed slow decrease o f P0Ut in the Hg-1212
microstrip lines suggested a different dissipation mechanism, most probably due to
formation o f local hot spots nucleated at various defects o f the sample. This argument
is based on similar observations made on YBaiCuaCh (YBCO) devices [50] in which
it was found that the microscopic defects in the material develop into local hot spot
that ultimately lead to device failure. In these studies, two different mechanisms o f
heating were identified using a thermal imaging technique. The first heating occurs at
lower power levels. This heat generation is due to intrinsic dissipation associated with
Rs o f the film and sample heating is minimal in most cases. Since the microwave
47
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
current density is generally non-uniform in a device, this lower-power heating was
observed typically in film areas o f high current density. In the second mechanism,
which only takes place at higher power levels, nucleation of a local hot spot with
temperature above Tc occurs and provokes the final breakdown of the device. Note
that the location o f the hot spot is not necessarily on areas with high current density.
Rather, various material defects in the devices serve as the nucleation centers for hot
spots. This suggests that high-power heating be most probably caused by material
defects, not the intrinsic limitation o f the material. In Hg-1212 microstrip lines, hot
spot formation was evident from the slow decrease of Pout at the Pc. Once these spots
are formed, heat diffuses into their neighborhood causing continuous power
absorption. The heat generation and diffusion become faster with increasing Pa. This
was confirmed experimentally at a further increased Pa and it was observed that the
Pout dropped faster and the overall sample temperature began to rise as detected by the
temperature sensor placed in the ground plane of the sample. This was very harsh to
the film. Actually, one sample was permanently damaged because o f the excessive
applied power.
Figure 3.8 summarizes the critical input powers o f D-l and D-2 at several
temperatures. At 80 K, sample D-l has a Pc close to 30 dBm and D-2, -24 dBm. In
fact, most Hg-1212 samples, which have been measured have Pc’s in the range of 20
dBm to 30 dBm although some have lower Tc values close to 110 K. At 100 K, D-l
exhibits the highest Pc o f 23 dBm.
This is in fact the best result that has been
obtained at 100 K for Hg-1212 microstrip transmission lines. At 110 K, the samples
still have moderate power handling capability as indicated by their Pc above 15 dBm.
Figure 3.9 compares the power handling capability of Hg-1212 with that of
YBCO and Tl-2212 patterned into the same type of microstrip lines. The YBCO films
have Tc - 90K and the Tl-2212 films, 105-108 K.
Both have Jc greater than
IMA/cm2 at 77 K and self-field. Their microwave properties are comparable to the
best reported in the literature [35]. On the real temperature scale (inset o f Fig. 3.9),
the three systems have comparable Pc’s in the range o f 27-30 dBm at temperatures
close to 77 K. At higher temperatures, Hg-1212 shows higher Pc due to its higher Tc.
48
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
30-
—• — D-1
—o —D-2
25-
l * 20'
<9
22 r
a-°is10-
75
80
85
90
95
100
105
110
115
120
T(K)
Fig. 3.8: Critical input power of two Hg-1212 samples (D1 and D2) at 1 GHz is plotted as a
function of temperature. Inset: typical behavior of Pout measured at the end of the pulse vs ?m.
30
25
- • - Hg-1212
-A-Tl-2212
-▼-YBCO
r
■O
o 15
0.
10
15 -
IQ100
5
0.6
110
120
T(K)____________
0.7
0.8
0.9
1.0
T/Tc
Fig. 3.9: Comparison of Pc for Hg-1212, Tl-2212 and YBCO films when plotted against
reduced temperature. Inset: Pc of three superconductors plotted against real temperature.
49
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
At 100 K, D-l exhibits the highest Pc o f 23 dBm, comparable to that o f Tl-2212 films
at 92K. and YBCO at 82 K. When Pc is plotted against the reduced temperature (T/Tc),
the curves for YBCO and Tl-2212 samples coincide nicely, suggesting equivalent
sample quality in terms of material defects and structural imperfection. The Pc vs.
T/Tc curve for Hg-1212, however, shifted downwards from that of the YBCO and Tl2212 samples, indicating the necessity for sample improvement in Hg-1212. Based on
the hot spot model, the power handling capability could be primarily limited by
heating effects near local defects, which causes device performance degradation
before reaching their intrinsic power limit. The results here suggest that a higher Pc
may be obtainable for Hg-1212 when the film is optimized. One possibility is to
improve the morphology o f Hg-1212 films. SEM studies showed that Hg-1212 films
have rougher surfaces than YBCO and Tl-2212 films have. This could result in higher
microwave losses. This argument is supported by the observation of a relatively
higher Rs for Hg-1212 at 77 K. So far, the lowest reported R* at 77 K and 10 GHz is
around 0.1 m fi for YBCO and Tl-2212, whereas Rs ~ 0.25 mQ for Hg-1212.
3.3 Limitations of traditional techniques
Microwave characterization o f HTS
superconductors using traditional
techniques is a sensitive way o f probing sample properties because it is based on the
collective response o f the sample to the microwave fields. However, measurement
results reflect only the average property of the entire sample, which could be
influenced by small fraction o f impurities or by the spatial variation o f the physical
property being investigated. Another limitation of traditional technique is the
generation of large screening currents, especially near the edges, and in the case of
superconductors, the subsequent admission of RF magnetic flux into the sample. It is
therefore desirable to develop new techniques that can break the limitations of
traditional methods. One promising technique is the near-field scanning microwave
microscopy (NSMM), which will be discussed in the next chapter.
50
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 4
Near-field scanning microwave
microprobe
4.1 Principle of operation
Any instrument that is based on propagation o f electromagnetic fields has a
classical limit in spatial resolution of X/2 over distances greater than wavelength (X).
This limit, which is called Abbe barrier, is a spatial version o f the sampling theorem,
which states that in order to fully recover a signal, one has to retrieve all o f its spatial
frequency components. However, the spatial frequencies higher than l/X, which is
known as evanescent waves, decay exponentially. In 1928, Synge proposed a
technique that surpasses the Abbe barrier [SI]. His idea was to closely scan a point­
like field source over an object so that the evanescent field is still strong enough to
interact with the object. Such field, known as near field, is still substantial enough for
electromagnetic excitation o f a sample. The interaction between the near field and the
sample has a perturbative effect on the probe. In resonant type probes, such effects
are a shift in the resonant frequency (fa) and a decrease in the quality factor (Q). Since
this perturbation relates directly to the surface properties o f a sample, measurement of
these quantities yields the information about the microwave property of the sample.
51
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.1.1 Probe designs
Near-field scanning microwave microscopy (NSMM) was first demonstrated
by Soohoo in 1962 [52]. Ten years later Ash andNicholls used it in a test rig to show
its imaging capability [53]. They obtained 0.5 mm spatial resolution on a conducting
sample. One popular design of an NSMM probe employs a metallic tip. It acts as
microwave field emitter and receiver. Immediately under the tip, this field is confined
to a very small spot that is approximately equal the tip diameter. In the near-field
regime, the spatial resolution is no longer determined by the wavelength of the
microwave, rather by the tip size. In Figure 4.1, three different techniques of
implementing a metallic tip are presented. Azar et al. [54] used microstrip and
stripline resonator design terminated by a sharpened wire tip to improve the spatial
resolution (Fig. 4.1(a)). The resonator is capacitively coupled to a short feedline using
a three-fingered interdigitated capacitor. The feed line is connected to a three-port
circulator, which circulates the signal from the microwave source to the resonator and
directs the reflected wave to a crystal detector. In the work o f Takeuchi et al. [55] the
resonating element is enclosed in a cylindrical shielding and the sharpened tip that is
connected to the center conductor, protrudes out from the aperture of the shielding
plate (Fig. 4.1(b)). Microwave signal is coupled in and out o f the resonator via the
coupling loops. Another design o f a near-field microwave probe was demonstrated by
Steinhauer et al. [56]. They used a commercial coaxial transmission line, which they
configured as a half-wavelength resonant line (Fig. 4.1(c)). By using a directional
coupler, microwaves from the source are fed to the probe via a decoupler and the
reflected signal is guided to a microwave diode detector. Aside from using metallic
tip to confine microwave fields to sub-wavelength resolution, apertures and slits are
also feasible. However, this configuration requires higher operating frequency, which
also increases the cost o f electronic equipment needed to operate the system. An
example o f this design was developed by Lann et al. [57]. They used
cylindrical waveguide terminated by 90 GHz
52
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
a W-band
Coupling
capacitor
cavity
coupling gap
w re (jp
m
source
substrate
detector
sample
sample
Fig. 4.1: Different techniques of implementing a metallic tip in scanning near-field
microwave probe, (a) Microstrip resonator terminated by a wire tip. (b) Quarter-wavelength
resonator shielded by a cavity, (c) Resonant coaxial transmission line.
asBuita-aufia
Fig. 4.2: Microwave probe design employing a slit to confine emitted fields at sub­
wavelength resolution.
53
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
transmitting/receiving resonant slit antenna (Fig. 4.2). This antenna is responsible for
localizing the fields and receiving the reflected signal from sample.
4.1.2 Signal measurement
The microwave near-field probe is similar to a sensor in a way that it needs an
electrical signal as an input and generates an output, which varies according to the
microwave properties o f the sample under it. In one-port resonant probes (i.e. Fig.
4.1a, 4.1c and 4.2) the output signal to be characterized is the reflected power. For
probes with two signal couplings (Fig. 4.1b), one is always for the microwave input
and the other is for the output. With resonant configurations, measurement can be
made either in frequency or in time domain. In the frequency domain, the input to the
probe is a frequency sweep containing one of the resonant frequencies. One can then
measure the shift in
the decrease in Q, and the change in reflection coefficient (T)
in the output signal when a sample is placed underneath the tip. The other mode of
measurement is the time domain. In this mode, the input signal is fixed to f 0 or very
close to it and the change in the power level o f the output signal is the one being
measured. The advantage o f frequency domain measurements is that it yields more
information. The shift in f 0 gives the change in the impedance of the probe-sample
system. f 0 is observed to be more sensitive to the tip-to-sample distance (TSD) than
the microwave sheet resistance (Rx) o f a conducting sample. The change in Q and T
are directly related to the microwave absorption of the sample. In time domain
measurements, where input frequency is fixed to f 0, the change in the output power
does not necessarily indicate change in real microwave dissipation. If there is a slight
change in f 0 due to some TSD variation, it will cause significant change in the output
power. Thus, interpretation o f data is not straightforward. The only advantage o f the
time domain measurement over frequency domain is its fast response. The power
level of the output signal can be monitored in real-time. For frequency domain
measurements, speed is limited by the sweep times of the microwave sources. Data is
54
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
obtained after one frequency sweep, which has a typical period of 10 ms to 1000 ms,
depending on the speed o f the microwave sweeper.
4.2 Open-end coaxial resonant probe
The microwave probe that has been developed in this work is classified as a
quarter-wavelength open-end resonator. This is a one-dimensional resonator based on
a coaxial transmission line structure. The length o f the transmission line determines
the operating frequency o f the fundamental mode and its higher harmonics. For a
resonator whose length is S.5 cm (X = 22 cm), the calculated fundamental resonant
frequency in free space (f0 = c/A.) is 1.36 GHz. However, the actual f a deviates a little
bit from this value because o f the different permittivity o f the dielectric material in
the coaxial cable.
Open-end resonator has only one feed line that capacitively
couples the resonator to the external electronics. This means that only one port
measurement can be performed. Since the coaxial resonator is a simple quarterwavelength resonator with open termination, it may be modeled as a series LCR
lumped circuit with complex input impedance Z m= R + 2j'L(co - to0) and Q = cOoL/R
[58]. At resonance (f0 = a sjln ), the input impedance becomes real and it is equal to
R. Thus, it is necessary to set R equal to the characteristic impedance o f the signal
source to get maximum power transfer (or minimum reflected power) at resonance.
