Full Text
Full Text
Full Text
2012
Sarinya Pasakawee
ACKNOWLEDGEMENT
There are numerous people who I wish to thank for their assistance during this PhD
study. Firstly, I want to take this chance to give a special thanks to my supervisor,
Dr. Zhirun Hu, who has inspired this work. I also want to thank him for his
utmost gratitude to my sponsor, the Royal Thai Government, who’s funding has
I would also like to thank Dr. Cahyo Mustiko for his assistance discussions and
advice over my PhD both with respect to lab work and theory. Thanks are also due to
Dr. Abid Ali and Dr. Mahmoud Abdalla for their previous works without which the
Finally, I would like to thank sincerely to my husband, Mr. Ross Allen, and the rest
of my family for their support and sacrifices for me to achieve what I have.
This thesis is an opportunity to express my thanks for all their support and
encouragement.
3
ABSTRACT
A new equivalent circuit of left-handed (LH) microstrip transmission line loaded
magnetic coupling into the equivalent circuit, the new equivalent circuit presents a
more accurate cutoff frequency than the old one. The group delay of CSRRs applied
with microstrip transmission line (TL) is also studied and analyzed into two cases
which are passive CSRRs delay line and active CSRRs delay line. In the first case,
the CSRRs TL is analyzed. The group delay can be varied and controlled via signal
frequency which does not happen in a normal TL. In the active CSRRs delay line,
the CSRRs loaded with TL is fixed. The diodes are added to the model between the
voltages from -10V to -20V, the group delay can be moved from 0.6ns to 5.6ns.
frequency at 3.9GHz. Moreover, a high rejection in the lower band and a low
Finally, it is shown that CSRRs applied with planar antenna can reduce the antenna
size. The structure is formed by etching CSRRs on the ground side of the patch
antenna. The meander line part is also added on the antenna patch to tune the
operation frequency from 1.8GHz downward to 1.73GHz which can reduce the
antenna size to 74% of conventional patch antennas. By using the previous antenna
structure without meander line, this proposed antenna can be tuned for selecting the
and ground. The results provide 350MHz tuning range with 35MHz bandwidth.
4
CONTENTS
Chapter 1: Introduction
1.1 Overview…………………………………………………………………… 17
1.4 References………………………………………………………………….. 22
2.1 History……………………………………………………………………… 24
2.5 References………………………………………………………………….. 55
5
Introduction…………………………………………………………………58
3.3 Electric, Magnetic Coupling and the new Equivalent Circuit Model…….... 83
3.5 Conclusion…………………………………………………………………. 94
3.6 References………………………………………………………………….. 95
Ring Resonators
Introduction……………………………………………………………… 99
4.2 Passive CSRRs Model and Delay Line Design Procedures……………… 103
4.4 Conclusion…………………………………………………………………116
6
4.5 References…………………………………………………………………117
Introduction………………………………………………………………. 121
Introduction………………………………………………………………. 149
Appendix…………………………………………………………………... 181
8
LIST OF FIGURES
Figure 2.1 (a)The first artificial dielectrics lenses: Lattice of conducting disks
arranged to form lens by Kock and Cohn in1948 [2.2], and (b) J.B. Pendry designed
Figure 2.2 The diagrams of electric field, magnetic field, wave vector (E, H, k) and
Poynting vector (S) for electromagnetic wave propagation in right-handed and left
Figure 2.3 Energy flow and wave vector diagram between RH and LH interface…34
Figure 2.4 Material classifications according to є, μ pairs and η, type of waves in the
Figure 2.6 (a) Planar lenses with the interface of negative refractive index material
(arrows are the wave vector in the medium) [2.11] and (b) the negative refraction
Figure 2.7 The Doppler effect (a) Conventional RH medium (Δω>0), and (b) LH
Figure 2.8 (a) Photo graph of the Metamaterial cube [Physic today, June 2004], and
(b) Generic view of a host medium with periodically placed structures constituting a
MTM……………………………………………………………………………….. 41
Figure 2.9 Metamaterial structures: (a) a medium composed with metallic wire and
(b) thin wire lattice exhibiting negative εeff if E applied along the wires………….. 42
Figure 2.10 Metamaterial structures: Split ring resonators lattice exhibiting negative
Figure 2.11 Some geometries of SRR used to realise artificial magnetic materials
[2.7]……………………………………………………………………………….. ..46
Figure 2.12 The first DNG metamaterial structures [2.5] Smith et al., 2000-2001: (a)
Figure 2.13 Topology of CSRRs and the stack CSRRs, E is parallel to the CSRRs
Figure 2.14 Geometry of the CSRRs (a) with and (b) without capacitive gap
Figure 2.15 The simulated results of (a) scattering parameters, (b) the output phase,
and (c) the effective permittivity of a combined CSRRs structure and microstrip
line………………………………………………………………………………......50
Figure 2.16 The simulated results of (a) insertion, return loss, and (b) phase of S21
for a unit cell of CSRRs combined structure with microstrip line and a series gap
0.3mm…………………………………………………………………………….... 51
Figure 2.19 (a) Photograph of the patch antenna loaded with CSRRs and (b)
[2.22]……………………………………………………………………………….. 53
Figure 2.20 (a) The ultra-wideband antenna with quadruple-band rejection and (b)
Figure 3.1 Topology of (a) SRR and (b) CSRRs with relevant dimensions…….. ..60
Figure 3.2 Planar equivalent circuits for a LH transmission line periodic cell
[3.1]……………………………………………………………………………….. ..63
10
Figure 3.3 The circuit model of a LH TL per unit length [3.1]………………….. ..65
Figure 3.4 Example of ω-β diagram in a purely LH (red) and RH (blue) TLs…... ..67
Figure 3.8 Dispersion diagram of the 30-stage CRLH TL balance casein Figure
3.7………………………………………………………………………………….. 74
Figure 3.9 (a) Typical index of refraction for the balanced (green) and unbalanced
3.7………………………………………………………………………………… ..77
Figure 3.10 Basic cell of CSRRs-based transmission line (a) and equivalent circuit
model, (b) The upper metallization is depicted in black; the slot regions of the
Figure 3.11 (a) The topology of CSRRs unit cell and (b) 2-cells CSRRs microstrip
Figure 3.12 Field variations with ring separation of the two CSRRs (a) Electric field
Figure 3.13 Equivalent circuit model of a two adjacent-cell CSRRs TL………... ..86
Figure 3.14 (a) The S-parameters of the 2-cells CSRRs TL with cell separation of
0.2mm and (b) Smith Chart, equivalent circuit model and full wave simulation... ..87
Figure 3.15 The prototype of 4-cell CSRRs TL, (a) backside view and (b) top
view……………………………………………………………………………….... 89
Figure 3.16 The measured results of (a) S-parameters of 4 CSRRs cells and (b)
Phase of S21……………………………………………………………………….... 90
11
Figure 3.17 The S-parameters (S21 and S11) by circuit simulation and
measurement……………………………………………………………………….. 91
Figure 4.2 (a) The topology of a unit cell CSRRs TL and Photographs of the
designed 4 –cell CSRRs TL ;(b) ground view, (c) microstrip top view………….. 104
Figure 4.3 The S-Parameters of the 4-cells CSRRs TL by both simulations and
measurement…………………………………………………………………….. 105
Figure 4.4 The group delay on passband of the 4-cells CSRRs TL by both
simulations (dot and dash blue) and measurement (solid blue) as well as a
Figure 4.5 The dispersive delay line in a simple RF system [4.13]……………… 108
Figure 4.6 The three continuous wave (CW) output signals in Time domain of 4-
Figure 4.7 The modulated output signals with different carriers ( fc1=2.25GHz in dot
Figure 4.8 The Envelopes of RF Pulse fc1 and fc2 in time domain compared to input
after travelling through the 4-cell CSRRs by the measured S-parameters……….. 111
Figure 4.9 (a) The equivalent circuit of two adjacent CSRR cells with varactor
diodes, (b) and (c) the photographs of the fabricated 4-cells CSRRs active delay line
Figure 4.10 The insertion response (dB) after DC bias voltages of 4-cell CSRRs
TL………………………………………………………………………………… 114
12
Figure 4.11 Measured delay time of the 4-cells active CSRRs delay line at the
Figure 5.1 Two-port network with the input reflection coefficient and the output
Figure 5.2 Interconnection of two-port network (a) Series, (b) Parallel, and (c)
Figure 5.4 The layout of a balanced load capacitance without CSRRs…………. 134
Figure 5.5 The insertion and return loss of balanced load capacitance structure
Figure 5.6 (b) The unit cell of CSRRs loaded with transmission line topology and
Figure 5.7 The layout of proposed filter. (a) Basic cell, (b) Topology of rectangular
CSRRs……………………………………………………………………………. 138
Figure 5.8 The current distributions of the proposed bandpass filter at (a) 0.72GHz
Figure 5.9 The magnetic field distribution on ground plane at (a) 0.72GHz and (b)
Figure 5.10 The photograph of the fabricated filter (a) Top view, (b) Bottom view.
total dimension of the filter is 40x41.5mm2 and dielectric constant (r) =10.2…… 140
Figure 5.12 The ADS simulated S-parameters (a), and (b) comparison of ADS and
Figure 5.13 Measured and HFSS simulated frequency response on the designed
Figure 6.3 The feeding methods of a microstrip antenna (a) Coaxial feed, (b) Inset-
feed, (c) Proximity-coupled feed, and (d) Aperture-coupled feed. [6.5]………… 154
[6.5]………………………………………………………………………………. 157
Figure 6.5 (a) The unit cell of CSRRs TL and (b) its T-equivalent circuit……… 160
Figure 6.7 Layout of the meander line (4 turns), a=0.6mm and the capacitive
gap=0.2mm………………………………………………………………………. 163
Figure 6.8 The simulated return loss (S11) of each CSRR patch antenna……… 164
Figure 6.9 The simulated field distributions at resonant frequencies (a) E field of
antenna1 at 1.8GHz and (b) E field of antenna3 at 1.73GHz, while (c) H field of
respectively.............................................................................................................. 165
Figure 6.10 The surface’s current distribution at 1.73GHz of antenna3 (a) on patch
Figure 6.11 The photos of antenna3 (a) the meander line patch with gap (b) The
Figure 6.12 The return loss in dB by full wave HFSS simulation and VNA
measurement of meander line loaded patch antenna with CSRRs etched on the
ground……………………………………………………………………………. 168
Figure 6.13 The simple structure of the proposed antenna with varactor diode… 169
14
Figure 6.14 The layout of tunale patch antenna (a) ground view and (b) patch
view………………………………………………………………………………. 170
Figure 6.15 The measured return loss of the tunable CSRRs patch antenna with
LIST OF TABLES
Table 3.1 Extracted Element Parameters for the 2-cell CSRRs TL……………… 88
Table 5.1 The overview of the relationship between two-port network parameters
and incident and reflected wave variables [5.1, 5.23, 5.24]……………………… 127
Table 5.2 Summarization of the terminated two port circuits [5.1, 5.25, 5.26]….. 129
antennas…………………………………………………………………………... 156
LIST OF ABBREVIATIONS
ADS Advanced design system
BW Bandwidth
EM Electromagnetic
LH Left-handed
MTM Metamaterial
RF Radio frequency
RH Right-handed
TL Transmission line
TW Thin wire
CHAPTER 1
INTRODUCTION
1.1 Overview
Left-Handed (LH) MTMs have negative electric permittivity (ε) and permeability (μ)
parameters, LH MTMs are said to have anti-parallel phase and group velocities,
therefore gives a negative refractive index (n) [1.2]. Left-handedness was theorized
first by Veselago [1.3] and was proved experimentally by Pendry [1.4] and Smith
[1.5]. Building on this initial work, LH structures can be categorized into main
configurations: thin wires (TW) and composite materials made from an array of split
ring resonators (SRRs) [1.6]; and periodically loaded transmission lines (TLs)
usually using shunt shorted inductance (stubs, meandered or spiral lines), and series
mentioned, in the last decade, there are many new frontiers of microwave circuits
Microwave industry for making mass produced components of high frequency, and a
(CPW) structure. Its specific effect is not noticeable by maintaining size [1.12-1.14].
excites the particle, and thus has been introduced [1.13-1.14]. Complementary split
ring resonators (CSRRs) are the dual form of SRRs. Since this structure is etched on
the ground plane, under the conductor strip position, and is excited by the electric
field, induced by the conductor line. This coupling can be modeled by a series,
connecting the line capacitance to the CSRRs, and therefore, due to these benefits,
The aim of this research is to use the unique properties of MTMs via CSRRs and
As mentioned, there are three main objectives of this study one of the objectives is to
novel CSRRs transmission line, which can minimize the length as well as enhance
degrade the system quality. Delay suppressions are the recommended part of
communication systems. However, this thesis presents a new method concerning the
group delay of LH passband of CSRRs transmission line. The group delay can be
managed, as a result more signals can be sent simultaneously without any concern
Filter is a component used to select the wanted signal and suppress the unwanted
noise. The filter performance is a crucial part resulting in system quality, therefore, it
is essential to design a high performance filter which is covered all criterion of the
selectivity, high and wide rejection out-of band, spurious suppression, as well as low
insertion loss. In planar microstrip filter, the structure and length of transmission line
band with high selectivity, therefore, with these attractive specific properties, it is the
selected particle to fulfill the planar microstrip filter criteria, as it exhibits good
Antenna is the front end in transmitter and receiver part. In antenna technology, the
patch antenna size is larger when there is a lower operating frequency. Therefore,
The thesis is organized into seven chapters. Chapter 2 presents the general theory of
metamaterials by explaining the Maxwell’s equations and how they are applied. The
MTMs’ specific properties, such as negative refraction and reversal Doppler Effect,
are described. The general MTMs formed in planar technology are presented; for
example, thin wires, split ring resonators (SRRs) and its dual part called
“complementary split ring resonators (CSRRs)”. In order to make a clear point with
20
regard to MTMs, the chapter will end up with some applications of MTMs in today’s
work.
structures are presented and analyzed, which includes the purely left-handed (LH)
structure and the composite left-right handed (CRLH) structure. The left-handed
analyzed. In addition, the new equivalent circuit of the left-handed particle structure
with the magnetic coupling effect between the two close rings included is also
passband region of CSRRs loaded TL structure is also indicated and analyzed by its
Since there is very limited work to date in the delay property on planar microstrip
circuitry using MTMs, the study of delay lines using left-handed Materials (LHMs)
dispersive properties, will be explored in chapter 4. There are two separated cases in
this chapter. The passive CSRRs delay line is first analyzed. The group delays are
studied via the displayed dispersive delayed signals, as both continuous waves
(CWs) and pulse signals. In active CSRRs delay line, the presence of varactor diodes
studied.
demonstrated. The filter theory is initially presented. Then, the prototype microstrip
ground plane of this prototype structure, which results in the novel microstrip filter
bandpass filter to a wide bandpass filter. Other filter criteria are also displayed and
In chapter 6, a criterion for ‘electrically small’ size, using the negative phase
property has been developed via antenna design. The etching of CSRRs on the group
plane part of a microstrip patch antenna is presented for specific propagation and
minimization. The meander line is another option for fulfilling further size reduction.
between the ground and the patch side. The frequency selective antenna is proposed.
