You are on page 1of 28

crystals

Article
Analysis of Longitudinal Waves in Rod-Type
Piezoelectric Phononic Crystals
Longfei Li and Yongqiang Guo *
Key Laboratory of Mechanics on Disaster and Environment in Western China, Ministry of Education,
and School of Civil Engineering and Mechanics, Lanzhou University, Lanzhou 730000, China; lilf14@lzu.edu.cn
* Correspondence: guoyq@lzu.edu.cn; Tel.: +86-931-8914-560

Academic Editors: Victor J. Sanchez-Morcillo, Vicent Romero-Garcia and Luis M. Garcia-Raffi


Received: 14 February 2016; Accepted: 12 April 2016; Published: 18 April 2016

Abstract: Phononic crystals can be used to control elastic waves due to their frequency bands. This
paper analyzes the passive and active control as well as the dispersion properties of longitudinal
waves in rod-type piezoelectric phononic crystals over large frequency ranges. Based on the Love
rod theory for modeling the longitudinal wave motions in the constituent rods and the method
of reverberation-ray matrix (MRRM) for deriving the member transfer matrices of the constituent
rods, a modified transfer matrix method (MTMM) is proposed for the analysis of dispersion curves
by combining with the Floquet–Bloch principle and for the calculation of transmission spectra.
Numerical examples are provided to validate the proposed MTMM for analyzing the band structures
in both low and high frequency ranges. The passive control of longitudinal-wave band structures is
studied by discussing the influences of the electrode’s thickness, the Poisson’s effect and the elastic
rod inserts in the unit cell. The influences of electrical boundaries (including electric-open, applied
electric capacity, electric-short and applied feedback control conditions) on the band structures are
investigated to illustrate the active control scheme. From the calculated comprehensive frequency
spectra over a large frequency range, the dispersion properties of the characteristic longitudinal
waves in rod-type piezoelectric phononic crystals are summarized.

Keywords: piezoelectric phononic crystals; periodic composite rods; longitudinal waves; modified
transfer matrix method; method of reverberation-ray matrix; frequency spectra; band structures; high
frequency analysis; poisson’s effect

1. Introduction
Phononic crystal, which refers to natural or artificial materials characterized by periodic
modulations of elastic properties, was put forward about two decades ago [1]. Since then, it has
been widely investigated for applications based on controlling elastic waves [2], because it possesses
frequency bands.
To control longitudinal waves, rod-type phononic crystals have been proposed. For example,
Asiri et al. [3] and Asiri [4] devised periodic passive struts and mounts, which are essentially
rod-type elastic phononic crystals, to isolate undesirable longitudinal vibrations in the helicopters
and automobiles, respectively. The investigations of longitudinal waves in rod-type elastic phononic
crystals up to 2014 were reviewed in Guo and Fang [5]. Unfortunately, these periodic elastic rods can
only control the propagation of longitudinal waves in a passive mode. That is, their band structures
are totally fixed as long as the geometrical, material and boundary configurations of the unit cells
are determined. The frequency bands may not always fit to the external excitations. Therefore,
rod-type phononic crystals containing smart material components have been presented in order
to control the longitudinal waves in an active mode. Their frequency bands can be adjusted by
tuning the temperature [6,7], magnetic [8] or electric [9–26] field to suit the external excitations.

Crystals 2016, 6, 45; doi:10.3390/cryst6040045 www.mdpi.com/journal/crystals


Crystals 2016, 6, 45 2 of 28

Among these smart periodic rods, the periodic piezoelectric rods [9–26] have particularly been
paid more attention, because the piezoelectric materials are widely used and the electric field is
comparatively easier to control. Thus far, two kinds of rod-type piezoelectric phononic crystals
have been presented to actively control longitudinal-wave bands via electromechanical coupling.
One kind is formed by periodically bonding shunted piezoelectric patches on the host rod, and the
other is shaped through periodically arranging the piezoelectric constituent rods [9]. The attenuation
of longitudinal vibrations/waves in rods with periodic shunted piezoelectric patches have been
studied by Thorp et al. [10], Chen et al. [11] and Lossouarn et al. [12,13] using the transfer matrix
method (TMM). Here, we focus on the longitudinal vibrations/waves in periodically arranged
piezoelectric rods. After Baz [9] proposed the seminal concept about active control of periodic
structures, Singh et al. [14] illustrated the effectiveness of the periodic rod with piezoelectric actuators
by analyzing the frequency bands and responses with the TMM. Asiri et al. [15] and Asiri [16,17]
predicted the attenuation of longitudinal waves in periodic binary piezoelectric-elastic composite
rods using the combination of spectral element method and TMM, and validated their analysis by
experiment. Using the same TMM, Li et al. [18] and Wang et al. [19] analyzed the localization of
frequency bands in randomly disordered periodic binary piezoelectric-elastic rods without and with
initial stress, respectively. Recently, Degraeve et al. [20] proposed to use periodic electrical boundary
conditions realized by periodically embedded electrodes to tune Bragg bands in uniform homogeneous
piezoelectric rods. This kind of electrical charge Bragg band gap in cases of electric-open, external
capacitances and electric-short boundary conditions was studied by an analytical method (AM),
finite element method (FEM) and experimental measurements. The model of Degraeve et al. [20]
was extended by Kutsenko et al. [21] to consider the negative capacitance and by Ponge et al. [22]
to devise a tunable Fabry–Perot resonator whose performance was validated by TMM and FEM
analysis as well as experiment. Degraeve et al. [23] also extended their previous model by considering
the different electrodes in the unit cell with different electrical boundaries and studied the electrical
charge Bragg bands of the new phononic crystal with a semi-analytical method and experiment. The
extensions to a model with elastic rod insert in the unit cell and with negative external capacitance
were conducted in Reference [24] using the previous AM, FEM and experiment. Kutsenko et al. [25]
further studied the band structures of more general rod-type piezoelectric phononic crystals where
the unit cell consists of several piezoelectric or elastic-piezoelectric rods by TMM. The effective
constitutive parameters were also derived. Vasilenko and Rogacheva [26] provided the similar effective
modulus and the electroelastic state equations of periodic piezoelectric and adhesive rods by the
homogenization method.
Although the above-mentioned studies make the longitudinal-wave band structures of
periodically arranged piezoelectric or elastic-piezoelectric rods relatively clear, to the authors’
knowledge, pending problems still exist in four aspects: (1) Few studies have analyzed the longitudinal-
wave bands in the high frequency range. The first reason is that the analytical method, like the
conventional TMM, is proposed on the basis of classical rod theory valid only within pretty low
frequency range. The second one is that the numerical method, like the versatile FEM, may be
inefficient for the high frequency analysis since large amount of elements and very small time step
are required in this case. (2) The mechanical vibration of the electrodes and the Poisson’s effect of the
constituent rods are neglected without understanding their influences. The influence of the elastic rod
inserts on the band structures is not clear enough. (3) Two electrical control manners in general have
been proposed, i.e., setting the external voltage through the feedback control gain [14–17] and tuning
the external electric capacity [18–26], since the electric-open and electric-short conditions correspond
to zero and infinity electric capacities, respectively. The connection and difference of the influences
of the applied feedback control and of the applied electric capacity on the frequency bands are not
known. (4) The wave dispersion property has been solely represented by the frequency–wavenumber
dispersion curve. However, other forms of dispersion curves such as the frequency–wavelength and
Crystals 2016, 6,
Crystals 2016, 45
6, 45 33 of
of 29
28

Aiming at solving the above four pending problems, this paper presents a modified transfer
frequency–phase velocity spectra are also vital to convey wave dispersion properties in different
matrix method (MTMM) to analyze the longitudinal waves in rod-type piezoelectric phononic
perspectives and engineering practices [5].
crystals, whose unit cell consists of any number of piezoelectric or elastic-piezoelectric rods with
Aiming at solving the above four pending problems, this paper presents a modified transfer matrix
electrodes covered on the ends of piezoelectric rods. The proposed MTMM combines the Love rod
method (MTMM) to analyze the longitudinal waves in rod-type piezoelectric phononic crystals, whose
theory [27,28], [29] (pp. 139–142), [30], the method of reverberation-ray matrix (MRRM) [5] and the
unit cell consists of any number of piezoelectric or elastic-piezoelectric rods with electrodes covered
transfer matrix method (TMM) [14–19,21,22,25] to analyze the dispersion curves by considering the
on the ends of piezoelectric rods. The proposed MTMM combines the Love rod theory [27,28], [29]
Floquet–Bloch theorem [31,32] and to calculate the transmission spectra. The Love rod theory is
(pp. 139–142), [30], the method of reverberation-ray matrix (MRRM) [5] and the transfer matrix
adopted to model the longitudinal wave motions in the piezoelectric, elastic and electrode
method (TMM) [14–19,21,22,25] to analyze the dispersion curves by considering the Floquet–Bloch
constituent rods of the unit cell. The MRRM is utilized to derive the member transfer matrices of the
theorem [31,32] and to calculate the transmission spectra. The Love rod theory is adopted to model
constituent rods. Therefore, the proposed MTMM is effective over a large frequency range as long as
the longitudinal wave motions in the piezoelectric, elastic and electrode constituent rods of the unit
the Love rod theory is valid, as illustrated by the numerical examples. The influences of the
cell. The MRRM is utilized to derive the member transfer matrices of the constituent rods. Therefore,
electrode’s thickness, the rods’ cross-sectional dimension, the elastic rod inserts and the applied
the proposed MTMM is effective over a large frequency range as long as the Love rod theory is
electric capacity and the applied feedback control on the band structures are studied numerically.
valid, as illustrated by the numerical examples. The influences of the electrode’s thickness, the rods’
All the frequency-related dispersion curves are calculated to show comprehensively the dispersion
cross-sectional dimension, the elastic rod inserts and the applied electric capacity and the applied
properties of characteristic longitudinal waves in rod-type piezoelectric phononic crystals. This
feedback control on the band structures are studied numerically. All the frequency-related dispersion
paper is organized as follows. Section 2 derives the formulation of the proposed MTMM. Numerical
curves are calculated to show comprehensively the dispersion properties of characteristic longitudinal
examples are given in Section 3 to validate the MTMM, to demonstrate the influences of the
waves in rod-type piezoelectric phononic crystals. This paper is organized as follows. Section 2 derives
electrode’s thickness, the rods’ cross-sectional dimension and the elastic rod inserts for the passive
the formulation of the proposed MTMM. Numerical examples are given in Section 3 to validate the
control, to illustrate the influences of the applied electric capacity and feedback control for the active
MTMM, to demonstrate the influences of the electrode’s thickness, the rods’ cross-sectional dimension
control, and to indicate all the frequency-related dispersion curves in both low and high frequency
and the elastic rod inserts for the passive control, to illustrate the influences of the applied electric
ranges. Section 4 draws conclusions.
capacity and feedback control for the active control, and to indicate all the frequency-related dispersion
curves in both
2. Analysis low and high frequency
of Longitudinal ranges. Section
Waves in Rod-Type 4 draws conclusions.
Piezoelectric Phononic Crystals with the MTMM
2. Analysis of Longitudinal Waves in Rod-Type Piezoelectric Phononic Crystals with the MTMM
2.1. Basic Model

2.1. Basic Consider


Modela periodic piezoelectric composite rod with its unit cell consisting of alternate
homogeneous piezoelectric and elastic rods as well as electrodes between them. The schematic of the
unit cell Consider
is shown a periodic
in Figure piezoelectric compositeand
1, where m members rodNwith its( Nunit
joints = m cell consisting
+ 1 ) are denotedoffrom
alternate
left to
homogeneous piezoelectric and elastic rods as well as electrodes between
right by 1,2,, i, j,,(m − 1), m and 1,2,, I , J , K ,,( N − 1), N , respectively. All of the components them. The schematic
of the unit cell is shown in Figure 1, where m members and N joints (N “ m ` 1) are denoted from
are assumed to be rigidly connected. The piezoelectric rods are polarized along the length direction.
left to right by 1, 2, ¨ ¨ ¨ , i, j, ¨ ¨ ¨ , pm ´ 1q, m and 1, 2, ¨ ¨ ¨ , I, J, K, ¨ ¨ ¨ , pN ´ 1q, N, respectively. All of the
The electrodes are covered at the ends of the piezoelectric rods, to which any of the four electrical
components are assumed to be rigidly connected. The piezoelectric rods are polarized along the
boundaries (electric-open, applied electric capacity, electric-short and applied feedback control
length direction. The electrodes are covered at the ends of the piezoelectric rods, to which any of the
conditions) is imposed. Therefore, the longitudinal effect of the piezoelectric rods is utilized. Figure
four electrical boundaries (electric-open, applied electric capacity, electric-short and applied feedback
1 also shows the pertinent physical and geometrical parameters of some typical constituent rods,
control conditions) is imposed. Therefore, the longitudinal effect of the piezoelectric rods is utilized.
where the Voigt notation (compressed matrix notation) is used for reducing the orders [33]. c (pqj ) ,
Figure 1 also shows the pertinent physical and geometrical parameters of some typical constituent
erq( j ) , αwhere
rods,
( j) ( j)
p, q = 1,notation
rs , ν rsthe( Voigt , r, s = 1,2,3 )matrix
2,3, 4,5,6(compressed and ρnotation)
( j)
are theis elastic
used forconstants,
reducingthethepiezoelectric
orders [33].
p jq p jq p the
constants, jq pdielectric
jq constants, the Poisson’s ratios and
p j q the material density of the piezoelectric
c pq , erq , αrs , νrs (p, q “ 1, 2, 3, 4, 5, 6, r, s “ 1, 2, 3) and ρ are the elastic constants, the piezoelectric
( j)
rod j , respectively;
constants, A constants,
the dielectric , I y( j ) , I z( jthe
)
and l ( j ) areratios
Poisson’s the cross-sectional area,density
and the material secondofmoment of inertia
the piezoelectric
p jq p jq
rod j, respectively;
about y and z axisAp jq ,and
Iy , the
Iz and p jq are the cross-sectional
rod llength, , ν rs( h )second
respectively; c (pqh ) area, , ρ ( h ) ,moment A( h ) , I of , I z( h ) and
( h ) inertia
l ( hy)
about
y
p hq p hq p hq p h q p hq p hq p h q
and z axis
are the and the rod length,
corresponding respectively;
parameters c pq , νrsrod
of the elastic , ρ h ,; A
and , Icy( i ) ,, Iνz ( i ) ,and
ρ (i )l, Aare, the
(i )
I ( i ) corresponding
, I ( i ) and l ( i )
pq rs y z
pi q
pi q pi q
pi q pi q pi q pi q
parameters
are those of of
thethe i . h; and c pq , νrs , ρ , A , Iy , Iz and l are those of the electrode i.
elastic rod
electrode

Figure
Figure 1.
1. The
The unit
unit cell
cell of
of general
general periodic
periodic piezoelectric rods.
piezoelectric rods.
Crystals 2016, 6, 45 4 of 28
Crystals 2016, 6, 45 4 of 29

2.2.
2.2. Coordinate
Coordinate Systems
Systems and
and Physical
Physical Variables
Variables
For
For the the convenience
convenience of ofsystem
systemand andjoint
jointdescription,
description,aaglobal coordinates(pX,
globalcoordinates X , YY,
, ZZq
) is
is set
set up
up as
as
shown
shown ininFigure Figure 2a.2a. To facilitate
To facilitate the member
the member description,
description, a pair
a pair of dual localofcoordinates
dual localpxcoordinates
JK , y , z JK q
JK

( x JK ,pxy KJKJ,, zyJK


and K J), zand x KJ up
K J q is(set , y KJfor
, z KJ
any) is set up
typical for any
member typical
j (also JK or K J jin(also
member
called called JKcoordinates),
the pertaining or KJ in
as also
the indicated
pertaining in Figure 2a.
coordinates), as also indicated in Figure 2a.

