Reconfigurable and MIMO Antenna Systems for Mobile Communications

Emad Abdullah Abbas Al Abbas

B.Sc., M.Sc.

A thesis submitted for the degree of Doctor of Philosophy at

The University of Queensland in 2019

The School of Information Technology and Electrical Engineering

Abstract

The use of bandwidth for mobile and cellular networks has been increasing at an unprecedented rate,
with applications in emergent areas such as smart vehicles and civil infrastructure, telemedicine, and
remote education. The predicted increase of mobile network utilisation requires a high data rate
capacity to transfer information through the network. The current limited capacity of the fourth
generation (4G) mobile network is not expected to satisfy the predicted high data demand by 2020.
To meet this requirement, a new fifth generation (5G) mobile network has been proposed. In order to
reach the high capacity demands, the 5G requires a wide frequency bandwidth, which is not available
in the sub-3 GHz frequency spectrum. Therefore, many international organisations have assigned the
millimetre-wave band (mm-Wave), which includes 24, 28, 37, 39, and 60 GHz, as a frequency band
for 5G. Using mm-Wave raises many challenges such as an increase in the path loss, radiation field
absorbed by the human body, signal polarisation misalignment, and MIMO antenna footprint inside
the mobile handset. Therefore, designing beam-steering, reconfigurable circular polarization, and
compact MIMO antenna systems for 5G applications is important to cope with these challenges.
Although there are some antenna systems were reported recently for mm-Wave 5G, they had limited
steering angles, one polarization types or bulky antenna size. This thesis aims to solve the current
challenges mentioned above by developing mm-Wave 5G antenna systems and in doing so makes
three main contributions to the field of mm-Wave 5G antenna systems.

As the path loss and radiated fields absorbed by surface layers of the human body increase at mm-
Wave, emphasis is given into the development of the beam-steering antenna system, which is the first
contribution of this thesis. The beam-steering technique aims at reducing the effects of
electromagnetic field exposure on the human head, as well as maintain a consistent connection
between the handset and the base station. A novel compact three-dimensional (3-D) beam-steering
system of mobile handsets at 28GHz is proposed. It consists of dipole antenna arrays, power dividers
and digital phase shifters. Nevertheless, a novel technique of connecting p-i-n diodes to build a phase
shifter is proposed for better performance. The proposed design has the ability to steer the beam in a
wide angle in 3-D. it relies on using both sides of a printed circuit board for a two-dimensional antenna
array. Moreover, to further improve the performance of the beam-steering system, a tunable power
divider for mm-Wave 5G is presented. It is used to tune the feeding amplitude of the beam-steering
antenna array, and thus reduces the side-lobe level (SLL) of the antenna radiation pattern.

The second contribution is the development of two reconfigurable circular polarisation (CP)
structures for 5G applications at mm-Wave. CP is an effective technique to mitigate signal
polarisation misalignment, phase deviation and multipath effect problems. Electrical reconfigurable
methods are used to switch the CP type by using p-i-n diodes. Firstly, an open-ended stub concept
I
with a simple feed circuit is used to build a novel reconfigurable CP patch antenna. Then the same
concept is developed to build a reconfigurable CP antenna array system using branch-line couplers,
power divider, and 4 p-i-n diodes. The open-ended stub concept is found to be efficient in both designs
by reducing the complexity and the number of the p-i-n diode switches compared to the reported
existing CP antenna systems in the literature.

MIMO is a major technology of 4G mobile antenna system and also has disruptive potential for mm-
Wave 5G. Therefore, the third contribution is the development of new design approaches to integrate
the current 4G with futurist 5G MIMO antenna systems. Integrating these two technologies in one
structure leads to a reduction of the total MIMO antenna footprint significantly within handsets, where
the available space is limited. To that end, two novel 5G/4G MIMO antenna systems are built. First,
5G/4G MIMO antenna system is proposed by using one structure; and it covers multiband mm-Wave
spectrums (28, 37 and 39) GHz for 5G as well as sub 3 GHz for 4G. Moreover, to further improve
the gain especially at mm-Wave band, an antenna array structure is deployed to build another 5G/4G
MIMO antenna system. This design integrates 5G and 4G operations into a single structure; and
covers a wideband 23-29 GHz for 5G and frequency-tuning 2.05-2.70 GHz for 4G.

The developed antenna systems reported in this thesis positively contribute to the antenna design
domain and are expected to encourage safety, compactness, reliability, and a multiband antenna
system for the mm-Wave 5G industry in near future.

II
Declaration by author

This thesis is composed of my original work, and contains no material previously published or written
by another person except where due reference has been made in the text. I have clearly stated the
contribution by others to jointly-authored works that I have included in my thesis.

I have clearly stated the contribution of others to my thesis as a whole, including statistical assistance,
survey design, data analysis, significant technical procedures, professional editorial advice, financial
support and any other original research work used or reported in my thesis. The content of my thesis
is the result of work I have carried out since the commencement of my higher degree by research
candidature and does not include a substantial part of work that has been submitted to qualify for the
award of any other degree or diploma in any university or other tertiary institution. I have clearly
stated which parts of my thesis, if any, have been submitted to qualify for another award.

I acknowledge that an electronic copy of my thesis must be lodged with the University Library and,
subject to the policy and procedures of The University of Queensland, the thesis be made available
for research and study in accordance with the Copyright Act 1968 unless a period of embargo has
been approved by the Dean of the Graduate School.

I acknowledge that copyright of all material contained in my thesis resides with the copyright
holder(s) of that material. Where appropriate I have obtained copyright permission from the copyright
holder to reproduce material in this thesis and have sought permission from co-authors for any jointly
authored works included in the thesis.

III
Publications included in this thesis

1. E. A. Abbas, A. M. Abbosh, and K. Bialkowski, "Tunable In-Phase Power Divider for 5G Cellular
Networks," IEEE Microwave and Wireless Components Letters, vol. 27, pp. 551-553, 2017. – Partly
incorporated in Chapter 3.

Contributor Statement of contribution

Emad Al Abbas (Candidate) Designed and tested (70%)

Wrote the paper (70%)

K. S. Bialkowski Designed and tested (10%)

Wrote the paper (10%)

A. Abbosh Designed and tested (20%)

Wrote the paper (20%)

2. E. Al Abbas, A. T. Mobashsher, and A. Abbosh, "Polarization Reconfigurable Antenna for 5G

Cellular Networks Operating at Millimeter Waves," 2017 IEEE Asia Pacific Microwave Conference
(APMC), pp. 772-774, 2017. – Partly incorporated in Chapter 4.

Contributor Statement of contribution

Emad Al Abbas (Candidate) Designed (75%)

Wrote the paper (75%)

A. T. Mobashsher Designed (15%)

Wrote the paper (15%)

A. Abbosh Designed (10%)

Wrote the paper (10%)

IV
Submitted manuscripts included in this thesis

1.E. A. Abbas, N. Nguyen-Trong, A. T. Mobashsher, and A. M. Abbosh, "Polarization-

Reconfigurable Antenna Array for Millimeter-wave 5G," IEEE Antenna and Wireless Propagation
Letters. Partly incorporated in Chapter 4.

Contributor Statement of contribution

Emad Al Abbas (Candidate) Designed and tested (60%)

Wrote the paper (60%)

N. Nguyen-Trong Designed and tested (20%)

Wrote the paper (30%)

A. T. Mobashsher Designed and tested (15%)

Wrote the paper (5%)

A.M. Abbosh Designed and tested (5%)

Wrote the paper (5%)

2. E. A. Abbas, M. Ikram, A. T. Mobashsher, and A. M. Abbosh, " MIMO Antenna System for Multi-
Band mm-Wave 5G and Wideband 4G Applications," IEEE Transaction on Antenna and
Propagation. Partly incorporated in Chapter 5.

Contributor Statement of contribution

Emad Al Abbas (Candidate) Designed and tested (55%)

Wrote the paper (30%)

M. Ikram Designed and tested (30%)

Wrote the paper (50%)

A. T. Mobashsher Designed and tested (10%)

Wrote the paper (10%)

A.M. Abbosh Designed and tested (5%)

Wrote the paper (10%)

V
3. M. Ikram , E. A. Abbas, N. Nguyen-Trong, K.H. Sayidmarie, and A. M. Abbosh, "Integrated
Frequency-Reconfigurable Slot Antenna and Connected Slot Antenna Array for 4G and 5G Mobile
Handsets," IEEE Transaction on Antenna and Propagation. Partly incorporated in Chapter 5.

Contributor Statement of contribution

M. Ikram Designed and tested (50%)

Wrote the paper (45%)

Emad Al Abbas (Candidate) Designed and tested (30%)

Wrote the paper (30%)

N. Nguyen-Trong Designed and tested (10%)

Wrote the paper (15%)

K.H. Sayidmarie Wrote the paper (5%)

A.M. Abbosh Designed and tested (5%)

Wrote the paper (5%)

Other publications during candidature

Peer-reviewed Journal Papers

1. E. A. Abbas, A. M. Abbosh, and K. Bialkowski, "Tunable In-Phase Power Divider for 5G Cellular
Networks," IEEE Microwave and Wireless Components Letters, vol. 27, pp. 551-553, 2017.

Peer-reviewed Conference Papers

1. E. A. Abbas, A. T. Mobashsher, and A. Abbosh, "Polarization reconfigurable antenna for 5G

cellular networks operating at millimeter waves," in 2017 IEEE Asia Pacific Microwave Conference
(APMC), pp. 772-774, 2017.
2. E. A. Abbas and A. M. Abbosh, "Tunable millimeter-wave power divider for future 5G cellular
networks," in 2016 IEEE International Symposium on Antennas and Propagation (APSURSI), pp.
1715-1716, 2016.

VI
3. E. A. Abbas and A. Abbosh, "Adjustable millimeter wave in-phase power divider for 5G using a
three-line coupler structure," in 2016 IEEE International Symposium on Antennas and Propagation
(APSURSI), pp. 1713-1714, 2016.
4. E. A. Abbas and A. Abbosh, "Millimeter wave tunable power divider using modified Wilkinson
design," in 2016 IEEE 2nd Australian Microwave Symposium (AMS), pp. 3-4, 2016.

VII
Contributions by others to the thesis

I acknowledge the contribution by others in this thesis. Prof. Amin M. Abbosh as the main supervisor
had contributed closely in defining the research problem, overall conception and direction of the
thesis. Dr. Konstanty S. Bialkowski as the second supervisor had worked closely on conducting some
measurements and given many technical advice. Dr. Ahmed Toaha Mobashsher as the third
supervisor had provided massive support on designing planar antennas specifically in Chapter 3
(Section 3.2) and Chapter 5 (Section 5.3). Dr. Nghia Nguyen –Trong had contributed in the design of
circular polarization antenna array. His contribution is specifically acknowledged in sharing his
experience in extracting the equivalent lumped-element models of p-i-n diodes at millimeter-wave in
Chapter 4.

VIII
Statement of parts of the thesis submitted to qualify for the award of another degree

No works submitted towards another degree have been included in this thesis.

Research Involving Human or Animal Subjects

No animal or human subjects were involved in this research.

IX
Acknowledgements

Studying PhD is a long journey that comes with its own excitements and frustrations along the way.
It takes a lot of effort and patience to complete each step till ending at the final milestone where you
submit your findings as a thesis. I take this opportunity to express my sincere gratitude to those
gracious people who have helped me shape my thesis.

First, I would like to express my deepest appreciation and sincerest gratitude to my principal advisor,
Professor Amin Abbosh, for his encouragement, advice and financial assistance that he offered in my
PhD project. His encouragement, constructive feedback, critical comments, travel funds and
responsiveness are all essential factors that lead me throughout my candidature. This thesis would
not have been possible without his continuous guidance.

I would also like to acknowledge valuable discussions and time that I was privileged to have with my
associated advisors, Dr. Konstanty Bialkowski, and Ahmed Toaha Mobashsher in resolving several
technical matters with testing, and designing antennas.

I am also express my gratitude to Dr. Nghia Nguyen –Trong, who has offered me much useful
information about designing and testing antenna array, and academic writing.

I thank my colleagues at Microwave group, in particular, Dr. Yifan Wang, Dr. Beadaa Mohammed,
who has offered me many tips in simulations techniques. Also, I would like to thank my PhD
colleagues He Zhu, Mohamed Manoufali, Muhammad Ikram, Arman Afsari, Amin Darvazehban,
Anthony Stancombe, Md Shahidul Alam, and Syed Naqvi for their enjoyable discussions about
technical and non-technical matters. I also especially thank Ms. Maissoun Abou Hamra, our Research
Project Officer, who has helped me through administrative work. Many thanks are also to the staff of
Engineering and Technical Support Group, Dr. Karl Bertling, and Mr. Dennis Bill for their technical
assistance in conducting antenna measurement and soldering PCB prototype.

My Ph.D. studies would have not been possible without the financial support of the Iraqi Ministry of
Higher Education and Scientific Research. I am very grateful for the chance that was given to me to
complete my study.

I wish to give my deep appreciation to my dear wife, Gofran, who has always been by my side in any
situations and supported me unconditionally. Her love, patience and continuous support have guided
me through this challenging journey. I would like to thank my children; Ali, Narjis, and Aya for their
love and patience.

X
Last but not least, I would like to give my endless love and appreciation to my deceased father, my
mother for their great role in my life, their numerous sacrifices, and moral support, and my siblings
for their continuous encouragement throughout my PhD journey.

Financial support

This research was supported by the Iraqi Ministry of Higher Education and Scientific Research.

Keywords

reconfigurable antenna, beam-steering antenna system , mobile communication, multiple-input

multiple-output, tunable power divider, circular polarization antenna

XI
Australian and New Zealand Standard Research Classifications (ANZSRC)

ANZSRC code: 100501, Antennas and Propagation, 50%

ANZSRC code: 100505 Microwave and Millimetrewave Theory and Technology, 50%

Fields of Research (FoR) Classification

FoR code: 1005 Communications Technologies, 100%

XII
Table of Contents

Abstract I

Publications included in this thesis ................................................................................................ IV

Submitted manuscripts included in this thesis ............................................................................... V

Other publications during candidature ........................................................................................ VI

Acknowledgements ........................................................................................................................... X

List of Figures ................................................................................................................................XVI

List of Tables .................................................................................................................................XXI

Chapter 1 – Introduction ................................................................................................................. 1

1.1 Background and Motivation ................................................................................................ 1

1.2 Background and Motivation ................................................................................................ 1

1.2.1 Beam-steering Antenna System Motivation and Advantages...................................... 1

1.2.2 Reconfigurable Circular Polarisation Antenna System Motivation and Advantages .. 3

1.2.3 5G/4G MIMO Antenna System for 5G Handsets Motivation and Advantages .......... 4

1.3 Aim of the Thesis ................................................................................................................. 4

1.4 Original Contributions of the Thesis.................................................................................... 4

1.5 Thesis Organisation.............................................................................................................. 5

Chapter 2 – Literature Reviews of Reconfigurable and MIMO Systems at mm-Wave 5G...... 7

2.1 Design Technology for Beam-steering Antenna Systems at mm-Wave 5G Handset.......... 7

2.2 Design Technology for Reconfigurable Circular Polarisation Antenna at mm-Wave for 5G
Applications ................................................................................................................................... 10

2.3 Design Technology for MIMO 5G/4G Antenna Systems of Mobile Handheld Devices .. 11

2.4 Summary ............................................................................................................................ 13

Chapter 3 -Three-Dimension Beam-steering System for Mobile Handsets at 28 GHz ........... 14

3.1 Introduction ........................................................................................................................ 14

XIII
 3.2 Millimeter-Wave Three-Dimensional Beam-Steering System for 5G Handsets ............... 15

3.2.1 Design of the 3-D Beam-Steering System ................................................................. 15

3.2.2 Digital Phase Shifter .................................................................................................. 19

3.2.3 Two-Layer Power Divider ......................................................................................... 22

3.2.4 Layout of the Antenna Array System ........................................................................ 24

3.2.5 Fabrication, Measurement and Discussion ................................................................ 27

3.3 Tunable In-Phase Power Divider for 5G Cellular Networks ............................................. 31

3.3.1 Proposed Design ........................................................................................................ 31

3.3.2 Results and Discussion .............................................................................................. 35

3.4 Tunable Power Divider for SLL Suppression .................................................................... 37

3.5 Summary ............................................................................................................................ 39

Chapter 4 – Reconfigurable CP Antennas for 5G Applications at Millimeter-wave............... 41

4.1 Introduction ........................................................................................................................ 41

4.2 Polarisation-Reconfigurable Antenna for Mm-wave 5G ................................................... 41

4.2.1 Design Geometry and Consideration ......................................................................... 41

4.2.2 Results and Discussions ............................................................................................. 43

4.3 Polarisation-Reconfigurable Antenna Array for mm-Wave 5G ........................................ 45

4.3.1 Antenna Array Structure and Operating Principle ..................................................... 45

4.3.2 Result and Discussion ................................................................................................ 53

4.4 Summary ............................................................................................................................ 55

Chapter 5 - MIMO Antenna System for mm-Wave 5G and Wideband 4G Applications ...... 56

5.1 Introduction ........................................................................................................................ 56

5.2 Antenna Design Requirements for 5G and 4G Mobile Antennas ...................................... 57

5.3 MIMO Antenna System for Multi-Band mm-Wave 5G and Wideband 4G Applications 58

5.3.1 Proposed MIMO Antenna System Design ................................................................. 58

XIV
 5.3.2 Result and Discussion ................................................................................................ 64

5.4 Integrated Frequency-Reconfigurable Slot Antenna and Connected Slot Antenna Array for
4G and 5G Mobile Handsets .......................................................................................................... 71

5.4.1 Antenna Element Design and Performance ............................................................... 71

5.4.2 Measurement and Result ............................................................................................ 79

5.5 Summary ............................................................................................................................ 86

Chapter 6 – Conclusion and Future Work .................................................................................. 88

6.1 Thesis Conclusion .............................................................................................................. 88

6.2 Suggestions for Future Work ............................................................................................. 89

References 91

XV
List of Figures

Figure 1.1 Typical mobile network. ..................................................................................................... 2

Figure 1.2 (a) The penetration depth (mm) with frequency [14] © 2017 IEEE, and (b) power density
(SAR) concentration in the skin at 60 GHz. ........................................................................................ 3

Figure 2.1 Block diagram of beam-steering antenna array system. ..................................................... 8

Figure 2.2 (a) Multiple feed port antenna array [26] © 2016 IEEE, and (b) large feeding circuit of
antenna array [28] © 2018 IEEE.......................................................................................................... 9

Figure 2.3 (a) 1 × 4 CP reconfigurable antenna array in [40] © 2018 IEEE, and (b) diode layout that
produces the polarisation switching mechanisms. ............................................................................. 11

Figure 2.4 Illustrative diagram of the future handset incorporating mm-Wave 5G antennas with 4G
and 3G antennas at sub-6-GHz [14] © 2017 IEEE. ........................................................................... 12

Figure 3.1 Dipole antenna array structure. Dimensions (mm) are D1 = 2.18, D2 = 0.3, D3 = 0.85, D4
= 0.59, D5 = 0.5, D6 = 0.65, D7 = 0.1, D8 = 0.2. ................................................................................ 16

Figure 3.2 (a) Simulated reflection coefficients of the antenna array, and (b) radiation pattern. ...... 16

Figure 3.3 3-D radiation pattern at different beam steering angles. .................................................. 17

Figure 3.4 Simulated power density (PD) distribution in the head phantom for the proposed antenna
array system at 28 GHz. ..................................................................................................................... 18

Figure 3.5 (a) T- Junction phase shifters, (b) Y- Junction phase shifters, (c) simulated spurious
radiation for the Y-Junction and T-Junction phase shifter, and (d) simulated S-parameter results for
the Y-Junction and T-Junction phase shifter...................................................................................... 20

Figure 3.6 Digital phase shifter. (a) 3-D structure, (b) top- layer. Dimension (mm) are: D9 = 3, D10 =
2.45, D11 = 2, D12 = 0.579, D13 = 0.422, D14 = 1.22, D15 = 2.15, D16 = 3, D17 = 0.08, D18 = 1.27, (c)
layout of SLPF, and (d) simulated performance of SLPF.................................................................. 21

Figure 3.7 Digital phase shifter. (a) 3-D structure, (b) top- layer. Dimension (mm) are: D9 = 3, D10 =
2.45, D11 = 2, D12 = 0.579, D13 = 0.422, D14 = 1.22, D15 = 2.15, D16 = 3, D17 = 0.08, D18 = 1.27, (c)
layout of SLPF, and (d) simulated performance of SLPF.................................................................. 23

Figure 3.8 3-D topology of the antenna array system ........................................................................ 25

Figure 3.9 Reflection coefficients of mode1s, 2 and 3. ..................................................................... 25

Figure 3.10 Simulated H-plane radiation of modes 1, 2 and 3. ......................................................... 26

XVI
Figure 3.11 The electric field snapshot of the design in E and H planes ........................................... 27

Figure 3.12 The electric field snapshot of the design in H planes. (a) Without absorber, and (b) with
two layers of absorbers. ..................................................................................................................... 27

Figure 3.13 Fabricated prototype. (a) Top view, (b) bottom view, and (c) prototype with absorbers.
............................................................................................................................................................ 28

Figure 3.14 Measured and simulated reflection coefficients of mode1s, 2 and 3.............................. 29

Figure 3.15 Normalised 28 GHz radiation pattern of mode1s, 2 and 3. (a) E-plane, and (b) H-plane.
............................................................................................................................................................ 30

Figure 3.16 Simulated gain and directivity of the design. ................................................................. 30

Figure 3.17 (a) Schematic diagram of the proposed tunable power divider, and (b) details of the
utilized L-shaped stub. Dimensions (mm) are: W0 = 0.4, W1 = 0.46, L1 = 2, W2 = 1.15, L2 = 2.67, Ws
= 0.29, W3 = 0.44, L3 = 0.25, W4 = 0.3, L4 = 1.27, W5 = 0.44, L5 = 0.45, W6 = 0.06, LC = 1.8, W7 =
0.14, S1 = 0.08, LS = 1.59. .................................................................................................................. 33

Figure 3.18 The required values for 𝑍2and 𝑍𝑇𝑆 versus𝜃2. ............................................................... 33

Figure 3.19 (a) Layout of the top layer, and (b) layout of bottom layer. ........................................... 35

Figure 3.20 Fabricated prototype. ...................................................................................................... 36

Figure 3.21 The performance at equal power division. ..................................................................... 36

Figure 3.22 The performance at the maximum tunable power division. ........................................... 36

Figure 3.23 Configuration of the proposed tunable feeding network with the dipole antenna subarray
in the top layer.................................................................................................................................... 38

Figure 3.24 Simulated S-parameters of the proposed feeding network with (a) equal output power,
and (b) unequal output power, which follows the Dolph-Chebyshev for 20-dB SLL suppression. .. 39

Figure 3.25 (a) |S11| of dipole antenna sub arrays with the tunable feeding circuit, and (b) normalised
E-plane radiation pattern at 28 GHz. ................................................................................................. 39

Figure 4.1 Schematic diagram of the polarisation-reconfigurable patch antenna. Dimensions (mm)
are: L1 = 3.3, L2 = 3.89, L3 = 1.1, L2 = 1.92. ..................................................................................... 42

Figure 4.2 Simulated surface current distributions of the proposed antenna with a period T at 28 GHz:
(a) t = 0, (b) t = T/4, (c) t = 2T/4, and (d) t = 3T/4. ........................................................................... 43

XVII
Figure 4.3 Simulated reflection coefficient results for LHCP and RHCP states of the proposed
antenna. .............................................................................................................................................. 44

Figure 4.4 Simulated axial ratio versus frequency plots for LHCP and RHCP states of the antenna.
............................................................................................................................................................ 44

Figure 4.5 Simulated radiation patterns for (a) RHCP, and (b) LHCP states of the antenna at 28 GHz.
............................................................................................................................................................ 45

