PhDThesis SunilSudhakaran

Negative Refraction from
Electromagnetic Periodic Structures and

Its Applications

A Dissertation Submitted for the Degree of Doctor of
Philosophy at the University of London

Sunil Sudhakaran

Department of Electronic Engineering
Queen Mary
University of London

June 2006

i

Abstract
Left Handed Metamaterials (LHMs) are materials with negative permittivity
and permeability. Negative refraction occurs when electromagnetic waves pass through
LHMs at an oblique angle. It was reported that such a phenomenon can be found in
Electromagnetic Bandgap structures (EBGs) at the bandgap edges. However, the
effects of spatial dispersion, which is inherent in all periodic structures, have not fully
been probed yet. These effects are negligible only when the period is much smaller than
wavelength. Also, the conventional approaches for analysing EBGs ignore incident
field coupling and evanescent fields.
In this thesis, spatial dispersion effects in finite EBGs on negative refraction and
the possibility of obtaining analogous LHMs are investigated. It is observed that high
order Floquets harmonics in finite periodic structures affect the overall dispersion of
the EBGs, and hence a super-cell model analysis is proposed and validated through
numerical simulation and experiments.
Verification of negative refraction from an EBG prism consisting of thick
metallic wires at multiple frequency bands is presented. Enhancement of negative
refraction at certain frequencies by selection of negatively refracted harmonic is
demonstrated. High resolution focusing at low microwave frequencies in the passband
from a slab structure consisting of thick metallic wires is detailed. Numerical
simulation and experimental results are presented in good agreement. Finally, the sub-
wavelength imaging using the canalization regime in a thin metallic wire medium is
experimentally investigated. The source and wire medium interaction in near field
imaging is studied and reduction in bandwidth is noticed.

ii

Acknowledgements
I express my deep sense of gratitude for my supervisor Dr. Yang Hao. This dissertation
is truly an outcome of his constant support, valuable ideas and suggestions. The
discussions with him always provided me great motivation.
I also thank Prof. Clive G. Parini for his suggestions and support.
I am thankful to the Department of Electronic Engineering, Queen Mary, University of
London for awarding me the scholarship and The Royal Academy of Engineering, UK
for the financial support for the research and travel grants.
I sincerely thank Mr. John Dupuy for his constant help in measurements.
I would like to thank Dr. Pavel Belov, working with him always provided me an
opportunity to learn more.
There are no words to express my thanks towards my colleagues at the Department of
Electronic Engineering. I express my sincere thanks to all of them.
A special line of thanks is due to my colleagues Mr. Akram Alomainy and Miss
Mariana Setta for their suggestions.
I am deeply indebted to my wife and parents; their love and support were the pillars for
the success of my research work.
Above all I thank Almighty God for showering his infinite bounties and grace upon me,
which has seen me successfully complete my research work.

iii

Publications
JOURNAL PUBLICATIONS
1. Pavel Belov, Yang Hao, Sunil Sudhakaran, Subwavelength microwave
imaging using an array of parallel conducting wires as a lens, Physical Review
B, Vol. 73, , pp. 033108-0033112, 2006
2. S. Sudhakaran, Y. Hao, C. G Parini, Focusing of waves at low microwave
frequencies using metallic wire media, Microwave and Optical Technology
Letters, , Vol. 48, pp.133-138, 2006
3. S. Sudhakaran, Y. Hao, C. G Parini, 'Negative refraction phenomenon at
multiple frequency bands from electromagnetic crystals, Microwave and
Optical Technology Letters, Vol.45, pp. 465-469, 2005
4. S. Sudhakaran, Y. Hao, C. G Parini, 'An enhanced prediction of negative
refraction from EBG-like structures', Microwave and Optical Technology
Letters, Vol.41, Issue 4, pp. 258-261, 2004
CONFERENCE PUBLICATIONS

1. Pavel Belov, Yang Hao and Sunil Sudhakaran, Sub wavelength imaging
using a lens formed by an array of parallel conducting wires, Proc.
Loughborough Antennas and Propagation Conference 2006 (LAPC 2006)
Loughborough, UK, April 2006.
2. S. Sudhakaran, Y. Hao, D. Nyberg, C.G Parini, 'Negative refraction from thick
metallic wire medium', Proc. URSI General Assembly, October, New Delhi,
India, October 2005
3. Yang Hao, S. Sudhakaran, D. Nyberg, C. G Parini, H. Cory, 'Metamaterial
realisation and its applications to antennas', Proc. Progress in Electromagnetics
(PIERS) 2005, Hangzhou, China, August 2005.

iv
4. S. Sudhakaran, Y. Hao, C. G Parini, 'Experimental verification of wave
focussing from metallic wire media', Proc. Loughborough Antennas and
Propagation Conference (LAPC) 2005, Loughborough, UK, April 2005
5. Y. Hao, S. Sudhakaran, C.G Parini, 'Spatial harmonics selection in
metamaterials for antenna applications', Proc. of Bianisotropics 2004 (BIAN'
04)-10th Conference on Complex Media and Metamaterials, Ghent, Belgium,
September 2004
6. S. Sudhakaran, Y. Hao, C. G Parini, 'An experimental verification of spatial
harmonics effects on negative refraction', Proc. IEEE AP-S International
Symposium on Antennas and Propagation and USNC/URSI National Radio
Science, Monterey, California, USA on June 2004.
7. S. Sudhakaran, Y. Hao, C. G Parini, J. Dupuy, ' Negative Refraction from
Generalised Periodic Structures - Measurement Verification', 27th ESA
Antenna Technology Workshop on Innovative Periodic Antennas, Spain, March
2004
8. Y. Hao, L. Lu, S. Sudhakaran and C. G. Parini, 'Numerical Techniques in Left-
handed Materials Modelling', Proc. Progress in Electromagnetics Research
Symposium (PIERS 2004), Pisa, Italy, March 2004
9. Y. Hao, L Lu, S Sudhakaran and C G Parini, 'From electromagnetic bandgap
to left handed metamaterials: modelling and applications', Proc. Meta materials
for Microwave and (Sub) millimetre Wave Applications: Photonic Bandgap and
Double Negative Designs, Components and Experiments, Seminar, London,
UK, November 2003
10. Y. Hao, S. Sudhakaran, C. G Parini, ' Spatial harmonic effects on
characterisation of left handed materials, Proc. Asia Pacific Microwave
Conference (APMC '03), Seoul, South Korea., November 2003

v
11. S. Sudhakaran, Y. Hao, C. G Parini, ' Characterisation of negative meta
materials based on Generalised periodic structure concept', Proc. 8th IEEE High
Frequency Post Graduate Student Colloquium, Belfast, Ireland, September 2003

vi
Table of Contents
Abstract i
Acknowledgements ii
Publications iii
Table of contents vi
List of figures xi
List of tables xxiii
List of abbreviations xxiv
Chapter 1: Introduction
1.1 Introduction 1
1.2 Left Handed Metamaterials (LHMs): A new class of materials 1
1.3 Photonic or electromagnetic Bandgap Structures (PBG/EBG) 4
1.4 Negative refraction from EBGs 6
1.5 Motivation for the research 8
1.6 Organisation of the thesis 10
References 12
Chapter 2: A Review of Negative Refraction Phenomenon
2.1 Introduction 16
2.1.2 Negative Permittivity and Permeability of materials 17
2.2 Dispersion in materials 18
2.3 Phase and group velocities 19
2.4 A brief history of negative refraction phenomenon 20
2.5 Maxwell equations and left handed phenomenon 21
2.6 Backward radiation in LHMs 23
2.7 Inverse Doppler Effect in LHMs 24
2.8 Terminology on negative refraction phenomenon 25

vii
2.9 Recent importance of LHMs studies 26
2.10 Realisation of LHMs 27
2.10.1 The Resonance Model- Split Ring Resonator (SRR) and Wire 28
2.10.1.1Inverse Snells Law verification 32
2.10.2 A planar equivalent high pass C-L medium 35
2.10.3 Composite right/ left handed (CRLH) transmission lines 39
2.10.4 Negative refraction from EBGs- The analogous LHMs 42
2.10.4.1 Analogy between EBGs and electronic crystals 42
2.10.4.2 All Angle Negative Refraction (AANR) 48
2.10.5 Drawbacks of the realisations 49
2.11 Applications of LHMs 53
2.11.1 Perfect lens 53
2.11.2 Scattering reduction using LHMs 56
2.11.3 Microwave components using LHM 58
2.11.5 Optical receivers using EBG based analogous LHM 62
2.12 Conclusions 64
References 66
Chapter3: Periodic Structures: Modelling and Negative Refraction Phenomenon
3.1 Introduction 73
3.2 Analysis of periodic structures 75
3.2.1 Circuit or network analysis 75
3.2.2 Wave analysis 75
3.3 Floquets theorem and spatial harmonics 78
3.4 - Diagram (Brillouin diagram) 81
3.5 Unit cell and Super cell analysis 83
3.6 Computational Electromagnetics (CEM) and simulation packages 87

viii
3.6.1 Finite Element Method (FEM) 89
3.6.1.1 Finite element method for electroamgentic problems 90
3.6.2 Finite Integration Technique (FIT) 91
3.7 EBG modelling in commercial packages 97
3.8 Numerical simulations of LHMs and analogous LHMs 95
3.9. Lumped element circuit simulations 103
3.9.1 C-L High pass filter circuit 104
3.9.2 L-C low pass filter circuit 106
3.10 Spatial harmonics suppression in periodic structures 108
3.11 Conclusions 111
References 112
Chapter 4: Negative Refraction Phenomenon from Metallic Wire Medium
4.1 Introduction 117
4.2 Braggs Law and bandgaps EBGs 118
4.3 Metallic electromagnetic bandgap structures (MEBGs) 123

4.3.1 Design of the MEBG Slab 120
4.3.2 Numerical simulation and discussion 125
4.3.3 Prism structure for Inverse Snells law verification 128
4.3.3.1 Design of the MEBG prism 128
4.3.3. 2. Measurement verification of refraction phenomenon 136
4.3.3.3 Inclusion of defects for enhancing negative refraction 142
4.4 Conclusions 148
References 149

ix
Chapter 5: Wave Focusing at Low Microwave Frequencies Using Metallic Wire
Medium
5.2 Design and numerical simulation 157
5.3. Measurement of focusing from MEBG slab 163
5.4 Discussion of results 166
5.5 Conclusions 170
References 172
Chapter 6: Experimental Investigation of Canalization and Imaging Using
Metallic Thin Wire Medium
6.2 The canalization regime for near field imaging 174
6.3. Design of Wire Medium Lens (WML) 178
6.4. Experimental verification 179
6.5. Interaction between source and lens 184
6.6 Conclusions 193
References 195
Chapter 7: Conclusions and Future work
7.1 Summary 197
7.2 Conclusions 197
7.3 Future work 198
7.3.1 Novel applications 199
7.3.1.1. Dispersive EBG for demultiplexing and beam scanning
applications 199
7.3.1.2 Sub-wavelength antennas 200
7.3.2 Theoretical extensions 200

x
7.3.2.1 Transmission enhancement 200
7.3.2.2. All angle negative refraction 200

xi
List of figures

Fig.1.1.A diagram showing the electric, magnetic, Poynting and wave vector directions
in a conventional right handed material (left) and in a left handed material (right). It can
be seen that the wave vector and Poynting vector are in opposite directions in
LHM 2
Fig.1.2.A ray diagram showing the wave propagation direction in a slab of RHM (left)
and LHM (right) when an electromagnetic wave obliquely incidents on it ....3
Fig.1.3.Diagram illustrating the application of EBG as a mirror and its comparison with
a metal reflector The EBG reflector allows propagation for waves at passband
frequencies and the metal inhibits waves at all frequencies.....6
Fig.2.1. A diagram showing the possible domains of electromagnetic materials and
wave refraction or reflection directions based on the signs of permittivity and
permeability. The arrows represent wave vector directions in each medium. There is
wave transmission only when both parameters are of same sign. Wave refracted
positively in conventional materials and negatively in LHMs
(Reproduced from [9])18
Fig.2.2. Energy flow and wave vector diagrams in a slab having negative refractive
index n = -1. The energy flow vectors are in same direction (shown in red) and flow
wave vectors are in opposite directions (shown in green). Since n = -1 the angles of
incidence and refraction are having same magnitude. (Reproduced from [21])...22
Fig.2.3.Doppler effect in a right handed medium (top) and in left handed medium
(bottom). A is the stationary source and B is the detector which moves with a velocity
v. The reference point is indicated in green....25
Fig.2.4: Unit cell of split ring contributing negative permeability (left) and wire (right).
The wires contribute negative permittivity. SRR can be characterised by the parameters

xii
w, d, g and c. Wire array can be characterised by radius a and the spacing between
each wires. The EM wave propagation direction is along the z
direction..29
Fig.2.5.A split ring and wire array having negative refractive index. The split rings
contribute negative permeability and wires contribute negative permittivity.
(Reproduced from [29]).31
Fig.2.6. Transmission through a medium consisting of SRRs only, wires only and
composite structure consisting of SRRs and wires showing increased transmission at
the frequency band 8.4-9.3GHz where the composite material posses negative
permittivity and permeability. The plasma frequency of the wire medium in this case is
around 20GHz. (Reproduced from [32)]31
Fig.2.7.A ray diagram showing refraction from a prism of RHM and LHM. For an
RHM prism the refracted ray bends towards the base making a positive angle with the
normal and in LHM prism the refracted ray bends away from the base making a
positive angle with the normal.
1
is the angle of incidence and
2
is the angle of
refraction.33
Fig.2.8.An experimental set-up used for Inverse Snells Law verification from a prism
composite medium consisting of SRRs and wires. (Reproduced from [24])34
Fig.2.9.A plot of refractive index vs. frequency for SRR-Wire medium. The red curve
is the real component of theoretical LHM data and the red dotted curve is the imaginary
part. The black curve gives the measured value in which the dotted line represents the
non reliable region of measurement due to experimental limitations. (Reproduced from
[24])34
Fig.2.10. Measured transmitted power at 10.5GHz as a function of refracted angle from
an LHM (SRRWire) prism and a Teflon prism. The Teflon prism shows an angle of

xiii
refraction of +27 (dashed curve) and for the LHM prism the angle of refraction is -61
(solid curve) (Reproduced from [24]).35
Fig.2.11.A unit cell of 2D loaded C-L high pass transmission line consisting of
capacitor and inductor. Each capacitor is having impedance per unit length Z
C
and each
inductor has admittance per unit length Y
L
. The unit cell length (d) is equal in both x
and y directions. (Reproduced from [35]).36
Fig.2.12. A backward antenna based on highpass C-L circuit (above) and the enlarged
view of the unit cell shown in the inset (below). (Reproduced from [36])38
Fig.2.13.Radiation pattern of a backward radiating CL structure. The backward
radiation is strong by 12dB at an angle of -40. (Reproduced from [36]).39
Fig.2.14.The CR-LH bandpass structure consisting of right handed transmission line
(RH-TL- high pass) and left handed transmission line (LH-TL- low pass)..40
Fig.2.15. Propagation () and attenuation () diagrams for a dominant mode leaky
wave antenna consisting of four regions. The LH and RH guided wave regions exist
when | | > k
0
(region I-LH and IV-RH), and leaky wave region exists when | | < k
0
.
When < 0 the structure supports LH backfire (region II) and when > 0 the structure
supports RH endfire (region III). (Reproduced from [40])41
Fig.2.16.EFS diagram showing wave propagation through a dielectric material. The
incident wave direction shown by red arrow and the refracted wave direction is denoted
by blue arrow. The propagation direction is determined using the normal to the EFS at
the k point. .43
Fig.2.17. EFS in an EBG in the long wave length limit and the medium can be
considered homogeneous44
Fig.2.18. Dispersion surface in a photonic crystal at high frequency where the medium
becomes inhomogeneous. The two EFS circles overlaps and form a bandgap. Around
the region L
o
Braggs law is satisfied and this region the bandgap exists..45

xiv
Fig.2.19.Schematic diagram of wave propagation through a weakly modulated
hexagonal photonic crystal (left) and for a strongly modulated photonic crystal at
different frequencies. The bandgap frequency is at = 0.48. ( Reproduced from
[45])46
Fig.2.20.A diagram showing analogy between EBGs and electronic crystals. The
effective mass in electronic crystal (right) is analogous to refractive index in EBG
(left)47
Fig. 2.21.Several iso frequency surfaces for the dielectric photonic crystal. The
frequency is written in units of 2c/a. It can be seen that the surfaces are convex in the
vicinity of M point. (Reproduced from [44]).49
Fig.2.22.Dispersion diagram for the photonic crystal. The all angle negative refraction
region is highlighted in red. (Reproduced from [44]).49
Fig.2.23.An illustration of the perfect lens formation from an LHM slab. The first focus
point is inside the slab and the second focus point is outside the slab at the right side.
(Reproduced from [57]).54
Fig.2.24. Experimental result showing focusing of electric field from an EBG due to
negative refraction [46]. The yellow point at the left side is the source point and yellow
point at right side is the image point (Reproduced from [42])...55
Fig.2.25. Experimental result showing focusing of electric field from an EBG due to
negative refraction when the source is moved 4cm upward. It can be seen that the
image at the left position also moved by the same distance (Reproduced from
[42]..56
Fig.2.26. Scattering pattern for a circular cylinder with coating. Curve l corresponds to
scattering when surface of the cylinder coated with parameters = -1-i*0.6 and = -1-
i*1, curve 2 represents the formal sign reversal for the real parts = 1-i*0.6 and = 1-
i*1. Curve 3 is the scattering when the material parameters are that of absorbing foam

xv
glass ( = 1 and = 1- i*0.6). Curve 4 shows scattering from an uncoated cylinder
(Reproduced from [64])..57
Fig. 2.27.A two layer structure consisting of positive and negative refractive index
materials. The left layer is a dielectric with > 0, >0 and the right layer is a lossless
material with negative permittivity and permeability ( < 0, < 0). In the left layer the
Poyting vector (S) and wave vector (k) are parallel. In the second layer they are anti
parallel....58
Fig.2.28.The two layer resonating structure consisting of RHM (green) and LHM
(yellow) slabs backed by conductors at both ends (shown in red) 60
Fig.2.29. Focusing from a crystal at negative refraction frequency. All the rays being
emitted from point A pass through point B and gather again at point C. (Reproduced
from [50])63
Fig.2.30. Schematic drawing of a field of view expander. The boundaries have the
curvatures r
1
and r
2
with the same centre of origin. The inset shows a magnified part of
the crystal surface. (Reproduced from [50])..63
Fig.2.31. Schematic drawing showing the reduction of illumination area obtained by
applying a photonic crystal with negative refraction characteristics (Reproduced from
[50])64
Fig.3.1. Diagram showing wave amplitudes in periodic structure. c
n
+
, c
+
n+1
are forward
propagating waves at the n
th
and (n+1)
th
terminals respectively and c
n
-
, c
-
n+1
are the
corresponding backward waves..76
Fig.3.2.- diagram of a C-L circuit showing phase velocity and group velocity at a
point x on the curve. The / at the point x gives phase velocity at that point and the
slope of the curve at the point x gives group velocity82
Fig.3.3. A Diagram illustrating the case of an infinite photonic crystal. The entire space
is occupied by the photonic crystal85

xvi
Fig.3.4.A Diagram illustrating the case of a finite periodic structure. The surrounding
extends to infinity and there exist Bloch modes and evanescent modes86
Fig.3.5.A Diagram illustrating the numerical approach for a finite periodic structure.
The boundaries are terminated by absorbing boundary conditions86
Fig.3.6. Categories within computational electromagnetics and the domain of the
packages CST Microwave studio and Ansoft HFSS.88
Fig.3.7.A cell V
i,j,k-1
of the cell complex G with the allocation of the electric grid
voltages e on the edges of A and the magnitude facet flux b through this surface8
Fig.3.8. The allocation of the six magnetic facet fluxes which have to be considered in
the evaluation of the closed surface integral for non-existence of magnetic charges
within the cell volume91
Fig.3.9. The spatial allocation of of a cell and a dual cell of the grid doublet {G, G1}
Fig.3.10. Near field intensity through multilayer LHM structure with refractive index n
= -1. (Reproduced from [34])..100
Fig.3.11: Simulated field showing refraction from a wedge with positive refractive
index (left) and negative refractive index (right). Refracted rays are shown in dotted
arrows. (Reproduced from [42])..101
Fig.3.12. Simulated electric field intensity showing positive refraction at 9GHz and
positive refraction at 7GHz from a metallic EBG (Reproduced from [43]).103
Fig.3.13. A unit cell of a high pass C-L filter with C= 0.5pF and L= 1.0nH...104
Fig.3.14. Response of the C-L filter showing pass band from about 5 GHz104
Fig.3.15. Dispersion diagram for the high pass C-L filter with unit element...105
Fig.3.16.Dispersion diagram for the high pass C-L filter with three elements105
Fig.3.17. A unit cell of a low pass L-C filter with C= 1.5pF and L= 1.5nH106
Fig.3.18 Response of the L-C filter showing pass band up to about 4.8 GHz.107
Fig.3.19. Dispersion diagram for the high pass L-C filter with unit element..107

xvii
Fig.3.20.Dispersion diagram for the high pass L-C filter with three elements108
Fig.3.21.A Pseudo Periodic Waveguide: A comb type structure (Reproduced from
[46])..109
Fig.4.1.A diagram showing passband and bandgap behaviour of an EBG..118
Fig.4.2.A diagram illustrating Braggs law..119
Fig.4.3.Complete bandgaps for TE modes as functions of r/a for an EBG with square
lattice consisting of metallic wires (Reproduced from [11])120
Fig.4.3.Schematic diagram of square lattice of period a consisting of metallic wires
of radius r..121
Fig.4.4. A diagram showing a metallic wire array which acts like a metallic surface at
low frequencies (since the wavelength is large in this case the spacing between the
wires are very small compared to the wavelength) and when frequency increases
(wavelength decreases) the spacing between the elements becomes comparable to the
wavelength122
Fig.4.5. A schematic diagram of the MEBG structure used for numerical simulation
studies and vector directions of electric field E, magnetic field H and propagation
vector k.124
Fig.4.6: Transmission response of the wire periodic structure showing bandgap around
8.5 GHz.126
Fig.4.7: Dispersion diagram for the wire periodic structure with period 3.15 for a
direction of propagation along x axis....126
Fig.4.8: A plot indicating magnetic field propagation directions outside and inside the
periodic structure at 7.5 GHz(left) and 6.5 GHz (right). The propagation directions in
each media are indicated by arrows. Surface normal is shown in dotted lines127
Fig.4.9. Diagram of the prism consisting of metallic wires with period 1.75
cm.129

xviii
Fig.4.10. Unit cell assignment in CST simulation model for the simulation of bandgap
diagram for EBGs consisting of square lattice and wave propagation directions , X
and M130
Fig.4.11.CST Simulation model for bandgap determination. Four side walls are defined
as periodic boundaries and bottom and top walls are defined as magnetic
boundaries131
Fig.4.12.Bandgap diagram for the metallic wire structure with r/a = 0.36 (TE
polarisation). Green region indicates complete bandgap.132
Fig. 4.13. Measured transmission response for the prism structure showing bandgap
around 5.7GHz.133
Fig.4.14: Dispersion Diagram for the prism periodic structure with metallic wires
having a period of 1.75 cm. The bandgap region is marked in blue134
Fig.4.15. Electric field plot showing positive refraction at 6.8 GHz. Units are
in V/m..135
Fig.4.16. Electric field plot showing negative refraction at 7.4 GHz. Units are
in V/m..135
Fig.4.17. A diagram showing the setup used for experimental studies136
Fig.4.18. Two prism structures used for measurement: Prism structure consists of
metallic wires (left) and Teflon prism (right)...137
Fig.4.19. Simulated electric field showing positive refraction from Teflon prism at
4.7GHz..138
Fig.4.20. Simulated electric field showing positive refraction from Teflon prism at
4.7GHz..139
Fig.4.21. Measured refracted field for Teflon prism at 4.7 GHz. The angle of refraction
is +23.139

xix
Fig.4.22: Measured refracted field for wire prism showing positive refraction at 4.7
GHz and negative refraction at 5.6 GHz, 7.2 GHz and 7.4 GHz..140
Fig.4.23. Diagram of radiating dipole inside an EBG with defects for changing the
radiation pattern and corresponding radiation patterns. The blue dots indicate the defect
positions. (Reproduced from [32])...144
Fig.4.24. Simulated electric field at 4.3 GHz on prism with defects showing negative
refraction..145
Fig.4.25. Measured refracted field from wire prism structure showing positive
refraction at 4.3 GHz, 4.7 GHz and negative refraction at 4.3 GHz after the introduction
of the defects into the structure146
Fig.4.26. Dispersion diagrams for the wire structure without the defects and with the
defects has been introduced.146
Fig.5.1. Wave focusing in a slab with negative refractive index. n = -1. A is the source
point, B and C are first and second image points respectively.154
Fig.5.2. Simulated electric field plot showing focusing from a slab of refractive index n
= -1155
Fig.5.3. A graphical representation simulated and measured average power intensity
distribution at the image plane with the EBG and the intensity without the EBG
structure. The red curve shows the FDTD simulation result and the blue line represents
the experimental result. The green dotted curve shows the field intensity without the
EBG. It can be seen that there is strong field intensity at the centre corresponding to the
image point. (Reproduced from [3])156
Fig.5.4. Schematic diagrams of the structure used for focusing studies (left) and the
experimental setup used for the experiment. There are four elements in z direction and
8 elements in y direction in the wire structure.158

xx
Fig.5.5.Dispersion diagram obtained for incident field perpendicular to the wire axis.
The negative refraction regions and bandgap regions for the incident field direction
along the z axis are marked in the frequency region starting from 5.5GHz.159
Fig.5.7.Simulated electric field intensity along the yz plane at 2.1GHz showing
focusing. The focus point is marked in the rectangular box.160
Fig.5.8.Simulated electric field intensity along the yz plane at 3.4GHz showing no
focusing.161
Fig.5.9.Simulated electric field intensity along xz plane at a distance of 2.4cm from the
source161
Fig.5.10. Simulated electric field intensity along xz plane at a distance of 3.4cm from
the source.162
Fig.5.11. Simulated total electric field intensity along xz plane at a distance of 7cm
from the source.162
Fig.5.12.A schematic diagram of the probe used (left) and a picture of the coaxial fed
probe used in the experiment for transmission and reception (right) ..163
Fig.5.13. Measured electric field intensity along xy plane from the slab consisting of
metallic wires at 2.1GHz showing focusing.164
metallic wires at 3.4GHz showing no focusing165
metallic wires at 4.2GHz showing focusing.166
Fig.5.16. Simulated electric field intensity along yz plane at 2.1 GHz with one layer of
wires..168
Fig.5.17. Simulated electric field intensity along the yz plane at 2.1GHz without any
slab structure showing no wave focusing 168

xxi
Fig.5.18. Simulated electric field intensity along the yz plane at 2.1GHz with a slab of
permittivity 2.0 showing no wave focusing..169
Fig.5.19. Simulated electric field intensity long the yz plane at 2.1GHz with a slab of
permittivity 80.0 showing no wave focusing170
Fig.6.1. Geometry of a slab lens consisting of metallic wires. The length of the wires is
an integer multiple of /2.175
Fig.6.2. A photograph of the wire medium lens with the P loop source (left) and a
diagram showing the dimensions of the P loop source.179
Fig.6.3. Distribution of electric field and its absolute value: (A), (C) in the vicinity of
the source (at 2.5mm distance from the front interface; (B), (D) at 2.5mm distance from
the back interface; (E) the transverse plane and (F) the intensity along the transverse
plane (reproduced from [9])180
Fig.6.4. A photograph of the two probes used for the measurement studies181
Fig.6.5. A diagram of the straight probe and the bend probe used for the measurements
and corresponding electric field directions...181
Fig.6.6. Measured near field intensity distribution along the yz plane: Absolute values
of x, y and z components of the electric fields (A, C, E, G) in source plane in image
plane (B, D, F, H).183
Fig.6.7.A photograph of the Meander shaped radiating source used for the experiment
and a diagram showing dimensions of the loop185
Fig.6.8.Measured electric field intensity distribution in the yz plane at 0.85GHz (A) and
0.86GHz (B). Source fields with and without lens (left and right respectively) and the
image field (middle) are given..187
Fig.6.9.Measured electric field intensity distribution in the yz planes at 0.87GHz (A),
0.88GHz (B), 0.89GHz (C) and 0.9GHz (D). Source fields with and without lens (left
and right respectively) and the image field (middle) are given188

xxii
Fig. 6.10. Measured electric field intensity distribution in the yz plane at 0.91GHz (A)
and 0.92GHz (B), 0.93GHz (C) and 0.94GHz (D). Source fields with and without lens
(left and right respectively) and the image field (middle) are given..189
Fig.6.11.Measured electric field intensity distribution in the yz plane at 0.95 GHz (A),
0.96 GHz (B), 0.99 GHz (C) and at 1 GHz (D). Source fields with and without lens
(left and right respectively) and the image field (middle) are
given...190
Fig.6.12.Measured electric field distribution at 155 mm away in front of the source
showing no imaging....192
Fig.6.13.Measured electric field distribution at 3 mm away from the foam sheet at the
backside of the source (12 mm from the source) and at 3 mm away from the foam sheet
at the front side at 0.88 GHz. In both cases the near field distribution is the same and
there is no considerable transmission loss193
Fig.7.1. A schematic diagram illustrating the application of MEBG prism as a spatial
frequency demultiplexer. The red arrow represents the polychromatic wave and the out
coming beams of different frequencies are shown in different colours and the
corresponding angles from the normal are marked..199

xxiii

List of tables

Table.3.1. A comparison of FIT, FEM methods.97
Table.4.1. Three steps for obtaining the complete bandgap diagram and
their corresponding phase shifts for the parameter sweep.131
Table: 4.2. Comparison of measurement results at 7.2GHz with similar works for
verification of inverse Snells law using different structures142

xxiv

List of abbreviations

AANR All Angle Negative Refraction
CR-LH Composite Right Handed- Left Handed
DNG Double Negative
EBG Electromagnetic Bandgap
FEM Finite Element Method
FIT Finite Integration technique
LHM Left Handed Metamaterial
RHM Right Handed Material
RH-TL Right Handed Transmission Line
LH-TL Left Handed Transmission Line
PC Photonic Crystal
MEBG Metallic Electromagnetic Bandgap
SRR Split Ring Resonator
TL Transmission Line
WML Wire Medium Lens

1

Chapter 1

Introduction

1.1 Introduction
Electromagnetics has received great attention among researchers all over the
world because of its immense civilian and defence applications. During the Second
World War, the use of radar and thereafter the wide use of microwave communication
systems facilitated the transformation from radio to microwave frequency. This
dramatic change demanded more advanced materials for high frequency performance
and opened up new dimensions in the field of electromagnetic materials. Artificial
materials were developed with desired dielectric and magnetic characteristics. Now
modern fabrication facilities can enable more advanced materials with superior
characteristics which cannot be obtained in nature. Such artificial materials with the
properties which cannot be found in ubiquitous materials are called Metamaterials.
Nano-composites and electromagnetic bandgap structures are examples of
metamaterials.

1.2 Left Handed Metamaterials (LHMs): A new class of materials
In 1968 Russian physicist Vaselago proposed the possibility of materials with
both dielectric parameter permittivity () and magnetic parameter permeability ()
negative [1]. He referred these materials with negative parameters as Left Handed
2
Metamaterials (LHMs). When an electromagnetic wave passes through these materials,
the electric field vector ( E
v
), magnetic field vector ( H
v
) and the wave vector ( k
v
) obey
a left handed rule whereas in natural materials these vectors obey a right handed rule
and they have positive material parameters. Such materials with positive parameters are
termed as Right Handed Materials (RHMs). Fig.1.1 shows the wave vector directions in
RHMs and in LHMs. It can be seen that the pointing vector S
v
and wave vector k
v
are
in opposite directions in an LHM.