This is necessary for maximizing the energy stored in the resonator. A plot o f the
reflected power from the probe versus frequency can be described as an inverted
Lorentzian curve, where the minimum reflected power occurs at f 0 (see Fig. 4.5). The
unloaded Q (Qu) of the probe can be determined using the standard formula for oneport measurement. That is, Qu = Ql(1+P), where Qi.(loaded Q) = //H P B W and the
coupling coefficient is
(3 = (1-|F| )/(l+|r|). Ql and P are
determined experimentally
using a vector network analyzer. HPBW is the half-power bandwidth and T is the
reflection coefficient at resonance. The actual implementation of this microwave
probe was carried out using commercially available components. This made the
assembly procedure and optimization easy and affordable. The probe has only three
55
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
major components: a semi-rigid coaxial cable, a copper block and an SMA connector.
Figure 4.3 shows a photograph of the constructed probe. The copper block is used to
secure the SMA and coaxial cable together. A hole is drilled through the copper block
so that the cable can be inserted. The pin of the SMA and the center conductor o f the
cable should be on the same axis. The center conductor of the coaxial cable serves as
the resonator. Because of its modular design, different coaxial cables with different
center conductors can be used. Moreover, a center conductor can be pulled out of a
cable and be replaced with other conducting wire, or it may be tapered to obtain better
spatial resolution. This adds flexibility to the probe because its resolution can be
Tunable gap for
optimizing sensitivity
Replaceable
coaxial probe
Fig. 4.3: Design and implementation of the microwave probe based on a quarter wavelength
open-end coaxial resonator.
56
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
adjusted for different applications. The cable is allowed to slide within the drilled
hole so that the gap between the SMA pin and the center conductor can be adjusted.
This gap serves as the capacitance coupling o f the resonator. With this gap adjustment
feature, capacitance can be tuned experimentally to achieve critical coupling. This
results in high Q values and good sensitivity. Tuning is performed with a network
analyzer. Once the cable position, which gives the highest Q, is found the set screw
on the Cu is tightened. For Cu cables, with 0.141” outer diameter, Qu as high as 8000
can be achieved at 1.7 GHz. The compact and rigid design o f the probe makes it
suitable for integration with vacuum systems and cryogenic enclosures. This is an
important feature because ultimately, it is desirable to implement the probe for
characterization of superconducting properties of HTS. It is important to mention that
this specific design of near-field microwave probe is original.
The complex input impedance (Zin) o f the open-end resonator was analyzed using
a vector network analyzer (HP 8722). Figure 4.4 is a plot o f the measured real and
imaginary parts, which correspond to the resistance (R) and reactance (X)
respectively, versus frequency. This plot includes the fundamental resonant frequency
o f the resonator. The black curve represents the resistance and its axis is labeled R.
The red curve represents the reactance and its axis is labeled X. Observe that R and X
oscillate with frequency. This behavior is not expected in the simple series LCR
model because in that model the effect o f coupling the resonator to the signal source
is not considered. In the actual probe, the input impedance that is seen by the analyzer
includes the capacitive coupling, which is frequency dependent. For the probe to be a
resonator, Zi„ = Zo and X = 0 at certain frequency as described in the series LCR
model. In Figure 4.4, notice that this condition exists at / = 1.735 GHz. At that
frequency, R = 50 Q and X = 0 corresponding to Zin = Z„. Therefore, the probe has a
resonant frequency at 1.735 GHz. At this frequency, the reflection coefficient,
expressed mathematically as,
_
Zm- Z o
z
tn +
z
o
57
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(3*l >
250
-100
200
-
-50
150^
cn
100 -
--50
50-
--100
01.5
1.6
1.7 / .
1.8
1.9
2.0
/(GHz)
Fig. 4.4: Plot of the real (black) and imaginary (red) terms in the complex input impedance
(Zin) of the open-end resonator as a function of frequency as measured by a vector network
analyzer.
goes to a minimum as expected. This agrees very well with the actual measurement o f
T(/) (see Fig. 4.5).
When a conducting sample is placed in the vicinity o f the resonator tip, the E
fields are directed perpendicular to the sample surface (Fig. 4.5(a)). Thus, if p is the
microwave resistivity along the c-axis o f the sample and t is the thickness then the
probe can measure Rx, which is the ratio between p and U The sample interacts with
the probe via the effective capacitance Cs arising from the gap between the probe tip
and the sample surface (Fig. 4.5(b)). As the tip approaches the sample, Cs increases
and causes a decrease in the resonant frequency o f the probe. From the standard
58
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. 4.5: (a) Electric field configuration at the tip of the probe when a sample is placed very
close to it. (b) Interaction between tip and sample via effective capacitance, C,.
transmission line theory, this is equivalent to increasing the length of the line [59].
Figure 4.6 is a plot showing the reduction off Qwhen a copper film was placed under
the probe. Effectively, this adds to the dissipation o f the resonator and it is manifested
by a drop in Q. Going back to Figure 4.6, it can be noticed that the resonance curve
became broader as compared to the original curve. This indicates a drop in Q. In this
specific case o f Cu sample, the decrease in Q is from 8000 to 1800. The sample
property that governs this power loss is the Rx. In general, sample interaction with the
probe tip is not trivial. Measured response of the probe (i.e. A f 0 and AQ) is strongly
dependent on the microwave properties o f the sample as well as on the TSD and
cross-sectional area o f the tip. Care must be taken in analyzing the data. The choice of
tip diameter is crucial because it influences sensitivity and resolution. A bigger tip
will yield higher sensitivity because Cs is proportional to the cross-sectional area but
will degrade resolution since the field will emanate from a larger source and interacts
with larger sample area. In principle, there is always a trade off between resolution
and sensitivity. The choice is generally determined by the nature of the sample and
the desired information.
59
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
-10
£
-
-
15-
-
20-
-
25-
-30
—• — no sample
—o— with Copper j
-
-35
1.725
1.730
1.735
1.740
1.745
f( GHz)
Fig. 4.6: Change in reflection property of the resonator when a conducting sample is placed
under the probe tip with distance of 0.5 mm. This measurement was performed at room
temperature and the corresponding change in unloaded Q was from 8400 (no sample) to 1800
(with sample).
4.3 Experimental setup for mapping
To perform mapping it is necessary for the probe to move relative to the sample.
To do this, three motorized stages (Newport) were assembled in an XYZ fashion.
Figure 4.7 shows the diagram of the XYZ assembly. The X and Y are identical
horizontal linear stages with a maximum range o f 25 mm and minimum step o f 0.07
pm. X is mounted on top o f Y in such a way that Y can carry it and move the whole
X-stage in the y-direction while it can move itself independently in the x-direction.
The Z is a vertical linear stage with a maximum range o f 4 mm and minimum step of
0.01 pm. It carries Y (thus X, too) and moves it in the z-direction. As a result, the top­
most stage can be moved in three dimensions. The XYZ assembly is secured on a
home-built platform with fine leveling adjustment The platform is then mounted on
an optical table to minimize vibration. Fine leveling adjustment platform was
60
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
necessary because the optical table is not self-leveling. For mapping without tip to
sample distance feedback, it is very important that sample is always leveled. A
LAB VIEW™ program was developed to fully control the three stages during
scanning and to synchronize data acquisition. The microwave probe was fixed on top
o f the sample stage, which has a coarse adjustment mechanism. On the X stage, a
spring-loaded holder for a standard microscope slide was added. For this system, the
sample is mounted on the X stage on a microscope slide. This manner is more
convenient because sample mounting can be made separately using conventional
optical microscopes or other working tables. In addition, if measurement is not
finished the microscope slide containing the sample can be removed easily and can be
stored in a dry box to avoid contamination or degradation.
Fig. 4.7: Configuration of the three motorized stages to move the top-most stage in three
dimensions.
61
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.4 Sensitivity and spatial resolution
NSMM probe has two figures of merit, namely: sensitivity and spatial resolution.
In principle, both can be measured experimentally but their interpretation could be
misleading. For example, it is well known that spatial resolution o f near-field probes
is determined by the tip diameter. Consider two patterns, A and B, made of different
materials that have feature size comparable to the tip diameter. Pattern A has higher
sheet resistance than pattern B. It is probable that only pattern A could be resolved by
the probe because its sensitivity is not enough to sense pattern B. For the sample
whose pattern is not resolved the expected resolution of the probe is not applicable.
Also recall that sensitivity depends on the TSD. The smaller is the TSD, the better is
the sensitivity. However, it is not always true that a sample can be placed as close as
possible to the tip during scanning. Some samples can distort the resonance so much
when they are very close that measurement becomes extremely difficult. Another
complication when talking about sensitivity is that it also depends on sheet resistance,
dielectric permittivity and surface impedance. As mentioned earlier, care must be
taken in analyzing the measured response o f the probe because it is not a direct
measure o f a specific sample property.
In this dissertation, the objective was the development of a prototype NSMM, so
absolute and systematic calibration o f sensitivity and spatial resolution for different
samples was not carried out. This procedure will be done in the future when system
optimization is performed. At this stage, the only samples o f interest are HTS films.
Resolving HTS film on a dielectric background (substrate) was the only test
performed to evaluate the sensitivity and resolution. Figure 4.8 shows a linescan of
the microwave probe with a tapered tip (~ 15 pm) on patterned HTS line made of
ThBa^CaCuiOfr+s (Tl-2212) whose width is 150 pm. The thickness o f this film is
about 0.3 pm. It can be observed that the response of the probe changed suddenly as
it crosses the HTS material. This indicates that it can recognize the pattern because it
has a different microwave resonance than the substrate.
62
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Tl-2212 film
(linewidth:150 um)
GAP
3
CQ
ha
u
o.
Q
•o
U
U
u
cu
a:
u>
CB
*
2u
i
150
X(um)
300
450
Fig. 4.8: Linescan of a tapered microwave probe with tip diameter around 15 pm on a
patterned superconducting Tl-2212 line whose width is 150 pm.
Another evaluation procedure of the near-field probe was carried out on a
patterned sample. This probe was not tapered and the tip diameter was 0.5 mm. The
sample used was a piece (around 10 mm x 10 mm) of standard printed circuit board
(PCB). PCB is an insulator with a Copper layer. This layer can be etched out with
Ferric Chloride (FeCh) solution. If a certain pattern is protected from the FeCh, it
will remain on the insulating board during etching. Thus, a desired pattern can be
defined. For this particular sample, “KU” letters were patterned on the PCB. Before
scanning, the probe was placed at four comers of the sample with approximate tip to
63
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
sample distance of O.S mm and then f Qwas checked. Note that iff 0 measurements at
four comers are not equal, it means that the sample is not well-leveled so adjustment
o f the fine leveling platform is necessary. The near-field probe was scanned over the
sample at constant height. Microwave input to the probe was fixed at single value,
which was equal to f Qat the comers. As the sample moved with respect to the probe,
the magnitude of the reflected power (Su) was acquired continuously by a voltmeter.
These values were recorded in a two-dimensional array form so that each one would
correspond to certain x,y position on the sample. The array was then plotted using
Origin
data analysis software. Figure 4.9 shows the photograph o f the sample and
corresponding microwave image obtained by the near-field probe. It is clearly seen
that the image generated by the near-field scanning microwave probe is consistent
Y(mm)
with the actual pattern.
(b)
Fig. 4.9: (a) Photograph of the sample with “KLT’ letters made of copper. The letters were
patterned on a printed circuit board (PCB). (c) Microwave image of the same letters as
generated by the near-field scanning probe.
