An audible demonstration of the speed of sound in bubbly liquids

Preston S. Wilson and Ronald A. Roy

Citation: American Journal of Physics 76, 975 (2008); doi: 10.1119/1.2907773

View online: http://dx.doi.org/10.1119/1.2907773
View Table of Contents: http://scitation.aip.org/content/aapt/journal/ajp/76/10?ver=pdfcov
Published by the American Association of Physics Teachers

Articles you may be interested in

Correction: “Bubbles that Change the Speed of Sound” (Phys. Teach.50, Nov. 2012)page 459
Phys. Teach. 51, 72 (2013); 10.1119/1.4775519

Bubbles that Change the Speed of Sound

Phys. Teach. 50, 458 (2012); 10.1119/1.4758141

Interaction between acoustic and non-acoustic mode in a bubbly liquid

AIP Conf. Proc. 1433, 531 (2012); 10.1063/1.3703243

A mechanism stimulating sound production from air bubbles released from a nozzle
J. Acoust. Soc. Am. 123, EL126 (2008); 10.1121/1.2908198

Sound–ultrasound interaction in bubbly fluids: Theory and possible applications

Phys. Fluids 13, 3582 (2001); 10.1063/1.1416502

This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.83.205.78 On: Mon, 11 May 2015 21:48:30
 APPARATUS AND DEMONSTRATION NOTES
Frank L. H. Wolfs, Editor
Department of Physics and Astronomy, University of Rochester, Rochester, New York 14627

This department welcomes brief communications reporting new demonstrations, laboratory equip-
ment, techniques, or materials of interest to teachers of physics. Notes on new applications of older
apparatus, measurements supplementing data supplied by manufacturers, information which, while not
new, is not generally known, procurement information, and news about apparatus under development
may be suitable for publication in this section. Neither the American Journal of Physics nor the Editors
assume responsibility for the correctness of the information presented.
Manuscripts should be submitted using the web-based system that can be accessed via the American
Journal of Physics home page, http://www.kzoo.edu/ajp/ and will be forwarded to the ADN editor for
consideration.

An audible demonstration of the speed of sound in bubbly liquids

Preston S. Wilsona兲
Mechanical Engineering Department and Applied Research Laboratories, The University of Texas at Austin,
Austin, Texas 78712-0292
Ronald A. Roy
Department of Aerospace and Mechanical Engineering, Boston University, Boston, Massachusetts 02215
共Received 5 January 2007; accepted 18 March 2008兲
The speed of sound in a bubbly liquid is strongly dependent upon the volume fraction of the gas
phase, the bubble size distribution, and the frequency of the acoustic excitation. At sufficiently low
frequencies, the speed of sound depends primarily on the gas volume fraction. This effect can be
audibly demonstrated using a one-dimensional acoustic waveguide, in which the flow rate of air
bubbles injected into a water-filled tube is varied by the user. The normal modes of the waveguide
are excited by the sound of the bubbles being injected into the tube. As the flow rate is varied, the
speed of sound varies as well, and hence, the resonance frequencies shift. This can be clearly heard
through the use of an amplified hydrophone and the user can create aesthetically pleasing and even
musical sounds. In addition, the apparatus can be used to verify a simple mathematical model known
as Wood’s equation that relates the speed of sound of a bubbly liquid to its void fraction. © 2008
American Association of Physics Teachers.
关DOI: 10.1119/1.2907773兴

I. INTRODUCTION a drinking glass filled with a “frothy liquid” produces a dull

sound. Perhaps inspired by his former laboratory assistant,2
The interaction of sound and bubbles has been studied in Rayleigh published a paper in 1917 that described the sound
modern science for almost 100 years and has been part of produced by water boiling in a tea kettle.3 In 1933, Minnaert
human consciousness for many hundreds of years, perhaps studied the sounds produced by running water, such as the
longer. It is a scientifically interesting and complex phenom- flowing stream and the crashing of ocean waves.4 The latter
enon with a long and diverse history, yet our understanding two studies both introduced relationships between the size of
of the phenomenon is still incomplete. Part of the intrigue a bubble and its period of oscillation with water as the host
comes from the fact that minute parameter changes in a bub- medium, but Rayleigh considered steam bubbles while Min-
bly fluid can have large and unexpected effects. For example, naert considered air bubbles.
consider a column of water contained within a rigid tube. The acoustics of bubbly liquids now has a number of im-
The addition of just a small number of bubbles, representing portant practical applications. These include a variety of in-
as little as 0.1% by volume, can transform the once simple dustrial processes utilizing multiphase flow,5 microfluidic
medium into a highly dispersive and nonlinear one, with a manipulation,6 and the design of nuclear reactor containment
speed of sound less than 20% of the speed of sound in the systems.7 In the ocean, various acoustic techniques are used
host medium. That corresponds to a 4 parts in 5 change in in conjunction with oceanic bubble populations to make
the speed of sound for a 1 part in 10 000 change in the measurements relevant to meteorology, chemical oceanogra-
density. It is the added compressibility of the gas phase that phy, marine fluid dynamics, and marine biology.8 By shroud-
causes the reduction of the speed of sound. ing propellers and hulls with man-made bubble screens, un-
The early scientific work on the acoustics of bubbles was wanted noise produced by ships and submarines is reduced.9
motivated primarily by the desire to understand certain man- Bubbles are used in medical applications to enhance ultra-
made and natural curiosities. In 1910, a paper was published sound imaging,10 hemostasis,11 and therapy.12
by Mallock1 that sought to explain the fact that when struck, Both theory and measurement reveal that for acoustic ex-

