ARCHIVES OF ACOUSTICS

Vol. 45, No. 4, pp. 699–707 (2020)

DOI: 10.24425/aoa.2020.135276

Research Paper

Optimal Selection of Multicomponent Matching Layers

for Piezoelectric Transducers using Genetic Algorithm

Tadeusz GUDRA(1)∗ , Dariusz BANASIAK(2)

(1)
Department of Acoustics and Multimedia, Wroclaw University of Science and Technology
Wroclaw, Poland
∗
Corresponding Author e-mail: [email protected]
(2)
Department of Computer Engineering, Wroclaw University of Science and Technology
Wroclaw, Poland
(received April 24, 2020; accepted June 22, 2020)

One major problem in the design of ultrasonic transducers results from a huge impedance mismatch be-
tween piezoelectric ceramics and the loading medium (e.g. gaseous, liquid, and biological media). Solving
this problem requires the use of a matching layer (or layers). Optimal selection of materials functioning as
matching layers for piezoelectric transducers used in transmitting and receiving ultrasound waves strictly
depends on the type of the medium receiving the ultrasound energy. Several methods allow optimal se-
lection of materials used as matching layers. When using a single matching layer, its impedance can be
calculated on the basis of the Chebyshev, DeSilets or Souquet criteria. In the general case, the typically
applied methods use an analogy to a transmission line in order to calculate the transmission coefficient T .
This paper presents an extension of transmission coefficient calculations with additional regard to the
attenuation coefficients of particular layers. The transmission coefficient T is optimised on the basis of
a genetic algorithm method. The obtained results indicate a significant divergence between the classical
calculation methods and the genetic algorithm method.
Keywords: acoustic impedance; matching layers; ultrasonic transducers.

1. Introduction particular medium. One of the parameters for evalu-

ating the operating efficiency of a transducer is the
Broad application of ultrasound waves in numer- transmission coefficient T , which defines the relation-
ous fields of technology requires ultrasound transduc- ship between the power dissipated at the load and the
ers to be adjusted to operating in various media: gases, power at the source. Table 1 shows the value of this
liquids, and solids. In the case of piezoceramic trans- coefficient for several piezoceramic materials operated
ducers, one of the main problems to be solved results in air (Z = 427 Rayl) and water (Z = 1.5 ⋅ 106 Rayl).
from huge acoustic impedance differences between the The data shown in Table 1 indicate that the prob-
ceramic material and the operating environment of lem of impedance mismatch is important for efficient
the transducer. This impedance mismatch has a sig- energy transmission from the transducer (especially
nificantly negative impact on the efficiency of ultra- when the transducer is operated in air). The literature
sound energy transmission from the transducer to the mentions a number of methods used to solve this

Table 1. Values of T for transmission from various piezoelectric materials into water and air (Nakamura et al., 2012).
Density ρ Velocity cL Acoustic impedance Z
Material T (water) T (air)
[kg/m3 ] [m/s] [MRayl]
Quartz SiO2 2650 5760 15.3 0.325 10.8 ⋅ 10−5
PZT-5A 7750 3880 30.0 0.181 5.5 ⋅ 10−5
PbNb2 O6 5800 2800 16.0 0.313 10.3 ⋅ 10−5
PVDV 1780 2260 4.6 0.742 35.9 ⋅ 10−5
700 Archives of Acoustics – Volume 45, Number 4, 2020

