Professional Documents
Culture Documents
I. I NTRODUCTION
• For non-piezoelectric materials: the two equations gov-
again, if we take partial derivative of (2a) with respect to combining this results we obtain
x and combine it with (2c), we get h i
F + hQ = sZc −Ae−sx/v + Besx/v
∂Γ ∂2ξ ∂D ∂2ξ Y ∂2ξ h ∂D
=Y 2
−h =⇒ 2 = −
∂x ∂x ∂x ∂t ρ ∂x2 ρ ∂x
If we are considering a thickness mode, this is, we assume We can also find the voltage across the piezoelectric trans-
the transducer to vibrate predominantly in one dimension ducer by integrating (2b) from x = 0 to x = L (all the
(thickness dimension), we can perform the following thickness of the transducer):
derivation:
∂D ∂D Z L Z L
ξ D
= =0 V = Edx = −h + dx =
∂y ∂z ∂x ε
0 0
and using Gauss’ law together with the fact that there Q
is no free charge inside the transducer (all the possible = −h [ξ(L) − ξ(0)] +
C0
charge would be located at the electrodes) yields to
F where C0 = Sε/L is the clamped or static capacitance of
∂D
∇ · D = ρinside = 0 =⇒ =0 the transducer. We can rewrite our expression as follows
∂x
hence in our previous equation we can get rid of the last sC0 Ze
V = −h [ξ(L) − ξ(0)] U , where U =
term: 1 + sC0 Ze
∂2ξ Y ∂2ξ Y
2
= , where v 2 = (3) being Ze the impedance of an arbitrary electrical load
∂t ρ ∂x2 ρ
connected across the transducer electrodes.
which is an identical wave equation to the one obtained
for non-piezoelectric materials. Note that if we had multiple piezoelectric layers, the overall
To solve the wave equation it is useful to use the Laplace voltage across the whole transducer would be the sum of the
transform, together with the assumption that all functions are correspondent voltages for each of the layers
zero at t = 0 Z b1 Z bn
Z ∞
L[f (t)] = f (t) = e−st f (t)dt V = E1 dx + . . . + En dx
0 a1 an
F s h
T HE boundary conditions that apply for the wave equation
i
Γ= = Y −Ae−sx/v + Besx/v =⇒ solution are: continuity of particle displacements, ξ and
S v
h i continuity of forces across the interface, F , at the front and
=⇒ F = Zs −Ae−sx/v + Besx/v (4) rear face, resulting in four boundary equations:
where Z = ρvS is the characteristic acoustic impedance
of the material (note that in some texts the cross sectional
ξ1|x=0 = ξc|x=0
ξ
area S is omitted). c|x=L = ξ2|x=L
• For a piezoelectric material, using (2a)
F1|x=0 = Fc|x=0
sh
Fc|x=L = F2|x=L
i
Γ=Y −Ae−sx/v + Besx/v − hD
v
and applying again Gauss’ law to the surface of the where each of the expressions has the following form:
piezoelectric material (in our case, where the electrodes
−sx/v1
are located, this is front and rear faces) ξ1 (s) = A1 e
+ B1 esx/v1
Z Z
Q ξc (s) = Ae−sx/vc + Besx/vc
D · ds = DS = ρcharge = Q =⇒ =D
ξ2 (s) = A2 e−sx/v2
S V S
M1 MINIPROJECT, ERASMUS MUNDUS MASTER IN COMPLEX SYSTEMS SCIENCE, JUNE 2014 3
Force input at Z2
the front face
h/2sZE FB (s) = F|x=L = −hQ(s)KB (16)
Zc + Z2
F1
+
+
TF 1/sZc
The final step will be, as we did for the receiver mode,
to find a transfer function relating the force generated at the
KF Voltage
front face to the voltage across the transducer. Notice that we
output! have obtained forces in both rear and front faces, but we are
+ V
+
h U −1 primarily interested in the front face.
TB 1/sZc
Z0
h/2sZE
IT
Loading medium Piezoelectric Backing layer
We are now particularly interested in the front and back IV. L ATTICE MODEL
faces of the transducer. Thus, set x = 0 and x = L ONSIDER the setup outlined in figure 7, where medium
respectively:
Z1
C 2, with a thickness l2 , is located between media
1 and 3, which are semi-infinite. We decompose forces
FF (s) = F|x=0 = −hQ(s)KF (15) into their backward Bn and frontward Fn components. The
Zc + Z1
M1 MINIPROJECT, ERASMUS MUNDUS MASTER IN COMPLEX SYSTEMS SCIENCE, JUNE 2014 5
the the first and the second coming from (20a) and the last
h/sZc TF /2
Front face
two ones from (20b). This can be written in matrix form:
force output!