Figure 4.6 Proposed antenna array arrangement for CP operation. (a) RHCP phase delay sequence,
and (b) RHCP phase delay sequence. 𝜃 is the reference angle. ......................................................... 46

Figure 4.7 Construction of branch-line couplers using an ideal diode with S-parameters and
transmission phase difference. ........................................................................................................... 47

Figure 4.8 Construction of feeding network using equivalent circuits of the diode with the (non-
optimised) S-parameters and transmission phase difference. ............................................................ 48

Figure 4.9 (a) The design of branch-line coupler with T-junction power divider, (b) simulated S-
parameters, and (c) simulated transmission phase. ............................................................................ 50

Figure 4.10 Orientation of diodes and SLPF. .................................................................................... 51

Figure 4.11 The geometry of the 4-element patch antenna array and its simulated reflection
coefficients. ........................................................................................................................................ 51

Figure 4.12 (a) Polarisation-reconfigurable antenna array, and (b) 3-D exploded view of layers. ... 52

Figure 4.13 Photograph of fabrication prototype. (a) Top view, and (b) bottom view...................... 53

Figure 4.14 (a) Simulated and measured S11 of the proposed antenna array, and (b) simulated and
measured AR. The shaded area is the targeted bandwidth. ............................................................... 53

Figure 4.15 Normalised simulated and measured radiation patterns at 28 GHz. (a) x-z plane, and (b)
y-z plane ............................................................................................................................................. 54

Figure 4.16 (a) Measured broadside gain of the proposed antenna array, and (b) measured efficiency.
The shaded area is the targeted bandwidth. ....................................................................................... 54

Figure 5.1 Key design requirements for 5G/4G MIMO antenna system. .......................................... 57

Figure 5.2 The geometry of a 4-element MIMO antenna system. (a) Top view (full design), (b) Ant-
1 (large view), and (c) tapered slot in the ground plane: Final optimised parameters with their values

XVIII
for MIMO antenna system are:𝐿𝑠𝑢𝑏 = 158, 𝑊𝑠𝑢𝑏 = 77.8, 𝑓𝑙1 = 1.89, 𝑓𝑙2 = 3.5, 𝑙1 = 13, 𝑙2 =
4.95, 𝑙3 = 27, 𝑙4 = 2, 𝑑1 = 2 and 𝑤1 = 1.17, all values are in mm. ............................................. 59

Figure 5.3 The geometry of 5G antenna. (a) Top view, (b) large view, and (c) 3D realised gain in dBi.
............................................................................................................................................................ 60

Figure 5.4 Construction of tapered slot (design cases and reflection coefficients). .......................... 60

Figure 5.5 (a) Integrated antenna structure without a low pass filter, (b) integrated antenna structure
with a low pass filter, (c) 5G reflection coefficient without/with low pass filter, and (d) 4G reflection
coefficient without/with low pass filter. ............................................................................................ 61

Figure 5.6 Current distributions at different stages (Stage1-Stage4) of the 4G feeder at 38 GHz. ... 62

Figure 5.7 Performance of low pass filter at (a) 4G band, and (b) 5G band. ..................................... 63

Figure 5.8 Photograph of the fabricated prototype. (a) Top view, and (b) bottom view. .................. 63

Figure 5.9 Simulated and measured S-parameters for 5G band. (a) 𝑆𝑖𝑖, and (b) 𝑆𝑖𝑗. ........................ 65

Figure 5.10 Simulated and measured S-parameters for 4G band. (a) 𝑆𝑖𝑖, and (b) 𝑆𝑖𝑗. ...................... 65

Figure 5.11 Normalised simulated and measured radiation patterns at different frequencies for 5G
bands (Ant-1). .................................................................................................................................... 66

Figure 5.12 Normalised simulated and measured Radiation patterns at different frequencies for 5G
bands (Ant-2). .................................................................................................................................... 67

Figure 5.13 Normalised simulated and measured Radiation patterns for 4G band. .......................... 67

Figure 5.14 Effect of LCD and hand/head phantoms on proposed design (a) 4G band, (b) 5G band at
different spacing between the LCD and the ground plane, (c) hand/head model, and (d) radiation
efficiencies at 4G (top) and 5G (bottom) bands. ................................................................................ 69

Figure 5.15 (a) Geometry of frequency-reconfigurable slot antenna, (b) 3D radiation patterns of slot
antenna with symmetry ground plane, and (c) 3D radiation patterns of slot antenna with modified
ground plane....................................................................................................................................... 73

Figure 5.16 (a) Simulated reflection coefficients of the 4G reconfigurable slot antenna with different
varactor capacitance Cv , and (b) simulated reflection coefficients at the lower (0.15 pF) and upper
limit (1.02 pF) of Cv with different values of Ledge. ........................................................................... 73

Figure 5.17 (a) Simulated reflection coefficient of each feeder for the CSAA without feeding network,
and (b) simulated current distributions on feeders and the ground plane. ......................................... 75
XIX
Figure 5.18 Performance of the 1 ˟ 8 power divider. (a) S-parameters, and (b) transmission phase. 76

Figure 5.19 (a) Geometry of connected slot antenna array, and (b) simulated reflection coefficient
with 3D radiation patterns after adding power divider at 28 GHz. .................................................... 76

Figure 5.20 The geometry of 2-element MIMO antenna system. ...................................................... 77

Figure 5.21 (a) Simulated isolation with and without filter when Cv = 0:15 pF, and (b) simulated
reflection coefficients for different values of Cv after adding the CSAA feeders. ............................ 78

Figure 5.22 Performance of the high pass filter incorporated into the CSAA feeding network. (a) Low
frequency, and (b) high frequency. .................................................................................................... 78

Figure 5.23 Photograph of fabrication prototype. (a) Top view, and (b) bottom view...................... 79

Figure 5.24 Simulated and measured S-parameters for low frequency band. (a) Reflection coefficient
curves, and (b) isolation. .................................................................................................................... 80

Figure 5.25 Simulated and measured S-parameters for high frequency band. (a) Reflection coefficient
curves, and (b) isolation. .................................................................................................................... 81

Figure 5.26 Simulated and measured normalised radiation patterns at low frequency band for Ant. 1.
............................................................................................................................................................ 82

Figure 5.27 Simulated and measured normalised radiation patterns at low frequency band for Ant. 2.
............................................................................................................................................................ 83

Figure 5.28 Simulated and measured normalised radiation patterns for Ant. 1 and Ant. 2 at 24 GHz.
............................................................................................................................................................ 83

Figure 5.29 Simulated and measured normalised radiation patterns for Ant. 1 and Ant. 2 at 28 GHz.
............................................................................................................................................................ 84

Figure 5.30 Simulated and measured, (a) Gain, and (b) efficiency at low frequency band............... 84

Figure 5.31 Simulated and measured, (a) Gain, and (b) efficiency at high frequency band. ............ 85

XX
List of Tables

Table 3.1............................................................................................................................................. 25

Table 3.2 Comparison between proposed and reported mm-Wave antenna beam-steering designs. 31

Table 4.1............................................................................................................................................. 47

Table 4.2............................................................................................................................................. 50

Table 4.3 Optimised Parameters of the final design (all values in mm) ............................................ 52

Table 5.1 Simulated and measured peak gain for 5G band ............................................................... 68

Table 5.2 Simulated and measured peak gain for 4G band ............................................................... 68

Table 5.3 Comparison between proposed and reported 5G/4G MIMO antenna systems.................. 70

Table 5.4 Optimised parameters used in the final design (all values are in millimetres). ................. 74

Table 5.5 Comparison between proposed and previous related works. ............................................. 86

XXI
Chapter 1 – Introduction
1.1 Background and Motivation
This introductory chapter presents the background and underlying motivation of this research for
building reconfigurable and MIMO antenna systems for 5G at millimeter-wave (mm-Wave). A
description of the current antenna systems for mm-Wave 5G and existing techniques of building such
systems are provided. The aims and original contributions of the thesis are also described. The
organisation of the thesis is outlined in the final section of this chapter.

1.2 Background and Motivation

The use of bandwidth for mobile networks has been increasing at an unprecedented rate, with
applications in emergent areas such as smart vehicles and civil infrastructure, telemedicine, and
remote education [1]. The predicted increase in mobile network utilisation requires a high data rate
capacity to transfer information through the network. The current limited capacity of the fourth
generation (4G) mobile network is not expected to satisfy the predicted high data demand by 2020
[2, 3]. To meet this expected requirement, a new fifth generation (5G) mobile network has been
proposed [4-6]. As the available bandwidth at low frequency is overloaded and unable to satisfy the
required wide bandwidth of 5G, many international organisations have suggested or licensed the mm-
Wave spectrum to be the operating frequency of 5G including at 24, 28, 37, 39, and 60 GHz [7-10].
Using mm-Wave raises many challenges such as an increase in the path loss, human body exposure
to and absorption of radiation, signal polarisation misalignment, increased MIMO antenna system
footprint inside the handset.

1.2.1 Beam-steering Antenna System Motivation and Advantages

One of the biggest challenges for 5G at mm-Wave will be the high propagation loss incurred [4, 11].
According to Figure 1.1 and the Friis transmission equation (Equation (1.1)), mm-Wave produces
high path loss compared with lower frequency band [12].

 C 
Pr  Pt  Gt  Gr  20log   dBm (1.1)
 4 Rf 

where Pr and Pt are the received and transmitted power respectively, in free space; Gr and Gt are the
receiver and transmitter antenna gains respectively; C is the speed of light, R is the distance between
the transmitter and receiver sources (base station and handset); and f is the carrier frequency.
Consequently, when f is higher (using mm-Wave), Pr will be smaller. Therefore, to compensate for
this reduction in signal strength of Pr several approaches can be suggested. First, R can be reduced–

1
but this is an unsuitable approach as it requires more base station to be built, and that increases the
infrastructure cost. Therefore, increasing Gt and Gr is the more reasonable option for enhancing Pr
and this can be done by building antenna arrays inside the handset instead of using a single antenna
as in current standard devices.

Figure 1.1 Typical mobile network.

The antenna arrays in mm-Wave need to be smart to direct the beam signal to different directions to
maintain a consistent connection between the handset and the base station. This capability can be
realised by adopting the beam-steering technique to control and tune the pattern direction of the
antenna array with wide scanning angle capability at mm-Wave 5G. However, realising a beam-
steering system with a wide scanning angle causes an increase in the side-lobe level (SLL) of the
antenna pattern. Therefore, to compress the SLL of the beam-steering system, a tunable power divider
is needed.

Another important challenge in using mm-Wave is the biological effect from the human body’s
exposure to electromagnetic (EM) radiation from the handset. The EM exposure level for mobile
technology absorbed by the human body can be examined by measuring two quantities: the specific
absorption rate (SAR) and equivalent power density quantities [13]. As the frequency increases, the
concentration of absorbed energy at superficial parts of the human body increases [14]. For instance,
as shown in Figure 1.2(a), the penetration depth in mm is reduced with the increase in frequency. As
a result, it is found that more than 90% of the power is concentrated in the skin layers (epidermis and
dermis) at 60GHz (Figure 1.2(b)) [14].

2
Figure 1.2 (a) The penetration depth (mm) with frequency [14] © 2017 IEEE, and (b) power density
(SAR) concentration in the skin at 60 GHz.

Because of the above, reducing the human head’s exposure to EMF is one of the most important
factors for ensuring the safety of emerging 5G technologies. As such, using the beam-steering
technique to reduce the amount of EM radiation directed into the user’s body is required. Thus,
building a beam-steering antenna system inside the mobile handset for mm-Wave 5G is a key factor
for maintaining a reliable connection with the base station at a high data rate, while also reducing the
human head’s exposure to EMF at the same time.

1.2.2 Reconfigurable Circular Polarisation Antenna System Motivation and

Advantages
One of the technical challenges of using mm-Wave for 5G is the polarisation mismatch effect between
the transmitting and receiving antennas. Generally, there are different types of polarisation in mobile
communications, including linear polarisation and circular polarisation. Circular polarisation is
divided into two rotating modes: right handed circular polarisation (RHCP) and left handed circular
polarisation (LHCP). Unlike linear polarisation, circular polarised radiation is an effective technique
to mitigate the transmitting and receiving antenna misalignment problems or any polarisation
mismatch. Moreover, circular polarisation offers advantages for mm-Wave 5G over linear
polarisation. It suppresses multi-path effects and phase deviations due to rain and snow. [15, 16].

More importantly, in the case of 5G device-to-device (D2D) communication [17] between two mobile
devices, polarisation efficiency can be extremely low if these devices work in opposing rotating
modes. To resolve this issue, antennas with reconfigurable circular polarisation modes should be
3
used. To this end, reconfigurable CP antennas are an important factor in mitigating the mismatch in
high mobility situations.

1.2.3 5G/4G MIMO Antenna System for 5G Handsets Motivation and Advantages
The multiple-input-multiple-output (MIMO) antenna system is seen as a key technology for an
upcoming 5G and the 4G mobile antenna system to enhance the data rate [5, 18, 19]. As mentioned
earlier, since 5G will work at mm-Wave, the current mobile handset needs to evolve to work at mm-
Wave as well as at the current sub-3 GHz band. This means more MIMO antennas need to be
integrated inside the handset chassis to cover different technologies and spectrums. In other words,
antenna footprints become large, considering the limited space inside the handset. This disadvantage
can be mitigated by developing a MIMO antenna system consisting of antenna elements that can
cover multi-bands for 5G and 4G by using a single structure rather than two structures [14, 20-22].
To this end, MIMO 5G/4G is one of the key factors in the future 5G handset industry.

1.3 Aim of the Thesis

This dissertation is aimed at developing new reconfigurable and MIMO antenna systems for mm-
Wave 5G. For state-of-the-art technologies at mm-Wave 5G, although some promising designs have
been reported, they are still at an early-stage and more novel improvements are required. Since 5G
will work at mm-Wave, the following problems are investigated in the thesis and can be summarised
as follow:
1- Increased path loss and EM absorption by superficial parts of the human body;
2- Increased polarisation misalignment at mm-Wave 5G devices;
3- Increased necessary MIMO antenna system footprint for 5G and 4G inside the handset
chassis.
Multiple solutions will be provided in this thesis to cope with the mm-Wave at 5G challenges. After
investigating previous work undertaken by other researchers, new designs have been proposed and
verified.

1.4 Original Contributions of the Thesis

This research work undertaken to resolve the problems listed above, results in the following original
contributions:

1- Design of a 3-D Beam-steering System for mobile Handsets at 28GHz. This contribution
includes
a. a 28 GHz 3-D beam-steering system. The proposed system has the ability to steer the beam
in three dimensions by utilising both sides of a printed circuit board for a two-dimensional

4
 antenna array. Additionally, the beam-steering technique aims at reducing
electromagnetic fields exposure from the mobile handset to human head as well as keep a
reliable connection with the base station;
b. a tunable in-phase power divider for 5G at 28 GHz. Apart from the beam-steering system,
this work presents a tunable power divider that can be used in the feeding network to
reduce the SLL of antenna radiation from the antenna array.
2- Design of two reconfigurable CP antennas for 5G at 28 GHz. This includes
a. design of a single patch antenna with the functionality to switch between LHCP and RHCP
using two p-i-n diodes and a simple feed circuit;
b. design of a 4-element antenna array that can switch its operation between two rotating
modes (RHCP and LHCP) using only 4 p-i-n diodes.
3- Design of MIMO Antenna Systems for mm-Wave 5G and 4G handheld devices. The
contribution includes
a. a MIMO antenna system for multi-band mm-Wave 5G and wideband 4G applications. The
design integrates the requirements of both 5G and 4G antennas using only a single structure;
b. an integrated frequency-reconfigurable 2-element slot antenna and connected slot antenna
array for 4G and mm-Wave 5G mobile handsets. This design integrates 5G and 4G operations
into a single structure. The proposed antenna system is compact and planar in structure, and
so has potential for future mobile handheld devices.

1.5 Thesis Organisation

The research work undertaken in this thesis is organised as follows:

Chapter 1 presents the background and underlying motivation behind the current research. The
chapter starts with a discussion of the necessity of inventing 5G technology at mm-Wave.
Additionally, the challenges of using mm-Wave bands for mobile communication are discussed and
compared with the existing mobile technologies at low frequency. The aims and original contributions
of the thesis are also outlined in detail.

Chapter 2 overviews the different antenna systems for mm-Wave 5G applications, including beam-
steering, reconfigurable CP and MIMO antenna systems. The existing beam-steering antenna
systems are discussed first. The existing power divider designs in beam steering systems are also
reported. Afterwards, the state-of-the-art development of reconfigurable CP antennas is discussed,
including their limitation at mm-Wave 5G. The chapter ends with a discussion of the challenges that
5G and 4G MIMO antenna systems face in achieving a compact structure inside the handset.

5
Chapter 3 presents the design 3-D beam-steering antenna system at 28 GHz. It also investigates the
effect of EM on the human head. A design for a tunable power divider for the beams-steering system
is also introduced. Lastly, a feeding tunable circuit is proposed to validate the proposed tunable
divider’s anticipated suppression of the beam-steering antenna system’s SLL.

Chapter 4 introduces two designs of reconfigurable CP antennas applicable to mm-Wave 5G. First, a
reconfigurable CP patch antenna is proposed. Second, a 4-element antenna array with a
reconfigurable feeding circuit is designed and prototyped. Both designs present a novel technique for
building reconfigurable CP antenna structures, using fewer p-i-n diodes.

Chapter 5 presents the design and development of two 5G/4G MIMO antenna systems combined in
a single antenna structure. The chapter starts with the design of a 4-element 5G/4G MIMO antenna
system. Later, a design for a 2-element antenna array for a 5G/4G MIM system is presented. Both
designs for future mobile handheld devices are prototyped and measured to validate their operating
concept at mm-Wave and sub-3 GHz bands.

Chapter 6 concludes the thesis with suggestions for future work.

6
Chapter 2 – Literature Reviews of Reconfigurable and MIMO Systems
at mm-Wave 5G
This chapter presents an overview of different antenna systems for 5G applications at millimetre-
wave (mm-Wave), beginning with a discussion of existing beam-steering antenna systems and their
limitations at mm-Wave 5G mobile handset. Additionally, the existing power divider designs in beam
steering systems at mm-Wave 5G are reported and the challenges in making the power division ratio
tunable are elaborated. In the next section, the state-of-the-art development of reconfigurable circular
polarisation (CP) antennas is discussed, including their limitation at mm-Wave 5G. Afterwards, the
current MIMO antenna array technology for mobile handheld devices is presented. The challenges
facing the current MIMO technologies when using mm-Wave 5G handsets are described as well.
However, when investigating all literature related to this topic for this chapter, the number of research
projects to be considered was found to be limited due to the near-standardisation of 5G at mm-Wave
band.

2.1 Design Technology for Beam-steering Antenna Systems at mm-Wave 5G

Handset
One of the key factors in designing a 5G mobile handset at mm-Wave, is building a beam-steering
antenna system. The beam-steering system must maintain a reliable connection with the base station
at high data rate [14] while reducing the electromagnetic field (EMF) exposure to the human head at
the same time [23].

Generally, there are several kinds of beam-steering techniques for an antenna array, including:
manual, mechanically steerable antennas; digital tuning arrays; and electrically controlled phased
arrays. Of all these techniques, beam-steering based on a phased array is the most widely used
technique for the mm-Wave 5G handset. The advantages of using phased arrays include their fast
radiation patterns, controllability, compact size, and seamless integration with other components of
the mobile handset. A block diagram of a typical phased array antenna system is shown in Figure 2.1.
Following the sequence from the input RF signal to the antenna array, there are power dividers and
phase shifters to control the amplitude and the feeding excitation phase of the antenna array elements.
The phase is used to tune the direction of the beam in the desired direction and the amplitude is used
to suppress the side-lobe level (SLL). It is worth mentioning that planar antenna arrays can generate
three dimensional (3-D) beam-steering, while linear antenna arrays generate two dimensional (2-D)
beam-steering.

7
Figure 2.1 Block diagram of beam-steering antenna array system.

In a mobile network, when users use the handset, they can move or orient the handset unpredictably.
Compared to two dimensional (2-D) beam-steering, three dimensional (3-D) beam-steering in
handsets enables better immunity to EMF exposure of the head, and maintains consistent network
quality with the base station regardless of mobile orientation.

Recently, several promising beam-steering designs have been proposed for mm-Wave 5G
applications. Among them, 2-D phase shifter based designs have been proposed. A piezoelectric
transducer controlled shifter (±2.3°) is deployed for mm-Wave 5G [24]. A proposed 28 GHz system
with simulated results is reported in [25]. It uses two eight-element slot antenna arrays in a metallic
handset frame, fed by an RF integrated circuit (RFIC chip) with beam steering in 2-D (±60°). A
number of promising 3-D designs also exist, that utilise multiple feed ports [26, 27] (Figure 2.2(a)),
or large feed networks [28] (Figure 2.2(b)), with wide 3-D steering angles.

However, integrating a mm-Wave beam steering design in a mobile phone handset is a complicated
challenge, given the required wide scanning angle capability and limited space in the mobile phone.
Therefore, the aforementioned designs have their own limitations for 5G mobile handset applications.
The phase shifter based designs tend to have limited steering angles [24]. The RFIC chip base
approach shows the measured result of only a single antenna element. The 3-D designs rely on using
multiple port feeds [26, 27], which can be too large for mobile handset integration, as in [28].

The previous designs have focused on building a beam-steering antenna array for mm-Wave 5G.
However, there is another important aspect of using mm-Wave that requires considerable attention,
8
which is the strong potential interaction between the radiated fields that come from the antenna at
mm-Wave band [14], and the strongly absorbing surface layers of the human body [23]. As such,
investigation into using beam steering to reduce the amount of EM radiation directed into the user's
body, is of paramount interest. Hence, a 28 GHz 5G beam-steering array antenna was proposed [23]
to reduce EM exposure of the human head by using two subarrays located on the metal frame of the
handset. However, that proposed design was built with a fixed beam direction and 2-D beam-steering
was only performed in a simulated model. After reviewing all previously reported approaches, it has
become evident that there are still some challenges remaining in the building of a beam-steering
antenna array for the 5G handset that must be addressed.

Figure 2.2 (a) Multiple feed port antenna array [26] © 2016 IEEE, and (b) large feeding circuit of
antenna array [28] © 2018 IEEE.

In relation to the above, it is nevertheless realised that a beam-steering system with a wide scanning
angle causes an increase in the SLL of the antenna pattern. Therefore, to compress the SLL of the
beam-steering system, and thus reduce interference, a tunable power divider is needed. The function
of the tunable power divider is to control the power distribution of the antenna elements in a certain
rule, such as Dolph-Chebyshev distribution, binomial distribution, Villeneuve distribution, Taylor
distribution etc. [29]. As a result, the SLL can be suppressed to a low level at the expense of a slight
increase in the angle on either side of the main beam [29]. Several power dividers have recently been
proposed for mm-Wave band applications [24, 30-33]. In [24], an electromagnetic bandgap (EBG)
was integrated with a Y-junction divider to feed a 5G antenna array at 60 GHz. Additionally, a typical
1-4 Wilkinson power divider deployed to an antenna array operating at 28 GHz for 5G applications,
was presented [30]. A 4-element beamformer for a phased array was designed by using SiGe
BiCMOS technology at 28 GHz [31]. The beamformer included two cascaded 1:2 Wilkinson power

9
dividers as a part of the feeding circuit. A modified Wilkinson power divider with a frequency
dependent isolation network was utilised in [32] to improve the isolation between the output ports in
a wide band (25 to 45 GHz). However, these reported designs were all fixed with either equal or
unequal power division ratios. Building a tunable power divider requires tuning the equivalent
impedance of the power divider’s branch lines, whilst matching all of the ports and providing the
required isolation between the output ports across the whole tuning range throughout the operating
bandwidth. To fulfil these requirements, efficient matching techniques must be used. It should be
noted that working at mm-Wave band can be difficult with respect to manufacturing and avoiding the
parasitic effects of the lumped components at such a high frequency. Thus, when reviewing the
literature, there is no surprise at not finding any tunable power divider designed for 5G at mm-Wave
band [20].