Chapter 7 describes the conclusions of this thesis and future works of the CSRRs
applications which are resultant of the conclusions that were able to come about
1.4 References
Fundamental Principles and Applications, John Wiley & Sons, Inc., 2005.
negative values of ε and μ”, Soviet Physics Uspekhi, Jan-Feb 1968, vol. 10,
no.4, pp.509-514.
permittivity”, Physical Review Letters, May 2000, vol. 84, no. 18, pp. 4184-
4187.
metamaterials”, IEEE Microwave Magazine, Sept. 2004, vol. 5, no. 3, pp. 34-
50.
“Left Handed Coplanar Waveguide Band Pass Filters Based on Bi-layer Split
Ring Resonators”, IEEE Microwave Wireless Compon. Lett., vol. 14, no. 1,
[1.11] A. Ali and Z. Hu, “Compact Left-Handed Microstrip Line Based on Multiple
Layered Step Impedance Resonators”, Int. Journal of Electronics, vol. 94, no.
ring resonators”, IEEE Microwave Wireless Compon. Lett., vol. 14, no. 6, pp.
transmission lines”, IEEE Trans. Microwave Theory Tech., vol. 53, no. 4, pp.
CHAPTER 2
2.1 History
In 1968, the theoretical verification of negative index material and left handed (LH)
(μ) was postulated and pointed out by Veselago [2.1]. His research showed the
exhibition of the phase velocity direction opposite to the Poynting vector of the wave
Therefore, his study is to support the theory of negative index of refraction or the
permeability (μ<0) are still not done until three decades later.
(a) (b)
Figure 2.1 (a)The first artificial dielectrics lenses: Lattice of conducting disks
arranged to form lens by Kock and Cohn in1948 [2.2], and (b) J.B. Pendry designed
The first structure that showed negative permittivity in certain frequency bands was
proposed by Pendry et al [2.3, 2.4]. Owing to the ideas of Pendry, Smith [2.5]
permeability. Then more studies, such as Shelby et al [2.6] designed the periodical
structures by using a split-ring resonators and wire strips, have been investigated of.
Metamaterials (MTMs), their properties, and their applications have been appeared.
The Left-Handed Materials (LHMs), Double Negative (DNG) Materials, and Single
Negative (SNG) Materials, are the example names of the above mention. However,
Before study, it should be firstly clarified about the definition of MTM and its
than through composition [2.7]. Since the period is smaller than a wavelength,
study and control these unnatural material characteristics for the specific purpose in
magnetic field quantities are presented which describe the electromagnetic wave are
[2.8]:
B
E Ms (2.1)
t
D
H Js (2.2)
t
D e (2.3)
B m (2.4)
All quantities are real values in the functions of space and time. and are the
t
D E (2.5)
B H (2.6)
where ε0 is the free space permittivity (8.854x10-12 C/Vm2) and μ0 is the free space
Then, the time dependence is assumed as e+jωt. Therefore, the time derivatives in
equation (2.1) and (2.2) can be replaced as jω. Maxwell’s equations can be then
presented as [8,12]:
E jH M s (2.7)
H jE J s (2.8)
In the case of absence of sources, the equation (2.7) and (2.8) can be expressed
k E H (2.9)
k H E (2.10)
where the k is the wave vector along the direction of the phase velocity( k ).
28
Before the study of LHMs, the right hand rule in electromagnetism will be reviewed.
When the direction of both normal E and H fields are represented by the thumb and
the index finger of the right hand respectively, placed at right angles to each other
then the middle finger placed at right angle to both the fingers gives the direction of
propagation of the wave. In nature, all electromagnetic waves such as light are
followed this rule which can be stated in mathematical form as equation (2.11):
EH S (2.11)
S represents the Poynting vector and the direction of energy and wave
propagation.
Backward Waves
ε>0, μ>0 ε<0, μ<0
k
k
H H
S S
E E
General Materials Left-handed Materials
(Right-handed) n
Figure 2.2 The diagrams of electric field, magnetic field, wave vector (E, H, k) and
Poynting vector (S) for electromagnetic wave propagation in right-handed and left
As from the previous Maxwell equations, it can be assumed that in a medium where
the permittivity and permeability is negative (LHMs), the phase velocity will be anti-
parallel to the direction of wave propagation or energy flow. In another word, the
wave has a ‘negative phase velocity’ in that medium. Even though, the direction of
energy flow is always sent from the transmitter to the receiver, the phase moves in
the opposite direction, illustrated of the S vector and the anti-parallel of k vector by
Figure 2.2.
Assuming that the external supply of electromagnetic energy to the system is shut
off, the internal energy per unit volume of medium (W) is positive as illustrated in
equation (2.12-2.24).
In general medium, the internal energy per unit volume of the medium (W) by time
W .S dt (2.12)
S (r , t ) E (r , t ) H (r , t ) (2.13)
B D
.S E .H H .E H . E. (2.14)
t t
30
In general medium, the relations of electric flux density (D) and electric field
intensity (E) as well as magnetic flux density (B) and magnetic field intensity (H) are
defined as
t
D (t ) E (r , t ) (t t ') E (r, t ')dt ' (2.15)
t
B (t ) H (r , t ) (t t ' ) H (r , t ' )dt ' (2.16)
Then, the time varying electric field intensity at a quasi-harmonic field with median
E (r , t ) Re E e j0t 1
2
E E * (2.17)
D * D
*
B * B
*
E. E . H. H . 0 (2.18)
t t t t
1 D * * D B * B
.S E . E . H. H *. (2.19)
4 t t t t
where the field components can be analyzed via Fourier series expansion to
D d ( ) E jt
j ( ) E e (2.20)
t d t
B d ( ) H jt
j ( ) H e (2.21)
t d t
Finally, the Poynting vector and the internal energy per unit volume in medium can
be reformed as
31
1 d ( ) ( E .E * ) d ( ) ( H .H * )
.S (2.22)
4 d t d t
1 d ( ) 2 d ( ) 2
W E H (2.23)
4 d d
As a result, the entropy condition can be written as W > 0 or it can be rewritten in the
d ( ) d ( )
0 and 0 (2.24)
d d
From equation (2.24), it is impossible that the entropy condition can be happened
while the constitutive parameters are both negative. However, this can be occurred in
some frequency regions in order to compensate for their negative values. This
condition causes LHM as a causal and dispersive medium with frequency dependent
constitutive parameters.
As the wave number (k) is anti-parallel with the power flow in LHMs as well as
medium are reverse compared to the LH medium. The phase velocity is defined as:
~ ~ k
Vp k , where k (2.25)
k k
32
The power flows of LH medium are similar to the RH medium and can be presented
as
S E H (2.26)
1
2 S
P0 E H ds (2.27)
P0 = Power flow
In summary, the LH mediums provide the backward wave phenomenon due to the
reverse general direction of the phase velocity. The phase velocity is opposite of the
oriental direction while the power and group velocity are not affected [2.8].
E s H (2.29)
H s E (2.30)
33
while s is the sign that s=+1 in case of RH medium and s=-1 in case of LH medium,
respectively.
To understand a LH system, the equation of refraction index has been defined by the
r r (2.31)
The possible real number of the signs of ε and μ are (+, +), (+, -), (-, +), and (-, -)
which lead to a double positive (DPS), single negative (SNG) or double negative
will not generally show the negative refractive index. However, Ziolkowski [2.9]
used the mathematics to prove the square root choice that leads to a negative index
of refraction (NRI)
It can be noted that ‘η’ can still be positive while the values of μ and ε are negative.
However, in mathematics, ‘η’ for DNG materials is negative; for example η=-1 can
permittivity and permeability in terms of magnitude and phase are considered as:
34
j ,
r r e 2
where (2.32)
j
r r e
,
2
From equation (1.32), the refractive index and the wave impedance of the medium
r r e j n
(2.33)
Θ1 Θ1
Θ2 n = -1
Energy flow Wave vectors
(rays)
Figure 2.3 Energy flow and wave vector diagram between RH and LH interface
The energy flow vectors are in the same direction, while the wave vectors are in
Figure 2.4 represents the electromagnetic applications based on the signs of the
material permittivity, permeability, and refraction index at the interface between air
and each medium. There are four regions in the diagram. Plasma belongs to the
region with negative permittivity and positive permeability. Split rings belong to the
35
that when the two signs opposite, there is no wave transmission in medium. This is
because the wave vector becomes imaginary. When both parameters are positive,
μ
I. SNG Materials: ε<0, μ>0 I. DPS Materials: ε>0, μ>0
є
III.DNG Materials: ε<0, μ<0
DNG IV.SNG Materials: ε>0, μ<0
SNG
Figure 2.4 Material classifications according to є, μ pairs and η, type of waves in the
In the negative materials, the traveling wave passes from the air through the medium
and bends to the same side of the normal as the incident ray. This phenomenon is
called negative refraction index, given by the equation (2.31) and illustrated in
Figure 2.5.
36
n1 n2 n1 n2’
Θ2 RH LH
Θ1 Θ1 Θ2
The Snell’s law supports the wave propagation through LHMs that bended in the
wrong way [2.10]. In Figure 2.5, the refractive index of n2’=-n2 and the wave is
refracted to the opposite side compared to the ray propagating in Right-Handed (RH)
medium. Although the wave bends the opposite direction, the Snell’s law is still been
satisfied when a negative value of n is substituted and θ2<0 into the equation,
Due to its negative refractive index, the wave propagating travelling through a LHM
slab would be internally focused inside the slab and then create an image point after
leaving the slab. Pendry et al [2.11] stated the useful of this material property in
1stfocus
Source Image
(a)
(b)
Figure 2.6 (a) Planar lenses with the interface of negative refractive index material
(arrows are the wave vector in the medium) [2.11] and (b) the time averaged power
density of the Gaussian beam on xoy plane in lossy RH-LH (x is the distance in
Doppler Effect is the phenomenon that change in frequency of a wave when the
(a) (b)
Figure 2.7 The Doppler effect (a) Conventional RH medium (Δω>0), and (b) LH
Consider the source P moving along z direction, while the angular frequency of the
ej (t )
E ( z, t ), H ( z, t ) , where (, t ) t kr (2.35)
r
whereas k refers to the wave number in medium and r is the standard radial variable
In both cases, when source moves towards the positive z direction with velocity vs,
the wave position function of time is z=vst, which r=z in θ=0. Therefore, the seen
v
v s t 1 s s t
k
t kvs t 1 (2.36)
v p
39
vs
Doppler , with s (2.37)
vp
In RH medium, the sign s=+1 and Δω>0, the Doppler frequency is shifted downward
when the observer is at point P and shifted upwards when the observer position is at
the right hand of moving source, displayed in Figure 2.7(a). In LH medium, sign s=-
There are fundamental phenomena of DNG media by Veselago in 1968 [2.1] such as
negative.
In 1999, Pendry et al [2.3] proved that the negative effective permeability (μeff(ω))
can modify the permeability of the host substrate from an array of conducting non-
magnetic rings. The cause of bulk μeff(ω) variation for a very large positive value of
μeff(ω) at the lower resonance frequency and a significantly large negative μeff(ω) at
of μeff(ω) when the constituent unit cells are resonantly made. Schultz et al [2.5] are
the scientists who first realized the LH materials by creating a periodical array of
Before their success of realizing the LHMs, the attempts to produce the negative
permittivity materials were made earlier by Pendry [2.16]. In order to create the
structure to alter the permittivity with supporting substrate. The exhibition at the
frequency region in his experiment can show the simultaneously negative values of
both effective permeability μeff(ω) and effective permittivity εeff(ω). Then Schultz et
al [2.6] used these two concepts to create a LH structure. The wire strips and a mesh
of interspaced split-ring resonators are introduced for this achievement. The wire
strips generate ε while the split-ring resonators (SRRs) alter μ. Therefore, the
frequency dependent negative material with both negative parameters was realized.
It should be noticed that this would only happen under the condition that the size of
unit cell is considerably smaller than the smallest operation wavelength. As a result,
these periodic structures can give a uniform isotropic alteration of the base material
a « λ = 2πcoω-1 (2.38)
Figure 2.8 shows a genetic view of the periodical structure that gives an effective
λg>>a
(a) (b)
Figure 2.8 (a) Photograph of the Metamaterial cube [Physic today, June 2004], and
(b) Generic view of a host medium with periodically placed structures constituting a
MTM.
Pendry 1998 [2.16, 2.17] applied the individual properties of thin metallic wires
which can alter the effective permittivity of the host medium when excited
(a) (b)
Figure 2.9 Metamaterial structures: (a) a medium composed with metallic wire and
(b) thin wire lattice exhibiting negative εeff if E applied along the wires.
The array of thin metal wire, illustrated in Figure 2.9(a), presents the effective
negative є. If the electric field E is applied along the wires, the induced current along
the wires will generate equivalent to electric dipole moment. There are two factors
affecting the electron movement in a wire radius r; the average electron density neff
nr 2
neff (2.39)
a2
From Ampere’s law, a current flowing through a wire produces a magnetic field,
where the direction of the field is depended on the direction of the current. In
addition, the presence of the magnetic field alters the momentum of electrons. The
I r 2 nve
H ( R) (2.40)
2R 2R
0e 2 r 2 n a
meff ln (2.41)
2 r
It is noticed from equation (2.40) and (2.41) that the bigger radius of the wire the
more effective mass of the electrons. This observation is later investigated in the
neff e 2 2c02
2
(2.42)
0 meff a
p
a 2 ln
r
In equation (2.42), it is noticed that decreasing the effective mass provides a large
shift in the plasma frequency. Moreover, in order to maintain the wire array as a
homogenous material, the wire radius must keep small as compared to the lattice
dimension (a). Therefore, the effective ε of the composite medium can be evaluated
permittivity is derived as
p2
eff 1 (2.43)
2
From equation (2.43), it has been appeared that the permittivity is negative when
Because the specific property of SRRs embedded in a host medium can give the bulk
composite permeability and become negative in a certain frequency band, the study
of the region above the SRR resonant frequency has been widely observed.