Fig.1.1. A diagram showing the electric, magnetic, Poynting and wave vector directions in a
conventional right handed material (left) and in a left handed metamaterial (right). It can be
seen that the wave vector and Poynting vector are in opposite directions in LHM

LHMs show interesting characteristics such as Inverse Snells law, backward
radiation and Inverse Doppler effect [1]. Fig.1.2 shows the wave propagation directions
using ray diagrams in a slab of conventional RHM and LHM, when an electromagnetic
wave obliquely incident on it. It can be seen that the direction of wave refraction at the
first interface is upward (inside the slab) with respect to the normal in the RHM slab
and downward in the LHM slab. Recently, a composite structure consisting of split ring
resonators (SRRs) and conducting wires was identified as a medium with negative
refractive index [2]. In this composite medium, an array of split rings gives rise to
H
v

E
v

k
v
S
v
E
v

k
v

S
v
H
v
RHM LHM
3
effective negative permeability and an array of wires exhibit negative effective
permittivity.

Fig.1.2.A ray diagram showing the wave propagation direction in a slab of RHM (left) and
LHM (right) when an electromagnetic wave obliquely incidents on it.

When these structures are combined, an effective negative refractive index at the
overlapping frequency bands can be enabled. This was experimentally verified by
Shelby et al. [3] in demonstrating the Inverse Snells Law from a structure consisting of
SRRs and wires. Later, a transmission line model was proposed having LHM
characteristics. It was demonstrated in [4] that a periodic array of capacitor-inductor
(C-L) high pass filter network showed equivalent effective negative parameters at the
regimes of effective medium theory (EMT).
In this medium, the capacitors contribute an effective negative permittivity and
inductors give effective negative permeability. Lately, some interesting applications of
LHMs such as a perfect lens which is free from all aberrations, sub-wavelength
resonators, zero phase delay lines etc. were put forward [5, 6]. A detailed analysis of
LHMs realisations and its applications will be presented in chapter 2 of this thesis.

RHM

LHM
4
1.3 Photonic or Electromagnetic Bandgap Structures (PBG/EBG)
PBGs are periodic arrangements of dielectric or metallic elements in one, two or
three dimensional manners. PBGs inhibit the passage of electromagnetic wave at
certain angles of incidence at some frequencies. These frequencies are called partial
bandgaps. At a specific frequency band, a PBG does not allow the propagation of wave
in all directions and this frequency region is called the complete bandgap or global
bandgap [7-9]. The original idea of PBGs was put forward by physicists and some
recent studies revealed the interesting fact that PBGs exist in living organisms [10-12].
The well known examples are the butterfly scales [11] and eyes of some insects [10]. In
biological systems, there is no metal. In this case, a metallic like reflection effect is
obtained by using refractive index differences. A multilayer thin film with different
refractive indices in animals is a good example for this. Recently, these ideas were
undergone a preliminary study for its commercialisation such as paints for certain
applications [10].
PBGs can be realised in one, two and three dimensional forms. The
dimensionality depends on the periodicity directions. Three dimensional PBGs are
more appropriate for getting a complete bandgap because they can inhibit waves for all
incident angles. The bandgap in PBGs is analogous to a forbidden energy gap in
electronic crystals. Hence PBGs are also termed as photonic crystals (PCs). The first
attempts towards three dimensional structures were realised in the form of face centred
cubic (fcc) lattice structures [13]. At the initial stages of PBG research, due to the lack
of theoretical predictions, a cut and try approach was adopted in experimentally
predicting the bandgap. At the beginning, the investigations of PBGs were mainly on
wave interactions of these structures at optical frequencies and hence PBGs emerged
with the name of photonic bandgap structures. Now, vast extensions of PBGs at
microwave [14], millimetre [15] and sub-millimetre [16] wave frequencies are
5
common. At microwave frequencies, such structures are popularly called as
Electromagnetic Bandgap (EBG) structures [17]. In the next sections of this thesis,
these materials are addressed as EBGs.
A periodic structure can give rise to multiple bandgaps. However, it should be
noted that the bandgap in EBG is not only due to the periodicity of the structure but
also due to the individual resonance of one element. A study by S. John et al. [13]
revealed the mechanisms to form a bandgap in an EBG. He proposed that the bandgap
formation in EBGs is due to the interplay between macroscopic and microscopic
resonances of a periodic structure. The periodicity governs the macroscopic resonance
or the Bragg resonance. It is also called the lattice resonance. Microscopic resonance is
due to the element characteristics and it is called the Mie resonance [18]. When the two
resonances coincide, the structure possesses a bandgap having maximum width.
Depending on the structural characteristics and polarisation of the wave, one resonance
mechanism (i.e. either the multiple scattering resonance or the single element scatterer
resonances) can dominate over the other [18]. The characteristic property of stop bands
at certain frequencies enables many applications using EBGs. At this stop band, all
electromagnetic wave will be reflected back and the structure will act like a mirror [18].
The advantage over a metal reflector is that for an EBG, reflection takes place only at
stopband frequency. At other frequencies it will act as transparent medium. This
concept is illustrated in fig.1.3.
EBGs found their applications in improving the performance of microstrip
antennas. Microstrip antennas mounted on a substrate can radiate only a small amount
of its power into free space because of the power leak through the dielectric substrate
[20]. In order to increase the efficiency of the antenna, the propagation through the
substrate must be prohibited. In this case, the antenna can radiate more towards the
main beam direction and hence increase its efficiency.
6

Fig.1.3.Diagram illustrating the application of EBG as a mirror and its comparison with a metal
reflector The EBG reflector allows propagation for waves at passband frequencies and the
metal inhibits waves at all frequencies.

An antenna with an EBG substrate can radiate more efficiently at frequencies
corresponding to the bandgap region [17, 20-22]. At this region, the substrate will not
allow radiated power to pass through it and the efficiency of the antenna will be
increased. In [20] such a design was detailed for a patch antenna operating at 14GHz.
The gain was increased by 1.8 dB corresponding to a radiated efficiency increment of
2%. EBGs found interesting wave-guiding applications by applying line defects which
can be obtained by removing a line of elements from the EBG. This produces defect
propagation modes inside the bandgap. Using this approach, a 90 degree bend [23, 24]
can be obtained which is not possible with conventional waveguides. The
characteristics of EBGs are completely not obtainable from naturally occurring
materials and hence EBGs belong to the class of Metamaterials.

1.4 Negative refraction from EBGs
It was revealed that negative refraction phenomenon can be observed in EBGs
due to the strong dispersion characteristics at certain frequencies [25, 26]. In [27], it
was shown that such a phenomenon was possible in the regimes of negative group
7
velocity and negative effective index at higher bands. It was also identified that at
certain frequencies, due to the negative effective mass of the crystal which is analogous
to a negative effective refractive index, negative refraction can be obtained [28,29]. The
negative refraction phenomenon from a dielectric EBG was investigated in [30] and it
was observed that the EBG supports the phenomenon for all incident angles at a narrow
band frequency. This was due to the negative photonic effective mass of the crystal.
Wave focusing from a dielectric slab with periodic air holes was demonstrated using
numerical simulation in [31]. It was made clear that neither a negative group velocity
nor a negative effective index is a prerequisite for negative refraction [31, 32].
Negative refraction phenomenon in EBGs is different from that in LHMs [29].
The advantage in obtaining negative refraction from EBGs is that it can explore the
phenomenon at optical frequencies at which the resonance models [2] for LHMs will
show limitations due to effective medium theory approximations. It appears that such
approximations are possible in very narrow band frequencies in the microwave regime
[30]. An EBG at negative refraction frequencies may not have negative effective
parameters [29]. This narrow frequency band at which the EBG exhibits negative
refraction characteristics were identified near the bandgap edges [29, 30]. The wave
vector diagrams for infinite EBGs at a constant frequency, called the equi-frequency
surfaces [25-29] were used for all these investigations. However, in an EBG, the
characteristics can not be completely governed by a single wave vector but by a
combination of many [30]. Also, in a finite EBG the dispersion characteristics are not
only governed by Bloch modes inside the EBG but also the evanescent modes and
coupling factors between the incident field and the finite structure [30], which were
completely neglected in the infinite EBGs analysis. Since the mechanism of negative
refraction from EBGs is different from that in LHMs, it is more accurate to regard such
EBGs as analogous LHMs rather than LHMs.
8
1.5 Motivation for the research
In the last few years, the field of metamaterials (LHMs and EBGs) received
much interest. Novel realisations of LHMs and their applications are in great demand.
It is noted that the LHMs or the analogous LHMs realised so far belong to the general
class of EBGs. This leads to great enthusiasm on the roles of periodic structures in
negative refraction phenomenon and the factors that need to be addressed in finite
structures especially the spatial dispersion which is inherent in all EBGs. Even in the
larger wavelength limit, the spatial dispersion can be strong in some EBGs, for
example, a metallic wire medium [25].
The research carried out in this thesis investigates the possibilities of obtaining
the negative refraction phenomenon from EBGs. Based on the EMT approach, the
period of the structure should be very small compared to the operating wavelength
(typically 1/100
th
of the wavelength) [32]. This limits its applicability for most of the
practically realisable structures at high frequencies and even at microwave frequencies.
When the period is comparable to the wavelength, the spatial dispersion in the structure
will be very high [25]. Also, in this case the effects of high order Floquets harmonics
are not negligible [34]. So the medium can not be characterised by assigning a single
permittivity and permeability value for throughout the finite structure. In other words,
the parameters depend on the spatial coordinates and this causes spatial dispersion. As
a result of this, the medium will not be homogenous.
The conventional analysis by using a unit cell of the infinite model with
periodic boundaries has taken only the characteristics of Bloch modes into account. In
fact, in a finite EBG there exist evanescent modes and a coupling between the incident
field and the structure. The unit cell analysis will give only fundamental Floquets
harmonic interaction which is not adequate for the complete dispersion nature of finite
9
crystals. This indicates that finite EBGs need to be addressed separately for negative
refraction studies.
In this work, the possibility of negative refraction due to the spatial dispersion
and from Floquets harmonics interactions at frequencies in the passband is
investigated. Imaging applications using the metamaterial nature of the EBGs are also
presented. An approach to enhance the dispersion prediction of finite EBGs is
proposed. It has been identified that at certain narrow frequency bands the structure
gives rise to negative refraction. Introduction of defects into the structure for enhancing
negative refraction at certain frequencies is verified. Negative refraction at narrow
frequency bands in passband are identified and confirmed by numerical simulations and
experiments.
In this work, wave focusing from metallic EBGs is demonstrated at low
microwave frequencies. Earlier works [34, 35] in this subject considered dielectric
EBGs and the measurement was at higher frequencies in one dimensional plane. This
does not guarantee the exact resolution of the image. The proposed idea of negative
refraction using spatial dispersion enables wave focusing at low frequencies in the
passband. In order to investigate the resolution of the image formed by a finite EBG
consisting of thick metallic wires, focusing measurement was performed on a two
dimensional scanning plane and good resolution of an image is obtained. The
phenomenon is observed from 2.1 GHz at multiple bands. In this work, dispersion and
negative refraction studies for realisable, finite EBGs is carried out and identification of
frequency regions where the structure shows negative refraction is carried out.
The metamaterial behaviour of a medium consisting of periodic array of thin
metallic wires is demonstrated for sub-wavelength imaging. Experimental observations
illustrated sub-wavelength imaging over a broad bandwidth. This imaging regime is
10
called canalisation. A study on the interaction of source and the proposed wire medium
lens was studied experimentally.
Simulation results were obtained using commercial packages High Frequency
Structure Simulator (HFSS) from Ansoft Inc., Advanced Design System (ADS)
from Agilent Technologies and CST Microwave Studio. Measurements were carried
out and the results obtained were found to be in good matching with numerical
simulations.

1.6 Organisation of the thesis
The thesis is divided into seven chapters. The organisation of the chapters can be
outlined as follows.
Chapter 1: In this chapter, an introduction to metameterials, LHMs and EBGs are
presented with the motivation factors that lead to the work presented in the next
chapters of the thesis.
Chapter 2: Literature review on LHMs and their characteristics, realisations of LHMs
including resonance model, EBG model, high pass C-L model, the composite right
handed- left handed media (CR-LH) and negative refraction from infinite EBGs are
presented in this chapter. Interesting applications of LHMs such as perfect lens
focusing, backward radiating antennas, sub-wavelength resonators and phase shifters
are also detailed.
Chapter 3: This chapter describes general approaches on the analysis of EBGs; wave
amplitude analysis and unit cell approach for infinite structures. Spatial dispersion in
finite models and its effects on negative refraction in finite EBGs are also discussed.
An overview on computational approaches used in this thesis and a brief review on
numerical simulations on LHMs and analogous LHMs is also presented. The proposed
super cell analysis to include higher order spatial harmonics and dispersion has been
11
verified with numerical simulations of lumped circuit models using Agilent ADS. It
is observed that when more elements are included for analysis the backward radiation
frequency regions start deviating from that of the unit cell model.
Chapter 4: In this chapter, metallic electromagnetic periodic structures (MEBGs) are
studied in detail for negative refraction phenomenon. It is observed that such a
phenomenon can be obtained at narrow frequencies in passband due to the strong
spatial dispersion from the finite structures. Negative refraction is verified using
numerical simulations and measurements. It is shown that introduction of defects into
the structure can enhance the negative refraction at certain frequencies. Measurement
results are in good agreement with simulation results.
Chapter 5: Wave focusing from MEBGs is detailed in this chapter. High resolution
focusing at low microwave frequencies in the passband is presented. A resolution of
/7 is demonstrated. Two dimensional scanning is performed to verify the focusing
phenomenon from the structure. The measurement results are matching with that
obtained from numerical simulations. The metamaterial behaviour of the slab under
study has been further verified with numerical simulations using more cases. It is
illustrated that the observed phenomenon is due to the metamaterial nature of the
MEBG slab.
Chapter 6: In this chapter, experimental investigation on canalization, a near field sub-
wavelength imaging regime using metamaterials consisting of metallic wires is
presented. High resolution images were obtained using metallic wire media at a
resolution of /15. It is proved using experimental results that the source - wire medium
interactions in near field imaging systems reduce the bandwidth of the imaging regime
but the image resolution is maintained.
Chapter 7: This chapter presents the conclusions of the research carried out and the
proposed future work directions.
12
References
[1] V. G. Vaselago, The electrodynamics of substances with simultaneously negative
values of and , Soviet Physics USPEKHI, Vol. 10, pp 509-514, 1968.
[2] D. R. Smith, Wille J. Padillla, D.C Vier, S.C Nemat- Naasser and Schults,
Composite medium with simultaneously negative permeability and permittivity,
Physical Review Letters, Vol. 84, pp. 4184-4187, 2000.
[3] R. A. Shelby, D. R. Smith, S. Schultz, Experimental verification of negative
index of refraction, Science, Vol. 292, pp. 77-79, 2001.
[4] Anthony Garbic and George V. Eleftheriades, A backward wave antenna
based on negative refractive index L-C networks, Proc. of the IEEE Intl.
Symposium on Antennas and Propagation, San Antonio, TX, Vol. IV, pp. 340-
343, 2002
[5] J. B. Pendry, Negative refraction makes a perfect lens, Physical Review
Letters , Vol. 85, pp. 39663969, 2000
[6] N. Engheta, "An idea for thin, sub-wavelength cavity resonators using
metamaterials with negative permittivity and permeability," IEEE Antennas and
Wireless Propagation Lett. Vol. 1, pp. 10-13, 2002
[7] E. Yeblonovitch, Photonic band-gap structures, Journal of Optical Society of
America. B, Vol. 10, pp. 283, 1993
[8] E. Yeblonovitch, T. J Gmitter, Photonic band structure: The face-centered-
cubic case, Physical Review Letters, Vol. 63, pp.1950-1953, 1989
[9] Stavroula Foteinopolou, Electromagnetic wave propagation in two dimensional
photonic crystals, PhD Dissertation, Iowa State University, 2003.
[10] Andrew R. Parker, Natural photonic engineers, Material Today, Vol. 5, pp.
26-31, 2002
13
[11] A. Argyrose, S. Manos, M. C. J Large, D.R McKenzie, G.C Cox and D. M
Dwarte, Electron tomography and computer visualization of three-dimensional
photonic crystal in butterfly wing-scale, Micron, Vol. 33, pp. 483-487, 2002
[12] L. P Biro, Z. Balint, K. Kertesz, Z. Vertesy, G. I Mark, Z.E Horwath, J.
Balazs, D. Mehn, I. Kiricsi, V. Lousse, and J.P Vigneron, Role of photonic
crystal-type structures in the thermal regulation of Lycaenid butterfly sister
species pair, Physical Review E, Vol. 67, pp. 021907-021914, 2003
[13] S. John, E. Yeblonovitch, Strong localization of photons in certain disordered
dielectric superlattices, Physical Review Letters, Vol. 58, pp. 2486-2489, 1987
[14] E. Ozbay, A. Abeyta, G. Tuttle, M. Tringides, R. Biswas, C. T Chan, C. M
Soukoulis and K. M Ho, Measurement of a three-dimensional photonic band
gap in a crystal structure made of dielectric rods, Physical Review B, Vol. 50, pp.
1945-1948, 1994
[15] E. Ozbay, E. Michel, G. Tuttle, R. Biswas, M. Sigalas and K. M Ho,
Micromachined millimeter-wave photonic band-gap crystals, Applied Physics
Letters, Vol. 64, pp. 2059 -2061, 1994
[16] E. Ozbay, E. Michel, G. Tuttle, R. Biswas, K. M Ho, J. Bostak, and D. M
Bloom, Terahertz spectroscopy of three-dimensional photonic band-gap crystal,
Optics Letters, Vol.19, pp. 1155, 1994
[17] Y. Zhao, Y. Hao, C. G. Parini, Radiation Properties of PIFA on UC-EBG
substrates, Microwave and Optical Technology Letters, Vol. 44, pp. 21-24, 2005
[18] E. Lidorikis, M. M Sigalas, E. N Economou and C. M Saoukoulis, Gap
deformation and classical wave localization in disordered 2d photonic band gap
materials, Physical Review B, 61, pp. 13458-13464, 2000
[19] J. D Joannopoulos, R. D Mead and J. N Winn, Photonic crystals: Modelling
the Flow of Light, (Princeston University Press, Princeston), 2000
14
[20] E.R Brown, C. D Parker and E. Yeblonovitch, Radiation properties of a planar
antenna on a photonic-crystal substrate, Journal of Optical Society of America,
Vol. 10, pp. 404-407, 1993
[21] Yongxi Qian, Roberto Coccioli, Dan Sievenpiper, Vesna Radisic, Eli
Yablonovitch and Tatsuo Itoh, A microstrip patch antenna using novel
photonic band-gap structures, Microwave Journal, Vol. 42, pp. 66(6), 2000
[22] R. Coccioli, W. R. Deal, and T. Itoh, "Radiation characteristics of a patch
antenna on a thin PBG substrate", IEEE Transactions on Antennas and
Propagation, Vol. 45, pp. 656-659, 1998.
[23] Shawn-Yu Lin, Edmund Chow, Vince Hietala, Pierre R. Villeneuve, and J. D.
Joannopoulos, Experimental demonstration of guiding and bending of
electromagnetic waves in a photonic crystal, Science, Vol. 282, pp. 274-276,
1998
[24] S. Y Lin, V. M Hietala, S. K Lyo and A. Zaslavsky, Photonic band gap
quantum well and quantum box structures: A high-Q resonant cavity, Applied
Physics Letters, Vol. 68, pp. 3233-3235, 1996
[25] Masaya Notomi, Negative refraction in photonic crystals, Optical and
Quantum Electronics, Vol. 34, pp.133143, 2002
[26] Kosaka, H., T. Kawashima, A. Tomita, M. Notomi, T. Tamamura, T. Sato
and S. Kawakami, Superprism phenomenon in photonic crystals, Physical
Review B, Vol. 58, pp. 10096-10099, 1998.
[27] Chiyan Luo, Steven G. Johnson, and J. D. Joannopoulos, Negative refraction
without negative index in metallic photonic crystals, Optics Express, Vol. 11, pp.
746-754, 2003
15
[28] Chiyan Luo, Steven G. Johnson, and J. D. Joannopoulos, All-angle negative
refraction without negative effective index, Physical Review B, Vol. 65, pp.
201104-201107, 2003
[29] P.V Parimi, W. T. Lu, P. Vodo, J. Sokoloff, S. Sridhar, Negative refraction
and left-handed electromagnetism in microwave photonic crystals, cond-
mat/0306109, 2003 (Online archive)
[30] Boris Gralak, Stefan Enoch and Gerard Tayeb, Anomalous refractive
properties of photonic crystals, Journal of Optical Society of America, Vol. 17,
pp.1012-1020, 2002
[31] Vito Mocella, Negative refraction in photonic crystals: thickness dependence
and Pendellosung phenomenon, Optics Express, Vol. 13, pp. 1361-1367, 2005
[32] P. A. Belov, R. Marques, S. I Maslovski, I. S. Nefedov, M. Silverinha, C. R.
Simovski and S. A. Tretyakov, Strong spatial dispersion in wire media in the
very large wavelength limit, Physical Review B, Vol. 67, pp. 113103-113106,
2003
[33] R. Marqus, F. Medina, An Introductory overview on left-handed
metamaterials, Proc. the 27
th
ESA Antenna Workshop on Innovative Periodic
Antennas, Spain, pp. 35-41, 2004
[34] R. Chatterjee, Elements of Microwave Engineering, Chapter 9, John Wiley,
1986.
[35] E E. Cubuksu, K. Aydin, and E. Ozbay, S. Foteinpolou, C. M. Soukoulis,
Sub-wavelength resolution in a two dimensional photonic crystal based
superlens, Physical Review Letters., Vol. 91, pp. 207401, 2002
[36] Patanjali V. Parimi, Wentao T.U, Plarenta Vodo, Srinivas Sridhar, Imaging
by flat lens using negative refraction, Nature, Vol.426, pp. 404, 2003

16
Chapter 2

A Review of Negative Refraction
Phenomenon

2.1 Introduction

In mechanical, electrical and thermal scenarios, materials exhibit different
characteristics. The electrical and magnetic properties are used to describe the
characteristics of a material when it is exposed to an electromagnetic field. Permittivity
() and permeability () are two parameters to govern electromagnetic performance of
all materials. The permittivity is a measure of how much a medium changes to absorb
energy when subject to an electric field [1-3]. It is defined as the ratio
E
D
v
v
where D
v
is
the electric displacement by the medium and E
v
is the electric field strength. The
common term dielectric constant is the ratio of permittivity of the material to that of
free space (
0
= 8.85 * 10
-12
F/m). It is also termed as the relative permittivity [2, 3].
Permeability is a constant of proportionality that exists between magnetic induction and
magnetic field intensity. Free space permeability (
0
) is approximately 1.257 x 10
-6

H/m. In other materials, it is usually greater than the free-space value. Due to the recent
advances in material fabrication techniques, highly engineered materials that exhibit
superior properties which can not be observed in naturally occurring materials can be
made. These structured composite materials are known as Metamaterials [4, 5].
Chapter 2: A Review of Negative Refraction Phenomenon_
17
Left handed metamaterials (LHMs), EBGs and artificial magneto-dielectric materials
[6] are examples of metamaterials. Among them, LHMs received great research
interests in the past few years. In the first part of this chapter, an overview of LHMs,
their properties and realisations based on resonance, high pass C-L and dispersive EBG
models are presented. This is followed by some exciting applications of LHMs and
their analogous.

2.1.2 Negative Permittivity and Permeability of materials
Negative permittivity can be obtained from some naturally available materials.
It can be achieved from metals at frequencies near their plasmonic resonance [7, 8].
Free electron gas with permittivity = 1-
2
p
/
2
is negative when the frequency
<
p
, where
p
is the plasma frequency.

Some crystals such as SiC, LiTaO
3
, LiF and
ZnSe can have negative permittivity at certain frequencies governed by their dispersion
relations [8]. Materials with negative permeability do not occur in nature. They have to
be constructed artificially.
Fig.2.1 illustrates a graphical representation of different material possibilities
for electromagnetic applications based on the signs of their permittivity and
permeability values and refraction and reflection at an interface between air and each
medium [9]. There are four regions in the diagram. Plasma belongs to the region with
negative permittivity and positive permeability. Split rings belong to the region of
negative permeability and positive permittivity. It can be seen that when the two
parameters are of opposite signs, there is no wave transmission. This is because when
the parameters are of opposite signs, the wave vector becomes imaginary. When both
parameters are positive, refraction occurs positively and when they are negative,
refraction occurs negatively.

18

Fig.2.1.A diagram showing the possible domains of electromagnetic materials and
wave refraction or reflection directions based on the signs of permittivity and
permeability. The arrows represent wave vector directions in each medium. There is
wave transmission only when both parameters are of same sign. Wave refracted
positively in conventional materials and negatively in LHMs (Reproduced from [9]).

2.2 Dispersion in materials
Dispersion is a phenomenon that causes the separation of a wave into
components with different frequencies, due to a dependence of the velocity of wave on
its frequency [3, 10]. A well familiar example of dispersion is the splitting-up of white
light into its components when it passes through a glass prism. There are generally two
sources of dispersion. First is the material dispersion, which results from a frequency
dependent response of a material when a wave passes through it. The second source is
the waveguide dispersion, which arises when the transverse mode solutions for waves
confined within a finite waveguide. This depends upon the frequency of operation.
LHMs
19
There is also modal dispersion, which comes when a signal consists of a superposition
of multiple modes at single frequency. In this case, different modes generally travel at
different speeds resulting in dispersion of temporal features.

2.3 Phase and group velocities
The phase velocity of a wave is the rate at which the phase of the wave
propagates in a medium [11]. This is the velocity at which the phase of any one
frequency component of the wave will propagate. One particular phase of the wave (for
example the crest) would appear to travel at the phase velocity. The phase velocity is
given by

k
v
p
v
v
= (2.1)

where is the frequency and k
v
is the wave vector.
Group velocity is the velocity with which the envelope of the wave propagates [3, 11].
The group velocity
g
v
v
is given by

1
|
.
|
\
|
=
d
dn
n c v
g
v

Where n is the refractive index of the medium, c is the velocity of light and is the
wavelength. In most cases, this is true and the group velocity is same as the signal
velocity of the waveform. But in certain cases, where the wavelength of light is close to
an absorption resonance of the medium, it is possible for the group velocity to exceed
the speed of light (v
g
> c), leading to the conclusion that superluminal (faster than light)
communication is possible [3]. In such situations, the value of group velocity becomes
meaningless, and does not represent the true signal velocity of the wave. The group
(2.2)
20
velocity itself is a function of frequency. This results in group velocity dispersion
(GVD), which can be quantified as the group delay dispersion parameter D.

|
|
.
|
\
|
=
2
2
d
n d
c
D

Due to GVD, a short pulse of light spread in time as different frequency components
can travel at different velocities. If D is less than zero, the medium is having positive
dispersion. If D is greater than zero, the medium has anomalous dispersion. If light
propagates through a normally dispersive medium, the result is that the higher
frequency components travel slower than the lower frequency components. If a pulse
travels through an anomalously dispersive medium, high frequency components travel
faster than the lower ones. So frequency decreases when the time increases. Dispersion
management and control is very important in communication systems especially at
optical frequencies [9].
2.4 A brief history of negative refraction phenomenon
Even though negative refraction received great attention recently, its possibility
has been suggested in some other ways long time ago. The existence of backward
waves in mechanical systems was suggested by H. Lamb [12, 13]. He identified that for
these waves the phase moves in the direction opposite to the energy flow indicating
opposite signs for phase and group velocities. However, Lamb has not given insight
into electromagnetic systems. Backward waves in electromagnetic systems were first
investigated by A. Schuster [14]. He detailed an optical refraction phenomenon in such
a media, if a material with such properties was ever to be found. So the possibility of
negative refraction was first appeared in this work. Schuster has not detailed what
could be the characteristic properties of such a medium. Pocklington [15] illustrated
(2.3)
21
that in a specific backward-wave medium, a source produces a wave whose group
velocity is directed away from the source, while its phase velocity moves toward the
source. In 1968 V. G Veselago [16] predicted the electrodynamics of materials with
both negative parameters [16] and he showed that such materials can support reversed
optical phenomenon such as Inverse Doppler Effect, backward radiation etc. This was
a major milestone in the history of LHMs.
2.5 Maxwell equations and left handed phenomenon
If a material has negative parameters, it must hold the energy propagation
relations given by Maxwell Equations [20]. The four Maxwell equations are given by

0 =
=
+
=
B
D
j
t
D
H
t
B
E
v
v
v
v
v
v
v

In order to see the effect of sign in material parameters, the relations in which ,
appear separately rather than in the form of their product need to be taken. These
equations are the two Maxwell equations 2.4 and 2.5 given above and the constitutive
relations.

E D
H B
v v
v v
=
=

(2.4)

(2.5)
(2.6)
(2.8)

(2.9)
(2.7)
22
Applying these equations for a plane monochromatic wave, it can be seen that if < 0
and < 0, then E
v
, H
v
and k
v
form a left handed set of vectors [16]. If
i
,
i
and
i
are the
direction cosines for E
v
, H
v
and k
v
vectors respectively, then a wave propagating in a
given medium will be characterised by the matrix

|
|
|
.
|
\
|
=
2 1
2 1
1
3
3

3 2
G

The determinant of this matrix p equals +1 if E
v
, H
v
and k
v
vectors are right handed set.
It is -1 if this set is left handed [16]. If the material parameters are negative, the
determinant p is -1. Hence these types of materials are called Left Handed
Metamaterials (LHMs).

Fig.2.2. Energy flow and wave vector diagrams in a slab having negative refractive index n = -
1. The energy flow vectors are in same direction (shown in red) and flow wave vectors are in
opposite directions (shown in green). Since n = -1 the angles of incidence and refraction are
having same magnitude. (Reproduced from [21])

(2.10)
23
2.6 Backward radiation in LHMs
Consider an electric field E
v
polarized along the unit vector direction p
v
. The
electric field E
v
and magnetic field H
r
can then be expressed as [17, 18]

r k i
e p E
v
v
v
v

=

r k i
e p k H
v
v
v
v v

= ) (
1

From Maxwells equation, the permeability and permittivity are related with electric
and magnetic fields as follows [17]

From Poynting Theorem, the power flow direction can be obtained by the time
averaged Poynting vector and is given by [18]

) (
1

*) (
p k p
H E R S
v
v
v
v v v
=
=

where R is a real operator. On expanding the triple product in eqn.2.15, we get

| | p k p k p p S
v
v
v
v
v v
v
) . ( ) . (
1
=

and 0 ) . ( = k p
v
v
due to Gausss Law. Hence eqn.2.16 becomes
(2.16)
(2.11)
(2.12)
(2.14)
(2.13)
(2.15)
) ( ) ( - ) (
) ( ) ( ) (
r E r H k
r H r E k
v
v
v
v v
v
v
v
v v

=
=
24

Considering the fact that for wave propagation, both , should have the same sign (to
obtain a real wave vector), eqn.2.17 is valid only when both parameters have the same
sign. Hence it is clear from eqn. 2.17 that when both and are negative, then time
averaged Poynting vector S
v
will be in opposite direction of the phase propagation
vector. This leads to backward radiation characteristic of LHM.

2.7 Inverse Doppler Effect in LHMs
Doppler Effect is the noticeable change in frequency of a wave when an
observer moving away or towards the source of the wave. Let a detector of radiation
move relative to a source of radiation with a frequency
0.
During its motion, the
detector will observe points of the wave which correspond to some definite phase. For
example, crest of the wave. The formula for Doppler shift can be written as [16]

|
.
|
\
|
+ =
u
v
p 1
0

where v is observer velocity and u is the energy flux velocity. In this equation p denotes
that the medium is either RHM or LHM. If p = +1 the medium is RHM and if p = -1 the
medium is LHM. If the detector is in a medium of positive refractive index, the
frequency received by the detector will be larger than
0
since the argument in the
bracket in eqn. 2.18 is greater than 1 ( p = +1 in this case). From fig.2.3, it can be seen
that in an RHM, the detector B and k vector are moving in opposite directions. In this
case, the detector will pick up the reference points (shown in green) fast and this leads
k
S
v
v
=
(2.17)
(2.18)
25
to increase in frequency. If the detector is in a medium with negative parameters then p
= -1 giving the argument less than 1. So the frequency received by the detector will be
decreasing in an LHM when it travels towards the source.