64
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 5
Applications of near-field scanning
microwave microscope
After developing the near-field scanning microwave probe, it was utilized in
some important characterization o f HTS films. This chapter discusses the actual
applications at room and cryogenic temperatures where this new version of a near­
field microwave probe was used.
5.1 Uniformity of large area HTS film
Device fabrication requires deposition of HTS material on large-area substrates
ranging from 0.5” x 0.5” to 2” x 2”. A critical requirement is the homogeneity of
physical properties (i.e. film thickness, microwave conductivity, etc.) over the entire
sample surface. Non-uniformities on a microwave device can cause undesirable
reflections that will affect the actual performance o f the device. In industry, films
should be homogeneous over an entire wafer so that devices built on different parts of
the wafer behave as designed. This improves the efficiency o f production. In general,
uniformity of microwave properties on different length scales (from wafer level to
sub-micron level) is a major concern in HTS microwave applications. Many factors
contribute to the non-uniformity of HTS films. These include impurities, growth
defects, grain boundaries, surface microstructure, etc. The dimensions o f these non­
uniformities range from millimeters to few tens o f nanometers. Different non-
65
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
uniformities may dominate the microwave properties o f a sample in different
frequency and power regimes. It is necessary to correlate them with the overall
microwave performance of the device. A unique technique for such correlation is the
near-field scanning microwave microscopy (NSMM).
Figure 5.1 shows a microwave image o f an un-pattemed Tljl^CaCuiOg (Tl2212) film deposited on a single crystal LaAlCh substrate (10 mm x 10 mm). This
image was generated using a probe o f silver-plated Cu center conductor whose
diameter is 0.5 mm.
Reflected Power
high
low
Large area film
lO m m x 10 mm
Scanned area
8mm x 8mm
8mm
0
Fig. 5.1: Image scanned by the near-field scanning microwave probe showing the nonuniformity of microwave property of a large-area Tl-2212 HTS film on LaA103 substrate.
Resolution is 0.5 mm.
66
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The tip to sample distance (TSD) was around 0.4 mm. Before imaging, necessary
adjustment was performed in the XYZ platform to make it as level as possible. Only
the inner area (8 mm x 8 mm) o f the sample was scanned to eliminate the edge effects
of the substrate. Microwave measurements were done in time domain. The frequency
of the input signal to the probe was fixed at 1.7 GHz. This was the resonant frequency
(/o) when TSD was 0.4 mm.
It can be noticed that near the sides, the response o f the probe is different relative
to the center region of the sample. The high power reflection on these areas suggests
that film thickness is smaller. The f Q shift at these regions also indicates similar
interpretation. This is a common problem in large-area film deposition. The physical
properties, such as conductivity and morphology near the edges tend to be worse
relative to the middle. With this measurement capability, it is convenient to routinely
evaluate sample uniformity after fabrication. The result can help to improve
fabrication process. It can also screen out samples that have the potential to fail in
further device characterization. If only a partial region of the film is to be used, the
image generated by the microwave probe could determine which part o f the film is
the best. For example, a transmission line resonator on a LaAlC>3 substrate (er= 24.5)
with fundamental resonance at 5 GHz has a length o f about 7.8 mm and width of 0.5
mm [30]. This device could still be patterned on the center o f the sample shown in
Fig. 5.1.
5.2 Local measurement of superconducting properties
Before utilizing the microwave probe in low temperature scanning o f HTS
samples, it is necessary to assess its single-pixel capability in characterization of
superconducting properties such as the superconducting transition temperature (Tc)
and sheet resistance (Rx). Moreover, it is desirable to characterize local sample
properties because extrinsic effects are minimized on the measurements. Recall that
with traditional techniques, results are from weighted average over large areas of the
sample. If there are scattered defects, they can dominate the measured values. The use
67
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
of the microwave probe for local characterization is non-contact and non-destructive.
It is suitable for routine quality check of samples. Moreover, it does not require large
and perfectly cut samples. As long as there is a circular region on the sample, which
has a diameter very close to the tip diameter, then effective measurement can be done.
5.2.1 Integration of the near-field probe with cryogenic enclosure
A cryogenic vacuum enclosure was designed and constructed to accommodate
the microwave probe and the sample stage. Figure 5.2 illustrates the basic structure o f
this enclosure. The probe and the stage are secured on a brass canister, which is
connected to a feed through box via a long stainless steel tubing (~ 5 feet long). The
probe has good thermal connection with the brass canister while the stage is thermally
isolated. When the canister is pumped to a good vacuum (-10'4 Torr) and immersed
into liquid nitrogen (77 K), the probe temperature could be maintained very close to
77 K. On the other hand, the stage could be heated up independently without affecting
the probe. It has a heater resistor and a temperature sensor for controlling its
temperature. Its position relative to the probe can be adjusted mechanically to vary tip
to sample distance. The stainless steel tube has poor thermal conductivity that
minimizes the heat transport from the feedthrough box (295 K) to the brass canister
(77 K). This is necessary for maintaining a large temperature difference between
them for long period o f time. All the needed electrical wires and microwave cable
from the canister go to the feedthrough box via the stainless steel tube. The box
serves as a transition from the components inside the enclosure to the external
electronic equipment (i.e. temperature controller, microwave detector and source
etc.).
68
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Feed
through box
Stainless
steel tube
'- t
lT
_L
-Brass
canister
NSMM
probe Sample_
stage
Canister diameter = 2.5 inches
Fig. 5.2: Sketch of the experimental cryogenic vacuum enclosure used for single spot
characterization of microwave properties of HTS films.
69
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5.2.2 Addition of mutual inductance coil to NSMM probe
During the initial evaluation of the single-pixel measurement setup, calibration
was necessary to check the results obtained by the NSMM probe. For example, to
verify whether the measured Tc is reliable, the sample has to be characterized by
another method such as an electrical transport measurement. To circumvent this
difficulty, a pair of coil for mutual inductance measurement was added to the
experimental setup. One coil was wound on the coaxial NSMM probe to serve as the
primary coil and the other one was wound on the sample stage as the pick up coil.
Figure 5.3 illustrates this modification. A lock-in amplifier with an internal reference
is used for this inductive measurement. The drive coil is connected to the internal
Fig. 5.3: Sketch illustrating the addition of mutual inductance coil to the NSMM probe and
stage.
70
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
source (4 KHz sine wave) via a series resistor (~2000 Q) and the pick up coil is
connected to the lock-in amplifier input. When an HTS sample is above Tc, there is
no exclusion o f magnetic field so the in-phase signal detected by the lock-in amplifier
is strong. At T ~ Tc, shielding current is generated within the sample causing the inphase signal in the pick-up coil to decrease. With this setup, the two different
measurements (AC and microwave) can be carried out simultaneously. This
combination of inductive and near-field techniques in a single probe is a simple form
o f the dual probe, which will be discussed in the later chapters of this dissertation.
5.2.3 Measurement of local Tc
Determination o f Tc is perhaps the most important and fundamental
characterization on superconductors. In chapter 1, the traditional techniques for
measuring Tc have been discussed. They give the collective response of the entire
sample. However, using the NSMM probe implemented in a cryogenic vacuum
enclosure (as discussed in the previous sections), the Tc o f a local spot can be
measured. Here, a sample is mounted on the stage using a thin layer of vacuum
grease. If a different tip-to-sample distance is desired, the stage can be moved
vertically relative to the NSMM probe. For this particular experiment, tip to sample
distance is around 0.4 mm. After evacuating the enclosure (~ KT4 Torr), it is
immersed in liquid nitrogen. A sample is cooled down to the lowest attainable
temperature. Since the stage has poor thermal connection to the cryogen, sample
cooling is slow and it usually stops around 81 K to 83 K, depending on the vacuum.
Once it attains its lowest temperature, it is expected that the NSMM probe has already
stabilized near 77 K because o f its good thermal contact with the brass canister. The
sample temperature is then increased to few degrees above its Tc using the heater in
the sample stage. Once stable, heater current is removed and the sample is allowed to
cool down slowly while measuring either the in-phase voltage from the pick up coil
or the reflected power from the NSMM probe.
71
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.4 is a plot against temperature of the voltage from the pick up coil and
the Su at resonance o f the NSMM probe. Sn is defined mathematically as
I0/og(r),
where T is the reflection coefficient. The maximum Sn is zero, which corresponds to
100 % reflection of the incident microwave power. That data was taken on a
HgBaiCaCuiOe+s (Hg-1212) film. For this particular sample, Su as the sample cools
down continuously was fluctuating. It was not clear why for some samples the
behavior of Su as they cool down was smooth and for others, like this particular Hg1212 sample, Su was fluctuating. Hence, the sample was just set at several selected
temperatures. At each temperature several trials were made in measuring Su, then the
average was taken. From the plot, it can be observed that results from the two
measurement techniques coincide nicely. The solid squares represent the plot of the
reflected power versus temperature. The sudden drop near 120 K signifies the Tc o f
the sample. Above this temperature, the sample dissipates more power resulting in a
lower Su- The open circles represent the signal at the pick up coil. This signal has
increased abruptly near Tc indicating that the sample became transparent to the
incident AC magnetic field coming from the primary coil.
In the microwave measurement, the frequency domain was also used to examine
if there was a significant change in f 0. When the behavior of f , as a function of
temperature was investigated, it was found that when TSD is comparable or greater
than the tip diameter (0.5 mm), f 0 has very weak dependence on the changes of Rx.
But as TSD gets smaller, dependence on Rx becomes stronger. This agrees
qualitatively with the model that actual coupling o f the sample to the tip is via the
capacitance (Cs) between sample surface and the cross-sectional area of the tip and Rx
[54]. Moreover, it is the effective impedance that changes the overall response of the
NSMM probe. In practice, care must be observed in assuming that f 0 only depends on
Cs as implied by the transmission line theory. Figure 5.5 shows the data from local Tc
measurement of three different HTSs namely: Hg-1212, Tl-2212 and YBCO films.
The results for Tl-2212 (solid squares) and YBCO (solid circles) are plotted in the
inset. The Su(T) curves for Tl-2212 and YBCO were taken continuously as they
72
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
-
5.9 4 -
-ao2
-5.96-5.98-
2
-6 .00 -
-
0.06
-6 .0 2 -
-
0.06
-aio i
. - 6 . 04 -
-
6.0 6 -
-
6.0 6 -
~
-6 . 10 -6 . 1 2 - 6.14
100
106
110
115
120
125
T(K)
Fig. 5.4: Plot of the pick up voltage (open circle) and microwave reflected power expressed
as Sii versus temperature. These data were taken from Hg-1212 HTS film.
-
0 .9 4 -
-
0 .9 6 -
-
0 .9 8
-
1.00
-
1.02
3-L
-
1.04
*3
-
1.06
-
1.08
-
1.10
— .
CO
•
H g -1 2 1 2
110
115
- 1.1 2 -
1.14
85
90
95
100
105
120
T(K)
Fig. 5.5: Data obtained from local Tc measurement of three different superconductors. Solid
circles correspond to Hg-1212. The solid squares and circles in the inset correspond to Tl2212 and YBCO respectively.
73
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
cooled down. It can be noticed that the data for Tl-2212 is very smooth but for the
YBCO, there are slight fluctuations. From the microwave data, the Tc are 92 K, 98 K
and 115 K for YBCO, Tl-2212 and Hg-1212 respectively. They agree very well with
the Tc values obtained from the inductive technique. They are also comparable to the
values reported for these materials.
In this measurement, the reflected power is related to the sheet resistance Rx of
the HTS sample. This is not the surface resistance R* defined in chapter 2 and
commonly measured on HTS films using the traditional techniques. Microwave
resistivity o f HTS material is highly anisotropic. Rx is along the c-axis and R* is in the
a-b plane of the HTS film. Rs will only have a direct relation to Rx if the microwave
resitivity along c-axis (pc) is equal to the microwave resistivity along a-b plane (pab).
