Synthesis Lectures on Wave Phenomena in the

Physical Sciences

Series Editor
Sanichiro Yoshida
Department of Chemistry and Physics, Southeastern Louisiana University,
Hammond, LA, USA

The aim of this series is to discuss the science of various waves. An emphasis is
laid on grasping the big picture of each subject without dealing formalism, and
yet understanding the practical aspects of the subject. To this end, mathematical
formulations are simplified as much as possible and applications to cutting edge
research are included.
Sanichiro Yoshida

Fundamentals of Acoustic Waves and Applications

Sanichiro Yoshida
Department of Chemistry and Physics, Southeastern Louisiana University,
Hammond, LA, USA

ISSN 2690-2346 e-ISSN 2690-2354

Synthesis Lectures on Wave Phenomena in the Physical Sciences
ISBN 978-3-031-48199-4 e-ISBN 978-3-031-48200-7
https://doi.org/10.1007/978-3-031-48200-7

© The Editor(s) (if applicable) and The Author(s), under exclusive license to
Springer Nature Switzerland AG 2024

This work is subject to copyright. All rights are solely and exclusively licensed
by the Publisher, whether the whole or part of the material is concerned,
specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and
transmission or information storage and retrieval, electronic adaptation,
computer software, or by similar or dissimilar methodology now known or
hereafter developed.

The use of general descriptive names, registered names, trademarks, service

marks, etc. in this publication does not imply, even in the absence of a specific
statement, that such names are exempt from the relevant protective laws and
regulations and therefore free for general use.

The publisher, the authors, and the editors are safe to assume that the advice and
information in this book are believed to be true and accurate at the date of
publication. Neither the publisher nor the authors or the editors give a warranty,
expressed or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral
with regard to jurisdictional claims in published maps and institutional
affiliations.

This Springer imprint is published by the registered company Springer Nature

Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham,
Switzerland
This book is dedicated to the memory of my late father who introduced me to the
joy of engineering, and my mother, Sonoko Yoshida, and my wife, Yuko Yoshida,
for filling my life with love, joy, and happiness.
Preface
This book discusses the fundamentals of acoustic wave phenomena. I use the
word “discuss” because I intend to describe each topic through discussions from
various angles rather than to “provide” the information about them. I believe we
can deepen our understanding of natural sciences by discussing subjects. I hope
the readers apply the concept gained from this book to other fields of science
and engineering. The target audience is undergraduate students majoring in
physics and engineering.
I wrote this book as part of Synthesis Lectures on Wave Phenomena in the
Physical Sciences, which aims to describe various waves in a short book. It is
impossible to cover many topics with this short-book concept. I selected a few
and dug them deeper, exploring their physical meanings. Mathematical
operations play significant roles in this process. Often we can sense physical
phenomena intuitively by digesting the associated mathematics. I tried my best
to break down mathematical operations into small pieces. This sometimes
makes equations long. Please bear with me while I explain a long equation step
by step.
Acoustic technologies have numerous applications in a variety of
engineering and scientific fields. Recent advancements in hardware and
software technologies facilitate the use of devices and analyses of the outcomes.
On the other hand, this convenience tends to hide the operation principles of
hardware and software modules. This tendency sometimes leads to the
inefficient use of a hardware device and the misinterpretation of data. A proper
understanding of the science behind the scene is essential. Upon writing this
book, I always bore this factor in mind.
The first two chapters discuss the basics of waves, where Chap. 1 focuses
on the mathematical aspect of it and Chap. 2 on the physical. Chapters 3 and 4
discuss acoustic waves propagating in air and solids, with audible frequency in
air and ultrasounds in solids in mind. Chapter 5 describes the operation
principles of typical acoustic transducers.
Finally, I would like to express appreciation to people. I owe much to my
high school and college teachers because my understanding originates from
what they taught me decades ago. I am grateful to my parents for allowing me
to receive such an excellent education. I thank my wife, Yuko Yoshida, for her
continuous support. Often I spent much time on weekends for this book project.
Sanichiro Yoshida
Hammond, LA, USA
August 2023
Contents
1 General Discussions of Waves
1.1 Wave as a Movement of an Oscillatory Pattern
1.1.1 Terms to Represent a Wave
1.1.2 Longitudinal and Transverse Waves
1.2 Mathematics of Waves
1.2.1 Wave Function as a Periodic Function of Time and
Space
1.2.2 Wave Equations
1.2.3 Wave Solutions
References
2 Wave Dynamics
2.1 Oscillations and Waves
2.1.1 Harmonic Oscillations
2.1.2 Equation of Motion of Point Mass-Spring System
2.1.3 Oscillation to Waves
2.1.4 Wave Equations
2.2 Acoustic Wave Equations and Solutions
2.2.1 Acoustic Wave in Air
2.2.2 Longitudinal and Transverse Vibrations in Solid
2.2.3 Decaying Acoustic Wave Equations and Solutions
2.2.4 Nonlinear Wave Equations
2.3 Wave as a Flow of Energy
2.3.1 Acoustic Energy
2.3.2 Intensity of Acoustic Wave
2.4 Acoustic Wave Reflection and Transmission
2.4.1 Laws of Reflection and Refraction
2.4.2 Behavior of Acoustic Waves Near Boundary
References
3 Propagation of Acoustic Waves in Air
3.1 Dynamics of Air Particles and Acoustic Wave Equation in Air
3.1.1 Particle Motions in Longitudinal Waves
3.1.2 Wave-Generating Mechanism in Air
3.1.3 Decaying Pressure Waves in Air
3.2 Sound We Hear
3.2.1 Our Ear
3.2.2 Sounds of Speech and Music
3.3 Transmission of Audible Sound Waves
References
4 Propagation of Acoustic Waves in Solids
4.1 Acoustic Wave as Solutions to the Equation of Motion
4.1.1 Equation of Motion
4.2 Propagation of Acoustic Waves in Solid Under Various
Conditions
4.2.1 Plane Waves in an Elastic Medium
4.2.2 Plane Waves in a Medium with Free Surface
4.2.3 Surface Waves
4.2.4 Waves Propagating in Plates
References
5 Electrical-Mechanical Transduction
5.1 General Argument
5.2 Reciprocal Transduction
5.2.1 Electrostatic Transducer
5.2.2 Electrical Governing Equation of Electrostatic
Transducer
5.2.3 Mechanical Governing Equation of Electrostatic
Transducer
5.2.4 Physics Behind Electrostatic Transducer’s Operation
5.3 Anti-reciprocal Transducers
 5.3.1 Moving-Coil Transducer
References
Appendix A: Direction Cosines
Appendix B: Gradient Operation
Appendix C: Orthogonality of Sine and Cosine Functions
Appendix D: Notes on Standing Waves
Appendix E: Generalized Hooke’s Law and Equation of Motion for
Isotropic Media
Index
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024
S. YoshidaFundamentals of Acoustic Waves and ApplicationsSynthesis Lectures on Wave Phenomena in
the Physical Sciences
https://doi.org/10.1007/978-3-031-48200-7_1

1. General Discussions of Waves

Sanichiro Yoshida1
(1)
Department of Chemistry and Physics, Southeastern Louisiana
University, Hammond, LA, USA

Sanichiro Yoshida
Email: [email protected]

This introductory chapter discusses wave characteristics and related

terminologies in general. We first view waves as a motion of a spatial pattern (a
snapshot) of an entity on the time axis and as a temporal oscillation on the
spatial axis. This spatiotemporal behavior leads to the concept of phase velocity,
which characterizes the medium that the wave propagates through. In the
second half of the chapter, we discuss some mathematical aspects of wave
dynamics. We derive wave equations from the spatiotemporal characteristics
and discuss their solutions in general.
1.1 Wave as a Movement of an Oscillatory Pattern
Consider a sample wave in Fig. 1.1. This figure illustrates a wave pattern spread
along the x-axis at three different times, 0, 1, and 2 s. The wavy pattern at
each time step is a snapshot of the wave. With the elapse of time (shown by the
arrows in Fig. 1.1), the pattern moves to the right. The little circle put at the
leftmost trough indicates that each wave pattern moves rightward in the positive
x-direction with time. The solid line connecting the little circles shows that it
has an angle to the time axis. Here the dashed line is drawn parallel to the time
axis. The angle made by the solid and dashed lines clarifies the rightward
movement of the circle with the passage of time. We say that the wave has a
rightward velocity. This type of wave velocity is referred to as phase velocity
for the reason clarified shortly in this chapter.
Fig. 1.1 Schematic illustration of a wave as a function of space. Each wave pattern illustrates the snapshot
taken at 0, 1, and 2 s

1.1.1 Terms to Represent a Wave

Wavelength; spatial oscillatory periodicity
Figure 1.2 illustrates this rightward motion of the wave pattern more explicitly.
Notice that the pattern has a spatial periodicity. As the dashed box indicates, the
periodicity in this example is 10 mm in length. The spatial period of a wave is
referred to as wavelength. We can say that this wave has a wavelength of 10
mm. From the analysis of the motion of the periodicity and how long it takes in
time, we can calculate the phase velocity. In this example, the periodicity moves
0.5 mm in 1 s, and we find that the phase velocity is 500 m/s (0.5 mm
s 500 m/s).
Period of wave; temporal oscillatory periodicity
Observation of Fig. 1.2 at the same location on the x-axis reveals the
temporal oscillation of the wave. For instance, see the signal at mm
indicated by a vertical line in Fig. 1.2. The signal at this location is positive at
s, approximately 0 at s, and negative at s. It
is clear that the medium oscillates during this duration of 2 s. This oscillation
is called particle oscillation and the corresponding velocity is called the particle
velocity.
Figure 1.3 illustrates the temporal oscillation (the particle oscillation)
behavior of the same wave as Fig. 1.1 in the same form of three-dimensional
presentation. Here, the oscillations at three representative locations of 0,
1, and 2 mm are presented. You may notice the similarity in the oscillatory
pattern between Figs. 1.1 and 1.3. In fact, as Fig. 1.4 explicitly shows, the
oscillatory patterns as a function of space and time are identical to each other.
The temporal periodicity observed in Fig. 1.4 is referred to as period.

Fig. 1.2 Schematic illustration of a wave as a function of space emphasizing the motion of the spatial
periodicity
Fig. 1.3 Schematic illustration of a wave as a function of time. Each wave pattern illustrates the particle
oscillation at a fixed x location
Fig. 1.4 Schematic illustration of a wave as a function of time. Each wavy pattern is a snapshot of the
wave at the corresponding time

1.1.2 Longitudinal and Transverse Waves

Longitudinal waves
Figures 1.1, 1.2, 1.3 and 1.4 illustrate the oscillatory behavior of a wave as
a function of space and time. However, they do not indicate the direction of
particle oscillations. It can be parallel or perpendicular to the direction of the
wave propagation. We call a wave a longitudinal wave when the particles
oscillate parallel to the direction of the wave propagation. A wave traveling
through a sling is an example of a longitudinal wave. A sound wave traveling
through air is also a longitudinal wave. (See Sects. 2.2.1 and 3.1.)
Figure 1.5 illustrates the propagation of a sample longitudinal wave. The
regions where the dots are dense (sparse) represent the portion of the medium
where the density is high (low). The illustrations at each row present the pattern
at a given time, where the time elapses from the top towards the bottom row
with an increment of 0.25 s. Notice that the illustrations on the left column
indicate that the dens-sparse pattern moves to the right, whereas those on the
right column represent that it moves to the left. Since the direction of the wave
motion is parallel to the particle motions, these waves are longitudinal.

Fig. 1.5 Propagation of longitudinal waves. The wave moves parallel to the dens-sparse patterns of the
particles’ motions

Transverse waves
If the particle oscillation is perpendicular to the wave propagation, we call
the wave a transverse wave. A wave formed by fans in a soccer stadium is a
transverse wave. Figure 1.6 illustrates the propagation of a transverse wave in
the same fashion as Fig. 1.5. The dots represent particles. The rows represent
the progression in time, whereas the left and right columns illustrate the waves
moving rightward and leftward, respectively. In this case, the oscillatory pattern
does not exhibit a change in the density. Instead, it indicates a swinging motion
of the particles in the perpendicular direction to the wave motion. A sound wave
can travel through a solid as a longitudinal or transverse wave. (See Chap. 4.)
An ocean wave is a mixture of longitudinal and transverse waves [1]. The
particle oscillation (the motion of the water molecules) has longitudinal and
transverse components to the propagation.
Fig. 1.6 Propagation of transverse waves. The waves move horizontally whereas the particles oscillate
vertically

1.2 Mathematics of Waves

1.2.1 Wave Function as a Periodic Function of Time and Space
The fact that a wave has an identical pattern in its spatial and temporal
characteristics allows us to put a wave function in the following form.

(1.1)
Here, t is the time coordinate variable, x is the spatial coordinate variable, and

(1.2)
is called the phase of the wave. For the physical meaning of and k, see the
following paragraphs.
Since is a periodic function, we obtain the following relation.
(1.3)
Here is the period of the function and N is an integer. Equation
(1.3) literally indicates the fact that every time the phase changes by the
function takes the same value. We can always find the same change in by
fixing t and varying x or fixing x and varying t. This fact stipulates the above
observation that a wave has identical temporal and spatial patterns.
We can interpret the quantities and k as the temporal and spatial
frequencies of the wave by making the following argument. Find the total
derivative of by differentiating (1.2).

(1.4)
Consider the spatial periodicity using (1.4) in association with Fig. 1.2.
Since the spatial wave patterns in this figure are the snapshots taken at the
respective time, we can put in (1.4). Under this condition, express
(1.4) for one period. We can put the left-hand side as the phase period
and the right-hand side as the wavelength . Thus,

(1.5)
Since period is a constant for a given wave, (1.5) indicates that k is
proportional to the reciprocal of . Here the wavelength represents the
periodicity in length (m in the SI unit). The reciprocal of measures the
number of wavelengths in the unit length, i.e., how frequently the wavelength
appears as we move over the unit length on the spatial axis. Thus, it is called the
spatial frequency of the wave. In the case of sinusoidal waves (as we will
discuss in more detail in later sections), the periodicity in phase is . In this
case, the spatial frequency becomes as follows.

(1.6)
We can repeat the same type of argument on (1.4) in association with Fig.
1.4. In this case, and we obtain the following equation as the
temporal version of (1.5).

(1.7)
Here is the period of the wave.
Similarly to (1.6), we can define the temporal frequency for a
sinusoidal wave.

(1.8)
Repeating the same argument as for k, we can say that represents how
frequently the period appears as we move along the time axis.
To distinguish from the terms the spatial and temporal frequency in the
unit of 1/m or 1/s, we call k and the angular spatial frequency and angular
temporal frequency. They are in rad/m and rad/s, respectively.
Traveling wave and phase velocity
As indicated by (1.3), the phase determines the value of a wave function.
The motion of a wave is the trajectory of a constant phase value. In other words,
the wave velocity we discussed above is the velocity of a phase value; that is
why we call the velocity phase velocity. A surfer stays at a crest of an ocean
wave where the phase makes the vertical position of the water molecule at the
highest. So, we can say that the surfer moves at the phase velocity of the wave
he is on. Here we consider the concept of phase velocity quantitatively.
When we move along with a constant phase, the phase does not change
over time. Mathematically, we can express this situation by saying
. We can discuss it using (1.4) by differentiating the phase
with respect to time and setting it to zero.

(1.9)
From (1.9), we obtain the following equation.

(1.10)
 Consider the meaning of dx and dt that appear on the right-hand side of
(1.10). These are the changes in the coordinate variable x and t necessary to
keep the phase constant. In other words, if the time elapses
by dt, we need to change the location on the x coordinate by dx to keep the
phase unchanged. From this viewpoint, we can interpret that dx/dt is the
velocity of the constant phase location on the x-axis. Thus, we can say that the
quantity expressed by the left-hand side of (1.10) represents the phase velocity
of the wave .

(1.11)
Note that the negative (positive) sign on the right-hand side of this
equation corresponds to the positive (negative) sign on the right-hand side of
the phase expression (1.2). This means that corresponds to
, i.e., a negative phase velocity for positive dx and dt,
indicating that the wave travels in the negative x direction, and
represents a wave traveling in the positive x direction. In
other words, to keep the phase constant when the time
elapses by dt (i.e., t increases), we need to increase x by .
A wave expressed in the form of (1.1) is referred to as a traveling wave. As
clear from the above discussion, when the sign between the time and space
terms of the phase is negative ( ), the wave travels in the
positive x-direction. If the sign is negative, the wave travels in the negative x-
direction. According to (1.10) and (1.11), we can distinguish the direction of the
wave motion by the sign of the phase velocity .

1.2.2 Wave Equations

Consider the property of the wave expression (1.1) further by differentiating
the wave function f with respect to time and space.
(1.12)

(1.13)
Here denotes the second-order derivative of the function f(t, x), e.g., if
, . Equating
appearing in (1.12) and (1.13), we obtain the following wave equation.

(1.14)
Note that appearing on the right-hand side is the phase velocity (1.11).
Expressing the unit of the quantity represented by f with [f], we find that the

unit of the left-hand side of (1.14) is [f]/s and the unit of is [f]/m
, indicating that is in m/s. Indeed, it has the dimension of velocity.
We can express the wave equation using (1.11) in (1.14).

(1.15)
Extension to three dimensions
It is clear that the differential equation (1.14) yields the wave solution
. We call this equation a wave equation. As discussed in the
preceding section, (1.14) yields a wave solution that travels along the x-axis.
Since its traveling direction is limited in one axis, this type of wave is known as
a one-dimensional wave.
What if the traveling direction is not in line with the x-axis, e.g., it has
some angle to the x-axis? Even if the traveling behavior is physically the same,
we cannot express the phase term in the form of . With the passage
of time, the change in the spatial coordinate account for the same phase cannot
be expressed only by the change in x. In this situation, we say that the traveling
wave is three-dimensional, and call the spatial frequency k the propagation
constant. Here, the propagation constant has a directionality, i.e., we need to
move along an axis that has an angle to the x-axis to define the spatial
frequency. In other words, the propagation constant k is a vector referred to as
the propagation vector.
Figure 1.7 illustrates the situation where the same physical wave travels
through space represented by two different coordinate systems. In Fig. 1.7a, the
x-axis is in line with the direction of the wave propagation, hence the
propagation vector has one component. In Fig. 1.7b the direction of
propagation has angles of , , and with the x, y and z-axis,
respectively. The cosines of these angles are called direction cosines [2]. The
propagation vector’s x, y and z-components are the vector’s magnitude
multiplied by the corresponding directional cosine, e.g., .

Fig. 1.7 The same wave propagates through space represented by two different coordinate systems. a The

propagation is in line with the x-axis. b The propagation has angles of , , and with the x, y and

z-axis. is the unit vector along the x-axis. is the unit vector along the wave’s propagation

The conversion from a one-dimensional to a three-dimensional wave

expression leads to the following change in the phase term of wave function
defined in (1.1).

(1.16)
 Here is the vector that represents the spatial coordinates; the three-
dimensional version of x in (1.1).

(1.17)
Here etc. are the unit vectors for the corresponding directions. The
scalar product in the phase term of (1.16) represents the spatial phase change in
the direction of propagation vector , i.e., along the propagation path of the
wave.

(1.18)

Here k is the magnitude of , i.e., .

Defining vector as the unit vector in the direction of , we can express
(2.89) in a compact form as follows.

(1.19)
Here etc. are the components of .
Comparison of (2.89) and (1.19) reveals that has direction cosines as its
components.

(1.20)
In Appendix A we prove that is a unit vector.
Using (1.19) or (1.20), we can express wave function (1.16)

(1.21)
The extension to three dimensions requires one more procedural change
regarding differentiation. Differentiating a function is the process to find the
change in the value of the function associated with an infinitesimal change of
independent variables. When the wave function has one spatial independent
variable, like the above case of f(t, x), the change in the value of function f is
definitely determined by the function’s dependence on x, i.e.,
. However, if the function has three independent spatial variables like
f(t, x, y, z), the spatial derivatives become a vector quantity. We need to replace
the differential operation df/dx with the gradient operation [3]. In
Cartesian coordinates, it is given as follows.

(1.22)
Appendix B discusses some more description of the gradient operation.
By repeating the gradient operation twice, we can derive the following
expression as the three-dimensional version of the second-order spatial

differentiation .

(1.23)
The operator is called the Laplace operator or Laplacian.
Using (1.21) in (1.23) we find the following expression as the three-
dimensional version of (1.13).

(1.24)

Here is the unit vector in the direction of the

propagation vector, and we used that on the right-

hand side of (1.24).
From (1.12) and (1.24), we derive the following equation.
(1.25)
Wave equation (1.25) is the three-dimensional version of wave equation
(1.14).
It should be noted that above we derived wave equations based on the
temporal and spatial periodicities. As will be discussed in the following section,
this type of wave equation yields solutions that continue unlimitedly spatially
and temporally. It is because the periodic function we started with has such a
property. In reality, waves decay due to various causes such as viscosity. We
will discuss this topic in the next chapter by considering the physics behind
wave dynamics.

1.2.3 Wave Solutions

Consider solving the wave equations derived above. Since we derived the wave
equations by considering a periodic function in the form of (1.1), the solution
should take this form. As a simple example, we know that sine and cosine
functions are periodic. Let’s take a moment and confirm that a simple cosine
function in the following form satisfies the one-dimensional wave equation
(1.14).

(1.26)
Here A, a, and b are constant.
Differentiating this function with respect to time and space twice, we
obtain the following set of equations.

(1.27)

(1.28)
We find that with the following conditions, (1.26) satisfies wave equation
(1.14).

(1.29)
(1.30)
Adding a constant phase term does not affect the temporal or spatial
differentiation. Thus, we find the following function is a solution to the wave
equation (1.14).

(1.31)
Of course, we can extend solution (1.31) into three dimensions.

(1.32)
The freedom to add a phase indicates that a sine function of the same
temporal and spatial frequency can also be a solution.
Above, we casually used sinusoidal functions as examples. However, this
discussion does not lose generality if we remember Fourier’s theorem that state
“a periodic function can be expanded into a series of cosine and sine functions.”
In the next section, we briefly discuss this topic.
Fourier series
According to Fourier’s theorem [4, 5], the temporal periodicity of a wave
function f(t) can be expanded into a series of cosine and sine functions as
follows.

(1.33)

(1.34)
Here is the fundamental frequency, , , are the
second, third, harmonics. For a given function f(t) we can find the
coefficients using the orthogonality of the sine and cosine
functions. See Appendix C.
Consider these harmonics in the context of wave solution (1.31). We
discussed that a cosine sine function in the form of (1.31) satisfies the wave
equation (1.15). This wave equation indicates that the secondary temporal
differentiation of a wave function is proportional to the secondary spatial
differentiation, where the constant of proportionality is the square of the phase
velocity. As will be discussed later, the phase velocity is a material constant.
Each material has its unique value of acoustic phase velocity. We also learned
that phase velocity is the ratio of temporal frequency over spatial frequency (see
(1.11)). These facts lead to the following discussion. When an acoustic wave
travels through a material at the frequency determined by the source, the phase
velocity determines the spatial frequency, hence the wavelength. The
combination of and k in solution (1.31) satisfies this relation, i.e., its ratio is
the material’s unique phase velocity.
The above situation indicates that if an acoustic wave at the fundamental
frequency in the Fourier series (1.33) satisfies the wave equation, a
harmonic of frequency is also a solution to the same wave equation
where the spatial frequency is , i.e., N times greater than the
fundamental wave. Since the wavelength of the fundamental frequency is

, the wavelength of this harmonic is N times shorter. This

situation is the same for any N.
The above discussion indicates that the same physical system (the acoustic
source and the medium through which the acoustic wave travels) can generate
harmonics. Shortly, we will discuss the phenomenon known as resonance,
which is an example where the same physical system generates harmonics
(higher resonator modes).
Standing wave
In Sect. 1.2.1, we discussed that a wave function travels
in the positive x direction and travels in the negative x
direction. Call the former a forward-going wave and the latter a backward-going
wave. Since these two waves have the same phase velocity, , the
same physical system can generate both waves at the same time, raising the
following questions: “Can a pair of forward-going and backward-going waves
coexist in a medium? If so, do they add each other, and if that is the case, what
is the direction of the combined wave?” The answer is “Yes, they can coexist
and add to each other. The combined wave can be stationary. This type of
stationary wave is called a standing wave.” I say “can be stationary” because,
under some conditions, the combined wave can travel. We will discuss such
conditions in Appendix D. Here, we focus our discussions on standing waves.
 Before starting the discussions here, I would like to note some physical
significance of standing waves. As will be discussed later, a wave carries
energy. Therefore, if a wave stays at the same location, the energy carried by the
wave does not flow and thereby the local energy grows in time. Indeed this
phenomenon occurs under the condition of resonance. We can construct an
instrument to embody such a resonance phenomenon, and the instrument is
called a resonator [6]. Most musical instruments [7] (see Sect. 3.2.2) are
resonators [8] to generate standing waves [9].
Consider the formation of a standing wave by adding two oppositely
traveling waves. Assume the two waves have the same amplitude A.

(1.35)
Equation (1.35) indicates that when the forward-going and backward-going
waves have the same amplitude, the superposed wave oscillates without
traveling in either direction (forward or backward). The time function part (
) of the right-hand side of (1.35) tells us every time the time t makes
(N is an integer), the wave function becomes zero over the entire
x. This situation is referred to as complete destructive interference. The forward-
going and backward-going waves (call these waves component waves) interfere
with each other so that the superposed wave becomes zero at all x locations.
When time t satisfies , on the other hand,
takes and thereby the component waves interfere with each
other constructively, maximizing the amplitude of the superposed wave. The
“2” in front of A on the right-hand side of (1.35) indicates that under the
condition of complete constructive interference, the resultant amplitude is
doubled as compared with the component waves. Note that destructive and
constructive interference each repeats at a period of (an increase of 1 in the
integer N changes the phase by ). This period is half of that of the component
waves.
Fig. 1.8 An example of a standing wave

Figure 1.8 is an example of a wave expressed in the form of Eq. (1.35).

Here the left column illustrates oppositely traveling component waves and the
right column the standing wave resulting from the superposition of the
component waves. Each row represents time at every one-eighth of the period
. It is seen that the first complete constructive interference occurs at
, i.e., a quarter period after when complete destructive
interference is seen. At , i.e., half a period later the second
complete destructive interference occurs. We can easily imagine that the second
complete constructive interference occurs a quarter period later, i.e., at
. As mentioned in the preceding
paragraph, the completely destructive and constructive interference each repeats
every half period or in phase.
 The left column illustrates that the component waves are completely out of
phase under complete destructive interference. Under complete constructive
interference, they are completely in phase. The complete destructive and
constructive interferences alternate at every quarter period ( ),
meaning that destructive (constructive) interference occurs in the middle of
consecutive constructive (destructive) interferences.
In the above sections, we assumed that the amplitude A is a constant. This
assumption means that the wave has the same strength over the plane
perpendicular to the x-axis (the axis along which the component waves travel).
Since every wave carries energy, this in turn indicates that the energy is infinite.
Such a wave is unrealistic, but we often use this type of solution as it exhibits
important characteristics of waves. We call a wave with a constant amplitude a
plane wave.
The standing wave solution (1.35) has the form of the product of a
temporal function and spatial function. This reminds us of the variable
separation method. Let’s solve the wave equation (1.15) using the variable
separation method [10]. Put the solution as the product of time function T(t) and
space function X(x) as follows.

(1.36)
Substitution of (1.36) it into the wave equation (1.15) yields the following
set of equations.