Figure
Figure 2.
2. The
The description
description of
of the
the unit
unit cell:
cell: (a)
(a) the
the global
global and
and local
local coordinates;
coordinates; (b)
(b) the
the convention
convention of
of
generalized displacements and forces; and (c) the convention of wave amplitudes.
generalized displacements and forces; and (c) the convention of wave amplitudes.

To make the later derivation clear, hereafter we use u , v and w to express the
To make the later derivation clear, hereafter we use u, v and w to express the displacements of
displacements of a particle along x , y and z axes, respectively. f x , f y and f z are used to
a particle along x, y and z axes, respectively. f x , f y and f z are used to represent the specified forces
represent
along x, y andthe specified
z axes of aforces
surfacealong
particle, y and z axes
x , respectively. The of a surface
resulting axial particle,
force atrespectively.
a cross-section The
of
resulting axial force
the constituent rod isatdenoted
a cross-section
by N. uof the constituent
X and rod is denotedand
p X are the displacement N . u X and
by resulting externalp X force
are the
at
displacement and resulting external force at the unit cell ends along X axis, respectively. For the
the unit cell ends along X axis, respectively. For the electrical variables: ϕ is the electric potential at
a cross-section
electrical of theϕconstituent
variables: is the electric q is the electric
rod; potential charge density
at a cross-section of per
the surface
constituent area;rod; and qtheistotal
the
electric charge on a cross-section is denoted by Q.
electric charge density per surface area; and the total electric charge on a cross-section is denoted by
Q . In general, the axial displacements and forces are the fundamental physical variables to analyze
the system, which are shown in Figure 2b. u XL (u XR ) and p XL (p XR ) denote the axial displacement and
In general, the axial displacements and forces are the fundamental physical variables to analyze
force of joint 1 (N), respectively, which are described in coordinates pX, Y, Zq. The axial displacement
the system, which are shown in Figure 2b. uXL ( uXRJK) and JK p XL JK
( p XRJK) denote the axial displacement
and force of a typical member j at cross-section x in px , y , z q are denoted by u JK px JK q and
and force of joint 1 ( N ), respectively, which are described in coordinates ( X , Y , Z ) . The axial
N JK px JK q, respectively. Those physical variables can also be described in px K J , yK J , zK J q as uK J px K J q
displacement
and N K J px K J q,and force of aFor
respectively. typical member j constituent
a piezoelectric at cross-section
rod, sayxJKrod x JK , electric
in j,( the y JK , z JK )field
are intensity
denoted
JK
by ( x JK ) and asNvector
and udisplacement
JK
( x JK )quantities
, respectively.
as wellThose physicalpotential
as the electric variables andcan alsoasbe
charge described
scalar quantities in
JK JK JK
( x KJneeded
are , y KJ , z KJ to
) as u KJ ( x KJ ) and
be considered N KJanalysis.
in the ( x KJ ) , respectively.
They are represented q, Dx px JK
as Ex pxconstituent
For a piezoelectric rod,
q, ϕsayJK px JK q
rod
KJ KJ KJ
and
j , theQ JKelectric
px JK q infield
px JKintensity
, y JK , z JK qand
as well as Ex px K Jas
displacement q, D q, ϕK J px K J qasand
x px quantities
vector QKas
well
J px K J q in px K J , yK J , zK J q,
the electric potential
respectively, as also given in Figure 2b. It should be noted that all the physical variables are deemed as
and charge as scalar quantities are needed to be considered in the analysis. They are represented as
positive when they are along the positive direction of the pertaining coordinate. During the following
E xJK ( x JK ) , DxJK ( x JK ) , ϕ JK ( x JK ) and Q JK ( x JK ) in ( x JK , y JK , z JK ) as well as E xKJ ( x KJ ) , DxKJ ( x KJ ) ,
analysis, the generalized displacements and forces are expressed by the wave amplitudes, which are
ϕ KJ ( x KJ ) and Q KJ ( x KJ ) in ( x KJ , y KJ , z KJ ) , respectively, as also given in Figure 2b. It should be
depicted in Figure 2c.
noted that all the physical variables are deemed as positive when they are along the positive
2.3. Governing
direction of theEquations
pertaining andcoordinate.
Wave Solutions
During of a the
Constituent
following Rodanalysis, the generalized displacements
and forces are expressed by the wave amplitudes, which are depicted in Figure 2c.
Based on the basic idea of Love rod theory [27,28], [29] (pp. 139–142), [30], which considers the
Poisson’s effect, the time-domain equations governing the longitudinal wave motions in any of the
2.3. Governing Equations and Wave Solutions of a Constituent Rod
constituent rods described in its dual local coordinates can be derived by Hamilton principle [29]
(pp. Based
126–131), on the basic
[34]. Theidea of Love rod
assumptions theory
of the Love [27,28],
rod with[29]considering
(pp. 139–142), the[30], which effect
Poisson’s considers the
include:
Poisson’s
(1) The axial effect, the time-domain
displacement (electricequations
potential)governing
is uniformly thedistributed
longitudinal onwave motions in any
any cross-section of theof rod.
the
constituent rods described in its dual local coordinates can be derived
Hence, the axial displacement (electric potential) at the centroid is used to represent that of the wholeby Hamilton principle [29] (pp.
126–131),
cross-section. [34].(2)
TheTheassumptions
two lateralof the Love rod are
displacements with considering
linearly the Poisson’s
distributed effect include:They
on the cross-section. (1) Theare
axial displacement (electric potential) is uniformly distributed on any cross-section of the rod. Hence,
the axial displacement (electric potential) at the centroid is used to represent that of the whole
Crystals 2016, 6, 45 5 of 28

zero at the centroid and alter linearly with the corresponding coordinates. The proportional coefficients
are the multiplication of the corresponding Poisson’s ratios with the axial strain of that cross-section.
Apply these assumptions and the three-dimensional elasticity [33,34], [35] (pp. 57–60), the derivation
for a piezoelectric rod is detailed in Appendix A, and that for an elastic rod which representing any of
the elastic rods and the electrodes is briefed in Appendix B.
For any piezoelectric rod in the unit cell, say member j, applying Fourier transform [36] (p. 725) to
the time-domain governing equations and constitutive relations, as given in Equations (A12) and (A11),
leads to the frequency-domain counterparts as

” ı d2 û
pc ` e2 {αqA ´ ρpν12
2 2
Iz ` ν13 Iy qω 2 ` ρAω 2 û “ 0 (1)
dx2
” ı dû d ϕ̂ dû d ϕ̂
2 2
N̂ “ cA ´ ρpν12 Iz ` ν13 Iy qω 2 ` eA , Q̂ “ ´eA ` αA (2)
dx dx dx dx
where ω is the circular frequency and a caret over a physical variable signifies the corresponding
quantity in the frequency domain here and after. û ( N̂) and ϕ̂ (Q̂) are the axial displacement (force) and
the electric potential (charge) at any cross-section x, respectively. c, e and α, as given in Appendix A,
are the equivalent elastic, piezoelectric and dielectric constants, respectively; ρ is the material density;
ν12 (ν13 ) and Iy (Iz ) are the Poisson’s ratio and the cross-sectional moment of inertia with respect to
y (z) axis, respectively; and A is the cross-sectional area. When Q̂ “ ´ D̂x A is deemed as the scalar
electric charge at the initial end of the piezoelectric rod with D̂x the uniform electric displacement,
the second formula of Equation (2) leads to

d ϕ̂ e dû Q̂ rl d ϕ̂ e
rl D̂x e Q̂l
dx “ α dx ` αA , V̂ “ ϕ̂plq ´ ϕ̂p0q “ 0 dx dx “ α rûplq ´ ûp0qs ´ 0 α dx “ α rûplq ´ ûp0qs ` αA (3)

where V̂ is the electric potential difference (electric voltage) between the two ends of the piezoelectric
rod. Consider the associated mathematical formulas to the four electric boundary conditions, as shown
in Table 1, the expressions of V̂, Q̂ and d ϕ̂{dx as formulas of mechanical variables can be obtained
from Equation (3), which are also provided in Table 1. The specified electrical coefficients are external
capacitance C and control gain K g for the applied electric capacity and the applied feedback control
conditions, respectively, while the electric-open and the electric-short conditions do not have specified
electrical coefficient.
According to the theory of ordinary differential equations [36] (p. 503), the solution to the spectral
axial displacement can be obtained directly from Equation (1). Substituting the above expressions of
d ϕ̂{dx into the first formula of Equation (2) and then introducing the solution to the spectral axial
displacement, one obtains the wave solution to the spectral axial force. These solutions are expressed as

ûpxq “ a1 eik1 x ` d1 e´ik1 x , N̂pxq “ ζ 1 a1 eik1 x ´ ζ 1 d1 e´ik1 x ` B rûplq ´ ûp0qs (4)


?
where i “ ´1 is the imaginary unit. The first and the second terms in the two formulas of Equation (4),
when combining with the kernel function eiωt of the Fourier transform [36] (p. 725), represent waves
traveling along the negative and positive x-axis that are called as the arriving and departing waves,
respectively.
b a1 and d1 are the corresponding undetermined wave amplitudes. k1 “ ω{c1 and
c1 “ 2 I ` ν2 I qω 2 s{ρA are the wave number and wave speed, respectively.
rpc ` e2 {αqA ´ ρpν12 z 13 y
b
ζ 1 “ ik1 rpc ` e2 {αqA ´ ρpν12
2 I ` ν2 I qω 2 s “ iω ρArpc ` e2 {αqA ´ ρpν2 I ` ν2 I qω 2 s and B are the
z 13 y 12 z 13 y
axial-force influence coefficients of the piezoelectric rod that are irrelative and relative to the electrical
boundary condition, respectively. The expressions of B for the four electrical boundary conditions are
also shown in Table 1. It should be noted in Table 1 that as (C “ 8 or K g “ 0), all the solutions to V̂, Q̂,
d ϕ̂{dx and B for the applied electric capacity (applied electric capacity or applied feedback control)
condition are degenerated to the corresponding ones for the electric open (short) circuit condition.
Crystals 2016, 6, 45 6 of 28

Table 1. The expressions of involved electrical variables as formulas of mechanical variables.

Electrical Boundary Associated


Expressions of V̂ Expressions of Q̂
Conditions Mathematical Formulas
e
Electric-open Q̂ “ 0 α rûplq ´ ûp0qs 0
eA eCA
Applied electric capacity Q̂ “ ´CV̂ αA`Cl rûplq ´ ûp0qs ´ αA`Cl rûplq ´ ûp0qs
Electric-short V̂ “ 0 0 ´ eA
l rûplq ´ ûp0qs
αA e
` ˘
Applied feedback control V̂ “ ´K g rûplq ´ ûp0qs ´K g rûplq ´ ûp0qs ´ l α ` K g rûplq ´ ûp0qs
Electrical Boundary
Expressions of d ϕ̂{dx Expressions of B
Conditions
Electric-open e dû 0
α dx
e dû e C 2
Applied electric capacity α dx ´ α αA`Cl rûplq ´ ûp0qs ´ α2 eA`αCl
CA

e dû 2
Electric-short α dx ´ αle rûplq ´ ûp0qs ´ eαlA
´ eA
e dû 1
`e ˘ `e ˘
Applied feedback control α dx ´ l α ` K g rûplq ´ ûp0qs l α ` Kg

For any elastic rod (or electrode) in the unit cell, say member h (or i), the frequency-domain
governing equation and constitutive relation are obtained by Fourier transform [36] (p. 725) to their
time-domain counterparts, as given in Equations (A17) and (A18)

” ı d2 û
2 2
cA ´ ρpν12 Iz ` ν13 Iy qω 2 ` ρAω 2 û “ 0 (5)
dx2
” ı dû
2 2
N̂ “ cA ´ ρpν12 Iz ` ν13 Iy qω 2 (6)
dx
respectively. From these equations, the wave solutions to the axial displacement and force of an elastic
rod (or electrode) are directly obtained as

ûpxq “ a1 eik1 x ` d1 e´ik1 x , N̂pxq “ ζ 1 a1 eik1 x ´ ζ 1 d1 e´ik1 x (7)


b
where k1 “ ω{ rcA ´ ρpν12 2 I ` ν2 I qω 2 s{ρA and ζ “ ik rcA ´ ρpν2 I ` ν2 I qω 2 s should be noted,
z 13 y 1 1 12 z 13 y
and all the other quantities are identical to those for the piezoelectric rod.

