Current Applied Physics 10 (2010) 381–385

Contents lists available at ScienceDirect

Current Applied Physics

journal homepage: www.elsevier.com/locate/cap

Acoustic characteristics of composite materials as acoustic window at oblique

incidence of sound waves
So Jung Lee, Suk Wang Yoon *
Acoustics Research Laboratory, Department of Physics, Sungkyunkwan University, Suwon 440-746, Republic of Korea

a r t i c l e i n f o a b s t r a c t

Article history: This paper described a method for estimating the acoustic characteristics of composite materials at obli-
Received 25 May 2009 que incidence of sound waves. Composite materials are used as acoustic windows of SONAR to protect
Accepted 22 June 2009 the internal sensors and electronic parts from water. In this study the composite material of glass rein-
Available online 24 June 2009
forced plastic and polyurethane was used as the specimen. As the acoustic characteristics the velocities
and attenuation coefficients of sound waves through the composite material were measured in the high
PACS: frequency range. The insertion loss was also measured as a function of incident angle at 200 and 76 kHz,
62.65
respectively. The attenuation coefficients in the low frequency range were estimated by interpolating the
68.45.N
43.30
measured attenuation in the high frequency range with power-law form fitting. A four-medium layer
model was proposed to estimate the insertion loss of composite materials at oblique incidence of sound
Keywords: waves in the low frequency range. The four-medium layer model well described the experimentally mea-
Acoustic window sured insertion loss at the high frequency range. It suggests that the insertion loss of the composite mate-
Insertion loss rials can be well estimated as a function of incident angle in the low frequency range.
Sound wave Ó 2009 Elsevier B.V. All rights reserved.
Oblique incidence
Four-medium layer model
Longitudinal wave
Transverse wave

1. Introduction waves. The composite material of glass reinforced plastic (GRP)

and polyurethane (PU) was used as the specimen in this study.
Acoustic windows need to have good mechanical strength in or- Large composite materials are needed as the specimens to measure
der to protect acoustic sensors and internal electronic parts from the acoustic characteristics at low frequencies. However, it is diffi-
all types of external impacts. They should not distort the sound cult to manufacture such large composite materials and even in
waves and cause their energy loss [1,2]. Sound waves will pass low frequency range experiments with them are more difficult.
through the acoustic window of streamlined dome at various an- Therefore, it is necessary to estimate the acoustic characteristics
gles [3]. Therefore, it is necessary to understand the acoustic char- of composite materials at low frequency range with scale factor
acteristics of acoustic windows when sound waves are transmitted using those measured at high frequency range.
at oblique incidence. When sound waves travel through a solid at
oblique incidence, the transverse waves as well as longitudinal 2. Theoretical consideration of four-medium layers
waves propagate in the solid [4]. When the solid is surrounded
by liquid, the transverse wave in the solid can be converted and ob- Composite materials surrounded by liquid can be considered as
served as the longitudinal wave in liquid because the transverse the four-medium layers. Sound waves pass through liquid 0, solid
wave cannot propagate in liquid [5,6]. Composite materials are 1, solid 2 and liquid 3. When sound waves travel through compos-
considered to improve the acoustic window performance. How- ite materials at oblique incidence, the longitudinal and transverse
ever, there are not enough data for the practical applications of waves can be generated, and they can propagate in the composite
composite materials as acoustic windows. Sound wave propaga- materials. Fig. 1 shows the longitudinal and transverse waves in
tion is also complicated in composite materials. each medium. As shown in Fig. 1, the transverse wave in the com-
This paper describes a method to estimate the acoustic charac- posite materials can be converted to the longitudinal wave in li-
teristics of composite materials at oblique incidence of sound quid. It can be observed as the longitudinal wave in liquid since
the transverse wave cannot propagate in liquid. The propagation
* Corresponding author. Tel.: +82 31 290 7043; fax: +82 31 290 7055. direction of the waves in each layer can be determined by Snell’s
E-mail address: [email protected] (S.W. Yoon). law [7].

1567-1739/$ - see front matter Ó 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.cap.2009.06.031
382 S.J. Lee, S.W. Yoon / Current Applied Physics 10 (2010) 381–385

2.2. Insertion loss of four-medium layers

Fig. 1 shows the coordinate axes. The pressures of the longitu-

dinal and transverse waves, Li and Si are as follows:

Li ¼ ½L0i eihi ðsin hLi xþcos hLi zÞ þ L00i eihi ðsin hLi xcos hLi zÞ eixt ;
ð5Þ
Si ¼ ½S0i eiki ðsin hSi xþcos hSi zÞ þ S00i eiki ðsin hSi xcos hSi zÞ eixt ;
where L0i and S0i are the pressure amplitudes of the incident longitu-
dinal and transverse waves, and L00i and S00i are the pressure ampli-
tudes of the reflected longitudinal and transverse waves,
respectively.
The boundary conditions should be satisfied with the continu-
ities of particle velocities and stresses at the boundaries. These
conditions allow the quantities at the second solid medium to be
expressed in terms of the corresponding quantities at the first solid
medium. We obtain the equations satisfying the boundary condi-
tions expressed in matrix form:
2 3 2 32 _ 3
n_ 2 E11 E12 E13 E14 n0
6_ 7 6 6_ 7
6 f2 7 6 E21 E22 E23 E24 7
76 f0 7
6
6 7¼6 7 7 ð6Þ
6r 7 4E E34 56 7
4 25 31 E32 E33 4 r0 5
s2 s0
2G2 E41 E42 E43 E44 2G0

where Eij is an element of matrix satisfying the boundary

conditions.
Since liquid cannot sustain tangential forces, the shear stresses
s0 and s2 must vanish. In addition, the following can be written
Fig. 1. Longitudinal and transverse waves in the four-medium layers. The propa-
from Eq. (5) with S0 ¼ 0 and S3 ¼ 0 for liquid:
gation direction of longitudinal and transverse waves in each layer can be
f_ 2 ¼ c3 n3 u000 ;
determined by Snell’s law.
f_ 0 ¼ c0 n0 ðu0  u00 Þ;
2.1. General equations for displacement and stress ð7Þ
r2 ¼ c23 q3 u000 ;
The general equations for displacements can be described in r0 ¼ c20 q0 ðu0  u00 Þ;
terms of Li and Si as follows:
where u0 ; u00 and u000 denote the incident, reflected and transmitted
1 @Li 2 @Siz 1 @Siy waves in liquid 0 and liquid 2, respectively, and ci is the velocity in
ni ¼  2 þ 2  2 ;
hi @x ki @y ki @z the ith medium and ni is the direction cosine of the normal to inci-
1 @L 2 @S 1 @S dent wave with respect to the z axis. By substituting Eq. (7) into Eq.
gi ¼  2 i þ 2 ix  2 iz ; ð1Þ (6), the transmission coefficient becomes
hi @y ki @z ki @x
1 @Li 2 @Siy 1 @Six u000
T total ¼
fi ¼  2
þ 2  2 ; u0
hi @z ki @x ki @y      
2 E23  E21E41E43 E32  E31E41E42 þ 2 E22  E21E41E42 E33  E31E41E43
where ni , gi and fi are the particle displacements in the x, y and z ¼        :
axes of the ith medium, respectively, Li is the pressure of the longi- E22  E21E41E42 þ E33  E31E41E43 þ E23  E21E41E43 c0nq0 0 þ E32  E31E41E42 c0nq0
0

tudinal wave in the ith medium, and Si is the pressure of the trans- ð8Þ
verse wave in the ith medium.
The general equations for stress can be also described in terms The insertion loss (IL) is defined as
of displacements as follows: IL ¼ 20 log jT total j: ð9Þ
rix ¼ ki Li þ 2Gi ð@ni =@xÞ; In order to predict more accurately the insertion loss, the atten-
riy ¼ ki Li þ 2Gi ð@ gi =@yÞ; ð2Þ uation of composite materials needs to be considered. Generally,
riz ¼ ki Li þ 2Gi ð@fi =@zÞ; attenuation is included in theoretical equations by allowing the
wave number to be complex [3]. The longitudinal and transverse
k ¼ mE=½ð1 þ mÞð1  2mÞ; ð3Þ wave numbers are defined as

  x x
@ gi @fi kLi ¼ ; kSi ¼ ; ð10Þ
six ¼ Gi þ ; tLi tSi
 @z @y where tLi and tSi are the complex velocities of the longitudinal and
@f @n
siy ¼ Gi i þ i ; ð4Þ transverse waves in the ith medium, respectively. The complex
 @x @z  velocities, tLi and tSi , can be expressed as
@n @ g
siz ¼ Gi i þ i ; xcLi 1
@y @x tLi ¼ 2 1=2
ei tan ðaLi cLi =xÞ ;
½x þ ðaLi cLi Þ 
2
where ri and si are the normal and shear stresses in the ith med- xcSi 1
ð11Þ
ium, respectively, E is the Young’s modulus, m is the Poisson’s ratio, tSi ¼ 2 ei tan ðaSi cSi =xÞ ;
and G is the shear modulus of the ith medium.
½x þ ðaSi cSi Þ2 1=2
 S.J. Lee, S.W. Yoon / Current Applied Physics 10 (2010) 381–385 383

where cLi and cSi are the velocities of the longitudinal and transverse Table 2
waves in the ith medium, respectively, and aLi and aSi are the atten- Velocities of the longitudinal and transverse waves in GRP and PU.