Note that in DC measurement, pc/pab has been determined to be around 50 for YBCO
and few thousands for Tl-2212. It is not clear whether this ratio remains the same at
microwave frequencies. The relationship between Rx and Rs needs further
investigation.
5.2.4 Single-pixel measurement of power handling capability
As discussed in chapter 2, microwave power handling capability o f HTS is also
an important issue fundamentally and technologically. Due to nonlinear effects at
moderate power levels, these materials are limited to low power applications. Most
current techniques used for studying nonlinear microwave properties are global in
nature. They are usually dominated by the extrinsic properties and it is difficult to
observe the intrinsic power limitation. In this single-pixel measurement employing
the NSMM probe, a sample is set at a fixed temperature, then the resonance curve is
obtained using the vector network analyzer. The input power to the NSMM probe is
increased continuously while monitoring the resonance curve. The increase in the
input power results in an increase in the intensity o f the near field interacting with the
sample.
74
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The sample used in this particular measurement was Hg-1212. Its temperature
was kept at 80 K and the input power to the NSMM probe was raised continuously
while monitoring the resonance curve. The tip-to-sample distance was still 0.4 mm.
The result is presented in Figure 5.6. It can be observed that the resonance curve does
not change at input powers below 35 dBm (blue curve). At 35 dBm, the resonance
curve (black) shows that f 0 and Su(f0) have decreased, indicating that the critical
power (Pc) has been reached. To be sure that this was sample related, a power sweep
was performed on the NSMM probe without a sample. The resonance curve stayed
the same up to the highest power available (40 dBm). The stable behavior of the
resonance curve below Pc is consistent with the behavior o f Rs in the linear regime. It
suggests that microwave property of the sample remains the same in this range of
-
1-
-
2-
-3-4-
£ -5-
6-
-7-
BekJw35dBm
36 dBm
Altar few sec.
8-
1.72
1.73
1.74
1.75
1.76
Frequency (GHz)
Fig. 5.6: Behavior o f the resonance curve as input power to the NSMM probe is increased.
The sample used in this experiment was Hg-1212 at 80 K and tip-to-sample distance was 0.4
mm.
75
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
field intensity. However, at 35 dBm the reflected power drops, indicating that the
microwave dissipation in the sample has increased. It is a signature that the sample
has entered the nonlinear regime. The distortion of the peak at the curves for 35 dBm
(black and red) is an indication of sample heating. In fact, it can be noticed that when
the critical input power was kept for few more seconds, Sn and f 0 further dropped.
Note that when heating occurs in HTS samples, the rate of power absorption
continues to rise. This behavior has also been observed in power handling
measurements of Hg-1212 microstrip lines discussed in chapter 3. The decrease in Su
at Pc is related to the power dependence o f the Rx. The downward shift in f 0, which
also means lowering of Q (Q = /j/HPBW ), is related to the power dependence of
surface reactance. This is in good agreement with the observation made by Oates et
al. [38] on YBCO stripline resonators. They found out that both the real part (Rs) and
the imaginary part (Xs) exhibit strong power dependence above a critical input power.
Based on this qualitative agreement, it is fair to claim that the measurement technique
developed and presented in this dissertation could be reliable once proper calibration
is performed. In addition, quantitative relationship between the surface impedance
along c-axis and the surface impedance o f the a-b plane can be investigated by using
this setup together with the traditional techniques. This would be very useful in
understanding some fundamental properties o f HTS at microwave frequencies.
Using this newly developed technique, Pc for the three different samples (Hg1212, Tl-2212 and YBCO) used in section 5.2.3 was measured and results are shown
in Figure 5.7. Note that in this comparative measurement, the tip-to-sample distance
was slightly reduced from 0.4 mm to about 0.3 mm to obtain higher field intensity
with lower input power. This is the reason why the Pc o f Hg-1212 at 80 K shown in
Figure 5.7 (28 dBm) is lower than the value obtained from the first measurement (35
dBm) as shown in Figure 5.6.
It can be observed that the Pc decreases as the temperature increases. This is the
general behavior o f any HTS sample including the temperature dependence o f critical
current density (Jc). The critical field needed to break paired electrons decreases as
temperature increases. For the YBCO and Tl-2212, Pc seems to drop linearly but
76
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
28
£
-
•
•
■
26-
Hg-1212
YBCO
TL-2212
24-
Q-
20-
1680
85
90
100
95
105
110
115
120
T(K)
Fig. 5.7: Summary- of the local power handling measurements performed on three different
superconductors namely: Hg-1212 (black circles), Tl-2212 (blue squares) and YBCO (red
circles). In this experiment, the tip-to-sample distance was approximately 0.3 mm.
not for the Hg-1212. This is an indication that the mechanism behind the Rx power
dependence for YBCO and Tl-2212 might be similar but could be different for Hg1212. This interpretation is supported by the results obtained from the power handling
measurements using microstrip transmission lines (Fig. 3.9), the Pc for YBCO and Tl2212 coincides when plotted against reduced temperature (T/Tc). This should be
investigated in more details.
Another interesting result is observed when Pc o f the three samples are compared
at the reduced temperature (T/Tc) = 0.85. This corresponds to 77 K, 82 K and 100 (C
for YBCO, Tl-2212 and Hg-1212 respectively. At these temperatures, their Pc
coincides nicely at 22 dBm. Although for YBCO, Pc at 77 K is just an extrapolation
assuming linear temperature dependence. One interpretation of this result is that the
77
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
measured Pc using the single pixel technique is fairly close to the intrinsic value. By
localizing the measurement, extrinsic effects are minimized. Thus, the power
dependence mechanism of Rx or even Rs (requires detailed investigation) for the three
samples could be similar but it is just difficult to observe using traditional techniques
because extrinsic effects vary from sample to sample. For example, going back to the
microstrip transmission line measurements (Fig. 3.9), the Pc curve for Hg-1212
plotted against T/Tc is below the curves for YBCO and Tl-2212 but they all have
similar behavior. This suggests that the extrinsic properties of the Hg-1212 microstrip
line have worse effects on its power handling capability as compared to that of YBCO
and Tl-2212. This is not surprising because unlike Hg-1212, fabrication of YBCO and
Tl-2212 is already a mature process and they have been optimized for a long time.
78
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 6
Development of a dual-channel near­
field scanning microwave and optical
microprobe
In the previous chapter, several important applications o f single channel near­
field scanning microwave probe to characterization of high-Tc superconductors were
demonstrated. However, a single probe is limited to characterization o f just one
sample property. In materials research it is often desirable to correlate different
physical properties at the microscopic scale in order to understand how they affect
each other. This can be achieved by integrating two SPM techniques into one probe.
By doing this, two different sample properties are obtained simultaneously. For
example, adding an optical channel to the microwave probe will allow it to measure
the optical response of a sample at the same time it measures the microwave
response. The added information provided by the optical channel can be used for
interpretation of the measured microwave properties or vice versa.
79
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6.1 Overview of near-field scanning optical microscopy
Near-field scanning optical microscopy or simply NSOM was developed in 1984,
a few years after the development of near-field scanning microwave microscopy
[60,61].’ It is considered an extension o f near-field scanning microwave microscopy
(NSMM) to the visible spectrum. The interaction o f visible light (instead of
microwave radiation) with a local spot on the sample is characterized. By scanning all
the spots representing a specific sample area, a two-dimensional image can be
generated. In NSOM, local characterization can be performed in different
configurations. Four of them are shown in Fig. 6.1 [62].
light source
aperture
light detector
(a)
sample
Fig. 6.1: Major configurations in near-field scanning optical microscopy (NSOM). They are
named as follows: (a) illumination; (b) collection; (c) oblique illumination; and (d) oblique
collection [62].
80
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The eye represents the light detector and the arrows symbolize the light rays. In
Figures 6.1(a) and 6.1(c), where light is applied locally, high spatial resolution is
achieved by funneling the light through a small aperture at the end of a sharp and
otherwise opaque probe. If the diameter o f this aperture is much less than the
wavelength of the light, the light is diffracted by the aperture. As a result, the light
will diverge as the far field propagates in space while its intensity drops
exponentially. However, if a sample is brought very close to the aperture (much less
than the tip diameter) such that it only interacts with the near field, then only a local
spot on the sample whose diameter is almost equal to the aperture diameter is
illuminated and interacts with the field, which is still strong. This improves spatial
resolution, which only depends on aperture size.
The probe with a small aperture can also be used to measure light on a local spot
of the sample. This is illustrated in Figures 6.1(b) and 6.1(d). In these configurations,
the aperture is like a very small active window to a detector. If it is brought very close
to the sample, it can only measure the light coming from a local spot whose diameter
is approximately equal to the aperture diameter. Thus, spatial resolution is also
determined by the aperture size.
The principle of NSOM is simple and straightforward but implementation was
not easy in practice at early stages. For example, making an aperture that could
confine light to the smallest spot possible was not trivial. In 1991, the most important
technological breakthrough in NSOM occurred. Betzig and co-workers [63] invented
the tapered fiber optic probe coated with a thin layer o f metal for NSOM applications.
With this structure, apertures having diameters o f few nanometers are achievable.
Moreover, the metal coating helps improve the funneling ability o f the fiber tip in the
illumination configuration.
6.2 Integration of fiber optic NSOM probe to NSMM probe
Designing the dual-channel near-field scanning optical and microwave probe
(dual probe) was the most critical stage in this work. The main principle o f the design
81
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
was to employ the same concepts used in the NSMM probe developed in chapter 3,
and to also utilize the metal coating o f the fiber optic NSOM probe to carry the
microwave signal. To realize this concept, the metal coated fiber optic tip has to
become the end part of coaxial resonator so that the fiber serves as the channel for
light while the coating serves as the channel for the microwave. The NSMM probe
was re-designed (while preserving its equivalent circuit model) to accommodate a
fiber optic in the center conductor without losing resonance. Figure 6.2 shows the
design of the dual probe. The original center conductor o f the coaxial cable is
removed and two stainless steel tubes (or better conducting tube for higher
sensitivity) are put in place of the original center conductor. The tube diameter must
be comparable to the original center conductor. The longer tube (6 cm) is the
resonator, while the shorter one is just a part of the feed line. Their inner diameter is
sufficient for insertion of a metal-coated optical fiber whose diameter is
approximately 125 pm. The gap between the long and short tube is the capacitive
coupling o f the resonator. This gap can be adjusted by sliding the longer tube with
respect to the shorter one, allowing the critical coupling capacitance for the highest
microwave quality factor (Q) to be achieved. Instead o f connecting the short tube to
the SMA pin directly, a silver microstrip transmission line, which was deposited on a
sapphire substrate using DC magnetron sputtering, is inserted in series. Its width is
critical and must be properly calculated to achieve characteristic impedance close to
50 Q. That is necessary for minimizing the generation of standing wave patterns
within it. The fiber optic cable is inserted at the open end of the resonator and
protrudes out from the short tube/microstrip junction. This approach prevents the
tapered tip o f the fiber from being damaged during insertion. The metal coated part of
the fiber, which is about 12 mm, measured from the tip aperture, is allowed to peek
out from the end of the longer tube. Conductive silver paste is used to secure the
electrical connection between the stainless steel tube and the coated fiber. Thus, the
metal-coated part o f the fiber optic becomes a part o f the resonator and its tapered tip
serves as the dual scanning probe. A slight decrease in the resonant frequency results
from the effective increase in the length o f the resonator.