problem (Nakamura et al., 2012; Goll, 1979; With the use of the Chebyshev, DeSilets and Sou-
Alvarez-Arenas, 2004; Qian, Harris, 2014; Il- quet criteria, matching layers are selected only on the
ham et al., 2016). One of the most common meth- basis of acoustic impedance values for individual ma-
ods is to use a matching layer or an arrangement of terials. An analysis of a great number of scientific
such layers (Pedersen et al., 1982; Hamidimioglu, publications on the design of piezoceramic transduc-
KhuriYakub, 1990; Łypacewicz, Duriasz, 1992; ers indicates that their primary focus is on investi-
Gudra, Opielinski, 2002; Toda, Thompson, 2010; gating the influence of matching layer impedance on
2012; Fang et al., 2016; Wang et al., 2018). The op- transducer efficiency, and only secondary one is on
timal choice of materials functioning as matching la- other factors. One of such other factors is acoustic
yers for piezoelectric transducers used in the trans- attenuation of the matching layer material. Attenua-
mission and receiving of ultrasound waves strictly de- tion (amplitude decrease) of an ultrasound wave along
pends on the type of the medium into which ultra- its propagation path in an actual medium is char-
sound energy is transmitted and on the desired pa- acterised by the use of an amplitude or an energy
rameters of the pulse at the output of the transducer. attenuation coefficient. Amplitude attenuation coef-
These parameters may include ultrasound pulse shape, ficient α describes a relative decrease of the ampli-
pulse energy, amplitude-phase characteristics of trans- tude per unit of wave travel distance. Analogically,
ducer admittance, transmission bandwidth, and band energy attenuation coefficient is defined as a relative
shape. The simplest method to match the transducer decrease of wave intensity per unit of wave travel dis-
to the load is by using a quarter-wave layer. Multi- tance. Wave attenuation is characteristic of a arti-
layer systems may offer an improved match. How- cular medium (material) and is determined by direct
ever, in this case the choice of proper materials for measurement. The value of attenuation is expressed
the matching layer (or layers) becomes a new pro- in Np/m (or dB/m). Because attenuation depends on
blem. the frequency of a wave propagating in a particular
Values of acoustic impedances for matching lay- medium, other units are also used, e.g. Np/(m ⋅ Hz) or
ers, as offered by various authors in the literature, dif- dB/(m ⋅ Hz). The value γ, which describes attenuation
fer depending on the premises used by the authors in along a single wavelength (γ = α ⋅ λ), is a very useful
their calculations. The most commonly used formulas coefficient.
are those derived for long electric lines (the Cheby- The influence of the attenuation properties of the
shev criterion, Goll, 1979), in which matching layer matching layer material on the losses in the en-
impedance is calculated as a geometric average of the ergy transmitted by the transducer was analysed by
ceramic impedance Zc and the medium impedance Zm : Alvarez-Arenas (2004). Figures 1 and 2 show the re-
lationship between the transmission coefficient T and
Zlayer = (Zc ⋅ Zm )1/2 . (1) the acoustic impedance values, as well as the match-
ing layer attenuation coefficient for the transducer op-
The matching layer acoustic impedance calculated erated in air and in water. Although the transmission
on the basis of the Chebyshev criterion provides the coefficient was calculated from relationships obtained
transmission coefficient value equal to 1 for layer thick- on the basis of a schematic for a different equivalent
ness λ/4 (with no attenuation assumed). transducer than in (Alvarez-Arenas, 2004), the re-
The goal of the analysis performed by DeSilets sults are practically identical.
et al. (1978) was to obtain an optimal pulse shape. In
such a case, matching layer impedance is calculated on
the basis of the following relationship:

Zlayer = (Zc ⋅ Zm ) .
2 1/3
(2)

Souquet et al. (1979) demonstrated that the max-

imum amplitude may be obtained if a condition of
equal goodness between the electric branch and the
mechanical branch in the transducer equivalent circuit
is met when calculating matching layer impedance. In
that case, the matching layer impedance is as follows:

Zlayer = (2Zc ⋅ Zm ) .
2 1/3
(3)

Criteria (1)–(3) may be generalised to include the

case of a transducer with many matching layers, so as Fig. 1. Influence of the attenuation coefficient and acous-
to allow calculations of acoustic impedance for indivi- tic impedance of the matching layer for air (attenuation
dual layers. coefficient per wavelength γ = α ⋅ λ in Np).
 T. Gudra, D. Banasiak – Optimal Selection of Multicomponent Matching Layers. . . 701

(Thomson, 1950). The publication offered a matrix-

based solution to the analysed problem. In the analy-
sis, the attenuated, longitudinal plane sound wave is
assumed to propagate through a system of N layers of
materials having different acoustic impedances. Figu-
re 3 shows a physical model of a piezoelectric trans-
ducer with multiple matching layers.

Fig. 2. Influence of the attenuation coefficient and acous-

tic impedance of the matching layer for water (attenuation Fig. 3. Piezoceramic transducer with multiple matching
coefficient per wavelength γ = α ⋅ λ in Np). layers.