2.4. Transfer Matrix of a Member


Based on the concept of MRRM [5], the transfer matrix of a structural member, which relates the
axial displacement and force of one end with those of the other, can be derived.
For any piezoelectric rod in the unit cell, say member j, the wave solutions to the axial
displacements and forces in the dual coordinates px JK , y JK , z JK q and px K J , yK J , zK J q can be written
from Equation (4). When the constants satisfying Γ JK “ ΓK J “ Γp jq (Γ “ k1 , ζ 1 , B, l) is noticed, these
wave solutions are expressed as
# + « ff » pjq JK
fi # + « ff # +
ik1 x JK
û JK px JK q 1 1
– e 0 a1 0 0 û JK p0q
“ p jq p jq
fl JK `
N̂ JK px JK q ´Bp jq B p jq û JK pl p jq q
pjq JK
ζ1 ´ζ 1 0 e´ik1 x d1
# + « ff » pjq K J
fi # + « ff # + (8)
ik1 x KJ
ûK J px K J q 1 1
– e 0 a1 0 0 ûK J p0q
“ p jq p jq
fl KJ `
N̂ K J px K J q ´Bp jq B p jq ûK J pl p jq q
pjq K J
ζ1 ´ζ 1 0 e´ik1 x d1
Crystals 2016, 6, 45 7 of 28

Due to the uniqueness of the physical variables at any cross section (x JK “ l p jq ´ x K J ) of a typical
member JK, compatibility conditions exist between the displacement (force) expressed in the dual
local coordinates
û JK px JK q “ ´ûK J pl p jq ´ x JK q, N̂ JK px JK q “ N̂ K J pl p jq ´ x JK q (9)

Substituting Equation (8) into Equation (9) gives the phase relations of the typical member j
pjq pjq pjq pjq
JK l KJ KJ l JK
a1 “ ´e´ik1 d1 , a1 “ ´e´ik1 d1 (10)

By noticing û JK pl p jq q “ ´ûK J p0q and ûK J pl p jq q “ ´û JK p0q, letting x JK “ 0 and x K J “ 0 in


Equation (8), and then introducing Equation (10) into the resulting equation, one obtains the relations
between the displacements/forces at the member ends and the departing wave amplitudes
» fi #
# + pjq pjq +
l JK
û JK p0q 1 ´e´ik1 d1
“– fl KJ
N̂ JK p0q
pjq pjq p jq p jq pjq pjq
rBp jq e´ik1 l
´ pζ 1 ` Bp jq qs r´pζ 1 ´ Bp jq qe´ik1 l
´ B p jq s d1
# + »
pjq
fi # + (11)
pjq JK
ûK J p0q ´e´ik1 l 1 d1
“– fl KJ
N̂ K J p0q p jq pjq pjq pjqpjq p jq
r´pζ 1 ´ Bp jq qe´ik1 l ´ B p jq s rBp jq e´ik1 l ´ pζ 1 ` Bp jq qs d1

Eliminating the departing wave amplitudes from Equation (11), one obtains the transfer relation
of the typical piezoelectric member j
# + # + « ff # +
ûK J p0q p jq û JK p0q 1 t11 t12 û JK p0q
“T “ (12)
N̂ K J p0q N̂ JK p0q ∆ t21 t22 N̂ JK p0q

where Tp jq is the transfer matrix of piezoelectric member j. The quantities forming Tp jq are
pjq pjq p jq pjq pjq p jq pjq pjq p jq
∆ “ ´Bp jq pe´2ik1 l
´ 1q ` 2ζ 1 e´ik1 l
, t11 “ ´t22 “ ´pζ 1 ´ Bp jq qe´2ik1 l
´ pζ 1 ` Bp jq q
2 (13)
pjq pjq
l p jq 2 pjq pjq
l p jq pjq pjq
t12 “ e´2ik1 ´ 1, t21 “ ´pζ 1 q pe´2ik1 ´ 1q ` 2ζ 1 Bp jq pe´ik1 l
´ 1q

The transfer relation of any elastic rod (electrode) in the unit cell, say member h (i), can be derived
by starting from Equation (7) and then using a similar way as that for the piezoelectric member j. It can
be expressed as the same form as Equation (12), but the quantities building the member transfer matrix
TpΠq (Π “ h, i) should be

pΠq pΠq pΠq pΠq pΠq pΠq pΠq pΠq pΠq 2 pΠq pΠq
∆ “ 2ζ 1 e´ik1 l
, t11 “ ´t22 “ ´ζ 1 pe´2ik1 l
` 1q, t12 “ e´2ik1 l
´ 1, t21 “ ´pζ 1 q pe´2ik1 l
´ 1q (14)

Please note that if the single local coordinates (other than the dual local coordinates) is used
to describe a structural member, the conventional member transfer matrix can be derived. Its
components are
pjq pjq pjq pjq p jq p jq pjq pjq p jq pjq pjq
∆ “ ´Bp jq pe´ik1 l
´ eik1 l
q ` 2ζ 1 , t11 “ t22 “ pζ 1 ´ Bp jq qe´ik1 l
` pζ 1 ` Bp jq qeik1 l
(15)
pjq pjq
l
pjq pjq
l p jq 2 pjq pjq
l
pjq pjq
l p jq pjq pjq pjq pjq
t12 “ ´pe´ik1 ´ eik1 q, t21 “ ´pζ 1 q pe´ik1 ´ eik1 q ` 2ζ 1 Bp jq pe´ik1 l
` eik1 l
´ 2q

for the typical piezoelectric rod j and

pΠq pΠq pΠq pΠq


l
pΠq pΠq
l
pΠq pΠq
l
pΠq pΠq
l
∆ “ 2ζ 1 , t11 “ t22 “ ζ 1 pe´ik1 ` eik1 q, t12 “ ´pe´ik1 ´ eik1 q
(16)
pΠq 2 pΠq pΠq pΠq pΠq
t21 “ ´pζ 1 q pe´ik1 l ´e ik1 l
q
Crystals 2016, 6, 45 8 of 28

for the typical elastic rod h or electrode i (Π “ h, i). Although the derivation of member transfer
relations on the basis of members’ dual local coordinates in our MTMM seems tangled when compared
with the process based on single local coordinates, it is noticed from the comparison of Equations (13)
and (14) with Equations (15) and (16), respectively, that the so-derived member transfer matrices abstain
from exponentially function in form of e`ik1 l . Namely, in Equations (13) and (14), only exponential
function in form of e´ik1 l is involved. However, in Equations (15) and (16), the exponential functions
in form of e´ik1 l and e`ik1 l appear simultaneously.

2.5. Transfer Matrix of a Joint


At the external joint 1, where the current unit cell is connected to the left adjacent unit cell,
the compatibility condition between the member displacement and the joint displacement and the
equilibrium condition between the member force and the joint force are combined to give the joint
transfer relations
# + « ff # + # + « ff # + # +
û12 p0q 1 0 û XL 1 û XL 1 0 û XL r 1 û XL
“ “T “ “T (17)
N̂ 12 p0q 0 1 ´ p̂ XL ´ p̂ XL 0 ´1 p̂ XL p̂ XL

where T1 “ I2 “ă 1, 1 ą is the transfer matrix of external joint 1 with I2 denoting the identity matrix
of second order, and ă ¨ ą here and after denotes diagonal (or block diagonal) matrix. T r 1 “ă 1, ´1 ą
is a variant of T1 as p̂ XL other than ´ p̂ XL is used in the transfer relation.
At any internal joint that connects the neighboring constituent rods, the typical joint K (2 ď K ď
N ´ 1) for instance, the continuous conditions of the axial displacements and forces lead to the transfer
relations thereof
# + « ff # + # +
ûKL p0q ´1 0 ûK J p0q K ûK J p0q
“ “T (18)
N̂ KL p0q 0 1 N̂ K J p0q N̂ K J p0q

where TK “ă ´1, 1 ą is the transfer matrix of internal joint K.


The transfer relations of the external joint N connecting the current unit cell with the right adjacent
unit cell can be derived similarly as that of joint 1
# + « ff # + # +
û XR ´1 0 û N p N ´1q p0q û N p N ´1q p0q
“ “ TN (19)
p̂ XR 0 1 N̂ N p N ´1q p0q N̂ N p N ´1q p0q

where T N “ă ´1, 1 ą is the transfer matrix of external joint N and is obviously different from T1 .

2.6. Global Transfer Matrix of the Unit Cell


Considering the joint and member transfer relations alternately from joint 1, member 1¨ ¨ ¨ , till to
joint N gives the global transfer relations of the unit cell
# + # +
û XR û XL
“T (20)
p̂ XR ´ p̂ XL

where T “ T N Tpmq T N ´1 Tpm´1q T N ´2 ¨ ¨ ¨ Tpkq TK ¨ ¨ ¨ Tp2q T2 Tp1q T1 is the global transfer matrix of the
unit cell, and N “ m ` 1 is implied.

2.7. Dispersion Relation of Infinite Periodic Structures


Due to the Floquet–Bloch principle for infinite periodic structures [31,32], the axial displacement
and force of external joint N are related to those of external joint 1 by
# + # +
û XR iqL û XL
“e (21)
p̂ XR ´ p̂ XL
Crystals 2016, 6, 45 9 of 28

where q is the complex wavenumber of the characteristic longitudinal waves in the periodic
piezoelectric composite rod, and L is the length of the unit cell.
The combination of Equations (20) and (21) gives
# + # + # + # +
û XL iqL û XL ´ ¯ û XL 0
T “e or T ´ eiqL I2 “ (22)
´ p̂ XL ´ p̂ XL ´ p̂ XL 0

which indicates that eiqL ” µ is the eigenvalue of the global transfer matrix T with referring to the
definition of the eigenvalues of a matrix [36] (pp. 278–279). The eigenvalue eiqL ” µ can be solved
from the dispersion relation

eiqL “ EigenvaluesrTs or detrT ´ eiqL I2 s “ 0 (23)

by first specifying the frequency ω. Eigenvaluesr¨s and detr¨s denote solving the eigenvalues and
the determinate of a matrix, respectively. The real part q R L and the imaginary part q I L of the
dimensionless complex wavenumber qL, which are called as the phase constant and the attenuation
constant [5,32,37], are further computed from the phase and the amplitude of eiqL ” µ, respectively.
The wavelength λ “ 2π{q and the phase velocity c “ ω{q are then obtained through the corresponding
algebra calculations. Finally, the comprehensive frequency spectra including the |µ| pωq spectra, the
qpωq spectra, the λpωq spectra and the cpωq spectra can thus be drawn, where |¨| here and after
denotes the amplitude of a quantity. These frequency spectra represent the properties of characteristic
longitudinal waves in the periodic piezoelectric composite rod from different viewpoints. From them,
the pass-bands and the stop-bands are easily differentiated.

2.8. Global Transfer Relation of Finite Periodic Structures


For a finite periodic structure consisting of M unit cells, the spectral axial displacement û XM and
force p̂ XM of the very right end are related to the spectral axial displacement û X0 and force p̂ X0 of the
very left end by

M times
hkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkikkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkj
# + M th
hkkkkkkkkkkikkkkkkkkkkj M´1 th
hkkkkkkkkkkikkkkkkkkkkj 2 nd
hkkkkkkkkkkikkkkkkkkkkj 1 st
hkkkkkkkkkkikkkkkkkkkkj # +
û XM N pmq 2 p1q N pmq 2 p 1q N pmq 2 p1q N pmq 2 p1q r 1 û X0
“T T ¨¨¨T T T T ¨¨¨T T ¨¨¨¨¨¨T T ¨¨¨T T T T ¨¨¨T T T (24)
p̂ XM p̂ X0
# + « ff # +
û X0 TG11 TG12 û X0
“ TG “
p̂ X0 TG21 TG22 p̂ X0

where TG is the global transfer matrix of the finite periodic structure formed by multiple matrix
multiplications as shown.
The transmission R of the system can be obtained from

R “ 20log10 |p1{Tq| (25)

where T “ TGrs (r, s “ 1, 2), depending on the observed and referred quantities utilized. Thus, the
transmission spectra can be obtained by first specifying the frequency ω in required range and then
calculating the transmission coefficient R from Equation (25).

3. Numerical Examples

3.1. Validation of the Proposed MTMM


First, consider a periodic piezoelectric composite rod [20] consisting of alternate 10 mm long
circular PZT-5H (PZ29) rod and 0.025 mm long circular brass electrode all with 1 mm diameters. The
piezoelectric PZT-5H constituent rod is poled along the X axis in positive direction. The material and
Crystals 2016, 6, 45 10 of 28

geometrical parameters of the two constituent rods in the unit cell [35] (pp. 358–382) are listed in Table 2
except those specified otherwise. Four kinds of electrical boundaries, i.e., the electric-open, applied
electric capacity, electric-short and applied feedback control conditions, are considered to impose on
the electrodes. In order to compare to the results by Degraeve et al. [20], we neglect the mechanical
vibration and the thickness of the brass electrodes but retain its electrical function for calculating
the phase constant spectra (q R pωqL curves) of this PZT-5H piezoelectric rod with periodic electrical
boundaries in the low and high frequency ranges. The results are given in Figure 3. For the convenience
of description, the engineering frequency f “ ω{p2πq and the dimensionless wavenumber qL{π are
used hereafter. The results of the first three electric boundaries are also compared in Figure 3 with the
corresponding ones by Degraeve et al. using the analytical method (AM) [20]. It should be noted that
the applied electric capacity is specified as the clamped capacitance C “ αA{l “ 1.02 ˆ 10´12 F, which
is the same as that used in [20].

Table 2. Material and geometrical parameters.