Fig.2.3.Doppler Effect in a right handed medium (top) and in left handed medium (bottom). A
is the stationary source and B is the detector which moves with a velocity v. The reference
point is indicated in green.

In other words, the wave vector k and the detector velocity v are in the same direction
and hence the detector has to travel for more time to pick up the reference point. This
leads to a decrease in frequency. Hence it can be concluded that waves will observe a
reversed Doppler Effect in an LHM [16].

2.8 Terminology on negative refraction phenomenon
Vaselago [16] termed the materials with negative permittivity and permeability
as Left Handed Metamaterials (LHMs) or double negative materials (DNGs). Inverse
Doppler Effect and backward radiation are characteristic properties observed in these
materials. These are also referred as backward radiating metamaterials [19]. However,
one could find much literature using these terms frequently irrespective of the
uniqueness of LHMs. It is worth making these terms more clear. Metamaterials are
26
those materials whose properties can not be obtained in naturally occurring materials. It
is not necessarily to be an LHM. A medium with negative permittivity and permeability
can be accurately termed as LHM or double negative (DNG) metamaterial. In EBGs,
negative refraction phenomenon arises due to the dispersion nature of the medium at
some frequencies. Such materials can be accurately termed as negative refraction
medium or backward metamaterials or analogous LHMs rather than DNGs or LHMs.
Another possibility is a case where the material offers negative group delay (hence
negative group velocity) and negative refractive index as proposed by [20] and these
materials can be referred as negative anomalous refraction materials.

2.9 Recent importance of LHMs studies
Since Vaselagos prediction in 1968, it has taken more than three decades to
realise a composite material with LHM parameters. Following the ideas that an array of
split ring resonators (SRR) can posses negative permeability and an array of conducting
wires have negative permittivity [21,22], Smith et al. put forward a composite medium
which consists of a periodic array of interspaced conducting non-magnetic split ring
resonators (SRRs) and continuous wires [22,23]. Shelby et al. first experimentally
verified the negative refraction phenomenon and the existence of negative permittivity
and permeability from a composite medium consisting of two dimensional arrays of
split rings and wires [24].
The realisation of negative refraction and its experimental verification has been
questioned by some researchers. Valanju et al. [25] believed that the transmission of
energy or a signal is possible when the wave comes in a range of frequencies, which
combine to form energy packets. The phase velocity is given by the speed of each
frequency component and the group velocity is the speed of the energy packet. A wave
with just one frequency component could be bent in the wrong way, but this is
27
irrelevant because real light never has just one frequency. It was argued that if more
than one frequency were combined together to form a packet, the resulting packet
would bend and travel in the same direction. But this argument have been set away by
considering the fact that in [26], it has to be taken into consideration that the wave
packet does not extend to infinity in both directions. Even if the wave fronts of such a
finite pack will bend as suggested in [25], they still propagate in the negative
directions. This verified the existence of negative refraction phenomenon.
Since Shelbys first experimental verification of negative refraction, there has
been a lot of attention in the measurement verification of the characteristics of LHMs
for the last few years. Transmission measurements of LHMs constructed by combining
split ring resonators and wires were detailed in [27]. The advantage of this method was
that both split rings and wires were on the same substrate board. Transmission and
reflection measurements for a double negative medium consisting of SRRs and
discontinuous wires were investigated by Ozbay [28].
In free space measurements, it is difficult to determine the material parameters
accurately because of the complexity in proper calibration to avoid extraneous
contributions. The transmission performance in a T junction waveguide filled with
negative media was presented in [29]. The negative refraction verification by numerical
simulations and measurements from an EBG prism structure consisting of metallic
wires at frequencies near bandgap edges was done by Parimi et al. [30]. The negative
refraction from a prism structure consisting split ring resonators and wires was studied
experimentally and the near field refraction measurements at different distances from
the structure were reported in [31]. It will be interesting to have a close view on the
major milestones in the realisation of LHMs or their analogous obtained in the past few
years, which provided great motivation to perform the works detailed in forthcoming
sections in this thesis.
28
2.10 Realisation of LHMs
Materials with LHM performance have been realised in many ways. Various
theories have been put forward for the LHM performance from composite media. Some
of these realisations are not really an LHM but an analogous LHM medium which can
give rise to the same properties of an LHM. However, in some literature, for
convenience it was preferred to address these analogous materials as LHMs or negative
index media. The main strategies in obtaining LHMs can be categorised as follows.
a) Resonance model
b) High pass filter model
c) Negative Refraction from EBGs- the analogous LHMs

2.10.1 The Resonance Model- Split Ring Resonator (SRR) and Wire
In 1999 Smith et al. found that periodic structure containing SRRs and thin
wires can perform as a medium having negative refractive index [22, 23]. The SRR
array can perform as a medium with negative permeability () and the wire array can
perform as a medium of negative permittivity () near their corresponding resonant
frequencies. The split ring resonator (SRR) exhibits resonant magnetic response to
electromagnetic wave, with the magnetic field vector H
v
parallel to the axis of the SRR.
A periodic array of SRRs having an effective magnetic permeability given by the
following equation [22].

) (
1 ) (
2
0
2
2

i
F
eff
+
= (2.19)

where
0
is the resonant frequency given by

29

3
2
0
2
ln
3
) 2 (
b
d
t
c L
x
|
.
|
\
|
=

F is the filling fraction of the SRR and is the damping factor, 2 = 2L
x

/b
0
, where
is the resistivity of the metal and L
x
is the unit cell size along the x axis. c

is the
velocity of light and t, d and b are the parameters of the split ring (fig.2.4) and is the
operating frequency.

Fig.2.4.Unit cell of split ring contributing negative permeability (left) and wire contributing
negative permittivity (right). The wires contribute negative permittivity. SRR can be
characterised by the parameters w, d, g and b. Wire array can be characterised by radius a and
the spacing between each wires. The EM wave propagation direction is along the z direction.

From eqn. (2.19) it can be seen that real part of
eff
is negative at an interval around
the resonant frequency. If this negative permeability medium is combined with a
medium of negative real part of permittivity, the resulting structure would possess
negative refraction in a frequency band. An array of thin metallic wires, which acts as a
high pass filter for electromagnetic wave polarized with electric field parallel to wires,
can perform as a medium of negative permittivity. This can be considered as plasma
w
g
z
b 2r
y
x
t
d
(2.20)
30
medium. The plasma frequency
p
of this medium

is determined by the dimensions of
the wire lattice. It is given by [22, 23, 28]

|
.
|
\
|
=
r
a
a
c
p
ln 2
2
2
2

where a is the lattice period and r is the radius of the wire.
In this array the effective permittivity is given by [23]

2
2
1 ) (

p
eff
=

From eqn.2.22, it can be seen that the permittivity is negative when <
p
. By
combining the above two structures, a medium having negative permittivity and
negative permeability at a narrow frequency band can be realised. When both material
parameters are negative, the refractive index becomes negative. It is given by the
equation

= = | | | | n

Fig.2.5 shows the photograph of the composite medium consisting of split rings and
wires. Fig.2.6 shows the transmission through the left handed medium consisting of
SRRs and wires [32]. From the figure, it can be observed that the transmission through
an array of SRRs shows a bandgap at the region 8.5-11GHz. This is due to the negative
(2.23)
(2.22)
(2.21)
31
effective permeability at this frequency region. The transmission of wire array is very
low below the plasma frequency due to its effective negative permittivity.

Fig.2.5. A split ring and wire array having negative refractive index. The split rings
contribute negative permeability and wires contribute negative permittivity.
(Reproduced from [29]).

Fig.2.6. Measured transmission through a medium consisting of SRRs only, wires only and
composite structure consisting of SRRs and wires showing increased transmission at the
frequency band 8.4-9.3GHz where the composite material posses negative permittivity and
permeability. The plasma frequency of the wire medium in this case is 20GHz. (Reproduced
from [32)

0

-40

-80

-120

-160

-200

-240

-280

T
r
a
n
s
m
i
s
s
i
o
n

d
B

Frequency GHz
32

The plasma frequency of the wire array under study was 20 GHz. Thus the transmission
response will be very low below 20 GHz. For the structure obtained by combining
SRRs and wires, the transmission is increased at 8.4-9.3GHz band. This is due to the
fact that at this interval the material has both negative effective permeability and
permittivity. The limitation of this methodology was that it failed in probing the
effective index corresponding to the extremes of the left handed frequency band. At
these frequencies, the wavelength in the material was very large and at the extreme
frequencies it was larger than the dimensions of the sample. So ray optics will fail to
characterise the sample. In this case, a scattering cross section analysis [33] is needed
to characterise the phenomenon accurately. It is a method to characterise
inhomogeneous materials by considering the effective area of collision wave.

2.10.1.1 Inverse Snells Law verification
According to the Snells law, refractive index can be expressed as

2
1
sin
sin
= n

where
1
2
is the angle of refraction.
When both material parameters are negative, the refractive index becomes negative. It
is interesting to see the validity of Snells law in such a media. When the refractive
index n is negative, the electromagnetic wave observes a negative angle with the
normal. In this case, the parallel component of the wave vector is preserved in the
transmission but the energy flow and the wave vector directions are opposite. As a
consequence, in an LHM prism the wave observes refraction in exactly opposite
direction to that from an RHM prism.
(2.24)
33

Fig.2.7 illustrates positive and negative refraction from a prism. From the
figure, it can be seen that in the case of an LHM prism, the refracted ray is going in a
direction away from the base of the prism which is opposite to the direction of
refraction from an ordinary (positive refractive index) RHM prism. Shelby et al. [24]
performed the Inverse Snells Law measurement using the SRR and wire structure.
Fig.2.8 shows the experimental setup used in Shelbys experiment. Fig.2.9 shows
variation of negative refractive index with frequency [24]. Fig.2.10 is the measured
refracted field with a Teflon prism and an LHM prism showing positive refraction for
Teflon prism and negative refraction for LHM prism. The main disadvantages of the
split ring-wire model were large dissipation [34] and anisotropy in the structure.

Fig.2.7: A ray diagram showing refraction from a prism of RHM and LHM. For an RHM
prism, the refracted ray bends towards the base making a positive angle with the normal. In an
LHM prism, the refracted ray bends away from the base making a negative angle with the
normal.
1
2
is the angle of refraction.

RHM

LHM

2

1

2

2

34

Fig.2.8.An experimental setup used for Inverse Snells Law verification from a prism
composite medium consisting of SRRs and wires. (Reproduced from [24])

Fig.2.9. A plot of refractive index vs. frequency for SRR-Wire medium. The red curve is the
real component of theoretical LHM data and the red dotted curve is the imaginary part. The
black curve gives the measured value in which the dotted line represents the non reliable region
of measurement due to experimental limitations. (Reproduced from [24])

Microwave absorber

Detector
Sample
8 9 10 11 12

Frequency GHz

I
n
d
e
x

o
f

r
e
f
r
a
c
t
i
o
n

3

2

1

0

-1

-2

-3
35

Fig.2.10.Measured transmitted power at 10.5GHz as a function of refracted angle from an
LHM (SRRWire) prism and a Teflon prism. The Teflon prism shows an angle of refraction of
+27 (dashed curve) and for the LHM prism the angle of refraction is -61 (solid curve).
(Reproduced from [24])

2.10.2 A planar equivalent high pass C-L medium
It was illustrated that transmission line structures comprising of capacitors in
series and inductors in parallel exhibit high pass filter characteristics and they
demonstrate negative refraction as equivalent to a material having both negative
parameters [35-37]. This was based on the fact that the fundamental component of
propagation constant of this C-L network was negative (indicating negative phase
velocity) and the group velocity positive. So the structure supports backward radiation
with an effective negative refractive index. In the EMT limit, the capacitors can be
expressed as an analogous uniform dielectric and the inductors with analogous uniform
magnetic media. The unit cell of the high pass C-L filter is given in figure 2.11. Let C
and L are the capacitance and inductance per unit length respectively.
36

Fig.2.11: A unit cell of 2D loaded C-L high pass transmission line consisting of capacitor and
inductor. Each capacitor is having impedance per unit length Z
C
and each inductor has
admittance per unit length Y
L
. The unit cell length (d) is equal in both x and y directions.

The analogous expressions for permittivity and permeability can be written as

d C
d L
=
2
2
1
1

Let i be the current and v be the voltage of which the directions are denoted by suffixes
x and y. From the figure one can write down the equations for each node [35],

|
.
|
\
|
|
.
|
\
|
d C j
i
x
v
d C j
i
z
v
x
y
z
y

1
,
1

|
.
|
\
|
d L j
v
x
i
z
i
y
x z
1

(2.27)
(2.28)
(2.25)
(2.26)
37

Combining equations 2.27 and 2.28 gives

d LC
v
z
v
x
v
y
y y
= = +

1
, 0
2
2
2
2
2

where is the propagation constant. The phase and group velocities are given by

d LC v
p
= =
2

d LC v
g
+ = |
.
|
\
|
=

2
1

From equations 2.30 and 2.31 it can be seen that the phase and group velocities are in
opposite directions. Hence the structure supports backward radiation. The refractive
index is give by [35]

d LC
v
c
n
p

= =
0 0
2
1

and n is found to be negative. So this structure has a negative refractive index and
supports backward radiation. A coplanar waveguide implementation of the above
circuit has been presented at 15 GHz [36]. Fig.2.12 shows the backward radiating high
pass periodic structure studied in [36]. The capacitor and inductor were designed by
controlling the slot and conductor dimensions. Fig.2.13 gives the radiation pattern from
the structure [36] indicating very strong backward radiation at an angle of -40 degrees.
(2.29)
(2.30)

(2.32)
(2.31)
38
The radiation is 12 dB stronger than the forward radiated component. These facts are
valid only in the regime of which the Effective Medium Theory (EMT) is satisfied, that
is the spacing of the EBG is very small compared to the operating wavelength so that
the material can have an effective material parameters and refractive index.

Fig.2.12. A backward antenna based on high pass C-L circuit (above) and the enlarged view of
the unit cell shown in the inset (below). (Reproduced from [36])

The backward radiation, negative refraction and focusing effect were verified by
simulation and measurements. But the wide band negative refraction from this structure
still remains as a question. In theory, the above structure is a wide band negative
refractive planar transmission line structure. However, realising a planar high pass filter
is extremely difficult because of the deviations of transmission line approximations at
high frequencies and also because of the loss and material parameter deviation at high
frequencies.

39

Fig.2.13.Radiation pattern of a backward radiating CL structure. The backward radiation is
strong by 12dB at an angle of -40. (Reproduced from [36])

2.10.3 Composite right/ left handed (CRLH) transmission lines
Recently, Caloz et al. [38, 39] studied the fundamentals of ideal transmission
line metamaterials with their physical characteristics and the synthesis of practical
artificial transmission line metamaterials (TL-MM). He put forward the idea of
microstrip implementation of composite right/left handed (CRLH) structures. Novel
applications such as backfire to end fire leaky wave antenna, bandwidth enhanced
hybrid ring, a dual band harmonic branch line coupler and a negative
reflection/refraction phase conjugation interface were also analysed in [39] using the
CRLH approach.
The CRLH structure is a combination of right handed TL (RH-TL, a low pass
filter) and its dual the left-handed TL (LH-TL, a high pass filter) forming a band pass
TL. A diagram of the circuit is shown in fig.2.14. Let L and C be inductance and
capacitance respectively. In order to indicate its position in the high pass (left handed)
and low pass (right handed), suffixes
'
L
and
'
R
are used with L and C. Let
c

represents the propagation constant for the CRLH medium.
40
At low frequencies, the CRLH structure is dominantly LH with a hyperbolic
dispersion relation [39]

Fig.2.14. The CR-LH bandpass structure consisting of right handed transmission line (RH-TL-
high pass) and left handed transmission line (LH-TL- low pass)

' '
1
L L
L
C L

=

while at high frequencies it is dominantly RH with a linear non dispersive relation

' '
R R R
C L =

L R c
=

The transition frequency from LH to RH is given by

(2.33)
(2.34)
41

| | / 2 where , ) (
0 ) ( .
c 0 c
0
4
' ' ' '
0
'
'
c
c R R
L
L
and
C L C L

= =
= =

0
is the centre frequency of the bandpass filter. Matching (Z
0
=Z
L
=Z
R
) can be obtained
by properly adjusting the L and C values. A microstrip equivalent model of the circuit
in fig.2.14 has been studied for its radiation properties [40]. The dispersion behaviour
of the structure is shown in fig.2.15 in which there are four regions. These are LH-
guided, LH-leaky, RH-leaky and RH-guided. Based on these principles, an antenna
with backward to forward scanning capability was introduced in [40]. This antenna
does not have the drawback of conventional uniform transmission line leaky wave
antennas, which can be used only for scanning the half space from broadside to end
fire.

Fig.2.15. Propagation () and attenuation () diagrams for a dominant mode leaky wave
antenna consisting of four regions. The LH and RH guided wave regions exist when | | > k
0

(region I- LH and IV- RH), and leaky wave region exists when | | < k
0
. When < 0 the
structure supports LH backfire (region II) and when > 0 the structure supports RH endfire
(region III). (Reproduced from [40])
(2.35)
42
In fig.2.14, there are four frequency regions: region I (f <3.1 GHz) and region
IV (f > 6.3 GHz) are having | | > k
0
and represents left handed (LH) and right handed
guided (RH) wave regions respectively. The region between 3.1 GHz to 6.3 GHz
includes an LH backfire region with < 0 (region II) and an RH end fire region with
> 0 (region III). The transition from LH backfires to RH end fire occurs at = 0 at 3.9
GHz. The verification of backward radiation was detailed in [40]. A more practical
realisation of an LH-TL structure was done by ladder network as an ideal LH- TL does
not exist in nature. It can be noticed that lumped element implementation of such
structures are not wide band as it includes numerous stop bands.

2.10.4 Negative refraction from EBGs- The analogous LHMs
In the previous sections of this chapter, various realisations of LHMs have been
detailed. In this section, negative refraction phenomenon in EBGs is detailed. These
EBG based realisations can be termed as analogous LHMs, indicating that these EBGs
will show certain characteristics same as that of an LHM.

2.10.4.1 Analogy between EBG and electronic crystal
For the last few years, there has been a growing research interest on the
negative refraction phenomenon in EBGs [41-50, 53, 54]. The propagation of wave in
an EBG is different from that in conventional materials [51, 52]. Due to the periodic
nature in the structure, the wave vector k is not conserved in the EBG. Due to this
reason, it is too complicated to define a specific refractive index in weakly modulated
EBGs (Modulation in an EBG is the periodic disturbance on transmission due to
periodicity of the medium). Strong modulation takes place when the period is much
smaller than the wavelength and weak modulation occurs when the period is
comparable to the wavelength.
43
Most of the theoretical analysis of negative refraction in EBGs is based on
Equi- Frequency Surface (EFS) analysis. EFSs are contours in the wave vector space at
a specific frequency. The direction of normal to the tangent on the EFS gives the
Poynting vector. Since Poynting vector indicates the direction of energy flow, the
normal to the tangent gives the direction of energy propagation through the crystal. In
vacuum, the dispersion relation in k space is given by a sphere of radius k= 2/, where
is the wavelength in the medium. Fig.2.16 shows schematic diagram of light
propagation through a dielectric media using EFS. The smaller circle represents EFS in
vacuum and the circle with larger radius represents EFS in the dielectric medium. The
incident field direction is shown by the red arrow. When the wave enters into the
dielectric media, it undergoes refraction and hence the direction of propagation changes
as shown by the blue arrow.

Fig.2.16. EFS diagram showing wave propagation through a dielectric material. The incident
wave direction shown by red arrow and the refracted wave direction is denoted by blue arrow.
The propagation direction is determined using the normal to the EFS at the k point.

The EFS are drawn in k space i.e. in coordinates of k
x
and k
y
where k
x
and k
y
are the x
and y components of k vector. A vertical line is drawn to study the k vector
Incident wave k
x
k
y

k
y

Excited wave k
x

EFS for vacuum

EFS for dielectric
44
conservation. The normal to the tangent on the EFS at any k point gives the wave
propagation direction at that point. This schematic diagram (fig.2.16) illustrates the
Snells law in a dielectric medium.
The above case is for a homogenous dielectric medium. The EFS of an EBG is
different from that of a uniform dielectric. For an EBG in the long wavelength limit,
the medium can be considered as homogeneous. Let
av
be the average wavelength in
the medium. In this case, dispersion surfaces are spheres of radius k = 2/
av
. Fig.2.17
shows the EFS in an EBG in the long wavelength limit. When frequency increases,
av

decreases and the medium start losing its homogeneity. Then
av
becomes small and the
radius of the circles increase. They start partially overlapping each other as shown in
fig.2.18. Fig. 2.18 shows the dispersion surfaces in a crystal in the long wavelength
(when the medium can be homogenous) limit. This is the point at which the Braggs
law is satisfied. The overlapping of the spheres leads to complex dispersion surfaces
and the bandgap appears.

Fig.2.17. EFS in an EBG in the long wave length limit and the medium can be considered
homogeneous.

The circles are with radius k = 2/
av
. h
v
is a vector, the modulus of which gives the
average wave number in the crystal. Braggs diffraction law is satisfied around point L
0

and this results a bandgap.

O
H
h
v
2/
av

45

Fig.2.18. Dispersion surface in a photonic crystal at high frequency where the medium becomes
inhomogeneous. The two EFS circles overlaps and form a bandgap. Around the region L
o

Braggs law is satisfied and this region the bandgap exists.

In [48], it was shown that the effective mass in electronic crystals is analogous
to the effective refractive index of an EBG at narrow frequency band near the bandgap.
In strongly modulated EBGs, an effective refractive index can be defined near the
bandgap frequency [48]. For strong modulation in EBG, the period must be much
smaller than the wavelength. In such cases, the effective refractive index of photonic
crystal is analogous to the effective mass approximation in electron-band theory. In
solid state physics, a particle's effective mass is the mass it carries during its transport
in a crystal. It can be shown that, under most conditions, electrons and holes in a crystal
respond to electric and magnetic fields almost as if they were free particles in a
vacuum, but with a different mass. In this approximation regime, one could observe
EFS at certain frequencies near bandgap as circles and hence at this region the medium
is similar to a homogeneous dielectric. In this sense, the wave propagation inside the
crystal can be defined in the narrow frequency region. However, in the other frequency
2/
av
O H
L
0

X
h

X
o

Bandgap
Dispersion surface
46
regions, for a strongly modulated crystal, the excited waves inside consist of a mixture
of many diffracted wave components. This indicates that the EFSs can be decomposed
into circles and due to this reason the situation can be characterised as chaotic [48]. For
such crystals, the EFS shape becomes rounded as the frequency approaches the
bandgap.
Fig.2.19 shows the EFSs for a weakly modulated and strongly modulated
photonic crystal [48]. In both cases, two dimensional EBGs with hexagonal lattice
consisting of air holes in a dielectric (GaAs with dielectric constant 3.6) were
considered. From fig. 2.19, it can be seen that the EFS of the strongly modulated PC
has circular shape at frequencies near bandgap (at = 0.45). Frequencies are
normalised (a/2c) in the plot. The bandgap was at = 0.48.

Fig.2.19.Schematic diagram of wave propagation through a weakly modulated hexagonal
photonic crystal (left) and for a strongly modulated photonic crystal (right) at different
frequencies. The bandgap frequency is at = 0.48. (Reproduced from [45]).

It can be seen that the EFS becomes circular in shape at = 0.45. This suggests that at
this frequency region, very close to the bandgap there can be an effective refractive
index.
47
When the modulation is not strong, EFS generally has a star-like shape
consisting of arcs belonging to diffracted waves as shown in fig.2.19. When the
modulation becomes strong, EFS becomes circular near the bandgap. This is similar to
the EFS of a dielectric and at this region one could define an effective refractive index.
The direction of propagation inside the EBG can be determined from the tangent to the
EFS [48, 54]. For a conventional dielectric, the wave propagation direction is outward
tangent direction but in an EBG, it is inward, indicating a negative propagation
direction. At this point, the EBG is having a negative effective refractive index [48].
The bandgap phenomenon in a strongly modulated EBG is analogous to
electronic bandgap in semiconductors. Likewise, there is a promising analogy between
the effective mass state in electronic crystal and effective refractive index in an EBG
[48]. In a semiconductor, the negative effective mass state appears below the energy
gap and a positive effective mass state appears above the gap.

Fig.2.20.A diagram showing analogy between EBGs and electronic crystals. The effective mass
in electronic crystal (right) is analogous to refractive index in EBG (left).

Positive index state Positive mass state
(Electron band)

Photonic bandgap Energy bandgap

Negative index state Negative mass state
(Hole band)
E
Effective index approx. Effective mass approx.
EBG Electronic crystal
48
The negative mass state is the hole band and positive mass state is the electron band.
This is similar to the effective index states in EBGs. The effective mass state
approximation is valid near the gap whereas in EBGs, the effective refractive index
state is valid near bandgap edges. Fig. 2.20 illustrates the analogy between EBGs and
electronic crystals. It can be seen that at band gap edges, there exists a negative index
state for EBG same as negative mass state in electronic crystals.

2.10.4.2 All Angle Negative Refraction (AANR)
Inspired by the ideas of Notomi [48], Luo et al. [43-45] demonstrated all angle
negative refraction from photonic crystals without having an effective negative index of
refraction. It is observed that in the lowest photonic band near Brillouin-zone, where
positive group velocity and positive refractive index can be found and with a negative
effective mass. In this band, for all angles of incidence, a negatively refracted beam
will result.
In [44], a two dimensional array of dielectric photonic crystal consisting of air
holes in dielectric
r
= 12.0 with lattice constant a and hole radius r = 0.35a was
studied. The EFSs were obtained at different frequencies as shown in fig.2.21. Fig.2.22
shows the bandgap diagram for the photonic crystal under study in [44] and the all
angle negative refraction region is marked in red. The photonic effective mass is
j i
k k

2 which is negative at the M point. The frequency at which the EFS is convex
has got negative refraction. However, the medium will still have a positive refractive
index [44].

49

Fig.2.21.Several EFSs for the dielectric photonic crystal. The frequency is written in units of
2c/a. It can be seen that the surfaces are convex in the vicinity of M point. (Reproduced from
[44])

Fig.2.22.Dispersion diagram for the photonic crystal. The all angle negative refraction region is
highlighted in red. (Reproduced from [44])

The theoretical analysis of negative refraction in photonic crystals makes use of EFS. It
has to be mentioned that out of the many wave vectors that coexist, it is not accurate to
analyse the phenomenon by selecting one among the different possible wave vectors in

N
o
r
m
a
l
i
s
e
d

f
r
e
q
u
e
n
c
y

a
/
2
c

50
the crystal [54]. The difference between these wave vectors governs the positively and
negatively refracted power at the exit surface of the crystal. If one follows the EFS
approach for EBGs for negative refraction prediction, it is needed to incorporate case of
incident wave on the crystal. That means one has to consider the dispersion surfaces in
the two half spaces. In this case the thickness of the EBG plays a role in determining
the dispersion [54].

2.10.5 Drawbacks of LHM realisations
In the previous sections, various realisations of LHMs or analogous LHMs are
presented. It was noticed that SRR-wire medium works entirely on resonance and high
pass C-L network is based on the backward radiation characteristics of the C-L unit
cell. In the former work, the effective permeability depends on the filling fraction of the
SRR and hence periodicity plays an important role. In order to obtain a truly quasi-
continuous medium, unit element sizes must be at least one hundredth of the operating
wavelength [55]. It is very difficult to achieve in practice and therefore it is interesting
to investigate the effects of periodicity towards negative refraction. The photonic
crystal model for negative refraction has taken only infinite structures into account
which have the drawbacks which will be discussed in detail in chapter 3. This model
does not take into consideration some of the dispersion sources that exist in finite
structures. The need of finite model analysis for negative refraction studies are to be
presented in chapter 3.
In a periodic structure, the field of a Bloch wave repeats at every terminal plane
having a propagation factor e
-d
where d is the length of the unit cell and is the
propagation constant. When the loss in the structure is negligible, equals propagation
constant . By Floquets theorem, a periodic structure can give rise to spatial
harmonics (Floquets harmonics).
51
Each harmonic has a propagation phase constant
n
[56]
.

d
n
n

2
+ = (2.36)

Where n = ....-2,-1, 0,1,2,3
Depending on the sign and value of n, these waves can have positive and negative
phase velocities as shown in equations below.

n d
d
v
n
n p

2
0
,
+
= =
(2.37)

Equation (2.37) can be rewritten into

d
n
d
d
n d
c c
v
c
n
n p
0 0 0
,
2
2
|
.
|
\
|
+ =
+
= =
(2.38)

0
is a large value and when n is not a very large value, from eqn.2.38, it can be seen
that the term
pn
v
c
is independent from the value of n when the value is
d
0
high, then
the sign of
pn
v
c
only relies on the fundamental propagation constant
0
. Omission of
higher order spatial harmonics can be done when the ratio
d
0
>> 1, where
0
is the
resonance wavelength [56].
In contradiction, when
d
0
1, spatial harmonics contribution is considerable

and overall response of the structure can be considered as an aggregation of individual
52
responses from fundamental component (n = 0 ) and other spatial harmonics (here n
can be positive or negative) [72]. This leads to change in dispersion curve at some
frequency bands. Indeed, according to [35-37], one can design a high pass type
transmission line to achieve principal propagation constant
0
< 0 subject to the
condition that d/
0
is infinitesimally small. However, it is worthy mentioning that
negative dispersion can also be achieved through contribution from high-order
harmonics in any EBGs. As a common practice, higher order harmonics contribution
has been neglected and EBGs are analysed only in terms of unit cell concept. The unit
cell analysis is sufficient for studying the pass band behaviour of a periodic system but
inadequate to predict negative refraction frequencies.
As discussed before in section 2.10.1, LHM performance in a narrow frequency
band can be obtained from numerical simulation and experiments using the analysis of
a periodic structure consisting of SRRs and wires using resonance theory and EMT. In
a periodic structure consisting of SRR alone or wire alone, it can also have a negative
refraction frequency region because of the dispersion of the entire structure and the
frequency regions may differ from the prediction using unit cell approach. When the
period increases, the structure starts deviating from EMT approximations. However, the
wire medium and split ring medium are EBGs. Based on the previous discussion,
negative refraction can be obtained from these models without combining them.
Strictly speaking, in this case the structure does not posses a negative effective
refractive index but instead acts as an analogous LHM.
The approach to obtain an LHM medium based on high pass C-L structure was
based on the unit cell dispersion behaviour of the high pass periodic structure. In the
work detailed in this thesis, using numerical simulations of lumped circuit models it is
demonstrated that a high pass filter structure can demonstrate negative refraction only
at some frequencies in the pass band and hence the LHM characteristics are not wide
53
band as claimed in [35-37]. This is more pronounced when the spacing of the structure
is increased. The unit cell analysis can effectively predict the stop or pass band
behaviour but it is not accurate to predict the effects of spatial harmonics in the
dispersion diagram. In this case, we need to consider high order harmonic interactions.
In this way, one can obtain frequencies other than lower edges of band gap in an EBG
structure, at which it can give negative refraction.

2.11 Applications of LHMs
While the research of LHMs is focused on the realisation of such materials
using different approaches, a lot of effort has also been made in exploring the
applications of LHMs or their analogous. There have been many works done towards
this direction [18, 50, 57, 63, 64]. Applications can be found in the construction of
novel devices such as perfect lens [57], sub-wavelength resonating cavities and phase
shifters [63, 65], LHMs as radar absorbing materials [18, 64], LHMs in optical
receivers [50] etc. Some of these applications were experimentally verified. In the
previous section, it was shown that planar LHM media have been used as backward
radiating and scanning antennas. In the next sections, some of the interesting
applications of LHMs are introduced.