975 Am. J. Phys. 76 共10兲, October 2008 http://aapt.org/ajp © 2008 American Association of Physics Teachers 975
This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.83.205.78 On: Mon, 11 May 2015 21:48:30
 citation frequencies that are close to the individual bubble If the bubble radius is perturbed, equilibrium is lost and
resonance frequency, the acoustic attenuation within a bub- oscillations occur. The alternating compression and expan-
bly liquid is measured in tens of decibels per centimeter, and sion of the bubble is due to the transfer of energy between
that the sound speed can vary from as little as 20 m / s to as the inertia provided by the surrounding liquid and the com-
much as 10 000 m / s, which is supersonic relative to bubble- pressibility of the gas inside the bubble. For the frequency
free water. In the littoral ocean environment, this behavior is range between 100 and 1200 Hz and a bubble size of about
a dominant obstacle for effective sonar operation because 1 mm, it is sufficient to assume lossless conditions within the
bubbles and bubble clouds are produced in abundance by bubble, but dissipative effects are important in other
breaking waves. The acoustic contrast provided by bubbly regimes.20 At a hydrostatic pressure P⬁, the compressibility
liquid causes sound to be scattered from bubble clouds, and of the gas inside the bubble leads to an effective stiffness
hence naval sonar and mine hunting operations are hindered. 12␲ā␯ P⬁.21 The bubble behaves acoustically like a pulsating
Bubbles are also a source of ambient noise in the ocean. sphere and radiates sound uniformly in all directions. Since
Rain drops, spray from breaking waves, and even snow can the bubble size is always much smaller than the resulting
entrain bubbles when they impact the ocean surface. The acoustic wavelength, the low frequency limiting value of the
impact deforms the ocean surface and creates an air cavity. acoustic radiation reactance can be used to define an effec-
When this cavity closes, a bubble is formed and a second tive mass, known as the radiation mass 4␲ā3␳ᐉ, where ␳ᐉ is
impact that occurs when the air cavity closes causes the the liquid density. The bubble’s resonance frequency is then
bubble to oscillate and radiate sound.13–15 Breaking waves given by the square root of the stiffness to mass ratio, or