Stiffness Constants (GPa)


Materials
c11 c22 c33 c12 c13 c23
PZT-5H 117.0 126.0 126.0 84.1 84.1 79.5
Brass 162.46 162.46 162.46 82.58 82.58 82.58
Epoxy 6.98 6.98 6.98 3.76 3.76 3.76
Dielectric Constants
Poisson’s Ratios Piezoelectric Constants (C{m2 )
Materials (ˆ10´9 F{m)
ν12 ν13 e11 e12 e13 α11
PZT-5H 0.41 0.41 23.3 ´6.5 ´6.5 13.02
Brass 0.337 0.337 — — — —
Epoxy 0.35 0.35 — — — —
Cross-Sectional
Mass Density Length Cross-Sectional Area
Moments of Inertia
Materials (kg{m3 ) (mm) (ˆπ mm2 )
(ˆπ mm4 )
ρ l A Iy Iz
PZT-5H 7500 10 1/4 1/64 1/64
Brass 8320 0.025 1/4 1/64 1/64
Epoxy 1180 10 1/4 1/64 1/64

Figure 3a–c indicates that the phase constant spectra of the first three electrical boundaries in the
low frequency range f ď 500 kHz obtained by our proposed MTMM generally agree well with their
counterparts by Degraeve et al. using AM [20]. Nevertheless, the discrepancy becomes observable
in the relatively high frequency range 500 kHz ă f ď 600 kHz. This is because that in the AM by
Degraeve et al. [20], the classical rod theory applicable to pretty low frequency range is adopted. Our
proposed MTMM, which adopts the Love rod theory, is effective till to a relative high frequency
p 1q p 2q pmq
0.9ω0 “ 3.81 MHz with ω0 “ minpωC , ωC , ¨ ¨ ¨ , ωC q the minimum of critical frequencies of the
b
constituent rods (For piezoelectric PZT-5H rod, ωC “ rpc ` e2 {αqAs{rρpν12 2 I ` ν2 I qs “ 4.23 MHz;
z 13 y
b
2 2
For brass electrode, ωC “ cA{rρpν12 Iz ` ν13 Iy qs “ 4.79 MHz) [28]. Figure 3d shows that our proposed
MTMM works also well for analyzing the phase constant spectra of the applied feedback control
condition in the low frequency range. Figure 3a–d illustrates that, in any electric boundary case, the
width of stop-bands decreases with its order. The comparisons of Figure 3a–d indicate that the widths
of stop-bands become wider and the central frequencies of stop-bands become lower in sequence of
electric-open (without band gaps), applied electric capacity, electric-short and applied feedback control
boundaries due to the separate increasing of coefficient B. However, this difference between the band
structures of the four electrical boundaries is inconspicuous in the high frequency range, as shown
Crystals 2016, 6, 45 11 of 28

in Figure 3e, since the band gaps of the latter three electrical boundaries all become trivial in that
frequency range. This is because the axial-force influence coefficient B of the piezoelectric rod, which
is relative to the electric boundary conditions, do not change with frequency. However, the axial-force
influence coefficient ζ 1 , which is irrelative to the electric boundary conditions, obviously increases
near linearly with the frequency ω. In all the four periodically electrical boundary cases, since the B-ζ 1
ratio decreases very quickly with the increasing of frequency, the widths of band gaps determined by
this ratio thus approach to zero quickly even the folding of bands is identifiable. The above results and
the Figure 3e demonstrate that our proposed MTMM works very well in frequency range below 0.9ω0
Crystals 2016, 6, 45 11 of 29
for all four kinds of electrical boundaries.

(a) (b)
600

500

400
Frequency f (kHz)

300

200

Present results using MTMM

100

0
0.0 0.2 0.4 0.6 0.8 1.0
Dimensionless real wavenumber q R L/π

(c) (d)

Figure 3. Cont.
Crystals 2016, 6, 45 Crystals 2016, 6, 45 12 of 28 12 o

3.0

2.9

Electric open
2.8
C = 1.02 × 10−12 F

Frequency f (MHz)
Electric short
K g = 5 × 108 V / m
2.7

2.6

2.5

2.4
0.0 0.2 0.4 0.6 0.8 1.0
Dimensionless real wavenumber q R L/π

(e)
Figure spectra
Figure 3. The phase constant 3. The phase
of theconstant
homogeneousspectraPZT-5H
of the homogeneous PZT-5H
piezoelectric rod piezoelectric rod with periodic
with periodic
electrical boundaries: electrical boundaries:
(a) electric-open (a) electric-open
condition; (b) appliedcondition; (b) applied
electric capacity electric
condition capacity condition with
with
C “ αA{l “ 1.02 ˆ 10 C
´12 = α A / l = 1.02 × 10 −12
F
F; (c) electric-short condition; (d) applied feedback control condition with control condition with
; (c) electric-short condition; (d) applied feedback
8
K g “ 5 ˆ 10 V{m; and (e) 108 four
= 5of× the
K g all V/m ; and (e) all of the
abovementioned four abovementioned
conditions conditions
in high frequency range. in high frequency range.

Figure
Second, consider the same 3a–c indicates
periodic piezoelectric that composite
the phase rodconstant
as above spectra
except of that
the first three electrical boundaries
the diameters
range f ≤ 500 kHz
of all constituent rods are 10 mm [20] and the mechanical vibration of the brass electrodes is generally
the low frequency obtained by our proposed MTMM also agree well w
theirminimum
reckoned in. The current counterparts critical byfrequency
Degraeve of et
theal. using AM
constituent rods [20].
is ωNevertheless,
0 “ 423 kHz. Thus the itdiscrepancy becom
has 0.9ω0 “ 381 kHz.observable
Using ourin the relatively
proposed MTMM, highwefrequency
calculaterange
the phase 500 kHz < f ≤ 600
constants kHz of
spectra . This
thisis because that in
periodic PZT-5H piezoelectric rod covered
AM by Degraeve et by
al. brass electrodes
[20], the below
classical rod 450
theory kHzapplicable
for all the four electrical
to pretty low frequency range
boundary conditions.adopted.
Figure Our
4a–dproposed
providesMTMM, the results associated with the electric-open, applied
which adopts the Love rod theory, is effective till to a relative h
electric capacity, electric-short
frequency and 0.9ωapplied
0 = 3.81 MHz
feedback with ω0 = min(
control ωC(1) , ωC( 2) ,
conditions, , ωC( m ) ) the minimum
respectively, in which theof critical frequencie
lines of ω0 and 0.9ω0 the
are also depicted.
constituent In particular, the
rods results of the
(For electric-open and electric-short
piezoelectric PZT-5H r
boundaries are compared with the2 corresponding ones by Degraeve et al. using the FEM and AM [20]
ωC = [( c + e / α ) A] / [ ρ (ν 12 I z + ν 13 I y )] = 4.23 MHz
2 2
; For brass electro
in Figure 4a,c, respectively. In these two electrical boundary cases, the transmission spectra of a finite
periodic piezoelectricωcomposite
C = cA / [ ρ rod
(ν 122 Iconsisting
z + ν 13 I y )] = 4.79 MHz ) [28]. Figure 3d shows that our proposed MTMM works a
2
of 14 unit cells are also calculated by MTMM,
which are depicted and compared in Figure 5 with the corresponding results by Degraeve et al. [20].
well for analyzing the phase constant spectra of the applied feedback control condition in the l
In order to compare to the numerical transmission from FEM [20], the ordinate in Figure 5a is specified
ˇ frequency range.
ˇ( Figure 3a–d illustrates that, in any electric boundary case, the width of stop-ban
as 20log10 ˇ1{rpc ` e2decreases ˇ , where the parameters of PZT-5H are utilized. In Figure 5b, the
{αqATG12 swith its order. The comparisons of Figure 3a–d indicate that the widths of stop-ban
becomeRwider
ordinate is the transmission computed and the from TG22 by
central using Equation
frequencies (25), for the
of stop-bands become sake lower
of comparing
in sequence of electric-op
with the experiment measurements [20].
(without band gaps), applied electric capacity, electric-short and applied feedback con
boundaries due to the separate increasing of coefficient B . However, this difference between
band structures of the four electrical boundaries is inconspicuous in the high frequency range
shown in Figure 3e, since the band gaps of the latter three electrical boundaries all become trivia
that frequency range. This is because the axial-force influence coefficient B of the piezoelectric r
which is relative to the electric boundary conditions, do not change with frequency. However,
axial-force influence coefficient ζ 1 , which is irrelative to the electric boundary conditions, obviou
Crystals 2016,
Crystals 2016, 6,
6, 45
45 14 of
13 of 28
29

450

ω0

400
0.9ω0

350

300

Frequency f (kHz) 250

200

150 Present results using MTMM

100

50

0
0.0 0.2 0.4 0.6 0.8 1.0
Real wavenumber q R L/π

(a) (b)

Figure 4. Cont.
Crystals 2016,
Crystals 2016, 6,
6, 45
45 15 of 28
14 29

450

ω0

400
0.9ω0

350

300

Frequency f (kHz) 250

200

150
Present results using MTMM

100

50

0
0.0 0.2 0.4 0.6 0.8 1.0
Real wavenumber q R L/π

(c) (d)
Figure 4.
Figure 4. The
The phase
phase constant
constant spectra
spectra ofof the
the periodic
periodic PZT-5H piezoelectric rod
PZT-5H piezoelectric rod covered
covered by
by brass
brass
electrodes with periodic electrical boundaries: (a) electric-open condition; (b) applied electric
electrodes with periodic electrical boundaries: (a) electric-open condition; (b) applied electric capacity
capacity condition
condition with C “ αA{l = αˆA10
with“C1.02 = 1.02
/ l´10 10−10 F ; (c) condition;
F; (c)×electric-short electric-short
and condition;
(d) appliedand (d) applied
feedback control
8
feedback control
condition with K gcondition V{m.K g = 5 × 10 V / m .
“ 5 ˆ 10 with
8
Crystals 2016, 6, 45 15 of 28
Crystals 2016, 6, 45 16 of 29

(a)

(b)
Figure 5. The transmission (in dB units) in a finite periodic PZT-5H piezoelectric rod covered by
Figure 5. The transmission (in dB units) in a finite periodic PZT-5H piezoelectric rod covered by brass
brass electrodes consisting of 14 unit cells with electric-open and electric-short boundaries from our
electrodes consisting of 14 unit cells with electric-open and electric-short boundaries from our MTMM
MTMM and their comparison with the corresponding ones by Degraeve et al. [20]: (a) comparison
and their comparison with the corresponding ones by Degraeve et al. [20]: (a) comparison with the
with the numerical transmission; and (b) comparison with the experimental transmission.
numerical transmission; and (b) comparison with the experimental transmission.

In the electric-open and electric-short boundary cases, Figure 4a,c indicates that the phase
In the spectra
constant electric-open
belowand electric-short
200 kHz from our boundary cases,
proposed Figurein
MTMM 4a,c indicates
general thatwith
agree the phase
both theconstant
results
spectra
of FEM and those of AM by Degraeve et al. [20]. Figure 5a also shows that below 200 kHzand
below 200 kHz from our proposed MTMM in general agree with both the results of FEM the
those of AM byspectra
transmission Degraeve et al.
from our [20].
MTMMFigurein5ageneral
also shows
agreethat below
with 200 kHz the
the numerical transmission
transmission spectra
from FEM
from our MTMM
by Degraeve in [20].
et al. general
Theagree with the
observable numericalin
differences transmission
these figuresfrom
mayFEM by Degraeve
be caused by that et the
al. [20].
FEM
The observable differences in these figures may be caused by that the FEM considers
considers all modes and their coupling, while our MTMM considers only the longitudinal mode but all modes and
their
withcoupling, while
taking the our MTMM
Poisson’s effect considers onlyThese
into account. the longitudinal
differencesmode butdue
are also withtotaking thevibration
that the Poisson’sof
electrodes is not taken into account in the FEM (and AM) models by Degraeve et al. [20]. Figure 5b
manifests clearly that our predicated transmission spectra in the stop bands agree very well with the
Crystals 2016, 6, 45 16 of 28

effect into account. These differences are also due to that the vibration of electrodes is not taken into
Crystals in
account 2016,
the6, 45
FEM (and AM) models by Degraeve et al. [20]. Figure 5b manifests clearly 17 of 29
that our
predicated transmission spectra in the stop bands agree very well with the measured transmission.
measured transmission. However, in the pass bands, our MTMM results are about 20 dB bigger than
However, in the pass bands, our MTMM results are about 20 dB bigger than the experimental results.
the experimental results. The damping in the experimental specimen may be the primary cause.
The damping in the experimental specimen may be the primary cause. Figures 4c and 5 also indicate
Figures 4c and 5 also indicate that the first stop bands of the model with electric-short boundary
that the first stop bands of the model with electric-short boundary condition predicated by the phase
condition predicated by the phase constant spectra and by the transmission spectra are coincident. In
constant spectra and by the transmission spectra are coincident. In the range of 200 kHz ă f ď 300 kHz,
the range of 200 kHz < f ≤ 300 kHz , the phase constant spectra from our proposed MTMM give
the phase constant spectra from our proposed MTMM give better prediction to those of FEM than
better prediction to those of FEM than the results of AM. The reason is that our proposed MTMM
the results of AM. The reason is that our proposed MTMM considers the Poisson’s effect while the
considers the Poisson’s effect while the AM neglects it. The deviation between our MTMM results
AM neglects it. The deviation between our MTMM results and those from FEM may be due to two
and those from FEM may be due to two reasons. One is that the FEM considers all modes and their
reasons.
coupling,Onewhile
is thatour
theMTMM
FEM considers
considersallonly
modes the and their coupling,
longitudinal mode while our MTMMthe
with considering considers
Poisson’sonly
theeffect.
longitudinal
The other is that the electrode vibration is ignored in the FEM model. In all the four electrical is
mode with considering the Poisson’s effect. The other is that the electrode vibration
ignored
boundarythe
in FEMourmodel. In allworks
the four
wellelectrical boundary cases,
0.9ωour MTMM works well till to the
cases, MTMM till to the frequency 0 = 381 kHz with ω0 = 423 kHz
frequency
the minimum0.9ω0 critical
“ 381 kHz with ωof0 the
frequency “ 423 kHz the rods.
constituent minimum
In thecritical
higherfrequency
frequency of the constituent
range, the proposedrods.
In MTMM
the higher frequency range, the proposed MTMM loss efficacy since the Love
loss efficacy since the Love rod theory becomes invalid. In sequence of electric-open, rod theory becomes
invalid.
appliedInelectric
sequence of electric-open,
capacity, electric-shortapplied electric
and applied capacity,
feedback electric-short
control boundaries, andthe
applied feedback
increasing of
control boundaries, the increasing of stop-band widths and the reduction
stop-band widths and the reduction of the stop-band central frequencies still hold due to the of the stop-band central
frequencies still hold of
separate increasing due B . increasing of coefficient B.
to the separate
coefficient

3.2.3.2.
Passive Control
Passive ControlofofLongitudinal
LongitudinalWaves
Wavesin
in Rod-Type
Rod-Type Piezoelectric PhononicCrystals
Piezoelectric Phononic Crystals
Consider thethe
Consider same rod-type
same rod-typepiezoelectric
piezoelectricphononic
phononiccrystal
crystalwith
with1 1mmmmdiameters
diametersasasininSection
Section3.1,
but3.1,
now reckon in the mechanical vibration of the brass electrodes. The electrode’s thickness,
but now reckon in the mechanical vibration of the brass electrodes. The electrode’s thickness, thethe
rod’s
cross-sectional dimension
rod’s cross-sectional and the elastic
dimension and therodelastic
insertrod
are insert
individually varied to show
are individually variedtheir influences
to show theiron
theinfluences on the band
band structures, structures,
while the otherwhile the other
parameters parameters
remain remain
the same as inthe same
Table 2.as in Table 2.