82
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fiberoptic
microwave source
and detecror
ring-shape microwave tip
^ ^ .coptical
tic a l aperture
Cross-section of
dual probe tip
I9 0 A S
Microfftrip
Trsnsmimon
Line
^ *"1— Fiber opbc
___
Light from LASER
' Coaxial cable
Collection Mode
Configuration
Fig. 6.2: Design of the dual-channel near-field scanning optical and microwave probe. Inset:
exploded view o f the tip and the collection mode configuration and cross-sectional drawing
of the dual probe tip showing the multi concentric structure.
Fiber optic
Coaxial outer
conductor
Removable
conducting /
tube
I
Sliding outer
conductor
Microstrip
Transmission line
SMA connector for
the microwave input
/
Fig. 6.3: Photograph of the actual dual-channel near-field scanning optical and microwave
probe developed in this work.
83
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6.3 Tapering of the fiber optic tip
There are currently two methods employed in tapering fiber optic cables and both
are being used widely [64,65]’. One method is called the heating and pulling, in which
an optical fiber is pulled adiabatically during heating by either a CO 2 laser or a
tungsten filament. The other method is called acid etching. It is based on etching glass
fibers at the meniscus formed between hydrofluoric acid and an organic overlayer.
The first method is relatively fast but requires more expensive apparatus to perform.
On the other hand, acid etching is slow but only requires simple and inexpensive
accessories. In this work, the latter method was employed to taper the optical fibers
used in the development of the dual probe.
Etching takes place in a custom built Teflon container shown in Figure 6.4(a). It
contains 40 mL of hydrofluoric acid (50% concentration) covered with 20 mL o f
38
time (min)
(a)
(b)
Fig. 6.4: (a) Diagram of the designed Teflon container for etching fiber optic, (b) Qualitative
description of the tip diameter during etching. The critical time for getting the sharpest tip is
around 38 minutes.
84
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
silicon oil. An optical fiber, whose polymer coating has already been stripped by a
commercial fiber stripper, is inserted into a cut-off hole on the Teflon cover. The
optical fiber is secured tightly on the cover using magnetic clamping mechanism.
Once the cover is placed on the container, the protruding fiber is immersed in the acid
solution. The shape of the taper is formed by the acid meniscus at the liquid-liquidglass interface, which declines with time. Figure 6.4b represents an approximate
description o f the tip diameter during etching. The critical etching time is
approximately 38 minutes ±15 seconds. If the fiber is retrieved a few seconds earlier
from the etching container, a very thin overhang of excess material will be present. If
it is taken out at exactly the right time, a tip diameter around 100 nm or less can be
achieved. If it is taken out later than the critical time, the tip begins to get dull due to
the slow penetration of the acid to the oil, which further dissolves the fiber material.
6.4 Metal coating of tapered fiber optic tip
Metal coating o f the optical fiber tip was carried out using thermal evaporation.
Modification was made to a commercial evaporator (Denton 502A) to achieve
uniform coating on the cylindrical probe. A custom designed spool was constructed
to hold the optical fiber, of length as long as 3 m during cylindrical coating. The
filament for the metal source was placed at 90° with respect to the axis (vertical) of
rotation of the fiber and about 25 mm below the tip. Figure 6.5 is a representation o f
the coating setup. With this arrangement, the chance o f coating the tip aperture is very
slim due to the shadow effect. In fact, no case was observed in which light could not
pass through the tip because o f a metal layer blocking the tip aperture. Since the metal
coating of the optical fiber carries the microwave signal, it should be a low-loss
transmission layer. However, if the metal coating has poor adhesion to the optical
fiber surface, it becomes so sensitive to mechanical damage that even careful
handling can still cause
85
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Motor
Vacuum
chamber
Fig. 6.S: Schematic diagram of the setup for metal coating of the optical fiber using thermal
evaporation.
scratches and peeling of the metal coating. This results in a drastic increase in the
overall microwave loss and reduction o f the microwave sensitivity of the probe.
Figure 6.6 shows scanning electron microscope (SEM) pictures of some coated
tips, which have been fabricated with a single metal layer. Fig. 6.6(a) is a gold-coated
tapered fiber tip. For microwave application, the tapering of the tip should not be too
abrupt to avoid excessive signal reflection. The typical length of the tapered part
achieved with acid etching is about 300 pm to 500 pm. Figure 6.6(b) is a blow up of
the tip shown in Figure 6.6(a). The outer diameter is a little more than 1 pm and the
coating seems to have edges and is not uniform. These edges can cause non-uniform
microwave current density, which can degrade the microwave performance of the
probe. Another tip with silver coating (Fig. 6.6(c)) has bumps on the tapered part.
These tips and most of the tips coated with just a single layer of metal (gold or silver)
did not show any microwave transmission capability. In fact some optical fibers had
86
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(C)
Fig. 6.6: (a) Scanning electron microscope (SEM) picture of tapered tip coated with single
layer of gold, (b) Magnified SEM picture showing the bottom part of (a), (c) SEM picture of
another tip coated with a single layer of silver.
87
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
very high DC resistance when measured from end to end of the coated part. To
circumvent this problem, a two-step evaporation process was employed. Two
filaments were used to deposit two different materials without breaking the vacuum.
At first, a titanium (Ti) layer was deposited on the optical fiber with an approximate
thickness o f SO nra. Ti has very good mechanical properties, (i.e. Young’s modulus
and tensile strength) which are comparable to steel. However, the electrical resistivity
o f Ti is approximately 50 times higher than silver (Ag). Ti is known to stick very well
to surfaces when deposited. It serves as an adhesion layer for the next metal coating,
which is the good conductor for the microwave. Here, the second layer deposited was
Ag. Its thickness was approximately 150 nm. With this procedure, the metal-coated
fibers were able to function as microwave near-field probes in addition to serving as
optical probes.
6.5 Evaluation of the dual probe
The block diagram for the experimental scanning setup using the dual probe is
shown in Figure 6.7. It is configured to do collection mode NSOM. A 514 nm argonion laser or a 632 nm He-Ne laser is used as the light source directed toward the back
o f the sample that is mounted on a standard microscope slide. The slide is fixed on a
XYZ stage (Newport), where it can move in the XYZ directions with respect to the
probe. The maximum XY scan is 25 mm by 25 mm with a minimum step o f 0.074
pm. In the collection mode configuration, the probe has to be aligned with the laser
beam to get the maximum light input. Hence, a holder for adjusting vertical
orientation of the fiber optic accurately was also designed and constructed. The
holder is added to the setup to hold the dual probe. Each time a new tip is installed,
alignment is performed. The light collected by the tapered fiber optic goes directly to
a photo-multiplier tube (PMT), which is enclosed in an aluminum housing. This
enclosure prevents the PMT from detecting any stray light. Only the photons
collected by the fiber optic probe are detected. In addition, this setup does not require
a dark room. When the fiber tip is very close to the sample during scanning, the light
88
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Network Analyzer
Directional
coupler
PMT
Diode
detector
Probe holder with vertical
alignment mechanism x
LASER
Platform for XYZ stage with
fine adjustment leveling
}
mirror
■s\
(b)
Fig. 6.7: (a) Schematic diagram and (b) actual experimental scanning setup for the newlydeveloped dual probe.
89
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
coming from the ceiling o f the room is barely picked up by the tapered tip via
reflection off the sample.
For the microwave channel, a network analyzer (HP 8722C) was used to measure
the microwave reflection coefficient (r), resonant frequency ( / ) and quality factor
(Q). Since the imaging speed is limited by the sweep time of the analyzer at each
pixel, which for HP 8722C is 600 ms, imaging is generally slow when the
measurement is carried out in the frequency domain. In order to have a faster data
acquisition in this experiment, only the reflected power in CW mode was measured
using a bi-directional coupler and a high-sensitivity diode detector. For this particular
dual p ro b e,/, without sample is around 1.8 GHz. The initial test employed for a
newly installed dual probe tip was to test whether /
shifts downward when a
conducting sample is brought closer to the tip. This behavior is described in the
standard transmission line theory. If the tip is interacting with the sample via the
effective capacitance C*, arising from the gap between the probe tip and the sample
surface, then bringing the sample closer increases Cs. This is equivalent to increasing
the length of the resonator. Figure 6.8 is a plot o f / as the tip approaches the sample.
This sample is a silver film deposited on a LaAlCh substrate by thermal evaporation.
When the sample is far from the tip it does not perturb the resonator and / remains
more or less constant. At a tip-to-sample distance (TSD) around 4-5 pm, / starts to
decrease as predicted by the transmission line theory. It should be pointed out that the
TSD ~ 4-5 pm is comparable to the diameter o f the tip used in this experiment as
measured by a conventional far-field optical microscope. Based on the principle of
near-field microscopy, where resolution is determined by the tip diameter, we
assumed that our microwave probe has spatial resolution ~ 4-5 pm, which is much
less than the wavelength o f the microwave signal used (16 cm). Also, with TSD less
than the tip diameter, which is the common practice in near-field microscopy,
sensitivity is good enough to detect the sample. For the optical channel, it is expected
that resolution is better than 4 pm, which is still greater than the wavelength o f the
light used (514 nm). It is important to point out that this optical resolution is still
90
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1.830
1.838
1.837
Interaction between tip
and sample begins
1.836
&
1.835
■; 1.834
S
J
1833
1.832
- tip dameter
Z-mctar steps: 0.1 irri
1.831
0
1
2
3
4
5
6
7
Tip to Sarrpie D'stanoe (um)
Fig. 6.8: Behavior of resonant frequency as conducting sample approaches the tip.
-0.662
Collected light
Reflected power
-0.664
-
0.666
-
0.668
-0.670
-0.672
Reflected power (a.u.)
•
-0.674
I
-0.676
0
10 20
30 40
50 60
70 80
90 100
X (nm)
Fig. 6.9: Dual line scan of the probe on a patterned conducting strip with 20 pm line width.
91
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
achievable by far field microscopy. Figure 6.9 is a plot showing the dual probe
line scan on a Cu strip with a 20 pm line width. The solid line corresponds to the
voltage reading on the PMT as the dual probe crosses the linewidth. The flat region of
the curve, which goes to zero and has a width o f 20 pm, represents the line segment
o f the strip. Light collected there is expected to be minimum and almost zero because
there the sample is opaque. The noise in the optical signal is caused by the stains and
discolorations, which are present at the back of the substrate. The dotted curve
corresponds to the reflected microwave power collected from the dual probe as
detected by the diode detector. The dip indicates a sudden increase in microwave
absorption, which occurs right on the edge of Cu strip. It shows that the microwave
channel o f the dual probe is able to recognize the metallic circuit The two curves in
Figure 6.9 were taken simultaneously in one line scan. Although the response o f the
microwave and optical signals to the Cu strip are consistent, the spatial resolution of
NSOM is still far superior to that of NSMM at this stage.
At present, the resolution o f the dual probe is approximately 4-5 pm based on the
tip diameter. However, it is still necessary to determine the actual value by scanning
standard pattern with feature size smaller than 20 pm. Spatial resolution o f the probe
is limited by the sensitivity of the microwave channel. It was observed that the dual
probe tips with diameters less than 4 pm do not show measurable interaction like the
one shown in Figure 6.7, where f 0 decreases as sample approaches the tip. It should
be realized that even for the single-channel coaxial-type NSMM probes using the
combination o f a commercial cable and a pointed tip (i.e. STM tip), the highest
reported resolution is around 1-2 pm [66-68] . In this sense, the resolution o f the
microwave channel achieved on the dual-channel NSOM/NSMM is comparable to
the best reported so far for the NSMM single-channel probe. This resolution may be
further improved if the sensitivity of the NSMM is enhanced. Note that a metalcoated fiber optic NSOM probe can easily achieve spatial resolution better than 50
nm [63]. One may increase the thickness o f the metal coating to increase the tip area
A, which in turn increases the coupling capacitance Cs. For example, for a tapered
metal tip with tip diameter of 2 pm, the tip area is 7t pm2. In the case of a coated fiber
92
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
optic, the ring-area is n(a2~b2) where b is the tip radius o f the tapered fiber and a-b is
the coating thickness (CT). Assuming b is 0.25 pm, the calculated CT to obtain an
area of it pm2 is ~0.8 pm. It is not straightforward to deposit a metal with such
thickness on a very sharp fiber optic tip because mechanical stability is a concern. It
is possible that this fine tip is not strong enough to carry such a heavy metal coating.