2.11.1 Perfect lens
Pendry proposed that a slab of LHM can be used as a lens which is free from all
aberrations observed in a lens made with positive refractive index [57]. In contradiction
to the conventional lens, the perfect lens formed by LHMs will be flat. Evanescent
waves in a conventional lens lead to diffraction limitation and aberrations on the quality
of an image. Due to this, very small details of the object in the image can not be
obtained. Pendry found that these decaying waves would actually amplify as they pass
54
through a left-handed material, making a perfect lens possible. Fig. 2.23 shows the ray
diagram of the focusing mechanism using a slab of LHM. The proposed lens is free
from all aberrations. So it is called a perfect lens.

Fig.2.23.An illustration of the perfect lens formation from an LHM slab. The first focus point is
inside the slab and the second focus point is outside the slab at the right side. (Reproduced from
[57])

Some researchers believed that the perfect lens seems to be in contradiction
with fundamental physics laws [58-62]. The argument by Valanju et al. [26] was that
the negative refraction phenomenon is not possible due to the reasons stated in section
2.9. However, these arguments were shown to be wrong by Pendry [26]. The second
objection was on the effect of loss of materials on the perfection of the image
obtained by the lens. The limitation on perfect lens formation due to the material loss
has been detailed [59]. For a perfect lens formation, the absorption loss should be zero.
So the resolution depends on the loss but still one can obtain a super lens (if not
perfect) that can focus features of an object with a resolution well below the
wavelength leading to sub-wavelength imaging.
55
Taking into consideration all these facts it can be concluded that the negative
refraction phenomenon and the perfect lens formation can fundamentally exist. This
application can lead to very small integrated circuits and more digitally enhanced
CDs and DVDs [6] by making use of high resolution focusing of lasers.
As discussed before, an LHM or analogous LHM can be obtained in many
ways. Focusing using LHM based on a C-L high pass filter model is demonstrated in
[35]. Wave focusing using a composite medium consisting of metallic wires and
spirals, which was similar to the LHM medium proposed by Smith et al. [22, 23] was
presented in [67]. On the other hand, focusing can be obtained from analogous LHM
using EBGs. Such an approach was presented in [42] using dielectric EBG. Fig.2.24
shows measured results for the perfect lens focusing from a PBG panel of 37.5 cm X 30
cm, at a frequency of 9.3 GHz [42]. Now the source position is displaced by a distance
of 4 cm upward and it can be noticed that the focus point also moved by the same
distance [42] as shown in Fig.2.25.

Fig.2.24.Experimental result showing focusing of electric field from an EBG due to negative
refraction [46]. The yellow point at the left side is the source point and yellow point at right
side is the image point (Reproduced from [42])
Image
Source
56

Fig.2.25.Experimental result showing focusing of electric field from an EBG due to
negative refraction when the source is moved 4cm upward. It can be seen that the
image at the left position also moved by the same distance (Reproduced from [42])

2.11.2 Scattering reduction using LHMs
Negative refractive index materials can be used for the reduction of
electromagnetic wave scattering [18, 64, 72]. In [64], the scattering from a metallic
cylinder coated with RHM / LHM with same material parameters in magnitude but
opposite in signs was presented. Fig.2.26 illustrates the effective scattering surface
patterns of a cylinder with different coatings (the angle = 180 degrees corresponds to
the backscattering direction). From the figure, it can be seen that the use of a material
with a negative refractive index -n is much more advantageous than that of an RHM
with refractive index +n for reducing the back scattering from the metallic cylinder.
Image Source
57

Fig.2.26.Scattering pattern for a circular cylinder with coating. Curve l corresponds to
scattering when surface of the cylinder coated with parameters = -1-i*0.6 and = -1-
i*1, curve 2 represents the formal sign reversal for the real parts = 1-i*0.6 and = 1-
i*1. Curve 3 is the scattering when the material parameters are that of absorbing foam
glass ( = 1 and = 1- i*0.6). Curve 4 shows scattering from an uncoated cylinder

In [18], the reflection from a metallic surface with coatings of RHM and LHM
was presented and it was argued that LHM is superior from an application point of
view due to its scalability and reduced weight. The limitation of this comparison was
that in this case the comparison between RHM and LHM performance was done for
materials with parameters of different magnitudes and opposite signs. This was
comparing the performance of two materials that have refractive indices +n
r
and n
1
,
where n
r
and n
1
are different values. The subscript r represents RHM and l represents
LHM. Based on the comparison, it was shown that LHM was superior for absorber
58
applications. Recently a theoretical analysis based on Mie scattering was presented in
[72] which indicated that when a metal is coated with an LHM the scattering
coefficient will be very small.
2.11.3 Microwave components using LHMs
LHM materials can be employed as sub-wavelength resonators and zero phase
delay lines. The advantage over that of RHM materials is that the dimension of the
resonator will be very small [63, 65]. If there are two layers of materials for the phase
compensator, with thickness d
1
and d
2
, in the case where one layer is with LHM then
the resonator size will be proportional to d
1
/d
2
on the other hand the if they both are of
RHM, then the resonator size will be proportional to d
1
+d
2
. The schematic diagram of
such a two layer structure is shown in figure 2.27.

Fig.2.27. A two layer structure consisting of positive and negative refractive index
materials. The left layer is a dielectric with > 0, >0 and the right layer is a lossless
material with negative permittivity and permeability ( < 0, < 0). In the left layer the
Poynting vector (S) and wave vector (k) are parallel. In the second layer they are anti
parallel.

1,
1 > 0

2,
2 < 0

d
2

d
1
k
0
k
1
k
2 k0
S
0
S
1
S
2
S
2

I II
59
Consider a uniform plane wave incident on layer I and if the impedance of layer I is
same as that of air (i.e.
0 =
1
)

there will not be any reflection. When the wave reaches
the end of slab I, there will be a phase change by an amount
1 0 1 1
d k n = where
0 0 0
= k . An LHM with refractive index n
2
< 0 was placed next to slab I (i.e. slab
II in the fig. 2.27). In the second slab, the directions of Poynting vector and phase
velocity are anti-parallel. Hence the wave vector k
1
and k
2
in slabs I and slab II are in
opposite directions and the phase change occurred from the beginning and end of slab
II is given by
2 0 2 2
d k n = . This shows that the phase change that occurred during
the travelling of wave through slab I, can be decreased or can be cancelled by
traversing the slab II. When the thickness of the slabs are taken in such a way that they
satisfy the following equation [63] then the total phase difference becomes zero.

1
2
2
1
n
n
d
d
=

In this way, the structure can be used for phase compensation / conjugate applications.
The phase change can be controlled by the ratio of d
1
and d
2
. In principle, the
thicknesses can be any value as long as the ratio d
1
/d
2
satisfies the above relation.
The same structure can be converted into a one 1D sub-wavelength cavity
resonator by employing conductors at both ends (as shown in fig.2.28, conductor layer
is shown in red). From fig.2.28, it can be seen that the first conductor is at z = 0 and
the other is at z = d
1
+d
2
. For a 1D cavity resonator all quantities are independent of x
and y coordinates. In the slab I (i.e. in region the electric and magnetic fields can be
written as follows.

(2.39)
60
) cos(
) sin(
0 1 01
1
0 1
1
0 1 01 1
z k n E
i
k n
H
z k n E E
y
x
=
=

Fig.2.28. The two layer resonating structure consisting of RHM (green) and LHM (yellow)
slabs backed by conductors at both ends (shown in red).

In slab II the fields can be expressed as

)] ( cos[
)] ( sin[
2 1 0 2 02
2
0 2
2
2 1 0 2 02 2
z d d k n E
i
k n
H
z d d k n E E
y
x
+ =
+ =

where subscripts1 and 2 represent regions I and II respectively. To satisfy the boundary
conditions at the interface between regions I and II

I

1,
1 > 0

II

2,
2 < 0

d
2

d
1
x
z

(2.42)

(2.43)
(2.40)

(2.41)
61
1 1
1 1
| |
| |
2 1
2 1
d z y d z y
d z x d z x
H H
E E
= =
= =
=
=

This leads to

0 ) sin( ) sin(
2 0 2 02 1 0 1 01
= d k n E d k n E

0 ) cos( ) cos(
2 0 2 02
2
0 2
1 0 1 01
1
0 1
= + d k n E
k n
d k n E
k n

In order to have E
01
0 and E
02
0, the determinant in of the above equations must
vanish. i.e.

0 ) cos( ) sin(
) cos( ) sin(
1 0 1 2 0 2
1
1
2 0 2 1 0 1
2
2
=
+
d k n d k n
n
d k n d k n
n

This can be written as

0 ) tan( ) tan(
2 0 2
1
1
1 0 1
2
2
= + d k n
n
d k n
n

The permeabilities
1
and
2
can be written as
1
= |
1
| and
2
= -|
2
| and substituting
into eqn. (2.49)
(2.44)

(2.45)
(2.48)

(2.49)
(2.46)
(2.47)
62

0 ) tan(
| |
) tan(
2 0 2
1
1
1 0 1
2
2
= + d k n
n
d k n
n

This indicates that at a frequency , if
1
> 0,

1
>0 and
2
< 0,

2
<0, a non trivial
solution for a one dimensional cavity can be obtained as

| |
| |
) tan(
) tan(
1 2
2 1
2 0 2
1 0 1
n
n
d k n
d k n
=

Hence in principle d
1
and d
2
can be as thin or as thick as long as the above ratio is
satisfied. This proves the possibility of sub-wavelength resonators using LHMs. Some
applications such as backward wave couplers with high coupling coefficient, broad
band antenna feed network etc. [65] have also been reported.

2.11.5 Optical receivers using EBG based analogous LHM
The spreading of the illumination area in the focal plane of a lens can be
suppressed by using negative refraction behaviour of photonic crystals. In [50], a field
of view expander (FOVE) technique based on this approach for optical receivers was
proposed. In optical detectors, the light acceptance area is limited so that a reduction of
focal spots illumination area at reception of aberrated beam will take place. The
propagation angles of the received beam in front of optical devices need to be taken
into account. The proposed method [50] can significantly reduce the focal spot size
without expanding the propagation angles.
When a point source A is located on a slab of LHM, light entering will focus at
point B inside the crystal and then again gathers at point C outside the crystal as shown
in fig.2.29. The line connecting the points A, B and C are perpendicular to the

(2.50)
(2.51)
63
boundaries and each ray in front of the slab is parallel to the corresponding ray behind
the slab. If the boundaries are slightly bend, then the location of the point B where the
rays gather will be shifted. So boundaries properly designed by connecting many short
bent lines may contribute to a reduction of beam spreading due to wave front
aberrations after the light passed through the crystal. If the radius of curvature is large
enough, the propagation angles of the rays in front of and behind the crystal will not
change very much. Making use of the aforementioned approach a FOVE technique for
optical receivers can be developed. The following figures 2.29-2.31 illustrate this
concept clearly. In this case, illumination range behind the crystal is reduced to 31% of
that in front of the crystal [50].

Fig.2.29. Focusing from a crystal at negative refraction frequency. All the rays being
emitted from point A pass through point B and gather again at point C. (Reproduced
from [50])

EBG
Y
X
A
B
C
64

Fig.2.30. Schematic drawing of a field of view expander. The boundaries have the
curvatures r
1
and r
2
with the same centre of origin. The inset shows a magnified part of
the crystal surface (Reproduced from [50])

Fig.2.31. Schematic drawing showing the reduction of illumination area obtained by applying a
photonic crystal with negative refraction characteristics. Reduction in illumination area is
marked by W. (Reproduced from [50])

65
2.12 Conclusions
In this chapter, a background survey of negative refraction phenomenon has
been presented. The characteristic properties of LHMs such as backward radiation,
inverse Snells Law and inverse Doppler Effect have been identified. Recent
realisations of LHMs using resonance model, high pass C-L model and analogous
LHMs using EBGs, numerical simulations of the performance of LHMs, their
analogous media and applications have been reviewed. It has been observed that LHMs
are capable of potential electromagnetic applications such as in the construction of
perfect lens, beam scanning antenna, microwave and optical circuit components. The
resonance models are very narrow band and they work on EMT approximations. The
limitation of EMT is that in order to get a pure homogeneous medium, the period
should be very small compared to wavelength. This makes the design difficult at both
microwave and optical frequencies. Also, the resonance models have the disadvantages
of strong dissipation and anisotropy in the metamaterial. However, the anisotropy in
such LHMs has not been addressed in-depth. It has been noted that negative refraction
can be obtained from EBGs. The negative refraction from EBGs does not guarantee a
negative index of refraction and this phenomenon arises due to the dispersion nature of
EBGs. All aforementioned studies were proposed for infinite models, which have not
taken into consideration the higher order Floquets harmonics interactions and the
incident field coupling. These factors can affect the dispersion of the model especially
when the period is getting comparable to the wavelength. This has indicated the need of
considering Floquets harmonics interactions and finiteness of EBGs for negative
refraction studies.

66
References
[1] A. Von Hippel, Arthur R, Dielectrics and waves, MIT Press, 1954
[2] Constantine A. Balanis, Advanced Engineering Electromagnetics, pp.44-60, 1989
[3] http://en.wikipedia.org
[4] http://www.darpa.mil/dso/thrust/matdev/metamaterials/metamaterials.html
[5] http://www.ee.upenn.edu/~engheta/Metamaterials.htm
[6] Hossein Mosallaei, Kamal Sarabandi Magneto-dielectrics in electromagnetics:
concept and applications, IEEE Transactions on antennas and propagation, Vol.
52, pp. 1558-1567, 2004
[7] John Pendry, New electromagnetic materials emphasize the negative, Physics
World, pp.1-5, 2001
[8] Gennady Shvets, Photonic approach to making a material with a negative index
of refraction, Physical Review B, 67, pp. 035109-1-8, 2003
[9] M. C. K. Wiltshire, Bending of light in the wrong way, Science 292: pp. 60-61,
2001
[10] Edward C. Jordan, Keith G. Balman, Electromagnetic Waves and Radiating
Systems, Prentice Hall Inc., 1995
[11] Simon Ramo, John R. Whinnery, Theodore van Duzer, Fields and Waves in
Communication Electronics, John Wiley and Sons, 1984
[12] H. Lamb, On group-velocity", Proc. London Math. Soc. 1, pp. 473-479 ,1904
[13] www.wave-scattering.com/negative.html
[14] A. Schuster, An Introduction to the Theory of Optics, Edward Arnold, London,
1904, pp. 313-318
[15] Pocklington, Growth of a wave-group when the group velocity is negative",
Nature 71, pp. 607-608 ,1905
[16] V.G Vaselago, The electrodynamics of substances with simultaneously negative
67
values of and , Soviet Physics USPEKHI, Volume 10, Number 4, pp. 509-514,
January-February 1968.
[17] James Clark Maxwell, Thomas Forsyth Torrance A dynamical theory of the
electromagnetic field, Scottish Academic Press, 1982.
[18] Joe Pacheco Jr, Theory and Applications of Left-Handed Metamaterials, PhD
dissertation, Massachusetts Institute of Technology, February, 2004.
[19] I. V. Lindell, S.A. Tretyakov, K.I. Nikoskinen and S. Ilvonen, BW-media-
media with negative parameters, capable of supporting backward waves,
Microwave and Optical technology Letters, Vol. 31, pp.129-133,2001
[20] Siddiqui O. F, Mojahedi. M, Eleftheriades G.V, Periodically loaded
transmission line with effective negative refractive index and negative group
velocity, IEEE Trans. Antennas and Propagation, Vol. 51 , pp. 2619 2625,
2003
[21] Pendry J. B, Holden, A. J, Robbins, D. J. Stewart W. J. Magnetism from
conductors and enhanced nonlinear phenomena, IEEE Trans. Microwave Theory
and Techniques, Vol. 47, pp. 2075 2084, 1999
[22] D. R. Smith, Willie J. Padilla, D. C. Vier, S. C. Nemat-Nasser, and S. Schultz,
Composite Medium with Simultaneously Negative Permeability and
Permittivity, Physical Review Letters 84, pp. 41844187, 2000
[23] D. R Smith and N. Kroll, Negative Refractive Index in Left Handed Materials,
Physical Review Letters, Volume 85, pp. 2933-2936, 2000
[24] R. A .Shelby, D. R. Smith, S. Schultz, Experimental Verification of a Negative
Index of Refraction, Science, Vol. 292, pp.77-79, 2001
[25] P.M Valanju, R.M Walser, A.P. Valanju, Wave Refraction in Negative- Index
Media: Always Positive and very Homogeneous, Physical Review Letters, Vol.
88, pp.187401-187401-4, 2002
68
[26] J. B Pendry, Comment on Left Handed Materials Do Not Make a Perfect
Lens, Phys .Rev. Lett., Vol.91, pp. 099701-1, 2003
[27] Mehmet Bayindar, K.Aydin, E. Ozbay, P. Marcos and C.M Soukoulis,
Transmission Properties of composite metamaterials in free space, Applied
Physics Letters, Vol. 81, pp.120-123, 2002
[28] Ekmel Ozbay, Koral Aydin, Ertugrul Cubukcu, Mehmet Bayindir,
Transmission and Reflection Properties of Composite Double Negative Meta
materials, in free space, IEEE Trans. Antennas and Propagation, Vol.51,
pp.2592-2595, 2003
[29] Hongsheng Chen, Lixin Ran, Jiangtao Huangfu, Xianmin Zhang, Kangsheng
Chen, Tomasz M. Grzegorczyk, Jin Au Kong, T-Junction Waveguide
experiment to characterize left-handed properties of metamaterials, Journal of
Applied Physics, Vol. 94, pp.3712-3716, 2003
[30] P.V Parimi, W. T. Lu, P. Vodo, J. Sokoloff, S. Sridhar, Negative Refraction
and Left- handed electromagnetism in Microwave Photonic Crystals, cond-
mat/0306109, 2003 (Online archive)
[31] C. G. Parazzoli, R. B. Greegor, K. Li, B. E. C. Koltenbah and M. Tanielian,
Experimental Verification and Simulation of Negative Index of Refraction Using
Snells Law, Physical Review Letters, Vol.90, No.10, pp. 107401-1-107401-4,
2003
[32] P. Marcos and P.M Soukolis, Left Handed Materials, arXiv: cond-
mat/0212136 v1 6, 2002 (Online archive)
[33] http://hyperphysics.phy-astr.gsu.edu/hbase/nuclear/crosec.html
[34] Andrew A. Houck, Jeffery B. Brock and Isaac L. Chuang, Experimental
observations of a left-handed material that obeys Snell's Law, Physical Review
Letters , Vol. 90, pp.137401, 2003
69
[35] Ashwin K. Iyer, George V. Eleftheriades, Negative Refractive Index
Metamaterials supporting 2-D Waves, Proc. of the IEEE Inter National
Symposium on Microwave Theory and Techniques, Vol.2, pp. 1067-1070, 2002
[36] Anthony Grbic, George V. Eleftheriades, A Backward Wave Antenna Based
on Negative Refractive Index L-C Networks, Proc. of the IEEE Intl. Symposium
on Antennas and Propagation, Vol. IV, pp. 340-343, San Antonio, TX, pp.16-21,
2002
[37] Anthony Grbic, George V. Eleftheriades, Experimental verification of
backward wave radiation from a negative refractive index metamaterial, Journal
of Applied Physics, Vol. 92, pp. 5930-5936, 2002
[38] Christopher Caloz, Tatsuo Itoh, Novel microwave devices and structures
based on the transmission line approach of metamaterials, Proc. of IEE MTT
Symposium, pp.195-198, 2003
[39] C. Caloz, A. Sanada, T. Itoh, Microwave applications of transmission-line
based neagtive refractive index structures, Proc. Asia Pacific Microwave
Conference (APMC) , Seoul, pp. 1708-1713, 2003
[40] Lei Liu, C. Caloz, T. Itoh, Dominant mode leaky wave antenna with backfire -
to-endfire scanning capability, Electronic Letters, Vol. 38, No. 23, pp.1414-
1416, 2002
[41] Kosaka, H, T. Kawashima, A. Tomita, M. Notomi, T. Tamamura, T. Sato
and S. Kawakami, Superprism phenomenon in photonic crystals, Physical
Review B, Vol. 58, pp. 10096-10099, 1998.
by flat lens using negative refraction, Nature, Vol. 426, pp. 404, 2003
[43] C. Luo, S. G. Johnson, and J. D. Joannopoulos, All-angle negative refraction
in a three dimensionally periodic photonic crystal, Applied Physics Letters, Vol.
70
81, pp. 2352-2354 , 2002
[44] Chiyan Luo, Steven G. Johnson and J. D Joannopolous, J. B Pendry, All
angle negative refraction without negative effective index, Physical Review B,
Vol. 65, pp.201104, 2002
[45] C. Luo, M. Ibanescu, S. G. Johnson, and J. D. Joannopoulos, Cerenkov
radiation in photonic crystals, Science, Vol. 299, pp. 368-371, 2003.
superlens, Physical Review Letters, Vol.91, No. 20, pp.207401-1-207401-3,
2002
[47] Ertugrul Cubukcu, Koray Aydin, Ekmel Ozbay, Stavroula Foteinnopolou,
Costas M.Soukoulis, Negative refraction by phtonic crystals, Nature, Vol. 423,
pp.604-605, 2003
[48] Masaya Notomi, Negative refraction in photonic crystals, Optical and
Quantum Electronics, Vol. 34, pp.133143, 2002
[49] S. Foteinopoulou, E. N. Economou, and C. M. Soukoulis , Refraction in media
with a negative refractive index, Physical Review Letters, Vol. 90, No.10
107402-1.107402-4, 2003
[50] Yoshihisa Takayama, Werner Klaus, Application of negative refractive
behaviour to field-of-view expansion in optical receivers, Journal of Optics. A:
Pure Applied Optics, Vol. 5 , pp. S305S309, 2003
[51] Jonathan P. Dowling and Charles .M. Bowden, Anomalous index of refraction
in photonic bandgap materials, Journal of Modern Optics, Vol. 41, pp. 345,
1994.
[52] Ho, K.M., C.T. Chan and C.M. Soukoulis, Existence of photonic bandgap in
periodic dielectric structures, Physical Review Letters Vol.65, No.25, pp. 3152-
71
3155, 1990
[53] Chao-Hsien Kuo and Zhen Ye, Band gap or negative refraction effects?,
cond-mat/0310423 v1, 2003 (Online archive)
[54] Vito Mocella, Negative refraction in photonic crystals: thickness dependence
and Pendellosung phenomenon, Optics Express, Vol. 13, pp. 1361-1367, 2005
[55] R. Marqus, F. Medina, An introductory overview on left-handed
metamaterials, Proc. the 27
th
ESA Antenna Workshop on Innovative Periodic
Antennas, Spain, pp. 35-41, 2004
1986.
[57] John Pendry, Negative refraction makes perfect lens, Physical Review Letters
Vol. 85, pp. 39663969, 2000
[58] G. V. ' Hooft , Comment on "Negative refraction makes a perfect lens", Phys.
Rev.Lett., Vol.87, No.24, 249701-1, 2001
[59] N. Garcia, M. Nieto- Vesperinas, Left-handed materials do not make a perfect
lens, Physical Review Letters, Vol. 88, pp.207403-1- 207003-4, 2002
[60] John Michele Williams, Some problems with negative refraction, Physical
Review Letters Vol. 87, 249703-1, 2001
[61] N. Garcia, M. Nieto-Vesperinas, Is there an experimental verification of a
negative index of refraction yet?, Optics Letters, Vol. 27, pp. 885-887, 2002
[62] Manuel Nieto-Vesperinas, Problem of image super resolution with a negative-
refractive-index slab, Journal of Optical Society of America. A, Vol. 21, pp. 491,
2004
[63] N. Engheta, "An idea for thin, sub-wavelength cavity resonators using
metamaterials with negative permittivity and permeability", IEEE Antennas and
Wireless Propagation Letters, Vol. 1, pp. 10-13, 2002
72
[64] A. N Lagarkov, V. N Kisel, Electrodynamic properties of simple bodies made
of materials with negative permeability and negative permittivity, Doklady
Physics, Vol. 46, pp. 163-165, 2001
[65] G. V. Eleftheriades, M. A. Antoniades, A. Grbic, A. Iyer, R. Islam,
Electromagnetic Applications of negative refractive index transmission line
metamaterials, Proc. of ESA Workshop on Innovative Periodic Antennas, Spain,
pp.19-25, 2004
[66] A. N. Lagarkov and V. N. Kissel, Near-perfect imaging in focusing system
based on a left-handed-material plate, Physical Review Letters, Vol. 92, pp.
077401-1-4, 2004
[67] Y. Hao, L. Lu and C. G. Parini, Dispersive FDTD modelling on multi-layer
left-handed metamaterials for near/far field imaging at microwave frequencies,
Proc. of IEEE Antennas and Propagation Conference, Vol.4, pp. 639-642, 2003
[68] R. B. Greegor, C. G. Parazzoli, K. Li, B. E .C. Koltenbah and M. Tanielian,
Experimental determination and numerical simulation of the properties of
negative index of refraction materials, Optics Express, Vol. 11, pp. 688, 2003
[69] C. D. Moss, T. M. Grzegorczyk, Y. Zhang, J. A. Kong, Numerical studies of
left handed metamaterials, Proc. Progress in Electromagnetic Research, PIER,
35, pp. 315-334, 2002
[70] Pavel Kolinko and David R. Smith, Numerical study of electromagnetic waves
interacting with negative index materials, Optics Express, Vol.11, No.7, pp. 640-
648 , 2004
[71] R. E Collin, Antenna Theory Part 2, Chapter 19, McGraw-Hill, 1969
[72] Andrea Alu, Nader Engheta, Achieving transparency with plasmonic and
metamaterial coatings, Physical Review E, Vol.72,pp.016623,2005

73
Chapter 3

Periodic Structures: Modelling and Negative
Refraction Phenomenon

3.1 Introduction
Collin [1] defined periodic structures as waveguides and transmission lines
loaded at periodic intervals with identical obstacles. Periodic structures have many
applications at microwave and optical frequencies due to their inherent properties of
forming pass/stop bands and supporting propagation of waves with phase velocities
much less than velocity of light (slow wave). The first characteristic ensures their role
in frequency selective surfaces (FSS) or filters and EBGs. The second property leads to
the realisation of slow wave structures or travelling wave tubes (TWTs).
At present, periodic structures are realised in various forms in one, two and
three dimensions using various materials including dielectrics or metals. Therefore it
opens up more innovations such as photonic/electromagnetic bandgap (PBG/EBG)
structures and novel applications in microwave and antenna engineering. For example,
radiation pattern and gain enhancement of microstrip antennas can be achieved by
using EBGs as ground planes [2, 3] or above the radiating patch [4] to suppress
radiation in certain directions. There have been many approaches for the analysis of
EBGs both analytically and numerically [4-7]. Analytical models can be obtained only
for a limited class of electromagnetic periodic structures with certain approximations.
So majority of the analysis were based on numerical modelling.
Chapter 3: Periodic Structures: Modelling and Negative Refraction Phenomenon

74
Recently, it has been observed that negative refraction can be obtained from
EBGs at frequencies at the edges of bandgap [8-15]. However, all such numerical
analysis considered infinite photonic crystals which extended in the entire space and
there was no incident field excitation used in this approach. In this case, the properties
of the crystal were determined from wave vectors inside the photonic crystal [16]. The
analysis with unit cell alone will only take into consideration fundamental spatial /
Floquets harmonic. An analytical study on the excitation of a crystal consisting of
periodic dipole scatterers was performed in [7]. In comparison with the initial approach
[17] which considered only fundamental Floquets harmonic interaction, in [7] the
interaction between higher order Floquets harmonics was presented. In fact, the
Floquets harmonics interaction, the coupling between the finite structure and incident
field and evanescent modes which exist in finite EBGs can contribute towards the
dispersion behaviour. This dispersion can be classified as spatial dispersion in which
the effective parameters change with spatial coordinates inside the structure and also
with the finiteness of the structure.
In this chapter, the methodology based on the unit cell approach, dispersion
characteristics of infinite and finite periodic medium with special emphasis to negative
refraction phenomenon and Floquets harmonics effects and spatial dispersion in finite
EBGs are detailed. It is observed that negative refraction phenomenon exists at narrow
frequency bands due to the spatial dispersion in the structure. This dispersion arises
due to the Floquets harmonics interactions. Lumped element circuit simulations are
carried out for the initial study. The role of Floquets harmonics and backward
radiation from periodic structures will be presented.


75
3.2 Analysis of periodic structures
Periodic structures can be analysed by either lumped circuit/ network analysis
or wave analysis method. The selection of analysis method depends on nature of the
problem. For simple periodic structures, circuit approximations can be obtained without
much complexity. Circuit network analysis is preferred in this case. For complex
structures, wave analysis is much easier.

3.2.1 Circuit or network analysis
This method of analysis starts with constructing an equivalent network for a
single basic section or unit cell of the structure. It is followed by a process to determine
the voltage and current in the structure connecting an infinite number of basic networks
[1] by applying periodic boundary conditions. This method is complicated in periodic
structures where it is difficult to get an equivalent analytical model for the structure and
also the approximations to obtain an equivalent model limit the applicability. The
approximations mainly depend on the size of the unit cell and the frequency of
operation.

3.2.2 Wave analysis
Periodic structures can be analysed in terms of forward and backward
propagating waves which can exist in each unit cell of the structure. This approach is
based on wave amplitude transmission matrix. This transmission matrix relates the
incident and reflected wave amplitudes on the input side of the junction to that on the
output side of the junction. This matrix can hold the same relationship to the scattering
matrix as the voltage-current transmission matrix related with the impedance matrix.
Fig. 3.1 shows the diagram of a periodic structure consisting of identical elements.

76

Fig.3.1. Diagram showing wave amplitudes in periodic structure. c
n
+
, c
+
n+1
are forward
propagating waves at the n
th
and (n+1)
th
terminals respectively and c
n
-
, c
-
n+1
are the
corresponding backward waves.

Let the amplitudes of forward and backward propagating waves at the n
th
and (n+1)
th

terminal plane be c
n
+
,c
n
-
,c
n+1
+
,c
n+1
-
. They are related by the wave amplitude
transmission matrix by the relation [1]

(
=
(
+
+
+
+
1
1
22 12
12 11
n
n
n
n
c
c
A A
A A
c
c
(3.1)

where A is wave amplitude matrix.
In eqn.3.1 amplitude matrix A describes the circuit. In terms of scattering matrix S for a
single junction we have [1]

(
(
=
(
(
=
(
(
+
+
+
2
1
22 12
12 11
2
1
2
1
V
V
S S
S S
c
c
V
V

This gives [1]
(
(
=
(
(
2
1
22 12
12 11
2
1
c
c
S S
S S
V
V
(3.2)

On solving these equations for c
1
+
and c
1
-
give [1]

n

n+1

n+1
Unit
cell
c
n
+
c
+
n+1

c
n
-
c
-
n+1


77

(
(
(
(
(
(
=
(
(
+
2
2
12
22 11
2
12
12
11
12
22
12
1
1
1
) (
c
c
S
S S S
S
S
S
S
S
c
c

For a Bloch wave solution it requires that

+ +
+
=
n
d
n
c e c

1

+
=
n
d
n
c e c

1

Hence eqn. (3.1) becomes in the general form for n
th
and (n+1 )
th
element

0
1
1
22 21
12 11
=
(
(
(
(
+
+
+
n
n
d
d
c
c
e A A
A e A

A nontrivial solution for c
+
n+1
and c
-
n+1
can be obtained only when the determinant
becomes zero. i.e

0 ) (
22 11
2
21 12 22 11
= + + A A e e A A A A
d d

or
2
cos
22 11
A A
d
+
=

For a two port network, a conversion from scattering matrix S to amplitude matrix A
can be obtained using the following equations.
(3.3)
(3.5)
(3.4)

78

12
22 11
2
12
22
12
11
21
12
22
12
12
11
1
S
S S S
A
S
S
A
S
S
A
S
A
=
=
=
=

Using equations 3.5 and 3.6-3.9 the dispersion characteristics which relate propagation
constant and frequency can be obtained.