3.2.1. Influence
3.2.1. ofofthe
Influence theElectrode’s
Electrode’sThickness
Thickness
Specifyingthe
Specifying thethickness
thickness of electrodeasas 0.025
of the electrode 0.025mmmm,, 0.25
0.25mm
mmandand 2.5
2.5mmmm,, wewecalculate thethe
calculate
propagationconstants
propagation constantsspectra
spectraincluding
including the
the phase constant
constant spectra
spectraand
andthe
theattenuation
attenuation constant
constant
spectra
spectra qI (ω )curves)
(q I(pωqL L curves) of the
of the periodic
periodic PZT-5H
PZT-5H piezoelectric
piezoelectric rod
rod covered
covered bybybrass
brasselectrodes
electrodeswith
withthe
theperiodic
four four periodic electrical
electrical boundaries,
boundaries, respectively,
respectively, and compare
and compare them withthem with
those thoseasresults
results as the
the electrode’s
electrode’s
thickness thickness is
is neglected, asneglected, as given
given in Figure 6. in Figure 6.

600 600 600 600

500 Electrode: 0.0 mm 500 500 Electrode: 0.0 mm 500


Electrode: 0.025 mm Electrode: 0.025 mm
Electrode: 0.25 mm Electrode: 0.25 mm

400 Electrode: 2.5 mm 400 400 Electrode: 2.5 mm 400


Frequency f (kHz)

Frequency f (kHz)
Frequency f (kHz)

Frequency f (kHz)

300 300 300 300

200 200 200 200

100 100 100 100

0 0 0 0
-0.3 -0.2 -0.1 0.0 0.2 0.4 0.6 0.8 1.0 -0.3 -0.2 -0.1 0.0 0.2 0.4 0.6 0.8 1.0
Imaginary wavenumber q I L/π Real wavenumber q R L/π Imaginary wavenumber q I L/π Real wavenumber q R L/π

(a) (b)

Figure 6. Cont.
Crystals 2016, 6, 45 17 of 28
Crystals 2016,
Crystals 2016, 6,
6, 45
45 18 of
18 of 29
29

600 600 600 600


600 600 600 600

500 Electrode: 0.0 mm 500 500 Electrode: 0.0 mm 500


500 Electrode: 0.0 mm 500 500 Electrode: 0.0 mm 500
Electrode: 0.025 mm Electrode: 0.025 mm
Electrode: 0.025 mm Electrode: 0.025 mm
Electrode: 0.25 mm Electrode: 0.25 mm
Electrode: 0.25 mm Electrode: 0.25 mm
400 Electrode: 2.5 mm 400 400 Electrode: 2.5 mm 400
400 Electrode: 2.5 mm 400 400 Electrode: 2.5 mm 400

f (kHz)

f (kHz)
f (kHz)

f (kHz)
Frequencyf (kHz)

Frequencyf (kHz)
Frequencyf (kHz)

Frequencyf (kHz)
Frequency

Frequency
Frequency

Frequency
300 300 300 300
300 300 300 300

200 200 200 200


200 200 200 200

100 100 100 100


100 100 100 100

0 0 0 0
0 0 0 0
-0.3 -0.2 -0.1 0.0 0.2 0.4 0.6 0.8 1.0 -0.3 -0.2 -0.1 0.0 0.2 0.4 0.6 0.8 1.0
-0.3 -0.2wavenumber-0.1q I L/π
Imaginary 0.0 0.2 0.4 0.6q R L/π0.8
Real wavenumber 1.0 -0.3 -0.2wavenumber-0.1q I L/π
Imaginary 0.0 0.2 0.4 0.6q R L/π0.8
Real wavenumber 1.0
Imaginary wavenumber q I L/π Real wavenumber q R L/π Imaginary wavenumber q I L/π Real wavenumber q R L/π

(c)
(c) (d)
(d)
Figure 6.
Figure 6. Influence
Influence of
of the
the electrode’s
electrode’s thickness
thickness on
on the
the propagation
propagation constants
constants spectra
spectra of of the
the periodic
periodic
Figure
PZT-5H Influence ofrod
6. piezoelectric thecovered
electrode’s
by thickness
brass on the
electrodes propagation
with periodic constants
electrical spectra of(a)the
boundaries:
PZT-5H piezoelectric rod covered by brass electrodes with periodic electrical boundaries: (a)
periodic PZT-5H
electric-open
electric-open
piezoelectric
condition;
condition;
rod covered
(b) applied
(b) applied electricby
electric
brass electrodes
capacity
capacity
with C
condition with
condition with C= αA
periodic
=α A // llelectrical
= 1.02
=
−12
10boundaries:
×10
1.02 × −12 F ; (c)
F ; (c)
(a)electric-short
electric-opencondition;
condition; (b)(d)applied electric capacity condition with C “= αA{l 8“ 1.02 ˆ 10´12 F;
and applied feedback control condition with K 5
electric-short condition; and (d) applied feedback control condition with K = 5 × 10 V / m .
g × 10 8 V/m.
(c) electric-short condition; and (d) applied feedback control condition withg K g “ 5 ˆ 108 V{m.

Figure 66 illustrates
Figure illustrates that
that with
with anan increase
increase inin the
the electrode’s
electrode’s thickness,
thickness, the
the central
central frequencies
frequencies ofof
Figure
all the 6 illustrates
the stop-bands
stop-bands and that
and the with
the width
width ofan increase
of the
the first in the
first stop-band electrode’s
stop-band slightly thickness,
slightly decrease, the
decrease, although central
although this frequencies
this feature
feature is not of
is not
all
allapparently
the stop-bands
apparently clear as
clear and
as thethe
the width ofthickness
electrode’s
electrode’s the first varies
thickness stop-band
varies from 00 to
from slightly 0.025
to 0.025 mm .. The
decrease,
mm although
The maximum
maximum thisvalue
feature
value of is
of
notattenuation
apparentlyconstants
attenuation clear as in
constants thethe
in electrode’s
the first thickness
stop-band varies
increases from
with the 0 to 0.025
electrode’s mm.
first stop-band increases with the electrode’s thickness. The
thickness. maximum value of
attenuation constants in the first stop-band increases with the electrode’s thickness.
3.2.2. Influence
3.2.2. Influence of of the
the Rod’s
Rod’s Cross-Sectional
Cross-Sectional Dimension
Dimension
3.2.2. Influence of the Rod’s Cross-Sectional Dimension
As the
As the diameter
diameter of of all
all the
the constituent
constituent rodsrods isis 11 mm
mm ,, 55 mm
mm andand 1010 mm
mm ,, respectively,
respectively, the
the
propagation
As the constants
diameter of allspectra
the of the
constituent periodic
rods isPZT-5H
1 mm, 5piezoelectric
mm and 10 rod
mm, covered by
respectively,
propagation constants spectra of the periodic PZT-5H piezoelectric rod covered by brass electrodes brass
the electrodes
propagation
are calculated
are calculated
constants spectra by by ourperiodic
of the
our proposed
proposed MTMM.
PZT-5H
MTMM. Figure 7a–d
piezoelectric
Figure 7a–d provides
rodprovides
covered thethe results
by brass associated
electrodes
results associated with
arewith the
calculated
the
electric-open,
byelectric-open, applied
our proposedapplied electric
MTMM.electric capacity,
Figure capacity,
7a–d provideselectric-short
the results
electric-short and applied
andassociated feedback control conditions,
with the electric-open,
applied feedback applied
control conditions,
respectively.
electric capacity, electric-short and applied feedback control conditions, respectively.
respectively.

300 300 300 300


300 300 300 300

Diameter: 1 mm Diameter: 1 mm
Diameter: 1 mm Diameter: 1 mm
250 Diameter: 5 mm 250 250 Diameter: 5 mm 250
250 Diameter: 5 mm 250 250 Diameter: 5 mm 250
Diameter: 10 mm Diameter: 10 mm
Diameter: 10 mm Diameter: 10 mm

200 200 200 200


200 200 200 200
f (kHz)

f (kHz)
f (kHz)

f (kHz)
Frequencyf (kHz)

Frequencyf (kHz)
Frequencyf (kHz)

Frequencyf (kHz)
Frequency

Frequency
Frequency

Frequency

150 150 150 150


150 150 150 150

100 100 100 100


100 100 100 100

50 50 50 50
50 50 50 50

0 0 0 0
0 0 0 0
-0.3 -0.2 -0.1 0.0 0.2 0.4 0.6 0.8 1.0 -0.3 -0.2 -0.1 0.0 0.2 0.4 0.6 0.8 1.0
-0.3 -0.2wavenumber-0.1q I L/π
Imaginary 0.0 0.2 0.4 0.6q R L/π0.8
Real wavenumber 1.0 -0.3 -0.2wavenumber-0.1q I L/π
Imaginary 0.0 0.2 0.4 0.6q R L/π0.8
Real wavenumber 1.0
Imaginary wavenumber q I L/π Real wavenumber q R L/π Imaginary wavenumber q I L/π Real wavenumber q R L/π

(a)
(a) (b)
(b)

Figure 7. Cont.
Crystals 2016, 6, 45 18 of 28
Crystals 2016, 6, 45 19 of 29
Crystals 2016, 6, 45 19 of 29
300 300 300 300
300 300 300 300
Diameter: 1 mm Diameter: 1 mm
250 Diameter: 15 mm
Diameter: mm 250 250 Diameter: 15 mm
Diameter: mm 250
250 Diameter: 510mm
Diameter: mm 250 250 Diameter: 510mm
Diameter: mm 250
Diameter: 10 mm Diameter: 10 mm

200 200 200 200

f (kHz)

f (kHz)
f (kHz)

f (kHz)
200 200 200 200

f (kHz)

f (kHz)
f (kHz)

f (kHz)
Frequency

Frequency
Frequency

Frequency
150 150 150 150

Frequency

Frequency
Frequency

Frequency
150 150 150 150

100 100 100 100


100 100 100 100

50 50 50 50
50 50 50 50

0 0 0 0
0 -0.3 -0.2 -0.1
Imaginary wavenumber q I L/π
0.0 0.2 0.4 0.6 0.8
Real wavenumber q R L/π
1.0 0 0 -0.3 -0.2 -0.1
Imaginary wavenumber q I L/π
0.0 0.2 0.4 0.6 0.8
Real wavenumber q R L/π
1.0 0
-0.3 -0.2 -0.1 0.0 0.2 0.4 0.6 0.8 1.0 -0.3 -0.2 -0.1 0.0 0.2 0.4 0.6 0.8 1.0
Imaginary wavenumber q I L/π Real wavenumber q R L/π Imaginary wavenumber q I L/π Real wavenumber q R L/π
(c) (d)
(c) (d)
Figure 7. Influence of the rod’s cross-sectional dimension on the propagation constants spectra of
Figure 7. Influence of the rod’s cross-sectional dimension on the propagation constants spectra of the
Figure 7. Influence
the periodic PZT-5Hof the rod’s cross-sectional
piezoelectric dimension
rod covered by brasson the propagation
electrodes withconstants
periodic spectra
electricalof
periodic PZT-5HPZT-5H
piezoelectric rod covered by brassbyelectrodes with periodic electrical boundaries:
boundaries: (a) electric-open condition; (b) applied electric capacity condition with C = α A / l ; (c)
the periodic piezoelectric rod covered brass electrodes with periodic electrical
(a)boundaries:
electric-open condition; (b)condition;
applied electric capacity condition with C “ αA{l;8 C (c) electric-short
electric-short(a) electric-open
condition; and (d) (b) applied
applied feedback electric
control capacitywith
condition condition × 10 V=/ α
K g = 5with
8 V{m. m A. / l ; (c)
condition; and (d) applied feedback control condition with K g
electric-short condition; and (d) applied feedback control condition“ 5 ˆ 10 8
with K = 5 × 10 V / m . g

Figure 7a–d indicates that, in any electrical boundary case, the central frequencies of all the
Figure
Figure7a–d
stop-bands 7a–dindicates
decrease indicates that,
and thethat, in
in any
maximum any electrical
electrical
attenuationboundary
boundary
constantcase,case, the
in thethe central
central
first frequencies
frequencies
stop-band of
allall
thethe
slightlyofincrease
stop-bands
stop-bands decrease
decrease and
andthe
the maximum
maximum attenuation
attenuation constant
constant in
in the
the first
first stop-band
stop-band
with an increase in the rods’ diameter. These phenomena reflect the influence of the Poisson’s effect slightly
slightly increase
increase
with
with
onan an
the increase
increase
band in in
thethe
structures. rods’
rods’ diameter.
Thediameter.
bigger Thesephenomena
theThese
rods’ phenomena
diameter reflect
is, reflect
the thethe
the
larger influence
influence ofofthe
Poisson’s thePoisson’s
Poisson’seffect
effect. effect on
theonband
the band structures.
structures. The The bigger
bigger the rods’
the rods’ diameter
diameter is, the
is, the larger
larger thethe Poisson’s
Poisson’s effect.
effect.
3.2.3. Influence of the Elastic Rod Insert
3.2.3. Influence
3.2.3. InfluenceofofthetheElastic
ElasticRod
RodInsert
Insert
As an elastic rod made of epoxy, whose material and geometrical parameters are given in Table
2, is Asanan
Asinserted elastic
elastic rod
rod
into madecell
themade
unit ofofepoxy,
epoxy,
of whosematerial
the whose
periodic material and
PZT-5Hand geometrical
geometrical
piezoelectric parameters
covered byare
rodparameters are given
given
brass in
inTable
Table 2,
electrodes,
2, is inserted
is inserted into the unit
into the constants
the propagation cell of the
unit cell ofspectra periodic
the periodic PZT-5H piezoelectric
PZT-5H piezoelectric
are computed and comparedrod rod covered
covered
with by brass
by brass
the results electrodes,
of electrodes,
the periodicthe
the propagation
propagation
composite constants
constants
rod without the spectra
spectra epoxy are computed
are computed
component and cell,
andincompared
unit compared
with thewith the
results
as depicted results
of of the
in periodic
the periodic
in Figure 8a–d composite
the four
composite
rodelectrical rod
withoutboundary without
the epoxycases, the
componentepoxy component in unit cell, as depicted in Figure 8a–d in the four
in unit cell, as depicted in Figure 8a–d in the four electrical boundary
respectively.
electrical
cases, boundary cases, respectively.
respectively.
600 600 600 600
600 Without elastic rod 600 600 Without elastic rod 600
With elastic
Without rodrod
elastic With elastic
Without rodrod
elastic
With elastic rod With elastic rod
500 500 500 500
500 500 500 500