E
H
a k
Figure 2.10 Metamaterial structures: Split ring resonators lattice exhibiting negative
While applying the magnetic field H perpendicular to the plane of the ring, the
induced currents then are generated around the ring equivalent to the appearance of
the magnetic dipole moments [2.3]. The permeability frequency function is formed
as
F 2
eff ( ) 1 2 (2.44)
02m j
3a
0m c (2.45)
2wr 3
ln
d
F is the filling fraction of the SRR, while ς is the damping factor due to metal loss,
2
r 2aR'
F and (2.46)
a r 0
Figure 2.11 some geometries of SRR used to realise artificial magnetic materials
[2.7]
The SRR structure has a magnetic response due to the presence of artificial magnetic
dipole moments by the ring resonator. At a resonant frequency range, these artificial
magnetic dipole moments are larger than the applied field which leads to the
presence of the real part of negative effective permeability, Re (μeff). In lossless case
(ς=0), the range around the resonant frequency providing negative permeability is
0m
under the condition as 0 m p , where ωp is the plasma frequency of
1 F
the SRR particle. In other words, in the medium the discontinuity of the dispersion
relation of the permeability is occurred between ω0m and ωp because of the negative
produced by the splits of rings. The strong capacitance between the two concentric
rings helps the flow of current along the SRR configuration. Since in the SRR the
capacitive and inductive effects nullify, the μeff has a resonant form. At resonant
frequency, owing to the capacitive effects due to the gap interacts with the inherent
47
external magnetic field and the electrostatic fields within the capacitive structure.
Therefore, normally the experiments are focused on a certain frequency band which
is around and above the resonance frequency in order to get the negative effective
permeability [2.18].
Depicted in Figure 2.12, is the first LHM prototype designed by Smith et. al. [2.5].
In this work, the combined particles of the thin wire structure and SRR structure
appeared an overlapping frequency range that have both negative permittivity and
structure, the pass band is presented at the frequency range of interest that the
(a) (b)
Figure 2.12 the first DNG metamaterial structures [2.5] Smith et al., 2000-2001: (a)
Figure 2.13 Topology of CSRRs and the stack CSRRs, E is parallel to the CSRRs
plane [2.19].
In 2004, CSRRs are firstly introduced by Falcone et al.[2.20]. The CSRRs, a dual
of slots which is the same dimensions as the corresponding SRR. By the principle of
duality, the CSRRs properties are in dual relation of the SRRs properties. The SRRs
behave as a magnetic point dipole, whereas the CSRRs present an electric point
dipole with negative polarization. In CSRRs, the E field is applied parallel to the
CSRRs plane in order to generate a strong electric dipole which affects the CSRRs
resonant frequency [2.15]. The CSRRs, as shown in Figure 2.13, r can be used to
obtain the effective permittivity of a bulk medium. Both SRRs and CSRRs present
The CSRRs can be formed in planar transmission media by etching these resonators
Figure 2.14(a). The HFSS simulation is used for design. The 50Ω line is chosen to
49
match the port impedance. This CSRRs structure provides the inhibition of signal
(a) (b)
Figure 2.14 Geometry of the CSRRs (a) with and (b) without capacitive gap
h=1.39mm)
Considering on Figure 2.15(a) and (b), at the resonant frequency, the dip in
to zero degree. At the frequency band above its resonant frequency, the CSRRs
0.00
-5.00
-10.00 S11
-20.00
-25.00 S21
-30.00
-35.00
-40.00
1.00 2.00 3.00 4.00 5.00 6.00 7.00
Frequency [GHz]
(a)
SAS IP, Inc. XY Plot 3 HFSSDesign1 ANSOFT
0.00
-25.00
-50.00
phase S21 [deg]
-75.00
-100.00
-125.00
-150.00
-175.00
1.00 2.00 3.00 4.00 5.00 6.00 7.00
Frequency [GHz]
(b)
(c)
Figure 2.15 The simulated results of (a) scattering parameters, (b) the output phase,
and (c) the effective permittivity of a combined CSRRs structure and microstrip line
51
microstrip transmission line, CSRRs is widely used for applying to the artificial LH
series capacitive gap etched in the conductor strip and loaded with CSRRs which
etched in the ground plane. The unit cell CSRRs transmission line is shown in Figure
etched periodically on the microstrip line to nullify the tank inductance and obtain
LH wave propagation.
0.00
-5.00
-10.00
S11 and S21 (dB)
-15.00
S11
-20.00
-25.00
-30.00
S21
-35.00
-40.00
1.00 2.00 3.00 4.00 5.00 6.00 7.00
Frequency [GHz]
(a)
SAS IP, Inc. XY Plot 3 HFSSDesign1 ANSOFT
200.00
150.00
100.00
phase of S21 [deg]
50.00
0.00
-50.00
-100.00
-150.00
-200.00
1.00 2.00 3.00 4.00 5.00 6.00 7.00
Frequency [GHz]
(b)
Figure 2.16 The simulated results of (a) insertion, return loss, and (b) phase of S21
The LH passband range is then appeared, shown in Figure 2.16(a). At the resonant
frequency of CSRRs, the advance phase 90 degrees is obtained which is one of its
specific properties of the LHMs, illustrated in Figure 2.16(b) [2.13], [2.20] The
chapter.
Since CSRRs presents stop band characteristic when applied with microstrip, it has
been used to develop the filter functions. The recent work [2.21] presents a wide stop
band filter by using the co-operation of the different CSRRs dimensions. It is found
that the size of CSRRs relates to its resonant frequency in opposition. In the other
words, the small size of CSRRs is presented; the higher resonant frequency is
obtained.
Figure 2.17 demonstrates the two CSRRs lengths after tuning to merge their resonant
frequencies. The two resonant frequencies are now presented as one stopband,
This application can be further applied to bandpass filter to widen the stop band as
(a) (b)
Figure 2.19 (a) Photograph of the patch antenna loaded with CSRRs and (b)
The compact antenna loaded with CSRRs and reactive impedance surface (RIS) is
shown in Figure 2.19(a). The CSRRs are modeled as a shunt LC resonator which
creates resonant frequency. The photograph shows the patch size around
54
(a)
(b)
Figure 2.20 (a) the ultra-wideband antenna with quadruple-band rejection and (b)
The four rejection bands at 2.6, 3.5, 5.5, and 8.2GHz are selected by LC resonator of
the depicted SRRs and CSRRs on patch and ground. The resonant frequency of each
interferences.
55
2.5 References
negative value of ε and μ”, Soviet Physics Uspekhi, vol. 10, pp. 509-514, Jan
1968.
[2.2] W.E. Kock., “Metallic delay lenses”, Bell System Technical Journal, 17:58–
82, Jan1948.
[2.4] J. B. Pendry, “Negative refraction makes a perfect lens”, Phys. Rev. Lett, vol.
negative index of refraction”, Science, vol. 292, pp. 77-9, Apr 2001.
2005.
[2.10] M. C. K. Wiltshire, “Bending of light in the wrong way”, Science 292: pp.
60-61, 2001
[2.11] J. Pendry, “Negative refraction makes perfect lens”, Physical Review Letters
[2.12] T. J. Cui et.al., “Study of Loss Effects on the Propagation of Propagating and
2004
[2.14] J. D. Kraus and R. J. Marhefka, Antennas 3rd edition, McGraw Hill, 2001
ISBN: 007123201X.
Jun 1998.
plasmons in thin wire structures”, J. Phys. Condens. Matter, vol. 10, pp.
4785-4809, 1998.
Resonators”, IEEE Microwave and Wireless Components Letters, vol. 14, no.
[2.21] P. Su, X. Q. Lin, R. Zhang, and Y. Fan, “An Improved CRLH wide-band
2, Feb 2012.
[2.23] N. –I. Jo, C. –Y. Kim, D. –O. Kim, and H. –A. Jang, “Compact Ultra-
CHAPTER 3
INTRODUCTION
Metamaterial transmission lines (MTM TLs) are artificial lines where Right-Left
what can achieve in conventional transmission lines. These artificial lines, formerly
proposed in the early 2000’s [3.1-3.5], are inspired on MTMs which exhibits similar
properties and, in some cases, fabricated using identical constituent particles [3.6,
3.7].
reactive elements. Since these elements are electrically small, the conditions to
achieve homogeneity can also be achieved, that is a small period compared to signal
parameters μeff and εeff are studied. However, homogeneity is not a fundamental
From the point of view of microwave circuit design, the advantages of MTM-TLs
rely on miniaturization and on the possibility to control the dispersion diagram and
as artificial lines, consisting on a host line loaded with reactive elements, with
such lines [3.8, 3.9]. Notice that according to this, there is no need for a minimum
number of unit cells to implement these artificial lines. Indeed, in most of the cases,
a single cell is considered since this reduces line dimensions, as will be shown later.
With regard to the implementation of MTM-TLs, there are two main approaches:
LC-loaded lines or dual transmission line [3.1-3.5, 3.10], and resonant type
approaches [3.6, 3.7]. In LC-loaded lines, proposed by Eleftheriades [3.2] and Caloz
[3.1], the key element to achieve left-handedness consists of a host line loaded with
series capacitances and shunt inductances. These lines can be implemented by using
components such as series gaps, interdigital capacitors, grounded stubs or vias. The
Resonant type MTM-TLs can be implemented by loading a host line with SRRs and
shunt inductive elements [3.11, 3.12], or, alternatively, by loading a host line with
CSRRs and series capacitances [3.13-3.15]. Both LC-loaded lines and resonant type
negative values of the effective permeability and permittivity of the medium for both
elements dominate over the per-section capacitance and inductance of the line [3.1,
3.13, 3.14, 3.16]. Therefore, both line types are useful for the implementation of the
In this chapter, the CSRRs, a dual counterpart of SRRs, are formed to show a left
plane of the microstrip transmission line. Due to the presence of CSRRs, the
This phenomenon interprets the negative effective permittivity of the medium. The
series capacitive gap is added to the structure in order to obtain the negative
permeability as well as a band pass with left handed wave propagation [3.13-3.15].
(a) (b)
Figure 3.1 Topology of (a) SRR and (b) CSRRs with relevant dimensions.
Therefore, SRR and CSRRs based transmission lines act as the frequency selective
and Limitations
properties of LHMs and the physical realization of these materials are reviewed.
transmission line is explained in the next topic as well as its characterization. Then
the design and implementation of the CSRRs TL and their applications are
presented.
assumes that the periodic loaded elements are infinity. In this case, the dispersion
periodic structure analysis for microwave periodic networks. ABCD matrix is used to
describe the transmission characteristics of a two port network through its input and
Vn A B Vn1
I C D I (3.1)
n n1
For a wave propagating through a line, the voltage and current at the cells (n+1) and
j (3.4)
where α is the attenuation constant and β is the phase propagation constant. From the
A e d B Vn1 0
(3.5)
C D e d I n1 0
A e D e BC 0
d d
(3.6)
Making use of the relation of reciprocal for any two- port passive network,
AD-BC=1 (3.7)
Then to analyze the propagation of a very short LH TL model, the planar equivalent
circuit of a unit cell is introduced. The LH transmission line model has a section of a
transmission line of length d and loaded with a series impedance Z and a shunt
admittance Y. For a lossless cell the equivalent circuit is shown in Figure 3.2.
63
Figure 3.2 Planar equivalent circuits for a LH transmission line periodic cell [3.1]
In the following analysis, Z, Z0 and Y are the load series impedance, characteristic
hosting transmission line. The periodic length of the cells is assumed very small
compared to the guided wavelength; otherwise the effect of the hosting elements
A B 1 Z cos jZ 0 sin
2 2 1 0
C D 2 Y 1
0 1 jY 0 sin cos
2 2 (3.9)
cos 2 jZ 0 sin Z
2 1
2
jY sin 0
cos 1
0 2 2
A cos ZY cos 2 j Z 0Y ZY0 sin
1 1
(3.10)
2 2 2 2
Z 2 1 Z
2
B Z cos Y cos 2 Z 0 sin 2 j ZYZ 0 2Z 0 2 Y0 sin
2 2 2 2 2 2
(3.11)
C Y cos 2 jY0 sin (3.12)
2 2
D=A (3.13)
where Z0 and Y0 are the characteristic impedance and admittance of host transmission
respectively, and θ is the propagation angle of the hosting transmission line given as
d (3.14)
From (3.8) and (3.10) to (3.13), the propagation condition can be written as
For propagation criterion given when α=0, β≠0, the dispersion equation is
65
1
cosd cos Z LYL cos 2 i Z 0 Z L Y0YL sin
1
(3.16)
2 2 2 2
The dual of conventional transmission line is used to analyze the transmission line
approach of LHMs, which the equivalent circuit is shown in Figure 3.3. Comparing
pass band of frequencies in nature, whereas the LH transmission line exhibits high-
(1/jωC’)dz
(1/G’)dz
(1/jωL’)dz
Y
(1/R’)dz
The wave number γ(ω) of such a line is derived from per unit length impedance
j Z Y
(3.17)
G jC 1 R jL1
66
1
2
R G 2 LC 2 L G R C R G 2 LC
2 2
(3.18a)
RG 2
LC LG R C
2 2 2
1
2
R G 2 LC 2 LG R C R G 2 LC
2 2
(3.18b)
RG 2
LC LG R C
2 2 2
where G, C , R, L are the values per unit length quantities. The equations (3.18a)
and (3.18b) show the attenuation factor α(ω) and propagation factor β(ω) where the
negative sign in β(ω) indicates a negative phase velocity and its nonlinearity
given by
R jL
Z c Z / Y (3.19)
G jC
0 (3.20a)
1
(3.20b)
LC
β(ω) exhibits the dispersion diagram shown in Figures 3.4 with the anti-parallel
hyperbolic phase and group velocities which indicate the left-handedness [3.1].
will then be
d d p p 1
p 2 (3.22)
d d LC
where p represents the physical length of the line. The group delay is in the relation
v p v g 2 LC is in a LH TL.
