Screw Compressors

MANIPULATION OF SOUND
PROPERTIES BY ACOUSTIC
METASURFACE AND
METASTRUCTURE
JIAJUN ZHAO
(B.Eng., Nanjing University)
A THESIS SUBMITTED
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
DEPARTMENT OF ELECTRICAL AND COMPUTER
ENGINEERING
NATIONAL UNIVERSITY OF SINGAPORE
2015
Declaration
I, JIAJUN ZHAO, hereby declare that this thesis is my original work and
it has been written by me in its entirety. I have duly acknowledged all the
sources of information which have been used in the thesis.
This thesis has also not been submitted for any degree in any university
previously.
Signed: Jiajun ZHAO
Date:
January 2016
i
Acknowledgements
It is my great pleasure to thank everyone who has kindly helped me during

my PhD period. First and foremost, I would like to sincerely thank Dr.
Cheng-Wei Qiu (National University of Singapore) and Dr. Baowen Li
(University of Colorado Boulder). Dr. Qiu and Dr. Li are very supportive
of my research and have given me the freedom to pursue various projects
without objection. They have also provided insightful discussions on my
research. I am also very grateful to Dr Zhi Ning Chen (National University
of Singapore/Institute for Infocomm Research) for his scientific advice and
knowledge, and many discussions and suggestions.
I specially would like to thank all my colleagues at NUS for their invaluable
advice. A special thank you to Qasim Mehmood and Mohammad Danesh,
for the great memories together and the sincere friendship. My thanks also
go out to the support I am receiving from the University of Texas at Austin,
where I am now working as a research scholar. I am especially grateful to
Dr. Harry Swinney, Dr. Likun Zhang and Dr. Michael Allshouse for their
patience, motivation, and immense knowledge in science. I am also deeply
grateful to my parents for their constant encouragement and endless love.
I acknowledge the support of the Presidential Fellowship I am granted with

by NUS and Singapore. For the non-scientific side of my thesis, I par-
ticularly want to thank “Living Stone” in Nanjing, “CPBC” in Singapore
and “Gracepoint” in Austin for their encouragement. Special thanks to
Tim Fitz, Debbie Fitz, Hongning Zhu, Dr. Franco Lim, George Butron,
Susanna Lee and Dr. Jonathan Lee.
ii
Summary
Manipulation of sound properties has attracted the interest of acoustical

engineers for a long time. Using the mathematical method called trans-
formation acoustics, researchers have demonstrated a host of striking un-
precedented devices with unusual sound properties and phenomena, such as
negative density, negative bulk modulus, extraordinary sound transmission,
super resolution, etc. These devices that are engineered to have acoustic
properties not yet found in nature are called acoustic metamaterials. Here,
we extend the concept of metamaterial to metasurface and metastructure,
which are made from assemblies of various elements fashioned from tradi-
tional materials such as copper or membrane.
Different from acoustic metamaterial which enables abnormal sound prop-

erties inside itself, the proposed acoustic metasurface is a thin-layer struc-
ture which allows the change of sound properties right at its surface. Thus,
instead of gradual change inside metamaterial, acoustic waves seem to have
an abrupt change after touching metasurface.
Besides, inasmuch as acoustic metamaterial is based on transformation

acoustics, the conformal mapping of coordinates inevitably leads to com-
plex parameters. Realizing these parameters in metamaterial is usually
challenging. Here, we also propose the concept of acoustic metastructure,
which does not rely on transformation acoustics. By a unique design, the
iv
acoustic metastructure has a much accessible layout and does not have
components of microscopic or smaller scales.
The work done related to the thesis during my doctoral program includes:
theoretically and simulationally constructing the acoustic metasurface of
inhomogeneous acoustic impedance for various applications such as acous-
tic disguise, acoustic planar lenses, acoustic ipsilateral imaging, and the
conversion from propagating to surface acoustic waves; extending the prior
proposed structure to be three dimensional, resulting in the out-of-incident-
plane fluid-particle vibration; optimizing acoustic focusing for medical and
industrial applications such as focused ultrasound surgery, lithotripsy, and
nondestructive testing; proposing a density-near-zero metastructure to pro-
vide an accessible way for acoustic cloaking.
In this thesis, we elaborate the design of acoustic metasurface and metas-

tructure, and their functionality in manipulation of sound properties. In
Chapter 1, the background of the research is given; In Chapter 2, a metasur-
face with inhomogeneous acoustic impedance is proposed to achieve acous-
tic wavefront manipulation; In Chapter 3, a metasurface is designed to
tweak the vibrational orientation of sound; In Chapter 4, an active metasur-
face piezoelectric transducer is used to control acoustic focusing; In Chapter
5, the near-zero density is obtained by a metastructure; In Chapter 6, we
envision some possible future works.
Contents
Declaration i
Acknowledgements ii
Summary iii
List of Tables viii
List of Figures ix
List of Abbreviations xi
List of Symbols xii
1 Introduction 1
1.1 Reshaping wavefronts in optics and acoustics . . . . . . . . . 2
1.2 Vibrational direction in fluids and solids . . . . . . . . . . . 3
1.3 Acoustic focusing and piezoelectric transducer . . . . . . . . 4
1.4 Invisibility cloaking in optics and acoustics . . . . . . . . . . 6
1.5 Outlines of Chapters 2 - 5 . . . . . . . . . . . . . . . . . . . 7
1.6 My published journal articles related to thesis . . . . . . . . 9
v
Table of Contents vi
2 Manipulating acoustic wavefront with metasurface of inho-

mogeneous impedance 10
2.1 Extraordinary reflection and ordinary reflection . . . . . . . 11
2.2 Continuous and discontinuous impedance . . . . . . . . . . . 14
2.3 Acoustic illusion and ipsilateral focusing . . . . . . . . . . . 22
2.4 Conversion from propagating to surface waves . . . . . . . . 27
2.5 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.6 Conclusion and discussion . . . . . . . . . . . . . . . . . . . 28
3 Redirecting acoustic waves out of the incident plane 30
3.1 Out-of-incident-plane vibration of fluid particles . . . . . . . 31
3.2 Two-dimensional varying metasurface . . . . . . . . . . . . . 36
3.3 Three-dimensional control of extraordinary reflection . . . . 41
3.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
4 Manipulating acoustic focus with an active metasurface

piezoelectric transducer 45
4.1 Design of planar metasurface piezoelectric transducer . . . . 46
4.2 Generation of an acoustic focal needle . . . . . . . . . . . . . 51
4.3 Generation of acoustic far-field multiple foci . . . . . . . . . 54
4.4 Acoustic super-oscillatory super resolution . . . . . . . . . . 58
4.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
4.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
5 Realizing acoustic cloaking and near-zero density with acous-

tic metastructure 61
5.1 Metastructure for acoustic cloaking made by copper . . . . . 65
5.2 Spring-mass model and density-near-zero property . . . . . . 68

Table of Contents vii
5.3 Coupling model and geometrical dependence . . . . . . . . . 72
5.4 Deploying cloaked area in free space and in waveguides . . . 75
5.5 Conclusion and discussion . . . . . . . . . . . . . . . . . . . 79
6 Future work 81
6.1 Manipulation of band gap for sound transmission . . . . . . 82
6.2 Manipulation of acoustic properties in stratified fluids . . . . 83
A Detailed derivation of impedance-governed generalized Snell’s

law of reflection (IGSL) in acoustics 86
B Distinction of IGSL in acoustics 94
C Parameters of metasurface piezoelectric transducer 97
C.1 Parameters of Lead Zirconate Titanate PZT-5H . . . . . . . 97
C.2 Optimized configuration of ring pattern . . . . . . . . . . . . 98
D Details of the simulation for internal waves 100
Bibliography 102
List of Tables
C.1 Symmetric elasticity matrix (Pa) . . . . . . . . . . . . . . . 97
C.2 Relative permittivity εSr . . . . . . . . . . . . . . . . . . . . 98
C.3 Coupling matrix e (C/m2 ) . . . . . . . . . . . . . . . . . . . 98
C.4 Configuration of 30 rings to generate a focal needle (mm) . . 98
C.5 Configuration of 28 rings to generate multiple foci (mm) . . 99
C.6 Configuration of 19 rings to generate super-oscillatory super

resolution (mm) . . . . . . . . . . . . . . . . . . . . . . . . . 99
viii
List of Figures
2.1 A flat interface with inhomogeneous acoustic impedance . . 12
2.2 Ordinary, extraordinary and negative extraordinary reflection 16
2.3 Oblique incidence upon metasurface . . . . . . . . . . . . . . 17
2.4 Schematics of hard-sidewall tubes . . . . . . . . . . . . . . . 19
2.5 Acoustic illusion and ipsilateral focusing . . . . . . . . . . . 23
2.6 Conversion from propagating to surface acoustic waves . . . 26
3.1 Out-of-incident-plane cross vibration of fluid particles . . . . 32
3.2 Two-dimensional varying metasurface . . . . . . . . . . . . . 35
3.3 Out-of-incident-plane extraordinary reflection . . . . . . . . 42
4.1 Schematics of metasurface piezoelectric transducer . . . . . . 48
4.2 The designed finite-length far-field focal needle . . . . . . . . 52
4.3 The designed far-field multiple foci . . . . . . . . . . . . . . 55
4.4 Focal size of super-oscillatory super resolution . . . . . . . . 57
5.1 Schematics of the metastructure for acoustic cloaking . . . . 63
5.2 Sound transmission through metastructure and its resonance 64
5.3 Spring-mass model and its coupling model . . . . . . . . . . 67
5.4 Acoustic cloaking in unbounded space . . . . . . . . . . . . . 74
5.5 “NUS”-shaped cloaked area in unbounded space . . . . . . . 76
ix
List of Figures x
5.6 Acoustic cloaking in curved waveguides . . . . . . . . . . . . 77
6.1 Direct numerical simulation of internal wave radiation for

tidal flow over synthetic random topography . . . . . . . . . 85
A.1 Illustration for mathematical derivation . . . . . . . . . . . . 87
B.1 Difference between GSL in optics and IGSL in acoustics . . . 96

List of Abbreviations
1D one dimension; one-dimensional

2D two dimension; two-dimensional
3D three dimension; three-dimensional
SAI specific acoustic impedance
GSL generalized Snell’s law of reflection
IGSL impedance-governed generalized Snell’s law of reflection
PAW propagating acoustic wave
SAW surface acoustic wave
FEM finite element method
PT piezoelectric transducer
RSI acoustic Rayleigh-Sommerfeld diffraction integral
BPSO binary particle swarm optimization
FWHM full width at half maximum
PZT lead zirconate titanate
B.C. boundary condition
DNZ density near zero
EST extraordinary sound transmission
xi
List of Symbols
ω angular frequency rad/s

p acoustic pressure Pa
pro p of ordinary reflection Pa
θro angle of ordinary reflection rad
pre p of extraordinary reflection Pa
θre angle of extraordinary reflection rad
pi p of incident sound Pa
θi angle of incident sound rad
ρw density of water kg/m3
ρa density of air kg/m3
c0 speed of sound m/s
d discretized spacing m
Zn specific acoustic impedance Pa·s/m
λ wavelength m
k0 wave number rad/m
f frequency Hz
V0 amplitude of electric potential J/C
γ damping coefficient 1/s
xii
Chapter 1
Introduction
Engineering the wavefronts of both electromagnetic and acoustic waves has

long captivated the increasing interest and popularity, such as cloaking,
illusion, focusing, etc. Among various schemes proposed to manipulate the
light, the generalized Snells law has recently been reformulated [1]. This
mechanism opens up new initiatives in realizing negative reflection and
negative refraction, as well as other possibilities in optical wave engineering.
However, there are a few critical problems in this optical scheme: the abrupt
phase change relies on the metallic subwavelength structures, which is not
scalable into other wave forms; the interesting new result derived from
this generalized Snells law is only associated to the anomalous reflected/re-
fracted components, which is very low in intensity.
In principle, all those phenomena can be explained in terms of using a

phased antenna array to steer its beam directions in classical electromag-
netism. Hence, the only unexplored wonderland is in the acoustic domain,
which cannot be simply included in the antenna engineering perspective.
1
Chapter 1. Introduction 2
This chapter addresses the background of my thesis work in acoustics. Sec-

tions 1.1 to 1.4 correspond to the introductions to Chapter 2 - 5, respec-
tively. Section 1.5 addresses the outlines of Chapters 2 - 5.
1.1 Reshaping wavefronts in optics and acous-
tics
Refraction in classic optics was recently revisited from the viewpoints of

complex refractive index of a bulky medium [2], abrupt phase change of
an interface [1], and diffraction theory for gratings [3, 4]. Furthermore,
these works shed light on the relation between the reflection and incidence,
interpreted as the generalized Snell’s law of reflection (GSL) [1], a novel way
to optical wavefront engineering, resulting in promising accomplishments
[5–9]. In optics, the phase-inhomogeneous metasurfaces realized by thin
metallic nano-antennas conserve the wave number along an interface while
impose the extra phase accumulation [1]. Fundamental physics is explained
by a phased antenna array [10].
In principle, GSL is based on Fermat’s principle, which holds for all monochro-
matic waves. However, the luxury of using metallic metasurfaces [1, 5] to
fulfill the optical phase control is no more available in acoustics due to the
limited choice of acoustic materials. Thus, the variable in GSL: the phase
change on a flat surface becomes an abstract concept in acoustics without
any design principle and practical clue. Therefore, it is indispensable and
valuable to establish a different principle to manipulate acoustic waves.
1.2 Vibrational direction in fluids and solids
When an acoustic wave with a certain frequency is excited in fluids, the fluid
particles will experience a restoring force, hence oscillating back and forth
in a monochromatic way. The orientation of such longitudinal oscillation is
the vibrational direction of a fluid particle. The vibration is undoubtedly
an important characteristic of acoustic waves (like the polarization for elec-
tromagnetic waves). In electromagnetism, we can manipulate polarization
by conventional methods such as dichroic crystals, optical gratings, or bire-
fringence effects, etc [11, 12]. In elastic waves, we can also reach the mode
conversion because the molecules in solids can support vibrations in var-
ious directions [13, 14]. However, when sounds propagate freely in fluids,
few attempts were made so far toward tweaking the vibrational orientation,
since the compression mode along the incident plane is considered to be the
only possibility in acoustics. On the other hand, being enabled by the flex-
ible dispersion of metamaterials, acoustic metamaterials can have solid-like
transverse modes at density-near-zero [15] while conversely elastic meta-
materials can have a fluid-like longitudinal mode when the elastic modulus
goes negative [16] to allow polarization conversion. However, these meta-
materials require resonating units, which have to be specially designed to
balance possible loss.
Nevertheless, if one can tweak the reflected sound out of the incident plane,
the vibrational direction, though still longitudinal with respect to the re-
flected beam itself, can therefore be manipulated accordingly. In other
words, we can yield perpendicular vibration components in reflection with
respect to the incident vibration, and control the spatial angle of such out-
of-incident-plane vibration. In this connection, we propose a scheme by
designing an acoustic flat metasurface reflector to manipulate the vibra-

tional orientations generated by sound in fluids. Metasurfaces have drawn
much attention recently in electromagnetism, such as frequency selective
polarizers [17], the wave-form conversion [18], wavefront-engineering flat
lens [19], and polarization converter [20]. The concept of acoustic meta-
surfaces was not well investigated before, owing to the intrinsic nature of
compression modes and limited choices of natural materials.
1.3 Acoustic focusing and piezoelectric trans-
ducer
Research on acoustic focusing has led to various applications such as non-

destructive testing techniques that inspect materials for hidden flaws [21–
23]. Usually piezoelectric transducers (PTs) are the most commonplace
devices serving as the actuators of acoustic focusing. As the mechanism,
when an electric field is applied across piezoelectric materials, the polar-
ized molecules will align with the electric field, causing the materials to
change dimensions [24, 25]. Apart from the industrial applications, acous-
tic focusing utilized in medical science contributes significantly to thera-
peutic techniques as well. Ultrasound waves excited by PTs are capable of
transmitting energy inside a body for medical purposes such as diagnostic
sonography [26]. Other examples include focused ultrasound surgery that
generates localized heating to treat tumors [27, 28], and lithotripsy that
breaks up kidney stones [29].
To avoid the bulky size of a curved PT, the flat annular Fresnel PT has
been invented over decades, reducing the volume of piezoelectric portions
into a flat layer [30]. However as the trade-off, Fresnel PT intrinsically

cannot concentrate the excited acoustic energy completely, as it is always
accompanied with higher-order diffraction. In detail, a planar PT using
a Fresnel equal-spaced array will inevitably generate the parasitic multi-
ple divergent beams and higher-order convergent beams, making the focal
spots less applicable. Actually, there has been to date no such acoustic
design technique that allows us to achieve arbitrarily designed focal pat-
tern along the axis. For example, one expected focal pattern for ultrasonic
surgery is a specific segment of high acoustic energy along the axis (both
its distance away from the PT and its focal depth can be designed), i.e., an
acoustic far-field focal needle, which was never obtained by PTs. A finite-
length focal-needle pattern is also quite promising for particle operation
and acceleration, which was developed in optics [31, 32].
Additionally, the focal resolution created with traditional PTs is usually

low, whose focal size is much larger than one wavelength λ. Since the focal
resolution can be improved with the wave frequency increased, previous
researches usually ignore the consideration of improving the resultant fo-
cal resolution. However, it is noteworthy that an excitation of a higher
frequency demands more energy consumption and suffers from stronger at-
tenuation. Besides, acoustic aberration could also severely blur the focal
resolution. Thus, the rational improvement is to increase the relative focal
resolution with the same excited frequency remaining.
1.4 Invisibility cloaking in optics and acous-
tics
Various metamaterial-based invisibility cloaking has been demonstrated in

optics, acoustics [33, 34] and heat conduction by the theory of transforming
coordinates. As a trade-off, in optics, the spatially-tailored properties of
metamaterials, usually inhomogeneous and anisotropic, impose challenging
complexities in structural configuration and cloaking realization [33, 35].
As the acoustic analog of transformation optics [36], the experimental
realization of acoustic cloaking was reported [37, 38], but its inhomoge-
neous acoustic inertia and modulus caused by coordinate transformation
inevitably result in the same challenges as in optics. More recently, a
topological-optimization method was invented to cancel acoustic scatter-
ing by wave interference [39, 40], which only requires a specific optimized
distribution of rigid boundaries around the object to be hidden. Although
this scheme does not require considering a complex structure of artificial
metamaterials, topological acoustic cloaking highly relies on the shape and
the locus of the object to be hidden. Therefore, the object actually is a part
of the cloaking device itself. It implies that the cloaking structure designed
for one object has to be redesigned for another which has different shapes,
locus, or material composition.
To construct an isotropic acoustic cloak, independent of the cloaked objects

in two-dimensional space or in curved waveguides, could be meaningful in
both theory and application. For example, the isotropic acoustic cloak in
free space could inspire the way of designing stealth planes or submarines
for military purpose.
1.5 Outlines of Chapters 2 - 5
In Chapter 2, we unveil the connection between the acoustic impedance

along a flat surface and the reflected acoustic wavefront, in order to em-
power a wide variety of novel applications in acoustic community. Our
designed flat metasurface can generate double reflections: the ordinary re-
flection and the extraordinary one whose wavefront is manipulated by the
proposed impedance-governed generalized Snell’s law of reflection. The pro-
posed law of reflection is based on Green’s function and integral equation,
instead of Fermat’s principle for optical wavefront manipulation. Remark-
ably, via the adjustment of the designed specific acoustic impedance, ex-
traordinary reflection can be steered for unprecedented acoustic wavefront
while that ordinary reflection can be surprisingly switched on or off. The
realization of the complex discontinuity of the impedance surface has been
proposed using Helmholtz resonators.
In Chapter 3, we demonstrate a flat acoustic metasurface that generates an

extraordinary reflection, and such metasurface can steer the vibration of
the reflection out of the incident plane. When acoustic waves are impinged
on an impedance surface in fluids, it is challenging to alter the vibration
of fluid particles since the vibrational direction of reflected waves shares
the same plane of the incidence and the normal direction of the surface.
Our proposed flat acoustic metasurface can steer the vibration of the re-
flection out of the incident plane. Remarkably, the arbitrary direction of
the extraordinary reflection can be predicted by a Greens-function formu-
lation, and our approach can completely convert the incident waves into
the extraordinary reflection without parasitic ordinary reflection.
In Chapter 4, we demonstrate the manipulation of focal patterns in acous-

tic far fields. It has a pivotal role in medical science and in industry to
concentrate the acoustic energy created with piezoelectric transducers into
a specific area. However, previous researches seldom consider the focal res-
olution, whose focal size is much larger than one wavelength. Furthermore,
there is to date no such design method of piezoelectric transducer that al-
lows a large degree of freedom to achieve designed focal patterns. Here,
an active and configurable planar metasurface PT prototype is proposed to
manipulate the acoustic focal pattern and the focal resolution freely. By
suitably optimized ring configurations of the active metasurface PT, we
demonstrate the manipulation of focal patterns in acoustic far fields, such
as the designed focal needle and multiple focuses. Our method is also able
to manipulate and improve the cross-sectional focal resolution from sub-
wavelength to the extreme case: the deep sub-diffraction-limit resolution.
Via the acoustic Rayleigh-Sommerfeld diffraction integral cum the binary
particle swarm optimization, the free manipulation of focusing properties
is achieved in acoustics for the first time. Our approach may offer more
initiatives where the strict control of acoustic high-energy areas is demand-
ing.
In Chapter 5, isotropic acoustic cloaking is proposed using a density-near-

zero metastructure for extraordinary sound transmission. The metastruc-
ture for acoustic cloaking is made by single-piece homogeneous elastic cop-
per, which can be detached and assembled arbitrarily. We theoretically and
numerically demonstrate the cloaking performance by deploying density-
near-zero metastructures in various ways in two-dimensional space as well
as in acoustic waveguides. The density-near-zero material can make any in-
side objects imperceptible along sound paths. Individually and collectively,
the metastructure maintains both the planar wavefront and the nearly per-
fect one-dimensional transmission, in presence of any inserted object. The
overall cloaked space can be designed by adding metastructures without
the limit of the total cloaked volume.
1.6 My published journal articles related to
thesis
1. J. Zhao, Z. N. Chen, B. Li & C. W. Qiu, Journal of Applied Physics