3.3 Floquets theorem and spatial harmonics
In an infinite periodic structure, the field of a Bloch wave repeats at every
terminal plane having a propagation factor e
-d
, where d is the length of the unit cell
and is the propagation constant when the loss is negligible [1]. Thus, if the field in the
unit cell between 0 < z < d is ) , , ( ), , , ( z y x H z y x E .The field in the unit cell located in
the region d< z < 2d is

) , , ( ), , , ( d z y x H e d z y x E e
d d

)

Then the field in the periodic structure is described by a solution of the form

) , , ( ) , , ( z y x E e z y x E
p
d
=

) , , ( ) , , ( z y x H e z y x H
p
d
=
(3.6)

(3.7)

(3.8)

(3.9)
(3.10)
(3.11)

79

Where E
p
and H
p
are periodic functions of z with period d.

) , , ( ) , , ( z y x E nd z y x E
p p
= +

This is called Floquets theorem. The field at z
1
+d is related to the field at z
1
as

) , , ( ) , , (
1
) (
1
1
d z y x E e d z y x E
p
d z
p
+ = +
+

) , , ( ) , , (
1
) (
1
1
z y x E e d z y x E
p
d z +
= +

which gives
) , , ( ) , , (
1 1
z y x E e d z y x E
d
= +
(3.12)

This is the property of Bloch wave.
Using the property of Fourier series, any periodic function can be expanded into an
infinite number of Fourier series. Thus

d nz j
pn p
(x, y) e E (x, y, z) E
/ 2
=

where E
pn
are wave functions of x and y.

dz e (x, y, z) E
d
(x, y) E
mz/d -j
p
d
pn
2
0
1
=

Exponential functions from a complete orthogonal set, i.e.

80

n m
n m 0
0
2 2
d
dz e e (x, y, z) E
d
mz/d j nz/d -j
p

The field in a periodic structure can now be represented as

nz/d nz-j -j
pn
- n
(x,y) E E(x, y, z)
2 2
=
=

z/d -j
pn
- n
n
( x, y) e E E(x,y,z)
=
=

where = j and

d
n
n

2
0
=

Each term in the expansion is called spatial harmonic and has a propagation phase
constant
n
. The phase velocity of the n
th
harmonic is given by

nd
v
n
pn

2
0

= =

The group velocity of the n
th
harmonic is

d
d
v
gn
=

(3.13)
(3.14)

81
and is the same for all harmonics. From the relations, we can see that some of the
spatial harmonics have phase and group velocities in opposite directions. This leads to
backward radiations from the structure.

3.4 - Diagram (Brillouin diagram)
The frequency bands of operation of a periodic structure can easily be
determined by plotting the curves of versus k
0
where is the propagation constant
and k
0
is given by

=
0
k

It will give the frequency bands for propagation and attenuation for the structure. Such
k
0
d- d diagram is also called Brillouin diagram. It can also be represented in terms of
and . So they are also termed as - diagrams. The phase velocity and group
velocity at a frequency point can be determined from the Brillouin Diagram. The /
value gives the phase velocity and the slope of the curve d /d gives group velocity.
Brillouin diagram is a good tool for determining frequency regions where backward
radiation or negative refraction can be obtained from an electromagnetic periodic
structure. Brillouin diagrams are also known as dispersion diagrams [18-20]. This also
can be plotted in k
o
d- d axes. Figure 3.2 shows the dispersion diagram of a C-L
periodic structure. This diagram shows positive phase velocity and negative group
velocity. Since they are in opposite directions it will lead to backward radiations.

(3.15)

82

Fig.3.2. - diagram of a C-L circuit showing phase velocity and group velocity at a point x on
the curve. The / at the point x gives phase velocity at that point and the slope of the curve at
the point x gives group velocity.

The dispersion diagram can be obtained from the unit cell analysis of the
periodic structure. This represents the variation of fundamental spatial harmonic with
frequency. As discussed before, in a periodic structure a number of spatial harmonics
can exist and the interaction can change the behaviour of the structure at some
frequency points. Depending on the interaction among harmonics, different composite
dispersion characteristics will be obtained [21]. For example, in a band gap structure,
the fundamental propagation constant is not adequate to represent the dispersion nature,
especially at band gap edges. Single dominant mode (n = 0 spatial harmonic) is
sufficient to describe the field adequately for a transmission problem. However, near
the bandgap edges a single spatial harmonic does not satisfy the prescribed boundary
conditions on the structure and does not constitute the normal mode by itself. In this
case the field can be represented by the aggregate of all spatial harmonics [21]. For a
minimum accuracy, at least two higher order harmonics should be considered [21]. In
this case, the dispersion is determined by the aggregate of all the harmonics.
Depending on the periodic structure, there can be a coupling between the spatial
harmonic modes and this can change the - diagram at some frequencies. This is

x
Slope= v
g
< 0

83
called mode coupling [21]. In this study, it was observed that phase velocity and group
velocity can be different from the frequencies predicted by unit cell analysis at some
frequency bands. This gives the indication that negative refraction phenomenon can be
found in frequency points where phase velocity is negative and group velocity is
positive at these frequencies there may not be any backward radiation or negative
refraction when analysed with unit cell alone.
In the past, defects in EBGs were analysed by considering a finite number of
cells which contain the defect, as a repeating element and this method is called super
cell analysis [22]. In this research, a similar analysis is carried out with special
emphasis to negative refraction phenomenon from EBGs. When the spacing is very
small compared to the operating wavelength the spatial harmonics effects can be
negligible. This is not the case when the spacing is comparable to the operating
wavelength; the case where most of the Electromagnetic Band Gap structures (EBGs)
are realised. So in practical structures, the spatial harmonics effects are important in
determining the dispersion characteristics. It is interesting to see these effects and their
contributions towards negative refraction so that one can make use of negative
refraction phenomenon in the passband of EBGs which can lead to novel applications
in antenna and electromagnetic engineering.

3.5 Unit cell and Super cell analysis
This section will show how the dispersion behaviour of a periodic structure
changes from that of a unit cell in practically realisable structures. When periodic
structures are analysed using unit cell, it will not give spatial harmonics effects but only
the fundamental harmonic. Even though unit cell analysis can predict transmission
characteristics of the structure, it is not accurate to predict negative refraction
phenomenon completely. This is because at some frequencies the spatial harmonics

84
effects are considerable and thereby the dispersion behaviour changes at narrow
frequency bands.
Most of the EBGs realised are finite in the microwave frequency region and in
this case, there is a coupling between the incident field and Bloch waves inside the
structure. This coupling depends on numerous parameters and the field in a finite EBG
can never be reduced to a combination of Bloch waves [16]. Bloch theory can predict
the properties of the crystal to a great extent (e.g. transmission characteristics). It does
not take into consideration the fact that all realisable EBGs are of finite extent, which is
generally excited by an incident field. So it neglects evanescent waves and coupling
that exist in all real situations. This suggests that an analysis is needed which include
more realistic situations.
By considering many elements of the periodic structure excited by an incident
field, one could incorporate the interaction of higher order spatial harmonics and
evanescent wave appears in finite structures. Figures 3.3 and 3.4 illustrate the
difference between the infinite and finite periodic structures and the numerical domain
for finite structures. In the case of infinite crystal, it can be seen that the entire space is
filled with the crystal itself and there exist only Bloch modes inside the crystal and
there is no incident field. For a finite crystal which is limited in extend, there exist an
incident field excitation and this leads to evanescent modes in addition to the Bloch
modes. In a computational procedure for the finite crystal, the infinite extending
surroundings have to be terminated with absorbing boundaries (fig.3.5).

85

Fig.3.3. A Diagram illustrating the case of an infinite photonic crystal. The entire space is
occupied by the photonic crystal.

In the subsequent studies of this work, various periodic structures are
considered. As a first step, structures consisting of lumped elements will be discussed
in this chapter. In the proposed method, more elements are considered for the analysis.
It is observed that there can be narrow frequency regions at which one could obtain
backward radiation phenomenon. In the next chapter, metallic periodic structures are
analysed using plane wave illumination in certain directions and negative refractions at
frequencies other than lower edges of bandgap are identified. In this case, the
interaction of higher order Floquets harmonics and the coupling between the incident
field and the structure are taken into consideration as it appears in the calculated S
parameters.

Photonic Crystal

Propagating Bloch modes

No incident field

- +
Problem space

-

P
r
o
b
l
e
m

s
p
a
c
e


86

Fig.3.4. A Diagram illustrating the case of a finite periodic structure. The surrounding extends
to infinity and there exist Bloch modes and evanescent modes.

Fig.3.5. A Diagram illustrating the numerical approach for a finite periodic structure. The
boundaries are terminated by absorbing boundary conditions.

Air
Periodic
structure
Propagating Bloch
modes +
Evanescent modes

Incident Field
Absorbing boundary
A
b
s
o
r
b
i
n
g

b
o
u
n
d
a
r
y

Absorbing boundary

I
n
c
i
d
e
n
t

f
i
e
l
d

Air

Periodic
structure
Propagating Bloch
modes +
Evanescent modes

Incident Field

I
n
c
i
d
e
n
t

f
i
e
l
d

Problem space
- +

-

P
r
o
b
l
e
m

s
p
a
c
e

+


87
3.6 Computational Electromagnetics (CEM) and simulation packages
Numerical simulation tools are vastly used these days in all branches of both
engineering and biological sciences. Computational Electromagnetics (CEM) is the
discipline that intrinsically and routinely involves the use of digital computer to obtain
numerical results for electromagnetic problems [23]. It is not uncommon to employ
analysis and /or CEM to understand experimental results, nor is it uncommon to verify
analysis results and CEM results with experimental results. In this work, commercial
electromagnetic simulations tools Ansoft High Frequency Structure Simulator
(HFSS), CST Microwave Studio were used for various studies. Agilent Advanced
Design System (ADS) was used for lumped circuit simulations.
HFSS is the industry-standard software for S-parameter and full-wave SPICE
extraction and for the electromagnetic simulation of high-frequency and high-speed
components [24]. It can be used for the design of on-chip embedded passives, PCB
interconnects, antennas, RF/microwave components, and high-frequency IC packages,
scattering/radiation problems etc. It solves problems in frequency domain using Finite
Element Method (FEM). The recent versions of HFSS can solve coupled thermal and
stress analysis by incorporating ePhysics, a software from Ansoft which can take into
account of mechanical and thermal parameter variations [24]. Ansoft HFSS uses
FEM in frequency domain with some advanced numerical methods for matrix solving.
Another commonly used high frequency simulation tool is CST Microwave
Studio. This software uses the Finite Integral Technique (FIT) for solving
electromagnetic problems with combination of some advanced techniques to increase
the computational performance. These techniques are mainly the Perfect Boundary
Approximation technique (PBA), Thin Sheet Technique (TST) and Multilevel Sub
Griding scheme (MSS) [25]. This is based on time domain techniques.


88
The package can give frequency domain information by performing Fourier transform
of the time domain data.
All the electromagnetic computational tools fundamentally deal with solving
Maxwells equations at different problem scenarios. Since the Maxwell equations can
be written either in integral or differential forms, different computational methods can
be deployed. Fig.3.6 shows the categories in CEM and the domains. The domains
where CST Microwave Studio and Ansoft HFSS are also shown.

Fig.3.6. Categories within computational electromagnetics and the domain of the packages CST
Microwave studio and Ansoft HFSS

In order to avoid any misinterpretation of the results obtained from commercial
packages, it is advantageous to know about the computational schemes used in each
package. In the next sections, the basis of the main computational schemes; Finite
Element Method (FEM) and Finite Integration Technique (FIT) will be explained.


89
3.6.1 Finite Element Method (FEM)
Many popular electromagnetic simulation packages (e.g. HFSS and
EMSolve) employ the finite element method and its hybrid versions for robust and
adaptable modelling. The basis of FEM is to take a complex problem whose solution
may be difficult to obtain, and decompose it into pieces upon each of which a simple
approximation of the solution may be constructed, and then put the local approximate
solutions together to obtain a global approximate solution [27]. In order to find the
local solutions, the problem space is divided into small elements (meshes).
FEM is used for solving partial differential equations (PDE) approximately.
Solutions are approximated by either eliminating the differential equation completely
(steady state problems), or converting the PDE into an equivalent ordinary differential
equation (without any partial differentiation involved), which is then solved using
techniques such as finite differences. In solving partial differential equations, the
primary challenge is to create an equation which approximates the equation to be
studied. This equation should be numerically stable, meaning that errors in the input
data and intermediate calculations do not accumulate and cause the resulting output to
be of no use.
A PDE will involve a function u(x) defined for all x in the domain with respect
to some given boundary condition. The purpose of the method is to determine an
approximation to the function u(x). The method requires the discretisation of the
domain into sub regions or cells. For example, a two-dimensional domain can be
divided and approximated by a set of triangles (the cells). On each cell, the function is
approximated by a characteristic form. Here u(x) can be approximated by a linear
function on each triangle [27].


90
3.6.1.1 Finite element method for electromagnetic problems
In electromagnetic problems, Maxwells equations can be solved using FEM. In
Ansoft HFSS the first step is to mesh the problem space into tetrahedral elements. In
the second step, it solves the field equation derived from Maxwells equation. Consider
the electric field equation [48]

Obtain a basis function W
n
which define conditions between nodal locations in the
overall mesh of tetrahedra, based on problem inputs. Third step is to multiply the basis
functions with equation (3.16) and then integrate the result over the volume V. This
gives

Where n = 1, 2, 3... N and N is the total number of unknowns.
Rewriting equation (3.17) using Greens and Divergence theorems gives

Writing E field as a summation of unknowns using the same basis function used before
gives
0
1
2
=
|
|
.
|
\
|
E k E
r o
r
v v
(3.16)
=
(
|
|
.
|
\
|

V
n r o
r
n
dV E W k E W 0
1
2
v v
(3.17)
( )

=
(
|
|
.
|
\
|

V
S
n r o
r
n
dS term boundary dV E W k E W ) (
1
2
v
(3.18)

91

Substituting equation (3.19) in (3.18) yields

Now equation (3.20) is in the form of a matrix equation

AX = B

where A is the basis functions and field equation, in a known N x N matrix, X is the
unknowns to be solved for and B is the excitation. Equation (3.20) can be solved by
using matrix solution methods. HFSS uses an iterative algorithm for the matrix
solution.

3.6.2 Finite Integration Technique (FIT)
The FIT technique was developed by Weiland [29, 30] in 1977 which provided
volume type discrete reformulations of Maxwells equations in an integral form. The
first step of the finite integral technique is to restrict the electromagnetic problem to a
bounded space region R
3
, which contains the space region of interest. Next step is
to decompose the computational domain into a finite number of cells V
i
in a way that
all cells have to fit exactly to each other. This decomposition gives finite cells complex
G, which serves as computational grid. Fig. 3.7 shows the diagram of such a grid in the
=
=
N
m
m m
W x E
1
v
( )

=
|
|
.
|
\
|
(
|
.
|
\
|

=
S V
m n r o m
r
n
N
m
m
dS term boundary dV W W k W W x ) (
1
2
1

(3.19)
(3.20)
(3.20) (3.20)

92
cell complex G.

Fig. 3.7. A cell V
i,j,k-1
of the cell complex G with the allocation of the electric grid voltages e on
the edges of A and the magnitude facet flux b through this surface.

Assume a rectangular shaped computational domain . Also, each edge of the cells
includes a direction. Let i, j and k be the number of mesh points along x, y and z
directions. After the definition of grid cell G, the further introduction of the finite
integral theory can be restricted to a single cell volume V
n
. Starting with Faradays law
in integral form [30]
). , ( ). , ( A d t r B
t
s d t r E
A
A
v
v
v
v v
v

can be rewritten for a facet A
z
( i, j ,k) of V
n
as the ordinary differential equation

) , , ( ) , , 1 ( ) , 1 , ( ) , , 1 ( ) , , ( k j i b
t
k j i e k j i e k j i e k j i e
z y x y x
= + + + +

as shown in fig. 3.7.
The scalar values
) , 1 , ( k j i e
x
+
) , , ( k j i e
y
) , , ( k j i b
z
) , , ( k j i e
x
) , , 1 ( k j i e
y
+

(3.21)
(3.22)

93

s d E k j i e
k i i
k i i
z y x
z y x
x
v
v
. ) , , (
) , , (
) , , (
1
+
=

is the voltage along one edge of the surface A
z
( i ,j ,k), representing the exact value of
the integral over of the electric field along this edge. The scalar value

A d B k j i b
k j i A
z
z
v v
. ) , , (
) , , , (
=

represents the magnetic flux.
Assuming a lexicographical ordering of the electric voltages e(i,j,k) and the
magnetic fluxes b(i,j,k) over the whole cell complex G and their assembly into column
vectors in such a way, that we compose the degrees of freedom first in the x, y and z
directions. The eqn.3.21 of all grid cell surfaces of the complex G can be collected in a
matrix form using finite integral approximation similar to eqn. 3.22 for the three facets.

|
|
|
.
|
\
|
=
|
|
|
|
|
|
|
.
|
\
|
|
|
|
.
|
\
|
M
M
M
M
n
n
n
n
b
t
e
e
e
3
2
1
... ... ...
1 - ... 1 - ... 1 ... 1
... ... ...

C . e = b

The matrix C represents a discrete curl operator on the grid G. The second discrete
operator to be considered is the divergence.
(3.25)
(3.23)
(3.24)

94

Fig.3.8.The allocation of the six magnetic facet fluxes which have to be considered in the
evaluation of the closed surface integral for non-existence of magnetic charges within the cell
volume.

Maxwells equations describing the non existence of magnetic charges is [30]

0 . ) , ( =
A d t r B
v
v
v
v

which is considered for a volume cell V
i,j,k
as shown in fig.3.8.

The evaluation of the integral in eqn.3.26 for the cell shown above gives

0 ) 1 , , ( ) , , (
) , 1 , ( ) , , ( ) , , 1 ( ) , , (
= + +
+ + + +
k j i b k j i b
k j i b k j i b k j i b k j i b
x z
x y x x

This relation for the single cell can be expanded to the whole cell G and this yields a
discrete divergence matrix

) , 1 , ( k j i b
y
+
) , , ( k j i b
z
) , , 1 ( k j i b
x
+

) , , ( k j i b
x
) 1 , , ( + k j i b
z
) , , ( k j i b
y
(3.27)
(3.26)

95
0
... ...
. 1 1 - 1 1 - 1 1 - .
... ...
6
5
4
3
2
1
=
|
|
|
|
|
|
|
|
|
|
|
.
|
\
|
|
|
|
.
|
\
|
M
M
m
m
m
m
m
m
b
b
b
b
b
b

S . b = 0

In the above discussion, it has been shown that how to perform the discretisation of two
of the four Maxwells equations. For the discretisation of these two equations using
finite integration technique, another cell complex G1 which is the dual to the primary
cell G is considered. For the Cartesian coordinates, G1 is obtained by taking the foci of
the cells of G as grid points for the mesh cells of G1. Fig.3.9 shows the spatial
allocation of a cell G and its dual G1.
The complete integral of the charge density within a dual cell V1 can be related
to a discrete charge onto the single grid point of the primary grid G placed inside V1.
The discretisation of Amperes law in integral form

( ) A d t r J t r D
t
s d t r H
A
A
v
v
v
v
v
v
v
. ) , ( ) , ( . ) , (
1
1
|
.
|
\
|
+

(3.28)

96

Fig. 3.9. The spatial allocation of a cell and a dual cell of the grid doublet {G, G1}

can be performed for a facet A1 of dual grid cell V1 in the same way as Faradays law.
Finally, Gauss Law in integral form can be discretised for dual grid cells. These will
result in matrix equations. These equations are called Maxwell grid equations. Let C1
be the grid operator for the dual discrete curl and S1 for the dual discrete divergence.
For the cell complex pair {G, G1}, the complete set of discrete matrix equations are
given by [30, 31]

q d S b S
j d
dt
d
h C b
dt
d
e C
= =
+ = =
v v v
v
v v
v
0
1

This system of matrix equations can be solved using numerical techniques. This is the
basis of FIT.

Table 3.1 presents a comparison of FEM and FIT.

Dual Grid G1
Grid G

97
Method Basis Capabilities and limitations
Finite Element
Method (FEM)
Frequency domain
differential
equations
Can handle complex geometries,
Dispersive modelling will result in longer
simulation time, Gives approximate
solution
Finite Integration
Technique (FIT)
Time domain
integral equations
Can handle complex geometries and
dispersive materials, Having numerical
dispersion

Table. 3.1. A comparison of FIT, FEM methods

3.7 EBG modelling in commercial packages
Most of the commercial packages are capable of modelling periodic structures
using periodic boundary conditions. Periodic boundary conditions can be implemented
using the symmetry in the structure and this way the computational time can be reduced
for determining the transmission response of the system. Periodic boundary conditions
are the computational implementation of Bloch theorem as given in eqn. 3.12. In CST
Microwave Studio and Ansoft HFSS these problems can be solved by considering
it as an eigen mode problem. This gives the result for source free solution under given
boundary conditions. In this case, the periodic structure in the computational model is
similar to a cavity problem and the solutions are calculated in a source free
environment for electric or magnetic fields satisfying the given boundary conditions.
For a given structure, there can be multiple solutions and these solutions are termed as
the eigen solutions of the problem. For example, the gap between the first and second
eigen modes in the wave vector space can be identified as the first bandgap of an EBG.


98
3.8 Numerical simulations of LHMs and analogous LHMs
Following the theoretical and experimental verification of the LHMs, analogous
LHMs and their characteristics, there have been lots of interests in numerically
studying the effects on wave propagation through in LHMs. These investigations
provided deep insight into the phenomenon especially from an application point of
view. The main complexities in numerical studies of LHMs are the dispersive
modelling and computational resources. Both time and frequency domain methods
were employed in various studies.
Investigations of LHMs characteristics began with the numerical simulation of
Inverse Snells law, focusing and applications of LHMs such as phase shifters. Some of
these studies considered the dispersive material characteristics of LHMs by applying
the Drude model [40]. Dispersive material modelling in frequency domain needs large
computational resources. So time domain modelling is usually preferred. For an
accurate numerical study of LHMs, the simulation model has to take care of the
material dispersion. Using this dispersion, the existence of negative material parameters
can be explained [40]. Let W be the energy density, E, H be the electric and magnetic
fields respectively. Then

2 2
2
1
2
1
H E W + =

In a case where there is no dispersion and if both material parameters are negative, the
total energy becomes negative which is a meaningless concept. If one takes the fact that
there is dispersion eqn. 3.29 becomes [34, 39]

(3.29)

99
(
=
2 2
] ) ( [ ] ) ( [
2
1
H E W

From the equation (3.30), it can be seen that for the energy density W to be positive, the
material parameters must obey the following conditions.

0
)] ) ( [
>

d
d

0
)] ) ( [
>

d
d

Following this criteria , can be defined as a plasma medium model [40]

]
) (
1 [ ) (
2
0
e
pe
i

=

]
) (
1 [ ) (
2
0
m
pm
i

=

Where
pe
is the electronic plasma frequency,
pm
is the magnetic plasma frequency.
e
is electric collision frequency and
m
is the magnetic collision frequency. The above
relations (eqns.3.33 and 3.34) represent the Drude model for material parameters.
FDTD simulations of LHMs were widely employed by many researchers [32-39].
Wave focusing using an LHM slab is demonstrated using dispersive FDTD in [33, 37]
considering the material parameter variations with frequency. Wave propagation in a
multilayer stack which consisted of thin alternating layers of conventional materials
and lossy LHMs were demonstrated using dispersive FDTD in [34, 39]. This structure
(3.30)
(3.31)
(3.32)
(3.33)
(3.34)

100
can act as a wave transportation system for evanescent waves over a large distance. The
simulated near field intensity through a multilayer LHM structure with refractive index
n = -1 is presented in fig. 3.10. It can be noticed that the image is transferred through
four layers of LHM which are equally spaced apart. It was also shown in [34] that the
LHM slab

450 500 550 600 650
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Image
Plane
Object
Plane
LHMs slab: 3 wavelength long
1/10 wavelength wide
cell=wavelength/220

E
z
Cell

Fig.3.10. Diagram showing the simulated near field intensity through multilayer LHM structure
with refractive index n = -1. (Reproduced from [34])

compensates for the phase delay of the wave as it passes through the vacuum layers in
between.
Numerical studies of a composite medium consisting of split rings and wires
using Finite Integration Technique (FIT) were detailed in [41]. In this work, a unit cell
consisting of a conducting wire and a split ring resonator with periodic boundary
conditions was selected for modelling. The scattering parameters were calculated and
from this an effective refractive index was obtained. Numerical simulations were
performed to study the losses in the split ring wire model. The Ohmic and dielectric

101
losses in the structure were studied in detail and it was observed that in addition to the
two lossy mechanisms, the transmission will also be affected by the thickness of the
conducting layer. Low dielectric substrates were found to be causing conductive losses
high.
Frequency domain solvers can also be applied for LHMs problems to some
extend. The latest versions of HFSS can solve these problems to some extent [42]. In
[42], Inverse Snells Law verification and slab lens focusing were verified using HFSS.
However, due to the difficulties in implementing a non linear dispersive material
properties and the high computational resource requirement, this package has not
received much interest in LHMs studies. Fig.3.11 shows the Ansoft HFSS simulation
results demonstrating positive refraction from a wedge of permittivity = 2.2 and = 1
and negative refractions from a wedge of = -2.2 and = -1 [42]. It can be seen that
the refraction directions are opposite directions in each case.

Fig.3.11: Simulated field showing refraction from a wedge with positive refractive index (left)
and negative refractive index (right). Refracted rays are shown in dotted arrows. (Reproduced
from [42])

102
The above mentioned works were on the characteristics of an LHM media which is
homogeneous (if such a medium exists). The wave propagation in analogous LHMs
obtained from EBGs was also studied using numerical simulations. As discussed in
chapter 2, all these analysis considered infinite EBGs, showing negative refraction
phenomenon near the bandgap edges.
Inverse Snells Law verification using FDTD from a periodic structure
consisting of triangular lattice of metallic wires for TM polarization was presented in
[43]. The radius to period ratio was 0.2. The simulated electric field intensity plots for
positive and negative refractions are given in fig. 3.12. It can be observed that at 7
GHz, the EBG exhibits positive refraction. At 9 GHz, the EBG shows negative
refraction. However it has to be noted that at 9 GHz there exists a positively refracted
components, even though it is low in intensity compared to the negatively refracted
component. This establishes the fact that in an EBG there exist both positively and
negatively refracted components. There have been many works one could find in
literature on the numerical simulation of wave propagation at negative refraction
frequencies in EBGs and some of them were briefed in chapter2. The intuitive
observation is that it considered the possibility of getting negative refraction from
infinite EBGs.

103

Fig.3.12. Simulated electric field intensity showing positive refraction at 7 GHz and negative
refraction at 9 GHz from a metallic EBG. (Reproduced from [43])

3.9 Lumped element circuit simulations
In order to verify the above discussed facts, numerical simulations of lumped
element circuits using ADS were carried out. The aim of the simulations was to
establish that the super cell analysis can predict backward radiation frequencies due to
the spatial dispersion in the structure. In these simulations, the finite periodic structure
itself acts as a repeating element of an infinite periodic structure. The dispersion
diagram obtained revealed that backward radiation exists at narrow frequency bands
due to spatial dispersion in the structure when more elements were included in the
model. The dispersion behaviour was obtained from S parameters as detailed in section
3.2.2. Bandgap regions are those regions with group velocity zero (slope of the
dispersion diagram is zero) and phase constant is either zero or .


104
3.9.1 C-L High pass filter circuit
Consider a high pass filer consisting of unit cells of capacitors and inductors (C-
L) as shown in fig.3.13. As discussed in chapter 2 (sections 2.10.2), this structure was
identified as an equivalent LHM medium. Let C = 0.5 pF and L= 1 nH. The
transmission response of the unit cell with periodic boundary condition is given in
fig.3.14 which shows a high pass filter response with a cut-off frequency around 5.0
GHz. The dispersion diagram obtained using unit cell analysis is given in fig. 3.15.
From fig.3.15, it can be noticed that this structure is a backward radiating structure
throughout the passband [44] when unit cell analysis is used for the analysis.

Fig.3.13. A unit cell of a high pass C-L filter with C= 0.5pF and L= 1.0nH

Fig.3.14. Response of the C-L filter showing pass band from about 5 GHz


105

Fig.3.15. Dispersion diagram for the high pass C-L filter with unit element

From this diagram, it can be seen that from 5 GHz onwards there is a passband, which
agreed well with frequency response in fig. 3.14. From the unit-cell analysis, planar
structure analogy to such a lumped-element circuit is a backward radiating structure
throughout the pass band starting from 5 GHz, which can be observed from the
dispersion diagram. Next, more elements were included in the simulation model to see
the effect of spatial harmonics in the model. When more elements were considered as
repeating elements in numerical analysis, the high order spatial harmonics were taken

Fig.3.16.Dispersion diagram for the high pass C-L filter with three elements

P
h
a
s
e

c
o
n
s
t
a
n
t

P
h
a
s
e

c
o
n
s
t
a
n
t


106
into account. Fig.3.16 shows the dispersion diagram when three elements were
considered. From the figure, it can be seen that the backward radiation regions were
narrowed to a frequency region of 5.0 GHz-5.7 GHz. When high-order spatial
harmonics are taken into account by considering many elements, the dispersion curve
has changed and negative dispersion disappeared at some frequency bands. This
indicates that when high order harmonics are considered, the dispersion behaviour
changes from that of a unit cell analysis. Hence the high pass model can not satisfy the
claim [44] that the medium can act as an ultra wide band backward radiating antenna
throughout the passband.
3.9.2 L-C low pass filter circuit
A similar analysis was carried out for a low pass type L-C filter. The unit cell is
shown in fig. 3.17, in which C=1.5 pF and L = 1.5 nH. The transmission response of
the filter shows a passband up to 4.3 GHz as in figure 3.18. Dispersion diagram
obtained from a unit cell is given in figure.3.19. From fig.3.19, it can be seen that there
is no backward radiation throughout the passband for the unit cell dispersion diagram.
Fig. 3.20 shows the dispersion diagram with three elements. It can be seen that the
dispersion behaviour is changed in the three elements case. There is a backward
radiation frequency region starting at 2.5 GHz, which was absent in the unit cell
analysis case.

Fig.3.17. A unit cell of a low pass L-C filter with C= 1.5 pF and L= 1.5 nH

107

Fig.3.18 Response of the low pass L-C filter showing pass band up to 4.3 GHz

Fig.3.19. Dispersion diagram for the low pass L-C filter with unit element
2 3 4 5 6 7 8 9 1 10
-2
0
2
-4
4
freq, GHz
b
e
t
a
*
d
P
h
a
s
e

c
o
n
s
t
a
n
t

S

(
2
,
1

)

d
B


108

Fig.3.20.Dispersion diagram for the low pass L-C filter with three elements

3.10 Spatial harmonics suppression in periodic structures
The suppression of one or more spatial harmonics in a periodic structure can
lead to novel applications in electromagnetics. Solntsev [46, 47] has recently studied
the spatial harmonics selection in pseudo periodic waveguides depending on frequency
[46,47]. This is based on the coordinated variation of the period and the phase
distribution of elements along the waveguide and thereby giving constant phase
velocity to one spatial harmonic and suppressing other harmonics. The principle of
selection based in using electrodynamic systems with non periodic spacing of its
elements and a specified relation between the step L
q
and the field phase
q
of the
elements, which makes it possible to select one spatial harmonic and suppress the
others. In this method, the phase distribution is determined by the shape and
dimensions of the system elements which are to be chosen from the conditions of
spatial harmonics selection. This is analogous to a filter whose frequency
characteristics also determined by the choice of elements and the period.