Figure 3.4 shows the dispersion diagram of a purely LH TL and purely RH TL. The
LH area is illustrated on the negative area of dispersion (red), while the RH area is
βd
Figure 3.4 the ω-β diagram in a purely LH (red) and RH (blue) TLs
68
the dispersion from the PRH contribution is increased with frequency. From
dispersion graph, the backward wave propagation is existed along the LH line as
presenting in negative values (red line). In this range, the line becomes a backward
Both μ(ω) and ε(ω) values in LH material can be derived from the analogy between
the plane wave solution in homogeneous material (Maxwell equations) as well as the
wave along the LH TL (TL equations). In RH TL, the impedance and admittance of
the material are supposed to Z j and Y j , respectively, which they have
1 1
2 0 (3.23a)
C C
1 1
2 0 (3.23b)
L L
These above equations both demonstrate the negative of μ and ε in a LHM together
with the relative dispersion of the frequency dependence by 1/ω2. Although such a
material does not seem to exist in nature, it might be realized artificially as MTMs
can be concluded that the LHMs with the dispersive constitutive parameters given by
equations (3.23a) and (3.23b) satisfy the generalized entropy conditions for
1 1
0 and 0 (3.24)
C 2 L 2
Then, the refractive index is derived by taking the square root of the product in
For 1 / LC , it yields as:
c0
1
r r 0 (3.25)
c0 LC 2
LC
This last equation confirms the negative value of η(ω) associated with the reversal of
In this section, the transmission line approach to CRLH MTMs will be discussed.
Figure 3.5. For simplicity, only the lossless case of transmission line will be
Z’
L’RΔZ
C’L/ΔZ
Y’
C’RΔZ L’L/ΔZ
ΔZ
Homogeneous Case
The homogeneous model of a CRLH lossless transmission line shown in Figure 3.5
consists of an inductance L’R in series with a capacitance C'L and a shunt capacitance
by j Z 'Y ' , where Z’ and Y’ are the per-unit length impedance and per-
1
Z ' ( ) j L' R ,
C ' L
(3.26)
1
Y ' ( ) j C ' R
L' L
1 L' C'
( ) s( ) 2 L' R C ' R R R (3.27)
L' L C ' L L' L C ' L
2
71
where
1 1
1 if 1 min ,
L' C ' L' L C ' R
R L
s( ) (3.28)
1 1
1 if 2 max ,
L' C ' L' L C ' R
R L
The equation (3.27) of phase constant β can be made in the form of purely real or
purely imaginary which is upon the sign of radicand. In the case of frequency range
which β is purely real, a pass band is present since γ=jβ. While a stop band occurs in
This stop band is a unique characteristic of the CRLH transmission line, which is not
clearly presented from the Purely Right-Handed (PRH) and the Purely Left-Handed
(PLH) cases. In CRLH TL, the LH characteristics presents at lower frequencies and
the RH properties at higher frequencies. From the dispersion diagram, the group
can be extracted.
72
ωr2
ω0
ωr1
Considering this diagram, it is found that in a PRH transmission line, vg and vp are
stop-band that occurs when γ is purely real for a CRLH transmission line. In the case
of the series and shunt resonances of the CRLH transmission line are equal which is
This means the LH and RH contribution are exactly balanced, shown in Figure 3.7
The impedance matching is achieved at the frequency where the Bloch impedance
coincides with the reference impedance of the ports; while the phase matching
occurs at those frequencies where the phase shift of the structure is a multiple of π.
73
-20
dB(S(1,1))
dB(S(2,1)) -40
m3 m4 m2
-60
1 2 3 4 5 6 7 8 9 10
m3 m2
freq, GHz
freq=1.310GHz freq=7.710GHz
dB(S(2,1))=-81.857 m4 dB(S(2,1))=-64.596
freq=3.180GHz
dB(S(1,1))=-84.899
Valley
The ripple occurring in Figure 3.7 is consequence of periodicity and it may limit
filter performance, which means the ripples increase at the edges of the band since
the Bloch impedance takes extreme values. The electrical simulation has been
obtained through ADS. From the example, the 30-cell CRLH TL has the resonate
frequency at f0=3.18GHz. The left handed cut off (fL) is 1.31GHz and the right
f0
βd
Figure 3.8 Dispersion diagram of the 30-stage CRLH TL balance case in Figure 3.7
The dispersion graph of the LH and RH region are connected in one line in the case
of balanced CRLH TL. The LH band is presented on the βd<0, while βd>0 is
Under the condition in equation (3.29), the propagation constant is reduced to the
simpler form:
1
R L LR C R (3.30)
LL C L
where the phase constant of CRLH distinctly splits up into the RH phase constant βR
From the dispersion diagram, it can be noted that the phase velocity v p / of
diagram illustrates the dual characteristic of the CRLH TL which at low frequencies
dominantly RH. In balanced case of CRLH TL the dispersion diagram indicates that
unbalanced 1 balanced 1
0 (3.31)
4 LR C R LL C L LC
in the balanced case, the connecting of dispersion graph of the LH and RH line
appeared. This means the curve of balanced CRLH TL dispersion does not have a
advance 0 occurs in the LH frequency range 0 , and phase delay 0
characteristic impedance is
LL
ZL (3.32b)
C L
76
LR
ZR (3.32c)
C R
ZL and ZR represent the PLH and PRH impedances, respectively. Consider the
equation (3.32a), the characteristic impedance for the unbalanced case is frequency
set up:
2 Z Y (3.33)
Z
Z0 and (3.34)
Y
Z 1
LR 2 (3.35a)
j C L
Y 1
C R 2 (3.36b)
j LL
77
η
ωr1 ω0 ωr2
(a)
f0
(b)
Figure 3.9 (a) Typical index of refraction for the balanced (green) and unbalanced
CRLH TL (red-orange), (b) the refraction index of 30-stage CRLH TL in Figure 3.7
negative, while it shows the positive values in the RH range (a) [3.1]. In balanced
78
case CRLH TL, the refractive index is equal to zero on the resonate frequency,
transmission line technologies. The CSRRs particles are etched in the ground plane,
underneath the conductor strip so that the CSRRs can be excited by the time varying
electric field of the quasi-TEM signal propagating in the line, while the time varying
magnetic field in CSRRs is applied parallel to the plane of the particle [3.15, 3.18].
The topology of one cell CSRRs can be analyzed by lumped element equivalent T-
circuit model depicted in Figure 3.10 which ignores the losses and inter-resonators
coupling. The model is valid under the condition of small electrical size of CSRRs.
ZS ZS
L/2 Cg Cg L/2
C
ZP
LC CC
(a) (b)
Figure 3.10 Basic cell of CSRRs-based transmission line (a) and equivalent circuit
model, (b) The upper metallization is depicted in black; the slot regions of the
capacitance Cc [3.19-3.24], whereas their coupling to the host line is modeled by the
capacitance C. The series gaps etching above the CSRRs to enhance line-to-CSRRs
coupling are replaced by the capacitance Cg. The series impedance must be
may not be omitted to describe the accurate structure. These parameters will be used
to model the layout of the CSRRs unit cell to obtain some specific targets.
The study of phase shift of the elemental cell and Bloch impedance; as shown in
equations (3.37) and (3.38), confirms the highly dispersive of the structure:
Z s j
cos 1 (3.37)
Z p j
Z B Z s j Z s j 2Z p j (3.38)
where Zs and Zp are the series and shunt impedance of the equivalent T-circuit model
of Figure 3.10, respectively. The ZB and are the key electrical characteristics of
artificial LH lines by CSRRs method, where has to set as the condition of =βl,
β is the phase constant for the Bloch waves, at the operating frequency, while l is the
period of the structure. Therefore, the structure exhibits a band pass behavior with
backward wave propagation in the allowed band from the analysis of equations
(3.37) and (3.38). The flexibility for the artificial lines is dependent on these four
The analysis of the limitation of these circuit parameters are described as follows. In
order to obtain the limit of this LH transmission band, the ZB or have to be forced
according to:
1 1
fL (3.39)
2
4
Lc C c
1 4
C g C
1
fH (3.40)
2 Lc Cc
fL and fH represent the lower and higher frequency of the interval, respectively. At
these frequencies, both the phase and Bloch impedance take extreme values, for
instance; ZB→∞ and =0 at fH, whereas ZB=0 Ω and =π at fL. In order to wider
the bandwidth, these fL and fH have to be set apart from the cut off frequency (fc) as
described as below;
From equations (3.39) and (3.40), the capacitance value of the series gap can be
inferred by
1 1 cos c
Cg (3.41)
2c Z c 1 cos c
81
where c 2f c . Then, the other parameters as follows can be extracted by using
Lc
Z c 1 cos c c H2 L2 H2 c2
2 1 cos c H4 c2 L2 (3.42)
1
Cc (3.43)
Lc H2
C
2 H2 c2 L2 1 cos 2 c
Z c c 1 cos c L2 2 H2 c2 L2
2
c
2
H (3.44)
where L 2f L and H 2f H . As from equations (3.42) to (3.44), the values of
Cg, Lc, and Cc should be positive to provide the condition of L c H . Since the
C which has to force in the real positive number as shown in the condition;
H2 c2 L2
1 cos c 2 2 2
c H L2
(3.45)
The value of the operating angular frequency, ωc, and phase, øc, only limit the range
parameter, the intrinsic limits for ωL and ωH are 1 / 2c L c and H c which
might not possible for physical implementation. Thus at the operating frequency, the
82
possibility to control phase and line impedance with a single cell structure will be
analyzed later.
Another interesting point is the transmission zero frequency, fZ, which is extracted
from the simulation or experiment results. At the transmission zero frequency, the
after consider the mentioned condition, the latter condition can be demonstrated as;
Another frequency point, where the second allowed (right handed) band starts up, is
the operating frequency fc. At the operating frequency, the phase variation can be
1
fZ (3.47)
2 Lc C Cc
1
fc (3.48)
2 LC g
In order to extract the parameters at the resonant frequency of the CSRRs, the null
condition of the shunt admittance has been forced, while the impedance from the
input port is set by adding the output impedance (50Ω) and the reactive impedance
of the L and Cg. Using equations (3.40), (3.47), and (3.48), the three elements of the
shunt reactance can be extracted. Because the line inductance, L, cannot be ignored,
83
It is noted from all the previous equations that the bandwidth is limited by the
transition bands at the lower and upper edges of the LH band, in addition, the in-
frequency is forced to be nearer the lower edge of the band, the gap capacitance Cg
appearance of a highly selective filter is provided at the lower band edge. As refer to
equations (3.39) and (3.40), it appears that the bandwidth is dependent on the values
of Cg and C, and it is difficult to design wide (or even moderate) band structures
[3.22].
Model
All previously published works of CSRRs have assumed that the coupling between
the rings in adjacent cells is dominated by the electric coupling and magnetic
coupling can be neglected [3.19-3.24]. However, it is revealed that there are strong
circuit performance. This section presents a new equivalent circuit model. The model
has been developed and verified both numerically and experimentally to take into
account of the magnetic coupling between the rings in adjacent cells, which has been
neglected so far.
84
To observe this effect, a simple 2-cell CSRRs microstrip TL, shown in Figure 3.11
was designed and analyzed, aiming to understand the coupling effects between the
CSRRs.
Rext
d
C
(a)
Cell separation=0.2mm
(b)
Figure 3.11 (a) The topology of CSRRs unit cell and (b) 2-cells CSRRs microstrip
The example in Figure 3.11 demonstrates the qualification of the 2-cell CSRRs when
added the series gaps on the strip line. As shown in this figure, the narrow pass band
is occurred and this also generates the left hand wave. The Rogers RO3010 substrate
has been chosen (thickness h=1.27mm, εr=10.2) for this electromagnetic simulation.
85
In the design the transmission line strip width is of 3.2mm, the thickness of substrate
1.27mm, the external radius of the ring (Rext) 3.5mm, the metal ring width (c)
0.4mm, and the gap between the inner and outer rings (d) 0.3mm. The total length of
the 2-cell CSRRs line is 15mm. Separation between the two cells is initially set to
Electric_field
30000
25000
Volts per Metre
20000
15000
10000
5000
0
0.2 0.4 0.6 0.8 1
distance (mm)
(a)
Magnetic-field
250
Amperes per Metre
200
150
100
50
0
0.2 0.4 0.6 0.8 1
distance (mm)
(b)
Figure 3.12 Field variations with ring separation of the two CSRRs (a) Electric field
Electric and magnetic fields at the centre point between the two rings are shown in
Figure 3.12(a) and (b), respectively, illustrating that there exists strong magnetic
coupling between the adjacent cells apart from electric coupling especially when the
cell separation is reducing. This magnetic coupling, which has so far been neglected
in all previously published works, must be included in the equivalent circuit model
A new equivalent circuit model for a 2-cell CSRRs TL has been developed as shown
magnetic coupling effects between the cells. The results in Figure 3.14 prove that
magnetic coupling has an influence in changing S-parameters by giving the cut off
offer more accurate results (0.2mm is set as a gap between the two CSRRs).
Csub Csub
CM
L2 C2 L2 C2
It can be seen that the two are agreed reasonably well (loss mechanism was not
included in the circuit model). Extracted element values with three different
87
separations between the two cells are given in Table 3.1. It can be seen, as expected,
from the table that CM and N change their values whereas all other element values do
not vary while the distance between the two CSRRs cells changes.
-5
-10
full wave simulation
S11 and S21 (dB)
-15
circuit simulation with magnetic coupling
-20
-25
circuit simulation w/o magnetic coupling
-30
-35
-40
2.0 2.2 2.4 2.6 2.8 3.0
Frequency (GHz)
(a)
(b)
Figure 3.14 (a) The S-parameters of the 2-cells CSRRs TL with cell separation of
0.2mm and (b) Smith Chart, equivalent circuit model and full wave simulation.