(2015), DOI: 10.1063/1.4922120
2. J. Zhao, H. Ye, K. Huang, Z. Chen, B. Li & C. W. Qiu, Scientific

Reports (2014), DOI: 10.1038/srep06257
3. J. Zhao, B. Li, Z. Chen & C. W. Qiu, Scientific Reports (2013), DOI:

10.1038/srep02537
4. T. Han, J. Zhao, T. Yuan, D. Y. Lei, B. Li & C. W. Qiu, Energy and

Environmental Science (2013), DOI: 10.1039/c3ee41512k
5. J. Zhao, B. Li, Z. N. Chen & C. W. Qiu, Applied Physics Letters

(2013), DOI: 10.1063/1.4824758
Chapter 2
Manipulating acoustic
wavefront with metasurface of
inhomogeneous impedance
This chapter establishes the framework of acoustic wavefront manipula-

tion by resorting to the acoustic metasurface which has specific acoustic
impedance (SAI) inhomogeneity and discontinuity, rather than the phase
inhomogeneity in terms of wave propagation [1, 2]. SAI is one of the acous-
tic properties of materials, which is comparably more possible to be control-
lable in reality than propagation phase. More specifically, we find out that
the inhomogeneous SAI will generally give rise to one ordinary reflection
pro and one extraordinary reflection pre , i.e., double reflections. Further-
more, the flat inhomogeneous SAI surface is able to switch on or off pro
without the influence on its direction, but to tweak pre in the manner of
our proposed design principle: impedance-governed generalized Snell’s law
of reflection (IGSL) in acoustics.
10
Chapter 2. Manipulating acoustic wavefront 11
2.1 Extraordinary reflection and ordinary re-
flection
The inhomogeneous SAI Zn of the flat surface can be expressed as a com-

plex, whose real and imaginary parts may change spatially. In order to
reduce the complexity of modeling as the beginning attempt, we set the
real part as a spatial constant. Later we prove that the spatial varying
of the real part cannot support our results, which is derived in detail in
Appendix A. We consider

ψ(y)
Zn (y, ω) = A 1 − i tan , (2.1)
2
where A is an arbitrary constant irrelevant to any spatial change and ψ(y)

is the variable for the imaginary part.
Note that ω-dependency on the right hand side of Equation (2.1) has al-
ready been included in ψ(y). The total acoustic pressure p in the upper
space satisfies the integral equation:

k0 √ π
2
y +z − 4 ) 2
p(y, z, ω) ≈ pi + pro − e i(k0
2π y 2 + z 2
∞
ρw c0 cos θ∗
× ∗
× eiψ(y0 ) p(y0 , 0, ω)e−ik0 y0 sin θre dy0 ,
2A cos θ + ρw c0 −∞
(2.2)
where pi denotes the incidence; ρw and c0 are the density and the speed of
sound in the upper space in Figure 2.1(a); k0 = ω/c0 is the wave number;
θ∗ is constant; θre is the angle of pre . Both pro and pre exist for a general
(a) pi Z pro (b) pi Z pro =0
IGS
IGS
pre pre
Ti Ti Tre Ti Ti Tre
L
L
Uc
0 0 Uc
0 0
y y
inhomogeneous SAI inhomogeneous SAI
(c) Reflected Pressure Field (Pa) (d) Reflected Pressure Field (Pa)
1.379 1.174
0.8 Normal Normal
incidence incidence
0.6
pre pro 0
z(m)
0.4 pre
-60 Degree 0 Degree 0
-60 Degree
0.2
-1.386 -0.807
0
-0.4 -0.2 0 0.2 0.4 -0.4 -0.2 0 0.2 0.4
y (m) ( )
y (m)
(e) Reflected Pressure Field (Pa) (f) Reflected Pressure Field (Pa)
1.502 1.467
0.8
0.6
Impedance Impedance
z(m)
0
0.4 discontinuity discontinuity 0
d = 0.0125 d = 0.00886
0.2
-1.242 -0.890
0
-0.4 -0.2 0 0.2 0.4 -0.4 -0.2 0 0.2 0.4
y (m) ( )
y (m)
-0.10 -0.05 0.05 0.10 y -0.10 -0.05 0.05 0.10 y
0.010 0.010
0.015
l 0.015
l
Figure 2.1: (a) For a flat interface with an inhomogeneous SAI, the
angle of pro , i.e., θro , is not influenced, while pre occurs simultaneously
and θre is controlled by IGSL. (b) If SAI is properly controlled, pro is
null. (c) Ultrasound with unit amplitude and ω = 300Krad/s impinges
upon SAI surfaces in water. The SAI along the flat surface generates
both pro and pre when an arbitrary A is chosen in Equation (2.1). (d) √A
particular SAI is chosen according to Equation (2.7). ψ(y) = −100 3y
is selected throughout. (e,f) Simulation results based on impedance
discontinuity with relations between l and y enclosed, corresponding to
the cases (c) and (d) respectively. Figure adopted and reproduced with
permission from ref. [41].
A, implying the unusual double reflections:
2A cos θi − ρw c0
pro ∝ exp[ik0 (y sin θi + z cos θi )]; (2.3)
2A cos θi + ρw c0
∞
pre ∝ eiψ(y) eik0 y(sin θi −sin θre ) dy. (2.4)
−∞
Double reflection consists of one ordinary reflection pro and one extraor-
dinary reflection pre . Intuitively, the direct-current component of the sur-
face impedance contributes to pro while the alternating-current component
contributes to pro . After applying the first-order approximation and the
stationary phase approximation to Equation (2.4), the relation between θre
and the incident angle θi is unveiled:
⎧
⎨ k0 (sin θre − sin θi ) = dψ(y)/dy
. (2.5)
⎩ Z (y, ω) = A{1 − i tan[ψ(y)/2]}
n
Note that the extraordinary reflection can exist only when the inhomoge-
neous SAI along the flat surface is expressed in form of Equation (2.1),
on the basis of our derivation. Although IGSL’s appearance is similar to
GSL [1, 2, 5], its physical meaning of ψ(y) is dramatically different. Fun-
damentally, the variable of our IGSL Equation (2.5) is about the value of
surface acoustic impedance instead of the abrupt propagating phase change.
Moreover, IGSL only serves to steer pre at will, with no influence on the
direction of pro , as illustrated in Figure 2.1(a). In Appendix B we highlight
the irrelevance between GSL and our proposed IGSL. In addition, GSL
mentions the extra accumulated phases along wave-propagation paths, but
it is still relying on graphical methods to find out the relation between the
configuration of the passive antenna array and the needed phase in op-
tics [1]. However, we do not have the passive antenna in acoustics. Here,
IGSL Equation (2.5) and Equation (2.1), serving as an explicit design rule,
provide us the feasible way based on a different mechanism in acoustics.
Equation (2.5) also sheds the light on an extreme angle (similar to critical
angle): ⎧
⎨ arcsin(−1 − 1 dψ(y)
), if dψ(y) <0
k0 dy dy
θe = , (2.6)
⎩ arcsin(+1 − 1 dψ(y)
), if dψ(y) >0
k0 dy dy
above which pre becomes evanescent in the upper space. Equation (2.6)

1
dψ(y)
holds only when −1 ≤ 1 − k0

dy
≤ 1. Otherwise, pre becomes evanes-
cent.
Usually, both pro and pre will coexist as shown in Figure 2.1(a), suggesting
double reflections, while IGSL only controls θre . Hence, it is interesting
to eliminate pro as shown in Figure 2.1(b), by means of a particularly
selected value of A. Equation (2.3) suggests that A = (ρ0 c0 )/(2 cos θi ) can
make pro vanish, i.e., pro is switched off, as shown in Figure 2.1(b). The
corresponding SAI of the flat surface then becomes

ρw c 0 ψ(y)
Zn (y, ω) = 1 − i tan . (2.7)
2 cos θi 2
2.2 Continuous and discontinuous impedance
Supposing the gradient of ψ(y) along the flat interface is constant, we

notice that Equation (2.4) turns out to be a Dirac Delta without any ap-
proximation. From Equation (2.5) we predict that the wavefront of pre will
propagate in the form of a plane acoustic wave, independent of y. We select
water (ρw = 1000kg/m3 ; c0 = 1500m/s [42]) as the background medium,

ω = 300Krad/s as the circular frequency, e−ik0 z as the normal incident
√
plane ultrasound, and a linear form ψ(y) = −100 3y in Equation (2.7).
θre is theoretically found to be −60◦ by IGSL, validated by our simulation

in Figure 2.1(d). pro is thoroughly suppressed thanks to the specific A
chosen according to Equation (2.1). In contrast, in Figure 2.1(c), the same
parameters are kept except for another A, whose value is arbitrarily taken
to be ρw c0 . It clearly shows that pro occurs and interferes with pre , but
pre still keeps the same, verifying our theoretical formulation. In terms
of phenomena, the designed inhomogeneous SAI Equation (2.1) essentially
implies the changes of both the propagating phases and amplitudes, only
by which the effect of double reflections may occur. In terms of physics, the
extra momentum supplied by the metasurface is employed to compensate
the momentum mismatch between the incident acoustic beams and the
diffracted beams. Therefore, for the double backward propagating beams,
pro is the commonplace reflection, while pre is attributed to the diffraction
of higher order.
Figure 2.1(d) suggests the possibility of negative reflection for pre , which
is further verified for oblique incidence in Figure 2.2. In Figure 2.2(a),
because of the inhomogeneous SAI and the arbitrary A in Equation (2.1),
both pro and pre occur. Figure 2.2(b) depicts the same situation except for
pro being switched off as a result of the specifically chosen A according to
Equation (2.7), while the red line pre stays the same as that in Figure 2.2(a).
The blue braces represent the region of negative pre . It is noteworthy
that pre does not exist if θi is beyond the extreme angle θe = −30◦ in
Equation (2.6), corresponding to the purple dots.
1
a
SinT r
n
io
ect
efl
i
T
0.5 r n
in
ar
y io
=S
t
n ec
ro
di fl
nT
or re
re
nT
Si
y
ar
Si
n
di
r0.5
-1 -0.5 o SinT i 1
tra
ex e y
ativ nar
g i n
-0.5 ne aord tio
tr ec
ex refl
SinT e
-1
1
b
SinT r
0.5 n
c tio
efle
re
yr
nT
r
Si
ina
-1 -0.5 ord0.5 SinT i 1
1 d\ ( y ) tra
ex e y
k0 dy
ativ nar
g i n
-0.5 ne aord tio
tr ec
ex refl
SinT e
-1
Figure 2.2: sin θro,re versus sin θi when k0 = 10rad/m and ψ(y) = −5y.
pro and pre emerge simultaneously in (a). In (b), only pre occurs for
the same parameters of (a) except A. The purple dot denotes sin θe in
Equation (2.6). Figure adopted and reproduced with permission from
ref. [41].
Reflected Pressure Field (Pa)

0.495
8
prb
30 Degree
6
z(m)
0
4
2 -60 Degree
incidence
0 -0.503
-4 -2 0 2 4
y (m)
√
Figure 2.3: The SAI Equation (2.1) with ψ(y) = (10 + 10 3)y is set
along the flat surface z = 0. In the upper space, the medium is water
(ρw = 1000kg/m3 ; c0 = 1500m/s). An audible plane wave with unit
amplitude and ω = 30Krad/s is −60◦ obliquely incident. Only reflected
acoustic pressure is plotted. The propagating path of pre is noted as
an arrow with purple crossbars. Figure adopted and reproduced with
The field simulation for oblique incidence is shown in Figure 2.3. For this
simulation we assume water (ρw = 1kg/m3 ; c0 = 1500m/s [42]) as the
background medium in the upper space. The SAI with the linear parameter
√
ψ(y) = (10 + 10 3)y is set along the flat metasurface, and an audible
(ω = 30Krad/s) plane wave with a unit amplitude is obliquely incident
with the incident angle −60◦ . These parameters theoretically lead to the
angle of pre 30◦ according to our proposed IGSL. Furthermore, pro vanishes
thanks to the specific A chosen in Equation. (2.1). In Figure 2.3, we find
out the simulation confirms the prediction via IGSL accurately, and pro
disappears as expected.
Moreover, the incident audible plane wave and pre are at the same side of
the normal line, confirming the possibility of the negative extraordinary
reflection. The singularity due to tan[ψ(y)/2] in the imaginary part of
Equation (2.1) does not play a significant role because the mathematical
singularity ±i∞ just occurs at singular positions, physically meaning the
total reflection (reflection coefficient equals +1).
flat interface y
l(y)
d
side view
hard wall
thin film
water y
air
top view
Figure 2.4: Realization schematics by hard-sidewall tubes of designed

lengths. Figure adopted and reproduced with permission from ref. [41].
As depicted in Figure 2.4, we propose one plausible realization schematic

for the general SAI of Equation (2.1), where all hard-sidewall tubes with
one pressure-release termination are gathered and juxtaposed perpendicular
to the flat interface. Observed at the top view, each tube has a square
cross section whose width is d, with four enclosed hard sidewalls (black).
Then observed at the side view, the upside open termination of each tube
constitutes an effective SAI pixel of the interface, while the other end sealed
by a thin film (orange) serves as the pressure-release termination. [42] The
upper space and the interior of each tube are filled with water, without
separation. The light blue indicates air downside, which is isolated from
water by the thin film.
The SAI of each tube at the opening facing the upper space is [42]:
ρw c 0 k 0 2 d 2
Zt (y, ω) = − iρw c0 tan [k0 l(y) + k0 Δl] , (2.8)
2π
√
where l(y) is the length of each tube and Δl ≈ 0.6133d/ π is the effective
end correction. By comparison of Equation (2.1) and Equation (2.8), it is
required that A = ρw c0 k0 2 d2 /(2π) and A tan[ψ(y)/2] = ρw c0 tan [k0 l(y) + k0 Δl],
leading to the value of the spacing d for impedance discretized spacing and
the dependence between l(y) and ψ(y):
⎧
⎨ d = (2πA)/(ρw c0 k0 2 )
, (2.9)
⎩ l(y) = 1 arctan[ k0 2 d2 tan ψ(y) ] + nπ
− Δl
k0 2π 2 k0
where the arbitrary integer n is required to be set suitably to make l a

positive value. Thus, the change of ψ along y, representing the control of
pre , is interpreted as the change of l, implying one straightforward realiza-
tion based on discontinuous impedance. Thus, the inhomogeneity of the
acoustic impedance is strictly paraphrased into the inhomogeneity of the

tube-array structure, resulting in our acoustic metasurface. At the side
view in Figure 2.4, the solid red contour indicates one arbitrary function
of l(y) calculated from Equation (2.9). Based on the discretized spacing
d calculated from Equation (2.9) as well, we are able to find d and the
corresponding height l(y), marked with the yellow dots. The width of
tubes, d, is required to be in subwavelength, which means that the criterion
A < 2πρw c0 needs to be satisfied. The parameters d used in our simulations
satisfy this criterion after compared with wavelengths. Also, note that the
top of the tube array is aligned into a flat surface (red dashed line), above
which acoustic waves impinge. Thus, the change of tube lengths will not
affect the flatness of the surface. In addition, thanks to the property of the
arc-tangent in Equation (2.9), the tube-array metasurface is within a thin
layer without the space-coiling-up technique [43].
It is also noteworthy that because of the intrinsic differences between optics

and acoustics, so far we cannot obtain the mechanism-analog of the optical
metasurface, which is based on resonances and independent with the thick-
ness or effective propagating lengths, but we can achieve the phenomenon-
analog in acoustics using the tube array. In principle, because tubes can be
regarded as Helmholtz resonators, complex SAI at each pixel can be real-
ized by a suitable arrangement of resonators, as the analog of the complex
electric impedance realized by the combination of resistance, capacitance
and inductance. In addition, we know that only the real part, the electric
resistance, consumes energy while the imaginary part does not. In the same
manner in acoustics, the energy loss is theoretically only attributed to the
real part of the surface complex SAI in Equation (2.8), i.e., the loss in our
case is caused by the energy consumption from the tube array.
Using this method, we reproduce Figure 2.1(c,d) by realistic impedance dis-

continuity, so as to verify our proposed realization. In Figure 2.1(e,f), d =
0.0125 and 0.00886 are selected respectively according to Equation (2.9),
and the corresponding contours of the tube length l in terms of the lo-
cation y are illustrated as the red lines, respectively. Figure 2.1(e) shows
strong interference between pre and pro while Figure 2.1(f) shows the nearly
undisturbed pre , coinciding with Figure 2.1(c,d), respectively.
2.3 Acoustic illusion and ipsilateral focusing
To demonstrate IGSL’s capability of designing novel acoustic devices, we

metamorphose acoustic pressure fields everywhere through SAI manipu-
lation as simulated in Figure 2.5. This deceptive effect is obtained by
manipulating plane wavefronts into the wavefronts generated by a virtual
reflector or focusing illumination, governed by the control of pre , i.e., IGSL.
Under these scenarios, we need to consider nonlinear forms of ψ(y). New
phenomena are thus expected when θre becomes spatially varying.
It is found that the acoustic deception can be created via IGSL, e.g.,
ψ(y) = 0.7y 2 in Equation (2.7), resulting in pro = 0. Correspondingly, θre
in Figure 2.5(a) is a position-dependent function sin θre = 0.14y, in which
case pre fans out as demonstrated in Figure 2.5(a), verifying our theory.
Here the discretized spacing d for impedance is 0.1772 and the relations
between l and y derived from Equation (2.9) are enclosed in Figure 2.5.
Therefore, IGSL can be employed to camouflage a flat surface as if there
were a curved lens at the origin instead of the physical planar interface.
(a) Reflected Pressure Field (Pa)