FF F20 1 + R12 −R12 F10
KF AF /2 −1 =
Voltage input B10 R12 1 − R12 B20
e +
a + Y h
+
F30 1 + R23 −R13 F2l2
=
B2l2 R13 1 − R13 B30
KB AB /2 −1
FB
Back face
where
force output! Z2 − Z1 Z3 − Z2
h/sZc TB /2 R12 = & R23 =
Z2 + Z1 Z3 + Z2
Fig. 6: Transmitter transfer function are the reflection coefficients for waves of force travelling into
the piezo at the 1-2 and 2-3 interfaces respectively.
Medium 1,Z1 Medium 2, Z2 Medium 3, Z3
F1 F2 F3 A. Transducer model
(Af 3 ) (Af 1 ) (Af 2 )
(Bb3 ) (Bb1 ) (Bb2 )
Now consider a similar disposition, placing the piezoelectric
layer between two passive layers.
B1 B2 B3
x1 x2 x3
∞ ∞
Back face medium, Z1 Transducer, Zt Front face medium, Z−1
Fig. 7: Basic setup for a number of passive layers. Displacements are F−1 Ft F1
bracketed. (Af −1 ) (Af t ) (Af 1 )
(Bb−1 ) (Bbt ) (Bb1 )
same for the displacements, backward Bbn and frontward Aan . B−1 Bt B1
x−1 xt x1
∞ ∞
Vt
Again, using complex Laplace transforms and the same It
analysis as for the systems model, we obtain: ZE ∼
( VE
ξ n = Af n e−sxn /vn + B bn esxn /vn (19a)
Fig. 8: Piezolectric system. Displacements are bracketed.
Γn = sZn [−Af n e−sxn /vn + B bn esxn /vn ] (19b)
Define now the force components at position xn to be The voltage across the transducer will be
F n = −sZn Af n & B n = sZn B bn h i Q
t
Vt = −h (ξt )|xt =lt − (ξt )|xt =0 + (21)
so we can rewrite (19a) and (19b) as Co
ξ n = [F n e−sxn /vn + B n esxn /vn ]/sZn ] (20a) where Qt is the total electrical charge in the electrodes.
Γn = F n e−sxn /vn + B n esxn /vn (20b)
The mechanical boundary conditions are
We shall drop the over lines to make notation clear, recall (
that for this model everything is done in complex Laplace ξ−1|x−1 =0 = ξt|xt =0 & ξt|xt =lt = ξ1|x1 =0
domain. Consider now the boundaries xn = 0 and xn = ln of Γ−1|x−1 =0 = Γt|xt =0 & Γt|xt =lt = Γ1|x1 =0
this n-th layer:
n−sx /vn
Fn0 = Fn e|xn =0
sx /v
& Bn0 = Bn e|xnn=0n that is
−sx /v
n n sx /v
Fnln = Fn e|xn =l & Bnln = Bn e|xnn=lnn
n
(−F−10 + B−10 ) /Z−1 = (−Ft0 + Bt0 ) /Zt
hence we can write for the model in the figure
(−F + B ) /Z = (−F + B ) /Z
tlt −tt t 10 10 1
−sT
F2l2 e 2 0 F20 F−10 + B−10 = Ft0 + Bt0 − hQt
=
B2l2 0 esT2 B20
Ftlt + Btlt − hQt = F10 + B10 − hQt
Co sZE 1
− h(1 − e−sTt )(Ft0 + Btlt ) (26) Fig. 9: Pasive multilayer setup, forces F & B.