(1.37)

(1.38)

(1.39)
Divide (1.39) by TX.
(1.40)

Here is a function of time independent of x, and is a

function of x independent of t. (1.39) says these two terms are equal to each
other. This means that both terms must be a constant. Let represent that
constant and put (1.40) in the following form.

(1.41)

(1.42)
Put T(t) in the following form.

(1.43)
Differentiate T(t) twice with respect to time.

(1.44)
From (1.44), we can easily find that (1.43) satisfies (1.41).
Similarly, we find the following form of X(x) satisfies (1.42).

(1.45)
Thus, by substituting (1.43) and (1.45) into (1.36), we find that the wave
function takes the following form.
(1.46)
We now in a position to determine amplitudes , , , and ,
and the constant c. The values of these quantities depend on the initial and
boundary conditions.
Initial condition
In Fig. 1.8, we set the time origin when the component waves
interfere completely destructively. It is clear that we can set the time origin
freely by adjusting constant phase in (1.31). Here we use the same initial
condition as Fig. 1.8, i.e., at , the standing wave is null. Substituting
into (1.46) and setting the value of the function zero, we find
. So, now the wave function takes the following form.

(1.47)
Boundary condition 1
Use the following boundary conditions for X(x). This condition represents
the situation where the standing wave is null at and .

(1.48)
Substituting the first condition of (1.48) into the spatial function in (1.47),
we obtain the following equation.

(1.49)
Thus, we find . The second condition of (1.48) and (1.57) yield
the following equation.
(1.50)
It follows that the following condition holds.

(1.51)
where n is an integer.
Since the constant c is in the form of in the time function T(t),
we can interpret this constant as the angular frequency of the wave. Call it the
eigen frequency . From (1.51) it follows that satisfies the following
condition.

(1.52)
Note that the eigen frequency is derived from the boundary condition
(1.48), which indicates that the standing wave has a spatial periodicity whose
integer multiple is equal to L. The physical meaning of this statement becomes
clear shortly.
Now substitute (1.52) into the spatial function X(x) using the above-found
condition .

(1.53)
Here (1.52) is used in going through the last equal sign.
So in this case the standing wave function takes the following form.

(1.54)
Here .
Figure 1.9a illustrates standing waves under Boundary condition 1 in a
scenario where a pair of reflectors generate the forward-going and backward-
going waves that form the standing waves. The reflection at each of the
reflectors is referred to as fixed-end reflection.

Fig. 1.9 Phase condition for a passive resonator. In the case of an active resonator, a source is placed
inside the resonator

Boundary condition 2
Now consider the following boundary condition instead of (1.48) using the
same initial condition . This condition represents the situation
where the standing wave has crests or troughs at and .

(1.55)
Substituting the first condition of (1.55) into the spatial function in (1.47),
we obtain the following equation.

(1.56)
 Thus, we find . The second condition of (1.55) and (1.56)
yields the following equation.

(1.57)
It follows that the following conditions hold.

(1.58)
where n is an integer. Using (1.58) , we find the spatial function X(x) in
this case as follows.

(1.59)
Here whether the standing wave is at a crest or trough in the boundary
condition (1.55) depends on the integer n; when it is odd, ,
and when it is even .
Repeating the same procedure as above using the first condition of (1.58),
we find the same eigenfrequency as Boundary condition 1. Thus, in this
case, the standing wave function takes the following form.

(1.60)
Here .
Figure 1.9b illustrates standing waves under Boundary condition 2 in a
scenario where a pair of reflectors generate the forward-going and backward-
going waves that form the standing waves. The reflection at each of the
reflectors is referred to as open-end reflection.
Boundary condition 3
 The last boundary condition considered here represents the situation where
the standing wave is null at and a crest or trough at .

(1.61)
Here represents the peak amplitude when the time function
. Substituting the first condition of (1.61) into the spatial
function term X(x) in (1.47) and obtain the following equation.

(1.62)
It follows that . Substituting this condition and the second
condition of (1.61), we obtain the following equation.

(1.63)
It follows that

(1.64)
Condition (1.64) leads to the following expressions of the wave under
Boundary condition 3.

(1.65)
where the eigen frequency is as follows

(1.66)
 Figure 1.9c illustrates standing waves under Boundary condition 3 in a
scenario where a pair of reflectors generate the forward-going and backward-
going waves that form the standing waves. The reflection at the left reflector is
open-end reflection and that at the right reflector is fixed-end reflection.
Resonance of wave
The situation depicted in Fig. 1.9 is associated with a phenomenon called
Resonance [8]. Imagine that a longitudinal wave traveling in air is incident to a
medium, e.g., a tube of length L. Consider that the frequency of the incident
wave is equal to one of the eigen frequencies discussed in the preceding section.
We can easily imagine that one of the situations depicted in Fig. 1.9 is
established. Due to the difference in acoustic impedance (see Sect. 2.3.1), the
incident wave experiences partial reflection at the left end of the tube. Similarly,
reflection occurs at the right end of the tube. The wave reflected at the right end
travels back toward the left end, and there it experiences reflection. In this
fashion, reflected waves go back and forth in the tube.
It is instructive to consider the resonant condition for each of the three
boundary conditions discussed in the preceding section. First, rewrite (1.47)
using the wavenumber or wavelength. Since c represents the angular frequency
and the phase velocity is given as , we can express
with the wavelength and thereby rewrite the spatial function term as
follows.

(1.67)

(1.68)
Using and (1.51) we find the following equality
under Boundary conditions 1 and 2.

(1.69)
Solving (1.69) for or L we find the following equation.
(1.70)
Similarly from (1.64) under Boundary condition 3 we find as follows.

(1.71)

(1.72)
Using (1.68), the initial condition , and other conditions
used in the above section we find the following expressions of standing waves
under the three Boundary conditions.
Boundary condition 1

(1.73)
Boundary condition 2

(1.74)
Boundary condition 3
(1.75)
The integer n represents the order of harmonics. Sometimes this number is
called the resonator mode number. As will be discussed in Sect. 1.2.2, the
phase velocity is determined by the medium. The eigen frequency
expressions (1.71)–(1.75) indicate that for a given resonator length L and a
medium inside it, the eigen frequency increases with the mode number n. The
frequency determines the oscillatory cycle of particles. We can easily imagine
that the higher the frequency the greater the kinetic energy of the particles as
they move faster. Generation of higher-order modes usually requires higher
input energy [11]. For example, if you apply an acoustic transducer to a steel
structure and monitor the resultant frequency with an acoustic sensor, higher-
order modes are observed as the transducer’s energy is increased.

References

1. 1. Komen GJ, Cavaleri L, Donelan M, Hasselmann K, Hasselmann S, Janssen PAEM (1994)

Dynamics and modeling of ocean waves, Ch. 1. Cambridge University Press, pp 1–68
2. 2. Roithmayr CM, Hodges DH (2016) Dynamics: theory and application of Kane’s method,
Append. I. Cambridge University Press
3. 3. Harper PG, Weaire DL (2009) Introduction to physical mathematics, Ch. 28. Cambridge
University Press
4. 4. Prof. Brad Osgood (2014) Lecture notes for EE 261 the Fourier transform and its applications.
CreateSpace Independent Publishing Platform
5. 5. Bracewell RN (1999) The Fourier transform and its applications, 3rd edn. McGraw-Hill, Boston,
New York
6. 6. Rienstra SW, Hirschberg A (2023) An introduction to acoustics. https://www.win.tue.nl/~sjoerdr/
papers/boek.pdf (accessed on July 17, 2023)
7. 7. Fletcher NH, Rossing TD (1998) The physics of musical instruments, 2nd edn. Springer Link
8. 8. Standing waves and resonance in university physics vol. 1 (British Columbia/Yukon open
authoring platform) https://pressbooks.bccampus.ca/universityphysicssandboxbook1/chapter/
standing-waves-and-resonance/ (accessed on July 17, 2023)
9. 9. King GC (2009) Vibrations and waves. Wiley, Chichester, UK, pp 137–158
10. 10. Borden B, Luscombe J (2017) Essential mathematics for the physical sciences, vol. 1
homogeneous boundary value problems, Fourier methods, and special functions, Ch. 2 Separation
of variables. Morgan and Claypool, San Rafael, CA, USA
11. 11. https://courses.lumenlearning.com/suny-osuniversityphysics/chapter/16-4-energy-and-power-
of-a-wave/
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024
S. YoshidaFundamentals of Acoustic Waves and ApplicationsSynthesis Lectures on Wave Phenomena in
the Physical Sciences
https://doi.org/10.1007/978-3-031-48200-7_2

2. Wave Dynamics
Sanichiro Yoshida1
(1)
Department of Chemistry and Physics, Southeastern Louisiana
University, Hammond, LA, USA

Sanichiro Yoshida
Email: [email protected]

In this chapter, we focus on the physics behind wave dynamics. After reviewing
the harmonic oscillation of point mass-spring systems, we discuss how
oscillatory motions transition into a wave. We derive wave equations from the
equation of motion for oscillatory dynamics in air and solids. Considering some
solutions to the wave equation, we discuss the propagation of waves as a flow
of energy, along with related topics such as acoustic impedance and acoustic
wave intensity. Finally, after briefly reviewing the laws of reflection and
refraction, we discuss the propagation of acoustic waves through the boundary
of media having different elastic properties.
2.1 Oscillations and Waves
A vibrating object generates sounds. A tuning fork [1] (used by a musician)
generates sound because it vibrates at a specific frequency. The physical
mechanism underlying this vibration is the elasticity of the material of the
tuning fork. If you hit a tuning fork made of steel against something rigid (like
an edge of a table), the steel atoms near the contact with the table edge shift
from their equilibrium positions. Due to the elasticity of steel, these shifted
atoms return to the equilibrium locations after they reach the farthest point. Due
to inertia, they pass the equilibrium location and swing in the opposite direction
from the initial shifts. This motion pushes the atoms of the neighboring lattice,
shifting them from their equilibrium locations. In this fashion, the vibratory
motion transfers through lattices reaching the other surface of the tuning fork.
Due to the difference in acoustic impedance from the air, the vibratory motion
reflects toward the initial surface hit by the table edge. This process is an
example of the resonant phenomenon associated with the reflection of the open-
open ends discussed in the last chapter. The size (corresponding to L in Fig. 1.
9b) and the phase velocity of steel (see (1.74)) determine the frequency (the
eigen frequency) of the fork.
The sound you hear in this fashion has the following properties. If you hit
the fork more strongly you will hear the sound louder. However, the pitch of the
sound will most likely be the same. Here, I say “most likely” because
depending on how strongly you hit the fork, harmonics can be heard [2]. If you
hit another tuning fork of a different dimension, you will hear a sound of a
different pitch from the first tuning fork. It is because the second fork is
different in material and/or size from the first fork. In other words, eigen
frequency is different. This is how you tune a musical instrument using tuning
forks. In this section, we explore the physics underlying these phenomena.

2.1.1 Harmonic Oscillations

In Chap. 1 we discussed that a wave is a spatiotemporal oscillation. The
temporal oscillation discussed in Sect. 1.2 indicates that a recovery force acts on
the particle, making it oscillate about an equilibrium position. We can call this
type of recovery force an elastic force and the resultant oscillation is a
harmonic oscillation [3].
A harmonic oscillation occurs generally in two modes. In the first mode,
the system oscillates after the external disturbance is removed. The above
example of a tuning fork falls in this category. This type of oscillation is
referred to as unforced oscillation [4, 5]. In the second mode, the external agent
keeps applying an oscillatory force and the system oscillates in response to the
external force. This type of oscillation is referred to as forced oscillation [4,
6]. Unforced and forced oscillations occur when the system exerts an elastic
force. (Forced oscillation can occur without an elastic force but in that case, the
system simply follows the oscillatory motion provided by the external agent.
This type of oscillation is not a harmonic oscillation and will not be discussed.)
The elastic force is defined as the force whose magnitude is proportional to
the displacement from the neutral (equilibrium) position and the direction is
opposite to the displacement. Naturally, the neutral position is where the elastic
force is null. When the oscillating entity (called the particle) comes back to the
neutral position, it passes there due to inertia. Consequently, the particle goes to
the other side of the neutral position. As soon as the particle passes the neutral
position, the elastic force changes its direction because it is always toward the
neutral position. We can say that the acceleration of the particle is always
toward the neutral position (according to Newton’s second law). Thus, the
acceleration is opposite to the displacement of the particle.

Point mass-spring system

Spring-mass systems [7–9] are simple yet informative to understand harmonic
oscillations. Consider the above dynamics using a system of a point mass
connected to a horizontally placed spring. In this case, the spring exerts the
elastic force that causes the oscillation of point mass and the point mass is what
we called the particle in the above paragraph. Figure 2.1 illustrates a sample
spring-mass system. Here the object connected to the right end of the spring is
considered to be a point mass. The middle illustration represents the moment
when the point mass passes through the neutral point from the right. The top
and bottom illustrations represent the situation before and after the mass passes
through the neutral point.

Fig. 2.1 Spring oscillation. The mass (particle) is on the right side of the neutral point (top), at the neutral
point (middle) and on the left side of the neutral point (bottom). The thicker arrow represents the spring’s
elastic force and the thinner arrow represents the displacement from the neutral position

The arrows represent the elastic force exerted by the spring. Consider the
situation after the mass passes through the neutral point. Since the magnitude of
the elastic force is proportional to the displacement, the acceleration toward the
neutral position keeps increasing. This causes the velocity of the mass to
decrease. Eventually, the velocity becomes null. At that point, the acceleration
is still toward the neutral point. Therefore, the mass starts to move back towards
the neutral point with zero initial velocity (after a momentary stop). From this
nature, the point where the mass stops momentarily is called the turning point.
After some time, the mass comes back to the neutral point from the opposite
side from the previous time. The mass passes the neutral point again due to the
inertia and moves toward the other turning point. There, by the same
mechanism as at the first turning point, it switches the direction of motion after
a momentary stop. In this fashion, the mass keeps oscillating back and forth
around the neutral position.
It is instructive to analyze the above dynamics qualitatively. The questions
I would like to ask ourselves are (a) How long does it take the mass to come
back to the neutral position? and (b) How far is the turning point away from the
neutral position? The quantity asked by the first question is called the
oscillation period and the one by the second question is the oscillation
amplitude. Let’s consider what determines the oscillation period. Above I said
that the mass moves toward the neutral position due to the elastic force and that
it passes the neutral position due to the inertia. It is easily expected that the
stronger the elastic force the sooner the mass returns to the neutral position and
that the greater the inertia the slower the oscillatory motion of the mass. Thus, it
is speculated that the stronger the elastic force the shorter the oscillation period,
and the greater the mass the longer the oscillation period becomes. Indeed, as
we will discuss in the next section quantitatively, the oscillation period is an
increasing function of the elastic constant divided by the mass of the particle.
Here the elastic constant measures the strength of the elastic force for a given
displacement.

Energy of oscillatory systems

As for the oscillation amplitude, neither the elastic constant nor the weight of
the mass plays a role in determining it. It is determined by the initial position of
the mass. In fact, the amplitude is similar to the distance between the initial and
neutral positions of the mass. (I said “similar” for a reason, which will become
clear shortly). How can we rationalize it? The best way is to consider energy
conservation. Above, I said, “The mass keeps oscillating back and forth around
the neutral position”. This statement indicates that the oscillation continues
forever with the same period and amplitude. You can easily imagine that such a
motion is unrealistic. Indeed, it is impossible for any system to oscillate forever.
Eventually, the oscillation stops, and the mass becomes stationary at the neutral
position. This is because there is always some mechanism that dissipates the
oscillatory energy.
Many oscillatory motions are dissipated by the mechanism known as
velocity damping [4], which will be discussed quantitatively later in this
chapter. The velocity damping force is proportional to the velocity of the mass
in magnitude and opposite in direction. In a system in which the damping effect
is low, the rate that the oscillation energy is dissipated is low.
The oscillation energy consists of the elastic potential energy and the
kinetic energy of the mass. At the initial point where the mass starts the
oscillatory motion, its velocity is null. All the oscillatory energy is in the form
of the elastic potential energy which is determined by the elastic constant and
the initial elastic force. Since the initial elastic force is proportional to the initial
displacement, the initial mass position determines the maximum elastic energy.
At the neutral point where the elastic potential energy is null, the
oscillation energy is in the form of the kinetic energy. As the system loses the
oscillation energy due to the damping mechanism, both the elastic potential
energy and kinetic energy decrease. Consequently, the displacement at the
turning point decreases and the velocity at the neutral point decreases. This in
turn means that the oscillation amplitude keeps decreasing as the oscillation
repeats.
If the system is a low damping system, the oscillatory energy diminishes
slowly, and therefore the amplitude decrease slowly. When the damping effect
is significant, the amplitude decreases fast.

2.1.2 Equation of Motion of Point Mass-Spring System

Unforced oscillation
A system comprising of a spring and a point mass with a velocity-damping
mechanism is a simple yet very illustrative physical system to understand the
decaying oscillatory dynamics. Let be the displacement from the
equilibrium (neutral point). Being proportional to the displacement in
magnitude and opposite in direction, the elastic force can be put in the
following form. (See Fig. 2.1 and note that the displacement vector and spring
force vector are opposite and that the spring force is the only external force
acting on the mass.)

(2.1)
Here the constant of proportionality is called the spring constant or
stiffness. It represents the strength of the elasticity of the spring. From
Newton’s second law, the external force on the point mass, , is
proportional to the acceleration where the mass is the constant of
proportionality. This enables us to write the following equation.

(2.2)
If the spring force is the only component of the external force,
and (2.1), and (2.2) lead to the following equation of motion.

(2.3)
Equation (2.3) indicates that the solution is a function of time that
has the following property; if differentiated twice the result is proportional to
the original function with a negative proportionality constant. We know that
sine and cosine functions have this property. Thus, it is clear that the motion
resulting from a spring force is oscillatory, and can be represented by a sine or
cosine function. The oscillation is called the harmonic oscillation.
When the mass undergoes velocity damping (viscous) force, we need an
additional term on the right-hand side of the equation of motion. The resultant
equation of motion looks as follows.

(2.4)
Here b is the damping coefficient.
For simplicity, introduce the following parameters and .

(2.5)
(2.6)
Here is referred to as the natural (angular) frequency and as the
decay constant. It will become clear in the next section why these parameters
are called in these ways.
Dividing both-hand sides by m and using (2.5) and (2.6), rewrite Eq. (2.4)
in the following form.

(2.7)
Equation (2.7) is a linear differential equation. The displacement of the
point mass is given as a solution to this linear differential equation. Note that
this differential equation has no source term, i.e., the right-hand side of the
equation is zero. Hence, (2.4) is classified as a linear homogeneous differential
equation. Physically, this corresponds to unforced oscillation. In other words,
there is no driving force in the system. Being represented by a linear differential
equation, the system is called a linear system.
Forced oscillation
When an additional force (the driving force) is acting on the mass, the
solution represents forced oscillation. In this case, the differential
equation has a source term as below.

(2.8)
Here is the driving force and m is the mass. It is possible to view the
driving force as an input to the linear system. In this view, the solution is
the output of the linear system. As long as the coefficient and are
constant, the output can be viewed as the linear response to the input.
The linear system associated with elastic force can be argued from the
viewpoint of elastic (potential) energy. As a conservative force, the elastic force
can be characterized as the first-order spatial differentiation of the potential
energy. When the potential energy has quadratic dependence, the force has
linear dependence.
 In an actual physical system, it is possible that the potential energy has a
dependence on the displacement with a third-order or higher polynomial
function. In this case, the system is no more linear. Such a system can be
interpreted as the case where the stiffness is a function of displacement
. A typical example of such a nonlinear system is an inter-atomic potential. The
famous Lennard-Jones potential curve [10] is approximately quadratic near the
equilibrium (near the potential well). This means that as long as the
displacement from the equilibrium is small, the inter-atomic force can be
approximated as a spring-like force.
When the displacement is so large that it exceeds the quadratic part of the
potential well, the dynamics becomes nonlinear. A good example is the acoustic
response of atoms in a solid material having residual stress. Residual stress can
lock the atom in a range where the potential curve is not quadratic. Figure 2.2
illustrates the situation schematically. If the specimen is free of residual stress,
the atom is located near the bottom of the well (the equilibrium position in
Fig. 2.2) where the potential curve has a quadratic dependence on . In
this case, the inter-atomic force is a linear function of
, and therefore the stiffness is constant. (Here is not defined as the
displacement from the equilibrium. is the zero interatomic distance.
Therefore, here is different from used in the differential equations
(2.3), (2.7) and other equations associated with these two equations.) If residual
stress locks the atom far from the equilibrium position, and therefore the
potential energy cannot be approximated as a quadratic function of ,
the stiffness is not a constant anymore; it depends on . In this case, the
atom shows nonlinear behavior when excited by an acoustic wave. Analysis
making use of this non-linearity to probe residual stress is known as
acoustoelasticity [11].
Fig. 2.2 Stress energy potential with residual stress. The slope of the potential curve represents inter-
atomic force. Potential curve is approximately quadratic near equilibrium . On side
potential is steeper than quadratic and on side it is less steeper than quadratic

2.1.3 Oscillation to Waves

In the preceding section, we discussed the temporal oscillatory behavior of a
wave as a harmonic oscillation of a point mass. In Chap. 1 we discussed a wave
is a spatiotemporal oscillation of particles. We can view the phase velocity of a
wave as the motion of a spatial oscillatory pattern in time, or the motion of a
temporal oscillatory pattern through space. In other words, if the harmonic
oscillation experienced by a portion of a medium has a time lag from the
neighboring portion, the harmonic oscillation travels through the medium
forming a wave. In this section, we look into the physics behind this
mechanism.

2.1.4 Wave Equations

One dimensional wave
We first discuss how a harmonic motion of a point mass propagates as a
wave in one dimension. Consider the series of point masses connected with
springs in Fig. 2.3. Imagine you pull the rightmost point mass to the right from
its equilibrium point and release it. When you pull this point mass, the spring
connected to it stretches, pulling the next (second from the right) point mass. In
this fashion, the point masses to the left of these two will be pulled to the right
one after another. Now the question is “Are these all point masses pulled
simultaneously?”. Intuitively, we know that the answer is “No”. The ones
located to the left will be pulled after a certain delay, and the more to the left,
the longer delay time they experience.
When you release the rightmost point mass, a similar delayed behavior is
observed in other point masses. At the moment you release the rightmost point
mass, the next point mass to the left is still moving to the right. Shortly after,
this second point mass changes its direction being pushed by the spring on its
right. Thus, all the point masses change their direction with a certain delay time.
This pattern continues where each point mass oscillates around its equilibrium
point. We can view this oscillatory pattern with the delay time as a longitudinal
wave that propagates the harmonic oscillation of the point masses.
Then, the next question is what is the delay time? Again, intuitively we
know that the weaker the springs the longer the delay time. However, it is not
straightforward to express this delay quantitatively. In fact, this delay time
represents the wave’s phase velocity. Below, we analyze this phenomenon
quantitatively.

Fig. 2.3 Series of point masses. a1 unstretched; a2 rightmost point mass pulled; a3 rightmost point mass
released. b elastic force on nth point mass. c elastic force acting on portion of elastic continuum

Equation of motion
Consider the dynamics of the point mass. This mass receives spring

force from the two springs, one that connects it to the point mass
and the other that connects it to the point mass. The equation of
motion that governs this motion is as follows.

(2.9)
Here m is the mass of the point mass, is the spring constant of the
two springs, and and are the differential displacement of the

nth point mass relative to the and point masses,

respectively. These differential displacements represent the stretch of the
respective springs, and therefore their products with the spring constant
represent the spring force.
Now we extend the above argument into a limit case where the springs are
infinitesimally small so that the entire series of the point masses can be
considered as a continuous elastic medium. In this limit, we can write the
equation of motion (2.9) in the following form.

(2.10)
Here (kg/m) is the linear density of the medium and is the length
of the medium along the x-axis that we apply the equation of motion
(corresponding to the point mass in (2.9)). On the right-hand side of
(2.10), and represent, respectively, the
stretch of the section of the elastic medium neighboring to the section
represented by on the negative and positive sides. In the infinitesimal
limit, we can express the differential displacement (the stretch of the elastic
medium) as follows.

(2.11)
 Here, the quantity represents the length of the section we consider the
stretch. Using (2.11) on the right-hand side, we can rewrite (2.10) as follows.

(2.12)
Expressing the spatial derivative of the displacement as

we can write the quantity inside the parentheses on the right-hand side of
(2.12) as follows.

(2.13)
Here, we used the following expression on the rightmost-hand side of
(2.13).

(2.14)
The above mathematical process lets us rewrite (2.10) as follows.

(2.15)
Note that in deriving (2.15) we canceled from both sides of the
equation.
Spring constant and elastic constant
Let’s take a moment to consider the physical meaning of the spring
constant and elastic constant. These two constants are similar concepts
representing the elasticity of a medium. However, a spring constant is not a
material constant, whereas an elastic constant is a material constant. Each
medium has its unique elastic constant, which characterizes the propagation of
acoustic waves. On the other hand, a spring constant depends on the dimension
of the medium, and thereby it is not a material constant. Below, we consider the
difference between these two constants.
According to Hooke’s law [12], the elastic force at x is the product of the
spring constant and the local differential displacement .

(2.16)
Here, represents the elongation between x and . In Fig.
2.3c, the two arrows depict the elastic force at the two ends of the material
having the length of . Here, the elastic force is due to the elongation over
the segment of thickness .

Fig. 2.4 Spring constant and Young’s modulus

In (2.16) the quantity represents the stretch (the elongation per

unit length or ). We call this quantity the normal strain
. Figure 2.4 illustrates the situation where the same normal
force f acts on a segment of the same material with various thicknesses and
cross-sectional areas causing the same normal strain. In Fig. 2.4a, the segment
has a thickness of and a cross-sectional area of A; in (b) the thickness is
and the cross-sectional area is A; in (c) the thickness is unity ( )
and the cross-sectional area is A; and in (d) the thickness is unity and cross-
sectional area is unity ( ).

According to (2.16), the spring constant for (b) is half of (a), as

inserted in the figure indicates. This is because the segment in (b) is twice as
thick as (a), and therefore, the elongation is double when the same force causes
the same strain. This comparison indicates that the spring constant depends on
the thickness of the medium. The medium thickness of (c) is unity, i.e.,
. In this case, the quantity corresponding to the elongation in (a) and
(b) becomes , i.e. is the strain .
According to continuum mechanics [13], Hooke’s law for a unity thickness
is given as follows.

(2.17)
Here is the stress defined as the force per unit area and E is Young’s
modulus. In the present case, it is the normal stress as the force is normal to the
area. Figure 2.4d illustrates Hooke’s law represented by (2.17).

(2.18)
Since E is defined by the force per unit area and the elongation per unit
length, it does not depend on the dimension of the medium, hence it is a
material constant.
From (2.16), (2.17), and (2.18), we obtain the following set of equations.

(2.19)

(2.20)
From (2.19) and (2.20), we find the following relationship between the
spring constant and Young’s modulus.

(2.21)
 The first expression of (2.21) indicates that Young’s modulus is in N/m ,
and the second expression tells us the greater the cross-sectional area and the
less the thickness, the stronger the spring constant (stiffness) of the elastic
medium becomes.
This situation is similar to the relationship between the electric
conductivity and conductance, ; where A is the
cross-sectional area of the conductive material and l is the length. This
relationship tells us the greater the cross-sectional area and the shorter the
length, the same material exhibits higher conductance.
Phase velocity as a material constant
The above discussion lets us state that phase velocity is a material
constant. Consider the wave equation (2.15).