-0.966 0 1.572
6
Normal incidence
5
4
z(m)
3 pre
2
0
-4 -2 0 2 4
y (m)
-6 -4 -2 0 2 4 6 y
0.2
0.3
0.4 l Impedance
discontinuity
(b) Reflected Pressure Field (Pa) d = 0.1772
-2.684 0 2.563
Normal incidence
Focal point
pre
-4 -2 0 2 4
y (m)
-6 -4 -2 0 2 4 6 y
0.2
0.3
0.4 l
Figure 2.5: Wavefront metamorphosis via SAI interface, with
impedance discontinuity d = 0.1772. A plane acoustic wave of ω =
15Krad/s is normally incident in water. Only reflected acoustic pres-
sure is plotted. (a) The SAI of Equation (2.7) with ψ(y) = 0.7y 2 is set
along the flat surface. pre diverges into
a curved wavefront.
(b) The SAI
of Equation (2.7) with ψ(y) = −10 2 2
y + 4 − 4 is set. pre converges
to a focal point in the 2D case. Figure adopted and reproduced with
The dual effect allowing a curved reflector to mimic a flat mirror, by ma-
nipulating the convex wavefronts into planar wavefronts, was reported in
plasmonic regime [44].
Furthermore, the SAI can be designed to make acoustic waves reflected

by a planar interface focused as well. In optics, a flat lens with metallic
nanoantennas of varying sizes and shapes can consequently converge the
transmitted light to a focal point [6, 7]. Note that the optical focusing
controlled by optical GSL is on the other side of incoming lights, i.e., on
two sides of the flat surface [6, 7] in the transmission mode. In acoustics,
we employ an inhomogeneous SAI flat surface to focus pre , in the reflection
mode by IGSL without pro .
This ipsilateral focusing in Figure 2.5(b), is thus obtained in the planar ge-
ometry in acoustics for the first time. In Equation (2.7), a hyperbolic form

is set: ψ(y) = −k0 y 2 + f 2 − f (f being the given focal length for the
SAI of the flat interface. pre from different angles constructively interferes
at the ipsilateral focal point, as if the waves emerge from a parabolic sur-
face. The parameters in Figure 2.5(b) are the same as those in Figure 2.5(a)

except for the specific hyperbolic SAI form ψ(y) = −10 y 2 + 42 − 4 ,
with the designed focal point at (y = 0, z = 4) and pro suppressed. In
addition, the simulated acoustic pressure by discretized impedance at the
focal point is well confined at (y = 0, z = 4).
Interestingly, the imaging at the same side was previously presented for
electromagnetic waves [45, 46]. In [45], it demands strong chiral materials
filled in the whole upper space. The same-side imaging is only a partial
imaging, i.e., only one circularly polarized wave being imaged and the other
being reflected ordinarily. In acoustics, our ipsilateral imaging is achieved
by translating all the stringent requirements of the half-space chiral ma-

terials into an inhomogeneous impedance surface. In electromagnetism,
ipsilateral imaging can be achieved as well by surface gratings [46] or an-
tenna arrays. However, the polarization of incident electromagnetic waves
is always closely related to the effect of focusing. Therefore, the ipsilateral
imaging in acoustics by IGSL has no polarization constraints thanks to the
acoustic wave nature, i.e., longitudinal vibration.
(a) Reflected pressure field (Pa)

6
1.425
5
Normal incidence (1Pa)
4
z (m)
3
pre pre 0
2 Surface acoustic wave Surface acoustic wave
1
-0.849
0
-8 -6 -4 -2 0 2 4 6 8
y (m)
(b) Reflected sound pressure level (dB) (c) Reflected sound pressure level (dB)
6 70dB (0.09Pa) 6 70dB (0.09Pa)
5 29.39dB 94.575dB (1.43Pa) 5 67.332dB 85.801dB (0.55Pa)
4
z (m)
4
z (m)
3 3
2 2
1 1
0 0
-8 -6 -4 -2 0 2 4 6 8 -8 -6 -4 -2 0 2 4 6 8
y (m) y (m)
Figure 2.6: Conversion from PAWs to SAWs via SAI interface. The
PAW with unit amplitude and ω = 15Krad/s is normally incident in
water. Only reflected acoustic pressure is plotted. (a) The SAI of Equa-
tion (2.7) is set to be ψ(y) = −11y for y < 0 and ψ(y) = 11y for y > 0.
SAWs are bifurcated at the origin and confined near the surface. (b)
The reflected sound pressure level of (a). (c) The reflected sound pres-
sure level when a homogeneous SAI is adopted instead. Figure adopted
and reproduced with permission from ref. [41].
2.4 Conversion from propagating to surface
waves
Beyond the acoustic-field metamorphosis, we further establish a kind of

acoustic cognitive deception about an SAI surface converting propagating
acoustic waves (PAW) to surface acoustic waves (SAW) in Figure 2.6, by
means of IGSL. The extreme angle 0◦ in Equation (2.6) demands ψ(y) =
±10y. Therefore, we set the SAI of Equation (2.7) slightly over that ex-
treme, e.g., ψ(y) = −11y for y < 0 and ψ(y) = 11y for y > 0 are set along
the flat interface symmetrically with respect to the z. In Figure 2.6(a),
the bidirectional surface acoustic waves are attributed to the coupling ef-
fect governed by the diffracted evanescent pre which propagates along the
metasurface [47]. Owing to the inhomogeneous SAI interface, the ideally
perfect conversion comes true in acoustics except for a little diffraction.
Physically, the SAI along the flat surface provides an extra momentum
to compensate the momentum mismatch between propagating waves and
surface waves in acoustics, resulting in the high efficiency conversion. In
contrast, if one uses a constant SAI Equation (2.7) with ψ(y) = 11 along
the flat surface (the homogeneous SAI does not generate pre ; only pro oc-
curs), the reflected sound pressure level in Figure 2.6(c) is almost uniformly
spread over the space.
Figure 2.6(b) clearly demonstrates that the acoustic field is well confined in
the region close to the interface and attenuated quickly to around 0P a away
from the interface, revealing the nearly perfect conversion. Interestingly,
it shows in [18] that the electromagnetic-varying metasurface is able to
prevent the propagating electromagnetic waves from being reflected back to
the upper space. Hence, our PAW-SAW conversion in acoustics, originating

from a distinguished mechanism, is differentiated from [18].
In Figure 2.6, we notice that such technology is functional as an alternative

invisible acoustic cloak by trapping the acoustic field in the vicinity of the
coating, resulting in much lower signal levels of reflection. It may pave the
avenue to the large size acoustic invisibility since it is only dependent on the
surface technique instead of wave-interaction based metamaterial acoustic
cloaking [37]. It will also be promising to consider the time-varying surface
technique in acoustics with nonreciprocal diffraction [48] in the future.
2.5 Methods
For theoretical derivations, we used Green’s function, the integral equation

Equation (2.2) and Born approximation. The detailed theoretical develop-
ment is elaborated in Appendix A. For the numerical calculations, we used
the Finite Element Method by means of the commercial software COM-
SOL Multiphysics. The left, right and top sides of the meshed domain are
set as plane wave radiation conditions, while the bottom side is set as the
impedance boundary with a certain value.
2.6 Conclusion and discussion
We have proposed the acoustic metasurface of inhomogeneous acoustic

impedance. Impedance-governed generalized Snell’s law of reflection has
been established for manipulation of acoustic wavefronts. Due to the lack of
abrupt-phase-changing surface structures in acoustics, we resort to acoustic

impedance as the variable to tweak sound reflection. Impedance-governed
generalized Snell’s law of reflection, which can simultaneously generate the
switchable pro and the steerable pre , provides us the explicit connection
between our designed specific acoustic impedance and the reflected field,
serving as the design rule in acoustics. We not only demonstrate intrigu-
ing acoustic manipulations but also provide insightful realization schemes
of the metasurface. As a few examples, we demonstrate acoustic disguise,
acoustic planar lens, acoustic ipsilateral imaging, and the conversion from
propagating to surface acoustic waves.
Ultra-thin acoustic metasurfaces can also be constructed by the method of

coiling up space, so that sound trajectories can be altered by changing wave
propagating paths [49, 50]. This method provides a more accessible way to
steer acoustic wavefronts than our proposed approach, but the trade-off is
the lack of capability to generate rich phenomena such as double reflections
demonstrated in this thesis.
Chapter 3
Redirecting acoustic waves out

of the incident plane
This chapter addresses a flat metasurface to manipulate the extraordi-

nary out-of-incident-plane reflection and vibration in acoustics, validated
by the theoretical modeling and the numerical experiment. We theoreti-
cally demonstrate that in fluids, extraordinarily reflected sound waves can
be achieved along a three-dimensional spatial angle out of the incident plane
by manipulating the impedance distribution of a flat metasurface reflector.
In particular, the arbitrary manipulation can be unanimously predicted and
concluded by our three-dimensional impedance-governed generalized Snells
law of reflection (3D IGSL), which is rigorously derived from Greens func-
tions and integral equations. Consequently, the vibrations of the extraordi-
nary reflection and the incidence will form a spatial angle in between, rather
than sitting in one plane. Such an inhomogeneous flat metasurface can be
effectuated by means of impedance discontinuity, and further implemented
30
Chapter 3. Redirecting acoustic waves 31
by tube arrays with properly designed lengths. Finite-element-simulation

results agree with the theoretical prediction by 3D IGSL.
3.1 Out-of-incident-plane vibration of fluid
particles
The coordinate is illustrated in Figure 3.1, where the flat metasurface re-
flector is placed at z = 0 plane, i.e., x-y plane. In water (speed of sound
c0 = 1500m/s; density ρw = 1000kg/m3 ), an acoustic plane wave pi from
the space z > 0 is impinged upon the flat surface z = 0 with unit amplitude
and the frequency ω = 3 × 105 rad/s. Figure 3.1(a-d) are the simulated
acoustic fields in the upper space z > 0, which are the projections upon the
plane perpendicular to z axis. For the incident field in Figure 3.1(a), one
can notice that the vibrational direction of fluid particles (orange double-
headed arrow) excited by the incidence forms the incident plane (yellow
dashed line) with z axis. As shown in Figure 3.1(b) for the reflected field,
if the impedance reflector is homogeneous, the particle vibration excited by
the ordinary reflection pro (orange double-headed arrow) will be co-planar
with the incident vibration, as expected intuitively.
In order to steer the acoustic vibrations freely, a metasurface reflector com-

posed of the inhomogeneous specific acoustic impedance SAI, which can
be realized by different layouts of tube resonators with designed lengths, is
implemented in Figure 3.1(c,d), while the same incidence in Figure 3.1(a)
is used. The incident plane (yellow dashed line) is identical through-
out all cases in Figs. 1(a-d). It can be seen from the reflected fields in
Figure 3.1(c,d) that the particle vibration excited by the reflection (blue
(a) Incidence Field (b) Reflection Field

-1.00 1.00 -0.39 0.38
field unit: Pa
tube length unit: m
0.1 m
0.1 m
n
ce
y flat reflector
io
en
ct
cid
fle
at z=0
re
in
z x
incident plane (c) Reflection Field (d) Reflection Field

-0.71 0.59 -0.64 0.64
0.1 m
0.1 m
in-
reflection
re
incident-plane
fle
vibration
ct
io
n
out-of-
incident-plane (e) Tube Length (f ) Tube Length
0.0035 0.0192 0.0035 0.0192
vibration
(cross vibration)
0.1 m
0.1 m
l(x,y) l(x,y)
out-of-
incident-plane
vibration
Figure 3.1: (a) Observing along −z, an plane wave is propagating to-
ward the metasurface at z = 0. The vibration of fluid particles excited
by the incidence (orange double-headed arrow) is within the incidence
plane (yellow dashed line). (b) The ordinary reflection generated by
a homogeneous flat reflector excites the in-incident-plane particle vi-
bration. (c) Observing along −z, the flat metasurface reflector excites
the out-of-incident-plane cross vibration of fluid particles (blue double-
headed arrow). (d) Another metasurface reflector excites the extraor-
dinary vibration of fluid particles (green double-headed arrow). (e,f)
The realization schematics of the metasurface, and the tube lengths cor-
responding to (c) and (d) respectively are exhibited in (e,f). Figure
adopted and reproduced with permission from ref. [51].
double-headed arrow) deviates away from the incident plane by employing

the inhomogeneous impedance surface. Observing along −z, we manipulate
the x-y plane projection of the vibration (excited by reflection) perpendicu-
lar to the incident plane as shown in Figure 3.1(c), as named cross vibration.
Another example is shown in Figure 3.1(d), where the out-of-incident-plane
vibrational orientation (green double-headed arrow) is steered robustly by
the flat metasurface at z = 0. The corresponding reflected acoustic field at
z > 0 projected in the x-y plane is shown in Figure 3.1(d), verifying the
robust and precise manipulation of the out-of-incident-plane vibrational
orientations of fluid particles.
In order to provide a theoretical and systematic framework for precisely

manipulating the vibrational orientation in fluids, we thereby establish 3D
IGSL. Here we consider the reflection by a flat acoustic metasurface at
z = 0, and formulate the modified Snells law in acoustics for inhomo-
geneous two-dimensional SAI [42]. More specifically, the inhomogeneous
SAI will give rise to the out-of-incident-plane vibration excited by the ex-
traordinary reflection pre (uniquely controlled by 3D IGSL) as well as the
in-incident-plane vibration excited by an ordinary reflection pro , as shown
in Figure 3.2(a).
Interestingly, it is found that the acoustic metasurface at z = 0 designed

according to 3D IGSL cannot alter the direction of pro excited by the meta-
surface reflector, but surprisingly can turn off pro as shown in Figure 3.2(b).
In other words, our design can create the steerable pre as well as the switch-
able pro . This is unique for our acoustic metasurface while it is generally
difficult to eliminate the parasitic ordinary refraction or reflection for elec-
tromagnetic metasurfaces [1]. pre can be in principle steered along arbitrary
directions by the acoustic metasurface, simultaneously with pro eliminated,

resulting in the corresponding manipulation of cross vibration in the ab-
sence of the ordinary in-plane vibration. Therefore, 3D IGSL describes the
generalized reflection law regarding the acoustic metasurface of 2D inhomo-
geneous acoustic impedance, and provides a clear-cut way for manipulating
pre and its vibration along arbitrary spatial angles.
Physically, our realization of our-of-plane vibration in fluids is based on the

reference of the incident plane. Because the reflection plane is rotated by
our design method, the vibration of acoustic reflected waves is out of the
incident plane, but still within the reflection plane. This pseudo transverse
polarization in fluids is different from the authentic transverse vibration for
acoustic waves in solids or electromagnetic waves.
(a) (b)
incidence
(c) x y flat interface
z
water
air
y hard wall
l(x,y)
d
thin film a cross-sectional slice perpendicular to x
top view
Figure 3.2: (a) For a flat metasurface reflector with an inhomoge-

neous 2D SAI, the directions of pro , i.e., θro and φro , are not influenced,
while pre occurs simultaneously with the direction θre and φre controlled
by 3D IGSL. (b) If a heterogeneous SAI is properly designed upon the
reflector, pro will become null. (c) Realization schematics by tube ar-
rays, comprising the reflector (yellow dashed line). Figure adopted and
reproduced with permission from ref. [51].
3.2 Two-dimensional varying metasurface
Here, we will focus on the theoretical formulation. As depicted in Fig-

ure 3.2(a,b), pi , θi (the angle between the orange line and z) and φi (the
angle between x and the x-y-plane projection of the orange vector) stand
for the incident plane wave, the incident polar angle and the azimuth angle,
respectively. The similar notations are adopted for pro and pre .
The inhomogeneous 2D SAI Zn of the flat metasurface at z = 0 is the

extension of the one-dimensional SAI that only creates the in-incident-plane
vibration and redirection in acoustics, which was introduced in Chapter 2
[41]. For simplicity in modeling, we consider

ψ(x, y)
Zn (x, y, ω) = A 1 − i tan , (3.1)
2
where A is an arbitrary real constant and ψ(x, y) represents the spatially

varying component only existing at the imaginary part. Note that ψ(x, y)
in Equation (3.1) has already taken into account the circular frequency.
We assume pi in the upper space satisfies
pi (x, y, z, ω) = pi0 (ω) exp [ik0 (x sin θi cos φi + y sin θi sin φi − z cos θi )] , (3.2)
where k0 stands for the wave number in free space and pi0 for the amplitude
of the incidence. It is found that pro excited by the reflector at z = 0 with
Zn satisfies [41]
2A cos θi − ρw c0
pro (x, y, z, ω) = pi0 (ω) ×
2A cos θi + ρw c0
× exp [ik0 (x sin θro cos φro + y sin θro sin φro + z cos θro )] ,
(3.3)
where ρw , c0 are the density and the speed of sound in the upper space
z > 0 in Figure 3.2(a-b), θro = θi and φro = φi . pro is attributed to the
reflection by the properly-averaged value of the inhomogeneous 2D SAI
Equation (3.1), while the variance of the 2D SAI is the cause of pre [41].
Here, by virtue of Greens functions [41, 52], pre in the upper space, serving
as the result of the 2D SAI variation, can be expressed as an integral
equation:
∞ ∞
ρw c 0
pre (x, y, z, ω) = ik0 × dy0 dx0 eiψ(x0 ,y0 ) [pi (x0 , y0 , 0, ω)
2A
−∞ −∞
+ pro (x0 , y0 , 0, ω) + pre (x0 , y0 , 0, ω)]G(x, y, z, ω; x0 , y0 , 0),
(3.4)
where G stands for the Greens function accommodating the boundary con-
dition. In the far field approximation [53], G can be derived explicitly:
exp(ik0 |r|)
G(r, ω; r0 ) = × exp[−ik0 (x0 sin θre cos φre + y0 sin θre sin φre )]
4π |r|
2A cos θ∗ − ρw c0
× [exp(−ik0 z0 cos θre ) + exp(ik0 z0 cos θre )],
2A cos θ∗ + ρw c0
(3.5)
where r = (x, y, z), r0 = (x0 , y0 , z0 ), and θ∗ , a constant, describes the effec-

tive incident angle with respect to the Greens function Equation (3.5) [54].
Inserting Equation (3.5) into Equation (3.4) and using Born approximation
[55], we are able to determine pre , which includes the following term:
∞ ∞
pre ∝ dy dx × eiψ(x,y) exp[ik0 x(sin θi cos φi − sin θre cos φre ) (3.6)
−∞ −∞
+ ik0 y(sin θi sin φi − sin θre sin φre )].
Note that for the trivial case when ψ(x, y) = 0, Equation (3.6) is non-zero
which implies that pre propagates along the same direction as pro . That
is to say, if the flat metasurface is of uniform SAI which only generates
the common reflection, the contribution of Equation (3.6) should also be
taken into account besides Equation (3.3). In addition, we find that Equa-
tion (3.6) is a two-dimensional Dirac Delta when ψ(x, y) is a linear function
with respect to x and y, which imposes the directivity of pre to be:
(3.7)
Ψ(θre , φre ) ∝ δ[k0 x(sin θi cos φi − sin θre cos φre )
+ k0 y(sin θi sin φi − sin θre sin φre ) + ψ(x, y)].
Therefore, the spatial directivity of pre only makes sense when
⎧
⎨ sin θre cos φre − sin θi cos φi = 1 ∂ψ(x,y)
k0 ∂x
, (3.8)
⎩ sin θ sin φ − sin θ sin φ = 1 ∂ψ(x,y)
re re i i k0 ∂y
where ψ is a linear function with respect to x and y. Equation (3.8) unveils

the relation between the incident direction and the direction of pre , i.e., 3D
IGSL, which is regarded as the generalized law for acoustic metasurface
reflection. We note that if the metasurface is thin and allows transmission,
Equation (3.8) is the generalized law of refraction for the metasurface as
well, revealing the generality of our approach. It is noteworthy that if ψ
is a constant for a uniform 2D SAI, Equation (3.8) will be reduced to the

usual Snells law. 3D IGSL serves the manipulation of the vibration of fluid
particles excited by pre , theoretically via tuning the parameter ψ of the
inhomogeneous SAI flat reflector, with no influence on the direction of pro ,
as illustrated in Figure 3.2(a).
Other advantage of our Green’s function formulation also gives pro ampli-
tude as [41]
rro = pi0 (ω) × (2A cos θi − ρw c0 )/(2A cos θi + ρw c0 ). (3.9)
Usually, both pro and pre coexist, but 3D IGSL only tunes θre and φre .
In order to obtain purely cross vibration excited by pre with full control,
we need to eliminate pro . By particularly controlling the value of A in
Equation (3.1), we manage to switch off pro , as illustrated in Figure 3.2(b).
Based on Equation (3.9), if A = (ρw c0 )/(2 cos θi ), pro will be eliminated,
while the 2D SAI becomes