冑
produce clouds composed of large numbers of bubbles that
oscillate collectively and radiate sound. These bubble clouds 1 3␯ P⬁
␻0 = 共radians/sec兲. 共2兲
are responsible for part of the ambient noise in the ocean ā ␳ᐉ
below about 1000 Hz.16
The physics of bubble acoustics has been the subject of This relation was first derived by Minnaert4 in 1933, al-
previous “Apparatus and Demonstration Notes” within this though via a different methodology. Minnaert considered
journal,17–19 but the apparatus described here complements bubbles in the size range between 3 and 6 mm, and con-
and expands upon the previous work. The apparatus consists cluded that the period of oscillation would be small com-
of a transparent PVC pipe about 0.5 m in length. Air bubbles pared to the time required for the heat of compression to be
are injected at the bottom of the water-filled tube and a hy- conducted away. The bubble was considered to be undergo-
drophone is used to detect the resulting sounds. The appara- ing an adiabatic process, and hence the polytropic index ␯
tus can be used to demonstrate the bubble sound production was taken to be the ratio of specific heats ␥.
mechanism and the bubbly liquid’s speed of sound depen- The assumption of adiabatic behavior made by Minnaert
dence on the air/water volume fraction. The latter is observed was reasonable for the bubble sizes he was considering, but
indirectly via the pitch of the sound. If the sounds are re- it is instructive to consider a more accurate description.
corded, the data can be used to verify Wood’s equation, Leighton22 considered bubbles undergoing natural oscillation
which is a mathematical model that relates the speed of and those driven by an external acoustic field, and showed
sound in a bubbly liquid to the fractional volume of air. that because of the high thermal conductivity and specific
This demonstration has been conducted for audiences heat capacity of the surrounding liquid, the gas in direct con-
ranging from professional acousticians to prospective under- tact with the bubble wall can be considered isothermal. The
graduate students and their parents. It is always a hit and gas at the center of the bubble behaves adiabatically. There-
usually generates questions and discussion. The basic theory fore, the effective polytropic index for the bubble as a whole
governing bubble oscillation, sound propagation in bubbly can take on a range of intermediate values depending on the
liquids, and the acoustics of one-dimensional waveguides is bubble’s size relative to the size of the thermal boundary
presented in Sec. II. The demonstration apparatus is de- layer. The ratio of the bubble equilibrium radius to the length
scribed in Sec. III. In Sec. IV the use of the apparatus is of thermal boundary layer lth is given by
described. Finally, the technique used to verify Wood’s equa- ā ā冑2␻
tion with our apparatus is discussed in Sec. V. = , 共3兲
lth 冑Dg
II. THEORETICAL CONSIDERATIONS where Dg is the thermal diffusivity of the gas and ␻ is the
frequency of excitation.22 At sufficiently high frequencies,
Consider a spherical air bubble with equilibrium radius ā the bubble is much larger than lth and it behaves adiabati-
suspended in a liquid at room conditions: atmospheric pres- cally. At sufficiently low frequencies, the bubble is much
sure and 20 ° C. For ā ⬎ 10 ␮m, the surface tension and va- smaller than lth and it behaves isothermally. The value of Dg
por pressure can be ignored and the hydrostatic pressure of for air at room temperature is 2.08⫻ 10−5 m2 / s, and there-
the surrounding water is balanced by the gas pressure inside fore ā ⬇ 90lth for Minnaert’s bubbles, confirming the adia-
the bubble. Assuming the air is an ideal gas, it is governed by batic assumption.
the relation
PV␯ = constant, 共1兲 A. Low frequency sound propagation in bubbly liquids
where P is the pressure, V is the volume, and ␯ is the poly- Now consider propagation of sound through a collection
tropic index. For air bubbles in water, the value of ␯ can of bubbles, where the incident acoustic pressure is the source
range from ␯ = ␥, where ␥ is the ratio of specific heats of the of the perturbations. In 1910 Mallock1 applied mixture
air, to ␯ = 1. These limits bound a continuum of behavior that theory to what he termed “frothy liquids,” under the assump-
ranges from adiabatic to isothermal, depending on the bubble tion that the gas bubbles were of such number and separation
size and the frequency of excitation. that the medium could be considered homogenous, and he

976 Am. J. Phys., Vol. 76, No. 10, October 2008 Apparatus and Demonstration Notes 976
This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.83.205.78 On: Mon, 11 May 2015 21:48:30
 obtained an approximate expression for the speed of sound in
the bubbly liquid. In 1930, Wood23 began with a thermody-
namic argument and confirmed Mallock’s work, but cast the
results in a more useful form, and specifically considered air
bubbles in water. In this section, the equation governing the
low-frequency speed of sound in bubbly liquids is devel-
oped, based on the work of Wood.23
The following assumptions and definitions are used. The
bubbly fluid is composed solely of spherical gas bubbles
suspended in liquid under the influence of acoustic pressure
fluctuations. The subscripts ᐉ and g refer to the liquid and
gas phases, respectively. Effective quantities pertaining to
the mixture are denoted with the subscript m. Mass is con-
served such that the mixture mass is the sum of the liquid
and gas masses, with no relative translational motion be-
tween the two phases. The volume fraction of the gas phase,
known as the void fraction ␹, is given by
␹ = Vg/Vm , 共4兲
Fig. 1. The low-frequency speed of sound of sounds of bubbly liquid given
where Vg is the volume of the gas phase and Vm is the total by Wood’s equation 关Eq. 共12兲兴 for air bubbles in water, with ␳ᐉ
volume of the mixture, given by Vm = Vg + Vᐉ. The effective = 998 kg/ m3, ␳g = 1.21 kg/ m3, cᐉ = 1481 m / s and cg = 333 m / s. At zero void
mixture density is fraction, the speed of sound is that of the liquid alone. At unity void fraction,
the speed of sound is that of the gas alone. The minimum speed of sound is
␳m = 共1 − ␹兲␳ᐉ + ␹␳g , 共5兲 about 23 m / s and occurs at ␹ = 0.5.