1 + Co sZE sZt
Consider the interface between layers i & i+1, the boundary
conditions are
Equations (22), (23) and (26) completely control the be-
(
(−Fili + Bili )/Zi = (−Fi+1,0 + Bi+1,0 )/Zi+1
haviour of the transducer: we have three input parameters:
Fili + Bili = −Fi+1,0 + Bi+1,0
F−10 , B10 and VE , altogether with three output parameters:
F10 , B−10 and Vt . Then we can develop a transfer function hence
from input to output (
− Fi+1,0 + Bi+1,0 = Zi+1 /Zi (−Fili + Bili )
F10
B10
Fi+1,0 + Bi+1,0 = Fili + Bili
B−10 = (Pi,j ) F−10
if we add and subtract both equations we obtain the following
Vt Ve
Fi+1,0 1 1 + Zi+1 /Zi 1 − Zi+1 /Zi Fili
=
where
Bi+1,0 2 1 − Zi+1 /Zi 1 + Zi+1 /Zi Bili
together with
P11 P12 P13
e−sTi
P = P21 P22 P23 Fili 0 Fi,0
=
P31 P32 P33 Bili 0 esTi Bi,0
M1 MINIPROJECT, ERASMUS MUNDUS MASTER IN COMPLEX SYSTEMS SCIENCE, JUNE 2014 7
B−(m+1)0
B
U11 B
U12
B−10
At the left boundary of layer a:
= B B
F−10 U21 U22 F−(m+1),0
(
ua|LB = Aa eika da /2 + Ba e−ika da /2
what can be reexpressed as: pa|LB = za Aa eika da /2 − za Ba e−ika da /2 + hD
At the right boundary of layer b1:
F(n+1)0 B(n+1),0 (
B−(m+1)0 = Wij F−(m+1),0 ub1|RB = Ab1 e−ikb1 db1 /2 + Bb1 eikb1 db1 /2
Vt VE pb1|RB = zb1 Ab1 e−ikb1 db1 /2 − zb1 Bb1 eikb1 db1 /2
M1 MINIPROJECT, ERASMUS MUNDUS MASTER IN COMPLEX SYSTEMS SCIENCE, JUNE 2014 8
And the boundary conditions impose that both velocities In the very front and back layers there is no incoming
and pressures must be the equal: and returning wave (respectively), this is AbN = BaM = 0,
resulting in:
z0 d
Aa eika da /2 + Ba e−ika da /2 =
A 0
(28a) HaM aM hV AaM
=Ab1 e−ikb1 db1 /2 + Bb1 eikb1 db1 /2 xaM = = =
A0aM
FaM z0 d
ika da /2 −ika da /2
AaM
za A a e − B a e + hD = hV
(28b)
z 0d
−ikb1 db1 /2
− Bb1 eikb1 db1 /2 B
=z
b1 Ab1 e
0
HbN bN BbN
xbN = =
hV
= 0
FbN z0 d −BbN
we will denote from now on αj = kj dj . −BbN
hV
Now, the voltage across the only active layer of the trans-
Due to the form of the equations, we can now reduce them
ducer looks like
using transmission matrices, the procedure is detailed below.
Z da /2
We shall start with the front layers: consider first equation
V = E dx = −h ξda /2 − ξ−da /2 + D/ε da = (31d):
−da /2
Ma G− +
a xa + s = Ma1 Ga1 xa1
h +iαa /2
= e − e−iαa /2 (Aa − Ba ) + D/ε da
iω from the next equation we can write
(29)
xa1 = (Ma1 G−
a1 )
−1
(Ma2 G+
a2 )xa2
and rearranging we can find the expression for D:
Similarly, from the following one
εV 2hε
D= − sin (αa /2) (Aa − Ba ) xa2 = (Ma2 G−
a2 )
−1
(Ma3 G+
a3 )xa3
da ωda
Thus, following Huang’s calculations, we can now represent and so on. Substituting the obtained expressions we get to:
− −1
equations (28a) and (28b) in the following matrix form Ma G− +
a xa + s = Ma1 Ga1 [(Ma1 Ga1 ) (Ma2 G+
a2 )]·
1 0
cos (−αb1 /2) i sin (−αb1 /2)