The quantity on the left-hand side of this equation is linear density.

Expressing the cross-sectional area of the medium with A, we can relate the
linear density and volume density as follows.

(2.22)
Using (2.22) on the left-hand side of (2.15) and the first expression of
(2.21), we can rewrite the wave equation in the following form.

(2.23)
A comparison with (1.14). we can identify the quantity as the
square of the phase velocity .

Thus, we find the phase velocity of the longitudinal displacement wave as

follows.
(2.24)
This discussion tells us that the phase velocity is a material constant. Each
elastic material possesses a unique phase velocity. This causes reflection at a
boundary between elastic media having different phase velocities.
Here (2.25) represents the phase velocity of a longitudinal wave because
the underlying equation of motion represents the normal force acting on the
segment of the material and the resultant elongation (and strain) is parallel to
the force. We can repeat the same argument for shear force and resultant shear
strain to derive the phase velocity of the shear (transverse) wave in the
following form.

(2.25)
Here G is the shear modulus.
2.2 Acoustic Wave Equations and Solutions
In Sect. 2.1.4 we discussed that an equation of motion yields a mechanical wave
equation. We also discussed that an elastic force is necessary for the oscillatory
behavior that generates wave dynamics. These statements indicate that an
expression representing the elastic force and an equation of motion associated
with the elastic force expression are essential to derive an acoustic wave
equation. Elasticity is intrinsic to a medium and therefore the expression of the
elastic force including the expression of the elastic constant depends on the
medium. In this section, we consider wave equations in air and solids.

2.2.1 Acoustic Wave in Air

Hooke’s law in air
We can express Hooke’s law in the following form.

(2.26)
 Here, P is the pressure that causes the volume change , V is the
initial volume, and is the volume after compression. B (N/m ) is the bulk
modulus, which is the elastic constant corresponding to Young’s modulus or
shear modulus for solids. Note that, unlike Hooke’s law for solids (see (2.17)),
we need to use a negative sign. This is because in solids conventionally positive
stress is expansion whereas in air pressure is compression.

Fig. 2.5 Change in volume due to pressure at a single point

Referring to Fig. 2.5 consider the change in volume due to pressure P.

(2.27)
Here, . Therefore,

(2.28)
In the infinitesimal limit the fractions , etc. can be
replaced with the corresponding partial derivatives. Thus (2.26) becomes as
follows.
(2.29)
The quantity represents the divergence of particles from a single
point due to the local displacement field. In other words, , hence the
pressure P is defined at this single point. The fact that the dimension of
is m/m, i.e., unity indicates that this quantity is defined at a single
point. We can view as the volume expansion at the point.
Now consider that the pressure P has a gradient, meaning that it is not
uniform. From (2.29) we put Hooke’s law as follows.

(2.30)
Consider the meaning of in Fig. 2.6 that illustrates a cubic
unit volume of air. Due to the pressure gradient, has x, y, and z
dependence. In other words, at the corners of the cube, the air experiences
different levels of volume expansion illustrated in Fig. 2.5.

Fig. 2.6 Gradient of volume expansion due to pressure gradient

 The left-hand side of (2.30) represents the forces that apply on this unit
volume experiencing . By breaking into the x, y, and z
components, we can visualize that the differential pressure causes the
differential volume expansion .

(2.31)
Equation of motion
We can view the left-hand side of (2.31) represents the net force acting on
the unit volume characterized by mass . Thus, we can write the equation of
motion for this unit volume as follows.

(2.32)
Using (2.30) we can also write this equation of motion expressing with the
temporal and spatial derivatives of the displacement vector and bulk
modulus.

(2.33)
Equation (2.33) explicitly represents that the pressure gradient causes the
acceleration of the unit volume. Thus, we can view it as the equation of motion
governing the unit volume.
Note that we need a negative sign on the right-hand side of (2.32) because
the inward pressure is defined to be positive; if the pressure at is
greater than at x, the net force is in the direction of negative x (the pressure at
pushes the unit volume more strongly inward than at x).
Taking the divergence of both-hand sides of (2.32), we obtain the
following equation.
(2.34)
We can write this equation more compactly as follows.

(2.35)
We can view (2.34) as the equation of motion governing where
the right-hand side represents the acceleration (times the mass) and the left-hand
side the net force.
Volume expansion (compression) wave equation
Express the equation of motion (2.32) in the following compact form.

(2.36)
Hooke’s law is as follows.

Equate the right-hand sides of (2.35) and (2.30), and take the divergence of
the resultant equation using the identity . Then we obtain the
following equation.

(2.37)
We can view (2.36) as an equation that represents a volume expansion

wave that travels at the phase velocity .

Pressure wave equation
Using Hooke’s law (2.29), we can replace on both sides of (2.36)
with and obtain the following equation.
(2.38)
Equation (2.37) is a wave equation that represents that pressure P travels at

the phase velocity of , the same velocity as the volume expansion

(compression) wave.

Alternative way to derive pressure wave equation

We can derive the wave characteristics of the pressure P in the following
fashion as well. Differentiate (2.29) with respect to time twice.

(2.39)
The equation of motion yields the following equations.

(2.40)
Using three relationships in (2.39) on the right-hand side of (2.38), we obtain
the following equation.

or more compactly, in the following form.

(2.41)
Apparently, (2.40) is identical to the pressure wave equation (2.37). We derived
(2.37) by considering the spatial differentiation of the equation of motion
(2.32), and then using Hooke’s law (2.29). On the other hand, in deriving the
pressure wave equation (2.40), we differentiated Hooke’s law (2.29) with
respect to time first and then used the equation of motion. In either case, we
derived the pressure wave equation based on Hooke’s law and Newton’s second
law (the equation of motion).
Density-change wave equation
From the conservation of mass, we find the following equality.
(2.42)
Here and are the density before and after the volume expansion.
Define the relative volume change s as follows.

(2.43)
Substituting (2.41) into (2.42),

(2.44)
From Hooke’s law (2.26),

(2.45)
Substituting (2.44) into pressure wave equation (2.40), we obtain the
following wave equation.

(2.46)
Using (2.43) we can write wave equation (2.45)

(2.47)
We can view (2.46) as a wave equation of density change.
Note that all the above four wave-equations have the following same phase
velocity.
(2.48)
This phase velocity expression has the same form as the one for solids; the
square root of the ratio elastic constant over density, as will be discussed in the
next section.

2.2.2 Longitudinal and Transverse Vibrations in Solid

In elastic solids, longitudinal and transverse vibrations propagate as
longitudinal and transverse (shear) waves. We can discuss the dynamics of these
waves by considering equations of motion.

Fig. 2.7 Elastic force acting on central block

First, consider longitudinal wave dynamics. Figure 2.7 illustrates a block

of an elastic medium between x and (called the central block). This
block is part of a long material extended along the x axis. At the far-right end of
this material, an external agent exerts a tensile force. The other external agent
holds the opposite end (on the negative x side). Consequently, the entire
material experiences a stretch. Since the external force is dynamic, the block is
not under static equilibrium.
At the face at x, the central block pulls the neighboring block on the
negative x side in the positive x direction. In response, the neighboring block
exerts a leftward reaction force on the central block. This force pair
causes a local stretch at the boundary of the central and left neighboring blocks.
Figure 2.8 illustrates the situation at the boundary in detail. Referring to
the upper part of this figure, we can interpret this local stretch as differential
displacement . Here represents
the thin boundary area between the neighboring and central blocks, and is
the coordinate of the neighboring block side of the boundary. Thus, using the
spring constant , we can express the leftward force exerted by the
neighboring block on the central block as follows.

(2.49)
Here is the spatial derivative of , i.e.,
.

Fig. 2.8 Local stretch at and

On the other face at , the neighboring block located

on the positive x side exerts a rightward force . Repeating
the same argument as the face at , we can express the elastic force at the
as follows.

(2.50)
 Here we assume that and at are the same as
at .
Thus, the net external force acting on the central block is
. According to Newton’s second law, the
acceleration of the central block is equal to the net external force. From (2.48)
and (2.49), we obtain the following equation of motion.

(2.51)
Expressing the mass of the central block with the density and the cross-
sectional area A as , we can rewrite the equation of motion
(2.50) as follows.

(2.52)
In the infinitesimal limit, we can put . Thus,
canceling and rearranging terms, we can write (2.50) as follows.

(2.53)
As discussed with (2.18), we can express the normal stress with elastic
force as follows.

(2.54)
Normal stress is related to normal strain with Young’s modulus E.

(2.55)
 Comparing the right-hand side of (2.53) and (2.54), we find the following
relation. (This is the same argument as the derivation of (2.21).)

(2.56)
Using (2.55), we can rewrite (2.52) as follows.

(2.57)
Equation (2.56) is a wave equation that describes the dynamics of a
longitudinal wave traveling through a solid that has Young’s modulus (the
longitudinal elastic modulus) E.
A comparison of (1.14) and (2.56) indicates that the wave equation is, in
fact, the equation of motion, and moreover, the phase velocity of the
longitudinal elastic wave is given as follows.

(2.58)
Similarly, considering the displacement behavior perpendicular to the x
axis and its variation with x, we can derive a transverse wave equation. The
lower part of Fig. 2.8 illustrates the x dependence of the displacement in the y
direction, , at the and boundaries of the
central block.
In this case, the net force acting on the boundary area of thickness is
the shear force in the form of the spring constant times the differential
displacement . Thus, we can write the equation of motion for the shear
dynamics as follows

(2.59)
and stress-strain equation as follows.
(2.60)
Here is the shear stress and the other quantities are the same as (2.53).
Equations (2.58) and (2.59) are the shear dynamics versions of (2.50) and
(2.53), respectively.
Similarly to (2.54), we can relate the shear stress with shear strain .

(2.61)
Here G is the shear modulus and is related to the spring constant as
follows. This relationship is the shear deformation version of (2.55).

(2.62)
With the same procedure as the longitudinal wave case, we can derive a
wave equation and phase velocity for the shear dynamics as follows.

(2.63)

(2.64)
Note that as we discussed at the end of Sect. 2.1.3, the phase velocities of
the longitudinal and transverse (shear) wave are material constants. They
contain material constants E, G, and . When an acoustic wave of the
frequency ( ) determined by the source (e.g., the operation frequency of an
acoustic transducer) propagates in an elastic material, the wavelength ( ) is
determined by the material’s unique phase velocity and the frequency (
).
2.2.3 Decaying Acoustic Wave Equations and Solutions
We derived the wave equations (2.36) and (2.37) from the equation of motion
(2.35). While this equation of motion describes the mechanism in which the unit
volume of air is accelerated by the differential pressure (the pressure gradient),
it generates a solution that represents dynamics that lasts forever. In reality,
waves decay due to various mechanisms such as velocity damping associated
with collisions between air molecules. (Drag force in viscous media is
proportional to the velocity only for lighter objects and lower speed [14, 15].)
To the first-order approximation, we can represent this effect by including
a velocity-damping force term in the equation of motion (2.33).

(2.65)
Equation of motion (2.64) leads to the following decay wave equation of
volume expansion.

(2.66)
Here is the decay constant and is the phase velocity defined by
(2.47).

(2.67)
In the case of longitudinal and transverse waves in solids, we derived the
longitudinal and transverse (shear) wave equations (2.56) and (2.62) from the
corresponding equations of motion (2.50) and (2.58). Similarly to the wave
equation in air, we can include the effect of velocity damping by adding the
term proportional to the first-order time derivative of the displacement.

(2.68)
(2.69)

As clear from the above arguments, the physical mechanism that generates
acoustic waves in materials in air and solids, in common, is elasticity. The phase
velocity has the form of the square root of the ratio of elastic constant over the
density. When the medium is viscous, the velocity-damping term must be
included. Thus, we can express the wave equation and phase velocity in the
following general forms.

(2.70)

(2.71)
Here f represents a quantity that exhibits the wave characteristics in
general, is the damping coefficient, and is the elastic constant. In the
case of the waves in solids, is either longitudinal or shear displacement, and
is either Young’s modulus or shear modulus correspondingly. In the case of
waves in air, f can be the volume compression , pressure p, or density
change . The other types of elastic waves discussed in Chap. 4 can have
a decay effect in the same fashion.
Decaying acoustic wave solution
Above we discussed that linear acoustic wave equations with a velocity
damping mechanism can be expressed in the form of (2.69) commonly to in air
and solids. In this section, we discuss general solutions to this type of wave
equation.
In Chap. 1, we used the cosine function (1.31) as a solution to the one-
dimensional wave equation without a damping term (1.15).
 The cosine function in (1.31) represents the oscillatory characteristics of
the wave. In the wave equation, the combination of the secondary temporal and
spatial derivative of the wave function represents the oscillatory characteristics.
The square of the phase velocity that appears in front of the spatial secondary
derivative term represents the elasticity, which is the physical origin of the
oscillatory behavior. The damping term in (2.69) originates from the velocity
damping effect, and does not cause the oscillatory behavior. The decaying wave
equation (2.69) still has the secondary temporal and spatial derivative terms in
the same form as the non-decaying wave equation.
These observations indicate that a decaying wave solution can be
expressed with a sine or cosine function, with the addition of a term that
represents the decaying characteristics of the wave. So, it is natural to assume a
solution that has a sinusoidal function as part of it. In this section, we use the
exponential form of a sinusoidal function as it is more convenient to include the
damping effect.
Using Euler’s notation [16, 17], we can express (1.31) as follows.

(2.72)
We can view (1.31) as the real part of this exponential expression with
initial phase . We set because the initial phase does not affect
the gist of the discussion.
Substitution of (2.71) into (2.69) leads to the following equation.

(2.73)
It follows that the content of the curly bracket is zero.

(2.74)
In order for (2.73) to hold, it is necessary that either or k is a complex
number. Let’s assume k is complex, i.e., .

(2.75)
It follows that the following set of equations holds.
(2.76)

(2.77)
Elimination of from (2.75) and (2.76) yields the following equation for
.

(2.78)
Solve (2.77) as a quadratic equation of taking only the real root to
obtain the following expression.

(2.79)
From (2.75), we find as follows.

(2.80)
Thus, we find it as follows.

(2.81)
(2.82)
By separating the real and imaginary parts of k, we can express the
solution (2.71) as follows.

(2.83)
The exponential term multiplied by amplitude A
represents the decaying effect as follows. As we discussed in Sect. 1.2.1, when
the wave propagates in the positive direction, the phase term takes the form of
with a positive propagation constant k. We set the complex
propagation constant as . So, for a wave propagating in the
positive x direction, . This indicates as the wave propagates along the
x axis, i.e., with an increase in x, the exponential term
decreases. We can express a wave propagating in the negative x direction by
assuming k is negative. In this case, the wave decays as it travels, too. As an
increase of x in the negative direction (i.e., as x becomes more negative), the
negative makes the exponential term decay. So, regardless of the
direction of propagation the wave decays as it travels.
When the damping coefficient b is negligible, we can put and
(2.80) and (2.81) reduce to the following expressions.

(2.84)
(2.85)
Here (2.84) indicates that when , , hence .
Using this fact, we expressed the phase velocity as equal to in
(2.83). Thus, under the condition of , the decay wave solution (2.82)
reduces to the exponential form of the non-decaying solution (1.31).

Three dimensional case

Now quickly extend the above discussions regarding the decay wave solution
into three dimensions by solving the three-dimensional wave equation.

(2.86)
As clear from the general discussion of extending a one-dimensional wave to a
three-dimensional wave made in Sect. 1.2.2, we can assume the solution in the
following form.

(2.87)
Use the Laplacian operator for the spatial differentiation of (2.86), as we did in
Sect. 1.2.2. Expressing the magnitude of propagation vector with k, we can
express the second-order spatial differentiation as follows.

(2.88)
Substitution of (2.86) into (2.85) with the use of (2.87) yields the following
equation.

(2.89)
Clearly, (2.88) leads to the condition (2.73) for the one-dimensional case.
Thus, we find the same expressions of the complex propagation constant as the
one-dimensional case. In other words, except for the direction of propagation,
the three-dimensional wave (2.86) behaves exactly the same as the one-
dimensional wave (2.71).
It is instructive to express the spatial part of the phase that appears

in the exponent of the three-dimensional wave function (2.86) using , , and

(the unit vectors for x, y, and z directions). Referring to Fig. 2.9, express
in the following form.

(2.90)
Here is the unit vector in the direction of the propagation vector . (I
do not use for this unit vector to avoid confusion with the unit vector for the
z direction.)

Fig. 2.9 Propagation vector k and its unit vector whose components are direction cosines

As discussed in Appendix A, the unit vector ’s x, y, and z components

are the direction cosines of the respective axes.

Using (A.8) we can rewrite (2.89) in the following form.

(2.91)
We can interpret the quantities , , and that appear in the last
term of (2.90) as the x, y, and z components of the propagation constant. Thus,
this term indicates that the wave propagates in the direction

gaining the phase (including the decay

factor) along the x, y, and z axes with the respective components of k.
By separating the propagation constant into the real (oscillatory) and
imaginary (decaying) parts, we obtain the following explicit forms.

(2.92)
where

(2.93)
When the damping coefficient b is negligible, we can put and
(2.92) and (2.93) reduce to the following expressions.

(2.94)
(2.95)
As is the one-dimensional case, the condition reduces the
solution (2.86) to the non-decaying solution where is a real number.

2.2.4 Nonlinear Wave Equations

So far we assumed that the elastic modulus does not depend on the space
coordinate. If the elastic modulus is a function of the space coordinate, the wave
equation becomes nonlinear. Consider this effect for the case of longitudinal
waves in a solid.
Reconsider the equation of motion (2.50) allowing the stiffness (spring
constant) to depend on x. In this case, we need to include differentiation of the
stiffness in the evaluation of the net force on the central block.

(2.96)
Using (2.55) allowing E to depend on x, we can put (2.97) in the following
form.

(2.97)
This situation makes it impossible to put the wave equation in the form of
(2.56) where the secondary differentiation of the wave function ( )
with respect to time is proportional to the secondary differentiation of the
function with respect to space. Thus, the phase velocity is not a constant.
The situation is the same as the case when the displacement is a vector or
the medium has viscosity.

(2.98)
As an example, we can discuss nonlinear wave equations in solids by
allowing the spatial coordinate dependence in the spring constant in (2.50), i.e.,
Youg’s modulus or shear modulus depends on the coordinate variable. This type
of nonlinearity is caused by the fact that the restoring (elastic force) has second
or higher-order dependence on the space coordinates. A good example of such a
case is the recovery force of residual stress discussed in Fig. 2.2. When residual
stress causes the strain so large that the atom deviates from the bottom of the
inter-atomic potential where the potential curve is approximately a quadratic
function of the inter-atomic distance, its slope becomes nonlinear [18, 19]. Let’s
take a moment to consider this case.
Let U(r) be the inter-atomic potential and have a third-order dependence
on the distance from the equilibrium (the bottom of the potential curve) r.

(2.99)
Here a, b, c, and d are constant, and r corresponds to in Fig. 2.2.
The spatial first-order differentiation of potential energy is the restoring
force. So, in this case, the restoring force has a second-order dependence on
r.

(2.100)
Elastic force is expressed as the product of the stiffness (the spring
constant) and the displacement from the equilibrium. From (2.100), we find
the following expression of the stiffness for the potential (2.99).

(2.101)
Apparently, is not a constant.
Although nonlinear acoustics and related technology [20] are interesting
topics, we mostly discuss linear acoustic waves in this book.
2.3 Wave as a Flow of Energy
In this section, we discuss acoustic waves as a flow of acoustic energy.

2.3.1 Acoustic Energy

Consider in Fig. 2.10 that a cubic block of elastic medium experiences tensile
force applied on the planes at x and . This pair of forces stretch the
block by the differential displacement in the x direction . We can express
the work done by these forces as follows.

(2.102)
Here is the stiffness (spring constant) of the elastic medium.

Fig. 2.10 Work done by force f is stored as potential (elastic) energy of the elastic medium

Express the force in two different ways so that we can rewrite (2.102)
independent of the dimension.

(2.103)

(2.104)
Here is the normal strain, is the normal stress, E is Young’s modulus,
and A is the cross-sectional area of the volume element.
By equating (2.103) and (2.104), we find the following relationship
between the spring constant and Young’s modulus.

(2.105)
Substitute (2.105) into (2.102).
(2.106)
Noting that is the volume of the block, we find the following
expression for the energy density.

(2.107)
This energy density represents the elastic potential energy per unit volume.
The suffix p stands for the potential energy.

Instantaneous acoustic energy

Above, we made the analysis for static equilibrium. Now consider that the force
(stress) is dynamic in the context of an acoustic wave. As we discussed in Sect.
1.2.3, a one-dimensional wave solution can be put in the following form (see (1.
31)).

(2.108)
Here represents the displacement from equilibrium. (In Fig. 2.10 we use
to represent the particle displacement from the origin, not necessarily from the
equilibrium. Here we define to be from the equilibrium. This difference does
not affect the gist of the discussion here.)
The particle velocity is the time differentiation of (2.108), and the normal
strain is the spatial differentiation of (2.108).

(2.109)

(2.110)
Fig. 2.11 Snapshot of displacement (top), strain (middle), and velocity (bottom) waves at two time-steps.
The horizontal axis is common to all plots

Figure 2.11 illustrates sample snapshots of ,

, and at two instances corresponding to
and one-eighth of the period later. Note that the strain wave is a quarter
period shifted from the displacement wave, and the velocity wave is out of
phase to the strain wave. These phase shifts are explained by (2.110) and
(2.109), respectively.
From (2.109) and (2.110), we find the following relationship between
v(t, x) and .

(2.111)
 As a side note, the negative sign in (2.111) indicates that the velocity wave
is out of phase to the strain wave. Also, is the reciprocal of the phase
velocity. Equation (2.111) indicates that the ratio of the amplitude of the
velocity wave over the strain wave is equal to the phase velocity and that the
velocity and strain waves are out of phase by .
Substitution of (2.111) into (2.107) yields the following expression.

(2.112)
From (1.11) we know is the phase velocity and from (2.24) we
know the phase velocity of a longitudinal elastic wave is related to the density

as . Thus, we find the right-hand side of (2.112) as follows.

(2.113)
Expression (2.113) represents the kinetic energy of particles in the unit
volume, .
How about the acoustic potential energy density in air? Fig. 2.12 illustrates
the situation where the external force f acting normally on the plane of area A
increases the acoustic energy density. In (a) the external force stretches the
volume and in (b) it compresses the volume. In (a) the stress is and in
(b) the pressure is p/A. Let a and b represent the shorter and longer sides of the
cube in the direction parallel to the external force ( ). Using Hooke’s
law expressions in solids (2.17) and air (2.29), we obtain the following
equations. (For the in-air case, we use a one-dimensional expression of (2.29).)
Here, in both cases, the external force does positive work, hence the change in
the acoustic energy is positive.

(2.114)
(2.115)

Fig. 2.12 Potential energy density in a solid (a) and air (b). In both cases, the external force increases the
acoustic energy density

In summary, we find the instantaneous elastic potential energy and the

particle’s kinetic energy as follows.

(2.116)
The above discussions indicate that at a given moment, the acoustic wave
has potential energy density and kinetic energy density , and they are
equal to each other. In Sect. 2.3.2, we find that an acoustic wave carries these
two types of energy at the phase velocity. This situation is analogous to that an
electromagnetic wave carries electric and magnetic energy densities that are
equal to each other.
Average acoustic energy
What is the actual magnitude of the energy that flows as a wave? It is
determined by the initial condition, i.e., how much the external agent initially
pulls the unit volume, i.e., the actual value of in (2.109) and (2.110). Putting
the amplitude of the velocity and strain as and , we
can express the temporal variation of the velocity and strain as follows. These
expressions correspond the temporal oscillation of v and at a fixed x in
(2.109) and (2.110).

(2.117)
(2.118)
We can express the average values over one period ( ) for
these energies as follows.

(2.119)

(2.120)
Thus, the average total energy density becomes as follows.

(2.121)
Some note regarding equality of kinetic and potential energy
The equality between the kinetic and potential energy leads to some
intriguing insight into the acoustic field. From (2.116) we obtain the following
equality.

(2.122)
Rearranging (2.122) we find the ratio of v over as follows.

(2.123)

Here, from (2.57) we know is the phase velocity of the

longitudinal wave characterized by the material constants E and .
 Knowing that and , we can view the ratio
of the amplitude of the velocity wave over the amplitude of the strain wave as
representing the phase velocity via the following argument.

(2.124)
Note that v and in (2.123) represent the amplitude of the respective
waves. That is why a negative sign does not appear in this equality unlike
(2.111) which relates the instantaneous value of the velocity wave and strain
wave.
Acoustic impedance
In the theory of electronic circuits, impedance represents the difficulty of
current flowing through a medium. This concept corresponds to the resistance
in a DC (direct current) circuit, which is defined as the ratio of the voltage drop
across a resistor over the current. In an AC (alternating circuit) transmission line
or electromagnetic wave transmission in air, impedance is significant because
when an AC current or electromagnetic wave passes through a boundary where
the impedance varies before and after, reflection occurs. Consequently, the
power is not transmitted as expected. Therefore, various techniques to match the
impedance are used. This operation is called impedance matching [21].
The acoustic impedance is defined as the ratio of the amplitude of the
stress (pressure) wave to the amplitude of the velocity wave. In the case of a
longitudinal wave in a solid, we obtain the following expression for the acoustic
impedance. See (2.123) in going through the third equal sign in the below
equation.

(2.125)
This definition is analogous to that of the electrical impedance. In
electrodynamics, the scalar potential (voltage) is the spatial integration of the
electric field, which exerts an electric force on the electric charges that flow as
an electric current. Here in (2.125), the acoustic impedance is the ratio of stress
(force) over the velocity (flow).
In the case of air, the acoustic impedance takes the following form.
(2.126)
Applying the equality between the potential and kinetic energy density
(2.122) to the case of air, we obtain the following equation.

(2.127)
So,

(2.128)
When we excite an ultrasound in a solid with an ultrasound transmitter, it
is necessary to use a buffer, typically distilled water or ultrasound gel, to
increase the coupling. The role of the buffer is to match the impedance. (Tip: If
an ultrasound gel is not available, honey may work well as a buffer. [22])

2.3.2 Intensity of Acoustic Wave

In the preceding section, we discussed that acoustic energy density is the
product of pressure (stress) and velocity. From this discussion, we can envision
that a pair of pressure (stress) and velocity waves carry acoustic energy. In this
section, we discuss this topic quantitatively.
Let’s use the displacement wave in the form of (2.108) for this discussion.
In this case, the velocity and strain waves are as shown in (2.109) and (2.110).
 Remembering that is the phase velocity (see (1.11)), we find
the following relationship from (2.109) and (2.110).

(2.129)
The relationship (2.129) is nothing new. It is an iteration of (2.111) that
states that ratio of velocity wave over the strain wave is equal to the phase
velocity with a negative sign. As I mentioned under (2.111), the negative sign
indicates that the velocity and strain waves are out of phase with each other by
. In a sinusoidal wave, out of phase by is equivalent to flipping the sign of
the wave function.
Viewing as the one-dimensional version of volume
expansion and using (2.30) that relates pressure and volume
expansion, we obtain the following expression for pressure P.

(2.130)
Thus, using (2.129) we can express the product of the pressure and
velocity waves in two ways as follows.