2.2 Design Technology for Reconfigurable Circular Polarisation Antenna at

mm-Wave for 5G Applications
Circular polarisation (CP) is a key feature of 5G antenna design at mm-Wave to mitigate
misalignment and multipath interferences. As stated previously, to avoid low polarisation efficiency
in mobile communication, circular polarisation (CP) antennas with a switchable rotating mode feature
need to be used. Using electronic components such as p-i-n diodes, is one of the most popular ways
to realise reconfigurable circular polarisation antennas since they can be easily integrated into the
antenna structure. To this end, reconfigurable CP antennas using p-i-n diodes at low frequency have
been reported extensively in the literature. By integrating p-i-n diodes in the antenna structure and
using patch antennas [34], slot structures [35, 36], the waveguide concept [37, 38], or loop antenna
[39], reconfigurable polarisation antennas have been realised. Alternatively, the reconfigurable
polarisation feature can also be realised by using p-i-n diodes incorporated in the feeding circuit of
the antenna array.

10
Figure 2.3 (a) 1 × 4 CP reconfigurable antenna array in [40] © 2018 IEEE, and (b) diode layout that
produces the polarisation switching mechanisms.

In [41], a square patch with a T- shaped feeder and p-i-n diodes was proposed to realise polarisation
(LHCP). Additionally, an annular patch antenna was proposed in [42] to yield RHCP and LHCP using
a metasurface and a reconfigurable feeding network. The work in [43] proposed a reconfigurable
feeding network that generates two orthogonal circular polarisations. All these designs utilised p-i-n
diodes that exhibited nearly ideal performance at low microwave frequency range, i.e. nearly an open-
circuit in the OFF state and a short-circuit in the ON-state with low insertion loss and high isolation.
There are two critical issues when adapting these designs for 5G applications at mm-Wave. First, 5G
antennas require high gain; thus, often an array configuration is used, which requires a large number
of p-i-n diodes and biasing components, as shown in Figure 2.3. More importantly, p-i-n diode
performance is significantly degraded when it is used at mm-Wave [44], which severely affects the
antenna performance. It is noted that mm-Wave CP antennas have been studied; however, the
operation modes was limited to one mode of CP [45-48], or to multiple input ports to switch between
two CP modes, which adds complexity when integrating with a feeding circuit [49, 50]. From the
previous literature review, it is clear that reducing the number of diodes is desirable, and subsequently
improves matching and reduces the insertion loss of the 5G antenna at mm-Wave.

2.3 Design Technology for MIMO 5G/4G Antenna Systems of Mobile

Handheld Devices
In mobile communication, MIMO antenna systems deploy multiple antennas to send and receive
multiple independent signals at the same time. This helps in achieving high data density and
improving the link budget more than the traditional single-input, single-output (SISO) channels [51].
Therefore, for mm-Wave 5G, it is strongly expected that MIMO technology, which has already been

11
used by 4G, will be one of the most compelling components of the mobile network [52]. As mentioned
earlier, many international organisations have suggested or licensed the mm-Wave spectrum as the
operating frequency of 5G, including 24, 28, 37, 39, and 60 GHz [7-10]. To that end, it is paramount
to evolve smartphones to incorporate mm-Wave 5G antenna with existing sub-6-GHz antennas inside
the handset, as exemplified in Figure 2.4. However, mobile terminals have limited space and adding
more antenna elements requires extra space, which is a significant challenge in the smartphone
industry. Therefore, the solution is to combine 5G and 4G antennas in one structure rather than two
structures [14, 20-22].

Figure 2.4 Illustrative diagram of the future handset incorporating mm-Wave 5G antennas with 4G
and 3G antennas at sub-6-GHz [14] © 2017 IEEE.

Recently, many research efforts have been invested in trying to satisfy the 4G and/or 5G antenna
requirements for MIMO application at the sub-6 GHz band. Some researchers have built MIMO
antenna systems for 5G and 4G by using separate antenna structures [53-55]. Additionally, 5G and
4G antenna models for LTE bands 42/43 and LTE band 46 have been proposed, utilising an inverted
L-shaped open slot antenna [53] or the bent strip structure in [54]. C-shaped coupled feed and L-
shaped monopole slot approaches were investigated in [55] at the sub-6 GHz spectrum. Some other
researchers have further miniaturised the size of the MIMO antenna footprint by using a single
antenna structure with multi-band or wideband bandwidth to cover both 5G and 4G bands. For
instance, T-shaped or coupled-feed approaches were investigated in [19, 56, 57] at the sub-6 GHz
spectrum for 5G or 4G applications. However, there are critical issues when adapting these designs
to 5G at mm-Wave and to 4G at sub-3 GHz in the MIMO system. The operating frequency of 5G at
mm-Wave is far from the operating frequency of 4G, which makes it difficult when designing a single
12
antenna structure to cover such a wide bandwidth. More critically, due to the high path loss at the
mm-Wave band, the gain from the 5G antenna should be high compared with that from the 4G. Hence,
it is comparatively hard to utilise a single structure to generate a high gain at mm-Wave and low gain
at sub-3 GHz. To that end, it is noted that only a few MIMO antenna designs have been built for 5G
at mm-Wave and 4G at sub-3 GHz utilising two separate antenna structures [58, 59] or a single
antenna structure [60]. Unfortunately, those reported designs are large in terms of the scarce real
estate of the mobile terminal.

2.4 Summary
In this chapter, the existing state-of-the-art designs of beam-steering and reconfigurable CP and
MIMO antenna systems were described at mm-Wave 5G. Examples of antenna design at low
frequency were given, and the challenges of upgrading them to the mm-Wave band were outlined. It
has been clearly shown that although there has been some promising research work done recently in
designing antennas at mm-Wave, extensive research is still required to improve the performance of
the 5G antenna. In terms of beam-steering, the 5G antenna system needs improving with respect to
the beam-steering angle, reducing SLL, and the potential interaction between the radiated fields from
the antenna to the human body. Additionally, reducing the number of used diodes in building
reconfigurable CP antenna arrays is required at mm-Wave to improve the overall performance of the
antenna. Furthermore, it is inferred from the literature review that the current MIMO antenna systems
need to evolve to cover mm-Wave 5G and to be compact because of the scarce real estate of the
handset. In the following chapters, various antenna designs at mm-Wave that overcome these
limitations will be presented.

13
Chapter 3 -Three-Dimension Beam-steering System for Mobile
Handsets at 28 GHz
3.1 Introduction
The use of bandwidth for mobile and cellular networks has been increasing at an unprecedented rate,
with many emergent applications. The predicted increase in mobile network utilisation requires a
high data rate capacity. The current limited capacity of the fourth generation (4G) mobile network is
not expected to satisfy the predicted high data demand by 2020. To meet this requirement, a new fifth
generation (5G) mobile network has been proposed [4]. In order to reach high capacity demands, 5G
requires a wide frequency bandwidth, which is not available in the sub-3 GHz frequency spectrum,
one of the potential frequency band candidates for 5G mm-Wave around 28 GHz [2, 61].

At mm-Wave, the path loss can be particularly high, having a significant impact on the transmitted
or received mobile signal [4, 11]. To counteract this loss, a directive antenna could compensate the
high path loss of the link. Therefore, the concept of a directive antenna with beam-steering capability
has attracted considerable attention at mm-wave, compared with the sub-6 GHz frequency band.
Integrating an mm-wave beam steering design into a mobile phone handset is a complicated
challenge, with consideration needed for the limited available space, the required integration with
other components of the handset and the circuit complexity of the full beam-steering system. In
addition, at mm-Wave, the radiated fields from the antenna are strongly absorbed by the surface layers
of the human body. Therefore, using the beam steering to reduce the amount of EM radiation directed
into the user’s body is of paramount interest. Thus, a key factor in developing a complete beam-
steering system for mobile handsets is to keep a reliable connection with the base station at a high
data rate, while reducing the EMF exposure to the human head at the same time.

In this chapter, a novel 28 GHz 3-D beam-steering system is proposed first. It meets the requirements
necessary to perform 3-D beam steering at a wide angle; and it also investigates the effect of EM
exposure on the end user. The 3-D beam-steering capability is achieved by combining two subarrays.
For demonstration of the concept of the beam-steering capability, a full circuit of 2-D beam-steering
prototype is built. The full beam-steering circuit consists of a two-layer dipole antenna array at the
two sides of the main PCB, two Wilkinson power dividers, two digital phase shifters (which use a
novel technique of integrating the p-i-n diodes in the feeding circuit) microstrip-to-coplanar stripline
baluns and a novel two-layer slot-microstrip power divider. Moreover, to further improve the
performance beam-steering system, a tunable power divider for 5G beam-steering feeding circuit
forming, has been presented. The tunable power divider is used to tune the feeding amplitude of an

14
antenna array. Last but not least, a feeding tunable circuit is proposed to validate the potential of using
the proposed tunable divider to suppress the side-lobe level (SLL) of the antenna radiation pattern.

3.2 Millimeter-Wave Three-Dimensional Beam-Steering System for 5G

Handsets
3.2.1 Design of the 3-D Beam-Steering System
The geometry of the antenna arrangement for the proposed 3-D beam-steering system is illustrated in
Figure 3.1. The system uses (4×2) dipole antenna subarrays, microstrip-to-coplanar stripline baluns,
and 50 Ω microstrip line. It has 4.9 mm centre-to-centre spacing between adjacent dipoles (0.45 λ at
28 GHz). The system is built using two substrates with three layers (top and bottom layer that are
separated by a ground layer). Each of the top and bottom layers contains 4×1 dipole antennas.
Microstrip-to-coplanar stripline baluns are used to connect the antenna array to the feeders [62]. Each
antenna element is composed of printed dipole with a truncated ground plane that works as a reflector.
The structure is printed on dielectric substrate Rogers TMM6 with relative permittivity 𝜀𝑟 = 6, loss
tangent of tan 𝛿= 0.0023, and thickness = 0.381 mm. The two-dipole sub-array technique enables the
system to steer the beam in the H-(yz) plane, while the 4×1 antenna arrangement in each layer enables
the beam steering capability in the E-(xy) plane. Therefore, combining both planes gives 3-D beam-
steering capability. The antenna is characterised via HFSS to ensure good impedance matching at 27.5-
28.5 GHz. Figure 3.2(a) represents the simulated performance of the 4×2 antenna array. The bandwidth
for a reflection coefficient < −10 dB is (27.5–28.5 GHz). Figure 3.2(b) shows the normalised radiation
pattern of the antenna array for the E-(xy) and H-(yz) planes. It is noted that the array system produces
typical end-fire radiation when all the ports are excited with the same phase and amplitude.

To study beam steering capability of this design, the eight ports of the antenna array are fed with the
same power, while the steering is achieved by controlling the phase at each port. The 3-D radiation
pattern of the antenna with relevant gain values at different beam steering angles at 28 GHz are shown
in Figure 3.3. This design has the ability to yield beam steering of up to ±55° in the H-plane. It is
worth mentioning that the maximum beam steering for the antenna beam-steering system without
using balun in the E-plane is ±85. However, only ±35° is achieved once balun is employed to feed
the differential dipole antennas. The simulated maximum gain is 10 dBi.

15
Figure 3.1 Dipole antenna array structure. Dimensions (mm) are D1 = 2.18, D2 = 0.3, D3 = 0.85, D4
= 0.59, D5 = 0.5, D6 = 0.65, D7 = 0.1, D8 = 0.2.

Figure 3.2 (a) Simulated reflection coefficients of the antenna array, and (b) radiation pattern.

16
Figure 3.3 3-D radiation pattern at different beam steering angles.

3.2.1.1 Reduction of Electromagnetic Field Exposure by using beam-steering

Using a higher frequency (mm-Wave) band increases the concentration of absorbed energy at
superficial parts of the human body (including the head), which might cause superficial burns [14].
Therefore, reducing EMF exposure to the human head is one of the most important factors for
ensuring the safety of emerging 5G technologies. One of the main advantages of the proposed design
is to resolve the aforementioned EM exposure to the user head, by steering the beam towards the base
station.

17
The EM exposure level at low frequency for current technologies, including 3G and 4G, is examined
by using the specific absorption rate (SAR). However, at mm-Wave, the regulation regarding EM
exposure was changed from a SAR limit to power density (PD) as per the Federal Communications
Commission (FCC) and International Commission on Non-Ionising Radiation Protection (ICNIRP)
[13]. Therefore, in this design, the PD is evaluated rather than the SAR [63].

In order to investigate the EM exposure using the proposed antenna array design, an MRI-derived
realistic human head phantom with the proposed antenna array of Figure 3.1 is simulated by using
the electromagnetic tool, CST Microwave [64]. The setup for the investigation is shown in Figure
3.4(a). By considering the thickness of the handset’s chassis, a separation distance of 5 mm is
maintained between the antenna array and the head phantom. Different tissues of the head phantom
are defined by using their frequency dispersive dielectric properties over the band of 27.5-28.5 GHz
[65].

Figure 3.4 Simulated power density (PD) distribution in the head phantom for the proposed antenna
array system at 28 GHz.

The simulated PD on the skin of the head is calculated for five beam steering angles (modes) as shown
in Figure 3.4(b-f). These modes are selected in talk mode to cover any possible scenarios of using the
handset while the antenna beam is directed to the base station. PD distribution in modes 1, 2 and 3
are shown in vertical cross section, while modes 4 and 5 are in the horizontal one. In mode 1, when
the beam is directed to the Z axis, PD is in the range of 1000 to 1400 W/m2, as shown in Figure 3.4(b).
However, when the beam is tilted to 55° towards the human head in mode 2, the power density
18
becomes maximum at peak 1540 W/m2 as expected–see Figure 3.4(c). On the other hand, when the
beam is tilted to -55° in mode 3, the peak value does not exceed 50 W/m2, as indicated in Figure
3.4(d). This significant reduction is caused by tuning the beam to the opposite direction of the head.
Similarly, the values of PD of modes 4 and 5 are very close to those of mode 3, which implies that
beam steering in the H-plane is more critical than in the E-plane for reducing EM exposure value
towards the head. As for the H-plane, E-plane beam steering is important to keep consistent
connection between the handset and the base station. Based on information provided in Figure 3.4(g),
the beam steering technique of this design succeeds in redirecting around 92% of the power radiated
from the antenna array away from the head and toward the base station. If the beam steering technique
of this design is used, any possible adverse health effects on the human head are significantly reduced.

It is to be noted that the available thresholds of PD released by FCC or ICNIRP are just for far-field
scenarios. They do not provide guidelines to evaluate power density near human bodies, nor do they
provide relevant rationale [66]. For that reason, the evaluated PD in Figure 3.4(g) cannot be claimed
to satisfy the requirement of the standards threshold of PD in the near-field. However, the proposed
3-D beam-steering provides an indication to mobile manufacturers of the possibility and limits of
increasing the transmitted power from the handset to compensate the path loss at mm-Wave, while
maintaining the EM exposure within the safe levels.

For demonstration of the concept of this design, a 2-D beam antenna beam-steering is built, including
this 4⨯2 antenna array, a well-known Wilkinson power divider [67], proposed ±90° digital phase
shifter and multilayer power divider.

3.2.2 Digital Phase Shifter

The traditional switched-line phase shifter consists of four diodes to control the flow of signal
between two transmission lines of different lengths. The resulting difference in the effective length
then produces the required phase shift [67]. However, reducing the number of diodes (switches) at
mm-Wave provides a significant advantage in reducing the power consumption and design
complexity. Therefore, the stub switched phase shifter technique with just three diodes [68] is used,
as shown in Figure 3.5(a). In [68], T-type p-i-n diode connections are adopted to connect the diodes
with the transmission line. Using this junction type at high frequency (mm-Wave) produces more
spurious signals, narrow bandwidth and high loss. This is because of the discontinuity at the T-type
connections and the bends generating high order modes. In addition, the diodes at mm-Wave have
poor performance compared with those at low frequency [67].

19
Figure 3.5 (a) T- Junction phase shifters, (b) Y- Junction phase shifters, (c) simulated spurious
radiation for the Y-Junction and T-Junction phase shifter, and (d) simulated S-parameter results for
the Y-Junction and T-Junction phase shifter.

To cope with this problem, and after doing a parametric study of possible diode orientations, the use
of Y-type p-i-n diode connections, instead of the T-type connections, is proposed (Figure 3.5(b)), for
connecting the diodes with microstrip lines. Using Y-type p-i-n diode connections reduces the
spurious propagation emitted from the phase shifters that overlap with the desired radiation from the
antenna array, as shown in Figure 3.5(c). As seen in this figure, using T-type p-i-n diode connections
generates around a 5 dB spurious signal, while the Y-type reduces this signal to around 2 dB, although
this level still causes some adverse effects on the performance, as discussed in Section 3.2.2. In
addition, a significant improvement in matching bandwidth and insertion loss is obtained by using Y-

20
type rather than T-type p-i-n diode connections, as shown in Figure 3.5(d). The Y-type achieves a
better insertion loss (less than 2.12 dB) and wider impedance bandwidth than the T-type.

Figure 3.6 Digital phase shifter. (a) 3-D structure, (b) top- layer. Dimension (mm) are: D9 = 3, D10 =
2.45, D11 = 2, D12 = 0.579, D13 = 0.422, D14 = 1.22, D15 = 2.15, D16 = 3, D17 = 0.08, D18 = 1.27, (c)
layout of SLPF, and (d) simulated performance of SLPF.

21
Two switched-line phase shifters are designed at 28 GHz to feed the two layers of the antenna array
and produce 0° and ±90° phase differences between these two layers to enable steering in H-plane.
The phase shifter is built into each side of the substrate, as shown in Figure 3.6(a). Figure 3.6(b)
shows the top layer of the structure, which is in symmetry with the bottom layer. p-i-n diodes (MADP-
000907-14020W) from M/ACOM [69] are used in this design. The on-state equivalent circuit is
(𝑅𝑜𝑛 = 3Ω, 𝐿𝑜𝑛 = 0.045 𝑛𝐻 ) and the off-state equivalent circuit is ((𝑅𝑜𝑓𝑓 = 300𝑘Ω, 𝐶𝑜𝑓𝑓 = 42𝑓𝐹 )
[70]. As part of the biasing circuit, two DC power supplies 𝑉1 and 𝑉2 and DC block capacitor 𝐶1 of
0.01 µ𝐹, are used. The available RF chokes that work at mm-Wave are relatively large [4], therefore
four stepped-impedance low pass filters (SLPF) are employed instead. The SLPF works ideally like
a short circuit for DC power supply and an open circuit for RF signals at mm-Wave. The SLPF
consists of a quarter wavelength high-impedance transmission line and a low-impedance pad. The
SLPF structure and its simulated performance are shown in Figure 3.6(c). The RF signal is blocked
from flowing between ports 1 and 3 by a more than 22 dB insertion loss, whereas the required mm-
Wave flow is enabled between ports 1 and 2 with only 0.4 dB loss over the operating frequency band.
In addition, two quarter wavelength transformers are used at each end of the phase shifter to improve
the matching and reduce the insertion loss.

3.2.3 Two-Layer Power Divider

This proposed design consists of a three-conductor layer structure. The input port and one of the
output ports are located in the top layer; the middle layer represents the ground layer; and the second
output port is located in the bottom layer, as shown in Figure 3.7. These three layers are separated by
two dielectric layers. Unlike [71] that has the output ports in the same layer, this design has the output
ports in different layers (top and bottom layers), which is necessary for the proposed antenna array
structure. To feed the input power to the two output ports in different layers, microstrip to slotline
transitions are employed. The narrow slotline of this transition is used to direct the signal from the
input to the two arms of the output ports with the same phase, as can be seen in Figure 3.7(b). The
input microstrip line is ended with a capacitive circular stub, while the slotline is ended with an
inductive circular slot. These capacitive and inductive stubs are used to couple the signal between the
microstrip and slotline. The length of the slot is determined to be approximately a quarter of the
effective wavelength at the centre frequency of operation (28 GHz) [72]. To have high matching in
the microstrip ports, impedance of the slot as seen from the microstrip line, is equal to (𝑍𝑠 𝑁 2 ), where
the impedance transformation ratio 𝑁 and the slot impedance 𝑍𝑠 depend on the slot width and the
substrate characteristics as given in [73, 74]. In this design, the slot width is chosen to give an
impedance of 100 Ω and 𝑁 = 0.91. As a result, the impedance of the slot, as seen from the microstrip
22
line, is (100 ⨯ 0.912 = 82 Ω). This impedance is relatively high due to the fabrication limitation of
slot width at mm-Wave. To improve the matching, a quarter wavelength transformer is added to the
microstrip feeding transmission line, as shown in Figure 3.7(c). Figure 3.7(e) represents the
performance of this power divider. It yields reflection loss of more than 11 dB over the designed
bandwidth (27.5-28.5) GHz, with an insertion loss of 1.2 dB.

Figure 3.7 Digital phase shifter. (a) 3-D structure, (b) top- layer. Dimension (mm) are: D9 = 3, D10 =
2.45, D11 = 2, D12 = 0.579, D13 = 0.422, D14 = 1.22, D15 = 2.15, D16 = 3, D17 = 0.08, D18 = 1.27, (c)
layout of SLPF, and (d) simulated performance of SLPF.

23
3.2.4 Layout of the Antenna Array System
The geometry of the whole antenna array system, after integrating all components, is illustrated in
Figure 3.8. It is planar, compact and fits within (35.8 𝑚𝑚 ⨯ 24 𝑚𝑚). It includes 4⨯2 diploe antenna
arrays, balun, two Wilkinson power dividers, switched-line phase shifters, a multilayer power divider
and the feeding line. The planar feature allows other parts of the handset’s components to be easily
integrated into the design. Because the two digital phase shifters can produce 0° and ±90° phase
differences, the fabricated prototype can yield three modes of operation. These modes can steer the
main beam of the antenna in three directions in the H-plane, as shown in Table 3.1. The S-parameters
of the design for the three operating modes are shown in Figure 3.9. In all these modes, the reflection
coefficient is less than −10 dB across the 27.5–28.5 GHz band. Modes 1 and 3 have similar reflection
coefficient plots.

The simulated H-plane radiation patterns for the three modes are shown in Figure 3.10. The antenna
system successfully steers the beam in three directions (−55°, 0°, 55° ). However, there are some
undesired dips or back loops in the pattern radiation and these are denoted by the red circles in Figure
3.10. These do not exist in the simulated antenna radiation of the dipole antenna array without the
feeding circuit, as shown in Figure 3.2(b), which indicates that the feeding circuit, including the power
dividers and the digital phase shifter, cause these dips. To investigate the cause of these dips, the
electric field distribution of the whole design, including the antenna arrays and the feeding circuit, is
calculated using HFSS, as shown in Figure 3.11. By focusing on mode 1 in the H-plane as an example,
the spurious signal, denoted by a black circle, is noteworthy. It propagates from the feeding circuit to
overlap with the radiation of the main antenna arrays. The main component of this overlapped signal
is cross polarisation. This overlapping causes the unwanted dips and back loops in the main pattern
of the design. The spurious radiation increases with frequency, as noted in [75]. Since the proposed
design is tested at high frequency (28 GHz), this problem becomes explicit in the results.

24
Figure 3.8 3-D topology of the antenna array system

Table 3.1

Phase Gain Main beam direction in H-plane

Mode
difference (dBi) (𝜽𝒎𝒂𝒙 )
1 90° 6.2 −55°
2 0° 6.7 0°
3 −90° 6.4 55°

Figure 3.9 Reflection coefficients of mode1s, 2 and 3.

25
Figure 3.10 Simulated H-plane radiation of modes 1, 2 and 3.

To overcome the problem of spurious radiation, double layers of an absorber (SB1007-040-6P) are
added above the feeding circuit on each side of the substrate. Figure 3.12 shows the difference in the
electric field distribution with and without using absorber layers in the H-plane. It is clear that after
adding the two layers of absorbers, the spurious signals are suppressed on both sides of the substrate.
As a result, this reduces the overlapping with the main radiation of the antenna array.