uation coefficients of the longitudinal and transverse waves in the GRP PU

ith medium, respectively. Velocity of the longitudinal wave (m/s) 2529 ± 15 1962 ± 43
Velocity of the transverse wave (m/s) 1386 ± 3 772 ± 0.3
3. Results

The velocities, attenuation coefficients and insertion loss of the The attenuation coefficients can be obtained comparing the spectra
composite material were measured to verify the four-medium of the Fourier transforms of the signals received by the receiver
layer model. Table 1 shows the mechanical properties of GRP and with and without a test specimen. The attenuation coefficients of
PU, such as thickness, density and Young’s modulus. Fig. 2 shows the longitudinal and transverse waves, aL and aS can be described
a schematic diagram of the experimental setup. as follows:
cL and cS are the velocities of the longitudinal and transverse
waves. For the velocity measurements, the transmitted signals lnðTAw =As Þ
aL ¼ ;
were measured both with and without a test specimen [8]. d
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð13Þ
cw 1  ðcS sin h0 =cw Þ2 lnðT t Aw =As Þ
cL ¼ ; aS ¼ ;
1  Dtcw =d d
cw ð12Þ
cS ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ; where As and Aw denote the FFT amplitudes for the transmitted sig-
2
sin h0 þ ðcos h0  cw Dt=dÞ2 nals with and without a test specimen, respectively.
Figs. 3 and 4 show the attenuation coefficients of GRP and PU as
where cw is the sound speed in water, Dt is the time difference be- a function of frequency. The symbols indicate the attenuation coef-
tween the transmitted signals with and without a test specimen, d ficients measured for the longitudinal and transverse waves
is the thickness of a test specimen and h0 is the incident angle. Table around 200 and 500 kHz, respectively. The solid and dotted lines
2 shows the velocities of the longitudinal and transverse waves in are the fitted curves for the attenuation coefficients expressed in
GRP and PU. The velocity of the longitudinal wave in GRP is faster power-law form [10–12]:
than the sound speed in water, while the velocity of the transverse
wave in GRP is slower than that. The velocities of the longitudinal a ¼ a0  jf jy ; ð14Þ
and transverse waves of PU showed similar results to GRP. where a is the attenuation coefficient expressed in power-law form,
The attenuation coefficients of GRP and PU were measured, as f is the frequency, a0 is a constant, and y is a real positive finite
one of the acoustic characteristics of the composite materials as number. Table 3 shows a0 and y of GRP and PU, respectively. These
an acoustic window. The broadband pulse technique developed results were used to predict the attenuation coefficients of GRP and
by Sachse and Pao was used for attenuation measurements [9]. PU in the low frequency range.
Large composite materials are needed to measure the insertion
loss as a function of incident angle in the low frequency range.
Table 1
Mechanical properties of the Glass Reinforced Plastic (GRP) and Polyurethane (PU).
However, it is difficult to manufacture the large composite materi-
als, and even more difficult to experiment with them at low fre-
Thickness d (mm) Density q (kg/m3) Young’s modulus E (N/m2) quency. Therefore, it is necessary to have a theoretical model to
GRP 20 1494 9.25  109 estimate the insertion loss in the low frequency range.
PU 10 1141 0.423  109 The insertion loss of the composite material was measured as a
function of incident angle at the frequencies of 200 and 76 kHz,
respectively. The incident angle was changed from 0° to 50° with

Arbitrary waveform generator

((Agilent 33250A)
or
Digital Storage Oscilloscope
Pulser/receiver
((Panametrics pulser/receiver 5072 PR) (LeCroy
( LT322)

Specimen

Transmitter Receiver

Rotating table

Personal computer
Water bath

Fig. 2. Schematic diagram of the experimental setup.

384 S.J. Lee, S.W. Yoon / Current Applied Physics 10 (2010) 381–385

Fig. 3. Attenuation coefficients of the longitudinal and transverse waves as a

Fig. 5. Insertion loss of the composite material (GRP + PU) as a function of incident
function of frequency in GRP. The attenuation coefficients are expressed in power-
y angle at 200 kHz.
law form a ¼ a0  jf j .

Fig. 4. Attenuation coefficients of the longitudinal and transverse waves as a

function of frequency in PU. The attenuation coefficients are expressed in power- Fig. 6. Insertion loss of the composite material (GRP + PU) as a function of incident
law form a ¼ a0  jf jy . angle at 76 kHz.

medium layer model show the theoretical insertion loss as a func-

Table 3 tion of incident angle at 200 kHz in Fig. 5 and 76 kHz in Fig. 6,
a0 and y of power-law form a ¼ a0  jf jy of the attenuation coefficient a. respectively. The theoretical estimation agrees well with the
GRP PU experimental data. Overall, the four-medium layer model can rea-
sonably estimate the insertion loss of composite materials as a
Longitudinal wave a0 1.52  1016 3.30  106
Longitudinal wave y 2.66 0.92 function of incident angle at broad frequency ranges.
Transverse wave a0 5.77  1018 9.59  107
Transverse wave y 3.02 1.16
4. Conclusion