ρw c 0 ψ(x, y)
Zn (x, y, ω) = 1 − i tan . (3.10)
2 cos θi 2
Thus, it is discovered that metasurface cannot affect the direction of pro

but just keep or eliminate pro .
Equally important is the plausible realization schematic of the inhomoge-

neous SAI in Equation (3.1), implemented by discretized impedance. As
depicted in Figure 3.2(c), all hard-sidewall tubes are assembled and juxta-
posed perpendicular to the flat interface, illustrated in the top view. Each
tube, serving as one discrete 2D SAI pixel of the flat metasurface reflector,
has a square cross section whose width is d, with four surrounding hard
sidewalls (black). In the view of a cross-sectional slice in Figure 3.2(c),
one end of each tube constitutes the flat interface of the metasurface at
z = 0 (yellow dashed line), and the other end contacts with air (light blue).
The space z > 0 and the interior are filled with water (dark blue), without
separation. The water-air interface separated by a thin film (orange) is
regarded as the pressure-release termination of each tube.
In order to realize the 2D SAI inhomogeneity by discretized impedance,

d < 2π/k0 is required to eliminate higher diffraction orders. The SAI of
each tube at the opening facing z > 0 can be calculated [42]:
ρw c 0 k 0 2 d 2
Zt (x, y, ω) ≈ − iρw c0 tan [k0 l(x, y) + k0 Δl] , (3.11)
2π
where l(x, y) is the spatial distribution of the length of each tube and
√
Δl ≈ 0.6133d/ π is the effective end correction. By comparison between
Equation (3.1) and Equation (3.11), it is required that A = ρw c0 k0 2 d2 /(2π)
and A tan[ψ(x, y)/2] = ρw c0 tan [k0 l(x, y) + k0 Δl], leading to the value of
the discretized spacing d and the dependence between l(x, y) and ψ(x, y):

d= (2πA)/(ρw c0 k0 2 )
2 2
(3.12)
1 ψ(x,y)
l(x, y) = k0
arctan[ k02πd tan 2
] + nπ
k0
− Δl
where the arbitrary integer n is required to be set suitably to make l a

positive value. Thus, the change of ψ at the flat-metasurface reflector, rep-
resenting the manipulation of fluid-particle vibrations through the control
of pre , is now interpreted by the change of l [red dashed line in Figure 3.2(c)],
demonstrating one straightforward realization scheme based on impedance
discontinuity. In principle, tubes can be regarded as Helmholtz resonators,

and the complex SAI at each pixel can thus be realized by a suitable ar-
rangement of resonators.
3.3 Three-dimensional control of extraordi-
nary reflection
The simulation results verify the robustness in the manipulation of fluid-

particle vibrations according to our 3D IGSL. We first consider the ideal
case without using the tube-array configuration, by selecting water as the
√
background media and ψ(x, y) = −100 3y at the SAI metasurface in Equa-
tion (3.10). The incident plane ultrasound with ω = 3 × 105 rad/s, θi = 18◦
and ϕi = 180◦ is impinged upon the metasurface at z = 0. The spatial
angles for pre , i.e., θre and ϕre , are theoretically found to be 66.9◦ and
250.4◦ by 3D IGSL in Equation (3.8), respectively. The simulation in Fig-
ure 3.3(b) validates our theory, where pro disappears thoroughly owing to
the specific design of the coefficient according to Equation (3.10). The cut
slice at ϕre = 250.4◦ in Figure 3.3(b) clearly shows that pre is propagating
towards the predicted direction without any disturbance. In other words,
we realize this out-of-incident-plane reflection, and simultaneously achieve
the full manipulation of its fluid-particle vibration.
In particular, in Figure 3.3(a), the same parameters are kept except for
another selection for A, whose value is arbitrarily taken to be ρw c0 . It
clearly shows that pro coexists and interferes with pre , but pre still keeps
the same direction (θre = 66.9◦ and ϕre = 250.4◦ ), verifying our theoretical
prediction. Although such double reflections are predictable well by our
(a) Reflected Pressure Field (Pa) (b) Reflected Pressure Field (Pa)
y o y o
=18 o x p = 1 8o x
Ui 180 1.175
re
Ui 180 0.970
Ji =Pi Ji = P
i
x 0.2 0.2
Ure =66.9o
p J =250.4o
Pro re Ure =66.9o 0
re p
re 0
J =250.4o
re
int
er
0.1
0.1 0 0.1
0.
0 1 0
fer
en
0.1
01
0. 0
0.1
0.1
yzx yzx
0 -1.146 0 -0.885
ce
0 0
-0.1
0.1 unit: m -0.1
0.1 -0 unit: m
-0.1
-0 1 0.1
-0.1
(c) Reflected Pressure Field (Pa) (d) Reflected Pressure Field (Pa)
y y
p
re
1.178 1.064
0.2 0.2
Ure =66.9o Ure =66.9o
J =250.4o Impedance J =250.4o Impedance
re re
discontinuity discontinuity
0
d = 0.01253 d = 0.00909 0
int
er
0.1
0.1 0 0.1
0.1 0
fer
0.1
0.1
en
0.1
0.
0.11
yzx y z x0
0 -1.209 -0.962
ce
0 0
-0.1
0.1 unit: m -0.1
0.1 -0 1
-0.1 unit: m
-0 1
-0.1
-0.10 -0.05 z 0.05 0.10 y -0.10 -0.05 z 0.05 0.10 y

0.010 0.010
0.015
l 0.015
l
Figure 3.3: (a) The acoustically flat-metasurface reflector with an

inhomogeneous SAI excites both pro and pre when an arbitrary A is
chosen in Equation (3.1). The fluctuation of the interference verifies our
theory. (b) A particular SAI is suggested according to Equation (3.10) so
that only the pure out-of-incident-plane vibration is excited by pre with
an expected direction. (c,d) Simulation results based on the realization
of tube arrays with relations l(x, y) enclosed below, corresponding to
the cases (a) and (b) respectively. Figure adopted and reproduced with
theory, pro will disturb the manipulated out-of-incident-plane fluid-particle

vibration excited by pre in Figure 3.3(a). Therefore, we generally switch
off pro and make the out-of-incident-plane vibration excited by pre pure, as
demonstrated in Figure 3.3(b).
Next, we consider the realization when the metasurface reflector with realis-
tic discretized impedance is applied, in order to reproduce Figure 3.3(a,b).
In the reflected-field simulation of Figure 3.3(c,d), d = 0.01253 and d =
0.00909 are selected respectively according to Equation (3.12), and their
corresponding distributions of l(x, y) are enclosed. (In these two cases
there is no variation of l along x.) Figure 3.3(c) shows strong interference
between pre and pro , while Figure 3.3(d) shows the nearly undisturbed pre ,
coinciding with Figures 3.3(a) and 3(b), respectively, verifying our proposed
realization using the layout of tube arrays.
Recalling the given example in Figure 3.1, we set the oblique incident
angles as θi = 60◦ and ϕi = 225◦ . The flat acoustic metasurface with
√
ψ(x, y) = 100 6x in Equation (3.10) is placed as the reflector at z = 0,
whose tube-length distribution l(x, y) is illustrated in Figure 3.1(e) accord-
ing to Equation (3.12). Through 3D IGSL in Equation (3.8), we manage
to make pre arise with the direction θre = 60◦ and ϕre = −45◦ , and si-
multaneously make pro eliminated, corresponding to the simulation of the
reflected field in Figure 3.1(c). The perpendicular intersection between the
incident plane and the x-y-plane projection of the particle vibrations in
Figure 3.1(c) exhibits the so-called cross vibration of fluid particles excited
by reflection, leading to this intriguing tweak of vibrational orientations in
fluids. Figure 3.1(d) is another example to verify the robustness of our the-
√ √ √
ory. The flat metasurface reflector with ψ(x, y) = 50 6x−100 3y+50 6y
in Equation (3.10) is selected and the corresponding l(x, y) is illustrated in

Figure 3.1(f). Similarly, by the prediction from 3D IGSL, pre occurs to the
direction θre = 60◦ and ϕre = 270◦ with the suppression of pro , whose field
projection at the plane perpendicular to z is Figure 3.1(d).
3.4 Conclusion
In conclusion, we propose an acoustic flat metasurface reflector to manip-

ulate vibrational orientations of fluid particles in acoustics, and show that
a complete conversion between two perpendicular vibrations by deviating
the extraordinary reflection out of the incident plane. It is found that the
control of the metasurface parameter can keep the extraordinary reflection
only, while suppressing the ordinary reflection. We also theoretically unveil
the generalized rule of three-dimensional impedance-governed generalized
Snell’s law of reflection with respect to the specific acoustic impedance.
The out-of-incident-plane fluid-particle vibration and the arbitrary degree
of freedom in directional manipulation are numerically implemented using
the designed layout of tube arrays.
Chapter 4
Manipulating acoustic focus

with an active metasurface
piezoelectric transducer
This chapter establishes the prototype of acoustic metasurface piezoelectric

transducer (PT), whose piezoelectric elements are squeezed into a flat thin
layer compared to the scale of the entire device. The active planar inter-
face also extends the knowledge of acoustic metasurface engineering for the
deflection of sound beams using passive elements [49, 51]. Through the
optimized ring configurations of the active metasurface PT, we are able to
manipulate the focal pattern and the focal resolution in acoustic far fields.
Firstly, we design the far-field finite-length focal needle with the manipu-
lated distance and depth. Its focal resolution is subwavelength for the full
width at half maximum (FWHM), and it propagates without divergence
for a distance of 5.9λ as designed, longer than the depth 4λ of the re-
ported optical needle [56]. These two designed focusing properties created
45
Chapter 4. Manipulating acoustic focus 46
with PTs were never achieved in acoustics, to the best of our knowledge.
To further verify the robustness of our manipulation of the focal pattern,
via another optimized ring configuration, we obtain the designed far-field
multiple foci, whose FWHM ( 0.45λ) beats the Rayleigh diffraction limit
of conventional acoustic instruments (0.5λ). Besides, to demonstrate the
manipulation of the focal resolution, we design the extreme case of the
super-oscillatory super resolution, whose size is 0.3λ in acoustic far fields,
much smaller than the diffraction limit.
4.1 Design of planar metasurface piezoelec-
tric transducer
In the 3D view of the configurable planar metasurface PT prototype in Fig-

ure 4.1(a), piezoelectric rings (red) are unevenly spaced with hard-boundary
mask rings (blue) in between. A type of common artificial ceramic is em-
ployed as the piezoelectric material: lead zirconate titanate PZT-5H [57].
In the radial view, the thickness q is set identical for all PZT-5H rings, and
the ring configuration (r1, rr1, r2, rr2, · · ·) will be optimized according
to different focusing manipulation. The thin hard-boundary mask rings in
between, through which no sound can pass, are the complements of the
spaced gaps between PZT-5H rings, co-planar with z = 0. The entire PT
is axisymmetric with respect to +z, and the upper surface of the structure
at z = 0 can be regarded as a flat active metasurface according to the
radial cross-sectional view. In our following demonstrations in air (density:
ρa = 1.21kg/m3 ; speed of sound: c0 = 343m/s), we will show the designed
focal pattern and the focal resolution created with the PT prototype in
acoustic far fields, the simulation of which is carried out by the finite el-
ement method (FEM). In detail, COMSOL Multiphysics is the platform
we use, and the simulation is facilitated by the coupling of the embedded
acoustic module and the acoustic-piezoelectric interaction module concur-
rently.
The design method and the related physics are elaborated in this part. The
electromechanical constitutive equations governing the piezoelectric effect
of PZT-5H are written in the stress-charge form [59]:
T = cE S − eT E; D = eS + εS E, (4.1)
where T and S are the vectors of stress and strain; E = −∇φ is the
electric field, which is rephrased with the electric potential φ exerted on
the PZT-5H rings; cE the elasticity matrix evaluated at a constant electric
field; e the coupling matrix; D the electric displacement; εS the electric
permittivity matrix evaluated at a constant mechanical strain. Usually, cE
is straightforwardly given as an anisotropic symmetric matrix; εS = ε0 εSr
includes the relative permittivity matrix εSr . These parameters of PZT-5H
are listed in Appendix C. For each PZT-5H ring whose poling direction
is aligned toward +z, the boundary conditions (B.C.s) are indicated in
Figure 4.1(b). The structural B.C.s for the inner and the outer sides are
free of constraint, while the bottom is fixed with the structural displacement
u = 0. The top undergoes the interaction between sounds and structures
[59]:
(a) ĨĂƌͲĮĞůĚĞīĞĐƚƐ
ŝŶĂŝƌ ŚĂƌĚͲďŽƵŶĚĂƌǇ
ŵĂƐŬƌŝŶŐƐ
WdͲϱ,ƌŝŶŐƐ z
r
z ŚĂƌĚďŽƵŶĚĂƌǇ
r1 rr1 r2 rr2 r3 rr3
Ž r
q
ƌĂĚŝĂůĐƌŽƐƐͲƐĞĐƟŽŶĂůǀŝĞǁ
(b) ĞůĞĐƚƌŝĐĂů͗͘͘ĞůĞĐƚƌŝĐƉŽƚĞŶƟĂůI
ƐƚƌƵĐƚƵƌĂů͗͘͘ĂĐŽƵƐƟĐͲƐƚƌƵĐƚƵƌĞŝŶƚĞƌĂĐƟŽŶ
ĞůĞĐƚƌŝĐĂů͗͘͘ ĞůĞĐƚƌŝĐĂů͗͘͘
ǌĞƌŽĐŚĂƌŐĞ ǌĞƌŽĐŚĂƌŐĞ
ƐƚƌƵĐƚƵƌĂů͗͘͘ ƐƚƌƵĐƚƵƌĂů͗͘͘
ĨƌĞĞ ĞůĞĐƚƌŝĐĂů͗͘͘ŐƌŽƵŶĚ ĨƌĞĞ
ƐƚƌƵĐƚƵƌĂů͗͘͘ĮǆĞĚ
Figure 4.1: (a) Schematics of the configurable planar metasurface PT

prototype. Inside the dashed box is the radial cross-sectional view of
the ring configuration, showing the unevenly-distributed piezoelectric
elements and the hard boundaries. (b) The boundary condition of each
piezoelectric ring observed from the radial cross-sectional view. Figure
T · nS
= −nS p|z=0
z=0

∂2u
nS · − ρ10 ∇p
= nS · ∂t2
(4.2)
z=0 z=0
S = 12 (∇u)T + ∇u
where T and S are the tensors of stress and strain; nS is the outward-
pointing unit normal vector seen from inside PZT-5H. Briefly, the first
equality indicates the acoustic pressure load p; the second makes the normal
acceleration of p on the boundary z = 0 equal to that of the structural
displacement; the third is the intrinsic constitutive equation of S and u
inside solids. Meanwhile, the electrical B.C.s for the inner and the outer
sides are free of charge nS · D = 0, whilst the bottom touches ground
φ = 0. The top is assigned with monochromatic φ = V0 cos(2πf t). One
common technique to impose the voltage bias is a sandwiched structure
with PTH-5H elements between two Au electrode layers.
Next, in order to manipulate the far-field focal pattern or the focal reso-
lution created with the planar metasurface PT prototype, we propose the
acoustic Rayleigh-Sommerfeld diffraction integral (RSI) in conjunction with
the method of binary particle swarm optimization (BPSO) [60–62]. The
derivation starts from the Kirchhoff-Helmholtz integral [42]:

dS [p(r , ω)∇ g(|r − r |) − g(|r − r |)∇ p(r , ω)] · n = p(r, ω), (4.3)
S
where r ∈ Ω, g(x) = exp(ik0 x)/(4πx) and k0 = 2πf /c0 . p(·) and

g(·) are monochromatic wave functions defined throughout the domain Ω
bounded by a closed surface S, while n · ∇ denotes the differentiation
along the inward normal to S. Provided that Ω is the half space z ≥ 0,
S thus consists of the plane z = 0 and a hemisphere of infinite radius,

centered at the origin. After adopting the Sommerfeld radiation condition
lim [∂p(r, ω)/∂r − ik0 p(r, ω)] = 0 to Equation (4.3), we obtain:
r→∞

1 ∂ exp(ik0 R1 ) exp(ik0 R1 ) ∂p(r , ω)
dx dy [p(r , ω) − ] = p(r, ω),
4π z =0 ∂z R1 R1 ∂z
(4.4)

where R1 = (x − x )2 + (y − y )2 + (z − z )2 , z > 0. On the other hand,
if r situates in the lower half space z < 0, we will obtain [11]:

1 ∂ exp(ik0 R2 ) exp(ik0 R2 ) ∂p(r , ω)
dx dy [p(r , ω) − ] = 0, (4.5)
4π z =0 ∂z R2 R2 ∂z

where R2 = (x − x )2 + (y − y )2 + (z + z )2 , z < 0. Subsequently, mak-
ing use of the connection between R1 and R2 at z = 0, we incorporate
Equation (4.4) and Equation (4.5), bringing about the acoustic RSI:

1 ∂ exp(ik0 R1 )
p(r, ω) = p(r , ω) dx dy . (4.6)
2π z =0 ∂z R1
Applying the first-order Born approximation [55] to Equation (4.6), which

simplifies the vibration at the fluid-solid interfaces at z = 0 to be uniform
and binary, we finalize the equation as the acoustic RSI cum BPSO:

2

1 z exp(ik0 R1 ) 1
|p(r, ω)| =
2
puni (r , ω) ik0 −
dx dy
, (4.7)
2π z =0 R1 2 R1
where puni (·) is a binary function describing the ring configuration at z = 0,

optimized for the designed focal pattern or focal resolution |p(r, ω)|2 . Other
than the requirement of acoustic pressure as well as its normal derivative

in Equation (4.3), Equation (4.7) only requires the acoustic pressure on the
surface, suitably for BPSO.
In general, BPSO is a nature-inspired evolutionary algorithm for stochastic

optimization [62]. At first, the designed energy distribution |p(r, ω)|2 will
be preset to benchmark acoustic focal patterns or resolutions in respective
scenarios. Simultaneously, by means of Equation (4.7), BPSO will be im-
plemented to optimize the parameter r = [r1 , r2 , r3 , ..., rN ] in order to fulfill
the benchmark |p(r, ω)|2 . For simplicity, the ring width Δr of the planar
metasurface PT is fixed in our design. Once r is given, one can obtain
another parameter rr = [rr1 , rr2 , rr3 , ..., rrN ] by rr = r + Δr. Then, the
parameters r and rr determine the ring configuration in Figure 4.1(a). Al-
though the fixed Δr sets a constraint in designing the active metasurface
PT, this low-cost approach always works well in many applications [63], as
we will show later.
4.2 Generation of an acoustic focal needle
To vindicate the proposed method, i.e. the acoustic RSI cum BPSO, in the
manipulation of acoustic focusing, we first demonstrate the manipulation
of the acoustic focal patterns such as the designed focal needle and the
designed multiple foci. The arbitrary design of a focal pattern is impossible
if we simply resort to acoustic wavefront construction by the method of
effective medium [43]. In our case for the pattern of a finite-length focal
needle on axis, we conveniently select V0 = 5V and f = 100kHz that
generates acoustic waves of λ = 3.43mm in the space z > 0. Note that
(a) RPaR2

Normalized Squared ŽƐĐŝůůĂƟŶŐ RSI+BPSO ĨĂƌͲĮĞůĚ
absolute pressure focal
ŶĞĂƌĮĞůĚ
FEM ŶĞĞĚůĞ