Another consequence of this large CT is the decrease in the resolution of the NSMM
relative to the NSOM. With the assumed b = 0.25 pm, the expected resolution of the
NSOM is 0.5 pm while the NSMM has only 2 pm. In order to obtain a comparable
resolution in the microwave and optical probes, the CT needs to be minimized but it
should not sacrifice the desired sensitivity. Besides tip area, other factors that cause
additional loss to the probe should also be minimized to improve the sensitivity of the
NSMM. One topic deserving much attention is to improve the metal coating on the
probe tip by exploring better metals and more suitable deposition techniques such as
electron beam and sputtering. The goal is to make the metal coating a low-loss
transmission line for the microwave signal, which improves the Q of the resonator
and the sensitivity of NSMM. Moreover, the sensitivity of microwave detection can
also be improved by employing synchronized measurements, which have been done
by another group [69].
93
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 7
Applications of dual-channel near-field
scanning microwave and optical
microprobe
7.1 Degradation study of HgBa2CaCu20 6+s thin film
One problem encountered with HTS material is that it is susceptible to humidity.
Chemical stability o f high-Tc superconductors is an important issue in their
application. Understanding the mechanism o f their degradation process can lead to
the development o f an effective passivation layer appropriate to the specific
application. In addition, understanding degradation may play a role in engineering the
material to be more stable against moisture attack. Investigation of real-time
processes in a material such as degradation can be performed by scanning probe
microscopy (SPM)- This technique is popular and useful for surface characterization
o f materials. In this experiment the newly developed dual- channel near-field
94
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
scanning optical and microwave probe (dual probe) was employed to investigate how
HgBa2 CaCu2 0 6 +s(Hg-l 2 1 2 ) degrades when exposed to very high levels o f humidity.
With the dual probe, two physical properties could be monitored.
The Hg-1212 samples were prepared using the cation-exchange process
developed by Wu et al. [41]. In this method a Tl-2212 precursor film is annealed in
Mercury vapor to form Hg-1212. Moderate quality films were chosen intentionally
for this experiment. To induce degradation, a sample at room temperature was
exposed to a very high level of humidity (85%-90%) using a custom built humidifier.
The dual probe then mapped a fixed area on the sample (1.5 mm x 1.5 mm) at regular
time intervals using a cleaved fiber optic probe whose diameter was ~ 100 pm. A
computer program was developed to automate the scanning and data acquisition. In
this experiment, the time evolution of microwave and optical maps of the sample was
observed.
Figure 7.1 shows the microwave and optical images o f the sample before
degradation and after 6 minutes of exposure. Figure 7.1a is the spatial variation of
microwave absorption showing that there is non-uniformity in microwave reflection,
where the left side has relatively higher dissipation of the microwave signal. Since the
dual probe really measures sheet resistance (pit), where p is the microwave resistivity
and t is the film thickness, there are two possible causes o f this non-uniformity.
However, thickness variation can be ruled out because from the optical transmission
image (Fig 7.1c) o f the virgin sample, significant variation is not observed. After 6
minutes, the left side of the sample, which has higher p originally, degraded severely
as seen in the optical transmission image (Fig. 7.Id). This is consistent with the
common observation that good quality Hg-1212 films are more stable from moisture
attack. It can also be seen that there is a corresponding change in the microwave
absorption property of that region (Fig 7.1b). However, the decrease in microwave
absorption does not mean that the material improved its conductivity. The resonance
frequency of the microwave probe could have been shifted due to the abrupt change
in microwave property such as transition from a conducting material to insulator. In
fact, when we measured the microwave property on that left side using the frequency
95
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reflected pcMer
■
■
■
■
■
■
■
91-1(30%
81-90%
71-80%
61-70%
51-60%
41-50%
31-40%
21-30%
■ 11-20%
■ 0 -10%
X(mm)
X(mm)
Tranonlted Bgt*
91-100%
181-90%
|71-80%
61-70%
151-60%
j 41-50%
|31-40%
[21-30%
111- 2 0 %
0 -10%
(C)
(d)
Fig. 7.1: Spatial variation of microwave reflection before (a) and after 6 minutes of exposure
to humid environment (b). Lower images correspond to the spatial variation of optical
transmission before (c) and after 6 minutes of exposure (d). Scan area is 1.5 x 1.5 mm2.
96
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
domain measurement, we saw that the material there had turned into an insulator. By
comparing Figure 7.1(c) and 1(d), it can be observed that there is discoloration. The
average transmitted light over the scanned area has increased after exposure. This
effect also showed up in the microwave images before and after degradation. Since
the sample was actually scanned every two minutes, it was observed that for some
local spots there were fluctuations on the measured microwave absorption and optical
transmission. To get more information about the degradation process, taking place on
the sample surface, atomic force microscopy (AFM) was also employed. Results
showed that there was a continuous process o f surface deformation. Figure 7.2
contains images of surface roughness taken at different times during the sample
degradation process. At first, it can be seen that the immediate effect of moisture on
the Hg-1212 sample is the degradation o f surface morphology. Then it can be noticed
that the surface morphology changes from image to image. This effect also results in
a change in the light transmission because surface roughness causes light to scatter
away from the collecting fiber optic tip. Surface roughening might also affect the
microwave absorption since surface roughness has also an influence on the
mechanism o f microwave dissipation. Thus, the fluctuations of microwave absorption
and optical transmission, observed on some local spots might be caused by the
continuous process o f surface degradation.
In this study, some interesting features were observed on how Hg-1212 films
degrade. Using the dual probe two physical properties were investigated and
interestingly simple correlation was observed. For example, it was found out that
regions with poor conductivity are more susceptible to degradation. More
importantly, it was observed that the gradual transition of Hg-1212 from a black
colored conductor to a colorless insulator is accompanied by a continuous process o f
surface degradation. In other words, the surface morphology o f a degrading sample is
not stable. Another common observation when dealing with this material is that once
it exhibits a sign of degradation, for instance due to careless handling during
characterization, its rate o f deterioration speeds up. This could be explained by the
AFM results. Right after exposure to humidity, the sample becomes rougher, resulting
97
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fig. 7.2: AFM images showing the process of surface deformation occurring as sample
degrades. Scan area is 5 x 5 pm2 and height scale is from 0 nm to 300 nm.
98
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
in more surface area. With higher surface area, the sample could collect more
moisture or other undesirable particulates from the environment that will cause the
sample to degrade. In addition, the formation of new compounds, exhibited by the
growth of new structures, was clearly observed in the AFM images. These new
compounds are found to be rich in Ba and Ca as reported earlier [70].
7.2 Diagnostic tool for HTS microwave devices
The ability to scan the dual probe over a large area and the fact that we can
already make dual probe tips with diameter ~ 4-5 pm having detectable sensitivity led
us to study planar microwave devices. These devices have patterned structures having
dimensions of 100 pm to 500 pm. Thus, the probe can be used to locate defects at any
parts o f a device. Samples can be evaluated non-destructively right after patterning.
This step could lower production cost in an industrial environment because devices
showing signs of failure can be screened out before they are packaged and tested. If
the standard optical techniques are employed for this evaluation, they might be unable
to detect flaws that have similar optical characteristics as those exhibited by the good
material. The dual probe is a microscope that can see two different properties. Results
from the two probes can also be compared for consistency. It is also possible to use
one measured property to explain or interpret the other one.
In this experiment, the potential application of the dual probe as a diagnostic tool
is demonstrated. Two microstrip resonators having different performance were
selected. They were fabricated from a Tl-2212 thin film. One resonator (R l) had a
quality factor (Q) of 620 while the other one (R2) had a Q of 950, both at 80 K. They
were measured at the fundamental mode corresponding to a resonant frequency (fQ) o f
2 GHz. The whole device is on a 12 mm x 10 mm LaA1 0 3 substrate with a copper
ground plane. It is not practical to scan the whole area with the high resolution probe
tip. Instead, only the center segment o f the resonator was scanned because that is the
most critical region o f the resonator in the fundamental mode since the center is
99
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
where the microwave current peaks. Figure 7.3a shows the microwave image of R l.
The superconductor is seen from the microwave probe as a region of higher relative
dissipation as compared to the dielectric substrate. Note that from the image, the line
width of the resonator is around 150 pm. This agrees very well with its actual line
width. The corresponding optical image is represented in Figure 7.3b. The resonator
0.80
Microwave absorption 80
Light transmission
a
r
H
High
■ H ig h
■■
n
0.64
5
0.48
mm
■1
■
0.481
mm
m
■A*i
■
■
1'
L ow
Low
m M H
0.32
0.32
.
0.16
0.16
(a)
(b)
0.05 0.10
0.05 0.10
Fig. 7.3: (a) Microwave image obtained by the dual probe on a segment of a microwave
resonator Rl. (b) Corresponding optical image of the same segment mapped by the dual
probe.
100
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
center segment is generally seen from the optical probe as a region where light has
low transmission. Using two different signals, two images were generated
simultaneously by using the dual probe. In addition, they were performed in just one
raster scan. From the microwave image, it can be observed that there is non­
uniformity in the microwave reflection of R l. The lighter color on the lower part of
the strip could be an over etched region. Consistently, the optical image shows more
pinholes on this lower region, which may explain the cause of different microwave
reflection in that area. This non-uniformity could be the cause o f the poorer overall
performance of Rl relative to R2. When the same segment was scanned on R2,
uniformity was better both in optical and microwave reflection. These images are
shown in Figure 7.4.
o .s o
Microwave absorption
H High
■n
Light transmissio
■ High
;
;
1
0.64
0.64 ..."
\
■ i
*f.|
%fit m
i
T
m
. , ;
■ i
o.46
mm
0.48
m
m
. r , ' ‘ ' ••
i
.-iM
s * M*
m
m
■
mm
mm
Low
0.32
Low
0.32
0.16
0.16
00
(b)
0 OS 0 10
0 05 0 10
Fig. 7.4: (a) Microwave image obtained by the dual probe on a segment of a microwave
resonator R2. (b) Corresponding optical image of the same segment mapped by the dual
probe.
101
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7.3 Local observation of photo-induced effects: A potential
pump and probe application
7.3.1 Introduction
Photoconductivity is a well-known phenomenon in semiconductor materials and
its mechanism has been extensively investigated [71]. A similar effect also exists in
high-Tc superconductors although the mechanism behind it remain unclear [71].
There are two common effects that have been observed when a high-Tc
superconductor responds to photo-illumination or photo-doping, namely: persistent
photoconductivity (PPC) and photo-induced superconductivity (PISC). PPC is
exhibited as a slow increase in the conductivity o f the material during illumination
until it saturates and stabilizes at a certain value. When light is removed, the enhanced
conductivity remains for a period ranging from several hours to several days and
relaxes back to the original value. PISC is exhibited as enhancement of
superconducting properties, such as an increase in Tc, after the photo-doping. Several
groups have already observed and investigated these PPC/PISC phenomena on
different high-Tc superconductors. Most observations o f PPC/PISC are in oxygen
deficient samples such as RBajCusOx [72] GdBaiCuaOx [73] and YBaiCu3 0 x [74,75]'.