Hb1
·[(Ma2 G−
a2 )
−1
(Ma3 G+ −
a3 )] . . . [(Ma(M−1) Ga(M−1) )
−1
·
0 = − −1
0 zb1 i sin (−αb1 /2) cos (−αb1 /2) Fb1 ·(MaM G+ +
aM )]xaM = [(Ma1 Ga1 )(Ma1 Ga1 ) ]·
− −1
·[(Ma2 G+ ] . . . [(Ma(M−1) G+
1 0 cos (−αa /2) i sin (−αa /2) Ha a2 )(Ma2 Ga2 ) a(M−1) )
= +
0 za0 i sin (−αa /2) cos (−αa /2) Fa
(Ma(M−1) G−
a(M−1) )
−1
]MaM G+
aM xaM =
0
+
1 − kT2 µa Fa (T+ + +
a1 . . . Ta(M−1) )MaM GaM xaM =
+
(30) = T+ +
a MaM GaM xaM = Ma Ga xa + s
from the previous equation we can write a32 = −[Tb− (1, 1) + Tb− (1, 2)zbN
0
]
xb1 = (Mb1 G+
b1 )
−1
(Mb2 G−
b2 )xb2 a42 = −[Tb− (2, ) + Ta+ (2, 2)zbN
0
]
Similarly, from the previous one
xb2 = (Mb2 G+ −1
(Mb3 G− Thus we can finally write
b2 ) b3 )xb3
−c− −is−
0
and so on. Substituting the obtained expressions we get to: a11 0 a a AaM 0
a21 0 −is− a −c+ 2 0 1
a + kT µa BbN
c+ is+ · Ha = 0
−
0 a32
Ma G+ + −1
a xa + s = Mb1 Gb1 [(Mb1 Gb1 ) (Mb2 G−
b2 )]·
a a
0 a42 is+
a c+ 2
a − kT µa Fa 1
· [(Mb2 G+
b2 )
−1
(Mb3 G− +
b3 )] . . . [(Mb(N−1) Gb(N−1) )
−1
·
· (MbN G− − + −1 where c± ±
a = cos(±αa /2) and sa = sin(±αa /2).
bN )] · xbN = [Mb1 Gb1 (Mb1 Gb1 ) ]·
· [Mb2 G− + −1
b2 (Mb2 Gb2 ) ] . . . [(Mb(N−1) G−
b(N−1) ) We can easily solve this system using for example Cramer’s
(Mb(N−1) G+
b(N−1) )
−1
]MbN G−
bN xbN = rule.
= (T− − −
b1 . . . Tb(N−1) )MbN GbN xbN =
− To finish with this model, we can write the expression for
= T+ +
b MbN GbN xbN = Ma Ga xa + s the transducer impedance as follows: consider equation (29)
where T− bn is the transmission matrix from the back layer h +iαa /2
to the transducer rear face. Each of the single transmission V = e − e−iαa /2 (Aa − Ba ) + D/ε da =⇒
iω
matrices is as follows, in a completely analogous fashion as Dda
h2 ε 2 zo da
we did before: =⇒ =V 1− sin(αa /2) (Aa − Ba )
ε Y αa hV ε
T− −
b = (Mbn Gbn )(Mbn Gbn )
+ −1
=
Dda εV
= V 1 − kT2 µa Fa =⇒ D = 1 − kT2 µa Fa
0 =⇒
cos(−αn /2) i/zan sin(−αn /2) (33)
= 0
ε da
i/zan sin(−αn /2) cos(−αn /2)
thus the electrical current through the transducer is
εV
Using (32) and (33) we can rewrite equations (31a)-(31e) I = iωDS = iωS (1 − kT2 µa Fa ) = iωCo (1 − kT2 µa Fa )V
da
in the more compact matrix form
( and the impedance
T− − +
b MbN GbN xbN = Ma Ga xa + s
Ma G− + + V 1 1
a xa + s = Ta MaM GaM xaM Z= = 2 (35)
I iωCo 1 − kT µa Fa
−
but Ma , G+ aM and GbN can be set to the identity matrix,
yielding to VI. A NALYTICAL RESULTS
(
T− +
b MbN xbN = Ga xa + s
− +
Ga xa + s = Ta MaM xaM
(34)
I N the three previous sections we have introduced and
defined three different analytical models that varied in
capabilities and modelling techniques. For a considered
where transducer, all will provide analyical results for the
A0aM 0
BbN
xaM = & xbN = transmitter and receiver force-voltage relations and the
A0aM −BbN0
electrical impedance behaviour.