(2.131)

(2.132)
Here going to the second line of (2.131), we use the total energy density
expression (2.116).
Dimensional analysis indicates that the Pv wave is in (N/m )(m/s)
J/m ) (m/s) = W/m . This quantity is called intensity often expressed with a
symbol I. Expression (2.131) literally indicates that the pv wave carries the total
energy density at the phase velocity .
On the other hand, (2.132) indicates that the Pv wave oscillates at the same
frequency as the displacement wave and its amplitude is . By
expressing (2.132) with the pressure and velocity wave amplitude
and , we can find the average
intensity as follows. See (2.119) and (2.120) for the equivalent expression for
the velocity and strain waves.

(2.133)
The above discussions indicate that the pressure and velocity wave
together carry the acoustic energy. Figure 2.13 illustrates that the velocity and
pressure waves are in phase, oscillating at the same frequency as the
displacement wave.

Fig. 2.13 Displacement, stress, pressure, and velocity waves. Note that the pressure and velocity waves
are in phase. They together carry acoustic energy according to (2.131)
 If we repeat the same procedure to express the product of the normal stress
and velocity waves us , we obtain the following pair of equations.

(2.134)

(2.135)
This time, unlike the product of pressure and velocity wave, we have a
negative sign in front of the phase velocity . Using Fig. 2.13, we can explain
this discrepancy between the cases of the pressure wave and stress wave as
follows. The stress wave is out of phase from the displacement wave by
and from the velocity and pressure waves by . The phase shift of
between the displacement wave and the other three waves originates from the
fact that these three waves are temporal or spatial derivatives of the
displacement wave. The phase difference between the stress and pressure
waves, which have the same physical dimension of N/m , is simply due to the
difference in the definition; a stretch of an elastic medium corresponds to
positive stress and negative pressure, i.e., vs
. From this standpoint, we can say that carries
acoustic energy together with the velocity wave .
In any event, the intensity is defined as the absolute value of the time
average of the product of the component waves, i.e., the pressure and velocity
waves or the stress and velocity waves. So, it is always a positive value.
2.4 Acoustic Wave Reflection and Transmission
2.4.1 Laws of Reflection and Refraction
Refer to Fig. 2.14. The acoustic wave is incident to the boundary AB from the
left. The acoustic wave travels in the medium on the left of the boundary with a
phase velocity , and in the medium on the right of the boundary with .
The frequency of the wave is constant through the boundary. Therefore, the
wavelength on the left of the boundary, is different from that on the right,
 . Lines N A and N B are normal to boundary AB. The angle made by
these lines and AB is called the angle of incidence .

Fig. 2.14 Law of reflection and refraction

The dashed lines represent the wavefront of three waves, i.e., the incident
wave, reflected wave, and transmitted wave (the crest of the respective waves).
Lines S A and S B are perpendicular to the wavefronts, representing the
propagation of the incident wave. Figure 2.14 depicts the moment when the left
edge of the wavefront of the incident wave reaches the boundary at point A.
Therefore, line DB is equal to wavelength . Similarly, on the right of the
boundary, line AE is equal to the wavelength .
Consider triangle ABD. Angle ABD is complementary to angle DBN .
Angle BAD is complementary to angle ABD because angle ADB is a right
angle. Therefore angle BAD is equal to angle DBN , which is the angle of
incidence . Length BD is equal to the wavelength in medium 1, . We
obtain the following equation.

(2.136)
Repeating the same argument for triangle ABC, we find the following
equation.
(2.137)
From (2.136) and (2.137) we find the following equality, which is known
as the law of reflection; the angle of reflection is equal to the angle of
incidence.

(2.138)
Now pay attention to triangle ABE. We can easily find the following
equation.

(2.139)
From (2.136) and (2.139) we find the following equality, which is known
as the law of refraction, or Snell’s law.

(2.140)
Since the wavelength is proportional to the phase velocity for the same
frequency, we can express Snell’s law in the following form as well.

(2.141)
Here and are the acoustic phase velocity in the medium on the left
and right of the boundary.
Now consider that a wave is incident to a boundary from medium 1 to
medium 2 where the phase velocity in medium 1 is higher than medium 2, i.e.,
. Solving (2.141) for , we obtain the following expression.

(2.142)
Since the sine function cannot take a value greater than unity and
, (2.142) does not hold for all . The following condition is
necessary.
(2.143)
This angle is referred to as the critical angle.

2.4.2 Behavior of Acoustic Waves Near Boundary

When an acoustic wave passes through a boundary that separates two elastic
media of different elasticity and density, part of the wave energy is reflected and
the rest is transmitted. If the angle of incidence is oblique, the direction of the
transmitted wave is determined by Snell’s law. The amount of reflected and
transmitted energy is determined by the boundary conditions. In this section, we
discuss these topics.

Formulation of incident, reflected, and transmitted waves

We first consider acoustic waves in solids. In this section, we use the
exponential form as it avoids complexity in the formulation. Set the incident,
reflected, and transmitted displacement in the following form.

(2.144)
Here , , and , are the amplitude of the respective waves.
Subscripts 1 and 2 for Young’s modulus E and propagation constant k denote
medium 1 (the incident side) and medium 2 (the transmitted side).

(2.145)

(2.146)

(2.147)
Remembering the definition of acoustic impedance (2.125), we obtain the
following expressions.
(2.148)

(2.149)

(2.150)
Notice that the magnitude of the acoustic impedance is the same for the incident
and reflected waves as they are in the same medium. Their signs are opposite
because the propagations are opposite. The acoustic impedance of the
transmitted wave is different from the impedance of the incident or reflected
wave as the transmitted beam propagates in a different medium.
Now we make the same discussion as above for acoustic waves in air.
According to (2.29) and (2.54), the only difference between the in-air and in-
solids cases is the elastic modulus’s sign. (You may notice that while (2.29) is
expressed in three dimensions for air, (2.54) is in one dimension for solids. We
will discuss Hooke’s law in solids in a three-dimensional form in Chap. 4.)
Replace E with and with in (2.145)–(2.150) to obtain the
following equations.

(2.151)

(2.152)

(2.153)
(2.154)

(2.155)

(2.156)
Similarly to the case in solids, the acoustic impedance for the incident and
reflected waves have the same magnitude and opposite signs, whereas that for
the transmitted wave is different.

Boundary conditions and reflectivity

At the boundary, the following two conditions must hold for the pressure and
particle velocity.

(2.157)

(2.158)
Condition (2.158) represents the force balance at the boundary plane. If the
balance is broken, the boundary plane will have acceleration, and consequently,
it moves in the positive or negative x direction. Condition (2.157) represents the
law of momentum conservation along the x axis, i.e., perpendicular to the
boundary plane. Per condition (2.158), the net force acting on the particle along
the x is zero.
Using (2.148)–(2.150) in (2.158) we obtain the following equation.

(2.159)
Eliminate from (2.157) and (2.159).
(2.160)
Similarly, eliminate to obtain the following equation.

(2.161)
From (2.160) and (2.161), we find the reflection and transmission
coefficient for the velocity wave as follows.

(2.162)

(2.163)
Substituting (2.148)–(2.150) into (2.162) and (2.163), we can express the
reflection and transmission coefficient in terms of the stress waves as follows.

(2.164)

(2.165)
Note that (2.162)–(2.165) represent the reflection and transmission of the
stress and velocity waves. Therefore, each of these expressions does not
represent the reflection or transmission of the acoustic energy. The product of
the stress (N/m ) and velocity (m/s) represents the intensity (N/m m/s
W/m ), i.e., the area density of acoustic power. As we discussed in Sect. 2.3.2,
it is the product of the pressure (stress) and velocity waves that represents the
flow of acoustic intensity.
Energy conservation at boundary
 Now express the intensity of the incident wave striking the boundary plane
by considering the product of the stress and velocity waves. From (2.144) and
(2.145), we obtain the following expression.

(2.166)
In (2.166), the factor represents the component of the intensity
parallel to the boundary plane (see Fig. 2.14). Since we are considering the
event on a fixed x where the boundary is located, the intensity is a function of
time. In other words, the intensity oscillates with angular frequency

according to . Here represents the x

coordinate where the boundary is located.
Similarly, we can formulate the reflected and transmitted intensities.

(2.167)

(2.168)
So, the total reflected and transmitted intensity is

(2.169)
 Equation (2.169) indicates that through the boundary the acoustic energy is
conserved.
By defining the reflectance and transmittance
we obtain the following simple expression from (2.169).

(2.170)
In the case of acoustic waves in air, , , and are the same as
solids (compare (2.145)–(2.147) and (2.151)–(2.153)). Therefore and
are the same as solids. For and , we need to replace and with
and in the expression of the corresponding p waves. This
change results in the reflectivity and transmissivity as follows.

(2.171)

(2.172)
After all, both coefficients are the same as the solid’s case (2.164) and
(2.165), if expressed with acoustic impedances.

References

1. 1. Addink CCJ (1951) Precise calibration of tuning forks. Philips Tech Rev 12(8):228–232
2. 2. Harvard Natural Sciences Lecture Demonstrations, https://sciencedemonstrations.fas.harvard.
edu/presentations/tuning-forks (accessed on July 17, 2023)
3. 3. https://www.birmingham.ac.uk/teachers/study-resources/stem/physics/harmonic-motion.aspx
4. 4. Blake RE (1996) Basic vibration theory. In: Harris CM (ed) Shcok and vibration handbook, 4th
edn, Ch. 2. McGraw-Hill, New York
5. 5. Margenau H, Murphy GM (1956) The mathematics of physics and chemistry, 2nd edn. D. Van
Nostrand, Co. Inc., Toronto, New York, pp 50–53
 6. 6. King GC (2009) Vibrations and waves. Wiley, Chichester, UK, pp 33–48
7. 7. øAström KJ, Murray RM (2021) Feedback systems, an introduction for scientists and engineers.
Princeton University Press, Princeton, London, pp 27–29
8. 8. Obando L (2018) Mass-spring-damper system, 73 exercises resolved and explained: systems
dynamics and transfer function. Universidad Central de Venezuela, Caracas, Venezuela
9. 9. Yoshida S (2017) Waves; fundamental and dynamics. Morgan & Claypool, San Rafael, CA,
USA, IOP Publishing, Bristol, UK, Chap 1
10. 10. Jones JE (1924) On the determination of molecular fields.—I. From the variation of the
viscosity of a gas with temperature. Proc R Soc Lond Ser A 106(738):441–462
11. 11. Toupin RA, Bernstein B (1961) Sound waves in deformed perfectly elastic materials.
Acoustoelastic effect. J Acoust Soc Am 33:216
12. 12. Mavko G, Mukerji T, Dvorkin J (2020) The rock physics handbook, 3rd edn. Cambridge
University Press, New York, Chap. 2
13. 13. Reddy JN (2013) An introduction to continuum mechanics, 2nd edn. Cambridge University
Press, Cambridge, UK
14. 14. Tonya Coffey, Drag, https://www.youtube.com/watch?v=xKfZ25yBxu8 (accessed on July 19,
2023)
15. 15. Hooge C. 5.2 drag forces. In: BCIT physics 0312 textbook. https://pressbooks.bccampus.ca/
physics0312chooge/chapter/5-2-drag-forces/
16. 16. Stipp D (2017) A most elegant equation; Euler’s formula and the beauty of mathematics. Basic
Books, New York, USA
17. 17. Boas ML (2006) Mathematical methods in the physical sciences, 3rd edn. Wiley, London, p 61
18. 18. Hughes DS, Kelly JL (1953) Second-order deformation of solids. Phys Rev 92(5):1145–1149
ADSCrossref
19. 19. Yoshida S, Sasaki T, Usui M, Sakamoto S, Gurney D, Park IK (2016) Residual stress analysis
based on acoustic and optical methods. Materials 9:112
ADSCrossref
20. 20. Garrett SL (2020) Nonlinea acoustics. In: Understanding acoustics. Graduate texts in physics.
Springer, Cham. https://doi.org/10.1007/978-3-030-44787-8_15
21. 21. Rehman A (2018) Practical impedance matching, 1st edn. Signal Processing Group Inc.,
Tempe, AX, USA
22. 22. Miyasaka C (2012) Private communication
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024
S. YoshidaFundamentals of Acoustic Waves and ApplicationsSynthesis Lectures on Wave Phenomena in
the Physical Sciences
https://doi.org/10.1007/978-3-031-48200-7_3

3. Propagation of Acoustic Waves in Air

Sanichiro Yoshida1
(1)
Department of Chemistry and Physics, Southeastern Louisiana
University, Hammond, LA, USA

Sanichiro Yoshida
Email: [email protected]

This chapter specifically discusses the propagation of acoustic waves in air. By

considering the source of elasticity of air, we derive the equation of motion and
corresponding wave equations. After these discussions of wave dynamics, we
touch on the sounds we hear. We briefly discuss how human ears conceive and
process sound waves. We consider how musical instruments generate their
unique sounds. Toward the end of the chapter, we discuss amplitude and
frequency modulations as methods to transmit sound waves over a distance.
3.1 Dynamics of Air Particles and Acoustic Wave Equation in Air
3.1.1 Particle Motions in Longitudinal Waves
In Sect. 2.2.1 we derived the longitudinal wave equations for volume
compression and pressure from the equation of motion. There we expressed the
equation of motion. In this process, we used the displacement of air particles to
derive the equation of motion. These procedures indicate that the deviation of
air particles from the equilibrium is the underlying mechanism to generate the
wave dynamics of air particles. We start this chapter by visualizing the
relationship between displacement and pressure gradient.
Figure 3.1 is a snapshot of air particles where a speaker generates a sound
wave in the air. The diaphragm of the speaker moves sinusoidally. The dots
represent the position of layers of particles. In Fig. 3.1a, the horizontal arrows
represent the deviation of layers from the equilibrium position. In those
locations where the dots are dense (space), the pressure is high (low). Figure
3.1b illustrates the pressure variation as a function of x, the coordinate variable
along the axis of wave propagation.
Pay attention to the second positive peak of the pressure graph. In Fig.
3.1a, the horizontal arrows are inward to this x location, indicating that the
corresponding displacement vector changes the sign over this location. In fact,
on the negative-x (positive-x) side of this location, the arrow is rightward
(leftward), meaning that the displacement is positive (negative). Figure 3.1c
explicitly illustrates this change in the sign. Algebraically, we can express the
above relationship between pressure and displacement as ;
the negative slope of corresponds to positive pressure (see (2.29)).
The dashed lines in Fig. 3.1b, c are the spatial variation of the respective
quantities later. Notice that the pressure wave travels to the right so does
the displacement wave. Figure 3.1d illustrates this temporal change observed at
the leftmost x position. During the elapse of , the pressure at this point
decreases while displacement increases.
Fig. 3.1 Snapshots of air particles a, pressure wave b, and displacement wave c, and represent the
wavelength and period of the waves

3.1.2 Wave-Generating Mechanism in Air

Air may not sound like an elastic medium. So, in this section, we discuss what
causes elasticity in air. Here we analyze the dynamics in one dimension to make
the analysis easily visible.
In short, we can say as follows: The pressure gradient is proportional to the
density gradient (gas law). The density gradient is the displacement gradient.
Therefore, the pressure gradient is proportional to the displacement gradient.
This is how Hooke’s law applies to air. Below, we will discuss these dynamics
quantitatively.

Fig. 3.2 Change in air density due to pressure gradient

For simplicity, let’s consider the dynamics in one dimension. Figure 3.2
illustrates the situation where a volume of air expands along the x-axis. Before
the expansion, this volume has a cross-sectional area of S and a short length
along the x axis. Due to the pressure dependence on x, the cross-sectional plane
at is displaced more than that at ( ). Since we
consider a one-dimensional case, the cross-sectional area is unchanged during
the expansion. In the following sections, we discuss this phenomenon in detail.

Density change due to volume expansion

Let be displacement as a function of coordinate variable x and be
the length of the volume along the x-axis before the expansion. Thus, the
volume before the expansion ( ) and after (V) are given, respectively, as
follows.

(3.1)
(3.2)
So, the volume change becomes,

(3.3)
Since the process is an expansion, the mass of the volume is unchanged while
the volume increases. Therefore, we can express the density before the
expansion ( ) and after ( ) as follows.

(3.4)

(3.5)
Here m is the mass. From (3.4) and (3.5), we find the following expression for
the change in the density.

(3.6)
Now consider the infinitesimal limit where we can express and
assume . In this case, we can rewrite (3.3) and (3.6) as
follows.

(3.7)

(3.8)
Using (3.1), (3.7) and replacing with dx, we obtain the following equation.
(3.9)
In the infinitesimal limit, we can express as follows.

(3.10)
From (3.4), (3.8), (3.9), and (3.10), we obtain the following equation.

(3.11)
Gas law
From the ideal-gas law [1], we know that pressure and density are related
as follows.

(3.12)

(3.13)
Here, P is the pressure of n moles of air, R is the gas constant, T is the
temperature, and M is the molar mass. Using (3.12) and (3.13), we can express
the pressure as follows.

So, P is a function of , . Noting that the density change is small,

i.e., we can express the pressure after the expansion as follows.
(3.14)
From (3.14), we obtain the following relationship between the pressure
change and density change.

(3.15)
where

(3.16)
It is interesting to look into the relationship (3.15) based on Newtonian
dynamics. Dimensional analysis on (3.16) indicates that K is in (N/m )/(kg/m
) = (m/s) . Thus (3.15) reads (N/m ) =K (m/s)(m/s) (kg/m ).
Interpreting that the product of (m/s) and (kg/m ) as a momentum density (
) and the other (m/s) as the flow velocity of the momentum density, we can
view the right-hand side of (3.15) as the unit volume of air having momentum
passing through a plane perpendicular to its flow with the velocity
expressed by the other (m/s).
Figure 3.3 illustrates the concept. A momentum density (momentum per
unit volume) passes through a plane. Here v is the particle
velocity. Since the particle velocity has a gradient , the
momentum of the unit volume changes as it passes through the plane. Since this
is a change in momentum over time, a force is involved in the process
(Newton’s second law). The left-hand side of (3.15) represents this force.
This phenomenon is similar to the so-called optical radiation pressure [2],
in which optical radiation exerts pressure on a reflector. When light is reflected
by a reflector, the momentum carried by the photons changes its direction.
Consequently, each of these photons experiences a momentum change from
to . Here h is Planck’s constant and is the wavelength of
the light. This net change of causes the exertion of force on the mirror.
The unit of Planck’s constant is J s = (kg m/s )(ms) = kg (m /s) and is in
m. Thus is in kg(m /s)/m = kg(m/s), i.e., it represents the momentum
change. Since this change occurs during the contact time with the mirror
surface, its rate is kg (m/s)/s = N. In other words, we can explain this
phenomenon using the impulse-momentum theorem. Although this force is in
N, not N/m , conventionally it is called radiation pressure.
Radiation pressure is observed in an optical resonator of high intra-
resonator power configured with suspended resonator mirrors employed by a
gravitational wave detector [3]. The high optical momentum pushes the
resonator mirror out of position, causing the optical resonator to lose the
resonant condition. Consequently, the intra-resonator optical power decreases,
reducing the radiation pressure onto the mirror and making the resonator regain
the operational configuration. Once the resonator operates, the intra-resonator
optical power increases, resulting in high radiation pressure and causing the
same problem to repeat. Figure 3.3 illustrates the dynamics. Naively, you may
understand this phenomenon by imagining that you apply a water jet from a
hose onto a light plate suspended with a string (like a pendulum). While the
water jet hits the plate perpendicularly (with the normal incident), the plate is
pushed. As the plate swings away, the pushing effect by the water jet reduces
and gravity pulls back the plate making the angle of incidence back to normal.
This swinging motion repeats. (This analogy does not represent the resonant
effect observed in a gravitational wave detector but illustrates the gist of the
effect.)
Acoustic waves can cause radiation pressure in conjunction with
deformation in the reflecting object. Cantrell [4] discusses a detailed analysis of
this topic.

Fig. 3.3 a Momentum density change causes pressure . b Radiation pressure pushes the
resonator mirror. Photon changes its direction as reflected on the mirror surface flipping the sign of the
momentum. h: Planck’s constant; : wavelength of light

Substitution of (3.11) into (3.15) yields the following expression.

(3.17)
Interpreting on the right-hand side as the one-dimensional
expression of , we find that (3.17) is equivalent to Hooke’s law (2.29).

From the comparison of (2.29) and (3.17), we find the following

relationship between K and B.

(3.18)
Since B is in N/m and in kg/m , we find that the quantity K has a
dimension of velocity squared as (N/m )/(kg/m ) = (kg m/(s m ))/(kg/m )
= (m/s) . In fact, as we discussed in the last chapter (see (2.70), the square root
of the ratio of an elastic constant over density is the phase velocity. So, we can
identify K as the square of the phase velocity of the acoustic wave in air that we
will derive shortly.
Equation of motion
Since represents the pressure around the plane at x, we can interpret
the quantity as the net force per unit cross-sectional area acting on the
volume. Thus, from Newton’s second law, we obtain the following equation of
motion that represents the one-dimensional dynamics due to pressure gradient
along the x-axis. See Fig. 3.4.
Fig. 3.4 Net force acting on volume Sdx due to pressure gradient

(3.19)
Here, S and dx are respectively the cross-sectional area and the thickness of
the volume under consideration, and we replaced with
. Thus, eliminating the common factor Sdx, we obtain the
following equation.

(3.20)
From (3.17) and (3.20), we can write the equation of motion (3.19) in the
following form.

(3.21)
Thus we obtain the following displacement wave equation.

(3.22)
From (3.22), we can derive a wave equation for the pressure. Differentiate
(3.22) with respect to x.

(3.23)
Use (3.17) in (3.23)
(3.24)
Thus, we obtain the following pressure wave equation.

(3.25)

3.1.3 Decaying Pressure Waves in Air

Using (3.18), we can rewrite (2.64) in the following form, which we can
interpret as the equation of motion for the unit volume represented by density
.

(3.26)
By dividing by and using we obtain as follows.

(3.27)
Now take the divergence of (3.27) and obtain the volume expansion
(compression) wave equation.

(3.28)
By substituting (2.29) we obtain the pressure wave equation.

(3.29)
Remembering that represents the square of the phase velocity
(see (3.18)), we find (3.29) takes the general form of decaying elastic wave Eq.
(2.69) we derived in the last chapter.
3.2 Sound We Hear
3.2.1 Our Ear
Structure of our ear
Our ear consists of the outer ear, the middle ear, and the inner ear [5]. Figure 3.5
illustrates the structure of our ear. The outer and middle ears are filled with air
and separated by the eardrum. The middle ear consists of three bones called the
malleus, incus, and stapes. The middle and inner ears are separated by a tissue
membrane called the oval window, which connects the stapes and the duct
called the vestibular duct. As shown in Fig. 3.6, the vestibular duct lies parallel
to the cochlear duct, separated by Reisnner’s membrane. On the opposite side of
the cochlear duct, another duct called the tympanic duct lies. The cochlear and
tympanic ducts are separated by the basilar membrane, which is connected to
the auditory nerves via hair cells. The tympanic duct is terminated with the
round window (another membrane) on the middle ear side.

Fig. 3.5 Anatomy of ear. The eardrum separates the outer and middle ear and the oval window separates
the middle and inner ear. The middle ear consists of three bones, the malleus, incus, and stapes. The stapes
is attached to the oval window. The cochlea has a spiraled, hollow structure as the top-right insert
illustrates. (Illustration courtesy of Luc Allain)
 In short, our brain perceives sound by the following mechanism. The
sound we capture vibrates the eardrum. The three bones in the middle ear
transfer this vibration to the oval window. These bones act as an impedance-
adjusting mechanism to match the low impedance of air to the high impedance
of the fluid in the inner ear [6]. The vibration of the oval window moves the
fluid that fills the vestibular duct. As the oval window moves in toward the
vestibular duct, the round window moves out from the tympanic duct. These
movements generate mutually out-of-phase motions of the fluids in the
respective ducts, which causes the basilar membrane to vibrate. The vibration of
the basilar membrane causes the hair cells to experience oscillatory
displacement, which generates electric impulse signals that are transmitted
through the auditory nerves to the brain. We can say that our ear-brain system
converts a mechanical acoustic signal to an electric signal like a microphone.

Fig. 3.6 Anatomy of the cochlea. The top right insert is a cross-sectional view of the vestibular, cochlear,
and tympanic ducts. These three ducts inside the cochlea contain fluids. The bottom insert is a zoomed-in
illustration of the basilar membrane. The pair of dashed lines indicate the opposite directions of fluid
motions. The basilar membrane picks up the vibration of the fluid at a specific frequency depending on its
location from the base to the apex. (Illustration courtesy of Luc Allain)

Frequency of sound we hear

The frequency band that humans can typically hear ranges from 20 Hz to
20 kHz [7]. In speeches, the frequency of the vowels ranges from 250 Hz to 2
kHz, and those of voiced consonants and unvoiced consonants range from 250
Hz to 4 kHz, and 2 kHz to 8 kHz, respectively [8]. The frequency of the piano
ranges from 27.5 Hz to 4.186 kHz over the leftmost to the rightmost piano keys
[9]. The frequency of scratchy sound ranges from 2 kHz to 5 kHz [10]. Table
3.1 lists these frequency ranges.
Table 3.1 Frequency range of various types of sound. [8–10]
Type of sound Low frequency (Hz) High frequency (Hz)
Vowel 250 2,000
Voiced consonant 250 4,000
Unvoiced consonant 2,000 8,000
Piano 27.5 4186
Scratchy sound 2,000 5,000

Human ears process the frequency of sounds in the following fashion [6].
The basilar membrane varies its width and stiffness along its length. It is the
narrowest and stiffest at the base (closer to the middle ear) end, and is the
widest and least stiff at the apex (the other end). Remember that in Chap. 2 we
discussed that the natural frequency of a spring-mass system increases in
proportion to the square root of the ratio of the spring constant (stiffness) over
the mass. The higher the stiffness and lower the mass, the higher the natural
frequency of the system. This structure makes the membrane’s vibration
resonant with the highest frequency near the base side and shifts the resonant
frequency towards the apex as the sound frequency lowers. The hair cell at each
location along the length of the basilar membrane dominantly processes the
sound at the corresponding frequency. Thus, complex sounds consisting of
various frequencies (e.g., speech) are resolved into component frequencies and
processed by the corresponding hair cells.

3.2.2 Sounds of Speech and Music

In this section, we look at some sample sounds of speech and musical
instruments and make quick analyses of the signal in the time and frequency
domain.

Sample speech signal

Figure 3.7 shows the Fourier spectrum of “a”, “ma”, and “sa” sounds spoken by
a female speaker. The left graphs are logarithmic plots of the sounds, and the
right plots are linear plots zoomed in near the lower ends of the spectra. As
indicated in the left graphs, the respective spectra contain the typical frequency
ranges shown in Table 3.1. The spectrum of the “sa” sound contains both the
vowel and unvoiced consonant frequency ranges.
 On the right graphs, the “ma” sound indicates somewhat higher peaks near
200 Hz and 400 Hz than the “a” sound, indicating that the “ma” sound is the
superposition of the vowel “a” and the voiced constant “m”. As compared with
the spectra of “a” and “ma” sounds, the “sa” sound has more high-frequency
components, which is obviously due to the inclusion of the unvoiced consonant
“s”.