where ␳ᐉ is the density of the liquid and ␳g is the density of

the gas. The mixture compressibility ␬m is
1 1
␬m = 共1 − ␹兲␬ᐉ + ␹␬b , 共6兲 ␬g = = . 共10兲
␯ P⬁ 1.3P⬁
where ␬ᐉ is the compressibility of the liquid and ␬b is the
compressibility due to the gas in the bubbles. In general, the Substitution of Eqs. 共5兲 and 共6兲 into Eq. 共7兲 results in an
speed of sound c is defined as equation for the mixture sound speed,
1
1 = 关共1 − ␹兲␳ᐉ + ␹␳g兴关共1 − ␹兲␬ᐉ + ␹␬g兴, 共11兲
c2 = . 共7兲 2
cm
␳␬
The value of ␬b and the way it is eventually included in Eq. which is equivalent to Wood’s 1930 equation.23 For bubbles
共7兲 depends in part upon the conditions that prevail during well below resonance size where surface tension and dissi-
the compression phase. For air bubbles in water at frequen- pative effects are negligible, Eqs. 共7兲 and 共10兲 can be used to
cies that approach zero, the process can be considered iso- relate the previous result to the density and speeds of sound
thermal, as was shown rigorously by Hsieh and Plesset.24 of the two mixture components. After manipulation this re-
The response of an individual bubble is dominated by the sults in the equivalent relation
compressibility of the gas within the bubble, hence ␬b 1 共1 − ␹兲2 ␹2 ␳2gc2g + ␳2ᐉc2ᐉ
= 共␬g兲isothermal = 共␯ P⬁兲−1 with ␯ = 1. As the excitation fre- = + 2 + ␹共1 − ␹兲 , 共12兲
2
cmlf c2ᐉ cg ␳ᐉ␳gc2ᐉc2g
quency approaches the bubble resonance frequency, the dy-
namics of the bubble become important and ␬b → ␬b共f兲 ⫽ ␬g. where cmlf is the low-frequency speed of sound of the mix-
Well above resonance, the bubbles behave adiabatically and ture. The behavior of Eq. 共12兲 for air bubbles in distilled
␬b = 共␬g兲adiabatic = 共␯ P⬁兲−1 with ␯ = ␥ = 1.4 for air. In between, water is shown in Fig. 1. Note that as the mixture approaches
the polytropic index that governs the compression and ex- pure liquid, ␹ → 0 and cmlf → cᐉ. Alternatively, as ␹ → 1,
pansion of the gas inside the bubbles is frequency dependent, cmlf → cg. Upon approach to ␹ = 1, a change to the adiabatic
as shown by Prosperetti,25 and is given by ␯ = Re关⌽兴 / 3, speed of sound in air would be required. The speed of sound
where in bubbly liquid, given by Eq. 共11兲 or 共12兲, is known as
Wood’s low-frequency speed of sound or the Wood limit of
3␥ the speed of sound.
⌽= 共8兲
1 − 3共␥ − 1兲iX关共i/X兲1/2coth共i/X兲1/2 − 1兴 A minimum occurs at ␹ = 0.5, the speed of sound is lower
than the intrinsic speeds of sound in air or water across much
and of the void fraction range. This remarkable result can be
X = D g/ ␻ a 2 . 共9兲 qualitatively explained: bubbles greatly increase the mixture
compressibility yet effect the density much less. High com-
In this work, the excitation frequency is well below the pressibility and high density yields a low speed of sound.
bubble resonance, therefore ␬b = ␬g. In the frequency range Another interesting result is that in this regime, the bubble
between 200 and 1000 Hz and a bubble size of about size does not play a role. The primary environmental variable
1.0 mm, Eq. 共8兲 yields 1.26⬍ ␯ ⬍ 1.33. It is sufficient for our is the void fraction.
purposes to take the average value of ␯ and set the compress- Because of the great difference in density and speed of
ibility of the gas to be sound between air and water, Eq. 共12兲 can be approximated

977 Am. J. Phys., Vol. 76, No. 10, October 2008 Apparatus and Demonstration Notes 977
This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.83.205.78 On: Mon, 11 May 2015 21:48:30
 by a much simpler expression for a limited range of ␹. For
air bubbles in water, in the range 0.002⬍ ␹ ⬍ 0.94, the first
two terms on the right-hand side of Eq. 共12兲 are much
smaller than the third term. Neglecting them and noting that
␳2gc2g  ␳2ᐉc2ᐉ results in
1
2
cmlf ⯝ . 共13兲
␳ᐉ␬g␹共1 − ␹兲
Substitution of Eq. 共10兲 into Eq. 共13兲 results in a simple
approximate expression for the low-frequency speed of
sound of the mixture

cmlf ⯝ 冑 ␯ P⬁
␳ᐉ␹共1 − ␹兲
, 共14兲

and differs less than 1.2% from Eq. 共12兲. Extensions to Eq.
共14兲 accounting for relative motion between the phases, i.e.,
bubbly fluid flows, have been provided by Crespo26 and
Ruggles et al.27