2 3 4 5 6 7 8 9 1 10
-2
0
2
-4
4
freq, GHz
b
e
t
a
*
d
P
h
a
s
e

c
o
n
s
t
a
n
t


109

Fig.3.21. A Pseudo Periodic Waveguide: A comb type structure (Reproduced from [46])

The longitudinal electric field distribution along the system with Q steps and different
length L
q
(q=1, 2, 3,Q) as shown in fig.3.21 is given by [46]

)] , ( exp[ ) , ( ) , (
0
z i z f E z E
z
=
(3.35)

The type of the periodic system (uniform periodic or non uniform) governs the
distribution of real amplitude f (z, ) and phase (z, ). Applying Fourier transform,
the amplitudes E(h, ) of spatial harmonics is given by [46].

dz ihz z E
l
h E
l
z
) exp( ) , (
1
) , (
0
=

In general, the amplitudes E (h, ) are continuous functions of the wave number h and
frequency and differ in spectral density by a factor l, where l is the length of the
system. Now, expressing E (h, ) as a sum over Q steps of the system gives

(3.36)
0 Z
1
Z
2
Z
Q
= l Z
q=1 2---------------------------Q
d
q

L
q

110
)]] ) ( ( exp[
1
) , (
1
q q q
Q
q
q
hz i M U
l
h E =

=

Where
q
() = (z
q
, ) is the average field phase at the q
th
step. M
q
(h) is the local
electron interaction coefficient. U
q
is the RF voltage at the q
th
step, z
q
d
q
are the mean
coordinate and effective width of the q
th
gap. If the field is constant in the gap f(z) = f
q
,
we have the expression [46]
2
2
sin

q
q
q
hd
hd
M
|
|
.
|
\
|
=

From eqn. (3.37), we can get the maximal values E (h, ) for the wave numbers h = h
m

that satisfy Q conditions

... 3 , 2 , 1 , 2 = + = q qm z h
q q m

m is an integer m = ..-2,-1, 0, 1, 2determines the number of the field spatial
harmonic with maximal amplitude. Now, introducing a field phase shift
q
=
q+1

q

at the q
th
step taking into account that L
q
= z
q+1
-z
q
. This yields the equivalent
conditions of synchronism for every step as

Q q m L h
q q m
,... 2 , 1 , 2 ) ( ) ( = + =

In a periodic waveguide, L
q
= L,
q
= and
q
= q, so the conditions for
synchronism will meet for an infinite number of spatial harmonics m = m. In a non
uniform waveguide with different steps L
q
, the synchronism condition can only be
(3.37)
(3.38)
(3.39)

111
satisfied by one harmonic by appropriately choosing the phase
q
. In this way,
selection of spatial harmonics takes place [35]. In other words, the spatial harmonics
selection can be done by making the structure pseudo periodic or by introducing a
phase changing mechanism. This idea is applied to metallic periodic structures for
negative refraction enhancement and is presented in chapter 4 of this thesis.

3.11 Conclusions
In this chapter, a brief discussion on periodic structure analysis is presented. It
is shown that conventional unit-cell approach for infinite periodic structures is not
adequate to predict negative refraction phenomenon at certain frequency bands. It is
seen that there are forward and backward radiating spatial harmonics components in a
periodic structure. These spatial harmonics can interact with each other and change the
dispersion characteristics. It has been identified that in a realisable periodic structure
which is always finite, there exist dispersion sources which are not included in the unit
cell model analysis. The method of considering finite structure as a super cell with
incident field excitation will take into account the higher order spatial harmonic
interaction, incident field coupling with the structure and evanescent wave interactions
towards dispersion characteristics. Numerical simulations of structures consisting of
lumped capacitor and inductor models were performed to verify the finite structure
analysis and spatial harmonics effects on dispersion characteristics. The variation of
backward radiating frequency regions from the unit cell was observed when more
elements were included in the analysis.


112
References
[1] R. E Collin, Foundations for Microwave Engineering, Chapter 8, Mc-Graw-Hill
Kogakusha Ltd, pp. 368-398, 1966
[2] E.R Brown, C. D Parker, E. Yeblonovitch, Radiation properties of a planar
antenna on a photonic crystal substrate, Journal of Optical Society of America.
Vol. 10, pp. 404, 1993
[3] R. Biswas, S. D Cheng, E. Ozbay, S. McCalmont, W. Leung, G. Tuttle, K. M
Ho, Photonic Bandgap Materials, Vol. 315, NATO Advanced Studies Institute,
Series E, Applied Sciences, edited by C. M Soukoulis, Kluwer publications, pp.
377
[4] Alomainy A. H, Hao. Y, Parini. C. G, EBG Structure for indoor WLAN
Antenna with beam shaping, Proc. 27
th
ESA Antenna Technology Workshop,
Spain, pp. 377-383, 2004
[5] M. Qiu, S. He, A non orthogonal finite difference time domain method for
computing the band structure of a two dimensional photonic crystal with dielectric
and metallic inclusions, Journal of Applied Physics, Vol. 87, pp. 8268- 8275,
2000
[6] Ansoft High Frequency Structure Simulator-Users Manual, Ansoft Corporation,
2003
[7] P. A. Belov, C. R Simovski, Excitation of semi infinite electromagnetic crystal,
Proc. of Bianisotropy conference (BIAN), Belgium, 2004.
[8] E. I. Smirnova, C. Chen, M. A Shapiro, J. R Sirigiri and R. J Temkin,
Simulaton of photonic bandgaps in metal rod lattices for microwave applications,
Journal of Applied Physics, Vol. 91, pp. 960-968, Feb 2002.
[9] Masaya Notomi, Negative refraction in photonic crystals, Optical and Quantum
Electronics, Vol. 34, pp. 133143, 2002

113
refraction without negative effective index, Physical Review B, Vol. 65, 201104
without negative index in metallic photonic crystals, Optics Express, Vol.11, pp.-
746754, 2003
superlens, Physical Review Letters, Vol.91, pp. 207401-1-207401-3, 2002
Costas M.Soukoulis, Negative refraction by photonic crystals, Nature, Vol.
423, pp. 604-605, 2003
[14] P.V Parimi, W. T. Lu, P. Vodo, J. Sokoloff, S. Sridhar, Negative refraction
and left- handed electromagnetism in microwave photonic crystals, cond-
mat/0306109, 2003 (On line archive)
by flat lens using negative refraction, Nature, Vol.426, pp. 404, 2003
properties of photonic crystals, Journal of Optical Society of America, Vol.17,
No.6, pp.1012-1020, 2002
[17] G. D. Mahan, G. Obermair, Polaritons at surfaces, Physical Review B. Vol.
183, No.3, pp. 834-841, 1969
[18] Anthony Grbic and George V. Eleftheriades, Experimental verification of
backward wave radiation from a negative refractive index metamaterial, Journal
of Applied Physics, Vol. 92, pp. 5930-5936, 2002
[19] Anthony Garbic and George V. Eleftheriades, A backward wave antenna
based on negative refractive index L-C networks," Proc. of the IEEE Intl.

114
Symposium on Antennas and Propagation, Vol. IV, pp. 340-343, San Antonio,
TX, 2002
[20] Ashwin K. Iyer, George V. Eleftheriades, Negative refractive index
metamaterials supporting 2-D Waves, Proc. of the IEEE Int. National Symposium
on Microwave Theory and Techniques, Vol.2, pp.1067-1070, 2002
[22] Sakoda Kazuaki, Optical properties of photonic crystals, Springer, 2001
[23] E. K. Miller, A selective survey of Computational Electromagnetics, IEEE
Transactions on Antennas & Propagation, Vol.36, pp. 1281-1305, Sept. 1988
[24] www.ansoft.com
[25] http://www.cst.de/Content/Products/MWS
[26] www.agilent.com
[27] Peter Hunter, FEM/BEM Notes ( Online Resource: www.auckland.ac.nz)
[28] Warren L. Stutzman, Gary A. Thiele, Antenna Theory and Design, Wiley &
Sons, 1998
[29] Weiland. T, A discretisation method for the solution of Maxwells equations for
six-component fields, Electronics and communications AEU, Vol.31,1977,
pp.116-120
[30] M. Clements and R. Weiland, Discrete electromagnetism with the finite
integration technique, Progress in electromagnetic research, PIER, Vol. 32,
pp.65-87, 2001
[31] CST Microwave Studio Advanced Topics, Uses Notes, Chapter 4.
[32] X. S Rao, C. K. Ong, Amplification of evanescent waves in a lossy left-handed
material slab, Physical review B, Vol. 68, pp. 113103, 2003.
[33] Lu L, Hao Y, Parini C. G, Dispersive FDTD characterisation of no phase-delay
radio transmission over layered left-handed meta-materials structure, IEE Proc.

115
Science Measurement and Technology, Vol.151, No.6, pp. 403-406, 2004.
[34] L. Lu, Y. Hao and C.G. Parini, Dispersive FDTD modelling on near field
imaging in multilayer left handed metamaterials at microwave frequencies
Proc. 2003 High Frequency Postgraduate Student Colloquium, Queens
University, Belfast, Sept. 2003,
[35] S. Rikte, G. Kristensson and M. Andersson, Propagation in bianisotropic
media - reflection and transmission, IEEE Proc. Microwaves, Antennas and
Propagation , Vol. 148 , pp. 29-36, 2001
[36] F. J. Rachford, D. L. Smith, P. F. Loschialpo, and D. W. Forester ,
Calculations and measurements of wire and/or split-ring negative index media,
Physical Review E, Vol. 66, 036613, 2002
[37] R. W. Ziolkowski, E. Heyman, Wave propagation in media having negative
permittivity and permeability, Physical Review E, Vol 64, pp.056625-056629,
2001
[38] H. Mosallaei and Y. Rahmat-Samii., Composite materials with negative
permittivity and permeability properties: concept, analysis, and characterization,
Proc. IEEE Antennas and Propagation Symposium, Vol. 4, pp. 378-381, 2001
[39] Y. Hao Y, L. Lu, Parini C. G, Dispersive FDTD modelling on multi-layer left-
handed meta-materials for near/far field imaging at microwave frequencies,
Proc. of IEEE Antennas and Propagation Symposium, Vol. 4, pp.639-642, 2003
[40] V. G Vaselago, The Electrodynamics of substances with simultaneously negative values
of and , Soviet Physics USPEKHI, Vol.10, pp. 509-514, 1968.
Experimental verification and simulation of negative index of refraction using
Snells Law, Physical Review Letters, Vol.90, pp. 107401-1-107401-4, 2003
[42] Pavel Kolinko and David R. Smith, Numerical study of electromagnetic waves

116
interacting with negative index materials, Optics Express, Vol. 11, pp. 640-648,
2003
[43] P. V. Parimi, W. T. Lu, P. Vodo, J. Sokoloff, J. S. Derov, and S. Sridhar,
Negative refraction and left-handed electromagnetism in microwave photonic
crystals, Cond-mat/0306109, 2003 (Online resource)
[44] Simon Ramo, John R. Whinnery, Theodore Van Duzer, Fields and Waves in
Communication Electronics, John Willey & Sons, 1984
[45] P. E. Mayes, G. A Deschamps, W.T. Patton, Backward-wave radiation from
periodic structures and application to the design of frequency- independent
antennas, Proc. IRE, vol. 49, pp. 962-963, 1961
[46] V. A. Soltsev, Properties of spatial harmonics selection in pseudoperiodic
waveguides, IEEE MTT-S International Microwave Symposium Digest, Vol.2,
pp. 829-832, June 2003
[47] V. A. Solntsev, K. P. Solntseva, Mode selection analysis in pseudoperiodical
slow wave structures, Proc. of IEEE Trans Black sea Region Applied
Electromagnetism Symposium, pp. MMWS-13 M MWS-13, 1996
[48] Baofu Jia, Getting started: Ansoft HFSS 8.0, Presentation material, University
of Electronic Science and Technology of China

117
Chapter 4

Negative Refraction Phenomenon from
Metallic Wire Medium

4.1 Introduction
In the previous chapters, it was shown that periodic structures are capable of
backward radiation and negative refraction at certain frequencies in the passband due to
the strong spatial dispersion effects. In this chapter, further verification of the idea
using EBGs consisting of metallic wires is considered in depth. Such structures are
generally called Metallic Electromagnetic Bandgap Structures (MEBGs) or metallic
photonic crystals (MPCs). The extensive studies on electromagnetic bandgap (EBG)
structures began with dielectric EBGs [1-6] and most recently a growing interest in
metallic EBGs [7-12] has emerged. Negative refraction phenomenon at the lower edges
of bandgap frequencies in dielectric EBGs was presented in [13-17] and the same
phenomenon in MEBGs [18, 19]. The limitation of this work was that the observed
negative refraction phenomenon was at frequencies in the bandgap edges and this limits
the transmitted power. Fig.4.1 depicts the frequency region for bandgaps and bandgap
edges. Also, the observed phenomenon was not reported for multiple frequency bands
using an EBG.

118

Fig. 4.1. A diagram showing passband and bandgap behaviour of an EBG

All the aforementioned works were based on the unit cell analysis which is not
adequate for refraction studies. In this chapter, it is demonstrated that negative
refraction can be obtained at frequencies in passband in addition to the lower edges of
the bandgap. This enables higher wave transmission. However, as in these structures
there are no negative effective parameters, the observed phenomenon should be more
accurately termed as refraction like phenomenon. For simplicity, the term refraction is
used in the following sections of this chapter.
MEBG structures are studied for negative refraction verification. Numerical
simulations were carried out for a slab model and a prism structure. Experimental
results in good agreement with simulations are presented for the prism model and an
approach for enhancing the negative refraction at certain frequencies is proposed and
verified.

4.2 Braggs Law and bandgaps in EBGs
As discussed before, EBGs consist of periodic arrangement dielectric/metallic
shapes. The basic principle of an EBG is the Braggs reflection. This occurs in a crystal
when the period of the crystal obeys Braggs law [20]. It is given by the equation

T
r
a
n
s
m
i
s
s
i
o
n

Frequency

B
a
n
d
g
a
p

Passband
Passband
Bandgap edges

-15dB
0dB
119

sin 2d n =

where n is an integer
is the wavelength
d is the period
is the angle between the scattering plane and the incident wave
Figure 4.2 illustrates the Bragg reflection from a crystal. The red lines indicate atomic
layers in the crystal. The law expressing the condition under which a crystal reflects
electromagnetic wave with the maximum amplitude and, at the same time, denoting the
angle at which the maximum reflection occurs. This reflection causes partial and
complete bandgaps in EBGs.

Fig.4.2. A diagram illustrating Braggs law

Equation 4.1 determines the centre frequency of the bandgap. The bandgap properties
are governed by the symmetry of the crystal, periodicity and the thickness of the EBG
slab. For example, the complete bandgap of an EBG consisting of metallic wires with
radius r and period a depends on both r and a. Fig. 4.3 shows the bandgap variation
d

(4.1)
120
with respect to r/a ratio for an infinite metallic wire EBG having square lattice for TE
polarisation.

Fig.4.3. Complete bandgaps for TE modes as functions of r/a for an EBG with square lattice
consisting of metallic wires (Reproduced from [11])

4.3 Metallic electromagnetic bandgap structures (MEBGs)
MEBGs are of great interest in the electromagnetics community due to its ease
of fabrication and clear cut-off at bandgap frequencies compared to dielectric bandgap
structures. MEBGs consisting of square lattices of metallic wires are selected for this
study. Throughout the following discussions, the radius of wires is denoted by r and
the period by a. A schematic diagram of the square lattice is given in figure 4.3.The
filling ratio of the EBG is the ratio of radius to period (r/a), which is the wire radius
normalised by the period.
121

Fig.4.3. Schematic diagram of square lattice of period a consisting of metallic wires of radius r

For an MEBG consisting of infinitely long thin wires, the bandgap depends mainly on
the filling ratio (r/a). For transverse magnetic(TM) polarisation, MEBGs will act as a
high pass filter having a low frequency cut-off [11,21] and for transverse electric (TE)
polarisation it does not posses a bandgap at lower modes of frequencies unless the
filling ratio r/a >0.2 [11]. This can be observed from fig.4.2. This difference in
performance for the two polarisations can be explained by considering the polarisation
of the cylinders induced by an external field. If the electric field is directed along the
wire axis i.e. TM polarisation, the cylinder will get easily polarised and this leads to
strong interaction of the cylinders in the structure. This interaction makes the array to
act like a metal surface. When an incident field with electric field vector ( E
v
) is parallel
to the wire axis incident on, it will set up currents along the wires just like on a metal
surface. This current cancels the incident field. This leads to complete reflection [22] of
the wave. When frequency increases the period is getting comparable to the wavelength
and the metal surface approximation starts deviating. Fig.4.4 illustrates this concept
clearly.
2r
a
a
122

Fig.4.4. A diagram showing a metallic wire array which acts like a metallic surface at low
frequencies (since the wavelength is large in this case, the spacing between the wires are very
small compared to the wavelength) and when frequency increases (wavelength decreases) the
spacing between the elements becomes comparable to the wavelength.

When the electric field is perpendicular to the wire axis, there will not be any
current flow since the interaction between the wires is small. This is because, for thin
wires the wire diameter is small (so that current along the radial direction is negligible)
hence there will not be any bandgap for TE polarisation, if the elements are not too
close (i.e. when r/a is small). The bandgap starts appearing when r/a becomes large. For
a given radius, this can be achieved by making the period a smaller and thereby
making the wires more closely. Thus the interaction between the wires will increase.
When the filling ratio greater than 0.2 there exists a bandgap at higher modes. When r/a
>0.3 there exists a narrow bandgap in the first few modes also and the width of the
bandgap increases with further increase in r/a ratio [11].
The bandgap in finite wire MEBG for TE polarisation can arise not only due to
the lattice resonance but also due to the self resonance of each wire [19, 21]. Recently,
it was revealed that the bandgap due to self resonance have the same characteristics as

Metallic wires
Incident field
E
v

123
that due to periodicity [19]. Negative refraction can be observed at frequencies near
these bandgaps. The negative refraction frequency shifts in wires of finite length,
loaded with active elements is given in [21]. Let l be the length of each wire in a finite
length wire MEBG. Then the first bandgap will occur at the first resonance frequency
of the wire. The resonating wavelength
r
is given by l
r
=
2
.
Negative refraction in a structure can be observed by verifying the inverse
Snells law. Two structures were considered for numerical studies. One was a slab and
the other was a prism structure consisting of metallic wires. Negative refraction at
multiple frequency bands including the passband is demonstrated in the subsequent
sections and experimental verification for a prism structure is presented.

4.3.1 Design of the MEBG slab
Firstly, a periodic structure consisting of metallic wires of radius of 0.63 cm and
height 1.26 cm was considered. A diagram of the structure used for the numerical
simulation along with the field vector directions are given in fig.4.5. The spacing
between the centres of wires was 3.15 cm. The filling ratio (r/a) was 0.2. There were
three wires in the x direction and six in the y direction. TM polarisation (electric field
vector parallel to the axis of the wires) was employed. The wave propagation was along
the x direction as shown in fig.4.5.
The lower cut-off plasmonic frequency of the MEBG when the wires are
infinitely long is given by the following equation [9].
124

Fig.4.5. A schematic diagram of the MEBG structure used for numerical simulation studies and
vector directions of electric field E, magnetic field H and propagation vector k

|
.
|
\
|

a
r
a
2
1 2
0

where a is the period, r

is the radius of the wire and
0
is the wavelength corresponding
to the lower cut-off frequency. Substituting the values equation 4.1 gives the
0
as
1.225cm corresponding to plasmonic frequency of 24.4 GHz. The period is 3.15cm and
therefore the lattice resonance is at = 2*3.15cm= 6.3cm corresponding to a frequency
of 4.76 GHz. Since the fundamental lattice resonance is lower than the plasmonic
resonance, the structure exhibits a pass band from 4.76 GHz. The second lattice
resonance will occur around 9.4 GHz. In the present model, the wires are finite with a
length of 1.26 cm. Hence, there arises a bandgap due to the self resonance of the wire
[19]. This self resonance corresponds to a wavelength of 1.26*2 = 2.52cm, which gives
a frequency of 11.9GHz. The passband of the MEBG starts from the lowest possible
resonance frequency. In the present case, it is due to the lattice resonance at 4.7 GHz.
So the MEBG is having a passband from 4.7 GHz.
z
y
E
v

k
v
H
v

x
1.26cm
7.56cm
17.01cm
(4.1)
125
4.3.2 Numerical simulation and discussion
Numerical simulations were carried out using Ansoft HFSS. In the simulation
model, the metallic wire structure was enclosed with an air box having absorbing
boundaries except the faces at which the ports or incident field excitations were
assigned. For S parameter simulation, wave ports were used as the source. For refracted
field study, incident field excitations had been employed. The angle of incidence was
assigned by providing direction vector values for E
v
, H
v
and k
v
and vectors in the
HFSS incident field options. The S parameters were obtained for the wave
propagation direction perpendicular to the wire axis. i.e. along the x axis. The
refraction occurred when the wave incidents obliquely at the interface and leaves the
EBG was studied. If the refracted field inside the structure is propagating towards the
negative z direction, it is considered as positive refraction.
The S parameters were obtained from frequency sweep and the dispersion
diagram was calculated from the S parameters. The simulated S(2, 1) response is given
in Fig.4.6. From the S(2,1) it can be seen that the bandgap is from 7.7-8.8GHz. This
bandgap is the second lattice resonance bandgap. The shift in frequency towards the
lower side (from 9.4 GHz to 8.5 GHz) can be explained by the fact that the effective
wavelength increases in the MEBG.
The dispersion diagram obtained is presented in fig.4.7. From the dispersion
diagram, the frequency regions where the structure supports backward radiation and
negative refraction (regions with phase and group velocities in opposite directions)
were determined based on the criteria presented in the previous chapter. In the band gap
region, the dispersion diagram shows some propagation. This can be explained by the
fact that the structure used was a small structure (with less elements or with less
number of periodicity), and hence the band gap is not deep. In theory, deep band gaps
can be obtained with more elements because when the number of elements increases, a
126
greater amount of reflection of wave occurs at the bandgap frequencies [18].

Fig.4.6: Transmission response of the wire periodic structure showing bandgap around 8.5 GHz

-4
-3
-2
-1
0
1
2
3
4
6 7 8 9 10 11 12 13
Frequency (GHz)
P
h
a
s
e

c
o
n
s
t
a
n
t

Fig.4.7: Dispersion diagram for the wire periodic structure with period 3.15 for a direction of
propagation along x axis

Band
Gap
127
From the dispersion diagram, there is negative phase velocity and positive
group velocity at 6.5 GHz. This indicates that there is negative refraction at this
frequency. At 7.5 GHz there is a positive phase velocity and a positive group velocity;
hence the structure supports positive refraction. In order to study the negative
refraction, an oblique incident field at one face of the model with an angle of incidence
45 was used. The magnetic field propagation in the structure was observed at 6.5GHz
and 7.5GHz. The simulated field plots at these frequencies are given in Fig.4.8. The
propagation directions of the field are denoted by arrows. From this field view, it can
be noticed that at 6.5 GHz, there is bending of the magnetic field propagation direction
towards the negative side. This wave bending is in the opposite direction i.e. towards
the positive side at 7.5 GHz.

Fig.4.8: A plot indicating magnetic field propagation directions inside and outside the periodic
structure at 7.5 GHz(left) and 6.5 GHz (right). The propagation directions in each media are
indicated by arrows. Surface normal is shown in dotted lines.

From fig.4.8 it can be seen that the incident field at an angle 45 reach upon the
structure and it gets refracted at the air and wire medium interface. When it comes out
128
of the structure again refracted at the wire medium and air interface. At 7.5GHz, it
refracted in such a way that the structure supports positive refraction. At 6.5GHz, the
structure supports negative refraction with an angle of -41 as observed in fig.4.8. From
the transmission response in fig. 4.6, it can be seen that 6.5 GHz is in the passband, 1.2
GHz below the bandgap. In terms of normalised frequency (a / 2c), the bandgap is at
0.81 and the observed negative refraction is at 0.68. If this occurred due to the all angle
negative refraction, it should be very close to 0.81. The above results clearly show the
existence of negative refraction from the MEBG in the passband.

4.3.3 Prism structure for Inverse Snells Law verification
The studies in the previous section showed negative refraction from a slab
shaped structure for oblique incidence. In the following section, the Inverse Snells
Law verification at frequencies in the passband, in addition to the bandgap edges from
a prism shaped EBG formed by periodic array of metallic wires is detailed.

4.3.3.1 Design of the MEBG prism
A prism structure made with square lattices consisting of metallic cylinders with
length 12 cm and the radius 0.63 cm was selected. A diagram of the prism is shown in
fig. 4.9. The period was 1.75cm and the filling ratio was 0.36. In this prism structure
there were 10 elements both on base and vertical sides which give minimum side
dimension of 17.01 cm corresponding to 2.68 at 4.3GHz.

129

Fig.4.9.Diagram of the prism consisting of metallic wires with period 1.75cm.

The plasmonic cut-off frequency for the EBG structure for TM polarisation is given by
eqn.4.1 which gives a frequency of around 13GHz. Since the lattice resonance
frequency (corresponds to wavelength = 2a = 3.5cm) is at 9GHz which is far below
the plasma resonance frequency, the EBG acts as a high pass filter for TM polarisation
having cut-off frequency of 9GHz.
For TE polarisation the structure acts as a dielectric EBG and since r/a > 0.3 the
first complete bandgap occurs within the first and second modes [11]. The simulation
model for the structure used for the calculation of bandgap and wave propagation
directions , X and M is shown in fig 4.10. An eigen value solution model was selected
with periodic boundary conditions. TM and TE polarisations can be achieved by
changing the top and bottom walls as Electric (i.e. E
tan
= 0) and Magnetic (i.e. H
tan
=0)
respectively. The bandgap diagram was obtained from numerical simulation using CST
Microwave Studio. Fig.4.11 illustrates the CST simulation model used for obtaining the
bandgap diagram.

130

Fig.4.10. Unit cell assignment in CST simulation model for the simulation of bandgap diagram
for EBGs consisting of square lattice and wave propagation directions , X and M

In a bandgap diagram eigen frequencies are plotted against wave number which
is equivalent to eigen solutions against boundary condition phases [9]. From fig. 4.10 it
can be seen that there are three directional sweeps for the model. They are to X, X to
M and to M. So the dispersion diagram consists of these three regions. The dispersion
curve for each region was produced separately with two phase shifts. The phase shift
assignments for the parameter sweep function through the three steps are given in
table.4.1. The three sweeps were combined to get the complete bandgap diagram. In
step 1, phase shift 1 is varied from 0-180, keeping phase shift 2 as zero. This is
equivalent to the wave vector sweep from 0 to 180 degrees in the to X domain. In
step 2, phase shift 1 is kept at 180 and phase shift 2 is varied from 0-180. In step 3, both
phases 1 and 2 are swept from 180-0. Then step 3 is inverted because the required
phase shift is 0-180. The dispersion diagram is obtained by combining the three steps.
The bandgap diagram of the structure for TE polarisation obtained from the CST
simulation is shown in fig.4.12.
X
M
Phase shift 1
P
h
a
s
e

s
h
i
f
t

2

131
From this figure, it can be seen that the complete bandgap is from 9.3- 10.2 GHz as
shaded in green.

Fig.4.11.CST Simulation model for bandgap determination. Four side walls are defined as
periodic boundaries. The top and bottom walls are defined as magnetic/electric boundaries
depending on the polarisation.

Table.4.1. Three steps for obtaining the complete bandgap diagram and their corresponding phase
shifts for the parameter sweep.

Step 1 : X

Step 2: X M

Step 3: M

Phase Shift 1: 0-180

Phase Shift 1 : 180


Phase Shift 2: 0


Magnetic/Electric
Periodic
Periodic
Periodic
Periodic
132

Fig.4.12. Bandgap diagram for the metallic wire structure with r/a = 0.36 (TE polarisation).
Green region indicates complete bandgap

The prism structure used was with less number of elements and the number of
elements was not same throughout the wave propagation domain because the number
of elements is decreasing from base to top. Unsurprisingly, a clear and wide bandgap
can not exist except for an infinite number wire prism. The S parameters were obtained
in the incident field direction perpendicular to one side of the prism. The transmitted
fields were measured at different frequencies using two WR-187 waveguides; one for
transmission and one for reception. The waveguides have a lower cut-off frequency of
3.5 GHz. The measured transmission response S (2, 1)) obtained for the prism structure
is shown in fig.4.13. From the figure it can be seen that the bandgap is from 5.6 to 6.1
GHz.
0
2
4
6
8
10
12
14
16
F
r
e
q
u
e
n
c
y

G
H
z
X M

133

Fig.4.13.Measured transmission response for the prism structure showing bandgap around
5.7GHz. The bandgap region is marked in blue.

The dispersion diagram was obtained for incident field direction perpendicular
to the wire axis, from the simulated S parameters. Dispersion diagram in fig.4.14
indicates the bandgap region around 5.54 to 6.5 GHz, which agrees with the measured
transmission response. The measured transmission shows bandgap from 5.6 - 6.1 GHz.
The difference between the dispersion diagram in fig.4.12 and 4.14 is that the first one
predicts the complete bandgap (bandgap for all incident field directions) and the second
figure gives the bandgap for the incident field direction perpendicular to the wire axis.
It can be noticed that the bandgap width is narrowed in the finite prism. The backward
radiation frequency regions were identified from the dispersion diagram as in the
previous case. For studying the refraction properties of the structure, using numerical
simulation, a rectangular dielectric slab waveguide with opening dimensions of 5 mm
was used for the excitation. The dielectric constant of the slab was 40. The design
equation for the cut-off frequency less than 10 GHz this slab waveguide is given in by
the equation [35]

134
1 4
10 10
9

=
r
c
t
f

Where t is the side dimension of the slab in centimetres. This gives the lower cut-off
frequency of 0.8 GHz.
-4
-3
-2
-1
0
1
2
3
4
4 5 6 7 8
Frequency GHz
d

Fig.4.14. Dispersion Diagram for the prism periodic structure with metallic wires having a
period of 1.75 cm. The bandgap region is marked in blue.

From the dispersion diagram it was observed that there are multiple regions that
can yield negative refraction. These regions were 4.0 - 4.5 GHz and 7.2 - 8.0 GHz. It
can be seen from the dispersion diagram that positive refraction occurs at 4.7 GHz and
negative refraction at 7.4 GHz. Simulated refracted field obtained at these frequencies
verified this. The simulated field plots are given in figs.4.15 and 4.16. It is clear from
the field plots that the wave refracted at an angle of +37 at 4.7 GHz and -45 at 7.4
GHz. The wave propagation directions are indicated in dark blue arrows and the
surface normal is denoted by the dotted line. Different scales were used at the two
frequencies for better view of the refraction phenomenon.

Bandgap

P
h
a
s
e

c
o
n
s
t
a
n
t

(4.2)
135
The animation of phase obtained from simulation also supported the refraction at
positive and negative angles at the two frequencies.

Fig.4.15. Electric field plot showing positive refraction at 4.7 GHz. Units are in V/m

Fig.4.16. Electric field plot showing negative refraction at 7.4 GHz. Units are in V/m

136
4.3.3.2 Measurement verification of refraction phenomenon
The effects of Floquets harmonics and spatial dispersion on refraction
properties of EBG like structures and negative refraction phenomenon from prism
periodic structures were studied in the previous sections and the theoretical predictions
were verified using numerical simulations. In this section, the experimental work
carried out for the verification of simulation results for the prism structure with period
1.75cm is detailed. The main goal of the experiments was to verify the existence of
positive and negative refraction from prism structure at various frequencies.
Measurement results have shown good agreement with simulation results. Based on the
available literature, so far there has not been any extensive experimental verification on
the role of Floquets harmonics and spatial dispersion in the study of LHMs or
analogous LHMs.
From the numerical simulations discussed in the previous sections, it has been
demonstrated that negative refraction exists at narrow bands of frequencies in the
passband. A diagram of the experimental setup used is shown in fig.4.17.