Depth
=5.89O

19.21O 25.1O z/O

(b) 0 RPaR2 10 20 30 40 50 54.2 RPaR2

1.46O
(5mm)
0O
r (0mm)
1.46O
(5mm) z
18.08O (62mm) 22.16O (76mm) 26.24O (90mm)
(c) RPaR2 r=1O
absolute pressure
50 47.2
40 z=20O
Squared
30 FWHM=2 [ 0.32O
20 =0.64O
10
0
0.0 0.4 0.8 1.2 1.6 2.0 0 RPaR2
r/O
(d) RPaR2 r=1O
absolute pressure
50 38.8
40 z=24O
Squared
30 FWHM=2 [ 0.31O
20 =0.62O
10
0
0.0 0.4 0.8 1.2 1.6 2.0 0 RPaR2
r/O
Figure 4.2: (a) The normalized squared absolute pressure, display-

ing the pattern of the designed finite-length far-field focal needle. (b)
The field distribution of the squared absolute pressure around the focal
needle. (c,d) The radial distributions of the squared absolute pressure
at the cross sections z = 20λ and z = 24λ, with their respective field
distributions. Figure adopted and reproduced with permission from ref.
[58].
for the purpose of a finite-length focal-needle pattern in the far field, we

require a depth of continuous acoustic focal energy along the axis with the
low energy level at the rest, whilst the location, i.e., both the depth of the
needle and the specific positions of the two ends away from the transducer,
could be subtly designed as well. In Figure 4.2(a), the on-axis focal-needle
pattern, whose position is preset to extend from 19.2λ to 25.1λ, is designed
as the orange dashed curve |p(r, ω)|2 , while simultaneously the optimized
ring configuration is calculated by the acoustic RSI cum BPSO as described
above. The optimized ring configuration listed in Appendix C includes 30
PZT-5H rings with the maximum radius of ∼180 mm, while q is adjusted
optimally to be 4 mm.
In the same plot, the full-wave acoustic field generated by the planar and
active metasurface PT as the real case is simulated by the FEM as the blue
curve, using Equation (4.1) and Equation (4.2), and the result coincides
with the designed finite-length focal-needle pattern (orange dashed curve).
It is noteworthy that the standing-wave-like oscillation from 0λ to ∼10λ
along z direction agrees with the classic characteristic of acoustic near fields
[42]. Besides, we notice that most energy is focused into the designed far-
field focal region from 19.21λ to 25.1λ, implying the focal needle of the
depth 5.89λ as designed, longer than the depth ∼4λ of the reported optical
needle [56]. In Figure 4.2(b), the field pattern of |p(r, ω)|2 simulated by
FEM around the focal needle is displayed. The contrast between the intense
needle and the ambient quiet field is apparent, which meets the requirement
of the focal-needle shape.
It is necessary to mention the subwavelength focal resolution of the focal-

needle pattern in Figure 4.2(c,d). The radial distributions of the squared
absolute pressure are plotted on the left while their field distributions are
on the right. The FWHM of the focal size in the far field is measured
to be ∼0.64λ at z = 20λ and z = 24λ, smaller than one wavelength but
larger than the diffraction limit. We may also conclude from Figure 4.2(c,d)
that the acoustic needle pattern is formed of a nearly constant subwave-
length width throughout. Moreover, the intensity of the side-lobes in Fig-
ure 4.2(c,d) is drastically smaller than the central intensity, crucially for
potential imaging applications. Note that such subwavelength acoustic fo-
cusing is generated in the true far field without resorting to evanescent
acoustic waves limited to the near field only.
4.3 Generation of acoustic far-field multiple
foci
To further show the robustness of the acoustic-focusing manipulation, we

take the example of the acoustic far-field multiple foci as another designed
focal pattern. Here, the multiple foci are designed as the four discrete
foci along the axis in the far field. The corresponding normalized energy
pattern |p(r, ω)|2 designed by the acoustic RSI cum BPSO is the orange
dashed curve in Figure 4.3(a), using the ring configuration which is simul-
taneously optimized in this case. Note that the ring configuration here is
designed and optimized in the same way except for a different focal pattern
(benchmark). It includes 28 PZT-5H rings, whose parameters are listed
in Appendix C. q = 3mm is optimized here while V0 and f remain the
same. The blue curve indicates the full-wave simulation by FEM using
(a) RPaR2
Normalized Squared ŽƐĐŝůůĂƟŶŐ ĨĂƌͲĮĞůĚ
absolute pressure RSI+BPSO ŵƵůƟĨŽĐŝ
ŶĞĂƌĮĞůĚ
FEM

z/O

18.40O 24.20O
(b) 0 RPaR2 10 20 30 40 50.8 RPaR2
1.46O
(5mm)
0O
r (0mm)
1.46O
(5mm) z
16.03O (55mm) 21.14O (72.5mm) 26.24O (90mm)
(c) RPaR2 r=1O
absolute pressure
50 49.4
FWHM=2 [ 0.22O
40
=0.44O
Squared
30
;ƐƵďͲĚŝīƌĂĐƟŽŶͲůŝŵŝƚͿ
20
10 z=18.40O
0
0.0 0.4 0.8 1.2 1.6 2.0 0 RPaR2
r/O
(d) RPaR2 r=1O
absolute pressure
50 FWHM=2 [ 0.23O 50.8

40
=0.46O
Squared
30 ;ƐƵďͲĚŝīƌĂĐƟŽŶͲůŝŵŝƚͿ
20
10
z=24.20O
0
0.0 0.4 0.8 1.2 1.6 2.0 0 RPaR2
r/O
Figure 4.3: (a) The normalized squared absolute pressure, displaying

the pattern of the designed far-field multiple foci. (b) The field distribu-
tion of the squared absolute pressure around the multiple foci. (c,d) The
radial distributions of the squared absolute pressure at the cross sections
z = 18.4λ and z = 24.2λ, and their respective field distributions. Figure
Equation (4.1) and Equation (4.2). Again, the satisfactory agreement be-
tween these two outcomes confirms our pattern design. The corresponding
field distribution in Figure 4.3(b) is simulated with respect to |p(r, ω)|2
around the multiple foci. Also, we notice that the focal resolution (FWHM
∼0.45λ) of the multiple foci in Figure 4.3(c,d) is subwavelength and even
beats the Rayleigh diffraction limit of 0.5λ, which was never realized in
terms of PT technology.
As mentioned in the above two cases, by means of the properly-optimized

ring configurations, we are able to achieve the designed acoustic focal pat-
terns with a subwavelength resolution as well as a sub-diffraction-limit reso-
lution. The underlying physics of achieving the focal resolution beyond the
restriction of diffraction limit by means of the multiple-ring active meta-
surface transducer is the interference and the diffraction of excited sound
beams.
In essence, Rayleigh diffraction has Airy pattern proportional to J1 (kr N A),

where kr is the radial wavenumber and N A is the numerical aperture. The
pattern can be decomposed into ∝ [J0 (kr N A) + J2 (kr N A)]2 where the term
J2 (kr N A) makes the focal size larger. In our case, each ring only generates
J0 (kr N A) at the focal point. Therefore, by eliminating J2 (kr N A), the su-
perposition at the focal plane breaches the traditional Rayleigh diffraction
limit, while the algorithm BPSO simultaneously optimizes the ring config-
uration to further improve the acoustic focal resolution [63]. If we intend
to further enhance the focal resolution to the deep sub-diffraction-limit fo-
cal size substantially smaller than 0.38λ, the situation will turn into the
extreme case of the super-oscillatory super resolution [60, 61].
RPaR2 z=20.06O
EŽƌŵĂůŝǌĞĚ^ƋƵĂƌĞĚ
ĨĂƌͲĮĞůĚ r=0.4O

ĂďƐŽůƵƚĞƉƌĞƐƐƵƌĞ
ƐƵƉĞƌͲŽƐĐŝůůĂƚŽƌǇ
ƐƵƉĞƌƌĞƐŽůƵƟŽŶ

r/O
Z^/нW^K FEM 0 2.9
FWHM=2 [ 0.15O=0.3O RPaR 2
;ĚĞĞƉƐƵďͲĚŝīƌĂĐƟŽŶͲůŝŵŝƚͿ
Figure 4.4: The radial distribution of the normalized squared absolute

pressure at the cross section z = 20.06λ on the left, showing the focal size
of the far-field acoustic super-oscillatory super resolution. On the right
is the corresponding field distribution of the squared absolute pressure.
Figure adopted and reproduced with permission from ref. [58].
4.4 Acoustic super-oscillatory super resolu-
tion
Our design method, the acoustic RSI cum BPSO in terms of PT technology,
is able to manipulate the acoustic focal resolution as well and to increase
it to acoustic super-oscillatory super resolution. Super oscillation is a sort
of phenomenon associated with the fact that the band-limited functions
can oscillate arbitrarily faster locally than the highest Fourier components,
which makes it possible to break the diffraction limit and to reduce the
focal size to be infinitesimal. In this case, we choose V0 = 100V and
f = 40kHz for the electric potential exerted on all PZT-5H rings, which
produce acoustic waves of λ = 8.575mm in air. After setting the acoustic
super-oscillatory super resolution (FWHM = 0.3λ at z = 20.06λ) as the
designed focal resolution, we simultaneously optimize the ring configuration
of the active metasurface PT in the same way, which includes 19 PZT-5H
rings as shown in Appendix C, while q = 2mm is adopted here optimally.
The optimized result is the orange dashed curve in Figure 4.4, showing
the designed radial distribution of the normalized |p(r, ω)|2 at z = 20.06λ.
The blue curve shows the corresponding full-wave simulation. The field
distribution of the squared absolute pressure at the same cross section is
plotted on the right in Figure 4.4 to exhibit the result of the focal-resolution
manipulation.
We manage to control the focal size of the super-oscillatory super resolution

to be 0.3λ in acoustic far fields, substantially smaller than 0.5λ. Neverthe-
less, if the FWHM of the focal spot is smaller than 0.38λ, there is an
inevitable trade-off uniquely in terms of the super-oscillatory case, which
does not exist for a larger FWHM. Namely, the cost of the super-oscillatory
super resolution is that most on-axis focal energy would be unavoidably
squeezed away into the side-lobes as shown in Figure 4.4 [63]. In prin-
ciple, the deep sub-diffraction-limit super resolution can be manipulated
to be infinitesimal, but the sacrifice is the increasing side-lobes that make
the super-oscillatory super resolution less efficient. However in spite of the
naturally-inevitable trade-off in the super-oscillatory case, people also try
applying it in biological imaging [60].
4.5 Discussion
Physically, the proposed active and configurable planar metasurface PT

used for the manipulation of acoustic focusing is a sort of binary-amplitude
exciter for directly focusing the excited sounds into a certain pattern or
a designed resolution in the post-evanescent fields, by optimally modi-
fying the constructive interference of a large amount of acoustic beams.
The acoustic RSI cum BPSO is applied to optimize the ring configuration
on the active metasurface after certain focusing targets such as focal pat-
terns or focal resolutions are chosen. Additionally, the manipulation of
acoustic focusing is not the translation from optics, whose binary mask is
passive without energy feeding [64]. Here, we fundamentally change the
passive binary-amplitude baffle in optics into the active binary-amplitude
exciter in acoustics that transforms itself into the feeding. Furthermore, we
demonstrate the versatility of our design method, such as the realization
of the focal-needle pattern, the multiple-foci pattern and the manipulated
super-oscillatory super resolution in acoustic far fields. Consequently, this

approach to manipulate focusing properties is unique in acoustics.
4.6 Conclusion
In conclusion, we propose the method to manipulate the acoustic far-field

focusing created with the active and configurable planar metasurface piezo-
electric transducer. The designed focal patterns and focal resolutions are
demonstrated respectively. By means of the acoustic Rayleigh-Sommerfeld
diffraction integral in conjunction with the method of binary particle swarm
optimization, a large degree of freedom is obtained to manipulate various
far-field focusing phenomena via optimized ring configurations. Here, the
pattern of a finite-length focal needle of a subwavelength resolution is de-
signed and realized; another example is the creation of the multiple-foci
pattern of a sub-diffraction-limit resolution; for the focal-resolution manip-
ulation, the extremely-high resolution, i.e., the acoustic super-oscillatory
super resolution, is demonstrated as well using the prototype of the active
and configurable planar metasurface piezoelectric transducer. Due to these
significant advantages and versatility, the manipulation of acoustic focus-
ing may offer a reformative framework in medical and industrial technology
where the strict control of acoustic high-energy areas is demanding (bio-
medical actuator, focused ultrasound surgery, lithotripsy, nondestructive
testing, etc.).
Chapter 5
Realizing acoustic cloaking and

near-zero density with acoustic
metastructure
This chapter proposes an acoustic metastructure which has the property of

near zero density and is developed for acoustic invisibility cloaking. This
acoustic metastructure sustains the characteristics of the reported acoustic
cloaks derived by transformation acoustics, and is also able to overcome
the defect of topologically-optimized acoustic cloaks. Different from the
traditional acoustic metamaterials which have complex (inhomogeneous or
anisotropic) components of micro or subwavelength scales, our acoustic
metastructure for invisibility cloaking is only made of single-piece homoge-
neous elastic copper in an accessible layout, including one pressure absorber
and one pressure projector connected by an isolated energy channel.
61
Chapter 5. Realizing acoustic cloaking and near-zero density 62
The elastic material can be regarded at a certain resonant frequency as

an effective density-near-zero (DNZ) structure. Due to the mechanical res-
onance of the elastic structure, the phase velocity of sound waves in the
cloaking setup almost reaches infinite value, and consequently, extraordi-
nary sound transmission (EST) is expected [65, 66]. The cloaking perfor-
mance by our acoustic metastructure is explained by simplified theoretical
spring-mass model and verified by COMSOL numerical simulation in two-
dimensional unbounded space as well as in curved waveguides.
The theoretical explanation of the DNZ property of our proposed metas-

tructure is to be elaborated using the spring-mass models at Section 5.2.
In short, DNZ is the consequence of systematic resonances, which result in
simultaneous vibration of the entire structure. Therefore, it appears that
the vibrational status at one end of the structure can reach to the other end
without taking time. The seeming infinite speed of sound can be further
paraphrased as the DNZ property.
One limitation of our design is the 1D functionality, i.e., the capability to

work only for normal incident waves.
(a) (b)
cloaked
cloaked object
object
energy channel
distorted
field
inhomogeneous cloak absorber projector
(c) ŶŽŶͲĮǆĞĚ
copper
ĮǆĞĚ
copper
1
ĐůŽĂŬĞĚŽďũĞĐƚ air chamber w

w=1 outside:
30
6
p=93 water
ĐůŽĂŬĞĚŽďũĞĐƚ air chamber
35
1
unit: mm
2.5 7 Cross-Sectional View
Figure 5.1: Comparison of (a) traditional acoustic cloaking based on

coordinate transformation and (b) our cloaking by a density-near-zero
metastructure for extraordinary sound transmission. (c) Our metastruc-
ture consists of one single-piece copper with air chambers inside it, and
it is immersed in water. Figure adopted and reproduced with permission
from ref. [67].
10k 14k Frequency (Hz) 22k 26k 30k

Transmission
1
(a) w=1mm
p=93mm
0
13412 Hz 16848 Hz 23742 Hz
13450 Hz 23100 Hz
(b) (c)
-9 -9
0 6.54x10 0 6.20x10
Displacement (mm) Displacement (mm)
(d)
16300Hz
-8
0 1.50x10
plane wave
Displacement (mm)
Figure 5.2: Simulated sound transmission through the density-near-

zero metastructure. The transmission shown in (a) is a function of the
frequency of the normally-incident sound waves. The three purple dots
indicate the eigen-frequencies of the copper part, and the three red
squares indicate the frequencies at the three peaks. The displacement
of the structure at input frequencies (three peaks) is shown in (b-d),
respectively. Figure adopted and reproduced with permission from ref.
[67].
5.1 Metastructure for acoustic cloaking made
by copper
To illustrate the concept of our design, we compare it with acoustic cloak-

ing based on coordinate transformation in Figure 5.1(a), which renders an
object invisible by distorting its ambient flow. The scheme of the proposed
DNZ acoustic-metastructure for cloaking in Figure 5.1(b) can produce EST,
to hide arbitrary inserted objects as well as to preserve wavefronts and
phases.
Conceptually, the flow at the front of an object is concentrated into the

energy channel by an absorber. Then, acoustic energy is coupled out by a
projector to the back side where the flow is restored. The entire process
resembles the engineering optical camouflage: positioning cameras upon an
object wrapped by a retro-reflecting coat; taking pictures and transmitting
the signal; projecting the front scene onto the back of the coat [68]. In
Figure 5.1(b), acoustic cloaking for 1D invisibility can be achieved without
wave distortion along sound paths, and the length of the cloaking device is
designed to maintain the phase continuity at both sound inlet and outlet.
The DNZ metastructure for acoustic cloaking shown in Figure 5.1(c) is only
implemented by copper (density: 8900kg/m3 ; Youngs modulus: 122GPa;
Poissons ratio: 0.35) [42] with two fixed planks (blue parts). p and w
are the length and width of the energy channel, respectively. The two
hollow enclosed chambers are designed to fill with air (density: 1.21kg/m3 ;
speed of sound: 343m/s), where any objects can be placed inside for the
purpose of invisibility. The T-shaped protrusion at each end is used as
the locally resonant element to enhance sound transmission [69, 70]. The
entire structure is immersed in water (density: 998kg/m3 ; speed of sound:

1481m/s), and a monochromatic acoustic plane wave propagates from left
to right. The COMSOL software has been used to do simulation.
The acoustic power transmission through the DNZ metastructure for acous-
tic cloaking immersed in water is numerically simulated. In Figure 5.2(a),
there are two frequencies that allow EST (13.45 kHz and 23.1 kHz) with a
small peak (16.3 kHz) in between. We also calculate the vibration states
at the three frequencies in Figure 5.2(b-d). Figure 5.2(b,c) shows by sim-
ulation two opposite resonances of the DNZ metastructure. Inside the
metastructure, the internal motion is simply the longitudinal movement
of the elastic energy channel, which can be seen at displacement fields.
We furthermore individually calculate the eigen-frequencies of the copper
parts, shown as the purple dots in Figure 5.2(a). The similarity between
the frequencies that correspond to the red squares and that correspond
to the purple dots shows that by the systematic resonance the structured
metastructure is capable of total power transmission, as the requirement of
acoustic cloaking. The slight disagreement between the dots of different col-
ors is because that for the red squares that correspond to Figure 5.2(b,c),
the acoustic load from the ambient water has been taken into account, while
for the purple dots, we only consider the eigen-frequencies of the copper
parts.
(a) air (b)

b) k
water
M f(Z)
air
f(Z)
(c) (d)
air
m m
air k2 k1 k2
1
(e)
p=77mm
p=85mm
p=93mm
Transmission
0
1
(f) w=1.0mm
w=1.3mm
w=1.6mm
0
10k 14k Frequency (Hz) 22k 26k 30k
Figure 5.3: (a,c) Schematics of the metastructure for acoustic cloak-

ing. Different colors stand for different parts of the metastructure. (b)
The spring-mass model. (d) The coupling model. (e,f) The power trans-
mission as a function of the input sound frequency when the shape of
the energy channel is changed. Figure adopted and reproduced with
5.2 Spring-mass model and density-near-zero
property
Our design focuses on achieving the DNZ property by using the structural
resonance, and meanwhile isolating the objects to be hidden from the en-
tire resonance system. A spring-mass model, the mechanical translation
of acoustic systems, is illustrated here to expound the resonance and the
consequent DNZ property.
The metastructure immersed in water in Figure 5.3(a) can be regarded as

the damping spring-mass model in Figure 5.3(b). In Figure 5.3(a), different
parts of the metastructure are tinted with different colors to illustrate the
counterpart elements in Figure 5.3(b). Specifically, we regard the fixed
copper planks (black) as the fixed wall (black), the input sound as the
driving force (pink), and the main body of the metastructure as the mass
chunk (blue). The joints (red) as well as the elasticity of the metastructure
are rationally modeled as the ideal spring (red). The acoustic impedance
exerted upon boundaries (yellow) is modeled as frictional damping, because
when the metastructure vibrates it will fight against the resistance from the
outside water and the inside air. Therefore, the equation for this spring-
mass model becomes:
M d2 x/dt2 = −kx − γM dx/dt + F0 cos ωt, (5.1)
where k denotes the stiffness of the ideal spring in Figure 5.3(b); γ is

the damping coefficient; M is the mass of the chunk; x is the displace-
ment; ω is the driving frequency and F0 cos ωt is the driving force. The

resonant frequency in the case of no damping is ω0 = k/M . Then,
the solution of Equation (5.1) becomes x = A cos(ωt − δ), where A =