B. Mode shapes and eigenfrequencies of a

one-dimensional waveguide
The compressibility of a bubbly liquid is large in compari-
son to the pipe wall material; hence, the waveguide walls
appear effectively rigid. The top of the tube is a water-air
interface, hence it is approximated as a pressure release
boundary, with p = 0. Again, because of the high compress-
ibility of the bubbly liquid inside the tube, the water just
below the opening of the needles appears as a rigid surface,
with ⳵ p / ⳵x = 0. Consider a rigid tube of length L and circular
cross section with radius b, filled with a fluid possessing a
Fig. 2. A schematic of the demonstration apparatus.
speed of sound c0. The cutoff frequency of the lowest order
transverse mode is given by28
⬘ c0
␣11
c
f 11 = , 共15兲 The acoustic waveguide was constructed from transparent
2␲b
schedule 40 PVC pipe that is available from plumbing sup-
where ␣11⬘ = 1.841 and is the first root of J1⬘共x兲, the Bessel ply houses. The inside diameter was 5.2 cm and the length
function of the first kind of order one, and the prime indi- was 0.6 m. The pipe is typically filled with distilled water to
cates the derivative with respect to the function’s argument. a level of ⬇0.5 m. A bubble injection manifold was fabri-
At frequencies below f 11c
, only plane waves can propagate cated with 12 22-gauge needles29 embedded in an epoxy
within the tube. Solving the wave equation within this tube filled PVC pipe fitting that attached to the bottom of the pipe.
for the given boundary conditions yields the eigenfunctions The needles were positioned in a vertical orientation and
therefore released air bubbles directly upward. A small me-
␲共n + 1/2兲x chanical float was positioned at the top of the pipe, such that
pn共x兲 = an cos , 共16兲
L it floated on the upper surface of the water, as shown in inset
共a兲 of Fig. 2. The float was attached to a pointer that indi-
and eigenfrequencies
cated the height of the water surface on a centimeter scale.
c0 The overall length L of the water column in the tube could be
fn = 共n + 1/2兲, n = 0,1,2, . . . . 共17兲 measured to an accuracy of about 1 mm. The void fraction is
2L
The resonance frequency f n is a linear function of the mode ␹ = ⌬ ᐉ /共ᐉ + ⌬ ᐉ 兲 = ⌬ ᐉ /L, 共18兲
number n with the sound speed c0 as a parameter.
where ᐉ is the length of the water column in the absence of
III. DESCRIPTION OF DEMONSTRATION bubbles and ⌬ᐉ is the change in length when bubbles are
APPARATUS present.
Any convenient source of air can be used, as long as the
The apparatus, shown in Fig. 2, consists of five basic flow rate can be controlled by the user. In our implementa-
parts: an acoustic waveguide that is filled with bubbly water, tion, a canister of medical air was used. The air was directed
an air injection system that creates the bubbles, a hydro- through a flow meter with a needle valve to control the flow
phone that senses the acoustic pressure within the wave- rate, although a measurement of the flow rate is not required.
guide, a data acquisition system that records and displays the If available, one can use a video camera with a macro 共clo-
hydrophone signal, and an amplified loudspeaker system that seup兲 lens to observe the bubbles and hence determine the
broadcasts the hydrophone signal to the audience. bubble size. A typical bubble is shown in Fig. 3.