In the case of the first layer, symbols pi1 and vi1

The presented data indicate that the attenuation indicate respectively the pressure and the speed of par-
of the matching layer has a significant influence, espe- ticles in the incident wave. Analogically, the pressure
cially in the case of air. When selecting a material for and the speed of particles in the reflected wave are
the matching layer, its attenuation may have a greater indicated with symbols pr1 and vr1 . In the intermedi-
significance than the impedance value. ate layers, two waves propagating in reverse directions
may be distinguished: the transmitted wave (symbols
2. The model of the piezoelectric transducer ptk , vtk ) and the reflected wave (symbols prk , vrk ).
with multicomponent matching layers Acoustic impedances of the layers and their at-
tenuations are indicated with respective symbols zk∗
The basic parameter used in the evaluation of the and γk (1 ≤ k ≤ N ). In the general case, the acous-
match between the acoustic impedance of an ultra- tic impedance of the k-th layer is a complex value.
sound transducer and the acoustic impedance of the Symbols lk indicate the thicknesses of individual lay-
medium is the energy transmission coefficient T . It de- ers. The first layer (k = 1) is the material of the ce-
fines the ratio between the power dissipated at the load ramic in the transducer, while the last layer (k = N )
and the power at the source. In the case of a transducer is the medium into which ultrasound energy is trans-
with a single matching layer, the value of the coeffi- mitted from the transducer (thus an assumption that
cient T may be calculated from the following formula lN = ∞). At the boundary between each of the layers,
(Lynworth, 1965): the incident wave is partially reflected and partially
transmitted into the successive medium. In addition,
4Z1 when passing through successive layers, the wave is
T= 2
Re(ZWE ), (4)
∣Z1 + ZWE ∣ attenuated. Let symbols A∗k and Bk∗ indicate ampli-
tudes of the transmitted wave and of the wave reflected
where
in layer k, respectively. In the general case, these va-
ZT cos (kd) + jZ2 sin(kd) lues should be considered as complex values. The pro-
ZWE = Z2 [ ]. (5) pagation of an ultrasound wave through a system of
Z2 cos(kd) + jZT sin(kd)
N layers is then described by the following equation
In the above equations, symbols Z1 , Z2 , and ZT (Saffar, Abdullah, 2012; Saffar et al., 2014):
indicate respectively impedance of the ceramic in the
⎡ 1 1 ⎤⎥
−1
transducer, impedance of the matching layer, and ⎢
A∗N ⎢ 1 ⎥⎥ [TN −1 ] ... [T2 ] [T1 ] [ 1∗ ], (6)
A∗
the impedance of the medium (coefficient k is the wave [ ∗ ]=⎢ 1
BN ⎢ ∗ − ∗ ⎥ B1
constant, and d is the thickness of the matching layer). ⎢z ⎥
⎣ N zN ⎦
The above equations may be generalised to take into where
account the case of a transducer with multi-matching
⎡ e−jk1∗ l1 ∗
ejk1 l1 ⎤⎥
layers. However, these equations do not account for the ⎢
⎢ ⎥
attenuation in the matching layer. [T1 ] = ⎢ 1 −jk∗ l1 1 jk1∗ l1 ⎥⎥, (7)
⎢ e 1 − e
As it has been demonstrated in the previous sec- ⎢ z∗ ⎥
⎣ 1 z1∗ ⎦
tion, taking into consideration the attenuation of ⎡ t3 ⎤
⎢ ⎥
each of the matching layers may be important in 1 ⎢⎢ t1 − kk t2 ⎥⎥
their selection. Transmission of elastic waves through [Tk ] = ⎢ k ⎥ for k = 2, ..., N − 1, (8)
2 ⎢ kt ⎥
a multilayer solid medium was first investigated in ⎢ t 2 t1 ⎥
⎣ 3 ⎦
702 Archives of Acoustics – Volume 45, Number 4, 2020