400 400 400 400


f (kHz)

f (kHz)
f (kHz)

f (kHz)

400 400 400 400


f (kHz)

f (kHz)
f (kHz)

f (kHz)
Frequency

Frequency
Frequency

Frequency

300 300 300 300


Frequency

Frequency
Frequency

Frequency

300 300 300 300

200 200 200 200


200 200 200 200

100 100 100 100


100 100 100 100

0 0 0 0
0 -1.0 -0.8 -0.6 -0.4
Imaginary wavenumber q I L/π
-0.2 0.0 0.2 0.4 0.6 0.8
Real wavenumber q R L/π
1.0 0 0 -1.0 -0.8 -0.6 -0.4
Imaginary wavenumber q I L/π
-0.2 0.0 0.2 0.4 0.6 0.8
Real wavenumber q R L/π
1.0 0
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
Imaginary wavenumber q I L/π Real wavenumber q R L/π Imaginary wavenumber q I L/π Real wavenumber q R L/π
(a) (b)
(a) (b)

Figure 8. Cont.
Crystals 2016, 6, 45 19 of 28
Crystals 2016, 6, 45 20 of 29

600 600 600 600


Without elastic rod Without elastic rod
With elastic rod With elastic rod

500 500 500 500

400 400 400 400

Frequency f (kHz)

Frequency f (kHz)
Frequency f (kHz)

Frequency f (kHz)
300 300 300 300

200 200 200 200

100 100 100 100

0 0 0 0
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
Imaginary wavenumber q I L/π Real wavenumber q R L/π Imaginary wavenumber q I L/π Real wavenumber q R L/π

(c) (d)
Figure
Figure 8. Influence
8. Influence of the
of the elastic
elastic rod rod insert
insert onpropagation
on the the propagation constants
constants spectraspectra
of theofperiodic
the periodic
PZT-5H
PZT-5H piezoelectric rod covered by brass electrodes with periodic electrical boundaries:
piezoelectric rod covered by brass electrodes with periodic electrical boundaries: (a) electric-open (a)
electric-open condition; (b) applied electric capacity condition with
condition; (b) applied electric capacity condition with C “ αA{l “ 1.02 ˆ 10 C = α A / l
´12= 1.02 × 10 −12
F ; (c)
F; (c) electric-short
8
electric-short condition; and (d) applied feedback control condition
condition; and (d) applied feedback control condition with K g “ 5 ˆ 10 K g 8= 5 ×
V{m. 10 V / m .

Figure 8a–d shows that inserting an elastic rod in unit cell adds many new band structures in all
Figure 8a–d shows that inserting an elastic rod in unit cell adds many new band structures in all
the four electrical boundary cases, which is due to the interaction between the elastic and
the four electrical boundary cases, which is due to the interaction between the elastic and piezoelectric
piezoelectric constituent rods. When an elastic rod is inserted into the unit cell, the differences
constituent rods. When an elastic rod is inserted into the unit cell, the differences between band
between band structures associated with the four electrical conditions are weakened. Since the
structures associated with the four electrical conditions are weakened. Since the inserted elastic rod is
inserted elastic rod is softer than the original piezoelectric rod, the values of the attenuation
softer
constantsthe
than original piezoelectric
enormously enlarge. In the rod, thethree
latter values of the conditions,
electrical attenuationafterconstants
the epoxyenormously enlarge.
rod is inserted
In in
thethelatter three electrical conditions, after the epoxy rod is inserted in the
unit cell, the original first stop-band turns into the third stop-band, with the central frequency unit cell, the original
first stop-band
increased turnsand
slightly into
thethe thirddecreased.
width stop-band, with the central frequency increased slightly and the
width decreased.
3.3. Active Control of Longitudinal Waves in Rod-Type Piezoelectric Phononic Crystals
3.3. Active Control of Longitudinal Waves in Rod-Type Piezoelectric Phononic Crystals
Consider the same rod-type piezoelectric phononic crystal as studied in Section 3.2.3. Its unit
Consider
cell, as referred the same
to in rod-type
Figure 1, piezoelectric
is composed phononic
of an epoxycrystal as studied
elastic in Section
rod, a brass 3.2.3.aIts
electrode, unit cell,
PZT-5H
as piezoelectric
referred to inrod Figure 1, is composed of an epoxy elastic rod, a brass electrode,
and a brass electrode. Their material and geometrical parameters are listed in Tablea PZT-5H piezoelectric
rod2. and
On the a brass electrode.
electrodes, Their
external material
electric and geometrical
capacitor and feedbackparameters
controller areareconnected
listed in Table 2. On the
to investigate
electrodes, external
the influences electric capacitor
of applied and feedback
electric capacity controller
and feedback are connected
control boundaries to on
investigate
the bandthe influences
structures,
of respectively.
applied electric capacity and feedback control boundaries on the band structures, respectively.
First,
First,specifying
specifyingthethe applied electric
applied capacitance
electric capacitanceas Cas“ C1 =ˆ110 ´−13
×10 13
FF, , C C“= 11׈ 1010−12 ´12
F and
F and
−11´11
C C “ = 1 ׈1010 F , F,individually,
individually, wewecompute
compute thethepropagation
propagation constants
constants spectra
spectra of ofthetheperiodic
periodic
Epoxy-Brass-PZT-5H-Brasscomposite
Epoxy-Brass-PZT-5H-Brass composite rod rod to study
study thethe influence
influenceofofapplied
appliedelectric
electric capacity
capacity ononthethe
band
band structures.The
structures. The results
results are
are given
given in Figure
Figure 9a,
9a, where
wherethey theyarearecompared
comparedwith with those
those of of
thethe
electric-open
electric-open (C( C“ =0)
0 )and
andthetheelectric-short
electric-short (C (C“=∞ 8)) boundary
boundaryconditions.
conditions.Second,
Second, when
when thetheapplied
applied
8 8
as K gas“ 1Kˆ g = 1× = 10
5 ×810V{m
feedback
feedback control
control is specified
is specified by control
by the the controlgain gain 10 8 10
V{m, V /Kmg ,“ K5g ˆ V / and
m and Kg “
1 ˆK10 9
= 1 ×
V{m,10 9
V / m ,
respectively, respectively,
the propagation the propagation
constants spectra constants
of the spectra
periodic of the
Epoxy-Brass-PZT-5H-Brass periodic
g

composite rod are calculated tocomposite


Epoxy-Brass-PZT-5H-Brass discuss therod influence of appliedto
are calculated feedback
discusscontrol on the band
the influence of structures.
applied
The results are
feedback depicted
control on theand bandcompared
structures.withThethose of the
results are electric-open
depicted and and the electric-short
compared with those of (K gthe
“ 0)
boundary conditions
electric-open Figure 9b. ( K g = 0 ) boundary conditions in Figure 9b.
and theinelectric-short
Crystals 2016, 6, 45 20 of 28
Crystals 2016, 6, 45 21 of 29

600 600 600 600


Electric-open
Electric-open (C = 0) Electric-short (K g = 0)

500 500 500 500


C = 1 × 10−13

C = 1 × 10−12 K g = 1 × 108
400 400 400 400
C = 1 × 10−11

Frequency f (kHz)

Frequency f (kHz)
K g = 5 × 108
Frequency f (kHz)

Frequency f (kHz)
Electric-short (C = ∞)
K g = 1 × 109

300 300 300 300

200 200 200 200

100 100 100 100

0 0 0 0
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
Imaginary wavenumber q I L/π Real wavenumber q R L/π Imaginary wavenumber q I L/π Real wavenumber q R L/π

(a) (b)
Figure
Figure 9. 9. Influenceofofthe
Influence theelectrical
electricalboundaries
boundaries on
on the
the propagation
propagationconstants
constantsspectra
spectraofof
thethe
periodic
periodic
Epoxy-Brass-PZT-5H-Brass composite rod: (a) applied electric capacity condition; and (b) applied
Epoxy-Brass-PZT-5H-Brass composite rod: (a) applied electric capacity condition; and (b) applied
feedback control condition.
feedback control condition.

Figure 9a indicates that when the applied electric capacity is increased from 0 to infinity, the
Figure
spectra of 9a
theindicates
propagation thatconstants,
when theasapplied
expected, electric
change capacity
graduallyis increased
from thosefromof the0 electric-open
to infinity, the
spectra of the
condition to propagation constants, as
those of the electric-short expected,
condition. Newchange gradually
stop-bands mayfrom those ofand
first appear, thethen
electric-open
most of
condition to those constants
the propagation of the electric-short
spectra altercondition.
towards the New stop-bands
frequency may first
reducing appear,
direction. and the
Thus, thencentral
most of
thefrequencies
propagation constants
of these spectraand
stop-bands alter towards the
pass-bands frequency
decrease reducing
obviously with direction.
an increaseThus,
in thethe central
electric
capacitance.
frequencies However,
of these some and
stop-bands bandpass-bands
structuresdecrease
in certain frequency
obviously withranges, 0 − 40
an increase inkHz and
the electric
270 − 490 kHz
capacitance. in Figure
However, some 9a,banddo not alter with
structures in the applied
certain electricranges,
frequency capacitance, which
0–40 kHz and are270–490
probablykHz
formed due mainly to the mechanical effect of the periodic composite rod. Note
in Figure 9a, do not alter with the applied electric capacitance, which are probably formed due mainly from Figure 9b that
to with the increasing
the mechanical of of
effect control gain from
the periodic 0 , the propagation
composite rod. Note from constants
Figurespectra
9b thatchange
with the from those of of
increasing
the electric-short condition towards the frequency reducing direction. The central
control gain from 0, the propagation constants spectra change from those of the electric-short condition frequencies of
most stop-bands
towards and reducing
the frequency pass-bands decreaseThe
direction. obviously
central with an increase
frequencies in the
of most control gain.
stop-bands andHowever,
pass-bands
the band structures in the same frequency ranges ( 0 − 40 kHz and
decrease obviously with an increase in the control gain. However, the band structures 270 − 490 kHz ) as in Figure 9a, same
in the do
not change at all. This again testifies that these band structures have nothing to do with the electrical
frequency ranges (0–40 kHz and 270–490 kHz) as in Figure 9a, do not change at all. This again testifies
boundaries.
that these band structures have nothing to do with the electrical boundaries.
If the influence of the electrical boundaries on the band structures is overall considered when
If the influence of the electrical boundaries on the band structures is overall considered when
comparing Figure 9a with Figure 9b, it can be concluded that the band structures of the electric-short
comparing Figure 9a with Figure 9b, it can be concluded that the band structures of the electric-short
condition play a referential role. The band structures of the applied electric capacity C (and
condition play a referential role. The band structures of the applied electric capacity C (and electric-
electric-open C = 0 ) condition lie on their lower-frequency side, while those of the applied feedback
open C “
control with0) condition lie on
control gain K gtheir lower-frequency
locate side, while side.
on their higher-frequency thoseConsequently,
of the appliedinfeedback control
the following
with control gain K g locate on their higher-frequency side. Consequently, in the following we focus
we focus on the electric-short condition to discuss the dispersion properties of the characteristic
on the electric-short condition to discuss the dispersion properties of the characteristic longitudinal
longitudinal waves in the rod-type piezoelectric phononic crystals.
waves in the rod-type piezoelectric phononic crystals.
3.4. Dispersion Properties of Longitudinal Waves in Rod-Type Piezoelectric Phononic Crystals
3.4. Dispersion Properties of Longitudinal Waves in Rod-Type Piezoelectric Phononic Crystals
In the case of electric-short boundary condition, comprehensive frequency-related dispersion
In the case of electric-short boundary condition, comprehensive frequency-related dispersion
curves, which include the eigenvalue’s amplitude spectra ( μ (ω ) curves), the wavenumber spectra
curves, which include the eigenvalue’s amplitude spectra (|µ| pωq curves), the wavenumber spectra
( q(ω) curves), the wavelength spectra ( λ (ω ) curves), the phase velocity spectra ( c(ω ) curves), of
(qpωq curves), the wavelength spectra (λpωq curves), the phase velocity spectra (cpωq curves),
the above discussed periodic Epoxy-Brass-PZT-5H-Brass composite rod are calculated in both low
of the above discussed periodic Epoxy-Brass-PZT-5H-Brass composite rod are calculated in both
( f ≤ 600 kHz ) and high ( 1.25 MHz ≤ f ≤ 1.85 MHz ) frequency ranges. Please notice that the
low ( f ď 600 kHz) and high (1.25 MHz ( ď f ď 1.85 MHz) frequency ranges. Please notice that
minimum critical frequency is ω = ω Epoxyp)Epoxy = 2.47
q MHz . Thus, our MTMM is valid till to
the minimum critical frequency is ω0 0 “C ωC “ 2.47 MHz. Thus, our MTMM is valid till
0.9ω0 = 2.22 MHz . The low and high frequency results are provided in Figures 10 and 11,
to 0.9ω0 “ 2.22 MHz. The low and high frequency results are provided in Figures 10 and 11
respectively.For
respectively. Forthe
the convenience
convenience ofof presentation,
presentation,the thedimensionless
dimensionless wavelength
wavelength λ / λ{L
L and phase
and phase
Crystals 2016, 6, 45 21 of 28
Crystals 2016, 6, 45 22 of 29

velocity c / clep are further introduced except for the engineering frequency f = ω / (2π ) and the
velocity c{clep are further introduced except for the engineering frequency f “ ω{p2πq and the
wavenumber qL / π . cclep = cepcep/ ρ
a
dimensionless wavenumber
dimensionless qL{π. lep “ {ρepep is the
the longitudinal
longitudinalwavewavespeed
speedinin
the epoxy
the rodrod
epoxy
without
without considering
considering thethe Poisson’s
Poisson’seffect, whereccepep= “
effect,where EepEepand
andρρepepare
arethe
theYoung’s
Young’smodulus
modulus and
and thethe
material density of epoxy, respectively.
material density of epoxy, respectively.