88
It has been found that the magnetic coupling which takes into account in the case of
the two adjacent CSRR cell can provide the S11 graph obviously closer to the S11
from HFSS full wave simulation at 2.27GHz. If considering the circuit simulation
without taking into account of the magnetic coupling, the S11 graph (dash line)
displays the centre frequency at 2.3GHz which is not exposed the precise centre
frequency. On Smith Chart, the S11 of the new equivalent circuit provides a better
match of the graph especially on the passband than the equivalent circuit without
magnetic coupling. The rational of transformer proves that there is strong magnetic
As mention at the beginning of this chapter, a CSRRs applied with series gap on host
line exhibits LH properties. In order to verify the Left handed characteristics, the LH
transmission line, shown in Figure 3.15, was represented for consideration. The 4-
89
cell CSRRs TL has been fabricated on the Rogers RO3010 substrate, with a
placed by cells 1-2 is 1mm, cells 2-3 2mm, cells 3-4 1mm, as shown in Figure 3.15
(a).
1 2 3 4
N N
(a)
(b)
Figure 3.15 The prototype of 4-cell CSRRs TL, (a) backside view and (b) top view.
90
CSRRs TLs has a total physical length of 35 mm and were fabricated The S-
Figure 3.16. The lower passband and higher passband of the CSRRs TL are
Figure 3.16(a), the centre of operating area gives highest transmitted signal with
lower loss; however there are some substrate and metal losses. The insertion loss at
this point is approximately 2.49dB. In addition, the phase compression has appeared
within the LH passband (2-3GHz) of the CSRRs TL, shown on Figure 3.16(b).
-10 S11
S11 and S21 (dB)
-20
S21
-30
-40
-50
-60
1 2 3 4 5 6 7
Frequency (GHz)
(a)
200
Phase S21 (degrees)
100
-100
-200
1 2 3 4 5 6 7
Frequency (GHz)
(b)
Figure 3.16 The measured results of (a) S-parameters of 4 CSRRs cells and (b) Phase
of S21.
91
The circuit model as shown in Figure 3.13 is used here again to extract the loading
parameters (C1, Csub, C2, L2, and L1). These parameters were used to model a cell of
CSRRs before cascading them as 4 cells to fit with the measurement. By fitting the
curve with the measurement results, the extracted parameters can be found as
CM=0.59pF), seen in Figure 3.17. The S parameters (S21 and S11) of both
measurement and simulation agree well. The passband has been correctly predicted.
0
S11 and S21 (dB)
-20
Circuit simulation
Measurement
-40
-60
1 2 3 4 5 6 7
Frequency (GHz)
Figure 3.17 The S-parameters (S21 and S11) by circuit simulation and measurement
The discrepancies between the simulated and measured data in the stop band are
probably due to the fabrication tolerance in our PCB laboratory, especially the
alignment between the metal strip on the top of the substrate and the CSRRs on
ground plane. However, the new equivalent circuit model that takes into account of
92
magnetic coupling effect between the CSRRs, which has been neglected so far in
criterion, shown in Figure 3.18. By taking the simulated S parameters, the dispersion
of the 4-cell CSRRs TL can now be plotted and be shown in Figure 3.18. It can be
seen that the ω-β dispersion diagram can be used to confirm LH and RH
(below lower cutoff frequency fc1) whereas the undesired passband (4.7-7GHz) is in
LH Passband
LH RH
fc1 fc2
Stop Band
The dispersion graph exhibits a very different behavior of the CSRRs TL especially
in the LH passband that provides higher dispersion slope comparing to the RH area.
This means when the input wave gets close to the left hand side of the lower cutoff
(fc1), the observed wave is obviously transformed the shape of the original wave
since the high dispersion which cannot be predicted as in the ideal TL case. This
behavior is explained by the fact that the backward wave effect of the periodic
(artificial) MTM transmission line which becomes dominant over the ideal TL. At
the dispersion effect higher than fc1; the dispersion ω-β curve exhibits a zero slope
3.5 Conclusion
In this chapter, we have derived the three-type TL; the purely LH TL, the CRLH TL,
and the CSRRs TL. The dispersion properties versus frequency of them have been
analyzed, including their phase and group velocities. In LH area, the sign of phase
backward wave propagation. Moreover, it is found that both CRLH and CSRRs TL
provide LH and RH area. In balanced CRLH TL case, the LH and RH area can be
has been analyzed in many years, the magnetic coupling is still not taken into
account to the equivalent circuit. This chapter also recommends the new equivalent
circuit model of the CSRRs TL under the two more unit cell adjacent. The simulated
scattering parameters are presented in the case of with and without the magnetic
coupling. The results show that the magnetic coupling should be added in the
3.6 References
[3.3] C. Caloz and T. Itoh, “Novel microwave devices and structures based on the
Line”, IEEE Trans. Antenna and propagation, vol. 52, no.5, pp. 1159-1166,
May 2004
and tunable group velocity”, Phys. Stat. sol., no. 4, pp. 1219-1226, Mar 2007
2006
metamaterials”, Optics Letters, vol. 29, no. 22, pp. 2623-2625, Nov 2004
97
Line”, IEEE Transaction on Antennas and Propagation, vol. 52, no. 5, pp.
2005
stacked complementary split rings resonators”, Optics Express, vol. 17, no. 3,
[3.19] M. Gil, J. Bonache, I. Gil, J. G. Garcia, and F. Martin, “On the transmission
split rings resonators”, Int. J. Numer. Model, pp. 87-103, John Wiley &Sons,
Ltd., 2006
[3.24] G. Siso, M. Gil, J. Bonache, and F. Martin, “On the dispersion characteristics
Oct 2007
99
CHAPTER 4
INTRODUCTION
Digital communications have become more and more important in our life; as a
result highly complex data processing is required. Group delay seems to be a main
There have been many works in analyzing and suppressing group delay on signal
distortion [4.1-4.8] such as a loaded transmission line with a varactor diode [4.5], a
surface acoustic wave (SAW) [4.6], and a magneto-static wave (MSW) [4.7, 4.8]. In
signal delay has been used in signal processing instead of simply suppressing the
lines in group delay has not been well developed. Due to its compactness and unique
electrical properties, the prototype of the 4-cell CSRRs TL has been investigated as a
passive delay line. The simulated and measured group delay and signal delay results
are well established with the dispersion relation of CSRRs TL. Also, a significantly
Furthermore, the active tunable delay line has been studied by embedding the 4-cell
100
CSRRs TL with varactor diodes. By observing a fixed frequency, the group delay
can be obtained.
measurement of how long the signal, in time, takes to transit a system. Normally, it
H(jω)
The group delay (G.D.); represented as τ, is a time dimension which relates to the
( )
G.D. ( )
(4.2)
where φ and ω are the total phase shift and the angular frequency, respectively.
Group delay; sometimes called “envelope delay”, also can be expressed as the form
can be described as the time delay of the amplitude envelope of a narrow group at
H ( j ) e j (4.3)
Which shows that the envelope of the signal delay is a form of group delay.
Dispersion is the phenomenon where the original wave change of waveform after
propagating through a medium which is dependent upon that wave’s frequency. This
In order to compare the group delay of CSRRs TL, the fundamental mode of
L
CSRR RH LH (4.4)
R
where ωR and ωL are the cut off frequency of right-handed and left-handed
passbands.
The first term is represented as the conventional part of the line, while the last term
phase velocity (vp) and the corresponding group velocity (vg) are
2
vp
L (4.5)
c2
vg c d / d
1
c
L
Then, the group delay in a CSRRs TL on a LH area can be given as [4.20, 4.21];
d d N
g (c ) N L2 (4.6)
d d c
In the case of a conventional TL, the group delay is given only by the RH effect
1
g l (4.7)
R
As shown in equation (4.6) and (4.7), the group delay of any signals travelled
through a CSRRs TL can be varied in the LH area which is dependent on the input
signal frequency (ωc) and a number of CSRRs cells (N). Whereas, the delay time of
certain frequency band, a CSRRs TL can provide different time delay by frequency
function. Therefore, by fixing the parameter N with 4-cell CSRRs in this chapter, a
TLs, the group delay and their signal delays of the 4-cell CSRRs microstrip TLs are
transmission line is fixed, both applied models are called “Passive model delay line”.
104
Rext
C d
(a)
(b) (c)
Figure 4.2 (a) The topology of a unit cell CSRRs TL and Photographs of the
designed 4 –cell CSRRs TL ;(b) ground view, (c) microstrip top view
Figure 4.2 illustrates the dimensions of designed 4-cell CSRRs delay line which is
implemented on the Roger (RO3010) substrate with εr = 10.2 and 1.27mm thickness,
In the design the transmission line strip width is adjusted to obtain 50Ω matching
network, the external radius of the ring (Rext) 3.4mm, the metal ring width (c)
0.4mm, and the gap between the inner and outer rings (d) 0.3mm. The total length of
the 4-cells CSRRs TL is 35mm. The equivalent circuit model of this optimized 4-cell
CSRRs delay line is shown in the previous chapter. The S-parameter measurement is
The simulated and measured frequency responses of the designed 4-cell CSRRs TL
are shown in Figure 4.3. The appearance of the LH passband, indicated in the
previous chapter, has the bandwidth of 300MHz from 2.2GHz to 2.5GHz, which
provides the limitation in the system of that frequencies can travel through this
0
S11 and S21 (dB)
-20
circuit simulation
-40 measurement
-60
1 2 3 4 5 6 7
Frequency (GHz)
Figure 4.3 The S-Parameters of the 4-cell CSRRs TL by both simulations and
measurement.
It should be noticed that the substrate and metal losses have been taken in account in
the HFSS simulation, which is not included in the equivalent circuit model of ADS
simulation; however, it can be seen that the passband has been correctly predicted.
The discrepancies between the simulated and measured data in the higher frequency
band are probably due to the fabrication tolerance in our PCB laboratory, especially
the alignment between the metal strip on the top of the substrate and the CSRRs
106
etched on the back of it. The measured transmission coefficient seems having
narrower bandwidth.
8E-9
7E-9
Group Delay (second)
6E-9
5E-9 measurement
3E-9
2E-9
circuit simulation
1E-9
(RH TL)full wave simulation
0
2.2 2.3 2.4 2.5
Frequency (GHz)
Figure 4.4 the group delay on passband of the 4-cells CSRRs TL by both simulations
(dot and dash blue) and measurement (solid blue) as well as a conventional TL at
Figure 4.4 shows the equivalent circuit model, full wave simulation and measured
delay time as a function of frequency within the passband of the designed 4-cell
CSRRs TL(blue). With consideration to the operating frequency band, there are
some differences on the level of delay time with respect to the measured result. This
provides a higher group delay than both simulations which is due to the losses by
measured results. However, the relationship of these two simulated group delay
graphs is parallel which supports the simulations and their measured results.
107
Especially, the HFSS full wave simulation results (dash blue), the group delay graph
seems to get closer to the measured group delay result (solid blue).
In both simulated and measured results, the slope of group delay obviously rises at
the frequency near the lower cutoff frequency which is 2.2GHz at 4.6ns, (by
measurement) 2.3GHz at 3ns and 2.4GHz at 2ns, respectively. At 2.4 and 2.5GHz,
the delay can be assumed to be stable at 2ns which supports the analyzed group
delay relation happening in LH area of the delay line that it is inversely proportional
For comparison purpose, the full wave simulated group delay of the conventional TL
with the same length (35mm) is represented as RH TL in Figure 4.4 (solid red). Its
group delay tends to be stable at 0.3ns all frequencies which presents as 10 times of
the CSRRs TL when observing at the mid passband 2.3GHz. These significantly
group delay results are very useful for size and cost reduction in communication
Delay lines can be very useful in signal processing such as for some radar or
the transmitter in the form of a single, complex signal and de-multiplex each
individual signal at the receiver end. The diagram of a simple RF system is shown in
Figure 4.5.
108
data
Mixer Mixer
RF
CSRRs TL LPF
LO
ωc
The analyzing delay signal was designed on Advanced Design System (ADS) which
was simulated by loading the measured S-parameters of 4-cell CSRRs TL to the data
block. Then, the three sinusoidal continuous waves (CWs) with magnitude 1V at
frequencies of 2.2, 2.3 and 2.4GHz were fed into the proposed CSRRs TL.
600
fcw3=2.4GHz fcw2=2.3GHz
400
fcw1=2.2GHz
CW output (mV)
200
-200
-400
-600
10.0 10.2 10.4 10.6 10.8 11.0
time, nsec
Figure 4.6 The three continuous wave (CW) output signals in Time domain of 4-cell.
At the output end, the different signals in phase and times are extracted and shown in
Figure 4.6. Because the CSRRs TL acts as a narrow bandpass filter, the delay of a
continuous wave cannot obviously be indicated. However, the signal properties and
Due to the relation between dispersion and group delay, the three sinusoidal waves
can clearly be seen both time differences of their amplitudes and phases at the output
port. The CW frequencies that are nearer the higher cutoff frequency (2.3 and
2.4GHz), presents less suffering from dispersion and consume less delay than the
In order to present the clear view of group delay in CSRRs TL, the pulse signals,
which are represented as digital data in communication systems, are fed in the
system in Figure 4.5. It is shown that the modulated pulse with carrier fc2=2.5GHz
(solid blue) is delayed less as it transmits with the peak propagated time at 5ns,
whereas the modulated pulse with the lower carrier fc1=2.25GHz (dot red) has a
time delay at 6ns which displays the peak voltage, illustrated in Figure 4.7.
110
200 fc2=2.5GHz
fc1=2.25GHz
150
50
-50
-100
-150
-200
0 5 10 15 20 25 30
time, nsec
Figure 4.7 the modulated output signals with different carriers ( fc1=2.25GHz in dot
In order to make a clear view of the signal delay and group delay, the envelope
detection of the two input pulses have been considered in the magnitude voltage and
It is obvious that the pulse signal of carrier frequency at 2.25GHz (dash pink)
consumes the time via the 4-cell CSRRs TL more than the pulse signal of the carrier
frequency at 2.5GHz (red) by 1ns delay difference and 3.5ns from the midpoint of
the original pulse. These time differences between the input and output pulses are
Input pulse
0.6 0.20
0.5
0.4 fc1=2.25GHz
0.3 0.10
fc2=2.5GHz
0.2
0.05
0.1
0.0 0.00
0 5 10 15 20 25 30
time, nsec
Figure 4.8 The Envelopes of RF Pulse fc1 and fc2 in time domain compared to input
Therefore, the signals at the output of the delay line can be functioned and extracted
in both CW and Pulse signals. It can be seen clearly that the signals sent through the
CSRRs TL provide different delays for different frequency signals and a much
longer delay with much compact size in comparison with a conventional TL, the
more complex data can be sent at the same time without increasing the system
complexity.