F0 /[M (ω0 2 − ω 2 )2 + (ωγ)2 ] and tan δ = ωγ/(ω0 2 − ω 2 ) . The maximum

amplitude occurs at ω1 = ω0 2 − γ 2 /2. If there is no friction γ = 0, the
driving frequency is ω0 and A goes up infinitely, and the mechanical energy
of the vibrating chunk is accumulated because of the driving force. How-
ever, if there is a damping effect, e.g. the friction between the ground and
the chunk, the accumulated energy at the resonance will be conveyed to dis-
sipation, making the system in its steady state. Under this circumstance,
A will remain a finite maximum when ω = ω1 .
Due to energy conservation, all input power will be consumed by the fric-
tion, and the driving energy is transferred wherever the damping (yellow)
is. As the comparison in Figure 5.3(a), the force from the input wave not
only drives the vibration of the DNZ metastructure, but also needs to over-
come the resistances at the solid-liquid interfaces (yellow). Therefore, at
systematic resonances, the momentum gain of the acoustic loads, i.e. the
inside air and the outside water, consumes all the cumulative input acoustic
power. Consequently, at the moment of systematic resonances, the entire
system is in its steady state. Note that in order to avoid disturbing the
resonances of the metastructure, the objects to be hidden are required to
be kept from touching the walls of the chambers.
Moreover, since the acoustic impedance of the outside water is extremely

higher than that of the inside air (3561 times), the acoustic load is almost
completely attributed by water. Thus, all input power is expected to be
transferred to outside water, which leads to EST. The objects inside the
cloaked space are isolated from the systematic resonance of the DNZ metas-
tructure, because the air chambers decouple the systematic resonance from
the existence of the inside objects. The decoupling effect is not considered
by the traditional design of membrane-induced DNZ metamaterials [66].
However, the proposed metastructure is not able to cloak objects from
airborne sound. Because the acoustic forces exerted upon the inside and
outside air are comparable, the vibration of the metastructure will result
in sound penetration to both inside and outside, which makes the cloaking
effect in airborne sound fail.
Instead of the major resonances, there are several other minor resonances
due to the rich oscillation modes of solids, which are not observable simply
based on simplified 1D spring-mass model. However, the spring-mass model
is the classic approach to explain the underlying mechanism of EST and
the cause of the acoustic cloaking effect in our design.
Additionally, the spring-mass model implies the DNZ property of the metas-
tructure at its resonance, which is the acoustic equivalent of an electromag-
netic epsilon-near-zero metamaterial [71]. We may define the effective mass
of the vibration system as Mef f (ω) = M − k/ω 2 , which intrinsically in-
cludes the acoustic inertance caused by its mass as well as the acoustic
compliance caused by its elasticity. The combination of the acoustic com-
pliance and the inertance is the exact analog of the combination of spring
compliance and substantial mass in a spring-mass model.
Since the restoring force from the elastic copper of our structure is able to
add a negative term to the effective mass, we can rearrange Equation (5.1)
considering the harmonic vibration [65]:
F0 cos ωt − γM dx/dt = Mef f (ω)d2 x/dt2 , (5.2)
which turns into the form of Newton’s second law: driving force resistance
= mass×d2 x/dt2 . At systematic resonance, Mef f (ω) = M −k/ω 2 becomes
zero, so that power transmission of input sound is expected to be extremely
enhanced [65]. The causality from the DNZ property of our structure to
the resultant EST was elaborated in Ref. [65] where a similar membrane-
mass model was proposed. (The detailed derivation of the transmission
coefficient was given in the Ref. [65] using the lumped element approach.)
Actually, the DNZ effect is an innate property of a dynamic structure with
a certain eigen-vibration excited. For our proposed structure, the DNZ
metastructure at resonance is explained by the aid of the damping spring-
mass model in Figure 5.2(a,b).
However, it is noteworthy that although the dynamic density of the in-

dividual metastructure is near zero, the acoustic impedance of the entire
acoustic cloaking setup is not near zero at all. When the system vibrates,
there will be additional radiation impedance exerted at both ends of the
metastructure. When the systematic resonances occur, EST surely implies
the impedance match between the proposed structure and the surround-
ing. The impedance match in such case is dominantly attributed by the
radiation impedance, apart from the minor contribution from the DNZ
structure itself. Thus, if taken into account the radiation impedance, the
overall acoustic impedance is no more near zero.
5.3 Coupling model and geometrical depen-
dence
After discussing the mechanism of the DNZ metastructure for acoustic

cloaking, we further examine the geometrical dependence of the power
transmission spectrum, by employing the coupling model in Figure 5.3(c,d).
As discussed in Figure 5.3(a,b), the fixed copper planks in Figure 5.3(c)

(black) can be modeled as the two hard walls (black) in Figure 5.3(d).
To investigate the dependence between the structural geometry and the
resonance, the main bodies in Figure 5.3(c) are modeled separately. In this
way, the two copper bodies (blue) are modeled as the two chunks (blue)
with mass m. The four joints (red) indicating elasticity are interpreted
as the two ideal springs (red) with stiffness k2 , connecting the chunks to
the walls. Additionally, owing to the narrowness of the energy channel
(orange), we model it as the spring with stiffness k1 and mass ms (orange),
which couples the two chunks. By the coupling model, we discover the
dependence between the structural geometry of the DNZ metastructure
and the working frequencies of EST.
The equations for the coupling model shown in Figure 5.3(c,d) are:
⎧
⎨ md2 x1 dt2 = −k2 x1 + k1 (x2 − x1 ) − γmdx1 /dt
, (5.3)
⎩ md2 x dt2 = −k (x − x ) − k x − γmdx /dt
2 1 2 1 2 2 2
where x1 and x2 are the displacements of the left and the right chunks.
The lower and higher resonant frequencies of Equation (5.3) are ωL =

k2 /(m + ms ) − γ 2 /4 and ωH = (2k1 + k2 )/(m + ms ) − γ 2 /4 , respec-
tively. If the energy-channel length p = 93mm in Figure 5.1(c) becomes
longer, ms will become larger but k1 will get smaller, similarly to the serial
connection of springs. Therefore, both ωL and ωH will become smaller,
which implies that all resonant frequencies of the DNZ metastructure will
be shifted lower when p becomes longer. Vice versa, if the energy-channel
length is shortened the resonant frequencies will be shifted higher. The
shift observed from the curves in Figure 5.3(e) validates the theoretical
spring-model analysis.
If the channel width w = 1.0mm becomes thicker, ms as well as k1 will

become larger, similarly to the shunt connection of springs. Therefore, ωL
will become even lower. As for ωH , the double increments of k1 at the
numerator (2k1 + k2 ) is empirically larger than the increment of ms at the
denominator, leading to the rise of ωH . Thus, we can predict that if w
gets thicker, the low resonant frequencies of the DNZ metastructure will
be shifted even lower, whereas the high resonant frequencies will be even
higher. The broadening of the power transmission spectrum in Figure 5.3(f)
verifies our analysis.
We investigate the geometrical dependence in terms of the sound-tunneling

channel inside the metastructure, which is the key component of our struc-
ture. However, the two terminal chunks of the metastructure cannot be
straightforwardly characterized by spring-mass model. The solid eigen-
vibration of the chunks are 2D, which is hard to be interpreted by 1D
model. The precise analysis of the solid chunks needs direct numerical
simulation.
(a) -1.34 -1 0 1 1.44(Pa) (b) -1.43 -1 0 1 1.38(Pa)

4
cloaked 20
y (cm)
object
-4 10
-12 x (cm) 12
(c) 2.47
y (cm)
10
1
y (cm)
shadow -10
(Pa)
0
zone
-1
-10 -20
-2.14
-20 x (cm) 20
-20 x (cm) 20
Figure 5.4: Applying the metastructure for acoustic cloaking in un-

bounded space filled with water. (a) Two objects are hidden inside the
metastructure. There is no sound inside air chambers, and the field
outside the metastructure is almost unperturbed. (b) The density-near-
zero array is immersed in water, and the wavefront and the phase are
restored at its back. Figure adopted and reproduced with permission
from ref. [67].
5.4 Deploying cloaked area in free space and
in waveguides
We numerically examine the acoustic pressure field distribution in the do-

main where the proposed DNZ metastructure is immersed in water, while
the monochromatic acoustic waves are normally incident from left with unit
magnitude and frequency 23.1 kHz. As expected, inside the air chambers
in Figure 5.4(a), there is no sound penetration, which means the inserted
objects are isolated and decoupled to the systematic resonance and the
outside field. It is remarkable that the power transmission nearly reaches
100%, which implies EST through the DNZ metastructure with no back-
scattering.
Also, we notice in Figure 5.4(a) that the phase at the inner side of the
T-shaped ends is not continuous because of the perturbation from the local
resonances inside the concave of the T-shaped ends. However, the phase
at the outer side of the T-shaped ends is almost the same as the adjacent
ambient phase. The length of the metastructure is also designed to maintain
the phase continuity at the sound outlet, which makes the plane wavefront
instead of other curved wavefront propagate out.
50
2.131 (Pa)
40
2
30 1.5
20 1
0.5
10
0
y (cm)
-0.5
-10
-1
-20
-1.5
-30
-2
-40 -2.368 (Pa)
-50
-40 -20 0 20 40
x (cm)
Figure 5.5: Multiple metastructures form an “NUS”-shaped cloaked

space. Figure adopted and reproduced with permission from ref. [67].
(c)
Energy transmission efficiency:

waveguide a: 0.998
waveguide b: 0.998
waveguide c: 0.971
cloaked object
in waveguides
(a)
(b)
-1.43(Pa) 1.38(Pa)
-1 -0.5 0 0.5 1
Figure 5.6: Applying the metastructure for acoustic cloaking in waveg-

uides. (a) Four objects in a straight waveguide are hidden inside the
metastructures. There is no sound inside air chambers, and the outside
traveling sound is largely transmitted. (b) The number of metastruc-
tures does not affect the resonant frequencies. (c) The high power trans-
mission shows that the metastructures are able to hide inside objects as
well as to bend sound in waveguides. Figure adopted and reproduced
with permission from ref. [67].
Furthermore, thanks to the rigid boundaries at the two fixed planks in

Figure 5.1(c), we can connect as many metastructures sharing the planks
as possible, to form an arbitrarily-designed cloaked space. In Figure 5.4(b),
the metastructures are aligned side by side to increase the acoustic-cloaking
volume. The cloaking effect is demonstrated that the bulky copper array
itself and the multiple objects inside the metastructure are imperceptible
from outside. Contrarily, if the energy channels are removed, strong back-
scattering will occur as shown in Figure 5.4(c). We can also design the
overall cloaked space with an arbitrary distribution of the metastructures
[see Figure 5.5 where an NUS-shaped cloaked space is formed]. Based
on the proposed detachable DNZ metastructures, we accomplish arbitrary
acoustic cloaking in 2D space only by a single kind of uniform isotropic
material.
Besides the scenario in a 2D space, sound manipulation in waveguides

showed significant applications as well, such as the acoustic circulator based
on non-reciprocity [72]. Here, the proposed DNZ metastructure is also func-
tional in cloaking objects in waveguides. As shown in Figure 5.6(a) where
acoustic waves propagate through the hard-wall waveguide filled with wa-
ter, the objects inside the air chambers are imperceptible. Note that one
characteristic of DNZ property is that energy tunneling occurs indepen-
dently of the number of DNZ segments, because each DNZ segment is able
to resonate at the same frequency without influencing each other [66]. Even
if we add more metastructures inside the waveguide, the resonant frequency
will not be shifted, as shown in Figure 5.6(b). The energy transmission is
almost 100% in both Figure 5.6(a,b), which indicates EST happens through
DNZ metastructures. In Figure 5.6(c), we further demonstrate the acoustic
cloaking effect along a curved waveguide, which is used to bend sound path
and to maintain the cloaking effect simultaneously as designed.
5.5 Conclusion and discussion
We design an acoustic metastructure for invisibility cloaking that has the

density-near-zero property, and is able to eliminate the perceptibility of
inside objects from underwater sound. Note that the density-near-zero
metastructure for acoustic cloaking is built only by a uniform material,
while its cloaking effect is independent of the objects inside it. It much
simplifies the traditional realization of acoustic cloaking by complex acous-
tic metamaterials.
The design is inspired by the combination of acoustic inertance and acoustic

compliance of the structure at systematic resonances. A plane wavefront is
maintained without distortion, and the reflection is dramatically suppressed
due to extraordinary sound transmission. Moreover, such density-near-zero
metastructures are detachable, and therefore robust in being assembled to
change or expand the overall cloaked space with no limit in volume and
distribution. The flexibility can be universally applied in unbounded space
as well as in waveguides. We believe that the proposed density-near-zero
metastructure may open a distinct and concise way to acoustic cloaking by
using natural bulky materials.
There is an alternative way of constructing acoustic cloaking devices with-

out using coordinate transformation, which is called the topological-optimization
method [40]. A topology-optimized cloak [73] can hide objects inside it from
impinging sounds, which is similar to the function of our device. However,
the shortcoming of a topology-optimized cloak is that the shape and dis-

tribution of the cloak components are highly sensitive to the positions,
quantities, and the shapes of the inside objects to be hidden. For our pro-
posed approach, the metastructure of a DNZ cloak is irrelevant to the inside
objects, though it requires two media of large impedance mismatch.
Chapter 6
Future work
Up to now, we have proposed different designs of acoustic metasurface (one-

dimensional varying and two-dimensional varying metasurfaces, and active
metasurface) and acoustic metastructure, in order to manipulate sound
properties as much as we want (sound wavefront, vibrational orientation,
acoustic focusing, and density-near-zero property). We also have demon-
strated many novel acoustic applications on the basis of our designs, such
as acoustic illusion and ipsilateral focusing, conversion from propagating to
surface waves, generation of an acoustic focal needle and acoustic far-field
multiple foci, super-oscillatory super resolution, and acoustic invisibility
cloaking.
In the thesis we demonstrated the theoretical work as well as the simula-

tion results. We are also in the process of experimentally realizing these
proposed devices and measuring their performances. One difficulty of en-
gineering implementation is the required high precision. Starting from the
metasurface structure that generates the focused acoustic beams, we are us-
ing 3D printers to fabricate the device, and the experiment is ongoing with
81
Chapter 6. Future work 82
the collaboration of the Institute of Acoustics in Nanjing, China. The ex-

periment setup includes the signal function generator Agilent 33250A, the
amplifier Electronics Innovation, and a movable hydrophone ONDA HNC-
1000 controlled by a LABVIEW program for scanning acoustic pressure
and phase.
This chapter addresses two possible future plans based on the current
research about acoustic metasurface and metastructure. Section 6.1 in-
troduces the possible manipulation of acoustic band gap (Bragg gap and
non-Bragg gap) by acoustic periodic metastructure; Section 6.2 extends
the manipulation of sound properties to the manipulation of general wave
properties in incompressible fluids.
6.1 Manipulation of band gap for sound trans-
mission
Our proposed metastructures and metasurfaces have the property of period-

icity. The working frequency, i.e., band gap, is also an important parameter
to adjust for improving the performances of our devices. The generation
of pass-bands and stop-bands in periodic structures is commonly attribute
to the Bragg resonances, which can be interpreted as the cumulative re-
flection from each unit of the structure. For a periodic acoustic waveguide,
the evanescence of the waves is caused by the typical Bragg resonances.
When the periodicity is comparable to the radial scale of a waveguide, the

non-Bragg nature resonances, involving high-order transverse mode inter-

actions, can also play a role in forbidding the wave propagation. Non-
Bragg resonances, caused by the interference between the guided modes
of different transverse standing-wave profiles in the periodic waveguide,
are theoretically predicted in 2003 in a planar electromagnetic waveguide
with corrugated walls [74]. For the future work, we may design and im-
plement acoustic periodic metastructures in realizing non-Bragg acoustic
band gaps, and investigate the efficient design of the band gap. In detail,
the non-Bragg band gap, as a result of the interference between two trans-
verse guided wave modes, could be investigated in an axis-symmetrical and
periodic metastructure. The manipulation of acoustic band gap by metas-
tructures may benefit the design of spectrum structures (the location, the
width and the depth of band gaps) in the “band gap engineering”.
6.2 Manipulation of acoustic properties in
stratified fluids
All the previous work are done in the environment of a uniform fluid. One
step forward is to carry out the manipulation of acoustic properties in
stratified fluids. The typical wave that is able to exist in stratified incom-
pressible fluids is called internal waves. In a global scale, internal waves are
commonly generated by tidal flow over seafloor topography, which plays an
important role in dissipation and mixing in the interior of oceans [75, 76].
Also, these waves contribute significantly to the global oceanic energy bud-
get [77, 78]. The main contributor to the internal waves is the M2 tide,
a lunar semi-diurnal tide of frequency ω = 1.4052 × 10−4 rad/s [79]. One
direct numerical simulation of internal wave radiation for tidal flow over
synthetic random topography is shown in Figure 6.1. The details of the
simulation are elaborated in Appendix D.
If acoustic metasurfaces are placed in deep water, internal waves will distort
acoustic signals, resulting in the inefficiency of device’s functions. The first
step from manipulation of acoustic properties in uniform fluids to manip-
ulation in stratified fluids needs the thorough study of internal waves. We
are considering to conduct the laboratory study of steering acoustic prop-
agation in the presence of internal waves. The experiment setup could be:
a tank filled with a stratified fluid containing a wave-maker that generates
internal waves, and an acoustic track crossing the internal wave beams.
The measurements can be designed to provide a benchmark for a better
understanding of the influence of internal waves on 3D sound propagation.
The future work lies in implementing metasurfaces and metastructures in
manipulating acoustic properties in stratified fluids.
<VerƟcal energy Ňux> t / J 0 -10 10
1.7
(a)
z (km)
Chapter 6. Future work
0
1.7
(b)
z (km)
0
0 x (km) 10
Figure 6.1: Direct numerical simulation of internal wave radiation for tidal flow over synthetic random topography,
resolved in (a) to 39 m [256 cycles/(10 km)] and in (b) to 312 m [32 cycles/(10 km)]. The vertical component of the
normalized internal wave energy flux averaged over a tidal period is shown in red for upward flux and blue for downward
(the normalization factor J0 is the total internal wave power generated in (a) divided by the 10 km length of the topography).
85
Appendix A
Detailed derivation of
impedance-governed
generalized Snell’s law of
reflection (IGSL) in acoustics
We mathematically derive the connection between the interface specific

acoustic impedance (SAI) and the manipulation of wavefronts, which gives
birth to the proposed impedance-governed generalized Snell’s law of re-
flection (IGSL) as the design rule of SAI. In addition, we mathematically
predict the double reflections and the situation when the ordinary reflection
can be switched off.
We assume the time-harmonic factor in this appendix is e−iωt , where ω is

the circular frequency, and the coordinate system is that in Figure. A.1(a).
The incident acoustic pressure can be expressed as:
86
Appendix A. Detailed derivation of IGSL 87
(a) S z pro dl
pi
Ti Tro
Tre pre
Uc
0 0 D
SAI Zn o y
n
(b) z 2
r
observing
r0 point
point
source T* 1
Uc
0 0
y
o = EZ
image
source
T*
r0 †
n
Figure A.1: (a) Illustration for some notations. The orange line indi-
cates the contour of the Green’s integral. S is the semicircular contour;
D is the flat one along the surface. pi , pro and pre denote the incidence,
the ordinary reflection, and the extraordinary reflection, respectively. n
is the unit vector opposite to z direction. Zn is set for the flat interface
(z = 0). (b) Schematic diagram for the effective paths of acoustic ra-
diation. The introduced θ∗ can be interpreted as the effective incident
angle. r, r0 , and r†0 are the location vectors for the point source, the
image source and the observer, respectively.
pi (y, z, ω) = pi0 (ω) exp[ik0 (y sin θi − z cos θi )], (A.1)
where k0 = ω/c0 is the wave number in free space, θi the incident angle and
pi0 (ω) the amplitude. Zn (y, ω) = p(y, 0, ω)/[n · v(y, 0, ω)] as the specific
acoustic impedance (SAI) [42] of a locally reacting boundary is laid at
the interface, where n is the unit vector opposite to z direction and v
is the acoustic velocity. The boundary condition of this problem can be
paraphrased as [80]:
∂
p(y, 0, ω) + ik0 β(y, ω)p(y, 0, ω) = 0, (A.2)
∂z
where β(y, ω) = ρ0 c0 /Zn (y, ω) (ρ0 and c0 being the given density and sound
speed respectively in the upper space) is the normalized acoustical admit-
tance of the locally reacting surface.
We expand β to be β(y, ω) = β̃(y, ω) + β0 (ω), where β0 is a real constant.