26
Figure 3.11 The electric field snapshot of the design in E and H planes

Figure 3.12 The electric field snapshot of the design in H planes. (a) Without absorber, and (b) with
two layers of absorbers.

3.2.5 Fabrication, Measurement and Discussion

To verify the antenna performance, a prototype was fabricated, as shown in Figure 3.13. The results
for the reflection coefficient and radiation patterns are shown in Figure 3.14 and Figure 3.15
respectively. A LOESS smooth technique in MATLAB was used to remove the noise from the
measured radiation pattern. The measured results are in good agreement with the simulated values.
The differences between them at this high frequency could be caused by the tolerance in the dielectric
constant of the substrate, the misalignment of the design’s layers, especially due to the use of plastic
screws to affix the layers together, and other possible fabrication errors [76]. The simulated directivity
and gain of the three modes across the band, are shown in Figure 3.16. The variation of gain across
the bandwidth of each mode is less than 0.8 dB. By comparing the gain with the directivity results, a
difference of around 4 dB can be observed. According to the results discussed in Section 3.2.3 and
Section 3.2.2, the insertion losses of the two-layer power divider and digital phase shifter on each
side of the substrate are 1.2 dB and 2.12 dB respectively. In addition, the insertion loss of the
27
Wilkinson power divider is 1.2 dB at each port, which is not plotted here for the sake of brevity.
Therefore, the whole insertion loss of the system is 4.5 dB (1.2 dB+2.12 dB+1.2 dB), which is
reasonable when compared with the difference between the directivity and the gain. The main
insertion loss comes from the digital phase shifter, as will be discussed in Section 3.2.4. Based on the
above calculations, the simulated radiation efficiency of this antenna array system is between 35%
and 40%.

Table 3.2 shows a summarised comparison between the proposed and reported antenna beam-steering
designs at mm-Wave. Beam-steering systems, but only of 2-D beam-steering capability, were
proposed [24, 25, 77, 78] using different techniques. In [26, 28] 3-D beam-steering is achieved by
using multiple feeding ports. The proposed design has 3-D beam-steering with one excitation port,
compact size and fully integrated feeding circuity features. Moreover, it can significantly reduce
exposure of the human head to EM when 5G mobile handsets are used.

Figure 3.13 Fabricated prototype. (a) Top view, (b) bottom view, and (c) prototype with absorbers.

28
Figure 3.14 Measured and simulated reflection coefficients of mode1s, 2 and 3.

29
Figure 3.15 Normalised 28 GHz radiation pattern of mode1s, 2 and 3. (a) E-plane, and (b) H-plane.

Figure 3.16 Simulated gain and directivity of the design.

30
 Table 3.2 Comparison between proposed and reported mm-Wave antenna beam-steering designs.

Using
beam-
Phase Fits
Array BW Steering steering
Ref. generation inside
size (GHz) capability to reduce
technique handset
EM
exposure
7 diodes
integrated in
[77] 1˟3 34-36 2-D (±60°) No Yes
the feeding
circuit
Using
[24] 4⨯4 27.3-29 2-D (±20°) No No
multiple ports
2 integrated
[78] 1⨯5 N/A 2-D (±16°) phase shifter No N/A
(MMIC)
Two
[25] subarray 27.5-30 2-D (±60°) RFIC No Yes
(1 ⨯8)
Using
[26] 3⨯3 27.3-29 3-D (±20°) No Yes
multiple ports
3-D (±80°) Using
[28] 4 ⨯4 28-32 No No
and (±88°) multiple ports
6 diodes
This 27.5- 3-D (±55°) integrated in
2⨯4 Yes Yes
work 28.5 and (±35°) the feeding
circuit

3.3 Tunable In-Phase Power Divider for 5G Cellular Networks

Previous sections focussed on realising a beam-steering antenna array system using tunable
(switchable) phase shifters. Such an approach might result in large SLL, especially at wide scanning
angles. To supress those side lobes, the excitation amplitude of the antenna array elements should be
adjusted [29]. A tunable power divider is needed. Thus, this section describes a tunable power divider
suitable for the beam-steering circuit of the future 5G antenna array of handset devices. The concept
of this power divider involves a modification of the Wilkinson design to operate from 27.5 to
29.5 GHz.

3.3.1 Proposed Design

The configuration of the proposed tunable power divider is shown in Figure 3.17(a). The design is
developed in microstrip technology on the substrate RT/duroid-5880 with dielectric constant = 2.2

31
and thickness = 0.127 mm. To enable tuning the output power level, a pair of varactor-loaded open-
ended L-shaped stubs is connected at the two branch lines of the structure. The utilised stubs enable
tuning the power ratio while providing the required matching at all of the ports, and thus acceptable
power division flatness. By varying the biasing of those varactors, their capacitances are changed.
Therefore, the equivalent impedances of the two branch lines, which are connected to these two
varactors, are changed and thus the power division ratio can be modified. To improve the matching
of the output ports at different power divisions, two asymmetrical coupled lines are included as
impedance transformers at those ports. Also, a stepped-impedance matching transformer is added to
the input port to improve its matching [67]. An isolation resistor (R) is connected between the output
ports to improve their isolation.
Details of the L-shaped stub, which is the main modification applied to the traditional Wilkinson
design, is shown in Figure 3.17(b). The varactors are shown to be connected effectively to ground
layer through a quarter-wavelength open-ended stub (𝐿𝑆 ) at the central frequency (28 GHz). For any
power division ratio between the output ports, (𝐾 2 ) the required input impedances of the two L-
shaped transmission lines, which are denoted by 𝑍𝑇𝑆(1) and 𝑍𝑇𝑆(2) , should be different. By using the
ABCD matrix technique [67, 79] , the following design equations are obtained:

Z2  70.7cot(2 / 2) (3.1)

1  K 2 cos2 (2 / 2)
ZTS 1   jZ 0 (3.2)
K 3 1  2sin 2 (2 / 2)

cos2 (2 / 2)
ZTS  2   jZ 0 K 1  K 2  (3.3)
1  2sin 2 (2 / 2)

where 𝑍0 is the impedance at the input and output ports (50 Ω). Initially, an equal power division
condition is considered in the analysis (𝐾 2 = 1), then (2) will be equal to (3) and 𝑍𝑇𝑆(1) and 𝑍𝑇𝑆(2)
will have the same value. Figure 3.18 shows the relationship between the values of 𝑍2 and required
𝑍𝑍𝑇𝑆 versus 𝜃2 (0°<𝜃2 <180°). When 𝜃2 becomes close to 90°, 𝑍2 approaches 70.7 Ω and the required
ZTS will be an open circuit. In this case, the design will work like a normal Wilkinson power divider;
as a result, the varactors will not have a significant impact on tuning the power division ratio, so this
value of 𝜃2 should be avoided. In addition, due to fabrication process limitations, and in order to have
reasonable circuit dimension at 28 GHz band, it is preferred that 𝜃2 not be less than 90°. Therefore,
the best possible values of 𝜃2 that satisfy (1), (2) and (3) will be between 120° and 150°. By assuming
the reasonable value of 𝜃2 = 135°,the required value of 𝑍2 from Figure 3.18 is 𝑍0 /2.

32
 (a) (b)

Figure 3.17 (a) Schematic diagram of the proposed tunable power divider, and (b) details of the
utilized L-shaped stub. Dimensions (mm) are: W0 = 0.4, W1 = 0.46, L1 = 2, W2 = 1.15, L2 = 2.67, Ws
= 0.29, W3 = 0.44, L3 = 0.25, W4 = 0.3, L4 = 1.27, W5 = 0.44, L5 = 0.45, W6 = 0.06, LC = 1.8, W7 =
0.14, S1 = 0.08, LS = 1.59.

Figure 3.18 The required values for 𝑍2 and 𝑍𝑇𝑆 versus𝜃2 .

33
Based on transmission line theory, ZTS can be calculated as a function of the varactor impedance,

ZT Z 3  j Z 32 tan 3
ZTS  (3.4)
Z 3  jZT tan 3

Z 4 ( Z5  wC Z52 tan 5 )
ZT  (3.5)
j Z 4 ( wC Z5  tan 5 )  j tan 4 ( Z5  wC Z 52 tan 5 )

where 𝐶 is the capacitance value of the varactor. While ferroelectric varactors could be a good option
for our design, the selected varactor is the available hyper abrupt varactor diode (MAVR-000120-
1411), which is suitable for the 28 GHz band used, and has a capacitance tuning range from 0.17 to
1.1 pF. For equal power, division ratio 𝐶 = 𝐶𝑉1 = 𝐶𝑉1 = 0.4 pF, which is equivalent to the middle
varactor impedance within the tuning range of the varactor, is selected. By applying 𝑍𝑇𝑆 from (2) or
(3) in (4) and (5), the remaining unknown parameters of the L-shaped structure are 𝑍3 , 𝑍4 and 𝑍5 .
Thus, no one solution can be directly obtained. Instead, several iterations (parametric solutions) are
calculated and the optimial values of these variables are then selected. For example, 𝑍3 = 𝑍5 = 𝑍𝑜 ;
with the aforementioned values, the even-odd mode analysis reveals that the required isolation
resistor is actually 𝑍0 . To change the power division, the capacitance of one varactor is decreased,
whereas it is increased for the other. Therefore, by applying 𝑍𝑇𝑆(1) from (2) in (4) and (5), the
equivalent capacitance value 𝐶 for the first varactor 𝐶𝑉1 is obtained; and by applying 𝑍𝑇𝑆(2) from (3)
in (4) and (5), the capacitance value 𝐶 for the second varactor 𝐶𝑉2 is obtained.

To enhance the matching in the output ports, two pairs of asymmetrical quarter-wavelength coupled-
line transformers are used. They can handle more impedance transformation across wide frequency
bands [71, 80, 81] and can eliminate the need for DC blocking capacitors. To design the coupled
lines, tight coupling (high even mode and low odd mode) is required across the tuning power division.
To achieve that, a narrow gap between those lines is needed, and that causes manufacturing
difficulties. Thus, two slots are created in the ground underneath the coupled lines to increase the
even mode impedance [82].

The initial dimensions of the main design parameters are obtained by using the previous analysis,
whereas the final dimensions are obtained using ADS and then optimized using HFSS as given in
Figure 3.17.

34
3.3.2 Results and Discussion
Figure 3.19(a) shows the layout of the designed tunable power divider. The top layer includes the
main structure of the power divider and the biasing circuit of the two varactors. The biasing circuit is
realised by using three RF chokes (𝑅𝐹𝐶) of 0.7µH and one DC block capacitor (𝐶𝑏 ) of 0.01 µF. The
two varactors are denoted by 𝐶𝑉1 and 𝐶𝑉2 and the capacitance tuning range can be changed by using
biasing voltage from 0 to 10 V [83].

Figure 3.19 (a) Layout of the top layer, and (b) layout of bottom layer.

A prototype, shown in Figure 3.20, is fabricated and tested. Figure 3.21 shows the simulated and
measured results at equal power division state (1:1) through the band from 27 GHz to 29 GHz with
𝐶𝑉1 = 𝐶𝑉1 = 0.4 pF and 𝑉1 = 𝑉2 = 5 V. To provide an accurate comparison with the simulations, the
insertion loss of the added test connectors, which is provided by the connectors’ manufacturer, and
the slight asymmetry between the transmission lines and cables connected to ports 2 and 3, are
calibrated out. The measured S21 and S31 are -4.5 dB and -4.14 dB respectively, compared with the
simulated ones of -3.5 dB for both S21 and S31. More than 12.5 dB input and output port return losses
are achieved across the same band with 10.7 dB measured isolation between the output ports. The
measured phase difference is 1.7° between the output ports.

35
Figure 3.20 Fabricated prototype.

Figure 3.21 The performance at equal power division.

Figure 3.22 The performance at the maximum tunable power division.

36
By varying the biasing voltages (V1 and V2) of the varactors independently, the equivalent varactors’
capacitances are also varied and thus a tunable power division ratio can be achieved. It is found that
the maximum obtained power division ratio is (2.3:1) when Cv1 = 1.1 pF and Cv2 = 0.2 pF as shown
in Figure 3.22 and the correct values for 𝑉1 and 𝑉2 are obtained from the varactors’ datasheet. The
measured S21 and S31 are 3.3 dB and -6.8 dB respectively, compared with the simulated ones of -2
dB and -6 dB across the 27-29 GHz band. More than 10 dB input and output port return losses are
achieved across the same band with 11 dB isolation between the output ports. The phase difference is
less than 5° in the simulations and 12° in the measurements. The slight discrepancies between the
simulated and measured results are due to the unavailability of varactors’ equivalent circuit models
and the inaccuracies of hand soldering.

3.4 Tunable Power Divider for SLL Suppression

One of the solutions for suppressing the side-lobe (SLL) of the antenna array is to control the exciting
power of each antenna element. By using a specific type of power distribution such as Dolph-
Chebyshev, Binomial, or Taylor in the feeding antenna array, the required SLL can be achieved. This
can be realised by using a tunable power divider in the feeding circuit rather than a traditional equal
power divider. Therefore, in this section, the proposed Wilkinson tunable power divider, presented in
Section 3.3 (Figure 3.17), will be discussed as a control for the feeding excitation of the two dipole
antenna subarrays described in Section 3.2.1 (Figure 3.1). By using this power divider, the required
SLL suppression level can be achieved. Dolph-Chebyshev distribution will be used to suppress the
SLL to 20 dB. Since the two dipole subarrays are identical, Figure 3.23 represents the configuration
of the proposed tunable feeding circuit with one antenna subarray only; the other tunable feeding
circuit with the second antenna subarray, is not mentioned for the sake of brevity. The proposed
feeding circuit consists of two stages, as shown in Figure 3.23. The first stage consists of a traditional
Wilkinson power divider, used to split the input signal (Port 1) into two identical output signals. The
second stage consists of two tunable power dividers with four output ports (Ports 2, 3, 4, and 5). The
four output signals are controlled by four voltages named V1, V2, V3, and V4. The power distribution
of this tunable feeding circuit is separated into two possible cases:

1) No SLL: when all output signals are identical (ideally -6 dB) as shown in Figure 3.24(a);
in this case, all biasing voltages are set to 0.4 V.
2) 20 dB SLL: when power signals are set based on Dolph-Chebyshev distribution to achieve
20 dB SLL suppression. The targeted power weight should be set at 0.57, 1, 1, 0.57,
which are equivalent to ideal power levels -7.3 dB, -4.9 dB, -4.9 dB , and -7.3 dB from
Ports 1, 2, 3, and 4 respectively (Figure 3.24(b)). The required biasing voltage values for
this distribution are (V1 = V4 = 10, V2 = V3 = 0).
37
To verify the objective of SLL suppression, the tunable feeding circuit is connected (co-simulated)
with the dipole antenna subarrays (as shown in Figure 3.23) by using the electromagnetic tool CST
Microwave Studio. The simulated result yields -12 dB impedance bandwidth through the targeted
frequency band (27.5-28. 5 GHz), as shown in Figure 3.25(a). The normalised radiation pattern in E-
plane at 28 GHz is plotted in Figure 3.25(b). It is clearly demonstrated that by using the proposed
tunable power divider, the SLL is suppressed from 13 to 19 dB, which is very close to the theoretical
values.

Figure 3.23 Configuration of the proposed tunable feeding network with the dipole antenna subarray
in the top layer.

38
Figure 3.24 Simulated S-parameters of the proposed feeding network with (a) equal output power,
and (b) unequal output power, which follows the Dolph-Chebyshev for 20-dB SLL suppression.

Figure 3.25 (a) |S11| of dipole antenna sub arrays with the tunable feeding circuit, and (b) normalised
E-plane radiation pattern at 28 GHz.

3.5 Summary
In this chapter, the development of a mm-Wave 3-D beam-steering system suitable for 5G mobile
communication handsets was presented. The system design and analysis were presented with
simulations and supporting test results obtained from a fabricated prototype. The 3-D beam-steering
concept is achieved by using two dipole antenna arrays, two Wilkinson power dividers, two digital
phase shifters and a two-layer power divider. Compared with other beam-steering systems, this design
has achieved 3-D beam-steering compactly and reduced the effects of electromagnetic field exposure
to the human head. Additionally, it has improved the link budget between the handset and the base
station. To further improve the performance of the beam-steering antenna system mm-Wave, a tunable
39
In-phase power divider for 5G is proposed to suppress the SLL of the antenna array pattern. The
utilised structure is based on modifying the Wilkinson design by using varactor-loaded open-ended L-
shaped stubs to tune the characteristic impedances of the two branches of the divider. Moreover, a pair
of asymmetrical coupled lines is connected to the output ports and a stepped impedance transformer
is connected at the input port to enhance the matching of the three ports over the band from 27 GHz to
29 GHz. A prototype was designed, fabricated, and tested. The results show that the power division
can be tuned from 1:1 to 2.3:1 across that frequency range. The return loss and isolation are more than
10 dB, and the simulated and measured phase difference is less than 5° and 12°, respectively, across
the whole range of tuning. Finally, to verify the benefit of the proposed tunable power divider, a
tunable feeding circuit was proposed to feed the two dipole antenna subarrays. This was confirmed:
by tuning the output power of the feeding circuit based on Dolph-Chebyshev distribution, the SLL
can be suppressed from 13 to 19 dB.

40
Chapter 4 – Reconfigurable CP Antennas for 5G Applications at
Millimeter-wave
4.1 Introduction
Circular polarisation (CP) is the key feature in antenna design to mitigate misalignment and multipath
interferences. This feature is also relevant to the forthcoming Fifth Generation (5G) mobile network.
Unfortunately, device-to-device (D2D) communication [17] between two CP radiators can suffer
from extremely low polarisation efficiency if the rotating modes of the two are set in opposite
directions. To resolve this issue, CP antennas with rotating modes that are switchable between right
hand circular polarisation (RHCP) and left hand circular polarisation (LHCP) should be used.
Electronic components, such as p-i-n diodes, are among the most popular to realise reconfigurable
circular polarisation antennas, as they can easily integrate within the antennas’ structures. However,
for 5G applications at millimeter wave (mm-Wave), the available p-i-n diodes have high insertion
loss and poor matching [44], which, in the literature, was claimed to limit the use of p-i-n diodes to
design reconfigured CP antennas at mm-Wave. In this chapter, two novel structures of polarisation-
reconfigurable antenna for 5G applications at mm-Wave are realised. First, a single patch antenna
with the functionality to switch between LHCP and RHCP using a simple feed circuit, is developed.
Second, a 4-element antenna array that can switch its operation between two rotating modes (RHCP
and LHCP) using only 4 p-i-n diodes, is presented.

4.2 Polarisation-Reconfigurable Antenna for Mm-wave 5G

This section reports the development of polarisation-reconfigurable patch antenna suitable for mm-
Wave 5G mobile systems. A simple feeding circuit can be used to achieve the switching function
between LHCP and RHCP modes.

4.2.1 Design Geometry and Consideration

The geometric structure of the proposed antenna is shown in Figure 4.1. It is designed on RT/duroid-
5880 substrate with dielectric constant = 2.2, thickness = 0.254 mm and loss tangent tanδ = 0.0009
at 10 GHz. The structure consists of three parts:

1- The radiating element is a corner-truncated square patch. Truncating the corners of a square
patch to produce circular polarisation is a well-known technique [84]. This antenna can also
be fed from two sides, as shown in Figure 4.1.
2- The two p-i-n diodes (D1 and D2) work separately as switches. They can be turned ON or
OFF by adjusting the applied bias voltage.

41
 3- The feeding transmission line is designed to feed the antenna. The input port is divided into
two arms through a T-junction connection. At the end of these arms, there are two quarter
wavelength transformers with length (L4) used to feed the antenna through D1 and D2. These
transformers are used to improve the matching between the antenna and the feeding line [4].
The total length of each arm (L2+ L3) is equal to a half wavelength of the central frequency
(28GHz).

When D1 and D2 are in ON and OFF-states respectively, the patch antenna is fed from the upper side
and consequently, RHCP is realised. By reversing the operating modes of D1 and D2, the position of
the patch antenna excitation can be reversed. As a result, LHCP is achieved.

Figure 4.1 Schematic diagram of the polarisation-reconfigurable patch antenna. Dimensions (mm)
are: L1 = 3.3, L2 = 3.89, L3 = 1.1, L2 = 1.92.

For both RHCP and LHCP, the transmission line is connected with the antenna through one arm and
one diode (in the ON-state) and the second arm is connected to the other diode (in the OFF-state). In
this case, the second arm works like an open-ended stub. Since, the length of this stub is a half
wavelength, the input impedance of the stub ( Z in ) becomes infinity by:

𝑍𝑖𝑛 = −𝑗𝑍0 cot(𝛽𝑙) (4.1)

where Z0 is the characteristic impedance of the input port, 𝛽 = 2𝜋/ λg is the phase constant, 𝑙 = λg or
λg/2 is the length, and λg is the guided wavelength at the operating frequency (28 GHz). This means
that the stub works as an open circuit and does not affect the impedance of the remaining part of the
design.

The initial values for the design are selected based on well-known half wavelength transmission and
patch antenna theory for the central frequency (28 GHz). The optimised dimensions are obtained
42
using the electromagnetic simulation tool HFSS. An ideal p-i-n diode circuit is used in the simulation.
For the patch antenna, the optimum value of length, L1 is found to be equal to 0.4λg. The total length
of (L2+ L3) is equal to 0.53 λg and L4 is equal to 0.5 λg.

To confirm the proposed antenna’s CP performance, the surface current distribution on the patch is
simulated and illustrated in Figure 4.2 for different time instants, i.e., t = 0, (T/4), (2T/4), and (3T/4)
where T denotes the time period of the electromagnetic wave at 28 GHz. The surface current on the
patch is distributed with respect to the diagonals and it rotates circularly with increasing T, which
indicates that the antenna is radiating CP waves properly.

Figure 4.2 Simulated surface current distributions of the proposed antenna with a period T at 28 GHz:
(a) t = 0, (b) t = T/4, (c) t = 2T/4, and (d) t = 3T/4.

4.2.2 Results and Discussions

Figure 4.3 and Figure 4.4 show a simulated reflection coefficient and axial ratio respectively. The
reflection coefficient covers bandwidth 27.6–28.6 GHz in both circular polarisation modes, while the
axial ratio, less than 3 dB, covers band 27.65-28.35 GHz, indicating that the antenna shows a good
CP performance.
43
Figure 4.5 depicts the corresponding E-plane and H-plane simulated radiation patterns of LHCP and
RHCP at 28 GHz. The antenna demonstrates directional radiation performance with wide beamwidth.
An average gain of around 7.2 dBic is found along the boresight direction. The 3-dB beamwidths are
83° and 100° in the E-plane and H-plane respectively.

Figure 4.3 Simulated reflection coefficient results for LHCP and RHCP states of the proposed
antenna.

Figure 4.4 Simulated axial ratio versus frequency plots for LHCP and RHCP states of the antenna.

44
Figure 4.5 Simulated radiation patterns for (a) RHCP, and (b) LHCP states of the antenna at 28 GHz.

4.3 Polarisation-Reconfigurable Antenna Array for mm-Wave 5G

A polarisation-reconfigurable antenna array using p-i-n diodes has been proposed at 28 GHz. The
design utilises the idea of deploying two feeding arms to reduce the number of used diodes reported
in Section 4.3.1.1. The main contribution of this design is the novel usage of the branch-line coupler
and open-ended stubs to reduce the number of used diodes, which consequently improves matching
and reduces insertion loss. Additionally, the proposed configuration allows the use of highly
imperfect p-i-n diodes at mm-Wave to switch between two operations, a feature that none of the
previously reported approaches could achieve. The targeted frequency band is 27.5-28.35 GHz,
which is approved by the U.S. Federal Communications Commission (FCC) for 5G applications [14].