Fig.4.17. A diagram showing the setup used for experimental studies.
Rx
Tx
Rotation path
30cm 30cm 30cm
Waveguide
Waveguide
Absorber Absorber
137
Two models were selected for the experimental study. One was a Teflon prism with
known parameters (
r
= 2.1 and = 1.0), having base dimensions of 7.4 cm, 10.5cm
and height of 30 cm and the other was the wire prism under study. For the excitation, a
WR-187 waveguide with 30cm length and opening dimensions of 4.68cm and 2.1 cm
was used. A waveguide of same dimensions as that at the transmitting side was used for
reception. A photograph of the two prism models used for the experimental studies is
given in fig.4.18. The waveguides and the equipments were surrounded by absorbers to
prevent any reflected wave reaching into the structure under study or to the receiving
waveguide.

Fig.4.18. Two prism structures used for measurement: Prism structure consists of metallic
wires (left) and Teflon prism (right)

In order to substantiate the experimental setup, the refracted field from a Teflon
prism was measured and positive refraction was identified. Figure 4.19 is the simulated
electric field plot for the Teflon prism showing positive refraction at an angle of +20.
For the measurement of refraction, the structure under study was placed in front of the
transmitting waveguide and then rotated the transmitting setup (equivalent to receiving
refracted power at different angles). First, the transmitted field was measured with the
138
waveguides at a distance of 30cm. The receiver was rotated from +90 degrees to -90
degrees and the electric field was measured at around 1200 steps (positions) at each
frequency. The received field thus obtained without any structure in between the
waveguides is given in fig.4.20. The figure shows a peak when the receiver is at the
boresight position i.e. at an angle of zero degrees. The negative angle side corresponds
to negative refraction and positive angle side corresponds to positive refraction. The
measured field with the Teflon prism is given in fig.4.21 which shows an angle of
refraction of +23.

Fig.4.19. Simulated electric field showing positive refraction from Teflon prism at 4.7GHz.
The angle of refraction is +20.

139
0
0.2
0.4
0.6
0.8
1
1.2
-90 -75 -60 -45 -30 -15 0 15 30 45 60 75 90
Angle (Deg.)
N
o
r
m
a
l
i
s
e
d

P
o
w
e
r

Fig.4.20.Measured pattern with flanges only at 7.4 GHz showing the peak amplitude at
boresight. The half power beam width is 34.

Fig.4.21.Measured refracted field for Teflon prism at 4.7 GHz. The angle of refraction is +23

Next, the angle of refraction was measured by replacing the Teflon prism with
the EBG. The metallic wires were fixed on two foam sheets of 1cm thickness as shown
in fig.4.22. There were 8 wires in base and vertical sides of the prism. This gives a side
dimension of 2.19 at 4.3GHz. The waveguide opening was kept very close (4mm
140
from the first air wire interface) to the structure to avoid field spreading at the incident
face. The two waveguides were at a distance of 30cm. At frequencies corresponding to
positive refraction in the dispersion diagram, the refracted wave will bend towards the
negative side. At positive refraction frequencies, it will bend towards positive side. The
refracted field was measured at different frequencies corresponding to positive and
negative refraction. The measured refracted field intensity at different frequencies is
presented in fig.4.22. It was found that at 4.7 GHz, the structure exhibited positive
refraction with a refraction angle of +34. From simulated field plot, it can be seen that
there is a positive refraction at +37. At 7.4 GHz the measured field showed negative
refraction with an angle of -44, where -45 was observed in numerical simulation.
0
0.2
0.4
0.6
0.8
1
1.2
-90 -75 -60 -45 -30 -15 0 15 30 45 60 75 90
Angle (Degree)
N
o
r
m
a
l
i
s
e
d

P
o
w
e
r
7.4 GHz
5.6 GHz
4.7 GHz
7.2 GHz

Fig.4.22: Measured refracted field for wire prism showing positive refraction at 4.7 GHz and
negative refraction at 5.6 GHz, 7.2 GHz and 7.4 GHz

The experiment was repeated for other frequency points where the dispersion diagram
predicted negative refraction (a potential 2% bandwidth around 7GHz) and found
dominant negative refraction phenomena. The measured field at 7.2GHz also gave
negative refraction with an angle of -64 (fig.4.22). Such a phenomenon has also
141
measured at 5.6 GHz which was at the lower edge of bandgap, shown negative
refraction with an angle of -19 as depicted in fig.4.22. The measurement results
indicated that there was a beam narrowing at negative refraction frequencies.
Results obtained from measurement at 7.2 GHz were compared with similar
results previously reported on Inverse Snells Law verification [6-8] using different
prism structures. Table.4.2 presents the comparison of measurement result at 7.2 GHz
with similar works using prisms consisting of metallic wires [18], split rings and wires
[28.29]. The parameters compared were the main beam refraction angle, 3dB beam
width in each case at negative refraction frequency and the strength of negatively
refracted beam by the positively refracted one. It was noticed from previously reported
results, that even though there was dominant negative refraction there associated a
positive refracted component. This can be explained by the fact that there always some
Floquets harmonics radiating towards the positive side [24-26]. In order to quantify the
amplitude of negative refracted field compared to that of positive refracted field, a
parameter Negative Refraction Excess (NRE) was introduced. NRE is defined as

% 100
=
n
p n
A
A A
NRE

Where A
n
and A
p
are the maximum amplitudes of negative and positive refracted beams
respectively.

(4.3)
142

Structure
under study
Negative
refraction
angle
(Deg.)
3dB Beam width
at NR
frequency(Deg.)

NRE

/d

Metallic wires
[18]

-35

15

70

1.05

SRR+ Wire
[28]

-61

33

81

6

SRR+Wire
[29]

-34

17

72

7.2
Metallic wires
[This work]

-64

17

85

2.3

Table: 4.2.Comparison of measurement results at 7.2GHz with similar work for verification of
inverse Snells law using different structures.

In [18], NRE was reported 70% from an EBG consisting of metallic wires,
while in [27], [28], it was 81% and 72 % for a structure of Split Ring Resonators
(SRRs) and wires respectively. In this measurement, an NRE of more than 80% (85%
at 7.2 GHz and 81% at 7.4 GHz) is achieved. This can be explained by the fact that the
measurement was performed in the pass band with less wave attenuation compared
with those achieved at the lower edge of bandgap. At 5.6 GHz NRE was 69%. The fifth
column in table.1 indicating the effective medium applicability on the basis of
operating wavelength () and the period (d) of the model used in each case. In the
present study, a /d ratio of 2.3 was used. This is comparable to the models which were
used previously for negative refraction measurement.

4.3.3.3 Inclusion of defects for enhancing negative refraction
According to the dispersion diagram in fig.4.14 there can be negative refraction
from 4.0 GHz to 4.5 GHz. However, from the measurement we observed that the
negative refraction phenomenon is not significant, for example at 4.3 GHz. In this case
143
the positive refracted component was strong by 42% then the negatively refracted
component. This can be accommodated by the fact that this frequency point is very
close to the positive refraction frequency band.
In [30] a motivating idea was presented to enhance anomalous refractions at
its bandgap of a periodic structure by introducing some defects into the structure. In
[31] the defect states in a dielectric periodic structure have been detailed. From the
dispersion diagram presented for a square array of GaAs dielectric rods with defects, it
was observed that the introduction of defects changes the refraction properties of the
band dispersion diagram even in the passband. In the conventional treatment of EBGs
using EFS diagrams, when we introduce anisotropy in the crystal the EFS shape can be
changed and one can obtain a case where the EFS circles become ellipsoids [32]. The
reason for this change is that the introduction of anisotropy causes a phase
displacement. The radiation patterns of a dipole surrounded by an EBG having different
line defects introduced on it to enhance the radiation pattern in certain directions is
experimentally studied in [33]. Fig.4.23 shows the two switching mechanisms used and
the corresponding radiation patterns. From the figure, it can be seen that the two
switches are mechanically controlled to turn it on or off. On or off position indicates
the removal or introduction of the rods. For example, ON-ON indicates that both top
and bottom defect lines are active. The radiation pattern is changed due to the removal
or introduction of the defect lines.

144

Fig.4.23. Diagram of radiating dipole inside an EBG with defects for changing the radiation
pattern and corresponding radiation patterns. The blue dots indicate the defect positions.
(Reproduced from [32] )

From the discussion in chapter 3 (section. 3.10), it was found that a phase
altering mechanism by introducing pseudo periodicity in the structure one could
suppress some of the Floquets harmonic components. In a prism structure like the one
under study this suppression can lead enhanced refracted field in one direction. A
similar approach was used to restore the negative refraction phenomena at the pass
bands at which both negative and positive refractions are equally significant. Four
wires from the prism structure were removed to enhance the negatively refracted beam.
Fig.4.24 shows that four metallic rods are missing from the prism, creating defects in
the prism structure. It is verified from the HFSS simulation (fig. 4.24) that the
introduction of defects indeed enhances the negative refraction phenomena in the way
of suppressing unwanted spatial harmonics.

145

Fig.4.24. Simulated electric field at 4.3 GHz on prism with defects showing negative refraction

This has been verified by measurement. Fig.4.25 shows that introduction of the defects
can lead to an increase of 27% in the negative refraction. This can be explained by the
frequency shift in the dispersion diagram of the prism structure with the defects. The
dispersion diagrams for both cases for an incident field direction normal to the wire
axis for TE polarisation is given in fig.4.26. From the diagrams it can be seen that
there is a frequency shift and that the frequency point of interest (4.3GHz) has shifted
more towards the negative phase velocity side.

146

Fig.4.25. Measured refracted field from wire prism structure showing positive refraction at 4.3
GHz and negative refraction at 4.3 GHz after the introduction of the defects into the structure.

-4
-3
-2
-1
0
1
2
3
4
4 5 6 7 8
Frequency (GHz)
d
with defects
without defects

Fig.4.26. Dispersion diagrams for the wire structure without the defects and with the defects
has been introduced

P
h
a
s
e

c
o
n
s
t
a
n
t

147
In [30], negative refraction at bandgap frequencies was obtained by the
introduction of defects. In the present work this was obtained in the passband. This is
the first realisation of negative refraction in the passband using defects. The change in
the dispersion diagram supports the enhancement of negative refraction by introducing
defects. In a periodic structure, the removal of the wires can be done in many ways. In
order to get an optimised performance, one needs to use optimisation routines in which
the removal of various wires with different degrees of freedom can be studied to predict
the phenomenon for better performance.
For a normal prism of a natural material the frequency dispersion is very small.
The transmitted power in the previously reported works based on split ring and wire
medium was found to be very low (less than 20%) [28, 16]. In the present work, the
measured transmission from the prism structure was 54% at 7.2GHz. This has shown
that the transmitted power obtained from the EBG prism is more than twice the
transmission from split ring and wire medium.
A comparison was carried out on the frequency regions of interest with the
negative refraction reported in [15] (fig. 2.22, chapter.2). In terms of normalised
frequency (a/2c), the negative refraction observed in this work was at 0.25 (at 4.3
GHz), 0.32 (at 5.6GHz) and 0.42 (at 7.2GHz). The complete bandgap for the EBG was
at 0.548. In [15] the bandgap was at 0.25 and the all angle negative refraction was
obtained at 0.19 (i.e. 0.06 below the bandgap). This indicated that the observed
negative refraction phenomenon in the present study was far below the complete
bandgap. If the observed phenomenon was due to all angle negative refraction proposed
in [15], it should be very close to the complete bandgap at 0.548.

148
4.4 Conclusions
In this chapter, negative refraction phenomenon from metallic EBGs has been
presented. It was found that negative refraction can be obtained from the EBG at
multiple frequency bands even though it does not posses a negative effective refractive
index. It has been identified that strong spatial dispersion in the structure due to high
order Floquets harmonic interactions, incident field coupling and evanescent modes
can give rise to negative refraction at certain frequencies. These frequencies are not
very close to complete bandgap. In conventional analysis of infinite EBGs, such a
phenomenon was obtained near the complete bandgap edges only. This limited the
transmitted power. In this work, transmitted power of 54% was obtained which was
much higher than that from a split rings and wires medium. A negatively refracted
beam which was 85% stronger than the positively refracted beam at 7.2 GHz was
obtained. This was higher than that of the previously reported works. To enhance the
negative refraction, a method is proposed by applying defects into the finite structure
and negative refraction enhancement at certain frequencies is verified. An enhancement
of 27% for the negatively refracted beam is obtained. Numerical simulations results
were verified with measurement results and both were in very good agreement. The
dispersion obtained from the MEBG can lead to novel applications in beam scanning
and spatial frequency demultiplexing. It is very difficult to obtain such type of
dispersion from natural materials at microwave frequencies.

149
References
[1] E. Yablonovitch, T. J. Gmitter, and K. M. Leung, Photonic band structure:
The face-centered-cubic case employing nonspherical atoms, Physical Review
Letters, Vol. 67, pp. 2295-2298, 1998
[2] J. D. Joannopoulos, R. D. Meade, and J. N. Winn, Photonic Crystals: Molding
the Flow of Light , Princeton University Press, Princeton, 1995
[3] Chonguin Jin, Bingying Cheng, Baoyuan Man, Zhaolin Li, Daozhong Zhang,
Shouzheng Ban, Bo Sun, Bandgap and wave guiding effect in a quasiperiodic
photonic crystal, Applied Physics Letters, Vol. 75, No.13, pp. 1848-1850, 1999
[4] S. L McCall , P.M. Platzman, R. Dalichaouch, David Smith, S. Schultz,
Microwave propagation in two- dimensional dielectric lattices, Physical Review
Letters, Vol.67, pp. 2017-2020
[5] E. Ozbay, E. Michel, G. Tuttle, R. Biswas, M. Sigalas and K. M Ho, Micro
machined millimetre-wave photonic band-gap crystals, Applied Physics Letters,
Vol. 64, pp. 2059-2061, April 1994
[6] Steven G. Johnson, Shanhui Fan, Pierre R. Villeneuve and J. D Joannopoulos,
L. A Kolodziejski, Guided modes in photonic crystals, Physical review B, Vol.
60, pp. 5751-5758,
[7] C. A. Moses, N. Engheta, Electromagnetic wave propagation in wire medium: a
complex medium with long thin inclusions, Wave Motion, Vol. 34, pp. 301-317,
March 2001
[8] Philippe gay-Balmaz, Claudio Maccio, Oliver J. F Martin, Microwire arrays
with plasmonic response at microwave frequencies, Applied Physics Letters, Vol.
81, No.15
[9] S. I. Maslovski, S. A. Tretyakov, P. A. Belov, Wire Media with Negative
Effective Permittivity: A Quasistatic Model, Microwave and Optical Technology
150
Letters, Vol.35, pp. 47-51, 2002
[10] Manvir S. Kushwaha, Stop-bands for periodic metallic rods: Sculptures that can
filter the noise, Applied Physics Letters, Vol.70, pp. 3218-3220, 1997
[11] E. I. Smirnova, C. Chen, M. A. Shapiro, J. R. Sirigiri, and R. J. Temkin,
Simulation of Photonic bandgaps in metal rod lattices for microwave
applications, Journal of Applied Physics, Vol. 91, pp. 960-968, 2002
[12] M. Bayinder, E. Cubukcu, I. Bulu, E. Ozbay, Photonic Bandgaps and
localisation in two dimensional metallic quasi crystals, Europhysics letters, Vol.
56, pp. 41-46,2001
[13] Dongmin Wu, Nicholas Fang, Cheng Sun and Xiang Zhang, Willie J. Padilla,
Dimitri N. Basov, David R. Smith and Sheldon Schultz, Terahertz plasmonic
highpass filter, Applied Physics Letters, Vol.83, pp. 201-203, 2003
without negative index in metallic photonic crystals , Optics Express, Vol. 11,
pp. 746-754, 2003
refraction without negative effective index, Physical Review B, Vol.65, 201104,
2003
superlens, Physical Review Letters, Vol. 91, pp. 207401, 2002
Costas M.Soukoulis, Negative refraction by photonic crystals, Nature, Vol.
423, pp. 604-605, 2003
[18] P. V Parimi, W. T. Lu, P. Vodo, J. Sokoloff, S. Sridhar, Negative refraction
and left- handed electromagnetism in microwave photonic crystals, cond-
151
mat/0306109, 2003
[19] P. A. Belov, Flat lenses formed by capacitively loaded wire media, Proc.
Progress In Electromagnetics Research (PIERS), August, 2005
[20] Tom Duncan, Advanced Physics, Fourth edition, pp.435-436, 1987
[21] P. Belov, C. Simovski, P. Ikonen, "Canalization of sub-wavelength images by
electromagnetic crystals", Physical Review B, vol. 71, pp. 193105, 2005.
[22] A-C. Tarot, S. Collardey, and K. Mahdjoubi, Numerical Studies of Metallic
PBG Structures, Progress In Electromagnetics Research, PIER , Vol. 41,pp.
133157, 2003
[23] http://www.st-andrews.ac.uk/~www_pa/Scots_Guide/mscct/part5/page3.html
[24] CST Training Material, Two Dimensional Photonic Crystals.
1986
[27] R.E Collin, Foundations for Microwave Engineering, Sec 8.9, McGraw-Hill
Kogakusha Ltd, 1966
[28] R. A .Shelby, D. R. Smith, S. Schultz, Experimental Verification of a Negative
Index of Refraction, SCIENCE, Vol.292,pp. 77-79, April 2001
Experimental verification and simulation of negative index of refraction using
Snells Law, Physical Review .Letters, Vol.90, pp.107401-1-107401-4, 2003
[30] Soumia Massaoudi, Frederique Gadot, Eric Akmansoy, Andre Lustrac,
Amplification of the Anomalous refraction in a PBG-Prism, Proc. of 27th ESA
Antenna Technology Workshop on Innovative Periodic Antennas, Spain ,March
2004, pp. 506-5-08
152
[31] J. Timothy Londergan, John P. Carini, David P. Murdock, Binding and
Scattering in Two-Dimensional Systems, Springer, pp. 190-192, 1999
[32] Mayasa Notomi, Negative refraction in photonic crystals, Optical and
Quantum Electronics, Vol.34, pp. 133-143, 2002
[33] Kourosh Mahdjoubi, Directive EBG Antennas, Presentation slides
[34] Martina Gerken and David A. B. Miller, Multilayer thin-film structures with
high spatial dispersion, Applied Optics, Vol. 42, 1330-1345, March 2003
[35] C. A. Balanis, Advanced Engineering Electromagnetics, John Willey, 1989

153
Chapter 5

Wave Focusing at Low Microwave
Frequencies Using Metallic Wire Medium

5.1 Introduction
Pendry [1] proposed the possibility of obtaining a perfect lens using a slab of
LHM. It was claimed that such a lens can lead to novel applications in medical imaging
and digital storage media with extremely high capacity due to its unique characteristic
of aberration-free image. Ordinary lenses are of curved shape, which suffer from
spherical aberrations, and hence low image resolution. A perfect lens formed by an
LHM enables focusing of electromagnetic waves to a size much smaller than the
wavelength [2-5]. This lens is free from all aberrations which limit the image quality
from conventional lens. Conventional lenses can not focus waves to a size significantly
sharper than /2 [2]. Fig.5.1 shows the focusing mechanism in a slab with negative
refractive index n = -1. Let the distance of the point source from the slab be l and the
thickness of the slab be 2l. In fig.5.1, A is a point source, B is the first focus formed
inside the slab and C is second focus formed outside the slab. Since the refractive index
is -1 the first focus B is formed at the centre of the slab.

Chapter 5: Wave Focusing at Low Microwave Frequencies Using Metallic Wire Medium
154

Fig.5.1.A Ray diagram showing wave focusing in a slab with negative refractive index. n = -1.
A is the source point, B and C are first and second focus points respectively.

After the pioneering proposal by Pendry [1], there has been growing curiosity in
realising such a lens. Due to the loss of all materials, obtaining a perfect lens is a
difficult task. However, getting a super lens is a more practical solution. There has
been a lot of work in studying focusing using an LHM slab by numerical simulations
and experiments. Fig.5.2 shows the numerical simulation result showing wave focusing
from a slab of refractive index -1. The simulation was carried out using HFSS. From
the figure, it can be seen that the source point at the right side is getting focused in the
middle of the slab and the second focus point is obtained outside of the slab on the left
side.
Recently, a medium consisting of resonant inclusions of conducting wires and
spirals was selected for LHM performance and focusing studies [4]. In this medium, the
spiral arrays contribute negative permeability and the wires contribute negative
permittivity [4]. One dimensional measurement of electric field was carried out using
two identical dipoles for transmission and reception. Since this model is based on EMT
approximations, the bandwidth is very narrow.

n = -1
A B C
l
l
2l
155

Fig.5.2. Simulated electric field plot showing focusing from a slab of refractive index n = -1

Wave focusing using negative refraction properties of EBGs was widely studied
in [3, 5-7]. Sub-wavelength focusing at microwave frequencies using an EBG
consisting of a square array of dielectric rods of dielectric constant = 9.61, the length
of rods l = 15cm and the radius r = 11.75mm was presented in [3]. The period of the
EBG was 4.79mm. TM mode polarisation was employed for the study and the wave
focusing was observed at a very narrow band with a centre frequency 14.27 GHz.
There were 21 layers in lateral direction and 15 layers in the propagation direction. The
source was at 0.7mm away from the surface of the EBG and an image was obtained at
0.7 mm on the other side. Figure 5.3 shows the results for focusing studies detailed in
[3]. It represents one dimensional power distribution showing focusing from the slab.
In [7], wave focusing from a two dimensional EBG slab consisting of cylindrical
alumina rods was detailed. The results are presented in chapter 2 (figs. 2.23 and 2.24).
The frequency at which the focusing observed was 9.3GHz.

156

Fig.5.3. A graphical representation simulated and measured average power intensity
distribution at the image plane with and without the EBG. The red curve shows the FDTD
simulation result and the blue line represents the experimental result. The green dotted curve
shows the field intensity without the EBG. It can be seen that there is strong field intensity at
the centre corresponding to the image point. (Reproduced from [3])

From the ray diagram in fig. 5.1 and the numerical simulation result of electric
field intensity in fig. 5.2, it can be seen that in addition to the focus point outside the
left handed slab, the refracted ray converges to another point inside the slab. This
indicates that objects in the left hand space produce a real image in the right hand space
[8]. Imaging using EBGs is close to the imaging by a mirror. However, in a mirror the
image is virtual one. The resolution of image obtained using an EBG depends on the
period. The maximum resolution obtained can not exceed the lattice spacing [8].
The experimental work detailed above considers only one dimensional
distribution of the field. This will not fully guarantee the exact resolution of the image
that is important in finite EBGs. There is also an issue that of the frequency at which
imaging was obtained was at higher frequencies. In [3], the frequency of focusing is
around 14GHz and in [7] it was at 9.3 GHz.
157
Next sections of this chapter will address the focusing from a finite MEBG at
low microwave frequencies. The design details of the EBG slab will be followed by
numerical simulations using HFSS and experimental verification. It is shown that the
phenomenon is due to the metamaterial nature of the slab other than any diffraction or
array antenna effect. This measurement was the first extensive measurement of wave
focusing from an MEBG along two directions; lateral and longitudinal to determine the
exact size and resolution of the focus, obtained from a finite EBG. This work is the first
realisation of focusing using metallic EBGs.

5.2 Design and numerical simulation
From the previous chapter, the complete bandgap frequency region was found
to be 9.3 GHz-10.2 GHz, for the MEBG consisting of metallic wires with square
lattice. The wires were having radius 0.63cm and length 12cm. The radius to period
ratio (r/a) was 0.36. The proposed slab was having a height of 13.4cm, thickness
6.51cm and width 12 cm (length of wire). It was formed with 4 wires in z direction and
8 in y direction (a total of 32 wires). Fig.5.4 shows the schematic diagram of the
structure and the setup used for experimental studies.
158

Fig.5.4. Schematic diagrams of the structure used for focusing studies (left) and the setup used
for the experiment. There are four elements in z direction and 8 elements in y direction in the
wire structure.

The dispersion diagram of the proposed slab obtained for a direction of incidence
perpendicular to the wire axis (the wave propagation is along z axis) with electric field
along the y axis is given in fig.5.5. This diagram gives the phase velocity versus
frequencies for a certain incident angle (perpendicular to the axis of the wire). In both
figs. 5.4 and 5.5, the incident wave propagation is along the z axis.
The difference between the figure 4.16 in chapter 4 and figure 5.5 is that the
first represents the eigen frequencies for different wave vector directions inside the
infinite EBG and the latter gives the phase constant over the structure for different
frequencies for a specific angle of incidence for the finite EBG. The latter diagram
shows a bandgap at 5.5 GHz and this bandgap exists for the incident wave direction
perpendicular to wires. The dispersion diagram of the structure gives negative phase
velocity and positive group velocity at some frequency regions. Phase velocity has the
159
same sign as phase constant and the slope of the dispersion curve gives the group
velocity.

-4
-3
-2
-1
0
1
2
3
4
1 2 3 4 5 6 7 8 9
Frequency GHz
P
h
a
s
e

c
o
n
s
t
a
n
t

NR NR
Bandgap

Fig.5.5.Dispersion diagram of the MEBG slab obtained for incident field perpendicular to the
wire axis (along z axis). The negative refraction regions and bandgap regions are marked. The
bandgap starts from 5.5GHz.

From the dispersion diagram in fig.5.5 it can be observed that there can be multiple
frequency bands with negative phase velocity and positive group velocity (e.g. 2.0-2.6
GHz and 4.0-5.0 GHz). Since at these frequencies the medium can act as an analogous
LHM, it can give rise to wave focusing.
Numerical simulation was performed using Ansoft HFSS to verify the
focusing phenomenon from the proposed slab given in fig.5.4. A wave port with a
small opening of 5mm at the plane containing the origin of coordinates at and a
distance of 2.4 cm away from the air-wire interface was used as the source with all
other boundaries as absorbing walls. Simulated electric field intensity along the yz
plane at 2.1 GHz is given in fig.5.7. The focus point is marked in the rectangular box.
At 3.4GHz was a positive refraction frequency point at which wave focusing was not
observed. Fig. 5.8 shows the simulated electric field at 3.4 GHz. It can be seen that
160
there is no focusing of waves and there is a low intensity point at the right hand side of
the slab.
It is shown that at 3.4 GHz the wave is spreading at the right side of the slab and
at the position corresponding to focus point there is no wave focusing while at negative
refraction frequency 2.1 GHz there is a high intensity point marked in a small
rectangular box in the field plot in fig.5.7. This high intensity point is at a distance of
11.31 cm from the source (i.e. from the origin of coordinates). The electric field
intensity along the xy plane at various positions along the z axis was taken. This
gradual movement of electric field plotting plane shows the wave is getting focused at
the middle of the slab, when it travels through the slab and then gets diverged. The
electric field intensity at 2.4 cm (at the left air slab interface), 3.4cm and 7.0cm from
the source are depicted in fig.5.9, 5.10 and 5.11 respectively.

Fig.5.7.Simulated electric field intensity along the yz plane at 2.1GHz showing focusing. The
focus point is marked in the rectangular box
161

Fig.5.8. Simulated electric field intensity along the yz plane at 3.4GHz showing no focusing.

This shows the progressive movement of the field along the wire slab. From these field
plots, it can be noticed that the wave propagates inside the model in a way shown in the
ray diagram in fig.5.1. So the focus obtained at outside the slab is the second focus
point.

Fig.5.9. Simulated electric field intensity along xy plane at a distance of 2.4cm from the source

162

Fig.5.10. Simulated electric field intensity along xy plane at a distance of 3.4cm from the
source

Fig.5.11.Simulated electric field intensity along xy plane at a distance of 7cm from the source

163
5.3 Measurement of focusing from an MEBG slab
The experiment was performed with a slab structure consisting of metallic wires as
used in the simulation. A monopole with resonant length 1cm was used for the
excitation. The antenna radiates almost uniformly in almost all directions except the
backfire. A probe with the same dimensions was used for receiving the transmitted
power. Figure 5.12 shows the schematic diagram and a picture of the monopole probe
used for the measurement.

Fig.5.12.A schematic diagram of the probe used (left) and a picture of the coaxial fed probe
used in the experiment for transmission and reception (right)

The source was placed at the origin which is 2.4 cm away from the air wire
interface and the field intensity was measured in the xy plane using a near field NSI
scanner. The receiving probe was fitted to the arm of the scanner which can move in x
and y directions. Fig.5.4 shows the schematic diagram of the measurement setup used
for the experiment. Calibration was done to eliminate the effects of surroundings. The
four sides of the slab were covered by absorbers. Electric field was scanned at points
with steps of 5 mm along the x direction and 2mm along y direction. The criteria
adopted for determining the focusing phenomenon was when there was a high field
intensity compared to the surrounding region at the position corresponding to the
source monopole in the image plane, it indicated a focus point. The measured field
164
intensity along xy plane at 2.1 GHz at a distance of 11.31cm from the source is shown
in fig.5.13. In the figure, the field values are given in dB. The reference was taken as
the maximum value measured in each case. It shows focusing at the right hand side of
the slab. The focus point is indicated in rectangular box. It can be seen that the
difference between the values at the focus point and the surrounding area was 6dB.
This indicated clear wave focusing. The focus point is marked in a rectangular box.

Fig.5.13. Measured electric field intensity along xy plane from the slab consisting of metallic
wires at 2.1GHz showing focusing

Measurement results in the region 2.7-3.8 GHz did not show any focusing since
it belonged to the positive refraction region in the dispersion diagram. The measured
field intensity along the xy plane at 3.4GHz which is a frequency point at which
positive refraction (hence no focusing) occurs is shown in fig.5.14. From the figure, it
165
can be seen that there is no wave focusing at 3.4 GHz. Measurements were carried out
for different frequencies and focusing was also observed at 4.2-5.0 GHz band. This
band corresponds to a negative refraction frequency region in the dispersion diagram.
The measured electric field intensity at 4.2 GHz is shown fig.5.15.

wires at 3.4 GHz showing no focusing

166

wires at 4.2GHz showing focusing

5.4 Discussion of results
The focusing in the MEBG at lower frequencies was due to strong spatial
dispersion in the structure. Unlike the results reported earlier in [6] and [8], this study
considers the scanning of field intensity in two dimensional planes. From simulation
and measurement results at 2.1 GHz, it can be seen that at 2.4 cm away from the right
edge of the slab, the electric field intensity at the centre is higher than that in the area
surrounding. In the measurement, an additional high intensity point at the edge was
observed and this can be explained by the edge effects from the slab and air interface.
The simulated and measured results clearly verified wave focusing from an EBG
consisting of metallic wires.
167
There may be an argument that the observed focusing was arisen by the
diffraction from the edges. The simulated electric field intensity at various xy planes
along the z axis indicated that wave is getting focused at the middle of the slab and then
diverged when it travels through the slab (fig.5.9-5.11). If the phenomenon observed is
due to diffraction of two sources separated by a distance, then there will be focusing
points at the centre of all xy planes at different z positions. In other words, if the
phenomenon is due to diffraction or due to an array antenna effect (where multiple
number of beams can be obtained depending on the number of elements used ), there
will be focusing with one layer of wires and this will repeat at multiple points as the
wave passes through the slab. This was not the case observed from simulation results.
In fig.5.9, the field plots at 3.4 cm away from the source (which is inside the wire
structure by 1cm from the left air- slab interface) indicating no focusing. Also the
simulated field distribution with one layer of wires is given in fig.5.16. No focusing
was observed in this case.
In order to verify that the focusing phenomenon was due to the metamaterial
behaviour of the MEBG slab, further numerical studies were carried out. In this aspect,
the MEBG slab was replaced with slabs of different materials, each having same
dimensions as that of the MEBG slab. The electric field distributions obtained from
numerical simulations was studied in three cases. In the first case, the MEBG was
removed (air box only). In the second case, a dielectric slab of permittivity 2.0 was
placed instead of the MEBG. In third case, another slab of permittivity 80.0 was
employed. In all of the above cases, high resolution focusing phenomenon was not
observed. Fig.5.17 shows the electric field distribution in the yz plane for the air box
with radiation boundaries only (i.e. no slab in between) and fig. 5.18 shows the electric
intensity distribution with a dielectric slab with low permittivity
r
= 2.0. Simulation
was performed with a dielectric slab with high dielectric constant (
r
= 80.0). Simulated
168
electric field at 2.1 GHz is presented in fig.5.19. From figures 5.16-5.19, it can be
observed that there is no wave focusing in all the cases. One may argue that the peaks
at the right side of the slab can be a focus but it is easily seen that the physical size of
the peak intensity area is three times larger than that in the case of metallic slab. In
other words, waves are not focused in this case.