978 Am. J. Phys., Vol. 76, No. 10, October 2008 Apparatus and Demonstration Notes 978
This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.83.205.78 On: Mon, 11 May 2015 21:48:30
 is not important at this stage.31 The demonstration begins
with no air flow from the needles, and hence, no bubbles in
the waveguide. The VSA and the PA system are turned on
and a flat spectrum of low amplitude is displayed, which
gives an indication of the background noise in the room. The
needle valve is then slowly opened to allow single bubbles to
breakoff the needles one at a time. The individual breakoff
events are audible through the PA system. Although not de-
scribed here, the time series of the hydrophone signal could
be presented using an oscilloscope. A bubble breakoff event
would appear as a transient signal that approximates a
damped sinusoid.
The flow rate is then slowly increased and a larger number
of bubbles is created. The acoustic excitation becomes effec-
tively continuous and standing waves develop in the wave-
guide. The excitation is sufficiently broadband to excite the
first several normal modes of the waveguide. One can then
continue to increase the flow rate and the number of bubbles
further increases. This reduces the speed of sound in the
waveguide, and hence reduces the frequency of the normal
modes. As the flow rate continues to increase, the pitch of the
sound produced continues to drop. Changing the void frac-
tion via flow rate from 0.1% to 2% over a few seconds re-
sults in the speed of sound in the tube changing from ap-
proximately 300 m / s to 100 m / s over the same time scale,
Fig. 3. A photograph of a typical bubble produced by one of the needles. and hence the pitch of the sound is changed from about
150 Hz to 50 Hz, which is an octave and a half. This is easily
detected by ear, but because the frequencies are relatively
The acoustic energy is supplied by the bubbles them- low, a large speaker is required to reproduce the sound. The
selves. When each individual bubble breaks off its needle, a overtones can still be heard if only a small speaker is avail-
transient acoustic pulse is produced. Continuous bubble pro- able, but the demonstration is more effective with a large
duction excites an acoustic standing wave field inside the speaker.
waveguide that is observed with a hydrophone. A Brüel and As the flow rate is increased, the height of the water col-
Kjær model 8103 miniature hydrophone was used in this umn also increases. This also plays a role in establishing the
work, but a small lapel microphone can also be used.30
frequencies of the normal modes, but it is small compared to
The hydrophone signal was amplified and bandpass fil-
the variation due to the void fraction induced change in the
tered between 10 Hz and 10 kHz by a Brüel and Kjær model
speed of sound. In addition to ramping the flow rate up and
2692 charge amplifier. The high-pass characteristic of the
filter suppresses very low-frequency 共⬍50 Hz兲 flow noise, then down, which produces decreasing and then increasing
pitch, the user can vary the flow rate in any manner desired,
which can in some cases be dominant. Suppressing the flow
noise allows the desired part of the signal 共the eigenmodes of which results in corresponding pitch changes. The user can
the waveguide兲 to be more easily heard by the audience. If also include a return to very low flow rates, which produces
the public address system has its own filter function, such as individual bubble breakoff events and their corresponding
a graphic equalizer, a separate high-pass filter is not required. sounds. With some practice, one can achieve at times humor-
The low-pass filter was used for antialiasing the signal prior ous, and at other times aesthetically pleasing and even mu-
to digitization. If the apparatus is used solely for an audio sical sounds.
demonstration, the low-pass filter is not required. During the demonstration, the VSA displays the spectrum
The hydrophone signal was directed to both a signal ana- of the hydrophone signal in real time. Near the beginning of
lyzer and a public address system. An Agilent 89410A vector the demonstration, we establish a flow rate sufficiently high
signal analyzer 共VSA兲 was used to digitize the conditioned to yield continuous acoustic excitation, such as that shown in
hydrophone signal and display and record the spectrum of Fig. 4共a兲. The demonstrator can point out that the peaks in
the signal via fast Fourier transform calculations performed the spectra correspond to the eigenfrequencies of the wave-