As a last step, we can write (34) as a single linear system To test our models against FEM and real data from the
Ax = b. From (34) we obtain: laboratory, we have considered three test cases whose specifi-
cations can be found in the following table:
0
−a32 BbN − [cos(αa /2)Ha + i sin(αa /2)Fa ] − 0 = 0
0 Test Case 1 Test Case 2 Test Case 3
−a42 BbN − [i sin(αa /2)Ha + cos(αa /2)Fa ] −
Material PZT PZT PZT
− 1 − kT2 µa Fa = 0
Dim 25x25mm 20x20mm 20x20mm
a11 A0aM − [cos(−αa /2)Ha + i sin(−αa /2)Fa ] − 0 = 0 Thickness 0.5mm 1mm 2mm
−a21 A0aM − [i sin(−αa /2)Ha + cos(αa /2)Fa ] −
The three transducers are made of Lead zirconate titanate
− 1 − kT2 µa Fa = 0
(PZT), one of the most common piezoelectric materials. Notice
where the lateral dimensions-thickness ratio for the three of them,
beginning with a reasonably high value and decreasing af-
0
a11 = [Ta+ (1, 1) + Ta+ (1, 2)zaM ] terwards. As the considered analytical models are 1D, we
will check how well this approximation holds for real world
0
a21 = [Ta+ (2, 1) + Ta+ (2, 2)zaM ] devices.
M1 MINIPROJECT, ERASMUS MUNDUS MASTER IN COMPLEX SYSTEMS SCIENCE, JUNE 2014 10
0.5
Test Case 1 (PZT5A) 25mmx25mm, 0.5mm thickness
1
0.4
Lab data
0.9 FEM
Lattice 0.3
Normalized impedance modulus (Ohms)
0.8 Huang
LSM 0.2
0.7
0.1
0.6
0
0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2
0.5 Frequency (Hz) ·106
1 0.6 Huang
Lab data LSM
0.9 FEM
Normalized impedance modulus (Ohms)
0.4
LSM
0.8 Lattice 0.2
Huang
0.7
0
0.6
−0.2
0.5
−0.4
0.4
−0.6
0.3
−0.8
0.2
−1
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6
0.1 Frequency (Hz) ·106
Test Case 2 (PZT5A) 20mmx20mm, 1mm thickness Fractional BW: 0.073993 . Test Case 1
1
0.8
Lab data 0.9
0.6 FEM
Normalized impedance phase (degrees)
LSM 0.8
0.4 Lattice
−0.6 0.2
−0.8 0.1
−1 0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 4.05 4.06 4.07 4.08 4.09 4.1 4.11 4.12 4.13 4.14
Frequency (Hz) 6 Frequency (Hz) ·106
·10
Fig. 16: Electrical impedance phase for test case 2 Fig. 18: Fractional bandwidth for test case 1
Test Case 3 (PZT5A) - 20mmx20mm, 2mm thickness Fractional BW: 0.587257 . Test Case 2
1
1
Lab data
FEM 0.9
0.8
Lattice
Normalized impedance phase (degrees)
Huang 0.8
0.6
Normalized force modulus (N)
LSM
0.7
0.4
0.2
0.6
0 0.5
−0.2 0.4
−0.4 0.3
−0.6 0.2
−0.8 0.1
−1 0
0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.9 1.92 1.94 1.96 1.98 2 2.02 2.04 2.06 2.08 2.1 2.12 2.14 2.16 2.18 2.2
Frequency (Hz) ·106 Frequency (Hz) ·106
Fig. 17: Electrical impedance phase for test case 3 Fig. 19: Fractional bandwidth for test case 2
Our results for the three transducers are presented below 0.3
A mismatch between transducer and loading medium Fig. 20: Fractional bandwidth for test case 3
acoustic impedances will result in poor bandwidth and energy
transfer, being them increased as both impedances get closer.
Hence, a preliminary test for our models will consist of shown below (figure 21).
varying the theoretical loading medium impedance and find
the maximum in the fractional bandwidth curve, which is The black vertical line shows the transducer impedance,
M1 MINIPROJECT, ERASMUS MUNDUS MASTER IN COMPLEX SYSTEMS SCIENCE, JUNE 2014 12
ideally we should get the maximum exactly at that point, but the first one, the model predicts as well high bandwidth peaks
due to the tendency of finding the maximum impedance peaks at the successive λ/2 jumps. The second one seems to have
a bit shifted, we find a maximum point which is close to the destructive wave interaction for early jumps, being maybe
theoretical result. The third test case, the one further from 1D stabilized at a later stage. However it is difficult to judge
approximation, presents some drift for higher impedances. because the dashed lines cut the bandwidth curve at very
steep points. The third case is more difficult to interpret: we
FBW optimization for a varying loading medium impedance definitely see the spike behaviour located around the dashed
120
TC1 lines, but it does not agree with the expected behaviour. This
110 TC2 may be caused by numerical instabilities or by a too small
TC3
100
lateral dimensions-thickness ratio, making theh 1D model
insuficient for this kind of analysis. In this case, further
90
analysis using laboratory instrumentation or FEM should be
Fractional bandwidth
80 used.