112
There are many reviews that present a varactor diode for a tuning part in
transmission lines. Especially in the recent years, the applied diodes for a large group
delay have been reviewed [4.23, 4.24]. Besides the passive tunable CSRRs delay
line, that is presented previously, the active tunable CSRRs delay line, by embedding
varactor diodes into the 4-cell CSRRs TL is now performed. Because the varactor
diode can act as a small inherence capacitor after changing the applied voltage, the
four varactor diodes (BB833 Infineon Technologies) are placed in the 4-cell CSRRs
the varactor diode and the strip, Rvar and Cvar representing the varactor loss and
The fabricated active tunable CSRRs TL for 4-cascaded cells is depicted in Figure
4.9(b) and (c) by following the dimensions of the 4-cell CSRRs TL. The
measurement is carried out using Agilent E8364A Vector Network Analyzer. The
Lchoke Lchoke
L2 C2 L2 C2
(a)
(b) (c)
Figure 4.9 (a) The equivalent circuit of two adjacent CSRR cells with varactor
diodes, (b) and (c) the photographs of the fabricated 4-cells CSRRs active delay line
on both sides
This new configuration can be tuned to provide varying delay time and passband.
When more negative DC voltages are applied, the diode capacitance reduces and the
group delay is prolonged. Figure 4.10 depicts the insertion loss for the different DC
bias.
114
Figure 4.10 The insertion response (dB) after DC bias voltages of 4-cell CSRRs TL
The measured delay time, of the tunable delay line, for DC bias voltages is illustrated
that the diodes begin to have significant effects as the DC bias varied from -10V to -
20V. The DC changes the delay from 0.6ns to 5.6ns. However, owing to the diode
observed that the passband of CSRRs TL has moved toward higher frequencies and
suffered from addition loss due to the inherit diode properties, as illustrated in Figure
4.10.
115
Figure 4.11 Measured delay time of the 4-cell active CSRRs delay line at the
It is worth pointing out that the group delay of the 4-cell CSRRs delay line can be
tuned without increasing the overall physical dimensions. This tunable CSRRs delay
line, with embedded varactor diodes, provides the tuning rate of 0.5ns/V from -10 to
-20V (bias).
116
4.4 Conclusion
In this chapter, the 2 types of CSRRs delay line have been presented; the passive
In passive CSRRs delay lines, the compact model, consisting of 4-cell CSRRs
particle, has been demonstrated. It is found that the group delay of 4-cell CSRRs TL,
are dependent on the frequencies over the left-handed passband. Then, the two
waveforms, represented as the continuous wave and digital signal, are investigated,
in order, to verify the signal delays in CSRRs TL and group delays. The measured
frequency. The 35mm length of 4-cell CSRRs delay line displays the delay of signal
2.25GHz modulated pulse, at 6ns and 2.5GHz modulated pulse, at 5ns, relating to
the group delay graph, respectively. In addition, the 35mm long 4-cell CSRRs TL is
presented 3ns delay at 2.3GHz which is approximately 10 times higher than those
can be provided by a conventional TL with the same length, which means the shorter
TL can be used in the system to obtain the same delay as in conventional case.
The later remarkable point in this work is the experimental demonstration of the
tunable delay line by embedding varactor diodes into the model of 4-cell CSRRs TL.
When DC bias varies from -10V to -20V, the average tuning rate 0.5ns/V is
impedance of transmission line, the loading of varactor diode can make a large
change of group delay with compact size, which can be proved very useful especially
4.5 References
Soft Computing and Engineering (IJSCE), Vol. 1, Issue 4, pp. 122-128, Sept
2011.
[4.3] S. Park, H, Choi, and Y. Jeong, ”Microwave Group Delay Time Adjuster”,
IEEE Microwave and Wireless Components Letters, Vol. 17, No. 2, pp. 109-
[4.5] S. –K. Park, J. –K. Lee, and C. –D. Kim, “Group Delay Adjuster Using
X/05IEEE.
[4.6] J. –P. Castera, “State of the art in design and technology of MSW devices”,
June 2009.
[4.11] H. V. Nguyen, and C. Caloz, “CRLH Delay Line Pulse Position Modulation
resonator for tunable pulse rate multiplication”, Radio Science, Vol. 44,
Transaction of Microwave Theory and Techniques, Vol. 57, No. 11, pp.
[4.16] F. R. Connor, WAVE TRANSMISSION, 1st edition, Edward Arnold Ltd. 1972.
Inc. 2005.
Ltd. 1995.
Conference, 2009.
[4.23] W. Tang, and H. Kim, ”Compact Tunable Large Group Delay Line”,
Microwave and Optical Technology Letters, Vol. 51, No. 12, pp. 2893-2895,
Dec 2009.
[4.24] Y. –M. Park and D. –W. Kim, “Tunable Composite Right/Left-Handed Delay
Line with Large Group Delay for an FMCW Radar Transmitter”, Journal of
Electromagnetic Engineering and Science, vol. 12, No. 2, pp. 166-170, Jun
2012.
121
CHAPTER 5
INTRODUCTION
Wireless communication has been an important role in human’s life for the last few
decades. Owing to the rapid growth of communication technologies, there have been
many researches to develop devices and circuits for its applications in multiple
between channels must be met for all communication systems. Microwave filter is an
essential part to get rid of unwanted signal from the system and allow the signal in
specific frequency to pass through. Planar filters are widely used because of the
integrated circuits.
It is essential to design compact filters with high frequency selectivity, high rejection
at the band edge as well as wide rejection band. To achieve these demands, many
design techniques have been reported [5.6-5.11, 5.17, 5.18]. Since resonators are the
basic components of planar filters and from the view point of this filter design
section, filters with CSRRs have been introduced to accomplish above requirements
[5.12, 5.13, 5.19-5.22]. Generally, CSRRs with microstrip line presents the narrow
band properties [5.14-5.16]. In addition, CSRRs on microstrip line with a series gap
presents a passband frequency [5.27]. For this reason, a key contribution of this
122
from 0.9-1.9GHz with good frequency selectivity by the aid of rectangular CSRRs
as “a transducer for separating waves on the basis of their frequencies”. There are 3
main types of filter; Active filter (need external source for operating), passive filter
(no requiring external source for operating), and hybrid filter [5.23]. The two-port
network and lumped elements have been used to distribute filters and their frequency
in Figure 5.1. The various parameters have been used to analyze the two-port
converted to each other. The S-parameters are used to present in this section which
The Forward Transmission (insertion) gain with the output port terminated in
The Reverse Transmission (insertion) gain with the input port terminated in a
load or S22
Figure 5.1 Two-port network with the input reflection coefficient and the output
The 2 port network, shown in Figure 5.1, has the parameters V1, V2 and I1, I2 which
are voltage and current variables in complex amplitudes at port 1 and port 2,
respectively. Z01 and Z02 are the terminal impedances and Vs is the voltage generator.
V1 V1 e j (5.2)
124
The wave variables a1, b1 and a2, b2 in Figure 5.1 are introduced for simplying the
voltage and current of the 2 port network which a is signified as the incident waves
and b is represented as the reflected waves. Thus, the voltage and current are
1
Vn Z0n (an bn ) and In (an bn ) , n=1 and 2 (5.3)
Z0n
1 V 1 V
or an n Z 0 n I n and bn n Z 0 n I n , n=1 and 2 (5.4)
2 Z 0 n
2 Z 0 n
The scattering parameters of 2 port network are represented in the terms of wave
variables as
b1 b1
S11 S12
a1 a2 0
a2 a1 0
(5.5)
b2 b2
S21 S22
a1 a2 0
a2 a1 0
While an=0 in the case of balanced impedance and no wave reflection at port n, the
Both S11 and S22 are the reflection coefficient, whereas S12 and S21 are the reverse and
Smn Smn e jmn while m and n are the integer 1,2,… (5.7)
While the insertion loss between port n and m (LA) and return loss at the port n (LR)
are
Which the return loss is related to the proportion of the Voltage Standing Wave
Ratio (VSWR) by
1 Smn
VSWR (5.10)
1 Smn
Beside the scattering parameters are used for microwave filter design, there are other
two paramters that use for analyzing the design. First, the Phase Delay (τp) represents
the difference phase of the wave between the input and output port in 2-port
network:
21
p (5.11)
Another one is the Group Delay (τd) which is considered the difference phase of the
d21
d (5.12)
d
In 2-port network, The reflection parameter (S11) can be analized in the term of
terminal impedance (Z01) by replacing with Zin1=V1/I1. Zin1 is called the input
b1 V1 / Z 01 I1 Z 01
S11 (5.13)
a1 a2 0
V1 / Z 01 I1 Z 01
Zin1 Z 01
S11 (5.14)
Zin1 Z 01
In the same means, if consider the port 2 as the input and Zin2=V2/I2, Zin2 is the input
Zin 2 Z 02
S22 (5.15)
Zin 2 Z 02
In lossless and passive network, the total the input power is from the transmission
(5.17)
Table 5.1 presents the summarization of the main four type of two-port network and
their parameters
Table 5.1 The overview of the relationship between two-port network parameters
b1 b1
S11 , S12
a1 a2
a2 0 a1 0 Smn Smn e jmn : m, n 1, 2
b2 b2 Amplitude : 20log Smn dB : m, n 1, 2
S 21 , S 22
a1 a2 at port m & n
a2 0 a1 0
Insertion Loss : LA 20log Smn dB : m, n 1, 2(m n)
b1 S11 S12 a1 at port n
Scattering
b S a
parameters 2 21 S 22 2 Return Loss : LR 20log Smn dB : n 1, 2
(S-
Z Z 1 Smn
parameters) S11 in1 1 Voltage Standing Wave Ratio : VSWR
Z in1 Z1 1 Smn
Z in 2 Z 2 21
S 22 Phase Delay : p seconds
Z in 2 Z 2
d
Group Delay : d 21 seconds
d
128
I1 I1
Y11 , Y12 ●Reciprocal network : Y12=Y21
V1 V 0 V2 V1 0
2
Short-circuit
●Symmetrical network : Y12=Y21 and
admittance I2 I2
parameters Y21 , Y22 Y11=Y22
(Y- V1 V 0 V2 V1 0
2
parameters) ●For lossless network, Y parameters are purely
I1 Y11 Y12 V1 imaginary
I Y Y V
2 21 22 2
V1 V1
Z11 , Z12 ●Reciprocal network : Z12=Z21
I1 I2 0
I2 I1 0
Open-circuit ●Symmetrical network : Z12=Z21and Z11=Z22
impedance V V
parameters Z 21 2 , Z 22 2 ●For lossless network, Z parameters are purely
(Z- I1 I2 0
I2 I1 0 imaginary.
parameters)
V1 Z11 Z12 I1
V Z [Z ] [Y ]1
2 21 Z 22 I 2
Table 5.2 Summarization of the terminated two port circuits [5.1, 5.25, 5.26]:
Z1 Z3
Z2
T Network
Z2
Z1 Z3
Pi Network
Z3
Z1 Z1
Z2
Z2
Z1 Z1
Z2
I1 I2
Z1
V1 V2
Z2
Y1
I1 I2
V1 V2
Y2
I1 I2
V1 A1 V2
A2
Figure 5.2 Interconnection of two-port network (a) Series, (b) Parallel, and (c)
Cascade [5.26]
If the 2 port networks with Z-parameters are connected in series, the equivalent port
is given by
[Z] eq=[Z] 1+[Z] 2+[Z] 3…+[Z] n , n is the number of 2 port network with Z-parameters
If the 2 port networks with Y-parameters are connected in parallel, the equivalent
port is
[Y] eq=[Y] 1+[Y] 2+[Y] 3…+[Y] n , n is the number of 2 port network with Y-parameters
If the 2 port networks are connected in cascade which each individual networks have
transmission parameters [A] 1, [A] 2, [A] 3,…,[A] n, the total equivalent 2 port
[A] eq= [A] 1*[A] 2* [A] 3*…*[A] n, n is the number of each individual network
132
line) are another alternative way for RF filter design. The characteristics of filter
(Butterworth, Chebyshev, Bassel and Elliptic) are described by their output response
0dB Rp
-3dB
BW
fL fC fH
upper and lower frequency (fH-fL) of the circuit at which the amplitude is
Insertion loss (IL): the loss of signal in power after transmitted to device,
IL 20log S 21 dB
Return loss (RL): the difference in dB between the forward and reflected
design, filter’s selectivity defines how much the filter will reject unwanted
frequencies.
fc
Q factor: The inverse fractional bandwidth defined as Q ; a high Q
BW
filter gives narrow passband, while a low Q filter will have a wide passband.
Group delay: is the amount of time for signal to pass through the filter, most
types of filter the group delay is defined as dϕ/df varied with frequency.
There are many methods to design RF/microwave filters. Because of the high
in planar filters [5.4-5.11]. In this chapter, a high rejection in the lower band and
improved wideband pass filters are designed using CSRRs and coupling stub
1. The coupling part on microstrip line width and length of the balanced-
capacitive stub and the inductive coupling line are provided to control its
2. The CSRRs which is etched on the ground plane and a coupling plate are used
zero. A lower band good rejection and the low loss in band operating frequency
The coupling structure with loaded capacitance is presented in Figure 5.4. The initial
structure of the design bandpass filter is comprised of the two coupling parts; firstly,
the coupling part between both stubs and their inductive lines while another coupling
is occurred in coupling plate. With these strong coupling parts, the structure provides
Linductance
Lstub
Coupling
plate
The simulation results by Ansoft HFSS are run on RO3010 substrate with thickness
h=1.27mm and dielectric constant (εr) =10.2 and tanδ=0.0035. The capacitive stubs
(Lstub) are 16mm. The lengths of both inductive coupling lines (Linductance) are 24mm.