t1 = cos k1k lk (ek2k lk + e−k2k lk ) The selection of an acoustic matching layer consists
+ j sin k1k lk (e −k2k lk
−e −k2k lk
), then in searching the state-space of possible solutions
(in this case, various combinations of materials for the
t2 = cos k1k lk (e−k2k lk − ek2k lk ) matching layer). When designing a transducer having
several matching layers and with a great number of
+ j sin k1k lk (e−k2k lk + ek2k lk ), available materials, the search space may be large and
t3 = zk [k1k − ηk2k + j (k2k − ηk1k )] . the task becomes complex. Solutions to optimisation
tasks with large search spaces may involve methods
Symbols Tk in Eq. (6) indicate transmission ma- based on artificial intelligence which include genetic
trices for the successive layers. If the value of A∗1 is algorithms. Publications (Saffar, Abdullah, 2012;
known and if BN ∗
= 0 (the last layer is limitless), rela- Saffar et al., 2014) provide examples of how genetic
tionship (6) provides a system of two linear equations algorithms are employed in the selection of matching
with two unknowns. Solving these equations provides layers for transducers operated in air. The mechanism
the values of A∗N and B1∗ . This may serve as the basis to behind the classical genetic algorithm can be shown as
calculate energy transmission coefficient for a system a series of the following steps:
of N layers TwN :
1) An initial population is generated randomly.
2
z∗ A∗N 2) The population is evaluated (selection). Individ-
TwN = ∣ ∗1 ∣ ∣ ∗∣ . (9) uals displaying the best fitness take part in the
zN A1
reproduction process.
In the above equations, the influence of attenuation 3) Genotypes of the selected individuals are sub-
in the matching layers is allowed for in the value of jected to evolutionary operators:
the wave constant, which is a complex value (Saffar,
Abdullah, 2012): a) crossover (recombination of the parent geno-
types),
ki
ki∗ = k1i + jk2i = 1/2
b) mutation (introduction of minor, random
(1 + η 2 ) changes to the genotypes).
1 1 4) Creation of a new generation (new population). If
⋅ [cos ( arctan(η)) − j sin ( arctan (η))], (10)
2 2 the solution is not satisfactory, return to point 2.
where Otherwise, a solution is obtained.
η = γ/π. Population is understood as a set of individuals
Detailed calculations for the above relationships are having a particular size. An individual is a set of task
provided in (Saffar, Abdullah, 2012). parameters which are a candidate solution to a prob-
lem and which are encoded in the form of chromo-
somes. In the case of the problem discussed here, its
3. Selection of matching layers using solution comprises such a combination of k matching
a genetic algorithm layers which ensures the optimisation of a particular
criterion. As the solution depends on the parameters
Classical methods allow calculations of acoustic of the ceramic material used in the transducer and on
impedances for a matching layer (or layers) depend- the parameters of the medium into which energy is
ing on the impedances of the transducer and of the radiated, the following structure of the chromosome
medium. However, the calculated values frequently (individual) can be assumed:
remain theoretical. Two aspects are noteworthy in
this case. Firstly, materials showing low acoustic oj = [w1 , (w2 ; l2 ), ..., (wn−1 ; ln−1 ), wn ], (11)
impedance have a high value of acoustic attenuation
coefficient. Secondly, for technological reasons, mate- where wi is the index of the material in the i-th layer
rials showing impedance values perfectly correspond- (I = 1, ..., n, n = k + 2), li is the thickness of the i-th
ing to the calculated theoretical values are difficult to matching layer.
produce. In addition, not all available materials can Symbol w1 denotes the ceramic material of the
be used as matching layers in transducers. A solution transducer, while wn is the medium into which energy
to the above problems may lie in treating the selec- is radiated. Symbols w2 − wn−1 indicate materials for
tion of matching values as an optimisation task. In the searched matching layers, while symbols l2 − ln−1
such an approach, candidate materials for a matching are thicknesses of individual layers, respectively. In the
layer should be selected from the available materials literature, the thickness of the matching layer is fre-
so as to optimise a selected parameter describing the quently accepted at 1/4 of the wavelength calculated
quality of the radiation from the transducer to a par- for the frequency of mechanical resonance fm (li = λ/4
ticular medium (e.g. the transmission coefficient T ). or uneven multipliers of this value). In such a case, the
 T. Gudra, D. Banasiak – Optimal Selection of Multicomponent Matching Layers. . . 703