70
60 Summarizing, nice agreement with theory is found using the
50
two first test cases, while the third one needs further testing. It
seems to work fine for transducers thoroughly satisfying our
40
assumptions. Deeper and broader analysis must be performed
30 for other transducers to check the validity of this models.
20
·10−2 Test Case 1 - Optimum matching impedance
10 5
0 4.5
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Front medium impedance Z1 (Rayl) ·108 4
2.5
C. Bandwidth and matching layers
2
When designing an application specific transducer with a
1.5
given loading medium, the matching layers come into play:
their purpose it to minimize the acoustic impedance mismatch 1
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Literature gives the ideal impedance value for a single Matching layer thickness (m) ·10−3
matching layer as the geometric mean √ of the transducer Fig. 22: Matching layer optimization for TC1
and loading medium impedances: Z = Zt Zl . The other
important aspect of the matching layers is their thickness:
optimal thickness for a single matching layer is a quarter Test Case 2 - Optimum matching impedance
of the operating wavelength (λ = c/f , being c the speed of 0.35
that waves that are reflected within the matching layer are
Fractional bandwidth
Test Case 3 - Optimum matching impedance considering the new index notation specified in the figure.
2.5
KF 2 TF 2 KB2 TB2 U (s)
1.5 1 − h2 +
2 2 s2 Ze Zc2
h · KF 2 TF 2 U (s)/sZc2
1 = ·
K F 2 TF 2 KB2 TB2 U (s)
1 − h2 +
2 2 s2 Ze Zc2
0.5
e · h · a(s)Z2 /(Zc1 + Z2 ) KF 1 (s)Y (s)
·
1 − h2 Y (s)/(sZc ) (TF 1 KF 1 /2 + TB1 KB1 /2)
0
0 0.5 1 1.5 2 2.5 3 3.5 4
Matching layer thickness (m)
To test the results we have used the first test case transducer
·10−3
as transmitter and receiver. The in-between medium is chosen
Fig. 24: Matching layer optimization for TC3 to be steel (Z2 = 45MRayl) with large enough thickness. The
results are presented below.
applications: we place two ultrasonic transducers at both sides
of an in-between medium, this can be a wall or other kind of Pitch - Catch system for TC1
object (see figure 18). 1 1
Air Piezo In-between medium Piezo Air Normalized force output - transmitter
0.8 0.8
VIII. ACKNOWLEDGEMENT
I would like to thank Nishal Ramadas for all the help and
training provided.
R EFERENCES
[1] M. Willatzen, Ultrasound transducer modelling – general theory and
applications to ultrasound reciprocal systems, IEEE Transactions on
Ultrasonics, Ferroelectrics and Frequency Control, vol. 48, No 1, January
2001.
[2] M. Redwood, Transient performance of a piezoelectric trasnducer, J.
Acoust. Soc. Am. 33 (4), April 1961.
[3] G. Hayward, C. J. MacLeod, A systems model of the thickness mode
piezoelectric transducer, J. Acoust. Soc. Am. 76 (2), August 1984.
[4] N. Ramadas, G. Hayward, C. J. C. Barbenel, Y. Estanbouli, A linear
systems model of the thickness mode piezoelectric transducer containing
dual piezoelectric zones , Ultrasonics Symposium, 2004 IEEE, Vol. 3.
[5] G. Hayward, M. N. Jackson, A lalttice model of the thickness-mode
piezoelectric transducer, IEEE Transactions on Ultrasonics, Ferroelectrics
and Frequency Control, vol. UFFC-33, No 1, January 1986.
[6] C Huang, M. V. Z. Marmarelis, Q. Zhou, K. K. Shung An analytical
model of multilayer ultrasonic transducers with an inversion layer, IEEE
Transactions on Ultrasonics, Ferroelectrics and Frequency Control, vol.52,
No 3, March 2005.
[7] T. E. Gómez Álvarez-Arenas and Luis Dı́ez, Novel impedance matching
materials and strategies for air-coupled piezoelectric transducers, SEN-
SORS, 2013 IEEE Conference, November 2013.