-5.00
-10.00
S11 and S21 (dB)
-15.00
-20.00
-25.00
Curve Info
dB(S(1,1))
-30.00 Setup1 : Sw eep1
dB(S(2,1))
Setup1 : Sw eep1
-35.00
0.00 1.00 2.00 3.00 4.00 5.00
Freq [GHz]
Figure 5.5 The insertion and return loss of balanced load capacitance structure
This coupling structure generates a narrow passband filter from 1.7-2GHz which has
the scattering parameters shown in Figure 5.5. The simulated S11 and S21 present that
the balanced load capacitance structure has the insertion loss in passband at -1dB and
give a sharp cut-off at the higher band edge which provide transition band
approximately 50MHz. In addition, the structure can present a high resolution down
to -30dB. The stopband area is clearly from 2.1GHz to 3.6GHz. These properties
The rectangular CSRRs loaded transmission line and its equivalent circuit model
(a)
Cc
Lr Cr
(b)
Figure 5.6 (b) The unit cell of CSRRs loaded with transmission line topology and (b)
the three factor as following air slot on ground plane (d), its conductance (c), and the
dimension length (a), respectively. Owing to the behavior of CSRRs which acts as
CSRRs is formed by the parallel combination of Lr and Cr and its coupling to the
host line is represented by the capacitance Cc. The series gap on transmission line is
represented by Cg.
137
As a result of input port at the resonant frequency, the parameter L and Cg can be
1
fr (5.18)
2 Lr Cr
1
fZ (5.19)
2 Lr (Cc Cr )
From equation (5.18) and (5.19), CSRRs has been applied in this design model for
selecting the passband of filter and controlling the transmission zero which is related
this filter is based on combining of rectangular CSRRs on ground plane and the
coupling structure formed by microstrip line on Figure 5.4. The layout of proposed
Linductance
Lstub
Coupling plate
(a)
(b)
Figure 5.7 The layout of proposed filter. (a) Basic cell, (b) Topology of rectangular
CSRRs
The width of strip (W) and stubs (Wstub) are 2mm while the lengths of balanced
inductive stubs (Lstub) are 14mm which is the same length as coupling plate,
respectively. The width of coupling plate is 4mm. The length of both inductance
lines (Linductance) is 29.5mm, whereas their width (Winductance) is 0.3mm and all gaps
are maintained at 0.3mm, respectively. The rectangular CSRRs size (a) is 10mm.
139
Figure 5.8 The current distributions of the proposed bandpass filter at (a) 0.72GHz
which is in the passband of the proposed filter, the wanted signal has been
Figure 5.9 The magnetic field distribution on ground plane at (a) 0.72GHz and (b)
1.4GHz, respectively
140
Figure 5.9. This shows that there is the existence of the H field on the ground plane
which is strongly concentrated on the CSRRs. It again illustrated that little magnetic
energy has been transferred to port2 at 0.72GHz, whereas the energy propagates to
port2 at 1.4GHz.
(a)
41.5mm
(b)
Figure 5.10 The photograph of the fabricated filter (a) Top view, (b) Bottom view.
total dimension of the filter is 40x41.5mm2 and dielectric constant (εr) =10.2.
141
Network Analyzer.
L2
Cs Cs
L1 L1
Cc C1 C
c
CM CM
Cr Lr
The equivalent circuit of the proposed filter, presented in Figure 5.11, is explained
by the combination model between the π model and T model. The extracted
L2=1.189nH, respectively.
142
(a).
(b)
Figure 5.12 The ADS simulated S-parameters (a), and (b) comparison of ADS and
The equivalent circuit simulated results of the proposed filter is shown in Figure
5.12(a). In Figure 5.12(b), the comparison of the two simulated results by ADS
equivalent circuit and HFSS are illustrated. Reasonably good agreements can be
found in the two simulated results. However, there are some discrepancies. The
143
equivalent circuit analysis provides wider band and lower transmission zero. This is
because the equivalent circuit analysis cannot provide the total effects from magnetic
Figure 5.13 shows the simulated and measured results of the return loss S11 and
insertion loss S21 in wideband. The results of the frequency response measured on
the fabricated band pass filter substrate show satisfactory agreement with the
Figure 5.13 Measured and HFSS simulated frequency response on the designed filter
at 0.9-1.9GHz.
The measured insertion loss is about 0.9dB at the centre frequency and the passband
return loss is less than -10dB. The filter presents wideband properties which covers
the frequencies from 0.9 to 1.9GHz. The 3dB FBW is approximately 77%. These
144
can be achieved due to the induced between a parallel inductive line coupling and the
applied CSRR on conventional filter that can be modified the filter properties
transition band with transmission zero deep down to -60dB by measurement at the
5.4 Conclusion
In this chapter, the design and fabrication of the wideband pass filter with the
been presented. It was found that the presence of CSRRs on conventional coupling
filter can modify the filter properties from 300MHz narrowband filter to 1GHz
wideband filter without changing any profile. In this work, the measured insertion
loss of this design filter at operation band is very low which presents <1dB at the
centre frequency as well as the passband return loss is less than -10dB. With CSRR
properties, the transmission zero and a sharp transition band on the lower cut-off
frequency has been demonstrated. At the out of band on higher frequency, this
conventional coupling filter at 3.9GHz. This design filter provides the excellent
5.5 References
[5.1] J. –S. Hong and M.J. Lancaster, Microstrip Filters for RF/Microwave
[5.2] L. Zhang, Z. –Y. Yu, and S. –G. Mo, ”Novel Planar Multimode Bandpass
[5.3] C. –Y Huang, Y. –X. Yu, and R. –Y. Yang, ”High Selective Dual-Band
Bandpass Filter design with Novel Feed Scheme”, Proceedings of the World
Theory and Techniques, Vol. 45, No. 12, pp. 2358-2365, December 1997.
Resonator”, Microwave and Optical Technology Letters, Vol. 53, No.4, April
2011.
172, 2011.
Pass Filter Using the Rectangular CSRRs”, PIERS Online, Vol. 5, No.5,
2009.
[5.14] M. Gil, J. Bonache and F. Martin, ”On the transmission properties of left-
Model for CSRR-Based Bandpass Filter Design With Even and Odd Modes”,
IEEE Microwave and Wireless components letters, Vo. 20, No.4, April, 2010
147
low-pass filter with sharper cut-off and reduced size”, IET Microwave &
219-226, 2007.
172, 2011.
[5.24] D. M. Pozar, Microwave Engineering, John Wiley and Sons, Inc., 2nd edition
1998.
[5.25] L. Zhu, S. Sun, and R. Li, Microwave Bandpass Filters for Wideband
[5.27] C. Li, K.-Y. Liu and F. Li,”Design of microstrip highpass filters with
Jan 2007.
149
CHAPTER 6
INTRODUCTION
permeability and negative permittivity, have led to the first Split Ring Resonators
(SRRs) been reported by Pendry [6.1, 6.2]. There are many applications of
metamaterials, which have been applied in the last decade, including antennas, in
order to enhance their performance [6.3, 6.4]. Due to the properties of metamaterials
that can manipulate the electromagnetic field, the higher impedance substrate can be
generated. This can be used to reduce the size as well as maintain its efficiency.
This chapter mainly consists of three parts. The first part will briefly mention about
the basic microstrip patch antennas and its properties. The second part will present a
new technique of antenna size reduction, by etching CSRRs on the ground plane as
well as the inductively meander line loaded patch. By this method, the size reduction
of the proposed antenna can reach 74% of the conventional rectangular patch
antenna. In last part, the technique of tunable CSRRs antenna has been introduced by
reversely biasing the varactor diode between its patch and ground side. The wide
tuning range has been achieved to 350 MHz without changing any dimension. The
The IEEE defines antenna as “The part of a transmitting or receiving system that is
antenna as “a usually metallic device (as a rod or wire) for radiating or receiving
radio waves”. In other words, antenna is a structure used to transit the signal between
free space and a guiding device e.g. transmission line. Then, transmission line will
transport electromagnetic energy from the antenna to the receiver or from the
Generally, there are two categories of antennas according to their application [6.5]:
Omnidirectional antennas are called the group that radiate or receive in all
directions. It is normally used when the relative position of the other station
is arbitrary or unknown.
If consider in terms of type, antennas can be divided into various types such as wire
antennas, aperture antennas, lens antennas, reflector antennas, and microstrip patch
antennas [6.5].
151
An antenna radiates its signal which is occurred basically due to the time varying
charges in a wire (no current), no radiation happens. The antenna radiation can be
sinusoidal voltage wave across the transmission line. A sinusoidal electric field is
generated which causes the appearance of electric lines of force which are tangential
The electric field’s magnitude is shown by the high concentration of the electric line
of force. The movement of the free electrons on the conductors under the electric
lines of force leads to the appearance of current flow which forms a magnetic field.
The time varies of magnetic and electric fields will then create the electromagnetic
waves traveled between the conductors. The free space waves are appeared at the
open ends of the electric lines which is connected to the open space. Because the
electromagnetic waves are continuously generated by the sinusoidal wave from the
source, its electric disturbance is constant which leads to these waves can propagate
through the transmission line and the antenna as well as the free space radiating.
Owing to the charges, inside the transmission line and antenna part, the
electromagnetic waves are continuing upheld. While in the free space; they are
Microstrip patch antennas are very popular in airborne systems, mobile radio, pagers,
radar systems and a variety of wireless and satellite communications because of its
attractive features which have advantages over other antenna structures such as low
profile, low fabrication cost, light weight, easy fabrication, simplicity and capability
microstrip antennas still have a bulky size especially in lower frequencies, therefore,
in order to respond for high demand of wireless communications with limited space
in the present-day; many techniques of minimization have been proposed [6.7, 6.8].
153
Patch antenna
Feeding point
h
substrate
Generally, a microstrip patch antenna comprises of the feed lines and a radiating
substrate and has a ground plane on the other side. The example of a rectangular
microstrip patch antenna, shown in Figure 6.2, has the length L, width W placing on
a substrate of height h. The radiating patch can be of many shapes and size, for
space wavelength. The patch thickness (t) is chosen to be quite thin by t<< λ0. The
constants (εr) of substrate of the design microstrip antennas can be ranging from 2.2
to 12 [6.5].
The feeding method is an important factor which can affect the performance of the
design patch antenna. In the other words, the different feeding methods exhibit
154
(a) (b)
(c)
(d)
Figure 6.3 The feeding methods of a microstrip antenna (a) Coaxial feed, (b) Inset-
The most common and easiest feed is coaxial type, while inset feed is widely used
for array antennas. The proximity-coupled feed is recommended to use for reducing
spurious radiation from the feed line in multilayer fabrication. The aperture-coupled
feed type, used in thick substrate, can eliminate feed line radiation [6.5, 6.6].
Patch antennas radiate mainly because of the fringing fields between the ground
plane and the edge of the patch. The antenna that has a thick dielectric substrate with
bandwidth and better radiation; however, it has a larger size [6.5]. In order to design
a compact patch antenna, higher dielectric constants must be used which are less
Since 1950s, microstrip antennas have been widely used, as compared as the
Several techniques have been proposed to overcome these limitations such as array
configuration techniques with power handling [6.9] and choke structure to gain the
Rrad
Xs
Rloss
Rs
Xin
Vs
in
Z
represents the source comprised of the internal impedance (Zs=Rs+jXs), while the
of antenna consists of the radiation resistance (Rrad) and the antenna losses (Rloss).
The input impedance can be used to analyze the reflection coefficient г by [6.5]
Zin Z 0
S11 (6.1)
Zin Z 0
antenna to the generator or source. The reflection coefficient shows the amount of
158
power reflected to the source. In ideal case, г should be zero. Therefore, the voltage
1
VSWR (6.2)
1
RL 20log (6.3)
frequencies on Smith Chart with less reflection to the source by matching with the
feed line. In order to receive the maximum power transfer to antenna, the input
impedance, ideally, contains only a resistive real part which is the same value as the
The substrate used for this case is RO3010 with a dielectric constant=10.2
The initial given values of the width (W) of the patch is calculated by using
1 2 v 2
W 0 (6.4)
2 f r 0 0 r 1 2 fr r 1
159
calculated by
1/ 2
r 1 r 1 h
reff 1 12 (6.5)
2 2 W
W
( reff 0.3) 0.264
L h
0.412 (6.6)
h W
( reff 0.258) 0.8
h
c
Leff (6.7)
2 f 0 reff
c
fr (6.9)
2 L reff
The CSRRs loaded transmission line, and its equivalent circuit model, have been
proposed [6.12, 6.13], shown in Figure 6.5 (a) and (b). Figure 6.5 (a) represents the
topology of CSRRs which is composed of three parameters: air slot on ground plane
160
(d), its conductance (c), and the average ring dimension (r0), respectively. Owing to
the behavior of a CSRRs which acts as LC resonator, the equivalent circuit can be
modeled by Figure 6.5 (b). It can be seen that CSRRs are formed by the parallel
combination of Lr and Cr and its coupling to the host line, which are represented by
d
r0
Zs/2 (a)
L/2 L/2
2Cg 2Cg
Cc
Zp
Lr Cr
(b)
Figure 6.5 (a) The unit cell of CSRRs TL and (b) its T-equivalent circuit.
The intrinsic parallel resonant frequency (fc) of the CSRRs can be given by [6.12,
6.13].
To obtain fc, the shunt patch to the ground is opened. This gives directly the value of
L from the negative permittivity line. Then, the additional condition is,
161
Zs and Zp are the series and shunt impedance of the T-equivalent circuit model of the
structure, respectively. Thus, the first factor to design is the radius of inner and outer
ring which mainly controls the resonant frequency of antenna. In this proposed
antenna, there is another factor which has been used to control the resonant
printed on one side of a dielectric substrate which has a ground plane on the other
side. The patch is normally made of conducting material such as copper. The
radiating patch and feed lines are usually etched on the dielectric substrate. In this
CSRRs on the ground plane with inductively meander line and capacitive gap on the
patch.
The idea of introducing a CSRRs particle on the ground plane of antenna structure is
to generate the negative permittivity at the design frequency which this coupling on
CSRRs is the first main key used to control the resonant frequency of antenna. Then,
the meandered line patch, acts as inductance, is applied for further controlling the
resonant frequency. The proposed rectangular microstrip patch antenna was designed
162
The antenna in this research was designed and adjusted to obtain a proper compact
simulation programme.