chromosome structure may be simplified in the follow- positions 2, ..., N − 1 (positions 1 and N must remain
ing form: unaltered). This operation is equivalent to the change
oj = [w1 , w2 , ..., wn−1 , wn ]. (12) of the material in one of the matching layers.
An important element of the genetic algorithm
The selection method should ensure that succes-
which has a significant influence on its effectiveness
sive populations have a higher average fitness func-
is the fitness function. It allows each individual in the
tion value. The basic selection method is the roulette
population to be evaluated and the results to be used
wheel method. Despite some disadvantages, it is fre-
to select the fittest (closest to the optimal solution)
quently used in genetic algorithms. In the roulette
individuals for further steps of the algorithm. In this
wheel method, each individual is assigned a propor-
paper, the evaluation of individuals was based on the
tion of the wheel corresponding to the value of the
transmission coefficient T , as defined with relation-
fitness function for a particular individual (higher fit-
ships (6)–(9).
ness function value means larger proportion of the
wheel assigned to the individual). The development
of evolutionary methods resulted in more effective 4. The results of the experiment
methods: rank selection method, tournament selection
method, etc. The purpose of the experiments is to evaluate the
A population resulting from the selection process potential of the genetic algorithm for the identifica-
is then subjected to genetic operators. The goal be- tion of matching layers for piezoceramic transducers
hind this step is to obtain a new generation which operated in air and in water. An important element of
should include better solutions to the problem than the experiments is to evaluate the influence of atten-
the solutions in the previous generation. Two opera- uation in the matching layers on the obtained results.
tors are used in a classical genetic algorithm: crossover Therefore, the individuals were evaluated on the basis
and mutation. The simplest version of crossover is sin- of the transmission coefficient T , calculated from rela-
gle point crossover. In this case, a pair of individu- tionships (6)–(9). An assumption was made that the
als (parents) and a crossover point are randomly se- transducer was manufactured from the PZT ceramic
lected from the population. The parent individuals are material with impedance Z1 = 33 MRayl. For simplifi-
then split in the crossover point (which is identical cation purposes, the thickness of individual layers was
for both individuals). The first fragments of the in- assumed at λ/4. The experiments were performed on
dividuals (before the crossover point) are not altered, a database comprising 96 available materials with var-
while the remaining fragments of the individuals (from ious acoustic properties. The database was developed
the crossover point) are exchanged (subjected to the with the use of data included in a number of scien-
crossover operation). This procedure leads to the cre- tific articles on designing ultrasound transducers with
ation of a pair of two new individuals (offspring). For matching layers and in catalogues from manufacturers
the structure of individuals described with relation- of ultrasound materials (Onda, 2003; Hung, Gold-
ship (12) and for the value of N = 5 (3 matching layers stein, 1983; Rhee et al., 2001; Alvarez-Arenas,
are assumed), the crossover operator works as illus- 2004; Trogé et al., 2010; Saffar et al., 2014; Qian,
trated below. Two individuals, op1 and op2 , were ran- Harris, 2014). Table 2 presents some of the ma-
domly selected from the parent population: terials having various acoustic properties. Figure 4
shows the relationship between the acoustic impedance
op1 = [w1 , w12 , w13 , w14 , w5 ], and the acoustic attenuation for all of the materials
(13) from the database used in the experiments. The analy-
op2 = [w1 , w22 , w23 , w24 , w5 ]. sis of the materials included in the database leads to
two important conclusions: the database contains a rel-
In the next step, a crossover point equal 2 was ran- atively small number of materials with low acoustic
domly selected. As a result, two new individuals (off- impedance, and the materials with low impedance dis-
spring) were created – oc1 and oc2 : play high values of attenuation coefficient (balsa wood
is a good example in this case).
oc1 = [w1 , w12 , w23 , w24 , w5 ],
(14) The first experiment consisted in identifying how
oc2 = [w1 , w22 , w13 , w14 , w5 ]. the attenuation of a material influences the atten-
uation coefficient in the case of a transducer with
The purpose of mutation is to introduce diversity one matching layer. The analysis was performed for
into the new population in order to prevent the loss a transducer operated in air and in water. In the case
of important components from the solution. Mutation when the transmission coefficient is optimised, match-
consists in the change of the value of a selected bit in ing layer impedance may be calculated on the basis of
the string representing an individual. For the structure the Chebyshev criterion. Optimum impedance value is
of individuals described in relationship (12), mutation then 119 kRayl for air and 7036 kRayl for water. The
may consist in a random change of the symbol found in database used in the experiments does not contain ma-
704 Archives of Acoustics – Volume 45, Number 4, 2020

Table 2. Acoustic properties of some materials.

Density Velocity Acoustic impedance Attenuation
Material
[kg/m3 ] [m/s] [MRayl] [dB/mmMHz]
Nylon 1100 1800 2.00 0.058
Polycarbonate 1220 2270 2.77 0.442
Polyethylene 920 1950 1.79 0.70
Polypropylene 880 2740 2.40 0.10
Polystyrene 1030 2200 2.28 0.17
PVC 1380 2380 3.27 0.224
Teflon 2140 1390 2.97 4.10
Vinyl 2230 1330 2.96 0.256
Paraffin 910 1940 1.76 1.05
TPX-DX845 830 2220 1.84 0.44
Balsa wood 100 800 0.08 22.4
ITAKOM 1 500 1980 0.99 3.0
ITAKOM 2 250 910 0.23 6.5