Fig. 3.7 Fourier spectra of “a” (vowel), “ma” (voiced consonant plus vowel), and “sa” (unvoiced
consonant plus vowel)

Figure 3.8 is another sample speech signal in the time domain (upper) and
the Fourier spectrum (lower). A female speaker (left) and a male speaker (right)
pronounce “chocolate”. The three envelopes observed in the time-domain
signals seem to represent the three syllables “cho”, “co”, and “late”. The female
speaker’s spectrum shows higher amplitude than the male spectrum in the
frequency range of 250 Hz–1000 Hz. This is because females have higher
fundamental voice frequency than males. The frequency ranges marked
indicate the typical fundamental voice frequency of 165 Hz–255 Hz (female)
and 85 Hz–180 Hz (male) [11, 12]. The spectral peaks in the further lower range
seem to represent syllabic and fluctuation characteristics of the speech [13].
Fig. 3.8 Time (upper) and frequency (lower) domain signals of sound “chocolate” spoken by a female
(left) and male (right). indicates typical fundamental voice frequency

The spectral shape in the frequency range above the fundamental voice
frequency results from the resonance and filtering characteristics of the vocal
tract [14] (the area from the nose and the nasal cavity down to the vocal cords)
[15]. Since the filtering characteristics depend on the shape and size of the vocal
tract [16], the same sound spoken by different people with the same
fundamental voice frequency sounds differently.
Figure 3.9 zooms in the three enveloped sections of the “chocolate” sound
by the female shown in Fig. 3.8 (left). Here the top row shows the time domain
signals of the sections, and the bottom row plots the Fourier spectrum in a
frequency range where most of the peaks appear in the respective envelopes.
These three envelopes correspond to the sound “cho”, “co”, and “late”. The
typical female’s fundamental voice frequency (165 Hz–255 Hz) is indicated for
the three envelopes. It is interesting to note that in the spectra of the “cho” and
“late” sounds, the fundamental frequency range covers prominent peaks. In the
case of “co” sound, higher peaks appear in the frequency range lower than the
fundamental frequency. The spectral shape in this low-frequency range seems to
represent the syllabic and fluctuation characteristics of the speaker. The time
domain signal clearly indicates that when the speaker pronounces “co”, the
amplitude fluctuates slower than the other two envelopes.
Fig. 3.9 Time (top) and frequency (bottom) domain signals of three envelopes of sound “chocolate”
pronounced by a female speaker. indicates typical female’s fundamental voice frequency

Figure 3.10 is the male speaker’s version of Fig. 3.9. Similar to the female
case, the fundamental voice frequency covers prominent peaks in the Fourier
spectrum. Also, the sound “co” has high peaks in the frequency range lower
than the fundamental frequency. A comparison of the spectra of “co” and “late”
sounds indicates that the energy of the latter is slightly weighted on the higher
frequency side. It is possible that this observation can be explained by the fact
that the fundamental frequency of “a” is higher than “o” [17].
Fig. 3.10 Time (top) and frequency (bottom) domain signals of three envelopes of sound “chocolate”
pronounced by a male speaker. indicates typical male’s fundamental voice frequency

Sample sound of music

Figure 3.11 shows the time domain (the left column) and frequency spectrum
(middle and right columns) of c2 sound generated by an electric base (upper
row), and an acoustic piano (lower row). The middle and right columns present
the same frequency spectrum at different frequency ranges.
The frequency of the c2 sound is 65.4 Hz [18]. The right column of Fig.
3.11 indicates that both musical instruments show a spectral peak at 65.4 Hz. At
the same time, they show overtones [19] (harmonics). While the bass shows a
higher peak at the fundamental frequency of 65.4 Hz, the acoustic piano
indicates that the peak at the second harmonic at 65.4 2=130.8 Hz is higher
than the fundamental.
Fig. 3.11 Comparison of sound from a Bass and b Acoustic piano. Each instrument is played at a pitch
corresponding to the piano c2 key. Time domain (left), Frequency domain (middle), and zoomed-in
Frequency domain (right) signals

The pattern of the overtone spectrum is unique for each musical

instrument, and it determines the characteristic sound. The mechanism of sound
generation determines the fundamental frequency. For instance, if you place
your finger at different positions on the string of a guitar, the pitch [20]
(frequency) changes. However, it is the body of the instrument that creates the
timbre. Compare the piano and guitar. Both are string instruments, and the
length of the string determines the fundamental frequency. Their distinctively
different bodies make their unique timbre. Figure 3.11 illustrates this situation
literally.
The generation of overtones results from the resonance of the standing
wave. From this standpoint, we can say that musical instruments are resonators.
Remember the discussion under “standing wave” in Sect. 1.2.3. The acoustic
wave generated by the sound-generation mechanism goes back and forth inside
the instrument, making standing waves. Here, its shape (the structure)
determines the characteristics of the standing wave. Some generate higher
frequency components with a high decay constant, whereas others have sound
lasting longer. Attack, decay, and sustain are the terminologies that characterize
the temporal characteristics of musical sound. The attack represents the rising
time to the maximum amplitude, the decay describes how quickly the sound
fades out, and the sustain indicates how long the sound stays at a constant
amplitude. (In Fig. 3.11, for example, the bass exhibits a longer decay and
shorter sustain than the piano).
If instruments have similar shapes, their length determines the fundamental
frequency. For instance, the sound of the flute is lower than the piccolo, and the
violin produces a higher pitch than the viola. The resonant wavelength
condition 1.70 precisely describes the situation. The longer the resonator length,
the longer the fundamental wavelength, hence lower the frequency.
3.3 Transmission of Audible Sound Waves
Sending an acoustic signal such as a speech over a long distance in the form of
acoustic waves is impractical. The speaking person needs to speak very loudly
and the listener must have a very good year even if they are apart by 100 m or
so. It is more efficient to convert the sound signal into an electric signal,
transmit the signal through air and receive it with an antenna.
The size of the antenna is proportional to the wavelength, i.e., inversely
proportional to the frequency of the receiving electromagnetic wave. The
optimal length for any antenna is around 1/4 of the wavelength it is receiving.
This type of antenna is known as a quarter antenna [21]. The audible frequency
to humans is 20 Hz to 20 kHz, which is in wavelength of 1.5 m to 1.5
m. Building an antenna of a quarter of these lengths in size is not
practical.
This is one of the purposes of sending sound information on a carrier
frequency that is high enough to make a practical antenna. For instance, if the
carrier frequency is 500 MHz, the wavelength becomes (m/s)
(1/s) = 60 (cm), i.e., its quarter is 15 (cm). In this scheme, the sound
information, the signal, is coded in the transmission signal in the form of
modulation. Depending on how to modulate, the modulation method is referred
to as amplitude modulation, phase modulation, and frequency modulation. In
this section, we consider amplitude modulation [22] and frequency modulation
[22], along with demodulations [23].
Amplitude modulation
Consider that we want to transmit signal S(t) as an amplitude-modulated
signal with carrier frequency . We can perform this operation by multiplying
the signal function S(t) to the carrier (a sinusoidal function oscillating at the
carrier frequency). We assume that S(t) has been converted from an acoustic
wave to an electric signal. We will discuss this sound-to-electric signal
conversion in Chap. 5.

(3.30)
Here is the angular frequency of the carrier.
Below we continue the discussion by replacing S(t) with ,
where is the angular frequency of the modulation signal. In other words,
we simplify the problem by assuming that the signal is a cosine function with a
single, constant frequency. The use of for the signal S(t) does
not lose generality because according to Fourier’s theorem, we can express S(t)
with a series of cosine and sine functions. We can view as a
term of the Fourier series. Note that a term of the sine series can be expressed
by the cosine counterpart of the same component frequency by shifting the
phase by ( ). In the process of
modulation and demodulation discussed below, we can repeat the same
procedure for other terms in the Fourier series as the term.
Let be the product of , and .

(3.31)
Using the following mathematical identity, putting and

(3.32)
we can express in the following form.

(3.33)
Here we assume that the , which is the condition generally
used for amplitude modulation. The necessity of this condition will be clarified
shortly when we discuss amplitude demodulation.
Fig. 3.12 Signals in time (left ) and frequency (right) domains. a carrier , b modulated signal

, and c signal

Figure 3.12 illustrates the signal , the carrier ,

and the modulated signal . The left column shows the carrier,
amplitude-modulated signal, and original signal along with demodulated signal
(from top to bottom). The right column shows the Fourier spectrum of the time
domain signal of the same row. The carrier’s frequency is 500 Hz (the period is
0.002 ms) and the frequency of the signal is 200 Hz (period of 5 ms), as Fig.
3.12 indicates. Notice that the carrier is not visible in the amplitude-modulated
signal in the time domain as the carrier frequency is beyond the resolution on
the scale of the signal. The Fourier spectrum of the amplitude-modulated signal
exhibits the carrier frequency of 500 kHz at the center and two sidebands at 500
kHz ± 200 Hz (the signal frequency). These sidebands correspond to the two
terms on the right-hand side of (3.33).
In the above case, the signal is fully modulated. It is possible to modulate
partially using the modulation depth M.

(3.34)
 Figure 3.13 illustrates the time series of modulated signals with different
modulation depths. Notice that as the modulation depth decreases, the
amplitude modulation becomes less prominent in the time-domain signals (the
left column), and the height of the sidebands decreases in the frequency domain.

Fig. 3.13 a ,b , and c

Amplitude demodulation
We need to retrieve the signal on the receiver side. There are
multiple methods to demodulate amplitude-modulated signals [24]. The
following operation of demodulation consisting of signal mixing
(multiplication) and low-pass filtering is one of them.
Consider multiplying the amplitude-modulated signal and the
carrier .

(3.35)
Here LF under the arrow stands for “low-pass filtering”, which eliminates
the high-frequency terms from the second line of (3.35).
 Frequency modulation
Frequency modulation is another common method to transmit audio
signals. In this case, the frequency of the carrier is modulated by the audio
signal. Let and be the amplitude and frequency of the carrier and s(t)

be the audio signal. We can express the frequency-modulated signal

as follows.

(3.36)
For the purpose of understanding the principle of frequency modulation
and demodulation, consider a simple case where the audio signal is a cosine
function.

(3.37)
In this case, the frequency-modulated signal can be expressed as follows.

(3.38)
Here, is called the modulation depth.

(3.39)
Figure 3.14 shows a sample of frequency-modulated signal in the time and
frequency domains. Here the carrier frequency is 500 Hz and the audio signal is
a cosine function of frequency 100 Hz. Figure 3.14a1, a2 indicate that the
carrier oscillates at the single frequency of 500 Hz. On the other hand, Fig.
3.14b1 reveals that the frequency-modulated signal changes the oscillatory
pattern with a periodicity of 10 ms; every 10 ms, the frequency alternates
between the high (dense) and low (sparse) oscillations. The periodicity of 10 ms
is equivalent to the modulation frequency of 1/(10 10 ) = 100 Hz.
 Figure 3.14b2 indicates that the Fourier spectrum of the frequency-
modulated signal has the central peak at the carrier frequency and the side lobes
at ± 100 Hz and ± 200 Hz from the central peak. In principle, these sidebands
appear at integer multiples of the signal frequency (in this case 100 Hz) away
from the carrier frequency on both sides. The greater the modulation depth, the
higher the sidelobe peaks towards the far ends of the spectrum.

Fig. 3.14 Frequency modulation time signal (left) and Fourier spectrum (right). The top graphs are for the
carrier and the bottom graphs are for the frequency-modulated signal

Frequency demodulation
Multiple methods are available to retrieve the audio signal on the receiver
side by demodulating the frequency-modulated signal [25]. Here, we discuss
one of them in which we retrieve the signal (modulation) frequency as a pattern
appears as amplitude modulation. An envelope detection technique [26] can be
used to retrieve the signal frequency.
Differentiate with respect to time.

(3.40)
In the case of the simple audio signal (3.37), the result of the
differentiation takes the following form.
(3.41)
In going to the second line in (3.41), we used (3.39).
We can view (3.41) as the is amplitude
modulated by the term . In other words, this sine
function’s amplitude oscillates in the range of .

Fig. 3.15 a Time derivative of the frequency-modulated signal. b The frequency-modulated signal, c
Fourier spectrum of a

Figure 3.15a, b compare the time-differentiated and the original frequency-

modulated signals on the same time axis. Notice that when the frequency-
modulated signal has a high frequency (where the oscillation is dense), the
amplitude of the time-differentiated signal has a higher amplitude. This
observation is something we can easily understand by examining (3.41); the
frequency of the modulated signal becomes highest when t in
 makes swing the largest. With the
same t, the cosine term in swings the largest from the
constant value .

References

1. 1. Tipler PA, Mosca G (2008) Physics for scientists and engineers, 6th edn. W. H. Freeman and Co.,
New York, Ch. 17, p 569
2. 2. Ma D, Munday JN (2018) Measurement of wavelength-dependent radiation pressure from
photon reflection and absorption due to thin film interference. Sci Rep 8:15930. https://doi.org/10.
1038/s41598-018-34381-z
ADSCrossref
3. 3. Goetz E, Kalmus P, Erickson S, Savage RL Jr, Gonzalez G, Kawabe K, Landry M, Marka S,
O’Reilly B, Riles K (2009) Precise calibration of LIGO test mass actuators using photon radiation
pressure. Class Quantum Gravity 26(24):245011
ADSCrossref
4. 4. Cantrell JH (2018) Acoustic radiation pressure, NASA Technical Report, NASA/TM–2018-
219806
5. 5. Balkany TJ, Brown KD (2017) The ear book. Johns Hopkins University Press, Baltimore
Crossref
6. 6. Purves D, Augustine GJ, Fitzpatrick D, Katz LC, LaMantia AS, McNamara JO, Williams SM
(eds) (2023)Neuroscience, 2nd edn. Sinauer Associates, Sunderland, MA, USA. The Middle Ear.
Available from: https://www.ncbi.nlm.nih.gov/books/NBK11076/ (accessed on July 19, 2023)
7. 7. Quam R, Martínez I, Lorenzo C, Bonmatí A, Rosa M, Jarabo P, Arsuaga JL (2012) Studying
audition in fossil hominins: A new approach to the evolution of language? In: Jackson MK (ed)
Psychology of language. Nova Science Publishers Inc, NY, USA, pp 1–37
8. 8. Ecophon, Generating and understanding speech
9. 9. http://shakahara.com/pianopitch2.php (Accessed on July 20, 2023)
10. 10. Science ABC (2023) Why do certain sounds make our skin crawl? https://www.scienceabc.
com/eyeopeners/why-do-certain-sounds-make-our-skin-crawl.html (Accessed on July 20, 2023)
11. 11. Oliveira RC, Gama ACC, Magalhães MDC (2021) Fundamental voice frequency: Acoustic,
electroglottographic, and accelerometer measurement in individuals with and without vocal
alteration. J Voice 35(2):174–180
Crossref
12. 12. Mahdi JF (2015) Frequency analyses of human voice using fast Fourier transform. Iraqi J Phys
13(27):174–181. https://en.wikipedia.org/wiki/Voice_frequency, https://www.studysmarter.us/
explanations/english/phonetics/fundamental-frequency/
13. 13. Edwards E, Chang EF (2013) Syllabic ( 2-5 Hz) and fluctuation ( 1-10 Hz) ranges in
speech and auditory processing. Hear Res 305:113–134
Crossref
14. 14. https://pronuncian.com/the-vocal-tract (Accessed on July 20, 2023)
15. 15. Facts about speech intelligibility, https://www.dpamicrophones.com/mic-university/facts-about-
speech-intelligibility
16. 16. Shadle CH (2006) Phonetics, acoustic. In: Brown K (ed) Encyclopedia of language and
linguisitcs, 2nd edn. Elsevier
17. 17. Maurer D (2023) Acoustics of the Vowel: Preliminaries, New ed. Material Part III, (Peter Lang
AG, Internationaler Verlag der Wissenschaften, 2016) available on www.oapen.org and www.
peterlang.com (Accessed on July 23, 2023)
18. 18. https://pages.mtu.edu/~suits/notefreqs.html (Accessed on July 23, 2023)
19. 19. Overtones B (2023) https://www.britannica.com/science/sound-physics/Overtones (Accessed
on July 23, 2023)
20. 20. Electronic Musician (2023) Understanding the difference between pitch and frequency. https://
www.musicradar.com/how-to/understanding-the-difference-between-pitch-and-frequency (Accessed
on July 23, 2023)
21. 21. Latorilla E (2000) Practical antenna design, online edn. LEDF Media (Accessed on July 23,
2023)
22. 22. Hagen JB (2009) Amplitude and frequency modulation. In: Radio-frequency electronics, 2nd
edn. Cambridge University Press, Cambridge, UK, pp 54–66. https://doi.org/10.1017/
CBO9780511626951.007
23. 23. Hagen JB (2009) Demodulators and detectors. Radio-frequency electronics, 2nd edn.
Cambridge University Press, Cambridge, UK, pp 227–241
Crossref
24. 24. Mouser Electronics (2023) How to demodulate an AM waveform. In: Practical guide to radio-
frequency analysis and design, Ch. 5. https://www.allaboutcircuits.com/textbook/radio-frequency-
analysis-design/radio-frequency-demodulation/how-to-demodulate-an-am-waveform/ (Accessed on
July 24, 2023)
25. 25. Mouser Electronics, How to demodulate an FM waveform. In: Practical guide to radio-
frequency analysis and design, Ch. 5. https://www.allaboutcircuits.com/textbook/radio-frequency-
analysis-design/radio-frequency-demodulation/how-to-demodulate-an-fm-waveform/
26. 26. Carr JJ (2000) Radio receiver basics. In: The technician’s EMI handbook, Ch. 16. Elsevier Inc.,
pp 163–195. https://www.sciencedirect.com/science/article/abs/pii/B9780750672337500168
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024
S. YoshidaFundamentals of Acoustic Waves and ApplicationsSynthesis Lectures on Wave Phenomena in
the Physical Sciences
https://doi.org/10.1007/978-3-031-48200-7_4

4. Propagation of Acoustic Waves in Solids

Sanichiro Yoshida1
(1)
Department of Chemistry and Physics, Southeastern Louisiana
University, Hammond, LA, USA

Sanichiro Yoshida
Email: [email protected]

In this chapter, we aim to discuss the propagation of sound waves in solids,

clarifying the dynamical difference in elastic behaviors between air and solids.
After briefly reviewing the theory of elasticity, we derive the equation of motion
for isotropic solid media. We derive the wave equations and discuss the
propagation of acoustic waves in solids under various boundary conditions. The
discussions include some nonlinear waves, such as surface waves.
4.1 Acoustic Wave as Solutions to the Equation of Motion
Sound propagating through a solid is an elastic wave. The elasticity of solids
causes oscillatory motion of the lattice and the oscillation travels as a wave. In
Chap. 2, we derived longitudinal and transverse wave equations from the
equation of motion. Here we combine the longitudinal and transverse effects
(i.e., the normal and shear forces) and make a similar discussion to Chap. 2. We
can start with the equation of motion for a unit block in an elastic medium. We
restrict our discussions here to isotropic materials.

4.1.1 Equation of Motion

The equation of motion for a unit volume of an isotropic, elastic medium can
be put as follows. (See Appendix E for the derivation of this equation.)
(4.1)
Here and are Lamé’s first and second parameters.
Equation (4.1) leads to the equations of compression wave and rotation
wave as follows. Take the divergence of (4.1) switching the order of and
operation for the first term on the right-hand side and using the identity
for the second term. These operations yield the following
wave equation of (the compression wave equation).

(4.2)
Equation (4.2) indicates that an elastic compression wave travels at the
following phase velocity.

(4.3)
Similarly, taking the curl of (4.1), we obtain the following equation of
rotation wave.

(4.4)

Here, rotation . (4.4) indicates that an elastic rotation

(deformation) wave travels at the following phase velocity.
(4.5)

Since , the compression wave is called the Primary

Wave (P-wave) and the rotation wave is called the Secondary Wave (S-wave).
Scalar and vector potentials
It is convenient to express the displacement vector using two potentials [1,
2].

(4.6)

(4.7)
Here and are called the scalar potential and vector potential of
elastic dynamics. Equations (4.6) and (4.7) are known as Helmholtz
decomposition [3]. As will be clarified shortly, represents compression
(spring) potential energy and the rotational potential energy of elasticity.
Substitution of (4.6) into (4.1) yields the following equation.

(4.8)
Here, in the last step of (4.8) the two terms in the first curly bracket are
combined and (4.7) is used in the second curly bracket on the right-hand side.
For (4.8) to hold arbitrarily, it follows that the following conditions hold.

(4.9)
(4.10)
Equations (4.9) and (4.10) are wave equations for the two potentials
indicating that and travel at the following velocity.

(4.11)

(4.12)
Notice that wave velocities (4.11) and (4.12) are identical to wave
velocities (4.3) and (4.5), respectively. These identities have physical meanings
as will be discussed in the following section. The wave and wave are
also called the Primary wave and Secondary wave, respectively.

Split (4.6) into two terms.

(4.13)
where

(4.14)

(4.15)
Equation (4.14) indicates that results from the differential elastic
energy, as the following argument explains. Dimensional analysis tells us is
in [m ]. If we multiply stiffness K [kg/s ], the dimension of becomes
energy as [m kg/s ] [kg m/s m] [Nm] [J]. Knowing that the
dimension of stiffness K is in [N/m] [kg/s ] and the elastic energy is

, we can interpret that is the elastic energy normalized to the unit of

[m ] (by dividing it by the stiffness K kg/s ).

Fig. 4.1 The product of stiffness K and vector H represents torque associated with the rotational wave.
represents the z-component of the displacement due to all torques along the circumference
around the reference point

Dimensional analysis on (4.15) indicates that potential is in [m ],

because is in m and is in 1/m. By multiplying the factor K defined
above for the other potential , we find the dimension of to be [kg m/s
m]=[N m]. Being associated with the rotational (deformation) wave via
(4.10), the dimension [N m] can be interpreted as representing torque.
Material-scientifically, we can interpret the above torque mechanism due
to and its association with as follows. Figure 4.1 schematically
illustrates the situation where a reference point in an elastic material is
surrounded by four infinitesimal segments. The four segments are on the
circumference of an imaginary circle set around the reference point. The plane
of the circle is parallel to the xy plane and each segment is represented by a disc
perpendicular to the circumference. The external force, f, generates torque on
each segment as illustrated in Fig. 4.1. Here the torque vector is in the tangential
direction of the imaginary circle, and the torque on each disc is clockwise when
viewed from the tail of the torque vector. Each torque vector is defined by
the local position vector representing the radius of the disc and the external
force vector, and related to with the factor K as follows.

(4.16)
The torques from the four segments all together generate upward
displacement at the reference point and downward displacement outside the
imaginary circle. Thus, we can interpret (4.15) as representing the displacement
associated with the rotational (torque) energy stored in the material surrounding
the point of interest.
The above interpretations indicate that the -wave (P-wave) and -
wave (S-wave) are generated by different mechanisms and therefore they
propagate at mutually different wave velocities. At the same time, however,
both wave-generation mechanisms are due to the elasticity of the material and
therefore they can be converted mutually (see Sect. 4.2.2 under Reflection at
the free surface with various incident waves). In the following sections, we
discuss the dynamics of these waves.
4.2 Propagation of Acoustic Waves in Solid Under Various
Conditions
Equations (4.2) and (4.4) indicate that an acoustic wave propagates through an
elastic medium in two forms. The first is the compression wave (P-wave) that is
given as a solution to (4.2) and travels at the wave velocity given by (4.3). The
second is the rotational wave (S-wave) that is given as a solution to (4.4) and
travels at the wave velocity given by (4.5). In this section, we discuss the
propagation of these two types of waves in various scenarios.

4.2.1 Plane Waves in an Elastic Medium

In general, a wave solution can be expressed as the product of functions that
represent amplitude and phase, respectively. In a Cartesian coordinate system,
let be the amplitude function and f be the phase function of a
displacement wave .

(4.17)
Here is the frequency, is the propagation vector and is the position
vector of coordinate point (x.y.z).
When the amplitude is a constant, independent of the spatial
coordinates, the value of is solely determined by the value of the phase
function f. When the spatial dependence of the phase function is linear, as is the
case of (4.17), the three-dimensional space that shares the same phase (called
the wavefront) is planar. If the amplitude is constant and the wavefront is
planar, the wave is called a plane wave.
A plane wave is an idealized solution to a wave equation. Due to
diffraction, a wave generally expands transversely to the direction of
propagation. At each point of the wavefront, the propagation vector is
perpendicular to the wavefront. Therefore, if a wave diverges or converges, the
wavefront is curved (curved inward for a divergent wave and outward for a
convergent wave when viewed behind the wavefront). So, a planar wavefront
does not exhibit a diffracting behavior. However, under various conditions, a
plane wave solution is a good approximation (e.g., the light wave from the sun
on the Earth). In addition, probably more importantly, the analysis of plane
wave solutions is very informative and helpful to understand the propagation of
acoustic waves.
Propagation vector
Above, I used the word “propagation vector” without explaining it. Since
this is an important concept in the following sections, take some time to
consider it. The phase function varies with time and space as . If
the phase function is sinusoidal, for instance, it varies in the range
as the phase varies with the period of . Since the phase
depends on both time and space, a temporal or spatial change can alter the
phase. If the phase change is an even integral multiple of the phase function
f takes the same value, and if the change is an odd integral multiple of , the
function takes the same absolute value with the opposite sign, and so on.
We can understand the concept of the wavefront most easily by freezing
the function with time. We can fix the time at without losing the
generality of the argument. With , the phase (argument) of the function
f is , and the equation of the plane
of a constant phase, , is given as follows.
(4.18)
Figure 4.2a illustrates this plane. Consider two points in the plane,
and . Let and represent the position
vectors of these two points, and a unit vector normal to this plane.
By definition, the following equations hold.

(4.19)

(4.20)
Using the coordinates expression, we can write as follows.

(4.21)
Using expression (4.21), make the scalar product of vector and
propagation vector .

(4.22)
Equations (4.19) and (4.20) were used in the last step of (4.22). Since
vector is in this plane and we can choose the second point
arbitrarily in the plane fixing the first point ,
derived by (4.22) tells us that the propagation vector
is normal to this plane, i.e., it is parallel the normal vector as .
Now consider another plane of constant phase . Obviously, this
plane is parallel to the plane of constant phase . Therefore, the propagation
vector is also normal to this plane. We can repeat the same argument by
adding a constant phase, and say that the propagation vector is always normal to
a plane of constant phase. Figure 4.2b illustrates the situation.
 The fact that the propagation vector of a wave is normal to the wavefront is
true even if the wavefront is curved. At a given point on the wavefront, we can
consider an infinitesimal plane (the plane tangential to the curved wavefront)
and apply the above argument. In conclusion, the propagation vector is always
normal to a wavefront.

Fig. 4.2 Wave front of a plane wave

Now we go back to the discussion of plane wave propagation. Naturally, a

plane wave solution can be found by solving the equation of motion (4.1). For
simplicity, we solve the equation of motion for a two-dimensional case. The
same arguments hold for three-dimensional cases. Define the vectors as follows.

(4.23)

(4.24)

(4.25)

(4.26)
Substitution of (4.24)–(4.26) into (4.17) yields the following expressions.
(4.27)

(4.28)
Since and are constant, we can differential the phase function to
find the derivatives of and .

(4.29)

(4.30)
Here represents the derivative of function f.
Thus, we obtain the following equations.

(4.31)

(4.32)

(4.33)
Similarly, the second-order derivatives are given as follows.

(4.34)
(4.35)

(4.36)

(4.37)

(4.38)
With (4.34)–(4.38), the derivative terms of (4.1) become as follows.

(4.39)

(4.40)

(4.41)
 Substitution of (4.39)–(4.41) into the equation of motion (4.1) yields the
following system of equations for and .