Eigenvalue's amplitude μ Frequency as qR L = 0 Frequency as qR L = π qL = qR L Frequency as qR L = 0 Frequency as qR L = π


1.0 100.0

c / clep
0.8
μ

Dimensionless phase velocity


Eigenvalue's amplitude

10.0
0.6

0.4
1.0
c0
0.2

0.0 0.1
0 100 200 300 400 500 600 0 100 200 300 400 500 600
Frequency f (kHz) Frequency f (kHz)

(a) (d)
qL = qR L qL = iqI L Frequency as qR L = 0 Frequency as qR L = π qL = qR L Frequency as qR L = 0 Frequency as qR L = π
1.0 5.0
qR L / π
0.8
Real wavenumber

c / clep
0.6 4.0
Dimensionless phase velocity

0.4

0.2 3.0

0.0
qI L / π
-0.2 2.0
Imaginary wavenumber

-0.4

-0.6 1.0
c0
-0.8

-1.0 0.0
0 100 200 300 400 500 600 0 100 200 300 400 500 600
Frequency f (kHz) Frequency f (kHz)

(b) (e)
qL = qR L Frequency as qR L = 0 Frequency as qR L = π
100.0

λ/L
Dimensionless wavelength

10.0

1.0

0.1
0 100 200 300 400 500 600
Frequency f (kHz)

(c)
Figure 10. Various frequency-related dispersion curves of the periodic Epoxy-Brass-PZT-5H-Brass
Figure 10. Various frequency-related dispersion curves of the periodic Epoxy-Brass-PZT-5H-Brass
composite rod in low frequency range: (a) eigenvalue’s amplitude spectra; (b) wavenumber spectra;
composite rod in low frequency range: (a) eigenvalue’s amplitude spectra; (b) wavenumber spectra;
(c) wavelength spectra in logarithmic coordinate; (d) phase velocity spectra in logarithmic
(c) wavelength spectra in logarithmic coordinate; (d) phase velocity spectra in logarithmic coordinate;
coordinate; and (e) phase velocity spectra in a small scale.
and (e) phase velocity spectra in a small scale.
Crystals 2016, 6, 45 22 of 28
Crystals 2016, 6, 45 23 of 29

Eigenvalue's amplitude μ Frequency as qR L = 0 Frequency as qR L = π qL = qR L Frequency as qR L = 0 Frequency as qR L = π


1.0 100.0

c / clep
μ 0.8

Dimensionless phase velocity


Eigenvalue's amplitude

0.6

10.0

0.4

0.2

0.0 1.0
1.25 1.35 1.45 1.55 1.65 1.75 1.85 1.25 1.35 1.45 1.55 1.65 1.75 1.85
Frequency f (MHz) Frequency f (MHz)

(a) (d)
qL = qR L qL = iqI L Frequency as qR L = 0 Frequency as qR L = π qL = qR L Frequency as qR L = 0 Frequency as qR L = π
1.0 10.0
qR L / π
0.8 9.0
Real wavenumber

c / clep
0.6
8.0

Dimensionless phase velocity


0.4
7.0
0.2
6.0
0.0
qI L / π 5.0
-0.2
Imaginary wavenumber

4.0
-0.4
3.0
-0.6

-0.8 2.0

-1.0 1.0
1.25 1.35 1.45 1.55 1.65 1.75 1.85 1.25 1.35 1.45 1.55 1.65 1.75 1.85
Frequency f (MHz) Frequency f (MHz)

(b) (e)
qL = qR L Frequency as qR L = 0 Frequency as qR L = π
100.0

λ/L
Dimensionless wavelength

10.0

1.0

0.1
1.25 1.35 1.45 1.55 1.65 1.75 1.85
Frequency f (MHz)

(c)
Figure 11.
Figure 11. Various
Various frequency-related
frequency-relateddispersion
dispersioncurves
curvesofofthe
theperiodic
periodicEpoxy-Brass-PZT-5H-Brass
Epoxy-Brass-PZT-5H-Brass
composite rod in high frequency range: (a) eigenvalue’s amplitude spectra; (b) wavenumber spectra;
composite rod in high frequency range: (a) eigenvalue’s amplitude spectra; (b) wavenumber spectra;
(c) wavelength spectra in logarithmic coordinate; (d) phase velocity spectra in logarithmic
(c) wavelength spectra in logarithmic coordinate; (d) phase velocity spectra in logarithmic coordinate;
coordinate; and (e) phase velocity spectra in a small scale.
and (e) phase velocity spectra in a small scale.

Figures 10a–e and 11a–e reflect that, in addition to the same dispersion properties of
Figures 10a–e
longitudinal wavesand
in11a–e reflect
general that, inrods
periodic addition to the
[5,37], the same dispersion
characteristic properties of
longitudinal longitudinal
waves in a
waves in general periodic rods [5,37], the characteristic longitudinal waves in
rod-type piezoelectric phononic crystal also possess the following dispersion properties:a rod-type piezoelectric
phononic crystal also possess the following dispersion properties:
(1) The eigenvalue’s amplitude spectra, which cannot be obtained by MRRM [5], demonstrate
clearly the width, the central frequencies and the bounding frequencies of the pass-bands
(1) The eigenvalue’s amplitude spectra, which cannot be obtained by MRRM [5], demonstrate clearly
( μ = 1 ) and the stop-bands ( μ ≠ 1 ). They also reflect the attenuation amplitudes e qI L of waves
the width, the central frequencies and the bounding frequencies of the pass-bands (|µ| “ 1)
in the stop-bands, which are verified by the attenuation constant ( qI L ) qspectra. The
and the stop-bands (|µ| ‰ 1). They also reflect the attenuation amplitudes e I L of waves in
eigenvalue’s amplitude spectra cannot indicate the properties of (q
waves in the pass-bands, but
the stop-bands, which are verified by the attenuation constant I L) spectra. The eigenvalue’s
the phase constant spectra do.
amplitude spectra cannot indicate the properties of waves in the pass-bands, but the phase
(2) constant
In these frequency-related
spectra do. dispersion curves, the bounding frequencies of the odd and even
order stop-bands correspond to qL = qR L = (2n + 1)π and qL = qR L = 2nπ ( n is a natural
Crystals 2016, 6, 45 23 of 28

(2) In these frequency-related dispersion curves, the bounding frequencies of the odd and even
order stop-bands correspond to qL “ q R L “ p2n ` 1qπ and qL “ q R L “ 2nπ (n is a natural
number), respectively. Within the stop-bands, the real part (q R L) of the complex wavenumber
qL, which cannot be computed by the MRRM [5] but obtained here by the MTMM, have the
same phases as their boundaries. In the wavelength spectra, the representations corresponding
to these two kinds of phases are horizontal lines λ{L “ 2{p2n ` 1q and λ{L “ 1{n. In the phase
velocity spectra, they correspond to inclined lines that pass through the origin and have slopes
c{ω “ L{rp2n ` 1qπs c{ω “ L{r2nπs, respectively. The lines determined by the above formulas
and the lines of bounding frequencies form the grids to draw the spectra in the corresponding
frequency-related dispersion curves.

4. Conclusions
A modified transfer matrix method (MTMM), which is based on the Love rod theory and the
method of reverberation-ray matrix (MRRM), is proposed in this paper for the analysis of characteristic
longitudinal waves by combining Floquet–Bloch principle and for the calculation of transmission
spectra in general rod-type piezoelectric phononic crystals with electric-open, applied electric capacity,
electric-short and applied feedback control conditions. Numerical examples are given to validate the
proposed MTMM, to study the passive control of the band structures by the electrode’s thickness, the
rod’s cross-sectional dimension and the elastic rod insert, and to investigate the active tuning of the
band structures by the electrical boundary conditions. The dispersion properties are summarized from
the comprehensive frequency spectra. From these analyses, we can draw the conclusions as follows:

(1) The proposed analytical MTMM provides an alternative analysis method for the complex band
structures and transmission spectra till to 0.9ω0 (ω0 is the minimum critical frequency) within
which the Love rod theory is valid. Its effectiveness is validated by some numerical examples.
(2) In passive mode, the electrode’s thickness and the rod’s cross-sectional dimension can be used
to slightly adjust the band structures of the rod-type piezoelectric phononic crystals, while the
elastic rod insert is able to enormously alter the band structures.
(3) In active mode, the switchable electrical boundaries among electric-short, applied electric capacity,
electric-open and applied feedback control conditions is effective for modulating some of the
band structures that are related to the electromechanical coupling of the rod-type piezoelectric
phononic crystals. The tunable capacity and control gain in the applied electric capacity and
applied feedback control cases, respectively, can also be used for tuning the propagation of
longitudinal waves. The band structures of the electric-short condition play a referential role for
designing the active control scheme.

Acknowledgments: The authors wish to thank the anonymous reviewers for their constructive comments that
greatly improve our manuscript. The work was supported by the National Natural Science Foundation of China
(No. 11372119).
Author Contributions: Y.Q.G. conceived and designed the main ideas; L.F.L. wrote the computer code; L.F.L. and
Y.Q.G. analyzed the data and wrote the paper together.
Conflicts of Interest: The authors declare no conflict of interest.

Abbreviations
The following abbreviations are used in this manuscript:
MTMM Modified Transfer Matrix Method
MRRM Method of Reverberation-Ray Matrix
TMM Transfer Matrix Method
FEM Finite Element Method
AM Analytical Method
Crystals 2016, 6, 45 24 of 28

Appendix A
Based on the basic idea of Love rod theory [27,28], [29] (pp. 139–142), [30], which considers the
Poisson’s effect, the displacement field u “ ru, v, wsT and the electric potential ϕ in a piezoelectric rod
are assumed as
Bu T
„ 
Bu
u “ upx, tq, ´ν12 y, ´ν13 z , ϕ “ ϕpx, tq (A1)
Bx Bx
where u, v and w are the displacements of particle at (x, y, z) in the rod along x, y and z axes, respectively;
ν12 and ν13 are the Poisson’s ratios with respect to y and z axes, respectively; and r¨sT denotes the
transposition of a matrix (or vector). The strain vector ε “ rε x , ε y , ε z , γyz , γzx , γxy sT and the electric
field vector E “ rEx , Ey , Ez sT can then be derived from the generalized strain-displacement relations
» B B B fiT $ ,
Bx 0 0 0 Bz By & u / T
B2 u B2 u
’ . „ Bu Bu Bu
ε “ LT u “ – 0 B 0 B 0 B
v , , , 0, z, y
— ffi
By Bz Bx “ ´ν12 ´ν13 ´ν13 ´ν12
Bx2 Bx2
fl
B B B

% w / Bx Bx Bx
0 0 Bz By Bx 0
- (A2)
„ T

E “ ´∇ ϕ “ ´ , 0, 0
Bx

where the nonzero shear strains γzx “ ´zν13 B 2 u{Bx2 and γxy “ ´yν12 B 2 u{Bx2 are obviously one order
higher than the three normal strains ε x , ε y and ε z . Consequently, we assume that γzx “ 0 and γxy “ 0,
and thus all the three shear strains are zero considering γyz “ 0 by itself. In order to derive the
governing equations with the Hamilton principle, the normal stresses σx , σy , σz corresponding to the
nonzero strains and the electric displacement along the x axis Dx associated with the nonzero electric
field should be obtained from the constitutive equations of the linear piezoelectric materials
# + « ff # +
σ C ´eT ε
“ (A3)
D e ff E

where σ “ rσx , σy , σz , τyz , τxz , τxy sT and D “ rDx , Dy , Dz sT are the stress vector and electric
displacement vector, respectively; C “ rc pq s is the 6 ˆ 6 symmetric and positive definite elastic
constant matrix, e “ rerq s is the 3 ˆ 6 piezoelectric constant matrix, and α “ rαrs s is the 3 ˆ 3 symmetric
dielectric constant matrix (p, q “ 1, 2, 3, 4, 5, 6, r, s “ 1, 2, 3). Substitution of Equation (A2) into
Equation (A3) gives
$ , » fi

’ σx /
/ pc11 ´ ν12 c12 ´ ν13 c13 q e11 # +
’ / Bu
& σ
y
. — pc21 ´ ν12 c22 ´ ν13 c23 q e12 ffi
Bx (A4)
— ffi
“— ffi Bϕ

’ σz /
/ – pc31 ´ ν12 c32 ´ ν13 c33 q e13 fl Bx
’ /
Dx pe11 ´ ν12 e12 ´ ν13 e13 q
% -
´α11

Consider Hamilton’s principle for linear piezoelectric continuum [34]


„ż t 
2
δ pT ´ H ` Wqdt “ 0, (A5)
t1

where T, H and W are the kinetic energy, the electric enthalpy and the external work in the system.
The kinetic energy T in the whole piezoelectric rod is obtained from Equation (A1) as
.
1y . T . 1y . 2 . 2 . 2
żl« ˆ ˙2 ff
1 .2 2 2 Bu
T“ ρu udΩ “ ρpu ` v ` w qdΩ “ ρAu ` ρpν12 Iz ` ν13 Iy q dx (A6)
2 2 2 0 Bx
Ω Ω

where Ω signifies the space region of the rod; ρ is the material density; l and A are the length and
the cross-sectional area of the rod, respectively; Iy and Iz are the cross-sectional moment of inertia
Crystals 2016, 6, 45 25 of 28

with respect to y and z axes, respectively; and the over dot signifies the derivative of quantities about
time here and after. The electric enthalpy H [33,34] in the whole piezoelectric rod is written from
Equations (A2) and (A4) as

1
t t
H “ 2 pσT ε ´ DT EqdΩ “ 12 pσx ε x ` σy ε y ` σz ε z ´ Ex Dx qdΩ
Ω Ω (A7)
şl ´ ¯ 2 şl Bϕ şl ´ ¯2

“ 12 0 cA BBux dx ` 0 eA BBux B x dx ´ 12 0 αA B x dx

where c “ c11 ` ν122 c ` ν2 c ` 2pν ν c ´ ν c ´ ν c q, e “ e ´ ν e ´ ν e and α “ α are