The antennas in three different meander line turns, as shown in Table 6.2, have been
studied and designed. The layout of the meander line and its geometrical parameters
are given in Figure 6.6 and 6.7. The introduced meander line with the capacitive gap,
acting as a series of inductors, by using the equation (6.10) to (6.12), helps to finely
match and tune the resonant frequency lower with maintain the patch size as
electrical small.
163
a
a
a
a
Figure 6.7 Layout of the meander line (4 turns), a=0.6mm and the capacitive
gap=0.2mm.
1 0
2 2
3 4
164
Figure 6.8 The simulated return loss (S11) of each CSRRs patch antenna.
The simulated antenna return losses are illustrated in Figure 6.8. The meander line
turns can move the resonant frequency lower and achieve better matching. However,
the CSRRs on the ground plane are a sensitive model, a few limited meander line
turn, as a fine tuning, can be accepted. The Bandwidth of antenna3 (dash blue line) is
After applied the voltage source to antenna, the electric field has been generated as
well as fringing zone polarizing through the patch. Figure 6.9(a) and (b) show the
electric field distribution on the patch surface of the proposed antenna1 and antenna3
(x-y plane), respectively. Moreover, there is another electric field occurred by the
coupling between the inner and outer of the two rings of CSRRs which is opposite
direction to the original one. As a result, the electric field distributions on the
coupling area on patch have been created which provides a new resonant frequency.
The magnetic field distributions, illustrated by Figure 6.9(c) and (d), are
(a) (b)
(c) (d)
Figure 6.9 The simulated field distributions at resonant frequencies (a) E field of
antenna1 at 1.8GHz and (b) E field of antenna3 at 1.73GHz, while (c) H field of
Figure 6.10(a) and (b) demonstrate the surface’s current distributions cover all on
patch and ground of antenna3. At the resonate frequency of CSRRs, the current
on ground. In addition, the inductive meander line on patch is also used to give better
coupling with CSRRs as tuning to get an optimum matching the antenna resonant
frequency.
166
(a)
(b)
Figure 6.10 The surface’s current distribution at 1.73GHz of antenna3 (a) on patch
The photos of the fabricated antenna are shown in Figure 6.11. The antenna was
fabricated on the RO3010 substrate, whose dielectric constant (εr) is 10.2, loss
167
tangent (tan δ) 0.02, and thickness (h) 1.27mm, respectively. The size of outer
ground plane on the front side of the substrate is Ws x Ls, which is 22 x 28mm2.
16.4mm
16.4mm
(a) (b)
Figure 6.11 The photos of antenna3 (a) the meander line patch with gap (b) The
The antenna was designed to have the centre frequency of 1.73GHz. The meander
line patch size (Wp x Lp) is controlled by the ring radius Ro. In this work Ro=8.2mm.
The CSRRs was etched on the ground plane with the parameters of c and d which are
0.2 and 0.2mm, respectively. The inductive meander patch gap is designed to get
Figure 6.12 The return loss in dB by full wave HFSS simulation and VNA
measurement of meander line loaded patch antenna with CSRRs etched on the
ground.
Figure 6.12 shows the simulated and measured return losses for the proposed
antenna3 which has a meander line of 4 turns. It can be seen that the measured
results are well agreed with the simulated one. As mentioned in the previous section,
the size of this proposed CSRRs resonant antenna with meander line can be reduced
to reach the patch size of 0.095λ0 x 0.095λ0. This is much smaller compared to the
conventional patch antenna which has the patch size of 0.21λ0 x 0.16λ0. The size of
the proposed antenna can be reduced dramatically because the resonant frequency
can be controlled by the reaction of inner and outer coupling of the CSRRs instead of
the physical length of conventional patch. Moreover, with the help of meander line
and the capacitive cap, acting as inductive and capacitive loadings, the antenna can
This section introduces a novel tunable compact antenna loaded with CSRRs. The
varactor diode (Infineon BB535) is used to tune the operation frequency. The
dimension of antenna is maintained the same as the previous section. The operation
section is 1.8GHz. The proposed varactor loaded CSRRs antenna can be tuned by
Zload
DC bias
Varactor
Diode
Vs
Figure 6.13 The simple structure of the proposed antenna with varactor diode.
Varactor diode can vary its inherit capacitance by reverse bias. The low reversed bias
voltage provides high inherit capacitance, while higher the reversed bias voltage will
decrease the diode capacitance, respectively. The equivalent circuit of the tunable
(a) (b)
Figure 6.14 The layout of tunable patch antenna (a) ground view and (b) patch view.
Figure 6.14(a) is the photograph of the varactor diode connecting on ground plane
via the second patch, whereas Figure 6.14(b) shows the patch view and the
connecting point of the varactor diode to the ground plane which is placed on the
second half of the patch in order to allow the constant electric wave pass through the
The bias can control the resonant frequency, shown in Figure 6.15. Without DC bias,
the antenna resonates at 1.15GHz with return loss of 33dB. After applied reversed
moving the resonate frequency higher. In the other words, the tunable antenna
provides the resonant frequency under the condition of the lower diode capacitance,
the higher resonant frequency. Moreover, in the frequency ranges of the working
antenna, the operated bandwidths are stable at 35MHz. After biasing DC to +12V,
there is no radiation from antenna (antenna is off) that because the completed
171
connecting of the depletion region inside diode presents a completed circuit loop and
Figure 6.15 The measured return loss of the tunable CSRRs patch antenna with
As a result of diode applied CSRRs patch antenna, a fine tuning range is covered
350MHz from 1.15 to 1.5GHz with maintaining bandwidth at 35MHz. This is the
first shown the co-operation of CSRRs as mention before as well as the work of
6.4 Conclusion
Novel compact CSRRs loaded patch antennas with tunability have been studied,
Microstrip patch antenna using CSRRs and capacitive loaded meander line can
reduce size by 74% in comparison with conventional rectangular patch antenna. The
distributions within and between the meander line patch and CSRRs. More turns of
the meander line can move the operating frequency lower without changing any
size of antenna is still compact by keeping the previous dimension of without diode
case. The operation frequency of CSRRs antenna with diode is at 1.15GHz which is
lower than CSRRs antenna without diode at 1.8GHz in same dimension. After
applied DC voltage, the inherit capacitance of diode changes, leading to the change
of the overall impedances of the antenna. Therefore, the tunable CSRRs antenna can
cover the frequencies from 1.15 to 1.5GHz with stable 35MHz narrow bandwidth.
173
6.5 References
[6.1] J.B. Pendry, A.J. Holden, D.J. Robbins,”Magnetism from Conductors and
2000.
[6.3] D.-O. Kim, N.-I. Jo, D.-M. Choi, ”Design of The Ultra-Wideband Antenna
August 2008
[6.5] C. A. Balanis, Antenna Theory: Analysis and Design, 3rd ed. John
[6.6] W.L. Stutzman and G.A. Thiele, Antenna Theory and Design, 2nd ed. New
[6.7] A. Kordzadeh and F. Hojat Kashani, “A New Reduced Size Microstrip Patch
[6.9] H. F. Lee and W. Chen, Advances in Microstrip and Printed Antennas, John
[6.11] W. G. Lim, H. S. Jang, and J. W. Yu, ”New Method for Back Lobe
France
circuit models for split ring resonators and complementary split ring
CHAPTER 7
7.1 Conclusions
The aim of this research was to propose and examine a novel approach of CSRRs
particle etched onto the ground plane. This CSRRs exhibit negative effective
The equivalent circuit and the left-handed area of CSRRs TL were analyzed in
Chapter 3. It was found that the magnetic coupling component should be added in
the lumped-element equivalent circuit model in the case of more than two adjacent
CSRRs cells. The analysis of the magnetic coupling effect was carried on by HFSS
and placed. By varying the distance between these two CSRRs, the detected
magnetic couplings were captured and analyzed. The graph result of magnetic
coupling presented that there were some currents that appeared from 200A/m to
50A/m between 0.2mm to 1mm distances of these two rings, respectively. By using
ADS extracted the parameters of the 0.2mm distance case, the S11 graph of the
proposed equivalent circuit can present the resonant frequency (fc) at 2.27GHz
matching the S11 from full wave simulation, which was better than the old equivalent
circuit, which provides fc at 2.3GHz. The S11s on smith chart also agreed with the
176
new equivalent circuit. Another noticeable outcome in this chapter was the indicating
LH operating area of CSRRs TL. The 4-cell CSRRs TL was fabricated. It was
noticed that there were two passbands that appeared (≈2.2-2.5GHz and 4.7-6.5GHz).
The dispersion graph clearly indicated the left-handed area (<2.55GHz) and right-
hand area (>4.7GHz) of this fabricated 4-cell CSRRs TL. This left-handed area was
conditions, the first application in this thesis was to study the delay property in
CSRRs applied to microstrip TLs (chapter 4). Initially, the group delay relation of
both LH and RH areas were analyzed. It was found that in LH areas the group delay
is inversely proportional to frequency. Then, the group delays of these two cases
were investigated, which are passive delay line and active delay line. The group
delay of the previous 4-cell CSRRs TL of 35mm length was measured and referred
to passive delay line. The group delay graph results of both simulations and
measurement agreed with the previous analyzed theory, whereas in 35mm long
normal TL the group delay remained the same overall frequency at 0.3ns. In
comparison, at 2.3GHz the group delay of this 4-cell CSRRs TL displayed 3ns which
was 10 times that of a conventional TL. Then, the signal delays on a simple RF
system were examined. The two waveforms, CWs and pulses, were studied. By
loading the measured S-parameters to the data block in ADS, the three different
sinusoidal waves of 2.2, 2.3 and 2.4GHz were fed in the 4-cell CSRRs TL
simultaneously. The results showed that these three signals perform differently in
times, phases, and amplitudes at the port end, relating to the group delay graph.
Next, the two modulated different pulses with carrier 2.25 and 2.5GHz were fed in
177
simultaneously. At the output port, the delayed pulses consumed the time in CSRRs
TL at 6 and 5ns of the carrier wave at 2.25 and 2.5GHz, respectively. In active delay
line, the 4 diodes were embedded in the 4-cell CSRRs TL. Because of the change of
characteristic impedance resulting from diode addition, the operating frequency was
shifted. The group delays were detected by monitoring one frequency. For example,
after applied the DC bias to diodes from -10 to -20V, the group delay displayed the
Both simulation and measurement results in this chapter prove that the group delay
can be varied, and controlled, without changing the transmission line length, which
The later application was to develop a wideband passband filter, by using rectangular
CSRRs (chapter 5). The new filter configuration was achieved by the modified
plane in the prototype model. The coupling between CSRRs and the coupling plate
of the host line generated a passband on the lower frequency side of the proposed
filter; as a result a high rejection in the lower band and the transmission zero were
presented. The novel filter exhibited a wide bandpass with a passband up to 77% of
the bandwidth covering the range from 0.9 to 1.9GHz. The insertion loss is less than
-1dB. Furthermore, the proposed filter presented the better performance in frequency
The validated results were agreed through both equivalent circuit simulation and
measurement.
178
The last application is to minimize the size of planar antenna, which was done by
etching a CSRRs on the ground side of the patch antenna. A capacitive gap was also
etched on patch side. The originated frequency of the initial antenna was controlled
by the LC resonator of CSRRs, then it was applied the meander line part on the
middle of the patch to 2 and 4 turns. The simulated results showed that the operating
frequency was lower when applying more meander turns from 1.75GHz to 1.73GHz,
respectively. The measured S11 of the antenna with 4 turns of meander line agreed
with the HFSS full wave simulation. This novel planar antenna has shown 35MHz
patch antenna. In addition, inserting a diode from patch to ground can change the
was applied from 0V to 8V, the antenna provided the tuning range to 350MHz
covering the frequency 1.15GHz to 1.5GHz and maintaining its physical dimensions.
This proposed antenna maintained 35MHz bandwidth over all tuning ranges.
In summary, this thesis presents three areas of work, one showing that the
transmission line can control signal delay and reduce size, which means more data
the CSRRs particle. Finally the development of an antenna of a smaller size is also
enhancement purposes.
179
This thesis provides the most fundamental part of communication systems. However,
there are some details that should be applied for more quality results. The future
analysis is needed. This simple system is includes the signal and carrier
generator, mixer, and filter. Each part has to be placed on the same board.
Therefore, the more correct delay in the real system should be analyzed.
work, it is hard to get the radiation pattern. Placing the antenna in the form of
LIST OF PUBLICATIONS
[1] S. Pasakawee and Z. Hu, “Compact Passive and Active Tunable Delay Lines
Accepted.
2012.
[3] S. Pasakawee and Z. Hu, “Electrical Small Meander Line Patch Antenna”, 6th
30, 2012.
20-23, 2009.
181
Appendix
Cs L2 Cs
2C1 L1 L1 2C1
CM Lr Cr CM
is in Pi-model, respectively.
2C1 L1 L1 2C1
Z1 Z3
C
Lr Cr
N1
Z2
L2
Cs Cs
Z5
N2
CM CM
Z4 Z6
Z Z2 Z2
[ Z ]T 1
Z2 Z 2 Z3
Z 2 Z3 Z2
Y11T , Y12T
Z1Z 2 Z1Z3 Z 2 Z3 Z1Z 2 Z1Z3 Z 2 Z3
Z2 Z1 Z 2
Y21T , Y22T
Z1Z 2 Z1Z3 Z 2 Z3 Z1Z 2 Z1Z3 Z 2 Z3
Z 4 (Z5 Z 6 ) Z4 Z6
Z Z Z Z 4 Z5 Z 6
Z
4 5 6
Z4 Z6 Z6 (Z 4 Z5 )
Z Z Z Z 4 Z 5 Z 6
4 5 6
Z6 (Z 4 Z5 ) 1
Y11 , Y12
Z 4 Z5 Z 6 Z5
1 Z 4 ( Z5 Z 6 )
Y21 , Y22
Z5 Z 4 Z5 Z 6
N1 and N2 are in parallel connection. Therefore, the proposed filter can be derived in
Y parameters as
Z 2 Z3 Z (Z Z5 )
Y11 6 4
Z1Z 2 Z1Z3 Z 2 Z3 Z 4 Z5 Z 6
Z2 1
Y12 Y21
Z1Z 2 Z1Z3 Z 2 Z3 Z5
Z1 Z 2 Z (Z Z6 )
Y22 4 5
Z1Z 2 Z1Z3 Z 2 Z3 Z 4 Z5 Z 6