(4.42)

(4.43)
Here, the common term is eliminated. Equations (4.42) and (4.43) can
be expressed in the matrix form as follows.

(4.44)
In order for (4.42) and (4.43) to yield non-trivial solutions, the determinant
of matrix M must be null. Thus, we obtain the following equations.

(4.45)
Equation (4.45) indicates that for displacement in the form of (4.23) to
be a solution to the differential equation (4.1), one of the following conditions
must hold.

(4.46)
(4.47)
Since is the wave velocity, (4.46) and (4.47) can be interpreted as
representing the following wave velocities.

(4.48)

(4.49)
Wave velocities (4.48) and (4.49) are identical to those of the compression
wave (P-wave) (4.3) and rotation wave (S-wave) (4.5), respectively. Equations
(4.48) and (4.49) indicate that the displacement wave travels in an isotropic
elastic medium as a P-wave or S-wave. In the following sections, we find that
these two types of plane waves can coexist in elastic media and can be
converted from one type to the other.
Another feature I would like to point out is that neither velocity (4.48) or
(4.49) depends on the frequency. They simply depend on the material constants
, and . This means that an acoustic wave propagating through an
isotropic media does not exhibit dispersion. Later in this chapter, we observe
that acoustic waves can exhibit dispersion in an isotropic media when certain
boundary conditions are imposed. As we will see in the next section, a free
surface is one of such boundary conditions that cause acoustic waves to exhibit
dispersive behaviors. This type of dispersion can be used in characterizing
structural and mechanical properties of solid objects and is therefore important
in various acoustic techniques of nondestructive evaluation [10].

4.2.2 Plane Waves in a Medium with Free Surface

Consider a plane acoustic wave travels in an isotropic elastic medium towards
its infinitely large free surface in Fig. 4.3. Here we set the free surface on the zx
plane and use the word “free” to mean that the external force acting on the
surface exerted by the surrounding medium (e.g., air) is null. By assuming that
the external force is null at for all (x, z) coordinate points, we can
express this free surface condition as the traction vector is null
as follows.

(4.50)
Here represents the m component of the stress vector on a plane l. For
condition (4.50) to hold at all (x, 0, z) coordinate points, it follows that

(4.51)
Figure 4.4 schematically illustrates the situation where a plane wave is incident
to the free surface at an angle of incidence . Here is the propagation
vector where and are its x and y components. Under these conditions,
derive wave equations for the potential and . Assume that the angle of
incidence to the free surface is and that plane of incidence is parallel to the
xy plane. This assumption does not lose the generality of the argument. In the
below discussion, we refer to planes parallel to the zx-plane as a horizontal
plane and planes parallel to the xy plane as a vertical plane (Fig. 4.3).

Fig. 4.3 Acoustic wave travels towards the free surface at an incident angle of
Fig. 4.4 Plane wave incidence to a free surface. is the propagation vector

From the plane wave condition, we can set . (Since the

plane wavefront is parallel to the z axis the displacement components are
independent of z.) Thus, according to (4.6), we can relate the displacement
components to the potentials as follows.

(4.52)

(4.53)

(4.54)
 We are now in a position to solve wave equations for plane wave solutions
of and . Remembering the discussion made in Sect. 4.1, express the
wave equations as follows.

(4.55)

(4.56)
As observed in the above sections, the acoustic wave can take two modes
(the P-wave and S-wave). (4.52)–(4.54) indicate that and components
depend on and , whereas component depends on and
and independent of . In this situation, it is convenient to discuss the wave
dynamics for the following two types; (1) the wave that involves and
and (2) the wave that involves . We call the first type (1) the plane strain
wave because the displacement components lie in the plane of incidence. (The
plane strain is the state where the non-zero strain components act in one plane
only.) From (4.52) and (4.53) we know that this wave is a mixture of a P-wave
and S-wave because it has the and components. (The wave is a P-
wave, and the wave is an S-wave, as we discussed in the paragraphs near
(4.16).) Here, the S-wave consists of the displacement vector components in the
plane of incidence, above referred to as a vertical plane. So, we call the S-wave
that constitutes the plane strain wave the Shear Vertical (SV) wave. Figure 4.5
illustrates a P-wave (a) and SV-wave (b) that constitute a plane strain wave.
Fig. 4.5 and SV-wave incidence to a free surface. is the propagation vector

We call the second type (2) the Shear Horizontal (SH-wave) because it is a
shear wave whose oscillation is polarized in the direction parallel to the plane of
the free surface, above referred to as a horizontal plane. The SH wave is an S-
wave. Figure 4.6 illustrates an SH-wave traveling along axis.

Fig. 4.6 SH wave incidence to a free surface. is the propagation vector

Plane strain wave solution

In this case, the displacement vector has x and y components (see Fig. 4.3).
We can solve wave equations (4.55) and (4.56) for and using (4.52)
and (4.53). Remember that from the plane wave condition we know that
and therefore and are independent of z. Wave functions
and have harmonic time dependence. It is reasonable to assume that
they have a harmonic dependence on x as well. Thus, using the method of
separation of variables we can put and in the following form and start
solving the wave equations.

(4.57)

(4.58)
Substitution of (4.57) and (4.58) into wave equations (4.55) and (4.56)
yields the following differential equations for the functions f and .

(4.59)

(4.60)
where

(4.61)

(4.62)
Equations (4.59) and (4.60) indicate that the functions f and are also
harmonic functions of y. Thus, absorbing these dependencies of f and on y
in expressions (4.57) and (4.58), we obtain the following forms of the
potentials.
(4.63)

(4.64)
Here, before proceeding further, consider the physical meanings of
etc. in (4.61) and (4.62). In these equations, and represent the
wave numbers of the respective waves. Referring to (4.17) and the discussions
under “Propagation vector”, we can interpret and as the x and y
components of the propagation vector , and and as the
components of .

(4.65)

(4.66)
If we call the wavelength of the respective waves as and , we find
the following relations (See Fig. 4.3).

(4.67)

(4.68)

(4.69)
From the second and third relations, we find
(4.70)

(4.71)
The same argument holds for the other wave and

(4.72)

(4.73)
Using (4.70)–(4.73), we can rewrite (4.63) and (4.64) as follows.

(4.74)

(4.75)
At this point, we can consider the boundary conditions. First, express the
stresses in (4.51) in terms of strain using constitutive relations.

(4.76)

(4.77)
(4.78)
Above, we used the plane wave condition in the right-hand
side of (4.77) and (4.78). Substituting (4.52) and (4.53) into (4.76) and (4.77)
we obtain the following conditions for . Note that for the analysis of
the plane strain wave, the boundary condition (4.78) is irrelevant because the
wave does not have the z component, .

(4.79)

(4.80)
By substituting (4.74) and (4.75) into (4.79) and (4.80) and setting
, we obtain the following equations.

(4.81)

(4.82)
Here in (4.82), is defined as the ratio of to as follows.
(4.83)
Using , we can factor out the time-dependent term
from (4.81) and (4.82). In order to make these two equations hold for any x, it is
necessary that the space-dependent term be factored out as well. It follows that
the following condition holds.

(4.84)
From (4.70) and (4.72), we can interpret the condition (4.84) as the spatial
frequency along the free surface is the same for the and waves so that
the boundary conditions (4.79) and (4.80) hold.
With (4.84) we can rewrite (4.74) and (4.75) as follows.

(4.85)

(4.86)
The wave is called the compression wave (P-wave) and wave is
called Shear-Vertical (SV) wave because it is a shear wave confined in the
vertical plane.
Shear-Horizontal wave solution
The Shear-Horizontal wave is a shear wave oscillating parallel to the plane
of the free surface. In the present case, a Shear -Horizontal wave has only the z-
component of the displacement vector. We can repeat the same type of
argument and analyze the shear mode wave. In this case, the displacement
vector has only the z component. This means that of the expression of the three
components of the displacement vector (4.52)–(4.54) only the expression for
(4.54) is relevant. Thus, the shear mode wave is characterized by and
. Similarly to (4.58), we can put the solutions in the following forms.

(4.87)
(4.88)
Note that we use the plane wave condition in (4.87) and
(4.88).
Substitution of (4.87) and (4.88) into the wave equation (4.56) yields the
following differential equations.

(4.89)

(4.90)
where

(4.91)
Notice that unlike the plane strain case discussed in the preceding section,

we use only one spatial frequency to characterize and waves.

This is because both waves share the same velocity .
Equations (4.87), (4.88) and (4.91) let us put the solution and in
the following forms.

(4.92)

(4.93)
There is an additional difference between the plane strain and shear-
horizontal cases. Whereas the compression ( ) wave and Shear-Vertical (
) waves are uncoupled in the plane strain case, the two components of the
Shear-Horizontal wave are coupled via (4.7). Substitution of solutions (4.92)
and (4.93) into (4.7) gives us the following equation.

Hence,

(4.94)
Since (4.94) holds for any y it follows that

(4.95)

(4.96)
Thus we can rewrite solutions (4.92) and (4.93) as follows.

(4.97)

(4.98)
Similarly to (4.72) and (4.73) we can put the spatial frequencies using the
angle of incidence as follows.

(4.99)

(4.100)
With (4.99) and (4.100), (4.97) and (4.98) become
(4.101)

(4.102)
The fact that and are coupled is not surprising if we remember
that the SH-wave has only one component of the displacement vector, , to
oscillate. In other words, and cooperatively generate .
It may be a good idea to express the SH-wave in terms of . By
substituting (4.97) and (4.98) into (4.54), we can easily find that the
displacement component can take the following form.

(4.103)
where

(4.104)

(4.105)
Of course, the above solution satisfies the governing (wave) equation for
.

(4.106)
where
(4.107)
Now we consider the boundary condition. In the case of the SH-wave, the
relevant boundary condition at the free surface is (4.78). Expressing
with the potentials, (4.78) becomes

(4.108)
Setting (4.108) to null for we obtain the following condition.

(4.109)
Hence,

(4.110)
Equation (4.110) indicates that the free surface makes the forward-going
and backward-going SH waves have mutually opposite signs. In other words,
upon reflection, the SH-wave undergoes a phase shift of .
Reflection at the free surface with various incident waves
In this section, we discuss reflection at the free surface when the incident
wave approaches the surface from the inside of the elastic medium with an
oblique angle of incidence. The behavior at the free surface differs among the
three types of wave, i.e., the P-wave, SV-wave, and SH-wave, as discussed in
the preceding section. In the following paragraphs, we will briefly discuss the
reflection for each of these waves. For more detailed discussions of these topics,
see Ref. [4].
(1) Incident SH wave
This is the simplest case of the three wave types. Substitute the SH-wave
expression with the displacement (4.103) into the boundary condition (4.78).
(4.111)
It follows that

(4.112)
Condition (4.112) indicates that the SH-wave undergoes a phase shift of
on the reflection. This is easily understood by noting that the boundary
condition (4.51) forces the oscillation along the z axis to be null at the surface.
The situation is similar to the reflection of an electromagnetic wave whose
electric field is parallel to the conductive boundary. Being conductive, the total
electric field at the boundary must be null. When there is no transmitted wave,
the amplitude of the reflected wave negates that of the incident wave so this
condition holds. In the present case, the transmission of the acoustic wave can
be neglected because the acoustic impedance of the air is much less than the
elastic medium.
(2) Incident P-wave
In this case, it is convenient to use the and -wave expressions
(4.85) and (4.86). Since we consider that the incident wave is a P-wave and an
H-wave is an S-wave, the first term on the right-hand side of (4.86) must be
null. This term represents a wave traveling in the negative y-direction, i.e.,
toward the boundary (see Fig. 4.3). Thus, we can set in (4.86).

(4.113)

(4.114)
Here (4.114) represents the -component in the wave reflected at the
boundary. Using , and (4.84), we can
express the boundary conditions (4.81) and (4.82) as follows.

(4.115)
(4.116)
By solving (4.115) and (4.116) for and , we can evaluate the
reflection coefficients for the P- and SV-waves. Before proceeding further it is a
good idea to express term as follows. Remember
,

(4.117)
Here, in going through the second equal sign, (4.84) was used. Using
(4.117) the two boundary conditions (4.115) and (4.116) become as follows.

(4.118)

(4.119)
Thus,

(4.120)

(4.121)
Here and are related through Poisson’s ratio.
(4.122)

Fig. 4.7 Reflection coefficient as a function of incident angle for several Poisson’s ratios when the
incident wave is a P-wave

Figure 4.7 plots the reflection coefficient (4.120) and (4.121) as a function
of incident angle for some Poisson’s ratio. Note that the incident P-wave
generates the reflected SV-wave. This phenomenon is referred to as the mode
conversion. When the incident wave is an SH-wave, there is no mode
conversion as we discussed in the preceding section.
The insert in Fig. 4.7 shows the rays of the incident and reflected waves.
Notice that when the incident P-wave is converted to the S-wave, the angle of
reflection is smaller than the angle of incidence. We can explain this
phenomenon as follows. As indicated by (4.11) and (4.12), the P-wave has a
higher velocity than the S-wave. From (4.65) and (4.66) this means that
, and from (4.84), that . It follows that
.
From the law of reflection, the angle of reflection for the P-wave is the
same as the angle of incidence. These arguments all together indicate that no
matter how great the angle of incidence may be the incident P-wave is reflected
back to the elastic medium. This situation is contrastive to the case when an
incident SV-wave is converted to a P-wave, as discussed below.
 (3) Incident SV wave
We can repeat the same type of argument as the preceding section to
discuss the case when the incident wave is an SV-wave. In this case, we can put
in (4.85) as there is no forward-going component in the P-wave.
Consequently, we obtain the following expressions.

(4.123)

(4.124)
The boundary conditions (4.81) and (4.82) become as follows.

(4.125)

(4.126)
By solving (4.125) and (4.126) we obtain the following reflection
coefficients.

(4.127)

(4.128)
Figure 4.8 plots the reflection coefficients (4.127) and (4.128). Here the
right vertical axis indicates the reflection coefficient when the mode is
unconverted on reflection whereas the left vertical axis indicates the reflection
coefficient when the SV mode is converted to the P mode.
Fig. 4.8 Reflection coefficient as a function of incident angle for several Poison’s ratios when the incident
wave is a SV-wave

The insert in Fig. 4.8 illustrates the rays of the incident and reflected
waves. Note that in this case, unlike the case when a P-wave is converted to an
SV-wave, the angle of reflection is greater than the angle of incidence on the
SV- to P-wave conversion. This indicates that if we increase the angle of
incidence, at a certain point the angle of reflection reaches . Beyond this
point, according to (4.84) exceeds unity and it becomes impossible to
find the angle of reflection from this equation. The angle of incidence that
makes the angle of reflection be is called the critical angle . (See (2.
143).) When the angle of incidence is greater than , the resulting P-wave
becomes confined in the subsurface region of the elastic material and decaying
exponentially away from the surface. We will discuss this phenomenon in the
next section.

4.2.3 Surface Waves

At the end of the last section, we touched upon the existence of a wave confined
near the free surface in association with the SV- to P-wave mode conversion
occurring at an angle of incidence greater than a critical angle. In this section,
we discuss such surface waves in more general.
 Let be the critical angle for an SV-wave. By definition, the
corresponding angle of refraction for the transmitting P-wave is .
According to (4.84) the situation can be expressed for the critical angle and
angle as follows.

(4.129)

(4.130)
From (4.70) and (4.84) we know

(4.131)

(4.132)

(4.133)
Substituting (4.131)–(4.133) into (4.61) and using condition (4.130), we
find as follows.

(4.134)
Equation (4.134) indicates that when the angle of incidence is greater
than the critical angle, is not a real number. Remember that
originally appeared in the differential equation (4.59) for the f(y) part of the
potential . Expressing
 with a real number as , we can rewrite (4.59) as
follows.

(4.135)
We can solve (4.135) and obtain the following form of solution.

(4.136)
Thus, we can put the potential (4.113) in the following form.

(4.137)
Here we discard the second term in the parenthesis because it is not
physical as representing an exponentially increasing amplitude. Solution (4.137)
represents a wave traveling along the x-axis with the amplitude exponentially
decreasing with the depth from the surface. We call this type of wave surface
waves.
We can extend the same idea to the other differential equation (4.60) to put
the function in the following form.

(4.138)
It is clear that displacement wave characterized by the two potentials in the
form of (4.137) and (4.138) represents a plane, surface wave. Express the x and
y components of such a displacement wave and analyze its feature.

(4.139)

(4.140)
 The boundary conditions (4.79) and (4.80) give us the
following equations.

(4.141)

(4.142)
In going through the second equal sign in (4.142),
((4.83)) is used.
Here

(4.143)

(4.144)
So,

(4.145)
Use (4.145) to rewrite (4.142) as follows.

(4.146)
Using (4.137) and (4.138) to express and at with
and , we obtained the following set of equations from (4.141) and (4.146).
(4.147)

(4.148)
From (4.147) and (4.148), we obtain the following equality.

(4.149)
From the second equality in (4.149) it follows that

(4.150)
hence,

(4.151)
Now compare the amplitude of and to find
out what the actual displacement vectors look like at the surface. At
(on the free surface), (4.139) and (4.140) become

(4.152)

(4.153)
Use (4.149) in (4.152) and (4.153) to eliminate .
(4.154)

(4.155)
Here (4.151) was used in going through the second equal sign for (4.154)
and (4.155). From (4.143) and (4.144),

(4.156)

(4.157)

Since , it follows . Thus (4.155) indicates that

(4.158)
The type of surface wave expressed by (4.154) and (4.155) is known as the
Rayleigh surface wave [5]. Figure 4.9 illustrates a sample Rayleigh surface
wave traveling in the positive x-direction. The spatial periodicity is seen to
move to the right. Here, can be interpreted as the
wave velocity. (See the exponential term of (4.154) and (4.155).) Note that
Rayleigh surface waves are non-dispersive. The insert to the right illustrates the
elliptical pattern of the displacement vector as a function of time. Notice that
when the Rayleigh wave travels in the positive x direction, the trajectory of the
displacement vector along the elliptical line is counterclockwise.
The above discussions cover only the gist of the Rayleigh wave. For more
information about Rayleigh waves, see Refs. [6, 7].

Fig. 4.9 Rayleigh wave traveling in positive x-direction. The illustration on the right represents the
elliptical pattern of the displacement vector as a function of time

4.2.4 Waves Propagating in Plates

In this section, we extend the discussion on acoustic waves in an elastic medium
with a free surface that we made in Sect. 4.2.2. The elastic medium has two free
surfaces, instead of one. The governing equation (the wave equation) is the
same as the one-free surface case but the other free surface gives an additional
boundary condition, which makes the resultant displacement vector behave
considerably different from the one free-surface case. The most important
difference here is that the displacement wave exhibits dispersion. This feature
has important engineering applications [8]. The aim of this section is to
understand the mechanism that generates dispersion through some simple cases.
For more general discussions about the dispersion of waves in plates, see Ref.
[9, 11].
SH waves in a plate
 As is the case of the one free-surface case, we start with the SH-wave case
since it is the simplest. The governing equation is the same as (4.106).

As before, we consider a plane wave with the condition.

The solution can be put as follows.

(4.159)
where

(4.160)
Now continue the analysis with a plate of thickness 2b having infinitely
large free surfaces at . The free surfaces are parallel to the zx plane.
Figure 4.10 illustrates the plate geometry. With the geometry shown in Fig.
4.10, the free surface boundary condition relevant to the SH-wave can be
expressed as follows.

(4.161)
Here is used for the plane wave solution. Substituting
solution (4.159) into (4.161) we obtain the following equations.
(4.162)

(4.163)
For the above conditions hold for any x, the inside the curly brackets must
be zero. This leads to the following condition.

(4.164)
It follows that

, or

(4.165)
leading to the following condition.

(4.166)

Fig. 4.10 Plate elastic medium with two free surfaces

Condition (4.166) means as follows. When an SH-wave of a given

frequency travels through a plate in the x-direction undergoing numerous
reflections at the two free surfaces, the spatial frequency in the y-direction
(along the thickness) must satisfy condition (4.166). For a given thickness b, the
spatial frequency can be any of . Here, the integer n
is called the mode number and the corresponding wave is called the mode n
wave.

When n is an odd integer, and in

(4.165). From (4.162) it follows that . This makes the cosine term
null in the solution (4.159). Similarly, when n is an even integer,
the sine term of the solution becomes null. Thus, we find the
following expressions for under the condition of free surfaces at
.

(4.167)

(4.168)
Apparently, solution (4.167) presents anti-symmetric waveforms with
respect to y-axis, and solution (4.168) presents symmetric waveforms. Figure
4.11 illustrates several anti-symmetric and symmetric waves for low mode
numbers. The displacement vector is polarized along the z-axis and the waves
travel along the x-axis (normal to the page).
Fig. 4.11 SH symmetric waves with even mode numbers and anti-symmetric waves with odd mode
numbers in the plate

Now consider the velocity of the SH plate wave. The wave solution
(4.159) indicates that the SH plate-wave travels in the positive x-direction at

velocity . From the frequency equations (4.160) and (4.166), the

spatial frequency along the x-axis takes the following form.

(4.169)
Note that represents the phase velocity of the SH-wave incident to and
reflected from the free surfaces parallel to the x-axis, and is a constant. On the
other hand, the phase velocity of the SH-wave traveling along the x-axis

depends on as shown below.

(4.170)
In other words, this SH-wave exhibits dispersion. We can find the group

velocity as [12]. Below, we consider this dispersion in

some detail.
First, normalize the spatial and temporal frequencies as follows so that
they become dimensionless.

(4.171)

(4.172)
By substituting (4.171) and (4.172) into (4.169), we obtain the following
dispersive relation.

(4.173)
Here, the dimensionless frequency and the mode number n are real.

Therefore, if , the dimensionless spatial frequency becomes

imaginary, and if , is real.

Since b is real, (4.171) indicates that an imaginary and a real

correspond to an imaginary and a real , respectively. According to
(4.167) and (4.168), the anti-symmetric and symmetric SH plate wave travels

when is real. If is imaginary, the oscillation from the source is not

transferred as a wave, and instead, it decays exponentially as a function of x.
 The condition or , which determines whether is
real or imaginary, tells us that for a given mode number the oscillation from the
source travels as a wave if the oscillation frequency is higher than a certain
value. We can intuitively understand this situation by imagining oscillating one
end of an elastic string at various frequencies with the other end fixed to the
wall. As we increase the oscillation frequency, the motion travels toward the
fixed end with the shorter wavelength, i.e., the higher mode number. In other
words, if we want to excite a wave of a certain mode, we need to oscillate the
open end at a certain frequency. We can imagine that if we oscillate the open
end very slowly the entire string oscillates up and down.
Figure 4.12 shows the dispersion curve for various mode numbers n. Here,
the solid lines are for symmetric waves and the dashed lines are for anti-

symmetric waves, and the right half is for the case when is real and the left
half is when it is imaginary. Notice that the dispersion curve depends on the
plate thickness via (4.171) and (4.172). This fact can be used to find a thickness
of a layer from the dispersion curve [13].

Fig. 4.12 Dispersion curves for SH-waves traveling through a plate

 P and SV waves in a plate and Rayleigh-Lamb frequency equation
In this case, it is more convenient to use the potentials and . Recall
the discussions made under Sect. 4.2.2, Plane strain wave solution. The
governing equations and the solutions we should deal with are as follows.

(4.174)
and

(4.175)

(4.176)

where and

. Similarly to the case of the

free surface condition discussed in the Plane strain wave solution section, the
boundary conditions relevant to the P and SV waves are as follows.

(4.177)
Here, the boundaries are assumed to be traction free and the relevant stress
are expressed as follows.

(4.178)
(4.179)
At this point, express the displacement vector components and
using potentials (4.175) and (4.176)

(4.180)

(4.181)
Of the two terms enclosed by the parenthesis in expressions (4.180) and
(4.181), the first one is symmetric about the x-axis and the second one is anti-
symmetric. The symmetric and anti-symmetric expressions of and are
shown below.
Symmetric displacement components:

(4.182)

(4.183)
Anti-symmetric displacement components:
(4.184)

(4.185)
Figure 4.13 illustrates the symmetry.

Fig. 4.13 Symmetric and anti-symmetric parts of and components of plane P and SV waves
traveling in a plate of thickness 2b

Substitute (4.182) and (4.183) into expressions (4.79) and (4.80) and find
the stress due to the symmetric waves as follows.
(4.186)

(4.187)
Here the following substitution was made for .

(4.188)
Substitution of (4.186) and (4.187) into boundary conditions (4.177) yields
the following equations.

(4.189)

(4.190)
Here the ± sign comes from the boundary condition at
.

We can view (4.189) and (4.190) as a system of equations for and

, and express them in the form of a matrix as follows.
(4.191)

For the matrix equation (4.191) to hold for a nontrivial pair of and

, it follows that the determinant of the matrix must be null. This observation
yields the following frequency equation.

(4.192)
Equation (4.192) is known as the Rayleigh-Lamb frequency equation for
the propagation of symmetric waves in a plate. Remember that , and
are related through angle as follows.

(4.70)

(4.71)

(4.73)

(4.84)
For a given elastic material ( and ) with a plate thickness (2b), once
frequency is determined (4.192) tells us the angle and for the
symmetric waves (4.182) and (4.183).
From the system of equations (4.189) and (4.190), we can derive the
amplitude ratio as follows.
(4.193)
Repeating the same argument for the anti-symmetric case, we obtain the
following equations.

(4.194)

(4.195)
Equation (4.195) is the Rayleigh-Lamb frequency equation of anti-
symmetric waves.

(4.196)
Equation (4.196) is the amplitude ratio for the anti-symmetric wave case.

A note on vibrations and waves in a thin plate

When a longitudinal wave is to be excited in a thin plate with the application of
an acoustic transducer normally to the surface, we need to be careful about the
relation between the wavelength and the thickness. If the thickness is
significantly smaller than the wavelength, say less than a quarter of the
wavelength, the phase difference between the leading and trailing edge of the
longitudinal wave is less than . The plate is more or less oscillating as a
whole. Recall that the wave velocity is intrinsic to the material. Therefore the
wavelength is determined by the frequency. To make the wavelength shorter for
a given wave velocity (material), it is necessary to increase the frequency.
The same can be said about the spatial resolution. When we try to detect a
void in a material, if the length of the void in the direction of the longitudinal
wave’s propagation, the void falls in a small portion of the wavelength. It is
necessary to increase the frequency. However, increase of frequency is not
straightforward as the attenuation can be proportional to the square of the
frequency [14–19].