The ordinary reflection is expressed as:
pro (y, z, ω) = pi0 (ω)R(θi , β0 ) exp[ik0 (y sin θro + z cos θro )], (A.3)
where R is the reflection coefficient and θro the angle of pro . Because pro
observes the usual Snell’s law, θro = θi . In order to find the expression of
R, we introduce the constant SAI:
ρ0 c 0 pi (y, 0, ω) + pro (y, 0, ω)

Z0 (ω) = = , (A.4)
β0 (ω) n · vi (y, 0, ω) + n · vro (y, 0, ω)
where n is the normal vector indicated in Figure. A.1(a), vi and vro the
acoustic velocities of pi and pro . Substituting Equation. (A.1) and Equa-
∂
tion. (A.3) into Equation. (A.4) and applying Euler equation ρ0 ∂t v = −∇p,
we obtain:
cos θi − β0 (ω)
R(θi , β0 ) = . (A.5)
cos θi + β0 (ω)
In Figure. A.1(a), the total acoustic field can be written in the integral
form:

∂
p(y, z, ω) = dl[G(y, z, ω; y0 , z0 ) p(y0 , z0 , ω) (A.6)
S+D ∂n0
∂
− p(y0 , z0 , ω) G(y, z, ω; y0 , z0 )],
∂n0
where dl(y0 , z0 ) is the infinitesimal segment along the integral contour, n0 =

n(y0 , z0 ) and G(y, z, ω; y0 , z0 ) is the Green’s function corresponding to the
following partial differential problem:
∇2 G + k0 2 G = −δ(y − y0 )δ(z − z0 ), z > 0

. (A.7)

[ ∂z∂ 0 G + ik0 β0 (ω)G]

=0
z0 =0
When the radius of the semicircular contour S approaches ∞, we can regard

the contour integral along S is mainly contributed by pi and pro . Therefore
Equation. (A.6) changes to be:
p(y, z, ω) = pi (y, z, ω) + pro (y, z, ω) (A.8)

∞
∂ ∂
− dy0 [G p(y0 , z0 , ω) − p(y0 , z0 , ω) G].
−∞ ∂z0 ∂z0
We can simplify Equation. (A.8) by substituting Equation. (A.7) and Equa-

tion. (A.2) into it. By defining the last part in Equation. (A.9) as the
extraordinary reflection pre (y, z, ω), which is the unique extra component
beyond pro , we obtain
∞
pre (y, z, ω) = ik0 β̃(y0 , ω)p(y0 , 0, ω)G(y, z, ω; y0 , 0)dy0 . (A.9)
−∞
The explicit solution of G(y, z, ω; y0 , z0 ) in Equation. (A.7) is
i (1)
G = H0 (k0 |r − r0 |) (A.10)
4 ∞
i 1 kz − ωβ0 /c0
+ exp[ikz (z + z0 ) + iky (y − y0 )]dky ,
4π −∞ kz kz + ωβ0 /c0
where r = (y, z), r0 = (y, z), and k0 2 = ky 2 + kz 2 . When r is away from the
surface D, kz ≈ k0 cos θ∗ holds, where θ∗ is introduced as a constant. Via
this approximation and another definition r†0 = (y0 , −z0 ), it turns out that
[54]
z − (−z0 )
cos θ∗ ≈

≈ constant. (A.11)
†

r − r 0
Through Equation. (A.11), it can be obtained that
kz − ωβ0 /c0 cos θ∗ − β0 (ω)

≈ ∗
≈ constant ≈ R(θ∗ , β0 ). (A.12)
kz + ωβ0 /c0 cos θ + β0 (ω)
Applying Equation. (A.12) into Equation. (A.10) and using the formula of
the cylindrical wave expansion in terms of plane waves, we approach a neat
form of the Green’s function:
i (1) i (1)
G(y, z, ω; y0 , z0 ) ≈ H0 (k0 |r − r0 |) +R(θ∗ , β0 ) H0 (k0

r − r†0
), (A.13)
4 4
(1)
where H0 (·) the Hankel function of the first kind. [81]
From the physical insight into Equation. (A.13), the first part of G is the
direct contribution of the point source to the observer through path 2 in
Figure. A.1(b). The second part is the product of the Green’s function
excited by the image source and the reflection coefficient R, denoting pro .
According to our interpretation, Figure. A.1(b) illustrates path 1 and path
2, visualized as pro and pre respectively[54]. Due to the expression of R, we
figure out that θ∗ is the effective incident angle regarding to Figure. A.1(b).
Furthermore, it is reasonable to say that the major contribution of the
integral in Equation. (A.10) is attributed to the vicinity of θ∗ , in which
way R can be regarded as a constant and put outside the integral.
By far-field approximation, we are able to get these expansions:
r · r0 = r(y0 sin θ + z0 cos θ)

r · r†0 = r(y0 sin θ − z0 cos θ) , (A.14)

(1)
2 i(x− π4 )
H0 (x)
≈ πx e
x→∞
where r is the length of r; sin θ = y/r; cos θ = z/r. Substituting Equa-

tion. (A.14) and Equation. (A.5) into Equation. (A.13), we obtain:

1 i(k0 r− π4 ) −ik0 y0 sin θ cos θ∗
G(y, z, ω; y0 , 0) ≈ i e e . (A.15)
2πk0 r cos θ∗ + β0 (ω)
After substituting Equation. (A.15) into Equation. (A.9), the extraordinary

reflection becomes:
∞
k0 i(k0 r− π ) cos θ∗
pre ≈ − e 4 β̃(y0 , ω)p(y0 , 0, ω)e−ik0 y0 sin θre dy0 .
2πr cos θ∗ + β0 (ω) −∞
(A.16)
Further, after applying Born approximation to Equation. (A.16) and ex-

panding it by Equation. (A.1) and Equation. (A.3), pre becomes:

2k0 pi0 (ω) exp[i(k0 r − π4 )] cos θ∗ cos θi
pre ≈ − × (A.17)
πr [cos θ∗ + β0 (ω)][cos θi + β0 (ω)]
∞
× β̃(y0 , ω)eik0 y0 (sin θi −sin θre ) dy0 .
−∞
Now we consider our proposed SAI:

ψ(y) ρ0 c 0
Zn (y, ω) = A 1 − i tan ; β0 (ω) = . (A.18)
2 2A
After substituting Equation. (A.18) into Equation. (A.3) and Equation. (A.17),
we obtain the ordinary reflection and the extra reflection: extraordinary re-
flection:
2A cos θi − ρ0 c0
pro ∝ exp[ik0 (y sin θro + z cos θro )], (A.19)
2A cos θi + ρ0 c0
∞
pre ∝ eiψ(y) eik0 y(sin θi −sin θre ) dy. (A.20)
−∞
Here note that in our case we are able to create double reflections by means
of SAI inhomogeneity.
Equation. (A.20) is a Dirac Delta if we consider ψ(y) to be a linear term

as the first order approximation. Or else, we know that the integral in
Equation. (A.20) will reach the maximum by imposing the stationary phase
approximation, i.e.,
1 dψ(y)
sin θre − sin θi = . (A.21)
k0 dy
Although Equation. (A.21) corresponds to the form of the generalized

Snell’s law of reflection (GSL) [1], the variables in the two situations are dif-
ferent. Starting from Equation. (A.18) and ending up with Equation. (A.21),
we provide the insight between our designed SAI and the direction of pre ,
without considering the phase in terms of wave propagation. We name
Equation. (A.21) as IGSL in acoustics, as the design principle of the SAI
Equation. (A.18).
According to Equation. (A.19), if A = (ρ0 c0 )/(2 cos θi ), we can switch off

pro . Therefore Equation. (A.18) becomes

ρ0 c 0 ψ(y)
Zn (y, ω) = 1 − i tan . (A.22)
2 cos θi 2
Appendix B
Distinction of IGSL in
acoustics
The same appearance of the impedance-governed generalized Snell’s law of

reflection (IGSL) and the generalized Snell’s law of reflection (GSL) may
cause the false impression that our IGSL is the same as GSL. Actually their
mechanisms are totally distinct.
In terms of phase inhomogeneity, the anomalous reflection pra actually cor-

responds to the situation when the ordinary reflection pro is steered toward
a “wrong” direction governed by GSL [1], illustrated in Figure. B.1(a).
There is only one single direction of reflection all the while. On the con-
trary in terms of SAI inhomogeneity, it is found that IGSL cannot alter
pro by an SAI interface, but can “turn off” pro so as to provide insight into
the engineering of special wavefronts by SAI interface, illustrated in Fig-
ure. B.1(b). Moreover, the extraordinary reflection pre governed by IGSL
is an additionally unique component in acoustics, which can be “geared”
94
Appendix B. Difference between GSL and IGSL 95
along arbitrary directions, simultaneously with vanishing pro . Therefore,

our proposed IGSL opens up rich effects and unprecedented applications in
the community of acoustics. Additionally, GSL can even be considered as
one subset of IGSL, when pro is turned off. In order to stress the irrelevance
between IGSL and GSL again, we list the differences:
1. GSL is initiated in electromagnetism with electric properties; IGSL

is initiated in acoustics with mechanical properties.
2. GSL is derived from Fermat Principle, i.e., the conservation of the

wave number along an interface; IGSL is derived from Green’s func-
tion. The fundamental physics is distinguished.
3. The variable of GSL is phase inhomogeneity; the variable of IGSL is

impedance inhomogeneity. The methods are independent.
4. GSL will only generate single reflection; IGSL not only can generate
single reflection, but also can generate double reflections.
5. GSL acts upon pro ; IGSL acts upon pre .
6. In GSL, the anomalous reflection corresponds to the situation where

pro is tweaked toward a different direction governed by GSL; in acous-
tics, IGSL cannot alter pro by SAI interfaces, but is capable of “turn-
ing on” or “turning off” pro .
Appendix B. Difference between GSL and IGSL 96
(a) GS
pi Z pro L
Ti Tro
Uc
0 0 proGSL pra
y
inhomogeneous
phase change
(b) Z on or off
pi pro
IGS
Ti Ti Tre pre
L
Uc
0 0
y
inhomogeneous
acoustic impedance
Figure B.1: (a) For a flat interface with an inhomogeneous phase

change, the angle of pro , i.e., θro , is tweaked in a fashion of GSL. The ma-
nipulated “ordinary reflection” is called to be the anomalous reflection
pra in terms of GSL. [1] (b) For a flat interface with an inhomogeneous
SAI, θro = θi without influence, while pre occurs simultaneously and θre
is controlled by IGSL, implying double reflections. If SAI is properly
controlled, pro can be switched off.
Appendix C
Parameters of metasurface
piezoelectric transducer
C.1 Parameters of Lead Zirconate Titanate
PZT-5H
The density of PZT-5H is 7500 kg/m3 .
Table C.1: Symmetric elasticity matrix (Pa)
1.27×1011 8.02×1010 8.47×1010 0 0 0

0 1.27×1011 8.47×1010 0 0 0
11
0 0 1.17×10 0 0 0
0 0 0 2.30×1010 0 0
0 0 0 0 2.30×1010 0
10
0 0 0 0 0 2.35×10
97
Appendix C. Parameters of metasurface transducer 98
Table C.2: Relative permittivity εSr
1704.4 0 0
0 1704.4 0
0 0 1433.6
Table C.3: Coupling matrix e (C/m2 )
0 0 0 0 17.0345 0
0 0 0 17.0345 0 0
-6.6228 -6.6228 23.2403 0 0 0
C.2 Optimized configuration of ring pattern
Table C.4: Configuration of 30 rings to generate a focal needle (mm)
r1 2.95 r7 36.68 r13 71.28 r19 102.62 r25 137.42

rr1 7.07 rr7 40.79 rr13 75.39 rr19 106.74 rr25 141.54
r2 8.19 r8 43.37 r14 76.35 r20 109.67 r26 143.69
rr2 12.31 rr8 47.49 rr14 80.47 rr20 113.78 rr26 147.81
r3 13.27 r9 49.20 r15 81.39 r21 114.91 r27 150.20
rr3 17.39 rr9 53.32 rr15 85.51 rr21 119.02 rr27 154.32
r4 19.32 r10 56.02 r16 86.77 r22 120.05 r28 156.36
rr4 23.44 rr10 60.13 rr16 90.89 rr22 124.17 rr28 160.48
r5 24.66 r11 60.79 r17 92.76 r23 126.30 r29 161.12
rr5 28.78 rr11 64.90 rr17 96.88 rr23 130.41 rr29 165.23
r6 30.33 r12 66.25 r18 97.84 r24 131.53 r30 167.04
rr6 34.45 rr12 70.36 rr18 101.95 rr24 135.65 rr30 171.16
Appendix C. Parameters of metasurface transducer 99
Table C.5: Configuration of 28 rings to generate multiple foci (mm)
r1 6.86 r8 49.05 r15 107.70 r22 171.84

rr1 10.98 rr8 53.17 rr15 111.82 rr22 175.96
r2 13.03 r9 55.22 r16 116.96 r23 183.50
rr2 17.15 rr9 59.34 rr16 121.08 rr23 187.62
r3 17.49 r10 63.80 r17 123.82 r24 194.14
rr3 25.73 rr10 67.92 rr17 127.94 rr24 198.26
r4 27.44 r11 72.37 r18 131.71 r25 201.34
rr4 31.56 rr11 76.49 rr18 135.83 rr25 205.46
r5 32.58 r12 79.92 r19 143.03 r26 216.09
rr5 36.70 rr12 84.04 rr19 147.15 rr26 220.21
r6 37.04 r13 86.78 r20 156.06 r27 230.49
rr6 41.16 rr13 90.90 rr20 160.18 rr27 238.73
r7 42.53 r14 95.69 r21 164.64 r28 244.90
rr7 46.65 rr14 99.81 rr21 168.76 rr28 249.02
Table C.6: Configuration of 19 rings to generate super-oscillatory su-

per resolution (mm)
r1 5.33 r6 46.56 r11 91.79 r16 136.87

rr1 9.61 rr6 50.84 rr11 96.08 rr16 141.15
r2 11.77 r7 55.50 r12 99.74 r17 145.58
rr2 16.05 rr7 59.79 rr12 104.03 rr17 149.87
r3 19.97 r8 64.13 r13 111.33 r18 151.41
rr3 24.26 rr8 68.42 rr13 115.61 rr18 155.69
r4 29.33 r9 72.18 r14 118.42 r19 159.96
rr4 33.62 rr9 76.47 rr14 122.71 rr19 167.56
r5 39.35 r10 81.53 r15 128.72
rr5 43.64 rr10 85.82 rr15 133.01
Appendix D
Details of the simulation for

internal waves
We conduct direct numerical simulations for tidal flow over synthetic ran-
dom topography. The computational domain of width 10 km has periodic
boundary conditions on the sides and a no-slip boundary condition on the
random bottom topography. Wave reflection from the top boundary of the
domain is avoided by adding a Rayleigh damping force that gradually in-
creases upward. The damping force starts from a height of 5 km and goes
up to the domain top of 15.5 km height; the approximately 3 × 106 control
volumes have unstructured grids.
The simulations use the CDP 2.4 code, which implements a fractional-step
time-marching scheme [82]. This code with disabled sub-grid modeling and
an addition of buoyancy forces has been used and validated in previous
studies of IW generation by tidal flow in stratified fluids [83–85]. Tidal
flow is produced by adding a horizontal force, F (t) = ρ0 U0 ωcos(ωt), to the
100
Appendix D. Simulation for internal waves 101
momentum equation. We use U0 = 0.14 cm/s, resulting in a tidal excursion

close to U0 /ω = 10 m. The small excursion relative to the autocorrelation
widths of the topographies is chosen to avoid overturning and turbulence,
since the interest here is in the radiation. The viscosity, ν = 0.01m2 /s, is
four orders larger than that of water to reduce the simulation cost (sparser
grids are needed for higher ν); however, the large viscosity has a negligible
effect on tidal conversion in the regime of laminar flow [84]. The salt
diffusivity κ is 2 × 10−5 m2 /s, which gives negligible diffusion for the large
Schmidt number, ν/κ = 500. Each simulation has 2000 time steps per
period and extends for 20 tidal periods to ensure a steady state. The
convergence is tested by doubling the spatial and temporal resolution; this
changes the computed IW power by less than 3%.
The radiated IW power is calculated from the horizontal integral of the

vertical energy flux averaged over a tidal period, Φz (x, z) ≡ p v t at z =
ztop , where ztop is the maximum topographic height, p is the wave pressure,
and v is the wave velocity, given by v (x, z, t) = v(x, z, t)−vbaro (z, t), where
v(x, z, t) is the fluid velocity and vbaro (z, t) is the barotropic velocity, which
in our simulations is approximated by U0 sin(ωt) because our domain height
is much taller than the topographic height (thus the flow acceleration over
the topography is negligible). An example of the time-averaged energy flux
field Φz (x, z) from our simulations is shown in Figure 6.1(a).
Bibliography
[1] N. Yu, P. Genevet, M. A. Kats, F. Aieta, J.-P. Tetienne, F. Capasso,

and Z. Gaburro, “Light propagation with phase discontinuities: gen-
eralized laws of reflection and refraction,” Science, vol. 334, no. 6054,
pp. 333–337, 2011.
[2] B. Sundar, A. C. Hamilton, and J. Courtial, “Fermat’s principle and

the formal equivalence of local light-ray rotation and refraction at the
interface between homogeneous media with a complex refractive index
ratio,” Optics Letters, vol. 34, no. 3, pp. 374–376, 2009.
[3] S. Larouche and D. R. Smith, “Reconciliation of generalized refraction

with diffraction theory,” Optics Letters, vol. 37, no. 12, pp. 2391–2393,
2012.
[4] J. Zhao and Z.-Y. Tao, “Fine-tuning of nonbragg bandgaps in ax-

isymmetric ducts via arbitrary periodic walls,” Journal of Sound and
Vibration, vol. 332, no. 25, pp. 6541–6551, 2013.
[5] X. Ni, N. K. Emani, A. V. Kildishev, A. Boltasseva, and V. M. Sha-

laev, “Broadband light bending with plasmonic nanoantennas,” Sci-
ence, vol. 335, no. 6067, pp. 427–427, 2012.
102
Bibliography 103
[6] M. Kang, T. Feng, H.-T. Wang, and J. Li, “Wave front engineering
from an array of thin aperture antennas,” Optics Express, vol. 20,
no. 14, pp. 15882–15890, 2012.
[7] F. Aieta, P. Genevet, M. A. Kats, N. Yu, R. Blanchard, Z. Gaburro,

and F. Capasso, “Aberration-free ultrathin flat lenses and axicons at
telecom wavelengths based on plasmonic metasurfaces,” Nano Letters,
vol. 12, no. 9, pp. 4932–4936, 2012.
[8] L. Huang, X. Chen, H. Muhlenbernd, G. Li, B. Bai, Q. Tan, G. Jin,

T. Zentgraf, and S. Zhang, “Dispersionless phase discontinuities for
controlling light propagation,” Nano Letters, vol. 12, no. 11, pp. 5750–
5755, 2012.
[9] P. Genevet, N. Yu, F. Aieta, J. Lin, M. A. Kats, R. Blanchard, M. O.