onboard the VSA. Alternatively, a consumer-grade sound guide and that these frequencies combine to produce the
card can perform the same functions on many personal com- acoustic sound we perceive during the experiment. As the
puters. Finally, the hydrophone signal must be amplified and frequencies of these peaks increase, the pitch is also per-
directed to the loudspeaker system. We used a musical in- ceived by a listener to increase. The spectra of these sounds
strument amplifier designed for bass guitar, since the lowest can be seen in real time on the signal analyzer, with the
eigenfrequencies can be below 100 Hz. peaks shifting up and down with the flow rate. Ideally, a
waterfall plot can be displayed showing a brief history, like
IV. DEMONSTRATION PROCEDURES that shown in Fig. 4共b兲. This demonstration has been per-
formed by the authors many times to the approval of observ-
The PVC waveguide is filled with water and the hydro- ers ranging from elementary school students to deans of
phone is placed midway down the tube. The exact placement large research universities.

979 Am. J. Phys., Vol. 76, No. 10, October 2008 Apparatus and Demonstration Notes 979
This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.83.205.78 On: Mon, 11 May 2015 21:48:30
 Fig. 5. Linear fits to the measured resonance frequencies as function of
mode number for three void fractions: the lowest, the highest, and an inter-
mediate void fraction.

Fig. 4. 共a兲 A single spectrum from the 0.4% void fraction case for which the
individual modes and their eigenfrequencies f n have been identified. The
units are decibels 共dB兲 referenced to 1 ␮Pa, where dB= 20 log10关p / pref兴, p is bubbles are fewer in number and less homogeneously spaced
the RMS acoustic pressure measured in the tube, and pref = 1 ␮Pa. 共b兲 A throughout the tube. This results in a speed of sound that is
waterfall plot of all the experimental spectra. inhomogeneous and resonance frequencies that are not per-
fectly harmonic, and hence less well-represented by a linear
fit.
V. VERIFICATION OF WOOD’S EQUATION The speeds of sound thus inferred for the entire dataset are
plotted versus void fraction in Fig. 6. Vertical error bars rep-
The bubbles produced in the apparatus are typically about resent the uncertainty in the speed of sound due to uncertain-
1 mm in radius, hence they have resonance frequencies near ties in the tube length, L, ⫾1 mm, and the finite bandwidth
3 kHz, according to Eq. 共2兲. As long as the eigenfrequencies resolution of the spectra, which was 2.5 Hz. The solid line is
of the waveguide are well below the bubble resonance fre- Eq. 共12兲 plotted using the following physical parameters:
quency, the bubbly liquid-filled waveguide will exhibit a
speed of sound nearly independent of frequency and bubble
size and the mixture will behave as an effective medium with
a speed of sound given by Wood’s equation 关Eq. 共12兲兴.
To experimentally verify Eq. 共12兲, spectra like the one
shown in Fig. 4共a兲 are obtained for a range of air flow rates.
At each flow rate, the height of the column L = ᐉ + ⌬ᐉ is
measured using the float at the top of the waveguide and the
void fraction ␹ is determined from Eq. 共18兲. Due to motion
of the water surface, the uncertainty in ⌬ᐉ is about
⫾0.5 mm, which leads to an uncertainty in the void fraction
␹ of ⫾0.001. The bubble-free column height was ᐉ
= 0.485 m ⫾ 0.005 m. A collection of spectra at void frac-
tions ranging up to ␹ = 2.5% is shown in Fig. 4共b兲.
As indicated by Eq. 共17兲, a plot of the modal frequencies
versus mode number ideally yields a straight line with a
slope directly proportional to the speed of sound in the bub-
bly liquid. Examples for three void fractions are shown in
Fig. 5. A least squares linear fit of the measured resonance
frequencies yields the slope from which the speed of sound
is obtained for each void fraction. At higher void fractions,
␹ = 2.46% and 0.6%, the bubbles are approximately homoge- Fig. 6. Experimental verification of Wood’s equation. The solid line is the
prediction of Eq. 共12兲 and the open circles are the measured values. Vertical
neously spaced along the length of the tube, which yields an error bars represent the uncertainties in the measurement of the speed of
effective speed of sound that is constant throughout the tube. sound due to uncertainties in the column length L and the finite bandwidth
The resulting resonance frequencies are nearly harmonic. At resolution of the spectra. Horizontal error bars represent uncertainties in the
the lower void fractions, exemplified by ␹ = 0.13%, the void fractions due to the uncertainty in the measured values of ⌬ᐉ.