References

1. 1. Graff KF (1975) Wave motion in elastic solids. Oxford University Press, Oxford, UK
2. 2. Rose JL (1999) Ultrasonic waves in solid media. Cambridge University Press, Cambridge, UK
3. 3. Rose JL (1999) Ultrasonic waves in solid media. Cambridge University Press, Cambridge, UK, p
25
4. 4. Graff KF (1975) Wave motion in elastic solids. Oxford University Press, Oxford, UK, Chap. 5
5. 5. Rayleigh JWS (1887) On waves propagated along the plate surface of an elastic solid. Proc Lond
Math Soc 17:4–11
MathSciNet
6. 6. Rose JL (1999) Ultrasonic waves in solid media. Cambridge University Press, Cambridge, UK,
chap. 7
7. 7. Graff KF (1975) Wave motion in elastic solids. Oxford University Press, Oxford, UK, Chap. 6
8. 8. Pilarski A, Rose JL (1988) A transverse wave ultrasonic oblique incidence technique for
interfacial weakness detection in adhesive bonds. J Appl Phys 63(2):300–307
ADSCrossref
9. 9. Graff KF (1975) Wave motion in elastic solids. Oxford University Press, Oxford, UK, Chap. 8
10. 10. Maev RG (2008) Scanning acoustic microscopy. Physical principle and methods. Current
development in R. G. Maev Acoustic microscopy: fundamentals and applications, Ch1. Wiley-
VCH, Verlag. https://doi.org/10.1002/9783527623136.ch1 (accessed on July 28, 2023), pp 9–19
11. 11. Rose JL (1999) Ultrasonic waves in solid media. Cambridge University Press, Cambridge, UK,
chap. 8
12. 12. Yoshida S (2017) Waves; Fundamentals and dynamics, Chap. 4. Morgan & Claypool, San
Rafael, CA, USA
13. 13. Parmon W, Bertoni H (1979) Ray interpretation of the material signature in the acoustic
microscope. Electron Lett 15(21):684–686
ADSCrossref
14. 14. Szabo TL, Wu J (2000) A model for longitudinal and shear wave propagation in viscoelastic
media. J Acoust Soc Am 107(5):2437–2446
ADSCrossref
15. 15. Szabo TL (1994) Time domain wave equations for lossy media obeying a frequency power law.
J Acoust Soc Am 96(1):491–500
ADSMathSciNetCrossref
16. 16. Chen W, Holm S (2003) Modified Szabo’s wave equation models for lossy media obeying
frequency power law. J Acoust Soc Am 114(5):2570–2574
ADSCrossref
17. 17. Chen W, Holm S (2004) Fractional Laplacian time-space models for linear and nonlinear lossy
media exhibiting arbitrary frequency power-law dependency. J Acoust Soc Am 115(4):1424–1430
ADSCrossref
18. 18. Carcione JM, Cavallini F, Mainardi F, Hanyga A (2002) Time-domain modeling of constant-Q
seismic waves using fractional derivatives. Pure Appl Geophys 159:1719–1736
ADSCrossref
19. 19. D’astrous FT, Foster FS (1986) Frequency dependence of ultrasound attenuation and
backscatter in breast tissue. Ultrasound Med Biol 12(10):795–808
Crossref
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024
S. YoshidaFundamentals of Acoustic Waves and ApplicationsSynthesis Lectures on Wave Phenomena in
the Physical Sciences
https://doi.org/10.1007/978-3-031-48200-7_5

5. Electrical-Mechanical Transduction
Sanichiro Yoshida1
(1)
Department of Chemistry and Physics, Southeastern Louisiana
University, Hammond, LA, USA

Sanichiro Yoshida
Email: [email protected]

This chapter discusses acoustic transduction. We focus on the operation

principle of typical acoustic transducers to understand how the devices work. To
this end, we discuss the physics underlying the operation.
5.1 General Argument
Acoustic transduction generally means the conversion of an electric signal to an
acoustic signal or an acoustic signal to an electric signal. The former is acoustic
transmission, and the latter is acoustic sensing. A loudspeaker is an example of
an acoustic transmitter, and a microphone is an example of an acoustic sensor.
In the application to experimental mechanics dealing with solid objects, often
acoustic sensors are vibration sensors, and acoustic transmitters are vibration
actuators.
Thus acoustic transduction is the interaction of an electric system and a
mechanical system. The following pair of equations conveniently represent the
process of the acoustic transaction [1].

(5.1)

(5.2)
 Here v is the voltage on the electric system, u is the velocity of the
mechanical load, is the electrical impedance, is the mechanical to
electrical transduction coefficient, f is the force on the mechanical system,
is the electrical to mechanical transduction coefficient, and is the
mechanical impedance. (5.1) and (5.2) are the governing equations that
describe the relationship between the electrical and mechanical systems. Call
the former the electrical governing equation and the latter the mechanical
governing equation. When the transducer acts as an acoustic sensor, the
electrical governing equation expresses the acoustic (mechanical) input
represented by the velocity of the sensing mechanism (e.g., a microphone’s
diaphragm) represented by u and the resultant current and voltage i and v
detected by the electric system. When the transducer operates as a transmitter
(actuator), voltage v represents the input electric signal, and i and u represent
the resultant current flowing through the electric system and the velocity of the
actuation mechanism (e.g., a loudspeaker’s diaphragm). The mechanical
governing equation describes the force acting on the vibrating part (e.g., the
diaphragm) where the first term on the right-hand side represents the electric
force and the second term represents the mechanical force.
As these equations indicate, the conversion from acoustic to electric or
electric to acoustic energy is not 100% efficient. For instance, when an acoustic
signal oscillates the diaphragm in a microphone, part of the generated electric
energy is lost in the circuit. The first term on the right-hand side of (5.1)
represents this loss. Similarly, in the opposite process, part of the generated
mechanical energy is lost by mechanical impedance expressed by the second
term on the right-hand side of (5.2).
Thus, the coupling factor K is defined as follows.

(5.3)
Here the numerator is the product of the active terms on the right-hand side
of (5.1) and (5.2), whereas the denominator is the product of the ineffective
terms.
5.2 Reciprocal Transduction
This is the case when the following condition (reciprocal condition) holds.
(5.4)
Define the transduction factor as follows.

(5.5)
(5.1) and (5.2) become as follows.

(5.6)

(5.7)
In this case, the coupling factor takes the following form.

(5.8)

5.2.1 Electrostatic Transducer

An electrostatic transducer is an example of a reciprocal transducer. Figure 5.1
schematically illustrates a sample electrostatic transducer. This transducer
consists of two stationary electrodes connected to an AC voltage supply and one
movable plate (diaphragm) located in between the two stationary electrodes.
Typically, the diaphragm consists of a thin plastic film coated with a conductive
material. Initially, the diaphragm is positively charged with at a DC voltage
. Throughout the operation, this DC voltage remains the same. (In other
words, the diaphragm is AC-isolated.) The diaphragm forms two capacitors
with each of the stationary electrodes, respectively.
Fig. 5.1 Electrostatic transducer

When this transducer operates as an acoustic transmitter, the AC voltage

supply applies a sinusoidal voltage (an electric input) to the system. As
explained below, the applied AC signal is proportional to the output acoustic
signal. Figure 5.1 illustrates the moment when the applied AC voltage is
negative on the left stationary electrode and positive on the right stationary
electrode. The absolute value of these negative and positive voltage changes are
the same. This application of the AC voltage causes two things; first, positive
charges q flow into the right electrode, and the same amount of positive
charges, q, flow out of the left electrode; second, these changes in the two
electrodes generate a leftward electric field in the space between the electrodes.
This electric field exerts Coulomb force on the positively charged diaphragm,
causing it to move to the left by x. Here, the charge q caused by the applied AC
voltage is much smaller than the total charge on the diaphragm, , and
the corresponding displacement of the diaphragm is much smaller than the
initial gap distance between the stationary electrode and diaphragm,
.
Below, we analyze the operation of this transducer as an acoustic
transmitter. When the same transducer operates as an acoustic sensor, the
oscillatory displacement of the diaphragm due to the acoustic pressure (input)
causes an AC voltage change that is read out as an output electric signal.
In the following analysis, we use the interaction between the left stationary
electrode and the diaphragm for simplicity. The same argument holds for the
interaction between the right electrode and the diaphragm. If the total charge on
the diaphragm is positive, the change in the charge q makes the right electrode
repel the diaphragm and the left electrode attracts it. Consequently, the
diaphragm moves to the left.

5.2.2 Electrical Governing Equation of Electrostatic Transducer

First, derive the electrical governing equation. Our goal here is to find the
expression of voltage change due to the application of AC voltage. Consider a
sinusoidal voltage v is superposed to the bias voltage, and consequently, the
charge on the capacitor changes by and the gap length changes by x (
and ).

(5.9)
Here, S is the area of the capacitor, is the electric permittivity of the space
between the electrodes and the diaphragm, and are the initial gap length
and charge, and is neglected. We can relate and with the
initial capacitance as follows.

(5.10)
(5.11)
From (5.9), (5.10) and (5.11), we can express the change in the voltage across
the capacitor as follows.

(5.12)
On the right-hand side of (5.12), the first term represents the effect due to the
change in the stored charge (with the initial capacitance), and the second term
represents the effect due to the change in the capacitor’s gap length (with the
initial voltage).
Since the superposed voltage is sinusoidal, the resultant displacement is
sinusoidal with the same frequency. Let be the frequency. We can express
the current to the capacitor and the velocity of the diaphragm as follows.

(5.13)

(5.14)
Here j is the imaginary unit .
Substitution of (5.13) and (5.14) into (5.12) yields

(5.15)
Comparison of (5.6) and (5.15) with the use of (5.5) identifies the
expressions for the electrical impedance and the mechanical to electrical
transduction coefficient as follows.

(5.16)
(5.17)
The physical meaning of the electric governing equation (5.15) is as
follows. The acoustic signal (input) oscillates the diaphragm when the
electrostatic transducer acts as a sensor. The resultant velocity u generates
current (the flow of charges) via the second term on the right-hand side of
(5.15) with the transduction coefficient (5.17). The generated current causes the
voltage drop via the first term on the right-hand side of (5.15). The left-hand
side of (5.15) represents the overall voltage that the transducer outputs.
When the transducer acts as a transmitter (actuator), the applied voltage
represented by the left-hand side of (5.15) causes charges to flow into or from
the stationary capacitor (depending on the polarity). This flow of charges
oscillates the diaphragm via the corresponding electric force, as discussed
above.

5.2.3 Mechanical Governing Equation of Electrostatic Transducer

Now we consider the mechanical governing equation (5.7). The first term on the
right-hand side of (5.7) represents the electric force on the diaphragm due to
current i. The second term represents the mechanical force on the diaphragm,
called , which causes the diaphragm to oscillate harmonically. Below, we
consider the second term first.
The explicit form of is not readily available. So, we find it out from
the equation of motion for the diaphragm. Let m be the mass, x be the
displacement from the equilibrium position of the diaphragm, and k be the
damping coefficient and the stiffness of the elastic mechanism that pulls the
diaphragm to the equilibrium position. We can put the equation of motion as
follows.

(5.18)
Expressing the velocity of the diaphragm and rearranging the
order of the terms, we can rewrite the equation of motion (5.18) as follows.

(5.19)
 Assuming that the motion of the diaphragm is sinusoidal with frequency

, we can express , and put the

equation of motion (5.19) in the following form.

(5.20)
Next, consider the electric force on the diaphragm, the first term on the
right-hand side of (5.7). Call this force . The electric field inside the
capacitor, E, acting on the change in the stored charge, q, generates . Since
, we can use the initial charge to find E (with the application of
Gauus’ law [2]).

(5.21)
Here (5.10) and (5.11) are used in going through the second equal sign.
Note that the rightmost-hand side of (5.21) represents the uniform electric field
inside the capacitor with the initial gap length . When charge
flows into the capacitor, the capacitor’s electrodes feel electric force as
follows.

(5.22)
Here i is the current flowing into the capacitor and can be expressed as
.
Using (5.20) and (5.22) we can express the total force on the diaphragm, f,
as

(5.23)
 Comparison of (5.7) and (5.23) identifies the expressions for the
mechanical to electrical transduction coefficient and the mechanical impedance
as follows.

(5.24)

(5.25)
Equations (5.17) and (5.24) confirm that the electrostatic transducer is
reciprocal, .

5.2.4 Physics Behind Electrostatic Transducer’s Operation

Consider first the physics behind the mechanical governing equation. The
mechanical force expression (5.20) tells us the following features of the force.
First, the greater the damping coefficient and the mass m, the stronger force
is necessary to excite the oscillation at a given velocity u. This makes sense
because higher means more mechanical energy dissipation in the mechanical
system. Second, the complex unit j in front of the parenthesis on the rightmost-
hand side indicates the elastic force represented by is in phase
behind the damping force. Finally, setting the quantity inside the parenthesis to
zero represents the situation where the mechanical force is null
regardless of the velocity u. This corresponds to the resonant condition where
the frequency of the mechanical input matches the resonant frequency of the
elastic mechanism (the diaphragm and its support). Setting the inside the

parenthesis to zero provides us with angular frequency .

This is the natural angular frequency of a harmonic system, see (2.5) in Sect. 2.
1.2.
Fig. 5.2 a Electrostatic acoustic actuator and b sensor as two capacitors connected in series

Now consider the electro-mechanical interaction from the viewpoint of

electrodynamics. Figure 5.2 illustrates the above electrostatic transducer as two
capacitors. Here the left capacitor represents the capacitor formed by the left
stationary electrode and the diaphragm, and the right one represents the
capacitor formed by the diaphragm and the right stationary electrode. The
diaphragm’s voltage is fixed at as mentioned above (Fig. 5.1).
First, consider the behavior of the charges stored in both capacitors when
the transducer acts as an acoustic actuator. Assume that initially the diaphragm
is in the middle of the space between the two stationary electrodes. Thus, the
two capacitors’ capacitance is the same as . The charge
stored in the two capacitors is , hence

(5.26)
Consider that the current source causes charge to flow from the
left capacitor to the right capacitor. We can implement this operation by raising
the voltage at the top-left corner of the circuit and lowering the top-right corner
of the circuit and turning on the switch. Charge q will flow as an electric current
due to the potential difference. At this moment the voltage across the two
capacitors commonly remains at . However, the stored charges become
different between the two capacitors. The left electrode of the left capacitor
loses q, so the new charge will be . The right electrode of the right
capacitor’s new charge will be . Since the voltage across both
capacitors is , this situation leads to the following expressions for the new
capacitance and .

(5.27)

(5.28)
Equations (5.27) and (5.28) indicate that , meaning that after
the flow of charge q, the left capacitor’s gap length decreases and the right
capacitor’s gap length increases. In other words, if charges flow rightward
through the current source, the diaphragm moves to the left.
Now consider that the transducer acts as an acoustic sensor. Assume that
the mechanical force pushes the diaphragm to the left by . The new
capacitance of each capacitor becomes as follows.

(5.29)

(5.30)
At the moment when the diaphragm moves, the charges on the two
capacitors are still whereas and . This situation
leads to the following expressions for the new voltage across the electrodes.

(5.31)
(5.32)
Since , (5.31) and (5.32) lead to the following equations.

(5.33)
Equation (5.33) indicates that when the diaphragm moves to the left,
current will flow to the right as .
It is interesting to discuss the role of the electric force in the force
governing equation when the transducer acts as a sensor. (When the transducer
is an actuator, obviously the electric force drives the diaphragm’s oscillation, as
we discussed above.) Consider the situation where the mechanical force acts on
the diaphragm while the AC circuit is open ( ). The right illustration in
Fig. 5.2b with the terminals labeled and open corresponds to this
situation. Under this condition, the electric force term in the mechanical
governing equation (5.23) is null. Call the total force on the diaphragm under
this condition . By setting in (5.23) we find the following
expression for .

(5.34)
We can define the mechanical impedance under this condition as

(5.35)
Now consider that the AC circuit is short-circuited, and the electric force
term becomes active in the mechanical governing equation. At the same time,
the voltage in the electric governing equation (5.6) becomes zero ( ).
From the latter condition, we obtain the following equation.
(5.36)
Substituting (5.36) into the first term on the right-hand side of (5.23), and
using (5.5), (5.16) and (5.17) under the reciprocal condition (5.4) to find
, we obtain the following expression.

(5.37)
Then defining as and using (5.35), we obtain the following
expression.

(5.38)
We can view the effect of short-circuiting the AC circuit as the decrease in
the stiffness by (i.e., from k to ). What does this
observation mean? It means as follows. When the mechanical force shifts the
diaphragm toward the left stationary electrode, the capacitance of the space
between the diaphragm and the left electrode increases. At the same time, the
capacitance of the space between the diaphragm and the right stationary
electrode decreases because the gap distance increases. Consequently, the
electric energy stored in the left capacitor increases, and the energy stored in the
right capacitor decreases. Nature is not fond of such energy imbalance, and
therefore it forces the left capacitor to discharge by carrying an electric current
flowing through the wire connecting the two stationary electrodes. If we short-
circuit the circuit under this condition, the charges on the capacitor’s electrodes
flow with no resistance to the other electrode. If we open the circuit, there is no
way for the charges to flow. In this case, nature tries to oppose the movement of
the diaphragm. The mechanical force feels more resistance from the system as
compared with the short-circuited case. This explains why the stiffness is lower
when the AC circuit is short-circuited.
 We can argue this extra stiffness based on the electric force on the
diaphragm as well. Under the “open circuit” condition, when the diaphragm
moves to the left the left electrode has higher voltage than the right electrode (
.) This means that the electric field vector is directed from the left
stationary electrode towards the right stationary electrode through the
diaphragm. Since the diaphragm is positively charged, this electric field exerts a
rightward electric force on the diaphragm. This is a resistive force for the
mechanical force. The external agent exerting the mechanical force feels a
higher stiffness.
5.3 Anti-reciprocal Transducers
In this case, the mechanical-to-electrical and the electrical-to-mechanical
transduction coefficients have mutually opposite signs.

(5.39)
Define the transformation factor as follows.

(5.40)
The subscript M denotes “magnetic” because the magnetic field plays the main
role in this transduction, as will be discussed later in this section. The electrical
and mechanical governing equations take the following form.

(5.41)

(5.42)
Similarly to the reciprocal case, consider short-circuiting the electric circuit to
make . From (5.41), the current under this condition becomes
. Substituting this into (5.42) yields and
as follows.
(5.43)

(5.44)

5.3.1 Moving-Coil Transducer

The most typical example of an anti-reciprocal transducer is the moving-coil
transducer. In this case, in (5.43) and (5.44) is the electrical impedance
consisting of the DC resistance and inductance .

(5.45)
Unlike the electrostatic case, the open-circuited condition makes the electric
force inactive. This is because while in the electrostatic case, the electric force
is the Coulomb force proportional to the voltage across the capacitor, in the
moving-coil case the electric force is the magnetic force generated in the coil in
proportion to the current.
According to electrodynamics [2], the magnetic force due to the magnetic
field B acting on current i flowing through a conductive wire of length l is
. On the other hand, when the electric current moves in the
magnetic field with velocity u, the magnetic induction induces an opposing
voltage of Blu.
Consider the above-mentioned mechanism in more detail by referring to
Fig. 5.3. In this figure, a permanent magnet applies a constant magnetic field
into the page. Two conductive wires are connected to either an AC power
supply (Fig. 5.3a) or a resistor (Fig. 5.3b). A conductor bar is placed on the two
wires. Here (Fig. 5.3a) models an acoustic actuator and (b) models an acoustic
sensor. The conductor bar represents a diaphragm in both cases. In the acoustic
actuator configuration, the transducer induces a force on the diaphragm
(represented with f) in response to an input electric current, whereas in the
acoustic sensor configuration, the transducer induces an electric current in
response to the motion of the diaphragm (represented with velocity u) due to
mechanical force. In Fig. 5.3, the dashed quantity depicts the induced quantity.
When the transducer operates as an actuator, the AC power supply
generates current. Figure 5.3a illustrates the moment when the current from the
power supply flows upward in the conductor bar (length l). Here, etc. are the
unit vector of the direction. The magnetic field due to the permanent magnet (

) exerts magnetic force leftward. The diaphragm moves to the left.

In the next cycle when the current flips the direction, the magnetic force exerts
rightward force on the diaphragm. In this fashion, the diaphragm oscillates. The
magnetic force at the moment when the electric current is , the

magnetic force is

(5.46)
When the transducer operates as a sensor, the mechanical force causes
velocity . Magnetic induction states that the induced voltage v is the change
in the magnetic flux over time. Here A is the area of the
magnetic field. Since the area A changes over time as the conductor bar moves
at velocity u while the magnetic field is constant, we can express the temporal
change in the magnetic flux as follows.

(5.47)
According to Faraday’s law, the induced current resulting from magnetic
induction flows in such a way that the resultant magnetic field opposes the
temporal change in the magnetic flux. When the magnetic force causes the
rightward velocity as shown in Fig. 5.3b, the magnetic force increases the
magnetic flux. Hence, the induced current flows in the direction that generates a
magnetic field in the out-of-page direction inside the loop formed by the
conductor bar and the two wires.
Fig. 5.3 Conceptual illustration of a magnetic force generated by an acoustic actuator and b magnetic
induction occurring in an acoustic sensor

Figure 5.4 illustrates a more realistic structure of a moving-coil transducer.

Typically, a moving-coil transducer consists of a solenoid wound around a
cylindrical core placed inside the inner space of a doughnut-shaped permanent
magnet. The diaphragm (not shown in Fig. 5.4 to avoid confusion) is attached to
the solenoid in such a way that the solenoid’s motion parallel to the axis of the
core cylinder (called the vertical motion) oscillates the diaphragm.
Fig. 5.4 Schematic view of a moving-coil transducer

In this configuration, as the solenoid experiences vertical motion, the

magnetic flux varies. In the situation illustrated in Fig. 5.3, the area that
penetrates the permanent magnetic field changes as the conductor bar moves
parallel to the plane. In the case of Fig. 5.4, the (active) area changes as the
conductor (the solenoid) moves perpendicular to the plane that the permanent
magnet penetrates. In this situation, the circular area formed by each turn of the
solenoid defines the active area involved in the magnetic flux. When the
solenoid is moving out of the vertical field line of the permanent magnet, the
induced voltage decreases as the circular area of each turn goes out of the
effective zone of the permanent magnet. This is the phase when the induced
voltage decreases. In the opposite phase, as each turn of the solenoid enters the
effective zone of the permanent magnetic field, the induced voltage increases
because the number of circular areas that contributes to the magnetic induction
increases.
Figure 5.5 illustrates another way of understanding the magnetic induction
in the configuration illustrated in Fig. 5.4. The top illustration of this figure
represents the situation where a solenoid carries an electric current provided by
an AC voltage supply. A permanent magnet is partially located inside the
cylindrical space of the solenoid with two poles as shown in the figure. In the
phase when the current increases in the direction depicted by solid arrows (the
current increases clockwise when the solenoid’s cross-sectional area is viewed
from the left side of the solenoid), the solenoid increases the magnetic field
vector rightward along the axis of the cylindrical space. Electromagnetic
induction reduces this change. Since the magnetic field due to the permanent
magnet is leftward, going deeper into the solenoid performs this task. Thus, the
permanent magnet is pushed to the right. The solid big arrow in Fig. 5.5
indicates the direction in which the permanent magnet is pushed.

Fig. 5.5 Explanation of temporal change in magnetic flux in moving-coil transducer

We can understand the above effect in terms of the interaction between two
magnets. When the current flows in the direction indicated by the solid line, the
solenoid forms an electric magnet with the S-pole on the left and the N-pole on
the right. (Consider this imaginary permanent magnet in the space of the
solenoid’s core on the right of the permanent magnet.) On the other hand, the
polarity of the permanent magnet is opposite, as the lower illustration of Fig.
5.5 indicates. The two S-poles facing each other repel the magnets from one
another. Hence, the permanent magnet is pushed to the left. When the solenoid
current increases, the electric magnet’s repelling force increases. Nature does
not like this abrupt change, so it exerts a counterforce. This force corresponds to
the above-mentioned opposing effect due to electromagnetic induction.
In the other phase, when the current flows in the opposite direction
(depicted by the thin dashed line in the top illustration), the direction of the
magnetic field generated by the solenoid flips, whereas the direction of the
permanent magnetic field remains the same. Consequently, the permanent
magnet is pulled into the solenoid, as the thick arrow with the dashed line
indicates.
The above description illustrates the dynamics when the solenoid carries
current. Hence it applies to the explanation of an acoustic actuator. Instead of
supplying current to the solenoid, consider that a mechanical force pushes the
permanent magnet into the solenoid. According to the law of magnetic
induction, it will induce an electric current. This situation corresponds to the
case when the transducer acts as an acoustic sensor.
In the above explanation, we assumed that the solenoid is stationary in
space, and the permanent magnet is movable. Since all the interactions between
the permanent magnet and the solenoid are relative, we can argue the same
effect by making the permanent magnetic field stationary and the solenoid
movable. The dynamics illustrated in Fig. 5.4 corresponds to that case.
The magnetic force and induced voltage yield the following two governing
equations. The above argument indicates that for a common direction of the
electric current, the direction of the induced force f and the velocity u caused by
the magnetic force are opposite to each other. This explains the negative sign
appearing in the first term on the right-hand side of the mechanical governing
equation (5.49).

(5.48)

(5.49)
Comparison of the set of equation (5.48) and (5.49) with (5.41) and (5.42)
indicates the following expression for

(5.50)
Repeating the same type of argument as the electrostatic transducer case,
we find the following force expression when the electric circuit is open. Note
that, unlike the electrostatic case, the electromagnetic force is inactive when the
electric circuit is open. This is because, in this case, the electromagnetic force is
magnetic force, which is proportional to the current. Thus, calling this force
we obtain the following expressions for the force and the corresponding
impedance .
(5.51)

(5.52)
When the electric circuit is closed and therefore , from (5.48) and
(5.50) we find . Substitution of this
current into (5.49) with the use of (5.25) yields the expressions for the
corresponding force and impedance as follows.

(5.53)

(5.54)
Here assuming , we use ((5.45). As
expected in the case of the moving coil transducer the short-circuited impedance
has greater stiffness than the open-circuited case (i.e., k vs. ).

5.3.1.1 Piezoelectric Transducer

As indicated by the name, the piezoelectric transducer uses the phenomenon
referred to as the piezoelectric effect [3] and the inverse piezoelectric effect [4,
5]. The piezoelectric effect refers to the property of certain materials that
exhibits voltage in proportion to the applied stress. The materials that exhibit
the piezoelectric effect are called piezoelectric materials. The inverse
piezoelectric effect refers to the reversed effect. In response to applied electric
voltage, piezoelectric materials are compressed or stretched in proportion to the
applied voltage. The applied stress or electric voltage can be in high frequency,
such as the ultrasonic frequency. Thus, a variety of ultrasonic transducers are
commercially available. Here, ultrasonic stress sensors use the piezoelectric
effect, and ultrasonic transmitters (actuators) use the inverse effect.
The piezoelectric effect is essentially the induction of electric dipole
moment in the material. The electric dipole is defined as a pair of unlike electric
charges, q and , separated by a small distance d. [2] The electric dipole
moment is defined as the product of them, . It is a vector
quantity whose magnitude is qd, and the direction is the displacement vector
drawn from the negative charge to the positive charge of the pair. The
polarization is defined as the density of the electric dipole moment.
We can explain the essence of the piezoelectric and inverse effects based
on the above explanation of dipole moment and polarization vectors. Take
quartz (SiO ) as an example since it is a well-known solid that exhibits
piezoelectric and inverse effects. Figure 5.6 illustrates a simplified, two-
dimensional configuration of quartz crystal. Figure 5.6a represents the stress-
free situation where in each pair of neighboring atoms the electrons are closer to
the oxygen atoms but the entire crystal is electrically neutral. The electrons are
closer to the oxygen atom because it has higher electron affinity than the silicon
atom. The entire crystal is neutral because the electron distribution makes it so.
Now imagine that you stretch the crystal vertically as indicated by Fig.
5.6b. In the enclosed region near the top of the crystal, the silicon atom has
moved away from the two oxygen atoms. Consequently, the polarity of this
region has become more positive as compared with the stress-free state. In other
words, the contribution of the two relatively negative oxygen atoms is lower in
the stretched state. On the other side of the crystal, the exact opposite
phenomenon occurs. The contribution of the relatively positive silicon atoms is
reduced and consequently, this end becomes negative. Overall, the entire crystal
is polarized with the top side positive and the bottom side negative.
Fig. 5.6 Unit cell of quartz a stress-free, b under vertical tensile stress, c under vertical compressive stress

If the crystal is compressed vertically, the situation is reversed (Fig. 5.6c.)