terials is also provided with the value of the attenua-

tion coefficient with consideration to its actual atten-
uation (value T2 ). The obtained results lead to two
important conclusions. Firstly, the attenuation coeffi-
cient value decreases significantly when the matching
layer attenuation is considered (this fact applies espe-
cially to air). Secondly, when selecting the material,
its attenuation may have a greater importance than
its impedance.
The next stage of investigations focused on a trans-
Fig. 4. Acoustic materials in the database.
ducer with two matching layers. Table 4 includes each
5 best results obtained with the use of the genetic
terials with such impedance. Table 3 presents 5 best algorithm, for a transducer operated in air. Two cases
results found with the use of the genetic algorithm were considered: one, in which the attenuation of the
for each of the criteria. Each of the matching layer matching layers was not considered in the calculations
materials is provided with its index in the database of the transmission coefficient (the solution selected
(value m), as well as its impedance and acoustic at- on the basis of value T1 ), and another one, in which
tenuation (Z and γ, respectively). The materials were the attenuation was considered (the solution selected
arranged by the obtained value of the transmission co- on the basis of value T2 ). In both cases, each of the
efficient without consideration to the influence of the solutions was provided with the values of both T1 and
matching value attenuation (value T1 ). Each of the ma- T2 . The analysis of the results indicates that allowing

Table 3. Effect of material attenuation for a transducer with a single matching layer (top 5 results).
Medium (ZT [kRayl]) Layer 1 m (Z [kRayl], γ [NP]) Transmission coefficient T1 Transmission coefficient T2
14 (109; 0.249) 0.99276 0.01077
3 (131; 0.130) 0.99035 0.03283
Air (0.427) 2 (100; 0.150) 0.97116 0.02583
17 (98; 0.216) 0.96414 0.01380
12 (95; 0.249) 0.95197 0.01072
90 (6970; 0.145) 0.99991 0.84657
64 (7300; 0.204) 0.99864 0.79391
Water (1500) 88 (6265; 0.116) 0.98666 0.86273
87 (5475; 0.067) 0.93965 0.87014
92 (9150; 0.014) 0.93401 0.91841
 T. Gudra, D. Banasiak – Optimal Selection of Multicomponent Matching Layers. . . 705

Table 4. Effect of material attenuation for the transducer with two matching layers (air, top 5 results).
Layer 1 Layer 2 T1 T2
Medium (ZT [kRayl])
Without attenuation
86 (4810; 0.043) 22 (17; 0.600) 0.99969 0.05544
67 (4930; 0.096) 22 (17; 0.600) 0.99821 0.04626
65 (5100; 0.138) 22 (17; 0.600) 0.99422 0.03973
93 (4300; 0.015) 22 (17; 0.600) 0.99113 0.07090
Air (0.427) 85 (4130; 0.032) 22 (17; 0.600) 0.98204 0.07086
With attenuation
39 (1790; 0.011) 22 (17; 0.600) 0.43888 0.14009
47 (1840; 0.028) 22 (17; 0.600) 0.45722 0.12957
43 (2520; 0.010) 22 (17; 0.600) 0.68949 0.12677
57 (2270; 0.024) 22 (17; 0.600) 0.60928 0.12579
42 (2420; 0.019) 22 (17; 0.600) 0.65833 0.12463

for the attenuation significantly influences the obtained should be based also on the value of the attenuation
solutions. The best solution obtained on the basis coefficient for the particular material. Therefore, in the
of the T2 value (materials 39 and 22) has a low value of final stage of the research, investigations focused on
coefficient T1 (0.43888). In reverse, the best solutions how an increased number of matching layers influences
obtained on the basis of the T1 value show a low value the value of coefficient T2 (with allowance for material
of coefficient T2 . The results for a transducer operated attenuation). The considered cases involved from 1 to
in water are shown in Table 5. 4 layers. The results are presented in Table 6. In the
The above results demonstrate that the selection case of air, the best result was obtained for two match-
of materials for the matching layers of a transducer ing layers, and in the case of water – for three layers.

Table 5. Effect of material attenuation for the transducer with two matching layers (water, top 5 results).
Layer 1 Layer 2 T1 T2
Medium (ZT [kRayl])
Without attenuation
92 (9150; 0.014) 46 (1950; 0.107) 1.000000 0.934749
53 (80; 2.063) 22 (17; 0.600) 0.999989 0.000096
85 (4130; 0.032) 72 (885; 0.446) 0.999974 0.760768
55 (14100; 0.200) 83 (3025; 0.021) 0.999961 0.867857
Water (1500) 56 (12900; 0.300) 36 (2770; 0.116) 0.999949 0.758927
With attenuation
63 (10300; 0.002) 60 (2270; 0.027) 0.998901 0.983166
63 (10300; 0.002) 42 (2420; 0.019) 0.990622 0.979150
91 (16200; 0.001) 35 (3000; 0.003) 0.980413 0.978420
63 (10300; 0.002) 43 (2520; 0.010) 0.981293 0.974784
63 (10300; 0.002) 40 (2400; 0.032) 0.992148 0.973983