22 13 33 12 13 23 12 12 13 13 11 12 12 13 13 11
the equivalent axial stiffness, piezoelectric and dielectric constants of the piezoelectric rod, respectively.
The equivalent axial stiffness constant c in the current anisotropic piezoelectric rod, like the Young’s
modulus E in the isotropic elastic rod, comprehensively reflects the contributions of anisotropic
stiffness coefficients to the axial stiffness. cA denotes the axial rigidity of the piezoelectric rod.
It can be deduced that the bigger c is, the faster the longitudinal wave in the rod. The equivalent
axial piezoelectric constant e and dielectric constant α comprehensively represent the contributions
of anisotropic piezoelectric and dielectric coefficients to the axial piezoelectricity and dielectricity,
respectively. e measures the axial stress generating from unit axial electric field or the axial electric
displacement resulting from unit axial strain, i.e., the transformation ability between electrical energy
due to axial electric field and mechanical energy due to axial vibration. α denotes the ability to store
electrical energy of the piezoelectric rod due to axial electric field. As will be seen at the end of
this section, e and α have direct and inverse correlations to the longitudinal wave speed in the rod,
respectively. The external work [34] imported to the piezoelectric rod consists of mechanical and
electrical work in form of x
W“ pfT u ´ qϕqdS “ pNu ´ Qϕq|0l (A8)
S

where S signifies the boundary surfaces with specified surface forces f “ r f x , f y , f z sT of the system; q is
the electric charge density per area; N and Q are the specified axial force and electric charge on the
ends of the piezoelectric rod, respectively. Substituting Equations (A6)–(A8) into Equation (A5) and
then applying the integration by parts to some terms so that the integral terms have common variation
δu or δϕ, one obtains

rl ˘ B u. ıˇˇl t2
" ” *
. ` 2 I 2 I
˘ B2 u. ı ” `
2 2
0 ρAu ´ ρ ν12 z ` ν13 y B x2
δudx ` ρ ν12 Iz ` ν13 Iy B x δu ˇ
0 t
1
r t2 !ş l ” 2 B2 ϕ ˘ B2 u.. ..
ı )
cA BB xu2 ` eA B x2 ` ρ ν12
2 I ` ν2 I
`
` t1 0 z 13 y B x2 ´ ρA u δudx dt

r t !r l ” 2 B2 ϕ
ı )
` t 2 0 eA BB xu2 ´ αA B x2 δϕdx dt (A9)
1

r t2 !” Bϕ ˘ B u.. ı )ˇˇl
N ´ cA BBux ´ eA B x ´ ρ ν12
` 2 2 I
` t1 Iz ` ν13 y B x δu ˇ dt
0
r t !” Bϕ
ı )ˇl
` t 2 ´Q ´ eA BBux ` αA B x δϕ ˇ dt “ 0
ˇ
1 0

Notice that in Hamilton’s principle, the generalized displacement functions at time t1 and time
t2 are given [29] (pp. 126–131), [34], i.e., the variations of these functions vanish at times t1 and t2 .
ˇ
Hence, on the basis of δu ˇt“t1 “ 0 and δu |t“t2 “ 0, the first term in Equation (A9) should vanish. The
second and third (fourth and fifth) terms in Equation (A9) represent the variations of energy functional
inside the rod (on the rod ends) due to the variations δu and δϕ, respectively. It is because that δu and
δϕ are independent each other both inside the rod and on the rod ends, the second to fifth terms in
Equation (A9) should vanish. Since the variations δu and δϕ are arbitrary as x P p0, lq, the vanishing of
Crystals 2016, 6, 45 26 of 28

the second and third terms of Equation (A9) leads to the governing equations (Euler equations) of the
piezoelectric Love rod
..
B2 u B2 ϕ ´
2 2
¯ B2 u .. B2 u B2 ϕ
cA ` eA ` ρ ν12 Iz ` ν13 Iy “ ρA u, eA ´ αA “0 (A10)
Bx2 Bx2 Bx2 Bx2 Bx2
On the rod end, x “ 0 (x “ l), if δu and δϕ are arbitrary, the vanishing of the fourth and fifth
terms of Equation (A9) gives rise to the constitutive relations (natural boundary conditions) of the
piezoelectric Love rod
¯ Bu ..
Bu Bϕ ´
2 2 Bu Bϕ
N “ cA ` eA ` ρ ν12 Iz ` ν13 Iy , Q “ ´eA ` αA (A11)
Bx Bx Bx Bx Bx
Substituting the latter formula into the former one in Equation (A10), one expresses the governing
equations of the piezoelectric Love rod as
..
´
2
¯B2 u ´
2 2
¯ B2 u ..
c ` e {α A 2 ` ρ ν12 Iz ` ν13 Iy “ ρAu, (A12)
Bx Bx2
where only the axial displacement u is involved.

Appendix B
The governing equations and the constitutive relations of an elastic rod of anisotropic material
based on the Love rod theory [27,28], [29] (pp. 139–142), [30] can be derived in a similar way as that
discussed in Appendix A. However, all the electric quantities should not appear. In particular, the
constitutive equations of the linear elastic materials

σ “ Cε (A13)

should be utilized to give the normal stresses σx , σy , σz


$ , » fi
& σx /
’ pc11 ´ ν12 c12 ´ ν13 c13 q " *
ffi Bu
.
“ – pc21 ´ ν12 c22 ´ ν13 c23 q fl (A14)

σy
’ / Bx
% σ -
z pc31 ´ ν12 c32 ´ ν13 c33 q

corresponding to the nonzero strains. Besides, Hamilton’s principle for linear elastic body [29]
(pp. 126–131) is identical to Equation (A5) except that the electric enthalpy H should be replaced by
the strain energy U

1y T 1y
żl ˆ ˙2
1 Bu
U“ σ εdΩ “ pσx ε x ` σy ε y ` σz ε z qdΩ “ cA dx (A15)
2 2 2 0 Bx
Ω Ω

and the external work should be x


W“ fT udS “ pNuq|0l (A16)
S

Substituting the kinetic energy T as given in Equation (A6), the strain energy U in Equation (A15)
and the external work in Equation (A16) into Hamilton’s principle for linear elastic body [29]
(pp. 126–131), and conducting the variation and integration by parts, one can write an equation
similar to Equation (A9) but without the electric terms. In the same way, the governing equations and
the constitutive relations of the anisotropic elastic Love rod are obtained as
..
B2 u ´
2 2
¯ B2 u ..
cA 2
` ρ ν12 Iz ` ν13 Iy 2
“ ρAu (A17)
Bx Bx
Crystals 2016, 6, 45 27 of 28

and ..
Bu ´ ¯ Bu
2 2
N “ cA ` ρ ν12 Iz ` ν13 Iy (A18)
Bx Bx
respectively. It should be pointed out that Equations (A17) and (A18) can also be easily degenerated
from Equations (A12) and (A11), respectively, by eliminating the electrically relative terms. If the
elastic rod is made of isotropic material, then the equivalent stiffness c and the Poisson’s ratios ν12 , ν13
will be automatically degenerated to the Young’s modulus E and the Poisson’s ratio ν, respectively.
The governing equations (A17) and the constitutive relations (A18) will accordingly be degenerated
to those of the isotropic elastic Love rod, as given by Love [27], Graff [28], Doyle [29] (p. 140) and
Ravindra [30].

References
1. Kushwaha, M.S.; Halevi, P.; Dobrzynski, L.; Djafari-Rouhani, B. Acoustic band structure of periodic elastic
composites. Phys. Rev. Lett. 1993, 71, 2022–2025. [CrossRef] [PubMed]
2. Deymier, P.A., Ed.; Acoustic Metamaterials and Phononic Crystals; Springer: Berlin, Germany, 2013.
3. Asiri, S.; Baz, A.; Pines, D. Periodic struts for gearbox support system. J. Vib. Control 2005, 11, 709–721.
[CrossRef]
4. Asiri, S. Vibration isolation of automotive vehicle engine using periodic mounting systems. Proc. SPIE 2005,
5760, 526–537.
5. Guo, Y.Q.; Fang, D.N. Analysis and interpretation of longitudinal waves in periodic multiphase rods using
the method of reverberation-ray matrix combined with the Floquet-Bloch theorem. ASME J. Vib. Acoust.
2014, 136, 011006. [CrossRef]
6. Ruzzene, M.; Baz, A. Attenuation and localization of wave propagation in periodic rods using shape memory
inserts. Smart Mater. Struct. 2000, 9, 805–816. [CrossRef]
7. Ruzzene, M.; Baz, A. Control of wave propagation in periodic composite rods using shape memory inserts.
ASME J. Vib. Acoust. 2000, 122, 151–159. [CrossRef]
8. Ding, R.; Su, X.L.; Zhang, J.J.; Gao, Y.W. Tunability of longitudinal wave band gaps in one dimensional
phononic crystal with magnetostrictive material. J. Appl. Phys. 2014, 115, 074104. [CrossRef]
9. Baz, A. Active control of periodic structures. ASME J. Vib. Acoust. 2001, 123, 472–479. [CrossRef]
10. Thorp, O.; Ruzzene, M.; Baz, A. Attenuation and localization of wave propagation in rods with periodic
shunted piezoelectric patches. Smart Mater. Struct. 2001, 10, 979–989. [CrossRef]
11. Chen, S.B.; Wen, J.H.; Wang, G.; Yu, D.L.; Wen, X.S. Improved modeling of rods with periodic arrays of
shunted piezoelectric patches. J. Intell. Mater. Syst. Struct. 2012, 23, 1613–1621.
12. Lossouarn, B.; Aucejo, M.; Deu, J.-F. Multimodal coupling of periodic lattices and application to rod vibration
damping with a piezoelectric network. Smart Mater. Struct. 2015, 24, 045018. [CrossRef]
13. Lossouarn, B.; Aucejo, M.; Deu, J.-F. Multimodal vibration damping through a periodic array of piezoelectric
patches connected to a passive network. Proc. SPIE 2015, 9431, 94311A.
14. Singh, A.; Pines, D.J.; Baz, A. Active/passive reduction of vibration of periodic one-dimensional structures
using piezoelectric actuators. Smart Mater. Struct. 2004, 13, 698–711. [CrossRef]
15. Asiri, S.; Baz, A.; Pines, D. Active periodic struts for a gearbox support system. Smart Mater. Struct. 2006, 15,
1707–1714. [CrossRef]
16. Asiri, S. Tunable mechanical filter for longitudinal vibrations. Shock Vib. 2007, 14, 377–391. [CrossRef]
17. Asiri, S. Broadband vibration attenuation using hybrid periodic rods. J. Eng. Res. 2008, 5, 7–19.
18. Li, F.M.; Wang, Y.S.; Chen, A.L. Wave localization in randomly disordered periodic piezoelectric rods.
Acta Mech. Solida Sin. 2006, 19, 50–57.
19. Wang, Y.Z.; Li, F.M.; Kishimoto, K.; Wang, Y.S.; Huang, W.H. Wave localization in randomly disordered
periodic piezoelectric rods with initial stress. Acta Mech. Solida Sin. 2008, 21, 529–535. [CrossRef]
20. Degraeve, S.; Granger, C.; Dubus, B.; Vasseur, J.O.; Thi, M.P.; Hladky-Hennion, A.-C. Bragg band gaps
tunability in an homogeneous piezoelectric rod with periodic electrical boundary conditions. J. Appl. Phys.
2014, 115, 194508. [CrossRef]
Crystals 2016, 6, 45 28 of 28

21. Kutsenko, A.A.; Shuvalov, A.L.; Poncelet, O.; Darinskii, A.N. Quasistatic stopband and other unusual features
of the spectrum of a one-dimensional piezoelectric phononic crystal controlled by negative capacitance.
C. R. Mec. 2015, 343, 680–688. [CrossRef]
22. Ponge, M.-F.; Dubus, B.; Granger, C.; Vasseur, J.O.; Thi, M.P.; Hladky-Hennion, A.-C. Theoretical and
experimental analyses of tunable Fabry-Perot resonators using piezoelectric phononic crystals. IEEE Trans.
Ultrason. Ferroelectr. Freq. Control 2015, 62, 1114–1121. [CrossRef] [PubMed]
23. Degraeve, S.; Granger, C.; Dubus, B.; Vasseur, J.O.; Thi, M.P.; Hladky, A.C. Tunability of Bragg band gaps in
one-dimensional piezoelectric phononic crystals using external capacitances. Smart Mater. Struct. 2015, 24,
085013. [CrossRef]
24. Degraeve, S.; Granger, C.; Dubus, B.; Vasseur, J.O.; Thi, M.P.; Hladky-Hennion, A.-C. Tunability of a
one-dimensional elastic/piezoelectric phononic crystal using external capacitances. Acta Acust. United Acust.
2015, 101, 494–501. [CrossRef]
25. Kutsenko, A.A.; Shuvalov, A.L.; Poncelet, O.; Darinskii, A.N. Tunable effective constants of the
one-dimensional piezoelectric phononic crystal with internal connected electrodes. J. Acoust. Soc. Am.
2015, 137, 606–616. [CrossRef] [PubMed]
26. Vasilenko, A.V.; Rogacheva, N.N. The effective characteristics and electroelastic state of active composite
elements. PMM-J. Appl. Math. Mech. 2011, 75, 463–471. [CrossRef]
27. Love, A.E.H. A Treatise on the Mathematical Theory of Elasticity, 4th ed.; Cambridge University Press:
Cambridge, UK, 1927; p. 428.
28. Graff, K.F. Wave Motion in Elastic Solids; Ohio State University Press: Columbus, OH, USA, 1975; pp. 108–125.
29. Doyle, J.F. Wave Propagation in Structures, 2nd ed.; Springer: New York, NY, USA, 1997.
30. Ravindra, B. Love-theoretical analysis of periodic system of rods. J. Acoust. Soc. Am. 1999, 106, 1183–1186.
[CrossRef] [PubMed]
31. Brillouin, L. Wave Propagation in Periodic Structures: Electric Filters and Crystal Lattices, 2nd ed.; Dover
Publications: New York, NY, USA, 1953.
32. Mead, D.J. Wave propagation in continuous periodic structures: Research contributions from Southampton,
1964–1995. J. Sound Vib. 1996, 190, 495–524. [CrossRef]
33. ANSI/IEEE Std 176-1987. IEEE Standard on Piezoelectricity; IEEE: New York, NY, USA, 1988; pp. 6–8.
34. Tiersten, H.F. Linear Piezoelectric Plate Vibrations; Plenum Press: New York, NY, USA, 1969; pp. 41–50.
35. Auld, B.A. Acoustic Fields and Waves in Solids; John Wiley & Sons: New York, NY, USA, 1973; Volume I.
36. Bronshtein, I.N.; Semendyayev, K.A.; Musiol, G.; Muehlig, H. Handbook of Mathematics, 5th ed.; Springer:
Berlin, Germany, 2007.
37. Mead, D.J. Wave propagation and natural modes in periodic systems: I. Mono-coupled systems. J. Sound Vib.
1975, 40, 1–18. [CrossRef]

© 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC-BY) license (http://creativecommons.org/licenses/by/4.0/).

You might also like