4.3.1 Antenna Array Structure and Operating Principle

4.3.1.1 Operating Principle and Feeding Circuit
The radiators consist of 4 patch antennas (Figure 4.6) resonating at 28 GHz. The design uses
RT/duroid-5880 substrate (relative permittivity εr = 2.2 with estimated tanδ ≈ 0.0025 at 28 GHz).
The simulation, modelling and optimisation are conducted using the ANSYS High Frequency
Structure Simulator (HFSS).

45
For CP operation, the patches are excited with equal amplitude signals and 90° phase delay in either
anti-clockwise, i.e. right-hand (RH), or clockwise, i.e. left-hand (LH), sequence (Figure 4.6).
Therefore, in order to realise a reconfigurable CP array, a feeding structure that can provide ±90o
phase switching is required. The proposed feeding structure consists of a branch-line coupler, two
feeding arms and 2 p-i-n diodes, as shown in Figure 4.7.

As an initial demonstration, the p-i-n diodes are assumed to be perfect in the simulation, i.e. operating
as the RF short-circuit in the ON-state and RF open-circuit in the OFF-state. In this case, the branch-
line coupler is designed to be ideal, i.e. to consist of 4 quarter-wavelength (L4 = L1 = λg/4) segments
with equivalent impedances of Z 0 /√2 and Z 0 respectively (Z0 is the characteristic impedance of the
input port) [67]. The two feeding arms have the length of λg and λg/2 respectively (λg is the guided
wavelength at the operating frequency). The branch-line coupler is connected to these two arms
through 2 p-i-n diodes, D1 and D2. The following operation cases are considered:

1) Case A: D1 and D2 are in the ON and OFF-state respectively. The branch-line coupler generates
equal amplitude and 90° phase difference between the output ports (ports 2 and 3).

2) Case B: D1 and D2 are in the OFF and ON state respectively. The branch-line coupler generates
equal amplitude and -90° phase difference between the output ports (ports 2 and 3).

Figure 4.6 Proposed antenna array arrangement for CP operation. (a) RHCP phase delay sequence,
and (b) RHCP phase delay sequence. 𝜃 is the reference angle.

46
Figure 4.7 Construction of branch-line couplers using an ideal diode with S-parameters and
transmission phase difference.

Table 4.1

State of diode Frequency (GHz) |S11| |S21|

OFF-state 28GHz -1.8 dB -1.5 dB
ON-state 28GHz -9.8 dB -1.4 dB

In both cases, since the length of each input arm is either λg or λg/2, the arm connected to a diode in
the OFF-state works as an open-ended stub, and based on Equation (4.1), the input impedance of the
open-ended stub ( Z in ) becomes infinity. Thus, it does not influence the other parts of the feeding.

Figure 4.7 shows satisfactory simulated results for this structure with either 90o or -90o phase
difference between port 2 and port 3 in the targeted bandwidth. Compared to the conventional ±90°
line phase shifter [67, 85], the technique of using a branch-line coupler with two feeding arms allows
reduction in the number of used diodes by half, which is a big advantage–especially at mm-Wave,
where the diode performances in terms of matching and insertion loss are poor.

47
In practice, the p-i-n diodes are far from perfect, especially at mm-Wave frequency. Therefore, it is
first necessary to extract the equivalent circuit of the p-i-n diodes, so it can be used in the simulation
rather than perfect diodes. The TRL calibration method [86] is conducted next to extract the
equivalent lumped-element models of a chosen diode (MADP-000907-14020W). The measured
reflection and transmission coefficients at 28 GHz are presented in Table 4.1. The electrical
parameters of the ON and OFF-state equivalent circuits are derived based on fitting the measurement
and simulation results [86]. These parameters (Figure 4.8) are Rs =5.81 Ω, and Lf = 0.0895 nH for the
ON-state, and Rr = 435.7 kΩ, Ct = 33.6 fF, and Lr = 0.131 nH for the OFF-state. At 28 GHz, the return
and insertion losses in the OFF-state are 1.8 dB and 1.5 dB respectively. Hence, the diode has a
critically poor performance and does not work as an open-circuit exactly like an ideal switch in the
OFF-state. Nonetheless, the diode has reasonable performance in the ON-state mode, with return and
insertion losses 9.8 dB and 1.4 dB respectively. The derived p-i-n diode’s equivalent circuit is applied
to the previous design (Figure 4.8). As expected, a significant degradation in performance is observed.
The main reason is that the two arms (λg and λg/2) no longer work as open-ended stubs. Instead, they
work together as an asymmetrical T-junction power divider, feeding the branch-line coupler from two
ports.

Figure 4.8 Construction of feeding network using equivalent circuits of the diode with the (non-
optimised) S-parameters and transmission phase difference.

48
4.3.1.2 Feeding Circuit Performance Improvement
A parametric study was carried out to improve the performance of the two operation modes of the
branch-line coupler (Cases A and B), based on the initial parameters used in the previous section.
Two sets of key parameters are found to play a key role in improving the performance:

Set 1) L3 and SS: L3 is the length of the longer arm λg and SS is the length of the transmission line
connecting the diode to the branch-line coupler. Since the two arms work as an asymmetrical power
divider, the lengths L3 and SS mainly affect the operating frequency of Cases A and B. By reducing
the values of these two parameters, the operating frequency of Cases A and B can be shifted to the
targeted frequency (28 GHz) and the performance of both cases can be improved.

Set 2) W5 and W11: These two parameters control the impedance of a branch of the coupler and the
feeding arms. It is found that by reducing the values of these two parameters, the impedance matching
for Case A and Case B can be improved.

For the sake of brevity, the optimised performance of a single feeding circuit is not shown here.
Instead, we show the optimised performance of the complete feeding circuit for the 4-element patch
array (Figure 4.9). In this configuration, the two branch-line couplers of Cases A and B are located
vertically and connected to a T-junction power divider [67]. The operation modes are controlled by
switching the states of the two pair of diodes, (D1, D3) and (D2, D4), to achieve a 90° phase delay in
a clockwise or anti-clockwise sequence between the output ports of the branch-line couplers
(summarised in Table 4.2). For demonstration, the optimised S-parameters and phase differences of
the LHCP operating mode are shown in Figure 4.9(b) and (c) respectively.

4.3.1.3 Biasing Circuit of p-i-n Diodes

The biasing circuit consists of two 0.01 μF chip capacitors (denoted by C1 and C2 ) to block the dc
currents between the two arms of the T-junction power divider (Figure 4.9(a)). Since the available
RF chokes working at mm-Wave are relatively large [4] and not easily integrated with the proposed
design, the stepped-impedance low pass filters (SLPF), described in Section 3.2.2, is employed
instead. The SLPF works like a short circuit for DC power supply and an open circuit for RF signals
at mm-Wave. The orientation of the p-i-n diodes and SLPF is shown in Figure 4.10. Using this
configuration, only one bias voltage V is required to switch between RHCP and LHCP (Figure 4.10).

49
Figure 4.9 (a) The design of branch-line coupler with T-junction power divider, (b) simulated S-
parameters, and (c) simulated transmission phase.

Table 4.2

(D1, D3) (D2, D4) Phase delay sequence Polarisation

ON OFF Anti-clockwise RHCP
OFF ON Clockwise LHCP

50
Figure 4.10 Orientation of diodes and SLPF.

4.3.1.4 Final Antenna Array Design

The patch antenna array in Figure 4.6 is excited by the optimised feeding circuit in Figure 4.9 through
H-shaped slots in the ground plane (Figure 4.11). The H-shaped aperture is used to improve the
resonant impedance and the coupling between the top and bottom layers [29, 87]. The final multi-
layered design is shown in Figure 4.12. Two RT/duroid-5880 substrates are adopted, each for the
patch and feeding circuit. An HT 1.5 bonding film (relative permittivity εr = 2.35, tanδ = 0.0025 and
thickness 0.0381 mm) is used to attach the top substrate layer to the ground. The final parameters of
the design are shown in Table 4.3.

Figure 4.11 The geometry of the 4-element patch antenna array and its simulated reflection
coefficients.
51
 Table 4.3 Optimised Parameters of the final design (all values in mm)

Parameter Value Parameter Value Parameter Value

Ls 34 L4 2.26 L2 2.58
Ws 31 W1 0.8 W2 0.6
L3 3.33 W11 0.23 SS 0.47
W5 0.219 L1 2.18 L6 L3 – L1/2
St 0.539 S2 0.35 Ht 1.42
W0 0.72 Tr 1.4 T1 17.5
Lt 0.53 Ta 7.7

Figure 4.12 (a) Polarisation-reconfigurable antenna array, and (b) 3-D exploded view of layers.
52
4.3.2 Result and Discussion
The proposed design was fabricated (Figure 4.13) and then measured to validate the proposed design
concept. The reflection coefficients for both states RHCP and LHCP are shown in Figure 4.14(a).
The design yields a measured -20 dB impedance bandwidth from less than 27 GHz to 29 GHz. A
near-field chamber is used to measure the axial ratio of the proposed design, where the axial ratio can
be derived from principal and cross-polarisation values across the targeted frequency band. The
prototype exhibits a measured 3-dB AR operation from 27.2 to 28.35 GHz for both CP modes, as
shown in Figure 4.14(b). Both impedance and AR bandwidth cover the targeted frequency band, i.e.
27.5 to 28.35 GHz.

Figure 4.13 Photograph of fabrication prototype. (a) Top view, and (b) bottom view.

Figure 4.14 (a) Simulated and measured S11 of the proposed antenna array, and (b) simulated and
measured AR. The shaded area is the targeted bandwidth.

53
Figure 4.15 Normalised simulated and measured radiation patterns at 28 GHz. (a) x-z plane, and (b)
y-z plane

The normalised simulated and measured radiation patterns for RHCP and LHCP in the x-z and y-z
planes at 28 GHz are plotted in Figure 4.15. The measured radiation patterns mostly follow the
simulated ones with slight differences due to the measurement errors at mm-Wave.

Figure 4.16 (a) Measured broadside gain of the proposed antenna array, and (b) measured efficiency.
The shaded area is the targeted bandwidth.

54
The measured gain and the radiation efficiency for both CP modes are illustrated in Figure 4.16(a)
and (b) respectively. The maximum measured gain is about 6 dBic, while the measured radiation
efficiency is around 51% across the targeted BW. This is reasonable for a reconfigurable antenna
operating at mm-Wave. Compared to simulated results; the lower efficiency is possibly attributable
to higher losses in the connector and substrates [88]. It should be noted that efficiency measurement
at mm-Wave is challenging and an error of 1 dB in gain measurement can cause the efficiency to drop
by roughly 20%.

4.4 Summary
In this chapter, two types of reconfigurable-polarisation antennas were presented. The novelty of
these designs is based on the switching of the feeding circuit, including the two arms, to reduce the
use of p-i-n diodes at mm-Wave. The first design is a polarisation-reconfigurable patch antenna,
suitable for mm-Wave 5G mobile systems. The design is based on a patch antenna with a switchable
transmission line feed. Using two independently biased p-i-n diodes, the proposed design can produce
left-hand or right-hand circular polarisation. The simulation results demonstrate that the antenna has
a good impedance matching across the 27.6-28.6 GHz band and 3 dB axial ratio performance across
27.65-28.35 GHz for both switchable polarisation states. The second design is a polarisation-
reconfigurable antenna array using p-i-n diodes at 28 GHz. A branch-line coupler and open-ended
stub are adopted in the feeding circuit to reduce the number of used diodes. This configuration also
allows a p-i-n diode with non-ideal performance to be used at mm-Wave. RHCP and LHCP
operations are realised using two pairs of diodes. The measurements have confirmed that the proposed
design can support CP reconfigurability across 27.5-28.35 GHz for 5G applications. Compared with
other exiting designs at mm-Wave, the two proposed designs have great potential to be used for 5G
wireless communication.

55
Chapter 5 - MIMO Antenna System for mm-Wave 5G and Wideband
4G Applications
5.1 Introduction
The multiple-input-multiple-output (MIMO) antenna system is seen as a key technology for the
upcoming fifth generation (5G) and the existing fourth generation (4G) mobile antenna system to
enhance data rate [14, 52, 89, 90]. Along with the MIMO antenna system, 5G requires a wide
bandwidth to maximise the data rate. Unfortunately, the current available bandwidth at sub-3 GHz is
congested [20]; thus, many international organisations are assigned the millimeter wave band (mm-
Wave), which includes 28, 37 and 39 GHz, as a potential frequency band for 5G [21, 91]. Therefore,
current mobile terminals need to support 5G at mm-Wave band in addition to the existing 4G at sub-
3 GHz band. However, mobile terminals have limited space, and it is preferable not to add many
antennas. Thus, the solution is to develop a MIMO antenna system which has antenna elements
covering different bands of 5G and 4G by using a single structure rather than two structures. In other
words, the usage of multi-band antenna elements in the MIMO system is a key factor for reducing
the antenna footprint inside the mobile terminal [14, 20-22]. However, integration of 5G and 4G
antennas in the same structure produces many challenges in terms of antenna size, isolation between
antennas at different bands and antenna gain.

In this chapter, two novel structures of 5G/4G MIMO antenna system are realised. First, an integrated
4-element 5G/4G MIMO antenna system using a novel single antenna structure is presented. It works
at triple bands (28, 37 and 39 GHz) for 5G and the wideband (1.8-2.6) GHz for 4G. Each MIMO
element consists of a slot in the ground plane and two microstrip feeding ports in the top layer; a low
pass filter is also integrated into the structure for isolation enhancement. By properly feeding the
antennas from one of two ports, the antenna can work in two different operating modes. One of the
feeding ports corresponds to a tapered slot antenna with end-fire radiation for 5G, whereas the other
feeding port corresponds to an open-ended slot antenna with omni-direction radiation at 4G. MIMO
performance factors are also calculated and discussed. Second, a compact 2-element 5G/4G MIMO
antenna system for handheld mobile terminals is presented. It works for both microwave and mm-
wave bands using a single ground slot structure. The proposed design works as a frequency
reconfigurable slot antenna with 27% of tuning range at microwave band (from 2.05 to 2.70 GHz)
and a wideband connected slot antenna array (CSAA) operating from 23 to 29 GHz. A compact high
pass filter is also integrated into the feeding structure to isolate the operation of 5G and 4G.

56
5.2 Antenna Design Requirements for 5G and 4G Mobile Antennas
With respect to antenna design, the transition from 4G to 5G at mm-Wave brings many new
requirements in addition to the existing ones [92, 93]. For instance, compared to the 4G mobile signal
at the sub-3 GHz band, 5G at mm-Wave suffers from higher path loss [94]; therefore, 5G antennas
must feature a high gain to compensate for the higher path loss. Moreover, high isolation between
antenna elements is important to ensure that 5G antennas do not interfere with 4G antennas. In
addition, using planar antennas and placing them on the side edges of mobile terminals is important
consideration for retaining sufficient space for other components [55]. Most likely, 5G will work
across multiple spectrums, according to providers and regions; therefore, building multiple-band 5G
antennas, similar to the existing multiple-band 4G antennas, is preferred for covering these bands and
thus reducing antenna size. Based on the aforementioned requirements for 5G and 4G antenna design,
integrating 5G and 4G antenna structures will lead to a new generation of 5G/4G antenna designs
whose key elements are categorised, as shown in Figure 5.1. These elements are used as goals in the
design and discussion of the 5G/4G antenna designs.

Figure 5.1 Key design requirements for 5G/4G MIMO antenna system.

57
5.3 MIMO Antenna System for Multi-Band mm-Wave 5G and Wideband 4G
Applications
This section reports the development of a 4-element MIMO antenna system for 5G/4G mobile
technology. The design can work as a tapered slot antenna for 5G covering 27.5-40 GHz or an open-
ended slot antenna for 4G covering 1.8-2.6 GHz.

5.3.1 Proposed MIMO Antenna System Design

The geometry of the complete MIMO antenna system is shown in Figure 5.2. It is modelled and
simulated using the High Frequency Structured Simulator (HFSS). An RT/duroid-5880 substrate
(𝜀𝑟 = 2.2, tan 𝛿= 0.0009, and thickness 𝑡 = 0.381 mm) with overall size of 𝐿𝑠𝑢𝑏 ⨯ 𝑊𝑠𝑢𝑏 𝑚𝑚2 is
used to print the antenna structure. All antenna elements (Ant-1-Ant-4) are realised on the four sides
of the substrate as a proof of concept. All microstrip feeding lines are placed on the top layer, while
tapered slots are etched in the ground plane located at the bottom layer of the substrate.

The ground slot has a dual function, which makes the antenna structure more suitable for future
5G/4G applications. First, when the slot is fed from the edge (𝐹2 for example), it works as a tapered
slot antenna and gives end-fire radiation patterns [95, 96] with high gain, thus satisfying the
requirements of 5G at mm-Waves. Second, when it is fed from the centre (𝐹1 for example), it works
as an open-ended slot antenna [97, 98] and gives omni-directional radiations satisfying the
requirements of 4G at 2 GHz band. Accordingly, a single antenna, having a small footprint compared
with separate antennas, can be designed to meet the requirements of low and high frequency bands.
A detailed discussion of the design procedure and construction approach of the proposed antenna
structure is given in the following sections.

58
Figure 5.2 The geometry of a 4-element MIMO antenna system. (a) Top view (full design), (b) Ant-
1 (large view), and (c) tapered slot in the ground plane: Final optimised parameters with their values
for MIMO antenna system are:𝐿𝑠𝑢𝑏 = 158, 𝑊𝑠𝑢𝑏 = 77.8, 𝑓𝑙1 = 1.89, 𝑓𝑙2 = 3.5, 𝑙1 = 13, 𝑙2 = 4.95,
𝑙3 = 27, 𝑙4 = 2, 𝑑1 = 2, 𝐿𝑒𝑑𝑔𝑒 = 13.5 and 𝑤1 = 1.17, all values are in millimetre (mm).

5.3.1.1 Multi-band 5G Antenna Structure

The structure of the 5G antenna is shown in Figure 5.3. It consists of a tapered slot and microstrip
feed line [99]. The total size of the single element with the ground plane is 70 ⨯ 77.8 mm2 (𝐿 ⨯ 𝑊).
The narrower edge of the tapered slot ends with a 𝜆4 /4 circular stub. The diameter of the stub is
𝑑2 (Figure 5.2(b)). The tapered slot is fed via a circular sector microstrip line with diameter 𝑑1 . The
circular sector and a 50 Ω transmission line are connected through an impedance matching
transformer that has a total length of 𝑇𝑙 [100]. The circular stub of the slot and circular sector of the
feed line contribute to the provision of a broadband impedance matching bandwidth [101, 102].

A parameter study is conducted to introduce stepped tapering into the slot. Four cases (Case1-Case4)
have been studied to give the final shape to the slot using a number of different tapers, as shown in
Figure 5.4. It is clear from the reflection coefficient curves that for the single tapered slot (Case1),
impedance matching is poor, whereas the three step tapered slot provides a wider bandwidth.
Therefore, the proper selection and tuning of the tapers can provide wider impedance bandwidth. The
final tapered slot (Case4) covers the 25-40 GHz band. The main beam is directed towards the negative
x-axis in an end-fire radiation pattern (Figure 5.3(c)) with maximum gain of 9.14 dBi at 38 GHz. The
59
observed gain variation is only 1 dBi within the covered band. Final optimised parameters (as shown
in Figure 5.2(c)) for the tapered slot are as follow: 𝑆𝑙1 = 15, 𝑆𝑙2 = 5, 𝑆𝑙3 = 1.5, 𝑆𝑙4 = 1.42, 𝑑2 = 2,
𝑆𝑤 = 4.2, 𝑇𝑙 = 2 and 𝑙𝑚 = 1.9; all values are in mm.

Figure 5.3 The geometry of 5G antenna. (a) Top view, (b) large view, and (c) 3D realised gain in dBi.

Figure 5.4 Construction of tapered slot (design cases and reflection coefficients).
60
5.3.1.2 Integration of 4G Feeding Structure to 5G antenna Structure
In this part, the same tapered slot which was used for the 5G band in the previous section is also used
to cover the low frequency band for 4G. This dual feature of the slot meets dual wireless standards
(5G/4G). The feeder (𝐹1 ) is used to excite the tapered slot, which works as an open-ended slot antenna,
as shown in Figure 5.5. First, the 4G feeding structure (F2) is a simple 50 Ω transmission line, as
shown in (Figure 5.5(a)), while the final optimised feeding structure consists of a low pass filter
(LPF). A modified transmission line, as shown in Figure 5.5(b), along with the 5G antenna. 𝐹1 is used
to excite the slot as an open-ended slot antenna and 𝐹2 is used to excite the same slot as a tapered slot
antenna. The design covers a low frequency band from 1.8-2.6 GHz and a high frequency band from
25-40 GHz. The length of the open-ended slot is 24.92 mm (𝑆𝑙1 + 𝑆𝑙2 + 𝑆𝑙3 + 𝑆𝑙4 + 𝑑2 ), which is
𝜆𝑔 /4 at 2 GHz (𝜆𝑔 is guided wavelength) [29, 103].

Figure 5.5 (a) Integrated antenna structure without a low pass filter, (b) integrated antenna structure
with a low pass filter, (c) 5G reflection coefficient without/with low pass filter, and (d) 4G reflection
coefficient without/with low pass filter.
61
Due to the addition of the 4G feeder (𝐹1 ) to the tapered slot, the performance of the 5G antenna is
degraded in terms of impedance matching, isolation and gain. To understand the performance
degradation of the 5G antenna, a study is conducted in four stages (Stage1-Stage 4) based on the
current distributions given in the following sub-sections.

1) Stage 1: The 5G antenna current distribution, before the addition of the 4G feeder, is shown in
Figure 5.6(a). It can be seen that the current is mainly propagating towards the open end of the slot
and radiates as a tapered slot antenna. It covers the 25-40 GHz band with maximum gain of 9.14 dBi
at 38 GHz.

Figure 5.6 Current distributions at different stages (Stage1-Stage4) of the 4G feeder at 38 GHz.

2) Stage 2: After adding the 4G feeder, the coupling is found to be high – between the 4G and 5G
feeders, as shown in Figure 5.6(b). The 4G feeder works as a shunt impedance or a signal blocker for
the 5G antenna signal propagation. As a result, the main current propagates toward the upper and
lower side of 4G micorstrip feeder instead of propagating toward the open-end of the slot. At this
stage, the s-parameters of the antenna at high and low frequencies are shown in Figure 5.5(c) and
Figure 5.5(d) without the LPF. The antenna covers 28.5-32.5 and 36-40 GHz while the maximum
gain is observed to be 5.87 dBi. It can be noticed that after adding the 4G feeder, gain and impedance
matching is reduced, and the isolation is also poor.

3) Stage 3: To minimise the effect of the 4G feeder on the 5G band, it is necessary to increase the
equivalent shunt impedance value of the 4G feeder, to work as an open circuit for the 5G signal. By
doing this, the coupling between 𝐹1 and 𝐹2 can be reduced significantly [104] and the performance
of the antenna can be improved. The shunt impedance can be increased by reducing the width of the
4G feeder above the slot, as shown in Figure 5.6(c). However, due to fabrication process limitations,
the width of the 4G feeder is reduced to only 0.17 mm, as shown in Figure 5.5(b), and an improved
gain of 7.72 dBi is achieved.
62
4) Stage 4: To improve the isolation between the 5G and 4G feeders, an LPF is added, as shown in
Figure 5.6(d). The LPF design is based on a stepped impedance method [105, 106]. The performance
of the LPF is shown in Figure 5.7. The simulated reflection coefficient and isolation curves for the
5G/4G bands after adding the LPF are shown in Figure 5.5(c) and Figure 5.5(d). It is clear that
isolation and impedance matching bandwidth is improved significantly after adding the LPF. It covers
the 25-40 GHz band with maximum gain of 8 dBi at 38 GHz. The maximum realised gain at stages
1 to 4 is 9.1, 5.9, 7.7 and 7.9 dBi, respectively.

Figure 5.7 Performance of low pass filter at (a) 4G band, and (b) 5G band.