Table 6. Acoustic impedances of matching layers obtained with the genetic algorithm.
ZT [kRayl] Number of layers Layer 1 Layer 2 Layer 3 Layer 4 T2
1 5 (254; 0.073) – – – 0.06784
0.427 (air) 2 39 (1790; 0.011) 22 (17; 0.600) – – 0.14009
3 93 (4300; 0.015) 5 (254; 0.073) 22 (17; 0.600) – 0.13419
4 35 (3000; 0.003) 43 (2520; 0.010) 39 (1790; 0.011) 22 (17; 0.600) 0.13342
1 92 (9150; 0.014) – – – 0.91841
1500 (water) 2 63 (10300; 0.002) 57 (2270; 0.024) – – 0.98461
3 63 (10300; 0.002) 93 (4300; 0.015) 35 (3000; 0.003) – 0.98858
4 63 (10300; 0.002) 35 (3000; 0.003) 43 (2520; 0.010) 39 (1790; 0.011) 0.98582
706 Archives of Acoustics – Volume 45, Number 4, 2020

In the case of air, the transmission coefficient is most Sonics and Ultrasonics, 26(6): 385–393, doi: 10.1109/
influenced by the layer with the lowest impedance (it T-SU.1979.31122.
has a high attenuation value). As mentioned above, 5. Gudra T., Opielinski K.J. (2002), Influence of
due to technological constraints, materials with low acoustic impedance of multilayer acoustic system on
impedance have high attenuation. Also, the number the transfer function of ultrasonic airborne transducer,
of such materials is limited. Importantly, the results Ultrasonics, 40(1–8): 457–463, doi: 10.1016/S0041-
were obtained only for a specific set of materials (the 624X(02)00159-2.
database comprised 96 materials). The results may be
6. Hamidimioglu B., KhuriYakub B.T. (1990), Poly-
improved with a larger database.
mers films as acoustic matching layers, IEEE Sym-
posium on Ultrasonics, Honolulu, HI, USA, Vol. 3,
5. Conclusions pp. 1337–1340, doi: 10.1109/ULTSYM.1990.171581.
7. Hung B.H., Goldstein A. (1983), Acoustic pa-
This paper presents a genetic algorithm based rameters of comercial plastics, IEEE Transactions on
method for the selection of materials functioning as Sonics and Ultrasonics, 30(4): 249–254, doi: 10.1109/
matching layers in piezoelectric transducers. The in- T-SU.1983.31415.
troduction of a genetic algorithm allows the selection
of matching layer materials from the available mate- 8. Ilham H.M., Salim M.N., Jenal R.B., Hayashi T.
(2016), Guided wave matching layer using a quarter
rials so as to optimise a selected parameter related
of wavelength technique, Applied Mechanics and Ma-
to the transmission of energy from the transducer to
terials, 833: 59–68, doi: 10.4028/www.scientific.net/
a particular medium (this paper has focused on the AMM.833.59.
transmission coefficient T ). Depending on the imple-
mented fitness function, the calculations may include 9. Lynworth L.C. (1965), Ultrasonic impedance match-
various factors related to the selection of matching ing from solids to gases, IEEE Transactions on Son-
layers. The research focused on identifying the influ- ics and Ultrasonics, 12(2): 37–48, doi: 10.1109/T-SU.
1965.29359.
ence of the acoustic impedance and the acoustic at-
tenuation of matching layer materials on the energy 10. Łypacewicz G., Duriasz E. (1992), Design princi-
attenuation coefficient for transducers operated in air ples of transducers with matching layers base on ad-
and in water. The investigated case was a transducer mittance measurements, Archives of Acoustics, 17(1):
with multiple matching layers (from 1 to 4). The re- 117–131.
sults reveal that the matching layer acoustic attenu- 11. Nakamura K. et al. (2012), Ultrasonic transducers.
ation (overlooked in many publications) has a signif- Materials and design for sensors, actuators, and med-
icant influence on the transmission coefficient, espe- ical applications, Woodhead Publishing Series in Elec-
cially in the case of a transducer transmitting energy tronic and Optical Materials: Number 29, Elsevier,
into air. pp. 185–219, 374–407.
12. Onda Corporation (2003), Tables of Acoustic Proper-
This article is an extended version of the paper pre- ties of Materials, USA, http://www.ondacorp.com (ac-
sented during 16th Winter Workshop on Acoustoelec- cessed on 10 January 2020).
tronics, 24–27 February 2020, Szczyrk, Poland.
13. Pedersen P.C., Tretiak P., He P. (1982), Impe-
dance – matching properties of an inhomogeneous
References matching layer with continuously changing acoustic
impedance, The Journal of the Acoustical Society of
1. Alvarez-Arenas T.E.G. (2004), Acoustic impedance America, 72(2): 327–338, doi: 10.1121/1.388085.
matching of piezoelectric transducers to the air,
IEEE Transactions on Ultrasonics, Ferroelectrics, and 14. Qian Y., Harris N.R. (2014), Modelling of a novel
Frequency Control, 51(5): 624–633, doi: 10.1109/ high-impedance matching layer for high frequency
TUFFC.2004.1320834. (>30 MHz) ultrasonic transducers, Ultrasonics, 54(2):
586–591, doi: 10.1016/j.ultras.2013.08.012.
2. DeSilets C.S., Fraser J.D., Kino G.S. (1978), The
design of efficient broad-band piezoelectric transducers, 15. Rhee S., Ritter T.A., Shung K.K., Wang H.,
IEEE Transactions on Sonics and Ultrasonics, 25(3): Cao W. (2001), Materials for acoustic matching in ul-
115–125, doi: 10.1109/T-SU.1978.31001. trasound transducers, 2001 IEEE Ultrasonics Sympo-
sium. Proceedings. An International Symposium (Cat.
3. Fang H.J. et al. (2016), Anodic aluminum oxide– No.01CH37263), Atlanta, GA, USA, 2001, Vol. 2,
epoxy composite acoustic matching layers for ultra- pp. 1051–1055, doi: 10.1109/ULTSYM.2001.991900.
sonic transducer application, Ultrasonics, 70: 29–33,
doi: 10.1016/j.ultras.2016.04.003. 16. Saffar S., Abdullah A. (2012), Determination of
acoustic impedances of multi matching layers for nar-
4. Goll J.H. (1979), The design of broadband fluid- row band transmitter ultrasonic airborne transduces
loaded ultrasonic transducers, IEEE Transactions on with frequencies < 2.5 MHz – application of a genetic
 T. Gudra, D. Banasiak – Optimal Selection of Multicomponent Matching Layers. . . 707