Fig.5.16. Simulated electric field intensity along yz plane at 2.1 GHz with one layer of wires
indicating no wave focusing.

Fig.5.17.Simulated electric field intensity along the yz plane at 2.1GHz without any slab
structure showing no wave focusing
169
This confirmed that the observed wave focusing from the MEBG is due to the
metamaterial behaviour of the slab and can not be obtained with ubiquitous materials.
Compared to the field intensity distribution at 2.1GHz, the resolution of the focusing at
4.2 GHz was reduced. At 2.1GHz the image size was observed as 2cm corresponding to
a resolution of /7 and at 4.2 GHz the image size was 2.7cm yielding a resolution of
/2.64. In both cases, an image size of 6mm was noticed in the y direction. From the
above discussions, it was shown that high resolution focusing can be obtained from an
MEBG at low microwave frequencies in the passband of the EBG. The observed
focusing was far below the complete bandgap. The complete bandgap of the EBG starts
only at 9.2 GHz.

Fig.5.18.Simulated electric field intensity along the yz plane at 2.1GHz with a slab of
permittivity 2.0 showing no wave focusing
170

Fig.5.19. Simulated electric field intensity along the yz plane at 2.1GHz with a slab of
permittivity 80.0 showing no wave focusing

5.5 Conclusions
Focusing at low microwave frequencies from a slab structure consisting of thick
metallic wires is studied in this chapter. The structure behaves like a slab having
negative refractive index or an LHM slab. So it can be considered as an analogous
LHM. Compared to the infinite model analysis which predicts negative refraction at
bandgap edges, in this study wave focusing at low frequencies in the passband is
demonstrated. At 2.1 GHz the phenomenon is observed with good resolution and this is
the lowest frequency reported for focusing using EBGs. This is far below from the
frequencies where focusing was obtained in previous work. Numerical simulation and
measurement results were presented with good agreement. In order to establish further,
that the observed phenomenon is due to the metamaterial nature of the MEBG, various
cases have been studied numerically. For this, the MEBG slab was replaced with
dielectric slabs with low and high dielectric constants. It has been demonstrated using
simulation results that the observed phenomenon is not due to array antenna or
171
diffraction effects. The two dimensional focusing phenomenon observed at lower
microwave frequencies can lead to applications in high resolution imaging.

172
References

[1] J. B. Pendry, Negative refraction makes a perfect lens, Physical Review
Letters, Vol. 85, pp. 39663969, 2000
[2] Gennady Shvets, Photonic approach to making a material with a negative index
of refraction, Physical Review B, Vol. 67, pp. 035109-1-8, 2003
[3] E. Cubukcu, K. Aydin, E. Ozbay, S. Foteinpolou and C. M Soukoulis, Sub-
wavelength resolution in a two-dimensional photonic-crystal-based super lens,
Physical Review Letters, Vol.91, pp. 207401-1-4, 2003
[4] A. N. Lagarkov and V. N. Kissel, Near-perfect imaging in a focusing system
based on left-handed material plate, Physical Review Letters, Vol. 92, pp.
077401, 2004
[5] P. A. Belov, Flat lenses formed by capacitively loaded wire media, Proc.
Progress in Electromagentics Research, 2005, China
refraction without negative effective index, Physical Review B, Vol. 65, 201104,
2003
by flat lens using negative refraction, Nature, Vol. 426, pp. 404, 2003
[8] Mayasa Notomi, Negative refraction in photonic crystals, Optical and Quantum
Electronics, Vol. 34, pp. 133-143, 2002
properties of photonic crystals, Journal of Optical Society of America, Vol. 17,
pp. 1012-1020, 2000

173
Chapter 6

Experimental Investigation of Canalization
and Imaging Using Metallic Thin Wire
Medium

6.1 Introduction
In the previous chapter, wave focusing at low microwave frequencies using
thick metallic wire medium was demonstrated. It was shown that the wire medium act
as a metamaterial at certain frequency bands and the phenomenon observed can not be
obtained using natural materials. A sub-wavelength imaging regime, recently proposed
by Belov et al. [1] using periodic structures which acts as a metamaterial with superior
imaging performance is investigated in detail. In this study, periodic structures
consisting of thin metallic wires were used. In the canalization imaging regime, the
source should be very close to the periodic medium so that the near field information
can be carried out to the image side without distortion. From the extensive
measurement studies, it was remarkably noted that some factors needed to be
considered for a precise study of the imaging systems. The first is related to the lens
performance due to the interaction of the source with the lens and the second is the
imaging resolution and bandwidth of the lens. However, these two issues are
interrelated.
Chapter 6: Experimental Investigation of Canalization and Imaging Using Metallic Thin Wire
Medium
174
In this chapter, the high resolution imaging regime proposed in [1, 2] is briefly
discussed with its merits over other near field imaging approaches. This is followed by
experimental studies of the imaging phenomenon. A wire medium slab lens was
constructed and used for the imaging studies. Imaging measurements were carried out
in two steps. In the first case, a source in the shape of letter P was used for the study.
The results obtained from the P source imaging measurements verified the applicability
of the wire medium lens for sub-wavelength imaging. In the second case, the
measurement was carried out with a meander like periodic radiating source. Compared
to the P loop source, the meander-like source has more complex near field distribution.
From this experiment, the source-lens interaction and bandwidth issues were studied
experimentally. Very high resolution with broad bandwidth was obtained in both cases
and it was found that there will be a reduction in bandwidth due to the interaction of
source and lens.

6.2 The canalization regime for near field imaging
Pendry [3] proposed imaging using a slab of left handed material. However, the
major challenge is still to obtain the homogenous materials possessing negative
magnetic properties at very high frequencies (optical and terahertz) and to overcome
the issues related to losses. The other option to reach the sub-wavelength resolution
was suggested by Wiltshire et al. [4]. The idea was based on the use of an array of
magnetic wires which were called as Swiss rolls [5-7] to transfer sub-wavelength
information directly from the source to the image plane (pixel-to-pixel imaging
principle). The lens formed by the Swiss rolls has to be placed in the near field of the
source since it is capable to transport rather than amplify evanescent harmonics. This
realization of sub-wavelength imaging systems experiences similar problems as those
Medium
175
in the development of left-handed medium. It is essential to obtain a metamaterial with
magnetic properties with negligible losses. The metamaterials consisting of Swiss rolls,
and the composite medium consisting of split rings and wires with negative parameters
were found to be very lossy at high frequencies.
To overcome these limitations of the existing models, Pavel et al. [1] proposed
the alternative approach to construct the sub-wavelength lens, which does not require
magnetic properties. The imaging device is formed by an array of parallel conducting
wires, so-called wire medium, a typical geometry of which is shown in Fig.6.1. At the
first sight it seems that this structure is an electrical analogue of Wiltshire's system [4-
7]. An array of Swiss rolls, being similar to magnetic wires, is capable to transmit s-
polarized (transverse electric, TE) spatial harmonics of the source spectrum. An array
of wires operates in the same manner, but for p-polarized (transverse magnetic, TM)
waves. In the other words, an array of Swiss rolls restores at the back interface normal
components of magnetic field produced by the source.

Fig.6.1. Geometry of a slab lens consisting of metallic wires. The length of the wires is an
integer multiple of /2.
n/2
Medium
176
An array of wires restores normal components of electric field. At the same
time, there is a considerable difference between the Wiltshire's system and the wire
medium slab. The Swiss rolls are artificial resonant structures which behave as
magnetic wires only at the frequencies in vicinity of the resonance. This fact restricts
the Swiss rolls to be narrow-band and very lossy. However, the conducting wires in this
sense are natural electrical wires. It means that they are wide-band and practically
lossless. The absence of strong losses (inherent in Swiss rolls) in ordinary wires
removes restriction on the lens thickness. It allows the creation of sub-wavelength
lenses of nearly arbitrary thickness that deliver images with sub-wavelength resolution
into an image plane in the far-field region of the source and beyond. The imaging
system effectively works as a telegraph formed by a multi-conductor transmission line.
Different spatial harmonics incident to the front interface of the lens formed by
wire medium experience different reflection/transmission properties. It happens due to
impedance mismatch between air and wire medium. The wire medium has surface
impedance for p-polarization which is independent of incidence angles in contrast to
the air which the surface impedance varies for different angles of incidence. The
reflections from the thin slabs are negligibly small, but become significant for thick
layers. This problem can be solved by choosing an appropriate thickness of the slab in
order to fulfil condition for the Fabry-Perot resonance and reduce reflections. Actually,
the reflections can be eliminated completely in the present case in contrast of those in
the classical Fabry-Perot resonator where nonzero reflections are inevitable for oblique
incidences.
For any incidence angle, the wire medium supports propagating modes, so-
called transmission line modes, which travel across the slab with the same phase
velocity equal to the speed of light. If the slab thickness is chosen to be integer number
Medium
177
of half-wavelengths then the Fabry-Perot condition holds for any incidence angle and
hence such a slab experiences total wave transmission. This phenomenon of collective
reduction of reflections for all angles of incidence together with the fact that the waves
are allowed to transfer energy only across the slab (along the wires) with a fixed phase
velocity is called as canalization regime [1]. This regime can be observed in various
electromagnetic crystals which possess flat iso frequency contours or the equi
frequency surfaces (EFS) at certain frequencies [7, 8].
The wire medium is a unique example of electromagnetic crystals with such
properties observed at very long wavelengths as compared to the period of the crystal,
which opens up a possibility to obtain nearly unlimited resolution of sub-wavelength
imaging. The transmission line modes exist in the slab have TEM polarisation and
travel along the wires with speed of light [1, 2]. Let the longitudinal component of the
wave vector is q
l
and transverse wave vector be q
t
. Then for the TL modes q
l
is equal to
k. This means that the medium has a flat equi frequency contour for all values of
transverse component of wave vector [1]. When the thickness of the slab equals half
wavelength, it acts as a Fabry-Perot resonator with complete transmission for p
polarised (TM) wave.
The resolution of a lens formed by a wire medium is restricted only by its
period which can be made as small as necessary for certain applications. In the present
case, the resolution is equal to the double period of the lattice [1]. This means that two
different objects can be distinguished if they are located close to two different wires,
but their location within one elementary cell cannot be determined. The sub-
wavelength imaging regime using canalization can be applied in optical microscopy
and near field scanning applications.

Medium
178
6.3 Design of Wire Medium Lens (WML)
A wire medium lens (WML) consisting of copper wires with radius 1mm was
selected for the studies. The array consists of 21 elements both in horizontal and
vertical directions. The period of the WML was 1cm. The frequency of interest was
1GHz (which corresponds to a wavelength of 30cm in free space). Based on the
discussion in the previous section (section 6.2), the length of the wires should be a
minimum of 15 cm (i.e. half the wave length at 1GHz). The wires were fixed on two
thin foam sheets. The foam sheets offer transmission without any considerable loss
around 1GHz. The source used was a loop in the shape of the letter P, formed from a
copper wire of radius 1mm. One arm of the loop was connected to the inner conductor
of a 50 Ohm coaxial cable and the other arm to the outer conductor. The P loop was
placed firmly at the surface of the foam (approximately 3 mm away from the
terminations of wires). A picture of the WML with the P shaped source used for
experimental study and the dimensions of the P loop are given in fig.6.2. The source
was expected to give a good near field distribution at the desired frequencies. The
measured return loss was -5 dB which confirmed that the source would serve its
purpose. The WML transports the sub-wavelength distribution of electric field from the
front interface (source plane) to the back interface (image plane).

Medium
179

Fig.6.2. A photograph of the wire medium lens with the P loop source (left) and a diagram
showing the dimensions of the P loop source.

6.4 Experimental verification
Numerical simulation results obtained from CST Microwave Studio verified the
imaging using WML. Fig.6.3 shows the simulated electric field intensity at the source
and at the image planes. From fig.6.3, it can be noticed that the energy at the source
plane was transported to the image plane (back interface). The details were not lost
during the transmission. The corners of the loop produced local maxima as shown in
fig.6.3.D. The p-polarized harmonics are guided by wires and canalized from the front
interface to the back one. The trace of their propagation is visible inside the slab. These
waves form an image at the back interface (fig.6.3.D). The resolution of the image
obtained was noted as 2cm which was /15 at 1 GHz. Very good agreement between
theoretical predictions and results from numerical simulations proves the validity of the
canalization regime. This indicates the canalization of images and the high resolution
imaging using WML.

8cm
10cm
6cm
1cm
Medium
180

Fig.6.3. Distribution of electric field and its absolute value: (A), (C) in the vicinity of the source
(at 2.5mm distance from the front interface; (B), (D) at 2.5mm distance from the back interface;
(E) the transverse plane and (F) the intensity along the transverse plane (reproduced from [9])

Further verification of the simulated results was carried out by measurements.
The main aims of the experiments were to demonstrate the canalization regime using
WML and to determine the imaging efficiency in comparison with simulation results.
In order to verify the canalization regime, the electric field components coming out of
the WML at the image plane needs to be considered separately; E
x
, E
y
and E
z
where E
x
,
Medium
181
E
y
and E
z
are the electric field components along x,y and z respectively. The WML
predicted to transport the E
x
component effectively, which is TM polarised and the E
y
,
E
z
components will suffer much loss since they are TE polarised. The source field
distribution was measured at a distance of 5mm from the source at the front interface
and the image was measured at 5 mm from the image plane at the back interface. To
measure various field components accurately, two different electric field probes were
used (Fig.6.4). Fig.6.5 shows the direction of electric field in each probe.

Fig.6.4. A photograph of the two probes used for the measurement studies

Fig.6.5. A diagram of the straight probe and the bent probe used for the measurements and
corresponding electric field directions
E
v
E
v

Medium
182
To scan the x component of the electric field, a coaxial probe of length 1cm was used.
This probe was held in parallel to the wire axis. Field intensity was measured at various
points in a rectangular area of 400cm
2

(70 points along y and z direction with scanning
centre at the centre of the air WML interface). To measure the y component of the
electric field, a bent probe which was similar to the previous one but with a bend as
shown in fig.6.4 was employed. Fig.6.5 shows the electric field directions in each
probe. This bend was introduced due to two reasons. One reason was the ease of fixing
the probe to the scanner without having mechanical distortions during the scanning
process. The second reason was that the reduction of source or image field interaction
with the nearby conducting cable. The probes were tested for good transmission in the
required frequency band. The measured electric field intensity distribution along the yz
plane at 0.98 GHz is given in fig.6.6. The wire medium was designed at 1 GHz.
However, in measurement the best quality of image was obtained at 0.98 GHz. This can
be explained by the fact that there can be a small shift in the Fabry- Perot resonance
due to the fabrication tolerance of the WML.
From the measurement results given in fig.6.6, it can be seen that the x
component of electric field is completely recovered at the back interface. The other two
components suffered image degradation since they are not in the TE polarisation. The
plots of the absolute values of source and image fields (fig.6.6.G, H) were in very good
agreement with the simulation results (fig.6.3.C, D). The probes used for the
measurements only give an average value over a very small volume (approximately
1cm
3
surrounding the probe) and due to this reason some of the local maxima
corresponding to the termination of the wires that appeared in simulation results, got
diminished in measurement.

Medium
183

Fig.6.6. Measured near field intensity distribution along the yz plane: Absolute values of x, y
and z components of the electric fields (A, C, E, G) in source plane in image plane (B, D, F, H).

Medium
184
From the measurement, it was clear that there was a resolution of 2 cm at 0.98GHz.
This corresponds to resolution of about /15. From the extensive measurements, it was
made clear that sub-wavelength imaging occurred from 0.92- 1.1 GHz. The bandwidth
can be calculated using the formula

c
l u
f
f f
BW

= (6.1)

where f
c
is the centre frequency, f
u
and f
l
are upper and lower frequencies respectively.
Substituting the values equation 6.1 gives a bandwidth of 17.8%. This is higher than the
bandwidth of imaging obtained using Swiss rolls in [4].

6.5 Interaction between source and lens
In the previous discussions, the canalization and imaging using WML was
studied experimentally and good agreement with numerical simulation results were
observed. However, the interaction of the source and the lens was not adequately
studied in the P loop imaging work. The basic principle of mutual induction explains
that there exist an induced current on a conductor when it is placed near a current
carrying element. This leads to close by near fields to interact and this can reduce the
bandwidth of the imaging regime. It depends on the source and the period of the lens.
When the source contains more complex elements, there will be much more interaction
compared the case of a P loop source. In the case of a P loop, the source was very
simple and it was not showing much interaction with the WML. In the following
discussions, the experimental observations of the resolution of images and bandwidth
of imaging from the WML, when the P loop source was replaced with a meander-like
Medium
185
radiating loop are detailed. The loop was constructed on a foam sheet using a thin wire
of 0.22 mm radius, by shaping it in a Meander like pattern. A photo of the Meander like
radiating source and a diagram showing its dimensions is given in fig.6.7. The spacing
between each folding of the wires was 2 cm. The corners were fixed using metallic
pins.

Fig.6.7.A photograph of the Meander shaped radiating source used for the experiment and a
diagram showing dimensions of the loop.

These fixing pins may affect the near field distribution. However, the main interest was
to create a complex near field pattern which can interact with WML effectively and to
study the canalization efficiency and bandwidth. So the presence of the pins was found
not affecting the aims of the studies. The source was fed using a coaxial cable as in the
previous case.
The WML proposed in the previous experiment was used in this study as well.
Near fields were measured at source and image planes in similar manner as in the
previous experiment. Also, fields were measured at the source plane with and without
the lens. In the previous experiment, it was noted that the x component of the electric
Medium
186
field was canalised very effectively. The source plane was at 3 mm from the first
interface and the image plane was 3mm distance from the back interface. As mentioned
earlier, the aim of the experiment was to accurately determine the imaging bandwidth
and resolution and to study the effect of the source WML interaction. The
measurement observations revealed the practical applicability of the WML. As it will
serve the purpose of the experiment, only the x component of the electric field was
measured in this case.
The design frequency of the WML was same as before i.e. at 1 GHz. The
electric fields were measured using the horizontal probe employed in the previous
experiment. Fields were measured in the frequency range 0.85- 1.1 GHz with a step of
0.01GHz. For each frequency, there were three field measurements; the source
distribution at the first interface at a distance of 3 mm from the backside of the foam
sheet without lens, the source distribution at the same point with the WML in front of
the source and the field distribution at the image plane at a distance of 3 mm from the
back interface (image plane). The source was placed very close (3 mm) to the WML.
The first set of measurements yields the near field distribution which is the
source for the WML. The second set of measurements gives the near field distribution
after the source field gets interacted with the WML. The last set of measurements
yields the image obtained from the WML. The measured results at 0.85-0.96 GHz and
at 0.99 and 1GHz at frequencies are given in figs.6.8 - 6.11. In the figures, the field plot
at the left is the source with lens and the plot at right is the source distribution without
the WML. The image field distribution is given in the middle.
Fig.6.8 shows that source distribution got distorted due to the insertion of WML
in front of the meander-like radiator. Also, from the image field distribution it can be
seen that there is no imaging takes place till 0.86 GHz. Fig. 6.9 A-D shows the field
Medium
187
distribution from 0.87- 0.9 GHz. The field distributions at different frequencies are
given in figs. 6.10- 6.11. From fig.6.8 it can be noticed that no imaging takes at 0.85
GHz and 0.86 GHz. It can be seen from fig.6.9 that the source with and without lens
(left and right in the figures) are different and this indicates that source has a strong
interaction with the WML at 0.87 -0.88 GHz. When the frequency is increased in steps
of 0.01GHz, the gradual change in the field distributions at typical frequencies can be
observed (Figs.6.9-6.11).

Fig.6.8.Measured electric field intensity distribution in the yz plane at 0.85GHz (A) and
0.86GHz (B). Source fields with and without lens (left and right respectively) and the image
field (middle) are given.

A
B
Medium
188

Fig.6.9.Measured electric field intensity distributions in the yz plane at 0.87GHz (A),
0.88GHz (B), 0.89GHz (C) and 0.9 GHz (D). Source fields with and without lens (left and
right respectively) and the image field (middle) are given.

A
B
C
D
Medium
189

Fig. 6.10. Measured electric field intensity distribution in the yz plane at 0.91GHz (A) and
0.92GHz (B), 0.93GHz (C) and 0.94GHz (D). Source fields with and without lens (left and
right respectively) and the image field (middle) are given.

A
B
C
D
Medium
190

Fig.6.11.Measured electric field intensity distribution in the yz plane at 0.95 GHz (A), 0.96
GHz (B), 0.99 GHz (C) and at 1 GHz (D). Source fields with and without lens (left and right
respectively) and the image field (middle) are given.

A
B
C
D
Medium
191
From the results it can be noticed that the canalization regime is from 0.89-0.94
GHz. At 0.92-0.93 GHz (Figs.6.10.B, C), the best quality image was obtained.
Resolution in this case was 2.1cm. This gives rise to a resolution of about /15 at
0.92GHz. At frequencies lower than 0.89 GHz, the source field plot with lens is
different from that without the lens. This indicates that the source field was reflected
back and distorted due to the introduction of the WML. It has to be noted that the
canalization regime is active till 0.87 GHz because whatever be the field at the source
plane got transported to the image plane. This disappears below 0.87 GHz. A
frequency above 0.94 GHz, the source is distorted due to the interaction of the WML
(fig.6.11 A, B). However, the canalization regime exists till 0.96 GHz. This indicated
that the actual canalization bandwidth was reduced from the actual one due to the
interaction between the source and WML. Here an imaging bandwidth of 5.3 % was
noted whereas in the case of P loop source, a bandwidth of 17.8% was obtained. This
reduction in bandwidth due to the source field alteration and can be explained by the
effect of induced field interaction.
The measurement results revealed that the actual bandwidth of the imaging
system was influenced by the WML and the type of source. However, in both cases the
resolution was maintained as /15. The resolution and bandwidth was found to be in
good agreement with the theoretical predictions given in [10]. It is proved in [10] that
the minimum bandwidth of imaging for the WML is 5%.
Next, the electric field was measured at the image plane without the WML in
between the source and scanning probe. This was to further establish the fact that the
observed imaging can not be obtained without the WML. The WML acts similar to a
transmission line medium supporting TL modes (for the x component of the field). The
measured field had not given a distribution similar to the one at the source, but instead
Medium
192
almost a uniform distribution was obtained. Measured electric field distribution at a
distance of 155mm from the source at 0.92 (which was at the centre of the canalization
regime) is given in fig.6.12, indicating no imaging.
Finally, the effect of the foam sheet used for the support was verified. Fig.6.13
shows the measured x component of the electric field in front of the source and at the
back, very close to the foam sheet (3 mm away from the foam sheet). In both cases, the
near field distribution is same (with very low transmission loss with the foam sheet).
This confirmed that the effect of the foam sheet was negligible. It was not reducing the
transmitted power considerably and it was not distorting the field distribution.

Fig.6.12. Measured electric field distribution at 0.92 GHz at 155 mm away in front of the
source showing no imaging.

z (cm)

y

(
c
m
)

Medium
193

Fig.6.13. Measured electric field distribution at 3 mm away from the foam sheet at the backside
of the source (12 mm from the source) and at 3 mm away from the foam sheet at the front side
at 0.88 GHz. In both cases the near field distribution is the same and there is no considerable
transmission loss.

The imaging can be observed at the same frequencies with longer WML as long
as the wire length is an integer multiple of half wavelength. The same phenomenon can
be obtained at higher frequencies when the frequency of operation is multiples of
Fabry-Perot resonance frequency.
6.6 Conclusions
In this chapter Canalization, a regime in which the wire medium acts as a
transparent slab for a near field source was studied experimentally. Images with sub-
wavelength resolution were obtained. The imaging of a P letter shaped source was
demonstrated with experimental results. The bandwidth of the observed phenomenon
was 17.8%. The observed phenomenon can not be obtained from slabs of naturally
available materials. In the present case, the wire medium is a metamaterial or it behaves
like a left handed slab (without any field amplification). Near field interaction of the
source with the proposed lens was an issue, which was not considered in previous
studies. For this purpose, a complex source compared to the P loop source was used.
The results obtained from the second experiment revealed that for complex sources the
Medium
194
bandwidth of the imaging regime was reduced but the imaging was obtained with the
same resolution as in the case of a simple source. For the meander-like source, the
bandwidth obtained was 5.3%. Extensive experimental results confirmed the existence
of canalisation regime. The WML has the best possible imaging resolution (/15)
among other near field imaging approaches based on periodic structures.

Medium
195
References
[1] Pavel A. Belov, Constantin R. Simovski, Pekka Ikonen, Canalization of sub-
wavelength images by electromagnetic crystals, Physical review B, Vol. 71, pp.
193105-193108, 2005
[2] Pavel A. Belov, Flat lens formed by a capacitively loaded wire media, Proc. of
the Progress in Electromagnetics Research (PIERS), China, August, 2005
[3] J. B Pendry, Negative refraction makes a perfect lens, Physical Review Letters,
Vol. 85, pp.3966-3969, 2000
[4] M. C. K. Wiltshire, J. B. Pendry, I. R. Young, D. J. Larkman, D. J.
Gilderdale, J. V. Hajnal, Microstructured magnetic materials for RF flux guides
in magnetic resonance imaging, Science, pp. 849 851, 2002
[5] M. C. K. Wiltshire, E. Shamonia, L. Solymar, Experimental and theoretical
study of magneto- inductive waves supported by one dimensional arrays of swiss
rolls, Journal of Applied Physics, Vol. 95, pp. 4488-8893, 2004
[6] S. Anantha Ramakrishna, J. B. Pendry, M C K Wiltshire, W. J. Stewart,
Imaging the near field , Journal of Modern Optics, Vol. 50, pp.1419 1430,
2003
[7] C. Luo, S.G Johnson, J.D Joannopolous and J.B Pendry, All-angle negative
refraction without negative effective index, Physical Review B, Vol. 65, pp.
201104, 2003
[8] C. Luo, S. G Johnson, J. D Joannopolous and J. B Pendry, Sub-wavelength
imaging in photonic crystals, Physical Review B, Vol. 68, pp. 045115, 2003
[9] Pavel. A Belov, Yang Hao, Sunil Sudhakaran, Sub-wavelength microwave
imaging using an array of parallel conducting wires as a lens, Physical review B,
Vol. 73, pp. 033108 - 033111, 2006
Medium
196
[10] Pavel A. Belov, Resolution of sub-wavelength transmission devices formed by a
wire-medium, ArXiv:Physics/0511139v2, March 2006 (Online resource)

197
Chapter 7

Conclusions and Future work

7.1 Summary
Study of Metamaterials in particular LHMs is an emerging and interesting topic
of research so an in-depth understanding of LHMs and analogous LHMs and their
characteristics is very challenging. In this thesis, substantial effort has been made to
understand the concepts and realisations of negative refraction or refraction like
phenomenon.
The main objectives of the research were to investigate the possibilities of
obtaining negative refraction-like phenomenon at frequencies other than the complete
bandgap edges of EBGs and to extend the concepts of EBG based metamaterial into
high resolution wave focusing and imaging. These goals were achieved in this work.

7. 2 Conclusions
The following conclusions can be drawn from this study:
Numerical simulation results confirmed the spatial dispersion effects due to the
finiteness of the periodic structure. Change in the backward radiation frequency
regions was noticed between unit cell analysis for infinite models and the
proposed super cell model.
Negative refraction from slab and prism models were presented with numerical
simulation results and the occurrence was further established with measurement
Chapter7: Conclusions and Future Work
198
results using a prism model. Extensive measurement results indicated that there
were multiple frequency regions in the pass band at which the structure gave
negative refraction. A transmission of 54% was observed which was higher
than that reported from the split ring and wire medium.
It was observed that the obtained result had higher negative refraction excess
rate (NRE) of 85%. This was higher than that from previously reported values.
Enhancement of negatively refracted beam in the passband by introducing local
defects into the structure was demonstrated. An enhancement of 27% for the
negatively refracted beam was obtained.
Wave focusing from an MEBG slab was demonstrated at low frequencies in the
passband with numerical simulation and experiments. A resolution of /7 was
obtained at 2.1GHz. Numerical simulation results confirmed that such a high
resolution focusing cannot be obtained from ubiquitous materials. The
phenomenon observed was at a lower frequency compared to similar works
reported before.
Sub-wavelength imaging using canalization was experimentally verified using a
P Loop source and a resolution of /15 with a bandwidth of 17.8% was
obtained. It was observed that the interaction between the source and WML
degrades the bandwidth of the system. The bandwidth is reduced to 5.3% with a
complex source.

7.3 Future work
The study presented in this thesis opens up a lot of possibilities in future
research. The outcomes obtained in the study can be extended to novel applications of
metamaterials. Also, novel approaches can be developed for transmission enhancement
and all angle negative refraction from EBGs.
199
7.3.1 Novel applications
7. 3.1.1. Dispersive EBG for demultiplexing and beam scanning applications
The MEBG used in the study shown capability of forming beams at different
angles at different frequencies. For beam scanning application or in a frequency
demultiplexer, only the properties of the exiting wave are important. So the structure
can find its applications as a frequency spatial Demultiplexer. Fig.7.1 shows the
diagram illustrating the mechanism of the spatial demultiplexer using the MEBG prism.

Fig.7.1. A schematic diagram illustrating the application of MEBG prism as a spatial frequency
demultiplexer. The red arrow represents the polychromatic wave and the out coming beams of
different frequencies are shown in different colours and the corresponding angles from the
normal are marked.

The refracted fields were obtained from measurements discussed in chapter 4. This also
can be used as a beam scanner at microwave frequencies. Another possibility is to
consider hemi-circular shaped EBGs for multiple beam applications. This can lead to
novel EBG based antennas. The advantage is the EBG can act as a filter and antenna
with the same hardware leading to joint filter/antenna applications.
EBG Prism
4.5GHz, +32

5.6GHz, -21

7.4GHz, -45

7.2GHz, -58

S
p
a
t
i
a
l

d
e
t
e
c
t
o
r

Normal
Polychromatic wave
200
7.3.1.2 Sub-wavelength antennas
It was noticed that spatial dispersion exists in all periodic structures and this can
cause backward radiation at certain frequencies. This concept can be applied to planar
periodic structures and novel antennas with beam scanning capability can be obtained.
In conventional backward wave antennas, backward radiation can be obtained only
when the period is comparable to the wavelength of operation. This makes the antenna
electrically large. The spatial dispersion in EBGs leads backward radiation at low
frequencies. This reduces the size of the antenna resulting sub-wavelength antennas.
Due to its reduced size, these antennas can be promising candidates for future on body
and indoor communication antennas.

7.3.2 Theoretical extensions
7.3.2.1 Transmission enhancement
From the studies performed, it was noted that the transmitted power obtained
from the MEBG was 54%. The main cause of the reduced transmitted power is the
impedance mismatch at the air and the MEBG interface. Methods can be developed for
impedance matching at the interface and transmission increase. However, it has to be
mentioned that the available matching techniques will cause reduction in bandwidth as
the reported techniques such as Fabry Perot resonator model for transmission
enhancement are for narrow band applications. So a broadband matching technique can
be developed.

7.3.2.2 All angle negative refraction
The study presented in this thesis has not considered all angle negative
refraction. All angle negative refraction is not a mandatory for beam scanning
applications. However, this could lead to improved performance in focusing and
201
imaging applications. As a possible extension of the work, all angle dispersion analysis
can be obtained by considering different angles of incidence.

PhDThesis SunilSudhakaran

Uploaded by

Copyright:

Available Formats

PhDThesis SunilSudhakaran

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

PhDThesis SunilSudhakaran

Uploaded by

Copyright:

Available Formats

Negative Refraction from

Electromagnetic Periodic Structures and

1, spatial harmonics contribution is considerable

You might also like