On the top side of the crystal, the contribution of the two oxygen atoms
increases and therefore this side becomes negative. On the bottom side, the
contribution of the silicon atom increases, and this side becomes positive. If the
stretching and compressing stresses are applied alternatively, the polarity
changes accordingly. If we place a quartz crystal on the surface of a solid
specimen undergoing compression and stretching, the quartz crystal exhibits
alternative positive and negative electric polarities. This quartz crystal behaves
as a stress sensor. It outputs the electric signal sensing the stretching and
compressive stress caused by the specimen.
It is now easy to understand the same quartz crystal as an ultrasonic stress
transmitter (actuator). When an alternating electric voltage is applied to the
crystal, the dipole moment is induced. Consequently, the crystal is stretched and
compressed, following the change in the electric polarity. This demonstrates the
quartz crystal as an ultrasonic transmitter.
The efficiency of the piezoelectric effect is indicated by the piezoelectric
coefficient (usually denoted by ) in the unit of m/V (meter/volt). The
piezoelectric coefficient of quartz is approximately 3 m/V [6].
Many ceramics that have a perovskite structure (ABX ) [8] exhibit the
piezoelectric effect. Here A atoms are larger metal cations, B atoms are smaller
metal cations, and X atoms are anions (usually oxide). PZT (lead zirconate
titanate Pb [ ] is one of the most used
piezoelectric materials. It is a binary solid-solution of PbZrO and PbTiO
having a perovskite structure with Pb for A, O for B, and Ti or Zr
for B.
Figure 5.7a illustrates the structure of a unit cell of PZT when the cell is
above the Curie point. With the Ti or Zr ion at the center of the cubic
cell, the polarization is null. Below the Curie temperature, the Ti or Zr
ion shifts from the center generating a dipole moment as illustrated in Fig. 5.7b.
When external force stretches or compresses the unit cell, the position of the Ti
or Zr changes, exhibiting the piezoelectric effect. When the unit cell is
under an external electric field, the electric force changes the location of the Ti
or Zr ion, exhibiting the inverse piezoelectric effect.
The piezoelectric coefficient of PZT is two orders of magnitude higher
than quartz [6]. Normally, PZT cell units are randomly oriented. To make an
efficient acoustic transducer, it is necessary to pole the crystal by applying a
high electric field [9].

Fig. 5.7 Unit cell of lead zirconate titanate PZT; a above Currier temperature and b below Currier
temperature

References

1. 1. Kinsler LE, Frey AR, Coppens AB, Sanders JV (1980) Fundamentals of acoustics, 3rd edn,
Chap. 14. Wiley, New York, USA
2. 2. See, for example, Griffiths DJ (1999) Introduction o electrodynamics, 3rd edn. Prentice Hall,
Upper Saddle River, NJ, p 74
3. 3. Curie J, Curie P (1880) Bulletin de la Société Minérologique de France 3(4):90–93
Crossref
4. 4. Lippmann G (1881) Principe de la conservation de l’électricité (Principle of the conservation of
electricity). Annales de chimie et de physique (in French) 24:145
5. 5. Curie J, Curie P (1881) Contractions et dilatations produites par des tensions dans les cristaux
hémièdres á faces inclinées (Contractions and expansions produced by voltages in hemihedral
crystals with inclined faces). Comptes Rendus (in French) 93:1137–1140
6. 6. Bottom VE (1970) Measurement of the piezoelectric coefficient of quartz using the Fabry-Perot
dilatometer. J Appl Phys 41:3941
ADSCrossref
7. 7. Kholkim AL, Kiselev DA, Kholkine LA, Safari A (2008) Piezoelectric and electrostrictive
ceramics transducer and actuators. In: Schwartz M (ed) Smart materials, 1st edn, Chap. 9. CRC
Press, London, New York
8. 8. Tilley RJD (2016) Pervoskites, 1st edn. Wiley, Chichester, UK
9. 9. Waller D, Iqbal T, Safari A (1989) Poling of lead zirconate titanate ceramics and flexible
piezoelectric composites by the corona discharge technique. J Am Ceramic Soc 72(2):322–324
Crossref
Appendix A
Direction Cosines
Assume be a unit vector in the direction that makes angles of , , and
to the three coordinate axes.

(A.1)
Consider the following scalar products based on the definition

. Here and are arbitrarily chosen

vectors, A and B are their magnitude, and is the angle between the two
vectors.

(A.2)

(A.3)

(A.4)

Here we used the fact that , , , and are all unit vectors, hence their
magnitudes are all one.
Now substitute the component expression (A.1) into the left-hand sides of
(A.2)–(A.4).

(A.5)
(A.6)

(A.7)
Comparing (A.2)–(A.4) with (A.5)–(A.7), we find that the components of
are the direction cosines.

(A.8)

Appendix B
Gradient Operation
Consider the differentiation of f(t, x, y, z) referring to Fig. B.1. The total
differential df is given by the following expression.

(B.1)
As indicated in Fig. B.1, the addition expressed by the right-hand side of (B.1)
consists of three portions each corresponding to one of the following three
steps; finding the change in the value of f associated with the parallel movement
along the x-axis by dx, the change associated with the parallel movement along
the y-axis by dy, and that associated with dz. While we can find the scalar
quantity df with (B.1), each of these three steps has a directionality. The short
answer to the question “What is the three-dimensional version of df/dx?” is the
gradient operation (vector) given by (1.22). Since this concept is not
straightforward, at least for me, I would like to take a moment to consider it.

(B.2)

Here the operator is called the gradient operator, and , , and are the
unit vector for x, y, and z directions. This operator applies to a scalar function
and produces a gradient vector. We use the case of a Cartesian coordinate here
but we can express the gradient operator for other coordinate systems.

Fig. B.1 Total differential df consisting of partial differential with respect to x, y and z

Let’s start the discussion with a simple case where the wave function f(t, x)
depends on a single spatial coordinate variable x. For this discussion, we
assume that the function f represents the gravitational potential to make the
argument intuitive. The slope indicates the path of an object to fall by gravity if
it is released. Figure B.2a illustrates an example of such a case where the
potential curve is expressed as follows.

(B.3)
The potential curve represents an arc of a circle with a radius of 10.
Consider the slope at two representative points, . As the arrows
in this figure indicate, the slope at these two points has the same magnitude and
mutually opposite signs. Thus, even in this case, the slope shows a vector-like
behavior. Since we are dealing with only one-spatial coordinate, the slope
appears to be a signed scalar.
Fig. B.2 a One-dimensional gravitational potential; b Two-dimensional potential depending on x and
independent of y; c Two-dimensional potential having quadratic dependence on x and y

Figure B.2b is the case where the wave function has two independent
spatial variables as f(t, x, y). Its x-dependence of the wave function is the same
as Fig. B.2a and y-dependence is null, i.e., . Imagine you put a
tennis ball at a representative point and release it
from your hand. It will fall in a direction parallel to the x-axis. If you repeat the
same thing on the other side at , the tennis ball will fall
in the opposite direction but still parallel to the x-axis. This observation is the
same as the one we made for Fig. B.2a, i.e., the slope has an opposite sign
(direction). In this case, since the function f has two independent spatial
variables (x and y), the slope is literally a vector quantity.1
Consider the situation in Fig. B.2b in terms of (B.2). Substituting (B.3)
into (B.2), we obtain the following gradient vector.

(B.4)
As expected from the above intuitive observation of the tennis ball, the
gradient vector has only x-component, corresponding to the ball falling parallel
to the x-axis. The two arrows in Fig. B.2b indicate this coordinate dependence
of the gradient. The arrow pointing downward represents the x-component, and
the one on the ridge of the f(x, y) indicates the zero y-component. Equation
(B.4) also reveals the fact that the slope (gradient) is proportional to the
coordinate value x with a negative sign; if the reference coordinate is positive (
), it is negative. This situation corresponds to the fact that the
gradient decreases with , i.e., its positive value decreases as increases
positively; consequently, the ball falls faster as increases. Similarly, we can
explain the observation that the falling speed of the ball decreases as the
negative reference coordinate increases toward , i.e., when
, the gradient is positive.
Now in Fig. B.2c, the gravitational potential has the same dependence on x
and y. Consider the motion of a tennis ball when you release if at
. You can easily imagine that the ball falls along a
longitudinal line because in the tangential direction, the height does not change.
As is the case of Fig. B.2b, let’s consider the gradient vector quantitatively. In
this case, the function f(x, y) has a dependence on x and y and can be expressed
as follows.

(B.5)
Substituting (B.5) into (B.2), we obtain the following expression for the
gradient vector.

(B.6)
This time, the gradient vector has both x and y components. This fact is
consistent with the above qualitative observation that the ball should fall along a
longitudinal line. In a top view, a longitudinal line contains both x and y
components. However, (B.6) is not convenient to find the longitudinal direction
that the ball falls from a reference point. It is better to use spherical coordinates.

Using , and to represent the unit vector for the radial, polar, and
azimuthal component of we can express (B.2) and (B.5) as
follows.

(B.7)

(B.8)
 Here we use that the surface shown in Fig. B.2c is part of the surface of a
sphere with a radius of 10.
Substituting (B.7) into (B.8), we find the gradient vector as follows.

(B.9)
The arrows in Fig. B.2c represent a tangential vector (the one parallel to
the xy-plane), and expressed by (B.9) at a reference point. The tangential
vector corresponds to the one on the ridge seen in Fig. B.2b.
Referring to Fig. B.3 we find the following relationship between the
spheric and Cartesian coordinate variables.

Fig. B.3 Coordinate point expressed with a Cartesian and b Spheric coordinate systems

(B.10)

(B.11)
Using and noting that from (B.10)

, we
can write (B.6) as follows.
(B.12)

(B.13)
Now use (B.11) in (B.9) to derive the following equation.

(B.14)
Comparison of (B.12) and (B.14) indicates that these two equations are
identical.

Appendix C
Orthogonality of Sine and Cosine Functions
According to Fourier’s theorem [1, 2] a periodic function can be expanded into
a series of cosine and sine functions.

(C.1)
Here is the lowest angular frequency. If function f(t) is defined on [0 T], the
lowest frequency and the corresponding angular frequency are as follows.

(C.2)

(C.3)
With this notation, we can express (C.1) as follows.
(C.4)
The next step is to determine the coefficients and . We can use the
following property to determine the coefficients. This property is known as the
orthogonality of the trigonometric functions.

(C.5)

(C.6)
Here the following identities are used.

(C.7)

(C.8)

(C.9)

(C.10)
Multiply to both sides of Eq. (C.1) and integrate for one period
(from to ). From the orthogonality, all terms vanish
except on the right-hand side. Thus, using Eq. (C.6), we find
that coefficient can be expressed as follows.

(C.11)
It is clear that other coefficients can be evaluated with the same procedure.
Generally, coefficients and can be expressed as follows.

(C.12)

(C.13)

Appendix D
Notes on Standing Waves
When the amplitude of the forward-going wave ( ) is different from that of
the backward-going wave ( ), the superposed wave can be expressed as
follows.

(D.1)
Equation (D.1) indicates that in this case the expression of the superposed wave
consists of the standing-wave term (the first term on the right-hand side of Eq.
(D.1)), and the traveling-wave term (the second term on the right-hand side of
Eq. (D.1)). Because of this traveling-wave term, the superposed wave is
behavior differs from the case when the forward-going and backward-going
waves have the same amplitude. First, the superposed wave is not zeroed out
even when the phase difference between the forward-going and backward-going
waves is an odd integer multiple of . Figure D.1 illustrates the situation.
Here, Fig. D.1a shows the forward-going (top), the back-ward-going (middle)
and superposed (bottom) waves for three representative moments when the
interference is destructive (left), constructive (right) and intermediate (center).
Notice that under the destructive interference the superposed wave is
completely zeroed out whereas under the constructive interference, the peak of
the superposed wave is doubled as compared with the forward-going wave. The
superposed wave oscillates as a standing wave. Figure D.1b exhibits the
situation when the amplitude of the backward-going wave is 25% smaller than
the forward-going beam ( ). It is seen that even when the two
wave are completely out of phase (the phase difference is ), the superposed
wave is not zeroed-out.

Fig. D.1 Superposition of two waves propagating in mutually opposite directions. a When forward-going
and backward-going waves have the same amplitude. b When the forward-going wave has greater
amplitude
 Second, the superposed wave travels. According to the traveling-wave
term of Eq. (D.1), the superposed wave travels forward if the amplitude of the
forward-going wave is greater than the backward-going, and it travels backward
if the amplitude of the forward-going wave is smaller than the backward-going
wave. The superposed wave in Fig. D.1b travels but is not obvious. This is
because the amplitude of the traveling-wave term in Eq. (D.1) increases in
proportion to the difference in amplitude between the forward-going and
backward-going waves. Figure D.2 shows that the superposed wave travels
forward clearly when the amplitude of the backward-going wave is 25% of the
forward-going wave ( ). With the increase in the difference
between and the amplitude of the traveling-wave term increases.
Thus, the traveling feature of the superposed wave is much clearer as compared
with case shown in Fig. D.1b.
Fig. D.2 Superposition of two waves where backward-going wave’s amplitude is 25% of the forward-
going wave

Appendix E
Generalized Hooke’s Law and Equation of Motion for Isotropic
Media
E.1 Strain and Stress Matrices
The strain matrix consists of normal strain components and shear strain
components.
(E.1)
Here the normal strain and shear strain components are defined as follows.

(E.2)

(E.3)
where and and
.
Similarly, we define the stress matrix as follows.

(E.4)
Here represents the stress on the plane i in the direction of j as
indicated in Fig. E.1. The plane i is the plane normal to the i axis. Stress is
defined as a force acting on a plane divided by the area of the plane. For
instance, is the force applied in the positive x direction on the plane
located at x perpendicular to the x axis. is the force acting on the plane
perpendicular to the y axis at y in the direction of positive x.
Fig. E.1 Normal and shear stress components. The first subscript denotes the plane and the second the
direction of the stress (force) vector

E.2 Compliance and Stiffness Matrices

The compliance or stiffness matrix constitutes the basis of generalized Hooke’s
law. These two matrices are mutually the inverse of the other and essentially
represent the same stress-strain relationship.
The compliance matrix S and stiffness matrix [C] are defined as follows.

(E.5)

(E.6)
Here and are the strain and stress matrices in Voigt’s notation [3].
In this notation the 3 3 matrices (E.2) and (E.4) become a vector with 6
components, respectively. Voigt’s notation is applicable to a symmetric matrix
in general.

(E.7)

(E.8)
Here represents transpose.
For later uses, we express the compliance and stiffness matrices in the
following component forms.
(E.9)

(E.10)

E.3 Principal Axes

Before proceeding to further analysis, take a moment to consider the coordinate
system to express stress and strain. Refer to Fig. E.2 and consider that an
external force is acting on an elastic cubic object. In three dimensions, we can
set up an xyz coordinate system in any direction. In this figure, we consider two
coordinate systems, xyz and . Notice that the xyz has the three
coordinate axes normal to the corresponding plane of the cube. Consequently,
there is no shear stress component. The coordinate axes that satisfy this
condition are called principal axes [4]. The axes of the other coordinate system
are not principal axes because when decomposed with this coordinate
system, the force vector has shear components.

E.4 Isotropic Case

Now we are in a position to express the strain tensor components with the xyz
and coordinate systems. We restrict ourselves in an isotropic case.
Here the isotropy means that the elasticity is independent of the orientation.
Under this condition, the elasticity is plane-symmetric.

Fig. E.2 When expressed with the coordinate system of principal axes xyz the shear stress is zero

E.4.1 Axis Inversion

Using plane symmetry, we can reduce the number of independent compliance
matrix elements. First, consider axis inversions. The elastic response is
symmetric about the xy plane. This means that if the z axis is inverted the
displacement components behave as follows.

(E.11)
On the other hand, the spatial derivatives behave as follows.

(E.12)
Consequently, the strain components behave as follows through this inversion
operation.

(E.13)
(E.14)

(E.15)

(E.16)

(E.17)

(E.18)
Repeating the same procedure for the stress matrix components, we obtain the
following equations.

(E.19)
Now express the first row of (E.5) using the components of the compliance
matrix in both xyz and systems, respectively.

(E.20)

(E.21)
(E.22)
By changing the sign of the fourth and fifth rows of (E.5) we obtain the
following equation.

(E.23)
Comparing (E.22) and (E.23), we obtain the following equation.

(E.24)
Note that on the left-hand side, the fourth and fifth columns (corresponding to
the non-inverted axes) the sign is flipped, and on the right-hand side, the fourth
and fifth rows (corresponding to the non-inverted axes) the sign is flipped.
From (E.24) we find as follows.
(E.25)
Note that the first and second lines of (E.25) list the matrix elements at mutually
symmetric positions. Thus, by considering inverting the z axis, we find the
compliance matrix has the following form.

(E.26)
Similarly, by considering inverting the x axis, we obtain the following equation.
Those matrix elements found to be 0 from (E.25) are set to 0 in the below
equation.

(E.27)
From (E.27), we find as follows.

(E.28)
Substituting (E.27) into (E.26), we can further simplify the compliance matrix
as follows.
(E.29)

E.4.2 Rotation Around Axis

Next, consider rotating the xy plane around the z axis by angle . The plane
symmetry condition indicates that the elastic property is intact by this operation.
Call the coordinate system before this rotational operation xyz and the one after
. By expressing the same point on the plane by expressing it with the
two coordinate systems, we obtain the following equation.

(E.30)
Similarly, we can express the displacement vector components , , and
written in the old coordinate system with the new coordinate system as
follows.

(E.31)
Next, express the spatial derivatives of the new coordinate system with the old
one using (E.30) and (E.31).
(E.32)

(E.33)

(E.34)
Using the above relationship between the quantities expressed with xyz and
systems, we can express in terms of the strain tensor components
expressed with the xyz system.

(E.35)
Similarly, we obtain the following equations.

(E.36)

(E.37)

(E.38)

(E.39)
(E.40)
With the same procedure, we can obtain the following equation for .

(E.41)
We can express etc. on the right-hand side of (E.35) with stress matrix
components by using the compliance matrix expression (E.29).

(E.42)
Now this time first express with using (E.22) and the compliance
matrix (E.29). Note that because of the isotropic condition, the elastic behavior
does not change by the rotational operation, hence we can use the same
compliance matrix.

(E.43)
Comparing (E.42) and (E.43) for the same terms, we find as follows.

(E.44)

(E.45)
(E.46)
Repeating the same procedure for rotation around other axis, we find the
following conditions.

(E.47)
Using (E.44)–(E.46) in (E.29), we obtain the following expression of
compliance matrix.

(E.48)
Expression (E.48) is the general form of compliance matrix for an isotropic
elastic medium.

E.5 Expression of Compliance Matrix with Elastic Constants

We next express the compliance matrix with elastic modulus. Consider in Fig.
E.3 that external force acts on a block of an elastic object along the x axis. The
elastic block experiences stretch in the x direction and compression in the y and
z directions according to Poisson’s effect. Clearly, there is no shear strain if we
use thexyz coordinate system.

Fig. E.3 Elastic block experiencing normal strains

As shown in Fig. E.3, the external force along the x axis causes three
normal strains; tensile strain along the x axis, and compressive strains along the
y and z axes. The compressive strains are due to Poisson’s effect. The
compressive strains can be related to the tensile strain via Poisson’s ratio .
The overall strain (the addition of all three strains) constitutes . Thus, we
obtain the following equation.

(E.49)
Here the first term on the right-hand side represents that the tensile strain
in the x direction is related to via Young’s modulus E. The other terms
represent that the normal strain is related to the normal stress and that the
compressive strain is lower than the tensile strain by a factor of Poisson’s ratio
where the negative sign indicates the strain is opposite to the one along
the x axis, which is the direct consequence of the applied force.
Comparison of the compliance matrix (E.48) and (E.49) allows us to
identify and as follows.

(E.50)
Using (E.50), we can express the compliance matrix for an isotropic elastic
medium with Youg’s modulus and Poisson’s ratio, and thereby write the strain-
stress relationship (E.5) as follows.

(E.51)
By finding the inverse matrix to the compliance matrix in (E.51), we can
express the stress-strain relationship (E.6) in the following form.
(E.52)
Conventionally, the stiffness matrix is expressed with Lamé’s first and
second parameters and .

(E.53)
Table E.1 summarizes the relationship among Lamé’s parameters, Young’s
modulus, shear modulus, and Poisson’s ratio. Notice that we can always express
an elastic constant with a pair of other elastic constants.
Table E.1 Various elastic moduli and their relationships

( ) (E, G) ( ) ( )
E E E

G G

E.6 Equation of Motion

Refer to Fig. E.4 and consider the force acting on an elastic block of density
having three sides of dx, dy, and dz. The net force in the x, y, and z directions is
the addition of the normal and shear force in line with the respective direction.
Thus, we obtain the following equation of motion for this elastic block.

Fig. E.4 All external force acting on an elastic block. The three cubes in this figure represent the same
elastic block. Each of the three cubes illustrates the net force exerted in the corresponding direction

(E.54)

Here, , , and are the unit vector for the corresponding direction.
Remember that the first subscript i of the stress tensor component
represents the plane on which the force is acting and the second subscript j
represents the direction of the force. The volume dxdydz is canceled on both-
hand sides on deriving (E.54).
Using (E.53) we can substitute the stress tensor components with the
pertinent strain tensor components and obtain the following equation.
(E.55)
The next step is to express the right-hand side of (E.55) in terms of
displacement . Remember the following expressions of the strain tensor
components.

(E.56)

(E.57)
Using (E.56) and (E.57), consider expressing the x component of the right-
hand side of (E.55) expressing with the displacement vector components.

(E.58)
We can view the first term on the last line of (E.58) as the x component of
, and the second term as the x component of . By repeating
the same term rearrangement for the y and x components on the right-hand side
of (E.55), we obtain the following expression.
(E.59)
Equation (E.59) is the equation of motion we wanted to derive.

References

1. 1. Prof. Brad Osgood (2014) Lecture notes for EE 261 the Fourier transform and its applications.
CreateSpace Independent Publishing Platform
2. 2. Bracewell RN (1999) The Fourier transform and its applications, 3rd edn. McGraw-Hill, Boston,
New York
3. 3. Brannon RM (2018) Rotation, reflection, and frame changes, Chap. 26, Voigt and Mandel
components. IOP, Bristol, UK
4. 4. Chou PC, Pagano NJ (1967) Elasticity: tensor, dyadic, and engineering approaches. Dover
Publications, New York

Index
A
Acoustic energy conservation at boundary 62
Acoustic impedance 55, 60, 63
Acoustic radiation pressure 69
Acoustic wave intensity 56
Acoustoelasticity 29
Amplitude demodulation 81
Amplitude modulation 79
Angular spatial frequency 7
Angular temporal frequency 7
Audible frequency 73
Average acoustic energy 53
Average intensity of acoustic wave 57
B
Boundary condition 17–19
Bulk modulus 35
C
Complex propagation constant 45
Constructive interference 14
Critical angle 59, 108
D
Damping coefficient 27
Deaying, longitudinal displacement wave equation 44
Deaying, transverse displacement wave equation 44
Decay constant 27, 43
Decay wave equation of volume expansion 43
Decaying acoustic wave solution 44
Decaying pressure wave in air 71
Destructive interference 14
Direction cosines 9
Dispersion 112
E
Eigen frequency 17, 23
Elastic constant 32
Elastic force 24
Elasticity 23
Elasticity of air 66
Elastic potential energy 26
Elastic potential energy density 51
Electrical impedance 55, 123, 127, 133
Electric conductivity 33
Equation of motion due to pressure gradient 70
Equation of motion for longitudinal vibration 40
Equation of motion governing series of point masses 30
Equation of motion of air compression 36
Equation of motion of isotropic solids 87
Equation of motion of spring-point mass system 26
Equation of motion of unit volume 37
Equation of motion of unit volume with velocity-damping force term 43
Euler's notation 45
F
Faraday's law 134
Fixed-end reflection 18
Forced oscillation 24, 28
Fourier series 12
Fourier's theorem 12
Frequency demodulation 83
Frequency modulation 82
Frequency ranges of vowel and consonants 74
G
Gas law 68
Gradient operation 11
H
Harmonic oscillation 24, 27
Harmonics 13
Hooke's law 32
Hooke's law in air 35
Hooke's law of continuum 33
Human ear 72
I
Impedance matching 55
Initial condition 16
Inverse piezoelectric effect 137
K
Kinetic energy density 53
L
Laplacian 11
Law of reflection 59
Law of refraction 59
Linear differential equation 28
Longitudinal and transverse vibrations in solid 39
Longitudinal elastic wave equation 42
Longitudinal elastic wave velocity 42
Longitudinal wave 4
M
Magnetic force 133
Mechanical impedance 123
N
Natural frequency 27
Normal strain 32
Normal stress 33
O
One-dimensional wave equation of series of point masses 29
Open-end reflection 19
Overtones 77
P
Particle velocity 2
Period 2
Phase velocity 1, 8
Phase velocity as a delay of oscillation in a series of point masses 30
Phase velocity as a material constant 33, 43
Phase velocity of air expansion (compression) wave 38
Phase velocity of air pressure wave 38
Phase velocity of compression wave in isotropic media 88
Phase velocity of longitudinal displacement wave 34
Phase velocity of rotation wave in isotropic media 88
Phase velocity of transverse displacement wave 34
Piezoelectric effect 137
Plane strain wave solution 98
Plane wave 15
Plane wave in elastic media 91
Primary (P-) wave 88, 89
Propagation constant 9
Propagation vector 9, 92
P-wave velocity 96
R
Radiation pressure 69
Rayleigh-Lamb frequency equation 116
Rayleigh surface wave 111
Reflectance 63
Reflection at free surface with various incident waves 104
Reflection coefficient 62
Resonance 14
Resonance of wave 20
Resonator 14
Resonator mode number 21
Rotation wave equation of solids 88
S
Scalar potential 88
Scalar potential wave equation 89
Secondary (S-) wave 88, 89
Shear-Horizontal wave solution 102
Shear modulus 34
SH waves in a plate 112
Snell's law 59
Spatial frequency 7
Spring constant 27
Spring-mass system 24
Standing wave 13
Stiffness 27
S-wave velocity 96
T
Temporal frequency 7
Transmission coefficient 62
Transmittance 63
Transverse elastic wave equation 42
Transverse elastic wave velocity 42
Transverse waves 5
Traveling waves 8
U
Unforced oscillation 24, 26
V
Variable separation 15
Vector potential 88
Vector potential wave equation 89
Velocity damping 26
Volume compression wave equation of solids 88
W
Wave equation 9
Wave equation of air density change 39
Wave equation of air pressure 38
Wave equation of air volume expansion (compression) 37
Wave equation of density change 39
Wavefront 92
Wavelength 2
Y
Young's modulus 33

Footnotes
1
You may wonder what if we place the tennis ball at a point where the x coordinate is 0. It is unclear if the
ball stays there forever or fall in either the positive or negative x direction. In theory, the ball does not fall.
In reality, it will fall either direction as another force, such as air pressure acts on it.
 Table of Contents
Front Matter
1. General Discussions of Waves
2. Wave Dynamics
3. Propagation of Acoustic Waves in Air
4. Propagation of Acoustic Waves in Solids
5. Electrical-Mechanical Transduction
Back Matter