Figure 5.8 Photograph of the fabricated prototype. (a) Top view, and (b) bottom view.
63
5.3.2 Result and Discussion
In this part, the results are discussed for the 4-element MIMO antenna system. Optimisation is studied
so as to achieve the best performance at low and high frequencies. A prototype is fabricated to verify
the proposed antenna structure. The photograph of the fabricated prototype is shown in Figure 5.8.

5.3.2.1 Integrated 5G/4G MIMO Antenna System

The geometry of the proposed MIMO antenna system is shown in Figure 5.2. It consists of 4 antenna
elements (Ant-1-Ant-4) and 8-feeder (𝐹1 , 𝐹3 , 𝐹5 and 𝐹7 for 4G and 𝐹2 , 𝐹4 , 𝐹6 and 𝐹8 for 5G). All
antenna elements are printed on the four sides of the substrate. The size of the substrate is similar to
the size of the latest smart phones. The length of both feeders (5G and 4G) is extended to integrate
50 Ω SMA and mm-Wave connectors and to have enough space for antenna testing. While adding all
the antenna elements to the sides of the substrate, performance remains unchanged in the 5G bands.
However, in the 4G bands, Ant-1 and Ant-2 have slightly different reflection coefficients, as shown
in Figure 5.10(a) (𝑆11 and 𝑆33 ). This difference is caused by the size of the ground plane [107]. These
antennas have different ground plane sizes, i.e. Ant-1 has a width 𝑊𝑠𝑢𝑏 , while Ant-2 has a width 𝐿𝑠𝑢𝑏
(as shown in Figure 5.2). However, the proposed design still has good impedance matching below -6
dB, which is acceptable for 4G applications [22].

1) Port Parameters: The simulated and measured reflection coefficient curves for the 5G band are
shown in Figure 5.9(a). The results show that the proposed 5G antennas cover the 25-40 GHz
frequency band. Figure 5.9(b) shows isolation curves between F1 and F2. A minimum of 25 dB
measured isolation is observed across the whole covered band. Isolation between the 5G ports (F2,
F3, F5 and F7) is observed to be more than 50 dB. The simulated and measured reflection coefficient
curves for the 4G band are shown in Figure 5.10(a). The results show that the proposed design covers
the 1.8-2.6 GHz band with -6 dB impedance matching bandwidth. Figure 5.10(b) shows the isolation
curves between the various ports of 4G and 5G. A minimum of 16 dB measured isolation is observed
between F5 and F7.

2) Radiation Patterns: The simulated and measured radiation patterns are obtained when one antenna
is excited and others are terminated with a match load of 50 Ω [53].

64
Figure 5.9 Simulated and measured S-parameters for 5G band. (a) 𝑆𝑖𝑖 , and (b) 𝑆𝑖𝑗 .

Figure 5.10 Simulated and measured S-parameters for 4G band. (a) 𝑆𝑖𝑖 , and (b) 𝑆𝑖𝑗 .

65
Figure 5.11 Normalised simulated and measured radiation patterns at different frequencies for 5G
bands (Ant-1).

The normalised 2D radiation patterns for 5G bands are shown in Figure 5.11 and Figure 5.12 for Ant-
1 and Ant-2 at 28 and 38 GHz frequencies, respectively. Due to the symmetry in the design, only the
radiation patterns for Ant-1 and Ant-2 are shown. Both antennas have strong radiation patterns in
different directions in the xy-plane, i.e. Ant-1 has a maximum value at   180 , while Ant-2 at
  90 . Such radiation patterns are desirable for MIMO applications and provide un-correlated
communication channels.

66
Figure 5.12 Normalised simulated and measured Radiation patterns at different frequencies for 5G
bands (Ant-2).

Figure 5.13 Normalised simulated and measured Radiation patterns for 4G band.

67
The normalised 2D radiation patterns for the 4G band are shown in Figure 5.13 for Ant-1 and Ant-2
at 2.0 GHz. Both antennas have directional radiation patterns in the xy-plane, i.e. Ant-1 has directive
beam towards   180 , while Ant-2 has it at   90 ; however, both the xz and yz planes show
omni-directional radiation patterns.

The simulated and measured peak gain values are shown in Table 5.1 and Table 5.2 for 5G and 4G
bands, respectively. The maximum measured value is observed to be 7 dBi at 38 GHz for the 5G
band; whereas for the 4G band, the maximum measured gain is 3.6 dBi at 1.8 GHz. A slight difference
between simulated and measured values is observed due to measurement errors.

3) Envelope Correlation Coefficient (ECC): As well as evaluating the reflection, isolation and
radiation pattern performances of the MIMO antenna system, it is also important to evaluate its
diversity performance. The Envelope Correlation Coefficient (ECC) parameter is used to serve this
purpose [53]. It is calculated based on the 3D measured far field radiation patterns between antenna
pairs at low and high frequencies by using the following formula, given in [58]:

4 2

 [ F1 ( ,  )  F2 ( ,  )d ]
e 
0
4 2 4 2 (5.1)
0
F1 ( ,  ) d   
0
F2 ( ,  ) d 

where, ρe is the ECC, F1 ( , ) and F2 ( , ) are the 3D radiated fields of Ant-1 and Ant-2,

respectively, while Ω is the solid angle [58]. The maximum measured value of ECC was found to be
lower than 0.001 at both bands (5G/4G), which is much lower than the acceptance criteria of below
0.5 for the ECC. Therefore, the proposed antenna design has good diversity performance.

Table 5.1 Simulated and measured peak gain for 5G band

Peak gain (dBi) 27.5 GHz 28 GHz 38 GHz 39 GHz

Simulated 5.8 6.2 7.6 7.2
Measured 5.5 5.1 7 7.2

Table 5.2 Simulated and measured peak gain for 4G band

Peak gain (dBi) 1.8 GHz 2.0 GHz 2.3 GHz 2.6 GHz
Simulated 3.7 3.3 2.5 3.6
Measured 3.6 3.1 2.4 2.9

68
5.3.2.2 Effects of LCD and Hand/Head Phantom
Since the positions of MIMO antennas are located on the edges of the ground plane, the influence of
adding an LCD is investigated through HFSS simulations. A large-size touch screen LCD with
dimension 136 × 68 mm2 is located underneath the ground plane. The LCD is assumed to be a perfect
electric conductor (PEC) for the proposed design [108]. The simulated and measured reflection
coefficient and efficiency of MIMO antenna are investigated for 3 mm, and 5 mm gap between the
LCD and ground plane. For the sake of brevity, only the performance of Ant-1 is shown in Figure
5.14(a) and Figure 5.14(b), and the antenna performance without LCD is included for comparison
purpose. As seen in Figure 5.14, the performance after placing the LCD is still preserved at the 4G
and 5G bands. However, it can be noticed that the 4G bandwidth is shifted towards the higher
frequency by 0.2 GHz, compared to the ones without LCD (Figure 5.14(a)).

Figure 5.14 Effect of LCD and hand/head phantoms on proposed design (a) 4G band, (b) 5G band at
different spacing between the LCD and the ground plane, (c) hand/head model, and (d) radiation
efficiencies at 4G (top) and 5G (bottom) bands.

69
This shift can be fixed by reducing the value of Ledge from 13.5 mm to 10.5 mm. In addition, it is also
noticed that the measured efficiency is degraded by 2-10% at the 4G band and by only 2% at the 5G
band.

To analyse the influence of head/hand phantoms, CST Microwave Studio and the simulation model
is used as shown in Figure 5.14(c) while efficiency curves are shown in Figure 5.14 (d). It can be
seen that Ant-2 and Ant-4, which are closely related to the head/hand phantom, are most affected,
while the influence of head/hand phantoms on Ant-1 and Ant-3, which are relatively far from the
phantoms, is less significant. Ant-1 and Ant-3 still work well with maximum radiation efficiency of
65% and 60% at 28 and 2.6 GHz, respectively. Nevertheless, Ant-2 is the most affected in terms of
antenna efficiency (a similar behaviour was observed in [53]). Compared with Ant-1, Ant-2 has a
reduction on the efficiency of about -4 dB and -7 dB in the 5G and 4G bands respectively. This is due
to the significant amount of field absorption into the hand/head phantom and the locations of four
antennas with respect to the phantoms.

Table 5.3 Comparison between proposed and reported 5G/4G MIMO antenna systems

Integrated Support Size of each Realised

5G/4G mm- antenna BW gain (dBi) Isolation
Ref.
in same Wave element (GHz) at mm- (dB)
structure band (mm2) Wave
1.9-3.5
[22] No No N/A N/A ≥5
16.5-17.8
15 ⨯ 2.5 3.4-3.8
[56] Yes No N/A ≥ 11
(5G/4G) 5.15-5.92
0.82-0.96
16 ⨯ 3 (5G)
[54] No No 1.7-2.69 N/A ≥ 10
30 ⨯ 9 (4G)
3.4-3.6
15 ⨯ 12 (5G) 3.4-3.8
[53] No No N/A ≥1 2
7 ⨯ 3 (4G) 5.15-5.92
35 ⨯ 18 (5G) 2-2.2
[58] No No N/A ≥ 13.5
24 ⨯ 6 (4G) 12.2-12.7
23 ⨯ 22 (5G) 1.85-2.53
[59] No Yes 8 ≥ 10
25 ⨯ 8 (4G) 26-28.4
This 25 ⨯ 12 1.8-2.6
Yes Yes 7.2 ≥ 17
work (5G/4G) 25-40

5.3.2.3 Performance Comparison with Current 4G and 5G Antennas

To make the merits of the proposed design explicit, a comparison table is drawn up of recently
published 5G/4G MIMO antenna systems for mobile applications, as shown in Table 5.3. The
proposed design is compared with other MIMO antenna systems in terms of the integration of 5G and
4G in the same structure, support for the mm-Wave band, MIMO element size, bandwidth, realised

70
gain and isolation. Table 5.3 makes it clear that the proposed design has the unique feature of
integrating mm-Wave 5G and 4G antennas in the same structure, while other designs do not.
However, the integration of the 5G and 4G structure is achieved [56] at small size and only covers
sub-7 GHz band. Although [59] covers similar bands (mm-Wave and 2 GHz band), it deploys separate
structures having larger sizes, 23 ⨯ 22 mm2 for 5G and 23.5 ⨯ 9 mm2 for 4G, compared with the
proposed design that has only a 25 ⨯ 12 mm2 size for both 5G and 4G. All other antenna designs [22,
54, 58] cover different bands (excluding mm-Wave) bands with variations in the realised gain and
isolation of MIMO elements. Overall, this design provides, to the best of author’s knowledge, the
first novel antenna integration strategy for future mm-Wave 5G with existing 4G technologies in a
single structure.

5.4 Integrated Frequency-Reconfigurable Slot Antenna and Connected Slot

Antenna Array for 4G and 5G Mobile Handsets
In this section, we will continue the investigation into design approach for the 5G/4G MIMO antenna
system. However, instead of using a single antenna element, this section will present an antenna array
structure to further improve the gain, especially for mm-Wave 5G. In addition, this design will add a
frequency-reconfigurable feature to 4G to enhance the out-of-band rejection performance.

5.4.1 Antenna Element Design and Performance

The proposed antenna is printed on RT/duroid-5880 substrate (𝜀𝑟 = 2.2, tan 𝛿 = 0.0009, and
thickness 𝑡 = 0.381 mm). It is modelled and simulated using the High Frequency Structured
Simulator (HFSS). The structure includes a narrow rectangular slot. At low frequency, the structure
works as a typical narrow-band antenna. By integrating a varactor, the low frequency resonance can
be tuned to cover different channels. At mm-Wave frequency, the slot is properly fed to operate as a
CSAA to produce a higher gain. This section starts by describing the frequency-reconfigurable slot
antenna, followed by the 5G CSAA. Finally, the combination of two structures with an integrated
filter for high isolation between two feeding ports, is presented.

5.4.1.1 Frequency-Reconfigurable Slot Antenna Design

The geometry of a frequency-reconfigurable slot antenna at microwave band is shown in Figure
5.15(a). The size of the ground plane is Lsub ×Wsub mm2 . When the slot is etched the system ground
plane at the center and side, and the equivalnt simulated 3D radiation patterns are shown in Figure
5.15(b) and Figure 5.15(c) respectively. Where, Figure 5.15(b) shows a slot type radiation pattern,
Figure 5.15(c) shows a tilted radiation pattern due to the asymmetrical ground plane. However, for
handheld devices, the location of the slot is preferred on the side of the ground plane, as shown in

71
Figure 5.15(a) [18, 109]. For this reason, the slot is etched on the side of the ground plane for the
final structure.

To cover a wide operating frequency range with a compact structure, a frequency-reconfigurable

antenna is desired [110, 111]. The tunability also enhances the out-of-band rejection performance
[112]. The tunable functionality can be included in this design by loading a short-circuited varactor
diode to the slot [113-115]. When the varactor capacitance C is very low, the varactor works as an
RF open circuit and the slot resonates at its unloaded resonance frequency of

𝑐0
𝑓𝑟 = (5.2)
2𝑆1 √𝜀𝑒𝑓𝑓

here 𝑐0 is the free-space speed of light, 𝑆1 is the length of the slot and 𝜀𝑒𝑓𝑓 is the effective permittivity.
In contrast, if C increases, the resonance frequency of the slot decreases as analysed in [113]. In this
design, the varactor diode (VAR) MA46H120 model is used. It has a measured capacitance (Cv) range
of (0.15, 1.34) pF and an internal resistance of Rv = 2 Ω. The length of the slot is selected as Sl = 35
mm to provide a resonance frequency of fr = 3 GHz in case Cv = Cmin, i.e. approximated as an open
circuit. The width of the slot is not critical, however it will be used for 5G application, so it should
be selected at less than half the guided wavelength at the mm-Wave band, i.e. 28 GHz. In our design,
the slot width (Sw) is kept fixed at 2.5mm.

The slot is fed from 50 Ω microstrip line that has a width of 1.18 mm and length of 9.5 mm. The
length of the microstrip line is optimised to obtain satisfactory impedance matching across the tuning
range. Figure 5.15(a) represents the biasing circuit that consists of two shorting vias, one large resistor
Rb = 20 kΩ, a choke inductor Lb = 220 nH, and a DC-blocking capacitor Cb = 39 pF. The two shorting
vias are used so that the biasing circuit can be printed on the top layer of the substrate while the
bottom layer is the ground.

The location of the varactor from the edge of the slot Ledge is optimised to obtain a wide frequency
tuning range. The results for different values of Ledge when Cv = 0.15 (solid curves) and 1.02 pF
(dashed curves) are shown in Figure 5.16(b). It can be seen that when the varactor is placed further
away from the slot edge, a larger tuning range can be achieved. In fact, when the varactor is placed
in the middle of the slot, the largest tuning range is obtained. However, this causes an issue in the
integration with the 5G module as will be shown in the next section. Here, Ledge = 2.69 mm is chosen,
which can already achieve a large tuning range from 2.2-3 GHz, i.e. 30%, with Cv varying in the 1.02-
0.15 pF. The optimised parameters are shown in Table 5.4. The simulated reflection coefficients are
shown in Figure 5.16 (a). Here, it is worth mentioning that only a part of varactor capacitance range
is utilised in the design to keep the simulated efficiency across the tuning range always more than

72
50%. When Cv is increased further, the resonance frequency only slightly decreases, while the
efficiency significantly decreases. The reduction of efficiency in the frequency-reconfigurable
antennas, which is due to the internal resistance of the varactor, is consistently reported in the
literature [110, 116-119].

Figure 5.15 (a) Geometry of frequency-reconfigurable slot antenna, (b) 3D radiation patterns of slot
antenna with symmetry ground plane, and (c) 3D radiation patterns of slot antenna with modified
ground plane.

Figure 5.16 (a) Simulated reflection coefficients of the 4G reconfigurable slot antenna with different
varactor capacitance Cv , and (b) simulated reflection coefficients at the lower (0.15 pF) and upper
limit (1.02 pF) of Cv with different values of Ledge.
73
5.4.1.2 Connected Slot Antenna Array (CSAA) Design
The slot antenna shown in the previous section can be deployed to work at mm-Wave with higher
gain to obtain a compact 4G/5G integration. Such compact and multifunctional designs are highly
desirable in mobile handsets due to the size constraints [14]. In order to achieve this, the concept of
the connected antenna array is utilised, where the antenna elements are connected or tightly coupled
[58, 120, 121]. The arrangement of the proposed CSAA is shown in Figure 5.17(a) where 8 feeding
lines feed the slot periodically with the spacing of Fs, which is around 𝜆𝑔 /2 at the targeted operating
frequency (𝜆𝑔 is the guided wavelength). This configuration is similar to the 8-element slot antenna
array where each element’s length is 𝜆𝑔 /2. The elements in the CSAA configuration are connected
with each other and they appear as a single antenna with multiple periodic feeders. It can be noted
here that after the feeder F8, there is enough space for the reconfigurable 4G band module as shown
in the previous section (see Figure 5.17(a)). This spacing only slightly increases the size of the slot at
mm-Wave band without significantly affecting the performance. Figure 5.17(a) also shows the
simulated reflection coefficient at the input (simulated as lumped ports) of the 8-feeders before
connecting them to the feeding network. As expected in a CSAA configuration, two features can be
immediately noticed:

1) The feeders F2 to F7 have the same response in terms of resonance frequency at 32 GHz, while the
feeders F1 and F8 have a different resonance frequency at 24 GHz. This difference is caused by the
mutual coupling between these feeders. Unlike F2 − F7, F1 and F8 are at the edges of the slot so they
only strongly couple with one adjacent feed. Additionally, the two edges of the slot are asymmetrical,
which leads to a different impedance matching across the targeted band.

2) For all feeders, |Sii| < −7 dB within a band from 22 − 36 GHz. Therefore, this design has the
potential to achieve a wide bandwidth.

Table 5.4 Optimised parameters used in the final design (all values are in millimetres).

Parameter Value Parameter Value Parameter Value

Lsub 70 t 0.381 Sw 2.5
Wsub 60 SL 35 Ledge 2.69
Fs 3.36 tL 4 wf 1.18
w1 1.18 w2 0.34 sf 0.1

74
To feed the slot as a CSAA, 1 ˟ 8 T-junction power divider with compact size is proposed at 28 GHz
as shown in Figure 5.18(a). The simulated S-parameter and transmission phase difference for only
the power divider are shown in Figure 5.18(a) and Figure 5.18(b) respectively. The results validate
that the input port (F9) of the power divider has a good impedance matching at the designed frequency
of 28 GHz (|S99| = -20 dB), and the transmission coefficients |Si9| from the input port to the 8 feeders
are equally around -9 dB ≅ 10 log(1/8). The 8 feeders have nearly identical transmission phases as
shown in Figure 5.18(b). Figure 5.19(a) presents the simulated reflection coefficient of the CSAA,
including the power divider feeding network (see Figure 5.19(a)). The proposed design has a wide
bandwidth from 23 to 29 GHz. Most importantly, the antenna has perfect matching at 5G bands 28
and 24 GHz announced recently by the Federal Communications Commission (FCC) [10]. These two
well matched frequencies, 28 and 24 GHz, occur due to the main resonance of the inner feeders (F2-
F7 at 32 GHz) and the side feeders (F1 and F8 at 24 GHz), respectively. However, the resonance of
the inner feeder shifts from 32 GHz (see Figure 5.17(a)) to 28 GHz due to the effect of the power
divider. The optimised parameters for CSAA with the power divider are shown in Table 5.4.

Figure 5.17 (a) Simulated reflection coefficient of each feeder for the CSAA without feeding network,
and (b) simulated current distributions on feeders and the ground plane.
75
Figure 5.18 Performance of the 1 ˟ 8 power divider. (a) S-parameters, and (b) transmission phase.

The simulated 3D radiation pattern in terms of total realised gain is also shown in Figure 5.19(b). The
directivity of an N = 8-element array of isotropic sources at f0 = 28 GHz can be calculated as 2 ˟ N ˟
Fs/ 𝜆0 ≅ 7.0 dBi. The simulated gain of a single slot is about 4.5 dBi, giving a calculated gain value
for the designed CSAA of about 11.5 dBi. This value agrees well with a simulated maximum gain of
10.9 dBi for the proposed mm-Wave antenna (measured results will be shown in Section 5.4.2.2).

Figure 5.19 (a) Geometry of connected slot antenna array, and (b) simulated reflection coefficient
with 3D radiation patterns after adding power divider at 28 GHz.

76
5.4.1.3 Integration of the Slot Antenna with an Array of Connected Slots
The two designs shown in previous sections can be merged to build a 4G/5G integrated antenna with
wide frequency tunability at microwave band and wide bandwidth at the mm-Wave band. To further
enhance the functionality of this structure, the MIMO arrangement is also deployed and investigated.
Figure 5.20 shows the layout of the final design with two identical slots arranged orthogonally to
achieve polarization diversity. The two slots are located at the edge of the geometry. All dimensions
of previous optimised designs are remained the same. MIMO performance will be presented in
Section 5.4.2.3.

The feeders affect the microwave band, where the resonance frequency of the slot antenna shifts to a
lower value. Compared to Figure 5.16, it can be noticed that the whole tuning range shifts from 2.2-
3.0 GHz to 2.05-2.60 GHz, as shown in Figure 5.21. This shift is expected and can be attributed to
the 5G feeders. These feeders perform as lumped capacitances which lower the resonance frequency
of the slot, like adding varactors. However, no change in the performance of the CSAA at mm-Wave
band is observed after integrating both structures. This is also as expected due to the simulated current
distribution indicated in Figure 5.17, where the current is very small in the region of feeding and
reconfiguring the 4G band.

Figure 5.20 The geometry of 2-element MIMO antenna system.

77
Figure 5.21 (a) Simulated isolation with and without filter when Cv = 0:15 pF, and (b) simulated
reflection coefficients for different values of Cv after adding the CSAA feeders.

Figure 5.22 Performance of the high pass filter incorporated into the CSAA feeding network. (a) Low
frequency, and (b) high frequency.

Another factor to be observed when integrating the 4G and 5G modules is the isolation between slot
antenna feeding port (Port 1) and array feeding port (Port 3) (Figure 5.20). It is found that the isolation
is quite poor at low frequency band. Figure 5.21(a) shows isolation of 5 dB between Port 1 and Port
3 at 2.6 GHz. To improve the isolation, a compact high pass filter (HPF) is designed and added into
the feeding network of the CSAA. The geometry of HPF is shown in Figure 5.20. It is designed based
on the interdigital capacitance method [122]. Five fingers with the width of Wf and the inter-finger
spacing of sf are implemented. These values are optimised to attain the best filtering performance for
low and high frequency bands as shown in Figure 5.22. By adding HPF, the isolation between Port 1

78
and Port 3 is improved significantly, to reach 23 dB between Port 1 and Port 3 at 2.6 GHz, as shown
in Figure 5.21(a).

Figure 5.23 Photograph of fabrication prototype. (a) Top view, and (b) bottom view.

5.4.1.4 Design Procedure

Based on the discussed and presented results, a brief design procedure for the proposed antenna can
be described as

1) Initially, the length of the slot is chosen based on the upper operating frequency of the lower
frequency range, i.e. 4G band. The width of the slot is not a critical parameter and can be tuned later
for better impedance matching.

2) The feeder for the 4G band, i.e. Port 1 and varactor is added. Then, the position of the varactor is
optimised to achieve a satisfactory frequency tuning range.

3) The array of feeders for the CSAA at 5G band is added.

4) The design parameter at 4G band is tuned (if required) to obtain the required tuning range again.
This is due to the influence of the 5G feeders on the resonance frequency of the slot.

5) The high pass filter is added and the feeding network parameters are optimised.

5.4.2 Measurement and Result

The proposed MIMO antenna is fabricated (Figure 5.23) and tested for verification. The two slots are
denoted as Ant. 1 and Ant. 2 for convenience. The antenna performances are presented as follows.