For these samples, the most popular model that explains the observed effects is the
oxygen defect mechanism described by Hansen et al. [72]. In this model, an electronhole pair is created during illumination. The electron is trapped in an oxygen vacancy
in the CuOx chain layer where normally O' ion would be located in a fully oxygenated
sample. The hole in turn is transferred into an extended state in the Cu02 plane,
where it participates in the current transport. As a result, conductivity o f the sample is
improved. The trapped electron can be released by providing sufficient thermal
energy. It recombines with the hole and the PPC effect vanishes. This thermal
deactivation process is slow. For example, the conductivity of a photo-doped sample
will only return to un-illuminated value after 33 hours at 295 K. [76]. If the
102
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
temperature is increased to 320 K, deactivation time reduces to 24 hours [77]. In
another report, illumination by infrared radiation was used to photo-excite trapped
electrons. It resulted in partial quenching of the PPC state [71]. One exception was
reported by Hoffmann et al. on the observation of PPC in oxygen-overdoped
TbBaiCuOe+s [78]. They claimed that depending on the doping level and illumination
wavelength, the normal state conductivity and Tc may either increase or decrease.
They concluded that the mechanism of PPC in this material is purely electronic. It
should be realized that in all HTS films showing PPC or PISC effects, their Tcs and
conductivity either before or after photo-illumination were much lower than that of
the same material with an optimized oxygen content. On HTS (including YBCO and
Tl-based HTS) samples with optimized oxygen content, and therefore optimized Tcs,
both PPC and PISC effects are negligible. This seems to suggest that both PPC and
PISC have an intimate relation with oxygen disorder due to the oxygen deficiency in
the HTSs.
7.3.2 Photo-induced superconductivity (PISC) in HgBa2CaCu20 6+§
thin films
An interesting story in this project was the accidental discovery of the photo­
induced superconductivity (PISC) on Hg-1212 films with optimal oxygen content and
therefore optimal Tcs around 120 K. When the NSMM/NSOM dual-probe was used
to study Hg-1212 films, strange instabilities were observed in the Rx after long-time
exposure of red light from the optical channel. This motivated us to investigate the
effect of photo-doping on the Tc of a Hg-1212 sample using the transport
measurement. We performed the experiment in a cryostat with an optical window so
that a light beam from a red laser (632 nm) can be directed on the sample surface. The
first step in the experiment was the measurement o f resistance versus temperature
curve, R(T), before photo-doping. After that, the sample temperature was set to 295 K.
and the light beam was applied for 12 hours. During this photo-doping, the cryostat
was maintained at low pressure (-lO"4 Torr) to avoid sample degradation and the
103
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
sample had 6 pA bias current. After 12 hours, laser was turned off and the R(T) curve
was measured again. The results are shown in Fig. 7.5. In this experiment, we
observed that the Tc increased from 120 K to 130 K. It should be realized that highest
Tc so far obtained for Hg-1212 films is around 124 K. while the observed Tc ~ 130K.
on photo-doped Hg-1212 is well above the original Tc of Hg-1212. In fact, this is the
first time a Tc higher than the intrinsic value o f the HTS has been observed, indicating
a higher-Tc superconductor can be obtained via photo-doping.
0.0008-1
r
0.0002
120
140
160
180
200
220
240
260
T(K)
(»)
(b)
Fig. 7.5: (a) Effect of 12-hr photodoping at 295 K to the Tc of Hg-1212 thin film, (b) Diagram
of transport measurement setup for studying photo-induced effects.
It should be noticed that this Tc enhancement is sample dependent and the
maximum increase we have observed is around 10 K. Moreover, this PISC effect was
not permanent. Six hours after the laser was turned off, Tc decreased to 127 K. After
two days, it went back to 120 K. We also observed that bias current is required to
induce such PISC. We have verified this effect using magnetization measurements
and the results are shown in Fig. 7.6. The increase in Tc as obtained by transport and
magnetization measurements are consistent with each other.
104
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
: 00*:
: ooig
-
~c{bef)= 1-5K
~ c i a t ) = 120 K
doping time I2 h rs
bias current 6 uA
J QQU4
too
200
290
300
Temperature (K)
(a)
- 0.001
-
•0 002
-
Tc(Def) = 11SK
T c(9tt)= 120 K
0.004 -
0
20
40
SO
30
ton
‘ 40
Temperature (K)
(b)
Fig. 7.6: Observation of PISC using (a) transport and (b) magnetization measurements.
Despite the exciting results we have achieved in investigation o f PISC on Hg1212 films, the mechanism remains a mystery. There are several differences distinct
this observation from most previous ones. First, the observed PISC is on Hg-1212
films with optimal oxygen content, suggesting the oxygen-vacancies trapping
proposed for YBCO case may not be the case here. Second, the observed
enhancement in Tc is much more significant than what reported for YBCO, and the
105
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
enhanced Tc after the photo-doping is much higher than that of the intrinsic Tc o f Hg1212. Thirdly, the conductivity change during this process is minimal. This differs
dramatically from the YBCO case in which, the PISC occurs simultaneously with the
PCC. Finally, the PISC effect on Hg-1212 films can only be induced when there is
bias current through the sample during the photo-doping. This is not the case for the
PISC on YBCO. In fact, when we eliminate the bias current during the photo-doping,
no PISC was observed on Hg-1212 films. This explains instability observed in R*
measurement using our NSMM/NSOM dual-probe, where the photo-illumination was
accompanied by the microwave bias. In fact, the observation o f this novel PISC effect
on Hg-1212 may very well be one of many new physical phenomena one may
discover using the multi-channel SPM. Such a quest will certainly be the center of my
research in next few years.
106
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 8
Conclusion and future direction
The basic accomplishment of this work is the demonstration of the idea that
different scanning probe microscopy (SPM) techniques could be combined into one
probing structure using a multi-layer design. Using this concept, different physical
properties o f a sample can be characterized simultaneously at very high spatial
resolution. This is an important tool in most practical research and applications
because correlation o f different physical properties is highly desirable.
In this dissertation, design, development and evaluation of a prototype SPM
probe was carried out systematically and successfully. For the first time, a single
probe that could perform near-field scanning microwave microscopy (NSMM) and
near-field scanning optical microscopy (NSOM) simultaneously was actually
implemented. In addition, the prototype probe was used in several real applications
wherein it was used to characterize high-temperature superconductors. From those
applications, it was realized how promising the dual probe can be as a tool in research
activities as well as fabrication type environment involving a wide range of materials.
This includes superconductors, semiconductors, dielectrics and even biological
samples. In developing the concept of multi-channel SPM, it should be realized that
this dissertation accomplished the first stage. Fundamental problems were identified
and critical issues were examined. Thus, helpful ideas for optimization o f this probe
are already being considered but addressing these issues will require more time.
107
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Other applications of the dual probe to fundamental research are also being
considered. One interesting application that was demonstrated in this dissertation was
the “pump probe”. This comes out naturally based on the dual-channel structure of
the probe. One channel can serve as the excitation probe and simultaneously the other
channel can measure or monitor the effects associated with the excitation. For
example, for the SPM probe developed here, the optical channel was used to apply
light locally on a sample while the microwave channel was utilized to measure
conductivity changes induced by the visible radiation. These photoconductivity
effects have already been observed in semiconductors and superconductors and might
also exist in other materials. Studying these materials locally with this dual probe
might give very interesting information.
Another potential “pump and probe” application is local heating on the sample
using the microwave channel and then probing the effects using the optical channel.
For instance, the fiber optic channel can be utilized to serve as an optical pyrometer.
It can measure local temperature changes due to microwave absorption by the
material. Since only a local spot is heated up, the dual probe can be used to study heat
diffusion.
In the future, the idea of integrating different SPM techniques through the multi­
layer approach could be adapted to develop new combinations. It could be a
combination o f existing techniques or a combination o f existing and non-existing
techniques. In addition, SPM probes with more than two channels might be developed
in the future. At present, development o f a dual probe, which can be used for
scanning tunneling microscopy (STM) and NSOM is underway. A very sharp
protrusion will be created beside the aperture o f a fiber optic NSOM probe. Once the
fiber is coated with the metal layer, this sharp point will become conducting and can
be used as an STM tip. The bias voltage will be applied to the metal coating, which
extends from the protrusion to the untapered part o f the optical fiber.
108
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Bibliography
[1]
A. de Lozanne, Supercond. Sci. Technol., 12, R43 (1999).
[2]
M.A. Paesler, P.J. Moyer, Near-Field Optics Theory, Instrumentation and
Applications, John Wiley & Sons (1996).
[3]
D.A. Bonnell, Scanning Tunneling Microscopy and Spectroscopy, VCH
publishers (1993).
[4]
H.D. Hallen, R. Seshadri, A.M. Chang, R.E. Miller, L.N. Pfeiffer, K.W.
West, C.A. Murray and H.F. Hess, Phys. Rev. Lett. 71, 3007 (1993).
[5]
J.C. Vickerman, Surface Analysis-the principal techniques, John Wiley &
Sons Ltd. (1997).
[6]
R.J. Hamers, R.M. Tromp and J.E. Demuth, Phys. Rev. Lett. 56, 1972
(1986).
[7]
E. Betzig, P.L. Finn and J.S. Weiner, Appl. Phys. Lett. 60,20 (1992).
[8]
C.E. Talley, M.A. Lee and R.C. Dunn, Appl. Phys. Lett. 72,2954 (1998).
[9]
W.E. Moemer, T. Plakhotnik, T. Imgartinger, U.P. Wild, D.W. Pohl and B.
Hecht, Phys. Rev. Lett. 73, 2764 (1994).
109
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
[10]
H.J. Mamin, P.H. Gueturer and D. Rugar, Phys. Rev. Lett. 65, 2418
(1990).
[11]
J.G. Bednorz and K.A. Muller, Z. Physik B 64, 189 (1986).
[12]
M.K. Wu, J.R. Ashbum, C.J. Tomg, P.H. Hor, R.L. Meng, L. Gao, Z.J.
Huang, Y.Q. Wang and C.W. Chu, Phys. Rev. Lett. 58,908 (1987).
[13]
H. Maeda, Y. Tanake, M. Fukutomi and T. Asano, Jpn. J. Appl. Phys. 27,
L209 (1988).
[14]
Z.Z. Sheng and A.M. Hermann, Nature 332, 55 (1989).
[15]
A. Schilling, M. Cantoni, J.D. Guo and H.R. Ott, Nature 363, 56 (1993).
[16]
G. Bums, High- Temperature Superconductivity, an Introduction,
Academic Press Inc. (1992) p. 16.
[17]
M.E. Hawley, J.G. Beery, F.H. Garzon and R.J. Houlton, Appl. Phys. Lett.
59, 3177(1991).
[18]
L. Luo, M.E. Hawley, C.J. Maggiore, R.C. Dye, R.E. Muenchusen, L.
Chen, B. Schmidt and A.E. Kaloyeros Appl. Phys. Lett. 62,485 (1993).
[19]
Ch. Gerber, D. Amselmetti, J.G. Bednorz, J. Mannhart and D.G. Schlom,
Nature 350, 279(1991).
[20]
H.L. Edwards, A.L. Barr, J.T. Markert and A.L. de Lozanne, Phys. Rev.
Lett. 73, 1154(1994).
110
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
[21]
A. Moser, H.J. Hug, I. Parashikov, B. Stiefel, O. Fritz, H. Thomas, A.
BaratofF, H J. Guntherodt and P. Chaudhari, Phys. Rev. Lett. 74, 1847
(1995).