Scully, Z. Gaburro, and F. Capasso, “Ultra-thin plasmonic optical
vortex plate based on phase discontinuities,” Applied Physics Letters,
vol. 100, no. 1, p. 013101, 2012.
[10] N. Engheta, “Antenna-guided light,” Science, vol. 334, no. 6054,

pp. 317–318, 2011.
[11] M. Born and E. Wolf, Principles of optics: electromagnetic theory of

propagation, interference and diffraction of light. Cambridge university
press, 1999.
[12] E. Hecht, Optik. McGraw-Hill, 1987.
[13] J. Achenbach, Wave propagation in elastic solids. Elsevier, 2012.
[14] G. Clement, P. White, and K. Hynynen, “Enhanced ultrasound trans-

mission through the human skull using shear mode conversion,” The
Bibliography 104
Journal of the Acoustical Society of America, vol. 115, no. 3, pp. 1356–
1364, 2004.
[15] J. Li and C. Chan, “Double-negative acoustic metamaterial,” Physical

Review E, vol. 70, no. 5, p. 055602, 2004.
[16] Y. Lai, Y. Wu, P. Sheng, and Z.-Q. Zhang, “Hybrid elastic solids,”
Nature Materials, vol. 10, no. 8, pp. 620–624, 2011.
[17] F. Falcone, T. Lopetegi, M. Laso, J. Baena, J. Bonache, M. Beruete,

R. Marqués, F. Martin, and M. Sorolla, “Babinet principle applied
to the design of metasurfaces and metamaterials,” Physical Review
Letters, vol. 93, no. 19, p. 197401, 2004.
[18] S. Sun, Q. He, S. Xiao, Q. Xu, X. Li, and L. Zhou, “Gradient-

index meta-surfaces as a bridge linking propagating waves and surface
waves,” Nature Materials, vol. 11, no. 5, pp. 426–431, 2012.
[19] X. Chen, L. Huang, H. Mühlenbernd, G. Li, B. Bai, Q. Tan, G. Jin, C.-

W. Qiu, S. Zhang, and T. Zentgraf, “Dual-polarity plasmonic metalens
for visible light,” Nature Communications, vol. 3, p. 1198, 2012.
[20] J. Hao, Y. Yuan, L. Ran, T. Jiang, J. A. Kong, C. Chan, and L. Zhou,

“Manipulating electromagnetic wave polarizations by anisotropic
metamaterials,” Physical Review Letters, vol. 99, no. 6, p. 063908,
2007.
[21] S.-J. Song, H. J. Shin, and Y. H. Jang, “Development of an ultra

sonic phased array system for nondestructive tests of nuclear power
plant components,” Nuclear Engineering and Design, vol. 214, no. 1,
pp. 151–161, 2002.
Bibliography 105
[22] S. Chatillon, G. Cattiaux, M. Serre, and O. Roy, “Ultrasonic non-

destructive testing of pieces of complex geometry with a flexible phased
array transducer,” Ultrasonics, vol. 38, no. 1, pp. 131–134, 2000.
[23] S. W. Shin, A. R. Qureshi, J.-Y. Lee, and C. B. Yun, “Piezoelectric

sensor based nondestructive active monitoring of strength gain in con-
crete,” Smart Materials and Structures, vol. 17, no. 5, p. 055002, 2008.
[24] B. Jaffe, Piezoelectric ceramics, vol. 3. Elsevier, 2012.
[25] W. Liu and X. Ren, “Large piezoelectric effect in pb-free ceramics,”

Physical Review Letters, vol. 103, no. 25, p. 257602, 2009.
[26] E.-K. Kim, C. S. Park, W. Y. Chung, K. K. Oh, D. I. Kim, J. T.

Lee, and H. S. Yoo, “New sonographic criteria for recommending fine-
needle aspiration biopsy of nonpalpable solid nodules of the thyroid,”
American Journal of Roentgenology, vol. 178, no. 3, pp. 687–691, 2002.
[27] J. E. Kennedy, “High-intensity focused ultrasound in the treatment

of solid tumours,” Nature Reviews Cancer, vol. 5, no. 4, pp. 321–327,
2005.
[28] E. A. Stewart, W. M. Gedroyc, C. M. Tempany, B. J. Quade, Y. Inbar,

T. Ehrenstein, A. Shushan, J. T. Hindley, R. D. Goldin, M. David,
et al., “Focused ultrasound treatment of uterine fibroid tumors: safety
and feasibility of a noninvasive thermoablative technique,” American
Journal of Obstetrics and Gynecology, vol. 189, no. 1, pp. 48–54, 2003.
[29] M. A. Averkiou and R. O. Cleveland, “Modeling of an electrohydraulic

lithotripter with the kzk equation,” The Journal of the Acoustical So-
ciety of America, vol. 106, no. 1, pp. 102–112, 1999.
Bibliography 106
[30] B. W. Drinkwater and P. D. Wilcox, “Ultrasonic arrays for non-

destructive evaluation: A review,” Ndt & E International, vol. 39,
no. 7, pp. 525–541, 2006.
[31] J. Wang, W. Chen, and Q. Zhan, “Engineering of high purity ultra-

long optical needle field through reversing the electric dipole array
radiation,” Optics Express, vol. 18, no. 21, pp. 21965–21972, 2010.
[32] M. Friese, T. Nieminen, N. Heckenberg, and H. Rubinsztein-Dunlop,

“Optical alignment and spinning of laser-trapped microscopic parti-
cles,” Nature, vol. 394, no. 6691, pp. 348–350, 1998.
[33] N. H. Gokhale, J. L. Cipolla, and A. N. Norris, “Special transforma-

tions for pentamode acoustic cloaking,” The Journal of the Acoustical
Society of America, vol. 132, no. 4, pp. 2932–2941, 2012.
[34] A. Norris and A. Shuvalov, “Elastic cloaking theory,” Wave Motion,

vol. 48, no. 6, pp. 525–538, 2011.
[35] D. Schurig, J. Mock, B. Justice, S. A. Cummer, J. B. Pendry, A. Starr,

and D. Smith, “Metamaterial electromagnetic cloak at microwave fre-
quencies,” Science, vol. 314, no. 5801, pp. 977–980, 2006.
[36] A. N. Norris, “Acoustic cloaking theory,” in Proceedings of the Royal

Society of London A: Mathematical, Physical and Engineering Sci-
ences, vol. 464, pp. 2411–2434, The Royal Society, 2008.
[37] S. Zhang, C. Xia, and N. Fang, “Broadband acoustic cloak for ul-
trasound waves,” Physical Review Letters, vol. 106, no. 2, p. 024301,
2011.
Bibliography 107
[38] B.-I. Popa, L. Zigoneanu, and S. A. Cummer, “Experimental acous-

tic ground cloak in air,” Physical Review Letters, vol. 106, no. 25,
p. 253901, 2011.
[39] V. M. Garcı́a-Chocano, L. Sanchis, A. Dı́az-Rubio, J. Martı́nez-Pastor,

F. Cervera, R. Llopis-Pontiveros, and J. Sánchez-Dehesa, “Acoustic
cloak for airborne sound by inverse design,” Applied Physics Letters,
vol. 99, no. 7, p. 074102, 2011.
[40] L. Sanchis, V. Garcı́a-Chocano, R. Llopis-Pontiveros, A. Climente,

J. Martı́nez-Pastor, F. Cervera, and J. Sánchez-Dehesa, “Three-
dimensional axisymmetric cloak based on the cancellation of acoustic
scattering from a sphere,” Physical Review Letters, vol. 110, no. 12,
p. 124301, 2013.
[41] J. Zhao, B. Li, Z. Chen, and C.-W. Qiu, “Manipulating acoustic wave-
front by inhomogeneous impedance and steerable extraordinary reflec-
tion,” Scientific Reports, vol. 3, 2013.
[42] D. T. Blackstock, Fundamentals of physical acoustics. John Wiley &

Sons, 2000.
[43] Y. Li, B. Liang, X. Tao, X.-f. Zhu, X.-y. Zou, and J.-c. Cheng, “Acous-
tic focusing by coiling up space,” Applied Physics Letters, vol. 101,
no. 23, p. 233508, 2012.
[44] J. Renger, M. Kadic, G. Dupont, S. S. Aćimović, S. Guenneau,

R. Quidant, and S. Enoch, “Hidden progress: broadband plasmonic
invisibility,” Optics Express, vol. 18, no. 15, pp. 15757–15768, 2010.
Bibliography 108
[45] C. Zhang and T. J. Cui, “Negative reflections of electromagnetic waves

in a strong chiral medium,” Applied Physics Letters, vol. 91, no. 19,
p. 194101, 2007.
[46] D. Fattal, J. Li, Z. Peng, M. Fiorentino, and R. G. Beausoleil, “Flat

dielectric grating reflectors with focusing abilities,” Nature Photonics,
vol. 4, no. 7, pp. 466–470, 2010.
[47] J. Zhu, Y. Chen, X. Zhu, F. J. Garcia-Vidal, X. Yin, W. Zhang,

and X. Zhang, “Acoustic rainbow trapping,” Scientific Reports, vol. 3,
2013.
[48] D.-D. Dai and X.-F. Zhu, “An effective gauge potential for nonrecip-
rocal acoustics,” Europhysics Letters, vol. 102, no. 1, p. 14001, 2013.
[49] Y. Li, B. Liang, Z.-m. Gu, X.-y. Zou, and J.-c. Cheng, “Reflected wave-
front manipulation based on ultrathin planar acoustic metasurfaces,”
Scientific Reports, vol. 3, 2013.
[50] Y. Li, X. Jiang, R.-q. Li, B. Liang, X.-y. Zou, L.-l. Yin, and J.-c.
Cheng, “Experimental realization of full control of reflected waves with
subwavelength acoustic metasurfaces,” Physical Review Applied, vol. 2,
no. 6, p. 064002, 2014.
[51] J. Zhao, B. Li, Z. N. Chen, and C.-W. Qiu, “Redirection of sound

waves using acoustic metasurface,” Applied Physics Letters, vol. 103,
no. 15, p. 151604, 2013.
[52] F. Mechel, “On sound absorption of finite-size absorbers in relation to

their radiation impedance,” Journal of Sound and Vibration, vol. 135,
no. 2, pp. 225–262, 1989.
Bibliography 109
[53] F. P. Mechel, Formulas of acoustics, vol. 2. Springer Science & Busi-

ness Media, 2002.
[54] G. Taraldsen, “The complex image method,” Wave Motion, vol. 43,
no. 1, pp. 91–97, 2005.
[55] J. J. Sakurai and J. Napolitano, Modern quantum mechanics. Addison-

Wesley, 2011.
[56] H. Wang, L. Shi, B. Lukyanchuk, C. Sheppard, and C. T. Chong,

“Creation of a needle of longitudinally polarized light in vacuum using
binary optics,” Nature Photonics, vol. 2, no. 8, pp. 501–505, 2008.
[57] R. Castellano and L. Feinstein, “Ion-beam deposition of thin films of

ferroelectric lead zirconate titanate (pzt),” Journal of Applied Physics,
vol. 50, no. 6, pp. 4406–4411, 1979.
[58] J. Zhao, H. Ye, K. Huang, Z. N. Chen, B. Li, and C.-W. Qiu, “Ma-
nipulation of acoustic focusing with an active and configurable planar
metasurface transducer,” Scientific Reports, vol. 4, 2014.
[59] B. A. Auld, Acoustic fields and waves in solids, vol. 2. RE Krieger,

1990.
[60] E. T. Rogers, J. Lindberg, T. Roy, S. Savo, J. E. Chad, M. R. Dennis,

and N. I. Zheludev, “A super-oscillatory lens optical microscope for
subwavelength imaging,” Nature Materials, vol. 11, no. 5, pp. 432–
435, 2012.
[61] H. Ye, C.-W. Qiu, K. Huang, J. Teng, B. Lukyanchuk, and S. P. Yeo,

“Creation of a longitudinally polarized subwavelength hotspot with an
Bibliography 110
ultra-thin planar lens: vectorial rayleigh–sommerfeld method,” Laser

Physics Letters, vol. 10, no. 6, p. 065004, 2013.
[62] N. Jin and Y. Rahmat-Samii, “Advances in particle swarm opti-

mization for antenna designs: Real-number, binary, single-objective
and multiobjective implementations,” IEEE Transactions on Anten-
nas and Propagation, vol. 55, no. 3, pp. 556–567, 2007.
[63] K. Huang, H. Ye, J. Teng, S. P. Yeo, B. Luk’yanchuk, and C.-W. Qiu,

“Optimization-free superoscillatory lens using phase and amplitude
masks,” Laser & Photonics Reviews, vol. 8, no. 1, pp. 152–157, 2014.
[64] E. T. Rogers, S. Savo, J. Lindberg, T. Roy, M. R. Dennis, and N. I.

Zheludev, “Super-oscillatory optical needle,” Applied Physics Letters,
vol. 102, no. 3, p. 031108, 2013.
[65] J. J. Park, K. Lee, O. B. Wright, M. K. Jung, and S. H. Lee, “Gi-

ant acoustic concentration by extraordinary transmission in zero-mass
metamaterials,” Physical Review Letters, vol. 110, no. 24, p. 244302,
2013.
[66] R. Fleury and A. Alù, “Extraordinary sound transmission through

density-near-zero ultranarrow channels,” Physical Review Letters,
vol. 111, no. 5, p. 055501, 2013.
[67] J. Zhao, Z. N. Chen, B. Li, and C.-W. Qiu, “Acoustic cloaking by ex-
traordinary sound transmission,” Journal of Applied Physics, vol. 117,
no. 21, p. 214507, 2015.
[68] S. Tachi, “Telexistence and retro-reflective projection technology,”

in Proceedings of the 5th Virtual Reality International Conference
(VRIC2003), vol. 69, pp. 1–69, 2003.
Bibliography 111
[69] Z. Liu, X. Zhang, Y. Mao, Y. Zhu, Z. Yang, C. Chan, and P. Sheng,

“Locally resonant sonic materials,” Science, vol. 289, no. 5485,
pp. 1734–1736, 2000.
[70] L. Yong, L. Bin, Z. Xin-Ye, and C. Jian-Chun, “Broadband acoustic

transmission enhancement through a structured stiff plate with locally
resonant elements,” Chinese Physics Letters, vol. 29, no. 11, p. 114301,
2012.
[71] Z. Yang, J. Mei, M. Yang, N. Chan, and P. Sheng, “Membrane-type

acoustic metamaterial with negative dynamic mass,” Physical Review
Letters, vol. 101, no. 20, p. 204301, 2008.
[72] R. Fleury, D. L. Sounas, C. F. Sieck, M. R. Haberman, and A. Alù,

“Sound isolation and giant linear nonreciprocity in a compact acoustic
circulator,” Science, vol. 343, no. 6170, pp. 516–519, 2014.
[73] J. Andkjær and O. Sigmund, “Topology optimized cloak for airborne

sound,” Journal of Vibration and Acoustics, vol. 135, no. 4, p. 041011,
2013.
[74] V. Pogrebnyak, “Geometric resonance in a periodic waveguide,” Jour-

nal of Applied Physics, vol. 94, no. 10, pp. 6979–6981, 2003.
[75] W. H. Munk, “Abyssal recipes,” Deep Sea Research and Oceanographic

Abstracts, vol. 13, no. 4, pp. 707–730, 1966.
[76] C. Wunsch and R. Ferrari, “Vertical mixing, energy, and the general
circulation of the oceans,” Annual Review of Fluid Mechanics, vol. 36,
pp. 281–314, 2004.
Bibliography 112
[77] S. R. Jayne and L. C. S. Laurent, “Parameterizing tidal dissipation

over rough topography,” Geophysical Research Letters, vol. 28, no. 5,
pp. 811–814, 2001.
[78] C. Garrett and E. Kunze, “Internal tide generation in the deep ocean,”
Annual Review of Fluid Mechanics, vol. 39, pp. 57–87, 2007.
[79] M. R. House, “Orbital forcing timescales: an introduction,” Geological

Society, London, Special Publications, vol. 85, no. 1, pp. 1–18, 1995.
[80] M. A. Nobile and S. I. Hayek, “Acoustic propagation over an

impedance plane,” The Journal of the Acoustical Society of America,
vol. 78, no. 4, pp. 1325–1336, 1985.
[81] C. Chien and W. Soroka, “Sound propagation along an impedance

plane,” Journal of Sound and Vibration, vol. 43, no. 1, pp. 9–20, 1975.
[82] F. Ham and G. Iaccarino, Annual Research Briefs 2004, pp. 3–14.
Stanford, CA: Center for Turbulence Research, 2004.
[83] B. King, H. Zhang, and H. L. Swinney, “Tidal flow over three-

dimensional topography in a stratified fluid,” Physics of Fluids, vol. 21,
no. 11, p. 116601, 2009.
[84] A. Dettner, H. L. Swinney, and M. Paoletti, “Internal wave and bound-

ary current generation by tidal flow over topography,” Physics of Flu-
ids, vol. 25, no. 11, p. 116601, 2013.
[85] L. Zhang and H. L. Swinney, “Virtual seafloor reduces internal wave

generation by tidal flow,” Physical Review Letters, vol. 112, no. 10,
p. 104502, 2014.

Screw Compressors

Uploaded by

Copyright:

Available Formats

Screw Compressors

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Screw Compressors

Uploaded by

Copyright:

Available Formats

MANIPULATION OF SOUND

(B.Eng., Nanjing University)

FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

DEPARTMENT OF ELECTRICAL AND COMPUTER

NATIONAL UNIVERSITY OF SINGAPORE

Signed: Jiajun ZHAO

It is my great pleasure to thank everyone who has kindly helped me during

I acknowledge the support of the Presidential Fellowship I am granted with

Manipulation of sound properties has attracted the interest of acoustical

Diﬀerent from acoustic metamaterial which enables abnormal sound prop-

Besides, inasmuch as acoustic metamaterial is based on transformation

In this thesis, we elaborate the design of acoustic metasurface and metas-

List of Tables viii

List of Symbols xii

1.1 Reshaping wavefronts in optics and acoustics . . . . . . . . . 2

1.2 Vibrational direction in ﬂuids and solids . . . . . . . . . . . 3

1.3 Acoustic focusing and piezoelectric transducer . . . . . . . . 4

1.4 Invisibility cloaking in optics and acoustics . . . . . . . . . . 6

1.5 Outlines of Chapters 2 - 5 . . . . . . . . . . . . . . . . . . . 7

1.6 My published journal articles related to thesis . . . . . . . . 9

2 Manipulating acoustic wavefront with metasurface of inho-

2.1 Extraordinary reﬂection and ordinary reﬂection . . . . . . . 11

2.2 Continuous and discontinuous impedance . . . . . . . . . . . 14

2.3 Acoustic illusion and ipsilateral focusing . . . . . . . . . . . 22

2.4 Conversion from propagating to surface waves . . . . . . . . 27

2.6 Conclusion and discussion . . . . . . . . . . . . . . . . . . . 28

3 Redirecting acoustic waves out of the incident plane 30

3.1 Out-of-incident-plane vibration of ﬂuid particles . . . . . . . 31

3.2 Two-dimensional varying metasurface . . . . . . . . . . . . . 36

3.3 Three-dimensional control of extraordinary reﬂection . . . . 41

4 Manipulating acoustic focus with an active metasurface

4.1 Design of planar metasurface piezoelectric transducer . . . . 46

4.2 Generation of an acoustic focal needle . . . . . . . . . . . . . 51

4.3 Generation of acoustic far-ﬁeld multiple foci . . . . . . . . . 54

4.4 Acoustic super-oscillatory super resolution . . . . . . . . . . 58

5 Realizing acoustic cloaking and near-zero density with acous-

5.1 Metastructure for acoustic cloaking made by copper . . . . . 65

5.2 Spring-mass model and density-near-zero property . . . . . . 68

5.3 Coupling model and geometrical dependence . . . . . . . . . 72

5.4 Deploying cloaked area in free space and in waveguides . . . 75

5.5 Conclusion and discussion . . . . . . . . . . . . . . . . . . . 79

6.1 Manipulation of band gap for sound transmission . . . . . . 82

6.2 Manipulation of acoustic properties in stratiﬁed ﬂuids . . . . 83

A Detailed derivation of impedance-governed generalized Snell’s

B Distinction of IGSL in acoustics 94

C Parameters of metasurface piezoelectric transducer 97

C.1 Parameters of Lead Zirconate Titanate PZT-5H . . . . . . . 97

C.2 Optimized conﬁguration of ring pattern . . . . . . . . . . . . 98

D Details of the simulation for internal waves 100

C.1 Symmetric elasticity matrix (Pa) . . . . . . . . . . . . . . . 97

C.2 Relative permittivity εSr . . . . . . . . . . . . . . . . . . . . 98

C.3 Coupling matrix e (C/m2 ) . . . . . . . . . . . . . . . . . . . 98

C.4 Conﬁguration of 30 rings to generate a focal needle (mm) . . 98

C.5 Conﬁguration of 28 rings to generate multiple foci (mm) . . 99

C.6 Conﬁguration of 19 rings to generate super-oscillatory super

2.1 A ﬂat interface with inhomogeneous acoustic impedance . . 12

2.2 Ordinary, extraordinary and negative extraordinary reﬂection 16

2.3 Oblique incidence upon metasurface . . . . . . . . . . . . . . 17

2.4 Schematics of hard-sidewall tubes . . . . . . . . . . . . . . . 19