79
5.4.2.1 Scattering Parameters
The simulated and measured reflection coefficient and isolation curves for both slot antennas at low
frequency band are shown in Figure 5.24(a) and Figure 5.24(b), respectively at different values of
applied reverse biased voltages (0.7-18 V). The bias voltages remain identical for both antennas. The
measured results indicated that the prototyped MIMO antenna covers the band from 2.05 GHz to 2.4
GHz, with a tuning range of 27%. It is noted that the minimum isolation between Port 1 and Port 2 is
19 dB. A low discrepancy is observed between the simulated and measured results. The reflection
coefficient and isolation curves for Ant. 1 and Ant. 2 at high frequency band are presented in Figure
5.25(a) and Figure 5.25(b), respectively. |S33| and |S44| are below -20 dB at the two targeted bands 24
GHz and 28 GHz. Overall, a wideband performance from 23 to 29 GHz is achieved. Slightly high
reflection is observed at around 26 GHz, nevertheless, such performance still satisfies the antenna
requirements of mobile application where the input power is low. A minimum of 20 dB measured
isolation between Port 1 and Port 3 is observed across the covered band. Again, for the high frequency
band, the measured results follow the simulated ones quite thoroughly.

Figure 5.24 Simulated and measured S-parameters for low frequency band. (a) Reflection coefficient
curves, and (b) isolation.

80
Figure 5.25 Simulated and measured S-parameters for high frequency band. (a) Reflection coefficient
curves, and (b) isolation.

5.4.2.2 Radiation Patterns

The radiation patterns are measured when one slot is active and the other slot is terminated with a
matched load (50Ω) for both bands at low and high frequencies. The normalised simulated and
measured radiation patterns 2D cuts at θ = 90° (azimuth cut), φ = 0° (elevation cut) and φ = 90°
(elevation cut) are depicted in Figure 5.26 and Figure 5.27 for Ant. 1 and Ant. 2, respectively at 2.6
and 2.1 GHz. The results show that both antennas provide different radiation patterns which are
desired for MIMO applications. The discrepancy in the simulated and measured radiation patterns is
noticed in several planes, which is mostly due to the scattering effect of the measuring connectors
and cables. This discrepancy is expected to be due to the effect of the connector, which is placed in a
close proximity to the radiating slot; the slot also radiates towards both +z and -z direction. The
normalised simulated and measured radiation patterns for Ant. 1 and Ant. 2 at high frequency bands
(24 and 28 GHz) are depicted in Figure 5.28 and Figure 5.29, respectively.

81
Figure 5.26 Simulated and measured normalised radiation patterns at low frequency band for Ant. 1.

In both figures, 2D cuts (elevation cuts) for Ant. 1 at φ = 0° and φ = 75°are presented to show the
direction of the main beam. Likewise, 2D cuts for Ant. 2 at φ = 90° and φ = 25° are also presented.
A good agreement is observed between the simulated and measured plots when the effects of
connectors are considered in the simulations. Some ripples are also observed in the radiation patterns
due to the effects of cables and connectors as discussed in the previous paragraph.

The simulated and measured gain and efficiency plots at low frequency band are depicted in Figure
5.30(a) and Figure 5.30(b), respectively. The peak measured gain varies from 4.5 dBi to 2.6 dBi,
while the measured efficiency varies from 70% to 45%. The simulated and measured gain and
efficiency curves at high frequency band are shown in Figure 5.31(a) and Figure 5.31(b), respectively.
The measured maximum gain is found to be 12.5 dBi, while the measured maximum efficiency is
95% at 24 GHz.

82
Figure 5.27 Simulated and measured normalised radiation patterns at low frequency band for Ant. 2.

Figure 5.28 Simulated and measured normalised radiation patterns for Ant. 1 and Ant. 2 at 24 GHz.

83
Figure 5.29 Simulated and measured normalised radiation patterns for Ant. 1 and Ant. 2 at 28 GHz.

Figure 5.30 Simulated and measured, (a) Gain, and (b) efficiency at low frequency band.

5.4.2.3 Envelope Correlation Coefficient (ECC)

To evaluate the proposed MIMO antenna system diversity performances, ECC values are calculated
based on 3D radiation patterns using Equation (5.1). The maximum value of ECC between Ant. 1 and

84
Ant. 2 is found to be 0.01 at 2.6 GHz, which is much lower than the acceptable value of ECC, of
below 0.5 for 4G applications [44]. Therefore, the proposed design has good diversity performance.

Figure 5.31 Simulated and measured, (a) Gain, and (b) efficiency at high frequency band.

5.4.2.4 Performance Comparison with Existing work

A comparison chart including some previous work on MIMO antenna arrays for 5G or dual bands
5G/4G is also drawn up to further illustrate the merits of the proposed design, as is shown in Table
5.5. Indeed, it is clear from the table, that the majority of reported designs worked only at mm-Wave
for 5G applications [61, 109, 123-126] and has a relatively large occupied size compared with the
size of mobile handhelds. The gain, efficiency and bandwidth of these works are comparable with
those of the proposed design. The remaining designs can serve 5G and 4G by using two separated
structures [58, 59], or single structure as in [22]. However, by combining the total required size for
the 4G and 5G antenna footprint together, it can be seen they require a considerably large profile
antenna making theme inappropriate for handheld devices. Therefore, a new design for dual-band
5G/4G microwave and mm-Wave applications is proposed. While maintaining the compact size, high
gain, efficiency and isolation compared to other designs, this design succeeds in integrating 5G and
4G antenna in only one structure.

85
Table 5.5 Comparison between proposed and previous related works.

Isolation
Integration No. of Size of Realised
(dB)
of 5G/4G antenna antenna BW gain (dBi)
Ref. between
in same array array (GHz) at mm-
feeding
structure elements (mm2) Wave
ports
27-29
[109] No 8˟1 42.3 ˟10.5 8.0 N/A
[61] No 16 ˟ 1 99 ˟ 17.47 24-31 19.8 N/A
28,33
[124] No 3˟3 26 ˟ 21 9.8-13.5 N/A
37-39
[125] No 8˟1 60 ˟ 28 24–28 7-11 N/A
[126] No 8˟1 33 ˟ 25 26-38 10.4–12.5 N/A
2 ˟ 2 (5G) 100 ˟ 60 1.9-3.5
[22] Yes 8.0 10
1 ˟ 2 (4G) (4G/5G) 16-17
8 ˟ 1 (5G) 1.8-2.5
[59] No 23⨯22 (5G) 8.0 10
26-28
2.4
[127] No 4 ˟ 1 (5G) 70⨯40(5G) 9.0 10
27-29
This 35 ⨯16.5 2-2.7
Yes 8 ˟ 1 (5G) 12.5 20
work (5G/4G) 23-29

5.5 Summary
In this chapter, two types of MIMO antenna systems for 5G/4G wireless handheld devices for
microwave and mm-Wave applications, were presented. The first design was a 4-element MIMO
antenna system for 5G/4G mobile technology. The design meets the requirements of both 5G and 4G
antennas using only a single structure. Each MIMO element consists of a slot-based antenna fed by
two microstrip feeders for 5G and 4G bands. The design works as a tapered slot antenna at mm-Wave,
offering end-fire radiation for 5G; and it works as an open-ended slot antenna for the 2 GHz band,
offering omni-direction radiation for 4G. The measured results from a prototype show a wide
impedance bandwidth of S11< -10 dB, covering the band 27.5-40 GHz for 5G; and impedance
bandwidth of S11 < -6 dB, covering the band 1.8-2.6 GHz for 4G. The measured gain values were 7
and 3 dBi for 5G and 4G MIMO, respectively. The envelope correlation coefficient did not exceed
0.5 for all antennas, showing good MIMO performance. To further improve the gain at mm-Wave
5G, a second design for 5G/4G handheld devices was proposed. It consists of a 2-element MIMO
antenna system. A dual-functional slot antenna was built to work as a frequency-reconfigurable slot
antenna at the 4G band and connected slot antenna array at the 5G band. Two separate structures were
used to feed the slot and connected slot antenna array. The proposed design provides frequency-
tuning of 2.05-2.70 GHz with a peak gain of 4.5 dBi at low frequency band and a wideband 23-29
GHz with a peak gain of 12.5 dBi at high frequency band. Isolation and ECC values were also

86
calculated to show good MIMO performances. The minimum measured isolation was 20 dB, whereas
the maximum value of ECC was 0.01. Compared with other exiting designs at mm-Wave, the two
proposed designs are planar and meet the requirements of dual wireless standards, making them a
suitable candidate for future 5G/4G mobile terminals.

87
Chapter 6 – Conclusion and Future Work
The previous chapters of this thesis have reported the research work for the design and
implementation of new reconfigurable and MIMO antenna systems for mm-Wave 5G. All the work
conducted in this research project was aimed at developing novel solutions with the capability of
overcoming the main challenges outlined in Chapter 1 (Section 1.1). This chapter summarises the
main contributions of the research work undertaken in this thesis and presents suggestions for future
investigations.

6.1 Thesis Conclusion

Breakthroughs in wireless innovation are creating unprecedented challenges for mobile industry.
Cellular communication has become a part of most everyday human activities and has led to rapid
increases in wireless data traffic requirements. However, the current mobile technologies have almost
reached a mature state and the forecast data traffic demands cannot be further satisfied by using
existing mobile technologies and the current limited carrier frequency spectrum. Therefore, the
solution is to implement a new technology for mobile systems, 5G; and to assign to it a new, higher
frequency spectrum that contains a much larger allocated signal bandwidth than the current ones.
Therefore, many international organisations have suggested or licensed the mm-Wave spectrum as
the operating frequency of 5G at 24, 28, 37, 39, and 60 GHz. However, using mm-Wave raises many
challenges such as high propagation loss and radiation absorption by the human body, polarisation
misalignment, and integrating mm-Wave MIMO into future handset devices. The invention of new
design approaches to overcome these challenges was the underlying background motivation in this
thesis. Notably, the main contributions of this thesis are as follows:

In Chapter 3, a 3-D beam-steering system design at 28 GHz was provided. This design has a compact
and planar profile and is suitable for mm-Wave 5G handsets. By utilising two stacked PCBs, the
beam-steering system uses two antenna arrays producing a 3-D beam steering pattern. The proposed
system redirects around 92% of the radiated power from the handset away from the head toward the
base station. In addition, this steering feature ensures a reliable connection between the handset and
base station despite the high path loss at 28 GHz. A prototype was built and tested as a proof of
concept. It successfully confirmed the beam steering system operating concept at mm-Wave for the
5G handset. Apart from the beam steering system, a tunable power divider was proposed at 28 GHz.
The structure utilised was based on a modification of the Wilkinson design via the use of varactor-
loaded open-ended stubs. The proposed design can attain a tuning power ratio ranging from 1:1 to
1:2.3. Afterwards, the tunable power divider was used in the beam-steering feeding circuit at 28 GHz

88
to tune the amplitude of the feeding circuit. The co-simulation demonstrates that by using the proposed
tunable power divider, the SLL is suppressed from 13 to 19 dB. These results show the potential of
this tunable device in the beam-steering circuit for mm-Wave 5G.

In Chapter 4, a polarisation reconfigurable patch antenna suitable for mm-Wave 5G was first
introduced. The design was based on a patch antenna with a switchable transmission line feed. Using
two independently biased p-i-n diodes, the proposed design can produce LHCP or RHCP. In this
design, the open-ended stub concept was introduced to reduce the number of diodes in the feeding
circuit. Afterwards, the concept of the open-ended stub was extended in the design of an antenna
array with polarisation-reconfigurable for mm-Wave 5G. The antenna array design allows p-i-n
diodes with non-ideal performance to be used at mm-Wave. RHCP and LHCP operations are realised
using two pairs of diodes. The measurements have confirmed that the proposed design can support
CP reconfigurability for mm-Wave 5G applications.

In Chapter 5, a MIMO antenna system for multi-band mm-Wave 5G and wideband 4G applications
was demonstrated first. By utilising a ground tapered slot and two microstrip feeders, the design can
serve triple bands at mm-Wave (28, 37 and 39 GHz) for 5G in addition to a 2 GHz band (1.8-2.6) for
4G. Furthermore, a frequency-reconfigurable slot antenna and connected slot antenna array for 5G/4G
mobile handsets was presented. The antenna structure uses a short-circuited varactor diode to achieve
frequency tunability from 2.05 to 2.7 GHz (4G). For mm-Wave band, by using 8 periodic feeders,
the slot works as a connected slot antenna array (CSAA) with a wide bandwidth of 23-29 GHz (5G).
Both mentioned designs fulfil the requirements of a 5G/4G MIMO antenna system at low and high
frequency bands. They use a single antenna structure approach and are thus suitable candidates for
future 5G handheld devices.

6.2 Suggestions for Future Work

This thesis has solved a number of important challenges regarding the use of mm-Wave for the 5G
mobile network. However, extensive future investigations are still required, including possible future
work as follows:

1. Reducing the size of the 5G/4G MIMO antenna system footprint inside the handset.
Miniaturising the 5G/4G MIMO antenna system size allows integrating more MIMO antennas
inside the handset. This achieves high capacity for mm-Wave 5G and can be considered as
future work stemming from this thesis.
2. Building a frequency-reconfigurable antenna for mm-Wave 5G. As mentioned earlier,
spectrum allocation for mm-W 5G will be wide and include 24, 28, 37, 39, and 60 GHz. In
89
 addition, the licensed frequency spectrum will be different among service providers and
regions. Theoretically, it is possible to build an antenna solution for each specific region to
cover the licensed 5G spectrum. However, this kind of solution is costly in terms of
measurement and antenna hardware interface. To alleviate this challenge, it is necessary to
devise a frequency-reconfigurable antenna at mm-Wave with wide frequency tuning range,
and this may constitute future work.
3. Considering the effects of the other mobile components such as the conductive edges antenna
frame, speakers or USB interfaces. Most of the proposed designs in the literature for mm-
Wave considered a design environment where the antenna structure was verified in isolation.
The effects of other components of the mobile handset were not considered on the design.
Therefore, a simulated and experimental study is necessary to investigate the possible effects
of mm-Wave 5G antennas on other hardware parts of the handset platform.
4. Increasing the scanning angle of the 3-D beam-steering antenna system. Further research
needs to be performed on beam-steering antenna system to achieve a wider scanning angle
with the associated challenge of designing and building low-loss low-cost tunable mm-Wave
phase shifters, and reduced side-lobe level (SLL).

90
References

[1] F. González-Landero, I. García-Magariño, R. Lacuesta, and J. Lloret, "PriorityNet App: A

Mobile Application for Establishing Priorities in the Context of 5G Ultra-Dense Networks,"
IEEE Access, vol. 6, pp. 14141-14150, 2018.
[2] T. S. Rappaport, S. Sun, R. Mayzus, H. Zhao, Y. Azar, K. Wang, et al., "Millimeter Wave
Mobile Communications for 5G Cellular: It Will Work!," IEEE Access, vol. 1, pp. 335-349,
2013.
[3] E. A. Abbas, A. T. Mobashsher, and A. Abbosh, "Polarization reconfigurable antenna for 5G
cellular networks operating at millimeter waves," in 2017 IEEE Asia Pacific Microwave
Conference (APMC), 2017, pp. 772-774.
[4] E. A. Abbas, A. M. Abbosh, and K. Bialkowski, "Tunable In-Phase Power Divider for 5G
Cellular Networks," IEEE Microwave and Wireless Components Letters, vol. 27, pp. 551-
553, 2017.
[5] J. G. Andrews, S. Buzzi, C. Wan, S. V. Hanly, A. Lozano, A. C. K. Soong, et al., "What Will
5G Be?," Selected Areas in Communications, IEEE Journal on, vol. 32, pp. 1065-1082, 2014.
[6] C. L. I, C. Rowell, S. F. Han, Z. K. Xu, G. Li, and Z. G. Pan, "Toward Green and Soft: A 5G
Perspective," IEEE Communications Magazine, vol. 52, pp. 66-73, Feb 2014.
[7] P. A. Dzagbletey and Y. B. Jung, "Stacked Microstrip Linear Array for Millimeter-Wave 5G
Baseband Communication," IEEE Antennas and Wireless Propagation Letters, vol. 17, pp.
780-783, May 2018.
[8] K. M. Mak, K. K. So, H. W. Lai, and K. M. Luk, "A Magnetoelectric Dipole Leaky-Wave
Antenna for Millimeter-Wave Application," IEEE Transactions on Antennas and
Propagation, vol. 65, pp. 6395-6402, Dec 2017.
[9] N. Shoaib, S. Shoaib, R. Y. Khattak, I. Shoaib, X. Chen, and A. Perwaiz, "MIMO Antennas
for Smart 5G Devices," IEEE Access, pp. 1-1, 2018.
[10] F. C. C. (FCC), "The FCC’s 5G FAST Plan, Spectrum.," http://www.fcc.gov./5G.
[11] S. Li, T. Chi, Y. Wang, and H. Wang, "A Millimeter-Wave Dual-Feed Square Loop Antenna
for 5G Communications," IEEE Transactions on Antennas and Propagation, vol. 65, pp.
6317-6328, 2017.
[12] W. Roh, S. Ji-Yun, P. Jeongho, L. Byunghwan, L. Jaekon, K. Yungsoo, et al., "Millimeter-
wave beamforming as an enabling technology for 5G cellular communications: theoretical
feasibility and prototype results," Communications Magazine, IEEE, vol. 52, pp. 106-113,
2014.
[13] K. Zhao, Z. Ying, and S. He, "EMF Exposure Study Concerning mmWave Phased Array in
Mobile Devices for 5G Communication," IEEE Antennas and Wireless Propagation Letters,
vol. 15, pp. 1132-1135, 2016.
[14] W. Hong, "Solving the 5G Mobile Antenna Puzzle: Assessing Future Directions for the 5G
Mobile Antenna Paradigm Shift," IEEE Microwave Magazine, vol. 18, pp. 86-102, 2017.
[15] K. X. Wang and H. Wong, "A Circularly Polarized Antenna by Using Rotated-Stair Dielectric
Resonator," IEEE Antennas and Wireless Propagation Letters, vol. 14, pp. 787-790, 2015.
[16] N. P. Lawrence, B. W. H. Ng, H. J. Hansen, and D. Abbott, "5G Terrestrial Networks:
Mobility and Coverage-Solution in Three Dimensions," IEEE Access, vol. 5, pp. 8064-8093,
2017.
[17] M. N. Tehrani, M. Uysal, and H. Yanikomeroglu, "Device-to-device communication in 5G
cellular networks: challenges, solutions, and future directions," IEEE Communications
Magazine, vol. 52, pp. 86-92, 2014.
[18] I. R. R. Barani and K. L. Wong, "Integrated Inverted-F and Open-Slot Antennas in the Metal-
Framed Smartphone for 2 x 2 LTE LB and 4 x 4 LTE M/HB MIMO Operations," IEEE
Transactions on Antennas and Propagation, vol. 66, pp. 5004-5012, Oct 2018.

91
[19] J. L. Guo, L. Cui, C. Li, and B. H. Sun, "Side-Edge Frame Printed Eight-Port Dual-Band
Antenna Array for 5G Smartphone Applications," IEEE Transactions on Antennas and
Propagation, vol. 66, pp. 7412-7417, Dec 2018.
[20] E. Al Abbas, A. M. Abbosh, and K. Bialkowski, "Tunable In-Phase Power Divider for 5G
Cellular Networks," IEEE Microwave and Wireless Components Letters, vol. 27, pp. 551-
553, Jun 2017.
[21] T. S. Rappaport, S. Shu, R. Mayzus, Z. Hang, Y. Azar, K. Wang, et al., "Millimeter Wave
Mobile Communications for 5G Cellular: It Will Work!," Access, IEEE, vol. 1, pp. 335-349,
2013.
[22] M. S. Sharawi, M. Ikram, and A. Shamim, "A Two Concentric Slot Loop Based Connected
Array MIMO Antenna System for 4G/5G Terminals," IEEE Transactions on Antennas and
Propagation, vol. 65, pp. 6679-6686, 2017.
[23] J. Bang and J. Choi, "A SAR Reduced mm-Wave Beam-Steerable Array Antenna With Dual-
Mode Operation for Fully Metal-Covered 5G Cellular Handsets," IEEE Antennas and
Wireless Propagation Letters, vol. 17, pp. 1118-1122, 2018.
[24] Z. Briqech, A. R. Sebak, and T. A. Denidni, "Low-Cost Wideband mm-Wave Phased Array
Using the Piezoelectric Transducer for 5G Applications," IEEE Transactions on Antennas
and Propagation, vol. 65, pp. 6403-6412, 2017.
[25] B. Yu, K. Yang, C. Sim, and G. Yang, "A Novel 28 GHz Beam Steering Array for 5G Mobile
Device With Metallic Casing Application," IEEE Transactions on Antennas and Propagation,
vol. 66, pp. 462-466, 2018.
[26] M. Khalily, R. Tafazolli, T. A. Rahman, and M. R. Kamarudin, "Design of Phased Arrays of
Series-Fed Patch Antennas With Reduced Number of the Controllers for 28-GHz mm-Wave
Applications," IEEE Antennas and Wireless Propagation Letters, vol. 15, pp. 1305-1308,
2016.
[27] N. Ojaroudiparchin, M. Shen, S. Zhang, and G. F. Pedersen, "A Switchable 3-D-Coverage-
Phased Array Antenna Package for 5G Mobile Terminals," IEEE Antennas and Wireless
Propagation Letters, vol. 15, pp. 1747-1750, 2016.
[28] J. W. Lian, Y. L. Ban, Q. L. Yang, B. Fu, Z. F. Yu, and L. K. Sun, "Planar Millimeter-Wave
2-D Beam-Scanning Multibeam Array Antenna Fed by Compact SIW Beam-Forming
Network," IEEE Transactions on Antennas and Propagation, vol. 66, pp. 1299-1310, 2018.
[29] C. A. Balanis, "Antenna theory analysis and design," USA: John Wiley and Sons, 2005.
[30] O. M. Haraz, A. Elboushi, S. A. Alshebeili, and A. R. Sebak, "Dense Dielectric Patch Array
Antenna With Improved Radiation Characteristics Using EBG Ground Structure and
Dielectric Superstrate for Future 5G Cellular Networks," Access, IEEE, vol. 2, pp. 909-913,
2014.
[31] A. Sarkar, K. Greene, and B. Floyd, "A power-efficient 4-element beamformer in 120-nm
SiGe BiCMOS for 28-GHz cellular communications," in Bipolar/BiCMOS Circuits and
Technology Meeting (BCTM), 2014 IEEE, 2014, pp. 68-71.
[32] K. Jui-Chih, H. Yuan-Hung, Y. Kuang-Sheng, C. Chau-Ching, L. Yu-Hsuan, L. Kun-You, et
al., "A 25-to-45-GHz 45ᴼ power divider," in Microwave Conference (EuMC), 2013 European,
2013, pp. 959-962.
[33] M. A. Islam and N. C. Karmakar, "A 4*4 Dual Polarized mm-Wave ACMPA Array for a
Universal mm-Wave Chipless RFID Tag Reader," IEEE Transactions on Antennas and
Propagation, vol. 63, pp. 1633-1640, 2015.
[34] W. Lin and H. Wong, "Wideband Circular-Polarization Reconfigurable Antenna With L-
Shaped Feeding Probes," IEEE Antennas and Wireless Propagation Letters, vol. 16, pp. 2114-
2117, 2017.
[35] K. M. Mak, H. W. Lai, K. M. Luk, and K. L. Ho, "Polarization Reconfigurable Circular Patch
Antenna With a C-Shaped," IEEE Transactions on Antennas and Propagation, vol. 65, pp.
1388-1392, 2017.