algorithm, Ultrasonics, 52(1): 169–185, doi: 10.1016/ ducer impedance matching and backing absorber ap-
j.ultras.2011.08.001. plications, IEEE Transactions on Ultrasonics, Ferro-
electrics, and Frequency Control, 57(12): 2818–2827,
17. Saffar S., Abdullah A., Othman R. (2014), In- doi: 10.1109/TUFFC.2010.1755.
fluence of the thickness of matching layers on nar-
row band transmitter ultrasonic airborne transduc- 21. Toda M., Thompson M. (2012), Detailed investiga-
ers with frequencies <100 kHz: Application of a ge- tions of polymer/metal multilayer matching layer and
netic algorithm, Applied Acoustics, 75: 72–85, doi: backing absorber structures for wideband ultrasonic
10.1016/j.apacoust.2013.07.002. transducers, IEEE Transactions on Ultrasonics, Ferro-
electrics, and Frequency Control, 59(2): 231–242, doi:
18. Souquet J., Defranoud P.H., Desbois J. (1979), 10.1109/TUFFC.2012.2183.
Design of low loss wide band ultrasonic transducers
for noninvasive medical applications, IEEE Transac- 22. Trogé A., O’Leary R.L., Hayward G., Peth-
tions on Sonics and Ultrasonics, 26(2): 75–81, doi: rick R.A., Mullholland A.J. (2010), Properties
10.1109/T-SU.1979.31070. of photocured epoxy resin materials for application
in piezoelectric ultrasonic transducer matching lay-
19. Thomson W.T. (1950), Transmission of elastic waves ers, The Journal of the Acoustical Society of America,
through a stratified solid medium, Journal of Applied 128(5): 2704–2714, doi: 10.1121/1.3483734.
Physics, 21(2): 89–93, doi: 10.1063/1.1699629.
23. Wang Y. et al. (2018), Magnesium alloy matching
20. Toda M., Thompson M. (2010), Novel multi-layer layer for high-performance transducer applications,
polimer-metal structures for use in ultrasonic trans- Sensors, 18(12): 4424, doi: /10.3390/s18124424.