980 Am. J. Phys., Vol. 76, No. 10, October 2008 Apparatus and Demonstration Notes 980
This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.83.205.78 On: Mon, 11 May 2015 21:48:30
 ␳ᐉ = 998 kg/ m3, ␳g = 1.21 kg/ m3, cᐉ = 1494 m / s, and cg 8
of Illinois Institute of Technology 共1958兲.
= 334 m / s. The speed of sound in the liquid, cᐉ, was deter- H. Medwin and C. S. Clay, Fundamentals of Acoustical Oceanography
共Academic Press, Boston, 1998兲.
mined from the water temperature, 24 ° C, and Eqs. 共5兲, 共6兲, 9
Underwater Ship Husbandry Manual, Naval Sea Systems Command,
and 共8兲 of Ref. 21. The speed of sound in the gas, cg, was 1995, NAVSEA Publication S0600-AA-PRP-050.
determined from Eqs. 共7兲 and 共10兲 for a hydrostatic pressure 10
F. Calliada, R. Campani, O. Bottinelli, A. Bozzini, and M. G. Som-
due to the standard atmospheric pressure, 101.3 kPa, and one maruga, “Ultrasound contrast agents basic principles,” Eur. J. Radiol. 27,
half the column height of water, 0.24 m. Agreement within S157–S160 共1998兲.
11
the measurement uncertainty is found between the measured S. Vaezy, R. Martin, P. Mourad, and L. A. Crum, “Hemostasis using high
and predicted speeds of sound throughout the range of void intensity focused ultrasound,” Eur. J. Ultrasound 9, 79–87 共1999兲.
12
X. Yang, R. A. Roy, and R. G. Holt, “Bubble dynamics and size distri-
fractions. Below ␹ = 0.3%, the agreement is somewhat less butions during focused ultrasound insonation,” J. Acoust. Soc. Am.
than at higher void fractions. This is due to the inhomoge- 116共6兲, 3423–3431 共2004兲.
neous conditions inside the waveguide discussed before. The 13
H. C. Pumphrey and L. A. Crum, “Free oscillations of near-surface
effective medium theory breaks down under such conditions bubbles as a source of the underwater noise of rain,” J. Acoust. Soc. Am.
and can no longer be used to describe the system. To inves- 87共1兲, 142–148 共1990兲.
14
tigate lower void fractions, a bubble injection scheme ca- J. H. Wilson, “Low-frequency wind-generated noise produced by the im-
pact of spray with the ocean’s surface,” J. Acoust. Soc. Am. 68共3兲, 952–
pable of evenly spacing bubbles at low flow rates would be
956 共1980兲.
required. 15
L. A. Crum, H. C. Pumphrey, R. A. Roy, and A. Prosperetti, “The un-
derwater sounds produced by impacting snowflakes,” J. Acoust. Soc. Am.
VI. CONCLUSIONS 106共4兲, 1765–1700 共1999兲.
16
W. M. Carey and J. W. Fitzgerald, “Low Frequency Noise from Breaking
A simple laboratory apparatus is described that provides Waves,” in Natural Physical Sources of Underwater Sound, edited by B.
an audible demonstration of several aspects of the acoustics R. Kerman 共Kluwer, Boston, 1993兲, pp. 277–304.
17
of bubbly liquids. The speed of sound in a bubbly liquid F. S. Crawford, “The hot chocolate effect,” Am. J. Phys. 50共5兲, 398–404
contained within a PVC pipe is altered by varying the injec- 共1982兲.
18
S. Aljishi and J. Tatarkiewicz, “Why does heating water in a kettle pro-
tion rate of air bubbles. The bubble injection process itself duce sound?,” Am. J. Phys. 59共7兲, 628–632 共1991兲.
produces broadband noise that excites one-dimensional 19
V. Leroy, M. Devaud, and J.-C. Bacri, “The air bubble: Experiments on
standing waves within the apparatus. The variation of the an unusual harmonic oscillator,” Am. J. Phys. 70共10兲, 1012–1019 共2002兲.
20
speed of sound as a function of the void fraction is apparent K. W. Commander and A. Prosperetti, “Linear pressure waves in bubbly
in the associated change of the pitch of the sounds produced. liquids: Comparison between theory and experiments,” J. Acoust. Soc.
The apparatus can also be used to verify an effective medium Am. 85共2兲, 732–746 共1989兲.
21
L. E. Kinsler, A. R. Frey, A. B. Coppens, and J. V. Sanders, Fundamen-
model of sound propagation in bubbly liquid, which relates
tals of Acoustics, 4th ed. 共Wiley, New York, 2000兲.
the void fraction to the speed of sound. 22
T. G. Leighton, The Acoustic Bubble 共Academic Press, London, 1994兲.
23
A. B. Wood, A Textbook of Sound, 1st ed. 共MacMillan, New York, 1930兲.
24
ACKNOWLEDGMENT D.-Y. Hsieh and M. S. Plesset, “On the propagation of sound in a liquid
containing gas bubbles,” Phys. Fluids 4共8兲, 970–975 共1961兲.
This work was supported by the U.S. Navy Office of Na- 25
A. Prosperetti, “The thermal behavior of oscillating gas bubbles,” J. Fluid
val Research Ocean Acoustics Program. Mech. 222, 587–616 共1991兲.
26
A. Crespo, “Sound and shock waves in liquids containing bubbles,” Phys.
Fluids 12共11兲, 2274–2282 共1969兲.
a兲 27
Electronic mail: [email protected] A. E. Ruggles, H. A. Scarton, and R. T. Lahey Jr., “An Investigation of
1
A. Mallock, “The damping of sound by frothy liquids,” Proc. Royal Soc. the Propagation of Pressure Perturbations in Bubbly Air/Water flows,” in
84共A 572兲, 391–395 共1910兲. First International Multiphase Fluid Transients Symposium, edited by H.
2
Arnulph Mallock worked for Lord Rayleigh in the 1870s and conducted H. Safwat, J. Braun, and U. S. Rohatgi 共American Society of Mechanical
experiments in the hydraulic laboratory at Terling, Lord Rayleigh’s estate, Engineers, New York, 1986兲, pp. 1–9.
28
as discussed in R. J. Strutt, Life of John William Strutt, Third Baron D. T. Blackstock, Fundamentals of Physical Acoustics 共Wiley, New York,
Rayleigh, O. M., F. R. S 共University of Wisconsin Press, Madison, 1968兲. 2000兲.
3 29
L. Rayleigh, “On the pressure developed in a liquid during the collapse of Hamilton Company; 4970 Energy Way; Reno, Nevada 89502 USA.
a spherical cavity,” Philos. Mag. 34, 94–98 共1917兲. http://www.hamiltoncompany.com/
4 30
M. Minnaert, “On musical air-bubbles and the sounds of running water,” The lapel mic must be made water proof. This can be done by coating the
London, Edinburgh Dublin Philos. Mag. J. Sci. 16, 235–248 共1933兲. microphone with rubber, such as Plasti Dip, available at hardware stores.
5
C. Crowe, M. Sommerfeld, and Y. Tsuji, Multiphase Flows with Droplets http://www.plastidip.com/home_solutions/Plasti_Dip
31
and Particles 共CRC Press, Boca Raton, 1998兲. The field inside the waveguide is spatially dependent according to Eq.
6
L. E. Faidley and J. A. I. Mann, “Development of a model for acoustic 共16兲. One can accentuate or suppress a mode by placing the hydrophone
liquid manipulation created by a phased array,” J. Acoust. Soc. Am. near an antinode or a node, respectively. Since the excitation is broad-
109共5兲, 2363 共2001兲. band, a number of modes are excited, but the relative amplitude of a
7
H. B. Karplus, “The velocity of sound in a liquid containing gas particular mode received on the hydrophone depends on the hydrophone’s
bubbles,” Technical Report No. C00-248, Armour Research Foundation location inside the waveguide.

981 Am. J. Phys., Vol. 76, No. 10, October 2008 Apparatus and Demonstration Notes 981
This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.83.205.78 On: Mon, 11 May 2015 21:48:30