92
[36] W. T. Li, S. Gao, Y. M. Cai, Q. Luo, M. Sobhy, G. Wei, et al., "Polarization-Reconfigurable
Circularly Polarized Planar Antenna Using Switchable Polarizer," IEEE Transactions on
Antennas and Propagation, vol. 65, pp. 4470-4477, Sep 2017.
[37] F. Farzami, S. Khaledian, B. Smida, and D. Erricolo, "Reconfigurable Linear/Circular
Polarization Rectangular Waveguide Filtenna," IEEE Transactions on Antennas and
Propagation, vol. 66, pp. 9-15, 2018.
[38] J. Hu, Z. C. Hao, and Z. W. Miao, "Design and Implementation of a Planar Polarization-
Reconfigurable Antenna," IEEE Antennas and Wireless Propagation Letters, vol. 16, 2017.
[39] L. Zhang, S. Gao, Q. Luo, P. R. Young, and Q. Li, "Wideband Loop Antenna With
Electronically Switchable Circular Polarization," IEEE Antennas and Wireless Propagation
Letters, vol. 16, pp. 242-245, 2017.
[40] W. Lin, S. Chen, R. W. Ziolkowski, and Y. J. Guo, "Reconfigurable, Wideband, Low-Profile,
Circularly Polarized Antenna and Array Enabled by an Artificial Magnetic Conductor
Ground," IEEE Transactions on Antennas and Propagation, vol. 66, pp. 1564-1569, 2018.
[41] S. W. Lee and Y. J. Sung, "Simple Polarization-Reconfigurable Antenna With T-Shaped
Feed," IEEE Antennas and Wireless Propagation Letters, vol. 15, pp. 114-117, 2016.
[42] L. Ji, P. Qin, Y. J. Guo, C. Ding, G. Fu, and S. Gong, "A Wideband Polarization
Reconfigurable Antenna With Partially Reflective Surface," IEEE Transactions on Antennas
and Propagation, vol. 64, pp. 4534-4538, 2016.
[43] H. C. Sun and S. Sun, "A Novel Reconfigurable Feeding Network for Quad-Polarization-
Agile Antenna Design," IEEE Transactions on Antennas and Propagation, vol. 64, pp. 311-
316, Jan 2016.
[44] A. Clemente, L. Dussopt, R. Sauleau, P. Potier, and P. Pouliguen, "1-Bit Reconfigurable Unit
Cell Based on PIN Diodes for Transmit-Array Applications in X-Band," IEEE Transactions
on Antennas and Propagation, vol. 60, pp. 2260-2269, May 2012.
[45] M. Asaadi and A. Sebak, "High-Gain Low-Profile Circularly Polarized Slotted SIW Cavity
Antenna for MMW Applications," IEEE Antennas and Wireless Propagation Letters, vol. 16,
pp. 752-755, 2017.
[46] D. J. Bisharat, S. Liao, and Q. Xue, "High Gain and Low Cost Differentially Fed Circularly
Polarized Planar Aperture Antenna for Broadband Millimeter-Wave Applications," IEEE
Transactions on Antennas and Propagation, vol. 64, pp. 33-42, 2016.
[47] S. J. Park and S. O. Park, "LHCP and RHCP Substrate Integrated Waveguide Antenna Arrays
for Millimeter-Wave Applications," IEEE Antennas and Wireless Propagation Letters, vol.
16, pp. 601-604, 2017.
[48] K. X. Wang and H. Wong, "A Wideband Millimeter-Wave Circularly Polarized Antenna With
3-D Printed Polarizer," IEEE Transactions on Antennas and Propagation, vol. 65, pp. 1038-
1046, 2017.
[49] A. B. Guntupalli and K. Wu, "Frequency-steered directive beam with dual circular
polarization and two-dimensional scan capability for millimeter-wave imaging and sensing
systems," in WAMICON 2014, 2014, pp. 1-3.
[50] Q. Wu, J. Hirokawa, J. Yin, C. Yu, H. Wang, and W. Hong, "Millimeter-Wave Multibeam
Endfire Dual-Circularly Polarized Antenna Array for 5G Wireless Applications," IEEE
Transactions on Antennas and Propagation, vol. 66, pp. 4930-4935, 2018.
[51] M. S. Sharawi, "Current Misuses and Future Prospects for Printed Multiple-Input, Multiple-
Output Antenna Systems [Wireless Corner]," IEEE Antennas and Propagation Magazine, vol.
59, pp. 162-170, 2017.
[52] B. Q. Yang, Z. Q. Yu, J. Lan, R. Q. Zhang, J. Y. Zhou, and W. Hong, "Digital Beamforming-
Based Massive MIMO Transceiver for 5G Millimeter-Wave Communications," IEEE
Transactions on Microwave Theory and Techniques, vol. 66, pp. 3403-3418, Jul 2018.
[53] Y. Li, C. Sim, Y. Luo, and G. Yang, "12-Port 5G Massive MIMO Antenna Array in Sub-
6GHz Mobile Handset for LTE Bands 42/43/46 Applications," IEEE Access, vol. 6, pp. 344-
354, 2018.
93
[54] Y. Ban, C. Li, C. Sim, G. Wu, and K. Wong, "4G/5G Multiple Antennas for Future Multi-
Mode Smartphone Applications," IEEE Access, vol. 4, pp. 2981-2988, 2016.
[55] M. Y. Li, Y. L. Ban, Z. Q. Xu, G. Wu, C. Y. D. Sim, K. Kang, et al., "Eight-Port Orthogonally
Dual-Polarized Antenna Array for 5G Smartphone Applications," IEEE Transactions on
Antennas and Propagation, vol. 64, pp. 3820-3830, Sep 2016.
[56] Y. Li, C. Sim, Y. Luo, and G. Yang, "Multiband 10-Antenna Array for Sub-6 GHz MIMO
Applications in 5-G Smartphones," IEEE Access, vol. 6, pp. 28041-28053, 2018.
[57] S. C. Chen, P. W. Wu, C. I. G. Hsu, and J. Y. Sze, "Integrated MIMO Slot Antenna on Laptop
Computer for Eight-Band LTE/WWAN Operation," IEEE Transactions on Antennas and
Propagation, vol. 66, pp. 105-114, Jan 2018.
[58] M. Ikram, R. Hussain, and M. S. Sharawi, "4G/5G antenna system with dual function planar
connected array," IET Microwaves, Antennas & Propagation, vol. 11, pp. 1760-1764, 2017.
[59] R. Hussain, A. T. Alreshaid, S. K. Podilchak, and M. S. Sharawi, "Compact 4G MIMO
antenna integrated with a 5G array for current and future mobile handsets," IET Microwaves,
Antennas & Propagation, vol. 11, pp. 271-279, 2017.
[60] C. Lee, M. K. Khattak, and S. Kahng, "Wideband 5G beamforming printed array clutched by
LTE-A 4 × 4-multiple-input–multiple-output antennas with high isolation," IET Microwaves,
Antennas & Propagation, vol. 12, pp. 1407-1413, 2018.
[61] M. Khalily, R. Tafazolli, P. Xiao, and A. A. Kishk, "Broadband mm-Wave Microstrip Array
Antenna With Improved Radiation Characteristics for Different 5G Applications," IEEE
Transactions on Antennas and Propagation, vol. 66, pp. 4641-4647, 2018.
[62] D. S. Woo, Y.-K. Cho, and K. W. Kim, "Balance Analysis of Microstrip-to-CPS Baluns and
Its Effects on Broadband Antenna Performance," International Journal of Antennas and
Propagation, vol. 2013, p. 9, 2013.
[63] J. Li, S. Yan, Y. Liu, B. M. Hochwald, and J. Jin, "A High-Order Model for Fast Estimation
of Electromagnetic Absorption Induced by Multiple Transmitters in Portable Devices," IEEE
Transactions on Antennas and Propagation, vol. 65, pp. 6768-6778, 2017.
[64] A. T. Mobashsher, P. T. Nguyen, and A. Abbosh, "Detection and localization of brain strokes
in realistic 3-D human head phantom," in 2013 IEEE MTT-S International Microwave
Workshop Series on RF and Wireless Technologies for Biomedical and Healthcare
Applications (IMWS-BIO), 2013, pp. 1-3.
[65] S. Gabriel, R. W. Lau, and C. Gabriel, "The dielectric properties of biological tissues: II.
Measurements in the frequency range 10 Hz to 20 GHz," Physics in Medicine & Biology, vol.
41, p. 2251, 1996.
[66] W. He, B. Xu, M. Gustafsson, Z. Ying, and S. He, "RF Compliance Study of Temperature
Elevation in Human Head Model Around 28 GHz for 5G User Equipment Application:
Simulation Analysis," IEEE Access, vol. 6, pp. 830-838, 2018.
[67] D. M. Pozar, Microwave engineering, 4th ed. Hoboken, NJ: Wiley, 2012.
[68] K. Nakada, "Stub switched phase shifter," Japan Patent US09695879, 2000.
[69] "MACOM Technologies,” [Online]. Available: http://www.macom.com."
[70] L. D. Palma, A. Clemente, L. Dussopt, R. Sauleau, P. Potier, and P. Pouliguen, "1-Bit
Reconfigurable Unit Cell for Ka-Band Transmitarrays," IEEE Antennas and Wireless
Propagation Letters, vol. 15, pp. 560-563, 2016.
[71] U. T. Ahmed and A. M. Abbosh, "Three-octave band in-phase power divider using loosely
coupled microstrip lines and microstrip to slotline transitions," Electronics Letters, vol. 51,
pp. 2019-2021, 2015.
[72] U. T. Ahmed and A. M. Abbosh, "Modified Wilkinson power divider using coupled
microstrip lines and shunt open-ended stubs," Electronics Letters, vol. 51, pp. 838-839, 2015.
[73] J. B. Knorr, "Slot-Line Transitions (Short Papers)," IEEE Transactions on Microwave Theory
and Techniques, vol. 22, pp. 548-554, 1974.

94
[74] R. Janaswamy and H. D. Schaubert, "Dispersion Characteristics for Wide Slotlines on Low-
Permittivity Substrates (Short Paper)," IEEE Transactions on Microwave Theory and
Techniques, vol. 33, pp. 723-726, 1985.
[75] P. B. Katehi and N. G. Alexopoulos, "Frequency-Dependent Characteristics of Microstrip
Discontinuities in Millimeter-Wave Integrated Circuits," IEEE Transactions on Microwave
Theory and Techniques, vol. 33, pp. 1029-1035, 1985.
[76] Y. Li, J. Wang, and K. M. Luk, "Millimeter-Wave MultiBeam Aperture-Coupled
Magnetoelectric Dipole Array With Planar Substrate Integrated Beamforming Network for
5G Applications," IEEE Transactions on Antennas and Propagation, vol. 65, pp. 6422-6431,
2017.
[77] X. Ding and B. Wang, "A millimeter-wave pattern-reconfigurable antenna with a
reconfigurable feeding network," Journal of Electromagnetic Waves and Applications, vol.
27, pp. 649-658, 2013/03/01 2013.
[78] E. Topak, J. Hasch, C. Wagner, and T. Zwick, "A Novel Millimeter-Wave Dual-Fed Phased
Array for Beam Steering," IEEE Transactions on Microwave Theory and Techniques, vol. 61,
pp. 3140-3147, 2013.
[79] T. Wen-Hua and C. Kai, "Compact second harmonic-suppressed bandstop and bandpass
filters using open stubs," IEEE Transactions on Microwave Theory and Techniques, vol. 54,
pp. 2497-2502, 2006.
[80] R. Mongia and R. Mongia, RF and microwave coupled-line circuits, 2nd ed. Boston: Artech
House, 2007.
[81] A. M. Abbosh, "Analytical closed-form solutions for different configurations of parallel-
coupled microstrip lines," IET Microwaves, Antennas & Propagation, vol. 3, pp. 137-147,
2009.
[82] A. M. Abbosh, "Three-Way Parallel-Coupled Microstrip Power Divider With Ultrawideband
Performance and Equal-Power Outputs," Microwave and Wireless Components Letters,
IEEE, vol. 21, pp. 649-651, 2011.
[83] A. M. Abbosh, "Compact Tunable Reflection Phase Shifters Using Short Section of Coupled
Lines," IEEE Transactions on Microwave Theory and Techniques, vol. 60, pp. 2465-2472,
2012.
[84] Y. J. Sung, T. U. Jang, and Y. S. Kim, "A reconfigurable microstrip antenna for switchable
polarization," IEEE Microwave and Wireless Components Letters, vol. 14, pp. 534-536, 2004.
[85] R. V. Garver, "Broad-Band Diode Phase Shifters," IEEE Transactions on Microwave Theory
and Techniques, vol. 20, pp. 314-323, 1972.
[86] G. F. Engen and C. A. Hoer, "Thru-Reflect-Line: An Improved Technique for Calibrating the
Dual Six-Port Automatic Network Analyzer," IEEE Transactions on Microwave Theory and
Techniques, vol. 27, pp. 987-993, 1979.
[87] V. Rathi, G. Kumar, and K. P. Ray, "Improved coupling for aperture coupled microstrip
antennas," IEEE Transactions on Antennas and Propagation, vol. 44, pp. 1196-1198, 1996.
[88] F. Ren, W. Hong, K. Wu, D. Yu, and Y. Wan, "Polarization-Adjustable Planar Array Antenna
With SIW-Fed High-Order-Mode Microstrip Patch," IEEE Transactions on Antennas and
Propagation, vol. 65, pp. 6167-6172, 2017.
[89] X. Liu, Q. Zhang, W. H. Chen, H. G. Feng, L. Chen, F. M. Ghannouchi, et al., "Beam-
Oriented Digital Predistortion for 5G Massive MIMO Hybrid Beamforming Transmitters,"
IEEE Transactions on Microwave Theory and Techniques, vol. 66, pp. 3419-3432, Jul 2018.
[90] F. M. Barradas, P. M. Tome, J. M. Gomes, T. R. Cunha, P. M. Cabral, and J. C. Pedro, "Power,
Linearity, and Efficiency Prediction for MIMO Arrays With Antenna Coupling," IEEE
Transactions on Microwave Theory and Techniques, vol. 65, pp. 5284-5297, Dec 2017.
[91] S. Hwangbo, A. Rahimi, and Y. K. Yoon, "Cu/Co Multilayer-Based High Signal Integrity and
Low RF Loss Conductors for 5G/Millimeter Wave Applications," IEEE Transactions on
Microwave Theory and Techniques, vol. 66, pp. 3773-3780, Aug 2018.

95
[92] B. Bangerter, S. Talwar, R. Arefi, and K. Stewart, "Networks and devices for the 5G era,"
IEEE Communications Magazine, vol. 52, pp. 90-96, 2014.
[93] R. Garg and A. S. Natarajan, "A 28-GHz Low-Power Phased-Array Receiver Front-End With
360 degrees RTPS Phase Shift Range," IEEE Transactions on Microwave Theory and
Techniques, vol. 65, pp. 4703-4714, Nov 2017.
[94] M. H. Novak, J. L. Volakis, and F. A. Miranda, "Ultra-wideband array in PCB for millimeter-
wave 5G and ISM," in 2017 International Workshop on Antenna Technology: Small Antennas,
Innovative Structures, and Applications (iWAT), 2017, pp. 108-110.
[95] A. El-Bacha and R. Sarkis, "Design of Tilted Taper Slot Antenna for 5G Base Station Antenna
Circular Array," 2016 IEEE Middle East Conference on Antennas and Propagation (Mecap),
2016.
[96] Y. Yao, X. H. Cheng, C. W. Wang, J. S. Yu, and X. D. Chen, "Wideband Circularly Polarized
Antipodal Curvedly Tapered Slot Antenna Array for 5G Applications," IEEE Journal on
Selected Areas in Communications, vol. 35, pp. 1539-1549, Jul 2017.
[97] M. Gholamrezaei, F. Geran, and R. A. Sadeghzadeh, "Completely Independent Multi-
Ultrawideband and Multi-Dual-Band Frequency Reconfigurable Annular Sector Slot Antenna
(FR-ASSA)," IEEE Transactions on Antennas and Propagation, vol. 65, pp. 893-898, Feb
2017.
[98] Z. H. Cai, Y. H. Qi, Z. B. Weng, W. Yu, F. H. Li, and J. Fan, "DC ground compact wideband
omnidirectional vertically polarised slot loop antenna for 4G long-term evolution
applications," IET Microwaves Antennas & Propagation, vol. 12, pp. 1087-1092, Jun 13
2018.
[99] K. Ebnabbasi, D. Busuioc, R. Birken, and M. Wang, "Taper Design of Vivaldi and Co-Planar
Tapered Slot Antenna (TSA) by Chebyshev Transformer (vol 60, pg 2252, 2012)," IEEE
Transactions on Antennas and Propagation, vol. 60, pp. 3077-3077, Jun 2012.
[100] E. Al Abbas, A. T. Mobashsher, and A. Abbosh, "Polarization Reconfigurable Antenna for
5G Cellular Networks Operating at Millimeter Waves," 2017 IEEE Asia Pacific Microwave
Conference (Apmc), pp. 772-774, 2017.
[101] D. Q. Yang, S. H. Liu, and D. Gengi, "A Miniaturized Ultra-Wideband Vivaldi Antenna With
Low Cross Polarization," IEEE Access, vol. 5, pp. 23352-23357, 2017.
[102] U. T. Ahmed and A. M. Abbosh, "In-Phase Power Divider with Three Octaves Band Using
Microstrip to Slotline Transitions and Loosely Coupled Microstrip Lines," 2015 Asia-Pacific
Microwave Conference (Apmc), Vols 1-3, 2015.
[103] C. I. Lin and K. L. Wong, "Printed monopole slot antenna for internal multiband mobile phone
antenna," IEEE Transactions on Antennas and Propagation, vol. 55, pp. 3690-3697, Dec
2007.
[104] Y. H. Xu, J. P. Wang, L. Ge, X. D. Wang, and W. Wu, "Design of a Notched-Band Vivaldi
Antenna With High Selectivity," IEEE Antennas and Wireless Propagation Letters, vol. 17,
pp. 62-65, 2018.
[105] F. De Paolis and F. Frezza, "Simplified Prediction of Peak Power-Handling Capability for
Stepped-Impedance Low-Pass Filters," IEEE Transactions on Microwave Theory and
Techniques, vol. 61, pp. 1079-1085, Mar 2013.
[106] J. Ni and J. S. Hong, "Compact Continuously Tunable Microstrip Low-Pass Filter," IEEE
Transactions on Microwave Theory and Techniques, vol. 61, pp. 1793-1800, May 2013.
[107] K. L. Wong and C. Y. Tsai, "Low-Profile Dual-Wideband Inverted-T Open Slot Antenna for
the LTE/WWAN Tablet Computer With a Metallic Frame," IEEE Transactions on Antennas
and Propagation, vol. 63, pp. 2879-2886, Jul 2015.
[108] Q. X. Guo, R. Mittra, F. Lei, Z. R. Li, J. L. Ju, and J. Byun, "Interaction Between Internal
Antenna and External Antenna of Mobile Phone and Hand Effect," Ieee Transactions on
Antennas and Propagation, vol. 61, pp. 862-870, Feb 2013.

96
[109] S. Zhang, X. Chen, I. Syrytsin, and G. F. Pedersen, "A Planar Switchable 3-D-Coverage
Phased Array Antenna and Its User Effects for 28-GHz Mobile Terminal Applications," IEEE
Transactions on Antennas and Propagation, vol. 65, pp. 6413-6421, 2017.
[110] N. Nguyen-Trong, L. Hall, and C. Fumeaux, "A Frequency-and Polarization-Reconfigurable
Stub-Loaded Microstrip Patch Antenna," IEEE Transactions on Antennas and Propagation,
vol. 63, pp. 5235-5240, Nov 2015.
[111] H. A. Majid, M. K. A. Rahim, M. R. Hamid, and M. F. Ismail, "Frequency and Pattern
Reconfigurable Slot Antenna," IEEE Transactions on Antennas and Propagation, vol. 62, pp.
5339-5343, Oct 2014.
[112] S. N. Yang, C. N. Zhang, H. K. Pan, A. E. Fathy, and V. K. Nair, "Frequency-Reconfigurable
Antennas for Multiradio Wireless Platforms," IEEE Microwave Magazine, vol. 10, pp. 66-83,
Feb 2009.
[113] N. Behdad and K. Sarabandi, "A varactor-tuned dual-band slot antenna," IEEE Transactions
on Antennas and Propagation, vol. 54, pp. 401-408, Feb 2006.
[114] H. Li, J. A. Xiong, Y. F. Yu, and S. L. He, "A Simple Compact Reconfigurable Slot Antenna
With a Very Wide Tuning Range," IEEE Transactions on Antennas and Propagation, vol. 58,
pp. 3725-3728, Nov 2010.
[115] C. C. Hung and T. Chiu, "Dual-Band Reconfigurable Antenna Design Using Slot-Line With
Branch Edge," IEEE Transactions on Antennas and Propagation, vol. 63, pp. 508-516, Feb
2015.
[116] P. Y. Qin, Y. J. Guo, Y. Cai, E. Dutkiewicz, and C. H. Liang, "A Reconfigurable Antenna
With Frequency and Polarization Agility," IEEE Antennas and Wireless Propagation Letters,
vol. 10, pp. 1373-1376, 2011.
[117] H. Gu, J. P. Wang, and L. Ge, "Circularly Polarized Patch Antenna With Frequency
Reconfiguration," IEEE Antennas and Wireless Propagation Letters, vol. 14, pp. 1770-1773,
2015.
[118] S. N. M. Zainarry, N. T. Nghia, and C. Fumeaux, "A Frequency- and Pattern-Reconfigurable
Two-Element Array Antenna," IEEE Antennas and Wireless Propagation Letters, vol. 17, pp.
617-620, Apr 2018.
[119] N. Nguyen-Trong and C. Fumeaux, "Tuning Range and Efficiency Optimization of a
Frequency-Reconfigurable Patch Antenna," IEEE Antennas and Wireless Propagation
Letters, vol. 17, pp. 150-154, Jan 2018.
[120] D. Cavallo, W. H. Syed, and A. Neto, "Connected-Slot Array With Artificial Dielectrics: A 6
to 15 GHz Dual-Pol Wide-Scan Prototype," IEEE Transactions on Antennas and
Propagation, vol. 66, pp. 3201-3206, Jun 2018.
[121] I. Tzanidis, K. Sertel, and J. L. Volakis, "UWB Low-Profile Tightly Coupled Dipole Array
With Integrated Balun and Edge Terminations," IEEE Transactions on Antennas and
Propagation, vol. 61, pp. 3017-3025, Jun 2013.
[122] L. Zhu and K. Wu, "Accurate circuit model of interdigital capacitor and its application to
design of new quasi-lumped miniaturized filters with suppression of harmonic resonance (vol
48, pg 347, 2000)," IEEE Transactions on Microwave Theory and Techniques, vol. 50, pp.
2412-2413, Oct 2002.
[123] O. M. Haraz, A. Elboushi, S. A. Alshebeili, and A. Sebak, "Dense Dielectric Patch Array
Antenna With Improved Radiation Characteristics Using EBG Ground Structure and
Dielectric Superstrate for Future 5G Cellular Networks," IEEE Access, vol. 2, pp. 909-913,
2014.
[124] S. F. Jilani and A. Alomainy, "A Multiband Millimeter-Wave 2-D Array Based on Enhanced
Franklin Antenna for 5G Wireless Systems," IEEE Antennas and Wireless Propagation
Letters, vol. 16, pp. 2983-2986, 2017.
[125] S. Zhu, H. Liu, Z. Chen, and P. Wen, "A Compact Gain-Enhanced Vivaldi Antenna Array
With Suppressed Mutual Coupling for 5G mmWave Application," IEEE Antennas and
Wireless Propagation Letters, vol. 17, pp. 776-779, 2018.
97
[126] S. X. Ta, H. Choo, and I. Park, "Broadband Printed-Dipole Antenna and Its Arrays for 5G
Applications," IEEE Antennas and Wireless Propagation Letters, vol. 16, pp. 2183-2186,
2017.
[127] C. Lee, M. K. Khattak, and S. Kahng, "Wideband 5G beamforming printed array clutched by
LTE-A 4x4-multiple-input-multiple-output antennas with high isolation," IET Microwaves
Antennas & Propagation, vol. 12, pp. 1407-1413, Jul 4 2018.

98