[22]
C.W. Yuan, Z. Zheng, A.L. de Lozanne, M. Tortonese, D.A. Rudman and
J.N. Eckstein, J. Vac. ScL Technol. B 14,1210 (1996).
[23]
C.C. Tsuei, J.R. Kirtley, C.C. Chi, L.S. Yu-Jahnes, A. Gupta, T. Shaw,
J.Z. Sun and M.B. Ketchen, Phys. Rev. Lett. 73, 593 (1994).
[24]
A. Oral, J.C. Barnard, S.J. Bending, I.I. Kaya, S. Ooi, T. Tamegai and M.
Henini, Phys. Rev. Lett. 80,3610, (1998).
[25]
I. Takeuchi, T. Wei, Fred Duewer, Y.K. Yoo, X-D. Xiang, V. Talyansky,
S.P. Pai, G.J. Chen and T. Venkatesan, Appl. Phys. Lett. 71, 2026 (1997).
[26]
E.B. McDaniel, S.C. Gausepohl, C.-T. Li, M. Lee, J.W.P. Hsu, R.A. Rao
and C.B. Eom, Appl. Phys. Lett. 70, 1882 (1997).
[27]
G. Nieva, E. Osquiguil, J. Guimpel, M. Maenhoudt, B. Wuyts, Y.
Bruynseraede, M.B. Maple and I.K. Schuller, Phys. Rev. B, 46, 14249
(1992).
[28]
R.S. Decca, H.D. Drew, B. Maiorov, J. Guimpel and E. Osquiguil, Appl.
Phys. Lett. 73, 120 (1998).
[29]
S. M. Anlage, D.E. Steinhauer, B.J. Feenstra, C.P. Vlahacos and F.C.
Wellstood, Microwave Superconductivity, Kluwer Academic Publishers
(2001) p. 239.
Ill
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
[30]
C. Wilker, Z.-Y. Shen, P. Pang, D.W. Face, W.L. Holstein, A.L. Matthews
and D.B. Laubacher, IEEE Trans, on MTT 39, 1462 (1991).
[31]
D.E. Oates, A.C. Anderson and B.S. Shih, IEEE MTT-S Digest, 627
(1991).
[32]
D.E. Oates and A.C. Anderson, IEEE Trans, on Magnetics, 27, 867
(1991).
[33]
A.A. Valenzuela and P. Russer, Appl. Phys. Lett., 55, 1029 (1989).
[34]
N. Tellman, N. Klein, U. Dahne, A. Scholen, H. Schulz and H. Chaloupka,
IEEE Trans, on Appl. Supercond. 4, 143 (1994).
[35]
C. Wilker, Z.-Y. Shen, V.X. Nguyen and M.S. Brenner, IEEE Trans. Appl.
Supercond., 3, 1457 (1993).
[36]
N. Klein, U. Dahne, U. Poppe, N. Tellman, K. Urban, S. Orbach, S.
Hensen, G. Muller and H. Pierl, J. Supercond. 5, 195 (1992).
[37]
J. Slater, Rev. o f Modem Physics, 18,441 (1946).
[38]
D.E. Oates, A.C. Anderson, D.M. Sheen and S.M. Ali, IEEE Trans, on
MTT. 39, 1522(1991).
[39]
P.P. Nguyen, D.E. Oates, G. Dresselhaus, M.S. Dresselhaus and A.C.
Anderson, Phys. Rev. B. 51,6686 (1995).
[40]
P. Lahl, R. Wordenweber and M. Hein, Appl. Phys. Lett. 79, 512 (2001).
112
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
[41]
J.Z. Wu, S.L. Yan, and Y.Y. Xie, Appl. Phys. Lett. 74, 1469 (1999).
[42]
S.H.Yun and J.Z. Wu, Appl. Phys. Lett. 68, 862 (1996).
[43]
C.P. Bean, Phys. Rev. Lett, 8, 250 (1962).
[44]
C.P. Bean, Rev. Mod. Phys., 36, 31.
[45]
J. Hudner, O. Thomas, F. Weiss, D. Boursier, E. Mossang, J.P. Senateur,
J.C. Villegier, H. Moriceau, M. Schwerdtfeger, A. Jager and H. Ohlsen,
Supercond. Sci. Techn. 7, 195 (1994).
[46]
R.L. Ormeno, D.C. Morgan, D.M. Broun, S.F. Lee and J.R. Waldram, Rev.
Sci. Instrum. 68,2121 (1997).
[47]
I.S. Gergis, P.H. Kobrin, J.T. Cheung, E A . Sovero, C.L. Lastufka, D.S.
Deakin and J. Lopez, Physica C 175,603 (1991).
[48]
W.L. Holstein, L.A. Parisi, C. Wilker and P.B. Flippen, Appl. Phys. Lett.
60, 2014(1992).
[49]
Y.M. Habib, C.J. Lehner, D.E. Oates, L.R. Vale, R.H. Ono, G.
Dresselhaus and M.S. Dresselhaus, Phys. Rev. B. 57, 13-833 (1998).
[50]
G. Hampel, P. Kolodner, P.L. Gammel, P A . Polakos, E. de Cbaldia, P.M.
Mankiewich, A. Anderson R. Slattery, D. Zhang, G.L. Liang and C.F.
Shin, Appl. Phys. Lett. 69, 571 (1996).
[51 ]
E.H. Synge, Philos. Mag. 6,356 (1928).
113
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
[52]
R.F. Soohoo, J. Appl. Phys. 33, 1276 (1962).
[53]
E.A. Ash and G. Nicholls, Nature (London) 237, 510 (1972).
[54]
M. Tabib-Azar, P.S. Pathak, G. Ponchak, and S. LeClair, Rev. o f Sci. Inst.
70, 2783 (1999).
[55]
I. Takeuchi, T. Wei, Fred Duewer, Y.K. Yoo, X-D. Xiang, V. Talyansky,
S.P. Pai, G.J. Chen and T. Venkatesan, Appl. Phys. Lett. 71, 2026 (1997).
[56]
D.E. Steinhauer, C.P. Vlahacos, S.K.. Dutta, B,J. Feenstra, F.C. Wellstood
and S.M. Anlage, Appl. Phys. Lett. 72, 861 (1998).
[57]
A.F. Lann, M. Abu-Teir, M. Golosovsky, D. Davidov, S. Djordjevic, N.
Bontemps and L.F. Cohen,, Rev. o f Sci. Inst. 7 0 ,4348 (1999).
[58]
D.M. Pozar, Microwave Engineering (Addison-Wesley Publishing Co.
Inc. (1990) p.330.
[59]
C.P. Vlahacos, R.C. Black, S.M. Anlage, A. Amar and F.C. Wellstood,
Appl. Phys. Lett. 69,3272 (1996).
[60]
D. W. Pohl, W. Denk, and M. Lanz, Appl. Phys. Lett. 44, 651 (1984).
[61]
A. Lewis, M. Issacson, A. Haratounian, and A. Murray, Ultramicroscopy
13, 227.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
[62]
M.A. Paesler, P.J. Moyer, Near-Field Optics Theory, Instrumentation and
Applications, John Wiley & Sons (1996).
[63]
E. Betzig, J.K. Trautman, T.D. Harris, J.S. Weiner, ans R.L. Kostelak,
Science 251, 1468(1991).
[64]
C.W. Hollars and R.C. Dunn, Rev. Sci. Instru., 69, 1747 (1998).
[65]
R. Stockle, Ch. Fokas, V. Deckert, R. Zenobi, B. Sick, B. Hecht and U.P.
Wild, Appl. Phys. Lett., 75, 160 (1999).
[66]
T. Wei, X.D. Xiang, W.G. Wallace-Freedman and P.G. Schultz, Appl.
Phys. Lett. 68,3506 (1996).
[67]
A.F. Lann, M. Abu-Teir, M. Golosovsky, D. Davidov, S. Djordjevic, N.
Bontemps and L.F. Cohen, Rev. Sci. Instr. 70,4348 (1999).
[68]
S.M. Anlage, D.E. Steinhauer, C.P. Vlahacos, B.J. Feenstra, A.S.
Thanawalla, W. Hu, S.K. Dutta and F.C. Wellstood, IEEE Trans, on Appl.
Supercond. 9, 4127 (1999).
[69]
M. Tabib-Azar, D.P. Su, A. Pohar, S.R. LeClair and G. Ponchak, Rev. o f
Sci. Instru. 70, 1725 (1999).
[70]
T. Aytug, B.W. Kang, S.L. Yan, Y.Y. Xie and J.Z. Wu, Physica C 307,
117(1998).
[71]
D.C. Chew, J.F. Federici, J. Guitierrez-Solana, G. Molina, W. Savin and
W. Wilber, SPIE proceedings 2696, 654 (1996).
115
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
[72]
J. Hasen, D. Lederman, I.K. Schuller, V. Kudinov, M. Maenhoudt and Y.
Bruynseraede, Phys. Rev. B. 51, 1342 (1995).
[73]
G. Nieva, E. Osquiguil, J. Guimpel, M. Maenhoudt, B. Wuyts, Y.
Bruynseraede, M.B. Maple and I.K. Schuller, Appl. Phys. Lett. 60,2159
(1992).
[74]
W. Markowitsch, C. Stockinger, W. Lang, K. Bierleutgeb, J.D. Pedamig
and D. Bauerle, Appl. Phys. Lett. 71, 1246 (1997).
[75]
J.F. Federici, D. Chew, B. Welker, W. Savin, J. Guitierrez-Solana, T. Fink
and W. Wilber, Phys. Rev. B. 52, 15592 (1995).
[76]
V.I. Kudinov, I.L. Chaplygin, A.I. Kirilyuk, N.M. Kreines, R. Laiho, E.
Lahderanta and C. Ayache, Phys. Rev. B. 47, 9017 (1993).
[77]
W. Gob, W. Lang, W. Markowitsch, V. Schosser, W. Kula and R.
Sobolewski, Sol. St. Comm. 96,431, (1995).
[78]
A. Hoffmann, I.K. Schuller, Z.F. Ren, J.Y. Lao and J.H. Wang, Phys. Rev.
B. 56, 13742 (1997).
116
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Publications
1. R.S. Aga, Jr., J. Brookman and J.Z. Wu, “Development c f dual-channel near-field
scanning optical and microwave micrcprobe ", submitted to Appl. Phys. Lett.
2. R.S. Aga, Jr., Y-Y. Xie and J.Z. Wu, “Real-time degradation study c f
HgBaiCaCuiOe+s", Mat. Res. Soc. Symp. Proc. 689, E3.9.1 (2002)
3. R.S. Aga, Jr., S.L. Yan,Y-Y. Xie and J.Z. Wu, “Passivation t Jects c f polymethyl
methacrylate on superconducting HgBajCaCufDs^s thin films ”, Physica C 366,
216 (2002)
4.
R.S. Aga, Jr., Y-Y. Xie, S.L. Yan, J.Z. Wu and S. Han, “Microwave-power
handling capability c f HgBa^CaCuiOe+s superconducting microstrip lines",
Appl. Phys. Lett. 79,2417 (2001)
5. R.S. Aga, Jr., Y-Y. Xie, J.Z. Wu and S. Han, “Microwave characterization c f
HgBazCaCuiOe+s thin film s ", Physica C 341-348,2721 (2000)
6.
R.S. Aga, Jr., S.L. Yan, Y.Y.Xie, S. Han, J.Z. Wu, Q.X. Jia and C. Kwon,
“Microwave suiface resistance cfH gBajCaCu^O ^s thin film s ", Appl. Phys. Lett.
76, 1606 (2000)
117
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Документ
Категория
Без категории
Просмотров
0
Размер файла
5 272 Кб
Теги
sdewsdweddes
1/--страниц
Пожаловаться на содержимое документа