This paper described a method to estimate the acoustic charac-

1° increment. The distance from the composite material to the re- teristics of composite materials at oblique incidence of sound
ceiver was kept within 30 mm in order to exclude the effects of dif- waves. The four-medium layer model was proposed to estimate
fraction from edges of the composite material. The insertion loss the acoustic characteristics of composite materials at oblique inci-
changed as the incident angle was changed. The symbols show dence. To verify the four-medium layer model, we directly com-
the measured insertion loss of the composite material as a function pared it with the measured velocities, attenuation coefficients
of incident angle at 200 kHz in Fig. 5 and 76 kHz in Fig. 6, respec- and insertion loss of sound waves through the composite material.
tively. They show different patterns as a function of incident angle. The velocities of the longitudinal and transverse waves were
A four-medium layer model was proposed as a theoretical mod- measured at 200 kHz. In GRP, the velocity of the longitudinal wave
el. The model was tested by comparing the experimental data with was faster than that of the transverse wave. Polyurethane also
the theoretical estimation. The solid lines estimated by the four- showed similar results as shown in Table 2. The attenuation coef-
 S.J. Lee, S.W. Yoon / Current Applied Physics 10 (2010) 381–385 385

ficients of the longitudinal and transverse waves were also mea- References
sured around 200 kHz and 500 kHz, respectively. In both GRP and
PU, the attenuation coefficients of the transverse waves were lar- [1] Woo-Jin Jung, Seung-Jin Han, Won-Ho Kim, Ku-Kyun Shin, Jae-Jin Jeon, A study
on the measurement and analysis method for the acoustic transmission loss of
ger than those of the longitudinal waves as shown in Table 3. the material for the acoustic window of sonar dome, KSNVE 16 (7) (2006) 729–
The attenuation coefficients expressed in power-law form fitting 738.
can be interpolated to those at low frequency range. [2] Ji-Hye Lee, Byoung-Nam Kim, Ku-Kyun Shin, Jae-Jin Jeon, Suk Wang Yoon, A
study of acousic characteristics for sonar dome window materials, Proceedings
The insertion loss of the composite material was measured as of the Acoustical Society of Korea Conference 26 (2s) (2007) 173–174.
a function of incident angle from 0 to 50° at 200 and 76 kHz. At [3] G.R. Barnard, J.L. Bardin, J.W. Whiteley, Acoustic reflection and transmission
the oblique incidence of sound waves the insertion loss of the characteristics for thin plates, J. Acoust. Soc. Am. 57 (3) (1975) 577–584.
[4] L.M. Brekhovskikh, Waves in Layered Media, Academic Press, 1980 (Chapter 1).
composite material was determined by the total effects of the
[5] P.J. White, G.T. Clement, K. Hynynen, Longitudinal and shear mode ultrasound
longitudinal and transverse waves in the composite material. propagation in human skull bone, Ultrasound Med. Biol. 32 (7) (2006) 1085–
The experimental insertion loss well agreed with the theoretical 1096.
estimation predicted by the four-medium layer model. This sug- [6] S. Ye, J. Wu, J. Peach, Ultrasound shear wave imaging for bone, Ultrasound Med.
Biol. 26 (5) (2000) 833–837.
gests that the four-medium layer model can be used to estimate [7] D.T. Blackstock, Fundamentals of Physical Acoustics, John wiley & Sons, New
the acoustic characteristics of composite materials at oblique York, 2000 (Chapter 5).
incidence of sound waves for broad frequency ranges. [8] Junru Wu, Determination of velocity and attenuation of shear waves using
ultrasonic spectroscopy, J. Acoust. Soc. Am. 99 (5) (1996) 2871–2875.
[9] W. Sachse, Y.H. Pao, On the determination of phase and group velocities of
dispersive waves in solids, J. Appl. Phys. 49 (1978) 4320–4327.
Acknowledgments [10] Tae Sik Kim, Ki Suk Lee, Bongyoung Ahn, Zin Hyoung Lee, Measurement of
acoustic properties of polyurethane by the through transmission method, J.
This work was supported by the Agency for Defense Develop- Acoust. Soc. Kor 19 (1) (2000) 40–47.
[11] T.L. Szabo, Causal theories and data for acoustic attenuation obeying a
ment. The specimens used in this work were supplied by the frequency power law, J. Acoust. Soc. Am. 97 (1) (1995) 14–24.
Agency for Defense Development and the Department of Chemical [12] T.L. Szabo, A model for longitudinal and shear wave propagation in viscoelastic
Engineering, Sungkyunkwan University. media, J. Acoust. Soc. Am. 107 (5) (2000) 2437–2446.